Long-read RNA sequencing redefines the clear cell renal cell carcinoma transcriptome and reveals novel genes and transcripts associated with disease recurrence and immune evasion

Abstract

AbstractBackgroundLong-read direct RNA sequencing (DRS) and PCR cDNA sequencing (PCS) of tumour samples could lead to discovery of novel transcript isoforms, novel genes, and transcriptomic co-dependencies missed by conventional short-read sequencing. However, only a handful of reports using DRS or PCS in cancer exist with no direct comparison between the two methods. Clear cell renal cell carcinoma (ccRCC) is the most common form of kidney cancer. Following primary tumour resection approximately 30% of patients experience disease recurrence. Long-read RNA sequencing has not been applied to kidney cancer.Methods12 primary ccRCC archival tumours (discovery cohort), 6 from patients who went on to relapse, were analysed by Oxford Nanopore DRS and PCS. Results were validated in an independent cohort of 20 patients and compared to DRS analysis of RCC4 cells.ResultsDRS and PCS were successfully performed achieving high read length, with PCS achieving higher sequencing depth. Differentially expressed gene sets in patients who went on to relapse were determined with good overlap between DRS and PCS. Deconvolution analysis showed a loss of immune infiltrate in primary tumours of patients who relapse and revealed the CD8+T cell exhaustion markerTOXas a novel recurrence-associated gene. Notably, novel transcript analysis revealed more than 10,000 uncharacterised candidate novel transcripts detected by both methods and in a ccRCC cell linein vitro. This allowed the definition of the full exonic structure of ccRCC-associated splice variants, including variants ofMVKandHPCAL1. Remarkably, this also revealed a novel sPD-L1transcript encoding for the soluble version of the protein with a longer 3’UTR and lower stability in ccRCC cells than the annotated transcript. Levels ofsPD-L1transcripts are unchanged in primary tumours that go on to relapse, whereas membranePD-L1shows a trend towards down-regulation. Finally, both methods identified 414 novel genes, also detected in ccRCC cellsin vitro, including a novel non-coding gene over-expressed in patients who relapse.ConclusionsPCS and DRS can be used in tumour samples to uncover substantial yet unmapped features underpinning the plasticity and instability of cancer transcriptomes which are linked to disease progression and immune evasion.