KISS1 methylation and expression as tumor stratification biomarkers and clinical outcome prognosticators for bladder cancer patients

Abstract

KISS1 is a metastasis suppressor gene that is lost in several malignancies, including bladder cancer. We tested the epigenetic silencing hypothesis and evaluated the biological influence of KISS1 methylation on its expression and clinical relevance in bladder cancer. KISS1 hypermethylation was frequent in bladder cancer cells analyzed by methylation-specific PCR and bisulfite sequencing and was associated with low gene expression, being restored in vitro by demethylating azacytidine. Hypermethylation was also frequently observed in a large series of bladder tumors (83.1%, n = 804). KISS1 methylation was associated with increasing stage (P = 0.001) and tumor grade (P = 0.010). KISS1 methylation was associated with low KISS1 transcript expression by quantitative RT-PCR (P = 0.037). KISS1 transcript expression was also associated with histopathological tumor stage (P < 0.0005). Low transcript expression alone (P = 0.003) or combined with methylation (P = 0.019) was associated with poor disease-specific survival (n = 205). KISS1 transcript expression remained an independent prognosticator in multivariate analyses (P = 0.017). KISS1 hypermethylation was identified in bladder cancer, providing a potential mechanistic explanation (epigenetic silencing) for the observed loss of KISS1 in uroepithelial malignancies. Associations of KISS1 methylation and its expression with histopathological variables and poor survival suggest the utility of incorporating KISS1 measurement using paraffin-embedded material for tumor stratification and clinical outcome prognosis of patients with uroepithelial neoplasias.

Figure 1

KISS1 was frequently methylated in bladder cancer cells. A: Schematic depiction of the KISS1 CpG islands analyzed as nucleotide sequences of the CpG island region analyzed by BS-SEQ. The sequencing primers are indicated by italics, the primers used for MS-PCR are indicated by underlining, and the CpG islands are numbered and indicated in gray. B: Analysis of CpG island methylation status of KISS1 by bisulfite genomic sequencing in bladder human cancer cell lines. Three individual clones are shown per methylated cell line and per unmethylated normal lymphocytes (NL) control. CpG dinucleotides are represented as dark squares for methylated cytosines and open squares for unmethylated cytosines. For dark squares, the presence of methylation was confirmed in at least two of the clones that were sequenced for the cell lines analyzed.

Figure 2

CpG island methylation was associated with KISS1 silencing. A: Methylation-specific PCRs (MS-PCR) for KISS1 in human bladder cancer-cell lines. A PCR band in lane M indicates a methylated KISS1 gene; in lane U, it indicates an unmethylated gene. Normal lymphocytes (NL) and MEC1 were used as controls for unmethylated KISS1; in vitro methylated DNA (IVD) was used as control for methylated KISS1.From bisulfite genomic sequencing (BS-Seq), methylated cell lines are highlighted in dark color. With RT-PCR, GAPDH expression was used as transcript loading control. B: Treatment with the demethylating agent reactivated the transcript expression of KISS1 in methylated bladder cancer cell lines. Qualitative RT-PCR analysis of KISS1 expression, with GAPDH as transcript loading control. After 1 or 5 μmol/L AZA exposure, KISS1 transcript expression increased in all of the methylated bladder cancer cell lines. C: Quantitative RT-PCR analysis of KISS1 expression. ΔC_TKISS1 values for each sample were determined by subtracting the C_T value of the target KISS1 gene from the value of the CCT6A control gene. KISS1 transcript expression increased in all of the methylated bladder cancer cell lines, after 1 or 5 μmol/L AZA exposure. MEC1, an unmethylated cell line, was used as a positive control for KISS1 expression.

Figure 3

KISS1 hypermethylation is a frequent bladder cancer-specific event. A:KISS1 hypermethylation was shown to be a cancer-specific event, as indicated by representative MS-PCR analyses for KISS1 using primary bladder tumors and normal urothelium. Normal lymphocytes (NL) and MEC1 were used as controls for unmethylated KISS1; in vitro methylated DNA (IVD) was used as control for methylated KISS1. B:KISS1 methylation increased in bladder tumors, compared with their paired normal urothelium, as indicated by methylation status of KISS1 analyzed by MS-PCR in pairs of primary bladder tumors and normal urothelium. C:KISS1 was unmethylated in normal urothelium unrelated to bladder cancer. Representative MS-PCR analyses are shown for KISS1 in normal urothelium (NU) from cystoprostatectomized patients with prostate cancer. MEC1 cells were used as positive control for unmethylated KISS1. D:KISS1 methylation correlated with a low transcript expression. Representative MS-PCR analyses are shown for KISS1 and matching transcript levels of KISS1 in bladder tumors.

Figure 4

Transcript expression and KISS1 methylation are associated with clinical outcome in bladder cancer. A: Loss of KISS1 expression alone was associated with disease-specific survival. Kaplan-Meier survival curve analysis indicated that low transcript expression of KISS1 with a 2^{−(ΔCt KISS1)} value of < 0.1 measured by RT-PCR was associated with poor survival (log rank, P = 0.003). B:KISS1 methylation alone was not significantly associated with disease-specific survival. Kaplan-Meier survival curve analysis indicated that tumors methylated for KISS1 had poorer survival than unmethylated tumors, but this trend did not reach statistical significance (log rank, P = 0.238). C: Loss of KISS1 transcript expression in combination with KISS1 methylation was associated with disease-specific survival. Kaplan-Meier survival curve analysis indicated that tumors with a transcript expression of KISS1 with a 2^{−(ΔCt KISS1)} value of <0.1 and methylated for KISS1 had the poorest survival (log rank, P = 0.019).