Rachel Rose1, Osman Chaudhary2, Andrzej Kierzek1, Dennis Reddyhoff1, Aban Shuaib1, Piet van der Graaf1, Jochem Gokemeijer3, Vibha Jawa2, Pegy Foteinou2, Brian Schmidt2, Lora Hamuro2
1Certara UK Ltd, 2Clinical Pharmacology and Pharmacometrics, Bristol Myers Squibb, 3Molecular Discovery Technologies, Bristol Myers Squibb
Introduction: Immunogenicity (IG) is the ability of a substance such as a biotherapeutic to provoke an immune response, including anti-drug antibody (ADA) production. The development of ADAs is a substantial concern that influences both design and clinical development of new biotherapeutics. ADA can significantly alter the pharmacokinetics (PK), reduce efficacy and compromise safety. IG is currently mitigated in the preclinical phase by using in silico prediction tools and in vitro assays to guide protein engineering and compound selection, and in the clinic through dose optimisation, tolerance induction, close monitoring, and co-medication with immunosuppressants [1]. However, these mitigation strategies do not address biotherapeutics that are dosed in combination, where the IG risk can be even more uncertain through synergistic effects, particularly for immuno-modulatory drugs [2]. The IG Simulator is a quantitative systems pharmacology (QSP) model that integrates mechanistic models of the immune response and PK with in silico and experimental data from IG risk assessment, population-specific HLA allele frequencies and clinical study design factors including dosing regimens and co-medication [3,4]. The aim of the model is to predict ADA production and its impact to PK for management of clinical immunogenicity. Nivolumab, a PD-1 inhibitor, and ipilimumab, a CTLA-4 inhibitor, are immunotherapies approved for several oncology indications either as monotherapies or in combination. In the CheckMate-067 clinical trial for advanced melanoma (NCT01844505), treatment-emergent anti-nivolumab ADA incidence increased from 12.3% for nivolumab monotherapy to 44.0 % when administered with ipilimumab, with no impact of ADA on nivolumab PK, safety or efficacy [5,6]. While uncertainty exists for ipilimumab’s role in Treg depletion in the tumor microenvironment to enhance tumor immunity [7,8] there is evidence to suggest ipilimumab binding to CTLA4 on Tregs could increase CD4+ Tcell activation and antibody production and is a plausible mechanism to explain the increase in nivolumab ADA when dosed in combination [9,10,11,12]. A previously developed QSP IG model with suppressive Treg pathways was updated to test this hypothesis and to enable IG predictions and PK impact for biotherapeutic combinations Objectives: The objectives of this work are (1) to update the IG Simulator to model the contribution of CTLA4-dependent cell activities via Tregs to the development of nivolumab ADA, and (2) to use the updated model to test the hypothesis that inhibition of Treg proliferation by ipilimumab can explain increased nivolumab immunogenicity for combination versus monotherapy. Methods: The IG Simulator was expanded to include a mechanistic model of T regulatory cell activation and activity previously described in [13]. PK of nivolumab and ipilimumab were modelled using a physiologically-based PK (PBPK) model [14], with time-varying clearance captured by a sigmoid Emax function [15,16]. The inhibition of Treg proliferation rates as a result of CTLA-4 inhibition by ipilimumab was integrated using an inhibitory Emax model using the reported in vitro IC50 [10]. Results: The original model without Tregs, overpredicted the ADA incidence for nivolumab monotherapy by > 5- fold. Extension of the model to include Treg activation reduced the predicted ADA incidence to 11.2% transient ADA positivity, with no impact of ADA on PK following nivolumab monotherapy, consistent with clinical data. Concentration-dependent inhibition of T regulatory cells by ipilimumab resulted in an increase in predicted ADA incidence for combination therapy to 44%, in good agreement with the clinical study. Conclusion: Results from the model support that activation of inhibitory Tcell populations such as Tregs can attenuate CD4+ T cell activation and supports inhibition of Treg activity by ipilimumab as a potential mechanism to explain the increase in anti-nivolumab ADA incidence following co-therapy. This model framework can be utilized to evaluate other CTLA4-dependent effects on B cell activity and represents a significant advancement in developing a mechanistic basis for predicting synergistic ADA formation responses through the regulation of B cell activity. This understanding can inform predictions of clinical immunogenicity for other immunomodulatory biotherapeutics.
[1] Shakhnovich et al. (2020) Clin Transl Sci 13:219-223 [2] Zheng et al. (2022) Clin Transl Sci 15:2096-2104 [3] Kierzek et al. (2019) CPT:PSP 8:773-776 [4] Franssen et al. (2023) CPT:PSP 12:139-143 [5] https://www.fda.gov/news-events/fda-meetings-conferences-and-workshops/model-informed-drug-development-approaches-immunogenicity-assessments-06092021 [6] Zhao et al. AAPS J, submitted 2025 [7] Simpson et al. (2013) J Exp Med 210:1695-710 [8] Sharma et al. (2019) Clin Cancer Res 25:1233-1238 [9] Sage et al. (2014) Immunity 41:1026-39 [10] Williams et al. (2022) Front Immunol 13 :871802 [11] La Muraglia et al. (2021) Am J Transplant 21:73-86 [12] Keler et al. (2003) J Immunol 171:6251-9. [13] Hamuro et al. (2019) AAPS J 21: 94 [14] Li et al. (2014) AAPS J 16:1097-1109 [15] Zhang et al. (2019) (2019) CPT:PSP 8:962-970. [16] Sanghavi et al. (2020) CPT:PSP 9:29-39
Reference: PAGE 33 (2025) Abstr 11652 [www.page-meeting.org/?abstract=11652]
Poster: Drug/Disease Modelling - Other Topics