Editor's Note: This story is part one of a two-part series on new insights into p53 regulation and function. Part two, on the role of p53 in Huntington's disease, will run in Thursday's issue.

Never trust a test tube.

That's the take-home message from a paper by scientists from the Salk Institute in San Diego and the Institut Pasteur in Paris published in the July 8, 2005, early online issue of the Proceedings of the National Academy of Sciences.

In their publication, the scientists investigated how the tumor suppressor p53 is regulated in vivo. P53 has long been known as an important tumor suppressor gene.

Under normal conditions, the p53 protein is unstable and found only at very low levels in the cell. But when DNA damage occurs, p53 levels rise and either initiate protective measures or apoptosis, depending on the severity of the damage.

P53 is mutated in more than half of all human cancers and is the target of a number of drug development efforts by biotechnology companies including Cleveland Clinic spin-off Cleveland BioLabs; Austin, Texas-based Introgen Therapeutics Inc.; Gemin X Biotechnologies Inc., of Montreal; and Reata Discovery Inc., of Dallas.

Most drug development efforts to date target p53 directly or its regulation by the protein MDM2, a negative regulator of p53 that facilitates its degradation. But Geoffrey Wahl, professor at the Salk Institute and senior author of the study, told BioWorld Today that the oncogenic activity of p53 may occur "not through changes in p53 itself, but through changes in MDM2 and MDMX regulation."

MDMX is another regulator of p53. Specifically, Wahl said that "the MDM2/MDMX ratio is critical to the regulation of p53. It is disturbed in many human cancers. Targeting this ratio is really where the future is at, in our opinion."

Wahl based his opinion on results he achieved with a transgenic mouse that has mutations in a part of the protein long thought to be critical for its stability: a group of lysines near the C-terminal. Previous experiments in tissue culture had indicated that those lysines are critical to the stability of p53 - so much so, in fact, that the scientists created an inducible model for fear that the mutated p53 would cause early embryonic lethality.

Instead, the mice, which had a string of arginines in place of the usual lysines, were pretty much indistinguishable from wild-type littermates, prompting the scientists to further investigate the function of the mutated protein by going back to cell cultures. What they found was that the mutant protein had a normal half-life, was activated like its wild-type counterpart in response to DNA damage, and was degraded normally by the cellular machinery. Longer-term culture experiments spanning several cell divisions revealed that after about 10 generations, cells containing the mutant protein failed to divide further. Overall, the authors concluded from their data that "C-terminal modifications believed to be critical for proper p53 regulation are not essential, but contribute to a fine-tuning mechanism of homeostatic control in vivo."

Of course, particularly in cancer research, many people will assert that "never trust a mouse" is a good motto, too; if every miracle cure in mice panned out in the clinic, average life expectancy would be 120 years, cancer-free. Asked about the relative merits of different model systems, Salk scientist and first author Kurt Krummel told BioWorld Today that mouse models tend to be a roulette game in drug screening, where tumor cells are injected into mice and different drugs screened for their ability to retard the growth of those tumors. Such screening experiments are "very crude," Krummel said.

"But," he added, "that doesn't mean that mouse models can't be extremely powerful and predictive" when working out cellular pathways. What it takes, in Krummel's opinion, is "smart mouse models, and the ability to perturb the system very specifically."

As for cell culture experiments, Krummel noted that "we are not saying at all that they are a waste. What we are saying is it's necessary to think carefully about what sorts of experiments really make sense in a cell culture model, and to not over-interpret the data you can get from them."