Important note: Information in this article was accurate in February 2002. The state of the art may have changed since the publication date. 
American Foundation for AIDS Research, February 2002
Gretchen Schmelz
Researchers have long wondered why people respond differently to drugs. This question also occurred to Swiss investigators who have been tracking a 5,000-member observational study group that records the experience of people with HIV (the "Swiss HIV Cohort"). They noticed differences in peoples' drug levels despite good adherence. "Now, everyone gets the same amount of medication, but patients are well aware that they respond differently to drugs," said Amalio Telenti of the University Hospital of Lausanne, Switzerland. "Two patients taking the same drug, with perfect compliance, can have large variability – as much as ten times difference in drug response."
| A Snip Here, A Snip There |
|
Pharmacogentics Pharmacogeneticists believe that part of the explanation for treatment success and failure is traceable to the human genome. The DNA in genes is composed of precise sequences of the four building blocks. These are the nucleotides A (adenosine), C (cytidine), G (guanosine) and T (thymidine). A gene can vary slightly from person to person - at times, by a single nucleotide. This single nucleotide difference is a mutation, but when it occurs in more than 1% of a population, it’s called a polymorphism. Single nucleotide polymorphisms (SNPs, pronounced snips) can impact how well, or poorly, people respond to drugs. 
This is because genes code for proteins - with two independent genes for each protein - and certain proteins interact directly with drugs. There are three kinds of drug-response proteins: drug metabolizers, transporters, and receptors. Metabolizers, like proteins from the cytochrome P450 family, break down drugs. Receptors and transporters move them in or out of a cell. Mutations can moderately affect the protein, or render it useless. P-glycoprotein is an efflux transporter. It protects cells by pumping drugs and other substances out of cells. In the example shown here, polymorphisms in the p-glycoprotein gene affect gene expression. People born with the CC genotype (a) have twice as much p-glycoprotein on cell membranes as those who have the TT allele (c). Those born with both genes express intermediate amounts (b). |
Telenti and his colleagues decided to look at the way genetic variations influence drug processing within the body and, hence, an individual's response to therapy. (J Fellay et al., Lancet, January 5, 2002; pages 30-36) see sidebar for link to abstract Although the group's results are not pivotal, their approach marked a turning point: Pharmacogenetics promises to transform the way doctors treat people with HIV.
If genes did contribute to drug response, Telenti's group reasoned, then people with a particular genetic makeup (genotype) would improve faster when starting medication. (In this study, improvement was measured by looking at CD4 cell count increases after six months. Viral load decreases were also measured in participants who were starting therapy for the first time.) They chose seven well-documented genes implicated in drug response: four drug metabolizers from the cytochrome P450 family of liver enzymes (CYP3A4, CYP3A5, CYP2D6, and CYP2C19), a transporter (p-glycoprotein) that pumps drugs out of cells, and a cell membrane receptor protein (CCR5) that helps make cells susceptible to HIV.
The study followed 203 white patients with high adherence to dosing schedules, as measured by sustained viral suppression and/or consistent serum drug levels. Everyone in the study group took either nelfinavir 1,250 mg twice daily or efavirenz 600 mg once daily. (87% also took nucleoside analogs; and 13% followed a regimen that contained nelfinavir plus a second protease inhibitor). CD4 cell counts and viral loads were recorded at baseline, one, three, and six months of treatment. Everyone also underwent gene analysis.
Only genetic differences in CYP2D6 correlated with drug levels. People who had two genes that made standard amounts of this enzyme were able to metabolize nelinavir and efavirenz efficiently, and therefore had lower serum drug levels compared with those who were poor metabolizers. But these differences did not extend to differences in CD4 cell counts.
Alternatively, p-glycoprotein production did influence CD4 count increases after people began therapy. P-glycoprotein is an efflux transporter: it pumps drugs like nelfinavir out of cells. People who had two genes that had a thymidine nucleotide at a particular point in both members of the gene pair (the TT genotype) made less p-glycoprotein that those with a cytidine at the spot in one or both of the genes (the CT or CC genotypes).
Study participants who had the TT mutation and less p-glycoprotein on their cells also had lower drug concentrations. This association was true for nelfinavir, which is a known substrate of p-glycoprotein, and efavirenz, which has not been shown to be subject to p-glycoprotein's efflux pumping.
