Tumor and Stem Cell Biology Cancer Research High Aldehyde ... · Christel van den Hoogen 1, Geertje...

Published OnlineFirst June 1, 2010; DOI: 10.1158/0008-5472.CAN-09-3806

Tumor and Stem Cell Biology
Cancer
Research

High Aldehyde Dehydrogenase Activity IdentifiesTumor-Initiating and Metastasis-InitiatingCells in Human Prostate Cancer

Christel van den Hoogen1, Geertje van der Horst1, Henry Cheung1, Jeroen T. Buijs1, Jenny M. Lippitt2,Natalia Guzmán-Ramírez3, Freddie C. Hamdy4, Colby L. Eaton2, George N. Thalmann3,Marco G. Cecchini3, Rob C.M. Pelger1, and Gabri van der Pluijm1

Abstract

Authors' ACenter, LeMedicine aUnited KinUniversityUniversity o

Note: SupResearch O

C. van den

CorresponCenter, Dep2333 ZA L31-715248

doi: 10.115

©2010 Am

www.aacr

Do

Metastatic progression of advanced prostate cancer is a major clinical problem. Identifying the cell(s) oforigin in prostate cancer and its distant metastases may permit the development of more effective treatmentand preventive therapies. In this study, aldehyde dehydrogenase (ALDH) activity was used as a basis to isolateand compare subpopulations of primary human prostate cancer cells and cell lines. ALDH-high prostate can-cer cells displayed strongly elevated clonogenicity and migratory behavior in vitro. More strikingly, ALDH-highcells readily formed distant metastases with strongly enhanced tumor progression at both orthotopic andmetastatic sites in preclinical models. Several ALDH isoforms were expressed in human prostate cancer cellsand clinical specimens of primary prostate tumors with matched bone metastases. Our findings suggest thatALDH-based viable cell sorting can be used to identify and characterize tumor-initiating and, more impor-tantly perhaps, metastasis-initiating cells in human prostate cancer. Cancer Res; 70(12); 5163–73. ©2010 AACR.

Introduction

Prostate cancer is the most commonly diagnosed cancer inmales and the second leading cause of death. Although pros-tate cancer detected at early stage can be successfully erad-icated by radical prostatectomy and radiotherapy, nocurative treatment exists for metastatic disease. Bone metas-tases occur in ∼90% of patients with advanced disease andare leading causes ofmorbidity. The identification of the cell(s)of origin of prostate cancer as well as the neoplastic cellsinvolved in the formation of distant metastases is, therefore,fundamental to a full understanding of this disease and thedevelopment of treatment and preventive therapies.Most carcinomas comprise of a heterogeneous cell popu-

lation with marked differences in their ability to proliferateand differentiate as well as their ability to reconstitute thetumor upon transplantation. This led to the hypothesis that

ffiliations: 1Department of Urology, Leiden University Medicaliden, the Netherlands; 2Bone Biology Group, School ofnd Biomedical Sciences, University of Sheffield, Sheffield,gdom; 3Department of Urology and Clinical Research,of Bern, Bern, Switzerland; and 4Department of Surgery,f Oxford, Oxford, United Kingdom

plementary data for this article are available at Cancernline (http://cancerres.aacrjournals.org/).

Hoogen and G. van der Horst contributed equally to this work.

ding Author: Gabri van der Pluijm, Leiden University Medicalartment of Urology and Endocrinology, J3-100, Albinusdreef 2,eiden, the Netherlands. Phone: 31-71-5265275; Fax:136; E-mail: [email protected].

8/0008-5472.CAN-09-3806

erican Association for Cancer Research.

journals.org

Researcon October cancerres.aacrjournals.org wnloaded from

the entire population of tumor cells might arise from a raresubpopulation of putative cancer stem/progenitor-like cells,also known as tumor-initiating cells (TICs; refs. 1–3). TICsshare principle characteristics with adult stem cells namelyself-renewal, high proliferative potential, clonogenicity, andmultipotency. In addition, they have the ability to reproduc-ibly form the same tumor phenotype as in the patient and toundergo differentiation into nontumorigenic cells (4–6). TICswere first isolated from patients with hematologic malignan-cies in which a few cells could initiate a new tumor (7). Dur-ing the past few years, TICs were also identified and isolatedfrom solid tumors such as breast, brain, colon, pancreatic,and prostate tumors (8–12). The cancer stem cell hypothesishas provided a paradigm shift in our understanding of carci-nogenesis, metastasis, and tumor biology. The identificationof TICs has important implications in the way cancer treat-ment should be conceived and future therapeutic approacheswill be designed. Whether the subpopulation of TICs is in-volved in the formation of distant metastases, tumor dorman-cy and therapy resistance has remained poorly understood.Because of the heterogeneous nature of solid cancers, the

prospective selection of cells based on cell surface markerssuch as CD44+/α2β1high/CD133+ cells (2) as the sole way toisolate TICs remains controversial (3–6). It also suggests thatselection by (single) cell surface markers such as CD133 isunlikely to identify all putative stem or progenitor cell types(3, 4).A complementary strategy for the functional identification

and characterization of normal stem cells and their malig-nant counterparts involves the measurement of aldehyde de-hydrogenase (ALDH) activity. ALDH enzymes have importantfunctions in the development of epithelial homeostasis, and

5163

h. 18, 2020. © 2010 American Association for Cancer

http://cancerres.aacrjournals.org/

van den Hoogen et al.

