Download - Campylobacter colitis: histological immuno- histochemical and · Campylobacter jejuni were studied in order to obtain histological and immunohistochemical criteria to differentiate

Gut, 1985, 26, 945-951

Campylobacter colitis: histological immuno-histochemical and ultrastructural findingsJ P VAN SPREEUWEL, G C DUURSMA, C J L M MEIJER, R BAX,P C M ROSEKRANS, AND J LINDEMAN

From the Department ofInternal Medicine, Division of Gastroenterology, St. Antonius Ziekenhuis,Nieuwegein, Department ofMicrobiology and Pathology, Stichting Samenwerking Delftse Ziekenhuizen,Delft, Department ofInternal Medicine, St. Elizabeth Ziekenhuis, Leiderdorp, and the Department ofPathology, Vrije Universiteit, Amsterdam, The Netherlands

SUMMARY The colonic biopsy specimens of 22 patients with colitis and positive stool cultures forCampylobacter jejuni were studied in order to obtain histological and immunohistochemicalcriteria to differentiate Campylobacter colitis from chronic inflammatory bowel disease. Inaddition we tried to identify Campylobacter inclusions by means of immunohistochemistry andelectron microscopy as evidence for invasion of the colonic mucosa. The results show that themajority of patients with Campylobacter colitis have the histological picture of acute infectiouscolitis with increased numbers of IgA and IgM containing plasma cells in the colonic mucosa incontrast with patients with active chronic inflammatory bowel disease who show increases of IgAand IgG (ulcerative colitis) or IgA-, IgM and IgG containing plasma cells (M Crohn) in theircolonic biopsies. The results of immunohistochemical stainings with Campylobacter antiserumshow invasion of Campylobacter in the colonic mucosa. These findings were confirmedultrastructurally.

In the past decade Campylobacter has emergedfrom obscurity to become an important humanpathogen. Provided that proper techniques areapplied Campylobacter fetus subspecies jejuni canbe isolated from the stools of patients with diarrhoeaas often as Salmonella.1 2 In the laboratory formicrobiology of the SSDZ, Delft, Campylobacterjejuni was isolated as the causative organism in 41%of 732 consecutive positive stool cultures of patientswith diarrhoea. Campylobacter may cause a spec-trum of intestinal disease ranging from acute gas-troenteritis to toxic megacolon, lymphadenitismesenterialis, and even appendicitis and cholecysti-tis.35 Usually infection with this organism results inacute gastroenteritis with fever and frequent loss ofoften bloody stools. The disease resolves spon-taneously within one week but in 20% it runs a moreprolonged course or relapses resembling chronicinflammatory bowel disease.1 Endoscopically it may

Address for correspondence: Dr J P van Spreeuwel, Department of InternalMedicine, Division of Gastroenterology, Ziekenhuis St. Annadal, St. Annadal1, 6214 PA Maastricht, The Netherlands.Received for publication 1 November 1984

be indistinguishable from ulcerative colitis. Histo-logical examination of rectal biopsies has rangedfrom normal to inflammatory changes suggestive ofacute infectious colitis or ulcerative colitis orCrohn's disease.61' In the present study we haveinvestigated the colonic biopsies of 22 patients withdiarrhoea and stool cultures positive for Campylo-bacter jejuni. The aim of our study was to determinethe histological features and the number of immuno-globulin containing cells in colonic biopsies ofpatients with Campylobacter colitis and to identifyCampylobacter inclusions immunohistochemically.

Methods

PATIENTSOne hundred and twelve colonic biopsies from 22patients with positive stool cultures for Campylobac-ter jejuni were examined. One patient had beensuffering from chronic ulcerative colitis for manyyears whereas one patient had previously sufferedfrom non-specific proctitis. The others had nocoexisting gastrointestinal disease. Diarrhoea withcrampy abdominal pain was the presenting com-

945

on January 31, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.26.9.945 on 1 Septem

ber 1985. Dow

nloaded from

http://gut.bmj.com/

van Spreeuwel, Duursma, Meijer, Bax, Rosekrans, and Lindeman

plaint in 15 patients, 13 of them had noticed blood intheir stools, and fever was an accompanying symp-tom in seven of these 15 patients. Nausea wasrecorded frequently. Six patients merely complainedof loosing blood with their stools. One patientpresented with abdominal pain, nausea, and vomi-tus. The numbers of immunoglobulin containingcells in colonic biopsies of patients with Campylo-bacter colitis were compared with those of 10healthy controls, 10 patients with active Crohn'sdisease of the colon and 10 patients with activeulcerative colitis. All patients with chronic inflam-matory bowel disease had histologically abnormalcolonic biopsies.

