Blockade of the PD-1 axis has modest efficacy in diffuse large B-cell lymphoma (DLBCL), but data regarding LAG3 are sparse. The impact of LAG3 digital gene expression was tested in 309 patients with DLBCL treated with standard chemoimmunotherapy. Cellular distribution of LAG3 protein was determined by immunohistochemistry and flow cytometry. In tumor-infiltrating lymphocytes (TILs), LAG3 expression was highest on CD4+ regulatory T cells (Tregs) and was also highly expressed on CD8+ T cells compared with CD4+ non-Tregs (both P = .008). LAG3high TILs were enriched in PD-1 and TIM-3. LAG3 was also expressed on a proportion of malignant B cells, and these patients had significantly higher LAG3 messenger RNA in their biopsies (P = .03). LAG3high gene expression was associated with inferior survival in discovery/validation cohorts, independent of cell of origin and the international prognostic index. Patients who were PD-L1high were fivefold more likely to be LAG3high (P < .0001). Patients who were LAG3high/PD-L1high had an inferior progression-free survival (P = .011) and overall survival (P = .005) compared with patients who were LAG3low/PD-L1high. Digital spatial protein analysis confirms LAG3 expression on T cells and, surprisingly, tumor-associated macrophages (TAMs) at higher levels than found on CD20+ B cells in the tumor microenvironment. LAG3 is frequently expressed on CD4+ Tregs and CD8+ TILs, typically with other immune checkpoints, and is also present in a proportion of malignant B cells in DLBCL and in areas enriched for TAMs. LAG3high expression is associated with poor outcome independent of conventional prognosticators.

Purpose: Diffuse large B-cell lymphoma (DLBCL) is molecularly and clinically heterogeneous, and can be subtyped according to genetic alterations, cell-of-origin, or microenvironmental signatures using high-throughput genomic data at the DNA or RNA level. Although high-throughput proteomic profiling has not been available for DLBCL subtyping, MYC/BCL2 protein double expression (DE) is an established prognostic biomarker in DLBCL. The purpose of this study is to reveal the relative prognostic roles of DLBCL genetic, phenotypic, and microenvironmental biomarkers. Experimental Design: We performed targeted next-generation sequencing; IHC for MYC, BCL2, and FN1; and fluorescent multiplex IHC for microenvironmental markers in a large cohort of DLBCL. We performed correlative and prognostic analyses within and across DLBCL genetic subtypes and MYC/BCL2 double expressors. Results: We found that MYC/BCL2 double-high-expression (DhE) had significant adverse prognostic impact within the EZB genetic subtype and LymphGen-unclassified DLBCL cases but not within MCD and ST2 genetic subtypes. Conversely, KMT2D mutations significantly stratified DhE but not non-DhE DLBCL. T-cell infiltration showed favorable prognostic effects within BN2, MCD, and DhE but unfavorable effects within ST2 and LymphGen-unclassified cases. FN1 and PD-1–high expression had significant adverse prognostic effects within multiple DLBCL genetic/phenotypic subgroups. The prognostic effects of DhE and immune biomarkers within DLBCL genetic subtypes were independent although DhE and high Ki-67 were significantly associated with lower T-cell infiltration in LymphGen-unclassified cases. Conclusions: Together, these results demonstrated independent and additive prognostic effects of phenotypic MYC/BCL2 and microenvironment biomarkers and genetic subtyping in DLBCL prognostication, important for improving DLBCL classification and identifying prognostic determinants and therapeutic targets.

The XPO1 inhibitor selinexor was recently approved in relapsed/refractory DLBCL patients but only demonstrated modest anti-DLBCL efficacy, prompting us to investigate the prognostic effect of XPO1 in DLBCL patients and the rational combination therapies in high-risk DLBCL. High XPO1 expression (XPO1^high) showed significant adverse prognostic impact in 544 studied DLBCL patients, especially in those with BCL2 overexpression. Therapeutic study in 30 DLBCL cell lines with various molecular and genetic background found robust cytotoxicity of selinexor, especially in cells with BCL2-rearranged (BCL2-R⁺) DLBCL or high-grade B-cell lymphoma with MYC/BCL2 double-hit (HGBCL-DH). However, expression of mutant (Mut) p53 significantly reduced the cytotoxicity of selinexor in overall cell lines and the BCL2-R and HGBCL-DH subsets, consistent with the favorable impact of XPO1^high observed in Mut-p53-expressing patients. The therapeutic effect of selinexor in HGBCL-DH cells was significantly enhanced when combined with a BET inhibitor INCB057643, overcoming the drug resistance in Mut-p53-expressing cells. Collectively, these data suggest that XPO1 worsens the survival of DLBCL patients with unfavorable prognostic factors such as BCL2 overexpression and double-hit, in line with the higher efficacy of selinexor demonstrated in BCL2-R⁺ DLBCL and HGBCL-DH cell lines. Expression of Mut-p53 confers resistance to selinexor treatment, which can be overcome by combined INCB057643 treatment in HGBCL-DH cells. This study provides insight into the XPO1 significance and selinexor efficacy in DLBCL, important for developing combination therapy for relapsed/refractory DLBCL and HGBCL-DH.