5164


as a result, deregulation of this class of enzymes has beenimplicated in multiple cancers (13). ALDH activity is impor-tant for drug resistance, cell proliferation, differentiation, andresponse to oxidative stress (14–16). In humans, 19 isoformsof ALDH enzymes have been identified thus far. To date, highALDH activity has been used to identify and select stem-likesubsets in hematopoietic cells (17–19), endothelial pro-genitor cells, and mesenchymal and epithelial stem cells(20–22). It is becoming increasingly clear that ALDH activitycan be used, either alone or in combination with cell surfacemarkers, to identify TICs in hematologic malignancies and asteadily increasing number of solid carcinomas, includingthose of the pancreas, lung, colon, and breast (14, 15, 23–25).These studies showed the feasibility of viable cell sorting

based on ALDH enzyme activity for the isolation of “stem-like”cells based on a developmentally conserved stem/progenitorcell function. In this study, we show that a subpopulation ofhuman prostate cancer cells with high ALDH activity corre-lates with enhanced clonogenicity and invasiveness in vitro.Furthermore, evidence is presented that ALDH-based viablecell sorting can be used to select tumor-initiating as well asmetastasis-initiating cells in preclinical models of orthotopicgrowth and metastasis.

Materials and Methods

Cell lines and culture conditionsHuman osteotropic prostate cancer cell lines PC-3luc,

PC-3M-Pro4lucA6, and PC-3M-Pro4lucBIII (PC-3M-Pro4lucpassaged thrice in the bone to generate highly metastaticclones that are capable of forming bone metastases) weremaintained as previously described (26). LNCaP, C4, C4-2,and C4-2B (27) were maintained as described earlier (2).DU145 cells were grown in DMEM (Life Technologies) contain-ing 4.5 g glucose/L supplemented with 10% FCS, 100 U/mLpenicillin, and 50 μg/mL streptomycin (Invitrogen).

Primary cell culturesProstate tissue biopsies were obtained after radical prosta-

tectomy with informed consent (CME P05.085). Single-cellsuspension was generated by dissociating prostate tissuebiopsies with collagenase I (200 IU/mL) in stem cell mediaat 37°C overnight according to a modified protocol (28).The epithelial and stromal fractions were separated bycentrifugation and the epithelial cells were maintained insupplemented CnT52 stem cell medium (CELLnTEC). Tu-morigenicity of the isolated cells was confirmed by qPCR ofTMPRSS2/ERG (Supplementary Fig. S1 and SupplementaryMethods).

ALDEFLUOR assay and fluorescence-activated cellsorting isolation of cellsALDH activity of the cells was measured using the ALDE-

FLUOR assay kit (StemCell Technologies) according to themanufacturer's protocol (see Supplementary Fig. S2). ALDHsubstrate was added to the cells and converted by intracellu-lar ALDH into a fluorescent product. Cells were concurrently

Cancer Res; 70(12) June 15, 2010

Researcon October cancerres.aacrjournals.org Downloaded from

labeled with fluorescent antibodies CD44-APC, EpCAM-APC(R&D Systems), α6-APC, αv-PE, or CD133/1-APC (Miltenyi).For fluorescence-activated cell sorting, cells were labeled

with the ALDEFLUOR kit and sorted using a FACS ARIA cellsorter (BD Biosciences; ALDHhi, highest 10% ALDH+ cells;ALDHlow, lowest 10% ALDH− cells). Cell viability was con-firmed by trypan blue exclusion postsorting.

Soft agar colony assayCell suspensions (2,500 cells) were prepared using 0.4%

Noble agarose (Becton Dickinson) and overlayed onto a60-mm dish containing a solidified bottom layer of 0.6% aga-rose in medium. Once the top layer solidified, 1 mL of medi-um was placed on top of the cell layer. Plates were incubatedfor 1 to 3 weeks and colonies were measured and counted byusing microscopy (ZEISS Axiovert 200M). Colony area size ofthree representative fields was measured with the Image Proplus software (Media Cybernetics). Only colonies with anarea of >800 μm2 were counted.

Colony-forming assayOn average, one cell per well was seeded into a 96-well

plate and monitored as previously described by the use oflight microscopy (ZEISS Axiovert 200M; ref. 1).

Migration assayTumor cell migration was performed in 8-μm Transwell

migration chambers (Costar). Prestarved cells (6 × 104) wereseeded in the upper chamber and allowed to migrate towardserum-containing medium in the lower chamber. Cells werefixed after 16 hours with 4% paraformaldehyde and stainedwith 0.1% crystal violet (2 mg/mL, Sigma-Aldrich). Three ran-dom fields were counted for each well, and mean numbers ofmigrated cells per field were calculated (29).

ImmunohistochemistryFirst, antibodies were incubated overnight at 4°C (Supple-

mentary Table S1). After being washed, sections were incu-bated with biotinylated secondary antibodies (1:300, 45min; DAKO) followed by strepatavidin-conjugated peroxi-dase (1:200, 30 min room temperature; DAKO). Sections werewashed and stained with AEC reagents (2 mg/mL) andMayer's hematoxylin (Merck). Pictures were taken using aNikon Eclipse 610 microscope.

Tissue microarraysFor the tissue microarray (TMA), 30 tumor samples ob-

tained by radical prostatectomy were collected at the Univer-sity of Sheffield, United Kingdom. The TMA includestriplicate areas representing the largest carcinoma, high-grade prostate intra-epithelial neoplasia, as well as normal-appearing epithelium of each patient sample.

In vivo animal experimentsMouse strains. Male nude (BALB/c nu/nu) mice were

housed in individual ventilated cages under sterile conditionaccording to the local guidelines for laboratory animals(DEC07026 and 09052).

Cancer Research



ALDH and Prostate Cancer Metastasis


Intraosseous inoculation. A single-cell suspension of1 × 104 PC-3M-Pro4luc cells per 10 μL was injected intothe right tibiae of anaesthetized 6-week-old mice as prev-iously described (30).Orthotopic inoculation. A single-cell suspension of 1 × 104

PC-3M-Pro4luc cells per 10 μL PBS was combined with 10 μLMatrigel (growth factor reduced, BD Biosciences) and surgical-ly inoculated into the prostate of 6-week-old nude mice (26).Intracardiac inoculation. A single-cell suspension of

1 × 104 PC-3M-Pro4luc cells per 100 μL PBS was injected intothe left cardiac ventricle of anaesthetized 5-week-old malenude mice as previously described (26, 31).

www.aacrjournals.org


Subcutaneous inoculation. A 10-μL single-cell suspensionof either 100, 1,000 or 1 × 104 PC-3M-Pro4luc cells in PBS wascombined with 10 μL Matrigel and injected s.c. in anaesthe-tized 6-week-old nude mice.For all in vivo experiments, the progression of cancer

cell growth was monitored weekly by bioluminescent imag-ing using the IVIS100 Imaging System (Caliper LifeSciences;ref. 26).

