Portal de Periódicos da CAPES

Molecular Profiling of Prostatic Acinar Morphogenesis Identifies PDCD4 and KLF6 as Tissue Architecture–Specific Prognostic Markers in Prostate Cancer

2012; Elsevier BV; Volume: 182; Issue: 2 Linguagem: Inglês

10.1016/j.ajpath.2012.10.024

1525-2191

Chi‐Rong Li, Jimmy J.–M. Su, Wei-Yu Wang, Michael T. Lee, Ting-Yun Wang, Kuan-Ying Jiang, Chein-Feng Li, Jong-Ming Hsu, Chi-Kuan Chen, Marcelo Chen, Shih-Sheng Jiang, Valerie M. Weaver, Kelvin K. Tsai,

Reproductive Biology and Fertility

Histopathological classification of human prostate cancer (PCA) relies on the morphological assessment of tissue specimens but has limited prognostic value. To address this deficiency, we performed comparative transcriptome analysis of human prostatic acini generated in a three-dimensional basement membrane that recapitulates the differentiated morphological characteristics and gene expression profile of a human prostate glandular epithelial tissue. We then applied an acinar morphogenesis–specific gene profile to two independent cohorts of patients with PCA (total n = 79) and found that those with tumors expressing this profile, which we designated acini-like tumors, had a significantly lower risk of postoperative relapse compared with those tumors with a lower correlation (hazard ratio, 0.078; log-rank test P = 0.009). Multivariate analyses showed superior prognostic prediction performance using this classification system compared with clinical criteria and Gleason scores. We prioritized the genes in this profile and identified programmed cell death protein 4 (PDCD4) and Kruppel-like factor 6 (KLF6) as critical regulators and surrogate markers of prostatic tissue architectures, which form a gene signature that robustly predicts clinical prognosis with a remarkable accuracy in several large series of PCA tumors (total n = 161; concordance index, 0.913 to 0.951). Thus, by exploiting the genomic program associated with prostate glandular differentiation, we identified acini-like PCA and related molecular markers that significantly enhance prognostic prediction of human PCA. Histopathological classification of human prostate cancer (PCA) relies on the morphological assessment of tissue specimens but has limited prognostic value. To address this deficiency, we performed comparative transcriptome analysis of human prostatic acini generated in a three-dimensional basement membrane that recapitulates the differentiated morphological characteristics and gene expression profile of a human prostate glandular epithelial tissue. We then applied an acinar morphogenesis–specific gene profile to two independent cohorts of patients with PCA (total n = 79) and found that those with tumors expressing this profile, which we designated acini-like tumors, had a significantly lower risk of postoperative relapse compared with those tumors with a lower correlation (hazard ratio, 0.078; log-rank test P = 0.009). Multivariate analyses showed superior prognostic prediction performance using this classification system compared with clinical criteria and Gleason scores. We prioritized the genes in this profile and identified programmed cell death protein 4 (PDCD4) and Kruppel-like factor 6 (KLF6) as critical regulators and surrogate markers of prostatic tissue architectures, which form a gene signature that robustly predicts clinical prognosis with a remarkable accuracy in several large series of PCA tumors (total n = 161; concordance index, 0.913 to 0.951). Thus, by exploiting the genomic program associated with prostate glandular differentiation, we identified acini-like PCA and related molecular markers that significantly enhance prognostic prediction of human PCA. Prostate cancer (PCA) is a leading cause of cancer-related death in men. For