Specific Nucleotide Modifications Induce Neopeptides with High Major Histocompatibility Complex Binding in Myeloid Malignancies

Background: Single Base Substitutions (SBS) are linked to mutagenic endogenous and exogenous processes in human DNA. Substantial number of signatures [Wellcome Sanger Institute, COSMIC database] have already been uncovered. Smoking and aging clock-like signatures are characterized by C>A and C>T substitutions, respectively. The clonal evolutionary process to explain myeloid clonal expansion under “selective immune” pressure is not fully appreciated. However, pathogenic variants evoke T-cell clonality, suggesting that mutagenesis may represent an “escape” mechanism with genomic consequences. Our team hypothesized that SBS originates “neopeptides” with distinct binding abilities to the class 1 Major Histocompatibility Complex [MHC-1]. Here, we investigate SBS ability to generate different peptide bindings to MHC-1.Methods: After IRB approval, 48 patients (pt) were evaluated, with 37/48 (77%) and 9/48 (23%) Acute Myelogenous Leukemia (MDS) and Myelodysplastic Syndrome (MDS), respectively. After extracting the pathogenic variant protein fasta coding from National Center for Biotechnology Information (NBCI) web site, strength of “neopeptide” binding to MHC-1 was calculated by blasting mutated protein in NetMHCPan4.1. MHC-1 peptide binding score, defined as %Rank, below and above 0.5 were considered as strong [superior magnitude for neopeptide MHC binding] and weak, respectively. t-test was used to detect differential MHC-1 binding abilities for individual SBS. Descriptive statistics and analysis were performed with SAS software. Results: Mean age (range) was 69 years (y) (25-92). From the 48 cases, 267 SBS were extracted. These pathogenic variants resulted in 3473 neopeptides, and respective %Rank. To build the mean % Rank linked to individual SBS, only the mutated peptides were aggregated for calculation. From all SBS, mean %Rank was 0.79 (0-1.99). The SBS were C>T, 59/267 (22%); C>A, 15/267 (5.6%); G>C, 15/267 (5.6%); C>G, 15/267 (5.6%); T>C, 11/267 (4.1%); G>A, 57/264 (21.3%); A>C, 6/267 (2.2%); A>T, 11/267 (4.1%); G>T, 18/267 (6.7%); A>G, 23/267 (8.6%); T>G, 9/267 (3.3%); T>A, 12/267 (4.4%). Top scoring SBS with strong binding were C>T with mean %Rank 0.54 vs 0.79, p=0.0016 in pt with and without C>T, and A>C with mean %Rank 0.15 v 0.75, p= 0.008. We previously reported that G>A is highly prevalent in older AML pt [Patel. ASH.2022]. Interestingly, G>A mean %Rank was anticorrelated with MHC-1 binding [0.90 v 0.69, p=0.009]. In C>T SBS, the median number of peptide binding to MHC-1 was 3.7 v 2.7, p=0.01 and for A>C, 3.0 v 1.0, p=0.001. When accounting %Rank as a binary outcome [Strong MHC binding vs weak], age <>60y was equally associated with strong MHC binding, p=0.007 and p=0.008, respectively. In C>T SBS, MDS-like and Epigenetic (Epi) mutations [DNMT3A, TET2, IDH1] were observed in 13.1% and 31%, respectively. In A>C, RAS and MDS-like mutations were seen in 28.4% and 14.1%, respectively. Conclusions: Mutations retaining C>T and A>C modifications are the most prevalent in our hemopoietic malignancies cohort with strong binding. While C>T modification is linked to aging clock-like signature, we observed the SBS widely distributed among pt <>60y. It is possible that high cytidine deaminase activity, previously associated with C>T mutagenesis, induces neopeptides with high age independent MHC-1 binding efficiency.

No relevant conflicts of interest to declare.