Statistical analysisStatistical analysis was performed using the GraphPad

Prism4.0 software using either t test or ANOVA. Data are

Figure 1. Characterization of ALDHhi and ALDHlow subpopulations in human prostate cancer cells. PC-3M-Pro4luc cells were FACS sorted to isolate ALDHhi

and ALDHlow subpopulations using the ALDEFLUOR assay. A, mRNA expression levels of α2 integrin, CD44, and CD133 in ALDHhi and ALDHlow

subpopulations as detected by qPCR (n = 3; *, P < 0.05). B, representative images of PC-3M-Pro4luc cells showing heterogeneous expression of CD44and α2 integrin in PC-3M-Pro4luc cells. No staining of CD133 was detected. Scale bars, 80 μm. C, left, a representative color dot plot of PC-3M-Pro4cells stained for both ALDH activity and α6 integrin. Cell surface expression of CD44, αv integrin, α6 integrin, and EpCaM in ALDHhi versus ALDHlow cells asdetected by FACS analysis. Data are representative for three independent experiments. N.d., not detectable.

Cancer Res; 70(12) June 15, 2010 5165




5166


presented as the mean ± SEM. P values of ≤0.05 were re-garded as being statistically significant.

Results

ALDHhi prostate cancer cells display enhancedclonogenicity and migration in vitroA potential strategy for the identification and isolation of

cells with a stem cell–like phenotype in hematologic and sev-eral epithelial malignancies involves selection based onALDH enzyme activity (15, 18) using ALDEFLUOR assays.Before the functional testing of clonogenicity, subsets ofALDHhi and ALDHlow cells were isolated from PC-3M-Pro4-luc and C4-2B human prostate cancer cell lines by flow cyto-metry using the ALDEFLUOR assay. First, we investigatedthe presence of prostate cancer stem/progenitor-like cellmarkers α2 integrin, CD44, and CD133 (2). As depictedin Fig. 1A, mRNA expression levels of α2 integrin and CD44were increased in the ALDHhi compared with the ALDHlow

population, whereas CD133 expression was not detected.Immunohistochemical analyses revealed CD44 and α2 in-

tegrin staining, but no detectable CD133 in PC-3M-Pro4luc



and C4-2B cells (data not shown; Fig. 1B), which is in linewith recent findings (3, 4).In addition, relatively strong expression of other potential

prostate tumor stem cell markers, α6 integrin (32), αv integ-rin (33), and the Epithelial Cell Adhesion Molecule EpCAM(34) was observed in the ALDHhi subpopulation comparedwith ALDHlow subpopulation (Fig. 1C). A representative ex-ample of PC-3M-Pro4luc cells that were stained for bothALDH activity and α6 integrin is shown as a color dot plotin Fig. 1C.Functional differences in vitro between FACS sorted

ALDHhi and ALDHlow subpopulations for both cell lines wereassessed by different clonogenic assays (see SupplementaryMethods). The ALDHhi subpopulations of PC-3M-Pro4lucand C4-2B cells formed significantly more colonies than theirrespective ALDHlow cell populations when grown anchorageindependently in soft agar (Fig. 2A and B). In addition, theaverage colony size was significantly increased in the ALDHhi

compared with the ALDHlow subpopulation (respectively,3,254 ± 520 μm2 versus 1,350 ± 96 μm2). When plated in vitroat low density (1 cell/well), ALDHhi PC-3M-Pro4luc andALDHhi C4-2B subpopulations both showed significantlymore colony formation than the cells with low ALDH activity

Figure 2. ALDHhi prostate cancer cells show enhanced cell growth, colony formation, anchorage-independent growth, and migration in vitro. A, the numberand size (P < 0.05) of colonies growing anchorage independently in the ALDHhi and ALDHlow subpopulations. B, representative images of the ALDHhi

and ALDHlow PC-3M-Pro4luc colonies. Scale bars, 500 μm. C, the number of colonies per 96-well plate in the single-cell diluted cultures after 2 wk inthe ALDHhi and ALDHlow subpopulations of PC-3M-Pro4luc and C4-2B cells. D, mean numbers of migrated ALDHhi and ALDHlow PC-3M-Pro4luc cells perfield. Data are representative of three independent experiments; P < 0.001.

Cancer Research





(Fig. 2C). In addition to their clonogenic ability, ALDHhi andALDHlow subpopulations were compared for their ability tomigrate (Fig. 2D; ref. 29). ALDHhi PC-3M-Pro4luc cells werefound to be significantly more migratory than the ALDHlow

subpopulation.

Human prostate cancer cell lines and primarycultures of clinical prostate cancer samples contain asubpopulation of cells with high ALDH enzyme activityThe ALDEFLUOR assay was used to assess the presence

and size of the population with high ALDH enzymatic activ-ity in prostate cancer cell lines and primary prostate cancercultures derived from patients that underwent radical pros-tatectomy (Table 1). The LNCaP-C4 series can be used as amodel of human prostate cancer development and mimicsthe natural course of prostate cancer from androgen-dependentand nonmetastatic (LNCaP) to androgen-independent andmetastatic (C4-2B), with the development of mixed osteoblastic/osteolytic bone metastases (27). Interestingly, the size of theALDHhi population coincided with an increase in tumor-igenicity from 0.8% in the poorly tumorigenic androgen-responsive LNCaP to 13% in the highly invasive C4-2B cells.The androgen-independent human prostate cancer cell linePC-3 was repeatedly passaged in vivo to produce clones thatare highly metastatic and osteotropic, PC-3M-Pro4luc and PC-3M-Pro4lucBIII. Similar to the LNCaP series, enhanced tumor-igenicity is associated with an increase in the size of theALDHhi subpopulation in these cell lines.Next, we evaluated the presence of an ALDHhi subpopula-

tion in clinical specimens of primary human prostate cancer.Primary cell cultures from different clinical human prostatecancer samples, obtained from patients undergoing radicalprostatectomy, indicated the presence of an ALDHhi popula-tion with an average size of 8% of all cancer cells (Table 1,second part).

