Clinical and molecular features of Epstein‐Barr virus‐positive diffuse large B‐cell lymphoma: Results in a multi‐center trial

Abstract

Epstein-Barr virus (EBV) is an important human oncogenic virus and closely related to the pathogenesis of diffuse large B-cell lymphoma.1 This is, to our knowledge, the first study to systematically evaluate the clinical impact of EBV infection on DLBCL, based on a prospective clinical trial (NHL-001, NCT01852435). Meanwhile, genomic and transcriptomic features of EBV+DLBCL were further investigated by whole exome/genome sequencing (WES/WGS), targeted-, and RNA-sequencing. EBV-encoded RNA (EBER) in situ hybridization is the gold standard of EBV-infection status in tumors.2 Here 46 of 429 patients were positive for tumor EBER (10.7%) and EBER-positive patients had poor performance status (ECOG≥2, 35.9% vs 11.8%, P = 0.0338) and low rate of complete remission (CR, 45.7% vs 88.8%, P < 0.0001) (Table 1). In univariate analysis (Figure 1A), adverse prognostic factors for both progression-free survival (PFS) and overall survival (OS) included age, ECOG≥2, advanced stage, extranodal site≥2, elevated serum LDH, IPI > 2, and EBER positivity. In multivariate analysis, EBER positivity independently indicated adverse outcome (P = 0.0328 for PFS and P = 0.0182 for OS, Table S1). The 2-year PFS and OS were 47.8% and 69.6% for EBER-positive patients, and 86.4% and 92.2% for EBER-negative patients (P both < 0.0001, Figure 1B). This was observed not only in young patients (P both < 0.0001, Figure 1C), but also in elderly patients (P both < 0.0001, Figure 1D). The adverse prognostic effect of EBV infection was in consistent with cohorts from Japan,3 and Korea,4 but not in western countries.5 probably due to relatively low incidence of EBV infection in western population. Nevertheless, EBER-positive DLBCL should be referred to an individual clinical entity in both younger and elderly patients. Survival analysis in DLBCL patients according to tumor EBER. (A) Forest plot of PFS and OS in DLBCL patients according to different clinical characteristics. (B) Survival curves of PFS and OS for all patients, (C) for young patients and (D) for elderly patients according to tumor EBER. PFS, progression-free survival; OS, overall survival; HR, Hazard ratio; CI, confidence interval; GCB, germinal center B-cell; ECOG, Eastern Cooperative Oncology Group; IPI, International Prognostic Index; LDH, lactic dehydrogenase As for EBV-associated serum markers, EBV-DNA is usually negligible in cell-free body fluid like serum, but becomes detectable in EBV-related disease.2 Among 141 patients available for serum samples, serum EBV DNA were positive in 14 cases, correlated with tumor EBER (P = 0.0036, concordance rate 82.3%, Table S2) and shared similar impact on ECOG≥2 (35.7% vs 11.81% P = 0.0298) and response to immunochemotherapy (CR, 64.3% vs 89.8%, P = 0.0186, Table S3). In univariate analysis, serum EBV DNA predicted unfavorable 2-year PFS (66.7% vs 84.2%, P = 0.0167) and 2-year OS (72.2% vs 92.1%, P = 0.0420). As serum EBV DNA is mostly derived from tumor cells,6 our findings emphasized the advantage of serum EBV DNA as noninvasive biomarker for predicting tumor EBER status and prognostic factor of DLBCL. EBV serological tests refer to different humoral antibody patterns against viral proteins, including Epstein-Barr nuclear antigen (EBNA), early antigen (EA), viral capsid antigen (VCA) and Rta protein (Rta). Among these 141 patients, 116 cases were positive for EBNA IgG (116/141, 82.3%), 33 cases positive for EA IgG (33/141, 23.4%), 12 cases for VCA IgA (12/141, 8.5%), and 8 cases for Rta IgG (8/141, 5.7%). Although EBV serological markers were not correlated with tumor EBER (Table S2) and prognosis (Figure 1A), VCA IgA-positive patients tend to have elevated serum LDH (67.4% vs 33.3%, P = 0.0264) and higher IPI (58.8% vs 22.2%, P = 0.0022, Table S3), providing additional clinical information in EBV-related DLBCL. Genomic alterations play an important role on pathogenesis of EBV-associated malignancies.7 Here, by integrating WES/WGS, targeted- and RNA-sequencing, we described a comprehensive view of genomic alterations based on a large cohort of DLBCL. Gene mutations were first screened in 180 patients with WGS/WES/targeted sequencing data, including 40 cases of EBER-positive DLBCL and 140 cases of EBER-negative DLBCL. No significant difference was observed in EBER-positive DLBCL according to cell of origin or genotypes (Table S4). Seven highly mutated genes were identified (Figure 2A), including DDX3X (22.5% vs 