In this work, we show that the combination of radiotherapy (RT) and an IL15/IL15R alpha fusion complex (IL15c) fails to confer antitumor efficacy; however, a CD8-driven antitumor immune response can be elicited with the concurrent administration of an aCD25 regulatory T cell-depleting antibody. Using IL15-/- and Rag1-/- knockout mouse models, we show that the response to RT + IL15c + aCD25 is dependent on both IL15 and cytotoxic T lymphocytes. Furthermore, despite an equivalent survival benefit following treatment with RT + IL15c + aCD25 and combination RT and PD1-IL2v, a novel immunocytokine with PD1- and IL2R beta gamma-binding domains, cytotoxic T lymphocyte immunophenotyping and phosphoproteomics analysis of intracellular metabolites showed a significant upregulation of activation and functionality in CD8 T cells in the RT + PD1-IL2v regimen. Finally, we show that in the absence of functional IL15 signaling, the immunostimulatory response to RT + PD1-IL2v is significantly diminished with a concurrent lack of TCF+ CD8 T-cell generation, suggesting a necessity of IL15 for CD8 stem cells in mediating a durable response to treatment. Together, our results are illustrative of a mechanism wherein unimpeded effector T-cell activation through IL2R beta signaling and regulatory T-cell inhibition are necessary in mediating an antitumor immune response.

The development of anti-CD3 antibody-based T cell engager therapeutics has improved the treatment of various malignancies, yet the challenge of achieving tumor-specific targeting while minimizing on-target off-tumor effects in normal tissues remains a substantial hurdle. One promising strategy to address this issue involves engineering antibodies with conditional pH-dependent binding affinities, capitalizing on the acidic microenvironment characteristics of tumors (pH ~ 6.5-6.8) compared to the neutral pH of healthy tissues (pH ~ 7.4). In this study, we focus on the pH-engineering of antibody binders against the human CD3 antigen, a critical component of T cell activation, to achieve preferential binding at acidic pH. Using molecular dynamics (MD) simulations on the reported CD3ɛ antibody binder 40G5c, we shed light on possible molecular mechanisms of the pH-responsiveness of key mutations and their impact on the overall binder structure at physiological or acidic pH. Our study highlights how MD has emerged as a powerful tool to guide and explain intrinsic pH-dependent molecular mechanisms in antibody engineering. Lastly, we report that our engineered CD3 binders preferentially bind and activate T cells under acidic pH conditions and display favorable affinity and pH-window profiles.

Introduction Pretargeted radioimmunotherapy (PRIT) aims to improve the therapeutic index of systemic radiotherapy, typically involving a tumour-targeting bispecific antibody (BsAb) followed by a rapidly clearing small radiolabelled molecule. This strategy minimises healthy tissue exposure while increasing the absorbed dose to tumours. Rapid radioligand pharmacokinetics are particularly critical for efficacy and tolerability when using short-lived radionuclides like 212Pb (t1/2 = 10.6 h). While earlier PRIT regimens required an intermediate clearing agent to neutralise circulating BsAb, efforts are now focusing on developing clearing agent-independent approaches to decrease logistical complexity and mitigate safety risks. Objectives We developed a novel two-step, clearing agent-independent PRIT regimen for carcinoembryonic antigen (CEA)-positive tumours. This regimen involves a complementary SeParated v-domains LInkage Technology (SPLIT) antibody pair and a 212Pb radioligand. We asse Materials and Methods Two human SPLIT antibodies were developed that bind specifically and bivalently to human CEA, each fused to one half of a split high-affinity sub-pM 1,4,7,10-Tetrakis(carbamoylmethyl)1,4,7,10-tetraazacyclododecane (DOTAM) antibody variable region fragment: one antibody with the variable heavy (VH) and the other with the variable light (VL) domain. Bound to CEA on the cell surface, these split VH/VL domains assemble to form the active binding site for the subsequently administered therapeutic radioligand, 212Pb-DOTAM, capturing it significantly more efficiently than circulating SPLIT antibodies. To assess the functionality of this approach, we treated mice bearing subcutaneous CEA-expressing BxPC3 xenografts with the two pretargeting SPLIT antibodies, followed 7 days later by 212Pb-DOTAM. Another group of mice received a CEA-DOTAM BsAb, followed 7 days later by a dextran-based clearing agent to neutralise circulating BsAb, and then 212Pb-DOTAM after 24 hours. The two- and three-step CEA-PRIT regimens were compared for 212Pb biodistribution, tumour growth inhibition, and tolerability after three treatment cycles of 0.74 MBq (20 μCi). Results Both CEA-PRIT regimens achieved significant and comparable tumour growth delay. Only mild transient body weight loss was observed in both groups, confirming comparable tolerability. Excellent 212Pb tumour specificity was confirmed in both regimens. Tumour uptake for the two-step PRIT was 25–30% IA/g at 24 h p.i., compared to 36–45% IA/g for the three-step PRIT. Importantly, blood and kidney retention remained low for both regimens (<0.5% IA/g and <2 %IA/g, respectively). Conclusion The novel two-step CEA-targeted SPLIT PRIT approach demonstrated therapeutic efficacy and tolerability comparable to the three-step regimen. Biodistribution data confirmed the preferential retention of 212Pb-DOTAM in tumours and successful rapid excretion of the non-tumour bound radioligand. This optimised SPLIT PRIT approach offers potential improvements in clinical PRIT implementation, and its translation to a Phase I trial for patients with metastatic colorectal cancer is currently underway. Funding Acknowledgements This research was funded by and performed as a collaboration between F. Hoffman-La Roche and Orano Med.

