TRANSLATIONAL AND CLINICAL RESEARCH

Transcription

1 TRANSLATIONAL AND CLINICAL RESEARCH Brief Report: A Bioassay to Identify Primary Human Prostate Cancer Repopulating Cells ROXANNE TOIVANEN, a DAVID M. BERMAN, b HONG WANG, a JOHN PEDERSEN, a,c MARK FRYDENBERG, a,d ALAN K. MEEKER, b STUART J. ELLEM, a GAIL P. RISBRIDGER, a RENEA A. TAYLOR a a Prostate and Breast Cancer Research Group, Department of Anatomy and Developmental Biology; and d Department of Surgery, Faculty of Medicine, Monash University, Clayton, Victoria, Australia; b Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA; c TissuPath Laboratories, Hawthorn, Victoria, Australia Key Words. Cancer Stromal cells Experimental models Immune-deficient models ABSTRACT Cancer cells are heterogeneous in both their phenotypes and ability to promote tumor growth and spread. Xenografting is used to identify the most highly capable cells of regenerating tumors, referred to as cancer repopulating cells. Because prostate cancers (PCa s) rarely grow as xenografts, indentifying PCa repopulating cells has not been possible. Here, we report improved methods to xenograft localized primary PCa tissues using chimeric grafts with neonatal mouse mesenchyme. Xenograft survival of tumor tissue was significantly increased by neonatal mesenchyme (six of six patients, 66% of grafts, versus four of six patients, 41% of grafts) and doubled the proliferation index of xenografted cancer cells. When applied to isolated Disclosure of potential conflicts of interest is found at the end of this article. PCa cells, neonatal mesenchyme effectively reconstituted PCa s and increased xenograft survival (four of nine patients; 32% of grafts with mesenchyme and 0% without), and supported active cancer cell proliferation. Using this assay, we showed that unfractionated a2b1integrin hi and a2b1integrin lo cells from primary localized PCa s demonstrated tumor formation at comparable rates, similar to previous reports using metastatic specimens. Thus, this new protocol efficiently established tumors and enabled proliferative expansion of both intact tumor tissue and fractionated cancer cells, providing a bioassay to identify and therapeutically target PCa repopulating cells. STEM CELLS 2011;29: INTRODUCTION Prostate cancer (PCa) is one of the most common cancers in men [1]. The development of improved treatments requires advances to our fundamental understanding of three cell types; prostate stem cells, cancer cells of origin, and cancer repopulating cells. Prostatic stem cells and cancer cells of origin are benign populations that were isolated from basal [2 4] and luminal cell [5] compartments of primary prostate tissues. However, the imperative is to identify and target cancer repopulating cells, defined as cells from primary PCa s that can regenerate tumors when transplanted in vivo. To date, cancer repopulating cells have not been identified in PCa due to the inefficiency of xenografting primary tumor specimens. Existing human PCa models have been derived from genetically modified cancers or from advanced metastatic cancers [4, 6, 7]. However, more than 90% of PCa s removed at surgery are localized to the prostate [8], and in contrast to metastatic tumors, localized PCa s almost never grow in immune-deficient mice [9, 10]. Recent advances made by xenografting primary tumors in the renal capsule improved graft survival but did not support growth [11, 12]. The goal of this study was to develop a technique to grow primary human PCa s to assay cancer repopulating cell activities of subpopulations within these tumors. a2b1integrin is correlated with in vitro clonogenicity of PCa cells, although metastasis-derived PCa cells with low a2b1integrin expression levels (a2b1integrin lo ) were shown to be equally tumorigenic in vivo when compared with a2b1integrin hi cells [9]. Using our improved xenografting technique, we demonstrate that a2b1integrin is also somewhat dispensable for tumor regeneration by localized PCa cells. The methods described will also provide an essential tool for determining more