Pilot: Development of a k-mer based Bioinformatics Method for Characterizing the Pan- Cancer Endogenous Retroviruses

ABSTRACT: Cancer immunotherapy is a monumental breakthrough for cancer treatment and has revolutionized the field of oncology. The core of immunotherapy relies on the recognition of tumor-specific antigens (TSA) by T cells to elicit adaptive T cell immune response against tumor. Despite the durable anti-tumor responses observed in several malignancies, the demands for a high tumor mutational burden (TMB), the lack of efficacious tumor antigens and the inability to predict immunotherapy responses have impedes the widespread application of immunotherapy. Intriguingly, substantial studies have been suggesting HERVs as an important source of TSA to be harnessed by immunotherapy and effective biomarkers for clinical prognosis and tumor response to immunotherapy. HEVR genes in healthy tissues are generally silenced, but epigenetic dysregulation of the cancer genome gives rise to the aberrant expression of HERV-containing genomic regions in many cancers. These tumor-specific HERVs can provoke innate anti-tumor immune response via induction of viral defense pathways. Additionally, they can also contribute to an adaptive immune response through the production of tumor-associated T cell epitopes, which can increase tumor cell visibility to immune surveillance, resulting in cytotoxic T cell responses. However, almost all the studies rely on the conventional bulk tissue sequencing, e.g., bulk tumor RNA-Seq, for HERV expression profiling, overlooking the distinct tumor and immune cell populations in the complex tumor microenvironment (TME). Most importantly, the single-cell RNA-Seq (scRNA-Seq) has emerged as powerful tools for the dissection of the TME, nevertheless, HERV landscape study at single-cell resolution remains limited due to the lack of specific bioinformatics tools. The novelty of this proposal is to develop an efficient bioinformatics method to accurately survey the HERV landscape using the wealth availability of Pan-Cancer scRNA-Seq datasets. We will leverage novel algorithms in data compression and k-mer indexing to surmount the extreme needs for computation and storage resources in processing large-scale scRNA-seq datasets. This pilot award funded by the American cancer Society will allow the PI to generate critical preliminary data to support a U01 (RFA-CA-23-015) and K01 (PAR-21- 296) in October of 2024. The outcome of this project will provide more precise and higher-resolution approaches to deconvolute the cancer-specific and cell type-specific HERV expression, facilitate the elucidation of HERV-modulated immune response and accelerate development of HERV-focused immunotherapies. Institutional Research Grant American Cancer Society – September 2023