Many malignancies, including CaP, exhibit aberrant methylation within the promoter regions of genes associated with a loss of function 
. Presumably, this loss of function contributes to the development and progression of the disease. DNMTs are the major mediators responsible for the hypermethylation of the promoter regions of many genes encoding for signaling factors including the TβRs promoter 
, which may subsequently inhibits TβRs translation which ultimately results in the insensitivity to the normal inhibitory effects of TGF-β, uninhibited growth and progression of cancer 
. Although DNMTs are recognized as important regulators of transcription of carcinogenesis 
, and have been a topic of considerable interest in the last few years, their assessment in vivo and within human specimens remains uncertain. Our study findings demonstrate that high level of expression of DNMTs is associated with more aggressive phenotypes of CaP, lower expression of TβRs, and lower sensitivity to the inhibitory role of TGF-β.
The molecular mechanisms which govern regulation of DNMTs have been largely unknown 
, and the relationship(s) between DNMTs and TGF-β in CaP have yet to be explored. Although other factors like c-Jun may be involved in the process 
, ERK appears to be an obligatory switch for TGF-β-mediated expression of DNMTs in CaP, although the effect of TGF-β on ERK activation remains controversial 
. More recently we reported that there was a differential activation of ERK between benign and malignant cells in response to TGF-β 
. In our prior studies involving benign cells, we reported that TGF-β exposure, ERK inactivation and DNMTs down regulation contribute to the expression of Foxp3 in benign immune cells 
. In the present study, higher expression levels of DNMTs were found to be associated with CaP with higher invasive capabilities when compared with CaP cells with lower invasive capabilities. Interestingly, we found that increased levels of DNMTs were associated with increased levels of TGF-β and p-ERK, and decreased levels of TβRs. In contrast, our hypotheses were verified by a serial of blockade assays, blockade of TGF-β signaling using the TβRIIDN or neutralizing antibody 1D11, decreased the levels of DNMTs between 50%–90% in more invasive cell lines, and to a lesser degree (30–50%) in the less invasive cell lines. These findings indicate that tumor-derived TGF-β is a major mediator involved in the regulation of DNMTs and TβRs in human CaP cells, and this process correlates with more invasive phenotypes. Meanwhile, down regulation of DNMT expression by blockade of TGF-β is associated with an up-regulation of naïve TβRs expression. These findings, taken together with results from our previous study, suggest that tumor-derived TGF-β activates ERK, which mediates the expression of DNMTs (because blockade of ERK resulted in 50% decrease on DNMTs expression). DNMTs then methylate the TGF-β receptor promoter regions resulting in the loss of growth inhibition mechanisms which we reported earlier 
. Our present study also provides insight into the interaction between ERK and DNMTs in CaP. Exposure to the ERK inhibitor, UO126 results in >50% reduction in the expression of DNMTs, indicating that ERK is one of the major regulators of TGF-β induced DNMTs expression in CaP cells. Our observations of the co-localization of p-ERK and DNMTs also suggest that only cells which exhibit ERK activation can express DNMTs, which is evidence that they are in the same TGF-β activated signal pathway.
Importantly, we found direct evidence that blockade of DNMT by its inhibitor 5-Aza resulted in decrease in the invasive capabilities of CaP, as well as the blockade of either TGF-β by 1D11, or blockade of p-ERK by UO126. This data indicates that DNMT is a major promoter for CaP invasive capabilities. This procedure is regulated by TGF-β and mediated by p-ERK.
Based upon the above findings, we postulate that tumor-derived TGF-β can regulate its receptors by a potential feedback loop which is mediated by activation of ERK. Some other signaling factors like Serine/threonine protein phosphatases 2 (PP2A) 
may be involved in this procedure. P-ERK may subsequently activate the transcription factors in the DNMTs promoter and increases the expression of DNMTs which methylates TGF-β receptor promoter regions resulting in the loss of growth inhibition mechanisms that are normally induced by TGF-β. Simultaneously, the downregulation of TβR expression and low level of TGF-β signaling may act as a positive feedback mechanism to induce the reflexive stimulation of TGF-β secretion in CaP. These potential feedback loops could explain the reduced expression of TβRs and large amounts of TGF-β that have been observed in advanced CaP. Our in vivo xenograft model also demonstrated that inhibition of DNMTs correlated to a lower tumor weight and cancer proliferation rate. These results suggest that the expression of DNMTs is associated with aggressive malignant phenotypes, tumor growth, and progression in vivo. In combination with our previous findings 
, we found that DNMTs is an important factor and predictor related to CaP progression.
Furthermore, the close correlation between TGF-β, ERK and DNMTs in tissue microarray specimens indicates that this cascade of signal events is likely not only associated with aggressive malignant phenotypes in vitro, but may also be involved with progression of CaP in humans. Based on our results, during progression of prostate cancer, an attenuation of expression of TGF-β receptors facilitates tumor cells escaping from the growth inhibition by TGF-β which is Smad dependent. Meanwhile, the Smad-independent pathway, such as p-ERK and DNMT signaling could be induced by TGF-β and results in the more aggressive phenotypes.
Our data shows that increased expression of DNMTs is highly correlated with both the expression levels of TGF-β1 and p-ERK. Furthermore, there was a significant correlation between the levels of DNMTs and Gleason grade, which further supports our findings that DNMTs are associated with more invasive CaP phenotypes. This finding is similar to recent reports suggesting that DNMT1 is associated with lung cancer progression 
. The present results demonstrate that DNMT1 is associated with biochemical recurrence in CaP patients seven years following radical prostatectomy. Thus, patients with higher tissue expression levels of DNMT1 are at increased risk for biochemical recurrence compared to those with lower tissue expression levels. The relationship between DNMT1 expression and biochemical recurrence is independent of Gleason grade. Although other variables including TGF-β1, p-ERK, Gleason grade were also showed significantly associated with biochemical recurrence, the final Cox Proportional Hazards Model demonstrated that DNMT1, in combination with pathologic Gleason sum, are stronger predictors for disease outcome. The exact mechanism of this observation remains unclear, but variables involved in the signal pathway including tumor expression of DNMT1, TGF-β1, and p-ERK may be useful in predicting clinical outcome following radical prostatectomy. High expression level of DNMT1 was risk factors for biochemical recurrence in men with CaP, regardless of Gleason's score.
In summary, our findings indicate that DNMTs expression levels are correlated with invasive capabilities in cultured human CaP cell lines. Additionally, we found that tumor-derived TGF-β and ERK are involved in the regulation of DNMTs in these cell lines. Inhibition of TGF-β in vivo results in the corresponding inhibition of DNMTs, and appears to significantly decrease tumor growth. In addition, we confirmed that the expression levels of TGF-β, ERK and DNMTs in tissue specimens obtained at the time of prostatectomy mimicked our findings in cell culture. Finally, we found that high expression levels of DNMT1 may potentially be used to predict biochemical recurrence in patients following radical prostatectomy.