Abstract 2951: NI-2901, an affinity-optimized CD47xPD-L1 bispecific antibody for dual immune checkpoint blockade

To enhance efficacy of anti-PD-1/PD-L1 antibodies, many combinations with various therapeutic agents are being investigated. Blocking the CD47/SIRPα myeloid checkpoint with monoclonal antibodies (mAbs) or decoy receptors is emerging as an effective approach to mobilize dendritic cells and macrophages to support T-cell mediated antitumor responses. The benefit of combining CD47/SIRPα and PD-1/PD-L1 blockade to improve tumor control has been convincingly demonstrated in preclinical models and is now being explored in patients. However, CD47 mAbs are hindered by ubiquitous CD47 expression, leading to pharmacokinetic (PK) and safety issues.NI-2901, an IgG4 CD47xPD-L1 bispecific antibody (bsAb), was generated using the κλ-body platform. In vitro assays were used to characterize its binding profile and checkpoint inhibition as well as its capacity to enhance T-cell activation and macrophage-mediated phagocytosis of tumor cells. PD-L1-independent CD47 antitumor activity was assessed in vivo in a PD-L1-negative xenograft model and compared to the anti-CD47 magrolimab analog. PK and tolerability of NI-2901 were evaluated in non-human primates (NHP), allowing for translational modeling to predict PK and dosing regimens in humans. Consistent with its intermediate affinity to CD47, NI-2901 shows lower binding to RBC as compared to magrolimab analog and is still able to induce CD47/SIRPα blockade on PD-L1-negative tumor cells, that is significantly enhanced once PD-L1 is expressed. As a result, the bsAb is able to enhance the phagocytosis of PD-L1-negative and -positive tumor cell lines induced by mAbs targeting tumor-associated antigens (e.g. rituximab, trastuzumab and anti-CD19) and demonstrates in vivo activity in the Raji B-cell lymphoma xenograft model. Given its high affinity for PD-L1, NI-2901 triggers an effective blockade of the PD-1/PD-L1 interaction, inducing T-cell activation in vitro to a degree similar to anti-PD-L1 benchmark antibodies atezolizumab and avelumab. In immunocompetent huCD47/huSIRPα-transgenic mice engrafted with MC38 cells engineered to express human PD-L1 and CD47, NI-2901 displayed significant anti-tumor activity. In a NHP study, NI-2901 was well-tolerated after four weekly injections at 30mg/kg, showing no signs of hemotoxicity. In contrast, the magrolimab analog induced a significant drop in RBC already after a single injection at 10mg/kg. PK modeling and simulations in humans suggest a more favorable dosing regimen as compared to CD47 targeted approaches. In conclusion, NI-2901, a dual immune checkpoint inhibitor, triggered effective T-cell activation and enhanced phagocytosis of tumor cells. Also, NI-2901 demonstrated significant antitumor activity in vivo and is therefore expected to show improved clinical efficacy over PD-1/PD-L1 blockade alone. The bsAb was well-tolerated in NHP without inducing RBC or platelet depletion.

Citation Format: Xavier Chauchet, Sebastien Calloud, Pauline LLoveras, Nicolas Bosson, Margaux Legrand, Laurence Chatel, Laura Cons, Adeline Lesnier, Pauline Malinge, Guillemette Pontini, Christophe Guillamo, Dmitry Shchelokov, Oleg Demin, Ulla Ravn, Valéry Moine, Bruno Daubeuf, Giovanni Magistrelli, Yves Poitevin, Susana Salgado-Pires, Limin Shang, Nicolas Fischer, Walter Ferlin, Krzysztof Masternak. NI-2901, an affinity-optimized CD47xPD-L1 bispecific antibody for dual immune checkpoint blockade [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2951.

Receptor occupancy assessment and interpretation in terms of quantitative systems pharmacology: nivolumab case study

Receptor occupancy assays applied in clinical studies provide insights into pharmacokinetic-pharmacodynamic relationships for therapeutic antibodies. When measured by different assays, however, receptor occupancy results can be controversial, as was observed for nivolumab, a monoclonal antibody targeting programmed cell death 1 (PD-1) receptor. We suggested an explanation of results obtained and a mechanistic approach based on specific features of the receptor occupancy assays: measurement of the free or bound receptor, normalized to the baseline or at each time point. The approach was evaluated against controversial clinical data on PD-1 receptor occupancy by nivolumab. It was shown that receptor occupancy measured by different assays might vary substantially if the internalization rate of the bound receptor is higher than the rate of degradation of the free receptor. Equations proposed in this work can be applied in quantitative systems pharmacology models to describe target receptor occupancy by different therapeutic antibodies.

Abstract 5421: Prediction of intratumoral TIGIT receptor occupancy after the treatment with anti-TIGIT antibodies

Background: T-cell immunoreceptor with immunoglobulin and tyrosine based inhibitory motif domain (TIGIT) is a co-inhibitory immune checkpoint receptor expressed on several types of immune cells, which can suppress T-cell activation, promote T-cell exhaustion, and suppress natural killer cell mediated cytotoxicity. Recent clinical data with anti-TIGIT monoclonal antibodies (mAbs) indicate that TIGIT blockade is a highly promising therapy when combined with PD-1/PD-L1 blockade. However, unlike PD-1 receptor occupancy (RO), there is a lack of information regarding RO in peripheral blood and tumors at different dose regimens with anti-TIGIT therapies. This study aims to predict intratumoral RO for a series of anti-TIGIT antibodies with known pharmacokinetic (PK) and binding characteristics (ociperlimab [BGB-A1217], vibostolimab, domvanalimab, etigilimab, tiragolumab) utilizing a physiologically-based PK (PBPK)/RO model.

Methods: The PBPK/RO model describes biodistribution of mAbs within bodily fluids, detailed transport across the endothelial barrier, two-step binding with the membrane-bound TIGIT receptor (taking into account target expression level, number of cells expressing target receptor and internalization process), linear and non-linear clearance of mAbs (via uptake by endothelium and internalization of mAb:TIGIT complexes, respectively). Physiological parameters were taken from the literature, while other parameters were identified based on available in vitro and in vivo data. Clinical PK data of anti-TIGIT mAbs were used for model calibration.

Results: The model-predicted results for RO in peripheral blood demonstrated almost complete occupancy, which is supported by clinical data available for ociperlimab and domvanalimab (predicted 99.9% vs observed 100%). A similar tendency was observed for all studied cell types (CD8, CD4, regulatory T cells) despite the significant differences in TIGIT expression on different cell types. According to model predictions, the intratumoral RO was close to the values reported for blood over a range of doses close to the recommended Phase 2 dose (e.g. ociperlimab 900 mg Q3W median trough RO with 95% CI: blood 99.96% [99.89, 99.98] vs tumor 99.75% [98.88, 99.94]). The direct comparison of extended dosing regimens (e.g. ociperlimab 150/200/300 mg per week regimens) demonstrated a sustainable level of TIGIT blockade and comparable values of trough RO.

Conclusions: The PBPK/RO model accurately predicted the RO in peripheral blood and tumors for different anti-TIGIT mAbs by taking into account their PK and binding properties. Moreover, the model allowed a direct comparison of RO across different regimens and different anti-TIGIT mAbs. The predicted TIGIT receptor occupancy within the tumor may be useful for future dose selection or optimization in clinical trials.

Citation Format: Dmitry Shchelokov, Oleg Demin Jr, Oleg Demin, Ahsan Rizwan, Yun Zhang, Wei Tan, Srikumar Sahasranaman, Nageshwar Budha. Prediction of intratumoral TIGIT receptor occupancy after the treatment with anti-TIGIT antibodies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5421.

Abstract 5049: Prediction of pharmacokinetics and receptor occupancy of a bispecific antibody against CD40 and PD-L1

Background: Currently bispecific monoclonal antibodies gain traction in pharmacology as a way to simultaneously target different surface molecules on immune and tumor cells and modulate their activity. While antibodies against a single target, such as CD40 on antigen-presenting cells (APCs) and PD1 on T cells, are widely used in clinic to treat various forms of cancer, bispecific antibodies are still in the early stages of clinical development. The aim of this study was to predict pharmacokinetics (PK), receptor occupancy (RO) and optimal dose for a bispecific antibody YH008 against CD40 and PD1 using the PBPK/RO model.

Methods: The PBPK/RO model describes distribution of an antibody (mAb) between physiological compartments, transport across the endothelium, binding with CD40 on dendritic cells (DCs), B cells, and macrophages (Mph) and PD1 on CD4+ and CD8+ T cells in the immunological synapses (ISs). The model takes into account number of cells expressing target molecules and their ligands, expression levels, clearance of the antibody. Two scenarios of mAb binding with CD40 were considered: a case when the mAb competes with CD40L for the binding site and a case when the mAb binds to a different site. Binding affinities of YH008 for human CD40 and PD1 were taken from [1]. The model was validated using PK data on monospecific antibodies against CD40 and PD1. ROs of PD1 on CD4+ and CD8+ T cells and CD40 for Mph, DCs, B cells and percentages of the trimer complex (CD40-YH008-PD1) were calculated. Optimal dose of YH008 was selected on the basis of maximal exposure/area under curve (AUC) of the trimers in ISs between APCs and CD8+ T cells in the tumor.

