Tumor suppressor gene inactivation during cadmium-induced malignant transformation of human prostate cells correlates with overexpression of de novo DNA methyltransferase

Abstract

Background: Aberrant DNA methylation is common in carcinogenesis. The typical pattern appears to involve reduced expression of maintenance DNA methyltransferase, DNMT1, inducing genomic hypomethylation, whereas increased expression of de novo DNMT3a or 3b causes gene-specific hypermethylation.

Objectives: During cadmium-induced malignant transformation, an unusual pattern of genomic hypermethylation occurred that we studied to provide insight into the roles of specific DNMTs in oncogenesis.

Methods: Gene expression and DNA methylation were assessed in control and chronic cadmium-transformed prostate epithelial cells (CTPE) using reverse transcription-polymerase chain reaction (RT-PCR), Western blot analysis, methylation-specific PCR, and methyl acceptance assay.

Results: During the 10-weeks of cadmium exposure that induced malignant transformation, progressive increases in generalized DNMT enzymatic activity occurred that were associated with over-expression of DNMT3b without changes in DNMT1 expression. Increased DNMT3b expression preceded increased DNMT enzymatic activity. Procainamide, a specific DNMT1 inhibitor, reversed cadmium-induced genomic DNA hypermethylation. Reduced expression of the tumor suppressor genes, RASSF1A and p16, began about the time DNMT3b overexpression first occurred and progressively decreased thereafter. RASSF1A and p16 promoter regions were heavily methylated in CTPE cells, indicating silencing by hypermethylation, while the DNA demethylating agent, 5-aza-2'-deoxycytidine, reversed this silencing. DNMT1 inhibition only modestly increased RASSF1A and p16 expression in CTPE cells and did not completely reverse silencing.

Conclusions: These data indicate that DNMT3b overexpression can result in generalized DNA hypermethylation and gene silencing but that DNMT1 is required to maintain these effects. The pattern of genomic DNA hypermethylation together with up-regulation of DNMT3b may provide a unique set of biomarkers to specifically identify cadmium-induced human prostate cancers.

Keywords: DNA methylation; DNMT3b; RASSF1A; cadmium; carcinogenesis; p16; prostate.

Figure 1

Global DNA methylation status and DNA methyltransferase activity during cadmium transformation of human prostate epithelial cells. Cells were grown in the presence or absence of 10 μM of cadmium for up to 10 weeks. (A) DNA methylation in control and exposed cells was determined by the in vitro methyl acceptance capacity of DNA using [³H-methyl]SAM as a methyl donor and a prokaryotic CpG DNA methyltransferase as described in “Materials and Methods.” Arrows indicate the ability of cells to form tumors upon inoculation into Nude mice according our prior work (Achanzar et al. 2001). (B) DNA methyltransferase activity was assayed as described in “Materials and Methods.” Data represent mean ± SE (n = 3). *Significantly different from control.

Figure 2

DNMT isoform transcript levels during cadmium exposure in CTPE cells. Transcript levels of specific forms of DNMT were assessed in control and CTPE cells exposed for 10 weeks to cadmium. RNA was isolated and subjected to RT-PCR analysis using a set of primers designed to amplify DNMT1, DNMT3a, DNMT3b, and β-actin transcripts. PCR products were separated on a 1.7% agarose gel. Densitometric data are given as percent of control (A) and normalized to β-actin (B) and expressed as mean ± SE (n = 3). *Significantly different from control.

Figure 3

Analysis of the effect of chronic cadmium exposure of prostate epithelial cells on DNMT3b and DNMT1 mRNA expression. Cells were grown in the presence or absence of 10 μM cadmium for up to 10 weeks. RNA was isolated and subjected to real-time RT-PCR analysis using a set of primers designed to amplify DNMT3b, DNMT1, and 18S gene products. PCR products were separated on a 1.7% agarose gel. Densitometric data are given as percent of control and expressed as mean ± SE (n = 3). *Significantly different from control.

Figure 4

Expression and promoter region methylation status of p16 and RASSF1A during chronic cadmium exposure. Cells were grown in the presence or absence of 10 μM cadmium for up to 10 weeks. (A) Proteins were isolated and subjected to Western blot analysis using antibodies against p16, RASSF1A, and β-actin. Blots were analyzed densitometrically, normalized to β-actin, and expressed as percent of control. (B) Methylation of the p16 and RASSF1A promoter regions determined by MSP. The presence of visible PCR product in lanes marked “U” indicates the unmethylated genes, whereas the presence of product in lanes marked “M” indicates the methylated gene. The source of DNA is indicated above each lane. CpGENOME universal methylated DNA (CHEMICON International, Inc., Temecula, CA) was used as a positive control.

Figure 5

Effect of a demethylating agent 5-aza-2′-deoxycytydine on p16 and RASSF1A protein expression. (A) CTPE cells were grown in the presence or absence of 0.5 or 1 μM of 5-aza-2′-deoxycytydine for 6 days. (B) CTPE cells were grown in presence or absence of 1 μM of 5-aza-2′-deoxycytydine over 96 hr. Proteins were isolated and subjected to Western blot analysis using antibodies against p16, RASSF1A and β-actin. Blots were analyzed densitometrically, normalized to β-actin, and expressed as percent of control. Data represent mean ± SE (n = 3). *Significantly different from control.

Figure 6

DNA methylation status, p16 and RASSFA1 protein expression after exposure to procainamide in CTPE cells. CTPE cells were grown in the presence or absence of 0.5 mM procainamide over 96 hr. (A) DNA methylation in control and CTPE cells was determined by the in vitro methyl acceptance capacity of DNA using [³H-methyl]SAM as a methyl donor and a prokaryotic CpG DNA methyltransferase as described in “Materials and Methods.” Proteins were isolated and subjected to Western blot analysis using antibodies against (B) RASSF1A and p16, and β-actin. Blots were analyzed densitometrically and normalized to β-actin. Data represent mean ± SE (n = 3). *Significantly different from control.