Background
Prostate cancer is the most common non-cutaneous malignancy in men in the Western world. The etiology of prostate cancer is not well defined; however, the inhibition of various tumor suppressor genes and concomitant activation of oncogenes is a frequent occurrence in most cancers, including prostate cancer. DNA hypermethylation of promoters at CpG sequences often sterically hinders the binding of transcription factors, thereby represses gene transcription [
1]. The transfer of a methyl group to DNA at the fifth carbon position of cytosine residues by the DNA methyltransferase (DNMT) family of enzymes is one of the most common events for the establishment of epigenetic program [
2]. Three active DNMTs have been identified so far in mammalian cells; DNMT1 (Gene ID: 1786), DNMT3A (Gene ID: 1788), and DNMT3B (Gene ID: 1789). DNMT1 is the most abundant and methylates hemimethylated CpG di-nucleotides in the mammalian genome during DNA replication in a number of different cancer types [
3]. DNMT3A and DNMT3B are methyltransferases involved in de novo methylation of DNA following replication [
4].
Numerous reports have demonstrated the overexpression of DNMT1 in lung, hepatocellular, acute and chronic myelogenous leukemia, colorectal, gastric, breast and prostate cancers [
5‐
11]. Gravina and associates have shown that the hormone resistant prostate cancer phenotype is associated with an increase in DNMT expression and activity [
12]. The majority of gene silencing induced by DNA hypermethylation in prostate cancer can be attributed to DNMT activity [
13]. Since promoter hypermethylation results in the silencing of several tumor suppressor genes in cancer cells, the reactivation of these genes by demethylation of their promoters is a feasible approach to cancer therapy. DNMT inhibitors, such as 5-aza-deoxycytidine and 5-aza-2’-deoxycitidine, have been used extensively in clinical trials for cancer treatment. These nucleoside analogues bind covalently to the DNMTs and irreversibly inhibit their function, leading to the demethylation of silenced promoters and subsequently, the activation of gene expression. However, studies in rodent cell lines and bacteria have indicated that these azacitidine-DNMT adducts are toxic and mutagenic if not repaired [
14‐
16]. In addition, these compounds are unstable in neutral aqueous solution and disintegrate to yield more stable analogues such as 5, 6-dihydro-5-azacytidine and 5-fluoro-2’-deoxycytidine [
17], that have been shown to have toxicity related issues in clinical trials [
18,
19]. Therefore, there is an urgent need for the development of new drugs that will target DNMT with low toxicity.
Ras-association domain family 1A (
RASSF1A; Gene ID: 11186) gene has been found to be the most frequently methylated gene described thus far in human cancers [
20,
21]. Hypermethylation of the promoter of
RASSF1A gene at its CpG-island has been observed in 70% of prostate cancers [
22,
23]. Since the restoration of
RASSF1A expression in tumor cell lines impairs tumorigenicity [
22,
24], factors that restore
RASSF1A expression have immense importance in preventing tumor growth. We demonstrated earlier that mahanine induces
RASSF1A gene expression in a diverse range of cancer cell types, including epidermoid, lung, pancreatic, colon, breast, ovarian and prostate cancer cells [
25]. Although we showed a decline in DNMT activity upon mahanine treatment, the scope of our study did not include establishing a causative effect of DNMT inhibition on
RASSF1A re-expression by mahanine, and this mechanism remains to be explored.
The cellular levels of DNMTs in mammalian cells can be regulated by transcriptional events or posttranslational modifications of enzymes which ultimately affect the catalytic activity and degradation of the DNMT proteins [
26]. DNMT1 is known to be phosphorylated at several serine and threonine residues under physiological conditions. Sun and associates have reported that Akt enhances DNMT1 protein stability by inhibition of its ubiquitin–proteasome mediated degradation [
27]. Recently it has been demonstrated that Akt1 directly interacts and phosphorylates Ser143 of DNMT1 to increase its stability [
28].
In the present study, we sought to establish the mechanism by which mahanine inactivates DNMTs and thereby restores RASSF1A expression in prostate cancer cells. We show that mahanine induces the degradation of DNMT1 and DNMT3B via the ubiquitin-proteasome mediated pathway in both androgen-responsive LNCaP and androgen receptor-negative PC3 human prostate cancer cells. Interestingly, our data suggests that the inactivation of Akt by mahanine treatment is involved in inducing DNMT degradation, thereby restoring the expression of the epigenetically silenced tumor suppressor gene RASSF1A.
Discussion
The silencing of the
RASSF1A gene has been associated with advanced stages of prostate cancer [
22]. Our prior work demonstrated that mahanine possesses the ability to restore
RASSF1A expression in several different cancer cell lines, including androgen-responsive LNCaP and androgen-receptor negative PC3 prostate cancer cells, which have been derived from metastases to the lymph node and bone, respectively [
25]. However, the scope of that report did not include defining a molecular mechanism to explain the re-expression of
RASSF1A upon mahanine treatment, although it did show a decline in total DNMT activity in the presence of mahanine [
25]. Our current work stems from these findings, and in this report we establish a molecular mechanism for the inhibition of DNMT signalling by mahanine; furthermore, we delineate the correlation between DNMT and
RASSF1A expression in prostate cancer cells.
