Investigating the sequence grammar regulating antibody somatic hypermutation

• Grant number: 23/18320-3
• Support Opportunities: Regular Research Grants
• Start date: November 01, 2024
• End date: October 31, 2027
• Field of knowledge: Biological Sciences - Immunology
• Principal Investigator: Israel Tojal da Silva
• Grantee: Israel Tojal da Silva
• Principal researcher abroad: Rushad Rusi PAVRI
• Institution abroad: King's College London, England
• Host Institution: A C Camargo Cancer Center. Fundação Antonio Prudente (FAP). São Paulo , SP, Brazil

Abstract

Background: Somatic hypermutation (SHM) of immunoglobulin variable (V) regions in B cells is the molecular basis for the diversification of antibodies in response to pathogens and vaccines and, therefore, for robust long-term humoral immunity. Mutations are generated by activation-induced cytidine deaminase (AID) which converts cytosines to uracils preferentially at WRCH motifs (where W = A or T, R = A or G and H = A, C or T). Moreover, off-target AID activity at proto-oncogenes has been implicated in driving B cell lymphomagenesis, making SHM a major contributor to genome instability.Importance and state-of-the-art: Multiple studies in recent years have shown that mutation frequencies vary substantially between WRCH motifs with large variation even between identical motifs in different locations. This strongly implies that AID targeting is regulated by the sequence context of WRCH motifs. Therefore, understanding the mechanism by which sequence context regulates mutability is a central topic of research in SHM biology. In support, it was recently reported that mutation frequencies at weakly mutated AGCT motifs could be enhanced by introducing a flexible, pyrimidine-rich sequence upstream of the motif. However, our bioinformatics analyses of public human V gene datasets and experiments in the SHM model human B cell line, Ramos, show that many highly mutated motifs are not embedded in pyrimidine-rich sequences and suggest that sequence contexts favoring SHM differ between WRCH motifs.Specific aim: Learning the sequence grammar (rules) governing the differential mutability of WRCH motifs by combining artificial intelligence (AI)-based predictions with experimental validation.Rationale and approach: Our preliminary findings suggest that DNA context is necessary but not sufficient for conferring differential mutability, and that differential mutability arises via a complex interplay of multiple genomic features. Hence, we propose to build an AI-based supervised machine learning framework trained on large public datasets of non-productive human and mouse V gene sequences for comprehensive integration of multiple genomic features. Importantly, we will use Explainable AI (XAI) to understand which features contribute positively or negatively to the prediction, thereby enabling us to decipher which features may be responsible for mutagenesis in different sequence contexts across all WRCH motifs. The same AI-based analyses will be performed on whole-genome-sequencing datasets of B cell lymphoma samples to determine whether similar or different rules apply for mutagenesis at AID off-target loci. Importantly, the predictions of these models will be tested in our optimized a Ramos B cell system where we can express any V region, and variants thereof, endogenously from the immunoglobulin heavy chain promoter, coupled with a customized bioinformatics pipeline for mutational analyses. In parallel, we will test several hypotheses, based on our preliminary data, in Ramos cells. These include investigating whether there exist natural sequence contexts specific for certain WRCH motifs, the importance of pyrimidine-richness in some WRCH contexts and the possible effect of motif position on mutagenesis.Outcomes and impact: By unraveling new insights into the sequence grammar regulating mutagenesis, we expect these studies to substantially advance our knowledge of the molecular mechanism of SHM and B cell genome instability. (AU)