Download - PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

Transcript
Page 1: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

PRIMARY AND METASTATIC UVEAL MELANOMA:

towards a therapeutic approach

Page 2: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

Tekening omslag: Yolanda Eijgenstein

Druk & vormgeving: Vormgeving Rotterdam

ISBN 90 72206 09 6

Page 3: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

PRIMARY AND METASTATIC UVEAL MELANOMA:

towards a therapeutic approach

(Primair en gemetastaseerd oogmelanoom:

naar een therapeutische benadering)

Proefschrift

Ter verkrijging van de graad doctor

aan de Erasmus Universiteit Rottenhun

op gezag van de rector Illagnifïcus

Prof. Dr. P. W.C. Akkermans M.A.

en volgens besluit van het college voor promoties.

De openbare verdediging zal plaatsvinden op

woensdag 15 mei 1996 011113.45 uur

!loOI'

Gregorius Petrus Maria Luyten geboren te Etten-Leur

Page 4: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

Pro 111 0 ti eco Illllliss i e

Promotor:

Co-promoter:

Overige leden:

Prof'. Dr. P.T.V.M. de Jong

Dr. C.M. Mooy

Prof. Dr. D. Bootsm3

Prof. Dr. G. Stoter

Prof. Dr. W . .J. l'vIooi

Page 5: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

To m)' dearcsl Yolanda

Page 6: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CONTENTS

CHAPTER 1 Introduction 8

1.1 Clinical aspects of uveallllelallollla 8

1.1.1 Introduction to the subject 8

1.1.2 Treatment of the primary tumor 10

1.1.3 Metastatic disease and sllI'vival 12

1.2 AilllS aJul scope of the thesis 15

1.3 Illtroductioll to the studies 17

1.3.1 Chapter 2: Thc origins of metastases 17

1.3.2 Chapter 3,4 and 5: Clinical and histopathological factors 18

1.3.3 Chapter 6 to 9: Experimcntal studies 20

1.3.3.1 Chapter 6: Uvealmclanoma cclllines 20

1.3.3.2 Chapter 7: Tumor genetics and inullunology 21

1.3.3.3 Chapter 8 and9: Animalmodels 29

Relerences 32

CHAPTER2

CHAPTER3

No delllonstnlted effect of pre-ellucleation irradiation of

uveallllelallollla patiellts

Metastatic uveallllelanollla:

44

A lllol'phologÏcal and illllllunohistochelllical allalysis 56

Page 7: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER4

CHAPTER5

CHAPTER6

CHAPTER 7

CHAPTER8

CHAPTER9

CHAPTER 10

SUlunulry

Neuml eeIl a(I11esion molecule distribution in primary

and llletastatic llveallllelanOlna

The expression NM23 gene in uveal melanoma

Establishment mul eharaeterization of primary mul

metastatic uveal melanoma eelllines

Expression of MAGE, GP100 and tyrosillase in uveal

melanoma eelllines

Relationship between natural killer eeIl susceptibility and

metastasis of hmlUm uveal melanoma eeIls in

a lllurine lllodel

A chicken embryo model to study growth of

uveallnelauOlua

Discussion and future perspectives

Samenvatting in het Nederlands

List of abbreviations

Curriculum vitae

List of publications

Dankwoord

70

86

102

126

137

153

167

175

179

186

187

188

190

Page 8: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1

INTRODUCTION

1.1 Clinical aspects of uveal melanoma

1.1.1 Inlroduclioll 10 Ihe subject

Uveal melanoma is a unCOll111lon discase with on estimation 80-100, yeurly incident

cases in The Netherlands (6 per million) (De Jong, 1987). Neverlhe1ess, uvea1

melanoma is the most coml11on primary malignant intraocular tumors in adults (Egan,

1988). Morcover, this type of cancel' orten proves fatal to the patient. Although 50 % of

the patients treated for their primary lIveal tumor recovers eompletely, the other

50 % eventually dies of metastatic diseuse (Gamel,1993; MeLean, 1993; Diener-West,

1992; Jensen, 1982). Tn other 1V0rds, half of the patients suffering from a primary uvea1

melanoma c1evelops metastatie discase at a distant site, whieh is then lethal in all cases.

Tn contrast ta cutaneous melanoma, uvealmclanoma disseminates primarily

hacmatogenously anel with a slrong preferenee for the Iiver; once metastases have been

diagnosed, the patient's life expeetancy is only 2 to 9 months (Kath, 1993; Alberl,

1992; Gragouclas, 1991; Rajpal, 1983; Bedikian, 1981; ehar, 1978). These metastases

offen becoll1c c1inically manifest years after initial trealment, that is, llveal melanomas

metastasize relatively late. The patient's median survival time aftel' discovery of the

primary tumor is 6.5 years (McLean, 1993), with a peak of elanomH-related deaths

between the secOIld ano fourth year following initial treatment (Zimmerman, 1978).

Tn brief, the studies presented in this thesis basically aim at prevention of (death by)

metastatic uvea1melanoma. Before we continue, it should be noted that wherever the

term uvea1melanoma is used in this baak, it relers only to melanoma of the choroid

and the ci1iary body. Melanomas 10cated in the iris, the third part of the uvea, are

re1atively benign with a low incidence of developing metastases at distant sites and are

therefore left out of the study here. The reader interested in iris melanoma is referred to

other reports (Grossniklaus, 1995; Jensen, 1993; Brown, 1990).

8

Page 9: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INTRODtJCTION ____ _

Tu date, studies on uveal melanoma mainly focused on the rcfinement of

clinical eliagnosis anel treatmcnt of the primary tumor. This has resulteel in a dccrease

in the c1inicalmisdiagnosis ratc from 3.5 up to 20 % (COMS, 1990; Jensen, 1963) to

less than 1 % at present (Davidorf, 1983; COMS, 1990; pers on al data). The fact th at

tocltty, in specialized centers, the correct diagnosis of primary uvcal melanoma can be

made in 1110re than 99 % of the cases is mainly owed to the application of indirect

ophthalmoscopy in experienced hands together with the use of ultrasonography.

Fluorescenee angiography has made the diagnosis also more reliablc. These devices

are hclpful in diseriminating uvealmelanoma from other types of ocular tumors and

disorelers, sueh as choruidal nevi, haemangiomas and metastases, granulomas and

macular c1egeneration, For instanee, a tumor or biconvex or lllUshroom-like shape lhat

shows choroidal excavation on the B-mode and a low to moderate high intern al

rellectivity on the A-mode in an examination by ultrasonography will, when combined

with indirect ophthalmoscopy, certainly be recognized as a choroidalmelanoma.

However, metastatic choroidallesions, that usually show a high internal rellectivity on

the A-mode, can be highly variabie in their ultrasound pa((ern. In sueh cases

examination of fine-needle aspiration biopsies (FNAB) can be useful (Shields, 1993),

but then again, in the cytological evaluation ofFNAB, differentiating melanomas and

other intraocular tumors ean be hard too. Next, in cases in which it is difficult to

discriminate a small dormant melanoma from a choroidal nevus, tumor growth can be

followed by B-mode ultrasonography to evaluate the tumor height, and by serial

Iluorescence angiography to cvaluate thc tumor diameter. Finally, in diagnosing

primary uvea( melanomas, magnetic resonance imaging with contrast (gadolinium)

may be used. The problem with applying this techniqlle, however, is that the

characteristical image of thc paramagnctic melanin is not present in all cases (Fen'is,

1995; De Potter, 1994; Bioom, 1992).

When a primary uvealmelanoma has been diagnosed, the patient is usually

screcned for the presence of extraocwar malignancies anel distant metastases of the

uvealmelanoma by a complete physical examination, an X-ray of the chest, liver

funetion tests, and liver lIltrasonography. In about 8% of the cases another seeOlld

primary tumor can be found (l<:indy-Degnan, J 989; personal data). Screening of the

patient aftel' eliagnosis of a primary lIveal tumor brings related clistant metastases to

9

Page 10: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTEI~ .~I ___ _

light in approximately 2 % of the cases (Paeh, 1986; pers(lJlal data). Furthennore,

before the decision of treatment can be made, aU factors that may be of importanee to

the therapeutic choke have to be evaluated for each patient individually, sueh as visu,,1

aeuity, intraoclliar pressure and other concomitant disorders or the affected anel the

non-affeeteel eye; size, location anel episcleral extension of the tumor, and age anel

general health of the patient (Shields, 1993; id. 1991).

1.1.2 Treatmellt of the primary tumor

Ollee thc diagnusis of primary llvealmelanoma has been made) the choicc of

treatment is controversial (Shields, 1993; id. 1991).ln the past, thc only available

treatment was enueleation of the tumor-eontaining eye. Tt is generaUy agreed up on that

this therapy is still preferabie in cases of large melanomas (largest tumor diameter (LTO)

> 15 mm and/or tumor height > 5 mm) (De Jong, 19X7; Manschot 1980). However,

mainly with regm'd to small (LTD <10 m111) or medirull sized (LTO 10 - IS nuu) tumors,

newly developed so-calIcel "conservative" treatmcnts as upposcel to enucleation such as

photocoagulation, local resection and especiaUy radiotherapy are applied in an

increasing number of cases,

With regard to these "conservative" types or treatment, laser photocoagulation is

currently not that much ac1voçatec1, but is stillusec1 by Wllle authors in selective cases of

smalluvealmelanomas (Shields, 1991; Oamato, 1993; Hungerford, 1993). Seeondly, the

sm"gical intcrvention can sometimes be confined to local resection of the tumor in

particular cases or posterior ehoroidalmelanomas. Because of the technical diffieulties

of this surgieal procedure, local reseetion of these tumors is only performec1 in a few

spccialized centers (Oamato, 1993; Shields, 1991). Aner local rcsection, especially, of a

nasally located posterior choroidal tumor, a goml vision of over 6112 is retained in 57 %

of the cases (Oamato, 1993). Tn cases of uvealmelanoma with massive orbital

involvement, then, orbital exenteration is sometimes indicated for local tumor contro!.

Since an excnteration \vill probably have no influence on the patienCs survival, it can be

disputed whether an orbital exenteration has to be performed in cases of episcleral

extension that coulc1 also be removed by enuc1cation (Shields, 1993; Kersten, 1985;

Shammas, 1977; personal data). Finally, in cases ofsmall, dormant tumors (tumor

height< 3 111m) - tmuars that show 110 signs of growth in serial examinations by

JO

Page 11: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_---'I"N-'T R 0 D U CT JON --- -_._- --_.

llltrasonography and fluorescence ungiography, - some tlLllhors advocate mere

observation untillumor enlargemenl has been demonstrateu (Butler, 1994;

Augsburger, 1993). However, a recenl clinical study reports thai 3 % of lhe palients wilh

lumors less lhan 3 mm in height and up to 19% of them in case ol' proven tumor groWlh

will develop distant metastases af ter all (Shields, 1995). Therefore, even in cases of

apparently small dOfmanl llveal 1l1clanomas, prompt treatment appears ta he

recomlllcndabie.

In most cases, hOWeVel\ the choice of trealment is nne between cnucleation anel

radiolherapy of the tumor-containing eye. For several ophthalmologislS the principle

choice has become radiotherapy, applied by means ol' a radioactive episcleral plaque

with cobalt-60 (Stallard, 1959), ruthenium-lOG (Lommatzsch, 1973), or iodine-125

(Packer, 1980), or by irradiation wilh charged particles or proion beams (Gragoudas,

1980) or helium-ions (Hungerford, 1993; Shields, 1991; Char, 1980).

Yet, there is much discussion about specified inclications anel the choice of

lhe optimal dose.

One of the major reasons ta perfofm radiotherapy, or any conservative lreatment,

is to retain lhe patienl's vision. Most mllhors advocating raeliotherapy elescribe a useful

vis ion as a vision of over 20/200. However, although they reporl a vismd acuit)' of more

lhan 20/200 in about 90% of lhe cases aftel' 1 year, aftel' two to three years this has

decreased 10 73% and 50%, respectively (Young,1994; Quive)', 1993; Guyer, 1992;

Shields, 1991). Addilionally, na differenccs have been 1'00llld in vision-depending­

activities betwcen groups of patients eilher trealed by enucleation or by radiotherap)';

neither do patients encounter significanl visual problems during lheir vision-depending­

activities afler enucleation ol' anc e)'e (Augsburger, 1994; Edwards, 1991). In olher

words, the choice bet ween enucleation and radiotherapy would nOl have 10 be hindered

by argumenls of respective bolher to the patien!. Although, a few comparative sludies

could nol demonstmte a significant dit'fercnce in survival between the two types of

lreatment (De Potter, 1994; Augsburger, 1990; Adams 1988), one should keep in mind

lhat there are hardly any proper studies available with long-term follow-up data and a

fixed protocol. Recently, the collaborative ocular melanoma study group (COMS)

designed a multicelltre lrial 011 lhe topic of episcleral plaqlle themp)' versus ellucleation;

it is expected thai this COMS study will clarify some of the issues (Slraatsma, 1988).

11

Page 12: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPT)~R 1 ,-"---''----

Aftel' any type of inilial treatmcnl, lhe patienls have to be followed regularIl"

Patienls lreated by radiotherapy need a frequent ophlhalmic check-up in the post­

treatment period in consideration of direct complications and tumor control. Post­

radiolherapy complications, such as neovascular glaucoma, scleral nccrosis and tumor

regrowth, lead to an indication 1'01' eventual enuc1eation in 6 to 23 % of the patienls

(Egan, 1989; Shields, 1990; De Potter, 1994). In case of enuc1eation, the follow-up

protocol in our own clinic includes ophthalmoscopic examination of the non-affected

eye, inspection of lhe socket, palpation of preauricular and submandibular lymph

no des, and liver-enzyme tests. With regard to the latte,,, in case of demonstrated

changes in the liver-enzymes, a liver ultrasonography is performed in order la detect

possible liver melastases (De Jong, 1987). These examinations are perfofllled every

lhree months in lhe first half a year, then every six months during the first two years,

and thereafter annually (De Jong, 1987).

1.1.3 Metastlltic di se ase and survival

Despite all effort that has been put in rcfining diagnosis and treatment of the primary

tumor described in lhe preceding paragraphs, the life expectancy of uveal mclanoma

patients has not improved over the last decades. Still, 50 % of the patients sooner or

later develops mctastatic lesions th at lead to a sure dealh. Survival rates, assessed hy the

actuarial survival estimate (Kaplan-Meicr curves) or the multivariate analysis (Cox

proportional hazard model), show survival percentages 01'75 %,65 % and 55 %

respectively 5, 10 and 15 yems aftel' initial treatmeut (Lul' ten, 1995; Mclean, 1993;

Gamel, 1993; Dicner-West, 1992; Jensen, 1982). Self-evidently, these analyses arc

based on tumor-related deaths, while deaths by (lther causes simpil' count as the last

date of follow-up ofthe patient in Cjuestion (Seigel, 1990). Bl' comparing the

hypothetical survival rates of uvealme1anoma-related dealh with the aclual survival

time, the asl'mptote of the hypothetical survival rates reaches a certain hypothetical

value which is thc function of the survival time of the uncured patients. The median

survival time of 6.5 years that was mentioned above was assessed this way, in a large

studl' of 4726 uveaimelanollla patients that also delllonstrated a cure rale of 42 %

(McLean, 1993). In other words, 50 % of all patients who will die from metastatic uveal

melanoma die within 6.5 years aftel' initial treatlllent; the highest 1ll0l'lalitl' rate is fmllld

12

Page 13: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INTRODUCTION

in the secOild to fourth year following initial treatment. Pinally, a significant longer

survival time has been found among younger patients, that is, patients bel ow 50 years

of age (Gamel, 1993; Kath, 1993; Gragoudas, 1991; Rajpal, 1983), and comparcd to men,

women have a longer survival time (Luyten, 1995; Folbcrg, 1993; Rajpa1, 1983).

Uveal melanoma dissemin~tes almost exclllsively haematogenously wÎth a

strong prefercnce for the liver. Liver involvelllent is found in 56 to 88 % of the

patients as the first symptom of Illetastatic diseasc, while in the final stage the liver is

involved in up to 95% of thc cases. Gther organs are much less frequently affected: the

lung in 24A6%, the bones in 17-29%, and the skin in 15-34% of the cases (Kath,

1993; A1bert, 1992; Gragoudas, 1991; Rajpal, 1983). After metastatic uvealmelanoma

is diagnosed the patient's life cxpectancy at the time of diagnosis is 2 to 7 months

with, and 19 months without liver involvemcnt (Kath, 1993; Albert, 1992; Gragouelas,

1991; Rajpal, 1983; Bedikian, 1981; Char, 1978). The orbitalrecurrcnce rate ranges

from 0.5 to 3 %, but increases to as high as 23% in cascs of extraocular ex ten sion of

the primary uveal tumor. Death by an orbital recurrence itself is exceedingly rare;

howevcr, when a patient presents an orbital recurrence, his life expectancy is very

poor (0 to 8 %), because he or she is a1most sure to develop distant metastases

(Kersten, 1985; Affeldt, 1980; Shat1lmas, 1977). Since the eye itselflacks a Iymphatic

drainage system, lymphatic dissemination is only f0l111d as a late event in the

metastatic process in cases of uveallnelanoma with extraocular extension.

Considering the short survival time of uvealmelanoma patients suffel'Îng from

metastatic disease, especially whcn the liver is involved, we have to dcduce that the

success of treatment modalities is to date rathel' limited. Por the treatment of

metastatic lIveaI melanoma six approaches have been describeu: combined

chemotherapy (Kath, 1993; Gragoudas, 1991; Einhorn, 1974), chemoiml1lunotherapy

(Gragoudas, 1991; Bedikian, 1981), chemotherapy and surgery (Kath, 1993; Fournier,

1984; Rajpal, 1983), intrahepatic chemoembolization (Kath, 1993; Mavligit, 1988),

acljuvant chelllotherapy (Sellami, 1986) and immunotherapy (Mitchell, 1994). Some

authors advocate the application of adjuvant thcrapies, such as immullotherapy with

the mcthano1-extractionresidue of the Calmette-Guerin bacille, or chemotherapy

inclueling vinblastine, thiotepa, methotrexate, elcticene anel procarbazine (Mclean,

1990; Sellami, 1986). However, neithcr the Illain, nor any of the adjuvant therapies

13

Page 14: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______________ -'C",H"AI'T"-E"R--'-I __ _

have proved truly el'f'ective yet. ln only a few cases a partial response ta treatment

could be delllonstrated and then only when Chclllotherapy was used, whether or not

combined with an anothcr treatment lllodality; one such study reports an extension of

the mean survival time of 3 months (Gragoudas, 1991). Besides, one case report

describes the remarkable successful treatment of a prilllary uveal melanoma by the

sole application of aclive specific iml11UnOlherapy with cutaneOlls melanoma lysates

(DETOX), that induced a reduction or the tumor hcight rrom 4.2 to 2.4 lllm (Mitchell,

1994). A pilot study th at investigates the effects of treatment of metastatic uveal

lllclanoma by recombinant alpha-2b-interferon with dacarbazinc, vincristine,

bleomycin and lomustine has recently been initiated by the Ophthalmic Oncology

Group of the EORTC (Pyrhönen, 1992).

14

Page 15: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

I NT RaD u CI'~-"O"N,-' ____ _

1.2 Aims and scope of the thesis

In the preceding section of this chapter, the clinical aspects of uvealmelanoma have

been described; the present anel next sections are dedicated to a presentation of the

aims and background of the thesis itself, the collection of articles taken up in Chapters

2 to 9 that also have been published or submitteel elsewhere in the periocl bet ween

1993-1996. As will have become clear from seetion 1.1, less than 2 % ofthe uveal

melallOlna patients presents distant lnctastases elctcctable with the currcntly availablc

methoeIs at the time of eliagnosis of the primary tumor, while 50 % of them eventually

dies from it. The c1evelopment of new anticancer strategies is therefore particularly

significant for these patients.

Thc studies in this thesis ultimatcly aim at the elevelopment of a therapy

aeljuvant ta thc available primary trcatmcnt lllOdalitics, elirccted at counteracting any

aspect of possiblc metastatic disseminatian, as weil as at a more effective therapy of

metastatic uvcal melanoma itself. Evidently, metastatic dissemination involves a

complex series of genotypic anel phenotypic changes; elissemination of uveal

melanoma includes tumor cell motility, invasion of the tumor border, the sclera anel

the vascular system, survival in anel extravasation from the circulatory system,

colonization of a distant site, angiogenesis, and avoidancc of host illUl1Une responses

(Fidler, 1990). To date, the develoJlment of anti-metastatic therapeutic agents or, in

general, antimetastatic modalities, has been halllpered by incomplete knowleelge of the

metastatic process; apparently, qllite a large number of genotypic anel phenotypic

changes havc to be elucidateel hcfOl'C ncw anticanccr stratcgics can be practiced

(Liolla, 1991). With the 1Iltimate aim of eleveloping new therapeutical modalities, this

thesis intends to contribute to the knowleelge of the nature anel causes of metastatic

uveal melanoma by investigating some of the key elements of these changes involved

in thc mctastatic process.

In this context th ree (interrelateel) questions arise: First of all, what arc the

origins of thc mctastatic seedings? How can we localizc and craelicate these,

preferably in the very first stage of eleveloplllent and before they bccome clinically

manifest? Secondly, how do we identify which patients are especially at risk, i.e. what

clinical and histopathological factors are involved in the llletastatic process, so that we

may adapt their treatment accordingly and as early as possible? Thirelly, how can uveal

15

Page 16: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1

melanoma-related metastatic disease be suecessfully treated? Might its development

be bloeked, for instance by the early applieation of immunotherapy?

Regarding the first question, the eradieation of (micro-) metastases at the

earliest stage, we studied the procedure of pre-enucleatioll irradiatioll (Chapter 2).

Regardillg the secOird question, we investigated tlre morphological and

inllllunohistochemical properties of clilliealmallifest metastases and their

corresponding primary uveal tumors (Clrapter 3), as weil as the possible role of neural

eell adhesion molecules (Chapter 4) and of the putative metastasis suppressor gene

NM23 (Chapter SJ. In view of the third questioll, the treatment of metastatic disease,

we undertook several experimental studies. Firstly, we established and characterized

primary and Illetastatic uveal melanollla eeillines, vital to the study of the biological

properties of uvealmelanomas (Chapter 6). We then studied the expression of

melanoma-specific anel associateel allligens in uveal melanOlna celllines. These ceIl

lines Illight be used in speeifie illllllunotherapy (Clrapter 7). Furthenllore, the aetual

susceptibility of uveaimelanollla cells to natural killer cells was experimentally tested

in nude mice (Chapter 8). Finally, we experimented with growth of uvealmelanomas

in a chicken embryo model (Clrapter 9). Tlre background of these studies will be

introduced illmore detail in the fol1owing section.

16

Page 17: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_______ --'I"N-'T-"-R 0 [) U C T ION

1.3 Introduction to the studies

1.3.1 The ol'igins of metastases (Chaptel' 2)

Much discussion remains on the origins of the initialmetastatic seedings th at give rise

lo the distant mctastatic Iesions of uveal melanoma. Among others, Zinuncrman and

associates suggested lhat tumor cells are disseminated by lhe physical manipulation of

lhe tumor-containing eye during enucleation (Zimmennan, 1978). In view of lhis

possible risk, the procedure of pre-enuclealion irradialion was introduced, preslllning

that eraelication of these micro-melastases by irraeliation woulel prevent the eventual

devc\opmenl of distant melastases (Char, 1982). Yel, Manschol anel Van Peperzec\

(1980) argued th al possible micro-metastalic seedings inc!tlCed by enuc1eation would

not beeome manifest wilhin 6 years following lhis procedure, which would not explain

lhe high l110rlality rate in the second to fourth year aner the trealment (see seclion

1.1.3). They therel'ore suggested that the l11icro-metàstatic spreaeling oceurs even

before, anel independenlof treatmenl (Manschal, 1980). To date, elata on the subject

have mainly been in favor of Manschol's hypothesis (Manschol, 1992, id, 1980).

Firstly, since lIvealmclanoma metastases are rarely discovered prior 10 the primary

tumor and they can occur many years after enucleation (see section 1.1.3), in survival

rates the major function for the survival time is the growth rate of the cells, nol the

lumor size of lhe uvealmelanoma. In olher words, large tumors consist of more

aggressively growing cells cOl11pared to small lumors. Secondly, the procedure of pre­

enucleation irradiation, applicd in the several centres in various doses, has not proved

lo have any beneficial effecl on survival (Luyten, 1995; Augsburger, 1990; Bornfelel,

1989; Kreissig, 1989). On the contrary, in one study irradiation before enucleation

showeel an inverse effect on survival (Char, 1988). These finelings suggesl a minor role

of the physicalmanipulalion of the tumor-containing eye eluring enucleation. The cure

mte and survival time calculations as described in seclion 1.1.3 are in concordance

with lhe hypothesis of Manschot, lherefore we can possibly adopt the view yet th al

mctastatic uveallnelanoma and melanoma-related death is the eventual consequence

of micro-melastasis that takes place independent of the initial trealment. Chapler 2 of

this thesis discusses a comparative clinical study of a group ofl1Ve~ll melanoma

patients lreated by preoperalive irradialion anel a hislorical conlrol group trealeel by

enuc1eation alone, which considercd a considerably longer follow-up (71/2 years).

17

Page 18: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

________ ~C~'H~APTER I

1.3.2 Clinical mul histopathological factors (Chapter 3, 4 and 5)

Regarding thc choice of how to manage patients with a primary uveallnelanoma, the

risk of developing metastases has to be inferred mainly from clinical observations, The

detcrmination of these clinical prognostic factors has become more important with the

increase in assessment of radiotherapy and other conservative treatments (see I. 1.2),

because histologicalmaterial can not be obtained in these types of treatment. Several

clinical and histopathological prognostic factors have yet been proposed; obviously,

their clinical impact is of limited value as long as thcre are na treatment protocols

directed by these clinical prognostic factors, nor successful adjuvant therapies for

high-risk patients (Mitchell, 1994; Albert, 1992; Mclean, 1990; see also section Ll.3),

Fram clinical observation the following three clinical risk factors have been

elucidated: largest tumor dimension in contact with the sclera (LTD), location of the

anterior tumor margin, and the presencc of extraoclIlar ex ten sion of thc tumor

(Shields, 1995; RUlllmelt, 1995; Augsburger, 1990), Furthermore, as mentioned in

I. 1.3, patients of higher age and of the male sex have proved to have an increased risk

to develop metastases (Rummelt, 1995; Luyten,1995; Folberg, 1993; Coleman, 1993),

With respect ta smult uvealmclanomas, several ather clinical risk factors have been

appointed, of which the most important one is demonstrated growth of the tumor,

Growing slllallmelanomas cOIll]lared to non-growing resull in an eight times

increased risk to develop metastases (Shields, 1995), Bcsicles demonstrated tumor

growth, posterior tumor margin touching the optie disc and initial greater tumor

thiekness were found to be important predictive factors for metastasizing of small

melanomas (Shields, 1995), Predictive factors for tumor grawth itself are: initial

tumor thickness, posterior tumor margin touching the optie disc, symptoms of Ilashes,

floaters, and blurred vis ion, orange pigment on the surface of the tumor, and the

presencc of subretinal tluid (Shields, 1995),

In tumors of patients treated by enucleation, additional histopathological

parameters ean be evaluated, sueh as eell type, vaseular networks, eytOlllorphometrie

assessmcnts of the nucleoli and mitotic figures, as weil as cytogenetics and tumor­

infiltrating Iymphocytes, Of these histopathological parameters, when evaluated along

with the clinical prognostic factors rOl' the risk of developing metastases, the

assessment of the largest tumor diameter (LTD) is again the most reliable independent

18

Page 19: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

I NTROD UCTlQ.N ____________ _

one (Coleman, 1993; McLean, 1993; Diener-West, 1992). The use ofthe cell type

classification as an independent prognostic factor is less powerful. Although the

modificd Callender classification is mostly applied (spindie, mixed and epitheloiel),

several studies have simplified this classification to the presence or absence of

epitheloiel cells (Mooy, 1995; McLcan, 1983). To improve the prognostic value of the

cell morphology, more reproducible parameters have been elevelopeel, as for in stance

mean diameter of the ten largest nucleoli (MLN) (McCurdy, 1991). The assessment of

morphometric analysis is ho wever time consuming anel therefore not very feasible 1'01'

daily practice (Mooy, 1995; Rummelt, 1995).

Again in the multivariate analyses of eure rates and survival time, LTD has been

shown to proviele the greatest prognostic insight (Gamel, 1993). A meta-analysis of 39

publications showed the following 5-year mortality rat es - which are higher than is

generally presumeel - in l-elation to LTD: 16 % of the patients with a smallmelanoma

(LTD < 10 mm LTD); 32 % of those with a medium melanoma (LTD 10 - 15 mm), anel

53 % of those with a large melanoma (LTD > 15 mm) (Diener-West, 1992). Multivariate

analyses further show that the percentage of mitotic figures as weil as the occurrence of

tumor-infiltrating Iymphocytes (TIL) is significantly associatcd with uvealmelanoma­

related eleath (Whelchel, 1993; Folberg, 1993; De la Cruz, 1990). Thc prognostic value

of these parameters and some ather immunohistaehemicalmarkers for proliferative

aetivity were faund fa associate with uvealmelanoma-rclated death, sueh as the

expression ofMIBI, MYC (Mooy, 1995; id., 1995; id., 1990; Bardenstein, 1991), still

have to be investigateel in a prospcctive study (sec scction 1.3.3.2 below on

eytogenetics).

Although LTD is the most frequently used prognostic factor, recent reports show

that the presencc of closed vascular loops or networks in uvealmelanomas is highly

associated with melanoma-related death, and that these vaseular networks might be of

even more prognostic significance than LTD (Runuuelt, 1995; Pe'er, 1994; Rummelt,

1994; Folberg, 1993). Finally, in order to select patients at risk, a new minimal detection

technique using rcverse-transcriptase polymerase-ehain-reaction (rt-PCR) has been

dcveloped. t't -PCR detects tlte tyrosinase gene and is thereby able to demonstrate

disseminating melanoma cells in peripheral blood of patients suffering from both

primary and metastatic disease (Tobal, 1993). The clinical value of this technique for the

/9

Page 20: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1

detection of circulating melanoma cells is however not yet c1ear (Foss, 1995).

In Chapter 3, 4 and 5 three inullunohistochemical studies that ShOldd contribute

to the formulation of differences between primary anel metastatic uveal melanoma are

described. Clulpter 3, besicles a morphologieal study of primary and related metastatie

lesions, presents an cvahlation or the sensitivity or these lesions to three melanoma­

associated antibodics: HMB-45, S 100 and NKI-C3. Chapter 4 investigates, according to

thc gt'O\ving evidcnce th at cell adhesion molecules may influence the binding anti

invasiveness of hllman cancers (Niederkom, 1993, Nicolson, 1988), whether there is an

assoeiation between the percentage or neural cell adhesion molecules present in the

tumor eells and their metastatic behavior. Chapter 5, then, following studies on breast

caneer (Steeg, 1991), investigates whether an in verse relationship exists with regard to

the expres sion of the NM23 gene and uveal-Illelanoma related death. (For a detailed

description of the cytogenetics involved in uveal melanoma, see section 1.3.3.2 below).

1.3.3 Experimental studies (Chapter 6 to 9)

1.3.3.1 Uvealmelalloma eelllines (Chaptel' 6)

The study or biological and molecular properties of thc successive steps involved in

melanoma progression, for instanee in an experimental animal 1110dcl (see section

1.3.3.3 below and Chapters 8 and 9), requires the availability of cultured celllines. To

date, only a few continuously growing ceHlines have been established of primary

llvealmelanomas, and of metastatic llveal melanomas none (Ma, 1995; De Waard­

Siebinga, 1995; Massarelli, 1994; Aubert, 1993; Soulieres, 1991; Kan-Mitchell, 1989).

In contrast, several cell lines have successl'ully been established of cutaneOlIS

mclanomas. This contrast can possibly be eXplained by the relatively high prevalence

of cutaneous melanoma cOlllpared to llveal melanollla (10: I) (Albert, 1984).

Furthennore, cutaneous melanomas are more feasible to culture as they frequently

metastasize to Iymph noeles, where the lesions are easily obtained. Metastatic uveal

lnelanoma lnaterial, on the other hand, is e1itTicult to obtain because uveal melanomas

lead mainly to hepatic metastases, which are mostly confil'med by an examination of

FNAB mther than by a true biopsy. Aelditional biopsies or excisions fol' culturing

purposes have to be taken together with the diagnostic biopsies, because if they are

20

Page 21: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

I NTRODUCTlq"N'---__ _

merely taken for culturing purposes, patients of ten do not agree. Besielcs, these

patiellIs arc often treateel beyonel the scope of the ophthalmologist , which forms

another impediment to the obtaillment of fresh metastalic uveal melanoma material.

Chapter 6 reports on the successful establishment anel characterization of two

primary anel three metastatic uveal melanoma celllines. Although aculturing

technique using a fibroblast feeeling layer has also been described (Albert, 1984), our

protocol follows most others in the use of a culturing medium of 10% fetal calf serum

with penicillin anel streptomycin to select the most aggressive anel proliferating ceHs.

1.3,3,2 1\ulIor gene tics mul inllllunology (Chapter 7)

1\ullorgenetics

In recent years several genes, both oncogenes and tumorsuppressor genes are founel to

play an important rale in carcinogenesis and tumor progression. Mutations or loss of

these genes leads to a growth advantage of the ceU or in case of a metastatic process,

outgrowth of a tumor cell population. Although some of these mutations in oncogenes or

tumor suppresser genes can be used for e1iagnostic purposes, the unelerlying basis of

genomic instability, leading to these lIlutations in the tumor ccH is at largely unknown.

Ohvious candidate genes arc genes involveel in ceH cycle control or genes involved in

DNA rep air mechanislIls. An example of a gene involved in cell cycle con trol is the

CDKN2 gene which is mutated or deleted in familiar 111elanoma kinelrcd's and several

ot her human caneers (Cairns, 1995). This gene wiH he discusseel in next paragraph. A

seeond type of genes are involved in mechanism which lead to frequent mutations,

deletions and insertions, and c10ning of a human mismatch repair gene, ilMSH2, Iinked

to human non·polyposis eoloreetal careinoma (HNPCC) was major step forward. To

date several delects in other DNA repair genes are found in HNPCC (ilMSH2, ilMLHl,

ilPMS2 and ilPMS J) anel are assoeiateel with genomie instability (Liu, 1996).

It is unknown whether in hllman llvcal melanoma a similar mechanism leads to

genomic instability anel provide genetic changes whereby certain cells aequire new

phenotypic traits th at results in earcinogenesis, tumor outgrowth and further the

metastatic process. Tn order to identi!'y the genes and their proteins involveel in

carcinogenesis or tumor progression several approaches can he used.

21

Page 22: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__________ ~ CHArTER I ____________ _

Linkage studies

Familialuvealmelanoma is quite rare (Young, 1994; Singh, 1994; Canning, 1988). In

one report, the mean age of the patients with familial uveal melanoma at the time of

presentation is significantly younger than usual (Canning, 1988). Affected members

are always from onc or two generations; the pattern of inheritancc is suggested to be

autosomal dominant with partial pellctrance, or more polygenic. In a ClllTcnt study on

familial uvealmclanoma at our institute, threc families were fOUlld with two affected

mcmbers each, linked several generations from each athel'. In contrast, familiaI

cutaneous mclanomu is quite common: in about 10 % of thc cases, at least ane of the

family members of the cutaneOlIS melanoma patient is affected too.

Bergman and associates (1986) describe several families with thc so-ealled

familiar atypiealmultiplc mole and malignant melanoma syndrome (FAMMM), which

has been mapped on chromosome 9p21 (Cannon-Albright, 1992; Gruis, 1993).

Patients from FAMMM families develop multiple naevi, atypical naevi (dysplastic

naevi) and cutaneous melanoma. Although several authors have described FAMMM

patients having an ocular melanoma as we1l, there is still some controversy regarding

the association between FAMMM and oClllar 111elanol11a (Seregard, 1995; Bataille,

1993; Vink, 1990; Taylor, 1984). The putative FAMMM gene, CDKN2, or multiple

tumor suppressor I (MTS1) gene, that has been cloned, eneades for the protein

PI éNK4, whieh was previausly identified through its ability to inhibit eyclin­

dependent kinase-4 (CDK-4) (Cairns, 1995; Kamb, 1994). Currently, our department

of Ophthalmology in eooperation with the (MGC) Institute of genetics, Erasmus

University, Rotterdam and the Department of Antropogenetics and Department of

Dermatology, Leiden University, The Nethcrlands, investigates the oceurence of

CDKN2 gene lllutations in uveal melanoma in vivo and in uvealmelanOlna celllines

by use of the single-stralld-eonformation polymorphislll (SSCP) teehnique.

Cytogelletics

Cytogenetic analysis of tumors may provide information on nonrandolll chromosomal

aberrations Iike deletions, transloeations and amplifications, 1'01' it provides the

possibility of detailed studies of loss or gain of certain oncogenes anel tumor

22

Page 23: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

1 N T R 0 D U C.~·l~' I~O~N~ __ _

suppressor genes located on thc chromosomes of intercst. Cytogenetic analyses of

uveal melanOllla have been perfonncd onlimlted series of primary uveal melanomas,

but sofar not on lnetastatic llveal melanomas. The most common genetical changes

that were found in primary uveal melanoma arc: 10ss of chromosOlllc 3, 6q alld 8p, and

gain of chromosome 6p and 8q (Horsthemke, 1992; Horsman anel White, 1993;

Wiltshire, 1993; Dahlenfors, 1993; Magauran, 1994; Singh, 1994). Monosomy of

chromosome 3 anel isochromosome 8q have been found to be associateel with ciliary

boely melanomas anel high risk for metastatic eliseasc (Preschel', 1992; Sislcy, 1990).

Two stuelies on uveal melanoma that used a recently eleveloped technique,

comparative genomic hybrielization (CGH) (Kallionicmi, 1992) have confinned these

finelings (Speicher, 1994; GoreIon, 1994). Hence, loss of chromosame 3 in cells of the

primary tumor might he useel as a prognostic factor.

The role of known oncogenes anel tumor supprcssor genes that have previously

becn described with regard to other types of cancel' has been studied with respect to

uvcal mclanOllla too. Firstly, whcreas in cutaneous melanoma activating NRAS

mutatÎons that associated with both sun-cxposure and malignant progression were

frequently fOUlld (Bali, 1994), two studies of uvealmelanoma could not elemonstrate

NRAS, HRAS or KRAS mutations (Soparker, 1993; Mooy, 1991). Seconc1ly, the tumor

suppressor TP53 gene, locateel on chromosome 17p and encoding a Mr 53 Kd nuclear

phosphoprotein, is known to suppress cellular proliferation in its native form, whereas

the mutant form controls oncogenic potential (Harris, 1993; HolIstein, 1991), In one

study of primary uveal melanoma, an overexpressioll of the TP53-protein was

demonstrated in 12 out of 18 melanomas; two of them showeel a point mutation in

TP53 (Tobal, 1992), Furthcnnore, a Victorian family has been elescribed in which four

generations developeel uvealmelanomas together with breast cancel'. In this family,

inununohistochemical analysis showed a mutant TP53 in the tumors, Unfortunately,

genn !ine material was not tested (Jay, 1993), Thirelly, the protein product of the MYC

oncogene is kllOwn to play an important part in cellular pro!iferation, elifferentiation,

and tumor fonnation, presumably by modulating the expression of genes involveel in

these processes (Peltenburg, 1994). Prognosis-related cnhanced expression of this

gene has been observed in human carcinOInas (Peris, 1991). In cutaneOlIS melanoma,

high MYC expressioll inhibits the expression of class I HLA, which makes the cells

23

Page 24: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1 '--'------

less prune tu MHC-restricted cytotoxic T-Iymphocytes (CTL) immunity on the one

hand, and more sensitive to natural killer (NK) cells on the other (Verstecg, 1988). Wc

will come back to this bel ow. The MYC gene is locateel on chrol11osome 8q24 (Hu man

gene lllapping 11, 1991), Ivhich shows a high expression in hUlllan uveal melanoma

(Spcicher, 1994). In uveal melanulllas, the immunohistuchemical staining of the MYC

protein was found to correlate with a proliferative index assessed by Ilow cytometry

(Royels, 1992). Another immunohistochemical study shows an inverse correlatiun

bet ween the MYC anel the BCL2 protcin, which plays a role in apoptosis.

Furthennore, the expression of MYC is significantly associateel with uveal-melanollla­

related death (Mooy, 1995).

Finally, the NM23 gene, due to its e1ifferential expression on relateel low anel

high metastatie potentialmurine K-1735 melanuma celllines, is proposeel as a

metastatic suppressor gene (Steeg, 1988). To date, two human NM23 genes have been

ielentified (Stahl, 1991; Rosengard, 1989), each encoding t'or a 17 kDa protein; they

have been mappeel on chromosome 17q21.3-22 (Backcr,1993). Expression of NM23 in

human tumors is most extensivcly stuelied in breast cancel'. Reeluced mRNA and

protein levels in e1uctal breast cancer have proveel to be correlateel with Iymph node

metastases, decreased disease-free survival and decreascd overall survival (Royds,

1994; id, 1993; Hennessy, 1991; Barnes, 1991; Bevilacqua, 1989). A similar inverse

rel at ion between the expression of the NM23 gene and reduced metastatic potential

could be demonstrateel for several other types of human cancel' (Bertheau, 1994;

Koelera, 1993; Yamaguchi, 1993; Florenes, 1992; Nakayama, 1992). However, in

some types of cancer, NM23 expression appears not indicative or, in reverse, is

increased in highly metastatic tumors (Fishman, 1994; Leone, 1993; Engel 1993;

Raelinsky, 1992; Haut, 1991).

The role anel prognostic value of NM23 protein expression in uveal melanoma

is more elaborately described in Chaptcr 5 of th is thesis; Cytogenetic analysis of

prÎmary anel metastatic uveaimelanollla cell Iines is presented in Chapter 6.

Cytogenetic analysis by karyotyping and Iluoresence in situ hybridization (FTSH) on

fresh primary and metastatic uveal melanoma are currently performed at Department

of Genetics of the Eraslllus University Rotterdam.

24

Page 25: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_J NTRODUCTION

Immullology

The nl1uor role of lhe immune system in conlrolling uvealmelanoma has nol yel been

elueidated. This role ean be deduced from clinical observations: lïrslly, primary uveal

mclanomas sometimes remain loealized far long periods withoul showing any

progression, and metaslases arc often only abserved eleeades aftel' enucleation of the

primary tumor. Secondly, oleler patients having a primary uvealmelanoma have an

inereasecl risk to develop melaslases and when they c1evelop metastatic c1isease, they

have a significantly sharter survivallime (see 1.1). Thirdly, although only in rare

cases, spontaneous regression of primary and metastatic lesiolls has been elescribed

(Jensen, 1974).

FlII'lhel1nore, the immunogenicity of uvealmelanomas and their related

melastalic lumors has been associaleclwith the presence of Iymphoeytic infiltration. In

5 to 12 % of lhe primary tumors T-Iymphoeytes (CD3-positive) and ta alesser extent,

numbers of natural killer (NK) eells, monocytes and maerophages can be observed.

The Iymphocylic infiltration comprÎses cytotoxic/suppressor (CD8-positive) and

helper/indueer (CD4-pasitive) eells (Tobal, 1993; Meeeham, 1992; DUrÎe, 1990). The

presenee of Iymphoeytic infillralion has been associatecl with a better prognosis in

culaneous melanoma, but, cliriollsly enough, with a poor prognosis in uveal melanoma

(Whelchel, 1993; Folberg, 1993; De la Cruz, 1990). Most likc1y, this negalive

assoeialion with regard tn uvealmelanoma can be eXplained by the fact lhat the eye is

an immune-privileged site anel that, for the generation of a T-Iymphocyte mediated

immune response, disseminatecl tumors cells are neeessary (Whelchel, 1993).

However, the fact that c1eteetable quantities of Iymphoeytes can be fOlllld in

lIveal melanomas means that the immune system is somehow stimlilated anel hen ce,

that these eells must express certain antigens that are reeognized as foreign by the

tumor-bearing host's immune systelll. lt is commonly assumed that tumor cells, as a

result of the expression of mutated or viral genes or a deregulated expression of

normal genes, produce eertain proteins that are either not proclueed at all or in much

lower quantities in normal eells (Germain, 1994). These abnol1nal proteins became

tumor antigens at the ceIl surface that serve as target for effectors of natural immllnity,

sueh as NK-cells, or far specific immune responses by eytotoxie T-Lymphoeytes

(CTL's), if they are associated to miUor histocompatibility complex (MHC) class I

25

Page 26: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1 ______ _

molecules and Jl2 microglobulin (Germain, 1994). Once these surl'ace antigens or

peptidelMHC class I molecule complexes are recognized by NK cells or specific

CTL's, thesc kill the infected or the tumor cell. The immune system thus provides a

useful target 1'01' cancel' treatment modalities.

The tumor antigens expressed by melanoma cells can either be melanoma­

specific, th at is unique ta melanomas, or melanoma-associated, in which case they are

not only present in melanoma cells but in all melanocytic derived cells. The first

discovereel melanoma-specific antigen, MAGEI (melanoma antigen), was elucidated

by CTL cl0l1es isolated from cutaneous melanoma patients and cl0l1eel by use of

peripheral blood lymphocytes (PBL's) of a patient who had been immunizeel with

autologous tumor cells (Van Der Bruggen, 1991; Herin, 1987). The immune system of

the patient recognizcd the protein encoded by th is gene in association with HLA-Al

(Traversari, 1992). The CTL clones were subsequently used to select antigen-loss

variants anel this led to the definition of several other antigens that are expressed by

autologOlIs melanoma cells (Van De Eynde, 1989). Thus the MAGEI gene, that lacks

similarity ta ather sequences in Cl1rrent databases, was found to be a member of a

multigene family. The product of another mem bel' of this family, cal led MAGE3, has

also been shown to be recognized by HLA-A I-restricted T-cells (Gaugler, 1994).

MAGEI and MAGE3 genes are expressed in culaneous melanomas in a frequency

of respectively about 40 % and 69 %. They have been f0lllld in several other

types or tumors as weil (e.g. squamous cell carcinoma, small ceillung carcinoma

and sarcomas) (De Smet, 1994), but not in normal tissue except for the testis

(De Smet, 1994).

Two melanOlna-associated antigens, th at besides in tumor eelIs, are expressed

in normal melanocytes as weil, are Iyrosillase and GP 100. Both genes are involved in

the melanin synthesis and are also recognized by CTL's. The lyl'Osillase gene was

cloned by means of CTL clones which were oblained from PBL's of cutaneOlIS

melanoma patients aftel' repeated in vitro tumor stimulation (Wölfel, 1994; Brichard,

1993); ils protein product was found to be recognized at the cell surl'ace in association

with the HLA-A2 molecule. The product of lhe other melanoma-assodated antigen

GP 100 is recognized by HLA-A2-restrieted tumor infiltrating Iymphocytes (TIL's) as

weil (Bakker, 1994; Kawakami, 1994a; Aclema, 1994; id, 1993;). The antigen encodecl

26

Page 27: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

IN T R 0 0 U.C~'f~'I~O~N~ _____ _

by CP 100 cDNA is recognized by the monoclonal antiboelies NKI-beteb, HMB-50

and HMB-45 (Adem a, 1993); these melanocytic lineage-specific Mabs are among the

best diagnostic luarkers t'or cutaneous and uveal melanoma available to date (Steuhl,

1993; Vennegoor, 1988; Vogel, 1988; Van Der Pol, 1987; Gown, 1986), Recently, a

thirelmelanoma-associated antigen has been cloned, MART!, which has highly

immunogenic common epitapes that are recognized by the majority of HLA-A2-

restricted melanoma specific TIL's isolated t'rom HLA-A2 melanoma patients

(Kawakami, 1994b; iel, 1994c),

Besides expression of antigens at the cellular membrane, recognition anel

cytolytic response by CTL's reguires expression of MHC molecules, Cells in which

the MHC expression is down-regulateel are not recognized by CTL's, In 75 to 85% of

uvealmelanomas, MHC class I molecules are expresseel (Meecham, 1992),

Endocytosed antigens anel exogenous antigens are associated with MHC class II

molecules, However, recently it has been shown that these cndocytosed (exogenous)

antigens can also be presented in the context of MHC class I molecules (Kovacsovics,

1995), MHC class I expression in uvealmelanoma metastases has not been reported

on yel, In other types of canccr no correlation exists between the level ot'MHC class I

expression and immunogenicity, It appears likely that a cmtain MHC molecule has a

strong preference to binel a particular peptide, and that the immunodominant peptide

from a certain antigen binds only to a particlJiar MHC class I molecule; hencc, only

tumors th at express the allele of that peptide will be immunogenic (Germain, 1994),

Other MHC class I molecules will probably also bind this peptide, but much less

effective,

As mentioned above, the immune system appears as a promising target, or

tooi, in the trealment of melanoma, From clinical and experimental stuelies on

cutaneous melanoma, evidence was obtaineel that specific CTL's indeeel play an

important role in the antitumor response against the melanoma (Barth, 1990; Mule,

1987; Greenberg, 1981). In vitro studies with TIL's and PBL's isolateel t'rom cutaneous

melanoma patients have demonstrated that these CTL's are able to recognize

melanomas in a MHC-restricted mannel', indicating that CTL's recognize specific

tlll1l0r antigens.(Coulie, 1992; Topalian, 1989; Herin, 1987; Knuth, 1984; Mukhelji,

1983). In a variety of murine tumor modeis, the adoptive transfer of murine TIL's has

27

Page 28: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1

shown these to mediate in the regression of establishedmetastases (Spiess, 1987;

Rosenberg, 1986). TIL's can also mediate in cancer regression when adoptively

transferred to the autolog(lUs patient (Rosenberg, 1992; id, 1988).

