Printer Friendly

MicroRNA signature helps distinguish early from late biochemical failure in prostate cancer.

Only a small subset of prostate cancer (PCa) [7] patients has an aggressive tumor that needs adjuvant treatment. For prognostic monitoring, the elevation of prostate-specific antigen (PSA) concentration in blood after prostatectomy (called biochemical failure, or biochemical relapse) remains the only available marker (1). PSA monitoring, however, cannot predict relapse at the time of surgery. Moreover, it has been shown that biochemical failure is not equal to clinical relapse (2). There is currently no biomarker that can accurately determine the risk of relapse at the time of prostatectomy.

Predicting relapse early can have important effects on patient outcome owing to closer follow-up and adjuvant therapy initiated for those with aggressive disease. Different approaches are used to estimate the risk of recurrence, e.g., the risk stratification model by D'Amico et al. (3); probability tables, such as the Partin table (4); or risk scores [e.g., the CAPRA (Cancer of the Prostate Risk Assessment) score] (5) and nomograms (6). However, independent validations have shown that these models still lack accuracy (7).

MicroRNAs (miRNAs) are short, single-stranded RNA molecules that regulate expression of their targets. They play key roles in different biological processes, such as stem cell maintenance, differentiation, organ development, and cancer pathogenesis (8) and metastasis (9, 10). miRNAs are differentially expressed in PCa (11-15) and have been shown to be involved in PCa pathogenesis, tumor progression, and metastasis (16). They also correlate with stage (14), perineural invasion (17), and androgen dependence (18).

In this study, we identified an miRNA signature that can distinguish PCa patients with low-risk disease (no biochemical failure for [greater than or equal to] 36 months after radical prostatectomy) from those who experienced an early biochemical failure (< 36 months after radical prostatectomy). We developed logistic regression models to segregate the 2 groups. We also provide preliminary evidence showing that at least 1 of the miRNAs featured in the models, miR-152, affects PCa cell proliferation. Target prediction indicated that miR-152 can target v-erb-b2 erythroblastic leukemia viral oncogene homolog 3 (ERBB3). [8] We experimentally validated the miRNA-target interactions in cell-line models.

Materials and Methods


Patients were approached for written informed consent before undergoing radical prostatectomy and agreed to donate surplus tissue samples for use in a research protocol approved by a research ethics board. Patient samples were obtained from the Charite School of Medicine, Berlin, Germany, and St Michael's Hospital, Toronto, Canada. A total of 41 patients undergoing radical prostatectomy were included in the discovery phase; 27 suffered biochemical failure within 36 months of surgery (high-risk group), and 14 did not experience biochemical failure for at least 36 months following surgery (low-risk group) (clinical data are summarized in Table 1 in the Data Supplement that accompanies the online version of this report at http:// The validation set consisted of 64 cases, including 35 patients from the high-risk group and 29 patients from the low-risk group.

Pure tumor areas from the primary Gleason grade were punched with use of the manual tissue arrayer (Beecher Instruments Tissue Microarray Technologies), creating cores 0.6 mm in diameter and 5 mm in length. Total RNA was isolated from 5-6 formalin-fixed paraffin-embedded (FFPE) cores with the FFPE miRNEasy kit (Life Technologies), following the manufacturer's protocol. Total RNA from cultured cells was isolated with the miRNEasy kit (Life Technologies), following the manufacturer's protocol.

Three milliliters urine and 0.6 mL serum were used for RNA extraction. Briefly, samples were mixed with 1.5 volumes QIAzol lysis reagent (Qiagen) and were cleaned with chloroform. RNA was precipitated from the aqueous phase overnight in 2 volumes of icecold 100% ethanol in the presence of glycogen. RNA concentration was measured with a NanoDrop 2000 (Thermo Scientific).


PC3 and DU145 human PCa cell lines were purchased from American Type Culture Collection. Transfections were performed in 6-well plates with siPORT-NeoFX transfection agent (Life Technologies) as recommended by the manufacturer. Synthetic miRNA precursors were purchased from Life Technologies. We assessed cell proliferation by counting PCa cells 24 and 48 h after transient transfection, using a Vi-Cell automated viability counter (Beckman Coulter).


