NEW YORK (GenomeWeb) – After Aevi Genomic Medicine recently reported that a trial investigating a personalized drug for attention deficit hyperactivity disorder, AEVI-001, failed to meet its primary endpoint, the company has performed an additional analysis in an effort to better define predictive markers and home in on the subpopulation of adolescents most likely to benefit from therapy.
At the World Congress on ADHD in Vancouver today, researchers from Johns Hopkins University and Aevi presented post hoc analysis of the so-called SAGA trial, which had previously failed to meet its primary endpoint, and reported that mutations in nine genes involved in glutamatergic signaling and neuronal development were predictive of which patients had a statistically significant and clinically meaningful response to AEVI-001. Mutations in one gene, CNTN4, stood out in particular for its association with "robust responses" to the drug, researchers reported.
The finding with regard to CNTN4 in ADHD is novel and unexpected, and was enabled by the use of a higher-resolution microarray from Illumina. "In genomic medicine, it's a good lesson for people to hear," said Aevi CEO Mike Cola. "As the technology changes, you're going to find things that you didn't previously."
In the post hoc analysis, SAGA principal investigator Robert Findling from JHU and others evaluated 42 patients who had mutations in these genes and received either AEVI-001 or placebo, and reported that 89 percent of those on the drug had a 30 percent or greater reduction in an ADHD symptom rating scale compared to 21 percent on placebo. The company estimates that around 10 percent of pediatric ADHD patients harbor mutations in the nine genes.
The rationale for the AEVI-001 development program is based on the research of Hakon Hakonarson, director of the Center for Applied Genomics at the Children's Hospital of Philadelphia. Aevi, which was formerly Medgenics, inked a $5 million deal with CHOP in 2014 to gain exclusive access to the genomic discoveries that come out of the Center for Applied Genomics biobank for the development of drug and diagnostic targets.
Aevi had been discussing with Hakonarson his research showing that more than 20 percent of ADHD patients harbored copy number variations in the metabotropic glutamate receptor (mGluR) gene network. The receptor, mGluR, activates glutamate, a major excitatory neurotransmitter, and mutations in GRM genes that encode these receptors have been associated with various neuropsychiatric disorders.
The data were compelling enough that Aevi licensed AEVI-001, a pan-selective modulator of mGluR. While under Nippon Shinayku in the 1990s, AEVI-001 had showed lackluster efficacy in vascular dementia, but had a clean safety profile.
A month ago, the company reported that AEVI-001 was unable to meet its primary endpoint in the SAGA trial, a randomized, placebo controlled trial involving around 100 ADHD patients between ages 12 and 17. Study participants all had copy number variations in 273 genes of interest impacting the mGluR network.
Though patients on the drug did not experience significant reductions in a symptom rating scale, called ADHD-RS, compared to placebo, there were signs that AEVI-001 was improving inattention. Additionally, 70 percent patients on the drug had a 30 percent or more improvement in the ADHD-RS scale, while 42 percent on placebo experienced that level of improvement. The drug did meet a secondary endpoint, meanwhile, where 57 percent of patients on AEVI-001 had a clinically meaningful and statistically significant improvement in the Clinical Global Impression of Improvement Scale (CGI-I) compared to 33 percent on placebo.
Cola explained that CGI-I involves the physician evaluating the patient more holistically and represents a more stringent measure in some ways. As such, these top-line findings were surprising, he said, because "how can you miss on the overall magnitude and make it on a more stringent endpoint?"
The difficulty, he said, was that while the drug showed good efficacy in responders, unlike other ADHD drug trials 30 percent of non-responders in the trial had virtually no response to AEVI-001. This is "highly unusual," said Cola, because in ADHD a non-responder on treatment has some response, even a sub-clinical response.
Moreover, in SAGA, researchers cast a wide net in terms of defining the molecular characteristic of the enrollment population, since in order to be mGluR-positive, patients had to have copy number variations in 273 genes. But researchers were able to revisit the data and reevaluate patients in the SAGA trial to see if there were more precise predictive markers among the larger gene set.