ALDHhi prostate cancer cells show enhancedtumorigenicity and metastatic ability in vivoOur in vitro data showed that cells with an increased

ALDH activity in human prostate cancer cell lines have ahigher clonogenic and migratory capability than cells withlow ALDH activity. Subsequently, we analyzed and comparedthe tumorigenic and metastatic ability of the ALDHhi andALDHlow populations of human prostate cancer cells in vivo.First, the tumorigenic potential of ALDHhi and ALDHlow

cell populations was assessed by s.c. inoculation of the differ-ent prostate cancer cell populations in immunocompromisedmice. To monitor the progression of tumor growth, 100,1,000, and 10,000 cells were implanted s.c. and measuredweekly by bioluminescent imaging for 28 days (Fig. 3A).The tumor take after inoculation of ALDHhi PC-3M-Pro4luccells was strikingly higher compared with the ALDHlow cellpopulation (Fig. 3A). ALDHlow cells failed to reproduciblygenerate tumors. In contrast to ALDHlow cells, as little as100 cells of the ALDHhi population were already able to formtumors in vivo. Enhanced tumorigenicity and subcutaneousgrowth of the ALDHhi population was observed at all threecell dosages tested (Fig. 3A).



To compare tumorigenicity and metastatic ability of bothcell subpopulations in a more clinically relevant model sys-tem, we investigated the ability of the cells to grow orthoto-pically when inoculated into the mouse prostate. Total tumorburden was significantly elevated in the mice injected withthe ALDHhi cell population compared with ALDHlow cells(Fig. 3B).In addition, we investigated the capacity of both ALDHhi

and ALDHlow subpopulations to grow in bone marrow, thepreferred site of prostate cancer metastasis. We observed en-hanced tumor growth in the bone marrow after intraosseousinjection of 10,000 ALDHhi PC-3M-Pro4luc cells comparedwith ALDHlow cells. Total tumor burden of the mice injectedwith ALDHhi cells was higher at all time points compared withthe mice injected with the ALDHlow subpopulation (Fig. 3C).

ALDHhi subpopulation in prostate cancer cell displayincreased metastatic ability in vivoWe have previously shown that the inoculation of 100,000

PC-3M-Pro4luc cells into the left cardiac ventricle of nudemice, a model for prostate cancer bone metastasis, resultsin the formation of multiple bone metastases. After FACSsorting using the ALDEFLUOR assay, metastasis formationand tumor growth was significantly increased in mice inocu-lated intracardiacally with only 10,000 ALDHhi prostate cancercells versus ALDHlow cell subpopulation (Fig. 4A–D). It is im-portant to note that ALDHlow cells failed to reproducibly gen-erate metastases, although limited growth was observed in

h. 18

Table 1. Flow cytometric analysis of cell linesand primary prostate cancer cultures basedon ALDH enzymatic activity

PCa cell lines

C

, 2020. © 2010 American

ALDH expression (%) ± SEM

LNCaP
0.8 ± 0.4 C4 2.6 ± 0.2 C4-2 6.2 ± 0.6 C4-2B 13.7 ± 0.3 PC-3 6.3 ± 1.0 Pro4luc 25.7 ± 1.9 Pro4lucBIII 31.0 ± 1.3 DU145 2.4 ± 0.3
PCa primary cultures
ALDH expression (%) ± SEM
7123
8.0 ± 0.5 567 0.5 ± 0.1 025 5.5 ± 0.3 585 10.4 ± 0.8 021 12.4 ± 0.2 601 12.5 ± 0.2
NOTE: Human prostate cancer cell lines and primary cultureswere assayed with an ALDEFLUOR kit. Data are presentedas percentage of cells with high ALDH enzymatic activity.Means of three individual experiments ± SEM are shown.

ancer Res; 70(12) June 15, 2010 5167

Association for Cancer



5168


∼25% of the animals. In addition to metastatic tumor burden,the number of both bone and visceral metastases was signif-icantly elevated in the mice injected with the ALDHhi cells ver-sus the ALDHlow cell subpopulation (Supplementary Fig. S3).



Expression of ALDH isoforms in human prostate cancercell lines and primary tumorsWe evaluated the mRNA expression levels of different

ALDH isoforms in several cell lines (Supplementary Table S2).

Figure 3. ALDHhi cells show increased tumorigenicity in vivo. A, tumor take after s.c. injection of 100, 1,000, and 10,000 ALDHhi PC-3M-Pro4luc cells inALDHhi and ALDHlow subpopulations (n = 8/group; *, P < 0.05). Total tumor burden of the mice injected with 10,000 ALDHhi PC-3M-Pro4luc cells (•)compared with the mice injected with 10,000 ALDHlow cells (○). Inset, the first 20 d after inoculation. B, representative images of mice 7, 14, 21, and28 d after orthotopic injection with either 10,000 ALDHhi or ALDHlow PC-3M-Pro4luc cells. Total tumor burden for the mice injected with the ALDHhi

population (•) or the ALDHlow cell subpopulation (○) (n = 8/group; *, P < 0.05). Data are representative of two independent experiments. C, representativeimages of mice intraosseously inoculated with either ALDHhi or ALDHlow PC-3M-Pro4uc cells (10,000) at different days after inoculation. Total tumorburden of the mice injected with ALDHhi (•) or ALDHlow (○) subpopulation (n = 6/group; **, P < 0.01). Inset, the first 20 d after inoculation.

