Liver Toxicity in Gene Therapy Trials Could Merit Stricter Entry Criteria

An FDA advisory committee offered up lower vector doses, stricter exclusion criteria, and long-term animal studies to grapple with liver toxicities associated with adeno-associated virus (AAV) vector-based gene therapies.

During the second of four sessions on AAV safety, the Cellular, Tissue, and Gene Therapies Advisory Committee (CTGTAC) weighed evidence of hepatotoxicity — ranging from liver enzyme elevation to liver failure and death — observed with the products in clinical trials for hemophilia, spinal muscular atrophy (SMA), and X-linked myotubular myopathy (XLMTM).

“The bulk of hepatotoxicity observed in hemophilia trials at this point, thankfully, pose more of an efficacy limitation than an actual safety concern,” said Lindsey George, MD, of Children’s Hospital of Philadelphia, during an opening presentation of the safety data.

However, she noted, there have been reports of acute liver failure in SMA and fatalities in a trial involving XLMTM.

In the ASPIRO trial testing the gene therapy AT132 in XLMTM, three deaths were presumed to be related to complications of liver failure among 17 patients enrolled in the highest dosing cohort (3.5E14 vg/kg), despite the fact that the trial excluded patients with a history of clinically significant liver disease.

“With respect to the doses in this trial, they are essentially the highest I am aware of in any AAV trial,” said George.

The trial was subsequently put on hold, then lifted by the FDA late last year at a lower dose (1.3E14 vg/kg), but liver issues have continued. Earlier this week, trial sponsor Astellas voluntarily announced a hold on screening and dosing after a patient developed abnormal liver function within a month of dosing.

“I think it is important to acknowledge that morbidity and mortality as it relates to hepatotoxicity has only been observed with the delivery of systemic AAV doses that are greater than 1E14 vg/kg, and actually, only in pediatric patients,” George said. “So maybe one of the important points of this is to continue to strive for lower vector doses as part of the clinical development of these programs.”

In determining appropriate dosing, committee member Raymond Roos, MD, of the University of Chicago Medicine, noted that “there are different serotypes of the AAV vectors and different transgenes, so it’s hard to generalize about a particular vector dose.”

“I think it’s important to do animal work to get some idea about this,” he said. “But, ultimately in clinical trials, looking at dose response would be important — starting at a low dose and challenging that patient population with increasing amounts of AAV vector.”

CTGTAC Chair Lisa Butterfield, PhD, of the University of California San Francisco, said an upper limit does not appear to be the best approach, and that total vector genome dose per kg or factoring in some measure of body mass index for patients might be preferable. In addition, she said, it remains to be seen what role vector dose and capsids played in the deaths seen in the XLMTM study.

In discussing mitigating the risk of liver injury, Theo Heller, MD, of the National Institute of Diabetes and Digestive and Kidney Diseases, suggested that relying on liver enzyme measurements or ultrasound examinations is not adequate, and that stringent exclusion criteria in trials and careful evaluation of patients “is essential.”

“The concept in hemophilia of not enrolling people with bridging fibrosis or greater is something that could be emulated in other diseases,” he said.

Learning From Animal Trials

Committee member James M. Wilson, MD, PhD, director of the Gene Therapy Program at the Perelman School of Medicine in Philadelphia, said primates as a model provide several areas of agreement with human studies: “Importantly, the threshold for severe toxicity requiring at least euthanasia with primates is almost spot-on with the threshold for human studies, which is 2E14 vg/kg,” he said.

But, there are important differences, he noted. “The severe liver toxicity observed in humans is often delayed from what we are seeing in primates, which is a bit earlier.”

There are also several key limitations to the use of non-human primates — i.e., macaques — in these studies, Wilson said. For example, it is difficult to evaluate the impact of disease factors in these animals because they don’t have any comorbid conditions. In addition, when macaques get sick early in a study, researchers are required to euthanize them; Wilson suggested that if the animals were given the same kind of supportive care available to humans, it might allow for further evaluation of these toxicities.

Regarding the acute toxicity seen in primates, however, Caroline Zeiss, DACVP, DACLAM, of the Department of Comparative Medicine at Yale University in New Haven, Connecticut, said: “We don’t see quite the same thing in people.” But being able to follow these animals longer, and later, could provide the opportunity to learn more about the mechanism of acute and long-term effects, she added.