early-stage localized prostate cancer, radical prostatectomy offers the best opportunity to eradicate the disease. However, approximately 15% to 30% of patients with localized disease at diagnosis develop recurrence within 5 to 10 years, and most of these patients subsequently show poor therapeutic outcome.1Bill-Axelson A. Holmberg L. Ruutu M. Haggman M. Andersson S.O. Bratell S. Spangberg A. Busch C. Nordling S. Garmo H. Palmgren J. Adami H.O. Norlen B.J. Johansson J.E. Radical prostatectomy versus watchful waiting in early prostate cancer.N Engl J Med. 2005; 352: 1977-1984Crossref PubMed Scopus (618) Google Scholar, 2Pound C.R. Partin A.W. Eisenberger M.A. Chan D.W. Pearson J.D. Walsh P.C. Natural history of progression after PSA elevation following radical prostatectomy.JAMA. 1999; 281: 1591-1597Crossref PubMed Scopus (2733) Google Scholar Strategies to stratify the initially diagnosed disease into higher-risk patients with PCA would permit a more personalized targeted treatment strategy that could prevent recurrence. Moreover, a deeper understanding of the pathomolecular mechanisms underlying disease recurrence would help to identify new therapeutic targets. Such as most glandular cancers, the malignant transformation of prostatic epithelium involves a gradual loss of cell adhesion and normal glandular architecture.3Gorlov I.P. Byun J. Gorlova O.Y. Aparicio A.M. Efstathiou E. Logothetis C.J. Candidate pathways and genes for prostate cancer: a meta-analysis of gene expression data.BMC Med Genomics. 2009; 2: 48Crossref PubMed Scopus (46) Google Scholar, 4Gorlov I.P. Sircar K. Zhao H. Maity S.N. Navone N.M. Gorlova O.Y. Troncoso P. Pettaway C.A. Byun J.Y. Logothetis C.J. Prioritizing genes associated with prostate cancer development.BMC Cancer. 2010; 10: 599Crossref PubMed Scopus (24) Google Scholar, 5Zhang X. Fournier M.V. Ware J.L. Bissell M.J. Yacoub A. Zehner Z.E. Inhibition of vimentin or beta1 integrin reverts morphology of prostate tumor cells grown in laminin-rich extracellular matrix gels and reduces tumor growth in vivo.Mol Cancer Ther. 2009; 8: 499-508Crossref PubMed Scopus (49) Google Scholar, 6True L. Coleman I. Hawley S. Huang C.Y. Gifford D. Coleman R. Beer T.M. Gelmann E. Datta M. Mostaghel E. Knudsen B. Lange P. Vessella R. Lin D. Hood L. Nelson P.S. A molecular correlate to the Gleason grading system for prostate adenocarcinoma.Proc Natl Acad Sci U S A. 2006; 103: 10991-10996Crossref PubMed Scopus (228) Google Scholar, 7Gleason D.F. Histologic grading of prostate cancer: a perspective.Hum Pathol. 1992; 23: 273-279Crossref PubMed Scopus (840) Google Scholar Loss of the ability to form tissue architectures by prostate epithelial cells has been functionally linked to increased tumorigenicity.5Zhang X. Fournier M.V. Ware J.L. Bissell M.J. Yacoub A. Zehner Z.E. Inhibition of vimentin or beta1 integrin reverts morphology of prostate tumor cells grown in laminin-rich extracellular matrix gels and reduces tumor growth in vivo.Mol Cancer Ther. 2009; 8: 499-508Crossref PubMed Scopus (49) Google Scholar Because human PCA frequently displays considerable intratumoral heterogeneity in glandular differentiation, this spectrum of tissue morphological characteristics is widely used to classify PCA pathological features according to metrics, such as the Gleason grading system.7Gleason D.F. Histologic grading of prostate cancer: a perspective.Hum Pathol. 