AveXis, a part of Novartis, is working “diligently” with the U.S. Food and Drug Administration to identify steps necessary to lift the hold on its gene therapy trial in types 2 and 3 spinal muscular atrophy (SMA) patients, but the company does not know what caused the inflammation seen in primates given AVXS-101 by intrathecal injection in a preclinical study.
The FDA placed a partial hold on the open-label Phase 1/2 STRONG trial (NCT03381729) after being informed in late October of troubling safety findings in the recent primate study of AVXS-101, known as Zolgensma outside trials, using spinal canal (intrathecal or IT) injection.
This hold stopped both treatment and further enrollment for a high-dose STRONG group. Treatment by IT injection at low and middle doses is complete, with no evidence of safety concerns among the 32 children treated prior to the FDA order, including an unidentified number at the high dose.
Zolgensma’s commercial use and its approval, which is by intravenous (IV) infusion in SMA infants and toddlers up to age 2, is not affected by the hold. Ongoing IV administration trials are also untouched.
No evidence of similar inflammation has been seen in treated children in a thorough review of “all available sources,” AveXis said in an email response to questions by SMA News Today. That evidence likely goes back more than five years to those first treated in Zolgensma’s pivotal Phase 1 trial.
The primate study, reported to involve 12 monkeys, divided the animals into three groups, with a primary goal of testing a contrast agent for spinal delivery. In AveXis’ words, its main purpose was to see the impact of two agents “on lower motor neuron transduction efficiency in AVXS-101 intrathecal (IT) administration.” Safety was a secondary study concern.
One primate group was given an AVXS-101 IT injection without a contrast agent, another the injection with contrast agent “A,” and the final group with contrast agent “B,” the AveXis spokesperson said.
Inflammation appeared in dorsal root ganglia due to an infiltration of “mononuclear cells”— different types of immune blood cells like T-cells, B-cells, NK-cells, or monocytes; this inflammatory reaction can damage nerve cells and lead to their loss.
But such inflammation was found in all three monkey groups, so the contrast agent — one of which has been used in the STRONG trial — was not the culprit.
“[A]ll three groups showed comparable findings as it relates to DRG [dorsal root ganglia] inflammation and there is no evidence that the contrast agent influenced the findings. It is not known what caused the inflammation,” AveXis said. Dorsal root ganglia are a cluster (ganglia) of sensory nerve cell bodies at the posterior (dorsal) root of spinal nerves. These cell bodies help to transmit sensory messages of pain and touch.
Other primates in early studies not necessarily done by AveXis or Novartis have had similar inflammatory reactions traced to use of the viral vector that carries the gene therapy to cells, a virus engineered to not cause infection. The adeno-associated virus serotype 9 (AAV9) is the transport agent used with Zolgensma and AVXS-101; it transports the gene therapy to the specialized motor nerve cells that control muscle contraction, and are damaged and dying in SMA patients because of mutations in the SMN1 gene.
“Similar findings have been reported after administration of other adeno-associated virus serotype 9 (AAV9) or AAV9 derivative (AAVhu68) vectors in monkeys and minipigs. The clinical [patient use] significance of the DRG inflammation observed in this non-clinical study is not known,” the spokesperson said.
Likely among these reports was a University of Pennsylvania study in three monkeys and minipigs that found “degeneration of dorsal root ganglia sensory neurons” in the six animals and liver toxicity in a monkey. In the piglets, “sensory neuron findings” included “proprioceptive [movement] deficits and ataxia [poor coordination].” Penn researchers traced these findings directly to the use of viral vectors at high doses.
In a phone interview that followed that study’s 2018 publication, AveXis researchers spoke of distinct “differences” in the vector used in AVXS-101 trials and that was used by Penn, the AAVhu68 variant, noting changes in two amino acids between these vectors and the consistency of high-quality manufacturing processes used in AAV9 production that likely wasn’t available to Penn.
AveXis researchers were alerted to its own preclinical “red flag” findings when study safety results arrived in March, and planned to include them in a September annual update to trial investigators. “Unfortunately, a mistake was made,” AveXis said, and it did not.
While Novartis was reviewing information given the FDA earlier, under what is called a 483 response, “this omission was identified and the issue quickly assessed,” and “appropriate steps were taken to notify trial investigators and health authorities,” the spokesperson said.
In total, 140 patients have been treated with Zolgensma by either IV infusion or IT injection, mainly in trials and commercial use of the IV gene therapy, but also through expanded access programs that allow use outside clinical trials. The exact numbers for each method were not disclosed; “more than 100” have been given Zolgensma via IV delivery, AveXis said, and 32 via IT in the STRONG trial.
None have shown any signs of sensory messaging problems in follow-up examinations that can stretch back years.
“[W]e have completed a thorough review of safety data from all available sources to date and no adverse effects related to sensory changes have been seen,” the spokesperson said, adding later that “adverse events that might be expected from the pre-clinical findings” were part of the “thorough review of human safety data.”
AveXis researchers favored IT infusion for types 2 and 3 SMA patients, who tend to be older children, teenagers, and adults, because it targets spinal motor neurons and allows for lower dosing that might avoid an immune reaction to the viral vector.
IV infusion is systemwide, delivering the gene therapy via the AAV9 vector to motor neurons in the central nervous system (the brain and spinal cord) and elsewhere. Company scientists found that high levels of the SMN protein, severely lacking in SMA patients because of SMN1 mutations, are needed early in life in peripheral tissues as well, but this need drops after infancy.
“Our approach in using IV early in infants is to try to replicate the biological situation by giving high levels of SMN expression throughout the body,” Dave Lennon, president of AveXis, said in a May interview with SMA News Today. Older patients are given IT infusion to deliver SMN “where it’s needed most, which is in the motor neurons of the spine.”
AveXis, and Novartis by extension, came to the FDA’s attention earlier this year because of “data manipulation issues” that led to inaccuracies in the application for Zolgensma’s approval it filed with the regulatory agency in October 2018.
AveXis was reported to have learned of the manipulation and inaccuracies, in early animal tests that were included in its application, on March 14, 2019. Neither company informed regulators of these issues until June 28, five weeks after the FDA’s May approval decision.
Zolgensma’s quality is not in question, both the FDA and Novartis said in announcing the data manipulation in August. These data issues are “unrelated” to the safety concern seen in the one recent primate study, the spokesperson added.