Stools were cultured for Salmonella, Shigella,Yersinia and Campylobacter routinely. Campylo-bacter was cultured on a selective medium: bloodagar plates containing Campylobacter Growth Sup-plement and Campylobacter Selective Supplement(Oxoid LTD Basingstoke England nos. SR 84 andSR 85). Inoculated plates were incubated during 48hours at 42°C in an atmosphere containing 5-15%02 and 5-8% CO2 by using the GasPak of BBLwithout a catlyst in a jar.

Biopsies were fixed immediately either in aformalin sublimate mixture for four to six hours forroutine histology and immunohistochemical studiesor in cacodylate buffered glutaraldehyde for ultrastructural studies. Tissue samples were embedded inparaffin wax, cut in sections 4,um thick and mountedon glass slides. Besides histological stains includinghaematoxylin and eosin and periodic acid Schiff,biopsies were stained specifically for IgA, IgM, andIgG heavy chains using an indirect immunoperox-idase technique.12 Appropriate controls were doneaccording to Sternberger.13 Rabbit antisera againstIgA, IgM, and IgG heavy chains were purchasedfrom Dakopatts (Denmark). Their specificity wasconfirmed by immunoelectrophoresis, immuno-fluorescence and immunoperoxidase stainings onbone marrow preparations monoclonal for IgA,IgM, or IgG. Horseradish peroxidase labelled goatantirabbit Ig was obtained from Miles (Yedah,Israel). The IgA, IgM, and IgG stained sectionswere used for morphometrical analysis. They werephotographed with a standard magnification (100, 8x) and projected on a graphic tablet interfaced to acomputer (Ibas I Kontron Munchen). Every photo-graph contained approximately 1 mm mucosa at fullthickness. The lamina propria area was limited bytwo lines perpendicular to the muscularis mucosaeand this area was measured per mm muscularismucosae length. The number of immunoglobulincontaining cells was measured in two consecutivesections in approximately the same area and ex-pressed both per 0 1 mm2 lamina propria and per

0-5 mm stretched muscularis mucosae. The latter iscomparable with the 'mucosal tissue unit' describedby Brandzaeg and Baklien.14 Moreover, immuno-histochemistry was carried out with Campylobacterantiserum using both direct immunoperoxidase andimmunofluorescence techniques. The antiserum wasraised in a rabbit that was vaccinated intravenouslywith a Campylobacter fetus subspecies jejunisuspension containing microorganisms of serotypeLAU1'5 that were killed in 1-5% formalin. Anti-serum titres were determined by agglutination andindirect immunofluorescence methods. Afterseveral absorption steps with Salmonella g, Shigellaboidii, Shigella sonnei, Proteus vulgaris and Escher-ichia coli the antiserum showed no cross reactivitywith Salmonella, Shigella, or Escherichia coli. Theantiserum gave a positive immunofluorescence with37 out of 38 different Campylobacter LAU sero-types; representing all known serotypes at that timein the Netherlands. Immunofluorescence wasachieved by removing paraffin with xylol andalcohol, preincubation with normal swine serum1:10, incubation with rabbit Campylobacter anti-serum 1:40, once again incubation with normalswine serum 1:10 and incubation with swine anti-rabbit FITC. Sections stained with preimmuneserum from the same rabbit served as controls.Colonic biopsies from eight patients with Crohn'sdisease of the colon and nine patients with ulcerativecolitis stained with Campylobacter antiserum servedas controls as well and were all negative. In additionin three patients in whom immunohistochemistrywith Campylobacter antiserum had given positiveresults ultrastructural examination of the biopsieswas done using a Zeiss EM 109 electron microscope.