specific biomarkers that identify PCa repopulating cells. As the adult host mouse microenvironment is insufficient in stimulating proliferation of human localized PCa, our approach was to include neonatal mouse mesenchyme in Author contributions: R.T.: conception and design, collection and assembly of data, analysis and interpretation and manuscript writing; D.M.B.: conception and design, data interpretation, and manuscript writing; H.W. and A.K.M.: collection, analysis, and data interpretation; J.P. and M.F.: provision of patient tissues; S.J.E.: analysis and interpretation of data and manuscript writing; G.P.R.: conception and design, data interpretation and manuscript writing; R.A.T.: conception and design, data interpretation and manuscript writing and final approval of the manuscript. Correspondence: Renea A. Taylor, Ph.D., Prostate and Breast Cancer Research Group, Department of Anatomy and Developmental Biology, School of Biomedical Sciences, Monash University, Wellington Road, Clayton, Victoria, Australia Telephone: ; Fax: ; renea.taylor@monash.edu Received January 19, 2011; accepted for publication April 26, 2011; first published online in STEM CELLS EXPRESS June 14, VC AlphaMed Press /2009/$30.00/0 doi: /stem.668 STEM CELLS 2011;29:

2 Toivanen, Berman, Wang et al Table 1. Patient demographic information and details of prostate cancer specimen pathology Patient ID a Pathology (index tumor) b Age c Pregraft pathology d Alone Tissue With mouse stroma Number of grafts analyzed Cells Alone With mouse stroma Unsorted a2b1 hi a2b1 lo Unsorted a2b1 hi a2b1 lo 1 4 þ 5 (9) 64 4 þ 5 (9) 2 (0) 2 (2) 2 4 þ 3 (7) 54 4 þ 3 (7) 3 (0) 2 (0) 2 (0) 3 3 þ 4 (7) 70 3 þ 4 (7) 1 (0) 1 (0) 1 (0) 2 (0) 2 (0) 2 (0) 4 3 þ 4 (7) 62 3 þ 4 (7) 4 (0) 5 3 þ 4 (7) 62 3 þ 4 (7) 1 (0) 2 (0) 1 (0) 1 (0) 6 3 þ 4 (7) 62 3 þ 3 (6) 1 (0) 1 (0) 1 (1) 2 (1) 7 4 þ 5 (9) 53 4 þ 4 (8) 2 (0) 8 4 þ 3 (7) 65 3 þ 3 (6) 2 (0) 2 (1) 4 (0) 1 (0) 3 (0) 5 (3) 5 (1) 6 (1) 9 4 þ 5 (9) 61 4 þ 3 (7) 3 (2) 2 (2) 4 (4) 10 4 þ 3 (7) 57 3 þ 4 (7) 5 (0) 7 (2) 11 4 þ 3 (7) 62 3 þ 3 (6) 8 (1) 7 (5) 12 4 þ 3 (7) 51 3 þ 3 (6) 9 (6) 6 (6) 13 4 þ 3 (7) 59 3 þ 4 (7) 9 (7) 8 (8) 14 4 þ 3 (7) 65 4 þ 3 (7) 8 (3) 8 (3) Tumor incidence 41% 66%* 0% 0% 0% 27% 25% 42% 0% 32%** ( ), Parenthesis indicates number of grafts that contained prostate cancer (detected by CK8/18 positive foci with absence of p63 positive basal cells). a Length of grafting time for each patient: >4 weeks (n ¼ 6), patient ID 2, 3, 4, 10, 12, and 13; 4 8 weeks (n ¼ 6), patient ID 1, 5, 6, 8, 11, and 12; 8 14 weeks (n ¼ 5), patient ID 7, 9, 11, 13, and 14. b Indicates Gleason pattern reported for patient, not specific to tissues grafted. c Indicates age of patient at time of radical prostatectomy. d Indicates Gleason pattern in pregrafted tissues. *p >.05 or **p, >.005 using Fisher s exact test for comparison between grafts alone and grafts with mouse mesenchyme. chimeric grafts with human localized PCa tissues. We found that contribution of mouse mesenchyme is an essential component of the xenografting process. We report significant long-term (>14 weeks) survival and increased proliferation of chimeric grafts from tissues and/or isolated cells, while recapitulating the essential characteristics of parental PCa tumors. MATERIALS AND METHODS Animals Seminal vesicle mesenchyme was isolated from day 0 BALB/c or C57BL/6J mouse pups (Monash Animal Services, Clayton, Australia, marp.monash.edu) [13]. Graft recipients were male nonobese diabetic severe combined immune deficiency (NOD-SCID) mice (Animal Resources Centre, Canning Vale, Australia, Procedures were approved by the Monash University Standing Committee of Ethics in Animal Experimentation (MMCA/2008/33). Human Tissues Localized PCa specimens were obtained from 14 patients undergoing radical prostatectomy with informed consent. Human subjects research was approved by Cabrini Human Research Ethics ( ) and Monash University (2004/145MC). Specimens were cut into 4 mm 2 pieces and either fixed in formalin (pregraft tissues), grafted or chemically digested into epithelial