FDA Considers Altering Preclinical Testing Guidelines

'OPENING THE DOOR': FDA's Joseph Contrera claims the new testing guidelines will encourage innovation. The Food and Drug Administration (FDA) is considering a plan to revise a 25-year staple of preclinical pharmaceutical testing in the United States. Under the current procedure-criticized by some as lengthy and expensive-development of a new drug must include two-year bioassays conducted on two separate rodent species to determine whether it could cause cancer in humans. FDA recently closed

Oct 28, 1996
Thomas Durso

Joseph Contrera
'OPENING THE DOOR': FDA's Joseph Contrera claims the new testing guidelines will encourage innovation.
The Food and Drug Administration (FDA) is considering a plan to revise a 25-year staple of preclinical pharmaceutical testing in the United States. Under the current procedure-criticized by some as lengthy and expensive-development of a new drug must include two-year bioassays conducted on two separate rodent species to determine whether it could cause cancer in humans.

FDA recently closed the public-commentary period on a proposal, drawn up under the auspices of an international committee, that would replace that requirement with a more flexible guideline. While there is substantial praise for the plan, some observers argue that it relies on an unproved process that could fail to screen out dangerous drugs.

The FDA document, the "Draft Guideline on Testing for Carcinogenicity of Pharmaceuticals," would give companies the optional alternative of limiting the two-year bioassay to just one rodent-typically a rat-with a second, complementary test featuring an alternative model, such as a transgenic mouse, that can be completed in as little as half a year.

According to Joseph F. Contrera, associate director of the office of testing and research at FDA's Center for Drug Evaluation and Research, the result would be a cheaper, faster way of testing pharmaceuticals for carcinogenic properties that could lead to more clinically relevant testing models. "This guidance encourages innovation, opening the door to alternative models," he states.

Supporters of the measure claim the current system forces drug companies to bear excessive costs and endure lengthy delays while their products are being certified as safe. Ray Stoll is director of toxicology and safety assessment at Boehringer Ingelheim Pharmaceuticals in Ridgefield, Conn., and represented the Pharmaceutical Research and Manufacturers of America (PhRMA) during the international process that produced the new guideline. He estimates that each bioassay involves between 500 and 1,000 animals. "Ultimately, if you're using less animals and taking less time, there should be a savings to the company. That's the pragmatic answer," he contends.

THUMBS DOWN: Thomas J. Moore, a critic of FDA practices, believes that the proposal could allow dangerous drugs to slip through and reach the marketplace.
Critics argue that while the current system already approves too many dangerous drugs, the revisions would be even worse. "It ought to be of concern," declares Thomas J. Moore, a senior fellow at George Washington University's Center for Health Policy Research. "What we know about the proven human carcinogens is that the actual incidence rate is so low it would never show up at the sample sizes we normally use in drug testing."

Proponents of the change say advances in molecular biology, such as the ability to transport genes from one species to another, have led to the development of alternative models on which tests can be conducted. Contrera, who notes that the new guideline could not have been drawn up as recently as five years ago, feels that besides reducing costs and time, such tests could eventually "be more predictive of human risk than [tests] we now have."

In fact, in addition to cost and time, many researchers dislike the current system because they feel it is not a good predictor of whether a certain drug will cause cancer. The proposed guideline itself declares: "Since the early 1970s, many investigations have shown that it is possible to provoke a carcinogenic response in rodents by a diversity of experimental procedures, some of which are now considered to have little or no relevance for human risk assessment."

"Science changes. It advances," notes Clark Heath, vice president for epidemiology and surveillance research at the Atlanta-based American Cancer Society. "With respect to how we judge the toxicity of potential biologic activity of a given compound, animal tests are not necessarily the final word. They're probably misleading. High doses of a particular compound given over a period of time may suggest carcinogenicity not so much from their innate carcinogenicity but the fact that they have caused so much tissue damage and tissue [proliferation] that the risk of cancer increases."

FDA closed the guideline's comment period on October 21. The international group that developed the plan will meet in London in November, but Contrera expects that commentary results will not be ready until its March 1997 meeting in Japan, when the group will make revisions. "The document will proceed along to final ratification, we envision, by July of 1997," he predicts.

The revised guideline is the product of an industrial-governmental coalition called the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH), a group whose function is embodied in its name. A three-region entity covering the United States, Europe, and Japan, ICH developed a 13-year database to examine the utility of the two-rodent bioassay, a practice Contrera estimates has been in existence for 25 to 30 years.

"The conduct of only one long-term carcinogenicity study [rather than two] would, in part, allow resources to be diverted towards other currently evolving experimental approaches," the proposed guideline says. "The totality of the data derived from one long-term study and other appropriate experimental investigations contribute a 'weight of evidence' approach that should improve the assessment of carcinogenic risk to humans."

"The timing was right," Contrera declares. "The development of alternatives has reached a sufficient maturity to apply them right now."

BIG CHEESE: Alternative models, such as this p53 knockout mouse, would replace one of the two-year bioassays.
The response from the American pharmaceutical industry has been generally supportive. PhRMA, based in Washington, D.C., and an ICH participant, still was developing its official comment on the guideline at press time, but spokesman Jeff Trewhitt makes note of the organization's participation in the ICH proceedings. "PhRMA is as at the table as a player in that process," he says. "PhRMA, and our member companies in general, are in agreement with the process."

Heath agrees. "I think it's a step in the right direction," he offers. "It's been talked about for a good while about how the means for testing compounds for carcinogenicity should be modernized or made more flexible or brought up to date."

Supporters who gaze into the future believe that if biological advancements continue, the alternative models proposed as supplements in the guideline could some day entirely supplant the current bioassays.