In vitro studies of uveal mclanoma also demonstratedmelanoma-speeific CTL

reactivity in a TIL culture. The TIL culture was expanded by IL-2 and the Iymphocytes

were fOlll1d to destroy the patiellt's own uveal melanollla cells and not those derived

from other patients, thus showing a specific inll11unoreactivity ror the autologous

tumor cells (Ksander, 1991). AutologOlIs as well as allogenic celilines have been used

to generate CTL clones from PBL's by mixed IYlllphocyte/tumor cultures (MLTC). A

higher aI11Clunt of melanoma-specifie CTL clones could be derived from autologous

compared to allogenie MLTC's (42% versus 14%) (Huang, 1994; Kan-Mitchell,

1991). Sinee it is relatively di rficult to culture uvealmelanoma cells, and melanoma­

specific CTL ean also be isolated [rom allogenie MLTC's, aetive specific

immunotherapy of uvcalmelanoma patients should be reasible.

In addition to CTL's, natural killer (NK) cells are probably also effector cclls of

natlll'al and acquired immune responses ta melanamas. NK's con.stitute a heterageneaus

group of cells ex pressing CD 16 that can Iyse both autologous as weil as allogenic

tumor cells (Warren, 1993). The basis for their specificity is not yet understood. Unlikc

T-cells, NK-eells do not express T-cell antigen receptors, and they kill target cells in a

non major histocompatibility complex (non-MHC) rcstrictecl way. Low MHC class I

expression in tumor cells makes the target cells more susceptible to NK's (Ma, 1995;

Versteeg, 1989; id, 1988). The enhancement of NK-cells by alpha-interreron or

Linomide in a mudne ocular melanoma model demonstratcd a significant decrease in

pulmonary metastases. This shows that natural killer cells, likc CTL's, are probably

important primary effector cells ror the control of metastasis, which might be utilized in

trcatment moclalities of metastatic disease (Harning, 1989; Yokoyama, 1986).

The expression of the melanoma-specific MAGE-family genes as weil as the

two melanoma-assoCÎated genes, tyroSillGSe and GPIOO werc cvaluated in uveal

Illelanoma, using the celllincs that had been established in thc institute (see 1.3.3.1).

In order to nnd out whethcr uveal melanoma cells constitutc an effective target for

specilïc immunotherapy, the MHC Class 1 expres sion in the celllines was also studied.

This experiment is discussecl in Chapter 7.

28

Page 29: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INTRODUCTION

1.3.3.3 Animal models (Chapter 8 and 9)

Thc study of growth and metastalic capacity in an experimental animallllodel is one

of the main interests of the eslablishmenl of hu man uvealmelanoma celllines (see

1.3.3.1). The major problem wilh animal models is the host's immune response againsl

the donor material. Fll1'thermore, to be able la compare the metastalic process in the

model with the hUl11an situation, not only human uvcallnelanoma eell lines are

necessary but also the local organ-specific faclors are important; to study growth,

invasion and dissemination. Therefore ortholopic Iransplantation is advocaled (Ma,

1995; Niederkolll, 1993; id, 1981).

To date, the best experilllentalmodel is the nudc lllouse. The athymic,

immunodeficient oude mouse provides an in vivo model with low residual immunity

(Hi 11, 1991). Actually, il did not appeal' very successful for human uveal melanoma

initially, but the problems turned out to be due la the usc of uncultured melanoma cells

that were transplanled heterotopically in the subculis (Albert, 1980). Furlhennore, it

has become increasingly cic ar that tumors have to be transplanted at an orthotopic site

to farm metastases. Two other research groups did succeed in an orthotopic

transplantation of human uveal melanoma eells th at dClllOnstrated metastatic spreading

(Ma, 1995; Niederkolll 1993; Soulieres, 1991). A cOlllparable murine modcl has been

IIsed 10 sWdy Ihe effect of difluOl'omethlhylornithine and dacarbazine (DnC) in

prevenling melaslatic spreading of B 16 melanoma (Illurine, cutanalIs) cells that had

been injected subcutaneously and inlracamerally (Sanbolll 1992; id, 1992). This well­

charaeterized ani mal model make use of B 16 melanomas eells, which is a murine

cutancous melanolllas transplanled in the syngenic C5?BL/6 mice. These B 16

lllclanOlnas lnctastasize to the lungs as all eutaneous melanomas (Harning 1987; iel,

1989; Niedcrkorn lY 1984; id, 1984; id, 1984). This model is howcver not applicab1c

for human uveal melanoma cel1s.

Two other mouse models that should be quile appropriate for experimental

studies are the severely immunocompromised nude mouse (CB-l? SCID) and the

triple-inullune-deficient beige mouse (bg)/nude (nu)/xid (BNX). Both species lack

functional B- and T-cells, while in the nude mice only the T-cells are lacking (Hili,

1991; Mulé, 1991). Furthermore, the BNX lllouse does nol possess functional

lYlllphokinc activating killer (LAK) cells. However, these models have not yel been

29

Page 30: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______ (;H A PT ER

appliecl in research on hlllllan uvealmelanoma.

Another anill1alll1oelel to study growth of human uvealll1elanoll1a cells is the

rabbit immllnosllppresscel by cyelosporin. In one stuely, an OCM-l human uveal

melanol1u\ cel1line derived I'ro111 a primary uveal melanoma was xenografted in the

anterior chall1ber of the rabbit. Growth coulel be demonstrated, but no metastatic

seeding was obtained (Kan-Mitchell, 1989). The main problem of this moelel is that

the imll1unosuppressing cyclosporin can only be supplied for about 60 days beeallse of

its toxic effects on the rabbit, which seemingly limits its practical use in studies on

metastasis and tumor biology. For the implantation of the so-called Greene melanoma,

whieh is a murine cutaneous melanoma, the albino anel Dutch rabbi! have also been

used. This moelel is applied in the study of new treatment modalities of the prill1ary

tumor but cannot be used to stuely tumor biology, as blood-bom distant metastases

have sofar not been obtained, xenografts express alien histocompatibility complex

antigens anel aC-type retro-virus is capable of transfofming juxtaposed cells (Römer,

1992).

Iwamoto anel associates (1991) e1eveloped a transgenic mouse with Rel

oncogene expression under the control of a metallothionein promotor. Some of these

transgenie 1110use strains deveJop Denlar melanosis anel uveal melanomu. This is a

novel ma111l11alian lllOde! in which lnelanosis allel melanotic tumors develop step\vise,

triggered by a single gene; therefore, it has the aelvantage of providing the possibility

to stuely the effect of one oncogene on the e1evelopment of melanoma. Another

transgenie mouse moelel (TySv40) has been e1eveloped bearing the Sv40 oncogene

under the control of the mouse tyrosinase promoter. These mice develop intraocular

tumors that bear characteristics of retinal pigment epithelial eells and choroidal

melanoma cells. The great advantage of the last mentioned model for the stuely of

metastatic t1veall11elanoma is that haematogenously disseminated hepatic lnetastases

do occm in th is model (Anand, 1994).

In Chapter 7, as was indicated in the previous section, the expres sion of

melanoma-specific and associated antigens in llvealmelanoma cells is discussed as

weil as their MHC Class Iexpression. Tn sequel of this study, Chapter 8 discusses an

experiment with a murine model in whieh the susceptibility of uvealmelanoma cells

to NK's was determined (sce 1.3.3.2). Moreover, the author and associates succeeded

30

Page 31: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INTRODUCTION

in developing yet another animal model for the study of uveal melanoma, namely a

chicken embryo model. This experiment, that could provide new possibilities because

of the immature immune system that chicken embryos possess, will be presented in

Chapter 9.

31

Page 32: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1

REFERENCES

Adams KS, Abramson DH, ElIsworth, Haik BG, Bedford lvI, Packer S, Seddon J, Albert D, Polivogianis L.

Coball ph1lluC versus enuc\ealion tur uveal melanoma: eomparison of survival mIes. Br J Ophthalmol.

1988;72:494-497.

Adema OJ, de Boer AJ, van 'I Hullenaar, Denijn M, Ruiler DJ, Vogel A~1, Figdor CO. !vlc\anoeyte lineagespecific

antigens reeognized by Illonoc\onal :lntibodies NKI-beleh, fHvlB-50, alld I-11\1B-45 are encoded by a single

cDNA. Am J Pathol. 1993;143:1579-1585.

Adema GJ, de Boer AJ, Vogel Md, Loenen WA~-I, FigdorCO. !'vlo1ceulareharacterizatioll ofthe melanocyte

lineage-specific antigen gp 100. J Biol ehem. 1994;269:20126-20133.

Affeldt JC, I'I'fülCkter OS, Azen SP, Yeh L. Prognosis in u\'eal metanoma with extrasc1erat extension.

Areh Ophthatmol. 1980;9R: 1975-1979.

Alber! Dlvl, Shadduek JA, Liu H, Sundennan F\V, Wagoner 1,10, Dohlman HG, Papale JJ. Animal mudc\s tur the

study ufuveal melanum:l. Int Ophthalmol Clin. 1980;20:143-160.

Albert DrvI, RuZLo 11A, .McLaughlin MA, Robinson NL, Craf! JL, Epstein J. Eslablishment of celilines of uyeal

melanoma. Invest Ophthal Vis Sci. 1984;25: 1284-1299.

Albert D~vl, Niffenegger AS, Wi11son JKV. Treatment of llletastatÎC uyeal melanoma: review and recolllmendations.

Su!"\' Ophthalmol. 1992;36:429-438.

Ananu R, Ma 0, Alizadeh H, Cometford SA, Sambrook Hl, Gething M1H, l\kLean IW, Niederkorn JY. Characterization

of intraueular tumours arising in transgenic mice. Invest Ophthalmol Vis Sci. 1994;35:3533-3539.

Aubert C, Rouge F, ReiIlauuoli lvi, i'vietge P. Establishment anu characterizalÎon of human oeular melanuma celllines.

Inl J Cancer. 1993;54:784-792.

Aug.~hurger JJ, Ooel SD. Visual fUllc\ion following elluc\eatioll or episc\eral phlljUe radiotherapy for posterior uyeal

melanolllu. Areh Ophthalmol. 1994; 112:786-789.

Augsburger JJ. Yrabec TR. Impact of delayed trealment in growing posterior llvealllle!anomas. Areh Ophthalmol.

1993;111:1382-1386

Allgsburger JJ, Lauritzen K, Gamel JW, Lowry JC, Brad)' LW. i'vlatched group study of preenuc\eatiun mdiuthempy

versus cnuc\eation alonc for primar)' malignant melanoma of the churoid and ciliary body. Am J Clin Oneo!.

1990; 13:3B2-3B7.

Backer Ji'vf, ~\'Iet1dola CE, Ko\'esdi I, Fairhurst JI" O'Hara B, Eddy RIJ, Shows TB, Mathew S, Murt)' VV, Chaganti

RS. Chromosomallocalization and nucleoside diphosphate kinase activity of human metastasis-suppressor

genes nm23-1 and nm23-2. Oncogene. 1993;8:497-502

Bakker ABI-I, Sclueurs lvlWJ, de Boer AJ, Kawakami Y, Rusenberg SA, Adema Gl, Figdor CG. Melanocyte lineage­

specif1c antigen gplOO is recognizeu by melanoma uerived tumor infiltrating lymphocytes. J Exp t>Md.

1994; 179: 1005-1009.

Ball NJ, Yolm JJ, ~I-'forelli JG, Norris DA, Oolitz, lIoefiler JP. Ras rnutations in human rnelanoma: a marker

of malignant progression. JInyes! Dermatol. 1994; 102:2R5~290.

Bart!ensteÎn OS, Char DH, Kaleta-MichaelsS, Kroll SM. Ki-67 and bromodeoxyuridine labeling of human

churoiual melanom<l eells Curr Eye Hes. 1991 ;10:479-484.

Barncs R, lvlasoou S, Barker E, Rosengard AM, Coggin DL, Crowell T, Rkhte!"-King C, Porter-Jonlan K,

Wargotz ES, Liolla LA, Steeg PS. Low nm23 protein expression in infiltrating ductal breast

32

Page 33: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INTRODUCTION

carcinomas correlates with redllced patinet ,~llr\'i\'al. Am J Pal hol. 1991; 139:245-250.

Barth RJ, Bock S~oJ, l\'lule 11, Spiess PI, Rosenberg SA. Unique murinc tumor-associatcd antigens idclltil1ct!

by tumor infiltmting Iymphocytc~. J lmmunol. IlJlJO; 144: 1531-1537.

Bataille V, Pinncy E, Hungerford IL, Cuzick J, Bishop T, Newton JA Five cases of cocxistent primar)'

ocular and cutaneous melanoma. Areh Dermatol. 1993; 129: 1 98-201.

Bedikian AY, Kantarjian H, Young SE, Bode)' GP. Prognosis in metastalie choroidal mc!anoma.

South l\-Ied J. 1981;74:574-577.

Bergman W, Palan A, Went LN. Clinical aml genetic studies in Dutch kindreds with dysphlstic naevus

s)'ndrome. Aml Ihun Genet. 1986;50:249-258.

Hertheau P, De l.a Rosa, Steeg PS, i\'lerino l\u, Nm23 protcin in neoplastic and nonneoplastit: thyroid tissues.

Am J Pat hol. 1994;145:2(j-32.

Hevilacqua G, Sobell'l'fE, !,iotta LA, Steeg PS. Association of low I1m23 RNA levels in human primary

int1ltrating dl1ctall)f!~ast carcinomas with lymph node involvement aml otller histopathologie al

indicators of high mctnstastic potential. Cancer Res. 1989;49R:51 85-51 90

BIoom PA, Ferris JO, Laidla\\' OA, Goddard PR. l\Iagnetic resonance imaging. Diverse appearances of llveal

rnalignant melanomas. Arell Ophthalmol. 1992;110: 1105-1111.

Bornfeld N, Huser U, Sauerwein W, Wessing A, Snck H. Präoperative Bestrahlllng vor Enuklcation bei

malignem t ... Jclanom der Uvea. Klin Mbl Augenheilk. 1989;194:252-2fiO.

Brichard V, Van Pel A, Wülfe! '1', Wölfc1 C, De Plaen E, Lethé H, Coulie P, Boon Th. 111e tyrosinasc gene

codes for an antigen recognized by autologous cytolytic T Iymphocytes on J-ILA-A2 melanomas.

J Exp IVled. 1993;178:489-495.

Brown D, BonilIk 1\'1, Font RL. Diffuse malignant melanom<l of the iris with metastases. Clinical pathological

review. Sur\' Ophthalmol. 1990;34:357-364,

Van der Bruggen P, Traversari C, Chomez P, Lurquin C, De Plaen E, Van den E)'nde B, Knllth A, Boon Th.

A gene encoding an <lntigen recognized by eytol)'tic T [)'mphoc)'tes on a humaJ1 melanoma. Science.

1 YY 1 ;254: 1643-1647.

Butler P, Char DH, Zarbin M, Krol! S. Natural history of indetenninate pigmented ehoroidal tumors.

Ophtlwlmology.1994;101:710-716

Cairns P, Polascik TJ, Eh)' Y, Tokino K, Califano J, Merlo A, l\'iao L, Her<llh J, Jenkins R, Westra W, Rutter JL,

Buck1er A, Gahric!son E, loekman Cho KR, Hedrick Bova GS, lsaacs W, Koch \V, SL'hwah 0, Sidransky D.

Frequency of homozygolls deltions at pI6/CDKN2 in primaI)' hUlllan lumours. Nature genetics. 1995;! 1:210-212.

Canning CR, HUllgerford J. Pamilia1uveal Illelanllilla. [Review] Br J Ophthalmul. 1988;72:241-243.

Cannon-Albright LA, Goldgar DE, 1Ieyer U, 1ewis CM, Anderson, Fauw.in JW, Hegi ME, Wiseman RW, Pett)' Et-..J,

Ba[eAE, 010pade Or, Dim~ MO, Kwiatkowski DJ, Piepkorn MW, Zone JJ, Skolnick MH. Assignment of a

locus for familial melanorlHl MLM, to chromosome 9pI3-22. SCÎence. 1992;258: 1148-1152.

Char DH.l\ktastatic choroidal melanoma. Am J Ophthalmol. 1978;86:76-80.

Char DH, Castro JR, Qllivey JM, Cllen GTY, Lyman JT, Stone RD, trvine AR, Barrieks lVI, Crawford JR, Hilton OF,

Lonn LI, SchwartzA. Helium ion charged partitie Iherapy for choroit1al melalloma, Ophthallllology.

19RO:B7:565-570.

ehar DH, Phillips TL. The potential for adjuvant radiotherapy in choroidalmc1anoma. Arch Ophthalmol. 982; [00:247-248.

Char DH, Phillips TL, Andcjeski Y, Hrooks Crawford J, Kroll S, Failure of preenucleation radiatiotl to decrease llveal

33

Page 34: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1,

melanol11a l11ortality. Am J Ophthall11111. [988; 106:2 [-26.

The Collaborative Ocular r ... telanoma Study Group. Accumcy of diagnosis of chllroidal melanomas in the

collaborative ocular l11elanllrna stud)' COIviS report no. 1. Arch Ophthalrnol. 1990; 108: 1268-1273.

Colcman K, Baak JPA, Van Diest P, ivlullaney J, Farell tv1, Fenton lVI. Prognostic factors following cnucleation of IiI

uveal melanolllas. Br J Ophthalmol. 1993;77:688-692,

Cllulie PG, Somville rvI, Lehmenn P, Hainaut P, Bra~seur F. Devos R, Boon T. Precursor frquency analysis of hUlllan

cytolytic T IYll1phocytcs directed against autologous melalloma ceUs. Int J Cançet: 1992;50:2."\9-297.

De La Cruz PO, Specht CS, !\kLeall I\V. Lymphocytic infiltration in uveal melanoma. Cancer, 1990;65: 112-1 [5.

Daltlenfors R, Törnovist G, Weltrell K, Mark J. Cytogenetical obscf\'ation in ni ne ocular malignant Illelallomas.

AntÎcaneer Res. 1993; [3: 1415-1420.

Damato BE, Panl J, Foulds WS. Predictive factors of visual outcorne after local resection of choroidall11elanoma.

Br J Ophthalmol. 1993;77:616-623.

Damms T, Seltafer H, Guthoff R, Wink1cr P, Winter R. Correlation of histologieal tumor vascularization and doppier

sonogmphy in patlents with malignant melanoma of the choroid.Graefes Arch Clin Exp Opthalmol.

1995;233:257-260.

Davidorf FH, Letson AD, Weiss ET, I.eyine E. Incidence of misdiagnosed and unsuspectcd choroidalmelanomas.

Arch OphthalmoI1983;101:410-412.

Diencr-West 1'\'1, Hawkins BS, Markowitz JA, Schachat AP. Review of lllortality from choroidall11clanoma, A meta­

analysis of 5-ycar l110rtality rales fo!lowing enucleatÎon, 1966 through 1988. Arch Ophthalmol.

1992; 1 10:245-250.

Drattoff G, Jaffec E, Lazenby A, Golumbck P, Levitsky H, Brosc K, Jaekson V, Hamada H, Pardoll D, Mlllligan Re.

Vaccination with irradiatcd tumor cells engineered to secrete murine granlliocyte-macrophage colony

stimulating factor stimulates potent, specific, and long-lasting anti-tumor irmnunilY, Proc Nat! Acad Sci. USA,

1993;90:3539-3543.

Duric FIl, Campbell AI\·f, Lee WR, Damalo BE, Analysis of Iymphocytic infiltratiotl in llvcal melanoma.

Tnyest Ophthal Vis ScÎ. 1990;31 :2106-2110,

Edwards i'vIG, Schachat AP, Impact of enuc\eation for choroidal melanoma on the performance of vision-dependent

activitics. Arch OphtltalmoL 1991; 109:519-21.

Egan KM, Seddon JM, Glytltl R, Gragoudas ES,Albert DM. Epidemiologic aspects of uvealmelanoma.

Sur\' Ophthalmol. 1988;32:239-251.

Einhorn I.H, Burgess ÎvlA, Gottlieb JA. Metastatic patterns of choroidal melanoma. Canter. [974;34: 1001-1004.

Einhorn LH, Burgess MA, Vallejos C, Bodey OP, Gutlcrman J, Mavligit, Hers EM, Lucc JK, Prei E,

Freireich EJ, Gottlieb JA. Prognostic correlations and response to treatment in advanced metastatÎC

malignant Illelanom;\, Cancer Res, 1974;34: 1995-2005,

Engel i'v[, Theisinger B, Seib T, Seitz, Huwer H, Zang KD, Welter, Dooley S. High levels of nm23-HI aml

nm23-H2 tllessengcr RNA in human squamous-cell carcinoma arc associated with poor differentiation

;md advatlced tumor stages, Int J CanceL 1993;55:375-379.

Van der Eynde B, HainHut P, Herin A, Knuth A, Lemoine C, Weynants P, van der Bruggen P, Pauchet R, Boon T.

Presence on a hUlllatl melanoma of multiple antigctls recognizcd by auto10gous CTL.

Int J Cancer. 1989;44:634-640.

Perris J, Bioom P. Does the melanin content of uvc,1.l malignant melanornas corrclate with their magnetic resonance

34

Page 35: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INTRODUCTION

imaging appearance? Arch OphthalmoL 1995; 113:555-556.

Fidler IJ. Critical factors in the biology of human cancel' meta,~tasis: nth G.H.A. Clowes J'v1cmorial Awmd

lceturc. Cancer Res. 1990;50:6130-6138.

f-lishman JR, Gumerlock PH, :tvleyers FJ, Devere White RW. Quantitatioll of lltl123 exprcssion in human

prostatc tissues. J Urol 1994; 152:202-207.

Flnrenes VA, Aamdal S, .Myklebost 0, lvfaelandsmo mvt, Bruland 0S, Podstad 0. Levels of nm23 mcsscnger RNA in

metastalic rnalignant melanomas: inverse correlation of disease progression. Cancer Res. J 992;52:60158-609 I.

Polberg R, Rlll11mcll V, Parys-Van Ginderdeuren R, Hwang T, Wooison RP, Pe'er J, Gmman Ll\'l. Thc prognoslie value

of tumor blood vessel morphology in primary uveal melanoma. Ophthalmology. 1993; lOO: 1389-13915.

Fass Al, Guille IvlJ. üec\eslon NL, Hykin PG, Hungerford JL, Lightman S. The deteclioll of lllclanorna eells in

peripheral blood by reverse transcription-pulymerase ehain reaetion. Br J Cancel'. 1995;72: 155-J 59.

Fournier GA, Albert nrv1, Arrigg CA, Cohen Ai .... I, Lamping KA, Seddon Jivf. Resection of ,~olitary rnetasta.~is.

Approach 10 palliative trealment of hcpatic involvement with ehuroidalmelanoma. AfCh Ophlhalmol.

1984;102:80-82.

Gamel JW, Gleason J, Williams 11. Greenberg R. Reproducibility of nudeolar measurements in human

intraucular melallorna cells 011 standard histologie mÎCroslides. Anal Quant Cylol Histol. 1985;7: 174

Gamel JW, ~·It:Lean IW, ~'kCurdy JB. A eomparison of prognoslie covariates for tlvealmelanoma. Invest

Ophlhal Vis Sei, 1992;33: J 9 J 9- I 922.

Gamel JW, fvlcLean IW, MeCurdy JB. Biological distinctions between eure and time In dealh in 2892 patients

wilh inlraoeular melanoma. Cancer. 1993;71 :2299-2305.

Gaugler H, Van den Eynde B, van der Bmggen P, Romero P, Gaforio 11, De Plaen E, Lethe R, Rrasseur p.

Boon T. Hurnan geile 1'IAGE-3 codes for a antigen reeognized on a melanoma autologolIs cytolytic T

Iymphncytes. J Rxp ~\'lcd. 1994; 179:921-930.

Germain RN. i'vIHC-dependent antigen processing aml peptide presentation: providing ligands tor T IYlllphoeyte

aelivation. Cello 1994;76:287-299.

Gordon KR. Thompson CT, Char DH, O'Brien JM, Kroll S, Ghazvini S, Gray J\V. Comparative genomic hybridization

in thc delection of DNA eopy number abnormalities in llvealmelanoma. Cancer Res. 1994;54:4764-47615.

Gown AM, Vogel A 1>.'1. Hoak D. Gough F, McNutt Yo,·IA. Monoclonal anlibodies specifie for melanlH::ytic tumors

distinguish subpopulations of melanocytes. Am J Pathol. 1986; 123: 195-203.

Gragoudas ES, Goitein M. Vefhey L, Munzenrt'idwer JE, Suil HD, Koehier A. Proton be-am irradiation; an alternative

to enlldeation for intraoeular melanomils. Ophthalmulogy. 1980;87:571-58 I.

Gragoudas ES, Egan KM, Seddon JM, Glynn Rl, Walsh SM, Finn SM Munzenreider JE, Spar MD. Survival ofpatients

with metastases from uveaimelanollla. Ophthalrnology. 1991;98:383-390.

Greenberg PD. Chee\'er, Fefer A. H-2 restric\ion of adoptive imtllllnotherapy of advHnccd tumors.J Immunol.

J 9X J; J 26:2100-2103.

Grossniklaus HE, Oakman JH, Cohen C, Calhoun Fr Jr, DeRose PR, Drews-Botseh C. llistopalhology, morphoI11etry,

and nuclear DNA content ofiris melanocytic lesîons. In\'est Ophthahnol Vis Sei. J995;36:745-50.

Gmis NA, Sandkllijl, LA, Weber JL, Van Der Zee A, Borgstein Ai'vf, Bergillan W, Prants RR. I ,inkage analysis in Dutch

familial atypicalmultiple mole melanoma (FAMMM) syndrome familes. Effect on nae\'us eounl.

i'vfelanoma Res. 1993;3:271-277.

Guyer DR, Mllkai S, Egan KM, Seddun JM, Walsh SM, Gragoudas ES. Radiation maculopalhy after proton

35

Page 36: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1

beam irradiatiun for choruidal melanoma. Ophthalmology. 1992;99: 1278-1285.

Hall PA, Levisun DA, Woods AL et al. proliremting cell nuc1ear antigen (VCNA) im!I1l1nulocalization in

paraffin settiuns: an index of cell prolifemtion with the evidence of deregulated expression in some

neoplasms. J Palhul. 1990; 162:285-294.

Harning R. Szalay J. Ocular melastasis of in vivo and in vitro derived syngeneic murine melanuma. lnvest

Ophthalmul Vis Sd. 1987;28: 1599-1604.

Harning R. Kou GC, Szalay J. ReglJlation of the metastasis of murine ocular melanoma by natural killer cells.

lnvest Ophthalmol Vis Sci. 1989;30: 1909- [9 [5.

Harris cc. p53: at thc crossroads of mo1ccular carcinogenesis and risk assessment. Science.1993;262: 1980 1981.

Hallt tv1, Steeg PS, Willson JKV, !vlarkowitl SD. Induction of nm23 gene expressiun in hu man colon ie neoplasms anu

cqual expression in colon tumors of high and low metastatic potential. JNCI. 1991;83:712-716.

Hcnnessy C, Henry J, ivlay P, Wcstly Angus R, Lenard T. Expression ofthe aotimetastatic gcne nm23 in

human Iln~ast cancer: An association with good prognosis. JNCI. 1991;83:281-285.

Herin ivf, Lcmoine C, Weynants P, Vessiere, Van Pel, Knllth A, Devos R, Boon T. Prodllction of stabIe eytotoxic T-cell

cl olies directed against autologous hu man melanoma. Int J Cancer. 1987;39:390-396.

HilI LL, Korngold R, laworsky C, Mllrphy G, i\JcClIe P, Berd D. Growh amlmetastasis of fresh human

melanoma tissue in mke with severe combined immllnodefidency. Canccr Res. 1991 ;51 :4937-4941.

Hollstein r-"t, Sidransky D, Vogelstein, Harris cc. P53 mutations in hllman cancers. Science. 1991;253:49-53

Horsman DE, While VA. Cytogenetic analysis of llvealmelanoma; constituent OCCllrrence of monosomy 3

anu trisomy 8q, Cancer. 1993;71:811-819.

Horsthemke B, PresdIer G, Bornfeld N, Becher R. 1.05S of ehromosoille 3 alleles and multiplication of

chromosome 8 alleles in uveal mclanoma. Genes Chrolll Cancer. 1992;4:217-221.

Huang X-Q, .i\·1itchell IvIS, Liggen PE, !vturphee AL, Kan-Mitchell 1. Non-fastidiolls, melanoma-spccit1c CD8+

cytotoxic Iymphoeytes from choroidalmelanoma patients. Cancer Immllnul Imll1ullothcr. [994;38:399-405.

Human Gene Mapping. Cytogcnet Cel! Genet. 1991;58:1-2200.

Hungcrford 1. Uvealmelanoma. Eur J Cancer. 1993;29A:1368-1372.

Iwamoto T, 'I~'lkahashi M, Ho M, Hamatani K, Ohbayashi .M, Wajjwalku W, Isobe KI, Nakashima 1 Aberrant

melanogenesis aod melanocytic tumour de\'clopmcnt in transgenie miee that carry a

methallothiooeinlret fusion gene. EMllO J. 1991; 1 0:3167-3175.

Ja)' rvI, i\fcCartlley ACE. Familialmalignat Illclanoma of the uvea and 1'53: a Victori,in detective slory. Slln'

Ophthalmol. 1993;37:457-462.

Jensen OA. lvlalignant mclanoma of the uvea in Denmark 1943-1952: a clinical, histopatholigical and prognostic study.

Acta OphthalJ11ol. 1963;75S:1-220.

lensen OA, Andersen S. Spontaneous regression of a malignant melanuma or the choroid. Acta Ophthalmol.

1974:52: 173.

lensen OA Malignant melanomas of the uvea. 25-Year follow-up of cases in Denmark, 1943-1952. Act'l Ophthallllol

1982;60: 161-182.

Jensen OA. "j\·raligoant melanoma of the iris. A 25-year analysis of Danish cases. Eur J Ophthalmol. 1993;3: 19 1-188.

De Jong PTV1J. Hoe te handelen bij een melanoom van de ehoroidea of van het corpus ciliare? Net.! tijdschr Ueneesk.

1987:131:439-443.

Kallioniemi A, Kallioniemi P-O, Sudar D, Rutovitl 0, Gray JW, Waldman F, Pinkel D. ComparativI.' genomie

36

Page 37: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INTROpUCTlON

hybridization fur molecular cytugenetic unalysis uI' solid tumors. Science. 1992;258:818-821.

Kamb A, Gmis NA, Weaver-Feldhaus J, Liu Q, Harshman K, Tavtigian SV, Stoekert E, DayIII RS, Johnson BE,

Skolniek i\llI. A eell cyele regulator potentially involved in gencsis of man)' tumor types. Science.

1994;91 :7563-7567.

Kan-ivlitchell J, IvlitchelllvfS, Rao N, Liggett PE. Characterization ofuycal melanoma cclllines th at grow as

xenografts in rabbit eyes. Jnvcst Ophth Vis Sci. 1989;30:829-834.

Kan-ivlitchell J, Liggctt PE, Hare! W, Stcinman L, Nilla T, Oksenberg JR, Posner i\m, l\-litchelll.."IS.

Lymphocytes cytotoxic to uycal and skin mclanuma ce lis from pcripheral blood of ocular melanoma

patients. Cancer Immunollmmunutber. 1991 ;33:333-340.

Kath R, Hayungs J, Bornfeld N, Saucf\\'cin W, HöftKcn K, Seeber S. Prognusis and treatment uf disseminated llveal

mclanoma. Caneer. 1993;72:2219-2223.

Kawakami Y, Eliyalm S, Delgado CH, Robbins PF, Sakaguchi K, Appella E, Yannelli JR, Adema GJ, Miki

T, Rosenberg SA. Identification of ,\ hu man Illelanuma antigen recognized by tumor-infiltrating

Iymphucytes associated with in vivo tumor rejection. Proc Natl Cancer Inst. 1994a;91 :6458-6462.

Kawakami Y, E1iyahu S, Delgado CH, Robhins PF, Rivoltini L, Topalian SL, Rosenberg SA. Cloning of the gene

coding tur a slmfed human melanorna antigen recognized by autologolIs T eells infiltmting tumor. Prae Nutl

Acad Sci USA. 1994b;91:3515-3619.

Kawakami Y, Eliyahu S, Sakllguchi K, Robbins PF, Rivoltini L, Ymmelli JR, Appclla E, Rosenherg SA.

Identification of the immnuduminant peptides of the iviART-l hu man melanoma antigen recognized

by the rnajority of HLA-A2-restricted tumor infiltrating Iymphocytes J Exp ivled 1994c; I 80:347-352.

Kerneny N, Schneider A. Regional treatment of hepatic metastases anti heptocellul,tr cnrcinuma. Curr Probl Cllllccr.

1989; 13: 197-284.

Kersten RC, Tse D'I~ Anderson RL, Blodi Fe. The role of orbital exentemtion in choroidalmelanoma with

extmsderal extentiun. Ophthalmology. 1985;92:436-443.

Kindy-Degnan N, Char 011. KroB S11 Coïncident systemic malignant disease in llveal metanom<l patients.

Can J Ophthalmol. 1989;24:204-206.

Koden\ Y. Isobe KI, Yamauchi, Kondoh K, Kimura N,Akiyalllu S, 1Iuh K, Nakashirna 1, Takagi H. Expresion uf mu23

H-I RNA levels in hurnan gastric e,\Ilcer tissues. A negativc corre1ation with nodal metastasis. Cancer.

1994;73:259-265.

Kovacsovics-Bankowski lvi, Rock KL. A ph,lgosome-Io-cytosol pathway ror cxogenous antigens presented on

tvll-IC dass I molecules. Sciencc 1995;267:243-245.

Ksander BR, Rubsamen PE, Olsen KR, Cousins SW, Stcilcin J\V. Studies of IUll1ur-infiltratioJl lymphoeytes

frorn a llUman choroidalmelanoma. Invest Ophthalmol Vis Scî. 1991 ;32:3198-3208.

Knuth A, Danowski B, Oettgen HF, Old L. T-cell mediated cytotoxicity against autologous malignant melanuma:

analysis with inetrleukin-2-dependent T-cell cultures. Prac Natl Cancer ins!. 1984;81:3511-3515.

Kreissig I, Rohrbach !\'l, Lincoff H. Irradiation of choroidal melallomas before enucleation? Retina. 1989;9: lOl 104.

LeoneA, Seeger RC, Hong Cl\'f, Hu YY,Arboleda J, Brodeur Gi'.'l, Stram D, Slamon DJ, Steeg PS. Evidellce t'or nm23

RNA overexpression, DNA amplificatioll and mutatioll in aggressive childhood neurablastoma.

Oncogene. 1993;8:855-865.

Uotta LA, Steeg PS, Stetler-Ste\'enson WGo Cancer mctastasis and angiogenesis: an imhalanee of positi\'e and

negative regulation. Cel!. 1991 ;64:327-336.

37

Page 38: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1

Liu B, Parsons R, Papadopoulos N, Nicolaides NC, Lynch HT, Watson P, Jass JR, Dunlop rvi, Wyllie A, Peltomüki P,

De La Chapclle A, Hamlitun SR, Vogebtein D, KinzIer KW. Analysis of mismutch repair genes in hereditary

nun-polyposis coloreetal cancer patients. Nature medici ne 1996;2: 169-174.

Lommatzsdl PI<. Experienees in the treatment of maligant melanoma of thc choroid with IOóRu/l06Rh beta-ray

applkalors. Tmn.~ Ophthalmol SOl' UK. 1973;93: 119-132.

Luyten GP.1.,.I, r ... looy Ci\'I, De Jong PTVr ... 1, Hoogeveen AT, Luider Thi'vL A chicken embryo model to study growth of

lIvt"al mc\aTlornu. Biochem Biophys Res CUI11m. 1993; 192:22-29.

Luyten GNvI, r ... looy Ci\·l, Eijkenbuum Wr ... 1H, Stijl1en T, t-Iel1emons LP, Luider TMvl, de Jong PTVivL No demonstralcd

effed of prc-enucleatiun irradiation of l1veal mclanoma patienls Am J Ophthalmol. 1995;119:786-791.

Ma D, Luytell GPi\·I, Luider Ti\J, Nicderkorn JY. Rclationschip between natural killer susceptibility ant! metastasis of

hu man u\'callllelnaoma eells in a lUurine model. Invest Opthal Vis Sei.1995;36, 435-441.

i\Ja D, Luyten OPM, Luider Ti\.J, Jager MJ, Niederkorn JY. Assudatiull between nm23-H 1 gene exprcssion alld

metastasis of human lI\'eat melanoma in an animalmodel. (suhmitted)

Manschot WA, van Peperzee\ HA. Choroidal melnnoma; enucleation or observation'! A new approach. Arch

Ophthalmol. 1980;98:71-77.

Manschot WA, van Strik R. U\'cul melanoma:therapeutic eonscquences of doubling times ;lIld irradiation resu!ts; ti

review. Int Ophthallllo1. 1992;16:91-99.

r..,.lagauran RG, Gray D, Slllul1 KW. Chromosome 9 abnormality in choroidal melanoma. Am J Ophthalmol.

1994;117:109-11 J.

r..,.fassarelli R, nc\klHlu R, Dunel-Erb S, Chelavier C, Abbé J-C, Pignol J-P, t ... toutaouakkil M, Sahel J.

Caractérisation cellulaires de mélanume uvéal humain. C R Aead Sei Paris. 1994;317:25~33.

1.Javligit Gi\.J, Churnstll1gavej C, Carrasco CH, Wallnce S, I'apadopoulos NEJ, Palt YZ. Regression of ocular

Illclanomn metastatic to thc Jiver a!'ter hepatic mierial eheilloelllbolization with cisplatin and polyvinyl

sponge. JAMA. 1988;260:974-976.

McCurdy JB, Gamel JW, I .... lcLean IW. A simple.efficient alltl reproducible mcthod tOr estimuting the maligTHlHt

potentialof ll\'eal 1l1elanoma from routine H&E slides. Pathol Res Prae!. 1991 ;187: 1025-1027.

I\kLean I\V, Foster WO, Zimmeffil,lIl I.E. Uveal mdanoma: location, size, cell type, and enudeation as risk

factors in metastasis, Hum Patllo!. 1982;13:123-132.

{ ... [cLean IW, Foster WO, Zimmerman LB, Gamel J\V. r ... fodifications of Callender's clnssification of uveal

melanomu at the Armed Forces Institutc of Patholug)'. Am J Ophthalmol. 19R3;96:502-509.

i\kLcan IW, Berd 0, Mastrangelo r ... u, Shields JA, Davidorf FH, Grever M, Mnkley TA, Gamcl J\V. A randolllized

stud)' of methanol-extraction residue of baeille Calmcltc-Guerin as postsurgkal adjll\'unt therapy of lIveal

melanoma. Am J Ophthalmol. 1990;110;522-526.

Mclean IW. The biology of huematogennus metastasÎs in hlllllun uvealmalignant me1anomu. Vir Areh Path Ana!.

1993;422:433-437.

Meecllam WJ, Char DH, Kuleta-Michnels S. Inl1ltrating Iymphoeytes and antigen expression in ll\'cllimelanuma.

Ophthalmie Res. 1992;24:20-26.

Mitehell MS, Liggett PE, Green RL, Kan-i\litchdl J, MurphreeAL, Dean G, Spears L, Walonker F. Sustained

regression of a primary choroidal mdanomu under thc inlluence of a therapeutie mclanomu vaccine. J Clin

Olleo!. 1994;12:396-401.

Muo)' CM, De Jong PTVlVI, Van der Kwast TH, Mulder PGH, Jager MJ, Ruiter DJ. Ki-67 Immunostaining

38

Page 39: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INT R 0 IJ UJ-: T~I '"O"N'---_____ _

in llveal melanoma: the effect of pre-enucJeation ratliotherapy. Ophthallllology. 1990;97: 1275-1280.

1vloo)' Civ!, Van tier Helm ivU, Van tier Kwast TH, De Jong P.T.V.i\-!., Ruiter DJ, ZwartholJEC. No N-ras mlltations in

hum,ln tlveal melnaoma: the role of ultraviolet light revisÎted. Br J Cancer. 1991 ;64:411-413.

I\'looy Civ!, Luyten GPl'vl, de Jong PTV1J, LuiderTl\·l, Stijnen T, van de Hnm F, van Vroonhoven CCJ, Bos!I1nn FT.

Immunohistochemical anti prognostic annlysis of apoplosis and proliferation in uveal melanoma, Am J Patho!.

1995;147: 1097-11 04.

Mooy Civl, Luylen GN·,'f, de Jong PTVM, lensen OA, Luider TfI'I, van de Ham F, Bos!l\lIll FT. Neural ceU adhesion

molecule tlistribulion in primary and melastlltie uycal me1anom<l. Humnn pathology. 1995;26: 1185-1190,

Moo)' C~vl, Vissers K, Luyten GP!vf, rvlulder A, Stijnen '1', De Jong PTVI\·I, Bosman FC DNA tlow-cylometry inllveal

melunoma: Ihe effect of pre-cnucelation irradiation. Br J üphlhalmol. 1995:79: [47-177.

rvlooy CM. Tumor progression in uveal me1anoma. Thesis 1995, ISBN 90-9007931-9.

Mukherji B, MacAlhter TJ. Clonal m\Ulysis of cytotoxie T eell response against hu man melanoma. J Exp Mcd.

1983; 158:240-245.

I\'[ulé IJ, Yung Je, l.aCreniere R, Slm S, Roscnberg SA, Identificatioll of ccllular mechanisms operational in vivo during

the regressioJl of established pu[ponary metastases by the systelllic adminislralion of high-dose recombinant

inetrleukin 2. J Immunol. 1987;139:285-294.

i'vlulé JJ, Jkha DL, Aebersold prvl, Travis WD, Roscnberg SA. Dissemim\led hlltllan malignant melanoma in

congenitally immune-delïdent (bg/nu/xid) rnice. J Natl Cancer Inst. 1991 ;83:350-355.

Nakayama T, ühtstlru A, Nakao K, Shima M, Nakata K, Watallahe K, Ischii N, Kimura N, Nagataki S. Expression in

numan hepatocellular carcinoma of nucleoside diphosplwle kinase, a homologue of the nm23 gene product.

JNCI. 1992;84: 1349-1)54.

Nkolson GL. Organ specificity of tumor metastasis: role of preferntial adhesion, invasion and growth of

malignant eells al speeific seeondary sites. Cancer Met Re\'. 1988;7: 143-1 SS.

Niederkom JY, Streilein JW, Shaduuek JA. Deviant immune responses to allogeneic tumor injected intracamerally and

suhcutaneously in mice. Invest Ophthalmol Vis Sci. 1981 ;20:355-363.

Niederkorn JY. Suppressed cellulur immunity in mice harboring intraocular melanomas. Inves! Ophlhalmol Vis Sci.

19R4;25:447-454.

Niederkorn JY. ETlucieation-induced metastasis of introcular melanoms in miec. Ophthalmology.1984;92:692-700.

Nicderkorn JY. EnucleatÎon in eonsort with immunologie impairment promotes Jllclastasis of intraocular melanomas.

Invest Ophthalmol Vis Sei. 1984;251080- J 086.

Niederkom JY, l ... lelloll J, Pidherney M, Mayhew E, ATlant! R. Effect of anti-ganglioside anlibucHes on the metastatic

spreau of intraocular mclanomas in a nude JTlouse model of human m'enl melanolllH. ClIrr Eye Res.

1993; 12:347-35R.

Paeh JM, Roberlson DM. i'vletastasis form lInlreateu uyenl melanomus. Arcl1 Ophthalmol. 1986; 10.:.\: 1624-1625.

Packer S, Rotman M. Radiotherapy of choroidal melanoma with iodine-125. Ophlhalmology.1980:87:582-59ü.

ParmÎler AH, Nowell Pc. Cytogenetics of melnaoeytic tumors. JInvest Denntol. 1993: 100;254S-258S.

Peltenburg LTC, Schrier PI. Transcriptional supprcssion of HLA-B expression by c-m)'c is mediated through the core

promoter elements, Immunogenetics. 1994:40:54-61

Pe'er J, Rummelt V, ~ihlWn L, Hwang T, Wooisoll RF, Folberg R. 1Jean of the ten largcst nucleoli, microcirculation

architecture, and prognosis of eiliochoroidalmelanomas. Ophthalmology. 1994; lOL: 1227-1235.

Pe'er J, Gnessin H, Shargal Y, Livni N, PC-IO iml\lunostaining of proliferating cell nuc1car antigen in posterior uveal

39

Page 40: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 1

rnelanoma. Enucleation versus enueleatioll postirradiation groups. Ophthulmology. 191)4; I Ol :5fi~62.

Peris K, G~rroni L, Chimenti S, Soyer HP, Kerl H, Höfler. Proto-oncogene expression in uermal naevi and

mdmlOmas. Areh Oenll,\tol Res. 1991;2~3:.'iOO-505,

Vnn Der Pol JP, Jager t\'11, de Wolff-RouenthHlI D, Ringens PJ, Vennegoor C, Ruiter DJ. HelerogeneOlls expression of

melanoma-associated antigens in llvcal melanomas. Curr Eye Res. 1987;6:757-765.

Postel Eli, Berberich SJ, Flint J, Ferrone CA. HUlllal1 C-mye transcription factor PuF iuentit1ed as lIm23-H2

llucleoside diphosphate kinase, a c<lndidate suppressor of tumollf IlIetastasis. Science. 1993;261 :478-480.

Postel RIl, Fc[[one CA. Nucleoside diphmphatc kinase enzyme activity of Ilm23-h2/Puf is not required for its

DNA binding und in vitro transcription functiolls. J Biol Chem. 1994;269:8627-8630.

De Potter P, Shiclds CL, Shielus JA, Caler JR, Tm'dio DJ. Impact of enlleleatio!l verSLlS phlllllc radiotherapy in the

management of juxtapapillary dlOroidal melanoma on patient survival. Br J Ophthalmol. 1994;78: 109-114.

De Potter P, Flanders AE, Shields JA, Shiclds CL, Gonzales CF, Rao Vt\'1. The role of fat-suppression techniqlle ;md

gadopentetate dimeglumine in magnetic resonance imaging cvaluation of intraocular tumors and simulating

lesions. Arçh Ophthalmol. 1994; 112:340-348.

Prescher G, Bornfeld N, Hosrlthclllke B, Beçher R. Chruillosomal aberrentions detlning uveal melaoma of poor

prognosis. Lancet. 1992;339:691-692.

Pyrhönen, Hahka-Kcmppincn M, Muhonen T, A proillising ingcrfcron plus four-drug chernotherapy rcgimen

ror melastalic mclanoma. J Clin Ol1co1. 1Y92;1 0: 1919-1926.

Quivcy Jivl, Char DH, Phillips TL, Weaver KA, Castro JR, Krol! SM. High intensity 125-iodine (1251) plaque

Ircatment of uycalmelanoma, Int J Radiat Onçol Biol Phys. 1993;26:613-6 [8.

Radinsky R, Wcisberg HZ, Staoseisky AN, Fidlcr IJ. Expression levcl of thc nm23 gene in donal populations of

mctastntic murine and hunwn neopl;lSIns. Cancer Res. 19Y2;52:5808-5814.

Rajpal S, ivloore R, Karakousis CP. Survival in metastalic ocular me!anol\llI. Clincer. 1983;52:52:334-33fi,

Römer TJ, van Delft J, de Wolf-Rouendaat 0, Jager ~u. Hamster greene ll1elanoma implanted in the anterim

chamber of a rabbit cye: areliabIe tIJmor model? Ophthalmic Res. 1992;24: 119-124.

Rosenberg SA, Spiess P, Lafellicre R. A new approach to the adoptive immunotherapy of Cllncer with tumor­

int1ltrating Iymphocytcs. Science. (Washington DC) 1986;233: 1318-1321.

Roscnberg SA, Packard BS, Aebersold PM, Sololllon D, Topalian SL, Toy ST, Simon P, Lolze !vlT, Yang JC Seipp CA.

Use of tumor infiltrating Iymphocytcs and interleukin-2 in the iml11unotherapy of patients \\'ilh melastatic

melanoma. Preliminary report. Ncw Engl J Med. 1988;319: 1676-1680.

