Tumor Microenvironment and It's Prognostic Value Is Different in Testicular and Systemic Diffuse Large B-Cell Lymphoma

Abstract

Testicular diffuse large B-cell lymphoma (T-DLBCL) is an aggressive lymphoid cancer with distinct clinical and biological features. Majority of the cases represent non-GCB phenotype, but the genetic landscape with frequent aberrations in MYD88, NFKBIZ, PDCD1LG2, and CD274 differ from systemic non-GCB DLBCL and highlight the significance of immune-escape and tumor microenvironment (TME). We have previously shown that high T-cell gene expression signature and high tumor-infiltrating lymphocyte (TIL) content translate to better survival in testicular but not in systemic DLBCL. We have also described a subpopulation of T-bet+FoxP3+ Tregs that have a significant adverse impact on survival in T-DLBCL but not in in systemic DLBCL. In order to further characterize the differences in the TME between T- and systemic DLBCL, we used NanoString digital gene expression profiling to assess the expression of 770 immune response genes in 60 T-DLBCL and 84 systemic DLBCL samples. In addition, we used multiplex immunohistochemistry (mIHC) to characterize TILs, tumor-associated macrophages (TAMs), NK-cells, checkpoint molecules, cell proportions, and cell interactions in 80 T-DLBCLs and 192 systemic DLBCLs. Unsupervised hierarchical clustering of the gene expression data revealed clusters of genes with differential expression between T- and systemic DLBCLs. In T-DLBCL, 11 HLA genes were downregulated and genes related to B-cell receptor (BCR) signaling upregulated compared to systemic DLBCL. All T-DLBCL patients that clustered among the patients with systemic DLBCL had high T-cell signature expression, and systemic DLBCL patients that clustered among patients with T-DLBCL had significantly higher Ki-67 expression (p=0.00085). Based on the mIHC data, we observed not only a significantly lower proportion of FoxP3+ Tregs altogether but also a higher proportion of T-bet+FoxP3+ double positive Tregs in T-DLBCL compared to systemic DLBCL, confirming our previous findings on the significant role of T-bet+FoxP3+ double positive Tregs in the T-DLBCL TME. We also observed less TAMs in T-DLBCL TME. Unsupervised hierarchical clustering of the cell interaction data verified that distinct interaction clusters are found and separate the patients into different subgroups. Interactions of cytotoxic T-cells were more commonly seen in T-DLBCL whereas interactions of PD-1 expressing cells and interactions between TILs and TAMs were more frequent in systemic DLBCL. Taken together, our results reveal differences in the TME between T- and systemic DLBCL, and further emphasize the impact of immune escape and distinct immune cells in T-DLBCL. Future studies are ongoing in order to further characterize the cell-to-cell interactions, their association with survival, and TME differences between T- and systemic DLBCL, in particular non-GCB DLBCL.