(T1430-02-09) Bridging Bevacizumab PK from Adults to Pediatrics through PBPK Modeling

Purpose: There are challenges to studying pharmacokinetics (PK) in pediatrics due to ethical, physiological, and operational factors (1).To address this, the FDA encourages careful use of extrapolation from adults to pediatrics (2). Towards this end, physiologically based pharmacokinetics (PBPK) modeling can be used to mechanistically extrapolate drug PK from adults to pediatrics and has shown successes in small-molecule drugs (3,4). However, such application for monoclonal antibodies (mAbs) remains limited. For mAbs, the Neonatal Fc Receptor (FcRn) plays a crucial role in its disposition in humans, resulting in a prolonged half-life of mAbs. This prolonged half-life is primarily due to the FcRn-mediated recycling process of IgG. Endogenous IgG is the only class of antibodies actively transported from mothers to neonates, resulting in short-term passive immunity for children. At birth, the concentration of endogenous IgG is as high as an adult due to the maternal transport, and declines over the first few months after birth, and then gradually climbs up as the child begins to develop its own. The dynamics of endogenous IgG demonstrate age-related variations, resulting in a nonlinear relationship (5). In pediatrics, the competition between IgG and mAb for FcRn binding may lead to higher exposure of mAb due to age-related endogenous IgG concentrations. Previously, a published atezolizumab pediatric PBPK model incorporated IgG concentration changes and competitive FcRn binding, accounting for age-related variations (6). The results showed that reduced IgG levels and competition for FcRn binding enhances atezolizumab recycling efficiency, leading to lower drug catabolism in younger children and higher exposure (6). The objective of the study is to evaluate the capability of the PBPK model to predict the exposure of mAb in adults and extrapolate to the pediatric population by incorporating the IgG-mAb competition for FcRn, utilizing pharmacokinetics data from bevacizumab clinical studies.

Methods: A minimal mechanistic PBPK model structure within Simcyp version 22 was used to develop the model for bevacizumab. The model consists of compound-specific and physiological parameters. Some parameters were derived using the Simcyp predictor, while others were obtained from literature (7). To evaluate the initial PBPK model performance, simulations were run using 100 virtual subjects who weighed 70 kg with different bevacizumab dose regimens (15 mg/kg every 3 weeks, 10 mg/kg every 2 weeks, and 5 mg/kg every 2 weeks). The exposure metrics (the geometric mean of AUCtau, Cmax, and Cmin) for Cycle one and steady state were compared with PopPK model-simulated exposure metrics. The population pharmacokinetics (PopPK) simulation was done with parameters obtained from literature using NONMEM version 7.5 (8). The PopPK model analysis showed that sex is a covariate that affects the pharmacokinetic of bevacizumab, with males having a higher clearance and volume of distribution than females. In the final PBPK model, an additional clearance parameter was included, which accounts for other elimination pathways with unknown mechanisms. The additional clearance parameter was optimized against the PopPK model simulated AUCtau at steady state. To further evaluate the final PBPK model performance, simulations were run on 100 virtual adult (aged >18) subjects based on bevacizumab PK data obtained from clinical studies. The simulations included single dose and multiple doses for different bevacizumab dose regimens. The clinical studies included a total of 786 adult subjects, from whom 7,750 PK samples were collected. Out of the 786 adults, 605 were used to validate the model. The subjects participating in the studies included individuals with breast cancer, non-small cell lung cancer, metastatic colorectal cancer, and pancreatic cancer (9). Patient demographics for males and females at different doses were also included in the simulations.

Results: As shown in Figure 1A, the initial PBPK model overpredicts the bevacizumab PK compared to the PopPK model-simulated exposure metrics. However, the final PBPK model, which includes an optimized additional clearance parameter, shows a significantly better agreement with PopPK simulation (Figure 1B), where five out of six exposure metrics ratios fall within the 0.8-1.25 fold range. In addition, the final PBPK simulation accurately captured the adult observed data following 15, 10, and 5 mg/kg bevacizumab single and multiple doses (Figure 2).

Conclusion: The Simcyp PBPK model accurately characterized the geometric mean of AUCtau, Cmax, and Cmin for both males and females by optimizing the additional clearance parameter through PopPK model-based simulations of 70 kg individuals. Additionally, the model was able to demonstrate its generalizability by accurately capturing the pharmacokinetic profile of bevacizumab across various dose regimens. For future considerations, the Simcyp PBPK model for bevacizumab will be evaluated for predicting bevacizumab PK in pediatrics (under 18 years of age). Furthermore, the relevance and implications of the IgG-mAb competition for FcRn in the context of bevacizumab will be evaluated. This will help us expand our understanding of the potential impact of mAb pharmacokinetics and therapeutic outcomes.