cell suspensions and sorted based on expression of a2b1integrin [14]. Gleason pattern was confirmed in adjacent fixed sections of pregraft tissues. Chimeric Xenografts For each patient, PCa pieces (approximately 13 individual 4 mm 3 pieces) or isolated PCa cells were prepared with or without neonatal mouse mesenchyme and grafted under the kidney capsule of adult male NOD-SCID mice for 4 14 weeks [15, 16]. Up to six grafts of mixed types (three per kidney) were implanted per mouse (Table 1). Statistics Prism five software (GraphPad Software, San Diego, CA, was used for all analyses. Tumor incidence from PCa tissues or cells, grafted with and without neonatal mouse mesenchyme, was compared using a Fisher s exact test. Proliferative indices were compared using a repeated measures analysis of variance and Newman-Keuls post hoc multiple comparison tests. Data are expressed as mean 6 SEM. Analytical methods including Immunohistochemistry, Fluorescence in situ hybridization (FISH), and Stereology are included in Supporting Information text. RESULTS Neonatal Mouse Mesenchyme Improves Efficiency of PCa Xenografting Using six-patient primary localized PCa specimens, we demonstrated that mouse mesenchyme increased the rate and quality of engraftment (Table 1). Without mouse mesenchyme, 17 of 41 (41%) grafts (representing four of six patients) survived and yielded relatively small tumor foci (Fig. 1A). With mouse mesenchyme, engraftment increased such that 25 of 38 (66%) grafts contained tumor foci (p <.05), representing tumors from all six patients attempted. Chimeric grafts also showed expansion of human PCa tissues when compared with grafts without mouse mesenchyme (Fig. 1A), likely due to a doubling of the epithelial proliferation index, as compared to PCa grafted alone or pregraft tissues (Fig. 1B). Validation of PCa Phenotype in Xenografts Cancer grafts maintained glandular differentiation consistent with their original Gleason patterns and showed cardinal pathological characteristics of human PCa, including prominent macronucleoli (Fig. 1C and Supporting Information

3 1312 Bioassay for Prostate Cancer Repopulating Cells Figure 1. Chimeric xenografts of localized prostate cancer (PCa) tissues maintain proliferation and original tissue pathology. Representative photomicrographs from patient 13 grafted for 4 weeks. (A): Immunohistochemical staining with human-specific antibodies against cytokeratins 8/ 18 (detecting prostatic epithelial/tumor cells) revealed a visible increase in PCa tumor growth when chimeric grafts with mouse mesenchyme are compared with tissues grafted alone. (B): Ki-67 proliferative index quantified in pregraft tissues when compared with PCa tissues grafted alone or chimeric grafts with mouse mesenchyme (n ¼ 4 patients). Data are represented as mean 6 SEM with a repeated measures analysis of variance reporting a significance of p ¼.039 and post-newman-keuls multiple comparison test showed significance between pairing denoted as *. Representative photomicrographs of Ki-67 immunolocalization showed increased proliferation in tumor cells of chimeric grafts when compared with PCa tissues grafted alone. (C): Histopathology of chimeric PCa graft with mouse mesenchyme. H&E staining shows Gleason patterns 3 and 4 in chimeric graft with mouse mesenchyme as comparable to pregraft tissue (shown in inset) and prominent macronucleoli (arrow). Immunohistochemical staining shows loss of p63 (marker of benign basal cells), while dual a-methylacyl-coa racemase, androgen receptor, and prostate specific antigen expression was positive. (D): Combined centromere (green) and telomere (red) fluorescence in situ hybridization shows the presence of human tumor foci (green and red, indicated by *) and stroma derived from both mouse and human cells (indicated arrows). Scale bars ¼ 500 lm (A); 50 lm (B, C); 200 lm (C) (upper left panel) and 100 lm (inset); 50 lm (D) and 200 lm (inset). Abbreviations: AMACR, a-methylacyl-coa racemase; AR, androgen receptor; CK, cytokeratins; PCa, prostate cancer; PSA, prostate specific antigen.

