GPX1-associated prognostic signature predicts poor survival in patients with acute myeloid leukemia and involves in immunosuppression.

Treatment of acute myeloid leukemia (AML) remains a challenge. It is urgent to understand the microenvironment to improve therapy and prognosis. Bioinformatics methods were used to analyze transcription expression profile of AML patient samples with complete clinical information from UCSC Xena TCGA-AML datasets and validate with GEO datasets. Western blot, qPCR, RNAi and CCK8 assay were used to assay the effect of GPX1 expression on AML cell viability and the expression of genes of interest. Our analyses revealed that highly expressed GPX1 in AML patients links to unfavorable prognosis. GPX1 expression was positively associated with not only fraction levels of myeloid-derived suppressor cells (MDSCs), monocytes and T cell exhaustion, the expression levels of MDSC markers, MDSC-promoting CCR2 and immune inhibitory checkpoints (TIM3/Gal-9, SIRPα and VISTA), but also negatively with low fraction levels of CD4+ and CD8+ T cells. Silencing GPX1 expression reduced AML cell viability and CCR2 expression. Moreover, GPX1-targetd kinases were PKC family, SRC family, SYK and PAK1, which promote AML progression and the resistance to therapy. Furthermore, Additionally, GPX1-associated prognostic signature (GPS) is an independent risk factor with high area under curve (AUC) values of receiver operating characteristic (ROC) curves. High risk group based on GPS enriched not only with endocytosis which transfers mitochondria to favor AML cell survival in response to chemotherapy, but also NOTCH, WNT and TLR signaling which promote therapy resistance. Our results revealed the significant involvement of GPX1 in AML immunosuppression via and provided a prognostic signature for AML patients.