Modeling the CNS Tropism of Diffuse Large B-Cell Lymphomas in Vivo

Introduction: Central nervous system (CNS) manifestation accounts for dismal prognosis in patients diagnosed with systemic diffuse large B-cell lymphoma (DLBCL). Myc rearrangement, deletions encompassing the CDKN2A (a.k.a. INK4a/ARF) or ATM gene loci as well as NF-kB-hyperactivating mutations are frequently detected in primary and secondary CNS-tropic lymphoma, however, investigations that functionally link these lesions to CNS involvement in adequate in vivo model systems are missing. Methods: We generated primary Eµ-myc transgenic mouse lymphomas with and without distinct naturally occurring NF-kB mutations (within genes encoding for MyD88, CD79B, A20, IkBζ, IkBε or BIRC3) or deletions at the INK4a/ARF and ATM loci by retroviral gene transfer and crossbreeding to the respective knockout strains. A subset of the lymphomas was subjected to gene expression profiling and whole-exome sequencing (WES). Wild-type recipient mice propagated with lymphoma cells via tail vein injection were monitored for systemic lymphoma development, the time at which the brain was isolated and examined regarding lymphoma infiltration. Results: Underlining Myc's role as a putative co-driver of CNS involvement, we found in about 40% of primary Eµ-myc lymphomas (with no additional exogenous lesions) meningeal lymphoma manifestations, and transplantation of the same individual lymphomas into numerous recipients reproduced the CNS lymphoma status. Gene set enrichment analysis of genome-wide transcriptome profiles indicated NF-kB hyperactivation in the CNS-tropic lymphoma group, suggesting that constitutive NF-kB signaling may promote CNS-prone pathogenesis in vivo. Transplantation of Eµ-myc transgenic hematopoietic stem cells expressing a variety of NF-kB-activating mutants in myeloablated recipient mice resulted in a significant acceleration of Eµ-myc -driven lymphomagenesis, with some, but not all of these mutants conferring a CNS-tropic lymphoma phenotype. Global NF-kB suppression in CNS-tropic Eµ-myc lymphomas via the NF-kB-antagonizing IkBΔN super-repressor did not fully abrogate lymphoma infiltration of the brain, suggesting that additional factor(s) must contribute. Accordingly, targeted ablation of the INK4a/ARF and ATM loci robustly enhanced CNS tropism of Eµ-myc lymphomas. Conclusions: The Eµ-myc mouse lymphoma model is well-suited to genetically dissect and rebuild components of DLBCL-like CNS tropism. We identified CDKN2A or ATM deletions as critical determinants of CNS tropism in vivo. Our systematic analyses of different NF-kB mutants - so far rather recognized as functionally interchangeable - indicated that only distinct NF-kB mutants contribute to CNS tropism in B-cell lymphomas. WES data, results from compound genotypes (e.g. combining ATM deletions with an NF-kB-activating mutation), and the impact of the host's cellular immune status will be reported at the meeting. Our findings underscore the need for functional analyses of oncogenic network contexts, and provide important insights into candidate target lesions for personalized CNS-directed therapies in DLBCL patients in the future. Disclosures No relevant conflicts of interest to declare.