Transcriptomic splicing analysis reveals a gene-independent form of TE reactivation in tumors representing unrecognized source of oncogenes and neoantigens

Abstract

Transposable elements (TEs) are reactivated in tumors and serve as significant contributors to tumor transcriptomic complexity and heterogeneity. However, their repetitive nature and diverse transcriptional forms have hindered efforts to characterize TE-derived transcripts independent of host gene contexts. Here, using our self-developed splicing-junction analysis tool ASJA, we systematically interrogated TE transcription across 32 cancer types, identifying transcripts autonomously initiated from and exclusively spliced among TEs with high TE content, termed TEtrans. Our analysis revealed 5,361 TEtrans exhibiting pan-cancer prevalence, tumor-specific expression and features of mature transcripts, including canonical splicing, 5 caps and polyA tails. Mostly unannotated and enriched in intergenic regions, TEtrans are predominantly derived from primate-specific classes with conserved splice pattern. TEtrans burden demonstrates heterogeneous associations with prognosis and immune activity across cancers, while their expression can be epigenetically dual-regulated by stemness- and inflammation-associated transcription factors. Functional studies uncovered that TEtrans could act as oncogenes, exemplified by tsTE1, a HERVH-derived transcript that promotes colorectal cancer proliferation by enhancing TOP1-mediated DNA supercoil relaxation. Remarkably, [~]7.2% of TEtrans are identified to encode tumor neoantigens, including viral proteins and unannotated peptides, which are shared among patients and validated by proteogenomic analysis. These TEtrans-derived neoantigens are immunogenic both in vitro and in vivo, and exceed neoepitopes from genomic alterations in abundance per tumor, particularly in cancers with low mutational burden. Collectively, as a gene-independent form of TE-derived transcripts, TEtrans represents a unique source of oncogenes and tumor neoantigens, expanding the functional repertoire of TEs in cancer biology and offering new avenues for therapeutic targeting. HighlightsO_LITEtrans are tumor-specific transcripts with conserved splicing patterns, predominantly derived from primate TEs in intergenic regions. C_LIO_LITEtrans expression is epigenetically activated and dually regulated by stemness/inflammation transcription factors. C_LIO_LITEtrans function as oncogenes, exemplified by tsTE1 enhancing TOP1-mediated DNA relaxation. C_LIO_LITEtrans encode tumor neoantigens that elicit CD8+ T cell responses, providing targets for immunotherapy, especially in low-mutation tumors. C_LI Abstract figure O_FIG O_LINKSMALLFIG WIDTH=198 HEIGHT=200 SRC="FIGDIR/small/640928v1_ufig1.gif" ALT="Figure 1"> View larger version (78K): org.highwire.dtl.DTLVardef@1d2d4cforg.highwire.dtl.DTLVardef@918f14org.highwire.dtl.DTLVardef@489465org.highwire.dtl.DTLVardef@1069d17_HPS_FORMAT_FIGEXP M_FIG C_FIG