Statistical analysis was carried out according tothe Student's t test, p<005 was considered signifi-cant.

Results

The results are summarised in the Table.

HISTOLOGICAL FINDINGSIn 19 patients the colonic biopsies showed a histolo-gical picture consistent with acute infectious colitis.In the patient with longstanding ulcerative colitiswith superimposed Campylobacter enterocolitis thehistological features of ulcerative colitis dominatedwith marked distortion of crypt architecture. In twopatients histological examination revealed no abnor-malities. Based on our findings in patients in whomfollow up biopsies were obtained, and the time lapsebetween the onset of bloody diarrhoea and thebiopsy we believe the histological changes of Cam-pylobacter colitis can be divided in three stages.

946


http://gut.bmj.com

/G


ber 1985. Dow

nloaded from

http://gut.bmj.com/

Campylobacter colitis: histological immunohistochemical and ultrastructural findings

Table Histological and immunohistochemical findings in 22 patients with positive stool cultures for Campylobacterjejuni

ImmunohistochemistryCamp. antiserum

Crypt surfDays after IgA IgM lgG epitheliumonset of Histologicalfindings containing cells Glycocalyx lamina prop Histiocytessymptoms

1 Act. chronicinflam. T T T + + +? Act. chronic inflam. T T T + + +

2 9 Acute colitis II T - - +30 Normal + -

3 ? Acute colitis II T + -

4 ? Acute colitis II T + -

5 17 Acute colitis III + -

6 7 AcutecolitisI - +14 Acute colitis III T T +30 Normal

7 ? Acute colitis II T +8 6 AcutecolitisI T T + +

30 Normal T + + +9 ? Acute colitis II T + +10 6 Acute colitis I T +11 3 Acute colitis I T T +12 10 Acute colitis II T T +13 ? Acute colitis III T - + +14 >7<14 Acute colitis II T T + + +15 ? Acute colitis II T T16 ? Acute colitis II T17 14 Acute colitis II T T +18 ? Normal T - + +19 >7<14 Acute colitis II T + + +20 ? Acute colitis II +21 ? Acute colitis I T +22 >14 Normal

I = early stage, II = late stage, III = residual stage, T = increased, ? = unknown.

EARLY STAGES (FIVE PATIENTS)In the first seven days after the onset of bloodydiarrhoea histological changes are evident anddiagnostic for acute infectious colitis. Epithelialchanges are conspicuous. The surface epithelium isinfiltrated by neutrophils and superficial ulcerationsare present. Often the surface is covered withexudate. Transmigration of granulocytes in the cryptepithelium gives rise to cryptitis (incipient cryptabscesses) that may ultimately result in crypt absces-ses: crypts filled with granulocytes lined by a flatdegenerated epithelium. There usually is mucindepletion of the epithelium. The lamina propriashows oedema and is diffusely infiltrated withgranulocytes, histiocytes, plasma cells, and lympho-cytes. Granulocytes and histiocytes predominate.Occasionally the histiocytic component of the infil-trate is very pronounced and gives the infiltrate agranulomatous appearance. Granulomata with giantcells, however, were never observed. Apart fromseparation of crypts because of oedema and loss ofcrypts, crypt architecture is well preserved. Thepathological changes are confined to the mucosa

qand tend to be more pronounced in the distal partsof the colon (Fig. 3).

LATE STAGE (12 PATIENTS)From seven to 14 days after the beginning ofsymptoms the inflammation gradually subsides andthe histological findings become less characteristic.The epithelium shows features of regeneration:large pale nuclei with prominent nucleoli and thereis increased mitotic activity. Transmigration ofneutrophils in crypts is either absent or may be seenfocally adjacent to normal crypts giving the inflam-mation a discontinuous appearance. The laminapropria is oedematous especially in the upper layersand contains an inflammatory infiltrate consistingmainly or solely of mononuclear cells, especially inthe deeper layers (basal cryptitis). Lymphoid hyper-plasia may be found and crypt architecture is notdistorted.

