Of mice and men

Could 'human' mice have foiled
drug trial fiasco?

By Owen Dyer

While there's no prospect of getting an animal to perfectly reproduce human responses, there may be a shortcut, according to researchers at Israel's Technion Institute. In the latest issue of Cancer Research, Dr Karl Skorecki and colleagues report that by injecting human embryonic stem cells into mice, they were able to grow a teratoma — a growth made up of a mixture of human cartilage, blood vessels, fat tissue or connective tissue — allowing them to conduct "human" trials in animals. Dr Maty Tzukerman, PhD, one of the Technion researchers, said the research "helps address concerns raised by some in the scientific community about whether structural similarities between animals and humans actually indicate similarity on the functional level."

Of course, the first test in a human will always be necessary, but this mouse model may make it less of a leap into the unknown. Injecting cancer cells into the "human" microenvironment provided by the teratoma, the Israeli team was able to generate human tumours without resort to a human patient. This could make the first stages of testing cancer drugs far safer, enabling researchers to look at tumour development, drug efficiency and side effects in a human tissue environment.

WOULD IT HAVE HELPED?
It's hard to say whether testing the now-defunct TGN1412 in such a model would have prevented the multiple organ failure experienced by the six test subjects. Their problems stemmed from complex immune system components, leading many to believe the drug, which was developed by German biotech firm TeGenero, caused a disastrous overstimulation of T cells.

While other monoclonal antibodies have been safely developed for a variety of diseases, TGN1412 packs an extra punch. A so-called "superagonist," the drug binds to a receptor on the surface of T cells called CD28, triggering an immune response. But while most monoclonal antibodies require a second factor to bind CD28 to activate T cells, TGN1412 bypasses that co-stimulation step, provoking an immediate, powerful reaction.

Dr Janet Darbyshire, director of the clinical trials unit of the UK Medical Research Council, said that because TGN1412 is a fully humanized monoclonal antibody, its disastrous effects in humans might never have showed up in animal testing, even in primates. Genetically modifying mice to form human tissue is a possible answer, she agreed.

UNPREDICTABLE RISK
The British Medicines and Healthcare products Regulatory Agency (MHRA), which investigated the circumstances surrounding the drug trial, confirmed there were "no obvious errors" that could explain the catastrophic adverse events. Its inquiry confirmed the volunteers' reaction was unexpected and had not been seen in preclinical studies.

As is standard in phase I human trials, the six subjects started with a dose only one-five hundredth as strong as where the animal dosing had left off. But within three hours of getting their injections, the men began feeling cold, then developed severe headaches and swelling of the face and neck. Shortly thereafter, their organs began shutting down. The trial was taking place on the campus of a large teaching hospital, a lucky circumstance that probably saved their lives.

Though five of the six volunteers have been discharged from hospital, 20 year-old Ryan Wilson is still being treated for the complications he suffered, and may now lose parts of his fingers and toes as a result.

TRIAL OVERHAUL
While established procedures do seem to have been followed, they have since been called into question. The approval process and regulation of phase I trials is under review following the release of the MRHA's interim findings.

In a flash of the sort of insight that always seems to come with hindsight, several clinical trials experts have pointed out that it's probably not a good idea to give an untested drug to six people simultaneously. In fact, the trial did stagger the subjects' injections at ten-minute intervals, watching for adverse reactions, but the first apparently set in after the last patient had been injected. Even that tiny delay between injections was unusual in phase I trials, according to the MRHA. Eighty percent of phase I trials, the agency said, involve no dose staggering at all.

That's one change likely to come out of the debacle. But it would clearly be better to avoid endangering anyone. As Dr Michael Goodyear of Dalhousie University recently wrote in the BMJ, "with increasing sophistication of molecular targeting using specific human receptors, the applicability of the mouse as a model becomes questionable. The CD28 T cell surface receptor, the target of TGN1412, shares only 68% of identity of amino acids between mouse and man. Relative lack of severe toxicity in animal models should never be construed as a guarantee of safety in man."