The role of genetic polymorphisms in cytochrome P450 and effects of tuberculosis co-treatment on the predictive value of CYP2B6 SNPs and on efavirenz plasma levels in adult HIV patients

Highlights

• The role of 10 SNPs on EFV levels with or without tuberculosis (TB) treatment was studied.

• CYP2B6 516T/T is associated with high EFV levels when with or without TB treatment.

• CYP1A2 −739T/G is correlated to EFV levels during HIV and TB co-treatment only.

• CYP2B6 SNPs predict EFV levels with a higher positive predictive value (PPV).

• TB treatment decreases the PPV and EFV levels, but does not affect the significant reduction of viral load.

Abstract

Efavirenz (EFV) exhibits interindividual pharmacokinetic variability caused by differences in cytochrome P450 (CYP) expression. Most tuberculosis (TB) drugs interact with the CYP metabolizing enzymes, while the clinical validity of genotyping in predicting EFV plasma levels in Rwandan subjects is not known. We investigated in patients co-infected with human immunodeficiency virus (HIV) and TB recruited in Rwanda the effects of 10 SNPs in five drug-metabolizing enzymes on EFV plasma levels and treatment response when patients are treated with EFV-containing therapy alone (n = 28) and when combined with rifampicin-based TB treatment (n = 62), and the validity of genotyping for CYP2B6 single nucleotide polymorphisms in predicting supra-therapeutic EFV levels. There was a significant difference between CYP1A2 −739T/G and T/T genotypes when patients were treated with EFV-containing therapy combined with rifampicin-based TB treatment, but not when EFV-containing therapy was alone. CYP2B6 516T/T genotype was associated with high EFV levels compared to other CYP2B6 516G>T genotypes in the presence and in the absence of rifampicin-based TB treatment. Predictive factors of EFV plasma levels in the presence of rifampicin-based TB treatment were CYP2A6 1093G>A, CYP2B6 516G>T, and CYP2B6 983T>C accounting for 27%, 43%, and 29% of the total variance in EFV levels, respectively. There was a high positive predictive value (PPV) (100%) for CYP2B6 516T/T and 983T/T genotypes in predicting EFV plasma levels above the therapeutic range, but this PPV decreased in the presence of rifampicin-based TB treatment. Rifampicin-based TB treatment was also shown to affect EFV plasma levels significantly, but did not affect the significant reduction of HIV-RNA copies. These results indicate that genotyping for CYP2B6 SNPs could be used as a tool in predicting supra-therapeutic EFV plasma levels, which could minimize adverse drug events.

Introduction

Among people infected with human immunodeficiency virus (HIV), tuberculosis (TB) is the most common opportunistic infection, with almost 50% of TB patients in Africa being co-infected (WHO, 2012). Managing HIV and TB infections requires multidrug therapy. The standard regimen for the treatment of TB is a combination of two to four drugs, where a rifamycin agent (usually rifampicin) is combined with non-rifamycin agents (ethambutol, isoniazid, and pyrazinamide). Streptomycin may be indicated when one of these drugs is contraindicated (WHO, 2010). The recommended first line antiretroviral therapy (ART) for the management of HIV/AIDS involves a combination of three drugs consisting of two nucleoside reverse transcriptase inhibitors with a non-nucleoside reverse transcriptase inhibitor (NNRTI). Efavirenz (EFV) is the NNRTI commonly administered as part of ART either in patients infected with HIV alone or co-infected with TB (Pozniak et al., 2011).

EFV plasma levels below 1 μg/ml and above 4 μg/ml have been associated with increased risks of virologic failure and central nervous system side effects, respectively (Marzolini et al., 2001). EFV is mainly metabolized in humans through hydroxylation by cytochrome P450 (CYP) 2B6, and to a lesser degree by CYP1A2, CYP2A6, CYP3A4, and CYP3A5 (Bélanger et al., 2009, di Iulio et al., 2009, Ogburn et al., 2010, Pozniak et al., 2011, Ward et al., 2003). EFV has been shown to exhibit interindividual variability in pharmacokinetics, which is thought to be caused by interindividual differences in CYP activities and expression (Bélanger et al., 2009, di Iulio et al., 2009, Holzinger et al., 2012, King and Aberg, 2008, Sánchez et al., 2011). Thus, factors affecting the metabolism of EFV could influence EFV exposure and the treatment response. Genetic influence is the most important among multiple factors affecting the pharmacokinetics of EFV (Sánchez et al., 2011). To our knowledge, there is no study that investigated in Rwandan adult HIV subjects the relationship between EFV exposure and single nucleotide polymorphisms (SNPs) with respect to EFV metabolizing enzymes, and the utility of genotyping as a way of improving the prediction of EFV plasma levels.

Despite the important role of genetics, drug–drug interactions deserve a particular attention because it is difficult to predict the outcome of complex interactions, such as those that might occur when several substrates for same CYP enzymes are used concomitantly, some being inhibitors or inducers of the CYP system (CDC, 2007). Non-rifamycin agents have been reported to induce or inhibit the hepatic CYP enzymes (Court et al., 2013, Rodríguez-Nóvoa et al., 2006, Manzi and Shannon, 2005, Wen et al., 2002) and interactions between rifamycin TB agents and antiretrovirals have been investigated, with focus on the induction effect of rifampicin on CYP2B6 enzyme and its impact on EFV disposition in different populations (Kwara et al., 2011a, Rodríguez-Nóvoa et al., 2006). Thus, the therapeutic management of HIV/TB co-infection may result in complex pharmacokinetic drug–drug interactions taking place collectively between ART components and TB drugs, impacting on blood levels of both HIV and TB drugs. Here we hypothesized that in HIV/TB co-infected patients, the genetic influence on EFV levels when patients are treated with EFV-containing therapy alone may differ from when the patients are treated with EFV-containing therapy combined with rifampicin-based TB treatment.

Therefore, the aim of this study was twofold, first to investigate the effects of 10 SNPs in five drug-metabolizing enzymes on EFV levels and treatment response in adult HIV/TB co-infected patients recruited in Rwanda when they are treated with EFV-containing therapy alone and when combined with rifampicin-based TB treatment, and secondly, to evaluate the effects of rifampicin-based TB treatment including its impact on the validity of genotyping for CYP2B6 SNPs in predicting EFV plasma levels above 4 μg/ml.