An Immune Senescence and Exhaustion-Related RNA Profile Predicts Clinical Outcomes in Acute Myeloid Leukemia

Background Chemotherapy refractoriness and disease relapse continue to be significant obstacles to therapeutic success in AML. We have recently identified bone marrow (BM) transcriptomic profiles that stratify patients with AML into an immune-infiltrated and an immune-depleted subtype and that refine the accuracy of survival prediction in response to conventional chemotherapy beyond that afforded by cytogenetic and molecular prognosticators (Vadakekolathu J, et al. Sci. Transl. Med. 2020; 12: eaaz0463). Importantly, AML CD8+ T cells exhibit features of immune exhaustion and senescence (IES), including the upregulation of natural killer (NK) cell-associated transcripts, which persist only in chemotherapy non-responders (Knaus HA, et al. JCI Insight 2018; 3: e120974). The aim of the current study was to determine whether IES correlate with immune infiltration and with clinical outcomes in AML. Methods We used prior knowledge and gene set enrichment analysis (GSEA) to derive a 68-gene IES signature score (computed as the cohort-wide average of gene expression) from publicly available RNA-sequencing datasets (TCGA and Beat-AML Master Trial; 162 and 281 AML patients, respectively). The wet-lab AML cohorts used in this study included a total of 382 BM samples from children and adults with AML treated with curative intent. BMs were collected at time of diagnosis, complete remission (CR) and relapse (PMCC, SAL and CHOP series). BM RNAs were profiled on the nCounter platform using the PanCancer Immune Profiling Panel (NanoString Technologies, Seattle, WA). Results In the AML discovery cohort, the IES score significantly correlated with BM lymphoid infiltration (R=0.958) and with CD8+ T cell (R=0.88), cytotoxic T cell (R=0.90) and tumor inflammation signature scores (R=0.82, P