Methylation mediated silencing of TMS1/ASC gene in prostate cancer

Gene silencing associated with aberrant promoter methylation has been suggested as an alternate pathway for development of cancer [1]. This form of epigenetic change contributes to tumor initiation and progression by transcriptional silencing of tumor-suppressor genes. Several genes have been shown to be epigenetically inactivated in a wide range of tumors [2] and most neoplasms show hypermethylation of one or more genes [3‐5]. This has led to the concept of a 'hypermethylation profile' of tumors which could have potential clinical applications [5‐7]. The hypermethylated genes can be broadly classified as those involved in cell cycle regulation (p16INK4a, p15, Rb), genes associated with DNA repair (BRCA1, MGMT), apoptosis (DAPK), and drug detoxification (GSTPi), drug resistance (MGMT), cellular differentiation, angiogenesis (THBS1) and metastasis (E-cadherin) [2].

Epigenetic dysregulation of an apoptotic pathway appears to be connected to the development of many cancers as confirmed by several research communications [8, 9]. Apoptosis or 'programmed cell death' is essential for embryonic development and also plays an important role in the immune system and maintenance of cellular homeostasis [10]. A defect in apoptosis is implicated in neurodegenerative diseases, autoimmunity, and cancer and in chemo-resistance. Multiple genes direct the apoptotic pathway to restrain the inappropriate proliferation of cells and a defect in the signaling mechanism gives the cancer cells an added survival advantage leading to tumor initiation, progression and even drug resistance. Apoptosis is mediated by a family of cystine proteases called caspases. There are two groups of caspases – the initiator caspases (CASP8, CASP9 and CASP10) and the effector caspases (CASP3, CASP6 and CASP7). Caspases exist as latent pro-enzymes and are activated by proteolytic cleavage. Several key genes involved in apoptosis have been showed to be target of epigenetic changes [9].

Recently, TMS1 was shown to be aberrantly hypermethylated in breast cancer tissues and cell lines [11]. This gene is also known as ASC (Apoptosis Speck Like protein containing a CARD) [12]. TMS1/ASC encodes a 22-kDa CARD protein and promotes apoptosis in a caspase 9 dependent pathway. TMS1/ASC has been shown to bind with various proteins like Pyrin, Ipaf and copyrin/PYPAF1[12]. In addition to its role in cancer development there is ample evidence which suggests that TMS1/ASC is involved in immune responses and NF-κB and caspase1 activation [13, 14]. The downregulation of TMS1/ASC in breast cancer cell lines correlates with dense methylation on the CpG islands. Methylation of the promoter region of TMS1/ASC has also been identified in small cell lung cancer and non-small cell lung cancer [15], human glioblastoma [16], ovarian tumors [17, 18], colorectal cancer [19], neuroblastoma [20], and melanoma [21]. However in a study by Roman-Gomez et. al., on acute lymphoblastic leukemia patients no correlation was found with methylation of the TMS1/ASC gene [22].

So far no studies have been reported on the role of TMS1/ASC in prostate cancer. In this study we examined the mechanism of TMS1/ASC gene silencing in prostate cancer cell lines. We also studied the prevalence of TMS1/ASC gene methylation in prostate cancer tissue samples as well as the association of race and clinico-pathological features with TMS1/ASC gene methylation.

Our study shows a clear correlation between TMS1/ASC methylation and silencing of the gene suggesting a role in prostate cancer cell lines. Re-expression accompanied by partial demethylation of TMS1/ASC following 5-AZA-2'-deoxycytidine confirms that methylation is responsible for transcriptional silencing of this gene. Lack of response to TSA treatment suggests that histone acetylation does not play a role in downregulating the expression of TMS1/ASC. This finding is similar to the results obtained by Stimson et al[23]. Over expression of TMS1/ASC was shown to inhibit cellular proliferation and induce DNA fragmentation which can be blocked by a caspase inhibitor [24]. In addition, forced reduction in TMS1/ASC promotes cell survival perhaps in a NFκ-B dependent pathway [25]. This makes it a therapeutic target by use of demethylating agents alone or in combination with additional apoptosis inducing drugs.

