Pharmacogenetics of antiretrovirals

Abstract

The introduction of highly active antiretroviral therapy (HAART) as standard of care has changed the natural history of HIV infection into a manageable chronic disease requiring long-term antiretroviral (ARV) treatment. However, response to HAART is often limited by the occurrence of toxicity or by the emergence of drug resistance. Antiretroviral treatment is characterized by differing rates of adverse events and responses. Genetic variations between human beings account for a relevant proportion of this variability. A relevant number of associations between human genetic variants and predisposition to adverse events have been described and for some antiretroviral drugs a clear and casual genotype–phenotype correlation has already been established. The strong association between abacavir hypersensitivity reaction and HLA-B*5701 has been demonstrated in both observational and blinded randomized clinical trials in racially diverse populations and represents the best example of the clinical utility of pharmacogenetic screening in HIV medicine. Genotyping for HLA-B*5701 before prescribing an abacavir containing regimen has been introduced into routine clinical practice as the standard of care for all patients. Other well-established associations include CYP2B6 alleles and efavirenz central nervous system side effects, UGT1A1 alleles and atazanavir-associated hyperbilirubinemia and HLA class II allele HLA-DRB*0101 and nevirapine-associated hypersensitivity. Despite genetic associations having been described for peripheral neuropathy, lipodystrophy, hyperlipidaemia, pancreatitis and renal proximal tubulopathy, numerous barriers exist to the successful introduction of widespread genetic testing to the clinic. Future prospects point in the direction of individualization of antiretroviral therapy through insights from host genetics. The present paper is aimed to provide a comprehensive review of the published literature and to summarize the state of research in this area.

This article forms part of a special issue of Antiviral Research marking the 25th anniversary of antiretroviral drug discovery and development, Vol 85, issue 1, 2010.

Introduction

The introduction of highly active antiretroviral therapy (HAART) as a standard of care has considerably enhanced the life expectancy among HIV-infected individuals. Over the last 20 years antiretroviral (ARV) therapy have moved from almost ineffective monotherapy to combination multidrug regimens able to virtually suppress viral replication in most HIV-infected patients. As a consequence, the natural history of HIV infection has been changed into a manageable chronic disease requiring long-term ARV treatment. However, response to HAART is a complex phenomenon and is often limited by the occurrence of acute or chronic toxicities or by the emergence or drug resistance. Drug metabolism and toxicity may vary greatly between individuals, affecting both efficacy and toxicity. It is believed that genetic variations between human beings account for a relevant proportion of this variability. As a matter of fact HIV-infected patients consist of individuals who differ genetically with regard to their response to both the virus and the ARV drugs.

The term pharmacogenetics refers to the effects of polymorphisms within human genes on drug therapy outcomes. By definition, single-nucleotide polymorphisms (SNPs) are defined as sequence variations that occur in human DNA with single-nucleotide changes occurring at an allele frequency greater than 1% (Hoehe et al., 2003). Nucleotide changes occurring with a frequency lower than this are referred to as mutations. Recent advances in technologies for genetic analyses are generating great opportunities to disclose the role of sequence variation in human genome influencing disposition, metabolism, efficacy and toxicity of ARV drugs.

The possibility to perform genome-wide analyses has recently disclosed new possibilities of research (Syvanen, 2005). However, up to now most genotype–phenotype studies have employed a direct, hypothesis-driven candidate gene approach. In most cases the study hypotheses were to verify a plausible link between a genetic variations of probable impact on drug metabolism and/or on drug toxicity and phenotype under study. According to the hypothesis-driven candidate gene approach variability in the host genome influencing ARV efficacy and tolerability has explored genetic factors involved into immunological and pharmacokinetic determinants of responses.

A great number of associations have been reported between host genetic polymorphisms and responses to ARV drugs. These include drug pharmacokinetics and pharmacodynamics, hypersensitivity reaction syndromes, hepatotoxicity, central nervous system side effects, hyperbilirubinemia, peripheral neuropathy, lipodystrophy, hyperlipidaemia, pancreatitis and renal toxicity. However, it must be underlined that numerous barriers exist to the direct translation of this body of knowledge toward the ultimate goal represented by the individualization of ARV therapy. The risk of false discoveries caused by multiple testing is a well-known problem in statistical genetics. Thus, some early reports on genotype–phenotype associations must be considered with caution until replicated in additional studies. Most studies published to date have significant limitations represented by small study size, lack of adequate statistical power and presence of selection bias. Moreover, the carriage of variant alleles is very often linked to ethnicity. As a consequence, the risk of ethnic bias exists in almost all studies of genotype–phenotype association. Thus, several prerequisites exist for successful introduction of a pharmacogenetic test into routine clinical practice. The test must be clinically relevant with high sensitivity and specificity. The evidence showing genotype–phenotype association should be ideally based on randomized, double-blind, prospective studies involving patients of different ethnicities. Large clinical trials and observational cohorts conducted across racially diverse populations are also extremely useful for association studies. Moreover, the genotypic assay should be rapid and simple to interpret. Finally, robust cost-effectiveness data should be provided to support its reimbursement.The present paper is aimed to provide a critical and comprehensive review of the published literature, to summarize the state of research and to provide insights into future prospects in this area.