Circulating Biomarkers and CD8+ T-Cell Subpopulations Reveal Evolving Phases of Immune Response in Patients Receiving Mosunetuzumab for Previously Untreated B-Cell Lymphoma

Background: T-cell-engaging therapies, including CD20xCD3 bispecific antibodies (BsAbs), have revolutionized treatment of B-cell lymphomas. However, BsAbs rely on an artificial, “single-signal” T-cell activation, bypassing TCR specificity of the effector CD8+ cells. The characteristics of systemic immune response on BsAb therapy are poorly understood. Recent studies reported T-cell exhaustion in lymphomas relapsing on BsAbs, so far only in patients (pts) with refractory disease exposed to prior immunosuppressive therapies [Duell, Blood 2023]. We hypothesized that in pts without prior chemotherapy exposure, BsAbs will trigger a different immune response early on treatment, and during later CD3 engagement after depletion of target B-cells. Here, we evaluated circulating cytokines and CD8+ T-cell subsets during therapy with a CD20xCD3 BsAb mosunetuzumab (Mosun) in a correlative study accompanying an ongoing clinical trial (NCT04792502).

Methods: BrUOG-401 is an academic phase 2 trial of Mosun for previously untreated follicular (FL) or marginal zone lymphoma (MZL; Olszewski, ASH 2022). Pts receive fixed-duration Mosun for 8 three-week cycles (C) with disease assessments after C4 and C8; low-dose lenalidomide (Len) is added during C5-C8 if complete response (CR) is not attained mid-treatment (“MidTx”). Peripheral blood samples pre-treatment (“PreTx”); on day 8 of cycle [C] 1 (“C1D8”); day 1 of C2; MidTx (after C4); and 4-6 weeks after the end of treatment (“EOT”) were collected. Soluble plasma biomarkers related to CD8+ T-cell activation (GZMA, GZMB, PRF1, IL2, IL6, IL7, IL15, IFNg, TNFa, CCL3, CCL4, sCD137/4-1BB), modulation (IL10, IL13, IL2Ra, sCD27, TGFb1, FAS, and TNFa), and exhaustion (PD-1, TIM-3, LAG3, CTLA4, Galectin-9) were measured at each timepoint using a microsphere-based, multiplexed assay (Luminex 200). The cytokine assay was performed for 10 pts (8 FL / 2 MZL, 2 male / 8 female) PreTx, and longitudinal measurements for 7 pts. Three pts did not have CR MidTx and also received Len during C5-C8. CD8+ T-cell subsets were evaluated on freshly collected samples by flow cytometry for 34 study participants (21 FL / 13 MZL). Soluble biomarker levels were compared in mixed-effects log-gamma models, and differential cell type abundance in negative binomial models using edgeR-based moderated tests; uncorrected P-values are reported in this exploration.

Results: We observed distinct phases of systemic immune reaction over the 6 months of Mosun therapy (+/-Len). The immediate phase (C1D8) was characterized by a significant increase over baseline in T-cell activation markers, including IL2 (P=0.018), IL7 (P=0.027), GZMA (P=0.015), GZMB (P=0.018), as well as CCL3 (chemokine participating in T-cell recruitment, P=0.043) and IL13 (Th2 cell-derived, P=0.008). This effect subsided by the 3^rd week of therapy (C2) when only IL2 (P=0.043) and GZMB (P=0.046) remained significantly elevated. At C2, we observed an egress of CD8+ TEMRA (P<0.0001) and effector memory (EM) CD8+ T-cells (P=0.0035) from the blood, with an increase in central memory T-cells (P=0.0043).

At MidTx assessment, markers of CD8+T-cell activation normalized, while TIM-3 (P=0.020), LAG3 (P=0.028), and co-stimulatory sCD27 (P=0.041) increased. Increased expression of LAG3 (P=0.0038) and PD-1 (P=0.0019) on cells peaked at C2 and was particularly pronounced in a subset of CD27+ transitional EM T-cells (P<0.0001). At EOT, all plasma markers normalized, but the increased EM CD8+ T-cells persisted in the circulation (P=0.019).

All pts undergoing cytokine analysis attained a CR EOT. Early CR at C4 assessment was associated with a more robust early (C1D8) release of IL7, GZMA, GZMB, IFNg, and CCL3, and at MidTx—a persistent GZMA/PRF1/CCL3 elevation and lower CTLA4 (rank-sum test, P=0.034).

Conclusions: Our results suggest a model of evolving T-cell engagement during Mosun therapy, with an early cytotoxic activation, followed by accumulation of EM CD8+ T-cells bearing markers of potential exhaustion, and normalization of serum cytokines after cessation of BsAb therapy. Further evaluation of the detailed phenotype (by CyTOF) and transcriptome of the CD8+ T-cells during these phases is ongoing. The observed changes offer insight into future translational opportunities to maximize the anti-lymphoma activity of Mosun through rational combinations with immunomodulatory agents or alternative administration schedules.