Abstract

Background and aims: LY2189102, a humanized neutralizing IL-1β antibody, was studied in type 2 diabetes mellitus (T2DM) patients with C-reactive protein (CRP) ≥ 2 mg/L, who received weekly subcutaneous doses of LY2189102 (0.6, 18 or 180 mg) or placebo over 12 weeks, and were monitored for 12 additional weeks. This report describes the selection of dose level and frequency of LY2189102 treatment using modeling and simulation of LY2189102 PK/PD relationships.

Materials and methods: Data from 106 patients were used in this analysis. A 2-compartment mammillary model with dosedependent first-order input and bioavailability was fitted to pooled PK data from this study and a previous study conducted in rheumatoid arthritis patients receiving LY2189102 intravenously (IV). Bioavailability (F) was coded as F=1/(1+exp(-(FA+FS/Dose))), and the absorption half-life (TKa) was coded as TKa= α . F, where FA, FS and α were estimated parameters, and Dose is in mg. Concentrations predicted using the individual Bayesian estimates of PK parameters were used as a forcing function in the PK/PD model, which simultaneously fitted fasting glycemia, insulinemia and HbA1c data. In this model, glucose-dependent insulin secretion capacity (ISC) had a zero-order input and LY2189102-concentration-dependent (inhibitory Emax) first-order output. Glycemia and insulinemia were interdependently related through ISC and fixed insulin sensitivity and liver glucose output parameters. HbA1c was governed by an input rate that is dependent on glycemia and unglycated hemoglobin, and a first order output. Simulations of a wide range of doses and administration frequencies, from once weekly to once every 6 weeks, were conducted.

Results: Data from 5 patients whose PK profiles indicated presumed development of immunogenicity were excluded. The total clearance (CL), distributional clearance (Q), central volume (Vc), peripheral volume (Vp), FA, FS and α were (estimate (%SEM)) 9.45 (4.4%) mL/h, 25.1 (16%) mL/h, 3.06 (7.7%) L, 1.91 (7.7%) L, -0.42 (25%), 0.837 (21.1%) and 217 (10.9%) h, respectively. Inter-individual variability (%CV) in CL, Vc, Vp and Tka was 44.2%, 51.9%, 69.2%, and 57.9%, respectively. Placebo effect, LY2189102’s maximum inhibition of insulin secretion output (Imax), the concentration at 50% effect (IC50), and insulin secretion loss rate constant were 6.5% (45%), 10.4% (39%), 342 μg/L (277%), and 4.72 mo-1 (37%), respectively. Inter-subject variability in PD parameters was modest. Simulation, conditional on final estimates, showed that while total administered doses saturably determined the magnitude of response, all studied administration schedules were essentially equivalent, probably due to the long half-life of the compound. A dose of 60 mg, administered once every 6 weeks, sustains near maximum glycemic response (-0.9 mM from baseline, -0.55 mM from placebo), while an 18-mg dose sustains 75% of the maximal response. Glycemic response plateaus by 3 months, while the HbA1c response at 6 months was 14% higher than that at 3 months.

Conclusion: Dosing (SC) of LY2189102 can be as infrequent as once every 6 weeks, potentially offering a convenient therapeutic alternative for patients with T2DM.

European Association for the Study of Diabetes (EASD), Lisbon, Portugal, September 2011

By E. Abu-Raddad, A. DeGaetano, Sebastien Bihorel, Jill Fiedler-Kelly, and J. Sloan-Lancaster