We reverse transcribed 500 ng total RNA from each sample using a Megaplex Primer Pool Human Set v2.0 A+ B (Life Technologies) with a TaqMan[R] miRNA reverse-transcription kit as suggested by the manufacturer. cDNA samples of individual patients were analyzed by a TaqMan[R] low-density array human microRNA card set v2.0 A+B. For validation, miRNAs were quantified by miRNA-specific TaqMan miRNA assays (Life Technologies) using 500 ng template total RNA. Reverse-transcription quantitative PCR (RTqPCR) reactions were performed with TaqMan Fast Universal PCR mix (Life Technologies). Ectopic and endogenous miRNA concentrations were normalized against RNU48, and relative expression was calculated by the [DELTA][DELTA] cycle threshold (Ct) method. Reverse transcription and qPCR reactions were performed on an ABI 7900HT machine.

Total cDNA was prepared with a high-capacity cDNA reverse transcription kit, and target mRNA was quantified by RT-qPCR. Expression was normalized against hypoxanthine phosphoribosyltransferase 1 (.HPRT1) and ribosomal protein, large, PO (RPLPO) endogenous expression. The following primers were used for this study: (cDNA)UPLPO, 5'-GGCGACCTGGA AGTCCAACT; (cDNA)RPLP0, 3'-CCATCAGCACCAC AGCCTTC; (cDNA)HPRTl, 5'-TTGCTGACCTGCT GGATTAC; (cDNA)HPRT1, 3'-TCTCCACCAATTAC TTTTATGTCC; (cDNA)ERBB2 (v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2), 5'-TGATAGACACCAACCGCTCTC; (cDNA)ERBB2, 3'-GATTGGGCATGGACTCAAAC; (cDNA)ERBB3, 5'-CTCTGGGTTCTCTGCTGGAT; (cDNA)ERBB3, 3'-TGCACCATACCATGTTCCTC; (cDNA)AR (androgen receptor), 5' -GACCAGATGGCTGTCATTCA; (cDNA)AR, 3'-GGAGCCATCCAAACrCITGA.


The TargetScan 6.2 miRNA prediction program was used for target prediction. Functional annotation was carried out with the Database for Annotation, Visualization, and Integrated Discovery (DAVID). Pathway analysis was performed using DAVID and the DIANAmiRPath TargetScan5.0 PicTar or DIANA-microT-3.0 Strict. Functional annotation was done through the Gene Ontology databases.


A total of 371 miRNAs, or 49.2% of all investigated miRNAs, exhibited very low or undetectable expression levels in all samples (median Ct value of 35 or greater). To ensure accuracy and reproducibility, we focused subsequent analyses on a subset of 130 miRNAs (median Ct value 30 or lower). miRNA features were ranked by significance using 2 analytical techniques, the nonparametric ROC curve and the permutation t-test (using random permutations of participant status labels to derive empirical P values). Nonparametric Kruskal-Wallis tests were used to determine the significance of miRNAs among the biochemical failure categories. A subset of 130 miRNAs with higher expression levels (i.e., median Ct values < 30.0) were selected for further analysis. Study participants were dichotomized into high-risk (i.e., biochemical failure within 3 years of surgery; n = 27) vs low-risk (i.e., biochemical failure more than 4 years after surgery; n = 14) categories. Binary logistic regression models were built to combine the segregation power of individual miRNAs.

miRNA expression values of the study set were clustered by SOTA (the Self-Organizing Tree Algorithm), available through MEV 4.9 (http://www.tm4. org/mev.html).

To compare the effect of different normalization methods, we normalized data in 3 different ways: (a) against the Ct mean values of RNU48 and RNU44; (b) by use of geNorm software (QBasePLUS package); (c) by use of NormFinder v0.953 ( publicationsnormfinder.htm). NormFinder determined miR-186, miR-30c, let-7d, and miR-191 as the best miRNA candidates for internal control.


A Human Phospho-MAPK (mitogen-activated protein kinase) array kit was purchased from R&D. A total of 150 jug total protein was used for each membrane as directed by the manufacturer. Dot blots were quantified by VersaDoc[TM] Imaging Systems. Adjusted intensity values were obtained by use of Quantity One and Image Laboratory programs. [RT.sup.2] PCR profiler array for human phosphoinositide 3-kinase (PI3K) protein family signalling was purchased from SABiosciences.



Of 754 human miRNAs screened, 130 miRNAs showed expression Ct values of <30 and were used for subsequent analysis (see online Supplemental Table 2). Based on the permutation f-test rank, 25 miRNAs were significantly differentially expressed between the 2 risk groups. The ROC area under the curve (AUC) analysis identified 16 miRNAs statistically significant for differentiation between the low- and the high-risk biochemical failure groups (P < 0.05) (see online Supplemental Table 2). miRNAs identified by the 2 methods were highly overlapping. These miRNAs included multiple members of the miR-17-92 family (miR-19a, miR-19b, miR-20a, miR-20b) and the closely related miR-106a. miR-148a, miR-135a, miR-141, miR-19a, miR-19b, miR-374b, miR-26b, miR-20b, miR-374a, miR-29c, and miR-151-5p ranked among the top dysregulated miRNAs (Table 1). Correlations between miRNA expression and age, preoperative PSA concentration, and other clinical parameters were, assessed and no significant correlations were found (see online Supplemental Table 2).

