Epigenetic regulation of micrornas in prostate cancer

E EP PI IG GE EN NE ET TI IC C R RE EG GU UL LA AT TI IO ON N O OF F M MI IC CR RO OR RN NA AS S I IN N P PR RO OS ST TA AT TE E C CA AN NC CE ER R by JOÃO BARBOSA MARTINS Dissertation to a Master’s Degree in Oncology 2012 II III João Álvaro Barbosa Martins EPIGENETIC REGULATION OF MICRORNAS IN PROSTATE CANCER Dissertation for applying to a Master’s Degree in Oncology - Specialization in Molecular Oncology submitted to the Institute of Biomedical Sciences Abel Salazar – University of Porto Supervisor: Rui Manuel Ferreira Henrique, MD, PhD Guest Assistant Professor Department of Pathology and Molecular Immunology Institute of Biomedical Sciences Abel Salazar – University of Porto Director of Department of Pathology and senior researcher at the Cancer Epigenetics Group of the Research Center Portuguese Oncology Institute – Porto Co-Supervisor: Cármen de Lurdes Fonseca Jerónimo, PhD Guest Associate Professor Department of Pathology and Molecular Immunology Institute of Biomedical Sciences Abel Salazar – University of Porto Assistant Investigator and Coordinator of the Cancer Epigenetics Group Department of Genetics and Research Center Portuguese Oncology Institute – Porto IV V AKNOWLEDGMENTS In the first, place I would like to thank Prof. Rui Henrique for the opportunity provided of making research in such a recognized group such as Cancer Epigenetics Group, for receiving the supervision of this Thesis, for all the time and patience spent with me and my work and for all the support and dedication provided. I am mostly thanked for all comprehension, dedication, confidence and constant encouragement to achieve the aims purposed. In fact, this past year was very demanding as I had to conciliate laboratory research with my work as a pharmacist in I.P.O.F.G.-Porto Pharmacy, but now looking back I realize it worth it. Thank you for the great opportunity to enrich my knowledge on molecular biology and I am positively sure that it will have a great impact in my professional future. I also want to thank to my co-supervisor Prof. Carmen Jerónimo for sharing all her knowledge and I am certain that with her expertise in Epigenetics, this work will be much more enriched. Thank you both for all patience for reading and reviewing this Thesis. To Prof. Carlos Lopes former Director of Oncology Master’s Programme for all the precious knowledge and experience shared during the first year of this Master Programme and also to Prof. Berta Silva actual Director, for all the help provided. A very special thanks to João Ramalho Carvalho and Pedro Pinheiro, for all the assistance provided, not only in laboratorial settings but also in results interpretation and development of this work. Thank to both of you for all the knowledge, hours, patient and optimism provided to me and my research, without which its success couldn’t be achieved. Other very special thank to Ana Oliveira, Filipa Vieira, Inês Graça, Elsa Sousa, Tiago Batista, Natalia Costa, Susana Neto, Sara Reis and all other members of Cancer Epigenetics Group for your conviviality, availability, help and suggestions that much have contributed to the success of this project. To all other colleagues of the Department of Genetics for the great environment provided, especially to Nuno, Sofia, Isabel and Diogo. I would like to give a special praise to all my Oncology Master’s class, especially Diana Mesquita for all her patience and dedication, which made this step in our academic life so much easier. I also want to pay tribute to Ana Sarmento, Ana São João, Michel, VI José Fernandes, Rui, Ida, Isa and all the others for their friendship and to wish them the best luck in professional and personal life. More personally I would like to give a special thank to my entire family for all the love, education and support provided during my entire life, which allowed me to become what I am today. A great and special thank to my mother, once all I became and conquered in life I owe to her and her constant dedication to me. Finally, I would like to thank all my friends, especially those from F.F.U.P. and C.E.S.P.U., for having provided me wonderful moments, once life isn’t just science. A special thank to José Moura, for all the hours we spent learning with each other and discussing science during my degree and during this Master Programme. A special praise to you all, this Thesis is dedicated to you. This study was funded by grants from Research Center of Portuguese Oncology Institute-Porto (Project CI-IPOP-4), and from the European Community’s Seventh Framework Programme – Grant number FP7-HEALTH-F5-2009-241783. VII RELEVANT ABBREVIATIONS 5-AZA-DC 5-Aza-2'-Deoxycytidine ACTβ Actin β AR Androgen Receptor AUC Area Under the Curve CpG Cytosine-phosphate-Guanine DNA Deoxyribonucleic Acid DNMTs DNA methyltransferases GS Gleason’s Score HD Healthy Donors HDAC Histone Deacetylases HGPIN High Grade Prostatic Intraepithelial Neoplasia miRISC MiRNA-containing RNA-Induced Silencing Complex MiRNA MicroRNA mRNA Messenger RNA MSP Methylation Specific Polymerase Chain Reaction NPT Normal Prostatic Tissue NPV Negative Predictive Value PCa Prostate Cancer PCR Polymerase Chain Reaction PIA Proliferative Inflammatory Atrophy PPV Positive Predictive Value PSA Prostate Specific Antigen qMSP Real-time Quantitative Methylation-Specific Polymerase Chain Reaction qRT-PCR Quantitative Reverse-Transcriptase Polymerase Chain Reaction RNA Ribonucleic Acid ROC Receiver Operator Characteristics VIII TGF-β Transforming Growth Factor β UTR Untranslated Region IX SUMMARY BACKGROUND: Prostate cancer (PCa) is one of the most prevalent cancers worldwide, constituting a serious health problem. Clinically localized disease might be successfully treated whereas disseminated disease remains mostly lethal. PCa is thought to be the end product of the interaction of environmental, physiological and molecular/genetic factors. Over the last decade, the role of epigenetic alterations in prostate carcinogenesis has emerged and provided a new framework for the understanding of the mechanisms underlying the disease as well as for the development of novel PCa biomarkers. Although aberrant DNA methylation and post-transcriptional histone modifications have been the main focus of epigenetic-oriented research in PCa, the role of microRNAs (miRNA) deregulation has more recently surfaced. These are small, single-stranded, non-coding, untranslated RNAs that control gene expression at the post-transcriptional level, interacting directly with messenger RNA (mRNA). MiRNAs are globally downregulated in most cancers and although genetic mechanisms have been appointed as the main cause, a role for epigenetic disruption of miRNAs regulation has been recently emphasized. AIMS: The main aim of this Thesis was to identify new epigenetically downregulated miRNAs in PCa, using an expression profiling based approach, followed by validation in a larger set of clinical samples. In addition, we attempted to identify novel PCa biomarkers suitable for clinical application in early detection, diagnosis and prognosis assessment. METHODOLOGIES: In silico analyses were performed in ten PCa against four morphologically normal prostatic tissues (NPT) based on gene expression profiling data of 740 miRNAs. MiRNAs significantly downregulated in that analysis and re-expressed after treatment with an epigenetic-modulating drug in at least two of three cell lines, were selected for further analysis. Subsequently, candidate miRNAs were surveyed for the presence of a CpG island up to 5000 bp upstream of their mature sequence. Candidate miRNAs fulfilling all these requirements were then validated through DNA methylation analysis in a larger series of tissue samples comprising PCa, NPT and high-grade prostatic intraepithelial neoplasias (HGPIN). Real-time quantitative methylation-specific polymerase chain reaction (qMSP) analysis of 101 PCa, 14 NPT and 56 HGPIN allowed for the determination of promoter methylation levels of the selected miRNAs. Correlation between methylation levels, on the one hand, and expression levels and standard clinicopathologic parameters, on the other hand, was performed. Methylation levels were also used to assess miRNAs performance as PCa biomarkers in tissue samples, and the XVI DNA Extraction ……………………………………………….………………… …..... 31 Sodium Bisulfite Treatment of DNA ................................................................ ..... 31 Methylation Specific Polymerase Chain Reaction ……..….………..…………. ..... 33 Real-time Quantitative MSP……….. ……………………….………….………….... 34 Identification of Prostate Cancer Cellular Pathways Targeted by Epigenetically Deregulated miRNAs ……………..…………….…………….………….…….…...…....... 35 Bioinformatics’ Uncovering of miRNAs Targets …..……….………………...…..… 35 Statistical Analysis …………………………………………………….……..……............… 36 RESULTS …...………………………………………………………..……………..…. 39 Clinical and Pathological Characteristics ……...……………………………………..….… 39 Identification of Epigenetically Regulated miRNAs ……………………………….......….. 40 Analysis of miRNAs Promoter Hypermethylation in Tissues ………..…………………… 44 mRNA Relative Expression Levels in Tissue ………..……………………………...….…. 47 Evaluation of the Biomarkers Diagnostic Potential Using Tissue and Urine Samples ... 50 Performance of miR-130a, miR-205 and miR-145 Methylation as Tumor Markers in Tissue …………………………………………….………………….......... 50 Performance of miR-130a and miR-205 Methylation as Tumor Markers in Urine Samples ……..………………………………….……………......… 52 MiRNAs Potential Targets …..……………………………………………………….…....… 53 DISCUSSION ……………………………………………..………………..…......…...55 CONCLUSIONS AND FUTURE PERSPECTIVES ..……………...….……...…..... 59 BIBLIOGRAPHY ………………………………………............................................ 61 - 1 - INTRODUCTION P Pr ro os st ta at te e C Ca an nc ce er r Epidemiology Nowadays, cancer is one of the most common health problems, with a register of 12.7 million cases and responsible for 7.6 million cancer deaths all around the world in 2008 (Jemal et al., 2011). Worldwide, prostate cancer (PCa) constitutes one of the three most common cancers among male (Siegel et al., 2012), is the second most commonly diagnosed neoplasia and the sixth leading cause of cancer death in males (Jemal et al., 2011), despite all the recent improvements in diagnosis and treatment. Indeed, in spite of men’s long lifetime represent a high risk to develop this disease (about 16–18%), the corresponding risk of death is only about 3% (Fleshner et al., 2012). In fact, international data reveals that PCa accounted for 14% (903,500) of the total new cancer cases and 6% (258,400) of the total cancer deaths in men in 2008 (Jemal et al., 2011). Its incidence rates vary by more than 25-fold worldwide as the develop countries (Oceania, Europe and North America) record highest rates when compared to less develop countries of Africa or the Caribbean region. On the other hand, the highest mortality rates are verified in the less developed countries (Fig. 1 and 2) (Jemal et al., 2011, Center et al., 2012). Indeed, this different global distribution and prognosis may be connected to population’s genetic profiles or even to different diagnosis or detection methodologies (Jemal et al., 2011). Concerning Portugal, the last statistics available revealed PCa as the most incident neoplasia in men, with 5140 cases in 2008, however being the third most lethal cancer (Ferlay et al., 2010) (Fig. 3). Figure 1 – Most commonly diagnosed cancers among men worldwide. Adapted from (Center et al., 2012) - 2 - Figure 2 – Age-standardized PCa incidence and mortality rates by geographic area. Adapted from (Jemal et al., 2011) Figure 3 – Incidence and mortality of different types of cancers in Portugal (number of newly diagnosed cancers cases and proportion for each cancer comparing to all types of cancer in both genders). Adapted from (Ferlay et al., 2010) - 3 - Clinical Disease and Diagnosis Prostate is a male exocrine retroperitoneal organ, encircling the neck of the bladder and urethra, which in the normal adult weighs approximately 20 g and is devoid of a distinct capsule (McNeal, 1981). Prostatic parenchyma can be divided into four biologically and anatomically different regions or zones: the peripheral, central and transitional zones and the region of the anterior fibromuscular stroma (McLaughlin et al., 2005) (Fig. 4). Indeed, the central zone surrounding the ejaculatory ducts is the dominant zone for benign hyperplasia development, while the peripheral zone harbors the majority of prostate carcinomas (75%) (McLaughlin et al., 2005, Shen and Abate-Shen, 2010). This organ’s main function is to produce and secrete an alkaline fluid, named seminal fluid, which forms part of the ejaculate, aiding spermatozoids motility and nourishment (Dunn and Kazer, 2011). Figure 4 – Zonal anatomy of the normal prostate. (A) Young male with minimal transition zone hypertrophy. (B) Older male with transition zone hypertrophy, which effaces the pre-prostatic sphincter and compresses the periejaculatory duct zone. Abbreviations: AFS - Anterior Fibromuscular Stroma; CZ - Central Zone; PZ - Peripheral Zone; SV - Seminal Vesicle; TZ - Transition Zone. Adapted from (McLaughlin et al., 2005) Concerning the clinical features of