Chimeric antigen receptor (CAR)-T cells represent a recent clinically validated modality in cancer therapy and beyond. However, broad industrial implementation faces technological, logistical, regulatory and financial challenges. A major bottleneck is the ex vivo production process, where cytokines are essential and product-determining constituents. To better understand the complexity of cytokine utilization throughout the production process, we rigorously reviewed the existing literature, including extended manufacturing parameters from 292 available clinical reports. We found a progressive reduction of interleukin-2 exposure, driven by its association with unfavourable cell characteristics. Preclinically, this catalysed the evaluation of alternative cytokines with potential to preserve naive phenotypes, support cell expansion and enhance the efficacy of CAR-T cells. Here we illustrate the evolving use of cytokines, reveal non-standardized clinical CAR-T cell manufacturing parameters, and summarize preclinical and next-generation concepts, which may improve manufacturing efficiency, cost-effectiveness and therapeutic outcomes. Our observations may further guide the development of cytokine-armouring strategies to promote in-patient expansion and persistence, even as innovations such as shortened manufacturing and in vivo engineering techniques could reduce reliance on cytokines during ex vivo culture.

Abstract Background Tumor heterogeneity has been identified as a major roadblock for cancer immunotherapy. To overcome this, universal effector cell engagers with interchangeable tumor-targeting adaptors have been developed. Current concepts in this field consist of antibody-based adaptors. Small-molecule (SM) ligands, on the other hand, infiltrate tissues rapidly, have short half-lives, and are potentially orally available. Therefore, we hypothesized that utilizing SM adaptors, combined with effector antibodies, could represent an attractive off-the-shelf therapy. Methods Here, we introduce the development of target-agnostic, small-molecule-guided hapten- and T cell-bispecific (TCB) antibodies with high affinity between the adaptor-effector pair. Specifically, we designed SM adaptors based on known tumor-targeting ligands with specificity to the antigens folate receptor 1 (FOLR1), prostate-specific membrane antigen (PSMA) and carbonic anhydrase IX (CAIX), and conjugated them to Ca2+-loaded 1,4,7,10-tetrakis(carbamoylmethyl)-1,4,7,10-tetraazacyclododecane (DOTAM) as a hapten. As an effector antibody, we designed DOTAM-specific T cell bispecific antibodies (DOTAM-TCB) with femtomolar (fM) affinity for the hapten. Results In vitro experiments showed in-solution assembly of stable adaptor-effector complexes. Additionally, T cell-mediated tumor killing was initiated by the universal DOTAM-TCB when combined with adaptors across various tumor cells. Ex vivo study of CAIX-directed Ca-DOTAM-acetazolamide (Ca-DOTAM-AAZ) with DOTAM-TCB showed potent activation of T cells against murine-engrafted human HT-29 tumor. Conclusions The studies described here demonstrate proof-of-concept for using hapten-containing small molecules as adaptors for effective universal T cell engager-based cancer immunotherapy.

