Somatic PRAME Deletions Are Associated with Decreased Immunogenicity, Apoptosis Resistance and Poor Outcomes in Diffuse Large B-Cell Lymphoma

Background: The current standard of care in diffuse large B-cell lymphoma (DLBCL) consists of chemotherapy and therapeutic monoclonal antibodies that have significantly improved patient outcomes over the past 15 years. However, a large proportion of patients suffer from refractory or relapsed disease. Therefore, the development of new therapeutic strategies for this subgroup of patients, who are threatened by a high chance of disease-related death, represents an important unmet clinical need.

Methods: We enrolled into our study 347 de novo DLBCL patients uniformly treated with R-CHOP from the BC Cancer population-based cohort between September 2000 and January 2012. RNAseq and high-resolution copy number analysis were performed and correlated with clinical outcome data and tumor microenvironment composition. We also performed functional studies to investigate PRAME-mediated memory T-cell responses and gene expression changes.

Results: We discovered novel, highly focal deletions of 22q11.22, including the PRAME gene in 13% (44/338) of the cases. The deletions cluster in a narrow chromosomal region that includes a very small number of genes (VpreB1, ZNF280A/B, PRAME, GGTLC2, miR-650). Of clinical importance, 22q11.22 deletions were found significantly more frequently in germinal centre B-cell-like (GCB) type DLBCL (17% (31/180) vs. activated B-cell-like (ABC) type: 8% (8/98), P < 0.01), and were also significantly associated with worse outcome, which was specifically observed in GCB-DLBCL (5-year disease specific survival, non-PRAME-deleted: 84.5% vs. PRAME-deleted: 67.2%, P = 0.026). Homozygous deletions were more strongly associated with poor outcome than heterozygous deletions. Interestingly, 90% of PRAME-deleted cases were Ig-lambda restricted (P < 0.001).

PRAME is a prominent member of the cancer testis antigen (CTA) family of proteins that are expressed in various types of cancers, but not in normal tissues, including normal mature B-cells, apart from male germinal cells. Due to the cancer-specific expression of CTAs, these molecules are considered promising targets for cancer immunotherapy using cytotoxic T-cells and tumor vaccination approaches. To determine the association with tumor microenvironment composition, we analyzed CD4/CD8 flow cytometry data from DLBCL patient samples. The numbers of CD4 and CD8-positive T cells were significantly lower in PRAME-deleted cases compared to wild type (CD4: P < 0.001, CD8: P = 0.013). Notably, RNAseq analysis revealed that the HLA-A*0201 genotype was seen significantly more often in PRAME deleted cases (PRAME wt: 2.5% vs. PRAME deleted: 10.8%, P = 0.005). In order to functionally characterize its interaction with the immune microenvironment, we utilized enzyme-linked immunoSpot (ELISPOT) assays to investigate memory T-cell reactions of patient-derived T cells to PRAME antigens using patient-derived peripheral blood mononuclear cells (PBMC) and measured IFN-g production (7 control healthy donors, 4 PRAME-deleted and 4-wild type patients). While T cells from PRAME-replete patients had no reaction to PRAME antigens, PRAME-deleted patient-derived T-cells had significant reactions to 4 independent PRAME peptides. These data suggest that PRAME-deleted tumor cells can escape from cytotoxic T-cell attack to gain growth advantage.

Next, we performed PRAME knock-out (KO) experiments using CRISPR/Cas9 genome editing to clarify the cell autonomous effects of PRAME deletions. Using 2 different cell lines (Karpas422 and SUDHL-4), we found TNFSF10 (TRAIL) expression was significantly down-regulated in homozygous PRAME-KO cell lines compared to wild type. The soluble form of TRAIL (sTRAIL) was also reduced, as measured with enzyme-linked immunosorbent assays. These results suggest that PRAME downregulated cells may contribute to cell survival via TRAIL and sTRAIL reduction.

Conclusion: We identified recurrent PRAME deletions and characterized their clinical and functional role in DLBCL. Our findings contribute to the understanding of cell-autonomous and extrinsic roles of PRAME deletions in lymphomagenesis and may lead to the discovery of new therapeutic avenues to simultaneously treat the tumor and the host.