43 Epigenomic profiling of translocation renal cell carcinoma via liquid biopsy

Abstract Background Translocation renal cell carcinoma (tRCC) is an aggressive subtype of kidney cancer usually driven by a fusion involving the TFE3 gene. Due to histologic overlap with other subtypes of RCC, tRCC is frequently misclassified. Methods for accurate diagnosis and detection of this molecularly distinct entity are therefore a pressing need. Epigenomic profiling of ctDNA via plasma chromatin immunoprecipitation and sequencing (ChIP-seq) has recently emerged as a powerful tool to detect and molecularly subtype cancers and may offer a more sensitive and specific detection of molecular fusions. We aimed to detect tRCC in plasma and to discriminate tRCC from ccRCC based on epigenomic profiling of cfDNA. Methods We first identified differentially expressed gene, methylated regions (DMRs) and regulatory elements (REs) specific to tRCC vs. ccRCC via RNA-seq, methylated DNA immunoprecipitation sequencing (MeDIP-seq) and ChIP-Seq, of 4 tRCC and 5 ccRCC cell lines. We collected 16 plasma samples from metastatic patients with tRCC, 11 with ccRCC and 9 healthy patients (HP). Ultra low pass whole genome sequencing (ulpWGS) was performed to infer ctDNA fraction (TF), and cfMeDIP-seq, H3K4me3/H3K27ac cfChIP-seq for epigenomic profiling. Signal at tRCC-specific regions derived from cell lines profiling was aggregated for each mark and normalized to common active REs and DMRs, then compared between classes using a Wilcoxon rank-sum test. Classification performance was evaluated using the area under the receiver operating characteristic (AUROC) curve. Results Overall 8/16 tRCC and 5/12 ccRCC samples had >3% TF by ulpWGS. H3K4me3 and H3K27ac cfChIP-seq signal was significantly higher in all tRCC samples compared to healthy patients (p<10-6, AUROC =1), at tRCC-specific promoters and tRCC-specific REs respectively. Furthermore, H3K27ac signal in plasma was also significantly higher at tRCC-specific REs in tRCC samples compared to ccRCC (p<10-4, AUROC=0.95). MeDIP-seq could not discriminate between tRCC and ccRCC. Conclusions Although a majority of tRCC plasma samples profiled had TF < 3%, all could be distinguished from healthy samples on the basis of cfChIP. H3K27ac cfChIP-Seq was also discriminatory for tRCC vs. ccRCC. Epigenomic profiling of cfDNA appears as a powerful tool for both detection of tRCC and discrimination from ccRCC/healthy plasma, with potential implications for diagnosis and guiding therapy.