PCa, the localized disease is often asymptomatic, but occasionally it may have some of the same symptoms as the benign hyperplasia, including weak stream, hesitant, urgent and frequent need to urinate, nocturia, incomplete emptying and various degrees of incontinence (Dunn and Kazer, 2011). The clinical condition may also include hematuria, hematospermia, elevated PSA levels, erectile dysfunction (Dunn and Kazer, 2011) and the diagnosis is confirmed by rectal and physical examination and finally needle biopsy. Advanced clinical disease is characterized by bony pain, especially in the hips and pelvis has a cause of metastasis (Dunn and Kazer, 2011). (SV) - 4 - This malignancy has been recognized as a clinical problem, since ancient Egypt, when it was firstly described, however, effective treatment by surgical procedures (prostatectomy) were only developed in the last century (Capasso, 2005). Concerning PCa diagnosis, the highly accessible blood test for prostate-specific antigen (PSA) constituted the greatest improvement over the past three decades (Shen and Abate-Shen, 2010, Hernandez and Thompson, 2004). This kallikrein-related serine protease is produced in normal prostate secretions, however is released into the blood stream when the normal prostate architecture is disrupted (Lilja et al., 2008). Elevated PSA, which the upper limit that has been considered is 4.0 ng/mL (Hernandez and Thompson, 2004), is usually the primary suspicion criteria for digital rectal examination and undergoing biopsy. In fact, studies like European Randomized Study of Screening for Prostate Cancer where a group of men invited for PCa screening based on PSA was compared to a control group without no active intervention (Schroder et al., 2009), demonstrated that, after a median follow-up of 9 years, men randomized to active surveillance, had a significant reduction in PCa mortality; ratio rate (RR) 0.80 (95% CI 0.65–0.98; adjusted p=0.04) (Schroder et al., 2009). In the same line, other studies revealed the same, like the Göteborg Randomized Population-based Prostate Cancer Screening Trial, in which a group of 20,000 men was divided in half and randomized to a screening program for PSA testing every 2 years against the other half, which was not included in the screening program, serving as a control (Hugosson et al., 2010). In addition, men in the screening group whose PSA concentrations were elevated, were offered additional tests such as digital rectal examination and prostate biopsies (Hugosson et al., 2010). The results of this trial showed that during a median follow up of 14 years, PCa incidence was 12.7% in the screening group and 8.2% in the control group; hazard ratio 1.64 (95% CI 1.50–1.80; p<0.0001) (Hugosson et al., 2010). Also the absolute cumulative risk reduction of death from PCa at 14 years was 0.40% for the control group against the screening group, added to the RR for death from this disease which was 0.56 (95% CI 0.39–0.82; p=0.002) in the screening group compared to the control group (Hugosson et al., 2010). Results from these studies, provide strong evidence that PSA based PCa screening may reduce its mortality. In fact PSA isn’t only associated to diagnosis, as it also constitutes a clinical weapon to assay treatment response, once it can be used to evaluate the response to hormonal therapy and to predict disease recurrence, especially after radical prostatectomy (Lilja et al., 2008, Lange et al., 1989). However, screening based on PSA to diagnosis intends, may have some limitations because it may lead to overdiagnosis and overtreatment, due to its lack of sensitivity and specificity (Henrique and Jeronimo, 2004). Interestingly, some problems have been raised by the scientific community as only 1 in 4 men with PSA levels higher - 5 - than 4.0 ng/mL would be found to have PCa, while the other 3 here unnecessarily biopsied (Hernandez and Thompson, 2004). Moreover, other PSA values and parameters have emerge in order to overcome PSA lack of sensitivity such as PSA velocity, volumerelated PSA, transition zone PSA, PSA density and ratio of free-to-total PSA, however none of those provide a satisfactory sensibility and/or sensitivity (Hernandez and Thompson, 2004, Carter et al., 2007, Bunting, 2002). In addition, adjustments have been proposed, in accordance with age (Gustafsson et al., 1998). Despite all this, PSA is still widely used in the screening of PCa as other clinical biomarkers slow to emerge. In fact, in the last few years, several biomarkers have been suggested in order to promote diagnosis or predict prognosis, such as Human Kallikrein 2, Prostate-specific membrane antigen, presence of fusions genes, between others, however much is still to be known and verified before including them in clinical practice (You et al., 2010). After biochemical evaluation and physical confirmation, needle biopsy is performed, followed by pathological analysis based on histopathological grading of the tissue. This evaluation is performed by Gleason scoring, which classifies tumors from 2 to 10 concerning tissue architecture, with minimal consideration of tumor cell morphology (Gleason and Mellinger, 1974, Yu and Luo, 2007). This scoring allows two grades by tumor sample and after combination of them, as one reflects dominant architectural pattern and the other a minor architectural pattern, it generates a final Gleason’s score. In this way, a tumor sample with combined Gleason’s score between 2 – 4 is considered well differentiated, 5 – 6 moderately and 7 – 10 poorly differentiated. In fact, Gleason scoring may constitute a powerful tool in predicting outcome after radical prostatectomy, radiation and hormonal therapy and also a helpful instrument for choosing the best therapeutic approach (Shah, 2009). Indeed, patients with low Gleason score (6 or lower) are often recommended for active surveillance, in the meanwhile, those with a score of 7 are indicated for therapy of any kind, finally those with a score between 8-10 are candidates for adjuvant therapy or radiation treatment (Shah, 2009). The diagnosis also includes the status of the primary tumor, from organ confined to fully invasive (T1 – 4), with or without lymph node impairment (N0 or 1) and the presence of distant metastasis (M0 and 1) (Ohori et al., 1994). Treatment Concerning treatment, localized disease may be effectively suppressed by surgical excision of total organ – radical prostatectomy – or irradiation through external or internal/implanted beam radiation – brachytherapy, however metastatic disease remains - 6 - incurable and fatal (Shen and Abate-Shen, 2010, Kumar-Sinha and Chinnaiyan, 2003). In advance disease, the treatment regiments are usually based on androgen deprivation therapy which conducts to apoptosis of malignant tumor cells and reduction of tumor burden and/or circulating PSA levels. Nevertheless this is usually temporary, as in most cases tumor cells became resistant to this therapeutic option, and proliferate independently of androgens, which mechanism is still not fully elucidated (Kumar-Sinha and Chinnaiyan, 2003, Shen and Abate-Shen, 2010, Craft et al., 1999). Risk Factors Despite having a high prevalence and mortality, few data is certain about what causes this disease or even the best prevention strategy. In fact, there are several conditions that may compose a risk factor such as age, lifestyle, diet, African American race and familiar history of the disease or even genetic variants (Fowke et al., 2012, Shafique et al., 2012). Concerning age, which is the most significant risk factor, as its occurrence in patients aged below 50 is very low, with around 60% of all cases being registered in men over 70 years old (Macefield et al., 2009). Regarding familiar history, recently a novel HOXB13 G84E variant was associated with a significantly increased risk of hereditary PCa, however, it only accounts for a small fraction of all cases (Ewing et al., 2012). About diet and lifestyle, it seems completely evident that dairy foods constitute one of the most solid predictor for PCa, interestingly at least 7 of 9 cohort studies and 12 of 14 case-control studies observed a positive association between these two variables (Wolk, 2005). Regarding African American men increased odd of developing PCa, it’s still a controversy issue as it’s not well defined whether it correlates with physiological or socioeconomic factors (Major et al., 2012). Still some authors describe an incidence rate of 233.8/100,000 for African Americans against 149.5/100,000 for Caucasians (Major et al., 2012). In fact, PCa prevention or disease prediction is still a delicate matter, which nowadays has its more solid bases on early detection by PSA screening, with all the controversy already referred. On the mean time, important clinical advances have been reported on prevention of this disease, as recently the Food and Drug Administration concluded that finasteride may reduce the risk of low-grade cancer but doesn’t have complete advantages when broad administrated (Theoret et al., 2011). Besides being a controversy issue, studies claim that obesity may constitute a risk factor for developing this disease or at least may influence the grade of PCa, and consequently its aggressiveness and prognosis (Fowke et al., 2012). Other interesting risk factor pointed by several studies is serum cholesterol levels, however it’s also still an inconclusive matter - 7 - that may be connected with grade and aggressiveness and consequently with mortality (Shafique et al., 2012). Also several chemicals and physic agents have been associated with PCa, such as dioxins, cigarette smoking, some farming pesticides, ultra violet radiation and minerals connected to occupational exposure (Mullins and Loeb, 2012), however none of them as achieved significant and solid arguments yet. Molecular Pathways of Carcinogenesis As previously mentioned, PCa is thought to be the end-product of the interplay of environmental, physiological and molecular/genetic factors. Age seems to be the common denominator of all those factors as several associations between gene expression alterations and age progression, including genes related to inflammation, oxidative stress, and cellular senescence, have been pointed (Shen and Abate-Shen, 2010). Concerning inflammation, its estimated that approximately a fifth of all human cancers including those of the stomach, liver and large intestine arise in a background of chronic inflammation (Haverkamp et al., 2008). Concerning PCa, the lack of solid epidemiologic and histological data connecting it with chronic inflammation, makes this correlation still unclear, although chronic prostatitis may be the origin of proliferative inflammatory atrophy (PIA), which is commonly seen in cancerous prostates and may constitute a precursor state, although this is still a controversial issue (De Marzo et al., 1999). Recently, a study has proposed a link between PCa and sexually transmitted infectious agents like Neisseria gonorrhoeae, Chlamydia trachomatis, Trichomonas vaginalis, Treponema pallidum, Human papilloma virus, Herpes simplex virus and Human herpes virus type 8, all of which have been detected in prostatic tissue (Wright et al., 2012). That link was suggested by the observation that circumcision before the first sexual intercourse was associated with a 15% reduction in the relative risk of PCa (Wright et al., 2012). Nevertheless, further studies are needed to clarify those findings. Another link between inflammation and PCa derives from the downregulation of a GSTP1, which encodes for an enzyme involved in the detoxification of reactive species, which are generated by inflammatory cells (Nakayama et al., 2004). Moreover, oxidative stress and consequent DNA damage may be due to hormonal deregulation, diet and/or epigenetic alterations (Shen and Abate-Shen, 2010, Gupta-Elera et al., 2012, Crawford et al., 2012). Indeed, oxidative stress may play a key role in cancer initiation and progression by regulating DNA function enhancers, cell cycle regulators, transcription factors (GuptaElera et al., 2012) or by causing direct DNA damage, which may contribute to telomere shortening (Meeker et al., 2002). - 8 - In respect to genomic alterations, chromosomal rearrangements or copy numbers alterations are also involved in prostate carcinogenesis. The most commonly reported are the gains of 8q and losses of 3p, 8p, 10q, 13q and 17p (Dong, 2001, Lapointe et al., 2004). Loss of chromosome 8p is considered a major genetic alteration in PCa initiation as it occurs in about 80% of all PCa and is already present in high grade prostatic intraepithelial neoplasia (HGPIN) lesions, which are putative PCa precursor lesions (Bergerheim et al., 1991). Several molecular pathways have also been linked with to PCa initiation and progression (Fig. 5). . Figure 5 – Progression pathways for human PCa and its connection to clinical stages. Adapted from (Shen and Abate-Shen, 2010) An important gene that may be lost during this process is NKX3.1 which is thought to play a significant role in prostate carcinogenesis (Abate-Shen and Shen, 2000), as it was found to be downregulated in HGPIN lesions (Bethel et al., 2006) as well as in advanced stage disease (Gurel et al., 2010). The function of NKX3.1 seems to