Among 137 study members beginning therapy, those with the TT genotype had a significantly greater rise in CD4 cell count at six months (257 cells/mm3) than people with CT or CC genotypes (165 cells/mm3 and 121 cells/mm3). Similar results were found in a second group of 80 patients commencing therapy – with either abacavir plus amprenavir or abacavir, nelfinavir and a second protease inhibitor. The investigators then analyzed the naïve CD4 cells (newly matured cells issuing from the thymus gland) among these 80. They found that the TT genotype also correlated significantly with increases in this CD4 cell subpopulation.
It is strange that lower drug levels in the blood correlated with greater CD4 cell gains. Another apparent paradox is that there was no significant correlation between p-glycoprotein production and reduced blood viral loads. One possibility is that reduced p-glycoprotein allows drugs to leave the blood more quickly and penetrate tissues and cells more extensively, especially those that the drugs have difficulty penetrating due to p-glycoprotein action. Viral blood levels might be unaffected by this difference in drug levels, but observers would find that cellular and tissue viral loads are lower, if these could be measured. At the least, the suppression of HIV after starting therapy might be faster with the TT genotype.
"This study was a good attempt at trying to figure out what genetic information may predict drug exposure or efficacy," concluded Angela Kashuba, Assistant Professor at the University of North Carolina Chapel Hill and member of the Adult AIDS Clinical Trials Group Pharmacology Committee. "But the data are not compelling enough to do anything therapeutically right now."
Telenti is continuing to research the complexities of p-glycoprotein's role. He predicts that more pharmacogenetic studies will be under way soon. "Until now, there was not much activity in this field," said Telenti. "We have never, until now, looked in vivo. The human genome was published just last year. In principal, we are just beginning to use the information that is available. There will be more and more people [in HIV research] trying to look at genes in the same fashion."
Some HIV researchers are working to make more data available. Dr. Alastair Wood, a pharmacologist at Vanderbilt University and chair of the Pharmacogenomics Working Group for the NIH's Adult AIDS Clinical Trials Group, has proposed that everyone entering an ACTG trial sign a consent form indicating whether they would be willing to donate DNA for future genetic testing. Wood and Dr. David Haas (also of Vanderbilt University) are co-chairs of protocol A5128, "Plan for obtaining informed consent to use stored human biological materials for currently unspecified analyses." The idea is twofold: to obtain proper consent to use samples that have already accumulated from past trials, and to build the archive with DNA from newly enrolling participants. That would give researchers access to potentially 50,000 DNA samples from well-characterized patients – a gold mine of pharmacogenetic data for HIV drug therapy research.
Genetic variations affect drug response, and they differ in frequency across different ethnic groups. One example is the CC genotype for p-glycoprotein. As Telenti discovered, this CC gene pair predicts a poorer CD4 cell recovery while on therapy. The genotype occurs with a 25% frequency in whites; in African blacks, the frequency is estimated to be between 67% and 83%. Another example is a mutation in CYP2D6 that leads to poor drug metabolism. It is present in 3% to 10% of whites, compared with less than 2% in Japanese, Chinese, and African Americans. Consequently, drug efficacy varies among ethnic groups.
"We don't know if antiretrovirals function in the same way in different ethnic groups or if they will work better, or if they will produce more toxicities," explained Telenti. "We don't know much, and the data are confounded by social issues. When we analyze a particular group that has a poor response to a drug, we quickly blame it on social differences. Now, by looking at genes, we are saying there is one more factor."
Dr. Scott Wegner and colleagues of the U.S. Military HIV Research Program noticed that African Americans coming in to their clinic were not responding as well to efavirenz as Caucasians. They were in a unique position to look at race and drug response because their patient population was highly homogenous in terms of the socioeconomic factors that Telenti cites. In addition to having similar baseline medical characteristics, the military personnel had similar educational levels and equal access to health care.
The Wegner group formed a study cohort of 373 persons taking two nucleoside analogs plus either efavirenz, indinavir, or nelfinavir. The primary endpoint was time from start of regimen to virologic failure. Median time to failure for both blacks and whites taking protease inhibitors was the same. For the subgroup taking efavirenz, half the blacks had experienced treatment failure by 440 days. This 50% failure level had not been reached for the whites after 1400 days, the limit of follow-up. The researchers could not attribute the dramatic differences to efavirenz levels in the blood, HIV resistance to efavirenz, or genotypic differences in CYP2D6, perhaps because of the small sample size (there were 56 blacks and 43 whites in this subgroup). The investigators also plan to look at differences in the CYP3A4 and p-glycoprotein genes – when they can obtain further funding.
"It's a very inflammatory topic," said Wegner, who presented results of racial differences in response to therapy at the 5th International Workshop on HIV Drug Resistance and Treatment last year. "I had people get up and say, 'How can you say there are differences? There are no differences.' Their arguments are naïve. And it seems to be threatening only to white people, to be honest. I've spoken to groups of African American physicians who say, 'You've got a lot of guts to stand up there and talk about this.' Clearly, there are genes over-represented in the African American population – all you have to do is look at the sickle-cell [anemia] trait. It's a crazy argument to think that there aren't differences. However, over-represented does not mean uniform, and what's very true for a population is only questionably true for an individual. But I think that it's reasonable to say that drug A is more likely to be effective in somebody with a specific demographic, whether it's a woman or man, black or white."
Pharmacogenetics has a clear role in penetrating beyond race to the basic genetic variations that could explain differences in therapeutic response.
"There will be benefits to pharmacogenetics, and we already have some examples," observed Wylie Burke, Professor and Chair of the Department of Medical History and Ethics at the University of Washington. "Treatment for childhood leukemia represents a very good example of what pharmacogenetics can do, because the child who is sensitive to mercaptopurine therapy due to low metaboism of this agent can be given a lower dose. It's not that anybody's cut out, it's just that you are using a genetic test to prevent a very serious adverse reaction in a minority of individuals. It wouldn't surprise me at all that we will see more very good examples…and we probably will see examples that aren't so good."
Prof. Burke raised two ethical concerns: "The first is the possibility that pharmacogenetics research will lead to the preferential development of drugs for some segments of the population while excluding others. It could be easier to develop a drug for people with a certain response profile than for others."
Getting new drugs on the market could be faster and cheaper for pharmaceutical companies, which could recruit people with the particular genotype that yields the best reaction to the drug. When residents of a developing country disproportionately have other genotypes, pharmacogenetics could increase the already profound divide in access to health care. Would genetic information be used to develop drugs targeted to poor populations that even now cannot afford to pay full price for drugs, or will that information merely form the rationale for further excluding them from medical care? And should genotyping become an important tool for determining individual therapy, who will pay for it in settings where resources are highly restricted?
"The other general ethical concern is that getting general information about drug response may also carry with it information about prognosis," Burke continued. "That raises even more questions. Is that good for a person or not? Could that have negative effects, in terms of discrimination and psychological effects? These questions get more disconcerting when the prognostic information has nothing to do with the disease at hand. And there are some hints that that in fact will play out."
020210
AM020203
Copyright © 2002 by the American Foundation for AIDS Research (amfAR) and first displayed on amfAR's Treatment Directory web site (http://www.amfar.org/gl). They appear on AEGIS with amfAR's permission. Organizations wishing to reprint or redistribute these materials should request authorization from amfAR's Department of Treatment Information Services (212/806-1600).
AEGiS is made possible through unrestricted grants from Boehringer Ingelheim, iMetrikus, Inc., the National Library of Medicine, and donations from users like you. Always watch for outdated information. This article first appeared in 2002. This material is designed to support, not replace, the relationship that exists between you and your doctor.
AEGiS presents published material, reprinted with permission and neither endorses nor opposes any material. All information contained on this website, including information relating to health conditions, products, and treatments, is for informational purposes only. It is often presented in summary or aggregate form. It is not meant to be a substitute for the advice provided by your own physician or other medical professionals. Always discuss treatment options with a doctor who specializes in treating HIV.
Copyright ©1980, 2002. AEGiS. All materials appearing on AEGiS are protected by copyright as a collective work or compilation under U.S. copyright and other laws and are the property of AEGiS, or the party credited as the provider of the content.