Rosenberg SA. The immunotherapy ;md gene therapy of cailcer J Clin Olleol 1992;10: 180

Rosengard Ai'vI, Kmtzsch HC, Sheam KA, Biggs Jr, Barker E, ivlnrgulies I/'vl, King CR, Uotta LA, Steeg PS. RedllCed

nm23/Awd protein in tumor metastasis and aberrant Drosophila developmcnt. Nature. 1989; 342: 177-180

Royds JA, Sharraru RM, Parsons J\"it\, Lawry J, Rees R, Cottam, Wagner B, Rennie IG. C-myc oncogene

expression in ocular Illelanomas. Gmef An:h Clin Exp Ophthalmol. I Y92;230:3fi6-371.

Royds JA, StepheIlson TJ, Rees R, Shorthouse A, Silcocks PB. Nm23 product expression in in situ and

invasivc human breast cancel'. JNCI. 1993;85:727-731.

Royds JA, Cross SS, Silcocks PB, Scholeticld JIJ, Rces RC, Stephenson TJ. Nm23'anti-melastatic' gene product

expression in colorectal carcinoma. J Patho1. 1994;172:261-266.

Rummclt V, Folberg R, Woolson RF, Hwang T, Pe'er. Relation betwecn the microcÎrculation mehiteetme

and thc aggressive behavior of ciliary body melanomas. Ophthalillology. 1995;102:844-85 I.

40

Page 41: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______ ~I~N~' T~R~O J) lJ C T ION

Rumrnelt V, Folberg R, Rtll11mcll C, ürulllnn Ltd, Hwang T, Woolsoll RF, Yi H, Naumann GO.

tvlicrocirculation an.:hiteelllrc of melanocytic nevÎ amlmalignant melanomas of the dUury body anti dlOroid.

A comparative histopathologie and ultrastrllclural stud}'. Ophthalmology. 1994;101 :71 H-727.

Sanborn GE, Niederkorn lY, Kall-ivlitchcll J, Albeft Dt>.t Pfevcntioll of metastasis of intraocular mclnnoma

in miee treated with dit1uoromcthylornithine. Greaf An:h Clin Exp Ophthalmol. 1992;230:72-77.

Sanoorn GE, Niederkorn JY, Gamel JW. Efficacy of dacarbazinc (DnC) in preventing metastases arising

from intraocular melanomas in mice. Gn~·af Areh CHn Exp Ophthalmol. 1992;230: 192-1 %.

Sastrc-Uarau X, Lacombe i'vl, Veron I\'I, ivlagdelenat H. Nucleoside diphosphate kinase/Ni\I23 exprcssion in

hUlllan breast cancer: lack of correlation with Iymph-nodc rnetastasis. Int J Cancer. 1 ~92;5():533-538.

Seigcl DG. Cox regression analy:.is in clinical research. Areh Ophthalmol 1990;108:888-~0~.

Sellallli td, WeU r-,·t, Dhermy P, Aron-Rosa D, Jacquillat e. Adjuvant chemothewpy in oeular malignant

melanoma. Oncolugy. 1986;43:221-223.

Scregard S, Trampe E, i\li'tnsso!l-Brahme E, Kock E, Bergenmar tv1, Ringborg U. Preva\encc of primary <lc{jllired

melanosis <Iud nevi of thc conjunctiva and llvea in dysplastic nevus symlrome. Ophthnlmology.

1995:102: 1524-1529.

Shammas HF, Blmli Fe. Orbitnl extension of choruidal al\(l cilinry body melanomas. Arch Ophthalmol.

1977;95:2002-2005.

ShÎelds CL, Shilcds JA, Kiratli H, De Polter, Cater JR. Risk factors fur growth and metastnsis ofsmall choroidal

melanocytic 1esions. Ophthalmology. 19~5; J 02: 135 [-1361.

Shields JA, Shic\ds CL, Ehya H, Eagle RC, De Potter P. Fine-needle a~piration hiopsy of sllspected intraocular tumors.

The 1992 lJrwick lecture. Ophthalmology. [993; 100: 1677- J 684.

Shields JA, Shields CL. Cmrent management of posterior uveal melanollw. I\,[ayo Cl in Proc. 19~3;6X: 1[96-1200.

Shiclds JA, Shields CL, Donuso LA. Management of posterior llvcal melanoma. Sur\' Ophthalmo!.1991 ;36: 161 195.

Shields CL, Shields JA, Karlssoll U, Menduke H, Brady LW. Enucleatio!l after plaquc mdiotherapy for

poslerior tlvealmclanoma. Histopathologie findings. Ophthalmology. 199U;97: 1665- [fi70.

Singh AD, Boghosian-Sel1 L, War)' KK, Shields CL, Dc Potter P, Donoso LA, Shic\ds JA, Cannizzaro L.

Cytugenetic findings in primary llvcalmelanoma. Cancer Genet Cytogenet. [994;72: 109-115.

Sisley K, Rellilic 10, Cottam DW, Polier AM, Polier CW, Rees Re. Cylogenetic findings in six pmlerior uveal

Illclanomas: Involvement uI' chromosome 3, 6 and 8. Genes Chrom Cancer. 1990;2:205-209.

De Smet C, Lurquin C, Van der Bruggen P, de Plaen E, Brasseur F, Hoon 1'. Seql1ence aml expression

pattern of the hl1man ivlAGE2 gene. ImmunugenetÎCs. 1994;39: 121-129.

Soparker CN, O'Brien Ji\I, Albert Divl. Investigatiun of the rolc of the ras proluuncogene pointmutation in hUl1lan

uveal melanuma. !Jlvest Ophthalmol Vis Sci. 19~3;34:2203-2209.

Soulieres D, ROllsseau A, Deschienes J, Tremblay .M, Tardif ivf, l'elletier G. Characterization of gangliosides

in hUll1an Ilvealmelanoma ceiis. Int J Cancel'. 199 1;49:498-503.

Speicher l\lR, Prcscher G, Du rvI,lOuir S, JaHL'h A llorslhelllke B, Burnfeld N, Becher R, Cremer 1'. Chrommomill gains

;1Ilt! losses in uveal me1anomas detected by eomparative genumic hybridization. Cancer Res. 1994;54:3817-3823.

Spie ss PJ, Yang JC, Rosenberg SA. In \'ivo antituTllor activity uf tumor infiltrating lymphocytes expanded in

recombinant interleukÎn-2. J Natl Cancer Inst. 1987;7~: 1067-1075.

Stabl JA, Leone A, Rmengard AM, Porter L, King R, Steeg PS. Identirication ofa second Illlman IIm23

gene, nm23-H2. Cancer research. 199 I; 51: 445-449.

41

Page 42: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_______ -'C'-"H'-'A I'T-'cE'-'R'---"I _____ _

Slallard H8. Mulignant l11elanomu of het dlOroid Irc<lled with radioactivc app!ic<ltors. Trans Ophthalmol SOl'

UK. I !}Y);7~:37J-392.

Steeg PS, BeviJncqu<l G, Kapper L, Thorgcir~soll UP, Talmadge JE, Uolta LA, Sobc! ! ... lE. Evidente for a

noyel gene associated with Jow tumor Illcla~tatic potentiaL J. Natl. Call1'Cr lnst. 19H9; 80: 200-204

Steeg PS. Geuetic control of tbc metuslntic phcnotype. Seminars in Cancer Biology. 1991 ;2: 105-110

Steuhl KP, Rohrbach Hd, Knorf i'd. ThiclllJ. Signitkance, specifidty, and ultraslruclurat lucalizatÎon of

HtvlB-45 antigen in pigmented ocul<lf tumors. Ophthalmology. 19!}3; 100:203-215.

Stevens EJ, Jacknin L, Robbins PF, Kuwakmni Y, El Garni! :tv1, Rosenberg SA, YanncIli JR. Generation of

tumor-speciJ1e CTLs [rom melmlOlll paticnts by Hsing peripheral blood stirnulatcd with allogeneic

melanolll tumorcelilines. J Immullo1. 1995;154:762-771.

Straatsma BR, Fine SL, Earle JO, Hawkiins BS, Dicllcr-We~t .M, 1kLaughlin ÎvIA, Thc collaborative ocular

melanoma study research group. Enuc1eation verslIs plaque irradiation tor choroidalmclanoma.

Ophthalmology. 19BB;95: 1000-1004.

Tobal K, Warren W, Cooper CS, McCartney ACE, Hungerford J, Lightman S. Incrcased exprcssion anti

l11utatioJl of p53 in ehoruidal melanom<l. Br J Caneer. 1992;66:900-904.

Tobal K, Sherman LS, Foss AJ, Lightman SL. Detection of rnelanoeytes from uveal melnnoma in peripheral

blood using the polyrnerase chain reaction. Invest Ophthalrnol Vis Sci. 1993;34:2622-2625.

Tobal K, Deuble K, McCartney ACE, Lightman S. Characterization of cclllJiar inl1ltration in choroidnl melnnomn.

fvlclanoll1a Res. 1993:3:63-65.

Taylor i\IR, Gucrry D, Bondi EE, Shields JA, Augsburger 11, Lusk El, Eldcr DE, Clark WH, Van Hom i\1. Lack of

association bctween intraoculnr rnclnnoma ;md cutaneotls dysplastic nc"i. Am J Ophthalll1ol. 1984;98:478-482.

Topalian SL, Solomon D, Rosenberg. Tumor specific eytolysis by lymphocytes infiltrnting hunJan melanoll1as. J

Immunol. 1989; 142:3714-3725.

TravcrsarÎ C, Van der BmggcJl P, Luescher IE Lurquin C, Cholllcz P, Van Pel A, De Plaen E, Amar-Costesec A, Boon

T. A nonapeptide c!leotlcd by human gene i\-fAGE-1 IS rceognized on HLA-AI by cytotoxic T lyrnphocytes

directcd against tumor antigen lvIZ2-E. J Exp 1Jed. 1992: 17ó: 1453-1457.

Vennegoor C, H:lgeman P, van Noulmijs 1I, Ruiter DJ, Calafat J, Ringens, pj, RUmkc PH. A monoclonal

antihody spcdt1c for cells of thc mehlflocytic lioeage. Am J pathol. 1988; 130: 179-192.

Versteeg R, Noordermeer IA, Kruse-Wolters ivl, Ruiter DJ, Schrier PI. C-myc down-regttlates c!a~s I HLA expression

in human melnomHs. EivlBO J. 19RR;7: 1023- [()29.

Vcrsteeg R, Peltenburg LTC, Plomp AC, Schrier PI. High expression of the c-myc oncogene renden; melanoma eells

prune to lysis hy Ilatural killer cells. J Immunol. 19R9; 143:4331-4337.

Vink J, Crijns Ivm, Mooy Ci\l Bergman W, Oosterhuis JA, Went LN. Oeular melanoma in families with dysplastic

nevus syndrome. J Am "cad Demlatol. 1990;23:858-862.

Vogel At>.t Esclamado RM. Identitlcation of a secreted Mr 95.000 glycoprotcin in hum;m melanocytes ,md

melunolllus by aspecific mOlloclonal untobody. Cancer Res. 1988;48: 1286-1294.

De Waard-Sieblnga I, lllom DJR, Griffioen tv1, Schrier PI, Hoogendoorn E, Be\'crstock G, Dallen EHJ, Jager MJ.

stublishment anti characterization of an uveal-Tllclanoma celiline. Int J Cancer. 1995;62: 155-161.

Warren HS, Kinnear BF, Skipsey LJ. Human natural killer (NK) eells: requirements for cell prolifcratioIl HOlt

expansion of phcnotypically novel subpopuJations. Irnmunol Cell Biol. 1993;71 :87-97.

Whckhel JC, Farah SE, MeLean IW, Burnier MN. Immunohistochcmbtry of infiltrating Iymphocytes in llvcal

42

Page 43: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

INTRODUCTION

melunoma. In\'est Ophthal Vis Sci. 1993;34:2603-2606.

Wiltshire RN, Elner Vi\·I, Dennis T, Vine AK, Trent H\'L Cytogenetic analysis of posterior u\'ealmelanomu.

Cancer Genet Cytogenet. 1993;66:47-53.

Wölfel T, van Pel A, Brichurd V, Schneider J, Seliger B, -r..oJeyer Zum Blisehenfe1t1e KH, Boon T. Two tyrosinase

nonapeptides recognized on HLA-A2 melanomas by alltologous eytolytie T lymphoeytes. Eur J Illllllunol.

1994;24:759-764.

Yumaguchi A, Urano T, Fushida S, Fllmkawa K, Nishimura G, Yonemllra Y, I\'liyazaki I, Nakagawara, Shiku H. Inverse

associatiotl of nm23-H 1 expression by colorectal cancer with lh'er malaslasis. Br J Cancer. 1 993;6~: I ()20 1024.

Yokoyama T, Yoshi 0, Aso H, Ebina T, Ishitla N, lvIizllno K. Role of natural killer cells in intraoeular rnelunoilla

metastasis. Invest Ophthalmol Vis St:i. 1986;27:516-518.

Young LHY, Gragolldas ES . .i\Jaeular uvealmelanoma treateu with proton beam irradiation. lU-year follow-up

observation with hbtopathologie eorre1ation. Retina. 1994; 14:43-46.

YOllng LHY, Egan KM, Walsh SJ\l, Gragoudas ES. Familial uveal mclunomu. Am J Ophlhalmol. 1994;117:516-520.

Zimmerman LE, MeLean IW, Foster WD. Does enuc1eation of the eye eontuining a Illalignant melanoma pre vent or

aecclerate tbc disselllinatioll of tumour eells? Br J Ophthulmol. 197R;62:420-425.

43

Page 44: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

44

CHAPTER 2

NO DEMONSTRATED EFFECT OF PRE-ENUCLEATION IRRADlATION ON SURVIVAL OF PATIENTS

WITH UVEAL MELANOMA

Gregorius P.M. Luyten, Comelia M. Mooy, Wilhelmina M.H. Eijkenboom,

Theo Stijnen, Lallrentius P. Hellemons, Theo M. Luider and

Paulus T. V.M. de Jong

American JOllmal Ophthalmology 1995; 119:786-791

Page 45: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

PRE-ENUCLEATION )-,(g~DIATION OF UVEAL I\IELANOl\IA PATIENTS

ABSTRACT

PUl'pose: To evaluate the hypolhelical effect of pre-enucleation irradiation on survival

of patients with lIveal mclanoma,

Methods: In a prospective stuely betwcen 1978 anel1990, 145 uvealmelanoma patienls

were treateel by irraeliation in two fractions of 4Gy beforc cnucleation. A historical

control group of 89 patients lVith uvealmelanoma treateel by cnucleation alone was

operateel on between 1971 anel 1990. Patients were follolVeel up until December 1992

or until eleath. The mean follow-up period was 65 months in the irradiated group anel 88

months in the con trol group.

Results: Thc preoperatively irradiated group of patients showcd no significant

improvement of the survival rate aftel' 71/2 years (75.9%) compared with thc con trol

group (72.1 %). Preoperativc irradiation was nol associateel with survival (P = .93), as

assessed by Cox proportional hazard analysis, adjustcd for age, gender, tumor location,

tumor size, cell type, anel year of enucleation. Women in both the irraeliated anel control

groups had a better prognosis than men (P = .002).

ConclusÎons: Preoperative irradiation in this nonrandolllizeel stuely had no effect on

survival of paticnts with uveal melanoma,

INTRODUCTION

Whelher micrometastasis can be ineluced by physical manipulation of an cye with uveal

lllelanoma during the enucleation proceelure, resulting in increased risk for tumor-relateel

death, relllains to be detenllineel. I,2 In animallllodcls, lllclanoma cells can get into the

circulation by lllanipulation of the tumor-containing eye3 To reduce this hypothctical

effect of the surgical procedure on the survival of uveal melanoma patients, preoperative

radiotherapy has been introduced2,4-7 Irradiation of the tumor before surgery might alter

the tumor cells, thus prevellting the implantation and growth of lhe circulating tumor

cells after thc enucleation procedureg

-II In the reported clinical studies no beneficial

effect of pre-enucleation radiotherapy has been demonstrated4 ,12-14 Char and

45

Page 46: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 2

associates 4 studied a small series of nonrandomized patients and found that preoperative

irradiation with 20 Gy had a worse prognosis compared with patients treatcd by

enucleation alone.

The purpose of our study was to evaluate the long-term effect of low-dose pre­

enucleation irradiation on the survival of a large nonseleeted series of patients with uveal

melanoma compared with a con trol group of patients trcated by enucleation alone. To date

th is study is the hU'gest series of nonselcctcd uveal melanoma patients, treated

preoperatively with a similar and rclatively low oose irradiation, with the longest

follow-up.

PATIENTS AND METHODS

Between 1978 and 1990 145 patients scheduled to have an eye enuc1eated for a

choroidal or ciliary body melanoma were irradiated 48 and 24 hours before the

operation with two fractions of 4 Gy electrou beams (16 Me V) by means of a 5 x 5-cm

anterior field on a linear accelerator. Between 1978 and 1982 patients were not treated at

random either by pre-enuc1eation irradiation or by enuc1eation alone but were treated

depending on the personal prefcrence of the individual ophthalmologist in the institute,

without selection otherwise, From 1982 on a standard protocol was used whereby all

patients received preopcrative irradiation unless there were reasons not to apply

preoperative irradiatioll, sueh as acute angle closure glaucoma (onc case), unexpected

melanoma in a phthisic eye (one case), ano a period of breakdown of the irradiation

equipment (fivc cases). All patients treated by enuc1eation alone bctween 1971 andl990

were uscd without selection as a historical control group (N = 89). Fram 1971 until

1985, postopcrative irradiation was performed on the soekets of 15 patients who

showed, on histologie analysis, extrascleral tumor growth. Aftel' 1985, postoperative

radiotherapy was discontinued because of the lack of etfect on survival. A few patients,

with small to medium-sized melanomas not adjacent to the optie disk or macula and

more than 60 years of age, were referred elsewhere for ruthenium plaque or proton

beam irradiation from 1978 on. These cases were not inc1uded in this study.

All patients had a complete physical examination before surgery, inc1uding chest

x-ray, andliver function tests; froml978 on liver ultrasonography was routinely added.

46

Page 47: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ ----'P"'R"'E-ENUCLEATION IRRADIATJON OF lIVEAL J\IELAN01 .... rA PATIENTS

No clinical evidencc for metastalic disease was detected at tIle time of treatmenl.

Palienls were followed up Iwice a year in the first two postoperalive years; after Ihat

time, annually. Follow-up examinalion consisted of ophthahnoscopy of the remaining

eye, inspection of Ihe soekel, palpation of prcauriclIlar and submandibular Iymph nodes,

and Iiver-enzyme IesIs. Follow-up data on patienlS failing Iheir appointment were

obtained by annually eontacting the general practilioner, loeal ophthalmologist, or both.

To verify the dale and cause of dcath, the files from Ihe general practioner, hospital, or

both were recovered. Uveal-melanoma-relaled death was diagnosed by proof from

biopsy malerial or by strong clinical evidenee (laboratory and radiodiagnostic) of

melaslalie disease. Otherwise the palienls were eonsidered to have died from other

causes. Palients were followed up unlil death or unti! December 1992. rnlhe pre­

enuclealion irradiation group Ihere was no loss to follow-up; four patienls were IasIlo

follow-up in the control group. The lolal mean follow-up period was 65 monlhs in Ihe

irradialion group and 88 monlhs in Ihe control group.

For each patient the following infonnation was obtained: age of the patienl al

time of enucleation, gender, location of the tumor (ciliary body, poslerior), eelllype

(spindie, mixed, epithelioid)15, lumor size classified by tumor diameIer, prominence, and

exlraseleral growth, 16 year of enuclealion, follow-up time, and evenlUal eause of dealh.

The cause of death was divided into two categories, uvealmclanoma-related death anc!

dealh by other causes. Classificalion of tumor size was as follows: TI, diameIer of 10

mm or less and prominenee of 3 mm or less; T2, diameter of 15 mm or less and

prominenee of 5 mm or Icss; T3, diameIer greater than 15 nUll and prominence greater

than 5 mm; and T4, extrascIcral growlh.

All hislologie data were reviewed by an ophthalmie pathologist (CM.M.). From

each tumor at least ten consecutivc slicles were examincd.

The X' lest was used to eompare Ihe irradiated group with the control group for

gender and lumor locations, the x'trend lesl lor tumor size and cell type, and the Student

I-test for age. Statistical analysis to study the effect of preoperative irradiation on survival

was pcrformecl using Kaplan-Meicr survival curves I? anel Cox proportional hazard

analysis. U~ In bath methods, survival was analyzed by using the uveal melanoma-related

dcath. Death by ot hel' causes was censored and thus not used in the survival analysis as

dca th bul as last date of follow-up.

47

Page 48: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 2

RESULTS

Five palients or the irraeliateel group anel ten of the control group recciveel

postoperative irradiation because of episcleral tumor growth. The e1ata on genelel',

mean age, tumor size cJassi fication, 16 eell type, anel tumor loeation for the irraeliateel

group and control grallp are shown in TABLE 1. For agc anel tumor size, the two groups

diffcred significantly. The control grollp contained relalively older patients anel had

48

TABLE 1 DISTRIBUTION OF PATIENTS AND VARIABLES BETWEEN GROUPS

ENUCLEATION PRE-ENUCLEATION ALONE IRRADlATION P (N = 89) (N = 145) VALUE

1971- 1977 54 0

1978 - 1981 28 21

1982 - 1990 7 124

Average age 61.8 57.3

Gende!'

Male 51 82

FellmIe 38 63

Location

Ciliary body II 18

Posterior 78 127

Tumor size

TI 8 14

T2 11 37

1'3 38 62

T4 29 32

CclItype

Spindie 44 75

Mixed 37 52

EpitheIioid 7 18

Necrotic I 0

, t-test

t eh i-square test f ehi-sqlwre trend test (Tumor sizc of the preoperative irradiatian group is rclatively smaller

eampared ta the enucleatioll group)

.03'

.91'

.99'

mi

.99+

Page 49: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

PRE-EI'{UCLEATf0i'! IRRADIATION OF_ UVEAL J\IELA.NOI\.IA PATIENTS

relatively larger (T4) tumors with extraselerallu1l10r growth. After 71/2 ycnrs of follow-up

the survival rales for the preaperatively irradiated group ancllhe cantrol group were

respeetively 75.9% (95% confidence interval, 66.7% to 82.9%) and 72.1 % (95%

eonfidence interval, 60.0% ta 81.0%). No significant ditTerenee in survival was fOlmd with

the log-rank test (P = .26). According to Kaplan-Meier, lhe survival estimale curve of the

prcoperalively irradialed group and the conlral group is shown in FIGURE. I. Both in the

irradialed and in the contral group (N = 234), wolllen had a significant beller survival than

men (P = .03) (FIGURE. 2).

The Cox praportional hazard analysis (TABLE 2) was used to study the effect of

preoperative irradiatiol1 on thc survival of llveal meIanoma patients l adjusted for prognostic

covariates. The covariates used were irradiation, age, gendeJ\ tumor location, tumor size l

ccll type, and year of enucleation. There was na significant association with pre-enucleation

radiotherapy (P = .93). The estimated adjusted hazard ralio (pre-enuelealion radiotherapy l'S

cllUciealion alone) was 0.96 (95% eonlîdence interval, 0.42 to 2.19).

TABLE2 COX PROPORTIONAL HAZARD ANALYSTS

DETA STANDARD P HAZARD

COEFFICIENT ,~

ERROR VALUE RATIO Î

Irradiatioll vs -OM 0042 0.93 0.96

enucleation alone

Age (yrs) 0.05 0.01 <0.001 1.06

Gende!' (\Vümen vs -1.0 0.32 (l.O02 0.37

men)

Tumor location 0.19 0040 0.64 1.21

(anterior vs

posterior)

Tumor size 0.48 0.18 0.008 1.62

Cell type 0.36 0.17 O.tJ4 1.43

Year of treatment 0.06 0.259 0.818 1.06

,~ The Beta-coefficient is the natural logarithm of the hazard ratio. t A hazard ratio, for instance 0.37 for sex, means th at at cach time during follow-up women

have a chance of dying within onc Illonth equal to 0.37 times the chance of dying within one month for men.

49

Page 50: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 2 ____ _

Uveal melanoma-realted death was associateel with older age (P < .00 I), male geneIer

(P = .(02), larger tumor size (P = .008) anel epithelioid ceU type (P = .(4). No association

was !()unelwith the year of treatmcnt (P = .82). Postoperative radiotherapy on the wcket

showed no significant effect on survival (P = .44).

* E ." (f) w

ï :J

Cf)

<ll

1.0 ._, Pre-enucleation . -'71-.. Irradiation

~. - Contral >~ 0.8

0.6

0.4

0.2

0.0 o 12 24 36 48 60 72 84 96

Time (months)

FIGURE. 1 Kaplan-l'vfcicr survival curves of the prc-cnucJcation irradiated group and thc cuntrol group (P;;;; .26).

l.°r~~~==-------:1~~ .. - Females ~. -Males

0.8r-~~:::::j lij E 0.6-j----------------------1

~ ~ c: 0.4 +-----------------------1 :J

Cf)

50

0.2+----------------------1

0.0 -j---,-----,------.---..,------,---r----.----1 o 12 24 36 48 60

Time (months) 72 84 96

FIGURE. 2 Kaplan-Meier sllf\'ivnl curves t'or womcn and men in thc tutal group of234 paticnts (P = .03).

Page 51: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ J'RE-ENUCLEATION IRRADIATION OF UVEAL MELANOtl.lA PATJEi']TS

DISCUSSION

In this study wc found no long-term difference in survival of uveal melalloma palÎenls

treated by tlVO fractions of 4 Gy before enucleation eOl11pared to patients treated by

enuc1eation alane. In vitro allel experimental studies have shown lhat all irracliation

dose of 8 Gy in two fractions would be optimal8 Reduction of 90% and 95% of

proliferation activity of l11elanoma eells after 8 and 20 Gy irradiation dose was

demonstrated, respectively, by all in vitm stud)'.8 Fot' human UVealll1clanomas,

a significant reduction of the proliferalÎon activitY/l11itotic rate lVas fOlllld aftel'

.. d" '18 11 120G t920 B t"1 . 11" preoperatlve lITH wtlon wit 1 anc y.' ecausc 0 t le expcnmenla muts

of these studies, uo substantial difterence between 8 Gy and 20 Gy pre-enucleation

irradiation on the survival ean be cxpected,

As a result of l11icrometastatic tumor spreading before or possibly tluring

enucleation, about 50% of the uvcall11elanoma patients die within 15 years af ter initial

treatment2t -23 Tn our study the survival rate in the irradiated group was 75.9% af ter

71/2 years. After 15 years the survival rate deereased to 50%, but the nUl11ber of

patients was too small la give reliable rales.

In the Cox proportional hazard analysis, uveaJ melanoma-relatecl cleath was

associatetl lVi th age, tumOl" size, and eell type, as describeel by ather authors21

-2ó

lnterestingly, we founcl a significant diiTerence in survival between men anel women,

Compared to men, women had a more f"avorable prognostie outeome. This difference

in prognosis rcmained af ter acijustInent for irradiation, age, tumor size, tumor loeation,

eeU type, anel year of trcatment. In several uveal melanoma studies on prognostie

f . 'f' d·ft C I b d 21-2426-28 aetors no slgm leant 1 erenee was 10Ul1( etween wo men an men, '

In a diseussion on survival aftel' posterior uveal melanoma, Augsburger lllentionecl a

better prognosis in women cOl11pared to men in an adclitional eliscussion of the stuely

by Egan anel assoeiates?8 More reeently, th is more favorable survival for women was

confirmed by Folberg and associates25

Although studies on primary cutaneous

melanoma are somewhat eontraelietory, a long-term survival study of primary

eutaneous melanoma paticnls also showed lh at women have a better survival rale 2930 eompared to men. '

51

Page 52: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_____________ C",H A PTER 2

The relalive risk for e1evelopmenl of melaslascs in palients trealed by pre-enuclealion

irradiation has been reportcd to be fivefold greater compared to nonirradialed con trol

patients4 In lhis study by Char and associales,4 41 palients were lrcated Wilh 20 Gy

(five fractions of 4 Gy) before enucleation. Palients wilh lumors largel' th'lIl 15 nun in

diameter or more than 5 mm clevalion withoul evidence of extraocular extension (T3

tumors) were selected for irradiation. The uneCJual tumor sizc in both groups may be

the cause of lhe observed effect. It is known thaI two fraclions of 4 Gy irradialion on

the two days prior to enuclealion e1ecreases the HLA class TI antigen expression31

Perhaps a higher irradiation dose has a ncgative intluence on the balancc bet ween

tumoricidal irraeliation dose and the possibility 1'01' the immunc syslem to recognize

tumor cells.

In anolher study, 12 29 patienls wcre trealeel by pre-enuclealion irradialion and

compared with a l11atched group of palients, who were trealed by enucleation alonc.

The tumors were large (preelominanlly T3 tumors) and the median follow-up period

for thc surviving patients was 4.1 year in the irradiatcd group anel 3.8 year in lhe

enuclcation group. The palients from lhe irradialed group received a total radialion

do se of 20 Gy in five fraclions. The five year Kapian-lvleier survival curves in both

groups was nol significantly differenl (P> .5). The cUl11ulalive 5-year survival

probability was 63.9% for thc 29 palients in lhe pre-enucleation irradiation group and

was 57.9% for the conlrol group. Tl could be intercsting 10 determinc whether the

relative survival in these matched groups \ViII remain equal over a longer time span,

for instanee, ten ta 15 years.

In lwo other sludics, the authors concluded thaI prc-enucleation irradiation

produced na beneficia1 effect on survival; however, thc radiation dose varied within

each sludy (12 to 50 Gy13 and 8.5 to 20 Gy14), and also no conlrol group was used.

Although our study hael a largel' sample size and a longel' folkllv-up comparcd

with olher reports, lhere were also several shortcomings. The patients from the

historical control group were not treated in the same period as thc patients treated by

pre-enucleation radiothcrapy. The effecl of the year of treatment was sludicd in lhe

multivariatc analysis and showcd na significant associatioll with uveal melanoma­

relateel e1eath (P = .82). Patients lreated by postoperalivc irradialion were not excludcd

from the study because no effecl could be del110nstraled in the l11ultivariale analysis

52

Page 53: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

PRE-ENUCLEATIQJ'lJ:RRADIATION OF UVEAL i\JELAN01IA PATIENTS

(P = .44). Thcre was a significant differcnce in age anel tumor si ze between the study

group and conlrol group. Both the older age as lhe largel' tumors in the conlral group

could have had a negative influcnce on survival in the contral graup. This can be lhe

reason for lhe sm aU differcncc in the Kaplan-Meier survival estimale in favor for the

pre-cnuc1eatian graup. Aftel' adjusling 1'01' these pragnostic covariates in the

nmItivariate atlalysis! this ditl'erence was nlinimized.

Tn canclusion, na long-term beneficial effect on survival could be found aftel'

8 Gy pre~enuclcatiol1 irradiation. Life prognosis in wamen was more favorable than in

men. Langer foUow-up in this study wiU aU ow the survival estimates in this

comparison to stabilize. However a long-term prospective randomizecl tfial ,vith a

large!' sample size is needed ta study the hypothelical effect of pre-enuclcation

radiotherapy.

53

Page 54: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 2

REFERENCES

I. Zimmerman LE, T\'IcLean IW, Pmler WD. Docs enuclcatioll of the eye conlaining a malignant rnclanoma

pre"enl or accelerate the dissemination of tU1ll0ur ceIIs? Hr J Ophthalmol. 1978;62:420-425.

2. I\'fanschot WA, van Peperzeel HA. Choroidal melanorna; enuclealion or observatiun? A new approach.

An:h Ophthalmol. 19RO;98:71-77.

3. Niederkorn JY. Enucleation-induced ml'laslasis of intraoculnr melanomas in miec. Ophthalillology.

1984;91 :692-700.

4. Char DH, Phillips TL, Andejeski Y, Brooks Crawford J, Krol! S. Failure or preellucleation radiation 10

decrease uveal melanoma morlalit)'. Am J Ophthalmol. 1988;106:21-26.

5. Letter 10 Ihe Editor. Authurs uI' the Collaborative Oeular Î\'telanuma SImi)' SleefÎng cOll1mittee. Am J

Ophthalmol. 1989; 107:440-442.

6. Powers WE, Palmer LA. niological hasis of preoperative radiation Irealmcnt. Am J Roentgenol.

1%8; 102: 176-1n.

7. eh ar 011, Phillips TI" The potential for adju\'anl radiolherap)' in choroidal melanoma. Arch Ophlhahllol.

I ~X2, I OO;247-24~.

R. Hoye RC, Smith RR. The effectivene~s uI' slllaU amouills of pre-operative irradialion in prc\'eilling the

growlh of tumor cells disseminated at surgery. Cancer. 1%1; 14:284-295.

9. Kenneal!y ez, Farber i\'fG, Smith ME, Devincni R. In vitro melanoma eell growlh after preenlicJealion

radiation therapy. Areil Ophthalmol. 19öö; 1 (]6:22:1~224.

Hl. Sanborn GE, Ngyllen P, Gamel J, Niederkorn JY. Reduction of enucleation-indllced melastasis in

intruclliar melanom,1 by periurbitnl irradiation. Arch OphlhnlmoL 1987; 105: 1260-1264.

11. Ivlooy CM, de Jung PTVi\·I, van der KwasI TH, Mulder PGH, Jager i\U, Ruiter DJ. Ki-67

im1ll\l1\o,';taining in \1\'('<11 melaHoma; the effect of pre-enudeation radiotherapy. Ophthalmology.

1990;97: 1276-12ÖO.

12. Augsburger .IJ, Lauritzen K, Gamel JW, Luwr)' Je, Brady LW. i''o'latehed group study ofprecnlle1calion

radiotherapy versus enucleatiun alonc for prirnnry mnlignnnimelanoma of Ihe choroid and ciliary body.

Am J Clin Oncol. 1990;13:382-387.

13. Kreis~ig I, Rohrbach IvI, Lineo!'f 11. Irradialion of choroidal melanomas before enucleation? Retina.

1989;9: 101-104.

14. I3urnfckl N, Huscr U, Sauerwein W, \Vessing A, Sack H. Prüoperative Beslrahlllng vor Enukleatiun bei

malignern Melanorn der U\'ca. Klin 1'Ibl Augenheilk. 1989; 194:252-260.

1 S. l\'lcl.ean lW, Poster WD, ZilllmerJllan LE, Gamel JW. Modilieatiolls of Callender's c1assi1katiOll of uveal

Illelanoma at Ihe Arrned Porces Institute uI' P<IIhulug)'. Am J Ophlhalmol. 1983;96:502-509.

16. Göllnilz R, LOllllll[\lz.~ch PK. Die prognostische Re1evanz hi.slopalhologiseher Parameier beim malignen

MelanoJll der Aderhaut unter Anwendung der pTNM Klassifikation. Klin Mbl Augenhei1k.

1988; 192:296-301.

17. Char DH, Huhta K, Waldman F. DNAccl1 cycle studies in uveal melanomas. Am J Ophlhalmol.

1989; 107:65-72.

18. Kaplan EL, Meier P. Nonparametric estimation from incomplete observalions. J Am Stat Ass.

1958;53:457-4R I.

54

Page 55: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

PRE-ENlJCLEATION IRRADJAT!ON OF UVEAL i\fELANOi\IA PATIENTS

19. Cox DR. Regrl's~ion models and life tahles (with uisclission)J!{ Stut Soc R 1972;34:187-220.

20. Augsburger JJ, Eagle RC, Chiu 1' .... 1, Shiclds JA. The effect of pre-enuclcutioll radiutherapy on mitotic

activity of choroklal anu ciliary body melunolllas. Ophthalmology. 1987;94: 1627-1630.

21. Jensen OA. Malignant melanomas of the hUTllan u\'ea; 25-year follow-up of cases in Denmark, 1943-1952.

Acta Ophthalmul. 1982;ÓO: lól-I 82.

22. l .... kLean IW, FosterWD, Zimmennan LH. Uvealmelanoma: location, sizc, cel1typ!..', and enucleation as

risk factors inllll'lastasis. Hum Pathol. 1982;13:123-132.

23. Augsburger ]J, Gamel JW. Clinical prognostic factors in patients with posterior uvcalmalignant

melanoma. Cancel'. 1 t.J90;66: 15%-1600.

24. ShamJ11as HP, Alodi Fe. Prognostic factors in choroidal anu ciliary bouy melauolll<ls. Arch Ophthalmol.

1977;95:63-69.

25. Folbcrg R, Rummelt V, Parys-Van Ginuerdeuren R, Ilw,lIlg T, \Vuulson RF, Pe'er J, ct al. The

prognostie value of tumor blood \'essel morphology in prinwry ll\'t'almelanoma. Ophthnlmology.

I YY3; 1 00: 13R9- U9H.

26. Colcman K, Baak JPA, Van Die st p, ivIl1l1aney J, Farrel1l\l, PClllOn r .... 1. Prognostic factors following

enuclcation of Ijl uveal me!anomas. Br J Ophthalmol. 1993;77:688-6t.J2.

27. Gamel J\V, .McCurdey JB, l\kLean IW. A ('omparison of prognostic ('ovariales tor llveal me!<lllollln. Invest

Ophthalmol Vis Sci. 1992;33: 19 [9-1922.

28. Egan KM, \Valsh S.M, Seddon JM, Gragoudas ES. An evalliation of the inlluClll'e of reproulldive fadors

on the risk of metastases from uvealmelanoma. Ophthalmology. 1993; [00: 1160-1166.

29. Shaw Hi\f, MiltOll G\\', Famgo G, l\'IcCarthy WIL Endocrine intluences on su!'\'ival from malignant

melanoma. Cancel'. [978;42:669~677.

30, Thörn ivI, Adami HO, Ringhorg U, Bergström R, KrusClllO UB. Long-term survival illlllalignallt

melanoma with special reference to age and sex as prognostic factors. J Natl Cancel' Inst. [9R7;79:9ó9-674.

31. Jager MJ, van der Pol JP, de Wolff-Rouendaal D, de Jong PTVM, Ruiter DJ. Decreased expression of

HLA class II antigens on human u\'eallllclanoma cells after in vivo X-my irradiation. Am J Ophthalmol.

1988; 105:78-86.

55

Page 56: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

56

CHAPTER 3

METASTATIC UVEAL MELANOMA: A l1lorphological mul il1ll1lunohistochel1lical mmlysis

Gregorius P.M. Luyten, Cornelia M. Mooy, Jatmie Post, O.A. Jensen,

Theo M. Luider, Paul T.V.M. de Jong

SlIbmitted for pliblication

Page 57: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

ABSTRACT

Purpose: Uvealmclanomu metastasizes often late anel preferentially to thc liver in

contrast ta cutaneous melunoma, Thc objective of the present study was ta evaluatc

the histopathological and iml11unohistochemical changes in primary uveal melanomas

anel their eOITesponding metastases.

Melhods: Of29 primary uveal melanomas and their carrespanding metastases, the

morphology anel the iml11ul1ohistochemical reactivity for melanoma associatecl

antibodies, HMB-45, S 100 protein anel NKI-C3, were assesseel.

Results: A significant differenee in eell type of the primary anel the metastatie uveal

melanoma was found (P=.OOOI). The metastases derh'ed from the 29 patients revealed

82.5% epithelioid ar highly unelifferentiateel eells. Pasitive staining of the primary

uveal melanomas and the metastases was fmllld ta be 93 % and 91 % for HMB-45,

80 % and 66 % far SlOO, and 56 % anel 71 % for NKI-C3, respeetively.

ConclusiollS: Metastases of uveal melanomas consist of a higher grade of malignant

eell types. Highly undifferentiated, non-classifiable eells ean be observed in 40 % or

the metastatic lesians. HMB-45 praved to be the most sensitive immunahistachemical

marker in metastatic uveal melanoma anel should be used as part of a panel of

monoc1onal antibodies in any undifferentiated metastatic tumor of unknawn origin.

INTRODUCTION

Uvcalmclanomas mctastasize relatively late: the 5, 10 anel 15 year survival rales of

uvealmelanoma patients are 65%,52% anel 46% respeetively. t.2 These melanomas

spread preferentially to the liver anel onee the eliagnosis of hepatie metastasis clinically

is made, the life-expeetancy is extremely paar (2 to 7 1l10nths) anel fatal in all eases 3,4

Ivletastases call OCCUl' decades aftel' treatment of the primary tumor. 1,2 Thercfore,

c1inicians should considel' llletastatic uveaI melanOllla in the differentiaI diagnosis of a

hepatic metastasis from an unknowll primary tmllor.

57

Page 58: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 3

The mosl imporlanI independenl prognoslic faclors in paticnts with a primary

uvealmelanoma are celllype, largesIlumor diameIer (LTD), presence of doseel

vascular loops and age of lhe palienl. l,6 Presence of epilhelioiel cells in thc primary

I b 1 I ' I ' 1 I' 1 I' I 2 5-8 b . tumor las een s 10wn la )e assocwlec Wit 1 a l1g 1 morta Ity rate " , ut metastatlc

diseasc does occur in patients with a pure spindie ceIl melanoma alld in relatively

small «3 mm lhickness) lumors9 Though lumors of bolh cell types metastasize in

marc or lesser exlenl, lhe relalionship bel ween lhe cell lype of the primary tumor anel

I f · I' . '11 I lOll t lat 0 lts correspallC 111g metastases IS st! unc ear. '

In cutaneous melanoma the sensitivity for the melanoma associated antibodies

HMB-45, S 100 and NKI-C3 has been reporteel to be 92%, 100%, and 95%,

respectively.12 In primary uveal mclanOlnas, thc immunohistochemical reactivity of

these antibodies has proven to be a rcliable taal in elifferentiating uvealmelanoma

from other intraocular tumors. 13-16 The S100 protein is not specific for melal10cylic

I . . .. I . 91'" f I' 1 1 13141718 B I tumors, Jut It stmns posltlve yIn -/0 0 tIe pnmary uvea me anOJllas. ' " ot 1

the monoc!clIlal antibodies NKl-C3 and HMB-45 bind a cytoplasmic antigen that is . 15161920 .

produced by letalmelanocytes as weil as l11elal1oma cells. ' , , HMB-45 IS

sensitive in 99% of the primary lIveal melanomas and is expressed most strongly at the

tumor invasion fron1. 14,15 It may ho wever be menlioned th at HMB-45 reactivity has

been demonstrated in non-melanocytic tumors but still is consiclcrcel the most specific

f I . 21 A ,. b' .. f 1 or me anocyttc tumors. n assocwtlOll etwccn lmnmlloreacttvlty 0 t lese

antibodies anel cell type has not been elescribeel.

It seems therefore appropriate to evaluate the morphological and

immunohistochemical features of metastatic uvealmelanomas in a series of

cytological smears, Iiver biopsies and aulopsy specimens.

MATERlALS AND METHODS

From the records of lIveal melanoma patients, who died of uveal-melanoma-rclateel

causes, 29 cases were selected for whom the hislologicalmaterial of thc primary

tumor anel its corresponding metastases was available. All available sections, paraffin

blocks, and cytology slides were reqllested. The morphology of the primary uveal

melanomas and the metastases could be studied in all 29 patienls. From six autopsy

cases, multiple metastatic samples from different sites \Vere available leading totally to

58

Page 59: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

l\IETASTATIC MELAN9l\"'I'-A'--_____ _

40 metastases. In six out or these 40 metastases only staineel cytologieal material was

available anel as a consequence only the morphology could be stuclied.

For immullohistochemical investigatioll, paraffin blocks of metastases were

available in 23 out or the 29 cases. In some subjects metastatic tissue was availablc

rrom more than one site, so that totally 34 metastases eould be evaluated. In 16 out of

these 34 metastases the paraffin bloeks of the eorrcsponding primary uvealmelanoma

were also available for immunochemistry.

Histopathology

From 20 serial seetions of eaeh primary tumor the predominant eell type

(spindie, mixed, or epithelioicl) was reeorded aeeording to the moclified Callender

classifieation.22

The cell type of the eorresponeling metastases was similarly c1assified

in a masked way. If the eells were not definable as spindie or epithelioiel eell type and

showed highly unelifferentiated eharaeteristics, they were calleel non-c1assifiable, or a

combination of a non-classifiable anel a classifiable eell type.

Illllllunohistochelllistry

From all available rormalin-fixed and pararfin-embeddeel bloeks, seetions of 6

to 7 mm were cut, mO\lIlted on aminopropyltriethoxysilane coateel glass slides (APES,

Sigma, St. Louis, USA), and dried overnight at 3TC to ensure tissue adherence. Aftel'

deparaffinizing anel rehydrating, enelogenOlls peroxielase activity was blockeel by

incubation for 20 min in methanol containing 3% hydrogen peroxide. The sections were

tested with three eommercially available antibodies suited 1'01' fOflnalin fixed tissue:

HMB-45 (Dako, Glostrup, Denmark), SIOO (Dako, Glostrup, Denmark) anel NKI-C3

(Sambio, Uden, The Netherlands). They wcre e1iluteel with phosphate-burrered saline

(PBS); the monoclonalmouse-antihuman-melanoma anti body HMB-45 to a solution of

I: 100, the polyclonal rabbit-anti-S 100 anti body to 1:2500, anel the monoclonalmouse­

antibody NKI-C3 to I: I O. Subsequently the slieles were incubated with a biotinylated

seeonelary anti body (Biogencx; 1:50) and a streptavidin-biotin-peroxidase complex

(Biogenex; 1:50). The peroxidase activity was maele visible using hyelrogen peroxide

anel 3 amino-9 ethylcarbazole as chroll10gene substrate. The reslIlting red stain alloweel

59

Page 60: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ --'C~HAPTER 3

easy detection of immunoreactivity in pigmented lesions, Seetions were counterstained

with Mayer's hemaloxylin anel mounted in 1msol (Klinipath, Zevenaar, The

Nelherlands), Negative conlrals were performed wilhout the prÎmary anlibody; nerves

presenl in lhe tissue serve el as internal positive control l'or the SI 00, and a culaneous

melanoma for lhe HMB-45 and lhe NKI-C3 antiboelies, The results of the

immunohistochemical slainings were graeled accoreling to the percentage of positive

staining cells; negative (0), 5-25% (I), 25-50 % (2), 50-75% (3), and 75-100% (4),

Statistical analysis

In oreler to study lhe difference between lhe cell types of each primary uveal

melanoma and their corresponding melastases, the Wilcoxon matched-pairs signed­

rank teslwas used, The tumors were ranked fromlow grade malignant (spinelle = I)

to high grade malignant (non-c1assifiable = 4), whereby in cases of multiple metastatic

lesions the malignancy grade of lhe multiple lesions was average(1. The sensitivity for

lhe antibodies HMB-45, SIOO and NKI-C3 in both primary anel metaslalic lesions was

assesseel using the McNemar pair test. The antiboelies were compared for positive

staining in pairs of two, Probability level of P < ,05 was conskiered significant.

60

FIGURE 1. Uvcal melunomu metastasis in thc jiver, Non-classifiable, undiffcrentiated Il1clanoma cells. (original magnifîcation x 361)

Page 61: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

~ ______ --""I,,-,E 'J~A STAT 1 C ME L;\ N 0 i\I A

RESULTS

The distribution of the cell types of the primary and metastatic uveal melanomas of the

29 patients is shown in TABLE 1.

TABLE 1 PRIMARY AND METASTATIC UVEAL MELANOlvIA; CELL TYPE CLASSIFICATION

AND DISTRIBUTION ACCORDING 1'0 METASTATIC SITES

CELLTYPE

SPINDLE MIXED EPITHELIOID NDN- TOTAL

CLASSIFIAllLE

Primary 9 15 5 0 29

(31 %) (51.7%) (17.2%) (0%)

MelastastÎc* 7 (4) 5 (2) 24 (6) 4 40

(17.5%) (12.5%) (60%) (10%)

Site

Liver I (I) 3 (1) 16 (4) 3 23

Subcutis I (1) 4 (I) 5

LUl1g 2 I (I) 3

Pleura 1 (1) 1 2

Orbit I (I) 1 2

Dther 2 (I) I 2 5

'" main eell type (+ nOI1-c1assifiablc areas)

In the metastatic lesions, a significant higher percentage of high grade malignant

(epithelioid and non-c1assifiable) cells was fOlllld than in the primary lesions

(Wi1coxon matched-paired signed-rank test, p=.OOO I). Sixteen of the 40 metastatic

lesiolls showed highly ulldifferentiated components or consisted of pure non­

c1assifiable, unditferelltiated cells (n=4)(FIGURE I), whereas in the primary uveal

melanomas no areas of these cells wcre observed at all. A correlation betweell cell type

61

Page 62: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__________ ---'C'"-'H A !'TE R 3

TABLE2 EXPRESSION OF HMB-45, S-IOO AND NKI-C3 IN PRIMARY AND

METASTATIC UVEAL MELANOMA

ANTIBODY GRADE PRIMARY UVEAL METASTATIC UVEAL

MELANOMA MELANOMA

HMB-45 0 lII5 3/32 1 2/l5 5/32 2 4/15 5/32 3 4/l5 7/32 4 4/l5 12/32

5-100 0 3/15 12/34 1 2/l5 5/34 2 3/15 6/34 3 4/15 4/34 4 3/15 7/34

NKI-C3 0 7/16 10/34 1 0 5/34 2 1/l6 0

3 3/l6 6/34 4 5/l6 13/34

anel the metastalic sile cOlllel not bc maele becallse lhe majority of the metaslalic

lesions \Vere only obtaineel from the liver (23/40).