1992; 23: 273-279Crossref PubMed Scopus (840) Google Scholar Large-scale clinical studies have established the degree of glandular differentiation as a reasonable determinant to assess the clinical behavior of PCA. Specifically, poorly differentiated, high–Gleason grade tumors are typically associated with a higher probability of tumor recurrence, and patients with these tumors often show poorer prognosis.8Stamey T.A. McNeal J.E. Yemoto C.M. Sigal B.M. Johnstone I.M. Biological determinants of cancer progression in men with prostate cancer.JAMA. 1999; 281: 1395-1400Crossref PubMed Scopus (591) Google Scholar, 9Albertsen P.C. Fryback D.G. Storer B.E. Kolon T.F. Fine J. Long-term survival among men with conservatively treated localized prostate cancer.JAMA. 1995; 274: 626-631Crossref PubMed Scopus (393) Google Scholar Nevertheless, this morphological characteristic–based classification system is only modestly prognostic and does not allow for risk stratification of PCA with similar histopathological characteristics. Tumor classification based solely on tissue architecture has failed to provide functional or mechanistic insights into tumor variability. Accordingly, there is a critical need for molecularly based diagnostic assays that can increase the accuracy of disease prognosis and clinical outcome in PCA. Recently, high-throughput genomic profiling techniques have been applied to molecularly characterize several human malignancies, including PCA, with encouraging success.10Singh D. Febbo P.G. Ross K. Jackson D.G. Manola J. Ladd C. Tamayo P. Renshaw A.A. D'Amico A.V. Richie J.P. Lander E.S. Loda M. Kantoff P.W. Golub T.R. Sellers W.R. Gene expression correlates of clinical prostate cancer behavior.Cancer Cell. 2002; 1: 203-209Abstract Full Text Full Text PDF PubMed Scopus (2035) Google Scholar, 11Glinsky G.V. Glinskii A.B. Stephenson A.J. Hoffman R.M. Gerald W.L. Gene expression profiling predicts clinical outcome of prostate cancer.J Clin Invest. 2004; 113: 913-923Crossref PubMed Scopus (398) Google Scholar, 12Stratford J.K. Bentrem D.J. Anderson J.M. Fan C. Volmar K.A. Marron J.S. Routh E.D. Caskey L.S. Samuel J.C. Der C.J. Thorne L.B. Calvo B.F. Kim H.J. Talamonti M.S. Iacobuzio-Donahue C.A. Hollingsworth M.A. Perou C.M. Yeh J.J. A six-gene signature predicts survival of patients with localized pancreatic ductal adenocarcinoma.PLoS Med. 2010; 7: e1000307Crossref PubMed Scopus (190) Google Scholar, 13van't Veer L.J. Dai H. van de Vijver M.J. He Y.D. Hart A.A. Mao M. Peterse H.L. van der Kooy K. Marton M.J. Witteveen A.T. Schreiber G.J. Kerkhoven R.M. Roberts C. Linsley P.S. Bernards R. Friend S.H. Gene expression profiling predicts clinical outcome of breast cancer.Nature. 2002; 415: 530-536Crossref PubMed Scopus (7849) Google Scholar, 14van de Vijver M.J. He Y.D. van't Veer L.J. Dai H. Hart A.A. Voskuil D.W. Schreiber G.J. Peterse J.L. Roberts C. Marton M.J. Parrish M. Atsma D. Witteveen A. Glas A. Delahaye L. van der Velde T. Bartelink H. Rodenhuis S. Rutgers E.T. Friend S.H. Bernards R. A gene-expression signature as a predictor of survival in breast cancer.N Engl J Med. 2002; 347: 1999-2009Crossref PubMed Scopus (5307) Google Scholar, 15Henshall S.M. Afar D.E. Hiller J. Horvath L.G. Quinn D.I. Rasiah K.K. Gish K. Willhite D. Kench J.G. Gardiner-Garden M. Stricker P.D. Scher H.I. Grygiel J.J. Agus D.B. Mack D.H. Sutherland R.L. Survival analysis of genome-wide gene expression profiles of prostate cancers identifies new prognostic targets of disease relapse.Cancer Res. 2003; 63: 4196-4203PubMed Google Scholar The profound prognostic utility of these genomic markers underlines the intrinsic molecular characteristic of tumors as a crucial determinant to their clinical behavior and has laid the framework for personalized medicine. Genomic tools have also been used to molecularly define tumor phenotypes or subtypes. For example, transcript profiling of human PCA has supported the existence of distinct tumor subclasses that are associated with distinct tumor grades and stage.16Lapointe J. Li C. Higgins J.P. van de Rijn M. Bair E. Montgomery K. Ferrari M. Egevad L. Rayford W. Bergerheim U. Ekman