Silencing of claudin-11 is associated with increased invasiveness of gastric cancer cells

Abstract

Background: Claudins are membrane proteins that play critical roles in tight junction (TJ) formation and function. Members of the claudin gene family have been demonstrated to be aberrantly regulated, and to participate in the pathogenesis of various human cancers. In the present study, we report that claudin-11 (CLDN11) is silenced in gastric cancer via hypermethylation of its promoter region.

Methodology/principal findings: Levels of CLDN11 methylation and mRNA expression were measured in primary gastric cancer tissues, noncancerous gastric mucosae, and cell lines of gastric origin using quantitative methylation-specific PCR (qMSP) and quantitative reverse transcriptase-PCR (qRT-PCR), respectively. Analyses of paired gastric cancers and adjacent normal gastric tissues revealed hypermethylation of the CLDN11 promoter region in gastric cancers, and this hypermethylation was significantly correlated with downregulation of CLDN11 expression vs. normal tissues. The CLDN11 promoter region was also hypermethylated in all gastric cancer cell lines tested relative to immortalized normal gastric epithelial cells. Moreover, CLDN11 mRNA expression was inversely correlated with its methylation level. Treatment of CLDN11-nonexpressing gastric cancer cells with 5-aza-2'-deoxycytidine restored CLDN11 expression. Moreover, siRNA-mediated knockdown of CLDN11 expression in normal gastric epithelial cells increased their motility and invasiveness.

Conclusions/significance: These data suggest that hypermethylation of CLDN11, leading to downregulated expression, contributes to gastric carcinogenesis by increasing cellular motility and invasiveness. A further understanding of the mechanisms underlying the role of claudin proteins in gastric carcinogenesis will likely help in the identification of novel approaches for diagnosis and therapy of gastric cancer.

Figure 1. Promoter methylation and mRNA expression levels of claudin-11 (CLDN11) in clinical specimens.

This figure illustrates the promoter methylation and mRNA expression levels of CLDN11 in paired gastric cancer (GC) and noncancerous gastric mucosal (NS) tissues. A) Quantitative methylation-specific PCR (qMSP) for CLDN11 in primary tissues. Genomic DNA extracted from 18 GC and 18 matched NS tissues were subjected to qMSP analysis using MSP amplicon and TaqMan probe designed to include multiple CpG sites in the 5′-UTR region of CLDN11 gene. CpGenome Universal Methylated DNA (Chemicon International, Temecula, CA) was used as a fully methylated positive control. Duplex PCR with β-actin (ACTB) primer and probe sequences containing no CpGs were performed for normalization. Normalized methylation value (NMV) was defined as follows: NMV = (CLDN11-S/CLDN11-FM)/(ACTB-S/ACTB-FM) * 100, where CLDN11-S and CLDN11-FM represent CLDN11 methylation levels in the sample and fully methylated DNAs, respectively, while ACTB-S and ACTB-FM correspond to β-actin in the sample and fully methylated DNAs, respectively. This one-dimensional scatterplot demonstrates significantly high CLDN11 promoter methylation levels in the GC specimens when compared to the NS specimens (P<0.001). Whole-genome amplified DNA (WGA) used as an unmethylated negative control did not show any amplification. The P value was calculated using the paired Student's t-test. B) CLDN11 mRNA expression in gastric tissues. Total RNA extracted from 18 paired NS and GC specimens were subjected to CLDN11 specific RT-PCR. CLDN11 mRNA expression was normalized to β -actin mRNA expression in each sample. The P value was calculated using the paired Student's t-test. This plot demonstrates that the CLDN11 mRNA expression levels were significantly lower to non-detectable in the GC specimens, while most of the corresponding NS specimens had detectable CLDN11 mRNA levels (P<0.001). C) 2D-scatter plot of promoter methylation and mRNA expression values in GC tissues. CLDN11 mRNA expression silencing is associated with promoter hypermethylation in GC patients. This two-dimensional scatterplot demonstrates NEV values (Y-axis) and NMV values (X-axis) in the 18 paired NS and GC specimens.