The expression of any single miRNA, however, was insufficient to accurately predict the biochemical failure risk category. Therefore, we conducted stepwise binary logistic regression analyses to construct models with high sensitivity and specificity. To avoid overfitting of models, we included only 2 or 3 miRNAs. Three models were developed and the predictive ability of these models was assessed by applying leave-one-out cross-validation methods. The first model is based on the expression of miR-331-3p and miR-152 (see online Supplemental Figs. 1A and 2). The ROC AUC was 0.931(95% Cl, 0.807-0.987). At a cutoff probability of 0.14, the model correctly classified 26/27 high-risk individuals (96.3%) and 8/14 low-risk individuals (57.1%); with a negative predictive value (NPV) of 8/9 (88.9%). At a cutoff probability of 0.86, the model correctly classified 21/27 high-risk individuals (77.8%) and 14/14 low-risk individuals (100%); with a positive predictive value (PPV) of 21/21 (100%) in our training set.

The second logistic regression model was based on the expression of 3 miRNAs: miR-331-3p, miR-152, and miR-135a (see online Supplemental Fig. IB). In this model, the ROC AUC measured 0.929 (95% Cl, 0.803-0.985), which was very similar to the performance of the first model. At a higher cutoff probability of 0.35, the model correctly classified 25/27 high-risk individuals (92.6%) and 11/14 low-risk individuals (78.6%), NPV = 11/14 (84.6%), while at a cutoff probability of 0.9, the model correctly classified 22/27 high-risk individuals (81.5%) and 14/14 low-risk individuals (100%), PPV = 22/22 (100%), as seen with the first model.

To improve the accuracy of detection of the low-risk group, we also developed a third model, based on the expression of miR-148a and miR-429 (see online Supplemental Fig. 1C online Supplemental Fig. 2). In this case, the ROC AUC value was 0.788 (95% Cl, 0.633-0.900). Setting the cutoff probability at 0.25, the model correctly classified 27/27 high-risk individuals (100%) and 5/14 low-risk individuals (35.7%), NPV = 5/5 (100%). Meanwhile, at a cutoff probability of 0.898, the model correctly classified 11/27 high-risk individuals (40.7%) and 14/14 low-risk individuals (100%), PPV = 11/11 (100%). This model was inferior to the first model in overall accuracy; however, it could identify a subgroup of low-risk individuals with 100% NPV. The performance of the statistical models was also examined by using harmonized cutoff values, which did not result in substantial changes (see online Supplemental Box 1). In addition, different normalization strategies were applied and compared, including normalization against RNU44 and RNU48 mean Ct values, geNorm approach, and NormFinder. Results were comparable

overall, with few exceptions (see online Supplemental Table 4).


We validated our results using miRNA-specific probe-based RT-qPCR for 2 miRNAs that are featured in the logistic regression models: miR-331-3p and miR-152. miR-331-3p showed significant downregulation (P = 0.009) in the low-risk biochemical failure group. Also, in agreement with the initial screening, miR-152 expression was significantly lower in the high-risk group (P = 0.012) (Table 2). To further confirm our results, we validated the differential expression of miR-331-3p and miR-152 on a second independent set of PCa cases (n = 64). We confirmed significant downregulation of miR-331-3p in the low-risk group and significant downregulation of miR-152 in the high risk for biochemical failure patients (P = 0.01), as expected (Table 2).

To further validate our findings, we tested the performance of the first logistic regression model on the independent patient set. A univariate logistic model containing miR-152 alone was predictive for biochemical failure risk (P = 0.043). Inclusion of miR-331-3p to miR-152 provided significant improvement to the predictive ability of the model (P = 0.004). At a cutoff probability of 0.30, the model correctly classified 33/35 high-risk individuals (94.3%) and 8/29 low-risk individuals (27.6%). At the harmonized cutoff probability of 0.28, the model correctly classified 33/35 high-risk individuals (94.3%) and 6/29 low-risk individuals (20.7%).

To further validate differential miRNA expression between the risk groups, we quantified miR-29c expression across the validation set patients. miR-29c expression correlated with PSA-free survival. Moreover, we examined miR-29c expression in metastatic PCa lesions. miR-29c expression was significantly lower in metastatic tumors, confirming its downregulation in more aggressive lesions (see online Supplemental Fig. 3).