RESIDUAL STAGE (THREE PATIENTS)After 14 days inflammatory changes are minimal orabsent. There may be a slight increase of lympho-

947


http://gut.bmj.com

/G


ber 1985. Dow

nloaded from

http://gut.bmj.com/


Campylobctercolitis n = 22

200

1 x-=

150-

100-

50-

O0

Control Crohn's Ulcerativen_10 disease of colitis

the colon nz10n=10

Patient with ulcerativecolitis and superimposedcampylobacter colitis

;

3

0

x

L

6

A MG A MG A MG A M G

Fig. 1 Numbers ofimmunoglobulin containing cells per0-1 mm2 lamina propria ± standard deviation.

cytes and plasma cells in the lamina propria andcrypts may not reach to the muscularis mucosae. Inone patient we observed branching of crypts.

IMMUNOHISTOCHEMICAL FINDINGSThe numbers of IgA, IgM and IgG containing cellsare listed in Figs. 1 and 2. IgE containing cells wereobserved sporadically and therefore not included.The total number of immunoglobulin containingcells is increased in Campylobacter colitis comparedwith controls. This increase is because of IgA andIgM containing cells, both the number per 05 mmmuscularis mucosae (p=00002 and p=003 respec-tively) and the number per 0-1 mm2 lamina propria(p=0.0005 and p=0001 respectively) differs signifi-cantly. The number of IgG containing cells per 0.1mm2 lamina propria was lower in Campylobactercolitis (p=0.01) compared with controls but this isprobably due to oedema of the lamina propria as thenumber of IgG containing cells per 0-5 mm muscu-laris mucosae does not differ. Biopsies of patientswith Campylobacter colitis differ significantly frombiopsies of patients with Crohn's colitis and ulcera-

Campylobactercolitis n=22

(+ix 276),nnj .

150-

100-

50-

Controls Crohn'sn=10 disease o

the colonn=10

Ulcerativecolitis

n-10

x= Patient with ulcerativecolitis and superimposedcampylobacter colitis

x~~~~~~~~~~~

5~~~~~~~~~~~~~~~~~~~~~~

it

so GAM

Fig. 2 Numbers ofimmunoglobulin containing cells per0 5 mm muscularis mucosae ± standard deviation.

tive colitis with respect to the number of IgGcontaining cells. The number of IgG containing cellsis significantly lower in Campylobacter colitis com-pared with Crohn's colitis and ulcerative colitis,both the number per 0-5 mm muscularis mucosae(p=001 and 0-00007 respectively) and the numberper 0-1 mm2 lamina propria (p=0002 and 0-00003respectively) differs significantly. Immunohistoche-mical staining with Campylobacter antiserum gavepositive results in one or more biopsies in 19 out of22 patients. The results with immunofluorescencewere equal to that obtained by immunoperoxidase.Positive staining was found in the glycocalyx (Fig.4), in the surface epithelium especially goblet cellsand focally in the lamina propria. Occasionallypositive staining was also found in crypt epitheliumand in histiocytes in the lamina propria. Positiveimmunohistochemical results did not correlatequantitatively or qualitatively with the degree ofinflammation and were also obtained in one patientthat showed no histological abnormalities.

In order to confirm the immunohistochemicalfindings with Campylobacter antiserum serial sec-

948

ILw

O-1


http://gut.bmj.com

/G


ber 1985. Dow

nloaded from

http://gut.bmj.com/

Campylobacter colitis: histological immunohistochemil.ni*'8 ¾.tt'i

* N

Fig. 3 Colonic biopsy (seven days after the onset ofsymptoms) showing diffuse infiltration ofthe laminapropria with granulocytes, histiocytes, lymphocytes andplasma cells, erosion andfocal cryptitis.

cal and ultrastructural findings 949

tions of the biopsies of four patients giving positiveresults were examined ultrastructurally. These sec-tions were compared with the ultrastructural pictureof a Campylobacter jejuni culture (Fig. 5). In allthese biopsies characteristic features of Campylo-bacter could be demonstrated in either the glyco-calyx (Fig. 6) overlying the surface epithelium or inthe cytoplasma of the surface epithelium (Fig. 7)including Goblet cells.

