Abstract

Objectives:

CGRP inhibitors are a class of drugs that meet a significant unmet medical need for migraine treatments. However, two CGRP inhibitors, telcagepant and MK-3207, failed in clinical trials due to liver toxicity (1,2). Telcagepant, MK-3207, and the next-in-class compound ubrogepant were represented in DILIsym, a quantitative systems toxicology (QST) model of drug-induced liver injury, in order to predict whether ubrogepant  would be a safer alternative.

Methods:

In vitro experiments were undertaken determining the potential for the three compounds to inhibit bile acid transporters, cause mitochondrial dysfunction, and produce oxidative stress. The results of the in vitro assays were used to produce a representation of each compound in DILIsym, along with a PBPK model of each compound. The clinical doses of telcagepant and MK-3207 at which
liver toxicity was observed were simulated, as well as a range of potential ubrogepant clinical protocols.

Results:

Each of the molecules tested displayed signals in the in vitro assays for bile acid transporter inhibition, oxidative stress generation, and mitochondrial dysfunction. DILIsym correctly predicted the liver toxicity of telcagepant and MK- 3207, while ubrogepant was predicted to be safe even at doses 10X the proposed clinical dose. Subsequent clinical trials demonstrated that ubrogepant was indeed safe (3), and the drug was approved by the FDA for the acute treatment of migraine.

Conclusions:

QST modeling can prospectively differentiate between toxic and non-toxic drugs within the same class and identify which drugs within a class carry less risk of toxicity. These applications can contribute to success in the clinic and regulatory
approval of new drugs.

Presented at ACoP 11 Virtual Conference, Nov. 9-13, 2020

By Jeffrey L. Woodhead, Brett A. Howell, Brenda Smith, Josh M. Rowe, Scott Q. Siler