Moving Forward After Two Deaths in a Gene Therapy Trial of Myotubular Myopathy

The very recent report of deaths in a gene therapy trial for children with X-linked myotubular myopathy (MTM) is a tragic reminder of how difficult it is to predict outcomes in first-in-human studies. ¹

The sponsor of this clinical study, Audentes Therapeutics (which was acquired by Japan's Astellas Pharma in 2019), had generated impressive efficacy studies in a canine model of MTM ² and conducted nonhuman primate (NHP) safety studies before advancing into the clinic. ³ The initial cohort of six pediatric MTM patients dosed at 1 × 10¹⁴ genome copies (GC)/kg also showed encouraging results. However, all three patients administered a threefold higher dose (i.e., 3 × 10¹⁴ GC/kg) experienced severe hepatotoxicity, which has proven lethal in two of these subjects.

The success of the adeno-associated virus (AAV)-based product Zolgensma, approved in 2019 for the treatment of infants with spinal muscular atrophy (SMA) type 1, supported the safety and efficacy of infusing up to 1 × 10¹⁴ GC/kg of vector into infants. ⁴ However, potentially dose-limiting toxicities have been observed in some of the subjects who participated in gene therapy trials of Duchenne muscular dystrophy (DMD) using similarly high doses of AAV (Solid Bio, Sarepta, and Pfizer). ⁵

Some subjects from the SMA and three DMD trials developed an acute but transient decline in platelets, which required treatment in a few cases. Some of these trial subjects also developed reversible and asymptomatic liver damage, which in most cases was mild. However, two subjects in the Solid Bio trial and one subject in the Pfizer trial developed anemia due to hemolysis and renal failure, consistent with complement disorders associated with hemolytic–uremic syndrome. Several subjects in both DMD trials presented laboratory evidence of complement activation. Two patients with more severe renal failure required dialysis. Fortunately, all subjects recovered.

In early 2018, one of us (J.M.W.) reported in this journal severe toxicity in NHPs infused with high doses of AAV (i.e., 2 × 10¹⁴ GC/kg). ⁶ Similar to the observations in the SMA and DMD trials, the NHPs experienced acute reduction in platelets with modest elevations in transaminases. While most animals remained asymptomatic and their laboratory abnormalities resolved, a few progressed to a syndrome of severe liver damage and a coagulopathy that led to diffuse hemorrhage and shock, requiring euthanasia.

The details around the deaths of the two Audentes MTM subjects are limited at present, although it appears that the observed toxicity more closely resembles those of the NHP studies rather than those of the DMD trials. While the precise mechanisms that led to these toxicities remain unknown, some hypotheses have focused on the role of antibodies to AAV that either preexist or rapidly accumulate following vector infusion. The formation of vector–antibody complexes could activate complement through the classic pathway or activate innate immunity through Fc-dependent uptake in antigen-presenting cells. In fact, one of us (J.M.W.) recently showed that lethal systemic inflammation observed following infusion of an adenoviral vector into the ornithine transcarbamylase-deficient patient Jesse Gelsinger in 1999 was due, in part, to the presence of pre-existing antibodies to adenovirus that formed immune-activating complexes with the infused vector. ⁷

Next Steps

The complexity of host–vector interactions in gene therapy and the species-specific differences in those interactions have long limited our ability to use preclinical data to predict clinical outcomes. This is especially true with regard to immune-mediated vector toxicities. However, these limitations should not be seen as a reason to abandon animal modeling. Instead, the early clinical development of gene therapy vectors may require an iterative process, especially if unexpected toxicities are observed.

Once the mechanism of a specific type of immune-mediated toxicity is observed in patients, it may be useful to return to the preclinical setting in order to evaluate the potential variables associated with this response. In the case of intravenous administration of recombinant AAV at doses of ≥1 × 10¹⁴GC/kg, the determining variables may include host factors such as age, body weight as a basis for dose adjustment, the presence of even low titers of preformed antibodies, and underlying genetic predispositions, as well as features of the vector such as capsid, promoter, and method of purification. While not all of these factors may be addressable through preclinical modeling, the investment of time, effort, and resources toward this goal is warranted and could have implications across multiple clinical applications of AAV.

Children with MTM and their families bravely participated in the Audentes clinical trial, which had enormous promise in changing the lives of patients with this disabling and lethal genetic disease. Further development of this product is currently on hold pending further evaluation of these serious adverse events. Whether this product can be revived is unclear. The fact remains that patients with MTM and many other rare genetic diseases have no treatments.

It is imperative that the scientific community work together with full transparency and cooperation to learn from these tragedies to assure that we can deliver safe and effective treatments to individuals living with rare genetic diseases such as MTM.