4 Toivanen, Berman, Wang et al Figure 2. Chimeric xenografts of isolated cells from localized prostate cancer (PCa) specimens maintain proliferation and original tissue pathology. Representative photomicrographs from patient 9 grafted for 14 weeks. (A): Immunolocalization of human-specific cytokeratins 8/18 (detecting prostatic epithelial/tumor cells) showed significant expansion of tumor tissue in chimeric grafts with mouse mesenchyme when compared with isolated cells grafted alone. Chimeric grafts show active proliferation as indicated by immunolocalization of Ki-67. (B): Representative photomicrographs of chimeric grafts containing a2b1integrinlo PCa cells with mouse mesenchyme. H&E staining shows Gleason 3 and 4 patterns as comparable to pregraft tissue (shown in inset) and prominent macronucleoli (arrow). Immunohistochemical staining shows loss of p63 (marker of benign basal cells), while a-methylacyl-coa racemase, androgen receptor, and prostate specific antigen expression was positive. (C): H&E staining shows the regeneration of malignant tissue when grafting unfractionated; a2b1integrinhi and a2b1integrinlo sorted cells from the same patient with mouse mesenchyme. Scale bars ¼ 500 lm (A) and 100 lm (right panel); 50 lm (B), 200 lm (upper left panel), and 100 lm (inset); 100 lm (C). Abbreviations: AMACR, a-methylacyl-coa racemase; AR, androgen receptor; CK, cytokeratins; PCa, prostate cancer; PSA, prostate specific antigen Fig. 1) and expression of classic PCa biomarkers, including a-methylacyl-coa racemase (AMACR), androgen receptor, and prostate specific antigen (Fig. 1C and Supporting Information Fig. 1) as well as loss of p63 and high molecular weight cytokeratin positive basal cells. Benign foci identified in pregraft tissues were maintained in chimeric grafts (Supporting Information Fig. 1) and served as controls for biomarker staining. We demonstrated the contribution of mouse cells to grafted human tissues, regardless of whether mouse mesechyme was cografted, using protein nucleic acid FISH probes to dual-label centromeres and telomeres [17] (Fig. 1D). Without mouse mesenchyme, small human-derived tumor foci were surrounded by human stromal cells with sporadic infiltration by mouse host cells. The inclusion of neonatal mouse mesenchyme increased the proportion of mouse stromal cells in the grafts; however, the tumor cells were still of human origin. Collectively, these data demonstrated that the inclusion of neonatal mouse mesenchyme with localized PCa tissue increased survival and promoted proliferation in vivo, significantly improving on previous xenografting methods. Bioassay for Cancer Repopulating Cells To determine if this optimized xenografting method could be adapted to assay for cancer repopulating cell activity, we isolated and sorted cells from nine localized PCa specimens (Table 1). We initiated 53 grafts of isolated cells with mouse mesenchyme from nine patients and compared their growth with 16 grafts of isolated cells without mesenchyme from five patients. None of the 16 PCa cell-only grafts grew tumors, showing only small benign foci in some grafts. In contrast, chimeric grafts from four of nine patients (including 17 of 53 grafts, 32%; p <.01) demonstrated tumor formation; the

5 1314 Bioassay for Prostate Cancer Repopulating Cells remaining five patients grew only benign epithelium. When compared with grafts without mouse mesenchyme, chimeric grafts showed significant expansion of human PCa cells. As a basis for growth, we observed that active cancer cell proliferation was ongoing at the time of harvest (Fig. 2A). Tumor foci maintained classic human PCa morphology, Gleason patterns, and PCa biomarker expression as comparable to pregraft tissues (Fig. 2B and Supporting Information Fig. 2) and were surrounded by mouse stromal cells (Supporting Information Fig. 3). a2b1integrin hi and a2b1integrin lo Fractions of Localized PCa Contain Cancer Repopulating Cells We separated cells into a2b1integrin hi and a2b1integrin lo populations, as both these populations from metastatic cell and xenograft lines (Du145, LAPC4, and LAPC9) formed tumors in vivo, providing a baseline to compare the efficacy of our cancer repopulating model using primary localized specimens [6]. Similarly, all subpopulations of localized PCa examined (unfractionated, a2b1integrin hi,anda2b1integrin lo ) regenerated tumor tissue (Fig. 2C) at comparable rates (27%, 25%, and 42%, respectively) that were not significantly different. These data definitively show that both a2b1integrin hi and a2b1integrin lo subfractions contain cancer repopulating cells in localized