Combination of the study set and the validation set was used to assess possible dependence of miRNA expression on age. miR-152 and miR-331-3p did not correlate with age (see online Supplemental Fig. 4). Two other significant microRNAs were tested in normal prostate tissue, urine, and serum and were shown to be independent of age (see online Supplemental Figs. 5 and 6).


Two of the 3 models featured miR-152 and miR-3313p. Moreover, miR-148a, used in the third model, shares the same seed region (CAGUGCA) with miR152 and is therefore predicted to have overlapping targets. miR-148a and miR-152 are reported to regulate cell proliferation, migration, and invasion in cancer (19) and to promote differentiation (20). miR-331-3p has been shown to induce cell-cycle arrest in gastric cancer and to play a role in castration-resistant PCa (21).

We examined the effect of miR-152 on PCa behavior through gain-of-function experiments in cell line models. DU145 cells transfected with miR-152 exhibited a 57% decrease in cell proliferation, and PC3 cells showed a 35% reduction of proliferation compared to control (Fig. 1). Our results show a negative effect of miR-152 on cell proliferation and are in agreement with the decreased miR-152 expression seen in the high-risk biochemical failure group. These results are also consistent with recent reports showing that increased proliferative activity of PCa cells is a prognostic marker for recurrent PCa after radical prostatectomy (22). Independent studies have demonstrated that miR-148a, which is highly similar to miR-152, inhibits cell growth, migration, and invasion of PC3 cells (23).

Transfection of miR-331-3p did not lead to significant changes in the proliferative ability of DU 145 and PC3 cells (Fig. 1). This miRNA was upregulated in the highrisk biochemical failure group (Table 1) and was therefore not expected to reduce the rate of cell proliferation.


To understand the mechanisms by which miR-152 may contribute to PCa progression, we first conducted a functional clustering analysis and pathway analysis on their predicted targets. The most significant predicted pathways included the transforming growth factor-[beta] (TGF-[beta]) signaling, focal adhesion, extracellular matrix (ECM)-receptor interaction, and ERBB signaling pathways (Fig. 2; also see online Supplemental Table 5). The same results were obtained when using the top 5 or top 10 most significantly different miRNAs between the high and low risk for biochemical failure groups. As a control, we performed the same analysis using a random miRNA set and a set of miRNAs that are not significantly dysregulated between high- and low-risk groups. These pathways were not recognized as significant categories.


To experimentally confirm miR-152-(mRNA)fiRBB3 interaction, we overexpressed miR-152 in PC3 and LNCaP cells. The ERBB3 mRNA concentration remained unchanged in PC3 cells treated with siPORT transfection agent or transfected with a random pool of miRNA precursors (RPP). However, ERBB3 mRNA expression dropped by 44% upon miR-152 overexpression (Fig. 3A). We observed a similar decrease of ERBB3 mRNA expression levels when LNCaP cells were transfected with miR-152. AR mRNA expression also decreased upon miR-152 transfection, showing a possible indirect effect of miR-152 (Fig. 3B).

To validate our findings in vivo, we compared ERBB3 expression in patients' samples between the 2 biochemical failure risk groups using the validation set samples from St Michael's Hospital (Toronto). We dichotomized patients to uRBB3-expressing and fiRBB3-nonexpressing groups (Ct cutoff value 39). Only 1 of 8 (12.5%) of high-risk patients had no measurable (mRNA)ERBB3 expression; whereas 8/27 (30%) of low-risk patients failed to show measurable (mRNA)ERBB3 expression, indicating increased (mRNA)ERBB3 concentrations and an inverse correlation with miR-152 in the high-risk category.

To further validate our findings, we took advantage of the publicly available The Cancer Genome Atlas (TCGA) database. ERBB3 expression was significantly higher in PCa patients who experienced early biochemical failure (< 24 months, n = 56) compared with patients with late biochemical failure ([greater than or equal to] 24 months, n = 39, P = 0.009).