Dose–response curves of PD-1 and LAG-3 blockade in the PD-1/LAG-3 Jurkat reporter cell line.

BACKGROUND:This first-in-human clinical study explored lomvastomig, an immunoglobulin G1-based Fc-silenced bispecific antibody that simultaneously blocks the immune checkpoint receptors programmed cell death protein 1 (PD-1) and T-cell immunoglobulin domain and mucin domain-3. METHODS:Lomvastomig was characterized in cell cultures and preclinically in cancer mouse models. The phase 1, open-label, multicenter clinical study of lomvastomig included a dose-escalation part in patients with advanced and/or metastatic solid tumors and an expansion part with four tumor-specific cohorts, which enrolled checkpoint inhibitor (CPI)-experienced patients with melanoma and non-small-cell lung cancer (NSCLC) and CPI-naïve patients with SCLC and esophageal squamous cell carcinoma (ESCC). Primary and secondary objectives included safety/tolerability, maximum tolerated dose (MTD)/recommended dose for expansion (RDE), pharmacokinetics, drug receptor occupancy, and antitumor activity. RESULTS:39 and 95 patients were enrolled in the dose-escalation and expansion parts, respectively. Lomvastomig was well tolerated up to the highest tested dose of 2,100 mg every 2 weeks (Q2W). One dose-limiting toxicity was reported at 1,200 mg (grade 3 troponin T increase). No MTD was reached, and 2,100 mg Q2W was established as the RDE. Linear pharmacokinetics across the studied dose range suggested target saturation. Peripheral blood drug receptor occupancy on CD3+ and CD8+ was saturated at >90% throughout treatment for doses ≥70 mg. Objective responses were observed at 2,100 mg lomvastomig during dose-escalation (21%; n=19), and in the CPI-experienced melanoma (8%, n=38) and CPI-naïve ESCC (20%, n=15) expansion cohorts. CONCLUSIONS:Lomvastomig had a tolerable and manageable safety profile at 2,100 mg Q2W. Clinical activity was limited in CPI-experienced patients with melanoma and NSCLC, while an encouraging signal was observed in CPI-naïve patients with ESCC. TRIAL REGISTRATION NUMBER:NCT03708328 (registration date: 2018-10-09).

Abstract Background: Alpha particles (α) are exceptionally cytotoxic, inducing complex DNA damage and bystander effects. Targeted α-therapy demonstrated a favorable therapeutic index (TI) in certain cancers. To improve the TI and expand to other indications, we developed a 212Pb-based carcinoembryonic antigen-related cell adhesion molecule 5 pretargeted radioimmunotherapy (CEA-PRIT 2.0), involving two complementary SeParated v-domains LInkage Technology antibodies (SPLIT Abs) and 212Pb-DOTAM. Each CEA-targeted SPLIT Ab carries half of a DOTAM binding v-domain that, when combined upon target binding, forms concentration-dependent stable complexes with 212Pb-DOTAM. While unbound 212Pb-DOTAM undergoes rapid renal clearance, 212Pb-DOTAM captured by the SPLIT Abs leads to α emission at CEA-expressing cells with minimal systemic irradiation. Here, we report key preclinical data informing a planned first-in-human (FIH) CEA-PRIT 2.0 study in metastatic colorectal cancer (mCRC) patients. Methods: We assessed CEA-PRIT 2.0-induced tumor growth inhibition (TGI), biodistribution, and tolerability in 3 CEA-expressing human xenograft models. SCID mice bearing BxPC3 (pancreatic) or LS174T (colorectal) tumors received up to 5 treatment cycles, and BRGS-CD47 mice (humanized and non-humanized) bearing HPAF-II (pancreatic) tumors received up to 3 cycles. Each cycle consisted of the two SPLIT Abs (1-5 mg/kg each) given on day 1 to allow for accumulation on CEA-expressing cells before giving 212Pb-DOTAM (20 µCi) on day 8. Results: Average 212Pb tumor uptake in all models was 10-43% injected activity per gram of tissue (IA/g) at 24 h, with low blood and kidney retention (<3% IA/g). In the BxPC3 model, TGI was SPLIT Ab dose-dependent, while the LS174T model showed potent TGI already at the lowest dose (1 mg/kg). The SPLIT Ab dose-dependence of the TGI in the HPAF-II model was abrogated in huBRGS-CD47, suggesting secondary immune responses contribute to the therapeutic effect.Manageable body weight (BW) loss was observed in the SCID BxPC3 model. More pronounced BW loss in LS174T (SCID) and HPAF-II (BRGS-CD47/huBRGS-CD47) models was observed, but minimal BW gain or actual BW loss in controls suggests tumor burden toxicity or strain-specific sensitivity were contributing factors.Utilizing the preclinical data, we designed a FIH study, starting with 203Pb-DOTAM as a surrogate for therapeutic 212Pb-DOTAM tumor uptake and healthy tissue distribution, SPLIT Ab pharmacokinetics, and tumor CEA expression to inform optimal SPLIT Ab dosing and interval to 212Pb-DOTAM, before initiating the 212Pb-DOTAM activity escalation and potential cancer immunotherapy combinations. Conclusions: CEA-PRIT 2.0 showed favorable tumor-to-healthy tissue radiation exposure and potent TGI with a favorable toxicity profile. A FIH study in mCRC is planned in H1 2026. Citation Format: Sofia H. Frost, Alexandre Pichard, Annabelle Mouchotte, Agnès Colmont, Sara Colombetti, Alexander Haas, Hans Peter Grimm, Birgit Kittel, Stephen Fowler, Bernhard Reis, Vincent Wolowski, Uta Sweere, Michael Hettich, Wolfgang Jacob, Frederic Prince, Christian Klein, Pablo Umana, Julien Torgue, Axel Boehnke. 212Pb-based CEA pretargeted radioimmunotherapy demonstrates tumor targeting and potent TGI in immunodeficient and humanized mouse models, informing a FIH study in mCRC [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 7203.