Discussion

In this study we have shown that Campylobacterinfection produces the histological picture of acuteinfectious colitis and that the number of immunoglo-bulin containing cells in the colonic biopsies of thesepatients differs from those of patients with chronicinflammatory bowel disease with regard to thenumber of IgG containing cells. In addition we havebeen able to show Campylobacter immunohistoche-mically in colonic biopsies in 19 out of 22 patientswith positive stool cultures for Campylobacterjejuni.

Campylobacter colitis should be distinguishedfrom other bacillary dysenteries and from activechronic inflammatory bowel disease because it maymimic these diseases both clinically and endoscopic-ally. Stool cultures and colonoscopy with biopsy areimportant tools in the differential diagnosis. Histo-

Fig. 4 Indirect immunofluorescence with Campylobacter antiserum showing positive staining in the glycocalyx overlyingthe surface epithelium x 400.


http://gut.bmj.com

/G


ber 1985. Dow

nloaded from

http://gut.bmj.com/


Fig. 5-Electron micrograph ofa Campylobacter culturefixed in formalin and embedded in epon x 10 000. ShowingCampylobacter microorganisms (arrowhead).

logically considerable overlap with chronic inflam-matory bowel disease exists and differentiation maybe difficult especially in the late stages.Determination of the number of IgA, IgM, and

IgG containing cells may be helpful in the differen-tial diagnosis between Campylobacter colitis andactive chronic inflammatory bowel disease becausethe number of IgG containing cells is normal in theformer whereas the number of IgG containing cellsis increased in active chronic inflammatory boweldisease. 16 As these findings are non-specific thehistological diagnosis of Campylobacter colitis ulti-mately relies on demonstration of Campylobactermicroorganisms immunohistochemically. Campylo-bacter colitis cannot be distinguished histologicallyfrom other bacillary dysenteries nor is it possible todiagnose Campylobacter infection histologicallywhen it is superimposed on chronic inflammatorybowel disease.17 Immunohistochemistry with Cam-pylobacter antiserum may prove useful in thesecases. It also makes it possible to diagnose Campylo-

g.trcural picture of the glycocalyx showing atransversal section of Campylobacter microorganisms x10 000 (arrowhead).

bacter colitis retrospectively. Although mostpatients recover from their illness within one weekthe median duration of excretion of microorganismsin untreated cases is two to three weeks,18 this mayexplain by Campylobacter can still be shownimmunohistochemically after the patient has re-covered. Because immunohistochemistry withCampylobacter antiserum yields positive results in19 out of 22 patients with stool cultures positive forCampylobacterjejuni it can be considered a sensitivetechnique.The mechanism by which Campylobacter causes

disease is not known.'9 Although the organism wasoriginally isolated from the blood,20 positive bloodcultures have only been reported occasionally.2'Campylobacter jejuni has been shown to invadechicken embryo cells in vitro21 but evidence fortissue invasion in man is lacking. Alternatively tissuedamage could be produced by toxins. Our findingswith Campylobacter antiserum that were confirmedultrastructurally clearly show that Campylobacter

950


http://gut.bmj.com

/G


ber 1985. Dow

nloaded from

http://gut.bmj.com/

Campylobacter colitis: histological immunohistochemical and ultrastructural findings 951

I

V~~~~~~~~~~~~@. *; ................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........ :t.i.j!,Fig. 7 Electron micrograph ofmoderately preservedGoblet cell (the cell wall is marked by small arrowheads)showing several transversal sections of Campylobactermicroorganisms (large arrowheads) x 6000.

invades the colonic mucosa. Recently it has beenshown that Campylobacter jej'uni also may produce acholera like enterotoxin.22

It is concluded that Campylobacter infectionsproduce the histological picture of acute infectiouscolitis. The histological diagnosis of acute infectiouscolitis, however, is difficult especially in the laterstages. Immunohistochemical demonstration of im-munoglobulin containing cells and Campylobactermicroorganisms contributes significantly to the his-tological diagnosis of Campylobacter colitis.