The ERBB3 tyrosine kinase is activated by neuregulins (NRGs) or other ERBB and non-ERBB kinases and it connects to the downstream PI3K/AKT/ MAPK mitogenic pathway. ERBB3 can also activate AR independently of the PI3K/AKT pathway by stabilizing AR protein in complex with ERBB2 (24), and secreted ERBB3 promotes bone metastasis (25). The activated ERBB/PI3K/AKT/nuclear factor-[kappa]B pathways were suggested to predict biochemical recurrence of PCa. Moreover, ERBB3 is reported to be responsible for therapy resistance in different cancers and is a promising therapeutic target (26 J. To examine a possible downstream effect of miR-152 on PI3K and MAPK pathways, LNCaP cells were transfected with miR-152. PI3K pathway activity was monitored by RT-qPCR array while MAPK pathway activity was followed by protein profiling array. Preliminary results indicated that miR-152 overexpression interfered with the protein concentrations of p-AKTl and p-AKT2 (see online Supplemental Fig. 7 and online Supplemental Table 6). Changes in the PI3K pathway members were less evident, with no molecules reaching a 2-fold difference. It should be noted, however, that miRNA effects on a specific pathway can be the sum of an "miRNA network" rather than an individual miRNA effect.

We also used the TCGA database to compare expression of the downstream PI3K and MAPK pathway members. Our pilot analysis showed differential expression of several members of these pathways for aggressive tumors (high risk for biochemical failure, and/or high Gleason score) compared to less aggressive ones (see online Supplemental Table 7).

In summary, these data indicate that miR-152 is a likely regulator of the Erbb3 and may act through downstream PI3K and MAPK pathways; however, further investigation is required to establish its additional regulatory roles.


In current practice, the combination of PSA concentration, Gleason score, and clinical stage provides the best tool for prediction of PCa survival. Biochemical failure can indicate relapse only after it has occurred; therefore, biomarkers that predict PCa aggressiveness are urgently needed.

Exome sequencing for genetic alterations has confirmed frequent somatic mutations related to PCa. mRNA profiling and epigenetic markers are candidate indicators for biochemical recurrence and prognosis after surgery, and noninvasive markers such as serum-based biomarkers and circulating tumor cells are of emerging interest. miRNAs possess attractive characteristics that make them ideal biomarkers. Their relative stability in formalin-fixed tissues allows extraction and quantification from clinical samples, and a large number of miRNAs can be quantified in parallel by multiplex RT-qPCR. Finally, recent studies have shown that miRNAs are present in a stable form in body fluids like serum and urine and that these fluids can be used for noninvasive testing (10).

Among the 130 miRNAs that showed high expression in our present study, 18 have already been linked to PCa relapse (11-13, 27-30). Our results are in agreement with earlier reports that showed that miR-26b, miR-30a-3p, and miR-26a correlate with biochemical failure (12).

On the basis of our bioinformatics and experimental analyses, the PCa-related ERBB/PI3K/AKT/ MAPK pathway is a potentially important target of miR-152. ERBB signaling was reported as a major steroid-independent activator of AR. ERBB3 overexpression has recently been finked with resistance to the ERBB protein family member ERBB1 inhibitors in lung cancer (31), and mutation of ERBB4, a closely related receptor tyrosine kinase, inhibits the formation of drug-resistant colonies in PCa cell lines (32), making miR-152 interesting not only as a prognostic marker but also as a therapeutic tool. miRNA therapy offers the benefit of simultaneous downregulation of multiple genes and cancer-promoting signaling pathways, and recently miRNA-122 inhibitor has successfully completed the first phase II clinical trial for hepatitis C virus (Santaris Pharma A/S).

In summary, we established 3 statistical models that provide useful information to predict biochemical failure at the time of prostatectomy. These logistic regression models appear to be useful in identifying patients at a high risk for biochemical failure. Our study evaluated a large number of potential miRNA markers on a relatively small-sized training set (41 individuals). This design is inherently predisposed to biased marker selection and can lead to markers that are highly specific for the given patient population but perform less well on an independent or a larger set of patients. Further evaluation of potential miRNA markers on a larger patient population is required to determine their accuracy as markers of PCa progression. Our experimental data support that miR-152 targets (mRNA)f; RBB3 and has an indirect effect on (mRNA)AR expression at the mRNA level; its exact downstream regulatory effect, however, needs further confirmation. The ability to quantify miRNA expression from formalin-fixed tissues and body fluids makes this approach potentially useful in determining prognosis in the clinic. Exploration of the molecular events during PCa progression will likely lead to the identification of new and more effective therapeutic targets.

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.

Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:

Employment or Leadership: K.R. Evans, Ontario Cancer Biomarker Network (OCBN).

Consultant or Advisory Role: P. Kupchak, Ontario Cancer Biomarker Network (independent statistical consulting contract).

Stock Ownership: None declared.

Honoraria: None declared.

Research Funding: Prostate Cancer Canada, no. 2010-555, the Movember Foundation grant no. D2013-39 (G.M. Yousef, PI); K. lung, Wilhelm Sander-Stiftung; A. Fendler, Wilhelm Sander-Stiftung; C. Stephan, Wilhelm Sander-Stiftung.