be connected with the regulation of prostate epithelial differentiation and stem cell function (Bhatia-Gaur et al., 1999). MYC upregulation, usually associated with amplification (at 8q) has been recognized more than one decade ago (Jenkins et al., 1997) and may be present in HGPIN, suggesting a relevant contribution to PCa initiation and progression (Koh et al., 2011). The protein encoded by MYC is a transcription factor which is vital in the control of the expression of genes involved in DNA replication, protein synthesis, cell cycle progression, cellular metabolism, chromatin structure, differentiation and stem cell differentiation (Koh et al., 2011). The tumor suppressor gene PTEN is frequently mutated or deleted in PCa (Salmena et al., 2008) and this alteration has been associated with advanced tumor stage, high Gleason grade, presence of lymph node metastasis, hormone refractory - 9 - disease, presence of ERG gene fusion and nuclear p53 accumulation (Krohn et al., 2012). The same study also found an association between PTEN’s deletion and PSA recurrence (Krohn et al., 2012), suggesting that this genetic alteration may constitute a promising biomarker for PCa diagnosis and/or prognosis. Several other genes have been reported to be involved in prostate carcinogenesis, including TP53, ZFHX3 , RB1 and APC (Grasso et al., 2012), implicated in several key pathways. Likewise, Akt/mTOR, mitogen-activated protein kinase (MAPK) or EGFR signaling deregulation , have also been linked to this malignancy (Grasso et al., 2012). A strong enrichment of ETS transcription factor target genes involved in protein synthesis, especially during the transition from benign epithelium to HGPIN lesion has been described (Tomlins et al., 2007). Although this pathway is initially upregulated, it seems to be downregulated during the transition from localized to hormone refractory metastatic PCa (Tomlins et al., 2007), owing to its central role in androgen signaling. Indeed, the same study revealed increased androgen signaling in HGPIN, compared to benign epithelium, but decreased androgen signaling in localized PCa when compared to HGPIN, as well as in high-Gleason grade cancer contrasted with low-grade, achieving the lowest expression levels in hormone refractory disease (Tomlins et al., 2007). The role of androgens is also pivotal in prostate carcinogenesis. Androgens bind to the human androgen receptor (AR), promoting a cascade of ligand-dependent and protein-protein interactions that may be connected with remodeling of chromatin structure at target promoters, recruitment of basal transcription machinery and RNA polimerase activation (Chmelar et al., 2007, Heinlein and Chang, 2004). However, the precise contribution of AR to prostate carcinogenesis and/or disease progression requires further clarification. Interestingly, the common fusion genes derived from ETS family members (e.g., ERG and ETV1) and the strong androgen-regulated TMPRSS2 are involved in prostate carcinogenesis owing to androgen induced expression (Hendriksen et al., 2006). The frequency of TMPRSS2-ERG fusion gene is 15% in HGPIN lesions and 50% in localized PCa, suggesting that this genetic alteration may occur after cancer initiation or at early stages of disease progression (Albadine et al., 2009, Mosquera et al., 2008). It may also function as a prognostic marker as some studies have indicated that it may be associated with clinical stage at diagnosis, although no correlation with clinical recurrence or mortality has been found (Pettersson et al., 2012). Finally, in addition to genetic mechanisms, epigenetic events, including microRNAs (miRNAs) deregulation, have been recognized as critical players in prostate carcinogenesis and their role will be addressed in the following sections (Shen and AbateShen, 2010, van der Poel, 2007). - 16 - Non-conding RNAs Recent evidence indicate that non-coding RNAs may play an important role in controlling multiple genetic and epigenetic phenomena’s with a significant impact in normal cellular differentiation and organism development (Goldberg et al., 2007, Mattick and Gagen, 2001). Interestingly in mammals, noncoding RNAs are closely involved in dosage compensation, such as changes in chromatin structure induced by histone modifications (Bernstein and Allis, 2005). There are two major groups of non-coding RNAs, the small ncRNAs and the long ncRNAs (Hassler and Egger, 2012). Thus, small ncRNAs derive from longer precursors and include transfer RNAs (tRNAs), ribosomal RNAs (rRNAs), microRNAs (miRNAs), piwi interacting RNAs (piRNAs), small nuclear RNAs (snoRNAs) and other less characterized RNAs (Hassler and Egger, 2012). Conversely, long ncRNAs constitute a heterogeneous class of mRNA-like transcripts, yet non-coding, with 200 bp to 100 kb. There is still much to be elucidated about the mechanisms by which these interference RNAs regulate gene expression and its relation with cancer. These questions will be target of further discussion. - 17 - M Mi ic cr ro oR RN NA As s MicroRNAs are small (~22 nucleotides), single-stranded, non-coding untranslated RNAs that control gene expression acting post-transcriptionally by destabilization or translational repression of the messenger RNA (mRNA), inhibiting protein synthesis (Ostling et al., 2011, Choudhry and Catto, 2011). Specifically, the 5’ end of a miRNA (positions 2–8 nt) binds to a targeting sequence, located at the 3’ end of the mRNA3’ UTR regiondepending on the level of complementary between the two sequences (Catto et al., 2011, Betel et al., 2008). Nevertheless, most miRNAs induce a modest reduction (less than two-fold) in their target concentration (Bartel, 2009). In a historical perspective, miRNAs were first described in a work with Caenorhabditis elegans, where two regulatory RNA sequences where reported – lin-4 and let-7 – lately these regulatory sequences where also described in other species, including in humans (Bartel, 2009). Currently over 1,223 human miRNAs mature sequences have been reported in the http://www.mirbase.org base catalog (Mestdagh et al., 2012). The miRNAs are expressed from independent transcription units, because they do not contain an open reading frame and are expressed separately from the nearby genes (Lau et al., 2001). Their expression profile varies between species and in each specie during embryogenesis, suggesting that miRNA might be connected to both gene and protein expression and consequently to the regulation of a variety of pathways (Lau et al., 2001). The Biogenesis of MiRNAs Presently it is accepted that within the nucleus, miRNAs are transcribed by a polymerase II into a long primary transcripts (pri-miRNAs) which contain both 5’-cap structure (7MGpppG), as well as a 3’-end poly(A) tail, with about 70 nucleotides length (Takada and Asahara, 2012, Iorio and Croce, 2012). Then, miRNAs fold back on themselves to form distinctive hairpin-shaped pre-miRNAs by the action of nuclear RNase III Drosha (Kim, 2005), associated to a double stranded RNA-binding protein DGCR8, known as the microprocessor complex (Iorio and Croce, 2012, Carthew and Sontheimer, 2009). Alternatively, but less frequent, miRNA processing might occur through splicing of pri-miRNA transcripts to release introns which are structurally identical to pre-miRNAs (Carthew and Sontheimer, 2009). Following this nuclear processing, the pre-miRNAs are exported to the cytoplasm, where its maturation and action will take place, this transport is made via one of the nuclear Ran-GTP-dependent transport receptors exportin-5 (Kim, 2005, Iorio and Croce, 2012). Here in, a RNAse III enzyme Dicer, processes pre-miRNAs - 18 - into ~22-nucleotide miRNA duplexes (Kim, 2005). Indeed, the PAZ domains of Dicer are crucial to this process, as they interact with the 3’ overhang and determines the cleavage site in a ruler-like fashion (Carthew and Sontheimer, 2009). The maturation process is finalized by the cleavage of a precursor miRNA hairpin, into a transitory miRNA/miRNA* duplex, which includes a mature miRNA with a biological activity and a complementary strand (identified by adding a *) usually subject to degradation (Iorio and Croce, 2012), (Bhayani et al., 2012, Griffiths-Jones, 2004). These mature miRNAs are ready to regulate a variety of pathways, by interfering with the translation process of certain mRNAs. This process requires an incorporation of the miRNA mature sequence into miRNA-containing ribonucleoprotein complex, also called as mirgonaute ou miRISC (miRNA-containing RNA-induced silencing complex) (Kim, 2005), which contains AGO proteins and binds to target mRNA (Iorio and Croce, 2012). The whole miRNAs biogenesis and function process is illustrated on figure 10. Figure 10 – After synthesis, mature miRNA is incorporated into an RNA-induced silencing complex with Argonaut proteins. This complex targets mRNA through the miRNA seed region, inducing either complete mRNA degradation (by perfect annealing, as seen in plants) or alterations in translation (with imperfect miRNA/mRNA annealing, as seen in mammals). Adapted from (Catto et al., 2011) Target binding is made by complementarily, into the 3’ untranslated regions (UTR) of the target transcripted gene (Iorio and Croce, 2012, Ostling et al., 2011). In fact, as referred, this target interaction does not require complete complementarily between the two sequences, however near perfect base-pairing of the 5’ region of the miRNA seems to be determinant in target recognition (Betel et al., 2008). Remarkably, each miRNA might control hundreds of target genes and may modulate up to 60% of all transcripts (Ostling et al., 2011), accounting itself for ~1% of the genome (Kim, 2005). The main target of each miRNA and exactly how its regulation is performed, is still a matter study, however the - 19 - reduction of the target gene expression appears to occur by initiation of translation inhibition or by degradation of the target mRNA (Betel et al., 2008). Concerning this issue, there are several databases that provide miRNA target predictions based on complex mathematic algorithms and several criteria such as: sequence complementarily to target sites or calculations of mRNA secondary structure and energetically favorable binding between sequences (Betel et al., 2008). MiRNAs and Cancer MiRNAs have also been implicated in cancer, ever since a study revealed that the gene cluster containing the miR-15 and miR-16 was deleted in most patients with chronic lymphocytic leukaemia (Calin et al., 2002). These miRNAs were later described as acting as tumor suppressing genes by targeting the oncogene BCL2, then interfering with cell survival and apoptosis (Cimmino et al., 2005). Moreover, subsequent reports revealed that miRNAs expression are altered in many cancers and have been implicated in tumorigenesis (Catto et al., 2011). MiR-21 constitutes a good example of an oncogenic miRNA that is frequently overexpressed in several tumors, such as breast, colorectal, lung, and pancreatic cancer, as well as, in glioblastomas, neuroblastomas, leukemia and lymphomas (Catto et al., 2011, Kong et al., 2012). Indeed, miRNAs expression constitutes an important mechanism of regulation of several cancer-related genes, relevant for apoptosis avoidance, cell proliferation control, epithelial-to-mesenchymal transition and angiogenic signaling. Furthermore, the potential role of miRNAs as tumor biomarkers is being explored in several cancers. Mechanisms of MiRNAs Deregulation in Cancer Interestingly, miRNA may be targeted by several genetic alterations. In fact, nearly 50% of the known miRNAs are located inside or nearby fragile sites and minimal regions of loss of heterozigosity, minimal zones of amplification and common breakpoints which have been already linked to cancer (Kozaki and Inazawa, 2012). Additionally, mutations or polymorphisms on the interference binding site of mRNA coding oncogenes may increasing cancer risk, as described to happen in non-small-cell lung cancer for KRAS (Chin et al., 2008). In addition, several reports verified that most miRNAs have lower expressions in tumors compared to normal tissues, indicating that they may function typically as tumor suppressors (Lu et al., 2005, Agirre et al., 2009, Creighton et al., 2010). - 20 - Thus, miRNAs may be targeted by mutations themselves or amplification or methylation events, becoming over or underexpressed (Fig.11). Figure 11 – miRNAs may regulate tumorgenesis at different levels. Oncogenic miRNAs upregulation may reduce expression of tumor-suppressor proteins, contrarily to the downregulation of tumor-suppressing miRNAs, which may increase oncogenic protein levels. Mutations in tumor-suppressing miRNAs and/or on its mRNA binding sites can cause tumorgenesis, on the other hand mutations in oncogenic miRNAs or targets would reduce tumorgenesis. Adapted from (Kong et al., 2012) Importantly, some studies in recent years correlate miRNAs profile with disease outcome or response to therapy (Calin et al., 2002, Yanaihara et al., 2006). For example, after a median follow-up time of 50 months, miR-96 downregulation was associated with cancer recurrence after surgery (Schaefer et al., 2010). Indeed, miRNA profiles may become a useful tool in assessing clinical stage, as a study using a metastatic