Lipid metabolism-related gene signature predicts prognosis and unveils novel anti-tumor drugs in specific type of diffuse large B cell lymphoma.

Background: Diffuse large B-cell lymphoma (DLBCL) is the most common type of lymphoma which possess highly aggressive and heterogeneous. Despite advances in understanding heterogeneity and development of novel targeted agents, the prognosis of DLBCL patients remains unsatisfied. Lipids are crucial components of biological membranes and signal transduction while accumulating evidence has supported the vital roles of abnormal lipid metabolism in tumorigenesis. Furthermore, some related pathways could serve as prognostic biomarkers and potential therapeutic targets. However, the clinical significance of abnormal lipid metabolism reprogramming in DLBCL has not been investigated. In the current study, we developed a prognostic risk model for DLBCL based on the abnormal expressed lipid metabolism genes and moreover based on our risk model we classified patients with DLBCL into novel subtypes and identified potential drugs for DLBCL patients with certain lipid metabolism profiles.

Methods: We utilized univariate Cox regression analysis to identify the prognosis-related lipid metabolism genes, and then performed LASSO Cox regression to identify prognostic related lipid metabolism related genes. Multivariate cox regression was used to establish the prognostic model. Patients were divided in to high and low risk groups based on the median risk score. Immune cell infiltration and GSEA were used to identify the pathways between high and low risk groups. Oncopredict algorithm was utilized to identify potential drug for high-risk patients. In vitro cell apoptosis and viability analysis were employed to verify the specific tumor inhibition effects of AZD5153.

Results: Nineteen survival related lipid metabolism genes TMEM176B, LAYN, RAB6B, MMP9, ATAD3B, SLC2A11, CD3E, SLIT2, SLC2A13, SLC43A3, CD6, SIRPG, NEK6, LCP2, CTTN, CXCL2, SNX22, BCL6 and FABP4 were identified and subjected to build the prognostic model which was further verified in four external microarray cohorts and one RNA seq cohorts. Tumor immune microenvironment analysis and GSEA results showed that the activation of MYC targets genes rather than immunosuppression contribute to the poor survival outcome of patients in the high-risk group. AZD5153, a novel bivalent BET bromodomain inhibitor which could inhibit the transcription of MYC and E2F exhibited specific antitumor function for cells with high-risk score.

Conclusions: Our results provide the first lipid metabolism-based gene signature for predicting the survival of patients with DLBCL. Furthermore, by determining novel subtypes with our lipid metabolism prognostic model we illustrated that drugs that compromising MYC target genes rather than immune checkpoint inhibitors may be beneficial to DLBCL patients with certain lipid metabolism profiles.