The new year greeted Repligen Corp. with disappointing data from its Phase III trial of synthetic human secretin in autism.

RG1068 showed little difference from placebo when evaluated in 132 children, contradicting some data from an earlier Phase II trial. The same amount of Phase III patients - about 40 percent - in each the placebo and drug groups reported improvement.

"I think it's highly unlikely that it will be successfully developed for autism, based on these results," said Jason Kantor, an analyst with WR Hambrecht + Co. in San Francisco. "There was a lot of skepticism and negative data already available for secretin. This result may be the final nail in the coffin."

The company's stock (NASDAQ:RGEN) dropped $1.77 Monday, or 42.6 percent, to close at $2.39.

Secretin failed to meet its primary endpoint, which was measured by a psychologist using the Childhood Autism Rating Scale, in a Phase II autism trial in 2001. The company, however, found a statistically significant symptomatic improvement in the secretin-treated group compared to placebo according to parental assessment, prompting Repligen to move forward into Phase III studies. (See BioWorld Today, April 5, 2001.)

Now with disappointing Phase III results, the company is looking at a subset of children with higher function level, IQ, in which there were statistically significant results.

"There's a wealth of data [showing] that we have biological activity," Repligen's CEO and President Walter Herlihy said in a conference call Monday. "We just don't know how to produce it yet in this patient population."

The objective of the double-blind Phase III study was to evaluate six doses of secretin vs. placebo in autistic children aged 2 years, 8 months to 4 years, 11 months. It was carried out at 15 medical centers in the U.S. There were 66 evaluable patients in each the drug and the placebo groups. The Waltham, Mass.-based company aimed for success with two primary endpoints: improvements in social interaction as measured by the Autism Diagnostic Observation Schedule (ADOS), and the parental Clinical Global Impression of Change (CGI).

"For both endpoints there was a higher placebo effect than was observed in our Phase II trial," Herlihy said.

Analyzing the data, the company looked at two subgroups, one in patients less than 4 years old and another in patients with higher function level. The company found no significant improvement in either endpoint in the subgroup of children less than 4 years old. But in 68 patients with higher IQ, the company found significant improvement (p= 0.046) in the ADOS social scale, which was evaluated by a clinical psychologist. The company eliminated seven high-functioning patients not treated according to the protocol, giving a "p" value of 0.003 in the remaining 61 patients.

"These observations suggest that low cognitive function, which is a common co-morbidity in autism," Herlihy said, "may have hindered the ability to see a treatment effect in the whole group."

No drug-related safety issues or adverse events were reported for either the drug or placebo groups in the Phase III trial.

Herlihy pointed out some differences in the Phase II and III trials that might explain the contradictory results. Children in the Phase III trial were younger and had higher ADOS social scores than children in the Phase II trial. With a maximum score of 14, 39 percent of Phase III children had a baseline ADOS score of 13 or 14, compared to 12 percent of children participating in the Phase II trial.

Also, children in the Phase III trial were treated for 18 weeks, compared to nine weeks in the Phase II study. And the Phase III trial enrolled patients with or without gastrointestinal symptoms, while the Phase II trial exclusively enrolled patients with GI symptoms.

"Obviously we are disappointed that the trial did not meet its primary endpoints," Herlihy said. "Despite the setback, we have accumulated a wealth of preclinical and clinical data suggesting secretin may be an important modulator of anxiety and/or social interactions."

The company intends to initiate a study of RG1068 in an anxiety disorder in the second half of this year. It also plans to continue with a Phase II trial in schizophrenia.

"The [schizophrenia] data look promising," Kantor told BioWorld Today. "If they can repeat the results that an independent researcher found, then they may have a drug on their hands."

"We expect to have data [in] calendar 2004," in the schizophrenia trial, Herlihy told BioWorld Today. As for the fate of RG1068 in autism, Herlihy said the company would discuss, probably in March, with the FDA which direction to go from here.

Kantor said in a research note that the secretin news is a "significant financial setback" for Repligen.

"We had previously forecast profitability in fiscal 2006 based on first-year sales of Secretin for autism and significant off-label use of SecreFlo for autism," he said.

Aside from secretin, Repligen is developing CTLA4 for autoimmune disease and Uridine for the treatment of neurological diseases, bipolar and major depression. Its two marketed products are SecreFlo, secretin for injection, which is marketed to gastroenterologists for pancreatic assessment, and rProtein A, a consumable reagent used by the pharmaceutical industry to produce monoclonal antibodies.

The company has $25 million in cash and no debt, Herlihy said. With a burn rate of $8 million a year during the Phase III trial, the company expects that rate now will drop to about $5 million a year. "We certainly have an adequate amount of cash to move forward the next several years," Herlihy said, adding the company may in-license additional programs.

He also said the company expects to receive about $6.5 million in revenues from its two marketed products at the end of its fiscal year on March 31.