The graeles of inlll111nohistochemical reactivity for HMB-45, S 100 anel NKI­

C3 in primary anel metastatic llvealmelanomas are listeel in TABLE 2. The percentage

of posilively staineel cells for HMB-45 \Vas 93 % in primary anel 91 % in metastatic

lIvealmelanomas. The percentage of positive slaining for S100 and NKI-C3 was 80 %

and 56 %, respectivcly in the primary lesions anel 66 % anel 71 %, respeclively in the

metastases. In the primary uveallnclanomas, 110 significant differences in staining

bet\Veen lhe lhree antiboelics cOlllel be demonstrateel llsing the McNemar lest. In the

metastases, lhe HMB-45 was fOllllel to be significantly more sensitive than S 100

62

Page 63: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ -',"-, 'é-E T, A STA TIC l\1 E_ [, A N~'",O",',-,' ,,-' ___ _

(p=,0391) bul nol significanlly more sensilive lhan NKl-C3 (P=,0703); no significanl

elitferencc in positive staining was elemonslrateJ between NKI-C3 anel SIOO

(p=,687S), Of the thrcc HMB-4S negatively staineel metastalic lesions, one was found

positivc for bath S 100 anel NKI-C3, whereas the two other lesions eliel not slain tal' any

of thc antiboJies, These lhree HMB-4S negative mctastatic lesions silOweel no

relationship lVith a specilïc eeU type; i.e, one case was of lhe spinellc eeUlype, one of

the mixed eeU type, one of the epithelioiJ eeU type anel IIVO lVith aelelitional

unelifferentiateel eharacleristies (TABLE 3), Thc S!OO and NKI-C3 negative metastatic

lesions showeel na correlation lVith a speeific ecU type, although the spinelle eeU

tumors had thc highest frequency of negative staining (TABLE 3), AUI'our metastases

eomposed of pure non-classifiable eells showcel a positive staining for HMB-4S

(FlGURE 2A anel 2B), three for NKI-C3 anel two for S 100 (TABLE 3),

TABLE3 ANTlBODY EXPRESSION IN PRIMARY AND METASTATlC UVEAL

MELANOMAS VERSUS CELL TYPE

ANTIBODIES

CELLTYPE HMB-45 S-100 NKI-C3

+ - + - + -

Primary Uvenl Spindie 4 0 2 2 1 3

Melunomu Mixed 8 1 X I 6 4

Epithelioid 2 0 2 0 2 0

Total 14 I 12 3 9 7

.Mctastatic Uveal Spinelle 5 1 2 5 2 5

lvIelanoma Mixed 4 I 4 I 3 2

Epithelioid 16 1 14 4 16 2

Non classifiable 4 0 2 2 3 I

Tatal 29 3 22 12 24 10

63

Page 64: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

64

CHAPTER 3 ______ _

FIGUUE 2A and 2B. Uveal mclanomH metastasis in the liver. Undiffcrcntiated uveal melanoma eells among normal hepatoeytcs (A) WA staining (B) HMB-45 staining (original magnification x 361.)

Page 65: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

i\lETASTATIC l\lELANO~lA

DISCUSSION

The primary uvealmelanomas in our study were selected on lhe basis of available

specimens of metastases and showed a relatively high percentage of mixed and

epithelioid cells compared to the normal distribution of cell types in prilllary uveal

melanolllas (50% spindie; 50% mixed and epithelioid)5,23-25 Since lhe OCClllTence of

epilhelioid cells in primary uveallllelanomas has been associated with an increasecl

, k C dl' 'd" I" I d' I 2 5 23-25 I' l' I I' I fiS ~ lor eve opmg metastatlc lsease m C 111lCa stu les, ' ,-, t 11S re atlve y llg 1

incidence of epithelioid cells in prilllary uveallllelanomas in our sludy is probably

caused by our selection criteria, Nevertheless, we fOllllcl in metaslalic uveal

Illelanomas a significant higher percentage of epithelioid cells cOlllpared to the

primary melanomas, This is consistent with earlier findings: Fuchs and associates 11

found in a total of 31 metastases derived fl'Om 11 uvealmelanoma patients nine

spindIe (29 %), five mixed (16 %) and 17 epithelioid (54 %) cell types, The celilypes

of the 11 corresponding primary uvealmelanomas were however not specified, Zakka

and associates 10 fOlllld in an aulopsy study on uvealmelanoma no spindie, two mixed

and five epithelioid eell tumors in the metaslases from uveal melanomas; the primary

uvealmelanomas eonsistecl of spindle cells in one case, mixed in four and epithelioid

in two, The highly undifferenlialed cell types observed in the present study were nol

d 'b d' I d' 'd b 10 11 escn e 111 t 1e stu les mentlOne a ave. '

A laek of differentiation is eonsidered to be a hallmark of Illalignaney,

Tumor cells showing lack of differentiation are ealled anaplastic, Literally anaplasia

mealls Hto form backward", implying areversion fr0111 a high level of diffcrentiation to

a lower level. However, it ean be disputed whether the laek of differentiation of the

non-c1assifiable uvealmelanoma cells is the consequence of declifferentiation or of a

malignant transformation of immature melanoeytes without any sign of differentiation,

Therefore, these undifferentiated cells were called non-classifiable,

In our study we fOllllcl four metastases that were of pure undifferentiatecl

nature and another 12 metastases with additional nOll-classifiable areas ncxt to areas

with classifiable cells, These highly undifferentiated cells eOllld not be detectecl in lhe

primary tumor.

65

Page 66: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 3. ____ _

Uveal melanomas of all eell types have mctastatic polency to some extcnt. Howevcr,

in the Callender classification system (spindie, mixed and epithelioid),22 we cannot

exelude the existcnee of sm all areas of epithelioid or non-classifiable eells in a

melanOlna cJassified as,predominantly spindie. Current views on tumorigencsis

propose that neoplasms develop by a sequential aequisition of a number of genet ic or

epigenetic events, whieh lead to eell growth, aberrant differentiation, invasive growlh,

d I b·I· . 26-2R TI·· I· I . f . I an t le a 1 lty to metastaslze. 11S Hnp les t lat tumors may anse 1'0111 a Slllg e

eell, i.e th at they are monoclonal29 Tf the mutant subclone has a proliferative ability it

may lead to overgrowth of the original or the met as ta tic lesion. In our study we found

a high percentage of non-classifiable eells in the metastases, whieh may be uerived

from such a mutant subclone.

Furthennore, it is yet unclear whether eell-cell and cell-matrix interaction may

influence the tumor cellmorphology. From the present study, we could not

demonstrate any correlation between cell type and metastatic site. In a study on

eytokeratin expres sion in primary and metastatic lIveal melanoma, na correlatÎon

between tumor site and cytokeratin expression eould be demonstrated. 11 However, we

reccntly demonstrated a significant differenee in expression of nemal eell adhesion

I I . I· dl· . 30 mo ecu es 111 Ivel' metastases an ot lel' metastatle SItes.

In eoncordance with findings in cutancous melanomas, we demonstrated that

HMB-45, although not significantly different from NKI-C3, was the most sensitive

marker for metastatic lIveal melallomH. 12-15 In studies on cutaneolls and primary uveal

melanomas a high expression of SIOO was found. t2-14,t8 In contrast to these findings,

wc founu a signifieantly lower staining l'or S 100 in metastatic uveal melanoma; S 100

will thus he less reliable in the diagnosis of diffieult cases of undifferentiated

mctastases. Thc percentage of positive staining was not related to the cell type of the

metastatic lesions. A non-significant higher percentage of ncgative staining tumors for

SIOO and NKI-C3 was found in the spindie cellmetastatie lesions. The most

aggressive eell types kept their immunoreaetivity for HMB-45 and in alesser ex tent

for SIOO and NKI-C3; all non-classifiable metastases stainec! positively for HMB-45.

66

Page 67: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

~lETASTA'I:IC lIELANO!\.~I~A~ __ _

COllclllsion:

Metastatic uveal Jllclanomas contain a significantly high percentage of epithelioid and

highly undifferentiated, non-classifiable cclls compared to their corrcsponding primary

lesiolls. The illul1unohistochcmical stainillg of HMB-45 appeared to be the most

sensitive marker in characterizing metastases of uveal melanomas. Uveal melanoma

may metastasize decades after diagnosis of lhe primary lumor. Therefore, it is

important th at HMB-45 should be included in the panel of mOlloclonal antibodies for

the c1iagnosis of undiffcrentiated metastatic tumors, to identify delayed and unexpected

uveal melanoma metastasis.

67

Page 68: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_________ "C"H,,~ P TER 3

REFERENCES

I. Gamel JW, l\'kLcan IW, 1..,,1cCurdy JE. Biologie distinctions bctween CUfe and time la dcath in 2892 paticnts

with intraocular I11clanomu. Cancer 1993;7 J :2299-2305.

2. lensen OA. Malignant melanomas of the hunulIl uvea; 25-year tollow-up of cases in Denmurk, 1943 lY52.

Acta Ophthalmol 1982;60: 161-182.

3. Kalh R, Hayungs J, Bornfeld N, Sauerwcin \V, Höffkcn K, Seeber S. Prognosis aml tr('alment of disscminated

uveul melanol11a. Cancer 1993;72:2219-2223.

4. Rajpal S, Maarr R, Karakousis CP. Survival in mctastatic oeular mclnnoma. C;mcer 1883;52:334-336.

5. Luyten OPM, Moo)' 0. .. 1, Eijkenboolll WMH, Stijncn '1~ Hellemons LP, Luider TI'.'1, et al. No demonstrated

effect of prc-cnucleation irradiation on survival of patients with uyealmelanoo1<l. Am J Ophthalmol

1995:1l9:786~ 791.

6. Rummelt V, Polberg R, Woolson RH Hwang T, Pe'er J. Relation between the microcirculation arehitecture antI

the aggressive hehavior of cilimy body melanomas Ophthalmology 1995; 102:844-851.

7. Coleman K, Baak JPA, Van Diest P, Mullaney J, Farrell 1\'1, Fentonl..,,1. Prognostic factors following

enudeation of 111 live al melanomas. Br J OphthalmoI1993;77:688-692.

8. Seddon Jivf, Albert DJ\·[, Lavin PhT, Robinson N. A prognostic factor study of disease-free interval and

survival following enuc\eation for uvealmclanoma. Areil Ophthalmo11983; 101: 1894-18<;19.

9. Shiclds CL, Shields JA, Kiratli H, De Potter P, Catef JR. Risk factors tor growth and metastasis ofsmall

choroidal mehll10cytic lesions. Ophthalmology 19<;15; 102: 1351-1361.

10. Zakku KA, Foos RY, Ornphroy CA, Siraatsma BR. r"falignant melanoma; analysis of all autopsy population.

Ophthalmology 1980;87:549-556.

11. Fuchs U, Kivelti '1', SUlllmanen P, Jrnmonen I, Tarkkanen A. Au imTllunohistochemical and proguostic analysis

of cytokeraHn expression in malignant u\'ealmelanOilla. Am J Puthol 1992;141: 169-181.

12. Fernando SS, Johnson S, Bate J. Immunohistoehemical analysis of cutaneolls malignant melanoma:

comparison of S-100 protcin, Hl\'lB-45 monoc\onal antibody and NKI-C3 IlIonoclonal anlibody.

Pathology 1994;26:16-19.

13. Sumier MN, i\kLeun IW, Gamel J\\'. ITlltllunohistochetllÎCal evaluation of uveal me1anocytic tumors.

Rxpression of WvlB-45, S-100 protein, alld neuron-specific enolase. Cancer 1991;68:809-R 14.

14. Beckenkamp G, Schafer H, von DomUIus D. Immllnocytochemical parameters in ocular malignant rnelanoma.

Eur J Clin Oncol 1988;24:S41-S45.

15. Steuhl KP, Rohrbach JM, Knorr rvI, Thiel Hl. Significance, speeit1city, and u1trastructurallocalization of

HrvIB-45 antigen in pigmcnted ocular tumors. Ophthahllology 1993; 100:208-215.

16. Ringens PJ, van Haperen R, Vennegoor C, de Jong PTV! ... 1, van Duinen SG, Ruiter DJ, et al. Monoc\onal

,mtibodies in deteclion of choroidal melmlOnm. Gmefe's Arcil Clin Exp OphthalmoI1989;227:287-290.

17. Kan-Mitchell J, Rao N, Alhert DM, van Rldik LJ, Taylor CR. S-100 imll1unophenotypcs of uvealmelanomas.

hivest Ophthalmol Vis Sci 1990;31:1492-1496.

18. Cochran AJ, Lu HF, ti PX, Saxton R, Wen DR. S-100 proteio remains a practica1marker for melanoeytic aod

other tumours. Melanoilla Res 1993;3:325-330.

19. Vennegoor C, Hngeman Ph, van Nouhuijs H, Ruiter DJ, Calafat J, Ringens PJ, et al. A monoclonal antibody

specific I' Of cclls of the melanocytic lineage. Am J PatIlol 1988;130: 179-192.

68

Page 69: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

i\l ETA S TA TIC ME L A N O~""[,,A~ ________ _

20. Gown Aivt, Vogel AM, J-Ioak D, G{lUgh p, ivkNutt r-,·tA. MonodOilal antibodics specific for Illclanocytic

tumors distinguish suhpopulations of mclanocytes. Am J Pathol 1986; 123: 195-203.

21. Yates AJ, Banerjee SS, Bishop PW, Graham KE. Hi.,./B-45 in non-melanocytic tUll1ours. Histopathology

[ 993;23:477-478.

22. lv1cLean lW, Foster WO, Zimmerman LH, Gamel JW. Moditkations of Callendcr's classitlcation of uveal

melanoma at the Armed Forces Institute of Pathology. Am J Ophthalmol 1983;%:502-509.

23. Rummclt V, Folberg R, RUllunelt C, Grumnn U\'l, Hwang T, Woolson RF, et al. i\lkrodrculation nrchitecture

of melanocytic !levi and malignant Illelanomas of the dliary body aIll1 choroid. Ophthalmology 1994; 101 :718-727.

24. Damato BE, Paul J, Foulds \VS. Predictive factors ofvisual outcome af ter local rescclÎon of choroidal

melanoma. Br J OphthalollloI1993;77:61fi-623.

25. !vlarcus Di\-I, l"Iinkovitz JR, WardweIl SD, Albert 01..,,1. The "allIc of nudeolar organiLer regions in ll"eal

melanoma. Am J OphthalmoI1990;llO:527-534.

26. Knuuson AG. Hereuitary calleer, onl'Ogenes, and antioneogencs. Ca neer Res 1985;45: 1437-1443.

27. 1 jOlla LA, Steeg PS, Stetler-Ste\'enson \VG. Caneer metastasis and angiogenesis: an

imbalance of positive and negative regulation. Cell 1991 ;64:327-336.

2K Steeg PS. Genetie eontrol of the metastatic phenotype. Seminars in Caneer Biology

1991;2:105-110.

29. Talmadge JE, Wolman SR, FiJler IJ. Evidence from the donal origill of spolltaneous metastasis.

Scienee 1982;2l7:3fi 1-363.

30. 1100)' Cf"I, Luyten GPt"f, Jensell GA, LuiderT!vl, van der Hmn P, de Jong PTVivl, ct al. Neural cell adhesion

molecule distribution in primary and meta.~tatie uveal melanomH. Human Pathology 1995;26: 118S-1190.

69

Page 70: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

70

CHAPTER 4

NEURAL CELL ADHESION MOLECULE DISTRIBUTlON IN

PRIMARY AND METASTATIC UVEAL MELANOMA

e.M. Mooy, G.P.M. Luyten,

P.T.V.M. De Jong, O.A. Jensen, T.M. Luider,

F. Van Der Ham and F,T. Bosman

Human Pathology 1995;11: 1185-1190

Page 71: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ --'NEURAL CELL ADHES)ON MOLECULE DISTIUI"-'"U-'T-"'"O-'N'---____ _

ABSTRACT

Tumor ecU ac1hesion, detachmcnt anel aggregation play an important part in tumor

invasion anel metastasis, anel a varicty of eeU aclhcsion molecules have been fOlllld on

tumor eells. Cell adhesion molecules, inclueling those of the immunoglobulin

superfamily, are associatccl with the development of melastatic behavior in cutaneous

melanomas. The neural cell aelhesion molccule (NCAM) belongs to this family. To

investigate its possible role in the development metastatie behavior of uveal

melanomas, we stuelieel immunohistochcmieally the expres sion of NCAM by using an

antiboely that recognizes all three major isoforms of NCAM anel an antibody which

recognizes the HNK-l epitope present on some isoforms of NCAM. The authors

stuelieel 32 primary uveal melanomas from 32 patients (among these, 12 were rapielly

mctastasizing anel 16 slowly metastasizing) anel 29 metastases from 19 patients. From

13 palients thc primary, as well as thc metastatic, tumors were available. With unc

exception, all HNK-l positive primary anel metastatic tumors were also positive for

NCAM. NCAM was significantly more expresseel in aggressive, rapidly metastasizing

primary tumors (P = .02 anel .04, respcetively) anel in metastases. HNK-l was

signifieantly (P = .(4) more expressed in largel' tumors. In liver metastases HNK-l

imnlllnoreaetivity was significantly (P = .(05) less frequently cxpressed than NCAM.

Thcrefore, NCAM isoforms which lack the HNK-I epitope might play a wie in the

organ specific llletastatic behaviul' of uveal melanomas.

INTRODUCTION

There is increasing eviclence that changes in adhesiveness anel 1110tility are of

eonsiderable importance in tumor progrcssion anel may be the prime features

determining aggrcssiveness anel metastatic potential.' Aelhesive properties or

malignant cells must change repeateelly to allow them to e1etach from their primary

location, attach to the extracellular matrix, enter a blood vesscl anel eventually loelge at

a metastatic site2 Aelhesion molecules which comprise of scveral complex families,

mediate interccllular interactions and interactions between cells anel the extracellular

matrix 3 They are grouped into four main classes on tbc basis of their moleeular

71

Page 72: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ ~C~H~A P:',-,T,-" R'--"3 _____________ _

structure: the integrin family, the cadherin family, the ill1munoglobulin superfamily

and the selectin superfamily4 The neural cell adhesion molecule (NCAM) and the

intereellular aelhesionmoleeule (TCAM) belong to the immunoglobulin (Tg)

superfamily.3,4 Although the members of the Ig superfamily are funetionally elivers,

most are cell surface molecules involveel in recognition of other surface molecules.

Severalneural eell aelhesion molecules have been shown to mediate intercelllilar

aelhesion eluring the development of the nervous system, inclueling the neural cell

aelhesion molecule (NCAM), myelin associateel glycoprotein, L I, amalgam, contactin

anel fascicilin n5 The HNK-I carbohydrate epitope is associateel with several of these

aelhesion molecules, including NCAM6 Although the exact role the HNK-! epitope

plays is unknowl1, it seems to serve as a ligand in eell adhesion?

For cutaneOllS melanoma it has been sho\Vn that the development of metastatic

potential is associated with de novo expression ofICAM_!8,9 anel MUC!8, 10 an

antigen which shows sequenee similarity to NCAM. 10.11 In contrast, others showeel

expres sion of MUC!8 on a full range of benign anel malignant melanoeytic lesions. 12

Tn uveal mclanoma, ICAM-I couId not be detecteel in one study, 13 but using a

dilTerent anti-ICAM-! monoclonal antibody (MAb) most ofthe uveal melanomas

. I 14 15 . I f . I .. f I . d d . I I"d Il 15 I SlUIIlCC, ' WIt 1 a pre erentJa reactlvlty 0 t 1e I11lxe an eplt 1e lOl ce type. n

several hu man malignancies tumor progression was paralleled by changes in NCAM

expression.I.16-18 A role for NCAM in the development of malignant potentialof

uveal mclanomas has, howcver, not been reportcel.

This study investigates whether NCAM expression is correlated with the

development of metastatic potential in uveal melanoma. The authors studied primary

tumors with known clinical outcome (inclueling rapidly metastasizing and c1inically

non- or slowly metastasizing tumors) and all available metastases. The aLlthors report

here on NCAM, which \Vas stained by a polyclonal anti body that recognizes all three

major NCAM isoforms, and by MAb Leu-7, which recognizes the HNK-! epitope l9 of

the cell binding domain of sOllle NCAM isoforms20.21

72

Page 73: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NEURAL CELL ADHESION 110LECULE DJ~rr_R=IB"-U~T,,,IO=N _____ _

MATERIALAND METHODS

Patient Selection

From the files of patients with uvealmelanoma related death, the authors were

able to eol!eet metastatie tissue from 20 patients (nine liver biopsies, one skin biopsy,

one autopsy, nine fine neeelle aspirations). From 14 of these patients the paraffin

bloeks from the primary eiliary body or ehoroidal melanoma were available; their

follow-up varieel between 6 months anel 147 months. Frol11 five of these patients,

paral'fin blocks from the primary, as weil as the metastatic tumor were available. Uveal

melanomas ereate a peak incidence of mortality during the seeonel anel thirel years aftel'

enucleation, irrespective of the hU'gest tumor diameter22

Tumor relateel e1eath within 3

years was therefore eonsidered to be causeel by rapidly metastasizing melanoma (n =

8, mean foUow-up, 2 1,3 months) (TABLE I). Furtherl11ore, the atlthors seleeted nine

patients with a follow-up of at least 10 years aftel' enucleatioll without clinical

evielenee of metastatic disease, of whom paraffin blocks of the primary tumor were

available. These melanomas were consiclered to be of low metastatic potential

(TABLE 2). The totalmean follow-up of this group (n = 11 inclueling two cases of

tumor relateel e1eath aftel' more than 10 years) was 169.5 1110nths.

From nine patients, the paraffin blocks of the primary uveal melanomas and

eight correspollding metastases (4 autopsies, 4 biopsies) were obtailled from the Eye

Pathology Institute in Copenhagen: the follow-up varieel between 2 years anel 32

years. Four were rapielly metastasizing melanomas (TABLE I) and five were of low

metastatic potential (TABLE 2). All these patients e1ied of tumor related death.

The total material consisted of 32 primary tumors. Among these, 12 were

rapidly metastasizing (TABLE I) and 16 slowly metastasizing (TABLE 2); the

remaining four deceased due to tlveal melanoma related e1eath between three and 10

years. We investigated 29 metastases (TABLE 3) from 19 patients. Of 13 patients

tissue of the cOlTesponding primary anel metastatic tumors were available.

73

Page 74: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 4

Histology

Sections were cut at 5 to 6 mm anel stnined with hemntoxylin-eosin. Of the

primary tumor the preelominant cell type (spinelie, mixed or epithelioid) and largest

tumor eliameter (LTD) (<; 10, 10 to 15 mm, anel >15 mm) were recoreleel.

Imm Uil 0 his toch emis t ry

Paral'fin sections, 6 to 7 flm thick were cut anelmOllllted on

aminopropyltriethoxysilane (APES, Sigma, St. Louis, USA)-eoated glass slicles and

elrieel overnight at 37°C. Aftel' deparaffinizing anel rehyelrating, endogenO\ls peroxidase

activity was blockeel by ineubation ror 20 minutes in methanol containing 3%

hydrogen peroxiele. Aftel' rinsing the slicles in water, antigen retrieval was performed

by microwave irraeliation (Bio-Rad 37°C, 750 W; 2 x 5 minutes in 0.1 % pl'Onase). The

slicles were ineubateel with phosphate-buffereel saline (PBS) at 4°C for 10 minutes and

subsequently at 1'00m temperature (RT) for 5 minutes. To deteet all three major NCAM

isoforms, irrespective of the presence or absence of polysialylation,23 a polyclonal

antibody was used24 in a 30 minutes ineubation in a dilution of 1 : 100 at RT. This

antiboby was a generous gift from Prof. E. Bock (Research Center for Medical

Biotechnology, University of Copenhagen). Visualization of anti body binding was

performed as described below. HNK-I, a sulfated glueul'Onic carbohydrate epitope

which is present on the 145 and 180 kD isoforms of NCAM and in related

proteins21 ,25 was detcetedusing the anti-Leu-7 monoclonal antibody (Becton

Dickinson, Sunnyvale, CA). This antibody was originally generated against a hu man

T-cellline and is present on a subpopulation of natural killer eells, 19 but is also

associatedwith scveral adhesion moleeules 20,21 This antibody was used in a elilution

of 1 : 10. The slicles were incubated for 30 minutes at RT with biotinylated goat-anti­

mouse-rabbit-rat-guinea-pig Ig (Biogenex, San Ramon, CA) in a dilution of I : 50, in

PBS with 5% BSA. Aftel' washing in PBS/Tween 0.05%, the slieles were ineubated

with the streptavidin-biotin-peroxidase complex (Biogenex) in a dilution of I : 50. The

peroxidase was visualized using hydrogen peroxide in N-N-dimethylformamide with

3-amino-9 ethylcarbazole as chromogenic substrate. The red stain allowed easy

eletection of immunoreactivity in pigmented lesions. The sections were counterstained

74

Page 75: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NEURAL CELL ADHEfiJON l\IOLECULE DISTRlnUTION

with Mayer's hematoxylin anel mounted in glycerin gelatin.

As a negative conlrol, specimens werc stained following the same incubation

protocol without use of the primary antibodies. A human neuroblastoma served as a

positive control lor NCAM. For HNK-I, nerve lissue and tumor infiltrating

Iymphocytes present in the tissue section serveel as an internal positive contra!.

NCAM and HNK-l imlllunoreaclivily were scoreel sel1li-quantitatively as a

percentage of positive cells: score 0: 0%; score I: 0% to 5%; score 2: 5% 10 50%;

score 3: 50% to 100%. The il1ll1lunohislochemical staining was scoreel wilhoul

knowleelge of the clinical elata anel was repeated after 2 lllonths in oreler 10 assess

repraelucibility. Major (more than Iwo classes) eliscrepancies eliel nol occur.

Statistical analysis

Speannan's rank correlatio!l was used ta test the relation bet ween the variables

celllype, LTD anel rapiclly/slowly metastasizing tumors anel NCAM and HNK-I

staining, respectively. Fisher's exact test was useel to test correlalions of NCAM and

HNK-I inUllUtlOreactivity in Ihe metastatic sites.

75

Page 76: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 4

RESULTS

Primary tumors

Thirteen tumors were of the spindie cell type, 15 of the mixed cell type and

four of the epithelioid cell type. Four tumors were small (SlO nun), 22 were

intermediate size (10 to 15 mm), and six were large (> 15 nun).

NCAM positive tumors showed strong cytoplasmic staining; in three

cytoplasmic (FIGURE lA) and membrane bound staining was noted (FIGURE IB).

The results for NCAM expression in the different cell types are illustrated in

FIGURE 2. NCAM was signifieantly more expressed in mixed-cell type and

epithelioid-cell type melanomas (P = .02) and in rapidly metastasizing tumors

(P = .04). Rapidly and slowly metastasizing tumors are summarized in TABLE 1

and 2. In the HNK-l-positive tumors, st rong eytoplasmic staining was noted

(FIGURE 3). HNK-I was significantly more expressed inlarger tumors (P = .04).

76

FIGURE IA mullE, Field of uveal melanoma of the epithelioid cell type immunostained ror NCAM. (A) Cytoplasmic staining (original magnification x 361) (B) membranc-bound staining (original magnification x 880). (Hematoxylin counterstain with streptavidin­biotin-pcroxidase complex immunopcroxidase with 3-amÎno-9 ethylcarbazole dimethylfofmamide substrate.)

Page 77: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NEURAL,Ç.ELL ADHESrON r.1~)LECULE DISTRIBUTrO~.

SpIndIe n=13

NCAM expression in uvcal melanoma

Mixed tplS

Eplthellold MeIaslasis n=4 n=29

FIGURE 2. Results [or NCAM immunostaining in primary and metastatic uveal melanoma. Percentage of positive eells: 11 50% - 100%; 11 5% - 50%; <5%; 0 0%.

FIGURE 3. Field of metastatic uveal mclanoma inUllunostaineJ [or HNK-I. (Hematoxylin counterstain with streptavidin-biotin-peroxidase complex inulllmoperoxidase with 3-amino-9 ethy1carbazole dimethylfonnamide substrate; originalmagnification x 361).

77

Page 78: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______ C HAPTER 4 ____ _

Metastases

The staining pallern of all metastases for NCAM is illustrated in FIGURE 2;

the staining pattem in the difTerent organs is specified in TABLE 3. Fol' NCAM, 72%

of the metastases stained (score 2+3). For HNK-l, 27% of' the metastases stained

(score 2+3). HNK-I was significantly (P = .0(5) less expressed inliver metastases

compared to olher metastatic site. In five autopsies, different metastases fro111 the same

subject had a varying score (from negative to score 3) bath fol' NCAM and HNK-l.

78

PAT. CltLL

NO.t TYPE

J S

2 E

3 M

4 M

5 E

6 M

7 E

8 M

9 M

to M

J t M

t2 M

TABLE 1 IMMUNOHISTOCHEMICAL DATA OF MELANOMAS

METASTASIZING WITHIN 3 YEARS

PRIMARY UVEAL CORRESPONDING

IVIELANOMA METASTASIS

r;m NCAM HNK·l NCAM HNK·l

2 t 0 NA NA

2 2 0 NA NA

2 2 0 0 0 2 2 2 NA NA

2 2 0 NA A

2 NA 0 NA NA

3 2 2 NA NA

2 0 3 3 NA

2 3 0 NA NA

2 2 () NA 2

3 2 2 0 0

2 2 NA 0 NA

NOTK Score 0 = negativc; score 1 = 0% 10 5% tumor cells positive; score 2 = SClo 10 50% tumor cells positivc; score 3 = 50% 10 IOW'/o tumor eeIls positive. Abbrcviations: E, epithelioid cclltype; fit mixed eeU type; S, spindie reIl type; LTD, largest tumor dinmeter: I: <la mm, 2: 10-15 mm, 3: >15 mm; NA, paraftïn block not availab\e; NCAM, neural ccll adhesion molecllie. t: All patients died of tumor-relatcd dealh.

Page 79: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NEURA)_ CELL ADIIESION l\10LECULE DI_~,)TRIBUTION

TABLE2 IMMUNOHISTOCHEMICAL DATA OF MELANOMAS

WITH A FOLLOW-UP OF MORE THAN 10 YEARS

PRIMARY UV"A CORR"SPONDING

MELANOMA METASTASIS

PAT. CELL

NO." TYPE LTD NCAM HNK·l NCAM HNK·l

1 M I 2 '" '" 0 '" 2 S 2 n 0

3 S 2 0 '" * 0

4 S 2 0 '" * 0 ,~

5 S 2 0 0

6 S 1 0 ,~ ,

0

7 S 3 2 ':< '" 0

8 M 2 0 * '" ()

9 S 2 ,~

I IJ 10";" S 2 2 3 NA NA ut M I I 0 3 IJ

12t S 3 3 3 2 2 \3+ S 2 0 0 2,2 3,2

14t S 1 3 0 2 0

15+ S 2 0 0 n,2,3,3 0,0,0,0 16i" M 3 2 0 3 0

NOTE. Score 0= negative; score 1 = 0% to Y'lo tumor ceUs positive; score 2 = 5% to 50% tumor cells positive; score 3 = Socio to 100% tumor cells positive. Abbreviatiolls: E, epithelioid cel I type; M, mixed cell type; S, spindie cell type; LTD, largest tumor diameter: I: <10 nun, 2: 10-15 mlll, 3: > IS mm; NA, paraffin block not available; NCAM, neum! cell adhesion molecule. *: patients alive and free of metastatic disease after> 10 years; Î: tumor-related death af ter > 10 years. :j:: multiple metastatic sitcs obtained at autopsy.

79

Page 80: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

ClIArT ER 4'---_________ _

Correspoudiug primary >Iud metastatic tumors

Of 13 tumor pairs (partly reflected in TABLE 1 and 2), nine primary tumors

and their metastases showed positive (score 2+3) staining for NCAM; in four major

(negative vs positive) discrepancics bet ween the primary tumor and the metastases

were llOted. For HNK-I major discrepancies were llOted in threc tumor pairs.

Relatiollship between HNK-I and NCAM immunostaining.

Of the primary tumors, 71 % was NCAM positive, but only 20% was HNK-I

positive. Similarly, of the metastases 72% were NCAM positive, but only 27% was

HNK-I positive. However, with one exception, all HNK-I positive primary and

metastatic tumors werc also positive for NCAM.

80

TABLE3 STAINING PATTERN IN THE METASTASES

NCAM HNK-I

NEG. POS. SCORE R NEG. POS. SCORE

I 2 3 1 2 3

Liver 3 1 6 5 12/15 14 - 1 -

Lung I - 6 2 8/9 5 - 3 1

Skin* I - 2 - 2/4 1 1 2 -

Abdomen" - - - - Dil - - - 1

Total* 5 I 14 7 22/29 20 1 6 2

* Thc Beta-coeftïcienl is the nuturallogarithm of the hazard ratio. Î A hazard ratio, tor instanee 0.37 for sex, Illcans that al each time during follow-up I,VOl11en

have a chance of dying within one ll10nth cqual 100.37 times the chance of dying within one month for men.

R

1/15

4/9

3/4

lil

9/29

Page 81: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ ~NEURAL CELL ADHESION Iv[OLECULE DISTRIBUTION'-' ____ _

DISCUS SION

Three major isoforms of NCAM have been identified, whieh are genera led by

alternalive splicing of a single gene, Two of these isoforms are transmembrane

proteins, differing only in lheir eytoplasmie domain, NCAM·140 kD has a short

eytoplasmie lail, whereas NCAM·180 kD has a large eyloplasmic domain thoughllO

interact with lhe cytoskeleton, NCAM·180 expres sion is restrictedlo neural tissues;

the 120 kD and 140 kD isoforms also oeeur on other celllypes26 One rale of lhe

different isoforms may be to targel lhem to different cellular destinations: lt has been

fotll1d that glycosyl phosphatidyl inositol (GPT) membrane anchoring (small

cytoplasmic domain NèAM.120) acts as an apieal tm'geling signa 1 in epithelia,27 The

antibody against NCAM used in our study recognizes the three major forms of

NCAM, whereas MAb Leu·7 recognizes the HNK·l epitope present in the 145 and

180 kD isofonn of NCAM, Furthennore, NCAM exhibits special earbohydrate

eharaeterislics: glycosylation of NCAM seems to be regulated during development

and to intluence the adhesivc function of the moiccule,2R It has been suggested that

polysialylated NCAM present in early (embryonic and fetal) stages of development is

involved in cellular migration, whereas the expres sion of unsialylated NCAM in

tissues may be important for local differentiation and organization29 Polysialylation

of lhe NCAM molecule decreases its adhesion properties and may thereforc play a

fa1e in connection with tumor invasion anel metastasis. ImmUllohistochemical

investigation does not pro vide information about NCAM polysialylation,

The percentage of NCAM positive primary tumors in our study was higher

than has been reported for cutancous melanomas,12,17,30 This might be explaincd in

tenns of methodologieal differenees: The authors used a different (polyclonal)

antibody on formalin·fixed, paraffin·embedded tissue and also applied an antigen

retrievalmethod, NCAM immunostaining was mostly eytoplasmic andless frequently

both cytoplasmic and membrane bound, which is in keeping with previous

, " 173031 NCAM "f' 1 d' 'I I1lvesligatlOIls. ,. ,- was SIgIlI leant y more expresse Hl tumors WIt 1

presence of epithelioid eells (mixed and epithelioid ccll type), The presence of

epitheIioid cells is one of the factors associaled with progression of uveal melanomas

(developlllenl of lllctastatic potential)32 That NCAM is expressed in epithelioid eells

8J

Page 82: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_______ ~c HArTER -"4'---_______________ _

secms surprising because epithelioid eells are lllorphologically noneohesive. Tt is not

deal' how adhesion molecules on the surface of cancel' cells influence metastasis.

Aelhesion molecules may delay the escape of tUlllor cells f1'Om the primary site

because of an increaseel 'lelhesion to other cells and to extracellular matrix proteins.

However, an altered pattern of CAM expression or expression of aberrant CAM might

disrupt normal adhesion anel attaehment of these cells to a cndotheliulll or extraeellular . . . 16

matnx at a metastatlc sIte.

lmlllunoreactivity l'or HNK-I was fOllllel in approxilllately one third of the

NCAM positive lesions, whieh can be explaineel by the HNK-I epitope being present

on only two isot'ol'lllS of NCAM. Interestingly, HNK-I was significantly less

expressed in liver metastases, compared ta ather mctastatic sites. This discrepancy was

not found for NCAM. Uvealmelanomas metastasize relatively late33,34 and in contrast

to cutaneous melanomas primarily haematogenously, preferentially to the liver. Gnee

hepatic mctastases arc clinically present, the median survival is extremely poor: only 2

to 9 months 35 These findings suggest th at NCAM isoforms, lacking the HNK-I

epitope might play a mie in the organ-specific pattern of uveal melanoma metastasis.

Isoforms with alternative moeles of membrane association are targeted to different

surfaces of polarized epithelial cells; the 120 kD small eytoplasmie elomain is

expressed on apical surfaee anel is attachcd to the cell membrane via a GPI-linkage,

whereas the 140 kD small surface domain form and the 180 kD large dOlllain farm are 27 expresscd on the basolateral surface.

Furtherlllore, the authors found a significant inerease of NCAM and HNK-I

positive lesions alllong rapidly 11lctastasizing tumors anel large tumors, linel in

metastases. This is in keeping with the increasing evidenee that changes in

adhesiveness and motility are of eonsiderable importance to tumor progression. 1

Howcver, a de novo expressioll in tumor progression could not be showl1 as has been

fOlllld for related cellldar aelhcsion molecules ICAM-I anel MUC-18 in progrcssion of

eutaneOlIS melanomas. 8-10 The pattem of NCAM and HNK-I expression in the

primary tumors differed significantly from th at in the paircd metastases, suggesting

th at NCAM anel HNK-l exprcssion are probably modulated by the tumor cell

microenvironment.

In summary our results show that expression of NCAM is associated with the

82

Page 83: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NEURAL eELt ADHESION MOLECULE DtSTRIBUT-,-IO""N __

development of malignant potentialof lIvealmelanoma. FlIrthermore we fOllnd that

NCAM isoforms, lacking the HNK-I epitope may be associatec1 with the organ

specit'ic metastatic behavior of uveal melallomas. The prognostic value of NCAM allel

HNK-I expressioll in uvealmelalloma rcmains to be established.

83

Page 84: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 4 ------

REFERENCES

I. Rünger TM, Klein CE, Becker JC, et al. Thc role of genetic instability, aJhesion aml immune escape

medlanisrns in melanoma progression. Curr Opin Olleol. 1994;6: 188-196.

2. Poste G, Pi dier Y. The pathogenesis of eaneer metaslasis. Nature. 1980;283: 139-146.

3. Edelman Gi'vL CAivis and 19s: Cell adhesioll and the evolutionary origin of immunity. Illll11unol Re\'.

1987;1011: 11-46.

4. Bock E. CeU-eell adhesion molecules. Bioehem Soe Trans. 1991; 19: 1076- 1 980.

5. Covault 1. i\'Ioleeubr biology of eell adhesion in lleuraI developmeot. In: Glover or .... t, Hames HD

(eds) l' ... foleeular Neurobiology. IRL Press, Oxford, PI' 143-200, 1989.

6. Edelman Gi\l. r .... todulation of cell adhesion dllring indllction, histogenesis and perinatal uevelopment

of Ihe nerVOllS system. Annll Rev Neurosci. 1984;7: 339-377.

7. Hoffman S, Edelman Gr.·1. A proleoglycan with HNK-I anligenic detenllinants is a neuron-assocÎaled

ligand for eytotaetin. Proe Natl Aead Sei USA. 1987;84:2523-2527.

8. Johnson JP, Stade BG, Holtzmann B, ct al. De nova expres sion of ICA1\'I-1 in melanoma correlates

with increased risk of Illetastasis. Proc Nat! Aead Sd USA. 1989;86:641-644.

9. Natali PO, Nicotra .MR, Cavaliere R, et al. Differenlinl e:'\pression of ICA! ... l I in primary and

melastatic melanoma Icsions. Caneer Res. 1990;50: 1271-1278.

10. Johnson JP, Lehmann JI\'I, Stade HG, et al. J-IunelÎonal aspects of th ree molecules associated with

metaslasis deve!oprnenl in human malignant melanorna. Invasion "tvIelastasis. 1989;9:338-350.

11. Lehmann Ji\I, Ricthmül!er G, Johnson JP. p.'lUCI8, a marker oftul11or progression in hl1l11an

mclanoma, shows sequenec similarily 10 the neuml eell alihcsion molecules of the imml1noglobulin

sllperfamily. Proc Nat! Aead Sci USA. 1989;86:9891-9895.

12. Denton KJ, Stretch JR, Gatter KC, et al. A slud)' of adhcsion molecules as markers of progressioll in

malignant lllelanoma. J Pathol. 1992; 167: lö7-191.

13. Carre! S, Schreyer t-.t Gross N, ct al. Surface antigcnic pronIe of lIveal mclanoma lesions analysed

with a panel of monoclonal antibodies dirceted against cutailcous melanoma. Anticaneer Res.

1990;10:81-89.

14. Van der Pol JP, Jager i'v[J, dc Wolff-Rollendaal D, et al. HClcrogeneOlls expression of melanoma­

assoeiated :tntigens in uvcal melanomas. CUITcnt Eye Rcs. 1987;6:757-765.

15. Natali PG, Higotti A, Nicolm ~'1R, ct al. Analysis of thc antigenic profile of ll\'eal melanoma lesions

with allti-elltaneOl1s melanoma-associated antigen ano anti-HLA monoclonal antibodies. Canccr Res.

1989;49: 1269-1274.

16. Johnson JP. Cell adhcsion molecules of Ihc immlllloglobl1lin supergene family and their role in

malignant transformation alld progression la metastalic discase. Cancer melastasis Rev. 1991; 10: 11-22.

17. Micttinen M, Cupo W Neuml cel! adhcsioll molecule dislribution in soft tissue tumors. Hum Pathol.

1993;24:62-66.

1 K Roth J, Zl1bcr C, W,lgner p, et ,11. Rc-cxpression of poly(sialic acid) units of Ihc neural een adhesion

molecule in Wi1ms Tumor. Proc Natl Aead Sci USA. 1988;85:2999-3003,1988

19. Aho T, Ba1ch C. A diifercntiation antigen of human NK and K eel1s identitïcd by a mOTloclonal

anti body (HNK-I). J IIlllllunol. 1981; 127: 1024-1029

84

Page 85: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ ~NEURAL eELt ADHESJON I\·IOLECULE qISTRIBUT10~N,-' ______ _

20. Cole 01, Schachner :1\'1. Localization of the J,2 lllonoclonnl untibody binding site Dil N-CA!\J ,md

evidence fOf ils role in N-CAlvI-mediated eell ad hes ion. Neurosci Lelt. 1987;78:227-232.

21. HoffmUll1l S, Edelman G. A prolcoglycan wilh HNK-I untigenic dclcfll1Înants is a neuron-assoCÎated

ligand for eytoactin. Proe Nat! Acau SCT USA. 1987;84:2523-2527.

22. Zimmerrmlll LE, 1'kLeunIW, FusterWD. Does enuc!eation prevcnt or accelerate thc dhsemillation

of tumour eells? Br J Ophthnlmol. 1978;62:420-425.

23. !\'Ioolenaar CECK, I\-fuller El, Schol DJ, et al. Expression of neural eell adhesion molccule-related

sialoglycoprotcin in small ceillung cancer ond ncurobl,lstoma celllincs H69 and CHP-212. C;mcer.

Res 1990;50: I 1ü2-1lOó.

24. Rasmussen S, Berezin V, Nnrganrd-Peuersoll H, ct al. Pmificatioll of Ihe glycoprotein D2 Irom felal

anti auult human bmin. In: Protidcs of the Biological FlllÎUS, XXX Colloquium, Peeters, H. (etl.)

Oxfurd: Pergamon Pre ss, 1983, p 83

25. ~'icoazec J-Y, Racine L, Coulevard j\, ct al. Parenehymal innervation of norlllal and cirrhotÎl' hUIl1an

livcr: a light and electron microscopie study Llsing Illonodonal antibodies against the neuml cc11-

adhesion molecule. J IIi~tocheIl1 Cytochcm. 1993;41 :899-YOS

26. Cuntlingham BA, Edelman Gi\!. i\lorphorcglliatur)' ivlo1eculcs. etls Edelm,lJl Gr..'l, Cunninghalll BA,

Thiery JP. WHe)', New Vork, 1990

27. Puwell SK, Cllnningham BA, Edelman GM, ct al. Targeting oftmnsmembrane and UPI-anchoreu

forms of N-CAI\'1 tu opposite domains of a polarizetl epithelial cell. Nature. 1991 ;353: 76-77.

28. Krog L, Bock E: Glycosylatiotl of neuml cell adhesion molecules of thc immunoglobulin superfamily.

APJ\HS Suppl. 19Y2;27:53-70.

29. Linnenl<HH1 D, Bock E. Cell auhesion molecules in ncural development. DeY NeUfosci. 1989; 11: 149-173.

30. Garin-Chesa P, Fe11inger EJ, Huvos AG, ct al. ImmUllohistochemkal a[wlysis of Ileural cell adhesion

molecules. Differential cxprcssion in smal! rouml cell tumors of ehildhood and adolescence. Am J

Pathol. 19YI;139:275-286.

31. Carbone OP, Koros Al\-IC, Linnoila I, et al. Neuml cel! adhesion molecule expression and messenger

RNA splicing patlerns in lung cancer celllines arc corrdated with neuroenocrine phenotypc ;\Ild

growth morphology. Cam'er Res. 1991 ;51:6142-6 I 49.

32. Seddon Hv!, Albert DM, Lavin PT, et al. A prognostic f:1clor sludy of disease-frce interval and

survival following enuclcation lar lIveal melanoma. Arch OphthalmoI. 1983; 10 I: I gY4-1899.

33. RaÎvio r. Uvealmelanoma in Finland: All cpidcmiological, clinical and prognostic study. Acta

Ophthalmol (Suppl). 1977; I 33:3-64

34. Jensen OA. MaHgnant melanomas of the human lI\'ca: 25 year follow-up of cases in Denmmk:

1943-1952. Acta Ophthalmol. 1982;60:161-182.

35. Kath R, Hayungs J, Bornfeld N, el al. Prognosis and tre,tlment of disseminatcd uveal me!ano!lla.

Canccr. 1993;72:2219-2223.

85

Page 86: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

86

CHAPTER 5

THE EXPRES SION OF NM23 GENE IN UVEAL MELANOMA

Gregorius P.M. Luytcn, Nicolc C. Naus, Cornelia M. Mooy,

Sonja Kerkvliet, Theo Stijncn, M.E. Mei Li.Tang, Ttte cic Waarcl·Sicbinga,

Theo M. Luider, Paulus T. V.M. cle Jong

Submittecl for publication

Page 87: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ --'N-'-"1~'_f"'2,,3__"E_"_X I~ R E S SlO N I N U V E A L 11 E L A I\f q,,~,,1 A"--______ _

ABSTRACT

Purpose: Low NM23-protein expression in human tumors has been associated with

metastatic potential, short diseasc-free interval anel decreased survival. We

investigated the prognostic significance of NM23-protein expression in uveal

melanoma.

Materials and melhocls: A retrospective inununohistochell1ical analysis was

performed on 42 fonnalin-fixed paraflïn-embedded primary uveal melanomas using a

NM23/NDP kinase A antibody. The percentage of NM23-llegative-staining areas was

comparecl with conventional prognostic factors anel survival. An iml11U11ocytochemicai

anel Western blot analysis of five primary and two lnctastatic uveal melanoma cell

lines was made in order to study the specificity of the antibody.

Results: The NM23-protein expression was not correlated with the conventional

prognostie factors for uvealll1elanoma. The association between NM23 expression

and survival was of borderline significance. The NM23-protein was expressed in the

cytoplasm and to alesser extent in the nucleus or the cell membranes. Both the NM23-

Hl and H2 isol'orms were found in the primary as well as in the metastatie uveal

melanoma cell lines.

ConclusioIls: This study provides an inclication that NM23 might be involved in uveal

melanoma, but that the use of the anti-NM23/NDP kinase A is of limited prognostic

value.

INTRODUCTION

Malignant ll1elanomas of the choroid and ciliary body are the most conunon primary

intraocular tumors in adults with an annual incidence of 6 per ll1illion.1 Fifty percent

of the uveal melanoma patients die within 15 years due to metastatic disease? which

is most frequently located in the liver? Identif'ication of genes and proteins that arc

involved in the process of metastasis may be useful as prognostic markers. The protein

product of the NM23 gene has been proposed to be a putative metastasis suppressor

87

Page 88: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 5

protein. The NM23 gene has been isalateel by elifferential hybrielization of relateellow

anel high metastatic murine melanoma sublines of the cellline K- 17354 Two human

NM23 isofarms have been elescribed, knawn as Nlv!23-Hl and NM23-H2, which are

88 percent homalogclUs to each other4 .5 They have beenmapped ta chromosame

17q21.3-22 and were founel to encoele for two proteins of approximately 17 kDa6

These genes appear to have a sequence homology ta the lnlluan nucleoside

diphosphate kinase-A and -B (NDPK-A,NDPK-B),7 NM23-Hl anel NM23-H2,

respectively. The relation bet ween the elegree of expression of the NM23 protein anel

metastatic potcntial has been most extensively studied in breast cancel'. In patients

suffering from breast cancel', low levels of NM23 mRNA anel NM23 protein have been

founel to corrclate with metastases to the lymph nodes, eleereaseel elisease free

intervals, anel dccreased relapse-free survival.S-

11 Furthermore, 10\v levels of the

NM23 protein is of prognostic significanee in a variety of other tumors, 12-15 inclueling

cutancous mclanoma,16-19

Thc purpose of the present stuely was to investigate the prognostic value of

redueeel NM23 protein expression in primary uvealmelanoma. Ta this end, the NM23

protein expression was studied in histologie al specimens of primary uveal melanOlnas

allel in bath primary anel metastalic llvcallnelanoma celllines as weil.

