Introduction: Chronic lymphocytic leukaemia (CLL) is the most common haematologic malignancy and remains incurable. The mutational status of the immunoglobulin heavy chain variable region (IGHV) predicts outcome independently of genetic aberrations. Mutated-IGHV has better overall survival than unmutated-IGHV, however, the mechanism underpinning this is unknown. Neoantigens are immunogenic peptides, unique to malignant cells, that are presented to T-cells via HLA-I/II. Unlike other tumours, CLL cells are antigen presenting B-cells that express both HLA-I/II. Also, T-cell clonal expansions are well described in CLL. It is believed that CLL has a low tumour mutational burden (TMB), with few non-synonymous mutations (NSM) and therefore NSM-neoantigens. However, traditional concepts of TMB fail to recognise that immunoglobulin genes of the B-cell receptor (BCR) are also highly mutated in B-cell malignancies and can create ‘BCR-neoantigens’ (Khodadoust, Nature 2017). There remains no large-scale analysis of HLA-I/II NSM and BCR-neoantigens in CLL.
Methods: Pre-therapy PBMCs from 24 patients with CLL (M18:F6, median age 69 years, IGHV-mutated=12, IGHV-unmutated=12) underwent whole exome sequencing on germline (n=19) and tumour (n=24), and expression by RNA-sequencing on (n=22) tumour samples. NSM-neoantigens were called using pVACseq (TESLA consortium guidelines; Wells, Cell 2020). BCR-neoantigens were called from full-length BCR sequences with pVACbind.
Results: Initially the NSM-mutational landscape was characterized. WES revealed a median 63 NSM/case. However, mutations were expressed (determined by RNA-sequencing) in only 9 (14%) NSM/case. No differences between the number of expressed NSM-mutations in IGHV-mutated and IGHV-unmutated were observed. Mutations in the BCR were tested next. There were 9 BCR-mutations/case, but unlike NSM-mutations, interestingly the majority of BCR-mutations (55%) were expressed. There were 5-fold more BCR-mutations expressed in IGHV-mutated vs. IGHV-unmutated CLL (5 vs 1, p<0.01).
Neoantigens were then assessed from expressed mutations. Combining NSM and BCR-neoantigens, there were a median 31/case. Interestingly, the frequency of BCR-neoantigens was significantly higher than NSM-neoantigens (HLA-I: 6 vs 2, p<0.0005 and HLA-II: 21 vs 6, p<0.0005). There were ~3-4 fold-greater HLA-II than HLA-I BCR-neoantigens. Strikingly, there were ~2-3 fold-greater BCR-neoantigens in IGHV-mutated vs IGHV-unmutated CLL (HLA-I: 11 vs 2, p<0.001; HLA-II: 33 vs 15, p<0.001).
Conclusions: The reported low TMB in CLL does not reflect the true frequency of neoantigens as it fails to account for BCR-neoantigens. These contribute ~75% of the neoantigen pool. There were more HLA-II than HLA-I BCR-neoantigens. The increase of BCR-neoantigens in IGHV-mutated CLL implicates anti-tumoural surveillance by BCR-neoantigen targeting CD4+ T-cells is behind the superior outcomes observed in IGHV-mutated CLL.