P1254: baseline hla-dr-restricted neoantigen burden of aml blasts impacts on hla-dr gene expression at relapse after allogeneic stem cell transplant

Topic: 21. Stem cell transplantation - Experimental Background: Relapse of acute myeloid leukaemia (AML) occurs in 20-50% patients following allogeneic haematopoietic stem cell transplant (AHSCT), with limited treatment options. It has been shown that relapse is associated with immune evasion from the graft-versus-leukaemia effect, with a significant number of cases downregulating MHC class II expression. Whilst AML typically has fewer mutations than many other cancers, several of these mutations have been shown to result in neoantigen presentation and may contribute to the immunogenicity of AML blasts. The link between initial neoantigen burden and post-AHSCT relapse is unclear. Aims: To utilize a multi-omic approach in combination with in silico methods for neoantigen prediction to assess the contribution of the potential neoantigen burden to AML relapse following AHSCT. Methods: The discovery cohort was made up of nine patients with AML treated on the same reduced intensity, T-replete transplant platform who all achieved full donor chimerism post-AHSCT. AML blasts were sorted from six patients with matched samples taken at diagnosis and post-AHSCT relapse using fluorescence activated cell sorting. Blasts were assessed for cell surface expression of antigen presenting molecules by flow cytometry, and for their ability to provoke an alloreactive immune response in co-culture with healthy donor peripheral blood mononuclear cell responders. Targeted mutation analysis was carried out using the Oncomine myeloid gene panel, and transcriptome analysis by RNAseq. Mutation calling was also carried out on RNAseq generated sequences providing a genome wide mutation analysis. Neoantigen prediction was carried out using NeoPredPipe for HLA-A, HLA-B, HLA-C and HLA-DR restricted neoantigens (Fig.1a). Results: Four of the six cases analysed were assessed for their capacity to stimulate an in vitro allogeneic T cell response at both diagnosis and relapse, and this was compared to the cell surface HLA-DR expression. There was a significant correlation between HLA-DR expression and both CD4 (r = 0.95, p = 0.048) and CD8 (r = 0.97, p = 0.033) T cell alloproliferation at diagnosis, an observation which was lost at relapse (Fig.1b). As expected, both MHC class I and class II gene expression was downregulated at post-AHSCT relapse. At diagnosis all patients had mutations in known AML driver genes, and all showed evidence of clonal evolution at post-AHSCT relapse with either loss, gain or both loss and gain of mutations. The total number of predicted neoantigens was not correlated with the total mutation burden either at diagnosis (r = 0.403, p = 0.428) or relapse (r = 0.375, p = 0.464) (Fig.1c). The majority of neoantigens were HLA-DR restricted at both diagnosis and relapse (Fig.1d). HLA-DR restricted neoantigens at diagnosis were shown to be significantly positively correlated with time to relapse (r = 0.91, p = 0.012) (Fig.1e) but were negatively correlated with the fold change in HLA-DR gene expression between diagnosis and relapse (r = -0.84, p = 0.037) (Fig.1f). Summary/Conclusion: These novel findings demonstrate that higher HLA-DR-restricted neoantigen burden at diagnosis is associated with a longer time to relapse post-AHSCT, consistent with a contribution to immune mediated graft-versus-leukemia effects. Furthermore, the increased downregulation of HLA-DR at relapse in patients with higher baseline HLA-DR neoantigen burden suggests the latter may drive this mechanism of immune escape known to contribute to relapse post-AHSCT.Keywords: Immune response, Acute myeloid leukemia, Allogeneic hematopoietic stem cell transplant