versus a non-metastatic PCa xenograft line, found that 140 miRNAs were differently expressed, including miR-16, miR-34a, miR-145 and miR-205 (Watahiki et al., 2011). Owing to the fact that miRNA may also be detected in body fluids, mostly serum and plasma but also in urine samples, they might also serve as biomarkers for early detection, as proposed for miR-141 and miR-375 (Kuner et al., 2012). Concerning PCa, its first described miRNA profile was reported by Porkka et al. (Porkka et al., 2007), in which the authors have identified 51 miRNAs (37 downregulated and 14 upregulated) that were differentially expressed in PCa when compared with benign prostatic lesions. These results were further confirmed by several studies in which a higher frequency of downregulated miRNAs has been reported in PCa, versus to the - 21 - lower percentages of miRNAs found to be upregulated in the same malignancy (Schaefer et al., 2010, Catto et al., 2011) (Table 3). In fact, some miRNAS have already been widely described as being downregulated, specifically miR-145, which has been implicated on apoptosis by regulating TNFSF10, a pro-apoptotic gene, as its reconstitution induced cellular death (Zaman et al., 2010) and/or regulating FSCN1 gene which is related to cell growth, migration and invasion (McLaughlin et al., 2005). Other miRNA that has received wide attention is miR-205, indeed this miRNA seems to be downregulated in PCa and seems to be connected to apoptosis escape by possibly targeting Bcl-w, promoting pharmacologic treatment resistance (Bhatnagar et al., 2010). Table 3 – A summary of miRNAs with altered expression in PCa, including their targeted mRNAs and pathways. Adapted from (Catto et al., 2011) MiRNA Expression MRNA target Pathway miR-20a Up E2F1-3 Apoptosis miR-21 Up PTEN, AKT, androgen pathway Apoptosis, mTOR pathway, androgen independence miR-24 Up FAF1 Apoptosis miR-32 Up BCL2L11 (Bim) Apoptosis miR-106b Up P21, E2F1 Cell cycle control/apoptosis and proliferation miR-125b Up P53, BBC3 (Puma), BAK1 Apoptosis miR-148a Up CAND1 Cell cycle control miR-221 Up p27 (kip1) Cell cycle control and androgen independence miR-222 Up p27 (kip1) Cell cycle control and androgen independence miR-521 Up Cockayne syndrome protein A DNA repair miR-1 Down Exportin-6, tyrosine kinase 9 Gene expression miR-7 Down ERBB-2 (EGFR, HER2) Signal transduction miR-15a-16 cluster Down CCND1 and WNT3a Cell cycle regulation, apoptosis and proliferation miR-34a Down HuR/Bcl2/SIRT1- >p53/p21/BBC3 Apoptosis and drug resistance miR-34c Down E2F3, bcl2 Apoptosis and proliferation miR-101 Down EZH2 Gene expression miR-107 Down Granulin Proliferation miR-143 Down MYO6, ERK5 Cell migration, proliferation miR-145 Down MYO6, BNIP3L->AIFM1, CCNA2, TNFSF10 Cell migration, apoptosis, cell cycle control miR-146a Down ROCK1 – miR-148a Down MSK1 Proliferation, stress response and drug resistance miR-205 Down IL-24 and IL-32, Cepsilon Cell growth and invasion, EMT miR-331-3P Down ERBB-2, CDCA5, KIF23 Signal transduction, cell cycle control miR-449a Down HDAC-1 Gene expression miR-1296 Down MCM family DNA replication Let-7a Down E2F2 and CCND2 Cell cycle control and proliferation - 22 - Recently, miR-130a, miR-205 and miR-203 have been implicated in androgen receptor and MAPK pathways, however the molecular mechanism causing this downregulation have not, yet, been found (Boll et al., 2012). Similar to protein coding genes, one reliable explanation for miRNAs’ downregulation might be the aberrant methylation of their respective codifying genes. Indeed 13 to 28% of human miRNA genes are located within 3 and 10 kb from a CpG island, respectively (Choudhry and Catto, 2011). In addition, it has been suggested that 81.9% of predicted promoters of intergenic miRNA genes contain at least one CpG island (Wang et al., 2010). On the other hand, concerning intragenetic miRNAs, approximately 13.0% have been reported to be located within 500 bp downstream of a CpG island (Wang et al., 2010). However, due to the fact that the location of the promoter region of miRNAs codifing genes is not fully clarified this is still a controversial issue. Remarkably, tumor suppressor miRNAs’ regulation by hypermethylation, has been suggested since the re-expression of miR-9-1 and miR-127 was achieved after the exposure of human cancer cell lines to 5-AZA-DC (Lehmann et al., 2008) (Saito et al., 2006) (the mechanistic of these studies is elucidated in Figure 12). Figure 12 – Activation of coding or non-coding genes that might function as tumor suppressors using an epigenetic therapy with DNMT and/or HDAC inhibitors. The activation of tumor suppressor miRNAs may cause downregulation of target oncogenes. Adapted from (Saito and Jones, 2006) Concerning PCa, miR-145, miR-205, miR-132, miR-126 and miR-193b have been recently reported to be regulated through methylation (Rauhala et al., 2010, Suh et al., 2011, Saito et al., 2009, Bhatnagar et al., 2010, Formosa et al., 2012). However, most of these studies were only performed in cell lines or in a limited number of primary tumors. Thus, epigenetic regulation of miRNA expression is still a largely unexplored field of research in PCa. - 23 - AIMS OF THE STUDY The key objective of this Master Thesis, performed at the Cancer Epigenetics Group of the Research Center of the Portuguese Oncology Institute – Porto, was to identify new epigenetically downregulated miRNAs in PCa, using an expression profiling based approach. Furthermore, it was our purpose to validate these miRNAs in a larger set of clinical samples, in order to identify a putative tumor biomarker amenable to be used for diagnosis and prognostic assessment of this malignancy. The specific aims of this project were: 1. Identify miRNAs that are downregulated in PCa compared to normal prostatic tissues (NPT); 2. Identify miRNAs that are upregulated in prostate cancer cell lines exposed to demethylating agents compared to untreated cell lines; 3. Validate the miRNAs putatively regulated by methylation by quantitative methylation-specific PCR in a larger series of tumors; 4. Assess the methylation status of the identified miRNAs in prostatic pre-malignant lesions (HGPIN); 5. Evaluate the performance of the newly identified miRNAs as tumor biomarkers in clinical samples. - 24 - - 25 - METHODOLOGIES Clinical Samples Patients and Tissue Sample Collection A total of 105 men with clinically localized PCa, diagnosed and primarily submitted to radical prostatectomy in I.P.O.F.G – Porto, from 2002 and 2006, were included in this study [stage T1c and T2, according to TNM system (Hermanek et al., 1997)]. Of the total 105 PCa tissues available, 10 were randomly selected to perform the global miRNA expression, while 101 were used for individual validation studies. HGPIN lesions were identified in 56 cases and also collected for further analysis. As sample controls, 14 morphologically NPT specimens were collected from the peripheral zone of prostates that did not harbor PCa obtained from cystoprostatectomy specimens of bladder cancer patients. The 4 NPT used were also randomly chosen from the 14 available patients tissues, to assess the global miRNA expression. All specimens were frozen at -80ºC and then cut with a cryostat for microscopic evaluation and selection of potential areas for analysis. Cut sections were trimmed to maximize target cell content (>70%) and then DNA extraction was performed using phenol-chloroform. From each specimen, parallel fragments were collected, formalin treated and paraffin-embedded for histopathological examination. Gleason’s score (Gleason and Mellinger, 1974) and pathological staging (Hermanek et al., 1997) were evaluated by an expert pathologist (Rui Henrique, M.D., PhD). Relevant clinical data was collected from the clinical charts. Urine Sample Collection and Processing Morning voided urine samples (one per patient) were collected from 39 patients with PCa diagnosed and treated at the in I.P.O.F.G – Porto, Portugal. 15 male controls were randomly chosen among healthy donors (HD) with no personal or family history of cancer. Patients and controls were enrolled after informed consent. Urine storage and processing conditions were standardized: each sample was immediately centrifuged at 4000 rpm for 10 minutes; the pelleted urine sediment was then washed twice with phosphate-buffered saline, and stored at -80ºC. These studies were approved by the institutional review board (Comissão de Ética) of Portuguese Oncology Institute - Porto, Portugal. - 32 - Figure 16 – Diagram of bisulfate modification of Methylated and Unmethylated DNA. Adapted from (Esteller, 2009) In detail, the genomic DNA fragments are denatured to single stranded DNA, for a more effective bisulfite action, secondly cytosines form adducts, across the 5-6 double bond with an oxidant reagent such as bisulfate ion, which promotes deamination of the cytosine to give a uracil-bisulfite derivate. Thus, a subsequent alkali treatment will form a uracil by the removal of the sulphonate group (Fig. 17) (Tost, 2009, Clark et al., 1994). This reaction is highly specific and is controlled by pH, bisulfate concentration, temperature and the length of the genetic material (Clark et al., 1994). After this reaction the strands of DNA are no longer complementary, allowing its analyses by PCR methods (Tost, 2009). Figure 17 – Schematic diagram of bisulfite conversion reaction. Adapted from (Clark et al., 1994) Genomic DNAs from tissues and urine were modified by sodium bisulfite, using EZ DNA Methylation-Gold kit (Zymo Research, Orange, CA, USA) and this procedure was performed in accordance to manufacturer’s guidelines. Briefly, in a PCR tube,1 µg of DNA in a total volume of 20 µL in sterile distilled water (B. Braun, Melsungen, Germany) was added to 130 µL of the CT conversion Reagent of the above mentioned commercial kit and incubated in a Veriti® Thermal Cycler (Applied Biosystems, Foster City, CA,USA) for 10 minutes at 98ºC, and 180 minutes at 64ºC, in order to complete the cited chemical reaction. Following, the DNA was recovered in 600 µL using M-Binding buffer placed in a Zymo-Spin IC™ column and centrifuged for 30 seconds at 10,000 rpm, followed by a - 33 - cleanup step with M-Wash Buffer in the column. After eliminating the M-Wash Buffer from the column, 200 μL of M-Desulphonation Buffer were added and the DNA was submitted to desulphonation by this buffer for 20 minutes at room temperature. After a centrifugation to remove M-Desulphonation Buffer from the Column, DNA was washed twice with MWash Buffer. Finally, the column was placed in a new 1.5 mL-tube and DNA was eluted by incubation with 30 μL of sterile bidistilled water (B.Braun, Melsungen, Germany) for 5 minutes at room temperature followed by a centrifugation at 12,000 rpm for 30 seconds. This last step was performed again to obtain a total volume of 60 μL. CpGenome™ Universal Methylated DNA (Millipore, CA, USA) and CpGenome™ Universal Unmethylated DNA (Millipore, CA, USA) were also modified. CpGenome™ Universal Methylated and Unmethylated DNA (Millipore, CA, USA) were eluted in a final volume of 30 μL. Finally, bisulfite modified DNA was stored at -50ºC until further use. Methylation Specific Polymerase Chain Reaction (MSP) MSP constitutes both a sensitive and specific methodology for evaluating promoter hypermethylation of CpG islands (Herman et al., 1996, Tost, 2009). Hence, after bisulfite modification, the amplification is possible using specific primers that can distinguish methylated from unmethylated DNA. To insure specificity they must include at least two CG residues, with at least one of them located near the 3’ region and also include nonCpG cytosines to amplify only modified DNA (Tost, 2009). The primers sequences, chosen from regions containing frequent cytosines, were designed using Methyl Primer Express v 1.0, for miR-130a, while miR-205 (Bhatnagar et al., 2010) and miR-145 (Suh et al., 2011) have been already published elsewhere (Table 5). In the present study, MSP was used to assess primers’ specificity for the methylated sequence, to further be analyzed by real-time quantitative methylation-specific PCR (qMSP). For that, CpGenome™ Universal Mehylated DNA (Millipore, CA, USA), previously modified, was used as a positive control for methylation, while negative controls were modified from CpGenome™ Universal Unmehylated DNA (Millipore, CA, USA). Water blanks were also included in each assay. Thus, bisulfite modified DNA (2 uL) was amplified by PCR, using a primer pair for methylated and unmethylated CpG sequences, in a total volume of 20 uL, each. The amplification mix contained 0.4 µL of 10 mM of dNTPs mix (Fermentas, Ontario, Canada), 1 µL of each pair (forward and reverse) of methylated or nonmethylated primers at 10 µM, 2 µL of 10x DyNAzyme™ II Hot Start Reaction Buffer (Finnzymes, Finland), 0.24 µL of DyNAzyme II Hot Start (2 U/µL) (Finnzymes, Finland) and 14.36 µL of sterile distilled water (B. Braun, Melsungen, Germany), according with manufactures’ instructions. In each assay negative and positive controls were tested simultaneously for - 34 - each pair of methylated primers, as well as, a water-blank to assess possible contaminations. The amplification conditions were performed as indicated by DyNAzyme™ II Hot Start manufacturer’s