Stoll, for example, believes that officials at the National Toxicology Program at the National Institute of Environmental Health Sciences (NIEHS) in Research Triangle Park, N.C.; at FDA; at the Environmental Protection Agency; and at other groups-"people in our government in high places"-hope for full replacement within a half-dozen years or so, as scientists learn more about the alternatives' positives and negatives. "These carcinogenicity bioassays, the two of them that companies have to do, are the most controversial, and the most difficult tests to do," he declares. "If one could ultimately be able to use shorter-duration [testing] using newer scientific methodologies and getting just as good, if not better, predictive answers in making a risk assessment, then society will be the winner, as well as the governments as well as the pharmaceutical companies. That's a little idealistic, but I believe that a number of these tests will hold up."

However, not everyone feels the guideline is a good thing. Moore, for example, calls it "a serious mistake." He is the author of Deadly Medicine (New York, Simon and Schuster, 1995), a case study of a family of drugs to treat irregular heartbeat that actually caused cardiac arrest. Moore criticizes what he sees as FDA's lax attitude in allowing potentially toxic products to reach the market without proper testing.

According to Moore, because so many dangerous drugs make it through existing measures-even despite testing positive in preclinical testing-altering the current system as proposed would decrease the information available to assess a drug's cancer risk. "If we eliminated testing in mice, which is the practical effect of this, we would miss a warning flag in about one out of five drugs that would have carcinogenic potential," he asserts.

Stoll notes that some other toxicologists are skeptical because the alternatives, as yet, have not been validated as scientifically accurate. He dismisses that by stating that since the current system of two bioassays also has not been validated, it should not be used to gauge the proposed changes. "You can't validate new methods against old methods that haven't been validated," he argues.

Moore replies that the current system, while flawed, is still the best currently available. "I don't think anyone can know or say that a drug found toxic in animals will be toxic in humans, but it's the only tool we have," he says. "All other diseases that hurt people are declining in prevalence, and cancer is not."

TRAILBLAZER: Raymond Tennant of the National Institute of Environmental Health Sciences helped to create some of the alternative testing models.
Disagreeing with Moore is Raymond Tennant, chief of the laboratory of environmental carcinogenesis and mutagenesis at NIEHS and a pioneer in the field of transgenic mouse studies (R. Tennant et al., Environmental Health Perspectives, 103:942-50, 1995). His work helped pave the way for the proposed guideline.

Tennant states that in addition to the two years the current system spends testing the two rodent samples, about three more years are needed to conduct "a thorough postmortem examination." The final tally, he estimates, is $1 million to $2 million.

"As more genes that were linked to [carcinogenesis] became identified," he explains, "and as the tools for manipulating those genes were developed, it was possible to create transgenic mice-those that possess new foreign genes and those with knocked-out genes-that could be plausible targets for the actions of these chemicals.

"Now, out of the literally hundreds of transgenic models that have been developed worldwide, we think we have identified a couple of models that can improve significantly on our ability to discriminate between carcinogens and non-carcinogens," Tennant continues. "That's an important aspect. It has to effectively discriminate between them, not just recognize carcinogens."

But Moore is unconvinced that substantial time and money would be saved under the new guideline. He asserts that because both rodent assays can be performed simultaneously, the current system does not cause delays in drugs reaching the market, and he compares the relatively small cost of the bioassays with the hundreds of millions of dollars a company spends to bring a new drug to the market. The risk, he claims, is just not worth it. "For the loss of data, this is not a major consequence."

Regarding other objections, Tennant maintains that he has heard complaints during meetings "that the systems are artificial, that they're the product of genetic engineering and don't relate to conventional rodents." He answers that by saying the standard use of inbred rodents "is just a less sophisticated form of genetic engineering. You haven't been as selective of which genes you're targeting."

Others have been concerned that transgenic mice will be "overly sensitive," responding incorrectly to chemicals that are not carcinogenic risks, but Tennant says there have been no false positives yet. "They're legitimate questions," he acknowledges. "The only way to resolve some of them is to get more data. That's what's going on right now."

But to Moore, altering the current bioassay system in favor of the new guideline would reduce the available data. "No one can be sure quite what to make of the animal studies when they show a danger, and [the drug has] already been tested in humans. The problem is we don't know enough about the drugs that we give to millions of people as it is. It's unwise to start eliminating the sources of potential information."

Contrera is cheered by what he sees as his organization's unusually active role in streamlining requirements. That a regulatory agency such as FDA is spearheading a drive for revision, he feels, will spur industry to develop the alternative models. He adds: "The biological insight from these alternative models may give us better understanding of human risk. . . . Doing a new model, a transgenic, might give us new information we couldn't get from a current mouse study."

Tennant states: "Another important characteristic is that by being able to do the assays in a relatively short period of time, you diminish some of the variables, the confounding factors that make it so difficult to interpret the bioassays."

Tennant further predicts a great increase over the next year in the number of chemicals and drugs that will be tested according to the new models. "We'll have a much clearer picture of whatever uses and limitations there are. The models are definitely going to be valuable tools for drug and chemical safety evaluation."

A very significant implication of the new guideline, according to its supporters, is that a regulatory agency is using cutting-edge scientific advances as the basis for streamlining. "What the chemists and the toxicologists can do in judging particular molecular compounds has advanced and gotten a lot more sophisticated," Heath observes. "We need to keep abreast of those kinds of developments. If they can provide more accurate, more sensitive measures of predicting human toxicity than the longer, tedious animal experiments which many people nowadays think may well mislead, that's a step in the right direction."

But George Washington's Moore would be quite happy to scrap the revisions and work simply to improve the current system. He terms the alternatives a way to "make the bad news of carcinogenic properties of drugs go away by not doing the testing." He states: "I don't think that's a wise policy decision. To me, we ought to be moving toward higher toxicity standards for our approved drugs, not trying to make some of the uncomfortable news go away."