P. DeMarzo A.M. Tibshirani R. Botstein D. Brown P.O. Brooks J.D. Pollack J.R. Gene expression profiling identifies clinically relevant subtypes of prostate cancer.Proc Natl Acad Sci U S A. 2004; 101: 811-816Crossref PubMed Scopus (1064) Google Scholar Furthermore, gene expression patterns that correlate with Gleason score and distinguish low- from high-grade PCA have been described.6True L. Coleman I. Hawley S. Huang C.Y. Gifford D. Coleman R. Beer T.M. Gelmann E. Datta M. Mostaghel E. Knudsen B. Lange P. Vessella R. Lin D. Hood L. Nelson P.S. A molecular correlate to the Gleason grading system for prostate adenocarcinoma.Proc Natl Acad Sci U S A. 2006; 103: 10991-10996Crossref PubMed Scopus (228) Google Scholar, 17Bibikova M. Chudin E. Arsanjani A. Zhou L. Garcia E.W. Modder J. Kostelec M. Barker D. Downs T. Fan J.B. Wang-Rodriguez J. Expression signatures that correlated with Gleason score and relapse in prostate cancer.Genomics. 2007; 89: 666-672Crossref PubMed Scopus (78) Google Scholar These molecular patterns of PCA are instructive, and they can help to characterize tumor characteristics. However, the mechanisms underlying the genesis of these molecular variations in human PCA remain to be further explored. Knowledge-based approaches offer an opportunity to identify more rational markers or classification systems that benefit clinical decision making and therapeutic advancement. Such approaches have been used to establish the prognostic roles of gene profiles associated with tumor progenitor cells, stromal activation, or tissue differentiation in several types of solid tumors.18Sotiriou C. Wirapati P. Loi S. Harris A. Fox S. Smeds J. Nordgren H. Farmer P. Praz V. Haibe-Kains B. Desmedt C. Larsimont D. Cardoso F. Peterse H. Nuyten D. Buyse M. Van de Vijver M.J. Bergh J. Piccart M. Delorenzi M. Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis.J Natl Cancer Inst. 2006; 98: 262-272Crossref PubMed Scopus (1585) Google Scholar, 19Chang H.Y. Sneddon J.B. Alizadeh A.A. Sood R. West R.B. Montgomery K. Chi J.T. van de Rijn M. Botstein D. Brown P.O. Gene expression signature of fibroblast serum response predicts human cancer progression: similarities between tumors and wounds.PLoS Biol. 2004; 2: E7Crossref PubMed Scopus (770) Google Scholar, 20Fournier M.V. Martin K.J. Kenny P.A. Xhaja K. Bosch I. Yaswen P. Bissell M.J. Gene expression signature in organized and growth-arrested mammary acini predicts good outcome in breast cancer.Cancer Res. 2006; 66: 7095-7102Crossref PubMed Scopus (100) Google Scholar, 21Liu R. Wang X. Chen G.Y. Dalerba P. Gurney A. Hoey T. Sherlock G. Lewicki J. Shedden K. Clarke M.F. The prognostic role of a gene signature from tumorigenic breast-cancer cells.N Engl J Med. 2007; 356: 217-226Crossref PubMed Scopus (849) Google Scholar Whether a similar approach could be applied to improve the prognostic prediction of PCA has yet to be determined. In this study, we exploited the experimental merits of a physiologically relevant model of tissue organization, thereby identifying a gene expression program that associates with prostate epithelial acinar morphogenesis. We constructed a gene signature that identifies a subset of more differentiated acini-like human PCAs with a favorable outcome. Relative to clinical criteria and Gleason score, this constructed, biologically informed molecular classification scheme displayed a more robust and accurate ability to predict the prognosis of human PCA. Uniquely, this signature consists of gene markers whose gene expression pattern depends on tissue architecture. The strength of this signature was validated as having strong prognostic value using several target genes as surrogate markers of tissue differentiation. Thus, to our knowledge, our results provide the first example