Deciphering the Mechanism of Action of the Novel Celmod, Golcadomide, during Germinal Center B Cell Immune Response and in a Preclinical Mouse Model of Follicular Lymphoma

Follicular lymphoma (FL) is an indolent tumor in which epigenetic alterations and BCL2 translocations foster malignant transformation by reprogramming germinal center (GC)/memory B cell immune dynamics and remodeling the tumor microenvironment (TME). Lenalidomide and other cereblon (CRBN)-targeting agents, such as the novel CELMoD, golcadomide, have shown clinical efficacy in B cell neoplasms such as FL or DLBCL. However, because these molecules lack activity in mouse models, unraveling the biological mechanisms by which CELMoDs exert their activity in complex in vivo systems, such as the native GC B cell immune response or relevant preclinical mouse models, has been limited. Here we investigated the in vivo mechanism of action of golcadomide during the normal and pathological GC B cell immune response and tested whether early intervention with immunomodulatory agents may represent a promising therapeutic avenue against FL progression.

To study the in vivo effect of golcadomide during a native immune response, we developed a novel mouse model (^huCRBN) by inserting the human CRBN gene upstream and in frame of the mouse Crbn gene, allowing substrate recruitment and degradation not possible in a wild-type mouse. We immunized ^huCRBN animals to induce GC formation and treated the mice by gavage every other day with golcadomide (3mg/kg) or vehicle for 1, 2, 3 weeks and sacrificed 48h after treatment cessation. We analyzed Ikaros and Aiolos protein levels in T and B cell subsets by flow cytometry at 6, 24h after treatment. To evaluate the therapeutic activity of golcadomide in FL in vivo, we crossed our previously characterized Kmt2d/BCL2 (KB) preclinical FL model (Brisou et al, ASH 2023) with the ^huCRBN mouse (KBC). We then generated chimeric mice in which hematopoietic progenitor cells from CD19Cre⁺ Kmt2d^{f/+ hu}CRBN⁺ mice were used as donor cells for retroviral transduction with the ^huBCL2-Cd2^Tag vector. The engineered progenitor cells were then re-transplanted into sub-lethally irradiated immunocompetent WT (n=15) or ^huCRBN recipients (n=29). KBC animals at the pretumoral stages (8 months post-transplant) were randomized to receive golcadomide or vehicle alone. Spectral flow cytometry was conducted on immune subsets in spleen/lymph nodes and integrative single cell RNAseq and immune repertoire analysis was done on whole spleen in selected ^huCRBN and KBC animals.

During the normal GC immune response, we showed that in the ^huCRBN mice treated with golcadomide, all splenic B and T cell subsets exhibit a rapid degradation of IKZF1 and, to a lesser extent, IKZF3 substrates in 6 hours, which is not seen in wild-type mice, confirming the relevance of the ^huCRBN model for evaluating CELMoD activity. Golcadomide treatment acts on B cells in all lymphoid tissues with a progressive decrease of naive and memory subsets. We observed a global increase and activation of T cells with T regulatory cells reaching 25% of the CD4 pool and a phenotypic shift to CD4/CD8 central memory and effector cells consistent with previous observations with lenalidomide. We next evaluated the efficacy of golcadomide and impact on mutant B and immune T cells in the pretumoral KBC model. As previously reported, single-cell expression profiles of the untreated KBC-mutant B cells manifested mainly as a continuum of states from proliferating GC-like to quiescent mem-like phenotypes confirming progression to FL pretumoral stages. KBC mice receiving golcadomide presented with decreased memory-like subsets whereas the GC/GC-Mem intermediates persisted and expanded, suggesting their resistance to golcadomide alone and a capacity to form a pool of residual cells after therapy. Finally, golcadomide induced a significant increase in Tregs and CD4/CD8⁺ towards an activated phenotype, helping to restore immune surveillance. Spatial analysis to map cellular neighborhoods of malignant B cells, immune cells on treatment is ongoing. In addition, the preclinical evaluation of a combination therapy of golcadomide with mouse anti-Cd20 immunotherapy modeling the impact of a chemo-free regimen in FL will be reported.

Our original ^huCRBN and KBC mouse models allow the first in vivo study to unravel the efficacy and mechanism of action of CELMoDs during native immune response and early FL progression. We expect that this study will allow us to propose novel rational hypotheses for combination therapies with golcadomide to intercept FL.