The authors wish to express their gratitude to Mrs PP Blommers and Mrs R J J R Scholte who preparedthe manuscript.

References

1 Blaser MJ, Berkowitz ID, La Force FM, Cravens J,

Reller LB, Wang WLL. Campylobacter enteritis:clinical and epidemiological features. Ann Intern Med1979; 91: 179-85.

2 Drake AA, Gilchrist MJR, Washington JA, HuizengaKA, Van Scoy RE. Diarrhea due to Campylobacterfetus subspecies jejuni. A clinical review of 63 cases.Mayo Clin Proc 1981; 56: 414-23.

3 Skirrow MB. Campylobacter enteritis: a "new" dis-ease. Br Med J 1979; 2: 9-11.

4 McKinley MJ, Taylor M, Sangree MH. Toxic mega-colon with Campylobacter colitis. Conn Med 1980; 44:496-7.

5 Mertens A, De Smet M. Campylobacter cholecystitis.Lancet 1979; I: 1091-2.

6 Lambert ME, Schofield PF, Ironside AG, Mandal BK.Campylobacter colitis. Br Med J 1979; 1: 857-9.

7 Willoughby CP, Piris J, Truelove SC. Campylobactercolitis. J Clin Pathol 1979; 32: 986-9.

8 Price AB, Jewkes J, Sanderson PJ. Acute diarrhoea:Campylobacter colitis and the role of rectal biopsy. JClin Pathol 1979; 32: 990-7.

9 Dickinson RJ, Gilmour MM, McClelland DBH. Rectalbiopsy in patients presenting to an infectious diseaseunit with diarrhoeal disease. Gut 1979; 20: 141-8.

10 Vesterby A, Baandrup U, Jacobsen NO, Albertsen K.Campylobacter enterocolitis, an important differentialdiagnosis in ulcerative colitis. Acta Pathol MicrobiolImmunol Scand 1983; Sect. A 91: 31-33.

11 Kumar NB, Nostrant TT, Appelman HD. The histo-pathological spectrum of acute self-limited colitis(acute infectious-type colitis). Am J Surg Pathol 1982;6: 523-9.

12 Bosman FT, Lindeman J, Kuipers G, van der Wal AM,Kreuning J. The influence of fixation on immuno-reactivity of plasma cells in routinely processed intes-tinal biopsy specimens. Histochemistry 1977; 53: 57-62.

13 Sternberger L. Immunohistochemistry. EnglewoodCliffs, NJ: Prentice Hall 1979; 53.

14 Brandtzaeg P, Baklien K, Fausa 0, Hoel PS. Immuno-histochemical characterisation of local immunoglobulinformation in ulcerative colitis. Gastroenterology 1974;66: 1123-36.

15 Lauwers S, Vlaes L, Butzler JP. Campylobacterserotyping and epidemiology. Lancet 1981; I: 158-9.

16 Rosekrans PCM, Meijer CJLM, Van der Wal AM,Cornelisse CJ, Lindeman J. Immunoglobulin contain-ing cells in inflammatory bowel disease of the colon: amorphometric and immunohistochemical study. Gut1980; 21: 941-7.

17 Day DW, Mandal BK, Morson BC. The rectal biopsyappearances in salmonella colitis. Histopathology 1978;2: 117-31.

18 Blaser MJ, Reller LB. Campylobacter enteritis. N EnglJ Med 1981; 305: 1444-52.

19 [Editorial]. Lancet 1982; 2: 1437-8.20 King EO. Human infections with vibrio fetus and a

closely related vibrio. J Infect Dis 1957; 101: 119-28.21 Butzler JP, Skirrow MB. Campylobacter enteritis. Clin

Gastroenterol 1979; 8: 737-65.22 Ruiz-Palacios GM, Torres J, Torres N, Escamilla E,

Ruiz-Palacios B. Cholera-like enterotoxin produced byCampylobacter jejuni. Lancet 1983; 2: 250-2.


http://gut.bmj.com

/G


ber 1985. Dow

nloaded from

http://gut.bmj.com/