prostate tumors. SUMMARY These data provided proof that inclusion of neonatal mouse mesenchyme improved the efficiency of human localized PCa xenografts by maintaining cancer cell proliferation. Validation of xenografts confirmed the generation of tumors that maintained the integrity of the parental specimens, as demonstrated by AMACR expression, loss of basal cells, and recapitulation of Gleason grade. Maintaining active proliferation of human localized PCa tissues in vivo will facilitate major advances in studying the genetics and molecular biology of PCa, which were not possible in previous xenograft models which had low survival and growth rates. This optimized method can be applied to study the cancer repopulating activity of isolated cell populations, an advance that is essential to elucidate the cellular hierarchy of human PCa tumors [18]. Using this assay, we have definitively shown that both a2b1integrin hi and a2b1integrin lo populations from primary localized tumors contain cancer repopulating cells that can regenerate tumors. Future studies can use this assay to identify unique populations of cancer cells with repopulating activity that are potential therapeutic targets for PCa. ACKNOWLEDGMENTS We wish to thank Michelle Richards and Birunthi Niranjan for their skilled technical assistance, Ruth Patterson for consenting and collecting the patient tissues through the Australian Prostate Cancer Bioresource. This work was supported by a Project Grant (ID: ) and Fellowship Award (ID: ; to G.P.R.) from the National Health and Medical Research Council of Australia and funding from the Prostate Cancer Foundation of Australia (Fellowship to R.A.T.), and Victorian Prostate Cancer Research Consortium, and US Department of Defense (ID: PC073334). DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST The authors indicate no potential conflicts of interest. REFERENCES 1 Jemal A, Siegel R, Xu J et al. Cancer statistics, CA Cancer J Clin 2010;60: Lawson DA, Zong Y, Memarzadeh S et al. Basal epithelial stem cells are efficient targets for prostate cancer initiation. Proc Natl Acad Sci USA 2010;107: Leong KG, Wang BE, Johnson L et al. Generation of a prostate from a single adult stem cell. Nature 2008;456: Goldstein AS, Huang J, Guo C et al. Identification of a cell of origin for human prostate cancer. Science 2010;329: Wang X, Kruithof-de Julio M, Economides KD et al. A luminal epithelial stem cell that is a cell of origin for prostate cancer. Nature 2009;461: Patrawala L, Calhoun-Davis T, Schneider-Broussard R et al. Hierarchical organization of prostate cancer cells in xenograft tumors: The CD44þalpha2beta1þ cell population is enriched in tumor-initiating cells. Cancer Res 2007;67: Gu G, Yuan J, Wills M et al. Prostate cancer cells with stem cell characteristics reconstitute the original human tumor in vivo. Cancer Res 2007;67: Penson D, Chan J. Prostate cancer. In: Litwin MS, Saigal CS, eds. Urological Diseases in America. Washington, DC: US Government Printing Office, 2007: Pienta KJ, Abate-Shen C, Agus DB et al. The current state of preclinical prostate cancer animal models. Prostate 2008;68: Taylor RA, Toivanen R, Risbridger G. Stem cells in prostate cancer: Treating the root of the problem. Endocr Relat Cancer 2010;17: R273 R Wang Y, Revelo M, Sudilovsky D et al. Development and characterization of efficient xenograft models for benign and malignant human prostate tissue. Prostate 2005;64: Zhao H, Nolley R, Chen Z et al. Tissue slice grafts: An in vivo model of human prostate androgen signaling. Am J Pathol 2010;177: Donjacour AA, Cunha GR. Induction of prostatic morphology and secretion in urothelium by seminal vesicle mesenchyme. Development 1995;121: Collins AT, Habib FK, Maitland NJ et al. Identification and isolation of human prostate epithelial stem cells based on alpha(2)beta(1)-integrin expression. J Cell Sci 2001;114: Cunha GR, Donjacour AA. Mesenchymal-epithelial interactions: Technical considerations. In: Coffey DS, Bruchovsky N, Gardner WA, Resnick MI, Karr JP, eds. Assessment of Current Concepts and Approaches to the Study of Prostate Cancer. New York: A. R. Liss; 1987: Taylor RA, Cowin PA, Cunha GR et al. Formation of human prostate tissue from embryonic stem cells. Nat Methods 2006;3: Vander Griend DJ, Konishi Y, De Marzo AM et al. Dual-label centromere and telomere FISH identifies human, rat, and mouse cell contribution to multispecies recombinant urogenital sinus xenografts. Prostate 2009;69: Wang ZA, Shen MM. Revisiting the concept of cancer stem cells in prostate cancer. Oncogene 2010;30: See for supporting information available online.