Introduction

Studies of genetic and DNA methylation changes in urothelial carcinoma of the bladder (BC) have been largely in muscle invasive disease. Herein we studied the genetic and epigenetic changes in tumors of patients with recurrent, non-muscle invasive (NMI) BC to identify clonal composition.

Methods

30 tumors from 13 patients with recurrent NMIBC were examined from an IRB approved database. The median follow-up was 26 months (range=6-180). Whole-exome sequencing (WES) was used to identify mutated genes; 450K DNA methylation array was used to detect genome-wide DNA methylation alterations. Metachronous tumors were analyzed for a common ancestral clone. Mutations seen in 2 or more tumors in a given patient were defined as public.

Results

13 patients (100%) showed clonal evolution from an ancestral clone with a mean of 38% (0.3-68%) of the total mutations per patient in the public mutations branch (Figure 1). 30 genes were publically mutated in at least 2 patients and present in all tumors for a given patient including MLL2 (54%), MUC16 (38%), ATN1 (23%), HRNR (23%), SRCAP (23%), TP53 (23%), and WNK1 (23%). Public DNA hypermethylation (36% of alterations per patient) and hypomethylation (29% of alterations per patient) occurred more frequently than genetic mutations, and occurred independent of genetic mutations (see Figure 2). 64 genes showed promoter hypermethylation in ≥60% of tumors (18); pathway analysis revealed cadherin signaling pathway members PCDH8 (83%), PCDHAC1 (80%), PDCHB7 *80%), PCDHB15 (77%) and PCDHB4 (73%), implicating the potential silencing of multiple genes in this pathway. FASSF1 hypermethylation was seen in 66% of tumors. 43 genes with hypomethylated promoters occurring in ≥60% of tumors were identified without specific associated pathways.

Conclusions

Recurrent NMIBC tumors arise from an ancestral clone. Public epigenetic mutations are more common than public genetic mutations in this setting, suggesting that both genetic and epigenetic alterations are involved in tumor initiation and are potential targets for therapies.

Funding

none