PI-055 - A SEMI-MECHANISTIC PHARMACOKINETIC/PHARMACODYNAMIC (PK/PD) MODEL OF ABBV-383 CHARACTERIZING THE INTERPLAY BETWEEN SOLUBLE BCMA (SBMCA) AND FREE AND TOTAL PK.

V. Schmitt¹, B. Engelhardt², S. Mensing², A. Polepally³; ¹Abbvie, Ludwigshafen am Rhein, Germany, ²Abbvie, Ludwigshafen am Rhein, Rheinland-Pfalz, Germany, ³Abbvie, South San Francisco, California, United States.

Background: ABBV-383 is a differentiated BCMA x CD3 bispecific antibody T-cell engager composed of a bivalent BCMA binding domain with high avidity, a low-affinity CD3-binding domain, and a silenced Fc tail retaining the FcRn binding for an extended half-life. Here, results from a semi-mechanistic PK/PD model characterizing the impact of sBCMA on PK and its interplay with free and total PK are reported.

Methods: ABBV-383 total and free PK as well as sBCMA data were obtained from a phase 1 trial in patients with relapsed and refractory multiple myeloma (MM) (N=215). The study evaluated intravenous administration of 0.025 to 120 mg Q3W and 60 mg Q4W. The PK/PD model was developed in NONMEM 7.5 using the SAEM estimation method.
The model comprised 3 moieties, each described by a 2-compartment model with shared PK parameters, of ABBV-383 based on binding to sBCMA: 1) free, unbound to sBCMA, 2) partially free, bound to 1 sBCMA, and 3) fully bound, bound to 2 sBCMAs. The interaction was solely based on sBCMA. First-order proliferation and ABBV-383 dependent elimination was assumed for sBCMA dynamics. Patient-specific covariates were tested on relevant model parameters. Predictive performance was assessed by absolute individual prediction errors (|PE|s) from numerical predictive checks (NPCs) and external validation by predicting total PK using only free PK (including sBCMA data), and vice versa, from a study which was not included in the model development.

Results: Data consisted of > 7,000 each total and free and 1,596 sBCMA concentrations. The parameters CL and Vc estimated at 0.205 L/day and 3.72 L, respectively. Body weight, baseline albumin, ORR, sex, and IgG MM type were significant covariates. The dissociation rate of 1 sBCMA was fixed to 86.5/day. NPCs indicated |PE|s < 16% for free and total exposures, and ~ 40% for sBCMA. The higher |PE|s for sBCMA could be attributed to limited sampling, measurement of only total sBCMA, and highly variable data. External validation showed that observed median falls within the 95% prediction interval for free and total PK.

Conclusion: The semi-mechanistic PK/PD model adequately described the free and total ABBV-383 PK as well as sBCMA data with reasonable predictive performance. The model can be used to predict total PK based on observed free PK (including sBCMA data), and vice versa.