EXPERIMENTAL BIOLOGIC AGENTS CAN BE STUDIED IN COMBINATION WITHOUT PRIOR INDIVIDUAL PHASE I DATA IF COMPELLING RATIONALE IS GIVEN -- FDA's WOODCOCK

Two experimental biologicals can be studied as a combination treatment in clinical trials without prior human experience with the individual agents if the sponsor can provide "a really superb rationale of why these two should be used together," FDA Center for Biologics Evaluation and Research Acting Deputy Director Janet Woodcock, MD, said. Speaking at a June 24-26 International Biotechnology Conference in Baltimore cosponsored by FDA and the Maryland state government, Woodcock observed that "normally each agent should be studied separately first -- fairly extensively [in humans] before you would combine them." However, she added that "there might be compelling reasons not to do that." Woodcock suggested examples of situations that might preclude prior human studies, such as when existing evidence indicates that "one might not be effective without the other or one might be an extremely toxic agent that would require the other agent to be tolerated at all." Entering two experimental drugs directly into human clinicals will "require more extensive preclinical evaluation," she noted. In the future, Woodcock predicted that "combination regimens are going to be the rule rather than the exception" in biologic INDs. "We are going to have to deal with them and think about them, but they are complex." The rationale for using two agents in combination, Woodcock said, can vary -- from showing in vitro synergistic or protective effects to an animal model that suggests the agents are useful together, or clinical experience of the single agents that suggest their effects might be synergistic. Each of these reasons "can be completely acceptable as a rationale" to do combination studies, she said. Woodcock suggested that sponsors estimate the dose-related toxicity profile of the combination to determine if the toxicities are additive or decreased prior to beginning combination trials. In addition, Woodcock advised sponsors to establish a safe starting dose. Finally, sponsors should develop an understanding of the pharmacokinetic interactions of the two experimental agents, she explained. If the pharmacokinetic interactions of the two drugs have been sufficiently explored, she noted, "you may be able to do all this without any additional animal studies." Addressing study protocols, Woodcock remarked that clinicals "should be designed to establish efficacy of all the experimental agents." That requirement would probably lead to studies designed to test the efficacy of the individual agents versus the combination. However, she noted that studying single agents alone may not be always ethical or practical. In addition, combination trials may make it difficult to obtain full dose-response curves for many single agents. "When you have a combination of two or more [drugs] and you have to look at each dose of one compared to a range of doses of the other and estimate very precisely the optimal dose of each, this can require extremely large trials and is very difficult," Woodcock noted. "We understand that." Woodcock offered several examples of trial designs with some caveats. For example, if an experimental drug is being added to an approved drug to ameliorate the toxicity of the approved drug, "you have to make sure the approved drug to some extent is as effective in the combination arm," she commented. If an experimental drug is being added to an approved drug to improve its efficacy and allow for a lower dosage regimen, she explained that sponsors will "need to know that there is a dose response from prior data to the approved drug, or else you are going to need three arms in this trial where you have the lower dose of the approved drug by itself in the third arm." A panel of prominent clinical investigators at the conference explained the reasoning behind certain biologic combinations, some of which are currently the subject of ongoing trials. Memorial Sloan-Kettering Cancer Center researcher Janice Gabrilove, MD, discussed the rationale of combining hematopoietic growth factors. She pointed out that G-CSF and GM-CSF in combination would provide a logical synergy. While GM-CSF "prolongs the circulating half-life of the neutrophil from its usual six-to-12 hours up to four days," G-CSF "has no effect on circulating half-life of the neutrophil," Gabrilove observed. Meanwhile, G-CSF's "major effect is to decrease markedly the transit time [of neutrophils] from the mitotic to postmitotic compartment, from seven-to-14 days down to one-and-a-half days," Gabrilove pointed out. She noted that GM-CSF has no effect on this aspect of mobilizing neutrophils. "You can see from this that the complementary action of these molecules would project from this that they ought to be . . . tried in combination," Gabrilove maintained. Massachusetts General Hospital Cardiac Catherization Lab Director Herman Gold, MD, discussed combination trials in thrombolytic therapy. Gold described the use of Centocor's CentoRx (7E3 FAB) antiplatelet monoclonal antibody in conjunction with tissue plasminogen activator (Genentech's Activase) to accelerate clot lysis. CentoRx is a glycoprotein IIb/IIIa blocker that inhibits platelet aggregation. When added to TPA, Gold said, CentoRx has been shown to prevent reocclusion of the arteries in dogs. CentoRx is currently in Phase II trials. Gold noted that "part of the problem doing clinical work with multiple agents is you need to do this carefully first in the animal labs and then [do] very small well-controlled pilot trials in [humans(BRACKET)." He pointed to several problems encountered in the early Phase I trials of CentoRx. In one safety study in angina patients, the monoclonal antibody/TPA combination was found to be ineffective, with angina episodes reoccurring in half of the patients in the trial. "The explanation [for the lack of efficacy(BRACKET), we believe, was the fact that heparin was omitted in this trial," Gold said. A second Phase I trial that includes low-dose heparin in the combination treatment is now underway, he said. CentoRx also increases the bleeding time when added to TPA, Gold noted. As a result, Gold said his group has looked at IIb/IIIa blocking peptides that are similar in action to the monoclonal product but which may lessen the bleeding time. Gold's group has recently studied Genentech's G4120 IIb/IIIa blocker in combination with TPA and heparin in dog studies as an inhaled treatment. The combination produced complete lysis in "an incredibly short period of time," seven minutes compared to 52 minutes, while bleeding time returned to baseline upon discontinuation of treatment. "This may provide us a shorter time interval in which to accomplish lysis with maximum rapidity, prevent reocclusion, and at the same time limit the bleeding potential," Gold told the IBC meeting. Thrombin inhibitors are also being investigated to determine whether they can prevent reocclusion and reduce bleeding time following clot lysis. A Phase I study with Genentech's argatroban plus TPA found that reocclusion was prevented, although lysis was not as rapid, and there was "very little prolongation if any" of bleeding time, Gold said. "If this kind of approach holds up in [further trials] it raises the possibility that we will be able to undertake clot lysis using a IIb/IIIa agent with rapid resolution of clot, including the resistant clot, and then follow it up with a thrombin inhibitor," Gold observed.