of a biologically tractable and clinically instructive molecular signature for PCA that is based on criteria informed by tissue morphological characteristics. Primary human prostate epithelial cells (PrECs; Clonetics Cell Systems; Lonza, Walkersville, MD) were propagated on tissue culture plastics in chemically defined medium, according to the manufacturer's instructions. RWPE-1 cells (ATCC, Manassas, VA) were maintained in keratinocyte–serum-free medium (Invitrogen, Carlsbad, CA) supplemented with bovine pituitary extract, 10 ng/mL epidermal growth factor, and 0.5% penicillin-streptomycin (Invitrogen).22Bello D. Webber M.M. Kleinman H.K. Wartinger D.D. Rhim J.S. Androgen responsive adult human prostatic epithelial cell lines immortalized by human papillomavirus 18.Carcinogenesis. 1997; 18: 1215-1223Crossref PubMed Scopus (331) Google Scholar LNCaP, DU-145, and PC-3 cells (ATCC) were maintained in RPMI 1640 medium (for LNCaP cells) or Dulbecco's modified Eagle's medium (for DU-145 and PC-3 cells; Invitrogen) supplemented with 10% fetal bovine serum and 0.5% penicillin-streptomycin. The cells were embedded and grown within a thick layer of three-dimensional (3D) reconstituted basement membrane (rBM) gel (Matrigel; BD Biosciences, San Jose, CA), as previously described.23Weaver V.M. Petersen O.W. Wang F. Larabell C.A. Briand P. Damsky C. Bissell M.J. Reversion of the malignant phenotype of human breast cells in three-dimensional culture and in vivo by integrin blocking antibodies.J Cell Biol. 1997; 137: 231-245Crossref PubMed Scopus (1205) Google Scholar Whole culture immunofluorescent staining was performed as previously described.24Lee G.Y. Kenny P.A. Lee E.H. Bissell M.J. Three-dimensional culture models of normal and malignant breast epithelial cells.Nat Methods. 2007; 4: 359-365Crossref PubMed Scopus (1000) Google Scholar Confocal imaging was performed using a Nikon Digital Eclipse C1 confocal microscope system (Nikon, Tokyo, Japan). The antibodies used include rat anti-α6 integrin (clone GoH3; Millipore) and mouse anti-GM130 (clone 35; BD Biosciences). Cell nuclei were counterstained with Hoechst 33342 or DAPI (Invitrogen). Total RNA samples were extracted from cells that were embedded and cultivated within the 3D rBM gels for different lengths of time (36 to 48 hours or 7 to 8 days) using TRIzol (Invitrogen) and then purified using an RNeasy minikit and DNase treatment (Qiagen, Valencia, CA). Experiments were performed in triplicate. Whole-genome gene expression analysis was performed on an Affymetrix Human Genome U133A 2.0 Plus GeneChip platform, according to the manufacturer's protocol (Affymetrix, Santa Clara, CA). The hybridization intensity data were processed using the GeneChip Operating software (Affymetrix), and the genes were filtered based on the Affymetrix P/A/M flags to retain those that were present in all three of the samples in at least one of the experimental conditions. A false-discovery rate of <0.025 was used to select differentially expressed genes within a comparison group. The gene expression data have been deposited in the National Center for Biotechnology Information's Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo; accession number GSE30304). Total RNA (1.0 μg) extracted from the 3D culture was used as a template for cDNA synthesis using Moloney Murine Leukemia Virus Reverse Transcriptase (Promega, Madison, WI). cDNA (100 ng) was used as a template for PCR amplification using the LightCycler FastStart DNA MASTERPLUS SYBR Green I Kit, and quantitative real-time RT-PCR analysis was performed on the amplified RNA using the LightCycler FastStart DNA MASTERPLUS SYBR Green I Kit (Roche Diagnostics GmbH, Mannheim, Germany). Oligonucleotide primers were designed using Primer Bank (http://pga.mgh.harvard.edu/primerbank/index.html, last accessed June 10, 2012). The sequences are available on request. We median centered the genes identified from the gene expression profiling experiments and performed average linkage clustering using Cluster software version 2.11 and TreeView software version 1.60 (Eisen Lab, University of California, Berkeley). For functional clustering, the selected genes were uploaded to DAVID (http://david.abcc.ncifcrf.gov, last accessed March 28, 2012) and a ranked list of functional annotations was executed and generated by the system. We surveyed the functional annotations according to the gene ontology biological process, with P < 0.05, and the top 10 functional annotations, along with their associated genes, were displayed in a diagram output using Cytoscape software (http://www.cytoscape.org, last accessed March 28, 2012). Ingenuity Pathway Analysis (Ingenuity Systems, Redwood City, CA) was used to search for the enriched biological networks and functions of the selected genes. The tumor transcriptome data of 50 patients with PCA who underwent radical prostatectomy at Brigham and Women's Hospital (BWH cohort; Boston, MA) and the associated clinical information were previously reported and were kindly provided by Massimo Loda (Dana-Farber Cancer Institute, Boston).10Singh D. Febbo P.G. Ross K. Jackson D.G. Manola J. Ladd C. Tamayo P. Renshaw A.A. D'Amico A.V. Richie J.P. Lander E.S. Loda M. Kantoff P.W. Golub T.R. Sellers W.R. Gene expression correlates of clinical prostate cancer behavior.Cancer Cell. 2002; 1: 203-209Abstract Full Text Full Text PDF PubMed Scopus (2035) Google Scholar Gene expression profiles of PCA and associated clinical information from 29 patients who underwent radical prostatectomy at Stanford University (Stanford, CA), Karolinska Institute (Solna, Sweden), and Johns Hopkins University (Baltimore, MD) (the SU/KI/JHU cohort) were downloaded from Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo; accession number GSE3933).16Lapointe J. Li C. Higgins J.P. van de Rijn M. Bair E. Montgomery K. Ferrari M. Egevad L. Rayford W. Bergerheim U. Ekman P. DeMarzo A.M. Tibshirani R. Botstein D. Brown P.O. Brooks J.D. Pollack J.R. Gene expression profiling identifies clinically relevant subtypes of prostate cancer.Proc Natl Acad Sci U S A. 2004; 101: 811-816Crossref PubMed Scopus (1064) Google Scholar To measure the degree of resemblance between the gene expression profiles of a clinical tumor specimen and a prostatic organoid, we mapped the selected probe sets from different microarray platforms to the same Entrez-gene IDs and then normalized and mean centered the probe hybridization intensity levels (for Affymetrix microarrays) or fluorescence ratios (for cDNA microarrays) of each of the selected probes across all tumors or the prostate architectures. For each tumor, we calculated the Pearson's correlation coefficient between the tumor and the organoid based on the expression profile of the selected gene probes. To identify, from the differentially expressed genes, a set of gene markers that optimally predicted risk of recurrence after radical prostatectomy in human PCA, we used a previously described supervised approach with modifications.25Wang Y. Klijn J.G. Zhang Y. Sieuwerts A.M. Look M.P. Yang F. Talantov D. Timmermans M. Meijer-van Gelder M.E. Yu J. Jatkoe T. Berns E.M. Atkins D. Foekens J.A. Gene-expression profiles to predict distant metastasis of lymph-node-negative primary breast cancer.Lancet. 