"We had an easy way to do a post hoc analysis of the data," Cola said. "The distribution of those 273 genes is not linear. We had a heavy skewing around the core GRM genes that were first described by [Hakonarson's] GWAS analysis and within those 16 genes, one gene turned out to be much more highly enriched."
They found that CNTN4 gene mutations were the most prevalent, showing up in 18 out of 42 patients in the present analysis. All six patients with CNTN4 mutations treated with AEVI-001 had a more than 30 percent reduction in ADHD-RS, while only 25 percent on placebo had similar reductions. In this subpopulation, patients on the drug had a 20.8-point reduction in ADHD-RS while those on placebo had an 8.9-point reduction.
"This is a gene we don't usually talk about in ADHD," Cola said. CNTN4 mutations have been previously associated with autism spectrum disorders, but based on a genotype/phenotype study involving more than 1,800 ethnically diverse children and adolescents with ADHD, Aevi found these mutations in nearly 20 percent of the marker-positive population. The company estimates that around 5 percent of all ADHD patients likely harbor CNTN4 mutations.
It was an unexpected finding somewhat driven by the fact that the group used a relatively new technology, the Omni 2.5 BeadChip array from Illumina, which "gave us a different view of the CNTN4 gene and picked up a CNV that previously had been missed," Cola said.
Findling and colleagues reported in a poster at the ADHD meeting that "the clinically meaningful response observed in the nine-gene subset was largely attributable to the CNTN4 subset." These patients were more prone to disruptive behavior, risk taking, hyperactivity, inability to finish tasks, making inappropriate sounds, and demonstrating anger control issues.
The large genotype/phenotype study allowed researchers to better characterize the clinical features of 92 ADHD patients with CNTN4 mutations. The analysis showed "clustering around emotional dysregulation," Aevi CSO Garry Neil said. The disruptive behavior, anger control issues, and risk-taking behaviors reported in this subpopulation "has the biggest life consequence for these patients," he noted, "in terms of getting kicked out of school, bad social interactions, association with drug and alcohol abuse, and even brushes with the law."
These patients also reported having overlapping symptoms with autism, such as inappropriate movements and sounds and hyperactivity. "They look like ADHD kids, but they have these more severe symptoms," Neil said.
The genotype/phenotype analysis also revealed that CNTN4 mutations were particularly prevalent in African Americans, with 73 percent of patients in this racial group harboring copy number variations in the gene. "Almost everything that has been published on that gene has been in Europeans," Cola said. "So, that's a very important gene of interest to us."
A successful precision drug that improves ADHD symptoms for even a small subset of the population could become a blockbuster. More than 6 million children have been diagnosed with ADHD as of 2011. Patients are currently treated by stimulants that have low efficacy rates and cause anxiety, sleep disturbances, and hinder growth in children. Doctors and parents switch drugs often for ADHD children, because of side effects or due to disappointing efficacy.
AEVI-001 doesn't appear to cause similar adverse effects in studies and the company is hoping to demonstrate in Phase III studies, slated to start later this year, that it can improve symptoms in a genetically defined subset of ADHD patients. The latest data in nine genes makes later-stage studies for AEVI-001 much more "executable" than having to look across 273 genes, Cola said.
The 273-gene panel, developed at CHOP, is still valid for diagnosing ADHD, Cola and Neil said. In studies so far, Aevi has used the spit kit-based larger gene panel, but the company's plans are to advance AEVI-001 with a companion diagnostic that interrogates a smaller gene set.
The company has regulatory exclusivity with AEVI-001, and has been filing "test-and-treat" patents covering the method of use for the drug, the unique phenotype, the genotype of the patient, and the diagnostic, according to Cola. Aevi is also exploring the drug in autism and pediatric anxiety disorder using a precision medicine approach, and is expecting data in these indications in the second half of this year.
"The beauty of genomic medicine is once you understand the genes and the response, you can generally find those genes in many other places," Cola said.