Inherited Variants Of Three Genes Affect Breast Cancer Treatment Response And Outcome

If future studies show that these genetic variants can predict response and outcome, a simple blood test and genetic analysis might help physicians select the best treatment options for each patient, the researchers said. Besides breast cancer, this pre-treatment genotyping might be useful to tailor treatments for numerous other diseases, given the three genes' broad involvement in drug processing pathways, said lead author William Petros, assistant clinical professor of medicine at Duke Comprehensive Cancer Center.

"Despite giving doses based on patients' height and weight, the blood concentrations of drugs vary widely, and we think that's genetically based," said Petros, who prepared the results for presentation Monday at the annual meeting of the American Association for Cancer Research. "In cancer treatment, factors that affect blood concentration can affect the thin line between effectiveness and intolerable toxicity.

"The link we've found between these three genes and patient response is potentially translatable to other situations and may be clinically applicable, but confirmatory studies need to be done," Petros said.

The study was funded in part by the National Institutes of Health.

The researchers from Duke and a collaborating company, PPGx of La Jolla, Calif., discovered that inherited variations in two genes, called CYP3A4 and CYP3A5, were associated with poorer response to high-dose chemotherapy treatment and poorer survival. Variants of a third gene, called GSTM1, provided a better response and longer average survival, they reported.

"CYP3A4 and CYP3A5 play important roles in the metabolism of over half of the drugs used clinically, not just drugs used in treating cancer, while GSTM1 detoxifies and removes drugs that get inside cells," Petros said. "Whether the variations are 'good' or 'bad' would depend on the drug being given."

Duke researchers reviewed the records of 86 patients with metastatic or inflammatory breast cancer who had been treated as part of a Duke clinical trial between 1988 and 1991. Duke scientists also extracted DNA from each patient's stored white blood cells, which had been collected prior to any treatment, and shipped the DNA to PPGx, where scientists analyzed each sample for 22 variants, or single nucleotide polymorphisms, in 12 drug metabolism genes.

"Most genetic work in cancer uses cancer cells, but drug metabolism is a process that takes place throughout the body," said Petros. "We did not examine the patients' tumors for this study, but instead looked at overall effects these variants could have on drug metabolism and sensitivity. Clearly, acquired genetic mutations that appear in cancer cells and not in normal cells are important in response to treatment as well."

Three significant gene variants were identified by combining the genetic information with the outcome -- the patient's blood levels of the three chemotherapy drugs used, her response to treatment and her survival. The researchers also were able to explain how these gene variants might affect the chemotherapy drugs the patients had received -- cyclophosphamide, cisplatin and BCNU, which is also called carmustine.

Cyclophosphamide has a complicated journey in the body. It is called a "pro-drug" because it must be metabolized to the active form by CYP3A4 or CYP3A5 proteins, which are primarily located in the liver. Administration of cyclophosphamide stimulates production of these metabolism proteins, so subsequent doses are converted to the active form even more rapidly.

"Cyclophosphamide induces its own activation," Petros said. "By examining the blood levels of active and parent cyclophosphamide after each treatment, we get an estimate of how effective the drug might be in any given patient. It is not good when patients have high blood concentrations of the inactive parent drug after the last treatment."

Most cells in the body have two copies of each gene. The researchers found that patients with one or two variant copies of CYP3A4 or CYP3A5 had significantly more parent drug in the blood on the last day of treatment compared to those who had two normal copies of each gene.

This implies patients with variant copies would have less active drug available, the researchers said. "The inactive parent drug doesn't seem to be efficiently metabolized to the active form in patients carrying at least one variant copy of CYP3A4 or CYP3A5," Petros said.

The researchers also report that patients with two normal copies of CYP3A4 and CYP3A5 survived longer on average compared to patients with at least one variant copy. For carriers of CYP3A4 variants, average survival after start of treatment was 1.6 years, while carriers of the CYP3A5 variant had an average survival of 1.3 years. Patients with neither variant had an average survival of 2.9 years after start of treatment.

Patients with at least one variant copy of GSTM1 in their cells, however, had better outcomes than patients with two normal GSTM1 copies. GSTM1 variants don't function, so patients with two variant copies of GSTM1 can't detoxify drugs that get inside cells through this pathway.

Patients with two variant copies of GSTM1 had an average survival of 3.8 years after treatment start, while survival for patients with at least one normal copy of the gene was 1.8 years.

"Patients with two variant GSTM1 copies had a higher complete response rate, longer survival and lower blood levels of BCNU," said Petros. "The survival and response information makes sense because patients without GSTM1 can't detoxify BCNU, allowing the drug to have a greater effect."

He said because the patients had been treated as part of a clinical trial, all had very similar disease and had received identical treatment and detailed follow-up after treatment, making the analysis more straightforward.

Co-authors on the study are Dr. James Vredenburgh, Gloria Broadwater, Dr. Michael Colvin and Jeffrey Marks of Duke, and Penelope Hopkins, Susan Spruill and Jeff Hall of PPGx, which is now a fully owned subsidiary of DNA Sciences, Fremont, Calif.