West Coast Editor

About a week after AtheroGenics Inc. disclosed in SEC paperwork that the firm is saying goodbye to about half its work force, and slightly more than a month after partner AstraZeneca plc bailed out of a major deal for AGI-1067, the Phase III focus for that compound is shifting from atherosclerosis to diabetes.

In March, when the oral small molecule failed in the Phase III cardiovascular study, Wall Street took away more than 60 percent of the company's share value, and the impact of the latest news was almost nil. AtheroGenics' stock (NASDAQ:AGIX) ended Thursday at $2.39, up down 2 cents. (See BioWorld Today, March 20, 2007, and April 24, 2007.)

The much-whooped over deal with London-based AstraZeneca gave AtheroGenics $50 million up front in December 2005, with potentially $650 million in milestone payments and handsome royalties, but that went down the tubes with the blowup of the Phase III trial called ARISE (Aggressive Reduction of Inflammation Stops Events). AstraZeneca had just been burned in October by the Phase III failure of the stroke drug NXY-059, partnered with Renovis Inc., and dropped that one, too.

AtheroGenics said at the time of the ARISE fizzle that it would continue developing AGI-1067, an anti-oxidant that in earlier studies blocked VCAM-1 expression and prevented atherosclerosis. Now, pointing to encouraging data in Type 2 diabetes patients who took part in the 6,100-patient ARISE study, the firm plans to start a Phase III trial called ANDES (AGI-1067 as a Novel Anti-Diabetic Agent Evaluation Study).

"AGI-1067 on top of all the commonly used medications was able to show an improvement," noted Mark Colonnese, vice president of commercial operations and chief financial officer of Atlanta-based AtheroGenics. "There's an additive effect with the drug."

In ARISE, the drug gained a 22 percent reduction in the prespecified endpoint of cardiovascular death, resuscitated cardiac arrest, myocardial infarction (heart attack) and stroke. The study also found a 64 percent drop in new-onset diabetes when treated with AGI-1067.

Enrollment in ANDES is expected to start in the third quarter, with an interim peek at data in mid-2008. The study will compare AGI-1067 vs. placebo on glycemic endpoints, and is intended to confirm results seen in diabetes patients from ARISE, where the drug worked for glycemic control and in reducing cardiovascular events in more than 2,200 diabetes patients, AtheroGenics said.

That benefit amounted to a 0.3 percent drop in hemoglobin A1C (glycosylated hemoglobin, the marker for blood glucose) with AGI-1067 as compared with a 0.2 percent rise with placebo. Colonnese would not say whether the FDA had signed off on the study's design.

"We meet with the FDA and other regulatory agencies on a regular basis," he said. "We don't go into details of our interactions." Unlike ARISE, ANDES will not be conducted under a special protocol assessment, but "the regulatory pathway is well defined" for diabetes drugs, he said, and AtheroGenics will start discussions with potential new partners later this year.

The firm aims to enroll 1,200 patients in worldwide ANDES by the end of the year, and keep them on therapy for 24 weeks, testing three doses, including the 300 mg/day dose used in ARISE. The study is 90 percent powered to check for ARISE's 0.3 percent drop in A1C vs. placebo, and only patients with two earlier A1C readings higher than 7.5 percent will be let in.

"The higher the baseline A1C, the larger drop you get," Colonnese told BioWorld Today, adding that results on A1C levels are "right in the range of what they saw with Januvia." Whitehouse, N.J.-based Merck & Co. Inc. gained FDA approval for Januvia (sitagliptin) in Type II diabetes last fall.

In the new AGI-1067 trial, AtheroGenics will monitor liver function test (LFT) elevations, which showed up three times above the upper limit of normal in the ARISE study, though company officials downplayed the finding during a conference call in March. Russell Medford, president and CEO, said the levels landed "very much within the range that we see on Lipitor, for instance." Lipitor (atorvastatin) is the blockbuster anti-cholesterol drug from New York-based Pfizer Inc.

"With Lipitor, LFTs above three times the upper limits of normal is 0.1 percent on 10 mg and 2.3 percent on 80 mg, and I can tell you that with that kind of analysis, our data look well within that range," Medford said. "We don't see this as a major cause for concern."

What's more, the ARISE patients were "well controlled at baseline," he said. "Their [A1C] was 7.2 percent. They didn't really have much room to move. Other mainline anti-diabetic drugs in that setting produced similar reductions in [A1C], so this is certainly as good as we see with other drugs used in diabetes."

The standard for portraying liver-enzyme changes, which is "being solidified by the statins, is to look at the confirmed elevations above three times the upper limits of normal," Medford said. "This really reflects the fact that patients of this type frequently have increases in liver function tests that have nothing to do with drugs."

Colonnese said that "by and large, [AGI-1067] has a favorable effect on patients' livers. In a few, very rare cases, we saw some highly elevated liver enzymes. It all happened early in the ARISE trial," and when the patients were taken off the drug, their levels went back down.

ANDES will not include cardiovascular endpoints because the trial is too short. "We're running a standard glycemic control trial that the FDA requires for approval," Colonnese said. About a week ago, AtheroGenics said in an 8-K filing that it had reduced staffing by about 50 percent to 67 employees to streamline operations and focus on AGI-1067.