Figure 2. Analysis of promoter methylation, mRNA and protein expression of claudin-11 in gastric cell lines.

This figure illustrates the claudin-11 promoter methylation, mRNA expression levels and protein expression in gastric cells lines. A) Quantitative methylation-specific PCR (qMSP) for CLDN11. Genomic DNAs isolated from immortalized human normal gastric epithelial cells (HFE145) and GC cell lines AGS, SIIA, MKN28, KATOIII, and SNU-1 obtained from ATCC were analyzed by qMSP. This Figure illustrates that the promoter region of CLDN11 gene is hypermethylated in all GC cell lines relative to HFE145 cells. B) CLDN11 mRNA expression in gastric cell lines. Total RNAs from different gastric cell lines were subjected to quantitative real-time RT-PCR analysis. As can be seen in this figure, HFE145 cells expressed very high levels of CLDN11 mRNA, while all five cancer cell lines tested had very low or undetectable CLDN11 mRNA expression. C) Western blot analysis of claudin-11 expression in gastric cell lines. Total cell lysates obtained from various gastric cell lines were probed with the anti-claudin-11 antibody. This figure illustrates that while the immortalized normal gastric epithelial cell line, HFE145, expressed abundant claudin-11 protein, it could not be detected in various GC cell lines. Anti-β-actin antibody was used as a loading control.

Figure 3. CLDN11 mRNA re-expression in AGS GC cells after 5-aza-2′-deoxycytidine (5-aza-dC) treatment.

AGS is a GC cell line manifesting hypermethylation of the CLDN11 promoter in conjunction with absent CLDN11 mRNA expression. These cells were treated with 5-aza-dC, a global demethylating agent. Total RNA from AGS cells before and after 5-aza-dC treatment were subjected to quantitative real-time RT-PCR analysis for CLDN11 mRNA expression. The Y-axis represents the average fold change in expression levels of CLDN11 mRNA after 5-aza-dC treatment at various time points, when compared with the untreated cells. AGS cells, when treated with 5-aza-dC, exhibited a time-dependent increase in CLDN11 mRNA expression up to 72 hrs of treatment. This restoration of CLDN11 expression after 5-aza-dC treatment supports our hypothesis that claudin-11 is silenced by promoter hypermethylation in GC cells.

Figure 4. Small interfering RNA knockdown experiments.

HFE145 cell line, which is claudin-11 positive was selected to study the role of claudin-11 knockdown on migration and invasion properties of gastric cells. Cells were transfected with CLDN11 specific siRNA oligos. Mock transfections and nonspecific siRNA duplexes were used as the negative controls. A) Western blot analysis of siRNA mediated claudin-11 knock-down in HFE145 cells. HFE145 cells were transfected with CLDN11- specific and non-specific control siRNA duplexes, as described in Materials and Methods. After 48 to 72 hours, total cell lysates were prepared and analyzed for claudin-11 expression. Transfection of claudin-specific siRNA oligos resulted in>90% reduction in expression of the protein, whereas the levels of claudin protein in the control cells were not significantly altered. B) Boyden chamber cell invasion assay. The invasiveness of the siRNA-transfected cells were determined using matrigel coated invasion chamber, using a modified Boyden chamber assay. Experiments were repeated at least three times, with triplicates in each experiment. The data represented here is the average fold change in invasion of the siRNA-transfected cells when compared with the mock transfections. As is demonstrated in this figure, siRNA knockdown of claudin-11 leads to an increase in cell invasion of HFE145 cells. C) Two-chamber cell migration assay. The effects of claudin-11 knockdown on migration of the HFE145 cells were compared by measuring the number of cells migrating through the uncoated filters (instead of matrigel-coated filters). The bars in this figure represent mean fold change in migration of siRNA-transfected cells compared with the mock-transfected control cells. As can be seen in this figure, a reduction in claudin-11 protein levels is associated with increased motility of the cells.