Myeloid differentiation primary response protein 88 (MYD88) is a critical universal adapter that transduces signaling from Toll-like and interleukin receptors to downstream nuclear factor-κB (NF-κB). MYD88L265P (leucine changed to proline at position 265) is a gain-of-function mutation that occurs frequently in B-cell malignancies such as Waldenstrom macroglobulinemia. In this study, E3 ligase RING finger protein family 138 (RNF138) catalyzed K63-linked nonproteolytic polyubiquitination of MYD88L265P, resulting in enhanced recruitment of interleukin-1 receptor–associated kinases and elevated NF-κB activation. However, RNF138 had little effect on wild-type MYD88 (MYD88WT). With either RNF138 knockdown or mutation on MYD88 ubiquitination sites, MYD88L265P did not constitutively activate NF-κB. A20, a negative regulator of NF-κB signaling, mediated K48-linked polyubiquitination of RNF138 for proteasomal degradation. Depletion of A20 further augmented MYD88L265P-mediated NF-κB activation and lymphoma growth. Furthermore, A20 expression correlated negatively with RNF138 expression and NF-κB activation in lymphomas with MYD88L265P and in those without. Strikingly, RNF138 expression correlated positively with NF-κB activation in lymphomas with MYD88L265P, but not in those without it. Our study revealed a novel mutation-specific biochemical reaction that drives B-cell oncogenesis, providing a therapeutic opportunity for targeting oncogenic MYD88L265P, while sparing MYD88WT, which is critical to innate immunity.

Diffuse large B-cell lymphoma (DLBCL) is a heterogeneous entity of B-cell lymphoma. Cell-of-origin (COO) classification of DLBCL is required in routine practice by the World Health Organization classification for biological and therapeutic insights. Genetic subtypes uncovered recently are based on distinct genetic alterations in DLBCL, which are different from the COO subtypes defined by gene expression signatures of normal B cells retained in DLBCL. We hypothesize that classifiers incorporating both genome-wide gene-expression and pathogenetic variables can improve the therapeutic significance of DLBCL classification. To develop such refined classifiers, we performed targeted RNA sequencing (RNA-Seq) with a commercially available next-generation sequencing (NGS) platform in a large cohort of 418 DLBCLs. Genetic and transcriptional data obtained by RNA-Seq in a single run were explored by state-of-the-art artificial intelligence (AI) to develop a NGS-COO classifier for COO assignment and NGS survival models for clinical outcome prediction. The NGS-COO model built through applying AI in the training set was robust, showing high concordance with COO classification by either Affymetrix GeneChip microarray or the NanoString Lymph2Cx assay in 2 validation sets. Although the NGS-COO model was not trained for clinical outcome, the activated B-cell–like compared with the germinal-center B-cell–like subtype had significantly poorer survival. The NGS survival models stratified 30% high-risk patients in the validation set with poor survival as in the training set. These results demonstrate that targeted RNA-Seq coupled with AI deep learning techniques provides reproducible, efficient, and affordable assays for clinical application. The clinical grade assays and NGS models integrating both genetic and transcriptional factors developed in this study may eventually support precision medicine in DLBCL.

A noncoding polymorphism (rs78378222) in TP53, carried by scores of millions of people, was previously associated with moderate risk of brain tumors and other neoplasms. We find a positive association between this variant and soft tissue sarcoma. In sharp contrast, it is protective against breast cancer. We generated a mouse line carrying this variant and found that it accelerates spontaneous tumorigenesis and glioma development, but strikingly, delays mammary tumorigenesis. The variant creates a miR-382-5p targeting site and compromises a miR-325-3p site. Their differential expression results in p53 downregulation in the brain, but p53 upregulation in the mammary gland of polymorphic mice compared to that of wild-type littermates. Thus, this variant is at odds with Li-Fraumeni Syndrome mutants in breast cancer predisposition yet consistent in glioma predisposition. Our findings elucidate an underlying mechanism of cancer susceptibility that is conferred by genetic variation and yet altered by microRNA expression.