The combination of systemic immune checkpoint inhibitors with standard of care intravesical Bacillus Calmette-Guerin (BCG) therapy for non-muscle invasive bladder cancer (NMIBC) has shown potential in enhancing BCG therapeutic effects. However, challenges such as disease recurrence, progression, and immune-related toxicities remain a significant hurdle. Furthermore, the mechanism of action of BCG involves broad, non-specific immune activation. While this recruits bystander CD8+ and NK cells, it also triggers the undesired expansion of Tregs, which may ultimately hamper therapeutic efficacy. We aimed to explore the feasibility of a targeted approach designed to overcome intravesical tumor resistance and selectively enhance the expansion and effector functions of CD8+ TILs. We tested intravesical administration of BCG with murinized PD1-IL2v, a fusion protein that simultaneously targets PD-1 and IL-2Rβγ in cis on the same cell, in a preclinical mouse model of NMIBC. Both BCG monotherapy and co-treatment with a murinized PD1-IL2v improved animal survival. Notably, the intravesical combination of BCG and murinized PD1-IL2v significantly increased the number of CD8+ TILs with polyfunctional cytotoxic effector phenotype and avoided Treg expansion in contrast to BCG monotherapy. In addition, in patient-derived tumors, we observed a higher frequency of CD8⁺ TILs (18

IntroductionImmune checkpoint inhibitors (ICIs) targeting the PD-1/PD-L1 axis have shown considerable promise as a therapeutic modality in oncology. Despite their ability to target stem-like CD8+ T cells and give rise to exhaustion-fated effector CD8+ T cells, a significant subset of patients do not respond or eventually develop resistance, highlighting the need for more efficacious therapies. Eciskafusp alfa (PD1-IL2v) is a novel immunocytokine, engineered for avidity-driven, cis-delivery of IL-2R agonism to PD-1+ cells.MethodsThis study provides a comprehensive ex-vivo characterization of PD1-IL2v’s target landscape using matched peripheral blood mononuclear cells (PBMCs) and tumor-infiltrating lymphocytes (TILs) from patients across seven solid tumor indications.ResultsWe confirmed that the TIL compartment is significantly enriched with both stem-like CD8+ T cells and immunosuppressive regulatory T cells (Tregs). Notably, PD-1 receptor density was increased up to three-fold on CD8+ TILs compared to PBMCs, establishing the basis for preferential intra-tumoral targeting. Ex-vivo assays demonstrated that PD1-IL2v preferentially targets CD8+ TIL subsets (stem-like and effector) over Tregs. This preferential targeting translated into superior biological activity, with PD1-IL2v inducing higher STAT5 phosphorylation (STAT5-P) in stem-like and effector CD8+ T cells compared to Tregs, confirming the intended cis-targeting and enhanced IL-2R agonism. DiscussionThese findings provide translational validation for PD1-IL2v’s mechanism, demonstrating selective intra-tumoral immune stimulation while minimizing Treg activation. This characterization identifies PD-1 receptor density and subset prevalence as critical factors for drug activity and represents potentially useful biomarkers for predicting patient responsiveness and guiding patient selection.