9.3%), TET2 (22.5% vs 7.1%), MYC (20.0% vs 5.0%), STAT3 (17.5% vs 3.6%), TNFAIP3 (17.5% vs 3.6%), TNFRSF14 (15.0% vs 4.3%), and LYN (10.0% vs 1.4%) (Figures S1 and 2B). Previously revealed by targeted sequencing, epigenetic regulator TET2 and transcription factor MYC were frequently mutated in EBV + DLBCL. We not only confirmed the association of TET2 and MYC mutations with EBV infection in DLBCL, but also revealed significant enrichment of TET2- and MYC-targeted genes expression in EBER-positive DLBCL (Figure 2C). As for major signaling pathways (Figure 2B), EBER-positive DLBCL showed aberrant activation of Wnt, NF-κB, and JAK-STAT pathway and inhibition of BCR signaling pathway, as compared to EBER-negative DLBCL (Figure 2C). It is possible that the mutation changes in these pathways promote tumor cell proliferation and progression in EBER-positive DLBCL in a BCR-signaling-independent manner. These findings expanded the previous observation of different genetic background and offered potential target signaling cascades for treating EBER-positive DLBCL. As revealed by CNV analysis in 122 patients with WGS/WES data, EBER-positive patients represented a distinctive pattern of copy number alterations, involving frequently deletions (>10%) in chromosome 6p21.32-33, 8p23.1, and 6p22.1, as well as recurrent amplifications in chromosome 1q32.1 and 1q21.3 (Figure 3A). Interestingly, genes in deletion regions (> 10%) were mostly involved in processing and presentation of endogenous peptide antigen (P < 0.0001, Figure 3B). Given the impact of loss of antigen processing and presentation on antitumor immunity,8 we observed a significant relationship between antigen processing and presentation Z-score and Immune Score (P = 0.0003, n = 200; Figure 3C), obtained from RNA-sequencing data by ssGSEA and xCell method. Moreover, EBER-positive DLBCL presented significantly decreased antigen processing and presentation Z-score and Immune Score, as compared to EBER-negative DLBCL. Specifically, we focused on the interactions between MHC class I/II molecular genes and immune cell subtypes, ranked by their association strengths with Immune Score or antigen presentation and processing Z-score (Figure 3D). MHC class I-related molecules (B2M, HLA-C, HLA-B, NLRC5, MR1, and CD74) were significantly decreased in EBER-positive patients and correlated with dendritic cells (DC) and gamma delta/CD4+/CD8+ T-cell subtypes. Corresponding decreases in CD4+memory T-cells (P = 0.0084), total CD8+ T-cells (P = 0.0002), CD8+effector memory T-cells (CD8+Tcm, P = 0.0005) and CD8+ central memory T-cells (CD8+Tem, P = 0.0499), were also observed in EBER-positive DLBCL (Figure 3E). As MHC class I molecules play a critical role in presenting viral or tumor antigen for CD8+ cytotoxic T-cell activation,9 loss of MHC class I molecules could allow for persistent EBV infection and impair tumor immunosurveillance in EBER-positive DLBCL. Thus, it might be implications for immunotherapies targeting antigens, such as EBV-specific T cell therapy, in treating EBER-positive DLBCL.10 In conclusions, EBV contributed to tumor progression in DLBCL with distinct oncogenic mutations and tumor microenvironment alterations. The authors are grateful to Yan. Zhao for sample preparations, and Chaofu Wang, Xiaochun Fei Xiao-chun Fei for pathological analysis. This study was supported, in part, by research funding from the National Natural Science Foundation of China (81830007 to W-LZ, 81870315 and U2005204 to J-DH), Chang Jiang Scholars Program, Shanghai Municipal Education Commission Gaofeng Clinical Medicine Grant Support (20152206 to W-LZ and 20152208 to LW), Clinical Research Plan of Shanghai Hospital Development Center (SHDC2020CR1032B to W-LZ), Program for Innovative Research Team in Science and technology in Fujian Medical University (2017XQ2030 to C-XZ), Joint Funds for the Innovation of Science and Technology, Fujian province (2016Y9032 to J-DH). All authors declare that no conflict of interest exists. C.-X.Z. and J.-J.W. collected and analyzed clinical data, prepared biological samples, analyzed the omics data and wrote the article. D.F. and P.-P.X performed the experiments and analyzed the omics data. S.C., L.W., X.W., J.-F.Z., L.-P.S., Z.-W.C., J.-P.C., M.-Y.F., T.L., Y.-P.S., K.Y., Y.L., J.G. and M.H. gathered detailed clinical information for the study and analyzed the data. C.-F.W. conducted pathological analysis. W.-L.Z., and J.-D.H. conceived the study, directed, and supervised research and wrote the manuscript. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts