Search in the Abstract Database

Abstracts Search 2017

* = Presenting author

P113 Model based predictions of the PTG-100 pharmacodynamic responses in ulcerative colitis patients

Mattheakis L.*1, Fosser C.2, Saralaya R.2, Horsch K.2, Rao N.1, Bai L.1, Zhao L.1, Annamalai T.1, Liu D.1

1Protagonist Therapeutics, Milpitas, United States 2Cytel Inc., Cambridge, United States

Background

PTG-100 is a gut restricted oral peptide antagonist of the T cell homing integrin α4β7, and it alters trafficking of gut homing T cells in preclinical animal models. Its potency and selectivity are similar to that of the anti-α4β7 antibody vedolizumab (Entyvio®), which is approved for moderate to severe ulcerative colitis (UC) and Crohn's disease. In preclinical models using healthy or dextran sulfate sodium (DSS) induced colitis mice, we previously showed that PTG-100 causes a dose dependent increase in integrin receptor occupancy (RO) and downregulation of integrin receptor expression (RE) on peripheral blood memory T cells. Similar pharmacodynamic (PD) responses were observed in the peripheral blood of normal healthy volunteers (HV) after PTG-100 dosing in a randomized, double-blind, Phase 1 trial. The aim of this study was to evaluate modeling approaches for predicting the PD responses in UC patients.

Methods

Semi-mechanistic, nonlinear, mixed effects models were based on the PD responses observed in healthy and colitis mouse studies, and in the Phase 1 HV study. Standard model validation techniques were used including posterior predictive checks that compare model based predictions with observed data. In vitro integrin activation studies were done using memory T cells from human PBMC donors.

Results

There is a significant increase in PTG-100 RO in colitis mice compared to healthy mice at 4 h post dose at all tested doses. The colitis mouse model was structurally connected with the healthy mouse model by estimating colitis mouse multipliers. These multipliers were then used to extrapolate from the human healthy volunteer dose response (DR) relationship to make predictions of the UC patient response. As observed in mice, the model predicts a lower dose of PTG-100 will be needed in UC patients to achieve PD responses equivalent to those observed in healthy volunteers. We also tested the binding properties of PTG-100 to the different activation states of α4β7 integrin. We found that PTG-100 prefers binding to α4β7 that has been activated by incubation of T cells with MnCl2 or with retinoic acid.

Conclusion

Modeling approaches successfully characterized the observed PD responses in healthy and colitis mice and in healthy human volunteers. Extrapolation of the model to UC patients shows a pronounced lower dose shift in UC patients compared to healthy volunteers. Based on in vitro binding data, this shift can be explained by PTG-100's preference to bind the activated state of α4β7. These results suggest that the increased activity of PTG-100 under colitis conditions is caused by the corresponding increased proportion of memory T cells expressing the activated state of α4β7 integrin.