Antisense oligonucleotides (ASOs) represent a promising therapeutic modality for central nervous system (CNS) disorders, offering highly specific modulation of gene expression. However, their clinical utility is severely limited by their inability to cross the blood-brain barrier (BBB), necessitating effective shuttling strategies. While transferrin receptor (TfR1)-mediated shuttling has shown therapeutic promise, the fundamental mechanisms governing the delivery of antibody-ASO conjugates across the BBB remain poorly understood. This study directly addresses this critical knowledge gap by establishing a mechanistic understanding of how the ASO cargo impacts major cellular interactions during the Brainshuttle (TM)-mediated transport across the BBB. Using a panel of advanced in vitro assays developed specifically for this purpose, including quantitative transcytosis, detailed imaging-based intracellular trafficking, and binding assays with brain endothelial cells (BECs), the shuttling process was systematically investigated. We demonstrate that ASO conjugation profoundly alters the cellular fate of the Brainshuttle (TM). Specifically, conjugation increased the binding to BECs of low-affinity TfR1 shuttles via avidity effects while paradoxically reducing the binding strength of high-affinity shuttles. Functional assays confirmed the biological activity of the delivered ASOs; however, transcytosis of high-to-moderate affinity binders across the BBB model was significantly delayed upon ASO conjugation. Building on these mechanistic insights, we engineered TfR1 Brainshuttles (TM) with optimized affinity and explored the shuttling potential of an alternative BBB receptor, CD98hc. These efforts culminated in the development of a novel bispecific Brainshuttle (TM) targeting both CD98hc and TfR1. This dual-targeting strategy exploits distinct and potentially non-competing trafficking pathways to overcome ASO-induced delays and significantly enhance in vitro transcytosis efficiency. The in vitro findings in this study underscore the necessity of mechanism-driven design to overcome ASO-induced limitations in delivery across the BBB. The bispecific CD98/TfR1 approach presented here provides a promising new strategy for maximizing delivery efficiency and enabling more effective therapeutic outcomes for CNS diseases.

T-Cell Engagers (TCEs) have recently transformed the therapeutic landscape of hematological malignancies, including relapsed or refractory B-cell non-Hodgkin lymphoma (R/R B-NHL). However, the variability in patient responses underscores the need for a deeper mechanistic understanding of the factors driving efficacy. Immune cell composition and T-cell functional states are emerging as critical determinants of immunotherapy outcomes. Recent advances in scRNA-seq technologies have enabled high-resolution characterization of T-cell states, revealing a spectrum from highly activated effectors to exhausted or dysfunctional subsets within the tumor microenvironment (TME). In this study, we conducted longitudinal scRNA-seq analyses and functional assessments of peripheral blood immune cells (PBMC) from glofitamab-treated R/R B-NHL patients, achieving complete metabolic responders (CMR) or progressive metabolic disease (PMD). Our findings reveal that the maintenance of naive-like (“fresher”) T-cell states (particularly the fresher cytotoxic T cells) at early timepoints is associated with clinical efficacy. In line with molecular data, T cells from responders exhibited enhanced functional activity compared to non-responders. Furthermore, the analysis of patient PBMCs and intra-tumor T cells from preclinical tumor models after consecutive glofitamab treatments revealed sustained functional activity, underscoring the long-term durability of T-cell responses. Combination of glofitamab with 4-1BB co-stimulation translated into increased proportions of intra-tumor T cells having a “fresher”, naive-like phenotype, ultimately leading to stronger anti-tumor efficacy. Taken together, our findings underscore the therapeutic relevance of “fresher” naive-like T-cell states and the potential of leveraging 4-1BB co-stimulation to overcome TCE resistance and improve clinical responses in aggressive lymphomas.