Cancer Research





Different ALDH isoforms (e.g., ALDH3A2, ALDH4A1, ALDH7A1,ALDH9A1, and ALDH18A1) were found to be highly ex-pressed at the transcriptional level in the cell lines.Similar to established cell lines, the mRNA expression

levels of several ALDH isoforms were evaluated in clinicalspecimens and primary cultures of human prostate cancer(Supplementary Table S3). We observed relatively highexpression of certain ALDH isoforms (e.g., ALDH3A2,ALDH4A1, ALDH7A1, ALDH9A1, and ALDH18A1) in primarycultures in a comparable manner as in established humanprostate cancer cell lines.Next, we investigated the potential use of these ALDH iso-

forms as diagnostic and prognostic markers in prostate can-cer by immunostaining of paraffin-embedded sections ofprostate primary tumors and matched bone metastases.We investigated a TMA (n = 30 samples) as well as 10 prima-ry prostate tumors and their 10 matching bone metastasesfor ALDH expression. In contrast to previous published re-ports describing a role for ALDH1 in other solid cancers



(14, 15, 23–25), no significant ALDH1 expression was foundin primary prostate tumors, matching bone metastases andnoncancerous tissue (Supplementary Fig. S4; Fig. 5).ALDH7A1 was expressed in the majority of primary tu-

mors on the TMA (25 of 30; Fig. 5). Interestingly, strongALDH7A1 staining was found in other series of matchingbone metastases, whereas the surrounding bone marrowwas negative (Supplementary Fig. S5; Fig. 5B). In ourTMA, no correlation was found between Gleason scoreand ALDH7A1 expression (Supplementary Fig. S5B). Nosignificant immunohistochemical localization was observedin noncancerous prostate tissue, high-grade prostate intra-epithelial neoplasia, or prostate carcinoma of ALDH3A2 andALDH18A1 (Fig. 5). In addition to ALDH7A1, other ALDHisoforms can potentially contribute to the strong ALDH activ-ity as detected by the ALDEFLUOR assay (Supplementary Ta-bles S1 and S2). Immunohistochemical staining was indeedobserved for ALDH4A1 and ALDH9A1 in primary prostatecancer specimens (Supplementary Fig. S6).

Figure 4. ALDHhi cells show increased metastatic growth in vivo. A, representative images of mice inoculated with either ALDHhi or ALDHlow PC-3M-Pro4luccells (10,000) at day 21, 28, and 35 after inoculation (n = 15/group). B, total tumor burden of ALDHhi (•) and ALDHlow (○) subpopulation. Inset, thefirst 28 d. C, total number of metastases per mouse in the mice injected with either ALDHhi (•) or ALDHlow (○) cells (*, P < 0.05; ***, P < 0.001). D, bonesof mice intracardiacly inoculated with either ALDHhi or ALDHlow cells stained with Goldner staining. B, bone; T, tumor; BM, bone marrow;GP, growth plate. Data are representative of three independent experiments. Scale bars, 80 μm.

Cancer Res; 70(12) June 15, 2010 5169




5170


Discussion

Prostate cancer is a molecularly, phenotypically, and clin-ically heterogeneous disease. In addition to primary tumors,rapid autopsy programs have revealed a remarkable degree



of heterogeneity among tumor cells within bone metastaticsites when comparing different patients as well as multiplesites within individual prostate cancer patients (35, 36).Due to the observed heterogeneity in primary tumors and

metastases, it has been a major challenge to distinguish and

Figure 5. Expression of ALDH isoforms in human prostate tissue, primary prostate tumor, and bone metastases. A, representative images of TMAstained for several ALDH isoforms showing ALDH7A1-positive prostate carcinoma (25 of 30). No staining was found in ALDH1-, ALDH3A2-, or ALDH18A1-stained primary tumors (0 of 30) or normal prostate tissue. As a positive control, either liver or colon tissue was used. B, representative images of ALDH7A1staining in primary prostate cancer and matching bone metastases (right, magnifications of the images). Scale bars, 80 μm.

Cancer Research





select prostate cancer cells with tumor- and metastasis-initi-ating ability. Identification of the cell(s) of origin of prostatecancer as well as the neoplastic cell(s) involved in the forma-tion of distant metastases is, therefore, fundamental to the un-derstanding of carcinogenesis and metastasis. Furthermore,the functional identification of metastasis-initiating cells is aprerequisite for properly targeted therapy of metastaticdisease in advanced prostate cancer.The exact role and nature of TICs in prostate cancer (and

the supportive stroma) in the formation of distant metasta-ses has remained largely elusive. The development of moreeffective cancer therapies in advanced prostate cancer may,thus, benefit from the outcome of new studies and requireselective targeting of this specific subpopulation of metasta-sis-initiating cells.In prostate cancer, pioneering studies have led to the identi-

fication of CD44+/α2β1high/CD133+ prostate cancer stem cells(2). Because of the observed heterogeneity in prostate cancer,the use of single-cell markers for the selection, characteriza-tion/identification, and functional evaluation of stem/progeni-tor-like prostate cancer cells has been amajor impediment andthe reliability of cell surface markers such as CD133 as the soleway to isolate TICs remains controversial to date (3, 37). Dereg-ulation of ALDH enzyme activity is implicated in the patho-physiology of various hematologic and epithelial cancers (13).The introduction of FACS-based viable cell sorting for ALDHactivity (ALDEFLUOR assays) in tumor biology has further sub-stantiated a role of ALDHhi subpopulations of cancer cells incarcinogenesis (15, 18). High ALDH activity, as detected bythe ALDEFLUOR assay, can thus be used as a functionalmarkerto isolate TICs in several types of epithelial cancers, includingthose of breast, lung, and colon (14, 15, 25). The applicability ofthe ALDEFLUOR assay for the functional identification in hu-man prostate cancer is, however, not known.We show here for the first time that high ALDH activity

