Identification of potential biomarkers with alternative splicing landscape of HIPEC-treated ovarian cancer: An RNAseq sub-analysis of a HIPEC clinical trial.

person Thanh Hue Dellinger City of Hope, Duarte, CA info_outline Thanh Hue Dellinger, Nathaniel P. Hansen, Nora H. Ruel, Paul Henry Frankel, Susan Elaine Yost, Marta Invernizzi, Melissa Eng, Ernest Soyoung Han, Mustafa Raoof, Xiwei Wu, Hyejin Cho, Stephen Lee, Mehdi Kebria, Amy Hakim, Edward Wenge Wang, Isaac Benjamin Paz, Lorna Rodriguez-Rodriguez, Patrick Pirrotte

Authors person Thanh Hue Dellinger City of Hope, Duarte, CA info_outline Thanh Hue Dellinger, Nathaniel P. Hansen, Nora H. Ruel, Paul Henry Frankel, Susan Elaine Yost, Marta Invernizzi, Melissa Eng, Ernest Soyoung Han, Mustafa Raoof, Xiwei Wu, Hyejin Cho, Stephen Lee, Mehdi Kebria, Amy Hakim, Edward Wenge Wang, Isaac Benjamin Paz, Lorna Rodriguez-Rodriguez, Patrick Pirrotte Organizations City of Hope, Duarte, CA, Tgen, Phoenix, AZ, City of Hope Comprehensive Cancer Center, Duarte, CA, City of Hope National Comprehensive Cancer Center, Duarte, CA, City of Hope Cancer Center, Duarte, CA, City of Hope National Medical Center, Duarte, CA, City of Hope National Comprehensive Cancer Center, Madras, OR, TGen/City of Hope Comprehensive Cancer Center, Duarte, CA Abstract Disclosures Research Funding No funding sources reported Background: Cytoreductive surgery combined with hyperthermic intraperitoneal chemotherapy improves survival in ovarian cancer patients. Currently, no biomarkers exist to select for patients who best benefit from HIPEC. We compared the slicing landscape of ovarian cancer tumors of good and poor HIPEC responders in a Phase I clinical trial, to identify potential predictive biomarkers. Methods: A total of 35 ovarian cancer patients enrolled in a Phase I HIPEC trial (NCT01970722), in which nine patients had paired pre-and post treatment tissue samples collected. Pre-HIPEC tumor samples had RNA isolated, and whole-transcriptome library construction performed. FASTQ files from tumor samples were aligned to GRCH38 and ran through SplAdder (PMID:26873928) to identify alternative splicing events. Outlier splice events were ranked by mean log-likelihood score and converted to amino acid sequences using Bisbee (PMID:34031440). The top 15 outlier events were run through NetMHCpan (PMID:32406916) and DeepHLApan (PMID:31736974) to predict putative neoantigens. Results: Among nine HIPEC-treated ovarian cancer patients with available RNAseq data, five were considered good responders (≥12months PFS), and four as poor responders (<12 months PFS). Alternative splicing analysis identified 519splice events as outliers between good and poor responders, with 250reported to be novel. Of the 250 novel splicing events, we identified a novel protein coding 3’ end splice event in CPNE1 which was alternatively spliced in at a Percent-Spliced-In (PSI) of 99% in three out four poor responders. The same splice event was alternatively spliced at a PSI of 30.3% in four out of five good responders. CPNE1 plays a role in calcium mediated intracellular processes, and is involved in the TNF-alpha receptor signaling pathway, thus playing an important role in cell proliferation, differentiation, apoptosis, and modulation of immune responses and induction of inflammation. The encoded novel peptide from the CPNE1 splice event produced six predicted strong binding sites based on a binding score>0.80 and an immunogenic score>0.80. Conclusions: CPNE1 was identified as a differentially spliced event in poor HIPEC responders with ovarian cancer. Poor HIPEC responders had higher immunogenic score. Alternative splicing analysis may be a promising method to determine potential biomarkers for HIPEC treatment in ovarian cancer patients. Clinical trial information: NCT01970722.

Clinical