Basis Of Anticancer Drug Resistance In Childhood Leukemia Revealed

Besides its use in ALL, methotrexate is widely used to treat other cancers and some autoimmune diseases. However, until the new study there was no valid test for analyzing the genetic basis of resistance. Such genetic analysis is important in childhood ALL because, although 80 percent of children with the disease can be cured, determining the basis of drug resistance in the other 20 percent would help increase the cure rate.

Researchers have successfully used laboratory studies of leukemia cells to explore the basis of resistance in other anti-leukemia drugs. However, according to Evans, such in vitro tests have not worked with methotrexate. The researchers analyzed the genetic profiles of St. Jude patients undergoing methotrexate treatment for ALL to identify genes that governed their response to the drug.

In their study of 161 ALL patients, they measured the response to initial methotrexate treatment and then used gene microarray analysis to measure the activity levels of 12,357 genes in the patients. In microarray analysis, researchers apply genetic material from the patients' leukemia cells to small "gene chips" on which samples of thousands of genes are arrayed. Researchers can analyze the reactions on the gene chip to each gene to measure the level of expression for those genes in the patient samples.

"In our analysis, we identified a large number of genes in the treated patients that differed in their expression level at a very significant level statistically," Evans said. "We elected to focus on the 50 most highly significant genes."

Among the genes were those involved in DNA synthesis, its components and repair of DNA. The identity of some of these genes was not surprising because the drug kills leukemia cells by interfering with their ability to replicate their DNA.

When the researchers compared the gene expression patterns of patients who responded well to methotrexate to those who responded poorly, they found distinct gene expression "profiles" among the groups. In further analysis to validate their findings, they found that the profiles predicted methotrexate response in an independent group of patients: Patients with gene expression profiles indicating a good methotrexate response had significantly better five-year, disease-free survival than those with profiles indicating a poor response.

To confirm their findings, the researchers also analyzed the predictive effects of those distinctive profiles in an independent group of 18 patients. They found that the gene expression profiles for the top 50 genes also predicted methotrexate response in those patients.

Further exploration of the genes identified in this study could yield clinical benefits. "Some of these could become potential targets for developing other drugs that would make methotrexate more effective in those children who are resistant," Evans said.

For example, one gene they identified as relevant to resistance produces a protein that transports the drug out of the leukemia cell. "It might be possible to give a drug along with methotrexate that blocks this transporter, which would make methotrexate more effective without having to give another cytotoxic drug," Evans said.

In further studies, the researchers plan to search for such drug targets. They will also search for subtle genetic differences among patients in the response-related genes in search of inherited genetic differences that might explain gene expression and methotrexate response. Finally, the researchers will explore whether patients who respond poorly to methotrexate have specific gene deletions or other genetic alterations in their leukemia cells that cause such poor response.

The findings broadly confirm the value of such sweeping surveys of gene expression in understanding response to anti-cancer drugs.

"Studies such as these add another piece of evidence that this genome-wide approach is very insightful and helpful and informative," Evans said. "If you simply look for the genes that you think might be important, you are likely to miss a number of genes that are."

Journal reference: Sorich MJ, Pottier N, Pei D, Yang W, Kager L, et al. (2008) In vivoresponse to methotrexate forecastsoutcome of acute lymphoblasticleukemia and has a distinct geneexpression profile. PLoS Med 5(4): e83.

Other authors include Deqing Pei, Wenjian Yang, Cheng Cheng, Ching-Hon Pui, Mary Relling, John Panetta and Meyling Cheok (St. Jude); Michael J. Sorich (University of South Australia); Nicolas Pottier (Pole Recherche, France); Leo Kager (St. Anna Children's Hospital, Austria); and Gabriele Stocco (University of Trieste, Italy).

This work was supported by National Institutes of Health grants, a Cancer Center Support Grant, a FM Kirby Clinical Research Professorship from the American Cancer Society and ALSAC.