Expert Testimony: None declared.

Patents: A patent application describing the miRNAs that are differentially expressed between prostate cancer with early or late biochemical failure has been filed at the MaRS Discovery Centre (Toronto, ON). This is a joint application between St Michael's Hospital (Toronto, ON), Charite - Universitatsmedizin (Charite School of Medicine) (Berlin, Germany) and University Health Network (Toronto, ON) (A. Fendler, K. lung, C. Stephan, N.E. Fleshner, and G.M. Yousef).

Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.


(1.) Maffezzini M, Bossi A, Collette L. Implications of prostate-specific antigen doubling time as indicator of failure after surgery or radiation therapy for prostate cancer. Eur Urol 2007;51:605-13.

(2.) Pound CR, Partin AW, Eisenberger MA, Chan DW, Pearson JD, Walsh PC. Natural history of progres sion after PSA elevation following radical prostatectomy. JAMA 1999;281:1591-7.

(3.) D'Amico AV, Whittington R, Malkowicz SB, Fondurulia J, Chen MH, Tomaszewski JE, Wein A. The combination of preoperative prostate specific antigen and postoperative pathological findings to predict prostate specific antigen outcome in clinically localized prostate cancer. J Urol 1998; 160: 2096-101.

(4.) Partin AW, Kattan MW, Subong EN, Walsh PC, Wojno KJ, Oesterling JE, et al. Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer. A multi-institutional update. JAMA 1997;277:1445-51.

(5.) Cooperberg MR, Pasta DJ, Elkin EP, Litwin MS, Latini DM, Du CJ, Carroll PR. The University of California, San Francisco Cancer of the Prostate Risk Assessment score: a straightforward and reliable preoperative predictor of disease recurrence after radical prostatectomy. J Urol 2005; 173:1938-42.

(6.) Shariat SF, Karakiewicz PI, Suardi N, Kattan MW. Comparison of nomograms with other methods for predicting outcomes in prostate cancer: a critical analysis of the literature. Clin Cancer Res 2008;14:4400-7.

(7.) Aggarwal BB, Danda D, Gupta S, Gehlot P. Models for prevention and treatment of cancer: problems vs promises. Biochem Pharmacol 2009;78: 1083-94.

(8.) Lichner Z, Mejia-Guerrero S, Ignacak M, Krizova A, Bao TT, Girgis AH, et al. Pleiotropic action of renal cell carcinoma-dysregulated miRNAs on hypoxia-related signaling pathways. Am J Pathol 2012;180:1675-87.

(9.) Fendler A, Stephan C, Yousef GM, Jung K. MicroRNAs as regulators of signal transduction in urological tumors. Clin Chem 2011;57:954-68.

(10.) White NM, Fatoohi E, Metias M, Jung K, Stephan C, Yousef GM. Metastamirs: a stepping stone towards improved cancer management. Nat Rev Clin Oncol 2011;8:75-84.

(11.) Ambs S, Prueitt RL, Yi M, Hudson RS, Howe TM, Petrocca F, et al. Genomic profiling of microRNA and messenger RNA reveals deregulated microRNA expression in prostate cancer. Cancer Res 2008;68:6162-70.

(12.) Ozen M, Creighton CJ, Ozdemir M, Ittmann M. Widespread deregulation of microRNA expression in human prostate cancer. Oncogene 2008;27: 1788-93.

(13.) Porkka KP, Pfeiffer MJ, Waltering KK, Vessella RL, Tammela TL, Visakorpi T. MicroRNA expression profiling in prostate cancer. Cancer Res 2007;67: 6130-5.

(14.) Schaefer A, Jung M, Mollenkopf HJ, Wagner I, Stephan C, Jentzmik F, et al. Diagnostic and prognostic Implications of microRNA profiling in prostate carcinoma. Int J Cancer 2010;126:1166-76

(15.) Sun R, Fu X, LI Y, Xle Y, Mao Y. Global gene expression analysis reveals reduced abundance of putative microRNA targets In human prostate tumours. BMC Genomics 2009;10:93.

(16.) Spahn M, Kneltz S, Scholz CJ, Stenger N, Rudiger T, Strobel P, et al. Expression of microRNA-221 is progressively reduced In aggressive prostate cancer and metastasis and predicts clinical recurrence. Int J Cancer 2010;127:394-403.

(17.) Prueitt RL, Yi M, Hudson RS, Wallace TA, Howe TM, Yfantis HG, et al. Expression of microRNAs and protein-coding genes associated with perineural invasion In prostate cancer. Prostate 2008;68:1152-64.