2005; 365: 671-679Abstract Full Text Full Text PDF PubMed Scopus (1991) Google Scholar Briefly, for each gene, univariate Cox regression analysis was used to measure the correlation between the expression level of the gene (on a log2 scale) and the length of relapse-free survival of the patients with PCA in the BWH cohort. We constructed 1000 bootstrap samples of the patients in the cohort and performed Cox regression analysis on each of the samples. We then determined an estimated P value and an estimated standardized Cox regression coefficient for each gene by calculating the median P values and the median Cox coefficient of the 1000 bootstrap samples, respectively. To ensure the consistency of our model, we selected the genes whose expressional changes during prostatic cystic differentiation were associated with the expected positive (for genes with higher expression levels in cellular aggregates) or negative (for genes up-regulated in prostatic cysts) risk of relapse, as determined by the estimated standardized Cox regression coefficient. The selected genes were then rank ordered according to the estimated P values, and multiple sets of genes were generated by repeatedly adding one more gene each time from the top of the descendingly ranked list, starting from the first three top-ranked genes. We then calculated a recurrence score (Equation 1) to measure the risk of relapse of a patient for a gene set:Recurrencescore=∑i=3kbixi(1) where k is the number of probes in the probe set, bi is the standardized Cox regression coefficient for the ith probe, and xi is the log2 expression level for the ith probe. For each selected probe set, the concordance index (C-index) was used to evaluate the predictive accuracy in survival analysis, where an index of 1.0 is perfect discrimination.26Pencina M.J. D'Agostino R.B. Overall C as a measure of discrimination in survival analysis: model specific population value and confidence interval estimation.Stat Med. 2004; 23: 2109-2123Crossref PubMed Scopus (1149) Google Scholar Sustained knockdown of programmed cell death protein 4 (PDCD4) or Kruppel-like factor 6 (KLF6) in RWPE-1 cells was achieved by retrovirus or lentivirus-mediated RNA interference (RNAi) using oligonucleotide sequences previously described27Narla G. DiFeo A. Yao S. Banno A. Hod E. Reeves H.L. Qiao R.F. Camacho-Vanegas O. Levine A. Kirschenbaum A. Chan A.M. Friedman S.L. Martignetti J.A. Targeted inhibition of the KLF6 splice variant, KLF6 SV1, suppresses prostate cancer cell growth and spread.Cancer Res. 2005; 65: 5761-5768Crossref PubMed Scopus (137) Google Scholar in the pSUPER.retro.neo system (OligoEngine, Seattle, WA) or validated short hairpin RNA oligonucleotides (MISSION shRNA lentiviruses). Amphotropic retrovirus was produced in Phoenix ampho cells (a gift from Garry Nolan, Stanford University), with packaging vectors pCgp and pVSV-G to boost viral titer. Formalin-fixed, paraffin-embedded (FFPE) tissues of human PCA from 61 patients who underwent radical prostatectomy at Chimei Foundational Medical Center (the CFMC cohort) or 50 patients at Mackay Memorial Hospital were acquired and used in conformity with Institutional Review Board–approved protocols (Table 1). The biochemical recurrence of PCA was defined as a prostate-specific antigen (PSA) level of at least 0.4 ng/mL or two consecutive PSA values of 0.2 ng/mL and increasing.28Stephenson A.J. Kattan M.W. Eastham J.A. Dotan Z.A. Bianco Jr., F.J. Lilja H. Scardino P.T. Defining biochemical recurrence of prostate cancer after radical prostatectomy: a proposal for a standardized definition.J Clin Oncol. 2006; 24: 3973-3978Crossref PubMed Scopus (409) Google Scholar Tissue sections were deparaffinized, hydrated, and immersed in citrate buffer at pH 6.0 for epitope retrieval in a microwave. Endogenous peroxidase activity was quenched in 3% hydrogen peroxidase for 15 minutes, and slides were then incubated with 10% normal horse serum to block non-specific immunoreactivity. The antibody was subsequently applied and detected by using the DAKO EnVision kit (Dako, Carpinteria, CA). The antibodies used include anti-PDCD4 (1:4000, clone EPR3431), anti-KLF6 (1:50, polyclonal), and anti–ATP-binding cassette transporter 8 (ABCG1; 1:500, clone EP1366Y) (all from Epitomics, Burlingame, CA). All of the immunohistochemical (IHC) staining was evaluated by two expert pathologists (C.