Background:
Malignant pleural mesothelioma (MPM) is a rare cancer of the pleura, currently without curative treatment, and tumour neoantigens are of interest in the development of targeted immunotherapy for its treatment. Neoantigens derived from frameshift mutations are likely to generate longer aberrant sequences than those of point mutations, and thereby are potentially more immunogenic against tumour cells. We hypothesized that, although the point mutation frequency is relatively low in mesothelioma compared with other tumours, immunogenic neoantigens may arise from frame-shift mutations. Our aim was to identify putative neoantigens arising from somatic frameshift mutations, and predict their binding affinity to HLA phenotypes.
Method:
Frameshift aberrations were identified from whole-genome sequencing (WGS) of tumour biopsies (80x) and peripheral blood (30x) in three MPM cases (MM1, MM2, MM3). Putative neoantigen sequences were extracted by obtaining wildtype coding sequence from RefSeq, inserting the variant of interestand generating new amino acid sequence using EMBOSS TranSeq. HLA phenotypes of each case were predicted using OptiType. Binding affinities of each neoantigen sequence against corresponding HLA phenotypes were evaluated using netMHCpan version 3.0 and 4.0, setting the threshold for strong binders to 0.5%, and weak binders to 2.0%.
Results:
HLA–A, –B and –C serotypes were predicted for both HLA alleles of each tumour. Twenty putative neoantigen sequences were identified from SNVs and fusion events in the following genes – MM1: NF2, TCF7L2, HSP90AB1, WBSCR17; MM2: NF2, SESN3, CRMP1; MM3: NF2, LATS2, SUMF2, ME1, KLHL1, ASIC2, COLQ, TP53BP1, ITGA1, RFX6. netMHCpan server predicted strong- and weak-HLA binding sequences from the candidates. Minor variations in the predictions were found between versions 3.0 and 4.0 of netMHCpan, resulting in changes to predicted rank scores and classifications of strong- or weak-binding candidates.
Conclusion:
WGS and in silico analysis identified several putative neoantigens in MPM, and revealed candidate sequences that may be immunogenic against tumour cells. These candidates will be assessed further using other in silico approaches to determine their binding properties to HLA phenotypes, and by experimental methods.