Dynamic epigenetic changes to VHL occur with sunitinib in metastatic clear cell renal cancer

Abstract

Background: Genetic intratumoral heterogeneity (ITH) hinders biomarker development in metastatic clear cell renal cancer (mccRCC). Epigenetic relative to genetic ITH or the presence of consistent epigenetic changes following targeted therapy in mccRCC have not been evaluated. The aim of this study was to determine methylome/genetic ITH and to evaluate specific epigenetic and genetic changes associated with sunitinib therapy.

Patients and methods: Multi-region DNA sampling performed on sequential frozen pairs of primary tumor tissue from 14 metastatic ccRCC patients, in the Upfront Sunitinib (SU011248) Therapy Followed by Surgery in Patients with Metastatic Renal Cancer: a Pilot Phase II Study (SuMR; ClinicalTrials.gov identifier: NCT01024205), at presentation (biopsy) and after 3-cycles of 50mg sunitinib (nephrectomy). Untreated biopsy and nephrectomy samples before and after renal artery ligation were controls. Ion Proton sequencing of 48 key ccRCC genes, and MethylCap-seq DNA methylation analysis was performed, data was analysed using the statistical computing environment R.

Results: Unsupervised hierarchical clustering revealed complete methylome clustering of biopsy and three nephrectomy samples for each patient (14/14 patients). For mutational status, untreated biopsy and all treated nephrectomy samples clustered together in 8/13 (61.5%) patients. The only methylation target significantly altered following sunitinib therapy was VHL promoter region 7896829 which was hypermethylated with treatment (FDR=0.077, P<0.001) and consistent for all patients (pre-treatment 50% patients had VHL mutations, 14% patients VHL hypermethylation). Renal artery ligation did not affect this result. No significant differences in driver or private mutation count was found with sunitinib treatment.

Conclusions: Demonstration of relative methylome homogeneity and consistent VHL hypermethylation, after sunitinib, may overcome the hurdle of ITH present at other molecular levels for biomarker research.

Keywords: VHL; heterogeneity; methylation; mutations; renal cancer.

Conflict of interest statement

Thomas Powles has participated on advisory boards for Pfizer and GSK for he has they received financial compensation. All remaining authors have declared no conflicts of interest.

Figures

Figure 1. Hierarchical clustering dendrograms of methylation and mutational data

a. Unsupervised hierarchical clustering of patient sample mutations. 8/13 (61.5%) patient biopsy and nephrectomy samples clustered completely and 4/13 (30.8%) clustered partly together. Supplementary Figure 1 shows the mutational heatmap. b. Hierarchical clustering of DNA methylation data. The analysis was performed on 14 matched pairs of untreated (biopsy) and treated (nephrectomy tissue). The 1,000 loci featured by the largest variance (after quantile normalization and log transformation) were used for clustering, employing complete clustering based upon Euclidean distance. For all 14 patients their biopsy and nephrectomy samples were found to cluster. Figure amended from (12) with permission.

Figure 2. Methylation differences for targets following sunitinib treatment

a. Comparison of biopsy and nephrectomy for all patients. Target label displayed in each subplot. False discovery rate (FDR) is provided in parenthesis. NA = no methylation core was present, either because the target's regions were filtered due to low average counts, or because no methylation cores were present for the target in the methylome map. If there was more than one region for a certain target, the Figure only shows the most significantly differential region according to P-value. VHL is the only target that has FDR under the 0.1 significance level (i.e. 0.077). The P-value is 0.00086 and the logFC -0.8734. The latter implies that the post-treatment samples are more methylated in average than the pre-treatment ones. This is only the case for the methylation core in the VHL promoter region 7896829 located from nt 10183068 to nt 10183220 on chromosome 3; other VHL regions are not found to be differentially methylated under this significance level. b. Per patient methylation of VHL at region 7896829. For all samples methylation was greater in the post-treatment nephrectomy samples than the pre-treatment biopsy. Results divided into patients who had a good or poor response to treatment, there was no significant difference in the VHL hypermethylation seen in patients with a good vs poor response to sunitinib (P = 0.896, Student's t-test).

Figure 3. Driver mutation comparison between biopsy and nephrectomy samples

a. Mean number of SNV/indel candidate driver mutations per gene across all biopsy (15) and nephrectomy (44) samples. Some genes have multiple candidate driver mutations in some samples. Putative passenger somatic mutations are not included. There were no significant differences in mutation count between biopsy and nephrectomy samples (two-sided Wilcoxon rank sum test, P≥0.05 for all genes). b. Dot plot of private mutation frequency in biopsy and nephrectomy samples. Median value indicated. The number of mutations was greater in the biopsy sample for 7 patients, nephrectomy in 4 samples and equal between biopsy and nephrectomy in 2 samples. There was no significant difference in the number of private mutations in the biopsy samples compared with the median number of private mutations in the nephrectomy samples (P = 0.2, unpaired t-test).