Abstract
Introduction: Constitutive NF-κB activation is a hallmark feature of B-cell malignancies, including the ABC-subtype of diffuse large B-cell lymphoma (DLBCL), and is mainly due to gain-of-function mutations in components of the Toll-like receptor (MYD88) or B-cell receptor (CD79B, CARD11) pathways, inactivating mutations of cellular NF-κB suppressors (TNFAIP3/A20, BIRC3) or copy number gain (NFKBIZ), often in the presence of high Myc oncoprotein expression. In order to understand whether such genetic lesions may cooperate with Myc during lymphomagenesis, we took advantage of the Eµ- myc transgenic mouse in which aggressive B-cell lymphomas are induced by B-cell-specific overexpression of c-Myc.
Methods: Eµ- myc transgenic hematopoietic stem cells (HSC) were stably transduced with GFP-expressing retroviral vectors encoding the murine orthologs of human MYD88-L265P, NFKBIZ, CD79B-ITAM-Y196H, CARD11-L244P, CARD11-L225LI or shRNA against TNFAIP3/A20 or BIRC3 prior to their transplantation into lethally irradiated syngenic C57BL/6 mice. Recipient animals were subsequently monitored for the development of a palpable lymphadenopathy, and lymphomas were subject to immunophenotyping, analysis of growth parameters and transcriptome profiling by RNA-Seq.
Results : Ectopic overexpression of NFKBIZ, MYD88-L265P, CD79B-ITAM-Y196H or knockdown of TNFAIP3/A20 in Eµ- myc transgenic HSC gave rise to GFP(+) B-cell lymphomas (> 70% GFP+/B220+ cells), albeit at varying frequencies. All mutant-driven lymphomas manifested with reduced apoptotic TUNEL/B220 indices when compared to the empty-vector control cohort (ctrl. lymphomas). In contrast, neither knockdown of BIRC3 nor overexpression of the CARD11 mutants generated GFP(+) Eµ- myc lymphomas in vivo . Unexpectedly, we found senescent GFP(+) B- lymphocytes residing in the spleens of these healthy animals after transplantation. Genetic ablation of the senescence program (via allelic deletion of the histone demethylase Suv39h1 or the INK4a/ARF locus) in Eµ- myc HSC prior to transduction with the CARD11 mutants gave rise to palpable GFP(+) B-cell lymphomas (unlike the empty-vector), indicating that CARD11 missense mutations require a senescence-deficient background to promote Eµ- myc -driven lymphomagenesis.
RNA-Seq unvealed downregulation of MHC class II genes as a common feature of all mutant-driven Eµ- myc lymphomas when compared to ctrl. lymphomas. Since MHC/T-cell receptor interactions are required for an effective anti-tumoral adaptive immune response, this finding suggested immune evasion mechanisms to be exploited by all genetic lesions tested to promote Myc-driven lymphomagenesis. Further supporting this notion, MYD88-L265P-driven lymphomas exhibited elevated surface expression of the immune checkpoint ligands PD-L1 and PD-L2. Strikingly, knockdown of PD-L1 alone was sufficient to significantly delay MYD88-L265P-driven lymphomagenesis in Eµ- myc HSC transplantation experiments, suggesting that PD-L1-mediated immune evasion mechanisms contribute significantly to the aggressive phenotype observed under the MyD88-L265P mutant in vivo . Furthermore, overexpression of PD-L1 in Eµ- myc HSC alone also induced GFP(+) B-cell lymphomas that displayed a striking overlap in gene expression with MyD88-L265P-driven Eµ- myc lymphomas.
Conclusions: We continue to characterize the Eµ- myc transgenic mouse as a well-suited model platform to functionally dissect DLBCL-relevant oncogenic networks in general and NF-κB-activating mutations in the context of high Myc protein expression, a frequent feature of DLBCL, in particular. Our systematic in vivo analysis of different NF-κB-activating mutants - so far rather recognized as functionally interchangeable moieties - indicated that some of these genetic lesions directly cooperate with Myc (i.e. MyD88-L265P, NFKBIZ) - partly via utilizing immune evasion mechanisms - while others (e.g. CARD11 mutations) require upfront senescence-ablating lesions to license Myc-driven lymphomagenesis. Our findings might explain the pressure to select for senescence-ablating lesions like INK4a/ARF loss frequently found in human DLBCL, and could inform immune checkpoint blockade treatment strategies based on certain oncogenic/immune evasion network interconnections.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.