Contact

Duke Health News

919-660-1306

DURHAM, N.C. -- Genes underlying HIV-infected patients' immune systems can be used to predict their risk of progressing to AIDS, a new study shows. The finding represents the first application of a novel statistical technique to the study of disease, said a Duke Medical Center researcher.

What's more, the analysis found, the virus is less likely to progress in patients bearing rare immune system gene variants than in those with more common gene variants. The research team reports its results in the July 2003 issue of Nature Medicine, which appeared online on June 22.

The finding clarifies the interaction between HIV and patients' immune defenses and provides information that may ultimately help physicians tailor treatments to individual patients, said Thomas Kepler, Ph.D., interim director of Duke's Center for Bioinformatics and Computational Biology, part of Duke's Institute for Genome Sciences and Policy. Patients who lack protective alleles -- variations of a gene that differ slightly in structure and function -- might benefit from more aggressive therapy, for example.

"Specific combinations of alleles that humans carry appear to protect against HIV," Kepler said. "HIV-infected people who carry particular, rare gene variants have much lower viral loads than other patients do." Viral load refers to the number of viral particles in a patient's blood plasma, a measure frequently used to measure HIV disease progression.

Joining Kepler in the collaborative effort were study leaders Elizabeth Trachtenberg of Children's Hospital Oakland Research Institute, Bette Korber of Los Alamos National Laboratory and Steven Wolinsky of Northwestern University, among others.

The study focused on 996 men -- 562 of whom were HIV positive -- enrolled in the Chicago component of the Multicenter AIDS Cohort Study (MACS). Launched in 1984, MACS is a study of the natural and treated histories of HIV infection in men. Other study sites include Baltimore, Pittsburgh and Los Angeles.

The researchers genetically screened study participants' blood samples for two immune system genes, human leukocyte antigen A and B (HLA-A and HLA-B). HLA molecules orchestrate the response of T-cells -- agents that defend the body against infected cells and invading pathogens. The many different HLA alleles have different biological activities that affect the body's response to invasion.

In analyzing the dozens of HLA-A and HLA-B alleles in the patients, the researchers identified some that conferred protection against HIV, while others left patients more susceptible.

However, immune system alleles are notoriously diverse, making it difficult to identify gene combinations that are important to patients' disease outcomes using traditional methods, Kepler said.

The researchers developed a novel statistical method, called minimum description length (MDL), to partition patients in a more detailed manner into disease progression groups based on their immune system genes. Although MDL has existed in theory, the research by Kepler and his colleagues represents the first time the method has been used for analyzing biomedical disease-association data.

"The method allowed us to exhaustively go through all possible gene partitions and assign a score to each, while avoiding the pitfalls of traditional methods of analysis." Kepler explained. "We were therefore able to identify the optimum means of dividing patients into genetic categories that correspond to viral load."

The statistical classification of the patients revealed that the genetic combinations that conferred the greatest protection were also the least common gene variants in the patient population, the team reported. For example, African-American patients, who were underrepresented in the study and bore less common immune system alleles, exhibited lower viral loads.

"Greater protection against HIV was afforded by rare immune system alleles," Kepler said. "That suggests that HIV has adapted to attack the dominant alleles in the population. In other words, the virus goes after the bigger target." The results also suggest that the gene variants that confer protection against HIV will vary among human populations, according to their frequency, he said.

By screening patients' immune systems, physicians might ultimately identify those patients at the greatest risk for progressing to AIDS and prescribe treatments accordingly, Kepler added.

Other researchers included Cristina Sollars, Elizabeth Hayes, Michael Hsu and Henry Erlich of the Children's Hospital Oakland Research Institute in Calif.; Robert Funkhouser, Michael Fugate and James Theiler of Los Alamos National Laboratory in N.M., Peter Hraber of the Santa Fe Institute in N.M., and Kevin Kunstman, Samuel Wu and John Phair of Northwestern University in Chicago. The National Cancer Institute and the National Institute of Allergy and Infectious Diseases funded the study.