e14514 Background: Chimeric antigen receptor (CAR) T cell therapy has transformed the treatment of hematologic malignancies but remains less effective in solid tumors such as lung cancer due to antigen heterogeneity, on-target/off-tumor toxicities, and immunosuppressive tumor microenvironments (TME). Adapter CAR systems offer a strategy to improve flexibility and toxicity control by separating antigen recognition from CAR signaling. Anti-P329G adapter CAR T cells recognize a P329G point mutation in the Fc region of engineered human IgG1 antibodies, permitting precise and reversible retargeting through P329G-Fc–mutated antibodies against multiple tumor antigens. This work investigates the anti-P329G CAR platform in lung cancer, emphasizing its performance against the clinically relevant and therapeutically validated target EGFR. Methods: Anti-P329G CAR T cells demonstrated strong, antibody-dependent in vitro activation and effector functions against EGFR-positive lung cancer cells and recombinant EGFR protein. CAR T cell activity required presence of the P329G-Fc–engineered antibody, confirming reversible and modular antigen control. To further explore controllability, antibody displacement experiments with mock binders were conducted to assess reversibility. In vivo, EGFR-redirected anti-P329G CAR T cells induced significant tumor regression and prolonged survival in lung cancer–bearing mice without systemic toxicity. In both TCS-PCLS and TD-PCLS, EGFR-directed anti-P329G CAR T cells effectively reduced EdU+ tumor proliferation and enhanced TUNEL+ apoptosis, while sparing adjacent nonmalignant tissue despite EGFR co-expression. Results: Anti-P329G CAR T cells showed strong, antibody-dependent in vitro activation and effector functions against EGFR-positive tumor cells and recombinant EGFR protein. Activity required the P329G-Fc–engineered antibody, confirming modular and reversible antigen control. To explore controllability, antibody displacement experiments with mock binders were performed. In vivo, EGFR-redirected anti-P329G CAR T cells induced significant tumor regression and prolonged survival in lung cancer–bearing mice without systemic toxicity. In both TCS-PCLS and TD-PCLS, EGFR-directed anti-P329G CAR T cells reduced EdU+ tumor proliferation and increased TUNEL+ apoptosis, while sparing nonmalignant lung tissue despite EGFR co-expression. Conclusions: The anti-P329G adapter CAR platform provides a tunable, specific, and reversible approach for EGFR-targeted immunotherapy in lung cancer. Demonstrated efficacy across in vitro, in vivo, and ex vivo models, together with feasibility testing of antibody displacement, highlights its translational potential to overcome key barriers of CAR T therapy in solid tumors while maintaining improved safety and controllability.

Tobemstomig provides TGI and eradication in several mouse tumor models by favoring the expansion of cytotoxic CD4 and CD8 TILs over Tregs. In vivo efficacy studies in CD34+ humanized NSG mice bearing subcutaneous or intramammary fat pad tumors treated for 4 weeks with the indicated treatments. A, Tumor growth curves of WSU subcutaneous tumors in humanized control mice and mice treated with the indicated therapies (n = 10 mice per treatment group, mean ± SEM). aPD-1, anti–PD-1. B, Tumor growth curves of OCI-Ly18 subcutaneous tumors in humanized control mice and mice treated with the indicated therapies (n = 14 mice per treatment group, mean ± SEM). C, Tumor growth curves of BC004 intramammary fat pad tumors in humanized control mice and mice treated with the indicated therapies (n = 14 mice per treatment group, mean ± SEM). D, Tumor growth curves of BxPC-3 subcutaneous tumors in humanized control mice and mice treated with the indicated therapies (n = 10 mice per treatment group, mean ± SEM). E, Representative contour plot depicting the effect of different treatments on CD4 and CD8 frequencies (gated on CD3+ T cells), Tregs (FOXP3+) vs. cytotoxic (granzyme B+) CD4 TILs (gated on CD4+ T cells), and proliferating (Ki67+) vs. cytotoxic (granzyme B+) CD8 TILs (gated on CD8+ T cells) isolated from BxPC-3 tumors. F–Q, Immunopharmacodynamic study depicting the effect of different treatments on CD45+ cells, CD4 and CD8 frequencies (gated on CD45+CD3+ T cells), Tregs (FOXP3+) vs. cytotoxic [granzyme B+ (GrzB+)] CD4 TILs, and proliferating (Ki67+) vs. cytotoxic (granzyme B+) CD8 TILs, expressed either as frequencies within the CD4+ or CD8+ T-cell gate, respectively (I, K, and O–Q), as frequencies within CD45+CD3+ T cells (J and L), or as intratumoral ratios (M and N), isolated from BxPC-3 tumors. The Kruskal–Wallis with Dunn post hoc test was used to compare TGI across treatment groups in the various mouse tumor models. The Mann–Whitney U test was used to compare TGI between tobemstomig and parental anti–PD-1. One-way ANOVA with the Tukey multiple comparison test was used to compare treatment effects on TIL subsets (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