ChIP analysis showed binding predominantly with MBD3 and only minimal enrichment of the other MBDs and MeCP2_. Earlier studies have demonstrated that MBD3 does not bind to methylated DNA alone [26, 27]. However, it has a definite role in maintaining methylation. Recent studies have shown binding of MBD3 to several genes like cox6c, leng6, bat5 etc. [28]. Wade et al showed that MBD3 is a subunit of the NuRD complex that has nucleosome remodeling and histone deacetylase activities [29]. MBD3 forms a part of the multiprotein NuRD complex and probably has a role as a transcriptional co-repressor. Our finding that MBD3 binds to methylated TMS1/ASC is contrary to the known pattern of MBD binding by MBD proteins to methylated DNA. This could be because MBD3-containing NuRD complexes bind more specifically to the methylated TMS1/ASC gene [28].

A noteworthy finding in our study was the statistically significant difference in the methylation of TMS1/ASC among the cases and controls in Whites compared to Blacks. We observed an age adjusted odds-ratio of 7.6 (95% CI 2.1–27.3) in Whites as compared to only 1.1 (95% CI 0.2–5.5) among Blacks. Whether this finding reflects involvement of different pathogenetic pathways in different races will be interesting to study. In the US, the incidence and mortality of prostate cancer is about two-fold higher among Blacks compared to Whites, suggesting racial differences in prostate tumor occurrence and aggressiveness [30]. The reason for these racial differences is not well understood. It is possible that promoter-region gene hypermethylation may be influenced by environmental exposures.

There have been only a few studies on differences in gene methylation between different races. Two studies by Woodson et al showed a differential methylation pattern and expression of CD44 in Blacks and Whites [31, 32]. When we examined 5 genes frequently methylated in prostate cancer (GSTP1, CD44, ECAD, RASSF1A and EBR) in the same patient population we did not find any significant differences in the gene specific methylation pattern between the different races (Table 5). Ethnic group related differences in hypermethylation of promoter region of GSTP1 gene were reported in a recent paper [33]. The authors observed higher hazard ratio (HR) for pathogenesis among African Americans as compared to Caucasians. In contrast to the above observation, we found the hazard ratio of TMS1/ASC methylation (prostate cancer versus BPH) to be lower in Blacks as compared to Whites. Ethnic origin is an important determinant of prostate cancer risk, incidence, and disease progression. In the US, the African-American male group has the highest incidence rate for prostate cancer [30]. Differences in diet, socioeconomic environment, lifestyle between the two ethnic groups have been implicated as causative factors for the striking ethnic differences in the incidence and clinical behavior of prostate cancer. However, molecular mechanisms underlying the racial diversity are not well understood. The recent report by Fang et. al., shows that Genistein leads to reversal of hypermethylation and reactivation of p16INK4a, RARβ, and MGMT genes [34]. Thus, diet seems to be an important factor in affecting the methylation status of different genes implicated in cancer. It is also likely that genes are differentially methylated in different ethnic groups owing to the lifestyle and dietary differences. Methylation of promoter in controls (BPH) may reflect that epigenetic alteration of the gene has already occurred and that they have acquired epigenetic malignant potential even though the pathological diagnosis classifies them as benign [33]. Our results indicate that differences in methylation pattern of TMS1/ASC in BPH among ethnic groups might explain the differences among different racial groups in susceptibility to prostate cancer. Thus it seems that the epigenetic make up of different ethnic groups would determine the risk to prostate cancer pathogenesis.

There did not appear to be a significant relationship between TMS1/ASC methylation status and patient's age. On our limited dataset there was a trend towards association between TMS1/ASC methylation and Gleason score 7 or higher (odds-ratio 2.3). Though the relationship was not statistically significant it possibly suggests a worse prognosis as patients with Gleason score 7 do worse that those with Gleason score 6. Many of our patient samples showed partial methylation status – this could be due to presence of normal fibroblasts, endothelial cells, inflammatory cells and non malignant prostate tissue surrounding the tumor. Tumors where TMS1/ASC was not found to be methylated could involve other genes in the apoptotic pathway. Use of pathway specific cDNA microarrays may help in determining the individual genes affected.

In summary our study has shown that TMS1/ASC, a pro-apoptotic gene, is silenced by hypermethylation of the CpG islands in the promoter region. This transcriptional repression is relieved by treatment with a demethylating agent (5-Aza-2'-deoxycytidine). Frequent methylation of TMS1/ASC in prostate cancer suggests this gene may be important in pathogenesis of prostate cancer and can be a target of pharmacologic demethylation in clinical trials. TMS1/ASC methylation patterns show significant ethnic differences. Our findings provide a novel insight into the molecular determinants of tumor growth that may underlie the ethnic differences in prostate cancer incidence and clinical behavior. Further studies are needed to find out if this has any significant clinical implications in the development of novel diagnostic approaches for biologically aggressive prostate cancer from diverse racial origin.