88

Page 89: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ ~1\~Tl~H.23 EXPRESSION IN UyEAL 1IELAN9~"~1,~\ ___ _

MATERlALS AND METHODS

Patients

Conseclllive patients sllffering from llveal melanomas of whom the eye had

been enucleated beilVeen 1984 and 1992, enlered this sludy on basis of availability of

adequate histological matcrial. Thc follow-up data were obtaincd from our database or

by contaeting the local ophlhalmologisl and/or general practitioner until December

1994. To verify date and cause of death, Ihe files were recovered. Uveal-melanoma­

relaled dealh was diagnosed by proof from biopsy malerial or by clinical history

logelher with laboratory and radiodiagnostic evidence of metaslalic disease. Otherwise,

the patients were consielered 10 have died from other eauses. Forly-two formalin-fixed

paraffin-cmbedded primary uvealmelanomas were available. Of Ihese, the following

parameters were recordcd: locatioll (ciliary body, choroidai), celllype (spindIe, mixed

or epithelioid), and largest lumor diameter (LTD) of Ihe tumor, presence of extra-scleral

growth, age al the time of enucieation, gender, follow-up time and, if appropriale, eau se

of death.

Celllilles

The eelllines llsed in Ih is study were established from five primary (EOM-3?O,21 21 2122 2324 .

EOM-29 , OCM-I, , MEL202, 92-1 ), and two melastatlc llvealmelanomas 202125 21 . . (OMM-I, , . OMM-3 ); apart from one cell hne from a breast caremoma (T47D) and

one from a cutaneol1s meianoma (Bowes). The eellline OCM-I was generousiy provided

by Dr. June Kan-Mitchell (USC Schooi of Medicine, Los Angeles, CA), while the cellline

MEL202 was a generous gift from Dr. Bruce Ksander (Schepens Eye Research Tnstitute,

Harvard Medicai School, Boston, USA). The T47D cellline obtained ti'Om a ductai breast

. I' . i 26 B' . Ili' carCllloma, was usec as a posItlve contra. owes IS a metastatle clltaneOllS ce me

obtailled from ATCC (CRL-9607/ U.S. patent 4, 766, 075). Six of these cell lines (EOM-3,

EOM-29, OCM-i, OMM-l, OMM-3, Bowes) were cultured in a Falcon tlask (30i3E)

with Dulbecco's-modified-Eagle's-medium (DMEM) sllpplemented with 10% heat­

inactivaled fetai caifserum (FCS), penicillin (0.751l1g/1l1i) and strept01l1ycin (1.25

1l1g/1l1i)(PS) at 5% C02 ancl37°c. 19 This medium was changed twice a week and the cells

89

Page 90: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ eliA P TI",' R"--'S'----_

were passaged depending on their growth rate, For the other three celllines, T47D,

MEL202, and 92-1, the RPMI medhull (Gibco, LifeTechnologies BV, Breda, The

Netherlands) was used, supplemented with 10% FeS and PS, The medium ofT47D was

supplemented with insulin (0,2 lU/mi),

IlIlll1unohistoehemistry

Paraffin sections (5 mm) were mounted on 3-aminopropyltrioxysilane coated glass slides

(APES, Sigma, St. Louis, USA) and dried overnight at 37°e, Aftel' deparaffinization and

rehydration, the slieles were rinsed in a (J,01 M citrate buffer (pH 6,0); antigen retrieval

was performed by microwave irradiation (Bio-Rad) for 2 x 5 min at 100°e, The slides

wcre pre-incubated with normal goat serum in a dilution of I: 10 phosphate-buffered­

saline (PBS) with 5% bovine serum albumin (BSA) (Fraction Y, Sigma, Axel, Bclgium)

for IS min, Subsequently, the slides were incubated overnight with anti-NM23/NDP

kinase A (Boehringer Matlllheim Gmbh, Almere, Thc Netherlands) in a dilution of I: 1600

PBS with 5% BSA at 4°e, Aftel' being washed in PBS/0,05% Tween 20 (PBS-T), the

slides were incubated for 30 min at room temperature with biotinylated goat-anti-rabbit

immunoglobulin (DAKO, Glostl11p, Denmark) in a dilution of 1:50, 2% normal human

serum and 2% normal goat serum in PBS with 5% BSA. Aftel' being washed in PBS-T

again, the slides were incubated with streptavidin-biotin-alkaline-phosphatase cOlllplex

(Biogenex) in a dilution of 1:50 for 30 min at room temperature, Finally the slides werc

rinsed with PBS and Tris-HCl (pH 8,0), and they were incubated in the dark in a solution

of 18 mg naphtol AS-MX, 2 mi dimethylformamide, and a mix of 200 mi fuchsine, 200

mi Na2N03 and IS mg levamisoL Eventually the sections were counterstained with

Mayer's hematoxylin and mounted in glycerin gelatÎll. As a negativc con trol, the primary

antibody was replaced by normal rabbit serum (DAKO Glostl1lp, Denmark; dilution I: 10),

A breast cancel' with an extensive in situ component sen'ed as a positive control.

Based on the hypothesis th at only few cells of the primary tumor have to become

metastatic competent in order to kill a patient, a scoring system was devised by Steeg et

al.27 They quantified the percentage of NM23 negative-staining (NM23-) tumor cells in

the area of the section with the lowest overall staining, Along the same line we screened

for NM23- eells in primary uveal melanomas in a masked marmer. ln each scction, ten

fields \Vere screeneel anel scoreel for the percentage of NM23- cells,

90

Page 91: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_____ ~I~V~k~f2=3_E~·X~·~P~R=E~SION IN UVEAL t-.IELANOMA

Imll1Unocytoehemistl'Y

The celllines we re cultured on glass coverslips utltil they reached half confluency.

Then the coverslips were washed twice with PBS and during 10 min fixed in cold

aeetone (-20°C). Endogenous peroxidases were inhibited by 0.3% hydrogen peroxide

in methanol for 15 min. Subseguently, the eoverslips were ineubated in 2% FCS/PBS

for 5 min and with anti-NM23/NDP kinase A (dilution I :20) rOl' 135 min at room

temperature. As a seCOlId step antibody, swine-anti-rabbit immunoglobulin (DAKO,

Glostrul', Denmark) was used in a dilution of I: I 00 and these coverslips \Vere

ineubated for 60 min. Then they were ineubated with 3,3'-diaminobenzidine

tetrahydrochloride and ureum hydrogen peroxide (Sigma, St. Louis, USA) 1'01' 7 min,

to detect the peroxidase aetivity. Cells \Vere eventually eounterstained with Mayer's

hematoxylin fOl' 10 seeonds, rinsed with water 1'01' 10min, dehydrated and mOllllted in

Entelan. Between these steps the eoverslips \Vere washed \Vi th PBS-T. The NM23

prolein staining was scorcd with a four-point scale: intense, 3; medium, 2; weak, I; no

staining, O.

Western bi ot analysis

The cultured cells were washed twiee with warm PBS (37°C), detached with a eell

seraper and then collected in cold PBS. This cell suspension was washed with cold

PBS again and sonicated, anel could be stored at -20°C aftel' addition of

phenyhnethylsulphonyl fluoride (PMSF) in a final concentration of IlllM. The protein

concentratÎon of the eell suspension was detennined by applying the BCA-Illethod

(Pierce, Rockl'ord, TL, USA). Per lane 75 mg of total cellular proteins were

eleetrophoresed in a 12% SDS-polyacrylalllide gel, prepared according to the

manufacturer's instructions (Biorad, Veenendaal, The Netherlands), and then

transferred to a polyvinylidene difluoride mem bra ne (Illllllobilon P, Millipore, Etten­

Leur, The Netherlands) in a Biorad blot apparatus at 100 Volts for 2 hr. The blots \Vere

blocked by putting them 1'01' one hl' at room temperature in PBS-T with 2% BSA.

Subseguently, the blots we re ineubated overnight with NM23!NDP kinase A anti body

in a dilution of I :20 at 4°C, As a secO/ld step, peroxidase-conjugated swine-anti-rabbit

antibody was used in a dilution of I: 10,000 for one hr. Following each incubation step,

91

Page 92: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAI'TER 5

blots were rinsed five times with PBS-T for 5 min. The peroxidase aclÎvily IVas

visualized using an Enhanced Chemiluminescence (ECL) deteclion syslem

(Amershamlnternational, Uttle Chalfont, UK). Blots were stained with 3,3'­

diaminobenzidine tetrahydrochloride anellll'cum hydrogen peroxide (Sigma, St. Louis,

USA) 1'01' 5 min. The slaining inlensity was scored as in the coverslips.

Statistical alla lysis

The correlalÎon bel ween the differenl parameIers was sludied using the Spearman rank

correlation test. The influence of percentage ofNM23- cells and the other prognostic

factors on survival was studied by performing univariate and multivariate Cox

proportiOluli hazard analysis?8 Survival was analyzed by llsing the uveal mclanOlua­

related death. Dealh by other causes was consiclered as a censored observation anel

th us not used in the survival analysis as death but as lasl date of follow-up.

RESULTS

Patient data) histopathological c1assification anel immunohistochcmistry

Fifteen out of lhe 42 patienls died of metastatic disease before December 1994

wilh a lotalmeclian follow-up time of 37.5 (6 - 88) months. Twenty-six patients were

male and 16 female; their mean age at diagnosis was 60.8 yems. The mean LTD of the

42 primary tumors was 12.6 mm; anlOng these, there were 12 spindie eell tumors and

30 epithelioid or mixed eell tumors; we founel 33 choroidalmelanomas and 9 ciliary

body tumors; 13 of these tumors showeel extrascleral growth.

The inllllunoperoxidase staining revealeel that the NM23-protein was localizeel

in the eytoplasm and to alesser extent in the nucleus (FIGURE I). The ave rage

percentage of NM23- cells in all primary uvealmelanomas was 78%.

Illllllllnocytochelllistl'y mul 'Vestem blot analysis

The inllllunoeytochemical NM23 staining intensity of the uveal melanoma

celllines is shown in TABLE I. The staining pattern was again mainly cyloplasmic

and to alesser extent staining was observeel in the nucleus or membranes (FIGURE 2).

92

Page 93: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ --'1,,'1-'1",,1-=23 EX PRE S SlO N I N U V E A L J\ 1 E L A N,'"O"" .. !,,A'--_______ _

FIGURE 1. Primary tlveal mclanoma stained with anti-1l1ll231NDP kinase A. Positive nm23-staining cells in a

primary uveal melanoma.

93

Page 94: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______ -'CHAPTER 5

FIGURE 2. Cytoplasmie and I1uclear staining of I1m23 in lIveal melanomu cells (OrvlM-l) 011 coverslips.

94

0\ rr, .-< 01 cr, .-< , , en tl , , , ~ ~ ~ ~ ~

Ol) Q ~ .-< 0\ ~ ~ ~ r-,

0 0 U 01 0 0 "i" ~ 0\ 01 W W 0 0 0 ~ E-<

FIGUUE 3. Western blot atwlysis of the celllincs; I;me I, 92-1; lane 2, Mc1202; lane 3, EOM29; lane 4, EOM-3; laneS, OCM-I; lane 6, OMM-3; lane 7, OMM-I; lane 8, Bowes; lanc 9, T47D.

97 68

43

29

14

Page 95: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NiH23 EXPRESSI!~N IN UVEAL I\[ELANONcA'--______ _

TABLEI NM23 EXPRES SION IN CELL LINES

CELLLINE ORlGINOF EXPRESSION OFNM23 WITHANTI·NM23/NDP KINASE A

TUMOR CYTOCHEMISTRY WESTERN IILOT

92·1 UM 3 3

Me]202 UM 3 2

EOM29 UM 2 I

EOM·3 UM 3 3

OCM·] UM I 2

OMM·] MUM 3 3

OMM·3 MUM 3 ]

T47D Brcast carcilloma 3 3

llowes nadal metaslasis 1 1

cutanalIs melanoma

Uivl: uveal melanoma (primary); rvlUrvl: metastatic lIvealmelanoma Scoring: Cells on coverslips were scorcd according la cytoplasnlic staining intensity. A four-point scale was used: intense staining, sçore 3; medium staining, score 2; wenk staining, score 1; na stailling, score O. Tile same scoring scale was used for the Western bIDt analysis. nd:notdonc

All celllines clearly expresseel the NM23 protein except for two celllines (OCM·I,

Bowes) that showeel only a weak staining.

The Western blot analysis of the extracts of the celllines of seven lIveal

melanomas, olle breast carcinoma anel olle cutaneous melanoma, as determined by the

NM23/NDP kinase A antibody is shown in FIGURE 3. Two banels (Mr 17 kD anel Mr

18.5 kD) coulel be eletected in all celllines; in three (92·1, OCM·I, anel OMM-I) a

third band of Mr IS kD was seen (Jane 1,5 anel 7). Their staining intensity is shown in

Table 1. Extracts of the celllines 92-1, EOM-3, OMM-l, T47D showeel two

intensively NM23 staining bands (Jane I, 4, 7 anel 9); those of MEL202, OCM-I, anel

Bowes a less intensive (Jane 2,5 anelS), whereas in those of EOM-29 anel OMM-3

only a weak staining could be eletecteel (lane 3 and 6).

95

Page 96: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 5

TABLE2

SPEARMAN RANK CORRELATION COEFFICIENT (r,)

OF THE EXAMINED PARAMETERS.

PERCENTAGE OF NM23· CELLS

l's

Cell type"' .15

Extmscleral growth -.09

LTDt .05

Locationt -.04

GeneIer .05

Age -.03

" Cell type (spinJlc celll mixed anel epithelioid)

t Largest Tumor Diameter

+ Tumor location (choroid/ciliary hody)

TABLE3

P-value

.32

.57

.74

.78

.77

.86

COX PROPORTIONAL HAZARD ANALYSIS; UNI- AND MUr;rIVARTATE ANALYSES

VARIABLE DETA STANDARD P HAZARD

CORRECTED FOR COEFFICIENTt ERROR VALUE RATJO:j:

11m23* .033 .017 .057 1.034

Age .cl37 .019 .048 1.037

GenJer .037 .018 .040 1.038

Extrascleral .033 .017 .059 1.033

gruw lh

Tumor loc at ion .034 .018 .060 1.035

Largest tumor diameter .035 .019 .068 1.035

Cell type .032 .018 .074 1.032

*' Univariatc analysis

-j- Thc Beta-coefficicnt is the naturallogarithm of thc hazard ratio.

:'ç A hazard ratio, for instance 1.034 for nm23, mcans that at each time during follow-up

paticnts with an uvealmclanoma with I1m23- staining areas have a chance of dying within

onc month cqual la 1.034 times the chance of dying within ane lllonth for patients with

n11123+ staining are as.

96

Page 97: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

Statistical aJllllysis

The results of the Spearman rank correlation test are shown in TABLE 2.

No correlation between the percentage of NM23- cells and the other clinical and

histopathological parameters could be demonstrated.

In the Cox proportional hazard univariate analysis (TABLE 3), no significant

association bet ween survival and percentage of NM23- cells could be demonstrated

(P=.057). Because of the relatively small sample size and therein the small number of

uveal melanoma-related deaths (n= 15), only two parameters could be included

simultaneously in one l1mltivariate analysis. The results of the llluHivariate analysis of

the clinical and histopathological parameters with the NM23- expression showed a

significant prognostic value for the NM23- expression aftel' adding age (P=.047) and

gender (p=.040) to the llluitivariate analysis. However no significanee was I'ound by

adding extrasc!eral growth, tumor location, LTD and cell type.

DISCUSSION

Although the prognostic value of NM23/NDPK has been demonstrated in variOlIs

I .. I' I . I RNA . 8-19 I Hlman tumor types wang 11llll1UllO lIstoe lCtnlCa or m expreSSlOll, ot lel'

research groups found contlicting results showing no correlation between the

NM23/NDPK expression and survival 29,30 In two other tumor types even increased

levels of NM23 expression have been I'ound in tumors with a high grade of

malignancy31,32 In cutaneous melanoma, rcduced NM23 expression on mRNA level

has been reported in aggressive tumors. 17, 18 The NM23 mRNA expression was

variabie in metastatic lesions of cutaneous mclanoma, 17,19 while 1011' lllRNA levels in

the primary Icsion showed a significant correlation to a reduced disease-I'ree

interval. 17,20 These results found in cutaneOlIS melanoma suggested a promising role

of thc NM23/NDPK expression as a prognostic factor in uveal melanolllas. The

present study of prilllary uveal melanolllas showed no correlation between the NM23-

protein expression, using the NM23/NDP kinase A anti body, and the other prognostic

factors. Even the 1110st independent prognostic factors in llvealmelanoma, sllch as

LTD, cell type and age of the patient, showed no correlation with NM23/NDP kinase

97

Page 98: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______ . ___ . __ .~APTER .. _5 __ . ____ . __ . ___ .. __ .

A expression. A borderline significant association could be demonstrated between the

percentage of negative staining tumor cells for the NM23/NDP kinase A anlibody and

survival in uveal melanoma palients in the univariate and multivariate analysis. The

relatively small sample size of the study limiteel the multivariate analysis to incluele

only two parameters. In a study with a large sample size, the inverse relation between

NM23-expression and survival coulel possibly more significant.

In the present study we have used a polyclonal antiboely to NM23/NDP kinase

A anel this anlibody showed immunoreactivity mainly in the cytoplasm anel in alesser

ex tent in the nucleus, which is in agreement with other reports27 This anlibody was

chosen because it shoulel recognize only the NM23-H I isofol1n. The detection of the

NM23-HI isoform was founel to be of more significance to demonstrate the inverse

association between the NM23 and survival. 12,27 However the NM23/NDP kinase A

anti body we have useel in our experiments showeel cross-reaclivity with the NM23-H2

boform. The Western-blot analysis demonstrated both NM23 isofol1ns, NM23-HI and

H2. The M r 18.5 kD and Mr 17 kD bands that showed up are ielentical to the deseribed

positions ofthe NM23-HI anel NM23-H2 proteins.27,33 Even a third band of Mr 15 kD

was found in three celllines, but this eoulel possibly be the result of the high am ou nt

of protein extracts (>50 flg) used in the experiments27 It will probably important to

use more selective antisera, whieh recognize specifically the NM23 isofol1ns. Whether

this could also be the reason that we could not find a ditference in NM23 expression

bet ween the primary and metastatic celllines by the immunocytochemieal and

Western blot analysis can be questioned. A more selective mouse monoclonal

antibody, th at was used by Ma et al.(manuscript in preparation)34, showeel a

eomparable NM23 protein expression in the celllines EOM-3, MEL202 and OCM-1.

A discordance however was found in cellline 92-1. Although we found a high NM23

protein expression in the cellline 92-1, Ma et al 34 found only weak

immunohistochemical expression ofNM23/NDPK-A anti gel!. This discordance in

NM23-protein expression in cellline 92-1 can possibly explaineel bl' the selectivity of

the two ditferent antibodies or ineleed by in vitro seleetion of the melanoma eells,

which can bc the result of the culturing technique.

The exact biological role of NM23 as a metastasis suppresser protcin is still

lacking9,27 Nucleoside diphosphate kinases (NDPK) activity of the NM23 prolein has

98

Page 99: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ ---'-N'c1"'W"'2} EXPRESSION IN. UVEAL i\[ELANOi>.IA'--______ _

been proposed to be related to the anti-metastatic effect. However, it has shown that

the NOPK activity was not cOl'1'elated with NM23 protein levels nor with the anti­

metastatic effect33 Evidence of its metastasis suppressar function has since been

found to be on another level: firstly, on the post-transcriptionallevcl in the role of

autophosphorylation of the serine instead of histidine residue of the NM23-protein.

Thc acid-stable phosphoserine does cOl'1'elate with suppressioll of tumor metastatic

potential.33 Secondly, another presumed biochemical activity of the NM23 protein

(NM23-H2) has been found on transcription level. In vitro studies on the

transcriptional regulating function of NM23 have shOWll that the NM23-H2 isoform

exhibits a purine binding factor (PuF) which is able to bind to the oncogene MYC and

as a consequence is able to regulate its transcription35 MYC overexpression has been

shown to correlate with proliferation and metastatic potential in general and has been

demonstrated in uvealmelanoma as we1l 35 ,36 Whether the transeriptioual rcgulating

function of NM23-H2 acts separately from its NOPK activity and is cOl'1'elated with

I · . f' .. k 37 t lC antl-metastatIc unctlOl1 IS as yet un nown:

In condusion, NM23 protein expression using the NM23/NOP kinase A

antibody has limited prognostic value, but indicated a possible role of the NM 23 gene

in uveal melanoma. The exact biological role of the NM23-protein is still undear.

Further studies on post-transcriptional and transcription level of the NM23-protein are

necessary to investigate the relationship bet ween expression of NM23 and metastatic

potentialof uveal melanoma,

99

Page 100: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAI'TER 5 ---~ --------

REFERENCES

I. Egan KM, Scddon Jivl, Glynn RJ, Gragoudas ES, Albert D. Epidemiologie aspects of lIwal1l1elanoma.

SUf v Ophthalrnol. 1988;32:239-251.

2. lensen OA. ~vlalignant melanoma of tbc human uvea. Acta Ophlhalmol. 1988;60: 161- 1:32.

3. Rajpal S, [vloore R, Karnkousis CP. Survival in metastatic ocular 1l1elanoma. Cancer. 1993;52:334-336.

4. Steeg PS, Be\'ilacqua G, Kopper L, et al. Evidence for a novel gene associated with low tumor

metastatic potential. J Natl Cancer Ins\. 1988;80:200-204.

5. Slahl JA, Leone A, Rosengard AlvI, Porter L, King R, Steeg PS. Identifieation of a secotld human

J1m23 gene, nm23·H2. Cancer Res. 1991 ;51 :445-449.

6. Backer Hd, l\'lendola CE, Km'esdi I, et al. Chromosomallocali7ution alld nucleoside diphosphate

kinase activity of human metastasis-suppressor genes n1l123- land nm23-2, Oncogcne, 1993;8:497-502,

7. Gilles AM, Presecan E, Voniea A, Lascu I. Nlielcoside diphosphate kinase from hUll1an erythrocytes.

Struc!ural characterizatÎon of the two polypeptide chains responsible for heterogencity of the hexamerie

enzyme. J Biol Chem. 1991 ;266:8784-8789.

,.;:. Bevilacqua 0, Sobel ME, Uotta LA, Steeg PS. A~sociation oflow nm23 RNA levels in hllman

primary int11trating ductal breas! carcinom<ls with lymph node involven1Cnt and other histopathologieal

imlkators ofhigh Illetastatic potential. Caneer Res. 1989;498:5185-5190.

9. Royds lA, Rees RC, Stephenson Tl. nm23 - a metastasis suppressor gene? Editorial. Am 1 Pnthol.

1994;173:211-212.

10. Barnes R, i\'Ia.~ood S, Barker E, et al. Low nm23 protein expression in infiltrating ductal breast

enrcinomas correlates with reduced patient survival. Am 1 Pat hol. 1991; 139:245-250.

11. Hennessy C, Henry 1, .May F, Westly AB, Lenard T. Expression ofthe antimetastatic gene nm23 in

human breast cancer: An nssoeiation with good prognosis. 1 Natl Cancer Ins!. 1991 ;R3:281-285.

12. Kodera Y, Isobc KJ, Ymnauchi r .... I, et al. Exprcssion of nm23 H-I RNA levels in human gastric caneer

tissues. A negative correlation with nodnlll1etastasis. Cancer. 1994;73:259-265.

13. Yasuda D, Jgllehi H, Ikeda Y, et al. Possible association of nm23 gene expression and Ki-ras point

ll111tations with metastatic potential in hlllllan pancreatie cancer-derived celllines. Int 1 Oneol. 1993;3:641-644.

14. Nakayama H, Yaslli \V, Yokozaki H, Tahara E. Reducet! expressioll of nm23 is associated with

metastasis in human gastric carcinornas. Jpn 1 Canccr Res. 1993;84: 184-190.

15. Nakayama T, Ohtsufli A, Nakao K, et al. Expression in human hepatocellular carcilloma of nucleoside

diphosphate kinase, a h0l1101ogue of the nI1123 gene product. 1 Nat! Canccr Ins\. 1992;84: 1349-1354.

16. Fh.\rencs VA, Aamdal S, ~vIyklebost 0, j'vlaelandsmo mvI, Bntland 0S, Fodstad 0. Levels of Jlm23

messenger RNA in metastatÎC malignilllt melanomas: inverse correlation of disease progression.

Cancer Res. 1992;52:6088-6091.

17. Caligo l\·IA, Grammatico P, Cipollini 0, Varesco L, Del Porto 0, Bevilacqua G. A low nm23.HI gene

expression identifying high malignancy hllman mclanornas. Melanoma Res. 1994;4: 179-184.

18. Hamby CV, rvlendola CE, Potla L, Stafford G, Baeker Hvl. DifferentiaI exprcssion alld mutation of

Ni'vIE genes in autologOlIs cultured hllman I11elanoma cel1s with different metastatic potentiais.

Biochem Biophys Res COI11ll1. 1995;211:579-585.

19, Xeni L, Grob 11, Battyani Z, GOllvernet 1, Hassoun J, BOllerandi 11. Nm23 expression in metastasis of

100

Page 101: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NlvJ23 EXPRESSION IN UVEAI, /l.IELANO/l."I"A~ __ _

malignant Illelanoma is a predietive prognostie parameter corrc1ated with .~urvi\'a!. Br J Cancer.

1994;70: 1224-1228.

20. /I.'in D, Luyten GP~vf, LuiderTt'I, Niederkorn JY. Relationship between natural killer cell susceptibility

uml metastasis of hUll\a!l lIveal melanotl]a cells in a lllurine model. Invest Ophlhal Vis Sci.

1995;30:435-441.

21. LHyten GPi'vl, Naus NC, 1\'looy CM, et al. Establishment and characterizatioll of primar)' and metastatic Llveal

melanoma cclllines. Int J Cancer. 1990;06: I-H.

22. Kan-Mitchell J, Mltchell tviS, Rao N, Liggetl PE. Churacterizulion ofuveal melanoma cel1lines thaI grow as

xenogmfts in rabbit eyes. Invcst Ophthalmol Vis Sci. 1989;30:H29-834,

23. KsamIer BR, Rubsamen PE, Ol sen KR, COUSÏlls SW, Streilein J\\', Studies of tumor-infiltrating-Iymphocytes

from a hllmun choroidal melanoma. Invest Ophthalmol Vis Sci. 199 I ;32:3198-3208.

24. De WaanJ-Siehinga 1, Blom DJR, Griffioen M, ct al. Establishment aml characterization of an uveal­

melanoma cell line. Inl J Cancer. 1995;62: 155-161.

25. Luyten GP1J, t\'fooy CM, de Jong PTVM, Hoogeveen AT, Luider TM. A chicken embryo model to study Ihe

growth of hu man uveal melanollla. Biochem Biophys Res Comn1lln, 1993;192:22-29,

26. Keydar J, Chen L, Karhy S, et al. Establishment nnd dwracterisation of a cellline of hUlllan breast carcinorna

origin. Eur J Cancer, 1979;15:659-fi70,

27. Steeg PS, De la Rosa A, PlntO\\' U, MacDonuld NJ, Benedict M, Leone A. Nm23 and breast cam'cr

melastasis. Brcasl Cancer Rcs Treat, 1993;25: 175-187,

28. Cox DR. Regression modet.~ amllife tables (wilh discussion). J R Sial Soc B. 1972;34: 187-220,

29. Lascu ASI, Veron M, Anderson JJ, Wright C, Home CH\\', Angus B, NDP-K/nm23 expression in human

breast cancer in rclation 10 relapse, survival, and (lther prognostic factors: an illllllunohistochemical stud)'.

J Pathol. 1994;172:27-34.

30, Saslre-Garau X, Lacombe i\'lL, Jouve lvi, Veron 1\'1, l\'lagde1enat H. Nucleoside diphosphate kinase /nm23

expression in breast cancer: lack of correlation with Iymphe-node melastasis, Int J CanccL 1992;50:533-538.

31, Ibilat N, Kcim DR, !vlelhem Rl', et al, High levels of p19/nm23 protein in neuroblastoma are assodated with

advunced stage dise<lse and with N-myc gene amplification, J crin 1I1\'esl. 1991;88:341-345.

32, Enge! M, Theisinger B, Seib T, Seitz HH, Zang KD, Welter OS, High levels of nm23-HI and nm23-

H2 messenger RNA in hu man squamous-cell curciJloma are associated with paar diffcrentÎation and

advanced lumor stages. Int J Cancer. 1993;55:375-379,

33. Macdonald rvu, De La Rosa A, Benedict MA, Freije JivlP, Krulsch H, Steeg ps. A serinc phosphorylation of

nm23, and not lts nucleoside diphosphate kinase activit)', correlales with suppression of tumor metastatic

potential. J Biol Chem. 1993;268:25780-25789.

34, tvlu D, Lu)'ten GPM, Luider Th!VI, Jager 1U, Nierderkarn JY, Associalion between nm23-H 1 gene

expression and llletasl<\sis of hurnan IIveal melanoma in an anima! model. (l\'lanuscript in preparation)

35, Postel EI-I, Berberich SJ, Flint J, Ferrone CA. Human C-myc transcription factor Pul' identified as I1rn23-112

nucleoside diphosphate kinase, a candidate supprcssor of tllmour melaslasis. Science, 1993;261 :478-480,

36, Moay CM, Luyten GPI\-1, de Jong PTVl\'l, et al. An illlIl1l1nohistochemical and prognostic analysis of

apoptosis anti proliferation in uveal melunoIl1a. Am J Path, 1995~147: 1097-1104.

37, Postel EH, Ferrone CA. Nucleoside diphosphate kinase cnzyme activity of mn23_h2/Puf is not

rClJuired for its DNA binding and in vitro transcription functions, J Biol Chcm, 1994;269:8627-8630.

JOl

Page 102: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

J02

CHAPTER 6

ESTABLISHMENT AND CHARACTERIZATION OF PRIMARY AND

METASTATIC UVEAL MELANOMA CELL LINES.

Gregorius P.M. Luyten, Nicole C. Naus, Camelia M. Mooy,

Anne Hagemcijer, Junc Kan-Mitchell, Ellen van Drunen,

Vojislav Vuzevski, Paulus T.Y.M. de Jong and Theo M. Luider

International Journal Cancer, 1996;66:1-8

Page 103: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_______ -'U"-''-IEAL 1fELAN01IA CELL U,-,N-'oE",S _______ _

ABSTRACT

Wc report on the establishment and charaeterization of two primary (EOM-3, EOM-

29) and threc metastatic uvealmelanoma celllincs (OMM-l, OMM-2, OMM-3),

and further eytogenetie eharacterization of a prcviously described primary uveal

melanoma eell line (OCM- I). Only a few long-term growing primary uveal

melanollla eell lines have as yet been established, while of metastatic uveal

melanoma eelllincs we have found no deseriptions. The morphology of the ill vi/ra

cultured eells varicd from spindIe to epithelioid. The celllines werc charaeterized

by illlmunocytochemistry, electron mieroseopy and cytogenetical analysis. The

relative growth ra te was detennined by bromodeoxyuridine (BrdU) incorporation.

The melanoeytic origin of the eell lines was cIetermined by the positive staining

with antibodies identifying lllclanoma associated antigens. Melanosomes and pre­

melanosomes were indeed observed by electron microseopy in all eclllines. The

stem eell karyotype was fOlllld to be normal in three eelllines (EOM-29, OMM-2,

OMM-3) and abnorInal in thrce ot hers (EOM-3, OCM-I, OMM-I) showing a net

loss of chromosome 6. The OCM -I and thc OMM-l celllines even dcmonstrated a

large amount of structural chromosomal aberrations, the farmer being neal'

tetraploid and the latter triploiel. The EOM-29 eell line, cultured from an ciliary

body lllclanoma, did not show thc previously described chromosome 3 and 8

abnonnalities.

INTRODUCTION

Uveallnelanoma is thc most COilunon primary intraocular tUlllOUl' in adults. I These

tumours spread haematogeneously and preferentially to the liver. A 50 % overall

incidence of metastasis occurs within 15 years aftel' initial treatmenl by enucleation

or radiotherapy of the tumour-eontaining eye? After clinical diagnosis of hepatic

metastases, median survival is extrelnely pOOl': between t wo anel se ven months?

The biological and molccular properties of the sueeessive steps involved in uveal

melanoma progression and metastasis eould be studied in detail once established

uvealmelanoma celllines were available. A limitcd number of eell Iines obtained

103

Page 104: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_____ -,C,-,H,-A I'T E R 6

from primary llveall11eJanomH have incleed been described,3-6 but not from

11lctastatic uvcallnclanoma.

Tn this study, we report on the successful establishment of two primary

(EOM-3, EOM-29) anel three metastatic uveal melanoma celllines (OMM-I,

OMM-2, OMM-3) in our institute. These celllines were characterized by

immunocytochemistry, light and electron microscopy, and cytogenetic analyses.

The relative growth rate was detennineel by bromoeleoxyurielin (BreiD)

incorporation. Tn addition, a previously described primary uveallnelanoma eell line,

OCM-l,3 was further characterizeel by extended karyotypic analysis.

MATERlAL AND METHODS

1\ullour nUlterial

After enucleation of the uvealmelanoma containing eye, part of the ocular

tumour was taken fol' cell culture. The eye was further histologically examined to

COnfil'lll the diagnosis. When, during follow-up of the patient, metastatic elisease

was suspected, a biopsy was taken for diagnosis anel for cell culture. All patients

had given full informed consent. Between January 1992 and May 1993, cells of the

primary tUl110l\l'S of 16 patients anel of the llletastatic tUlllOUI'S of four patients were

processed for culture. The patients with prilllary uveaimelanollla ranged in age

from 22 to 87 (mean: 56.8); nine were fenwie anel seven \Vere male; eight patients

had a choroidal anel cight patients had a ciliary body melanoma; three patients had a

spindle-cell tumour, nine had a mixeel-ceU tumour and four had an epithelioid-cell

tUlllOur; tumour diameter ranged from 3 to 22 mm (mean: 11.6 111m). The patients

\Vere followed until May 1995; two patients died of metastatic disease and one is

alive, 6 montlls after detection of a metastatic lesion in the subcutis. From all four

patients with lllctastatic disease tU1110ur 1l1aterial was ubtained from sub-cutaneous

lesions. Prom anc patient lhe primary as weIl as thc mctastatic tumour were

available.

104

Page 105: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

UVEAL t ... IELANOl\IA CELL L1NES

Culturing meihoeIs

The tumom material was proeessed and cultured within one to three hours

aftel' dissection as described by Luyten el a/7 In brief, the tllmour-eontaining eye or

the mctastatic les ion was dissected under steriIe conditions. A full thickness biopsy

was taken and transported in Dulbecco's-modified-Eagle's meelium supplemcnteel

with 10% fetal calf serum, and 1% penieillin/streptomycin (full DMEM). The

tumour matcrialwas suspended with a small pair of scissors anel strained through ti

linen eloth by continuous irrigation with full DMEM. The resulting sus pension of

single cells and to alesser extent ol' small c1U1nps of tumour eells, was then washed

two or three times with full DMEM. This washing was repeated depending on the

amount of pigment. The cells were seeded in a culture Ilask (Fa1con plastic T-30)

with 5 mI full DMEM at 5% C02 (37°C). Aftel' one week the medium was renewed

while the non-attached eells \Vere diseardcd. Subseguently, the culture medium was

ehanged twice a week. At reaching confluence, the eells were e1etached with 0.05%

trypsin and 0.5 mM Na2 EDTA and subseguently subcultured. Depending on the

growth rate the cultures were passaged at 1:2 to I: 10 dilutions.

BrelU illcorporation

Ta measure the logarithmic growth ratc of the eells in culture, the cells wcre

ineubated in culture f1asks for six hours with I mM BrdU (Boehringer Mannheim,

Almere, The Netherlands). Aftel' being washed and detaehcd, the eells were

centrifuged onto object glasses, fixed in 2 M HCL at 37°C for 30 minutes and then

incubatcd in 0.1 M borate buffer pH 8.5 at room temperaturc 1'01' five minutes. The

eells were subseguently washed twice with PBS and incllbated 1'01' 30 mimltes with a

peroxidase-conjugateel immunoglobulin of a mouse monoclonal antibody specilïc for

BrelU (Boehringer Mannheim; dilution I :33) in 1% BSAIPBS. Aftel' washing with

PBS again, the incorporated BrdU lVas visualized with 3'3 diaminobenzidine and

ureum hydroxide (Sigma, St. Louis, MO). The percentage of BrelU incorporated

cells could then be eounted in one low power field Cl OX) lIsing an eye piece grid.

105

Page 106: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 6

IlIllnullocytochemistry

The cells we re cultured 011 a coverslip, washeu twice IVith PBS at 37°C and

subse'luently fixed in eold acetone at -20°C for 10 minutes. EndogenOlls peroxidase

was inhibited by incubation in 0.3% hydrogen peroxide/methanol for 15 minutes.

Subsequently the coverslip was incubated in 2% fetal calf serum/PBS for five

minutes. The primary antibodies, S100 (Dako, Glostrup, Denmark), NKI-C3

(Dako), HMB-45 (Biogenex, San Ramou, CA) and HNK-I (Becton Dickinson, San

Jose, CA) were ineubated at dilutions of 1:600, 1 :20, 1 :50, I: 10, respeetively for 1

hour. For SI 00, swine anti-rabbit peroxidase-conjugated immunoglobulin was used

as a second step alltibody (Dako; dilution I: 100). For the other three antibodies

rabbit anti-mouse peroxidase-conjugated immunoglobulin (Dako; dilution of I: 100)

was used. Both seeondary antibodies were ineubated for one hour. The activity of

the peroxidase was then visualized using tablets of 3,3' diaminobenzidine HCL and

hydrogen ureaperoxide (Sigma). Finally, the cells were counterstained with

haematoxylin for 10 seconels. Between eaeh of the incubation steps the covers lip

was rinsed with 0.1 % TlVeen-20IPBS.

Alltibodies

The SI 00 anti body recognizes an acidic intracellular Ca ++ binding protein

and is a sensitive but non-specific marker for lllelanoma cells,8 Monoclonal

antibody NKI-C3 anel HMB-45 binds both a cytoplasmic antigen produced by fetal

melanoeytes anel melanoma cells of adults. 9,IOThe monoclonal antibody HNK-I

recognizes a family of eell ad hes ion molecules (CD 57), migrating nemal erest cells

and a series of neuml crest derivatives including tmllours of neural crest origin, I 1

MHC class I expressioll

Sub-confluent eell cultures were detached with trypsinlEDTA, after whieh

cytocentrifuge slides eould be prepared. The major histoeompatibility complex

(MHC) class I expression was dertermined using the standm'd NIH micro­

cytocytoxity test with loeally obtained reagents.

106

Page 107: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ ~U-,V,-,E,-,' A"I. ~1 E L A N 01\[ A C -'~ L L L I NES

Electrollmicroscopy

Aftel' detaehment by use of trypsinfEDTA, the cultured eells \Vere

centrifugeel and fixcel in 3% glutaralelehyde at room temperature for 30 minutes,

Then they \Vere post-osmificateel for 30 minutes in 1% Os04, anel subsequently

embeeleled in Bpon, Ultrathin seetions \Vere prepareel and eounterstained with uranyl

acctate and lead ciO'ate.

Cytogenetic allalysis

In oreler to obtain metaphases, cultureel eells in the logarithmie phase of

proli reration \Vere treateel \Vith eolcemid for one to three hours, Af ter trypsinization,

the eells \Vere subjecteel to a hypotonie (0,075 M KCL) solution, Finally the eells

\Vere gradually fixeel with colel methanol/ace tic acid (3: 1 v/v), Air-dried slieles \Vere

baneled by the re verse methoc!.

To ielentify the marker chromosomes, fluorescenee in situ hybridization

(FISH) was useel, SIkies \Vere then hybrielizeelwith chromosome-specific

libraries, 12 kinelly provieleel by DI'. J, Gray, The results presented conform to the

reeommenelations of the International System for Human Cytogenetic

Nomenclature (ISCN), 13

107

Page 108: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 6

RESULTS

Prilllary aIHllIletastatic uveallllchmollla celllillcs

Out of the 16 cultures of the primary uveal melanomas, two celllines coulel be

establisheel successfully. The remaining cell cultures stoppoel grO\ving after being

passaged two to !'ive times. Of the cultures of the metastatic uvealmelanomas, three

out of four resulteel in a stabIe growing cellline. The eell cultures from one patient,

from whom both the primary anti eorresponeling metastatic tumour were available,

were not suecessrul. The patient data anel tumour data from the establisheel cclllines

are summarized in TABLE I.

TABLE 1 PATIENT AND TUMOUR DATA OF SUCCESSFULLY ESTABLISHED

UVEAL MELANOMA CELL LINES

CEL!' PATIENT

LINE SEX/AGE SURVIVAL(MO)

, GCM-I F /n.a. Tl.U.

EOM-3* tv1/62 29

Eotvl-29* ivl/87 14

Ol\'livl-I t tvI/74 343'

Ol\Hvr-2t Mln 40'

Ol\HvI-3 t Mln 41'

Primury uycul mclanoma cellline

t Metastatic uvealmelanomu cellline

* Tumour related ueath

TUi\IOUR

LOCATION

Posterior

Posterior

eiliur)' body

Subcutis

Subculis

SUbClltis

Abbrcviations: F, fcmale; .M, male; IJ.U. =:: not avt'lilable

108

CELLLINE

HlSTOLOGY MOHPHOLOGY(Li\I)

Spindie B l'vIixed

Mixed Epithelioid

Epithclioid Spindie

i'vlixed iVlixed

Epithelioid Spindie

ivIixed Spindie

PASSAGE

>50

20

15

42

21

19

Page 109: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_______ U'--'-V"'C"'AL MELANOl\lA CELL LINES

The primary uvealmelanoma ceU line OCM-l, described by Kan-MitcheU and

associates,3 has been growing for more than seven years without morphological

changes (FlGURE IA), The EOM-3 ceU line was established from a mixed cell type

primary uvealmelanoma. During the nrst six passages, the cultured cells had a spindIe

ceU type appearance. However, a small clone of epithelioid ceUs started to grow and

subsequently overgrew the rest of the tumour ceUs. The epithelioid-ceU culture

llllderwent no further morphological changes and has been growing now for two years

(FIGURE IB). EOM-29 was established from a large extra-oCldar extension of a

ciliary body melanoma (epithelioid ceU type) and could be passaged once a weck.

Gradually the passage time increased with a e1ecreasing proliferation rate. Aftel'

passage ten, the ceUs hael a spindIe like appearance in culture (FIGURE IC) anel

continueel to grow without demon st rating changes in their morphology.

A subcutaneous metastatic lesion (mixed ceU type) was excised from a uveal

melanoma patient (OMM-l), who hael undergone an enuc1eation of his primary

tumour 29 years eadier. The primary culture of this metastatic les ion and the first ten

passages consisted of adherent and non-adherent ceUs. Graelually more cells became

adherent, showing a heterogeneous morphology of spindIe anel epithelioid cells. This

morphological heterogeneous spectrum remained the same over 2.5 years (passage 42)

(FIGURE ID). The primary culture (OMM-2) of a sub-cutaneous metastatic lesion

(epithelioiel-cell type) had a spinelle cell pattern in cell culture, anel did not show

morphological changes for two years. The cells have been growing slowly but

continuously with a passage time of one month (FlGURE IE). The most recently

establisheelmetastatic cellline (OMM-3) was cultured from a large, partly amelanotic

sub-eutaneOlls metastatic lesion (mixed-cell type) with a spindIe cell appearance in

cell culture. This cellline has now been growing tor more than one year and na

morphological changes have been observed (passage 19) (FIGURE 1 F).

109

Page 110: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

110

CHAPTER 6

FIGURE 1 Light mieroseopy (seaJe har, 50 111111).

Iif ,~

Yf1

"'->,

'4"

lP "1

,< j

(A) The OCM-l eell Jine has smull and large rounded eells, with onc or more nuclei, growing lIpon a stellate and eJongated eelllayer with dendritie extensiolls;

(B) the EOM-3 cell Jinc shows a monolayer of pillmb polyhedral anapJastie, epithelioid-cells with abundant cytoplasm, large nuclei and prominent nucleoli;

Page 111: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

UVEAL lIELANOr-.IA CELL LINES. ___________ _

(C) The EOM-29 celiline forms spindie like palterns witlt large round nuclei;

(I) The OMM-1 cellline: cluster of large round cells (mono- and multi-nucleated), irregularly shaped stellate and elongatcd cells. Very large cells with abundant cytoplaslll and cytoplasmic inclusions were obscrvcd;

1/1

Page 112: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

112

CHAPTER 6

(E) The OMM-2 cellline showed a mOl1olayer eharaeterized by predominantly elongaled eells with large nuclei, abundant cytoplasm alld long dendritic extellsiolls, forming a bundle­like, spindIe cell pattern;

(F) OMM-3 cellline has spindle-cell appearance with dendritic extensiolls; Thc nuclei are large with abundant cytoplasm and prominent nucleoli.

Page 113: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ -,U,,-\,-,!I~,AL J\1ELANOJ\-1A CELL LINES

Electron microscopy, inullunocytochcmistry

Electron microscopical morphology of the six cell lines is shown in F1GURE

2A-F, All figures are representative for the cultureel cells examineel anel contained

melanosomes anel pre-melanosomes (TABLE 2). The immunocytochemical staining of

the five celllines with NKI-C3, HMB-45, S 100, HNK-I anel BrelU incorporation is

summarized in TABLE 2. The NKT-C3 staineel all five celllines; expression of the

other three markers appeared to be facultative. The celllines EOM-3 and OMM-l

showeel a relatively high BrdU incorporation, thus showing a high frcquency of cell

elivision. The more slowly growing celllines EOM-29, OMM-2 anel OMM-3 hael

indeeel a lower BrdU incorporation. The primary celllines OCM-I anel EOM-29 were

fOllllel to express MHC-c1ass-I molecules, whereas in the EOM-3 cellline, na

detectable MHC-c1ass-I expression was observed. All three metastatic celllines were

fOllllel to express MHC-c1ass-I antigens.

TABLE2 CHARACTERISTICS OF THE ESTABLISHED CELL LlNES

CELL ELECTRON II\1I\fUNOCYTOCHEi\lISTRY"'

LINE ~IIcnOSCOpy

l\Iehmosomes NKIIc3 HMn·45 8·100 HNK-l I)"pe

OCrvl-1 J I(Il) 1003 1003 [003 0

EOM·3 I 100 (pIS) 100 (plg) o (pIS) 100 (pIS)

EOM·29 111. IV 100 (pS) o (pS) o (pS) o (pS)

üM!vl-l 111. IV (Il) 100 (p26) 100 (p26) 100 (p26) 0.51 (p2?)

0rvHvl-2 Il. 111. IV 100 (pI?) o (pI?) o (pI?) g (pI?)

OMM-3 Il, 111 IDO (pS) o (pS) o (pS) 4 (pS)

Percentage cells which bind the indicated antibody (passage l1umbcr) ;\ Kun-Mitchell et alo, 1989 t No detectabJc MHC c1ass lexpression t Initially 50 percent of (he cclls expressed HNK-l antigens.

Abbreviations: Il.d., not done

UrdU MCH cla~ I expressioll on (eli lint'

n.d. A24,A2g

24 (p20) t

<I (pI5) A3,A30

49 (p42) A2.AII.B62

<I (p2J) AI

2 (pl3) AI,AII

liJ

Page 114: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

114

___ eH A PT E R,--,6oe, __

FIGURE 2 - Electron microscopy.

