A Systems Biology Approach Predicts Distinct Roles for NFkB Subunits cRel and RelA in DLBCL

Heterogeneity in therapeutic response presents a challenge to the successful treatment of Diffuse Large B-Cell Lymphoma (DLBCL). Despite the recognition that DLBCL cases have diverse genetic and transcriptional characteristics, standard first-line therapy has remained unchanged for more than a decade. Canonical Nuclear Factor KappaB (NFκB) is a dimeric transcription factor usually consisting of either cRel or RelA bound to p50. While aberrant NFκB activation is frequently observed in DLBCL, subunit composition in individual DLBCL cases is not routinely characterized but has the potential to improve stratification and identify novel molecular targets for treatment. Computational simulations of NFκB control over B-cell proliferation and apoptosis accurately predict experimental results with accuracy at both single-cell and cell-population scale. However, the key regulatory networks controlling B-cell differentiation were not factored into these predictive models. Simulations based on known regulatory interactions were insufficient to recapitulate healthy B-cell differentiation. Using a systems biology approach we found that although cRel drives B-cell proliferation, it also blocks terminal differentiation to antibody-secreting plasma cells; dynamic downregulation of cRel by Blimp1 was a pre-requisite for differentiation. Inclusion of this interaction into multiscale computational models enabled simulations to accurately predict B-cell population dynamics in wild-type (WT) and cRel knockout cells. Simulations of aberrantly increased NFκB activity, recapitulated the increased proliferation and cell survival seen in both ABC- and GC-DLBCL. In order to interrogate the subunit-specific roles of cRel and RelA in DLBCL we performed simulations in which each subunit was individually upregulated. Both of these models predicted hyperproliferation and apoptosis avoidance, but only upregulation of cRel resulted in an inability to exit the germinal centre as seen in GC-DLBCL. In contrast, RelA- specific upregulation resulted in population expansion without a block on differentiation, with cells predicted to take on a more differentiated state consistent with cell-of-origin classification of ABC- DLBCL. This subunit-specific control over DLBCL sub-type aligns with experimental observations of the less differentiated state of GC- compared to ABC-DLBCL. Other commonly occurring mutations in DLBCL affecting BCL2, IRF4, and MYC were simulated to recapitulate dysregulated apoptosis, differentiation and cell-cycle respectively, along with “double-hit” mutations. These mutation-specific simulations of DLBCL represent in silico laboratories where biomarkers can be identified to stratify and target lymphoma.