(18.) Xu G, Wu J, Zhou L, Chen B, Sun Z, Zhao F, Tao Z. Characterization of the small RNA transcriptomes of androgen dependent and independent prostate cancer cell line by deep seguencing. PLoS One 2010;5:e15519.

(19.) Zheng X, Chopp M, Lu Y, Buller B, Jiang F. MIR-15b and miR-152 reduce glioma cell invasion and angiogenesis via NRP-2 and MMP-3. Cancer Lett 2012;329:146-54.

(20.) Zhang J, Ying TL, Tang ZL, Long LQ, Li K. MicroRNA-148a promotes myogenic differentiation by targeting the R0CK1 gene. J Biol Chem 2012;287:21093-101.

(21.) Epis MR, Giles KM, Barker A, Kendrick TS, Leedman PJ. mi R-331-3p regulates ERBB-2 expression and androgen receptor signaling in prostate cancer. J Biol Chem 2009;284:24696-704.

(22.) Bettencourt MC, Bauer JJ, Sesterhenn IA, Mostofl FK, McLeod DG, Moul JW. Ki-67 expression is a prognostic marker of prostate cancer recurrence after radical prostatectomy. J Urol 1996;156: 1064-8.

(23.) Zheng B, Liang L, Wang C, Huang S, Cao X, Zha R, et al. MicroRNA-148a suppresses tumor cell invasion and metastasis by downregulating ROCK1 in gastric cancer. Clin Cancer Res 2011; 17:7574-83.

(24.) Mellinghoff IK, Vivanco I, Kwon A, Tran C, Wongvipat J, Sawyers CL. HER2/neu kinase-dependent modulation of androgen receptor function through effects on DNA binding and stability. Cancer Cell 2004;6:517-27.

(25.) Chen N, Ye XC, Chu K, Navone NM, Sage EH, Yu-Lee LY, et al. A secreted Isoform of ErbB3 promotes osteonectin expression in bone and enhances the Invasiveness of prostate cancer cells. Cancer Res 2007;67:6544-8.

(26.) Sithanandam G, Anderson LM. The ERBB3 receptor In cancer and cancer gene therapy. Cancer Gene Ther 2008;15:413-48.

(27.) Hudson RS, YI M, Esposito D, Glynn SA, Starks AM, Yang Y, et al. MicroRNA-106b-25 cluster expression is associated with early disease recurrence and targets caspase-7 and focal adhesion in human prostate cancer. Oncogene 2013;32: 4139-47.

(28.) Saini S, Majid S, Shahryarl V, Arara S, Yamamura S, Chang I, et al. miRNA-708 control of CD44(+) prostate cancer-initiating cells. Cancer Res 2012; 72:3618-30.

(29.) Selth LA, Townley SL, Bert AG, Strieker PD, Sutherland PD, Horvath LG, et al. Circulating microRNAs predict biochemical recurrence In prostate cancer patients. Br J Cancer 2013;109: 641-50.

(30.) Zhang B, Chen H, Zhang L, Dakhova O, Zhang Y, Lewis MT, et al. A dosage-dependent pleiotropic role of Dicer in prostate cancer growth and metastasis. Oncogene [Epub ahead of print 2013 Jul 15]

(31.) Engelman JA, Zejnullahu K, Mitsudomi T, Song Y, Hyland C, Park JO, et al. MET amplification leads to gefitinib resistance In lung cancer by activating ERBB3 signaling. Science 2007,316:1039-43.

(32.) Williams EE, Trout U, Gallo RM, Pitfield SE, Bryant I, Penington DJ, Riese DJ II. A constitutively active ErbB4 mutant inhibits drug-resistant colony formation by the DU-145 and PC-3 human prostate tumor cell lines. Cancer Lett 2003,192: 67-74.

Zsuzsanna Lichner, [1,2] Annika Fendler, [3] Carol Saleh, [1] Aurfan N. Nasser, [2] Dina Boles, [1,2] Sahar Al-Haddad, [2] Peter Kupchak, [4] Moyez Dharsee, [4] Paulo S. Nuin, [4] Kenneth R, Evans, [4] Klaus Jung, [5] Carsten Stephan, [5] Neil E. Fleshner, [6] and George M. Yousef [1,2] *

[1] Department of Laboratory Medicine, Keenan Research Centre in the Li Ka Siting Knowledge Institute St. Michael's Hospital, Toronto, Canada; [2] Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Canada; [3] Max Delbruck Centrum (MDC) for Molecular Medicine, Berlin, Germany; [4] Ontario Cancer Biomarker Network, MaRS Centre, Toronto, Canada; [5] Depart ment of Urology, University Hospital Chante, Berlin, Germany; [6] Department of Surgery, University Health Network, Toronto, Canada.