Our data demonstrate that the re-expression of RASSF1A upon mahanine treatment is a time sensitive event; while the demethylation of the RASSF1A promoter is apparent after 24 hours of mahanine treatment, RASSF1A expression can only be detected after 72 hours of mahanine treatment. This shows that the partial de-methylation of the RASSF1A promoter which occurs after 24 hours of mahanine treatment is not enough to restore its expression; the promoter region must be demethylated to a greater extent, which is only achieved approximately 72 hours following mahanine treatment. This suggests that mahanine could modulate the activities of certain factors involved in the maintaining the methylation pattern of the RASSF1A promoter region, which accounts for the time lag between mahanine treatment and RASSF1A re-expression. Since DNMT1, DNMT3A and DNMT3B are known to be involved in de novo methylation and the maintenance of methylation patterns of genes, we investigated the levels of expression of all three members of the DNMT family. Our data clearly indicates that mahanine treatment causes a decline in the levels of DNMT1 and DNMT3B, without affecting the levels of DNMT3A, in both LNCaP and PC3 prostate cancer cells. Interestingly, the time frame within which mahanine down-regulates DNMT1 and DNMT3B in LNCaP and PC3 cells correlates well with the re-expression of RASSF1A in these cell types, suggesting that mahanine could mediate RASSF1A re-expression via DNMT inhibition. Furthermore, these results indicate that mahanine selectively modulates the cellular levels of certain DNMTs, without ubiquitously down-regulating the levels of all members of the DNMT family. This data is further supported by our findings that the knock-down or over-expression of either member of the DNMT family is sufficient to restore or inhibit the expression of RASSF1A in PC3 or BPH1 cells, respectively; however to a lesser extent with DNMT3A. Therefore, mahanine selectively degrades the two DNMTs which appear to most strongly inhibit RASSF1A expression in PC3 prostate cancer cells. The ability of mahanine to selectively target DNMT1 and DNMT3B clearly differentiates it from other known DNMT inhibitors like 5-aza cytidine and 5-aza-2′-deoxycitidine, which ubiquitously bind to and irreversibly inhibit all members of the DNMT family. Therefore, anti-cancer agents like mahanine which selectively targets DNMT1 and DNMT3B could be beneficial in prostate cancer therapy.
While mahanine causes degradation of DNMT1 and DNMT3B by inducing the chymotrypsin-like activity of the proteasome, it is interesting to note that it does not potentiate the trypsin- and caspase-like activities of the proteasome. The decline in these particular enzymatic activities of the proteasome upon mahanine treatment could be to compensate for the significant induction of the chymotrypsin-like activity upon mahanine treatment and thereby ensure that the overall proteasomal activity is balanced and cellular homeostasis is undisturbed.
The activity of survival kinases such as Akt is known to be highly up-regulated in prostate cancer, which correlates with the high abundance of DNMTs in prostate cancer cells, as Akt is involved in the stabilization of DNMT1, and possibly other DNMTs, via site-specific phosphorylation on Ser/Thr residues. Interestingly, the degradation of DNMTs by mahanine is dependent on its ability to inhibit Akt activity; when Akt is constitutively active mahanine treatment does not result in proteasomal degradation of DNMT1 and DNMT3B. A recent report demonstrated that Akt phosphorylates the Ser143 residue of DNMT1 and thereby increases its stability [
28]. However, no such stabilizing phosphorylation events have been described to date for DNMT3B. Our data clearly indicates that Akt is involved in stabilizing not only DNMT1, but also DNMT3B, since constitutively active Akt renders both DNMTs resistant to proteasomal degradation induced by mahanine. The mechanism by which Akt imparts increased stability to DNMT3B remains to be explored. In addition, our immunoprecipitation data shows that mahanine causes a striking decrease in the overall serine phosphorylation of DNMT1, suggesting that mahanine might modulate the phosphorylation status of DNMT1 on more than one serine residue, via the inactivation of other Ser/Thr kinases, in addition to Akt. It will be interesting to further explore specific phospho-sites involved in imparting stability to DNMT1 and DNMT3B and the kinases responsible for this phosphorylation.
Acknowledgements
This work was supported by National Institutes of Health R01 CA131123 to Dr. Partha P. Banerjee. We thank Dr. Simon W. Hayward of Vanderbilt University, Nashville, TN for his generous gift of BPH1 cells. We also thank Dr. Gerd P. Pfeifer, Department of Cancer Biology, Beckman Research Institute of City of Hope, Duarte, CA for providing us RASSF1A expression vector.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SA is responsible for the execution, data interpretation, data analyses and drafting of the manuscript. KA contributed to experimental design, conducting various experiments for the revision and the editing of the manuscript. SJ performed several RT-PCR experiments. GB, PR, NB and SB were responsible for the purification of mahanine. PB directed the experimental design and provided insight for experimental execution, and editing the manuscript and figures. All authors have read and approved the final manuscript.