Immune population changes after glofitamab and obinutuzumab (GLOBIN) as first-line therapy for patients with follicular lymphoma and marginal zone lymphoma

Abstract Introduction: Glofitamab (glofit) is a CD3xCD20 bispecific antibody that achieved high response rates but also frequent grade 2 or higher cytokine release syndrome (CRS) (26%) in patients with relapsed/refractory follicular lymphoma (FL) in an initial phase II study (Morschhauser, ASH abstract 2021). Our institution is leading an investigator-sponsored, phase II multicenter trial (NCT05783596) testing time-limited first-line treatment with obinutuzumab (obin) and glofit for patients with high-tumor burden FL and marginal zone lymphoma (MZL), allowing us to evaluate immune dynamics without the biological effects of prior and concurrent cytotoxic chemotherapy. With the goal of lowering the risk of CRS and deepening responses, patients receive debulking therapy with 4 weekly doses of obin (Day -21, -14, -7, 0) followed by 12 four-week cycles of glofit, including glofit step-up dosing (SUD) during cycle 1 (C1). Preliminary results from the trial reveal high response rates (best overall and complete metabolic response rates of 100 and 88%, respectively) and lower than expected rates of grade ≥ 2 CRS (3%) in patients with FL. Outcomes for patients with MZL are immature. We conducted an initial analysis of peripheral blood to assess treatment-associated changes in T/NK and B-cell populations in patients with and without high-grade CRS. Methods: Peripheral blood mononuclear cells were analyzed using two spectral flow cytometry panels, comprising 18 and 16 colors, to profile the T/NK and B-cell compartments, respectively. Seven of the initial patients were selected based on CRS grade and initial response to therapy. When available, samples from 5 timepoints were analyzed: screening (C1D-21), C1D1 (after 4 doses of obin and before glofit), C1D8 (after first glofit SUD), and C2D1 (after 2 glofit SUDs). In addition, a progression timepoint sample was analyzed for one patient with progressive disease. Results: Following obin administration, an increase in CD4+ T cells (FL: 57.2 to 81.1%, MZL: 58.7 to 72.1%) and a decrease in CD8+ T cells (FL: 31.2 to 12.9%, MZL: 32.2 to 22.2%) were observed in both FL and MZL patients, with the most pronounced CD4+ T cell rise (~20%) in patients without CRS. T follicular helper (Tfh, CD4+ CXCR5+) and regulatory T cells (Tregs, CD4+ FoxP3+ CD127+) constituted a higher percentage of CD4+ T cells in FL (Tfh: 6.3-7.4%, Treg: 18.4-26.8%) compared to MZL (Tfh: 1.3-3.7%, Tregs: 4.6-6.32%). Following glofit treatment, CD8+ T cells increased in both lymphoma types (FL: 12.9 to 34.3%, MZL: 22.2 to 26.4%), with a gradual decline at C2D1 observed in patients achieving complete response (CR) with no evidence of CRS (19.5%), accompanied by increased TIM3 expression (FL: 11 to 77.8%, n=1, MZL: 7.7 to 36.4%, n=1), indicating heightened T cell exhaustion. Effector (47.8 to 55.8%) and cytotoxic CD8+ T cell populations (56.4 to 83.2%) steadily increased over time during C1 for both lymphoma types, with higher PD1+ CD8+ T cell levels noted in MZL patients (FL: 5.9-7.8%, MZL: 21.1-27.2%). Notably, when comparing patients with CRS to those without, we noticed a lower percentage of effector CD8+ T cells (32.9-44.5% vs. 61.8-78.4%) and higher effector memory subsets (44.2-51% vs. 5.7-21.4%). Preliminary NK cell analyses revealed increased cytotoxic NK cell populations in FL after the introduction of glofit (26.9 to 59.5%), albeit with increased activation over time as indicated by CD69+ expression (4.8 to 14.6%) and reduced Granzyme B+ NK cells (14.9 to 8.8%) post-obin and during glofit SUD. Analyses of the B-cell compartment showed that the majority of CD19+ CD20+ B cells were depleted by C1D1, highlighting the potency of obin pretreatment in eradicating circulating B cells. Kappa light chain (65.8%) was also highly favored in B cells compared to lambda light chain (1.3%), and IgM was highly present across naïve (95%) and switched (72.5%) memory B cell compartments, with minimal IgG. Conclusion: These preliminary findings suggest that immune responses to GLOBIN are characterized by dynamic alterations in T and NK cell subsets, with notable differences among FL and MZL patients. The expansion of cytotoxic T and NK cells correlates with favorable responses, particularly in patients without CRS. The depletion of circulating B cells after obin also supports pretreatment with monoclonal antibody. Overall, these are preliminary results, and analyses of the complete trial cohorts will be presented at the meeting.