(A) OCl\1-1: part of an lIvea! melunomu cell. Tbe cytoplasm cOl1tains a few Illclanosomcs (M) and pre-melullosomcs type I (Ptvl) and sporadic type 11 (seale bar, .350 ~tm)

Page 115: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ ~lJ~\~'EAL J\IELANO~lA CELL LINE,S'--__ _

(B) EOM-~'; very primitive, immature ceIl with a large nucleus and two prominent nucleoli. Only a few pre-melanosomcs type I (PM) afC present (scalc har, 1 ~lm)

115

Page 116: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

116

____ CHAPTER 6

(C) EUM-29: he,wily pigmented celI. ... with abundant cytoplasm in whieh there urc man)' melallosomes Uv!) auel prc-meianosoIllcs type Hl (PM) anti Iysosomal structures (L) (sc ale bar, 1 ~lll1)

Page 117: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ -"U'-\,,'"'EAL llELANOJ\IA CELt LINES

(D) OiV1[\:I-l: two melanomu cells contain in thc cytoplasl11 H large amount of osmiophylic (Jense) I11eklllOSOmes (M) ancl pre-mclunosomcs type III and IV, anel unly a few type n (PM), Note the presencc oflysosomcs (L) (se ale har, .583 ~tl11)

117

Page 118: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

118

CHAI'TER 6 ------

PM

(E) OJ\IiH-2: melanol1w eell with lUl aberrant cylopJasm in whieh densc fusiform melanosomes (M) and prc-melanosomcs (types 11, In :md IV) (PM) arc obserycd (se ale har, 1 ~llll)

Page 119: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

UVEAL MELANOi\fA CE_LL LINES

(F) 0.1\11\'1-3: spindie melanocytic cells. The cytoplasm contain large Uillount of rough cndoplasmatic retieulum and pre-Illelanosomcs types II and III (Piv!); a rew lysosomcs (L) are also present (seaJe bar, I )Jm).

119

Page 120: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CELL LINE

______ CHAPTER 6 _____ _ --------

Cytogenetic analysis

The karyotypes of the six eelllines are deseribed in Table 3, The karyotypes

wem obtained from JO to 25 metaphases frOI11 eaeh eellline at different passages,

Cells ti"Oll1 celilinc EOM,29 and OMM,3 displayeel a norm al karyotype, The stemline

karyotype of OCM, I was found to be tetraploid anel to have l11any structural

chrol11osomal aberrations (FIGURE 3A), These, among other changes, have resulted

in 11 net loss of 3'1 anel of the distal part of chromosome 8 (8q24'qter); monosomy 6

was observed in 60% of the metaphases of OCM, I. The stemline karyotype of EOM,3

was fOlllld to be pseudodiploid with numerical changes: a loss of Y, monosomy 6,

trisomy 5 and 18 (FIGURE 3B), The stemline karyotype of eellline OMM, I was ne ar

triploid with lllany struetural and numerical changes, a net gain of chrolllosome 3, 7,

12 anel 20, and loss of 4, 8p, 9, 11,15, 17,21, among others (FIGURE 3C), Cellline

OMM-2 showcd no clonal aberrutions, hut SOlne non-clona1 re arrangement, sn eh as a

l1l0nosomy 6 in 2 cells and an unbalanced translocation (14;22) in one cel!,

TABLE3 STAINING PATTERN IN THE METASTASES

PASSAGE NOOF MODAL KARYOTYPE

~lETAI'HASES CI!ROMOSOMES

ANALYZED

OCM-I >50 14 90-97 94,<4>,XX,-X,-X,dd( I )(1'21 )[40% ],dec(2)1( I ;2)(1'3 1 ;q37),add(3)(q 12),

+del(3)(q 13q23),del(4 )('112)[ 60%],-4[40% J,add(5)(p 11 )[40%]),add(5)('113),

-6[6U% ],add(7)(q35),del(7)('133),-7 ,add(S)(q24)x2,-9,-11 ,inv( 11 )(p 15q24),

dele 12)(p 12)x3,-12,atld{ 13)(q31 ),-I30r t( 13q; 14q)l60%],-15,

add( 16)(pI3),-16[40%],-18[60%J,-19[60%],add(20)(q I 3)x2 or x3,-20,

aud(2l )(q21),-21,-21,del(22)(q 11 )x2,iso(22q)x2,+6-12 markers

EOlvl-3 10 22 46 46,X,-Y,+5,-6,+18

EOivf-29 5 10 46 46,XY

W ... Hvl-] 8-10 25 60·68 64,<3>,X,-X,-Y,tIer( I )I( I ;3)(p31;p 13),oer(2)1(2;5)(q32;'1 14?)x2,

1(2; 10)('132;'125),+ 3[50%],-4,oel(5)(q 13'124), + 7,ndd(R)(p 11),,9,-11,

+ 12,itw( 13)('1 12q34),-15,add( 16)(p 12),,17,

del( 19)(p 12),+20,add(21)(p 13),-21,+ 1-3 markers

OMM-2 11-14 19 46 46,XYfll]/44-45, with random loss of 1 or2 chromosomcs,

[8 of which 2x -6]

OMM-3 1,4 21 46 46,XY

720

Page 121: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

UVEAL I\.IELANOMA CELL LINES

FIGUHE 3 - Karyotypes of the celllines.

(A) OCM-I

(TI) EOM-3,

121

Page 122: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_CIIAI'TER 6 ______ . ___ _

(C) OMM-I

DISCUSSION

Culturing of hUlllan uveallllelanollla eelIs was first atiempted in 1929 by Kirby, 14 but

until the present date only a few eelilines from primary melanomas,3-6, 15 and no eell

lines tium melastatic uvealmelanomas have been described, The relative paucity of

established uveal melanoma celIlines can be eXplained partly by the low incidenee

(I: 100,000) of uvealmelanomu und by cliffieulties in the culturing of primary uveal

melanoma tissue, 1 Lack of therapeut ie mOllalities for patients with hepatic metastases

has led to a high mortality among them within a very short survival time, and hence

provides a problem with respect to the availability of fresh metaslalie tissue as weiL In

this study we were able to establish two primary uveal lllelanollla celllines and thrce

metastatic uveallllelanoma eell lines, Although a lillliled overall Ilumber of tUlllours

eould be cultured at our institute, the sueeess rate in the establishment of long-term

122

Page 123: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

UVEAL_ ~IELANOl\lA CELL LIN(~S _____ _

gl"O\ving cclllines was relatively high for metastatic uveal melanomas (3/4) as

compareel with primary uveal melanoma (2/16). Mctastatic lesions originate from

selected cells that havc proved to be capable of growing in seconelary organs. This

could bc the reason for the high success ratc in cuIturing mctastatic celllines. The

present availability of primary and metastatic uveal melanoma ceJilines make it

possible lo study the dil'ferenecs between primat')' anel 111ctaslatie uveal me1anoma in

vitro.

To ascertain the l1lelanocytic Ol'igin or the cultureel cells, the cells were

il1l111Unocytochemically studieel by the use or melanoma-associateel antiboelies. SI 00,

NKI/C3 and HMB-45 are known to be powerfull1larkers rol' differcntiating tumours

of melanocytic lineage from other unaplastic tUlllOurS. In addition, electron­

mieroscopc studies revealcd melanosomes ano pre-melanosomes in all the celllines

elescribed. The celilines showed a consielerable variation in passage time as measured

by BrdU incorporation. The celilines EOM-3 anel OMM-I hael rclatively high BrelU

incorporation and were shown to have high gl'owth rates with short passage times in

culture. Thc celilines EOM-29, OMM-2 and OMM-3 had low levels of BrdU

incorporation ano were also more elifficult to maintain in culture. However, the eeU

lincs OMM-2 and OMM-3 were stabic over a long peri oei of culturc without showing

any morphological changes. The cellline EOM-29 has been in culture for a relatively

short pedod, showing no morphologieal changes, but immortality is not yet

established. No elil'ference in growth rate bet ween primary and llletastatic celilines

coulel be demonstrated by thc BrdU incorporation.

Although epithelioiel cells have of ten been regareled as the most malignant cell

type, three out of six celilines in our study, were of the spinelle cell type (EOM-29,

OMM-2, OMM-3). This in contrast to the histological cell type of the original tumour

of the patient. This can probably be explaineel by the selcctive outgrowth of certain

tumour cells in a I'avolll'abie culture meelium such that uscel by us. In casc of the cell

line EOM-3, the spindie cells of the fresh tumolll' culturc were graelually overgrown by

undifferentiatcel, anaplastic cclls at passage five. After that time the uniform

lllorphological cell type of EOM-3 rel1laineelunaltereel. Wh ether the phenotypical

change of thc cell line EOM-3 is accompanied by genctical instabilitics, reflecting one

or more steps in tUJl10lJl' progression, or by simple selection is not knowIl. Both cell

123

Page 124: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_______ ~C~H A "TER 6

lines OCM-l anel OMM-l consist of a heterogcnous population of cells lVith a high

growth mte, showing signs of selection of one of the morphological cell types present

in thc original tumour.

The normal expression of MHC class T molecules in five of the six celllines

indieates a possibility of a MHC class T restricted T-cell response in metastatic uveal

melanoma patients. In the EOM-3 cellline, MHC-c1ass-I expres sion is elown-regulateel

and, as a result of this, increaseel susceptibility to natural-killer-cell activity has been

elemonstrated. 16 These data indieatc th at enhancement of the T-cell response by the

use of allogenic irradiateel celllines might be one of the most promising ways to treat

patient~ with metastatic disease.

Tn cytogenetical analyses of p1'Îmary uvcal melanoma, the most frequently

described chromosomal abnormalities founel are: monosomy of chromosome 3, net

loss of genetic material of chromosome 6q anel Sp, and gain of 6p and 8q. 17,18 11 has

been suggesteel th at monosomy 3 anel the isochromosome 8g are associated with

ciliary-body mclanoma anel pOOI' prognosis. 1Y Cytogenetic analysis of Dur eelllines,

sIlOweel a normal karyotype in three celllines anel karyotypic abnormalities in the

othcr three, with partial loss of chromosome 6 among multiple ot her aberrations.

OCM-l anel OMM-l were near-tetraploid and t1'Îploid, respectively. In these celllines,

partialloss of 3q and Sp were seen, as a consequence of structural changes. In contrast

to other results on cytogenetical abnonnalities in ciliary-body melanoma, our cellline

EOM-29, whieh was eleriveel from a massive extrasc1eral growth of a ciliary-boely

melanoma, elid not elemonstrate the chromosomc-3 anel -8 abnormalities. We have

fOUllel hOlVevcr no systematic elifferenees in the karyotypes of the cultured cells of

primary versus metastatic tumollrs. The karyotyping was assesseel after establishment

of the celllines and thus, some chromosomal abnormalities coulel be the conseguence

Of;11 vitro c10nal seleclion and progression, instead of a representation of the situation

of the tumour ill vivo, Tn future stuelies, cytogenetic analysis of primary uveal

melanoma cell cultures will hopefully proviele more information on the karyotype of

the tumour cells. Furthellllore, complcmentary stuelies using mol ecu lar cytogenetic

approaches anel comparative genomic hybridization could give more detailec1

infonnation on chrOlllosomal abnormalities in fresh tumoUf and celllines in culture?O

124

Page 125: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

-_._-- UVEAL 1\1ELANOl\-IA CELL LJNE,cS'--__ _

REFERENCES

I. Egan KI\'I, Seuuon Jiv!, Glynn Rl, Gragoudas ES, Albcrt Dlv!. Epidemiologie aspects of uveal melanomu.

Sm\' Ophthalmol. 1988;32:239-250.

2. Kuth R, HaYllngs J, Bornfeld N, Sauerwein W, HöÜken K, Sccber S. Prognosis aod trealmen! of disseminated

lI\'eal melanol11a. Cancer. 1993;72:22 [9-2223.

3. Kan-r ... litchell J, i'vlitellell ~JS, Raa N, Liggett PE. Characterization of uveal melanom<l celllines thaI grow as

xenogr<ll'ts in rabbi! eyes. Invest Ophthalmol Vis sci. 19R9;30:S29-834.

4. Auber! C, Rouge F, Reillaudoll lvI, i\'Ietge P. Establishment am[ characteri7ation of hu man Ol'ular melanoma

ccUlitJes.lnt J Cancer. 1993;54:7R4-792.

S. l ... Iassarclli R, BeJkhou R, Dunel-Erh S, Chelm'ieT C, Abbé J-C, Pignol J-P, i'vfoutaouakkil ~I'f, Sahel

J. Caractérisation cellulaires de mélanome uyéal humain. C R AC.IU Sci Paris. 1~94;3l7:25-33.

6. De Waard-Sicbinga I, Blom DJR, Griffioen J'vl, Schrier PI, Hoogendoorn E, Beverstock G, Danen EHJ, Jager

!\U. Establishment anti characterization of an uveal-mclanom<l cell line. Int J Ctlncer. 1~95;62: 155-161.

7. Luyten GPM, ~vlooy Ci'vl, De Jong PTVM, Hoogeveen AT, Luider TM",L A chicken embryo model to study

growth of uvealmelanoma. Biochem Biophys Res Comm. 1 ~~3;1 ~2:22-29.

8. !vloore BW. A ~oluble protein characteristic of the ne-rvous system. Biochem Biuphys Res Comm. 965; 19,739-744.

9. Vennegoor C, Calafat J, Hageman Ph, Van BUÎtenen P, Jansen H, Kolk A, Rilmke Ph. Biochemical

characterization anti celllliar localization of a fonnnlin resistent lllelanoma-associateJ antigen rcacting

with mOn0c101lal antibody NKIIC3. Int J Cancer. 1985;85:287-295.

10. Gown Ai'v!, Vogel AM, Hoak 0, Gough F, i'dcNutt MA. ivlonoc1onal antibO(lies specifïc tor melanucytic

tumors distinguish .~ubpopulations of melanocytes. Am J Pathol. 1986; 123: 195-203.

11. Abu T, Ba!ch CM. A diiJerentiation antigen ofhuman NK and K eells identifled by a monoclonal anti body

HNK-1. J Immunol. 1981;127:1024-1029.

12. Pinkel 0, Landegcnt Jj Collitri C, Fuscoe J, Scgraves R, Lucas J, Gray J. Fluorescence in situ hybrirJization

with human chromosome-spcdl1c libraries: DetedÎon of trisomy 21 and translocations of chromosume 4.

Proc Natl Acad Sci USA. 1988;85:9138-9142.

13. Mitelman F, (ed). Guidelines for Canccr Cytogenetics, Supplement to an International System for

Hllman Cytogenetic Nomenclature. (JSCN) Karger, Basel (1991).

14. Kirby DB. Tissue culture in ophthnlmic research. Trans Am Ophthalmol Soc. 1929;27:334-383.

15. Albcrt DM, Ruzw MA, I\'Iclaughlin ~vfA, Robinson NL, eraf! JL, Epstein J. Establishment of eell

lincs uf uveal melanoma. Illvcst Ophthulillol Vis Sci. 1984;25: 1284-1299.

16. r-,'Ia D, Luytcll GP!>"1, Luider TM, Niederkorn n~ Relationschip between natural killer susceptibility and

metastasis uf hUlllall uveal melnaollHl cells in a Jllurine model. Invcst Opthal Vis Sci. 1995;36,435-441.

17. Horsman DE, White VA. Cytogcnctic analysis of U\'eal melanollm; constituent occurrenee of lllonosomy 3

and trisomy 8q. Cancer. 1993;71,811-819.

18. Singh AD, Boghosian-Sell L, War)' KK, Shields CL, De Potter P, Donoso LA, Shie!ds JA, Cannizzaro L.

Cytogenetic l1ndings in primary uveal melanoma. Cancer Genet Cytogcnet. 1994;72: I09-115.

19. Sistey K, Rennie 10, Cottam OW, Potter AM, Potter CW, Rees Re. Cytogenetic I1ndings in six postcrior uve,tl

melanomas: Involvement ofchromosomes 3, 6 and 8. Genes Chrom Cancer. 1990;2:205-209.

20. Speicher tvlR, Prescher 0, Ou y.,·Ianoir S, Jauch A Horsthemke B, Bornfeld N, Becher R, Cremer T.

Chromosomal gains and lusses in uveal melallOlllas detected by comparative genomic hybridization.

C,lIlccr Res. 1994;54:3817-3823.

/25

Page 126: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

/26

CHAPTER 7

EXPRESSION OF MAGE, GP100, AND tyrosillase IN UVEAL MELANOMA CELL LINES

Gregorills P.M. LlIyten, Comelia W. van der Spek, Tneke Brand,

Kees Sintnicolaas, Itte cle Waaru·Siebinga, Paulus TY.M. cle Jong,

Peter T. Schrier, Thco M. Luider

Submittecl for publication

Page 127: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

EXPI{ESSION OF ,\fACit', GPIOO AND n'IUJSINASE

ABSTRACT

Plll'pOse: 1b evaluate the expression ofthe hllnulll genes MAGEI, -2, -3, GPIOO anel

t.Vlvsinose in uveal melanoma.

Method: mRNA expression of MAGEI, -2, -3, GP 100 anel tyrosillase, and thc MHC

class Iexpression were detennined in five primary and three metastatic uveal

melanoma celllines.

Results: The expression of the examined genes was heterogenously in primary and

metastatic celllines. The celilines OCM-l anel OMM-l expresse el MA GEI, -2 anel-3,

whereas EOM-3, MEL202, 92-lunel OMM-3 were negative. GP IDa was expressed in

all cellline" anel tyrosillase in all but th ree (EOM-29, OMM-2 anel OMM-3). Except

for EOM-3, all celilines expresseel MHC class I molecules.

Conclusion: The tested lnelanoma-associated antigens can be expressed on uveal

melanoma celilines, as weil as MHC cia" lmolecliles.

INTRODUCTION

Uveal melanoma is the most common intraocl!lar malignancy in aelults. Although

several different therapies have been useel for the local treatment of the primary tumor,

lifc expectancy has not improveel over the last two decaeles. j At least 50% of thc

patients develops metastases in distant organs, especially in the Iiver, which is

eventllally fatal within a few months 2 New immunotherapy strategies using

autologOlIS or allogenie tumor cell vaccination are being eleveloped for the treatmenl

of cutaneous melanoma patients suffering from advanced lnetastat'ic çli.sease.3,4

Similarly these treatment moelalities can possibly applied to metastatic uveal

melanoma patients.

Ill1munohistochell1ical studies have shown that tumor-infiltrating IYll1phocytes

(TIL) in uveal melanoma are mainly composeel of helper (CD4+) anel cytotoxic

(CD8+) phenotypes. 5,6 From mixed lymphocyte/tumor cell cultures using TIL or

127

Page 128: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ ~ __ ----'c HAP TER 7 ---.--.-----

peripheral blood lymphoeytes (PBL), cytotoxic T-eell (CTL) clones have been isolatcel

which are able tu recognize uveal melanoma eells in a major histocompatibility

complex (MHC) class I restricted manner7 It has been shown that these CTLs

reeognize autologOlIs as weil as allogenic uvealmelanoma eells anel thus share the

same tumor-associateel antigens7 The anti-tumor response displayed by tumor-specifie

CTLs is the result of their ability to recognize "self"-proteins which are processeel in

the cytoplasm of tumor eells. Small peptides, processed from these proteins, are

transported to the eneloplasmatie reticulum, where they binel to newly synthesizeel

MHC class I molecules. These MHC/peptide complexcs are exported to the plasma

membrane, where they ean be reeognized by specific CTLs8

In recent years scveral cutaneous melanoma-specific antigens slich as the

MACE family,9-12 and melanoma-associated antigens (CP/OO, tyro.l'illase)13~15 have

been identified. MA CE genes are expressed in severalmalignancies but not in normal

tissue, exeept for testis. These melanoma-specifie antigens can be recognizeel by CTLs

in the context of specific MHC class T molecules. 10 The melanoma-associated

antigens, GP 100 and fyrosinGse, are not melanOlna specific since the)' ure a1so

expresseel in normal melanocytes. 13-15 The CP 100 anel tyrosillase antigens can be

recognized by CTLs in an MHC class I restricted malmer anel ean be recognized in the

context of HLA-A2.

The purpose of the present study was to investigate whether uvealmelanoma

ceillines do express melanoma-specific and melanoma-associated antigens, sueh as

MA CEl, -2, -3, CP /00 and tyl'Osillase. Since proper antigen presentation can only take

plaee in the presence of a specific MHC class I molecule, the MHC class lexpression

was eletermined in these celllines.

128

Page 129: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_____ r:XPRESSION OF MA,GE, GPIOO AND TYROSIN/1Sl!.

MATERlALS AND METHODS

Celllines

At the department of Ophthalmology, Erasmus University Rotterdam, we succeeded in

establishing celllines fram two primary uveal melanomas (EOM-3, EOM_29)16,17 and

fram three subcutaneously localized metastases of uvealmelanomas (OMM-I, OMM-

2, OMM_3)I6.18 The other primary uvcal me la noma ceillines used in this study were

established and kindly provided by Dr. 1. Kan-Mitchell, University of Southern

California, Sall Diego, California (OCM_I),7,16,19 Dl'. B. Ksander, Schepens Eye

Research Illstitute, Harvard Medical School, Boston (MEL202)20 and Dr. I. de Waard­

Siebinga, Department of Ophthalmology, University Hospital Leiden, The Netherlands

(92_1)21. Thc cells were cultured until eonfluency, washed with phosphate-buffered

saline (pH 7.4), and subsequentJy homogenizeel on ice with an ultra-turrax (TP 18-10)

with intervals 1'01' 2 minutes. For each eellline, three flasks of 25 cm2 (Faicon 3013E)

were used.

Reverse n'allscriptase Polymerase Chain ReaetiOll (RT·PCR)

The expression of the genes MAGEI, -2, -3, GP 100 and ty/'Osiuase was detenllined by

RT-PCR. RNA isolation was accomplished using the LiCl/ureum methad, as described

elsewhere.22 For cDNA synthesis, 3 pg of total RNA was aeldcd to a 50 ftl reaction

mixture containing five times concentrated reverse-transcriptasc buffer, 2.5 pg random

primers (0.09 OD260 Units/ftl) (Boehringer Mannheim), 0.5 ftM DIT, 25 pM dNTP,

and 20 units of RNAse inhibitor (GTBCO BRL). Tncubation was performed for I hour

at 37"C and subsequently for 10 min at 100°e. The thus praduccel cDNA could be

stored at -20°e. The RT-PCR amplification of the cDNA was accomplished by the use

of oligonucleotide primers, which are locateel in various exons of the MA GE,

t)'l'OsÎl1ase, GP 100 anel human f3ACTIN genes. The RT-PCR of the f3ACT1N gene was

used as a positive contra!. The seCjuences of these exons are: 5'-CGGCCGAAGGAA

CCTGACCCAG-3' (CHO-14) and 5'-GCTGGAACCCTCACTGGGTTGCC-

3' (CHO-12) f'or MAGE1;1l 5'-AAGTAGGACCCGAGGCACTG-3' (CDS-9) anel

5'-GAAGAGGAAGAAGCGGT

129

Page 130: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

C HAPTER 7 ______________ _

CTG-3' (CDS-7) lor MAGE2; 11 5'-TGGAGGACCAGAGGCCCCC-3' (AB 1197) and

5'-GGACGATTATCAGGAGGCCTGC-3'(BLE5) (or AB-913, 5'GGAGTCCTCATAGGA

TTGGCTCC3') for MAGE3;12 5'-TATTGAAAGTGCCGAGATCC-3' (149711516)

and 5'-TGCAAGGACCACAGCCATC-3' (1839/1857) for GP100;13 5'TTGGCAG

ATTGTCTGTAGCC-3' (HTYR1) and 5'-AGGCATTGTGCATGCTGCTT-3' (HTYR2)

for tyl'OsÎnase gene;23 5'-GGCATCGTGATGGACTCCG-3' (exon 3 sen sc) and 5'­

GCTGGAAGGTGGACAGCGA-3'(exon 6 anti-sen se) for human f3ACTIfil

(Phanllacia, Woerden, the Netherlands).

Tt was made sure that each polymerase chain reaetion (PCR) eonlained 0.2 ~ll

cDNA supplemented with 2.5!i1 of lOx PCR Buffer (GIBCO BRL), 2.5 pi 01'2.5 mM

dNTP (Sphaero Q, the Netherlands), 5 pM of eaeh primer, 0.5 units of Taq-polymerase

(GIBCO-BRL) and water, to a total VOIUllle of 25 lil. In PCR-tubes (0.5 mi; Eppendorf

safe loek) approximately 50 ftl of mineral oil (Sigma) was added to the reaction volume

to prevent evaporation. The eDNA was first denaturated in 5 min (MAGE1,-2, GPlOO,

f3ACTIN) or in 4 min IMAGE3, tylVsÎnase) at 94°C. The amplifications were then

induced; of MA GEl 36 cyèles (lmin at 94°C, 3 min at noC), of MAGE2 36 cyc1es (I

min at 94°C, 2 min 67°C, 1 min noC), of MAGE3 30 cycles (lmin at 94°C, 2 min noc,

2 min at noC), of.f3ACTIN gene 30 cycles (lmin at 94°C, 2 min 68°C, 2 min at noC),

and of fyl'OsÎnase and GPIOO 36 cycies (1 min 94°C, I min 60°C, I min noC). Afinal

incubation 01'7 minutes IMAGEl, GP100, f)'lVsÎnase) or 15 min (MAGE2,-3, f3ACTIN)

was performed at noc before the samples were cooled down to 4°C (Tribloek,

Biometra). A 15 mi aliquot of eaeh reaction was size-fractionated on a 2% agarose gel

stained with ethidium bromide. A Biorad Kb ladder was used as a calibration marker.

MHC Clnss Iexpression

To establish the expression of MHC c1ass I molecules, the ceUlines were tested in the

complement-dependent mierolymphocytotoxicity assay using a set of monoc1onal

antibodies direeted against the various HLA-A and HLA-B antigens (monoc1onal

typing trays LM In and LM2n, One Lambda lnc., Canoga Park, CA, USA) Some of

these monoc1onal antibodies are reactive with only a single HLA antigen, whereas

others recognize more than one. From the reaction pattern of the eells with these

monoclonal antibodies the HLA antigen composition of the ceUs has been establishecl.

130

Page 131: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

EXPRES SION Oy_MAGE, GPt/JO AND TYROSJNASf·~; ______ _

RESULTS

The expression of the genes uIlder examination in the various uvealmelanoma eell

lines is shown in TABLE I anel FlGURE I. These figures show that MA GEI , -2, anel -

3 transcripts were found in the ceillines OCM-I anel OMM-I, while no expression ol'

these genes coulel be demonstrated in the other celllines. Expression of GPIOO was

fOllllel in all examineel cell lines. The 1)'l'OsillGse transcript was 1'00lllel to be expressed

in five out of the cight celllines (not in EOM-29, OMM-2 anel OMM-3).

The MHC class I expression of the uvealmelanoma celllincs is also inelicateel

in TABLE 1. Thc primary uveal melanoma celllines OCM-I, EOM-29, MEL202 anel

92-1 expressed MHC class I molecules; the latter two expresseel HLA-A I or HLA-A2

molecules. In the primary uveal melanoma cellline EOM-3, the MHC class I

expression was eIown regulated anel no inll11Unocytochemical staining was e1etectable.

Of the metastatic uvealmelanoma celllines, all were found to express HLA-AI or

HLA-A2.

TABLEt EXPRESSION OF MELANOMA-SPECJFIC AND -ASSOCIATED ANTIGENS, AND MHC

CLASS J MOLECULES IN UVEAL MELANOMA CELL L1NES

ANTIGENS CELLLINES

OCM-1* EOM-3'· EOM-29·' ~deI202'!< 92-1* or .... Hvl-lt OJ\Hvr-2t OMi\·1-3t

MAGE-1 + - - - - + - -

MAGE-2 + - - - - + - -

MAGE-3 + - - - - + - -

OPIOO + + + + + + + + TyrosiJ/lise + + - + + + - -

/Jac/in + + + + + + + -

tvtHC dass I A24,A28 na HLA A3,A30 Al,A3 A2.A3 A2.AII,B62 AI AI,AlI

*' primary uvenl mcJanoma cellline; 1" mctastatic lIvea! melanoma celllinc; + positive RT-PCR product; - nu RT-PCR product

131

Page 132: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

1

CHAPTHR 7 -"-"--'-----

2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

FIGURE 1. RT-PCR products of MAGE-l (a), -2 (b), -3 (c), CP/OO (d), tyrosillase (e), and {3-Clclill (f) in human uveal mclanoma celllillcs; lane 1, EOM-3; lane 2, EOM29; lane 3, OMM-2; lane 4, lvle1202; lune 5, 92.1; lane 6, OMM-I; lane 7, OMM-} and lane 8, OCM-I.

DISCUSSION

For cutancous mclanoma paticnts with advanced mctastatic diseasc new

immunotherapeutic strategies are being developed. Patients arc vaccinated with

autologous tumor eells or allogenie celilines transfected lVith the IL-2 or Granulocyte­

Maerophage Colony-Stimulating Factor (GM-CSF) genes 3,4 Along the same line, a

similar trcatment l110elality can be eleviced for patients suffering from uvealmelanoma

lVith distant metastases. In order to apply these new immunotherapeutic strategies for

mctastatic uvcal melanoma paticnts, suitablc lIVeallllclanoma eeHEnes, whieh cxpress

melanoma-associated antigens and proper MHC class I molecules, are necessary. In

the present stuely, we demonstrated that the melanoma-specific MAGEI, -2 anel-3

/32

Page 133: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

EXPRI~SSION o.F MA(j/~_, GP/OO AND TYIUJSfNASE

genes and the melanoma-associated GP 100 and lyrosÎl/ase genes can be exprcssed in

uveal melanoma celiJines. Furthermore, MHC class I molecules can be expressed in

these uveal mclanoma ccll Jines.

In cutaneOllS melanoma, MA GEI is expresscd in about 40% of the cases,

MAGE2 in 85% and lvlAGE3 in 69%9-12 In our study, the mRNA transcripts of thc

MAGEI, -2, and -3 gene are expressed in one of the primary (I :5) and in one of the

metastatic (I :3) uveal melanoma celllines. Thc higher expression rate of MAGE2 and

MAGE3 found in cutaneous melanoma, could not be demonstrated in our uveal

melanoma celllines. The melanocytic-lineage spedfic antigen GP I 00 was found to be

expressed in all uveal melanoma celllines. The GP JOO antigen was found to be

recognized by the monoclonal antibodies NKI/beteb, HMBA5 and HMB_5024 These

antiboclies are highly sensitive markers for cells clltaneOllS as well as llvcal melanOlnas

and their metastases. 24•25 Recently it has been reported that the GP I 00 antigen is a

type 1 transmcmbrane glycoprotein that is processed intl"acellularly and presented as

peptides to CTLs in the context of HLA-A2. 14 Such CTLs, obtained from HLA-A2

positive melanoma patients, often display a broad reactivity not Jimited to melanoma

cells only, but also to normal cutaneous melanocytes as weil. Another antigen that is

recognized by CTLs appeared to be the transcript of the lyrosÎl/ase gene. The

tyrosÎl/ase gene is involved in the melanin synthesis and is expressed in all

melanocytic-lineage specific ceHs; peptides derived from this protein can bind to

HLA-A2 and HLA-A24 mOlecules. 23,2ó Sinee HLA-A2 is exprcssed in about 49 % of

the Caueasian poplilation and GPIOO as well as the lyrosÎl/ase gene are expressed in a

high ra te of all uveal melanomas, it eonstitutes a useful target for specific

immunotherapy. However, further studies are required to determine the side-effects of

this type of therapy, sueh as the apparent destruction of choroidal melanocytes and

retinal pigment epithelial cells.

The expression of the membrane-bound MHC/peptide complexes, reeognized

by CTLs, and of which the peptides are derived from the MAGE, GPlOO and

lyrosÎl/ase genes were formerly deseribed as being restricted to HLA-AI or HLA-A2

molecules. However, more recent studies have elucidated that the complex binding of

these peptides ean probably also oecur in other MHC class 1 molecules and is henee

not restricted to HLA-A I or HLA-A2 alone27

Considering this fact, it might therefore

/33

Page 134: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

________ . ______ CHAI'TER 7. __ ------

be pos si bie lhat CTLs can recognize MHC/peptide complexes of lhe OCM-I eeils,

although they express HLA-A24 and HLA-A28.

In lllore than 90% of the uveal mclanoll1a patients, liver metastases arc the

IÏrst signs of melastatie disease, with a median life-expectancy of only 2 to 7 11l0nlhs.z

Therefore specific il1ll1lunolhcrapy with Iransfeeled ailogenic uveal melanoma cel!

!ines mighl be a good approach. The ceil!incs OCM-l and OMM-l seem to be the

best candidates ta enhance immunoreactivity, because in these eeB lines all melanoma­

specific anclmclanoma-associated antigens, as weil as the MHC c1ass I molecules

were found 10 be expresscd. Additional studies on CTL responses to these uveal

melanoma cell targets are necessary 10 demons tra te effective eytolysis.

134

Page 135: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

EXPREssrON OF MAU~, GPIOO AND T)'ROSINAS/,:'

REFERENCES !. Egan KM, Cetant ]lvi, Glynn Rl, Gragol1uas ES, Albert mvt Epidemiologie aspect~ of uveal

melanoma. Surv Ophthalmol. [988;32:239-251.

2. Kath R, Hayungs J, Bornfeld N, Sauerwein W, Höflken K, Seeber S. Prognusis. and treutment uf

disseminated uwal melanoma. Cancer. 1993;72:2219-2223.

3. Dranoff G, Ja/ree E, Lazenby A, ct al. Vaccination with irradiated tUlllor cells engineered tu secrete

murine granulucyte-macrophage coluny-stimu[ating factor stimulates potent, specific, lInd lung­

lasting anti-tumor immuility. Proc Natl Acad Sci USA. 1993;90:3539-3543.

4. Panlull D~vr. Ncw strategies tor enlwncing the immunogenicity of tUlllors. Curr Opinion hml1unul.

1993;5:719-725.

5. Whelehel JC, Pandl SE, McLean IW, Bl1rnier ~vl. Immmunohistochernistry uf infiltrating

Iymphocytes illuvcal melanoma.ln"cst Ophthalrnol Vis ScÎ. 1993;34:2603-2606.

6. Tobal K, Deuhle K, 1kCartney A, Lightman S. Charw:terization of cell11lar inl1ltration in choroidal

rnelanoma. Melanoma Res. 1993;3:63-65.

7. Huang X-Q, :~vIitchell tvlS, Liggett PE, Î"hnphee AL, Knn-Mitchell J. Non-fastidiuus, melanolllU­

specific CD8+ cytotoxie Iymphocytes from choroidalmel:moma patients. Cancer Immunol

Immunother. 1994;38:399-405.

8. Gernmin RN. MHC-Dependent antigen processing and peptide presentation: Provkling ligands for T

Iymphocytc activation. Cello 1994;76:2R7-299.

9. Van der Bruggen P, Traversari C, Chumez P, et al. A gene encoding an antigen recogniLed by

cytolytic T lyrnphocytes on a hunHln melanoma. Science. 1991 ;254: 1643-1647.

10. Traversari C, Van der Bruggen P, Lucscher IF, et al. A nonapeptide encoded by human gene

MAGE~ 1 is recognized on HLA-A I by cytotoxic T Iymphocytes directed against tumor antigen

t't'tZ2-E. J Exp r.,'fed. 1992;176: 1453-1457.

11. De Smet C, Lurqllin C, Van tier Bruggen P, de Plaen E, Brasseur F, Boon 'I: Scquence and

expression pattern of the human JvIAGE2 gene. Immunogenetics. 1994;39: 121-129.

12. Gaugler B, Van tien Eynde B, Van tier Bruggen P, et al. Hurnan gene MAGE-3 codes fur an antigen

recugnized on a melanoma autologOLls cytolytic T lymphocytes. J Exp Metl. 1994;179:921-930.

13. Adema GJ, de Boer AJ, Vogel Atvl, Loenen WAI'."I, Figdor CG. Molccular characterizatiun uf the

melanocyte lineagc-specific antigen gplOO. J Biol Chem. 1994;269:20126~20133.

14. Bakker ABH, Schreurs MWJ, de Boer AJ, et al. Ivlclanocyte lineage-specit1c antigen gplOO is

recognized by melanomH derived tumor int1ltrating Iymphucytes. J Exp t-.kd. 1994; 179: [005-1009.

IS. Brichard V, Van Pel A, Wolfel T, et al. Thc tyrosinase gene codes for an antigen recognized by

autologOlIs cytolytic T lymphucytes on HLA-A2 melanomas. J Exp l'vfed. 1993;178:489-495.

16. Luyten GPlvI, Naus Ne, ivlooy C1'i, et al. Establishment am! characterization of primary and

metastatic llveal melanoma celllines. Int J CanceL in pre ss

17. 1."la D, Luytcn GPM, Luider TM, Niederkorn JY. Relationship between natural killer cell

sllSceptibility and metastasis ofhuman uveal melanoma ce lis in a mLlfine model. Invest Ophthalmol

Vis Sci. 1995;36:435-441.

IR. Luyten GPM, lVIooy CM, De Jong PTV:/vl, Hougeveen AT, Luider TM. A chicken embryo model to

135

Page 136: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______ CfIAPTf~ __________ _

study Ihe growth of lmman uveall11clnnoma. Biochem Biophys Rio's COI11I11. 1993; 192:22-29.

19. Kun-Ivlitdlcll J, l\'tijchell !\'IS, Rau N, LiggeLt PE. Charactcrization of uvealmelunom<l ccl11ines thaI

grnw as xcnugrafts in rabbi! eyes. hlvest Ophthalmol Vis Sci. 19B5l;30:829-B34.

20. Ksunder ER, Rubsamen PE, Ol sen KR, COl1sins, Streilein J\\'. Studies of tlllllor-infïltrating­

lyrnphocytes trom n hllman choroidal rnelanumll. Tnves! Ophthalmol Vis Sd. 1991 ;32:3198-320ö.

21. De WaanJ-Sicbingn I, Blom DJR, Griffioen 1-1, et al. Establishment and characlerizatiun of an tl\'cal­

melnnoma ccllline. Int J Cancer. 1995;62: 155-16 I.

n. Auffray C, Rüllgcon F. Pmifîcation of manse immunoglobulin heary-chain messenger RNAs from

Inlui myeluma tumor RNA. Eur J I3iochem. 198R; \07:303.

23. Brossart P, Kcilholz U, Schcibenbogcll C, 1·töhler T, WilIhauck M, Hunstein W. Detection of

residual tumor cells in paticlIts with malignant me~anoma respon~ing to imlllunotherapy.

J lmmunothcr. 1994;15:38-4l.

24. Adema Gl, dl' Boer AJ, vall 't Hullcnaar R, ct al. lvfelanocyte lincage-specific antigens recognizcd by

monodonal antibodies NKI-beteb, Hl'vfB-50, and H~m-45 are encoded b)' a single cDNA.

Am J Pat hol. 1993;143: 1579-1585.

25. Steuhl KP, Rohrbach JM, Knorr M, Thiel11J. Signil1cance, spccificity, and lIltra~tfllctllral

localization of HMB-45 antigen in pigmcnted oeular tumors. Ophthalmology. 1993; 1 00;208-215.

26. Robbins PF, EI-Gamill'vl, Kawakmni Y, Rosenberg SA. Recognition of tyrosinasc by tumor­

innItrating Iymphocytes from a patient responding 10 immllllllotherapy. Cancer Res.

1994;54:3124-3126.

27. Celis E, Tsai V, Crimi C, et al. 1nduction of anti-tumor cytotoxic T Iyrnphocytcs in norm al hUm<lllS

Ilsing primary cultures anti synthetic peptide epitopes. Proc Nat! Acad Sci USA. 1994;91 :2105-2109,

136

Page 137: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 8

RELATIONSHIP BETWEEN NATURAL KILLER CELL SUSCEPTIBILITY AND METASTASIS OF HUMAN UVEAL

MELANOMA CELLS IN A MURINE MODEL

D. Ma, G.P.M. Luyten, T.M. Luider, J.Y. Niederkorn

[nvest. Ophthal Vis Sci. 1995;36:435-441

137

Page 138: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 8

ABSTRACT

Plll'pOse: The pmpose of this StllUy was to eletermine the sllseeptibility of human

uveal melanoma eells to in vitro and in vivo natural killer (NK) cell-mediateel

cytolysis and to eletermine if NK cells intluenee metastasis from the eye.

Methods: Fom human uvealmelanoma ceillines anel one melanoma eeliline elerived

from a metastatic lesion from a patient with lIveaI meJanoma were tested tal' in vitro

and in vivo NK cell-mediateu lysis in a 1l10use model. Mf\jor histocompatibility

complex (MHC) class I antigen expres sion was evaluated by tlow cytometry. The role

of NK cells in controlling the metastasis of uvealmelanoma eells from the eye to the

liver was examined in nude mice.

Results: Sensitivity to in vitro and in vivo lysis by human anelmmine NK eells was

eorrelateelwith reduced expression of MHC class I antigens. Uvealmelanoma lines

expressing normal MHC class I antigen expres sion were insensitive to NK eell­

mediated lysis, both in vitro and in vivo. Metastasis of uveal melanoma cells was

inhibiteel by NK cell activity beeause elisruption of in vivo NK function produeeel a

sbarp inerease in the spontaneOlls metastasis of intraoclIlar melanomas in nude miee.

Conclusiolls: There is considerabie variation in the suseeptibility of human

melanomas to NK cell-mediated cytolysis. Susceptibility is closely correlated with

reduced expression of MHC class I antigen expression. Disruption of NK eell function

signifieantly increases the development of hepatic metastases from human uveal

melanoma eens.

INTRODUCTION

Uvealmelanoma is the most eOl11mon primary intraoeular tumor on adults. 1 Although

primary uvealmelanoma can be successfully treated by enucleation or radiotherapy,

approximately 50% of uveal tumor patients will elie from metastatic disease2 The liver

is the most frequently affected organ in patients with metastatic me\anoma from uveal

138

Page 139: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ ~N"'0A.J'URAL KILLER CELL SlJJ.3SCEPTIBILITY OF UVEA,L l\IELANOMA

tract, and hepatic metastasis remaim; the leading cause of cleath in uveal lllclanoma

patients.3-5 Thus, prcventing and treating hepatic metastases could have an important

impact in thc treatment of uveal melanoma patients.

In recent years, there has been a rekindled interest in the application of'

immunotherapy for a vadety of tumors, especially cutaneous melanoma. Clinical trials

involving the use of interleukin-2 (IL-2) and the adoptive transfer of tumor-infiltrating

Iymphocytes (TIL) have pravidcd evidence attesting to the significant, albeit Iimited,

ft· f" I' .. f I 6-9 e leac)' 0 1111111Ullot lerapy lor metastases anslllg rom cutaneOlis me anOlnas.

Although cutaneous melanoma has received considerable attention, only a limited

nu mb er of studies have focused on the potentialof inununotherapy of uveal

melanoma. Ksander and co-workers have shown that antigen-specific cytolytic T

Iymphocyte (CTL) activity was demonstrabIe in TIL isolated from human choraidal

melanomas. to Kan-Mitchell and co-workers were able to isolate and demonstrate

melanoma-specific CTL fram the peripheral blood of patients with choraidal

melanoma. 11 Although cells with NK activity are present in TIL populations isolated

f1'0111 human choroidalmelanomas, 10 na studies ta date have examined thc rclative

susceptibility of hlllnan uveal melanomas to NK cell-mediated cytolysis or the

capacity of NK cells to prevent the metastatic spread of uveal melanoma cells. The

present study was designed to address both of these crucial questions.

MATERlALS AND METHODS

Mice

Athymic nude BALB/c (H_2d) mice wcre purchased from the Jackson

laboratories (Bar Harbor, ME). The use of animals conformed to the ARVO Statement

for the Use of Animals on Ophthalmic and Vision Research.

Tumor Cell Lines

Fom human uveal melanoma celllincs, designated OCM-I, OCM-3, OCM-8,

and EOM-3, \Vere llsed. OCM-I, OCM-3, and OCM-8 were generausly pravidecl by

Jllne Kan-Mitchell (USC School of Medicine, Los Angeles, CA). OCM-! is a

predominantly spindie morphology while OCM-3 anel OCM-8 are predominantly

139

Page 140: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

eH A P T E ,,"--'8"--____ _

epithelioid tumors. tt OMM-I was isolated from a subcutancOlIS metastasis in a patient

with uvealmelanoma paticnt and displays a mixed cellmorphology.12 The EOM-3

cellline displays an epithelioid morphology in vitro but was derived fram a posterior

choroidal melanoma of mixed cell morphology. OCM-I, OCM-3 and OCM-8 cells

were cultured in HAM's F-12 medium containing 10% heat-inactivated fetal calf

serum, 1% L-glutamine, 1 % sodiu111 pyruvate, 1% nonesscntial amino acicls, 1 %

HEPES buffer, and 1% antibiotic-antimycotic solution. OMM-I and EOM-3 cells

were cultureel in Dulbecco's modifieel Eagle's medium eontaining 10 % heat­

inactivated fetal calf serum, 1 % L-glutaminc,l % sodiul11 pyrl1vate, 1 % vitamin

solution, and 1% antibiotie-antimycotie solution.

Analysis ol' MH C Class IExpression

The expression of major histocompatibility complex (MHC) class J antigens

was assessed by flow cytometry using a f1uorescenee-activated eell sorter (FACS) and

was detecteel using a mouse monoclonal antibody directed against nonpolymorphic

MHC class I determinants (IgG, clone BT-3d3; Aeeurate Chemical & Scientific Co.,

Westbury, NY). Mouse anti-H-2k monoc\onal antibody (IgG; clone 11-4.1;

PharMingen, San Diego, CA) serveel as an isotype-matched negative contral antibody.

Single cell suspensions were prepared anel washed in FACS buffer consisting of

phosphate-buffered saline (PBS; Ph 7.4) with I % BSA and 0.02% sodium azide. Cells

(106) were incubated with primary antibodies for 30 minutes on ice, washed th ree

times, and then incubated with FITC-Iabeled secondary antibodies for 20 minutes at

O°C and washed an aelditional three times. Cell suspensions were fixed in 1%

paraformaldehyele anel assayed for positive staining on Epics Profile Analyzer (Coulter

Eletronics Ine., Hialeah, FL). Gates were set at I % of total eells, baseel on staining by

seeondary antibodies alone.

In vitro Natural Killer Cel! Assay

Natural killer eell-mediated cytolysis was evaluated by a eonventional4-hour

slCr-release assay as elescribed previollsly.13 Human peripheral blood was used as a

SDuree of NK cells. Heparinized blood from healthy donors was layered onto

140

Page 141: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

Histopaque-1l19 (Sigma Chemical Co., St Louis, MO) anel centrifuged for 30 minutes

at 2,200 rpm at room temperature. The mononuclear celllayer was collecteeI. BALB/c

mouse spI eens serveel as a souree of murine NK cells. Spleen eells were presseel

through sterile stainless steel screens, washeel twice in HBSS, pas se el through stCl'ile

nylon mesh (Tetko, Elmsforel, NY), anel washed in HBSS an aelelitional time. Total

release C0l1l1tS per minute were determineel by treatment target cells with 0.05 mI

Hematall LA-Hgb Reagent (fisher Scientific, Pittburgh, PA) and spontaneous release

by incubating target cells with complete medium in the absence of effector cells.

Lh'er Localization Assay

Assessment of hepatic NI< cell-meeliateel immunity was performeel by an in

vitro clearance assay as described previously.14 Briefly, oeular tumor eclls were

labeleel in vitro with 100 IlCi of Na251Cr04 for I hour at 37"C, washed three timcs

with HBSS, and resuspended in HBSS at a concentration of I x 107 cell/ml.

51Cr-labeled melanoma cells (I x 106 in 0.1 mI) were injecteel into the lateral tail vein.

Clearance of 51Cr-labeled tumor cells was eletermineel by counting the radioactivity

(gamma emission) of livers, lungs, spI eens, kidneys, and 0.5 mI blood removed ti'om

each animal 24 hours after intravenous tumor injection. Percent of radiolabel

recovered was determined by dividing the eounts per minute (cpm) for each liver

divieleel by the total cpm recovereel for liver, lungs, kielneys, spleen and 0.5 mI blood

ror eaeh animal. The total cpm t'or each anima I nmged from 35.000 to 45.000 cpm. In

vivo NI< cell-mediateel elimination of radiolabeleel tumor cells is reflccted by a

reduction in the radioaetivity of the Iiver. 14

In vivo NK eell mediated elimination of radiolabeled tumor eells is rctlected by a

redllCtion in the radioactivity of the Iiver. t4

Intracameral n'ansplantation mul HelJatic Metastasis Determination

A moelified Cjuantitative techniCjue for the orthopie intracameral (IC) transplantion

of preeise numbers of tumor eells into the mouse eyc has been described previously. t5

Mice were deeply anaesthetized with 0.66 mg of ketamine hydrochloride (Vetalar; Parke­

Davis, Detroit, MI) given intramuscularly. Tumor cells (105/5/.11) were inoculated IC using

141

Page 142: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

----- ___ .. __ CHAPTER 8 ,, ___ . ______ . __ _

a J .O-ml Hamilton syringe fitted with a 35 gauge glass needie. Metaslatic hepatie lu mor

foei were readily demonstrabie by histopalhologieal examination of the live I' and were

scored as previously deseribed16,17 Severily of metastases was scored as: clear (0 = na

diseernible foei); minimal involvement (I + = melastatic lumors involved less than 10% of

the liver); moderate (2+ = melastatic lumors invoJved 10-25% of lhe liver); or extensive

(3+ = melastalic tumor mass involved ;:> 25% of the liver).

Cyclophosphamide lhatment

Cyclophosphamide (CY) was purchased from Mead Johnson Laboratories

(Princeton, NJ) and dissolved in slerile HBSS. Panels of miee received intraperiloneal

injeelions of CY (240 mg/kg) 24 hours befare lheir sacrifice and use in the in vitro NK

assays, For in vivo experiments, NK activity was inhibiled by intraperitoneal injection of

cyclophosphamide (240 mg/kg) on day 30 aftel' tumor lransplantation.