-F.Li and C.-K. Chen), and the staining patterns were quantified using the histological score (H-score).29Budwit-Novotny D.A. McCarty K.S. Cox E.B. Soper J.T. Mutch D.G. Creasman W.T. Flowers J.L. McCarty Jr., K.S. Immunohistochemical analyses of estrogen receptor in endometrial adenocarcinoma using a monoclonal antibody.Cancer Res. 1986; 46: 5419-5425PubMed Google ScholarTable 1Patient CharacteristicsClinical cohortBWH group (n = 50)SU/KI/JHU group (n = 29)CFMC group (n = 61)MMH group (n = 50)Mean (SD) age (years)59.4 (6.8)59.0 (6.2)65.1 (5.6)67.4 (5.4)Gleason score 4-746 (92.0)22 (75.9)29 (47.5)20 (40.0) 8-94 (8.0)7 (24.1)32 (52.5)30 (60.0)Mean (SD) baseline PSA (ng/mL)8.3 (5.8)NA18.8 (30.8)15.2 (3.6)Tumor stage I-II30 (60.0)19 (65.5)53 (86.9)42 (84.0) ≥III20 (40.0)10 (34.5)8 (13.1)8 (16.0)Recurrence9 (18.0)7 (24.1)11 (18.0)12 (24.0)Medium time to recurrence (months)42.911.587.970.2MMH, Mackay Memorial Hospital; NA, not available. Open table in a new tab MMH, Mackay Memorial Hospital; NA, not available. We used the statistical programming language R version 2.14.0 (http://cran.r-project.org) and SPSS software version 10.0 (SPSS, Chicago, IL), to conduct the statistical analysis of data. A two-tailed Student's t-test was used for simple significance testing, and two-tailed Pearson tests were used for correlation analysis. C-statistics analysis was conducted using the R survcomp package. A cutoff value that best discriminates between groups with respect to outcome was determined using the maximal Youden's index.30Pepe M.S. The Statistical Evaluation of Medical Tests for Classification and Prediction. Oxford University Press, New York2003Google Scholar Survival curves were generated using the Kaplan-Meier method. The curves were plotted and compared using the log-rank test using GraphPad Prism software version 5.02 (GraphPad Software, La Jolla, CA). Statistical significance was considered if P < 0.05. The study of tissue-specific differentiation and tumorigenesis was greatly enhanced by the exploitation of 3D organotypic culture assays that use an rBM to recapitulate glandular-like tissues ex vivo.5Zhang X. Fournier M.V. Ware J.L. Bissell M.J. Yacoub A. Zehner Z.E. Inhibition of vimentin or beta1 integrin reverts morphology of prostate tumor cells grown in laminin-rich extracellular matrix gels and reduces tumor growth in vivo.Mol Cancer Ther. 2009; 8: 499-508Crossref PubMed Scopus (49) Google Scholar, 31Yamada K.M. Cukierman E. Modeling tissue morphogenesis and cancer in 3D.Cell. 2007; 130: 601-610Abstract Full Text Full Text PDF PubMed Scopus (1381) Google Scholar, 32Bissell M.J. Rizki A. Mian I.S. Tissue architecture: the ultimate regulator of breast epithelial function.Curr Opin Cell Biol. 2003; 15: 753-762Crossref PubMed Scopus (342) Google Scholar, 33Debnath J. Brugge J.S. Modelling glandular epithelial cancers in three-dimensional cultures.Nat Rev Cancer. 2005; 5: 675-688Crossref PubMed Scopus (843) Google Scholar, 34Webber M.M. Bello D. Kleinman H.K. Hoffman M.P. Acinar differentiation by non-malignant immortalized human prostatic epithelial cells and its loss by malignant cells.Carcinogenesis. 1997; 18: 1225-1231Crossref PubMed Scopus (115) Google Scholar Therefore, we applied this 3D rBM assay to gain molecular insight into prostatic epithelial cell morphogenesis and transformation. We noted that PrECs within 3D rBM grew rapidly within the first few days and formed small cellular aggregates (Figure