Another round of disappointing data for NGD 2000-1 caused Neurogen Corp.'s stock to drop nearly 19 percent Wednesday, and made further development of the product in rheumatoid arthritis an unlikely plan.

The compound, an oral C5a antagonist, failed in the Phase IIa trial to meet its primary endpoint, a change in C-reactive protein, a biomarker of disease activity derived from patient blood samples. While it showed some positive trends and reached statistical significance in a secondary endpoint, Branford, Conn.-based Neurogen said the higher dose needed to reach therapeutic benefit raises too big a safety issue.

The company's stock (NASDAQ:NRGN) fell $1.76 Wednesday, or 18.9 percent, to close at $7.55.

William Koster, president and CEO of Neurogen, said the problem with the compound is its "ability to inhibit this 3A4 enzyme," especially at higher doses.

Phase I studies indicated that NGD 2000-1 inhibited the liver enzyme CYP 3A4, which metabolizes certain drugs. Therefore, the company had hoped that in the Phase IIa trial the compound would show benefit at 10-mg or 60-mg doses, rather than 100 mg.

"We were hoping to pick up a robust response at the lower doses that would allow us to have a greater therapeutic window," Koster told BioWorld Today.

The exploratory trial in rheumatoid arthritis was designed to evaluate the benefit of blocking the C5a receptor in patients with mild to moderate RA. NGD 2000-1 did not demonstrate an effect in the change in C-reactive protein, but it did demonstrate a statistically significant result (p=0.041) on a secondary responder analysis endpoint, Subject Global Assessment of Disease Activity, at the highest dose tested: 100 mg twice daily.

It did not reach significance in other secondary endpoints, but it did show positive trends.

"We're disappointed that this compound looks like it's not the compound to take forward in RA, but we're very much encouraged about the mechanism," said Steve Davis, Neurogen's chief business officer.

While the calculation of an ACR (American College of Rheumatology criteria) response was not a pre-specified endpoint in the study, a post hoc analysis revealed a statistically significant ACR 20 (20 percent improvement) response (p=0.014) at the highest dose tested. Data showed a trend toward significance on the ACR 20 scale at the lower 10-mg and 60-mg twice-daily doses.

The ACR criteria require an improvement in painful and swollen joint counts and improvement in three of five variables, including patient global assessment, physician global assessment, C-reactive protein, pain scale and a health assessment questionnaire. It is measured at the levels of 20 percent, 50 percent or 70 percent improvement.

"We're actually very interested that we were able to pick up this ACR 20 response," Koster said, adding that those results and other activity trends support the benefit of blocking the C5a receptor.

As the first company to reach the point of testing orally available C5a antagonists, Neurogen officials did not know what to expect with the Phase IIa studies. In January, NGD 2000-1 also failed in a Phase IIa trial for asthma, but the company saw the asthma and RA indications as mutually exclusive, with dissimilar underlying pathologies, Koster said. (See BioWorld Today, Jan. 14, 2004.)

Neurogen officials have not made a decision on whether to move forward with NGD 2000-1 in asthma. Like the RA indication, they still are analyzing data to look at other biomarkers and determine if there were any inherent biases. But they say further development in RA is "unlikely."

What is likely is that the company will look for another compound within the C5a antagonist program, and then develop it in both RA and asthma, and possibly other indications down the road. Koster also pointed out that the failure in meeting the primary endpoint in the RA trial might be due to something other than the C5a antagonist.

"Not seeing a biomarker moving could be a function of the size of the trial or the time that we actually ran the trial," he said. "So we still think that the mechanism is worthy of further exploration."

It's possible, Koster said, to separate out the CYP 3A4 enzyme activity from the C5a antagonist activity, allowing the company to develop a compound that could be moved to higher doses for higher potency.

The Phase IIa trial in RA involved 49 patients with mild to moderate RA treated in eight clinical centers. They were dosed twice daily for 14 days with either placebo, or one of three doses of NGD 2000-1 at 10 mg, 60 mg and 100 mg. No concomitant disease-modifying anti-rheumatic drugs were permitted during the trial due to NGD 2000-1's inhibition of the liver enzyme CYP 3A4.

In March, Neurogen raised $100 million in funds to expand its discovery platform and portfolio. As of March 31, the company had $67 million in cash and marketable securities. (See BioWorld Today, March 22, 2004.)

The company currently has five programs, three of which are partnered. In addition to the C5a antagonist program, Neurogen's other internal program is for MCH-1 antagonists as therapies for diabetes and obesity. The three partnered programs are for insomnia with New York-based Pfizer Inc.; for pain with Whitehouse Station, N.J.-based Merck & Co. Inc.; and for depression with Aventis Pharma SA, a unit of Strasbourg, France-based Aventis.