References: 1. Grimsrud, K. N., Sherwin, C. M., Constance, J. E., Tak, C., Zuppa, A. F., Spigarelli, M. G., & Mihalopoulos, N. L. (2015). Special population considerations and regulatory affairs for clinical research. Clinical research and regulatory affairs, 32(2), 47–56. https://doi.org/10.3109/10601333.2015.1001900
2. Kern S. E. (2009). Challenges in conducting clinical trials in children: approaches for improving performance. Expert review of clinical pharmacology, 2(6), 609–617. https://doi.org/10.1586/ecp.09.40
3. Zhuang, X., & Lu, C. (2016). PBPK modeling and simulation in drug research and development. Acta pharmaceutica Sinica. B, 6(5), 430–440. https://doi.org/10.1016/j.apsb.2016.04.004
4. Ince, I., Dallmann, A., Frechen, S., Coboeken, K., Niederalt, C., Wendl, T., Block, M., Meyer, M., Eissing, T., Burghaus, R., Lippert, J., Willmann, S., & Schlender, J. F. (2021). Predictive Performance of Physiology-Based Pharmacokinetic Dose Estimates for Pediatric Trials: Evaluation With 10 Bayer Small-Molecule Compounds in Children. Journal of clinical pharmacology, 61 Suppl 1(Suppl 1), S70–S82. https://doi.org/10.1002/jcph.1869
5. Pan, X., Stader, F., Abduljalil, K., Gill, K. L., Johnson, T. N., Gardner, I., & Jamei, M. (2020). Development and Application of a Physiologically-Based Pharmacokinetic Model to Predict the Pharmacokinetics of Therapeutic Proteins from Full-term Neonates to Adolescents. The AAPS journal, 22(4), 76. https://doi.org/10.1208/s12248-020-00460-1
6. Huang, W., Stader, F., Chan, P., Shemesh, C. S., Chen, Y., Gill, K. L., Jones, H. M., Li, L., Rossato, G., Wu, B., Jin, J. Y., & Chanu, P. (2022). Development of a pediatric physiologically-based pharmacokinetic model to support recommended dosing of atezolizumab in children with solid tumors. Frontiers in pharmacology, 13, 974423. https://doi.org/10.3389/fphar.2022.974423
7. Basu, S., Lien, Y. T. K., Vozmediano, V., Schlender, J. F., Eissing, T., Schmidt, S., & Niederalt, C. (2020). Physiologically Based Pharmacokinetic Modeling of Monoclonal Antibodies in Pediatric Populations Using PK-Sim. Frontiers in pharmacology, 11, 868. https://doi.org/10.3389/fphar.2020.00868
8. Han, K., Peyret, T., Quartino, A., Gosselin, N. H., Gururangan, S., Casanova, M., Merks, J. H., Massimino, M., Grill, J., Daw, N. C., Navid, F., Jin, J., & Allison, D. E. (2016). Bevacizumab dosing strategy in paediatric cancer patients based on population pharmacokinetic analysis with external validation. British journal of clinical pharmacology, 81(1), 148–160. https://doi.org/10.1111/bcp.12778
9. Han, K., Peyret, T., Marchand, M., Quartino, A., Gosselin, N. H., Girish, S., Allison, D. E., & Jin, J. (2016). Population pharmacokinetics of bevacizumab in cancer patients with external validation. Cancer chemotherapy and pharmacology, 78(2), 341–351. https://doi.org/10.1007/s00280-016-3079-6

Figure 1: Ratio of predicted PBPK to PopPK parameters for the evaluated bevacizumab dose regimens. The different colors represent the exposure metrics for Cycle one and steady state concentration. The Black dashed lines indicate the 0.5, 1-, and 1.5-fold prediction intervals. 1 representing a high correlation between the PBPK and PopPK simulations. The Red dashed lines indicate the 0.8- and 1.25-fold prediction intervals, representing the bioequivalence range.


Figure 2: The red circles represent the observed data  | The green lines represent 100 simulated individuals | The dashed black line represent the mean data for the simulated individuals | The green shaded area represent the 5th to 95th percentiles of the simulation