Identification of novel endocrine-metabolic and splicing machinery factors with diagnostic, prognostic and/or therapeutic potential in prostate cancer

Abstract

Prostate cancer (PCa) is the tumor pathology with the highest incidence among men in developed countries and represents one of the main causes of cancer-related death in this group. Despite the advances achieved during the last years in PCa diagnosis and treatment, the diagnostic and therapeutic tools currently available are not sufficiently efficient. Specifically, prostate specific antigen (PSA) is the current “gold standard” used for PCa screening. Unfortunately, the PSA test has controversial specificity, since certain physiological conditions or non-tumor pathologies (e.g. benign prostatic hyperplasia, prostatitis) are also associated with high circulating PSA levels. Furthermore, the diagnostic capacity and clinical utility of the PSA test is especially low in patients with PSA levels between 3-10 ng/mL (the so-called grey zone). For these reasons, PSA test is associated with PCa overdiagnosis and unnecessary biopsies. Moreover, this invasive method required to subsequently confirm or discard the presence of PCa can cause side effects such as pain, bleeding and infections, which significantly impact patient’s quality of live. Therefore, there is an urgent need to identify new diagnostic tools that complement or replace PSA in clinical practice. Similarly, the therapeutic armamentarium in PCa is limited and clearly insufficient. Specifically, the pharmacological treatment used as the first option for the treatment of advanced-stage PCa or tumor recurrences consists of androgen deprivation (e.g. LHRH antagonists, abiraterone), due to the marked dependence of androgenic signaling of PCa cells. However, about 80% of patients treated with androgen deprivation therapy will develop resistance to this treatment, generating the so-called Castration Resistant PCa (CRPC), the most aggressive phenotype of this disease, which remains lethal nowadays. Therefore, the search for alternative therapeutic targets useful for the development of new, more efficient, treatments than those currently available continues to be one of the main objectives of the biomedical scientific community. PCa develops mainly in older patients, who suffer, with a high frequency, endocrinemetabolic pathologies (i.e. obesity, diabetes), which can influence the development and/or progression of PCa, increasing the complexity of this pathology. Specifically, obesity is associated with an increased risk of developing PCa, and with higher PCa aggressiveness. In this context, it has been described that metformin and statins (drugs commonly used in patients with endocrine-metabolic pathologies such as obesity or diabetes) exert antitumor effects in numerous tumor types, both in vitro and in vivo. However, although the combination of both drugs has been shown to exert additive antitumor actions in some cancer cell types, this possibility has not been systematically studied in the case of PCa. An important limitation in this field is that the information available regarding the exact molecular mechanisms underlying the pathophysiological interaction between obesity and PCa is very limited and fragmentary. In this sense, it is well-known that numerous components of different hormonal systems (i.e. androgen, ghrelin, etc.) are clearly dysregulated in PCa and obesity, which may constitute a source of biomarkers in these pathologies. The ghrelin system is a clear example in this context, since it is a pleiotropic hormonal system consisting of various ligands (e.g. Ghrelin, In1-ghrelin), enzymes (GOAT), and receptors (GHSR1a, GHSR1b). Specifically, In1-ghrelin (an alternative splicing variant derived as the result of the intron-1 retention) is overexpressed and plays an oncogenic role in PCa. Likewise, GOAT levels are higher in PCa tissues compared to nontumor prostate tissue, but the putative pathophysiological role of this enzyme in PCa is still unknown nowadays. Furthermore, it has been reported that circulating levels of both In1- ghrelin and GOAT are higher in PCa patients compared to healthy patients, suggesting that both elements could represent non-invasive diagnostic biomarkers of PCa. However, to date, it is unknown whether In1-ghrelin or GOAT can be detected and whether they are dysregulated in urine (which constitutes an enriched source of proteins derived from the prostate) in patients with PCa compared with patients without PCa. Finally, it is important to mention that the alteration of the splicing process has become a common hallmark of a large number of tumor pathologies, including PCa. Splicing is a physiological mechanism involved in the maturation of pre-mRNAs, which can generate different mature mRNAs from a single gene, thereby increasing the transcriptomic complexity of eukaryotic cells. However, increasing evidence demonstrates the existence of different splicing variants with oncogenic potential which are associated to the development, progression and/or resistance of pharmacological treatments of multiples tumor pathologies, including PCa. These dysregulations of the splicing process may be due to an alteration of the machinery that regulate this cellular process (spliceosome) and/or of the proteins that regulate this mechanism (splicing factors), suggesting that they could play a relevant role in the development and/or progression of PCa. Although certain studies have shown the specific dysregulation of some of these elements in PCa, the possible involvement of the alteration of this machinery and the putative implication in the development and/or progression of PCa has not been systematically explored yet. Based on the all the information mentioned above, the general aim of this Doctoral Thesis was to determine the potential role that certain splicing-related elements (spliceosome components and splicing factors) and metabolic modulators (endogenous elements and exogenous treatments) play and/or exert in the development, progression and/or aggressiveness of PCa, with the ultimate goal of discovering novel biomarkers (diagnostic and/or prognostic) and therapeutic tools that could improve the diagnosis, treatment and the clinical management of PCa patients. The first section of this Doctoral Thesis was focused on exploring the putative dysregulation and the clinical and functional implications of the splicing machinery in PCa. Results presented herein indicate that the splicing machinery is drastically dysregulated in PCa, including pivotal elements of this machinery such as SNRNP200, SRRM1, SRSF3, RBM22 or SF3B1. Among all the dysregulated spliceosome components and splicing factors, SNRNP200, SRRM1 and SRSF3 were selected initially to be further explored because they were associated with all the relevant aggressiveness features available in the cohort of patients used in this study. Mechanistically, silencing of SNRNP200, SRRM1 or SRSF3 inhibited oncogenic pathways and altered the splicing process of genes previously reported to be involved in tumor aggressiveness and/or CRPC development. Consistently, SNRNP200-, SRRM1- and SRSF3-silencing sensitized the CRPC cells to an androgen receptor (AR)-inhibitor (enzalutamide). Therefore, these data demonstrate that SNRNP200, SRRM1 and SRSF3 could represent attractive novel diagnostic and prognostic biomarkers as well as therapeutic targets for PCa and CRPC. This Doctoral Thesis was also focused on exploring the potential dysregulation and role of RBM22 in PCa, a splicing factor tightly involved in the activation of the spliceosome. Specifically, we found that RBM22 levels were lower (at mRNA and protein levels) in PCa tissues compared to non-tumor prostate tissues, and that these levels were inversely associated to key clinical parameters of aggressiveness in patients and in a preclinical mouse model. Moreover, we observed that the aggressiveness features of PCa cells were reduced in response to RBM22 overexpression in vitro and in vivo (i.e. reduced tumor growth, proliferation, migration, etc.) Mechanistically, results from high-throughput techniques (i.e. RNAseq, CLIP-seq and signaling array) performed in samples from in vitro and in vivo PCa models pointed out that the inhibition of MYC, MYCN and E2F, as well as a profound dysregulation of the splicing process could be the molecular mechanisms associated to the relevant findings previously observed after the RBM22 overexpression in PCa. Therefore, these results suggest that RBM22 plays a relevant antitumor role in PCa through the regulation of the splicing process and the reduction of MYC, MYCN and E2F levels/activity. In the case of SF3B1, a key component involved in the spliceosome assembly, we demonstrated that SF3B1 is overexpressed in PCa and that its levels were directly associated with important aggressiveness features of PCa. Moreover, the downregulation of SF3B1 and the inhibition of its activity (using the selective inhibitor pladienolide-B) significantly reduced functional aggressiveness features and increased the apoptotic rate of PCa cells. From a molecular perspective, SF3B1 inhibition altered PI3K/AKT and JNK oncogenic signaling pathways, down-regulated key oncogenic splicing variants (AR-v7 and In1-ghrelin), and dysregulated the expression of several elements involved in mRNA metabolism [splicing and non-sense mediated mRNA decay (NMD)]. Taken together, these results suggest that SF3B1 could represent a new prognostic biomarker and a therapeutic target in PCa, providing convincing evidence for the putative utility of pladienolide-B as novel therapeutic tool for the treatment of this devastating pathology. Once demonstrated the key role of several splicing-regulating factors in PCa, and since metformin has been found to exert antitumoral effects in various cancer types and that alter the expression of specific splicing factors in certain cell types, we next aimed to explore the potential involvement of the splicing dysregulation in the antitumor actions of metformin in PCa. Our data indicate that metformin produced a profound downregulation in the expression of different spliceosome components and splicing factors in PCa cells, being some of these changes (i.e. downregulation of SF3B1, SRRM1 and NOVA1) further validated in vivo using a PCa xenograft model. Moreover, the antiproliferative response of metformin treatment observed in PCa cells was completely blocked when the expression of SF3B1, SRRM1 or NOVA1 were silenced. Therefore, these results suggest that the dysregulation of some components of the splicing machinery, especially SF3B1, SRRM1 and NOVA1, might represent an additional molecular mechanism underlying the wellknown antitumor effects of metformin in PCa. The second section of this Doctoral Thesis was focused on the analysis of the antitumor actions of metformin, statins and their combination, as well as the underlying molecular mechanisms in PCa. Specifically, PCa patients treated with metformin and statins in combination were associated to more beneficial effects in key clinical parameters of aggressiveness compared to those receiving individual treatments. Consistently, although biguanides and statins significantly decreased the aggressiveness of PCa cells in vitro, these antitumor effects were more pronounced when combining both treatments. Mechanistically, we found that the treatment with the combination of metformin and simvastatin in vitro resulted in a hyper-inactivation of AR and mTOR, as well as in an upregulation of the expression of different cyclin-dependent kinase inhibitors (CKIs; CDKN1A, CDKN1B, CDKN2A and CDKN2D), being some of those changes further validated in samples from patients with PCa (i.e. upregulation of CDKN1B and CDKN2A). Altogether, this study demonstrates that metformin and simvastatin exert additive antitumor effects in PCa, thus suggesting that this combination might be a novel and attractive therapeutic approach to tackle PCa. Finally, in the third section of this Doctoral Thesis, we analysed specific components of the ghrelin system (specifically, GOAT enzyme and In1-ghrelin variant) to unveil their potential role as diagnostic, prognostic and/or therapeutic tools in PCa. These results demonstrated that GOAT could be detected in urine samples, wherein its levels outperformed the capacity of plasma PSA levels to diagnose PCa, especially clinically significant PCa (SigPCa; defined as Gleason score ≥7) in patients with PSA in the grey zone. Moreover, high urine GOAT levels were associated to increased risk of developing PCa and SigPCa, and were positively correlated to key clinical and molecular aggressiveness parameters in PCa patients. Furthermore, we demonstrated that GOAT overexpression increased PCa aggressiveness in vitro and in vivo, while its silencing and the blockade of its activity (using a specific GOAT inhibitor) significantly decreased the aggressiveness features of PCa cells in vitro. Therefore, these results demonstrated that GOAT enzyme may represent an additional non-invasive diagnostic biomarker, as well as a potential prognostic and therapeutic tool for PCa. Moreover, the analysis of In1-ghrelin in urine from patients with PSA in the grey zone indicated that its levels were able to distinguish between patients with and without PCa, were associated to higher risk of PCa, obesity and diabetes, and correlated with metabolic and aggressiveness parameters. In fact, we found that the capacity of urine In1-ghrelin levels to diagnose PCa was especially elevated when only obese patients were considered, suggesting its potential utility as a noninvasive diagnostic biomarker from a more personalized perspective. Altogether, the general conclusion derived from this Doctoral Thesis is that the alteration of the splicing process and the dysregulation of certain endocrine-metabolic systems could contribute to the development, progression and/or aggressiveness of PCa, representing a source of novel diagnostic and prognostic biomarkers, as well as therapeutic targets that could be exploited to improve the diagnosis of PCa patients, and to develop more effective prognostic and therapeutic tools to tackle this devastating pathology.