Results: Model simulations show good agreement with the experimental PK and RO data for monospecific mAbs against PD1 (Pembrolizumab) and CD40 (Selicrelumab, BI 655064). For the bispecific antibody, the simulations show that competition for the CD40 binding site between CD40L and the mAb does not significantly affect PK and RO in blood or tissue. The model predicts 99% RO on antigen-presenting cells (APCs) and 98% RO on T cells in blood at dose 80 mg. In the tumor, 95% RO on APCs and 85% RO on T cells, as well as 5-to-6-fold decrease in the number of PD1-PDL1 complexes in ISs, are achieved at dose 240 mg. The dose dependence for YH008 was calculated, and the optimal dose based on the numbers of the trimers and signaling complexes in the ISs was estimated.

Conclusions: The PBPK/RO model for a bispecific mAb based on physiological levels of immune cells, target molecule expression, and binding parameters for YH008 antibody against CD40 and PD1 predicts PK and RO in peripheral blood and tumor. The model shows that competition for CD40 binding site between the mAb and CD40L does not significantly affect PK and RO. The optimal dose for maximal effect of YH008 was estimated.

Citation Format: Alexandra Diakonova, Oleg Demin, Dmitry Shchelokov, Oleg Demin. Prediction of pharmacokinetics and receptor occupancy of a bispecific antibody against CD40 and PD-L1 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5049.

Abstract 2233: Prediction and comparison of PD-1 receptor occupancy in the tumor after treatment with immune checkpoint inhibitors

Background: Measurements of receptor occupancy (RO) are believed to provide the necessary information on the pharmacodynamic efficacy of immune checkpoint inhibitors. However, only blood data are often available, whereas virtually nothing is known about RO at the site of action due to difficulties in tumor/tissue sampling. Besides, analysis of clinical data revealed highly controversial results of RO measurements for the same antibodies (e.g., maximal PD-1 occupancy by nivolumab 70-80% or 100% at the same doses) that raise the question about experimental procedures of RO determination and interpretation of the results. This work aims to predict and compare RO in the tumor for several anti-PD-1 therapeutics (nivolumab, pembrolizumab, INCMGA00012 (MGA012), dostarlimab (TSR-042), MEDI0680, sintilimab (IBI308), cemiplimab (REGN2810)) based on the previously developed PBPK/RO model.

Methods: Developed PBPK/RO model describes biodistribution of antibodies (mAb) within body fluids, detailed transport across endothelial barrier, two-step binding with membrane-bound PD-1 receptor (taking into account target expression level, number of cells expressing target receptor and internalization process), linear and non-linear clearance of mAb (via uptake by endothelium and internalization of mAb:PD-1 complexes, respectively). Physiological parameters were taken from literature, while other parameters were identified on the basis of available in vitro and in vivo data. Clinical data on PK of anti-PD1 mAbs, RO in blood and tumor were used only for model validation.

Results: Clinical PK data at various doses and regimens of anti-PD1 mAbs are captured by the model without any fitting. Moreover, the model allows to correctly describe heterogeneous data on RO in plasma using two different approaches: normalization to baseline quantity of receptor and normalization to the total quantity of receptor at each time point during measurements. To evaluate the impact of binding parameters (Kd, koff) on predicted trough PD-1 occupancy in tumor simulations were conducted over the wide range of Kd values at 3 mg/kg IV dosing Q3W. The model shows that mAbs against PD-1 receptor with a dissociation constant less than 3 nM demonstrate trough RO ≥ 90% at the chosen dosing regimen. Nevertheless, a direct comparison of approved doses for nivolumab (240 mg Q2W) and pembrolizumab (200 mg Q3W) revealed that predicted trough RO in the tumor is quite similar (93.98% vs 92.57%, respectively) despite the almost 100-fold difference in Kd values.

Conclusions: The proposed PBPK/RO model is able to predict PK of anti-PD-1 mAbs and RO in blood and tumor without the prior fitting of clinical data. Moreover, predictions of PD-1 occupancy in the tumor may be used for informed dose selection or explanation of clinical trial results. This approach can be expanded for mAbs targeting other receptors including bi- or multispecific antibodies with different structures and properties.

Citation Format: Dmitry Shchelokov, Oleg Demin Jr. Prediction and comparison of PD-1 receptor occupancy in the tumor after treatment with immune checkpoint inhibitors [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2233.