conditions, at 94°C for 10 minutes, followed by 35 cycles at 94°C for 30 seconds, for each pair of primers an optimal annealing temperature (60ºC for miR-130a, 64ºC for miR-145, and 59ºC for miR-205) was performed for 30 seconds and 72°C for 30 seconds, followed by a final extension for 5 minutes at 72°C. The amplification products were loaded on a 2% agarose gel, stained with ethidium bromide, and visualized under UV illumination. Table 5 – Oligonucleotide primers used for MSP and promoter methylation levels quantification by qMSP. Real-time Quantitative MSP (qMSP) QMSP was performed for the same miRNAs in order to quantify the levels of CpGs promoter methylation, of each identified miRNAs. To date, most of the studies detecting miRNAs hypermethylation have used conventional MSP, a sensitive but not quantitative assay. Thus, using the same primers previously used for MSP fluorescence based, qMSP were performed using Power SYBR® Green PCR Master Mix (Applied Biosystems, USA), as performed by Savva-Bordalo et al. (Savva-Bordalo et al., 2010). Briefly, 2 µL of modified DNA from tissue and urine samples were amplified in a final reaction volume of 20 µL, which consisted of: 10 µL of Master Mix, 1 µL of forward and reverse methylated primers at 10µM and 7 µL of sterile deionized nuclease free water (MP Biomedicals, LLC, OH, USA). Analyses was performed in a 96 well plate in a 7500 Sequence Detection System (Applied Biosystems, USA), using the following amplification conditions: 50°C for 2 minutes, followed by 95°C for 10 minutes, then 45 cycles at 95°C for 15 seconds and specific primer annealing temperature [miR-130a, miR-145 and ACTB (used as reference gene) at 60ºC, and miR-205 at 59ºC) for 1 minute (Savva-Bordalo et al., 2010). After all cycles were completed, a dissociation-curve analysis was performed by the following Target miRNA MSP primers miR-130a Forward Methylated - ATAAATTTTGTCGGGGAGAGC Reverse Methylated - AATACCCCGATCAACGAAAA miR-145 Forward Methylated - GGGTTTTCGGTATTTTTTAGGGTAATTGAAGTTTC Reverse Methylated - TAAAATACCACACGTCGCCG miR-205 Forward Methylated - GAGTTTAAGTTGCGTATGGAAGC Reverse Methylated - AAAACAAATATTTCTTTTATAATCCGAA - 35 - conditions: 95°C for 15 seconds, 60°C for 20 seconds and 95°C for 15 seconds. Samples were run in triplicate and multiple water blanks were used as control for contamination (negative control). To build the standard curve, five dilutions (dilution factor of 5x) of the same stock of bisulfite modified CpGenome™ Universal Methylated DNA (Millipore, CA, USA) were run in each plate. A run was considered valid when the further requisites were achieved: a) Slopes of each calibration curve above -3.60 corresponding to a PCR efficiency near 100%; b) R2 of at least three relevant data points ≥ 0.99; c) no template controls not amplified; d) The positive methylation control had to supply a significant methylation signal; e) Threshold cycle value for each gene ≤ 40; f) No negative template controls were amplified. For each sample, the triplicate with a standard deviation greater than 0.38, was removed. Also for quality control, all amplification curves were visualized and scored without knowledge of the clinical data. Finally, the ratio generated by the former mathematic operation, which constitutes an index of the percentage of input copies of DNA that are completely methylated at the specific primer site was multiplied by 1000 for easier tabulation (methylation levels = target gene/reference gene × 1000) (SavvaBordalo et al., 2010). To classify samples as methylated or unmethylated, an empirical cutoff value was selected based on the highest methylation ratio value obtained for NPT or HD samples, ensuring an absolute specificity of the assay. Identification of Prostate Cancer Cellular Pathways Targeted by Epigenetically Deregulated MiRNAs After the identification and validation of 3 epigenetically regulated miRNAs, we investigated their putative target genes and forecast its implication to prostatic carcinogenesis. Bioinformatics’ Uncovering of MiRNAs Targets The prediction of miRNA-mRNA interactions remains a challenging task due to the interactions complexity and still a limited knowledge of the entire process. There are numerous target prediction algorithms to find and speculate numerous targets that exploit different approaches and methods to predict such interactions. The current available algorithms can be divided in two categories based on the use or non-use of conservation comparison. The algorithms based mostly on conservation criteria are for example miRanda, PicTar, TargetScan and DIANA-microT while PITA and rna22 belong group of - 36 - those who use other parameters such as free energy of binding or secondary structures of 3’UTRs that can promote or prevent miRNA binding (Witkos et al., 2011). There is no consensus regarding the best algorithm, since all have advantages and disadvantages. Therefore, the algorithm that we have used was DIANA-microT, which has been already widely used in several studies (Albertini et al., 2011, Maragkakis et al., 2009, Formosa et al., 2012). This algorithm uses a 38 nt-long frame that is moved along the 3’UTR, and the minimum energy of potential miRNA binding is measured and compared with the energy of 100 per cent complementary sequence bound to the 3’UTR region (Witkos et al., 2011). Additionally, this database searches for sites with canonical central bulge, requiring 7-9 ntlong complementary in 5’ region of target miRNA (Witkos et al., 2011). This database considers mainly conservative alignment for scoring but also nonconservative sites and it provides a probability of existence for each result depending on its pairing and conservation profile (Witkos et al., 2011). Therefore, for the identification of the pathways that are targeted by the identified miRNA, a higher and specific threshold score of 0.9 was used according to previously reported (Vlachos et al., 2012). All the targets predicted by this algorithm will be properly validated in future studies. Statistical Analysis The Wilcoxon Signed Rank non-parametric test was performed for two-groups comparison of gene expression (Khan, 2004) and all miRNAs which showed a significant differential expression (p<0.05) were further considered for analysis. The frequency of methylated cases, as well as the median and interquartile range of miR-130a, miR-145 and miR-205 promoter methylation was determined in PCa tissues and NPT. Also, the frequency of methylated cases as well as the median and interquartile range of miR-130a and miR-205 promoter methylation levels were determined in PCa patients and HD urine samples. To classify each sample as methylated or unmethylated an empirical cutoff value was established based on the higher methylation level observed in NPT tissues or HD urines, respectively. Differences in quantitative levels between NPT, HGPIN and PCa tissues or PCa and HD urines were assessed by the nonparametric Kruskal-Wallis test, followed by Bonferroni-adjusted Mann-Whitney U test for pairwise comparisons, when appropriate. For statistical analysis purposes, PCa samples were divided into three Gleason’s Score (GS) categories (GS < 7, GS = 7, and GS > 7). Clinical stage at diagnosis - 37 - comprised two categories (stage II and III). The relationship between methylation or expression levels and other clinicopathological variables such as serum PSA levels at the time of diagnosis, GS and pathological stage were evaluated using the Kruskal-Wallis or the Mann-Whitney U tests, as appropriate. The Spearman non parametric correlation test was used to correlate methylation levels with age and also to correlate methylation levels with expression levels of each miRNA in the 101 PCa cases selected for both analyses. For the purpose of examining the biomarker potential, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of miR130a, miR-145 and miR-205, in tissue samples, and of miR-130a and miR-205, in urine samples, alone or in association, were calculated. In addition, multivariate logistic regression was used to examine associations between miRNAs with biomarker potential. For multivariate logistic regression the backward stepwise (Wald) selection method was used. Then, a receiver operator characteristics (ROC) curve was performed by plotting the true-positive rate (sensitivity) against the false-positive rate (1-specificity) and the area under the curve (AUC) was also calculated individually and for combination of the selected miRNAs. Two sided p-values were considered significant when inferior to 0.05 for all tests, with Bonferroni’s correction, when appropriate. Statistical analyses were performed using SPSS version 20.0. - 38 - - 39 - RESULTS Clinical and Pathological Characteristics Tissue samples from 105 PCa, 56 HGPIN lesions, and 14 NPT samples were tested. The clinical and pathological characteristics of the patients are illustrated in Table 6. No significant differences were found for the median age of NPT, HGPIN and PCa patients, using the Kruskal-wallis test. The clinical characteristics of the patient who provided urine samples are illustrated in Table 7. However significant differences were found in urine samples, for the median age of HD and PCa patients (p<0.001), using the Mann–Whitney U test. Table 6 – Clinical and pathological characteristics of patients from which tissue samples were obtained. Clinicopatholgical Features PCa HGPIN NPT Patients, n 101 56 14 Median age, years* 64 (49 - 74) 65 (53 - 75) 65 (45 - 80) PSA (ng/mL), median (range)* 7.6 (3.4 - 35.5) n.a. n.a. Pathological Stage, n (%) pT2 56 (55.4) n.a. n.a. pT3 45 (44.6) n.a. n.a. Gleason Score, n (%) < 7 30 (29.7) n.a. n.a. = 7 56 (55.4) n.a. n.a. > 7 15 (14.9) n.a. n.a. Abbreviations: NPTMorphologically Normal Prostate Tissue; HGPINHigh Grade Prostatic Intraepithelial Neoplasia; PCaProstate Cancer and n.a.- Not Applicable Table 7 – Gender and age distribution of patients which provided urine samples for this study. Abbreviations: PCaProstate Cancer; HDHealthy Donors HD PCa Patients, n 15 39 Gender, n (%) Male 15 (100%) 39 (100%) Female 0 (0%) 0 (0%) Median age, yrs (range) 52 (43-64) 65 (52-88) - 40 - Identification of Epigenetically Regulated MiRNAs After comparing the expression values between PCa tissues and NPT, only 173 miRNAs, out of the 740 analyzed, shown to be differently expressed with a p<0.05 by the Signed Rank non-parametric test. After applying the previously mentioned fold variation cut-off, 47 miRNAs were significantly downregulated, whereas 5 miRNAs were found to be upregulated (Fig. 18). From the downregulated miRNAs, the highest fold variation values were displayed by miR-187 (~3.0 fold) and miR-224, miR-31*, miR-548b-3p, miR193b*, miR-205 and miR-221*(~2.0 fold) (Table 8). Conversely, miR-449a and miR-32 were upregulated with a fold variation of 4.0 and 3.0, respectively. Figure 18 – Number of differentially expressed miRNAs in PCa tissues concerning respective fold variations. Expression<-1.0 Expression -1.0; 0 Expression 0; 1.0 Expression >1.0 - 41 - Table 8 – Downregulated miRNAs with a fold variation lower than -1.0. * Wilcoxon Signed Rank non-parametric test Ranking miRNA Mean Fold Variation (Log 2) Chromosome localization P value* 1 miR-187 -3.07 18q122 P=0.002 2 miR-224 -2.72 Xq28 P=0.013 3 miR-31* -2.47 9p213 P=0.027 4 miR-548b-3p -2.43 6q2231 P=0.013 5 miR-193b* -2.18 16p1312 P=0.002 6 miR-205 -2.13 1q322 P=0.004 7 miR-221* -2.02 Xp113 P=0.014 8 miR-27b* -1.91 9q2232 P=0.019 9 miR-204 -1.84 9q2112 P=0.002 10 miR-624* -1.81 14q12 P=0.002 11 miR-628-3p -1.79 15q213 P=0.036 12 miR-502-3p -1.78 Xp1123 P=0.024 13 miR-214 -1.73 1q243 P=0.036 14 miR-221 -1.72 Xp113 P=0.008 15 miR-555 -1.71 1q22 P=0.024 16 miR-139-5p -1.67 11q134 P=0.014 17 miR-100 -1.67 11q241 P=0.024 18 miR-505* -1.64 Xq271 P=0.006 19 miR-338-5p -1.60 17q253 P=0.013 20 miR-125b-2* -1.54 21q211 P=0.008 21 miR-145* -1.54 5q32 P=0.024 22 miR-328 -1.51 16q221 P=0.019 23 miR-1271 -1.49 5q35 P=0.002 24 miR-224* -1.49 Xq28 P=0.024 25 miR-145 -1.48 5q32 P=0.006 26 miR-455-3p -1.42 9q32 P=0.004 27 miR-30a* -1.42 6q13 P=0.024 28 miR-222 -1.40 Xp113 P=0.004 29 miR-193b -1.37 16p1312 P=0.028 30 miR-133b -1.35 6p122 P=0.036 31 miR-300 -1.29 14q3231 P=0.036 32 miR-143* -1.29 5q32 P=0.024 33 miR-1468 -1.29 Xq112 P=0.024 34 miR-296-5p -1.28 20q1332 P=0.036 35 miR-509-3p -1.27 Xq273 P=0.014 36 miR-29b-1* -1.19 7q323 P=0.024 37 miR-320a -1.17 8p213 P=0.014 38 miR-23b* -1.16 9q2232 P=0.004 39 miR-181a-2* -1.12 9q333 P=0.028 40 miR-675b -1.12 11p155 P=0.028 41 miR-193a-5p -1.10 17q112 P=0.014 42 miR-23b -1.09 9q2232 P=0.014 43 miR-138 -1.03 16q13 P=0.046 44 miR-130a -1.03 11q121 P=0.001 45 miR-378 -1.03 5q32 P=0.036 46 miR-1181 -1.03 19p132 P=0.036 47 miR-99b -0.99 19q1341 P=0.014 - 48 - Figure 26 – Distribution of miR-205 transcript expression levels in prostatic tissue. Abbreviations: NPTMorphologically Normal Prostate Tissue; PCaProstate Cancer; AvAverage Concerning miR-145, although the same trend has been observed, no statistical significance was attained (Fig. 25). No correlation was found between transcript expression levels of any of the analyzed miRs and respective gene promoter methylation levels. Moreover, no significant correlations were found between the transcript levels of miR-130a and miR-145 and clinical-pathological variables (age, pre-operative serum PSA, Gleason’s score or pathological stage). However, concerning miR-205, significant correlation was found between transcript levels and both Gleason score and pathological stage. Interestingly, miR-205 expression levels were lower in high Gleason’s scores tumors (Kruskall-Wallis test, p=0.001). Indeed, significant differences were found between GS<7 and GS=7 (p=0.018), between GS=7 and GS>7 (p=0.026) and between GS>7 and GS<7 (p=0.001) by the Mann-Whitney U test (Fig. 27). Furthermore, miR-205 lower expression levels inversely correlated with pathological stage by the Mann-Whitney U test (p=0.006) (Fig. 28). - 49 - Figure 27 – Expression levels of miR-205 in samples of 101 PCa prostatectomies according to Gleason’s Score. Abbreviations: GSGleason’s score; AvAverage Figure 28 – Expression levels of miR-205 in samples of 101 PCa prostatectomies according to clinical stage. Abbreviations: CSclinical stage; AvAverage - 50 - Evaluation of the Biomarkers Diagnostic Potential Using Tissue and Urine Samples Performance of MiR-130a, MiR-205 and MiR-145 Methylation as Tumor Markers in Tissue The diagnostic performance of the three miRNAs was assessed using the cutoff values of methylation levels previously referred and determined for each of these gene promoters (117.54 for miR-130a, 463.15 for miR-145 and 298.39 for miR-205). Validity and information estimates for each miRNA or the best combination of genes are displayed in Table 10. Table 10 – Diagnostic performance of selected miRNAs methylation as a PCa biomarker, alone or in association. Parameter - Value (%) Gene Sensitivity Specificity PPV NPV Accuracy miR-130a 82.18 100 100 43.75 84.35 miR-145 26.73 100 100 15.91 35.65 miR-205 64.36 100 100 28.00 68.70 miR-130a and miR-205 89.11 100 100 56.00 90.43 miR-130a, miR-145 and miR-205 91.09 100 100 60.87 92.17 Abbreviations: PPV - Positive Predictive Value; NPV Negative Predictive Value (NPV) Although, a specificity of 100% was apparent for all tested miRNAs, the sensitivity has ranged from 27% to 82%, for each miRNA alone. The best sensitivity was achieved combining in the same panel the methylation analysis of two miRNAs (miR-130a and miR205). According to the model of logistic regression applied, the inclusion of miR-145 in the panel, did no increment significantly the performance of the two miRNAs-panel. ROC curve analysis allowed for the determination of the AUC (95% CI) for each miRNAs gene: 0.956 (0.9170.996) for miR-130a, 0.828 (0.7100.946) for miR-145 and 0.907 (0.8320.982) for miR-205 (Fig. 29). The ROC curve based on the above mentioned panel of two - 51 - markers (miR-130a and miR-205) resulted in an area under the curve (AUC) of 0.970 (0.941-0.998), at a significance of P<0.0001, by the multivariate logistic regression Wald test (Fig. 30). Figure 29 – Receiver operating characteristic curve in PCa tissue for each individual gene (miR-130a, miR205 and miR-145). Figure 30 - Receiver operating characteristic curve in PCa tissue for the best combination of two genes (miR130a and miR-205). - 52 - Performance of MiR-130a and MiR-205 Methylation as Tumor Markers in Urine Samples We have additionally investigated the methylation status of the promoter region of the miRNAs panel which have shown the best performance in terms of sensitivity, specificity and ROC curve (thus miR-145 was not tested), in distinguishing PCa patients from controls in tissue samples, in urine samples. Nevertheless, due to time constrains, we have only performed a preliminary study in a few number of urine sediments: 39 PCa urine patients and 15 urine samples obtained from HD. Interestingly, the Bonferroni-adjusted Mann-Whitney U test revealed that miR-130a methylation levels were significantly higher in urine samples from PCa than from healthy donors (p=0.015) (Fig. 31), however the same was not demonstrated for miR-205. The possible diagnostic performance of the miR-130a alone was assessed using the cutoff values of methylation levels of 114.05. Validity and information estimates are displayed in Table 11. Figure 31 – Distribution of the methylation levels of miR-130a in urine samples. Abbreviations: HDHealthy Donnors; PCaProstate Cancer Table 11 – Diagnostic performance of miR-130a methylation in urine samples. Parameter Value (%) Sensitivity 25.64 Specificity 100 PPV 100 NPV 34.09 Accuracy 46.30 Abbreviations: PPV - Positive Predictive Value; NPV Negative Predictive Value (NPV) - 53 - MiRNAs Potential Targets In order to investigate possible target genes and respective signaling pathways of the three methylation regulated miRNAs previously identified, a well know database DIANA-microT was surveyed, using the described criteria in the material and methods section. Possible target pathways of miR-130a, miR-145 and miR-205 are listed in Table 12, Table 13 and Table 14, respectively. Globally, several critical pathways involved in tumor progression seem to be targeted by the three miRNAs herein identified as epigenetically deregulated in PCa. Table 12 – Possible pathways targeted by miR-130a. Putative Target Pathways Gene Name -ln(value) TGF-beta signaling pathway LTBP1, ROCK1, ZFYVE9, SMAD5, ACVR1, SKP1A, NOG, INHBB, GDF6, PPP2R1B, ACVR2B, TGFBR2, EP300 13.84 Calcium signaling pathway PDE1C, PLN, PDGFRA, LHCGR, ADCY1, ERBB3, ATP2A2, SLC8A1, ITPR1, ADCY2, SPHK2, GRIN2A, PLCB1, ERBB4 7.01 Wnt signaling pathway WNT2B, NFATC2, ROCK1, SKP1A, DAAM1, WNT1, FBXW11, PPP2R1B, NLK, PLCB1, EP300, FZD6 4.46 Ubiquitin mediated proteolysis UBE2D1, UBE2D2, HERC3, BIRC6, CUL3, UBE2W, SKP1A, FBXW11, UBE4B, ANAPC5, CUL5 4.35 Gap junction PDGFRA, GJA1, ADCY1, ITPR1, PRKG1, ADCY2, SOS1, PLCB1 3.15 Cell cycle E2F3, YWHAB, CDKN1A, GADD45A, E2F2, CDC14A, SKP1A, ANAPC5, EP300 3.07 mTOR signaling pathway TSC1, PRKAA1, ULK2, IGF1 2.94 Phosphatidylinositol signaling system SYNJ1, ITPR1, PTEN,PTENP1, ITPK1, PLCB1 2.28 MAPK signaling pathway NFATC2, PDGFRA, GADD45A, MAP3K4, PPM1B, SOS1, DUSP16, RPS6KA5, NLK, HSPA8, TGFBR2, RASA1, MAP3K12, MAX 2.22 Adherens junction PTPRM, MET, WASL, NLK, TGFBR2, EP300 2.21 Focal adhesion PDGFRA, MET, ROCK1, CAV2, ITGB8, SOS1, PTEN,PTENP1, PAK6, ITGA11, COL2A1, IGF1, ITGA5 2.09 ABC transporters ABCA1, ABCC5, ABCD3, ABCB7 1.86 Ether lipid metabolism ENPP6, LYCAT, PAFAH1B1 1.37 ErbB signaling pathway CDKN1A, ERBB3, SOS1, PAK6, ERBB4, EREG 1.34 Methionine metabolism DNMT1, MAT2B 0.93 - 54 - Table 13 – Possible pathways targeted by miR-145. Table 14 – Possible pathways targeted by miR-205. Possible Pathways Gene Name -ln(value) Adherens junction ACTB, IGF1R, PTPRF, YES1, NLK, TGFBR2, ACTG1, SMAD4, SMAD3 21.3 TGF-beta signaling pathway ZFYVE9, SMAD5, SKP1A, INHBB, TGFBR2, SMAD4, SMAD3 8.67 Wnt signaling pathway FZD7, CTNNBIP1, SKP1A, PPP3CA, NLK, CCND2, SENP2, SMAD4, SMAD3 8.04 Tight junction ACTB, MAGI2, IGSF5, VAPA, YES1, MPP5, ACTG1, EPB41L3 6.7 Focal adhesion ACTB, FN1, IGF1R, FLNB, ITGB8, CCND2, ACTG1, PAK7 3.45 Cell adhesion molecules MPZL1, CD40, CDH2, PTPRF, ITGB8, HLA-DRB5 3.38 p53 signaling pathway CDK6, CCND2, BAX, BBC3 3.02 MAPK signaling pathway EVI1, FLNB, MAP4K2, RASA2, PPP3CA, NLK, TGFBR2, DUSP6, RASA1 2.75 Cell cycle CDK6, SKP1A, CCND2, SMAD4, SMAD3 2.35 Calcium signaling pathway ADRB3, PTGFR, ERBB3, PPP3CA, NOS1, ERBB4 1.87 Possible Pathways Gene Name -ln(value) Tight junction MAGI3, MAGI2, PRKCA, YES1, CLDN11, PTEN,PTENP1, PARD6B, EPB41, MYH1 6.82 Adherens junction PTPRM, YES1, SORBS1, FGFR1, INSR, SMAD4 6.17 Ubiquitin mediated proteolysis UBE2G1, UBE2NL, UBE1, MAP3K1, UBE2N, SIAH1, ANAPC5 3.49 Notch signaling pathway APH1A, NOTCH2, PCAF 2.35 ABC transporters – General ABCC9, ABCD1, ABCB7 2.17 Phosphatidylinositol signaling PRKCA, INPPL1, PTEN,PTENP1, PLCB1 2.03 mTOR signaling pathway VEGFA, RPS6KA3, EIF4E 1.79 Inositol phosphate metabolism INPPL1, PTEN,PTENP1, PLCB1 1.79 Wnt signaling pathway PRKCA, NKD1, SIAH1, PLCB1, NFAT5, SMAD4 1.61 Cell cycle CDKN2B, CHEK2, CDC25B, ANAPC5, SMAD4 1.57 TGF-beta signaling pathway INHBA, CDKN2B, SMAD1, SMAD4 1.23 JAK-STAT signaling pathway TPO 1.23 Cell adhesion molecules PTPRM, NRCAM, CLDN11, NRXN1, PTPRC 1.22 Gap junction PRKCA, NPR2, PLCB1, GNAQ 1.09 - 55 - DISCUSSION The mechanisms involved in PCa initiation and progression are not fully understood at present, demanding the search for yet unidentified molecular alterations which underlie tumor heterogeneity. Moreover, the growing concerns about PCa overtreatment provide an opportunity for the discovery of novel biomarkers which not only are accurately able to detect PCa but are also capable of identifying the aggressive forms of the disease. For more than a decade, our research team has been involved in the identification of epigenetic-based markers for PCa. The initial research efforts were devoted to the characterization of the methylome but the fast evolution of Epigenetics has now made clear that other epigenetic alterations, such as histone onco-modifications and miRNAs deregulation, might play a critical role in prostate carcinogenesis. Because epigenetic mechanisms are closely inter-related, we aimed to identify miRNA genes deregulated by promoter methylation in PCa, in an attempt to further illuminate the biological mechanisms underlying PCa and, eventually, provide new PCa biomarkers or therapeutic targets. Thus, comprehensive in silico analyses were performed to identify miRNAs downregulated in PCa and simultaneously re-expressed in response to 5-AZADC exposure in PCa cell lines. Subsequently, the candidate miRNA genes were validated in a large number of clinical samples through a qMSP assay and their putative role as cancer biomarkers was assessed. In brief, we showed, for the first time, that miR-130a downregulation in PCa is due to an epigenetic mechanism, namely aberrant DNA promoter methylation, and that this is an early event in prostate carcinogenesis. We have also found associations between miR-205 expression and clinical-pathological variables. Concerning miR-145, we found that, contrarily to previous reports, it is not significantly downregulated in PCa, although it may still be regulated by epigenetic mechanisms. Finally, we demonstrated that selected miRNAs promoter methylation may provide useful biomarkers for accurate identification of PCa in tissue and urine samples. Expression profiling analysis identified several miRNAs differentially expressed in PCa samples compared to NPT and most were downregulated (47 vs. 5), a finding which is in line with previous reports (Porkka et al., 2007). These results are also in accordance with the more comprehensive observation that miRNAs are globally downregulated in most human cancers, probably reflecting the lower differentiation of malignant cells (Lu et al., 2005). The validity of this first approach is provided by the fact that several miRNAs which we found to be downregulated in PCa, including miR-130a, miR-145, miR-221, miR-100, miR-99b miR-224 and miR-205, have been already reported by other - 56 - researchers (Porkka et al., 2007, Szczyrba et al., 2010, Boll et al., 2012, Sun et al., 2011). Genomic deletions have been generally considered the cause of miRNA dowregulation (Calin et al., 2004), and, indeed, we found that some miRNAs (miR-548b-3p and miR30a* at 6q16-22, and miR-328 at 16q) located at frequently deleted regions in PCa (Lu and Hano, 2008, Carter et al., 1990) were downregulated in our series. However, epigenetic alterations have recently emerged as an alternative mechanism (Bandres et al., 2009), and these were the main focus of this study. We hypothesized that among epigenetic alterations involved in miRNA deregulation, aberrant promoter methylation would be an obvious mechanism, in similarity with protein coding genes. Among the relatively large number of candidate miRNA genes fulfilling strict selection criteria (re-expressed in at least two cell lines, dowregulated in PCa, expressed in normal prostate tissues, and having at the least one CpG island 5000 bp upstream its mature sequence), only three miRNA genes - miR-130a, miR-145 and miR-205 – emerged as candidates for deregulation by promoter methylation. Thus, it is likely that the re-expression of a proportion of miRNAs in cell lines may be due to cell death induced by 5-AZA-DC, which has cytotoxic properties, as previously suggested (Christman, 2002). Indeed, miR-520g and miR-497, which we found to be re-expressed in this study, have previously been found to be upregulated in response to treatment with 5AZA-DC in bladder cancer cell lines (Yoshitomi et al., 2011). Recent studies have also found a few miRNAs regulated by methylation in PCa in cell lines treated with 3 µM of 5AZA-DC and later compared to clinical samples (Formosa et al., 2012). The results of this study are not fully in line with ours but it must be emphasized that different technologies were used for miRNA profiling, as well as higher drug concentrations, a 10 kb upstream limit for CpG searching and lower cut-off values considered for downregulation (Formosa et al., 2012). These data not only demonstrate that miRNA deregulation in PCa is probably a relatively uncommon phenomenon but also that different methodologies are likely to yield quite dissimilar results. Whereas miR-145 and miR-205 have already been reported as deregulated by promoter