* Address correspondence to this author at: Department of Laboratory Medicine, St. Michael's Hospital, 30 Bond Street, Toronto, ON, M5B 1W8, Canada. Fax 416-864-5648; e-mail

Received February 22, 2013; accepted July 30, 2013.

Previously published online at DOI: 10.1373/clinchem.201 3.205450

[7] Nonstandard abbreviations: PCa, prostate cancer; PSA, prostate-specific antigen; miRNA, microRNA; FFPE, formalin-fixed paraffin-embedded; RT-gPCR, reverse-transcription guantitative PCR; DAVID, Database for Annotation, Visualization, and Integrated Discovery; Ct, cycle threshold; MARK, mitogen-activated protein kinase; PI3K, phosphoinositide 3-kinase; AUC, area under the curve; NPV, negative predictive value; PPV, positive predictive value; TGF-[beta], transforming growth factor-[beta]; ECM, extracellular matrix; RPP, random pool of miRNA precursors; TCGA, The Cancer Genome Atlas database; NRG, neuregulin.

[8] Human genes: HPRT1, hypoxanthine phosphoribosyltransferase 1; HPRT1, hypoxanthine phosphori bosyltransferase 1; RPLPO, ribosomal protein, large, PO; ERBB2, v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2; AR, androgen receptor; ERBB4, v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 4.

Table 1. Statistical significance of miRNAs
differentially expressed between the low and
high risk for biochemical failure groups.

rniRNA (a)      Upregulated   P (empirical
                risk group    t-test) (b)

miR-148a         Low-risk        0.006
miR-1274b (d)    Low-risk        0.009
miR-141          Low-risk        0.012
miR-135a         Low-risk        0.014
miR-19a          Low-risk        0.015
miR 19b          Low-risk        0.015
miR-374b         Low-risk         0.02
miR 26b          Low-risk        0.023
miR-20b          Low-risk        0.025
miR 374a         Low-risk        0.025
miR-151 -5p      Low-risk          NA
miR-29c          Low-risk          NA
miR-174b         Low-risk          NA
miR-196b         Low-risk          NA
miR-26a          Low-risk          NA
miR-331 -3p      High-risk         NA
miR-193a         High-risk         NA
miR 365          High-risk         NA
miR-125a         High-risk         NA
miR-125b         High-risk         NA

rniRNA (a)      ROC AUC      Number of
                 value     normalization
                          methods showing
                          significance (c)

miR-148a         0.78            5
miR-1274b (d)    0.75          NA (e)
miR-141          0.74            3
miR-135a         0.75            5
miR-19a          0.74            5
miR 19b          0.72            3
miR-374b         0.74            NA
miR 26b          0.74            5
miR-20b           NA             NA
miR 374a          NA             NA
miR-151 -5p      0.71            NA
miR-29c          0.73            5
miR-174b          NA             5
miR-196b          NA             3
miR-26a           NA             3
miR-331 -3p       NA             3
miR-193a          NA             3
miR 365           NA             3
miR-125a          NA             3
miR-125b          NA             3

(a) miRNAs in the table are named according
to the miRNA database miRBase.

(b) P values are given for the top 10
differentially expressed miRNAs.

(c) MicroRNAs significant with
[greater than or equal to] 3 normalization
methods are Indicated.

(d) miR-1274b Is likely a derivative of

(e) NA, not assessed.

Table 2. Validation of miR-152 and miR-331-3p
differential expression on the study set and an
independent set of prostate cancer cases.

                                    miR-152   miR-331-3p

Validation on the study set

  Mean expression fold change         2.2       0.51
    (low- vs high-risk group)
  P                                   0.012     0.009

Validation on an independent set

  Mean expression fold change         2.56      0.53
    (low- vs high-risk group)
  P                                   0.01      0.02
COPYRIGHT 2013 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Molecular Diagnostics and Genetics
Author:Lichner, Zsuzsanna; Fendler, Annika; Saleh, Carol; Nasser, Aurfan N.; Boles, Dina; Haddad, Sahar Al-
Publication:Clinical Chemistry
Article Type:Report
Date:Nov 1, 2013
Previous Article:Noninvasive prenatal methylomic analysis by genomewide bisulfite sequencing of maternal plasma DNA.
Next Article:Quantitative PCR measurement of tRNA 2-methylthio modification for assessing type 2 diabetes risk.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters