Testis specific Y-like 5: gene expression, methylation and implications for drug sensitivity in prostate carcinoma
In this study we have demonstrated that the presence of DNA methylation in the 5? region of the gene is negatively associated with expression of TSPYL5 mRNA and protein in PC cells, NT cells and clinical prostate tissue samples. Methylation induced TSPYL5 gene silencing was previously reported in glioma and gastric cancer types [15, 16]. The TSPYL5 protein expression mirrored the expression pattern of mRNA in the cells (DU145, LNCaP or RWPE-1). The TSPYL proteins are members of the NAP superfamily of proteins [9] that have been shown to bind to proteins involved in transcription, cell cycle regulation [25], and shuttling histones between nucleus and cytoplasm [26]. However, it is not clear whether such a function for TSPYL5 exists in PC cells.
Previous studies with colorectal HCT116 cells indicated that both DNMT1 and DNMT3B enzymes were essential to methylate TSPYL5 gene promoter regions [15]. While one or the other enzyme was observed in the cells tested in this study, we observed only DNMT3B protein was predominantly expressed in more advanced PC tissues in which TSPYL5 was absent. Earlier studies in prostate cancer have analyzed various methyltransferases and found that DNMT1 expression was found to be lower than DNMT3b. Further, de novo methylation remains in DNMT1 knockout embryonic stem cells and the role of DNMT1 in tumor methylation remains ambiguous [27]. Depending on the cellular context, the TSPYL5 gene might be differentially targeted for methylation by methyltransferases.
Previous studies have shown a correlation between methylation in chromosome 8 region (Chr 8: 97278129–97278175) and loss of TSPYL5 gene expression in lung carcinoma cells, although, no tissue studies or normal cell studies have been done [17]. While CpG islands are important to regulate gene expression [28], previous studies suggest that the lower density CpG shores of islands may also be important [29]. Our studies with MSP analysis of the CpG island identified methylation of the TSPYL5 gene in PC cells and tissues. As anticipated, PSQ analysis of CpG dinucleotides on the 5? shore of the CpG island revealed higher methylation of the four cytosine residues (Chr 8: 97278367–97278417) in DU145 cells relative to other cell lines. Only a subtle difference was observed in individual cytosine methylation between, LNCaP and RWPE-1. This is in keeping with our observations that DU145 had the least TSPYL5 expression due to methylation-induced gene silencing.
Planning treatment for prostate cancer patients relies on histopathological grading by GS [30] which currently lacks a precise molecular correlate [6, 7]. There is a critical need to identify companion biomoleules that distinguish more advanced phenotype tumors within intermediate GS-7 specimens. Our studies identified TSPYL5 mRNA and protein expression in benign and tumor tissues with a GS-6 or-7. High grade tumors with GS???8 had the least expression of TSPYL5, likely due to DNA methylation. Interestingly, a few GS-7 tumor samples with Gleason pattern (4?+?3) had no message for TSPYL5. At this time, it is not clear whether the absence of TSPYL5 mRNA expression in tissues with GS 7 (4?+?3) would indicate any undetected higher-grade disease. Further studies with more tissues are needed to assess this possibility. Taken together, these data suggest that the absence of TSPYL5 may be an indicator of more advanced prostate cancer disease.
MSP analysis of the TSPYL5 gene indicated a mixture of methylated and unmethylated bands in benign and intermediate-grade tumors with GS-6 or ?7, while GS-8 tumors had predominantly methylated bands suggesting a methylation induced TSPYL5 silencing in these tumors. Previous studies indicated that TSPYL5 could be an independent marker of poor outcome in breast cancer based on their high expression in aggressive basal-like breast cancers [31]. On the contrary, we observed both by mRNA expression and IHC that TSPYL5 expression diminishes in high grade tumors. Such a difference in TSPYL5 expression could be exploited to identify the clinical behavior of intermediate grade prostate tumors (GS-7). A recent report suggested the use of higher levels of SNPs-rs2735839 to stratify patients with GS-7 because of the association with aggressive PC [32]. However, to classify GS-7 patients based on diminished TSPYL5, large cohorts of prostate tumor samples will need to be investigated. Studies along this direction are in progress in our laboratory.
In addition to its anti-proliferative role, CDKN1A is also vital to proliferation and survival. A previous study reported that knockdown of TSPYL5 increased the endogenous expression of p53 and its downstream target CDKN1A in MCF7 breast carcinoma cells [31]. It has been reported that in lung carcinoma cells, TSPYL5 was able to suppress CDKN1A by modulating PTEN/AKT pathway [17]. Also, TSPYL5 gene silencing increased the CDKN1A protein expression and caused growth reduction in cells [17]. However, we observed that TSPYL5 gene silenced cells (DU145) exhibited very minimal expression of CDKN1A. Conversely, low or moderately TSPYL5 expressing LNCaP and RWPE-1 cells showed high and relatively low CDKN1A expression. We also observed a decrease in CDKN1Aprotein expression in TSPYL5 overexpressing LNCaP cells. However, in contrast to lung carcinoma cells [17] LNCaP cells lack PTEN, so any involvement of TSPYL5 in modulating CDKN1A must work by a PTEN-independent mechanism [33].
Our observations identified that even low TSPYL5 expressing cells (eg: LNCaP) had higher pAKT. Also, TSPYL5 expressing RWPE-1 cells expressed basal pAKT, albeit low levels compared to LNCaP cells. This in sharp contrast to the observation made in lung carcinoma cells [17] that high TSPLY-5 expression can activate AKT. The exact role of TSPYL5 is not clear in modulating AKT expression in PC cells and could vary depending on the cellular phenotype.
Mounting evidence suggests that TSG’s play an important role in the response to a variety of chemotherapeutic drugs such as px, 5-Fluorouracil, cisplatin and trastuzumab [34, 35]. It was reported that a decrease in retinoblastoma (Rb) protein, a TSP, in sarcoma cells conferred resistance to doxorubicin and cisplatin [36]. PC cells and glioblastoma cells deficient in Rb were resistant to cisplatin and doxorubicin, respectively [37, 38]. Similarly, p53 inactivation resulted in reduced sensitivity to cisplatin but not px in ovarian carcinoma cells, suggesting that the role of p53 in response to chemotherapy depends on both cellular context as well as the class of chemotherapeutic compounds [39]. Increased expression of CDKN1A leads to chemoresistance, and its loss sensitizes the cells to chemotherapy response [40, 41]. Interestingly, it was reported earlier that LNCaP cells were resistant for dtx and knockdown of p53 protein increases its sensitivity to the drug by decreasing CDKN1A [42]. Previous studies had shown that increase in TSPYL5 can compete with USP7, a deubiqutinylating protein thereby decreasing p53 in MCF 7 cells [31]. Our studies show that reduction in CDKN1A in TSPYL5 overexpressing LNCaP cells exhibit more sensitivity for dtx and px compared to WT cells. All the above studies highlight the possible roles of TSP’s in chemotherapy response. Our studies suggest that increased TSPYL5 enhances the sensitivity of the cells to chemotherapy drugs, likely by downregulating CDKN1A. While it is tempting to suggest that TSPYL5 status in PC cells could be indicative of predicting chemotherapy response, further studies are needed to substantiate this notion. In keeping with the previous studies regarding the role of TSP’s in chemosensitivity [40, 41], we speculate that the response to chemotherapy drugs in TSPYL5 expressing PC cells likely may vary depending upon the cellular context and the type of chemotherapy drug.