can be used to isolate human prostate cancer cells with sig-nificantly enhanced clonogenic and migratory propertiesin vitro as well as elevated tumor- and metastasis-initiatingabilities in vivo. The percentage of ALDHhi cells in prostatecancer cell lines also seems to be related to tumorigenicityand metastatic behavior.Our observations in human prostate cancer are in line

with recent data showing elevated ALDH activity in adultmurine prostate stem cells. Murine ALDHhi prostate epithe-lial cells also displayed a higher proliferative potential andwere more effective in generating prostatic tissue in anin vivo prostate reconstitution assay (38).Strikingly, our data show that the ALDHhi subpopulation of

human prostate cancer cells display not only enhanced clono-genicity, migration, and tumorigenicity but also readily formmetastases in vivo using preclinical models of intraprostaticgrowth and experimental metastasis. Importantly, significant-ly higher numbers of metastases were formed by the ALDHhi

subpopulation and (10- to 1,000-fold) less cells are required toinduce experimental metastasis in these models (26, 31). Ourobservations in prostate cancer are in line with the observeddecrease in tumor growth of the ALDHlow cell population inbreast cancer coinciding with diminished lung metastases



(39). Limited growth of the ALDHlow population in prostatecancer was observed in our in vivo models as ALDHlow cellsfailed to reproducibly generate large tumors and (bone) me-tastases. Our in vivo observations together with the in vitrodata further support the notion that the ALDHlow populationpredominantly contains restricted numbers of transit-ampli-fying cells (limited proliferative capacity) and postmitotic, dif-ferentiated cells, whereas the ALDHhi population is enrichedfor stem/progenitor-like prostate cancer cells (high prolifera-tive capacity). In agreement with other studies describing thefunctional identification of TICs in prostate cancer, we showthat the ALDHhi subpopulation expresses the potential stem/progenitor-like cell markers α2, α6, and αv integrins, andCD44 (2, 32), whereas CD133 was not detected (3, 4, 37). Itis believed that the lack of CD133 expression, as described re-cently also by others, may be due to the observed heterogene-ity in prostate cancer. Strikingly, CD44 and integrin adhesionreceptors (particularly α2, α6, and αv) are significantly upre-gulated in these tumor- and metastasis-initiating prostatecancer cells. It has been firmly established that α2, α6, andαv integrins, and CD44 (and their interplay with urokinase re-ceptor and matrix metalloproteinases) play pivotal roles inthe acquisition of an invasive phenotype and bone metastasisin several osteotropic cancers, including prostate tumors (40–43). Taken together, we strongly feel that the elevated expres-sion of the integrins and CD44 may underlie the observed en-hanced migratory, clonogenic, tumor initiation, andmetastasis initiation abilities of the ALDHhi/α2+/α6+/αv+/CD44+ subpopulation of human prostate cancer cells.To the best of our knowledge, our study is the first to ad-

dress the issue of multiple ALDH enzymes in tumor progres-sion and, more importantly perhaps, metastasis initiation. Atpresent, it is not entirely clear to what extent the variousALDH isoforms contribute to the high ALDH activity ob-served in highly tumorigenic and metastatic prostate cancercells. It is important to note that the ALDEFLUOR kit usedhere and in previous reports (15, 44) has only been validatedfor ALDH1A1 (StemCell Technologies), whereas 19 differentisoforms have been identified to date (13). In contrast to otherepithelial cancers, ALDH isoforms (other than ALDH1) maycontribute to ALDH activity and may be indicative of tumor-initiating and metastasis-initiating cells in human prostatecancer (14, 15, 25, 44). Moreover, ALDH3A2 and ALDH18A1are not likely involved in the observed high ALDH activityin human prostate cancer as well. Instead, relatively high ex-pression of ALDH7A1 in prostate cancer cell lines, primarycultures, and in primary prostate cancer tissue and matchedbone metastases was found.At present, it is unclear if high ALDEFLUOR activity is

functionally involved in stemness (normal and cancer) or ifit is a useful biomarker (“flag”) to identify tumor-initiatingand metastasis-initiating cells in prostate cancer. Further re-search is warranted to identify which of the 19 isoforms, be-sides ALDH7A1, are contributing to the observed highALDEFLUOR activity of highly tumorigenic and metastatichuman prostate cancer cells.Given the fact that there is no real consensus in stem cell

surface markers in prostate cancers, implementation of

Cancer Res; 70(12) June 15, 2010 5171




5172


markers of universal stemness based on intrinsic propertiesinstead of phenotype seems invaluable for the stem cell re-search field.It should also be noted that ALDH enzyme activity is fre-

quently required for the detoxification of compounds, thusserving to protect (stem) cells, and maintaining cellular in-tegrity. Indeed, several studies involving human cancersshowed the ability of this class of enzymes to cause resis-tance to chemotherapeutic agents such as cyclophosphamide(45, 46). In addition, ALDH enzymes also play important rolesin retinoic acid metabolism (47) and androgen receptor bind-ing (48), which are both involved in prostate development,prostate cancer progression, and metastasis formation.In summary, the overall results of this study show that

ALDH-based sorting of human prostate cancer cells byALDEFLUOR can be potentially used simultaneously to selectfor metastasis-initiating cells and TICs. Analysis of ALDH ac-tivity of clinical prostate cancer samples may thus becomeuseful for the stratification of prostate cancer patients at riskof developing metastatic disease.



Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank the Department of Pathology (University of Sheffield, Sheffield,United Kingdom) for supplying the TMA, Maggy van Wijk (Department ofEndocrinology, LUMC, the Netherlands) for the technical support, CarolineJong-Mom (Department of Urology, LUMC, the Netherlands) for the clinicalsampling, and Prof. N. Maitland and Dr. A. Collins (University of York, York,United Kingdom) for their technical support with primary prostate cancercultures and helpful suggestions.

Grant Support

PROMET, a European Specific Targeted Research Project.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received 10/15/2009; revised 03/23/2010; accepted 03/31/2010; publishedOnlineFirst 06/01/2010.