methylation in PCa, either in cell lines or clinical samples (Hulf et al., 2012, Bhatnagar et al., 2010, Hulf et al., 2011), as well as in other tumor models (Wiklund et al., 2011, Tellez et al., 2011), the main novelty of this study is the identification of miR-130a as an epigenetically-regulated miRNA. This finding has been indirectly corroborated by the recent demonstration that miR-130a was downregulated in PCa by other researchers, although the underlying mechanism was not identified (Boll et al., 2012). Concerning miR145, although it was previously found to be downregulated in PCa (Suh et al., 2011), we verified that the reduction of expression levels in primary tissues was less dramatic than expected. Although methylation is implicated in miR-145 dowregulation, it is likely that - 57 - other mechanisms are also involved, including monoallelic methylation and histone deacetylation. Indeed, PCa cell lines treated with demethylation agent plus a potent histone deacetylase inhibitor, display higher re-expression levels of miR-145 than cell lines treated with demethylation agent only (Zaman et al., 2010, Ke et al., 2009). The same mechanism might explain the lack of correlation between methylation and expression levels found not only for miR-145, but also for miR-130a and miR-205. Remarkably, miR-205 seems also to be regulated by histone acetylation at lysine 9 of histone 3 in PCa cell lines (Hulf et al., 2011). This might explain the inverse correlation between miR-205 expression levels with Gleason score and clinical stage and the lack of it with promoter methylation levels, in accordance with previous studies which found lower miR-205 expression in advanced PCa (Schaefer et al., 2010, Boll et al., 2012). This study is the first to demonstrate that promoter methylation of miR-130a, miR205 and miR-145 genes precedes the development of invasive PCa as they occur in HGPIN lesions, which are generally considered PCa precursors. Interestingly, methylation levels in HGPIN are intermediate between those of NPT and PCa, suggesting that this epigenetic alteration initially affects only a small subset of morphologically normal epithelial cells, which might benefit from a growth / survival advantage. This may foster the neoplastic transformation into HGPIN cells and further progression to an invasive phenotype, as previously suggested for other genes (Henrique et al., 2006), with which these alterations might act in concert. This trend is more obvious for miR-130a and miR205, although the higher miR-145 gene promoter methylation levels found in HGPIN vs. PCa might be due to epigenetic heterogeneity as suggested for other epigenetic regulated genes in prostate carcinogenesis (Henrique et al., 2006). A major aim of this study was to assess the biomarker capabilities of epigenetically deregulated miRNAs for PCa detection. We found that a panel comprising miR-130a and miR-205 promoter methylation is able to accurately discriminate PCa from NPT in tissue samples and may, thus, constitute an interesting ancillary tool for histopathological assessment of diagnostically challenging prostate lesions. It may also be used in patients with clinical suspicion of cancer, abnormal digital rectal examination and persistent elevation of PSA, who are to be submitted to a second biopsy. The inclusion of miR-145 in the gene panel would only increment sensitivity in 2%, and, thus, we considered that no relevant gain was obtained from extending the molecular analysis, especially in minute tissue samples. Remarkably, the analysis of miR-130a and miR-205 promoter methylation in tissue samples compares well with quantitative promoter methylation analysis of protein coding genes for PCa discrimination such as GSTP1 and APC (Henrique et al., 2006). However, the performance of miR-130a promoter methylation for PCa detection in urine has been disappointing owing to the low sensitivity and NPV (25.64% and 34.09%, - 64 - FOWKE, J. H., MOTLEY, S. S., CONCEPCION, R. S., PENSON, D. F. & BAROCAS, D. A. 2012. Obesity, body composition, and prostate cancer. BMC Cancer, 12, 23. FUSE, M., NOHATA, N., KOJIMA, S., SAKAMOTO, S., CHIYOMARU, T., KAWAKAMI, K., ENOKIDA, H., NAKAGAWA, M., NAYA, Y., ICHIKAWA, T. & SEKI, N. 2011. Restoration of miR-145 expression suppresses cell proliferation, migration and invasion in prostate cancer by targeting FSCN1. Int J Oncol, 38, 1093-101. GANDELLINI, P., FOLINI, M., LONGONI, N., PENNATI, M., BINDA, M., COLECCHIA, M., SALVIONI, R., SUPINO, R., MORETTI, R., LIMONTA, P., VALDAGNI, R., DAIDONE, M. G. & ZAFFARONI, N. 2009. miR-205 Exerts tumor-suppressive functions in human prostate through downregulation of protein kinase Cepsilon. Cancer Res, 69, 2287-95. GLEASON, D. F. & MELLINGER, G. T. 1974. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol, 111, 58-64. GOLDBERG, A. D., ALLIS, C. D. & BERNSTEIN, E. 2007. Epigenetics: a landscape takes shape. Cell, 128, 635-8. GRASSO, C. S., WU, Y. M., ROBINSON, D. R., CAO, X., DHANASEKARAN, S. M., KHAN, A. P., QUIST, M. J., JING, X., LONIGRO, R. J., BRENNER, J. C., ASANGANI, I. A., ATEEQ, B., CHUN, S. Y., SIDDIQUI, J., SAM, L., ANSTETT, M., MEHRA, R., PRENSNER, J. R., PALANISAMY, N., RYSLIK, G. A., VANDIN, F., RAPHAEL, B. J., KUNJU, L. P., RHODES, D. R., PIENTA, K. J., CHINNAIYAN, A. M. & TOMLINS, S. A. 2012. The mutational landscape of lethal castration-resistant prostate cancer. Nature, 487, 239-43. GRIFFITHS-JONES, S. 2004. The microRNA Registry. Nucleic Acids Res, 32, D109-11. GUPTA-ELERA, G., GARRETT, A. R., ROBISON, R. A. & O'NEILL, K. L. 2012. The role of oxidative stress in prostate cancer. Eur J Cancer Prev, 21, 155-62. GUREL, B., ALI, T. Z., MONTGOMERY, E. A., BEGUM, S., HICKS, J., GOGGINS, M., EBERHART, C. G., CLARK, D. P., BIEBERICH, C. J., EPSTEIN, J. I. & DE MARZO, A. M. 2010. NKX3.1 as a marker of prostatic origin in metastatic tumors. Am J Surg Pathol, 34, 1097-105. GUSTAFSSON, O., MANSOUR, E., NORMING, U., CARLSSON, A., TORNBLOM, M. & NYMAN, C. R. 1998. Prostate-specific antigen (PSA), PSA density and age-adjusted PSA reference values in screening for prostate cancer--a study of a randomly selected population of 2,400 men. Scand J Urol Nephrol, 32, 373-7. HASSLER, M. R. & EGGER, G. 2012. Epigenomics of cancer - emerging new concepts. Biochimie. HAVERKAMP, J., CHARBONNEAU, B. & RATLIFF, T. L. 2008. Prostate inflammation and its potential impact on prostate cancer: a current review. J Cell Biochem, 103, 1344-53. HEARD, E., CHAUMEIL, J., MASUI, O. & OKAMOTO, I. 2004. Mammalian X-chromosome inactivation: an epigenetics paradigm. Cold Spring Harb Symp Quant Biol, 69, 89-102. HEINLEIN, C. A. & CHANG, C. 2004. Androgen receptor in prostate cancer. Endocr Rev, 25, 276308. HENDRIKSEN, P. J., DITS, N. F., KOKAME, K., VELDHOVEN, A., VAN WEERDEN, W. M., BANGMA, C. H., TRAPMAN, J. & JENSTER, G. 2006. Evolution of the androgen receptor pathway during progression of prostate cancer. Cancer Res, 66, 5012-20. HENRIQUE, R. & JERONIMO, C. 2004. Molecular detection of prostate cancer: a role for GSTP1 hypermethylation. Eur Urol, 46, 660-9; discussion 669. HENRIQUE, R., JERONIMO, C., TEIXEIRA, M. R., HOQUE, M. O., CARVALHO, A. L., PAIS, I., RIBEIRO, F. R., OLIVEIRA, J., LOPES, C. & SIDRANSKY, D. 2006. Epigenetic heterogeneity of highgrade prostatic intraepithelial neoplasia: clues for clonal progression in prostate carcinogenesis. Mol Cancer Res, 4, 1-8. HERMAN, J. G., GRAFF, J. R., MYOHANEN, S., NELKIN, B. D. & BAYLIN, S. B. 1996. Methylationspecific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci U S A, 93, 9821-6. HERMANEK, P., WITTEKIND, C. & SEIB, H. J. 1997. [Current TNM aspects. 5. TNM classification and stage grouping]. Zentralbl Chir, 122, 934-7. - 65 - HERNANDEZ, J. & THOMPSON, I. M. 2004. Prostate-specific antigen: a review of the validation of the most commonly used cancer biomarker. Cancer, 101, 894-904. HUGOSSON, J., CARLSSON, S., AUS, G., BERGDAHL, S., KHATAMI, A., LODDING, P., PIHL, C. G., STRANNE, J., HOLMBERG, E. & LILJA, H. 2010. Mortality results from the Goteborg randomised population-based prostate-cancer screening trial. Lancet Oncol, 11, 725-32. HULF, T., SIBBRITT, T., WIKLUND, E. D., BERT, S., STRBENAC, D., STATHAM, A. L., ROBINSON, M. D. & CLARK, S. J. 2011. Discovery pipeline for epigenetically deregulated miRNAs in cancer: integration of primary miRNA transcription. BMC Genomics, 12, 54. HULF, T., SIBBRITT, T., WIKLUND, E. D., PATTERSON, K., SONG, J. Z., STIRZAKER, C., QU, W., NAIR, S., HORVATH, L. G., ARMSTRONG, N. J., KENCH, J. G., SUTHERLAND, R. L. & CLARK, S. J. 2012. Epigenetic-induced repression of microRNA-205 is associated with MED1 activation and a poorer prognosis in localized prostate cancer. Oncogene. IORIO, M. V. & CROCE, C. M. 2012. microRNA involvement in human cancer. Carcinogenesis. JEMAL, A., BRAY, F., CENTER, M. M., FERLAY, J., WARD, E. & FORMAN, D. 2011. Global cancer statistics. CA Cancer J Clin, 61, 69-90. JENKINS, R. B., QIAN, J., LIEBER, M. M. & BOSTWICK, D. G. 1997. Detection of c-myc oncogene amplification and chromosomal anomalies in metastatic prostatic carcinoma by fluorescence in situ hybridization. Cancer Res, 57, 524-31. JERONIMO, C. & HENRIQUE, R. 2011. Epigenetic biomarkers in urological tumors: A systematic review. Cancer Lett. JORDE, R., SVARTBERG, J., JOAKIMSEN, R. M. & COUCHERON, D. H. 2012. Plasma profile of microRNA after supplementation with high doses of vitamin D3 for 12 months. BMC Res Notes, 5, 245. KATADA, S., IMHOF, A. & SASSONE-CORSI, P. 2012. Connecting threads: epigenetics and metabolism. Cell, 148, 24-8. KE, X. S., QU, Y., ROSTAD, K., LI, W. C., LIN, B., HALVORSEN, O. J., HAUKAAS, S. A., JONASSEN, I., PETERSEN, K., GOLDFINGER, N., ROTTER, V., AKSLEN, L. A., OYAN, A. M. & KALLAND, K. H. 2009. Genome-wide profiling of histone h3 lysine 4 and lysine 27 trimethylation reveals an epigenetic signature in prostate carcinogenesis. PLoS One, 4, e4687. KHAN, H. A. 2004. ArraySolver: an algorithm for colour-coded graphical display and Wilcoxon signed-rank statistics for comparing microarray gene expression data. Comp Funct Genomics, 5, 39-47. KIM, V. N. 2005. MicroRNA biogenesis: coordinated cropping and dicing. Nat Rev Mol Cell Biol, 6, 376-85. KOH, C. M., IWATA, T., ZHENG, Q., BETHEL, C., YEGNASUBRAMANIAN, S. & DE MARZO, A. M. 2011. Myc enforces overexpression of EZH2 in early prostatic neoplasia via transcriptional and post-transcriptional mechanisms. Oncotarget, 2, 669-83. KONG, Y. W., FERLAND-MCCOLLOUGH, D., JACKSON, T. J. & BUSHELL, M. 2012. microRNAs in cancer management. Lancet Oncol, 13, e249-58. KOO, S., MARTIN, G. S., SCHULZ, K. J., RONCK, M. & TOUSSAINT, L. G. 2012. Serial selection for invasiveness increases expression of miR-143/miR-145 in glioblastoma cell lines. BMC Cancer, 12, 143. KOVALEVA, V., MORA, R., PARK, Y. J., PLASS, C., CHIRAMEL, A. I., BARTENSCHLAGER, R., DOHNER, H., STILGENBAUER, S., PSCHERER, A., LICHTER, P. & SEIFFERT, M. 2012. miRNA-130a targets ATG2B and DICER1 to inhibit autophagy and trigger killing of chronic lymphocytic leukemia cells. Cancer Res, 72, 1763-72. KOZAKI, K. & INAZAWA, J. 2012. Tumor-suppressive microRNA silenced by tumor-specific DNA hypermethylation in cancer cells. Cancer Sci, 103, 837-45. KROHN, A., DIEDLER, T., BURKHARDT, L., MAYER, P. S., DE SILVA, C., MEYER-KORNBLUM, M., KOTSCHAU, D., TENNSTEDT, P., HUANG, J., GERHAUSER, C., MADER, M., KURTZ, S., SIRMA, H., SAAD, F., STEUBER, T., GRAEFEN, M., PLASS, C., SAUTER, G., SIMON, R., MINNER, S. & SCHLOMM, T. 2012. Genomic Deletion of PTEN Is Associated with Tumor Progression and - 66 - Early PSA Recurrence in ERG Fusion-Positive and Fusion-Negative Prostate Cancer. Am J Pathol. KUMAR-SINHA, C. & CHINNAIYAN, A. M. 2003. Molecular markers to identify patients at risk for recurrence after primary treatment for prostate cancer. Urology, 62 Suppl 1, 19-35. KUNER, R., BRASE, J. C., SULTMANN, H. & WUTTIG, D. 2012. microRNA biomarkers in body fluids of prostate cancer patients. Methods. KWON, M. J. & SHIN, Y. K. 2011. Epigenetic regulation of cancer-associated genes in ovarian cancer. Int J Mol Sci, 12, 983-1008. LANGE, P. H., ERCOLE, C. J., LIGHTNER, D. J., FRALEY, E. E. & VESSELLA, R. 1989. The value of serum prostate specific antigen determinations before and after radical prostatectomy. J Urol, 141, 873-9. LAPOINTE, 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. 