ABSTRACT:T-cell engagers (TCEs) are transformative therapeutics in hematologic malignancies, including non-Hodgkin lymphoma. Initially approved for relapsed/refractory disease settings, TCEs are now explored in first-line and second-line settings, often combined with standard-of-care (SOC) treatments, including chemotherapy and antibody-drug conjugates. This study investigates glofitamab (CD20×CD3 TCE) combinations in preclinical humanized lymphoma models, addressing heterogeneity of tumor antigen expression, immune evasion, and T-cell exhaustion. Combining glofitamab with R-CHP-Pola (rituximab, cyclophosphamide, doxorubicin, prednisone, and polatuzumab vedotin) chemotherapy or Pola demonstrated strong synergistic antitumor efficacy with rapid tumor regression and reduced tumor cell proliferation. Glofitamab combination with gemcitabine/oxaliplatin also demonstrated strong efficacy, enhancing intratumor T-cell number, activation, and reduced exhaustion. These combinations were particularly advantageous in models with low and heterogeneous CD20 expression, facilitating rapid tumor debulking and elimination of CD20-low/CD20- cells. Translational studies with patient-derived peripheral blood mononuclear cells receiving glofitamab combination with chemotherapies demonstrated sustained T-cell functionality throughout extended treatment cycles. Novel chemotherapy-free combinations, including CD19-targeted 4-1BBL and CD19-CD28, amplified glofitamab activity, especially in CD20 high- and homogenous-expressing tumor models, with dual costimulatory approaches revealing synergy. In addition, the combination with checkpoint inhibitors (programmed cell death protein 1/Lag3-bispecific antibody) and regulatory T-cell depletion (α-CD25) emerged as promising approaches for enhanced efficacy and to sustain T-cell functionality. These findings highlight the versatility of glofitamab when integrated with SOC and innovative combinations, addressing resistance and improving patient outcomes. The preclinical investigations provide a strong foundation for ongoing and future clinical trials, emphasizing the need to tailor TCE-based combination therapies to maximize efficacy while minimizing toxicity in lymphoma treatment. These trials were registered at www.clinicaltrials.gov as #NCT04408638 and NCT03467373.

Supplementary Figure S10. Single cell RNA-sequencing of whole blood treated with 0.2 μg/mL CD20-TCB was performed using the BD Rhapsody platform

T cell bispecific antibodies (TCBs) have demonstrated promising results in patients with solid tumors, yet the immunological mechanisms influencing their efficacy require further investigation. T cell exhaustion, induced by prolonged antigen exposure, is known to compromise T cell-based immunotherapies, but its effect on TCB efficacy remains unclear. Herein, we assessed the TCB efficacy on tumor-specific T cells, emphasizing their functional status. Utilizing an immunocompetent mouse model with melanoma expressing an immunogenic antigen, we showed that tumor-specific T cells acquire an exhausted phenotype and fail to expand under TCB treatment. Both mouse and human tumor-specific T cells in vitro demonstrated that chronically stimulated T cells exhibit a reduced response to TCBs. The comparison of TCB efficacy in T cell-inflamed versus non-inflamed tumors in mice revealed TCB success depends more on T cell functional fitness than their initial abundance. These data underscore the importance of T cell exhaustion, suggesting that exhausted tumor-specific T cells are unlikely to be the primary effectors redirected by TCBs for tumor eradication. Our study highlights the need to maintain T cell fitness and prevent exhaustion to enhance TCB therapy outcomes, which may help identify patients who could benefit most from TCB treatments in clinics.