References
1. Li H, Chen X, Calhoun-Davis T, Claypool K, Tang DG. PC3 human
prostate carcinoma cell holoclones contain self-renewing tumor-initiating cells. Cancer Res 2008;68:1820–5.

2. Collins AT, Berry PA, Hyde C, Stower MJ, Maitland NJ. Prospectiveidentification of tumorigenic prostate cancer stem cells. Cancer Res2005;65:10946–51.

3. Guzman-Ramirez N, Voller M, Wetterwald A, et al. In vitro propaga-tion and characterization of neoplastic stem/progenitor-like cellsfrom human prostate cancer tissue. Prostate 2009;15:1683–93.

4. Pfeiffer MJ, Schalken JA. Stem cell characteristics in prostate cancercell lines. Eur Urol 2010;57:246–54. Epub 2009 Jan 19.

5. Clarke MF, Dick JE, Dirks PB, et al. Cancer stem cells-perspectiveson current status and future directions: AACR Workshop on cancerstem cells. Cancer Res 2006;66:9339–44.

6. Fillmore C, Kuperwasser C. Human breast cancer stem cell markersCD44 and CD24: enriching for cells with functional properties in miceor in man? Breast Cancer Res 2007;9:303.

7. Lapidot T, Sirard C, Vormoor J, et al. A cell initiating human acutemyeloid leukaemia after transplantation into SCID mice. Nature1994;367:645–8.

8. Al-Hajj M, Wicha MS, ito-Hernandez A, Morrison SJ, Clarke MF.Prospective identification of tumorigenic breast cancer cells. ProcNatl Acad Sci U S A 2003;100:3983–8.

9. Singh SK, Clarke ID, Terasaki M, et al. Identification of a cancer stemcell in human brain tumors. Cancer Res 2003;63:5821–8.

10. Lee CJ, Dosch J, Simeone DM. Pancreatic cancer stem cells. J ClinOncol 2008;26:2806–12.

11. Ricci-Vitiani L, Lombardi DG, Pilozzi E, et al. Identification and ex-pansion of human colon-cancer-initiating cells. Nature 2007;445:111–5.

12. Kelly K, Yin JJ. Prostate cancer and metastasis initiating stem cells.Cell Res 2008;18:528–37.

13. Marchitti SA, Brocker C, Stagos D, Vasiliou V. Non-P450 aldehydeoxidizing enzymes: the aldehyde dehydrogenase superfamily. ExpertOpin Drug Metab Toxicol 2008;4:697–720.

14. Huang EH, Hynes MJ, Zhang T, et al. Aldehyde dehydrogenase 1 is amarker for normal and malignant human colonic stem cells (SC) andtracks SC overpopulation during colon tumorigenesis. Cancer Res2009;69:3382–9.

15. Ginestier C, Hur MH, Charafe-Jauffret E, et al. ALDH1 is a marker ofnormal and malignant human mammary stem cells and a predictor ofpoor clinical outcome. Cell Stem Cell 2007;1:555–67.

16. Douville J, Beaulieu R, Balicki D. ALDH1 as a functional marker ofcancer stem and progenitor cells. Stem Cells Dev 2009;18:17–26.

17. Armstrong L, Stojkovic M, Dimmick I, et al. Phenotypic characteriza-tion of murine primitive hematopoietic progenitor cells isolated onbasis of aldehyde dehydrogenase activity. Stem Cells 2004;22:1142–51.

18. Cheung AM, Wan TS, Leung JC, et al. Aldehyde dehydrogenaseactivity in leukemic blasts defines a subgroup of acute myeloid leu-kemia with adverse prognosis and superior NOD/SCID engraftingpotential. Leukemia 2007;21:1423–30.

19. Storms RW, Trujillo AP, Springer JB, et al. Isolation of primitivehuman hematopoietic progenitors on the basis of aldehyde dehydro-genase activity. Proc Natl Acad Sci U S A 1999;96:9118–23.

20. Gentry T, Foster S, Winstead L, Deibert E, Fiordalisi M, Balber A.Simultaneous isolation of human BM hematopoietic, endothelialand mesenchymal progenitor cells by flow sorting based on alde-hyde dehydrogenase activity: implications for cell therapy. Cytother-apy 2007;9:259–74.

21. Povsic TJ, Zavodni KL, Kelly FL, et al. Circulating progenitor cellscan be reliably identified on the basis of aldehyde dehydrogenaseactivity. J Am Coll Cardiol 2007;50:2243–8.

22. Matsui W, Huff CA, Wang Q, et al. Characterization of clonogenicmultiple myeloma cells. Blood 2004;103:2332–6.

23. Pearce DJ, Taussig D, Simpson C, et al. Characterization of cellswith a high aldehyde dehydrogenase activity from cord blood andacute myeloid leukemia samples. Stem Cells 2005;23:752–60.

24. Jelski W, Szmitkowski M. Alcohol dehydrogenase (ADH) and alde-hyde dehydrogenase (ALDH) in the cancer diseases. Clin Chim Acta2008;395:1–5.

25. Ucar D, Cogle CR, Zucali JR, et al. Aldehyde dehydrogenase activityas a functional marker for lung cancer. Chem Biol Interact 2009;178:48–55.

26. Buijs JT, Rentsch CA, van der Horst G, et al. BMP7, a putative reg-ulator of epithelial homeostasis in the human prostate, is a potentinhibitor of prostate cancer bone metastasis in vivo. Am J Pathol2007;171:1047–57.

27. Thalmann GN, Sikes RA, Wu TT, et al. LNCaP progression model ofhuman prostate cancer: androgen-independence and osseous me-tastasis. Prostate 2000;44:91–103.

28. Collins AT, Habib FK, Maitland NJ, Neal DE. Identification andisolation of human prostate epithelial stem cells based on α (2) β(1)-integrin expression. J Cell Sci 2001;114:3865–72.

Cancer Research





29. Petersen M, Pardali E, van der Horst G, et al. Smad2 and Smad3have opposing roles in breast cancer bone metastasis by differential-ly affecting tumor angiogenesis. Oncogene 2010;29:1351–61.