2004. Gene expression profiling identifies clinically relevant subtypes of prostate cancer. Proc Natl Acad Sci U S A, 101, 811-6. LAU, N. C., LIM, L. P., WEINSTEIN, E. G. & BARTEL, D. P. 2001. An abundant class of tiny RNAs with probable regulatory roles in Caenorhabditis elegans. Science, 294, 858-62. LEHMANN, U., HASEMEIER, B., CHRISTGEN, M., MULLER, M., ROMERMANN, D., LANGER, F. & KREIPE, H. 2008. Epigenetic inactivation of microRNA gene hsa-mir-9-1 in human breast cancer. J Pathol, 214, 17-24. LI, Y. & TOLLEFSBOL, T. O. 2010. Impact on DNA methylation in cancer prevention and therapy by bioactive dietary components. Curr Med Chem, 17, 2141-51. LILJA, H., ULMERT, D. & VICKERS, A. J. 2008. Prostate-specific antigen and prostate cancer: prediction, detection and monitoring. Nat Rev Cancer, 8, 268-78. LOIZOU, J. I., MURR, R., FINKBEINER, M. G., SAWAN, C., WANG, Z. Q. & HERCEG, Z. 2006. Epigenetic information in chromatin: the code of entry for DNA repair. Cell Cycle, 5, 696701. LOPEZ-SERRA, L. & ESTELLER, M. 2008. Proteins that bind methylated DNA and human cancer: reading the wrong words. Br J Cancer, 98, 1881-5. LU, J., GETZ, G., MISKA, E. A., ALVAREZ-SAAVEDRA, E., LAMB, J., PECK, D., SWEET-CORDERO, A., EBERT, B. L., MAK, R. H., FERRANDO, A. A., DOWNING, J. R., JACKS, T., HORVITZ, H. R. & GOLUB, T. R. 2005. MicroRNA expression profiles classify human cancers. Nature, 435, 834-8. LU, S., LEE, J., REVELO, M., WANG, X. & DONG, Z. 2007. Smad3 is overexpressed in advanced human prostate cancer and necessary for progressive growth of prostate cancer cells in nude mice. Clin Cancer Res, 13, 5692-702. LU, T. & HANO, H. 2008. Deletion at chromosome arms 6q16-22 and 10q22.3-23.1 associated with initiation of prostate cancer. Prostate Cancer Prostatic Dis, 11, 357-61. MACEFIELD, R. C., LANE, J. A., METCALFE, C., DOWN, L., NEAL, D. E., HAMDY, F. C. & DONOVAN, J. L. 2009. Do the risk factors of age, family history of prostate cancer or a higher prostate specific antigen level raise anxiety at prostate biopsy? Eur J Cancer, 45, 2569-73. MAJOR, J. M., NORMAN OLIVER, M., DOUBENI, C. A., HOLLENBECK, A. R., GRAUBARD, B. I. & SINHA, R. 2012. Socioeconomic status, healthcare density, and risk of prostate cancer among African American and Caucasian men in a large prospective study. Cancer Causes Control, 23, 1185-91. MARAGKAKIS, M., ALEXIOU, P., PAPADOPOULOS, G. L., RECZKO, M., DALAMAGAS, T., GIANNOPOULOS, G., GOUMAS, G., KOUKIS, E., KOURTIS, K., SIMOSSIS, V. A., SETHUPATHY, P., VERGOULIS, T., KOZIRIS, N., SELLIS, T., TSANAKAS, P. & HATZIGEORGIOU, A. G. 2009. Accurate microRNA target prediction correlates with protein repression levels. BMC Bioinformatics, 10, 295. - 67 - MASCARENHAS, J., ROPER, N., CHAURASIA, P. & HOFFMAN, R. 2011. Epigenetic abnormalities in myeloproliferative neoplasms: a target for novel therapeutic strategies. Clin Epigenetics, 2, 197-212. MATTICK, J. S. & GAGEN, M. J. 2001. The evolution of controlled multitasked gene networks: the role of introns and other noncoding RNAs in the development of complex organisms. Mol Biol Evol, 18, 1611-30. MCLAUGHLIN, P. W., TROYER, S., BERRI, S., NARAYANA, V., MEIROWITZ, A., ROBERSON, P. L. & MONTIE, J. 2005. Functional anatomy of the prostate: implications for treatment planning. Int J Radiat Oncol Biol Phys, 63, 479-91. MCNEAL, J. E. 1981. Normal and pathologic anatomy of prostate. Urology, 17, 11-6. MEEKER, A. K., HICKS, J. L., PLATZ, E. A., MARCH, G. E., BENNETT, C. J., DELANNOY, M. J. & DE MARZO, A. M. 2002. Telomere shortening is an early somatic DNA alteration in human prostate tumorigenesis. Cancer Res, 62, 6405-9. MESTDAGH, P., DERVEAUX, S. & VANDESOMPELE, J. 2012. Whole-genome RT-qPCR microRNA expression profiling. Methods Mol Biol, 815, 121-30. MESTDAGH, P., VAN VLIERBERGHE, P., DE WEER, A., MUTH, D., WESTERMANN, F., SPELEMAN, F. & VANDESOMPELE, J. 2009. A novel and universal method for microRNA RT-qPCR data normalization. Genome Biol, 10, R64. miRCURY LNA™ Universal RT microRNA PCR, Instruction manual Version 4.1 #203300 (August 2011) MOSQUERA, J. M., PERNER, S., GENEGA, E. M., SANDA, M., HOFER, M. D., MERTZ, K. D., PARIS, P. L., SIMKO, J., BISMAR, T. A., AYALA, G., SHAH, R. B., LODA, M. & RUBIN, M. A. 2008. Characterization of TMPRSS2-ERG fusion high-grade prostatic intraepithelial neoplasia and potential clinical implications. Clin Cancer Res, 14, 3380-5. MULLINS, J. K. & LOEB, S. 2012. Environmental exposures and prostate cancer. Urol Oncol, 30, 216-9. NAKAYAMA, M., GONZALGO, M. L., YEGNASUBRAMANIAN, S., LIN, X., DE MARZO, A. M. & NELSON, W. G. 2004. GSTP1 CpG island hypermethylation as a molecular biomarker for prostate cancer. J Cell Biochem, 91, 540-52. OHORI, M., WHEELER, T. M. & SCARDINO, P. T. 1994. The New American Joint Committee on Cancer and International Union Against Cancer TNM classification of prostate cancer. Clinicopathologic correlations. Cancer, 74, 104-14. OSTLING, P., LEIVONEN, S. K., AAKULA, A., KOHONEN, P., MAKELA, R., HAGMAN, Z., EDSJO, A., KANGASPESKA, S., EDGREN, H., NICORICI, D., BJARTELL, A., CEDER, Y., PERALA, M. & KALLIONIEMI, O. 2011. Systematic analysis of microRNAs targeting the androgen receptor in prostate cancer cells. Cancer Res, 71, 1956-67. PEARSON, H. & STIRLING, D. 2003. DNA extraction from tissue. Methods Mol Biol, 226, 33-4. PETTERSSON, A., GRAFF, R. E., BAUER, S. R., PITT, M. J., LIS, R. T., STACK, E. C., MARTIN, N. E., KUNZ, L., PENNEY, K. L., LIGON, A., SUPPAN, C., FLAVIN, R., SESSO, H., RIDER, J. R., SWEENEY, C. S., STAMPFER, M. J., FIORENTINO, M., KANTOFF, P. W., SANDA, M. G., GIOVANNUCCI, E., DING, E. L., LODA, M. & MUCCI, L. A. 2012. The TMPRSS2:ERG Rearrangement, ERG Expression, and Prostate Cancer Outcomes: a Cohort Study and Meta-analysis. Cancer Epidemiol Biomarkers Prev. PHE, V., CUSSENOT, O. & ROUPRET, M. 2010. Methylated genes as potential biomarkers in prostate cancer. BJU Int, 105, 1364-70. PORKKA, K. P., PFEIFFER, M. J., WALTERING, K. K., VESSELLA, R. L., TAMMELA, T. L. & VISAKORPI, T. 2007. MicroRNA expression profiling in prostate cancer. Cancer Res, 67, 6130-5. RAUHALA, H. E., JALAVA, S. E., ISOTALO, J., BRACKEN, H., LEHMUSVAARA, S., TAMMELA, T. L., OJA, H. & VISAKORPI, T. 2010. miR-193b is an epigenetically regulated putative tumor suppressor in prostate cancer. Int J Cancer, 127, 1363-72. - 68 - SAITO, Y., FRIEDMAN, J. M., CHIHARA, Y., EGGER, G., CHUANG, J. C. & LIANG, G. 2009. Epigenetic therapy upregulates the tumor suppressor microRNA-126 and its host gene EGFL7 in human cancer cells. Biochem Biophys Res Commun, 379, 726-31. SAITO, Y. & JONES, P. A. 2006. Epigenetic activation of tumor suppressor microRNAs in human cancer cells. Cell Cycle, 5, 2220-2. SAITO, Y., LIANG, G., EGGER, G., FRIEDMAN, J. M., CHUANG, J. C., COETZEE, G. A. & JONES, P. A. 2006. Specific activation of microRNA-127 with downregulation of the proto-oncogene BCL6 by chromatin-modifying drugs in human cancer cells. Cancer Cell, 9, 435-43. SALMENA, L., CARRACEDO, A. & PANDOLFI, P. P. 2008. Tenets of PTEN tumor suppression. Cell, 133, 403-14. SAVVA-BORDALO, J., RAMALHO-CARVALHO, J., PINHEIRO, M., COSTA, V. L., RODRIGUES, A., DIAS, P. C., VEIGA, I., MACHADO, M., TEIXEIRA, M. R., HENRIQUE, R. & JERONIMO, C. 2010. Promoter methylation and large intragenic rearrangements of DPYD are not implicated in severe toxicity to 5-fluorouracil-based chemotherapy in gastrointestinal cancer patients. BMC Cancer, 10, 470. SCHAEFER, A., JUNG, M., MOLLENKOPF, H. J., WAGNER, I., STEPHAN, C., JENTZMIK, F., MILLER, K., LEIN, M., KRISTIANSEN, G. & JUNG, K. 2010. Diagnostic and prognostic implications of microRNA profiling in prostate carcinoma. Int J Cancer, 126, 1166-76. SCHMITTGEN, T. D. & LIVAK, K. J. 2008. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc, 3, 1101-8. SCHOLZ, B. & MARSCHALEK, R. 2012. Epigenetics and blood disorders. Br J Haematol, 158, 307-22. SCHRODER, F. H., HUGOSSON, J., ROOBOL, M. J., TAMMELA, T. L., CIATTO, S., NELEN, V., KWIATKOWSKI, M., LUJAN, M., LILJA, H., ZAPPA, M., DENIS, L. J., RECKER, F., BERENGUER, A., MAATTANEN, L., BANGMA, C. H., AUS, G., VILLERS, A., REBILLARD, X., VAN DER KWAST, T., BLIJENBERG, B. G., MOSS, S. M., DE KONING, H. J. & AUVINEN, A. 2009. Screening and prostate-cancer mortality in a randomized European study. N Engl J Med, 360, 1320-8. SHAFIQUE, K., MCLOONE, P., QURESHI, K., LEUNG, H., HART, C. & MORRISON, D. S. 2012. Cholesterol and the risk of grade-specific prostate cancer incidence: evidence from two large prospective cohort studies with up to 37 years' follow up. BMC Cancer, 12, 25. SHAH, R. B. 2009. Current perspectives on the Gleason grading of prostate cancer. Arch Pathol Lab Med, 133, 1810-6. SHEN, M. M. & ABATE-SHEN, C. 2010. Molecular genetics of prostate cancer: new prospects for old challenges. Genes Dev, 24, 1967-2000. SIEGEL, R., DESANTIS, C., VIRGO, K., STEIN, K., MARIOTTO, A., SMITH, T., COOPER, D., GANSLER, T., LERRO, C., FEDEWA, S., LIN, C., LEACH, C., CANNADY, R. S., CHO, H., SCOPPA, S., HACHEY, M., KIRCH, R., JEMAL, A. & WARD, E. 2012. Cancer treatment and survivorship statistics, 2012. CA Cancer J Clin. STEINER, M. S., ZHOU, Z. Z., TONB, D. C. & BARRACK, E. R. 1994. Expression of transforming growth factor-beta 1 in prostate cancer. Endocrinology, 135, 2240-7. SUH, S. O., CHEN, Y., ZAMAN, M. S., HIRATA, H., YAMAMURA, S., SHAHRYARI, V., LIU, J., TABATABAI, Z. L., KAKAR, S., DENG, G., TANAKA, Y. & DAHIYA, R. 2011. MicroRNA-145 is regulated by DNA methylation and p53 gene mutation in prostate cancer. Carcinogenesis, 32, 772-8. SUN, D., LEE, Y. S., MALHOTRA, A., KIM, H. K., MATECIC, M., EVANS, C., JENSEN, R. V., MOSKALUK, C. A. & DUTTA, A. 2011. miR-99 family of MicroRNAs suppresses the expression of prostate-specific antigen and prostate cancer cell proliferation. Cancer Res, 71, 1313-24. SZCZYRBA, J., LOPRICH, E., WACH, S., JUNG, V., UNTEREGGER, G., BARTH, S., GROBHOLZ, R., WIELAND, W., STOHR, R., HARTMANN, A., WULLICH, B. & GRASSER, F. 2010. The microRNA profile of prostate carcinoma obtained by deep sequencing. Mol Cancer Res, 8, 529-38. TAKADA, S. & ASAHARA, H. 2012. Current strategies for microRNA research. Mod Rheumatol. - 69 - TELLEZ, C. S., JURI, D. E., DO, K., BERNAUER, A. M., THOMAS, C. L., DAMIANI, L. A., TESSEMA, M., LENG, S. & BELINSKY, S. A. 2011. EMT and stem cell-like properties associated with miR205 and miR-200 epigenetic silencing are early manifestations during carcinogen-induced transformation of human lung epithelial cells. Cancer Res, 71, 3087-97. THEORET, M. R., NING, Y. M., ZHANG, J. J., JUSTICE, R., KEEGAN, P. & PAZDUR, R. 2011. The risks and benefits of 5alpha-reductase inhibitors for prostate-cancer prevention. N Engl J Med, 365, 97-9. TOMLINS, S. A., MEHRA, R., RHODES, D. R., CAO, X., WANG, L., DHANASEKARAN, S. M., KALYANASUNDARAM, S., WEI, J. T., RUBIN, M. A., PIENTA, K. J., SHAH, R. B. & CHINNAIYAN, A. M. 2007. Integrative molecular concept modeling of prostate cancer progression. Nat Genet, 39, 41-51. TOST, J. 2009. DNA methylation : methods and protocols, New York, Humana Press. VAN DER POEL, H. G. 2007. Molecular markers in the diagnosis of prostate cancer. Crit Rev Oncol Hematol, 61, 104-39. VLACHOS, I. S., KOSTOULAS, N., VERGOULIS, T., GEORGAKILAS, G., RECZKO, M., MARAGKAKIS, M., PARASKEVOPOULOU, M. D., PRIONIDIS, K., DALAMAGAS, T. & HATZIGEORGIOU, A. G. 2012. DIANA miRPath v.2.0: investigating the combinatorial effect of microRNAs in pathways. Nucleic Acids Res, 40, W498-W504. WANG, G., WANG, Y., SHEN, C., HUANG, Y. W., HUANG, K., HUANG, T. H., NEPHEW, K. P., LI, L. & LIU, Y. 2010. RNA polymerase II binding patterns reveal genomic regions involved in microRNA gene regulation. PLoS One, 5, e13798. WATAHIKI, A., WANG, Y., MORRIS, J., DENNIS, K., O'DWYER, H. M., GLEAVE, M. & GOUT, P. W. 2011. MicroRNAs associated with metastatic prostate cancer. PLoS One, 6, e24950. WIKLUND, E. D., BRAMSEN, J. B., HULF, T., DYRSKJOT, L., RAMANATHAN, R., HANSEN, T. B., VILLADSEN, S. B., GAO, S., OSTENFELD, M. S., BORRE, M., PETER, M. E., ORNTOFT, T. F., KJEMS, J. & CLARK, S. J. 2011. Coordinated epigenetic repression of the miR-200 family and miR-205 in invasive bladder cancer. Int J Cancer, 128, 1327-34. WITKOS, T. M., KOSCIANSKA, E. & KRZYZOSIAK, W. J. 2011. Practical Aspects of microRNA Target Prediction. Curr Mol Med, 11, 93-109. WOLK, A. 2005. Diet, lifestyle and risk of prostate cancer. Acta Oncol, 44, 277-81. WOODSON, K., O'REILLY, K. J., HANSON, J. C., NELSON, D., WALK, E. L. & TANGREA, J. A. 2008. The usefulness of the detection of GSTP1 methylation in urine as a biomarker in the diagnosis of prostate cancer. J Urol, 179, 508-11; discussion 511-2. WRIGHT, J. L., LIN, D. W. & STANFORD, J. L. 2012. Circumcision and the risk of prostate cancer. Cancer. XU, Q., LIU, L. Z., QIAN, X., CHEN, Q., JIANG, Y., LI, D., LAI, L. & JIANG, B. H. 2012. MiR-145 directly targets p70S6K1 in cancer cells to inhibit tumor growth and angiogenesis. Nucleic Acids Res, 40, 761-74. YANAIHARA, N., CAPLEN, N., BOWMAN, E., SEIKE, M., KUMAMOTO, K., YI, M., STEPHENS, R. M., OKAMOTO, A., YOKOTA, J., TANAKA, T., CALIN, G. A., LIU, C. G., CROCE, C. M. & HARRIS, C. C. 2006. Unique microRNA molecular profiles in lung cancer diagnosis and prognosis. Cancer Cell, 9, 189-98. YANG, X., NOUSHMEHR, H., HAN, H., ANDREU-VIEYRA, C., LIANG, G. & JONES, P. A. 2012. Gene reactivation by 5-aza-2'-deoxycytidine-induced demethylation requires SRCAP-mediated H2A.Z insertion to establish nucleosome depleted regions. PLoS Genet, 8, e1002604. YOSHITOMI, T., KAWAKAMI, K., ENOKIDA, H., CHIYOMARU, T., KAGARA, I., TATARANO, S., YOSHINO, H., ARIMURA, H., NISHIYAMA, K., SEKI, N. & NAKAGAWA, M. 2011. Restoration of miR-517a expression induces cell apoptosis in bladder cancer cell lines. Oncol Rep, 25, 1661-8. YOU, J., COZZI, P., WALSH, B., WILLCOX, M., KEARSLEY, J., RUSSELL, P. & LI, Y. 2010. Innovative biomarkers for prostate cancer early diagnosis and progression. Crit Rev Oncol Hematol, 73, 10-22. - 70 - YU, Y. P. & LUO, J. H. 2007. Pathological factors evaluating prostate cancer. Histol Histopathol, 22, 1291-300. ZAMAN, M. S., CHEN, Y., DENG, G., SHAHRYARI, V., SUH, S. O., SAINI, S., MAJID, S., LIU, J., KHATRI, G., TANAKA, Y. & DAHIYA, R. 2010. The functional significance of microRNA-145 in prostate cancer. Br J Cancer, 103, 256-64.