Poly I:C Tl'eatment

Polyeytidylie aeid (poly I:C) was purchased from Sigma Chemical Co. (SI louis,

MO) and dissolved in steriIe HBSS before trealment. Natural killer aelivity was

stimulated in experimental animals by injeeting poly I:C (100flg/mouse) inlraperiloneal at

weekly intervals aftel' IC lumor transplanlation aud continuiug untillhe time of necropsy.

Anti-asialo GMI Tl'eatment

Anti-asialo GMI antibody was purehased from WAKO Chemica I s( Dallas,TX)

and used tor in vivo depletion of NK eells, BALBIe nu de miee were trealed intravenously

wilh anto-asialo GMI (0.2 mi of a I: 10 clilution/mouse/injeelion) on day -3 and day -I

before the in vitro und vivo NK assays. For metastases cxperiments, anti-asÎalo GMI

antibody was injeeled on ce per week until the lime of necropsy,

Statistics

Sluden!'s t-test was used 10 tesllhe slalistical significanee of the data. Differenees

were consiclered significanl when P<0.05.

142

Page 143: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ N_'ATURAL KILLER CELI_, SUBSCEPTIHlLITY OF UVEAL 1fELANOi\IA

RESULTS

Human uveall11elanol11a cells eliffer in their susceptibility to NK cell-l11ediated

cytolysis

Although IYl11phoiel cells bearing NK-associated celll11embrane markers have

been ielentified in uveal melanamas,lO no studies to elate have specifically cxamined

the susceptibility of uveallllelanollla cells ta NK cell-meeliated cytolysis. Therefore,

fivc hUl1lan uvealmelanollla celllines wcre examined for their susceptibility to in

vitro cytolysis by murine and human NK cells. Two human uvealmelanoma cells

lines, OCM-3 anel EOM-3, were susceptible to cytalysis by both murine (FIGURE 1).

anel human NK cells (data not shown). By contrast, the other three celllines displayed

insignificant susceptibility to NK cell-mediateel lysis. Specific lysis by human NK

cells was less than 10% far OCM-I, OCM-8, and OMM-I (data not shown).

50

40

JO

20

10

0-'----'---'--OCM1 OCMJ OCMB OMM1 EOMJ

FIGURE 1. Susccptibilily of five humnn uvealmelanoma celilincs to cytolysis by murine NK cclls. Cytolysis was measufcd in a cOIlvcntional4-houf SIer-release <lssay lIsing murine spleen ceHs at an effector-ta-target ratio of 50: I. Results are exprcssed as menll ± standard deviatioll. NK, natural killer.

143

Page 144: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 8

The susceptibility of human uvealmelanoma eells to NK eell-mediated lysis

in vivo was tested using a previollsly described clearance assay.14 The elimination of

intravenously injeeted radiolabeled tumor cells is a well-establishcd method tor

evaluating NK eell-mediated lysis in vivo. 14 We have previously shown that hUlllan

uveal melanoma eells preferentially loealize in the liver following intravenOlIS

injeetion in nllde mice. 18 Thereforc, the in vivo climination of hllman lIveallnelanoma

eells was evalLUlteel by detennining the residllal radioaetivity of livers removeel from

nuele miee 24 hOlll's after intravenous injection of radiolabeleel tumor ceHs. In some

mice, NK runetion was stimulated by intraperitoneal injection of 100 ~Ig or poly I:C

24 hour prior to the assay. In other groups of miee, NI< eell aetivity was impaired by

intraperitoneal injection of either cyclophosphamide (240 mg/kg) or anti-asialo OM I

antiboely.

The in vivo clearance of intravcnous injeeted uveal melanoma eells paralleled

the in vitro results. That is, NK-sensitive OCM-3 and EOM-3 melanoma eells were

eliminated more efficiel1tly from the Iiver than lVere NK-insensitive OCM-I, OCM-8,

anel OMM-I eells (TABLE I).

TUMOR

OCM-I

OCM-3

OCM-8

OMM-I

EOM-3

TABLE 1 HEPATIC CLEARANCE OF RADIOLABELED HUMAN UVEAL

MELANOMA CELLS

PERCENT RADIOLAllEL RECOVERED' (PVALUE)

CONTROL POLYI:C CYCLO- ANTI·ASIALO

PHOSPHAMIIJE GMl

70.5 57.7 (.057) 80.3 (.18) 81.8 (.14)

64.2 52.3 (.089) 81.2 (.034) 83.7 (.02)

72.1 68.4 (.59) 76.2 (.54) 75.4 (.63)

67.5 64.3 (.61 ) 74.5 (.29) 73.6 (.37)

63.3 59.6 (.58) 79.4 (.035) 78.~ (.046)

Percent dadiolabel recovered in the liver based on total counts recovered from lungs, Iiver, soleen, kidneys, aud 0.5 mI of pherpheral blood. Probability (P) values for each group compercd to untrealed controls ure shown in parenthese.

144

Page 145: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ ~N"',,ATURAL KILLER eELt, SUBSCf~PTIBILTTY OF UVEAI" MELAN(?i\"IA"----__

Moreover, disruption of in vivo NK cell function with either anti-asialo OMI antibody

or cyc1ophosphamide signifieantly impaired the liver clearance of OCM-3 and EOM-3

cells but had an insignificant effect on OCM-I, OCM-8, and OMM-I eells. Thus

OCM-3 and EOM-3 cells displayed significant in vitro and in vivo susceptibility la

NK eell-Illediated lysis.

NK sensitivity of human uveal melanoma is invel'sely cOl'l'elated with MHC Class

I antigen expl'ession.

The NK sensitivity of murine fibrosarcomas and Illurine IYlllphomas has been

shown la be inversely correlated with MHC cla" T antigen expression. 14-

19

Accordingly, the expression of MHC class I anligens by human uveal melanollla cells

was evaluated by flow cylometry. The resuIts indieated thaI NK cell-mediated

cytolysis was significantly reduced in human uveal melanoma eelllines I MHC

antigens (FIOURE 2) By contrast, uveal melanoma celllines less than 10 % positive

for MHC cla" Texpression were vulnerable 10 NK cell-mediatecl cylolysis (FIOURE

2). Thus, the NK sensitivity of human uveal melanoma is inversely correlated wilh

expression of MHC class T antigens (TABLE 2.)6

TABLE2 CORRELATION BETWEEN MHC CLASS I ANTIGEN EXPRES SI ON AND

NK CELL SENSITIVITY

MHCCLASSI INVITRO NK CELL

MELANOMA EXPRESSION ACTIVITY

CELLLINE (% positive eells) (% Specific c)'toI)'sis)

OCM-I 90.12 3.5 ± 3.7

OCM-3 84.10 3.6 ± 4.6

OCM-8 76.85 5.9 ± 5.3

OMM-I 8.55 31.8 ± 9.8

EOM-3 4.71 35.2 ± 11.9

Abbrcviatiol1s: MHC, major histcompatibility complex; NK, natural killer.

145

Page 146: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______ . __ CHAPTI~, ___ . ___________ .

NK eells prevent thc metastasis of uveal melanoma eells

The possibiJity that NK eells ean limit or even prevent the metastasis of uveal

lnelanoma eel1s wa~ examined in mice. OCM-3 human lIveal melanoma was selectecl

for fmther study based on its in vitra and in vivo suseeptibility to NK eell-mediated

lysis. OCM-3 eells were transplanted into the anterior segments ol' BALE/c nude miee

on day O. In same miee, NK function was disl'l1pted by either injeeting

cyclophosphamide intraperitoneally 011 day 30 or through weekly injeetions of anti­

asialo GM I. Other groups of miee had their NK aetivity stimulated by weekIy

injections of the five untreated nuele miee and none ol' the poly l:C miee (TABLE 3).

However, 8 of the 10 nude miee lVith impaired NK eell aetivity harbored hepatie

metastase."; of OCM-3 human llvealmelanoma.

TABLE3 INHiBITlON OF UVEAL MELANOMS METASTASIS BY

NATURAL KILLER CELLS

EXPERIMENTAL GROUI' SEVERITY OF METASTATIC FOCI

Untreatcd 0 0 2+ 0 0

Cyclophosphamide 2+ 2+ 0 3+ 3+

Anti-asialo G!vl1 2+ 3+ 1+ 0 2+

Poly I:C 0 0 0 0 0

INCIDENCE

1/5

4/5 4/5 0/5

Severity of metastasis bascd on relative sizc of metastaic foei detected by mocroscopie examinatioll of histopathologie liver sections staincd with hcmatoxylin and cas in, as described in Matcrials and I'\'letllods, Allmicc were nccropsied on Jay 50.

146

Page 147: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NATURAL KILLER eELL 0UBSCEPTIBILITY OF UVEAL ;'.,.IELANOi'v[A

0 '" 0 50 ~ OCMl

~ OCMB

40 40

1;; 1;;

'" .~ 30 '" .~ 30 E ~ E 5 0 3- 0 z z jij MI (/120 jij MI (/1 20

" " 10 10

0 0

10' 10' 10' 10' 10' 10' 10' 10' 10' 10' Fluorescence Jntensity FluorescenCe lntensity

0 50 0 50

~ OMMl ~ EOM3

40 "" 1;;

.~ 30 " '" '" .~ 30 E ~ E 5 J 3- 0 z z 0; MI (/120 0; MI (/120

" " 10 10

0 0 0

w' 10' 10' 10' 10' 10' 10' 10' 10' 10' Ftuorescence tntensity FluQfeScence Intensity

~ 50

OCM3 40

1;; .~30 '" E

0 Z .J

0; MI I ~ 20

" 10

0

10' IQ' 10' 10' 10' Fluorescence Intensity

FIGURE 2. Inverse reJationship bctween expression of MHC class I antigen cxpression nnd slIsceptibiJity of human uve,tI mclanoma cclllines to NK cell-mcdiated cylolysis. Human peripheral blood sen'ed as a source of NK cells. For FACS profiles, thc abscissa is alogarithI1lic sc ale th at represents tluoresçcnçe illtensity with specitic antibody to MHC class I antigen. The ordinale is a Iillear scale represellting the relative number of cells staining positively with the same antibody. Thc bar graphs represent the menn NK­mediated cytolysis (± standaHI deviation) of thc respectivc target cell in a cOl1vcntional 4-hour SIef-release <lssay using hUlllan peripheral blood lymphocytcs as effector cells. Effcctor-to-target ratio was 50: 1. MHC = major histocompatibiIity complex; NK = natural killer~ }-<ÄCS = fluorcscence-activated cell sOfter

147

Page 148: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

C HAPT E R ~8~ _____ _

DISCUSSION

Numerous studies have demonstrated that NK eells can limit the metastasis of

experimental cutaneOllS B 16 melanomas, and more recent studies have suggested that

Iymphokine-activated killer eells provide protection against metastasis in some

patients with eutaneOlIS melanoma6-9 However, to our knowiedge, there are no

published studies that have examined the suseeptibility of human uvealmelanoma

eeUs to in vitro and in vivo NK eell-mediated eytolysis, Szalay and eoworkers20 have

reported that impairment of NK aetivity promotes the metastasis of intraeameraUy

transplanted BI6 melanoma eells, Using the same murine model, these workers also

reported that in vivo stimulation of NK eell aetivity with the biologie response

modifier, LS2616, redueed the metastasis of intraoeular BI6 melanoma2t However, it

should be pointed out that B 16 melanoma is a murine eutaneous melanoma that may

or may not aceurately refteet the behavior of human uveal melanoma, As shown here

and elsewhere, tumors of the same histologie origin can display vastly different

susceptibilities to NK eell-mediated cytolysis, and one must exereise eaution when

extrapolating results from studies in vol ving heterotopieally transplanted tumor eells,

To study the mie of NK eells in the prevention of metastases of intraoeular

melanomas in a prospeetive marmer, a suitable anima I model is required, We and

others have transplanted BI6FIO melanoma into the eyes of syngeneie C57BL/6 miee

as a model of intraocldar lnelanomH. 20-24 Greene hamster melanoma is perhaps thc

most widely used animal tumor in ocular melanoma research, However, like B 16

melanOlna, Greene melauOllla arosc as a cutaneous melanoma and therefore does not

display tbc same metastatic behavior as uveal melanomu. Human uveal melanoma

" '11 ' h I' 3-52526 I B 16 I pre 'crclltla y I11ctastaslzes ta telver, ' , w lereas me anoma

eharaeteristieally spreads to the lungs 25

lt is beeoming inereasingly clear that tumors transplanted to heterotopic sites

f d d ' I ' bi' , 'h h "I 27-3t o ten 0 not ISp ay metastatlc e lilVlor consistent Wit t e ongma tumor. .

Moreover many human tumors do not metastasize from heterotopic sites in nude mice

but will form metastases aftel' orthotopie transplantation,27-3 t The importanee of

orthotopic transplantation may be due, at least in part, to the int1uenee of loc al organ­

specitïe faetors,27 Thus, the importanee of orthotopie, rather than heterotopic,

148

Page 149: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ ---'N,ATURAL KILLER eELL SUBSCEPTIBILITY OF UVEAL l\IELANOi\IA

transplantation cannot be overemphasized.

The present model has several attributes that be ar noting. Unlike murine BI6

melanoma and Grcene hamster melanOlua, whieh are both cutaneous melanomHs faur

of the five tumor Iines were elerived from patients with uveal melanoma. Thus, the

intracameral transplantations in four of the t'ive tumor ceillines was orthotopic, not

hcterotopic. Although the in vivo studies involved murine NK cells, it should be

cmphasized th at the in vitro studies elemonstrated th at the NK cell-meeliateel cytolysis

paralleled the results obtained with human NK cells. The orthotopically transplanted

uvealmelanomas used in this model demonstrate astrong predilection to metastasize

to the Iiver and thereby mimic the human counterpart. By contrast, B 16F I 0

melanomas invariably metastasize ta the Iungs aftel' intravenolls, intracameral, or

subcutaneOlIS transplantation.

The clearance of radiolabeled human uvealmelanoma cells within the liver

paralleled their in vitro sensitivity to NK cell-mediated cytolysis. Depletion of NK

activity through the administration of either cyclophosphamide or anti-asialo OMI

significantly impaired in vivo clearance of radiolabelcel uvealmclanoma ceUs in the

Iiver anel resultcd in more severe and extensive hepatic metastases. Although in vivo

treatment with Poly I:C nonnally elevates NK activity, it failed to enhance the

clearance of radiolabeled OCM-3 anel EOM-3 cells within the Iiver and appeared to

have an effect, albeit statistically insignificant, in enhancing the clearance of NK­

insensitive OCM-I cells. However, these minor aberrations in the vivo clearance in the

Poly I:C experiments may be elue to the unique qualities of hepatic NK cells in nude

miee. Tzung and Cohen32 have shown that the liver nonnally displays significantly

greater activity than other organs because of the endogenOlls production of interferon­

a/f3 by Kupffer cells. In nude mice, hepatie NK cell function appears to be elevated

even further because liver clearance of radiolabeled tumor ceUs is 60% greater than of

the euthymic counterparts.33 Thus, our inability to enhance Iiver clearance of

radiolabeled uveal melanoma cells in nu de mice were already maximally activatcd, as

sllggested earlier. This proposition is consistent with the results from anti-asialo OM I

and cyclophosphamide treatments grollps. Both of these agents are known to inhibit

NK t' . . I I' t' I d 3233 TI 'k' .. f I' unctlOll 111 t le lver 0 t le nu e mouse: ,- - le stn mg unpmnnent 0 lver

clearance of radiolabeled uvealmelanoma cells anel the increased severity of hepatic

149

Page 150: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__________ CHAPTER 8

metastases in the eyclophosphamide and anti-asialo GM 1 groups support the

hypothesis that NT< eells restriet uveal melanoma metastases.

In summary, thc present findings indicate that human uveal melanomas differ

in their susceptibility to in vitro and in vivo NT< cell-mediated cytolysis. The degree of

susceptibility to NK cell-mediated cytolysis is inversely correlated with expression of

MHC class T antigen expression. That is decreased expressioll of class T antigens

correlates with illcreased NT< cell-mediated cytolysis. The results also indicate that

disnIption of in vivo NK activity can greatly increase the 1l1etastatic spread ofuveal

melanoma ceUs.

150

Page 151: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

NATlJRAL KILLER eELt SlJBSCEPTIBILITY OF UVEAL l\[ELA_N9"',,I~A,--__ _

REFERENCES 1. Egan K~'I, Seutlon nvl, Glynn R, Gragollda~ ES, Albcrt mvt Epidemiologie aspeets of uveal

lllelanoma. Sur\' Ophthalmol. 1988;32:239-251.

2. Jensen OA ivlalignant Illclanoma of the human uvca. Acta Ophthalmol. 1988;60: 161-182.

3. Rajpal S, Moore R, Karakousis CP. Survival in metuslatiL: ocular Illclanoma. Cnncer.

1983;52:52:334-336.

4. Char DH. l\klastatic choroidalmelanom:l. Am J Ophthalmol. 1978;86:76-80.

5. Donoso LA, Berd D, Augsburger JJ, Maslrangelo ivlJ, Shields JA.lvIelastatic m'eal melanomn:

prethenlPY scrum liver cn'Lyme and Jiver scan abnormalities. Arcil Ophthalmol. 1985; 103:796-798.

6. MuJe JJ, Slm SS, Schwmtz SL, Rosenberg SA. Adoptive immunotherapy of established pulmonaTy

metastases wilh LAK cells aml recombinant interleukin-2. ScÎence. 1984;225: 1487-1489

7. Y ran I, Wood TA Jr, Spiess PJ, Rosenberg SA. In virto growth of lllurine T cells: V: The isolatio!l

and growth of Iymphoid cc!1s in11ltrating syngeneic solid tumors. J IlllJl1Unol. 1980; 125:238-245.

8. Rosenberg SA, Lotze i\H, Yung JC, et uI. Experience \Vith Ihe Hse of high-dose inleTleukin-2 in the

treatment of 652 cancer patients. Ann Surg. 1989;210:474-485.

9. West WH, TaueT KW, Yanelli JR, el al. Constnnt-infusion recombinant interleukin-2 in adoptive

immunotherapy of advanced caneer. N Engl J Med. 197R;3 [6:898-905.

10. Ksander BR, Ruhs<lml'n PE, Olsen KR, Cousins SW, Steilein JW. Studies of tlllllor-infiltration

lymphocytes from a hllillan choraidal melanoma. IJlvest Ophthalmol Vis Sci. 1991 ;32:3198-3208.

11. Kan-Mitchcll J, Liggctt PE, Harel W, Steinman L, Nitta T, Okscnbcrg JR, PosnCf MR, Mitchc\l

1\'IS. Lymphocytes cytotoxie to llveal anti skin melanoma cells fram peripheral blood of ocular

lllclanoma palients. Cancer Immunolllllmunother. 1991 ;33:333-340.

12. Luyten GPI\'I, Moo)' CM, De Jong PTVM, Hoogeveen AT, Luider ThM. A chicken embryo model

10 study growth of lIveal rnelanoma. Biochclll Biophy.~ Res CO!l1rn. 1993;192:22-29.

13. Koo Gc, DunlOnt PJ, Tut! M, lIaekett J Jr, Kumar V. The NK-l.l (-) mOllse: A model to study

differentiation of lllurine NK ceIIs. J Immuuol. 1986; 137:3742-3747.

14. Algarra I, Öhlén C, Pcrez rvI, et al. NK sensitivity <lnd lung clearance of MHC-class-I-defident cells

within a heterogeneous fibrosarcoma. Int J Cancer. 1989;44:675-680.

15. Nicderkorn JY, Streilein nv, Shadduek JA. Deviant immune responses to allogeneic tumor injccted

intnlcamerally and sllbclltaneollsly in mice. Im'est Ophthalmol Vis Sei. 198 I ;20:355-363.

16. Quax PHA, van Mllijen GNP, Weening-Verhoeff EJD, cl al. Metastatie behavior of hu man

melanoma celllines in nude miee correlales with urokinasc-type plasminogen activHlor, its type-I

inhibitor, and urokinase-mediated matrix degradation. J eell Bio!. 1991; 115: 191-199.

17. Van tvluijen GNP, Janssen CPJ, Cornelissen JMAH, Smeets DPCi\I, Beek JLM, Ruiter DJ.

Establishment and characterization of a new l\llman melanoma ceilline (lVIV3) which is highly

metaslalic in nudc miee. Int J Cancer. 1990;48:85-91.

IR. Niederkorn JY, l'v1ellon J, Pidherney lvI, ~\'Iayhew E, Anand R. Effect of anti-ganglioside antibodies

on the llletastatie spread of intraoeular me1anom<ls in a nude mouse model of hll!11an u\'cal mehl11oma.

Cmr Eye Res. 1993; 12:347-35R.

/5/

Page 152: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_________ eH APTE~ ___________ _

19. Ljunggren H-O, Öh!én C, Höglunu P. Frankssoll L, Kiirrc K. The RivfA-S lymphumu mutant;

conscqucnces of a peptide luading defect on iIlllllunolugkal recognition and graft rejection.

Int J Cancel', 1991 ;6:3R-44.

20. Raming R, Sznlay 1. Ocular metaslasis of in vivo and in VIITO derÎvcd syngcneic omrine melanoma.

Invest Ophthalmul Vis Sci. 1987;28: 1599-160·:1.

21. Raming R. Kou Ge, Snlay J. Regulation of the metastasis of 111llrine oeulnr melanoma by natural

killer ee11s. Invest Ophthalmol Vis Sci. 1989;30: 1909-1915.

22. Niedcrkorn JY. Supprcssed cclltilar immunity in rnice harboring inlraocular melanomus.

Inycst Ophthalmul Vis Sci. 1984;25:447-454.

23. Niedcrkorn Jy. Enuc1eatioll-induced metaslasis of intraocular meJanomas in mice. OphthalnlOlogy.

1984;n:692-700.

24. Niedcrkorn JY. EnucJeation in consort with imlllunologie impairment promotes mctastasis of

intraocular mel<IIlOrnas in mice. Invcst Ophthalrnol Vis Sci. 1984;25: 1080-1086.

25. Pkl1cr IJ, GersiL'n Divt, Hart IR. Thc hiology ofenneer Învasion and melastasis. Ad\' C;lIlccr Res,

1978;28: 149-250.

26. Einhorn LH, Burgess i'vlA, Gottlicb JA. i'vlctastalic patterns of choroidalmelanomu, Canccr.

1974;34: 1001-1004.

27. Pidler IJ. Orthotopic implantation of hu man colon carcinomas inta nude miee provides a valuab1c

model tor thc biology and therapy of mclastasis. Cancer !vtelastasis Rc\', 1991; I 0:229-243,

28, StephensolI RA, Dinncy CPN, Gohji K, Ordonez NG, Killion H, Fitl1er lJ.1ktastatic lllodel fur

hllman prostate caneer using orthotopic implatantion in nude mice, J Nat! Caneer Inst.

1992;84:951-957.

29. FitIIer 11. The biulogy ofrenal canccr metaslasis, Semin Uro!. 19n; 10:3-11,

30. Gohji K, Nakajima 1' ... 1, Dinncy CPN, et al. Thc importanee of orthotapie implantation ta (he

isolation alld biological eharacteriLation uf a melastatic human c1ear eell renal c,1rcinoma in nude

mice. Int JOncol. 1993;2:23-32,

31. tvlanzotlÎ C. Alldisio RA, Pratesi G. Importnnee of orthotopic implantatian for hllll1an tumors as

model systems: Relevance ta metastasis and invasioll, CJin Exp Metastasis. 1 Y93;11 :5- 14.

32. Tzung S-P, Cohen SA. Endugcnous interferon alb produeed hy Kupffer eells inhibits interlellkin-I,

tumor necrosis faeor a produetion ant! inlerlelikin-2-inulleed aetivlition of Ilonparenehymal eel1s.

Cailcer Jmmunol Ill1munother. [991 ;34: [50-156,

33, Cohen SA, Tzung S-P, Doerr JU, Goldrosen ivlH. Rale of asialo-Gtvll positive Jiver eells from

athymic llude or polyinosinic-polycytidylic acid-treatcd rnice in suppressing eolon-derivcd

experimental hcpatic metastasis. Caneer Rcs, 1990;50: 1834-1840,

152

Page 153: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPTER 9

A CHICKEN EMBRYO MODEL TO STUDY THE GROWTH

OF HUMAN UVEAL MELANOMA

Gregorius P.M. Luyten, Cornelia M. Mooy, Paulus T.V.M. De Jong,

Andre T. Hoogeveen, Theo M. Luider

Bioehem. Biophys. Res Cam. 1993;192:22-29

/53

Page 154: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___________ CHAPTER~9 __

ABSTRACT

111 vilm cultured human uveal or skin melanoma eells were injeeted into the ehiekcn

embryonal eye at a stage when the immune system was not yet mature. The melanoma

eells \Vere aeeepted as part of the organism by the host. Even single melanoma eells

eOllid be traccd by morphologiealmethods as well as immunohistoehemical markers,

sueh as SIOO, HMB-45, NKI/C3 and HNK-l. We found tumours in 10 and 40 per cent

of the embryos injeeted with uvealmelanoma anti skin melanoma, respeetively. The

embryos did not exhibit abnormal development of the cye as a eonsequenee of the

mieroinjeetion and had a high survival rate (90 anti 60%, respeetively) during

embryogenesis. With this model for uvealmelanoma the growth and possible the

metastatic behavior of llveal melanoma cells can be stucliecL

INTRODUCTION

The development of ncw techniques 1'01' the treatment of uvealillelanollla has been

hampered by the diffieulties eneountered in the culturing of uveaimelanollla eells ill

vilra as weil as the laek of animallllodels to study the pathogenesis. Ta date several

eelllines of uvealmelanoma have been described. t,2 Albert anel associates were able

to culture hunulll uvealmelanoma tissue experimentally in nude miee by transplantillg

fresh (hu man) tumour tissue into the anterial' chamber orthe eye3 Kan-Mitehell and

a"oeiates were able to culture uveal melanoma eells in rabbit eyes2 The xenograftcd

'lnimals, however) nceded continolls immunosuppressive therapy.

The purposc of the present study is to develop an animal model to study the

growth of human mclanoma eelIs in the chicken embryonal eye without using

immunosuppressive agents. We injeeted ill vilm eulturecluveal or skin melanollla eelIs

into the eyes of chicken embryos before maturation of the immune system. The

injected eells were not rejeeted. We eleseribe a simple model for studying the growth

allel metastatic behavioul' of uvcal melanoma; the advantages of this model are

discussed.

154

Page 155: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ "-Ct-JICKEN Ei\IHRYO 1IODEL_ ~ ____ _

MATERlALS AND METHODS

Embryos

Fertilized eggs of GaI/lis g{/I/lls dOlllesliclIs we re obtained from alocal

supplier and incubated in a forced draught incubator at 3TC and 60 % hUl1lidity.

Embryos were staged according to the number of incubational or embryonal days (E).

From E2.5 (embryonal day 2.5) onwards, the eggs were not tilted any more.

Microinjection

At E3.5, the embryonal eye is loealized at the surface of the chorioallantoic

membrane and easily recognized because the retinal pigment epithelium has just

developed. The injection of melanol1la cells into the eye with a microinjeetion system

is therefore straightforward. Aftel' E3.5, the head of the chick moves farther away from

the surface and dips towards the bottom of the egg so that in vivo injection of cells

becOlnes technieally more complicated. Mieroinjection was perfonned with a glass

pipetle with an internal diameter of 50 mm. Each l1lieroinjection represented about

102 - 103 single eells of either skin or uvealmelanoma injeeted into the developing

eye (I - 10 mi injection).

Culture of skinmelanom>l

A human skin melanoma cellline derived from a Iymph node metastasis was

cultured in RPMI medium supplemented with 10 % heat-inactivated foetal ealf serum,

penicillin and streptomycin. This eellline was a fast growing monolayer culture,

which underwent passage every week at 1:4 dilution and was grown continuously for

more than two years. Confluent cell layers were trypsinized, washed with phosphate­

buffered-saline (PBS), centrifuged, and the resulting pellet was suspended in PBS (5 X

106 eells/ml) for mieroinjection.

Culture of uvealmelmlOma

A 46-year-old patiellt who had undergone enucleation for uvealmclanoma 28

years earlier had metastatic lesions in tbc liver, kidneys and subcutaneOlls tissue.

/55

Page 156: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ CHAPTER~9,,----________ _

Af ter obtaining informeel consent) a subcutaneous metastatic nodule was

excised when the palient was still alive anel the specimen was inullediately cultureel.

The metastalic tumours in livel" subcl1taneous tissue allel kidneys were consistent with

metastatic amelanotic and melanotic melanoma of the mixeel eell type (spinelle anel

epithelioid cells). The patient died within 3 months of diagnosis aftel' the clinical

diagnosis of metastatic disease. At autopsy na skin melanomas or alher primary

tumours were fOl\l1el.

The tumour of thc subclltaneous tissue was mineed with a pair of fine scissors,

irrigated continuously with Dulbecco's moelit'ied Eagle's medium and subsequently

strained through cheese cloth. The resulting suspension consisted mostly of single

eells anel to a les ser ex tent smaH c1ll111pS of epithelioid ceHs. The tissue was

eharacterizeel by a e1iversity of small ancllarge multinuclear cells. Most nuclei

contained multiple prominent nucleoli and numerous mitotic figures. The ceHs were

washeel twice with DMEM medium supplemented with 10 % heat-inactivateel foetal

calf serum, penicillin and streptomycin and seeeled in a culture Ilask at 5 % C02. A

fee der layer was not used.

The cultures, whieh consisteel of adherent and non-adherent ceHs, were

maintaineel in DMEM medium supplemented with 10% l'oetal cal I' serum. Cont1uent

cultures underwent passage every 2 to 3 weeks at 1:4 dilution. The adherent ceHs were

trypsinized, washed with PBS, centrifugcd. The peHet suspended in PBS (5 X 106

eells/ml) for microinjection.

Antisera

S 100,4.5 NKl/C3,6 HMB-457 and HNK- I 8 were used for the

inul1unohistochemical studies of both the cell cultures anel the histological sections of

embryonal chicken eye. The SI 00 monoclonal anti body recognizes an acidie

intracellular Ca++ binding protein and is a non-specific marker for melanoma ceHs9

Monoclonal antiboelies NKI/C3 and HMB-45 are specific markers for precursors of

melanoeytes anel melanoma eells. Monoclonal antibody HNK-l recognizes a family of

ceH ad hes ion molecules (CD57), migrating neuml erest cells and a series of neural

crest derivatives.

156

Page 157: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHICKEN El\lBRYO l\,10DF.,"-,L'"---___ _

IllllllUllohistochmnistry

The chicken embryo hcads were deeapitatecl at ElO and El9 and fixed in 10

percent formalinIPBS for at least two days, the eyes were enucleatecl, dehydrated and

embedded in paraffin, The sections, 5-7 mm thiek, were prepared by staining with

haematoxylinl eosin or by incubation with the primary anti body in a moist incubation

eh amber at room temperature for 30 minutes, Negative and positive controls for the

primary antibodies were included. For immunoperoxidase staining, rabbit-anti-mouse

peroxidase conjugated immunoglobulins were used as a second-step antibody.

EndogenOlIS peroxidases \Vere inhibited by incubation for 25 minutes in a

methanol/hydrogen peroxide (99: I viI') solution, Peroxidase was visualized with 0, I

% 3,3'diaminobenzidine HCL (Serva) and 0,01 % hydrogen peroxide. Sections were

counterstained with haematoxylin/eosin for one minute. PBS with 0.1 % Tween-20

was used for all rinses, Cytoeentrifuge preparations of cell cultures were analysed in a

man nel' similar to that used for histological sections.

RESULTS

Culture of skillmelanoma eells

The amelanotic cells grew as a uniform monolayer culture with severalmitotic

figures and possessed numerous lamellipodia,

Culture of llletastatic nveallllclanoma eells

One week aftel' initial plating of the metastatic melanoma cells about 40% of

the cells became adherent and nonadherent cells \Vere removed weekly. Aftel' 3 weeks

some adherent cells became rounded and had lost the ability to adhere. Some of thé

non-adherent cells (Le. 25% approx.) were vital, according to trypan blue cxclusion

analysis, and hence multiplied in suspension, becoming adherent again aftel' being

secded into ncw llasks. The culture was kept as a suspension culture and was grown

continuously for nine months, To date there have not been any signs of differentiation,

pigmentation or a decrease in the doubling time.

157

Page 158: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ CHArTER 9 --------.. -

Micl'oinjection ol' hunulll skin melanoma eells

Twenty cmbryos were injeeted with skin melanoma cells; twelve embryos

sllrvived and eight of thesc embryos developed an intraoclilar melanoma. The eyes

were removed rOl' histologieal evaluation on days EIO or E 19. The chicken eye was

macroscopically normal. Thc injection site hael undergone regeneration and cOllld not

be identified histologically. Histochemieal studies revealed roei of melanoma cells,

which appeared vital and mitotic figllres \Vere numerous. In all cyes, thc melanoma

cells grew on extra-cellular matrix structures in the choroid (FIGURE I). There was

no growth in the "itreous. At stage EIO, large numbers of infiItrating cells were

already present. The melanoma eells bOlllld the lllonoclonal antibodies SIOO and

HNK-l anel the lllonoclonal antibodies HMB-45 ano NKIJC3 were not bouncl

(TABLE I). Even single lllelanoma cells were easily cleteetable on clays EIO and E19.

TABLE 1 EXPRESSION OF CELL MARKERS ON IN VITRO CULTURED AND

XENOGRAFTED MELANOMA CELLS

8-100 HMB-45 NKIIC3 HNK-l

Metastatic uvealmelanoma H.d. ++ (100%) +++ (100%) ++ (100%)

Metastatic uveal melanoma ++ ( 20%) ++ ( 70%) ++ ( 70%) +++ ( 50%)

cultured in \'itro

Uveal mclanoma ,xcnografted into + ( 60%) +++ (100%) + ( 70%) H.d. chkken cmbryonal eye

Skin melanoma cultured + ( 90%) + (100%) + ( 90%) +(100%)

in I'ilm

Skin melanoma xenografted into ++ (100%) - ( 0%) - ( 0%) + (100%)

chicken cmbryonal eye

ll.d. ::: Ilot detcrmined; Intensity of staining: -::: 110 staining; + = slight staining; ++ = intense staining; +++ = very intense (saturatcd) staining; percentage in paracenteses is the numbcr of positive eells observcd,

158

Page 159: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

-_ ... _---- CHICKEN Ei\IBRYO i\10DE",,,L ____ _

c

lA

159

Page 160: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

160

CHAPTER 9

B FIGURE 1.A and IB. Growth of skin melunolllu cells injected into thc eye on E3.5 and anulysed on El 0, the tumour cells bind intensely the S 1 00 anti body (a) and E 19 (b). Infiltration of melanoma cclls are observed in the choroid (c), betwcen the carlilagc ossicles (0), and extra-sclerally (es). (Magnification 320, 420x, respectively)

Page 161: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

Micl'oinjectioll of human metastatic uvealmelanoma eells

Eighteen out of lwenty injected embryos survived wilh two embryos (E 19)

exhibited intra-oeular lumour growth (FIGURE 2). On histologieal examination lhe

ehicken eye had developed normally and the injeelion site eould nol be rceognized.

Clones of metastatic uveal melanoma cells were seen adherent to lhe eiliary body and

along the hyaloid artery which is very pronounced in the chiekcn embryo compared to

human ernbryos. The tumor nodules derived from uveal melunoma wcre smaller than

those originating frOlu cutaneous melanoma.

FIGURE 2. Growth ofuveal hlllllUl1l11Clallomu cells along thc ciliary body in an E19 eye. The tumuLIr cells stain intensei)' wijh the Hr ... 1B-45 marker (arrow). (Magnitïcation 750x)

161

Page 162: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_______ . ___ .. __ ------"HAPTI~. ___ . __ . __ .. __ . __

Immunohistochemistry

We eletermineel the expression of four markers (SlOO, HMB-45, NKIIC3, anel

HNK-I) on skin anel uveal melanoma eells from thc patient, on il1 vitra cultureel eells

and on the xenografteel l11e\anoma eells (TABLE I). Some examples are illustralecl in

FlGURE 3. For both skin and uveal melanoma we found a difference in marker

expression which elepenelecl on the growth system used. Skin melanoma eells in

parlicular lost all marker expression for HMB-45 anel NKT/C3 when xenograftecl into

the ehieken.

FIGUnE 3A. Human uveal melanoma meta,Stasis in subcutaneous ti.ssue stained with HMB-45 lIveal melanoma (u) infiltrating lymphocytcs (i) (Magnificalion 150x)

/62

Page 163: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____________ Ç:HICKEN. EI\InRYO J\fOD,-,E~L~_

In contrast to skin melanoma, xenografted hUI1Ulll uvealmelanoma retaincd the marker

expression for HMB-45 and NKJ/C3. Thc SI 00 marker revealed na large differenees

in marker expression for either type of melanoma irrespeetive of growth conditions.

We conclude from the restJlts summarized in TABLE 1 that xenografted human uveal

melanoma cells can be traeed on histological sections or the chicken eye with the four

markers and that, althollgh the expression ditl'ered depending on thc

microenvironment, marked change in morphology of the xcnograftedllveal tumollr

eells does not oeeur.

FIGUUE 3D. Cytoccntrifuge preparations of llveal melanom<l cells staÎncd with HNK-l. AbuLlt 50% of cells are positive, a divcrsity of eeIl types ean be scell. Sevcral eells are multinucleated (arrow). (Magni1ïcntion 460x)

163

Page 164: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

C HAPTE R 29'---___________ _

DISCUSSION

The prognosis of primary uveal melanoma has not improved despile rccent introduced

. d" I' 10-12 I . d I I f I" I I' d Unl latIon tee lmques. t IS assume t lat t lC presence 0 C ll11ca (ISeaSe an

micro-metastasis at the time of treatment of the primary tumunr is the main cause of

thc low survival rates,U Uveal melanomas grow slowly compared ta cutaneous

mclanomas, metaslasis is mostly observed 5 year aftcr the clinicalobservation of the

primary lumour. After the clinical detection of the metastatie disease, in general,

. d' . I' 1415. I d' . I .. f I" I 16 patlents Ie wit 1111 a year, ' me u lllg 111 t lC mHJonty 0 cases Ivcr mva vement.

The paucity of animalmodels for hu man uveal melanoma has limited progress

in our underslanding of the aeliology, tumourigcnesis, tumour progression, and

eventual trealment. Since knowlcdge about lhe groWlh and metastasis of uveal

melanoma is Iimited, lhe developmenl of efficient diagnostic and therapeulieal

modalities is slagnating. Until recently it was difficu1t 10 culIure primary and

metastatic uvealmelanoma in vitro. Albert and assocÎales used a feeder layer and

epidermal groWlh faclor to support cell attachment and groWlh. I The use of epidermal

growlh factor in combination with cholera toxin had a synergistic effect since

conlaminating eells were inhibiled and proliferalion of thc 1l1elanocyles was

still1ulated. Kan-Mitchell and associales did not use these additives in order 10 selecl

the most aggressive of lhe grO\ving melan01l1a eells as lhe poor culture medium was

sufficient to select aggressive cells2

\Ve cultured cells of a metastatic uveal melanoma containing fast-growing

epilhelioid eells with many 1l1ullinucleatcd cells. Except for a different mincing

procedure and lhe use of DMEM medium inslead of RPMI, lhe culluring procedure

was lhe same as that described by Kan-Mitchell and associales2 We obtained a cell

culture which eontained heterogeneous morphological cell types. We were able to

culture the melastatic uveal1l1elan01l1a cclls for at least 9 months. Seventy percenl of

the eens were positive lor two melanoma markers (HMBA5 and NKI/C3) and 50

percent wcre ]lositive for HNK-I.

The injeclion of primary tumour ce lis into lhe anterior chamber and vilreous

cavily of nude miec is described but no follow-up studies have been published3

Recclltly Kan-Mitchell anel associales describcd a rabbit moclellhat was suitable for

164

Page 165: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHICKEN E~IBRYO i .... /ODEL ------~

studying the growth of uvealmelanoma, in whieh uvealmelanoma was growll in the

rabbit anterior ehamber2 A disaclvantage of this animal model is thc daily injeetion of

cumulative doses of al1 inullunosuppressive agent.

We describe a new model for studying the gwwth of human uvealmclanoma

eells, which uses the chicken embryo, Aftcr eulturing, either skin or uveal melanoma

eells were injeeted into the developing eye of chicken embryos before maturation of

the immune system, Bath melanoma eell types grcw in the developing embryonal eye.

Skin melanoma eells grew l11ueh fastcr in culture and in the chicken embryonal eye

than nvcal melanoma eclls. The observed foei of bath tumour types were alive; mitotie

figures were frequently fOlmd implying aetive growth. The behaviour of the tumour

eells in the chicken eyc was rathcr consistent. Thc tumour eells grew along the tunica

vaseulosa of the lens, in the ehowid and along the sdem. Apparently, thcse eells

needed a matrix to start growth. At El91arge infiltmting tumours of skin melanoma

were observed. Probably due to the slower growth of uvealmelanoma, comparabie

infiltrations of uveal melanoma eclls were not found at E19. In general the uveal

melanoma eell culture grew mueh more slowly in the cmbryos and a mueh lower

percentage of embryos had histologieally deteetable uveal tumours.

Both, uveal and cutaneous tUIllOur eells, could be recognized easily on

histologieal seetions of thc chicken embryonal eye by means of their Illorphology and

by antibody markers. Separate tumour cells could be readily traeed. In this study the

skin Illelanomas eould be tmeed with SI 00 and uveallllelanoma cells with the

melanoma marker HMB-45 and to alesser ex tent with NKl/C3 and SI 00. Therc was a

differenee in expression of the markers between the observed tUlllour foei and the

injeeted eells, indieating that clonal selection had oecUlTed in vivo resulting in a

different marker make up. Another possibility is that the altered marker expression is

environmen t -dependent .

In addition to the potential to study the growth and metastasis of hUlllan uveal

melanoma cel Is in an environment that tolerates xenografted eells, this model has

several advantages: a) low cast b) large series can be studied e) easy teehnique d) high

survival rate l'or treated embryos e) the possibility of screening therapeutical agents.

Another advantage is that funduseopy in hatehed chicken and adult ehiekcn can be

performed. We hope this cxperimental model wil! contributc to the understanding of

the tU1l1ourigcnesis of uveallnelanoma and the metastasis of this neoplasm.

165

Page 166: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

----- _______ ~ A PTER _9 _________________ _

REFERENCES

1. Albcrt Di\[, Ruzzo ~,rA, r\'lcLaughlin I\'IA, Robinson NL, Craft JL, Epslcin J. Establishment ofccillines of

uveal rnelanomas. Illvest Ophthal Vis Sci. 1984;25: 1284-1299.

2. Kan-Î\'litchell J, ivlitcheU rVIS, Rao N, LiggclI PB. Characteriz,\lion of llveal melnnoma celllines that grow as

xenugrafts in rabbi! eyes. hlvesi Ophthal Vis Sci, 1989;30:829-834.

3. Albcrt mvI, Shadduck, JA, Lil! H, Sllllderl11an Jr FW, Wagoner :1,'10, Dohlman l-IO, Papale, JJ. Animal models

fOf Ihc stud)' uf llveal melanoma. Int Ophthalmol eHn. 1980;20: 143-16U.

4. r.'Îoorc HW, A soluble prolcin charncteristk of nerVütls system. Biochem Biophys Res Comrn. 1965; J 9:739-744.

5. NakajinHl T, Watanabe S, Satoh Y, Kameya OT, Hiruln Y, Shimosato Y. An iInIllulloperoxidase study uf S-\ 00

protcin distribution in normal am\ neoplastic tissue. Am J Surg Pathol. 1982;6:715-727.

6. !\'lacKie Rivl, Campbelll, Turbitt l\lL. Usc ofNKIC3monudonai anti body in the assessment ofbenign and

maligant melanocytic lesions. J CJin Pathol. 1984;37:307-72.

7. Colombari R, Bonetti P, Zamboni G, et al. Distribution of mclanoma specilk antibudy (HivIB-45) in bcnign

and lllalignant melanocytic tumours. Virchow An:h lA]. 19H8;413:l7-24.

8. Aho T, Balch CI\t A dillèrentiation antigen of human NK and K cclls identified by a monoclonal antibody

(HNK-I). J Illlmunul. 1981;127:1024-1029.

9. Gaynor, K, Herschman, H.R., Irie, R., et al. S-100 protein: a marker for human malignant mclanomas?

Lancet. 198\;1:869-71.

10. Loml11atzsch PK. Rcsults a!'ter b-irradiation (l06 Ru/106Rh) of choroidal melanoll1as; twenty years'

xpericnce. Br J Ophthalmol. 1986;70:844-H51.

11. Sedtlon JM, Gragoudas BS, Egan Ktvl, Glynn JU, Howard S, Pante RG, Albert DM. Relative survival mtes

a!'tcr alternativo therapies for uveaimelanollla. Ophthalmology. 1990;97:769-777.

[2. Shields JA, Shields CL, Donoso LA. 11anagement of posterior uvealmelanoma. Smv Ophthalmol.

1991 ;36: 161-195.

13. lvJansehot WA, Van Pcperzee1 HA. Choroidal mclanmna; cnucleation or observation? a new approach.

Arch Ophthalmul. 1980;98:71-77.

14. Wagoncr MD, Albcrt D~·1. Thc incidcnce of metastase from untreated ciliary body and choroidalmclanoma.

Arch Ophth;t1mol. 191\2; 100:939-940.

IS. Pach Ir,,!, Robcrtsoll DM. r-,'letastasis from untreatetl u\'ealmelanomas. Areh Ophthalmol. 1986;104: 1624- [625.

16. Jensen OA. i'vlalignant melanomas of the human u\'ea: 25 years follow-up of cases in Denmark, 1943- [952.

Acta Ophthalmologica. 1982;60: 16\-lH2.

166

Page 167: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHAPIfER 10

DISCUSSION AND FUTURE PERSPECTIVES

In this final Chapter we will turn back to the basic questions that fornled the

background to the studies presented in Chapters 2 to 9. We wil! discuss what these

studies have revealed concerning these questions, and propase what direction the

investigations should take next in order to arrive at a therapeutically more successful

approach of uvealmelanoma. As was explained in section 1.2, the present thesis set

out to investigate same of the key elements involved in the metastatic process of this

type of cancer, with the aim of providing insight in how to treat uvealmelanoma­

related metastatic disease as promptly as possible in order to decrease the overal!

mortality rate of 50 %. The three basic points of interest were directed at the origins of

the metastases, differences between primary anel metastatic lIveal melanoma, and thc

treatment (and prevcntion) of metastatic disease (see 1.2).

Firstly, we will take another look at the subject of enucleation versus

radiotherapy of the primary tumor-containing eye, that was introduced in section 1.1.2

(Chapter I). As was inclicated there, the relative indication 1'01' enucleation or

radiotherapy is not directed by the relative risk for metastases but by tumor contra!.

Preliminary results do not show a difference in survival bet ween either of these

therapies. A possible demonstrated difference in survival bet ween patients treated by

either one of the therapies cannot be expeeted within 6 years aftel' initial treatment

(Manschot, 1992). The COMS investigation will study this problcm in a large group of

patients with medium and large sized tumors. However, among these patients

specifically there is a relatively high mortality rate, and the majority of them is

probably not curable by applying any type of local treatment. In this context the

question rises whether not paticnts with a 8111all primary tumor, who in most cases arc

potentially curable, should be included in such a study. Apparently, regarding the

ascertaining of the safety of conservative treatment modalities, especially long-term

studies of cases of small primary tumors should be rewarding. On the other hand

prampt treatment of smallmelanocytic lesions willlead to overtreatment of potentially

167

Page 168: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ , __ . __ CHAPTE~ ________ . __ . __

benign lesions. In tlle meantime, that is, as long as these studies have not been

pcrformcd, it is guite certain that radiotherapy will often be applied in selected cases

in which vision can be prcserved and total tumor con trol can be established, as

radiotherapy with cither charged particles or episcleral plaques has become the first

choice of trcatment of uvealmelanoma patients in most of the large ophthalmic

oncology centers (sec 1.1.2). In addition, as was mentioned in the same paragraph, the

relativc advantage or lllerc observation in cases of sl11ull primary uveal melanomtls,

advocated by sOllle authors, appears questionable as long as we are not certain abuut

the long-term survival of patients having such a smallmelalloma (Shields, 1995).

Turning to the guestion of thc origins of metastases, in Chapter 2 the

controversial issue of prc-enucleation irradiation was dealt with (see also I. 3.1).

Zimmerman's hypothesis (1978) th at the high mortality rate in the secOlld to fourth

year following enucleation is the result of micro-metastatic secdings induced by the

physicalmanipulation of the tumor-containing eye during the surgical procedure does

probably not holc!. As Manschot pointed out, the high mortality peak might also be

explained by the natural history of uveal melanomas, in which case this peak \Viii be

the result of micro-metastatic seedings beforc enuc1eation; he argued, based on

doubling-times, that a dircct elfcct of cnuc1eation itself can only be seen artcr 6 years.