30. Wetterwald A, van der Pluijm G, Que I, et al. Optical imaging of can-cer metastasis to bone marrow: a mouse model of minimal residualdisease. Am J Pathol 2002;160:1143–53.

31. Buijs JT, Henriquez NV, van Overveld PG, et al. Bone morphogeneticprotein 7 in the development and treatment of bone metastases frombreast cancer. Cancer Res 2007;67:8742–51.

32. Xin L, Lukacs RU, Lawson DA, Cheng D, Witte ON. Self-renewal andmultilineage differentiation in vitro from murine prostate stem cells.Stem Cells 2007;25:2760–9.

33. Birnie R, Bryce SD, Roome C, et al. Gene expression profiling ofhuman prostate cancer stem cells reveals a pro-inflammatory pheno-type and the importance of extracellular matrix interactions. GenomeBiol 2008;9:R83.

34. Munz M, Baeuerle PA, Gires O. The emerging role of EpCAM incancer and stem cell signaling. Cancer Res 2009;69:5627–9.

35. Roudier MP, True LD, Higano CS, et al. Phenotypic heterogeneity ofend-stage prostate carcinoma metastatic to bone. Hum Pathol 2003;34:646–53.

36. Shah RB, Mehra R, Chinnaiyan AM, et al. Androgen-independentprostate cancer is a heterogeneous group of diseases: lessons froma rapid autopsy program. Cancer Res 2004;64:9209–16.

37. Eaton CL, Colombel M, van der Pluijm G, et al. Evaluation of thefrequency of putative prostate cancer stem cells in primary and me-tastatic prostate cancer. Prostate 2010;70:875–82.

38. Burger PE, Gupta R, Xiong X, et al. High ALDH activity: a novelfunctional marker of murine prostate stem/progenitor cells. StemCells 2009;27:2220–8.

39. Sobel RE, Sadar MD. Cell lines used in prostate cancer re-search: a compendium of old and new lines-part 1. J Urol 2005;173:342–59.

40. Desai B, Rogers MJ, Chellaiah MA. Mechanisms of osteopontin and



CD44 as metastatic principles in prostate cancer cells. Mol Cancer2007;6:18.

41. Ports MO, Nagle RB, Pond GD, Cress AE. Extracellular engagementof α6 integrin inhibited urokinase-type plasminogen activator-mediated cleavage and delayed human prostate bone metastasis.Cancer Res 2009;69:5007–14.

42. Van Slambrouck S, Jenkins AR, Romero AE, Steelant WF. Reorgani-zation of the integrin α2 subunit controls cell adhesion and cancercell invasion in prostate cancer. Int J Oncol 2009;34:1717–26.

43. Nemeth JA, Cher ML, Zhou Z, Mullins C, Bhagat S, Trikha M. Inhibi-tion of α (v) β3 integrin reduces angiogenesis, bone turnover, andtumor cell proliferation in experimental prostate cancer bone metas-tases. Clin Exp Metastasis 2003;20:413–20.

44. Croker AK, Goodale D, Chu J, et al. High aldehyde dehydrogenaseand expression of cancer stem cell markers selects for breast cancercells with enhanced malignant and metastatic ability. J Cell Mol Med2009;13:2236–52. Epub 2008 Aug 4.

45. Moreb JS, Mohuczy D, Ostmark B, Zucali JR. RNAi-mediated knock-down of aldehyde dehydrogenase class-1A1 and class-3A1 is spe-cific and reveals that each contributes equally to the resistanceagainst 4-hydroperoxycyclophosphamide. Cancer Chemother Phar-macol 2007;59:127–36.

46. Sladek NE, Kollander R, Sreerama L, Kiang DT. Cellular levels of al-dehyde dehydrogenases (ALDH1A1 and ALDH3A1) as predictors oftherapeutic responses to cyclophosphamide-based chemotherapyof breast cancer: a retrospective study. Rational individualization ofoxazaphosphorine-based cancer chemotherapeutic regimens. Can-cer Chemother Pharmacol 2002;49:309–21.

47. Zhao D, McCaffery P, Ivins KJ, et al. Molecular identification of a ma-jor retinoic-acid-synthesizing enzyme, a retinaldehyde-specific dehy-drogenase. Eur J Biochem 1996;240:15–22.

48. Pereira F, Rosenmann E, Nylen E, KaufmanM, Pinsky L, Wrogemann K.The 56 kDa androgen binding protein is an aldehyde dehydrogenase.Biochem Biophys Res Commun 1991;175:831–8.

Cancer Res; 70(12) June 15, 2010 5173



2010;70:5163-5173. Published OnlineFirst June 1, 2010.Cancer Res Christel van den Hoogen, Geertje van der Horst, Henry Cheung, et al. Prostate CancerTumor-Initiating and Metastasis-Initiating Cells in Human High Aldehyde Dehydrogenase Activity Identifies

Updated version

10.1158/0008-5472.CAN-09-3806doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2010/06/01/0008-5472.CAN-09-3806.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/70/12/5163.full#ref-list-1

This article cites 48 articles, 14 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/70/12/5163.full#related-urls

This article has been cited by 26 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/70/12/5163To request permission to re-use all or part of this article, use this link

Research. on October 18, 2020. © 2010 American Association for Cancercancerres.aacrjournals.org Downloaded from


http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-09-3806

http://cancerres.aacrjournals.org/content/suppl/2010/06/01/0008-5472.CAN-09-3806.DC1

http://cancerres.aacrjournals.org/content/70/12/5163.full#ref-list-1

http://cancerres.aacrjournals.org/content/70/12/5163.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/70/12/5163


Tumor and Stem Cell Biology Cancer Research High Aldehyde ... · Christel van den Hoogen 1, Geertje...

Documents

Transcript of Tumor and Stem Cell Biology Cancer Research High Aldehyde ... · Christel van den Hoogen 1, Geertje...