Therefore he advocated the pre-enuc1cation radiotherapy to study this potentially

metastatic effect of the surgical procedure itself. Clmpter 2 showed that pre­

enucleation radiotherapy in the first long-term follow-up study (7.5 years) did in fact

not demons tra te a beneficial effect on survival, which makes Manschot's hypothesis

on the origins of the micro-metastases more plausible. At any rate, a 10 to 15 )'eat'

follow-up study should settie this issue at an even more retiable levcl. As mentioned in

Chapter I and in continuity with the preceding paragraph, it is interesting from a

theoretical point of view that especially cases of small metanomas will (possibly)

benefit from pre-cnuc1eation irradiation because these can potentially be cured, in

contrast to cases of large tumors, for the latter are more liable to already having given

way to micro-metastatic seedings before the initial treatment (Hungerforel, 1993). With

regard to the survival rate of patients suffering T I tumors, the study reported in

Chapter 2 sholVeel a non-significant difference between the pre-enuc1eation irradiation

anel the control group, but this T I follow-up group was actually too slllall to answer

168

Page 169: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

DISCUSSI0l'{ AND FUTURE yERSPECTIVES_

this question. For that matter, the follow-up of patients, who have been treateel by pre­

enucleation raeliotherapy will be continued.

On the morphology of uvealmelanoma eells, Chapter 3 reporteel that primary

anelmetastatic uveal melanomas e1i1Ter signifieantly in eell type: metastatic lesions

consist of more aggressive, unelifferentiated tumor eells than the primary tumor they

derive from. In clinical practiee, this is important mainly with respect to patients

having uneli fferentiated liver metastases with unknowI1 primaries. In such cases the

relatively high percentage ofI-IMB-45 staining of uvealmelanoma-related metastases

can be helpful in differentiating these from ot her non-melanocytic lineage derived

tumors. Apart from this practical siele, it is interesting to hypothesize on the

differences in cell type bet ween the primary anel the metastatic lesions. On the one

hanel, it can be expecteel that epithelioiel cells are more metastatically potent than

spindie cells, because the epithelioid cell type is associated with a higher mortality rate

(Luyten, 1995; Gamel, 1993, Coleman, 1993). On the ot her hand, however, primary

uveal melanomas eonsisting purely of spindle eells are ab Ie to metastasize as weil.

Furlhermore, the cells of one metaslatic lesionmay show a epithelioiel-like

morphology while cells of another lesion in the same patient may display a spindie

cell pattern (sec Chapter 3). Whether these morphological changes are relaled la the

genet ic instability of the tumor cells and thus with tumor progression or to micro­

environmental changes is unknown. More likely is however that these morphological

changes are to a certain ex tent the result of genOluie abberations anel accur in

concordance with their intrinsic melastatic potency and proliferation rate. Whether or

not these changes can be explained by malignant transformation of differentiated

melanocytes, which will de-differentiate or by malignanl transformation of immature

stcm cells, which will differentiate to a certain extent, is unclear for uvealmelanomas.

The fact that an increasing number af uveahnelanoma patients is nowadays

treated by raeliotherapy implies that histopathological tumor material is often not

available for analysis. Therefore, prognostic factors, will have to be deduced mainly

from clinical observations. In sunllnary of section 1.3.2, investigations that were

performed to ielentify patients at a high risk of developing metastatic e1isease, carried

out in groups of palients treated by enucleation or radiotherapy anel in groups of

patients with sm all dormant anelnon-dormant uvealmelanomas, elucidated the

169

Page 170: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_____ ----'C"'HCCA 1''1' E RIO

following main clinical prognostie factors: largest tumor diameter (LTD), anterior

location anel prCSCllce or extraocltlar extensÎol1 of thc tumor, and advancecl age at the

time of the initial treatment andmale sex of the patient. or patients treated by

radiotherapy or other conservative treatment modalities, additional histopathologieal

parameters su eh as eell type anel proliferation markers (MIE I, MYC) could

hypolhetieally be obtained by means of fine-needle-aspiration biopsies (FNAB).

However, these parameters eoulelnot be used as prognostie factors anyway for these

cytologie al specimens are not representative for the wllDle tumor.

Another possibility to identify patients at risk as pointed out in 1.3.2, is the yel

hypothetieal possibility or the early detection of disseminated tumor eells, whieh

foremostly would be a great advanlage with respect to the ehoiee or to whether or not

assess systemie treatmen!. To deteet these eells, the minimal detection of tyrosinase in

peripheral blood by the polymerase-ehain-reaetion technique mighl be the solution

(Tobal, 1993), despite recent eontraelietory results (Foss, 1995;). Still it appears that the

minima I dctcctioll of a panel of circulating melanOl1HH;pecific anti associated antigens

in peripheral blood could possibly prove even more sensitive than that of tyrosinase

alone. Apart from this, ilwould be a eonsiderable gain if ultrasonic leehniques eould

be developed to determine the presenee of vaseular networks, whieh has been shown to

be a highly significant histopathologieal prognostic factor (Damms, 1995).

The value of anolher possible prognostie factor in uveal melanoma, that of

NM23 expression, was investigated in Chapter 5. In this study it appeared that the

association betwcen immunohistochcmically assessed NM23 expres sion anel survival

is only of "borderline significanee". The hypothesis that NM23 is related with

metastatie behavior in uvealmelanoma is supported by another study. Un[)ublished

data of Ma anel associates on llveal melanoma celllines transplallted intracamerally in

nude miee report a st rong inverse correlation between the level or NM23 mRNA

expression and NM23 antigen expres sion on the one hand and the development of

metastases on the other. Fmthennore, the sUlvival time of the nude miee in this study

shows to be eorrelatecl with the NM23 gene expression (Ma, 1996; submitted for

publication). At any rate, as was rurthel' mentioned in Chapter 5, the

inullunohistoehemical assessment of NM23 antigen expression might not altogether

be that inrormative, for it is likely that the NM23 gene is involvecl in the metastatic

170

Page 171: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

DISCUSSTON __ AND FUTURE PERSPECTIVES

process at another, transcriptional or post-transcriptionallevcl (Postel, 1994; id 1993).

Consequently, in order to determine the precise prognostic value of NM23,

investigations that focus on this particular level should be undertaken.

In vivo as weIl as in vitro studies show, turning to another subject of section

1.1.3, that in most cases of ll1etastatic uvealll1elanoma the liver is involved (Kath, 1993;

Niederkorn 1993; Rajpal, 1983;), whereas eutaneous melanoll1as preferentially

metastasize to the lungs and other orgäns (Einhorn, 1974). This difference in organ­

specific metastasizing between the uveal and cutaneous melanoma suggests that site­

designation in metastasis is not an at random cvent. The predilcction of certain tumors to

metastasize consistently to specific organs must be the result of organ-specific homing

and hence, reeognition of vascular endothelial receptors unique to those organs by

circulating tumor cells (Niederkorn, 1993; Nicolson, 1988). The association between

high NCAM expres sion with both rapidly metastasizing primary uvealll1elanomas and

ll1etastatic lesions themselves, demonstrated in Chapter 4, shows th at cell ad hes ion

molecules may play an important role in the homing capacity of disseminating uveal

ll1elanomas. Furtherll1ore, lack of HNK-I expression in the metastatic lesions was in this

study proved to be strongly associated with liver involvement, which may be an

indication that the NCAM isoforms lacking the HNK-I epitope is involved in the organ­

specific metastatic behavior of uveal melanomas.

We willnow turn to the experimental studies presentcd in the latter half of the

thesis. The establishment and characterization, described in Chapter 6, of two primary

and three metastatic uvealmelanoma celllines - of which until recently indeed only

very few respectively none were available (see 1.3.3.1) - opens the possibility to initiate

studies for a better understanding of the fundamental processes regulating the uveal

melanoma metastatic phenotype and the development of models that accurately reflect

the pathogenesis of ll1alignancy and predict responsiveness of uveal melanoma in

humans. Thereby these celllines might open the door to the development of new

therapeutical modalities of metastatic uvealmelanoma, which was the third and

eventually main aim of this thesis. Tn fact, apart from studies by the au thor and

associates, these eelllines are nolV being used by seventl other research groups

conquering various problems on carcinogenesis, tumor progression and tumor

inul1unology.

171

Page 172: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

elIA PT E" _ec1écO'--__________ _

As was elescribeel in Chapter 7 (sec also section 1.3.3.2), these celllines were

used to determine the expression of the melanoma-specific and associated antigens

MA GEI , -2 alld -3, GP 100 and tyrosÎllase in uvealmelanotlla. Two celllines that were

founel to express all five antigens, OCM-l and OMM-l, could for th is reason probably

be applieel in specific active immunotherapy, analogue to metastatic cutaneous

melanoma, which is now experimentally treated by means of inlll1Unotherapy

consisting of repeated injections of irradiated tumor cells expanded in tissue cultures.

These cutaneOlls melanOlna vaccines have proved, al least in I11urine moclels, ta i11Cleed

be able to inuuce a long-Iasting specific anti-tumor immunity (Dranoff, 1993). In order

to stimulate systemic anti-tumor activity, tumor cells deelicated for inlll1Unotherapy

have to be transfectecl with either 1L-2 or GM-CSF (Dranoff, 1993). Because the

wllOle procedure of culturing biopsiecl melanoma cells and carrying out the cytokine

gene transfection anel irradiation is rat her time consuming, some investigators of

cutaneous melanoma immunotherapy make use of allogenic insteacl of autologous cell

lines (Stevens, 1995; Mitchell, 1994). Accorelingly, as the survival time of uveal

melanoma patients with metastatic disease is limited, these patients will probably also

benefit most from allogencic tumOl' vaccines.

However, berore the celllines OCM-I and OMM-I can be used in practice for

specific immunotherapy, several aelditional studies will have to be carried out,

especially on the susceptibility of these celllines to the various vectors of the immune

systcm anel vice versa, on tbe reactivity of these vectors to thc celllines, for insta nee

in animalmodels (see sections 1.3.3.2-3). One such stucly was presenteu in Chapter 8,

which is also the secOlld example of the use of our cell lines. Jn this study, the cell

lines were transplanteel intracamerally into nude mice, anel thus it was showed that the

elown-regulation of MHC class I molecules in uveal melanoma cells raises the

susccptibility of these cells to natural killer (NK) cells. In other words, when CTL­

antitumor response is not ineluceel by lack of an appropriate MHC class I molecule,

uvealmelanoma cells still induce an immune reaction in the sense that they arc

recognized anu killeel by NK's. Whether it might inueed be possible to employ NK­

eell activity in the treatment of patients suffering distant uveal melanOlua metastases,

other than with a genera! stimulation of the cellular response by 1'01' example IL-2, is

yet unc1ear.

172

Page 173: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

DISCUSSION A.~D FUTURE PERSPECTIVES

In the lïnal Chapter, Chapler 9, thc eelllines were used in a novel chicken

embryo model in whieh growlh and metastalic process of uveal melanomas ean be

studied. The basic idea of developing this model is that human uvealmelanoma eells

ean grow in the eye of an immunoeompelent anima!. The human eeHs have to be

transplanled at an orlhotopie or a heterolopic site before the chicken embryo's immune

syslem has developed, that is amply before day 9; in this model, the human eells laken

from lhe uvealmelanoma eelllines were transplanled inlo the eye of lhe chicken

embryos at day 3.5. Although a high sueeess rate of lransplanted tumor eells was

reached, the eventually attained number of halched chicken with a growing intraoeular

lumor was relatively 1011'. Therel'ore, although lhis modelmight be useful to the study

of intraoeular growth of human uvealmclanoma cells, it is probably not reprodueible

enough to be elaborately engaged in the study of metastatie proeesses and prevention

strategies. Better results were obtaincd in another study, when lhe ceHlines were

transplanled into l1\lde mice in collaboration with Niederkorn an assoeiates (Ma, 1995,

id, 1996; submitted for pubIkation).

In conclusion, regarding thc therapeutic approach of uveal melanoma, it is

essentiallo reeognize those patients wilh a high risk of developing lethalmetastatic

disease al lhe time of presentation with a primary uvealmelanoma. With respect to

this, at present two 10pics appeal' to be of main importanee. Firstly, the evaluation of

more sensilive prognostic factors, so thallhe ehoice of treatment may appropriately be

tuned to every individual case. The prognostic factors that are seemingly most worthy

of further elaboration are the presenee of elosed vaseular nelworks, the carly deteetion

of disseminated tumor eells and the organ-specifie homing capacities of tumor eells.

Seeondly, evidently lhe treatment modalilies themselves have 10 be further developed.

As was argued above, regarding these, aelive speeifie immunotherapy for metastatic

diseasc or as an adjuvant therapy by use of lhe eeHlincs OCM-I and OMM-l appears

to be a promising oplion. Before this type of immunotherapy might be introclueed,

apparently more studies are required on uvealmelanoma eell Iines in relation to the

immune system, first of all direeted to the Cjuestion whether cytotoxie T-Iymphoeytes

(CTL) in deed rceognize and kill these tumor eeHs. Sueh sludies on CTL reeognilion

and the ability of CTL's to mediate anli-tumor responses are at present underway.

When a speeifie aelive immunotherapy wiH have been developed, its effects on the

173

Page 174: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

CHArTER -"-'0'--_____________ _

survival of patients suffering metastatie disease should be assessed in a clinical trial. If

an immunol1l0dulation ean indeed be indueed and the side-effects of sueh <i therapy

are minimal, specific active imlllunotherapy might in the future be applied as an

adjuvant therapy to eases of prilllary uvealmelanoma as weil. When the effects of

sueh adjuvant iml11unotherapy, especially with regani to the down-regulation of MHC

class I and TI molecules after radiotherapy, will stand the eomparison with those of

enucieation and radiotherapy, we may have taken a step towards a decrease in uveal

lllclanOlua-related mortality.

174

Page 175: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

SUi\l"''ÏARY

SUMMARY

Uveal melanoma is the most common primary inlraocular malignant tumor in adults

with an estimated annual incidence in The Netherlands of about 80 to 100. The

introductional Chapter points out that to date studies on this type of cancel' mainly

focused on diagnosis anel treatment of the primary tumor. This has reslllted in the very

low miscliagnosis rale of primary uvcal melanoma of less than I % today, owed to the

application of eliagnosis techniqlles like ultrasonography, fluorescence angiography

anel especially inelirect ophthalmoscopy. The choice of how to manage primary uveal

melanoma is quite eontroversial, anel often but to a limiteel extent directed by long­

term survival chances of the patient. In addition to the initially only available

lreatment of enucleation of thc primary tumor-containing eye, several alternative

(eonservative) treatment moelalities such as photocoagulation, local reseclÎon anel

principally radiotherapy have been developed with the major interest of retaining

visual acuity and tumor eontrol. However, despite all efforts to develop new therapies,

still at least 50 % of the patients with uvealmelanoma eventually dies of distant

metastatic e1isease. Once such metastases have been diagnosed, the life expectancy of

the patient is very low (2 - 9 months) as metastatic uvealmelanoma, which is mostly

located in the Iiver, is a particularly malignant disease 1'01' which practically no

rewarding treatment is available yet. Only a few therapies which inc1uded

chemotherapy have proved to extend the median survival time of the patients with

about 3 months. On the whoie, survival percentages a"essed by Kaplan-Meier curves

anel the Cox analysis are 75%, 65% and 55% aftel' respectively 5,10 anell5 years,

with a peak of melanoma-related deaths in the secOlle1 to fourth year following

treatment of the primary tumor.

The present thesis investigates same of the key elements involveel in the

metastatic process of uvealmc1anoma with the aim of contributing to the development

of a therapeutic approach that shOlIld decrease the high overallmortality rate of 50 %.

The three basic points of interest are the origins of the metastases, the formulation of

clinical and histopathological factors involveel in the metastatic proee", and the

treatment (and prevention) of uveal melanoma-related metastatic e1isease.

Chapter 2, regareling the nrst point, elaborates on the hypothesis that micro-

175

Page 176: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ -"SUM1[ARY

metastatic spreading is not induccd by the physical manipulation during the

enucleation of the primary tumor-containing eye, as Zimmcnnan suggested (1978),

but is due to micrometastasis prior to treatment, as Manschot pointed out (1980).

Evidence in favor of Manschot's hypothesis has already been given by several studies

on the effects on survival of pre-enucleation radiothcrapy. The present first long-term

follow-up study (7.5 years) demonstrated no difference in survival between a group of

patients treated by two fractions of 4 Gy prior to enuclcation and the control group,

patients who were treated by enucleation alone. This makes Manschot's hypo thesis

that micro-metastases originate independent of treatment even more Iiable. However,

to set tIe this issue another study with a follow-up of 15 years is required.

Previous investigation into the identification ofpatients at a high risk of

developing metastatic disease, our secOlId point of interest, clucidated the following

main clinical prognostic factors: largest tumor diameter (LTD), anterior location and

presence of extraocular extension of thc tumor, and advanced age at the time of initial

treatment and male sex of the patient. Notably, histologicalmaterial is of ten not

available now with the increase in conscrvative trcatment modalities, so that

prognostic factors indeed have to be deduced mainly from clinical observation.

In the studies presented in Chapters 3 and 4, ceHs from primary and metastatic

lIveal melanomas \Vere compared. A significant difference in l1lorphology was

demonstrated: metastatic tumors display a higher percentage of aggressive and

undifferenliated cells compared to lhe corresponding primary tumor. Furthennore, it

was found lhat staining by HMB-45 is the mos I sensitive immunohislochemical

marker for this type of melanoma. Next, the presence of neural ceH adhcsion

molecules in the primary and metastatic uveal melanoma cells was detennined. The

presence of thc adhesion molecule NCAM in bath aggressive, metaslasizing primary

and metaslatic lesions and tlle relalive paucity of HNK-I expression in metastalic

lesions of lhe Iiver together suggesl lhat the expression of lhe NCAM isoforms lacking

the HNK-I epitope might be an important faclor in the organ-specific, Iiver-directed

homing of uveal melanoma. These dala show that detailed expression of adhesion

lllolecuies in prilnary uveal melanomas lllay have prOlnising prognoslic value.

Chapter 5, mUIlogous to research on (mainly) mamma carcinoma, studies tlle

prognostic value of the expression of NM23, which has been deseribed as a putative

176

Page 177: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

__ ---"s"cu M 1'11 A R Y ____ _

metastasis suppressor gene. Howcver, in the present study on lIveal melanomu the

prognoslic value of the protein expression in lhe primary lesion appeared to be only of

"borderline significance". Yet, as the function of the NM23 gene and its produCls is to

date unclear, ils major function might be appointeel al (post-)transcription level, whieh

then has to be investigaled in anolher type of study.

The latter half of the thesis is dcelicateel to experimcntal studies. To date,

studies on biological beha vi or and therapies of uvealmelanoma, the third and main

aim of this thesis, have been hampered by a paucity of llvcal melanoma celllines.

Chapter 6 elescribes lhe suecessful establishmenl of two primary anellhree metastalic

uveaimelanollla cell lines, lhal were subsequcnlly characterized by light anel

electromicroscopy, inlll1unohistochemieal anel cylogenetic analysis and then used in

the several studies discussed below.

The study reported in Chapter 7 showed that two of these cell lines (OCM-l

and OMM-I) express the lhree MAGE, the fyrosÎllase and the GPlOO genes. Because

they express MHC c1ass I molecules as weil, lhese eelllines Illight be employed in

allogenic lumor vaccination, similar 10 illlll1unotherapies currently app!ied to palients

suffering mctastatic cutaneous melanoma. In cases of metastatic uveal mclulloma,

considering thc !imiteel survival time, lhe use of autologOlIS Illaterial appears Ie"

feasible. 11' such specifie aelive inllnunotherapies mighl be developed, these eould

ultimatcly he applied as adjuvant therapy in cases of primary uvcal melanoma too.

HOlVever, before such immunotherapies Illighl be introduced in clinical practice,

eylotoxic T-Iymphocyte (CTL) reeognilion and anti-tumor response have to be

delermined.

Chapter 8 describes lhe transplantation of uveal melanoma cell !ines inlo lhe

animalmodel of the inul1unodeficient nude mouse. FollolVing reccnl data on olher

modeis, human uvealmelanoma cells were transplantcd al the orthOlOpic site, thaI is

intracamerally, after whieh lhe model indeeel developed c1islant liver metastases. This

experimcnt further showed lhat the c1own-regulation of major histocompatibility

complex (MHC) cia" I molecules in uveal melanoma cells raises the susceptibility of

these cells 10 natura I killer (NK) cells. Thus, IVhen CTL antitumor response is not

induceel by laek of an appropriate MHC class T molecule, lIvealmelanoma cells still

inc1uce an immune reaction in the sen se that they are recognized and killec1 by NK's.

/77

Page 178: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

______ s lJ l\1i\[AR Y __ _

Chapter 9 reporls on the eslablishment of a novel animalmoelel, the chicken

embryo. The uvealmelanoma cells wcre transplanteel into the model's eye al e1ay 3.5,

hence amply before day 9 when it's immune system e1evelops. The lransplanls were

nol repulseel, the embryo's eye development was not disturbed anel the chickens even

halchcd, suffering a hu man uvealmelanoma. In spite of the many possibilities il

seemingly provides, lhis model proved 10 be less reproducible in larger series anel

mighl therefore not be practical lor elaborate inlervention studies.

Finally, the Discussion in Chapter 10 emphasizes lh at the management of a

primary uveal melanoma should be direcled by the fact that this disease is eventually

lcthal in 50% of the cases. In this regard, a large Sluely on lhe effects on survival of

radiotherapy versus enucleation is currenlly undertaken (COMS). Furthennore, the

evahultion of prognostic factors slich as vasclllar networks, dctection of circulating

e1isseminateelmelanoma cells and neural cell adhesion molecules is of major

importance, as weIl as the development of thcrapies for uveal melanoma-relatcd

metastalic e1isease. With respect to lhe latter, especially active immunotherapy using

allogenic cclllines, reporteel in Chapters 6 anel 7 above, appears to be a mosl

promising starting-point, while the nude Illouse appears to be a useful model for the

invcstigation of actlwl immune reponses to sneb therapy.

178

Page 179: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____________ SAMENyATTING ___ _

SAMENVATTING IN HET NEDERLANDS

Het oogmclanoom is de meest voorkomende primaire maligne tumor van het oog met

een jaarlijkse incidentie van circa 80 à 100 in Nederland. In paragraaf 1.1.1, IlIleidillg

tot het ollderll'elp wordt besproken dat met betrekking tot deze vorm van kanker het

stellen van de diagnose en de lokale behandeling van de primaire tumor de afgelopcn

30 tot 40 jaar de belangrijkste punten van discussie zijn geweest. Dit heeft

geresulteerd in een tegenwoordig zeer laag percentage «I %) van fout-positieven,

ofwel gevallen die ten onrechte worden gediagnostiseerd als oogmelanoom, te danken

aan de ontwikkeling van technieken als ultrasonografie, fluorescentie-angiografie en

met name indirecte ophthalmoscopie. Van oudsher werd het oogmelanoom behandeld

door middel van een enucleatie, dat wil zeggen het verwijderen van het gehele oog.

Met het doel de oogbol en gezichtscherpte te behouden is er gestart met

behandelingsmethoden zoals fotocoagulatie, lokale verwijdering van de tumor en

vooral radiotherapie, die in tegenstelling tot verwijdering van het gehele oog als

conservatieve behandeling verder geduid worden. Desondanks is de levensprognose

van patiënten met een oogmelanoom nauwelijks verbeterd. Meer dan 50% van alle

patiënten krijgt op termijn metastasen, welke meestal gelokaliseerd zijn in de lever.

Deze zeer kwaadaardige metastasen zijn tot op heden praktisch onbehandelbaar, en op

het moment dat zij klinisch zijn vastgesteld is de levensprognose van de patiënt zeer

kort (2 tot 9 maanden, atllankelijk van de lokatie van de metastasen). Er is

geëxperimenteerd met diverse therapieën, waarbij slechts enkele

behandelingsmcthoden waarin gebruik wordt gemaakt van chemotherapie - al dan niet

gecombineerd met een operatie of immuuntherapie - een kleine vermeerdering van de

gemiddelde overlevingstijd (3 maanden) lieten zien. Overlevingspercentages van

patiënten hij wie een primair oogmelanoom is vastgesteld, berekend aan de hand van

Kaplan-Meier curves en Cox analyses, bedragen 5, 10 en 15 jaar na de eerste

behandeling respectievelijk 75 %, 65 % en 55 %. De hoogste mortaliteit wordt

gevonden in het tweede en derde jaar volgend op behandeling van de primaire tumor.

Deze situatie vormt het uitgangspunt van dit proefschrift, zoals wordt

aangegeven in paragraaf 1.2, Doel ell bereik !'(JII de these. De gepresenteerde studies

179

Page 180: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,
Page 181: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ -"s A i\lENVATT J NG"-' ___ _

paragraaf 13.2 voorafgaand onderzoek besproken. Dit onderzoek heefluitgewezen dat

de grootte van de primaire tumor (LTD = grootste tumordiameter) verreweg de

belangrijkste parameter is met betrekking tot het risico op het ontwikkelen van

metastasen. Onder de klinische prognostische factoren vallen onder andere ook lokatie

en extrasclerale doorgroei van de tumor en gevorderde leeftijd en mannelijk geslacht

van de patiënt, die alle een verhoogd risico opleveren. Histopathologisch lijkt, naast

tumorgrootte, ook celtype van belang. Vanwege het toenemend aandeel van

conservatieve behandelingen is histologisch onderzoek van lUI110nnateriaal vaak

echter niet mogelijk. Op het moment wordt er gewerkt aan een rt-PCR-techniek die

middels het aantonen van tyrosinase circulerende lllclanoomcellcll detecteert.

Tenslotte is recent een nieuwe, zeer sensitieve en onatllankelijke prognostische factor

gevonden: het vóórkomen van vasculaire netwerken in oogmelanomen, waarvoor

echter op dit moment nog wel histologisch weefsel nodig is.

Tn de in Hoofdstuk 3 en 4 gcrapporteerde studies werd van een groep patiënten

die zijn overleden als gevolg van metastasen op afstand de primaire tUlllor vergeleken

met de corresponderende metastase(n). Allereerst werd in Hoofdstuk 3 een significant

morfologisch verschil geconstateerd: metastasen bevatten een veel hoger percentage

agressieve en ongedifferentieerde tumorcellen vergeleken met het primaire

oogmelanoom waar zij van afstammen. Een deel van de metastasen bestond zelfs

volledig uit ongedifferentieerde cellen die geen kenmerken meer vertoonden van een

oogmelanoom. Kortom, de morfologie van tumorcellen verandert in de loop van het

metastaseringsproces. Uit dit onderzoek bleek verder dat een immunohistochemische

kleuring met HMB-45 het meest sensitief is voor het gemetastaseerde oogmelanoom.

De studie besproken in Hoofdstuk 4 toonde vervolgens aan dat in zowel primaire als

gemetastaseerde oogmelanomen de neurale cel-adhesiemoleeulen NCAM en HNK-l

kunnen voorkomen. Adhesiemoleculen zijn essentieel voor de nesteling van

tumorcellen in organen op afstand vanwege hun specifieke herkenning van receptoren

in het vaatendotheel ter plaatse, zoals onder andere is gebleken uit onderzoek naar

huidmelano0111. Nu bleek een verhoogde expressie van NCAM significant

gecorreleerd te zijn met primaire oogmelanomen die vroeg metastaseren en met

metastasen zelf, terwijl de expressie van de HNK-l epitoop in de metastasen van de

lever vaak sterk verminderd was ten opzichte van de primaire tumor. Dit suggereert

18/

Page 182: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

SAMENVATTING ____________ _

dat een subvorm het adhesiemolecuul NCAM, die niet het HNK-l tot expressie

brengt, mogelijk verband houdt met de orgaanspecificileit, de voorkeur voor de lever,

van het oogmelanoom. Het oogmelanoom metastaseert immers specifiek naar de leVel\

in tegenstelling tot bijvoorbeeld het huidmelanoom, dat voornamelijk naar de long

metastaseert.

In andere typen tumoren is een gen gevonden waarvan het eiwitprodukt

mogelijk een metastase-onderdrukkend effect heeft, het NM23 (niet-metastaserende

kloon 23) gen. Verminderde expressie van dit gen blijkt in met name

mammacarcinomen sterk geassocieerd te zijn met een verhoogde kans op metaslasen

en verminderde overleving. Hoofdstuk 5 beschrijft in analogie met deze gegevens een

studie naar de expressie van het NM23-gen in primaire oogmelanomen en cellijnen. In

dit onderzoek werd de genexpressie bepaald op grond van het immunohistochemisch

aantonen van het gen produkt, het NM23-eiwit, waarbij een "borderline" associatie

gevonden werd tussen het percentage NM23-negalicve velden en overleving. Zoals in

Hoofdstuk 10 ook wordt bediscussieerd, zou de associatie tussen NM23 en

metastasering/overleving in werkelijkheid echter nauwer kunnen zijn gezien het Ceit

dat er aanwijzingen bestaan dat het NM23-gen op een ander, namelijk transcriptioneel

of post-transeriptioneel niveau werkzaam is. Het vaststellen van de werkelijke relatie

tussen NM23 en overleving vereist daarom nader onderzoek op deze niveaus.

Met het oog op het ontwikkelen van een behandeling voor gemetastaseerd

oogmelanooll1, het derde aandachtsveld van deze these, is het voor alles van belang

dat er mogelijkheden worden geschapen voor experimenteel onderzoek. Zoals in 1.3.3

uitvoerig wordt uiteengezet, kan het metastaseringsproces immers pas werkelijk

uitgebreid bestudeerd worden onder laboratorium-omstandigheden. Tot op heden werd

dit echter belemmerd door het Ceit dat er slechts enkele humane oogmelanoom

cellijnen beschreven waren en er daarnaast geen geschikt dierproefmodel beschikbaar

was aan de hand waarvan (ingrijpen in) het metastaseringsproces bestudeerd kon

worden. De auteur en medeonderzoekers zijn er recentelijk wel in geslaagd twee

primaire en drie metastatische oogmelanoomcellijnen te kweken. Hoofdstuk 6

beschrijft deze kweek en de karakterisering van de verkregen cellijnen. De

karakterisering werd uitgevoerd door middel van morfologisch onderzoek met licht­

en electronenmicroscopie, immunohistochemisch onderzoek en

182

Page 183: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

_______________ ---"S"'A'-'~l EN VA TT I N C,,-' __ _

chromosoomonderzoek door middel van karyotypering. De ontwikkeling van deze

cellijnen biedt goede mogelijkheden voor stndies naar de biologische eigenschappen

van het oogmelanoom; ook kunnen zij gebruikt worden voor studies aan de hand van

proe tûiermodellen.

In de in Hoofdstuk 7 besproken studie, ingeleid in 1.3.3.2, werd de expressie

van enkele melanoom-specifieke en geassocieerde antigenen onderzocht in

voornoemde oogmelanoom-cellijnen. Zoals wederom is gerapporteerd met betrekking

tot huidmelanoom, spelen deze antigenen, MACE1, -2, en -3 alsmede CP 100 en

Iyrosinase, een essentiële rol bij de herkenning van melanomen door cytotoxisehe T

lymphocyten, waardoor deze laatste ook in staat zijn de tumorcellen te doden. De

genen die coderen voor deze eiwitten komen in het geval van de MACE-genen alleen

bij melanomen tot expressie en worden daarommclanoom-specifiek genoemd. In het

geval van CP 100 en Iyrosinllse komen zij in alle melanocytaire cellen tot expressie,

dus ook in melanocyten, en worden daarom melanoom-geassocieerd genoemd. In de

cel wordt na transciptie van deze genen het eiwitprodukt gereduceerd tot een peptide

van 8 tot 10 aminozuren, dat in het endoplasmatisch reticulum een complex aangaat

met een MHC klasse I molecuul. Dit complex wordt getransporteerd naar de

cclmembraan, waar het herkend kan worden door eytotoxisehe T lymphocyten. De

huidige studie toont aan dat met name twee cellijnen, OCM-I en OMM-l, alle

antigenen in kwestie tol expressie brengen. Deze twee cellijnen zouden daarom bij

uitstek gebnlikt kunnen worden voor actieve immunisatie van oogmelanoompatiënten

luct metastasen op afstand, zoals thans al gebeurt in gevallen van gcmelastm;eerd

huid melanoom. Overigens is VOO I' oogmelanoompatiënten mel metastasen actieve

inullunisatie met autologe tumorcellen te prefereren boven allogenc immunisatic.

Gezien de korte levensprognose van de patiënten enerzijds en de voor het kweken,

transfecteren en bestralen van de tumorcellen benodigde tijd anderzijds is het

praktisch vrijwel onlllogelijk om autoloog materiaal te gebruiken. Zoals in Hoofdstuk

10, Discussie wordt besproken moet nog vastgesteld worden of eytotoxische T

lymphocyten de tUlllorcellen inderdaad herkennen, alvorens met een klinische proef

naar het effect van een dergelijke il1l111uuntherapie gestart kan worden. Studies naar de

cytotoxisehe T lymphocyten herkenning van de oogmelanoom cellen wordt op dit

moment nog verder onderzoek verricht. Als actieve immunisati~ van

183

Page 184: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

SAi\IENVATTING ----------

oogmelanoompatiënten met metastasen op afstand inderdaad mogelijk blijkt te zijn en

geen ernstige bijwerkingen sorteert, zou deze op den duur ook als ondersteunende

therapie naast enucleatie en radiotherapie kunnen worden toegepast in de behandeling

van primair oogmelanoom. Voor gevallen van gemetastaseerd huidmelanoom, althans

in diermodellen, is beschreven dat een dergelijke therapie een goede anti­

tU1110rilllllluniteit oplevert.

Hoofdstuk 8 beschrijft een studie met een proefdiermodel, het naakte

muismodcl, waarbij oogmelanoomcellijnen in het proefdier werden getransplanteerd.

In de naakte muis is de cellulaire lymphoeytaire afweer afwezig waardoor allogene en

xenogene transplantaten niet worden afgestoten. Uit andere studies is gebleken dat

metastasering alleen optreedt indien de tumorcellen worden getransplanteerd naar het

orgaan van oorsprong; in dit gevalllloesten cIe humane oogmelanoOlllcellijnen dus

getransplanteerd worden naar de voorste oogkamer van de naakte muis. In deze proef

werden behalve een geslaagde transplantatie van primaire tlllllorce11en inderdaad

reproduceerbaar metastasen verkregen. De studie toonde bovendien aan dat de

gevoeligheid van de getransplanteerde oogmelanoomeellijnen voor "natural killer"

ce11en groter is in geval van een verminderde expressie van major

histocompatibiliteitscomplex (MHC) klasse 1 moleculen; de proefdieren die

getransplanteerd waren met een eellijn die deze klasse MHC verminderd tot expressie

brengt ontwikkelden geen metastasen, terwijl deze wel ontstonden bij onderdrukking

van de "natural killer" cellen door middel van cyclophosphamide of anti-asialo GM I.

Hierbij is natuurlijk van belang dat bij verminderde MHC-expressie ook cytotoxische

T Iymphocyten de cellen niet herkennen en dus niet doden. In de Discussie in

Hoofdstuk 10 wordt over dit ondenverp nog opgemerkt dat het vooralsnog onzeker is

of NK-activiteit ook in de praktijk ingezet kan worden om metastasen in humane

patiënten te bestrijden

In Hoofdstuk 9 wordt een ander proefdiermodel beschreven: een voor de

studie van oogmelanoom nieuw model van kippenemblYo's. De belangrijkste gedachte

achter dit model is dat de transplantatie van de cellen plaatsvindt voordat het

im111lHlllsysteem van het kippcnemblYo zich heeft ontwikkeld, zodat zij als

lichaamseigen worden ervaren en zodoende niet worden afgestoten. Het bleek

mogelijk om oogmelanoomcellen te transplanteren in het oog van een 3.5 dag oud

184

Page 185: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ ~SAi\IENVATTING

kippenembryo, dat wil zeggen ruimschoots voor dag 9 wanneer het immuunsysteem

tot ontwikkeling komt. De tumorcellen konden worden opgespoord door

immunohistochemisehe kleuring met HMB-45, NKI/c3 en SI 00 en bleken aldus

inderdaad te groeien in het oog van het kippenembryo. Het bleek bovendien mogelijk

het ei te laten bebroeden en uit te laten komen, zodat er kuikens met een humaan

oogmelanoom geboren werden. Dit model, dat in principe uitstekende mogelijkheden

biedt om groei en afweer van oogmelanomen in vivo te bestuderen, bleek echter niet

in die mate reproduceerbaar dat er uitgebreide interventiestudies mee zouden kunnen

worden verricht.

Tenslotte wordt, aan het eind van Hoofdstuk 10, Discussie ell

toekomstperspectieven, het belang van het ontwikkelen van nieuwe

behandelingsvonncll voor oog melanoom nogmaals onderstreept. Bij constatering van

een primair oog melanoom moet het feit dat dit in 50 % van de gevallen tot de dood

leidt, nu nog theoretiseh, beslist moeten meewegen in de beslissing omtrent de wijze

van lokale en/of toekomstige systeem behandeling. Om beter te kunnen insehatten

welke behandeling voor welke patiënt het meest in aanmerking komt is met name

onderzoek naar de waarde en het vaststellen van de versehillende prognostische

factoren van belang, zoals het zichtbaar maken vall de aanwezigheid van vasculaire

netwerken in de tumor, het vroegtijdig aantonen van eirculerende tumorcellen, en het

bepalen van de in deze these besproken spiegels van neurale eel-adhesiemoleculen.

Voorts lijken de kweek en de eigenschappen van enkele cellijnen, besproken in

Hoofdstuk 6 en 7, veelbelovend met betrekking tot onder andere het ontwikkelen van

actieve immuulltherapieën, die ingezet kunnen worden om metastasen van het

oog melanoom te behandelen en wellicht te voorkomen. De gegevens uit en over de

verschillende dierproefmodellen kunnen mogelijk van dienst zijn bij de ontwikkeling

van een meer succesvolle therapeutische benadering van oognlclanoom.

185

Page 186: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

ATP

CDK

cDNA

COMS

CTL

___ ---"A"'B BJ{ E V I AT ION S

ABBREVIATIONS

: adenosine triphosphate

: cyclin-dependent kinase

: copy deoxyribonucleic acid

: collaburative ocular melanoma study

: cytotoxic T-l yphocytes

FAMMM : familiar atypical mole and malignant melanoma

FNAB : fine-needle aspiralion biopsy

HLA : humanleucocyle agglutinin

HNK-l : human natural killer-l

IL·2 : inlerleukin-2

LAK eells : lymphokine activating killer cells

LTD : largest tumor diameter

Mabs

MHC

MLN

mRNA

NCAM

NDPK

NK cells

NM23

PB

PCR

PBL

PuF

SCID

SSCP

TIL

186

: monoc1onal antibodies

: major histocompatibility complex

: mean diamcter of the len large st nucleoli

: messengcr ribOllUcleic acid

: neuml eeU adhesion molecule

: nucleoside diphosphale kinase

: natural killer cells

: nonmetastatic clone #23

: Phosphate-buffered saline

: polymerase chain reaction

: peripheral blood lymphocytes

: purine binding factor

: severe combined immunodeficient

: single-strand-conformation polymorphisll1

: tumor-infiltrating lymphocytes

Page 187: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ --"C"'U-"R"'R"I,,( U L U j\] V I TA E

CURRICULUM VITAE

Op 9 september 1960 werd de schrijver van dit proefschrift geboren te Etten-Leur en

behaalde in 1979 zijn eindexamen VWO aan de Katholieke Scholengemeenschap te

Etten-Leur. Van 1979 tot 1986 studeerde hij geneeskunde aan de Erasmus Universiteit

te Rotterdam. Gedurende zijn studie was hij werkzaam als student-assistent op de

afdeling celbiologie waar hij werkte aan een single cell detectie techniek voor de

ziekte van Fabry. In 1985 was hij werkzaam op de afdeling Ophthalmogenetica van

het Interuniversitair Oogheelkundig Instituut te Amsterdam waar hij in samenwerking

met de afdeling Genetica, Erasmus Universiteit Rotterdam een studie verrichtte naar

de verschillende uitingsvormen van de ziekte van von Hippel-Lindau. In dezelfde tijd

startte hij het vervolgonderzoek naar het effect van voorbestraling bij oogmelanoom

patiënten in het Oogziekenhuis Rotterdam. Hierna heen hij zijn militaire dienst

vervuld van 1986 tot 1988 op de afdeling Dermatologie van het Militair Hospitaal A.

Mathijsen. Van 1988 tot 1992 was hij in dienst van het Oogziekenhuis Rotterdam als

arts-assistent in opleiding tot oogarts. In deze tijd werd het vervolg onderzoek naar de

oogmelanoom patiënten voortgezet en Inaakte hij een start met het kweken van

oogmelanool11 cellijnen. De laatste 2 maanden van zijn opleiding voltooide hij op de

afdeling Oogheelkunde van het Academische Ziekenhuis Rotterdam, waar hij daarna

als oogarts werd aangesteld en tot op heden werkzaam is.

187

Page 188: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

___ -'L'-'IST OF PURL1CATIONS

LIST OF PUBLICATIONS

Luyten OPM, Hoogeveen AT, Oa1jam'd H. A F1uorescence staining method for the

demollstration anel measurcment or lysosOlllal enzyme activitics in single ceUs.

J. Histochem Cytochem. 1985;33:965-968.

Luyten OPM, Mooy CM, dc Jong PTVM, Hoogeveen AT, Luider ThM. A chicken

embryo model to study the growth of human uvealme1anomtl. Biochcm

Biophys Res Comm. 1993;192:22-29.

Luyten OPM, Mooy CM, Eijkenboom WMH, Stijnen Th, Hellemons LP, Luider ThM,

de Jong PTVM. No demonstrated effect of pre-enuc1eation irradiation on

survival of patients with uvea1melanoma. Am J Ophthalmol. 1995; 119:786-791.

Ma 0, Luyten OPM, Luider ThM, Niederkorn JY. Relationship between natura1 killer

cell susceptibility and metastasis of humun uveal melanoma eells in a 111urine

model. Invest Ophtha1 Vis Sci. 1995;36:435-441.

De Vries CJ, Mooy CM, van Balken MR, Luyten OPM, Quax PHA, Verspaget HW,

Weidie UH, Ruitcr DJ, van Muijen ONP. Components of the plasminogen

activation system in uveal me1anoma - a elinico-patho1ogica1 study.

J Patho1ogy 1995;175:59-67.

Mooy CM, Vissers CJ, Mulder AH, Stijnen 1', Luyten OPM, de Jong PTVM. DNA

t10w cytometry in uvea1me1anoma. The effect of pre-enucieation

radiotherapy. Br J Ophthalmol. 1995;79: 174-177.

MooyCM, Luyten OPM, de Jong PTVM, Jensen OA, Luider ThM, van Ham F,

Bosman FT. Neural cel! ad hes ion molecule distribution in primary and

metastatic uvea1mc1anoma. HunHIIl Pathol. 1995:26: 1185-1190.

Mooy CM, Luyten OPM, de Jong PTVM, Luider ThM, Stijnen T, van de Ham F, van

Vroonhoven CCJ, Bosman FT. Inullunohistochemica1 and prognostic analysis

of apoptosis and proliferation in uveal melanoma. Am J Pathol.

1995; 147: 1097 -1104.

188

Page 189: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

____ -"UST OF PVBLICATIONè'S __ _

Naus NC, Luyten GPM, Stijnen T, elc Jong PTVM. Phacoemulsi fication in

comparison with eXlracapsular cataract surgery: post-operative astigmatisl1l

anel visual recovery. Doe Ophthalmologiea 1995;90:53-59.

Creyghton WM, ele Waard-Siebinga T, Danen EHJ, Luyten GPM, Van Muijen

GNP, Jager MJ. Cytokine-mediateelmodulation of expression of integrin

ICAM-I anel CD44 on human uvealmelanoma eells in vitra.

Melanoma Res.1995;5:235-242.

Luyten GPM, Naus NC, Mooy CM, Hagemeijer A, Kan-Mitchell J, van Drunen E,

Vuzevski V, de Jong PTVM, Luider ThM. Establishment and charaeterization

of primary anelmetastatic uvcalmelanoma celilines. Int J Cancer 1996;66: 1-8.

189

Page 190: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

DANKWOORD

DANKWOORD

Het verrichten van wetenschappelijk onderzoek en het schrijven van een proefschrift is

een gezamenlijke inspanning van veel mensen, Alleen een goede samenwerking tussen

mensen en tussen diverse research groepen kan leiden tot de vollooiing daarvan, Ik

ben blij dat we betere samenwcrkingsverbanden tot stand hcbben gebracht met andere

mensen en groepen, Uiteindelijk zal kan dit onze inzichten over het oogmelanoom

vergroten.

Allereerst wil ik Professor DJ'. P,T, V.M, de Jong dankcn voor het gestelde

vertrouwen in mij, Beste Paulus, je was mijn opleider tot oogarts en je hebt mij

geïntroduceerd in de klinische aspecten van de oogheelkundige oncologie, In het

wetenschappelijk werk was je kritisch en altijd open voor discussie, Ik ben blij dat ik

dit proefschrift heb kunnen afronden en ik hoop dat we in de toekomst nog veel

research projecten samen kunnen opzetten nu je verhuisd bent naar je nieuwe

werkplek, het I.O,I, te Amsterdam,

Vervolgens ZOu ik DJ'. C,M, Mooy en DJ'. Th,M, Luider willen danken voor

hun wetenschappelijke inzichten, inspanningen en suggesties, Neellje, dankzij jou

pathologische kennis zijn veel manuscripten tot stand gekomen en nauwlettend

geredigeerd, Theo, jouw ongebreidelde enthousiasme voor onderzoek bracht vele

nieuwe aspecten aan het licht en leidde tot nieuwe mogelijkheden,

Vele studenten hebben hun onderzoek stages doorgebracht op de research

afdeling en de kliniek en een belangrijk aandeel gehad in realisatie van dit proefschrift

(Jannie Post, Rens Hellemons, Nicole Naus, Mei Lie Tang, Sara Naghib, Maleika

Fuchs, Riena Aleredjo), Naast dat ik jullie wil bedanken voor jullie bijdragen, heb ik

het zelf ook allijd plezierig gevonden onderwijs te geven en met jullie te werken,

Waarschijnlijk zullen er vele studenten volgen, Speciaal wil ik toch Nieole Naus

danken voor haar enthousiasme en werklust waardoor we veel research projecten

hebben kunnen at:sluiten,

Voor de grote analytische inzet en de prettige samenwerking zou ik graag

Corrie van der Spek van de afdeling medische oncologie, AZL (DI', p, Schrier) willen

danken, Daarnaast wil ik Sonja KerkvJiet voor al het histopathologische werk en Ineke

Brand voor het moleculair biologische werk danken,

190

Page 191: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

DAN K -,v.~a,-,O!.cR"-,,D __ _

Next I would Iike to thank Dr. Itte de Waard-Siebinga, Dl'. June Kan-Mitehell

and Professor Bruce Ksander that they providedme their uvealmelanoma eelllines.

Then I wanted ta thank Professor Jerry Niedcrkorn and Dr. Ding Ma for our

collaboration on the transplantation of the uvealmelanoma eell lines in nu de mice. I

hope we will continu the collaboration with your group together with ot her dutch

research groups on uveal melanomas. Op dit punt zou ik graag Dr. Martine Jager

danken voor onze vruchtbare discussies en onze gezamenlijke inspanning om tot een

hetere samenwerking te komen binnen Nederland.

Ook wil ik dankenll1evrouw T. Muijlwijk-Planting en lllevrouw C. Kruisheer.

Tineke ik wil je bijzonder danken voor al je secretarieel werk en je zorgvuldige

controles op de manuscripten. Clasien, dank voor je redigeerwerk en het minutieus

doorwerken van alle artikelen en het proefschrift. Daardoor is het een overzichtelijk en

leesbaar geheel geworden.

Verder wil de hele afdeling oogheelkunde van het Academisch Ziekenhuis

Rotterdam, Dijkzigt danken voor de ruimte die jullie mij hebben gegeven om dit

proefschrift af te ronden.

Tot slot mijn allerliefste Yolanda. Jouw ruimte, betrokkenheid en je

nimmeraflatende stilllulering zijn erg belangrijk voor mij.

/9/

Page 192: PRIMARY AND METASTATIC UVEAL MELANOMA · 1.1 Clinical aspects of uveal melanoma 1.1.1 Inlroduclioll 10 Ihe subject Uveal melanoma is a unCOll111lon discase with on estimation 80-100,

DANKWOORD

De studies werden financieel mogelijk gemaakt door grants van the "Ooglijders"

Foundation Rotterdam, the Erasmus University Trust Foundation, Rotterdam and the

Dutch Society for the Prevention of Blindness, Ulrecht, The Foundation of Ooglijders

and Blindenbelangen, lhe Hague The Nelherlands.

Dit proefschrift kwam lol stand door financiële ondersteuning van de firma's;

Allergan B.V., Bournonville Phanna, Lameris Oolech B.V., Medical Workshop,

Chibret, Oculenti B.V., Ophthec, Phannacia & Upjohn, Rockmed B.V.,

Tramedico B.V., e.a.

192