LITERATURE REVIEW 11 - Population pharmacokinetics and pharmacogenetics of antiretroviral drug

Abstract

Besides therapeutic effectiveness, drug tolerability is a key issue for treatments that must be taken indefinitely. Given the high prevalence of toxicity in HIV therapy, the factors implicated in drug-induced morbidities should be identified in order to improve the safety, tolerability and adherence to the treatments. Current approaches have focused almost exclusively on parent drug concentrations; whereas recent evidence suggests that drug metabolites resulting from complex genetic and environmental influences can also contribute to treatment outcome.

Pharmacogenetic variations have shown to play a relevant role in the variability observed in antiretroviral drug (ART) exposure, clinical response and sometimes toxicity. The integration of pharmacokinetic, pharmacogenetic and metabolic determinants will more likely address current therapeutic needs in patients.

This chapter offers a concise description of three classes of antiretroviral drugs’ metabolic profile and a comprehensive summary of the existing literature on the influence of pharmacogenetics on their pharmacokinetics and metabolic pathways, and the associated drug or metabolite toxicity.

Antiretroviral drug toxicity in relation to pharmacokinetics, metabolic profile and pharmacogenetics

M.Arab-Alameddine^1,2, A. Telenti³, T. Buclin¹, L.A. Décosterd¹and C.Csajka^1,2

1Division of Clinical Pharmacology and Toxicology, University Hospital Center, University of Lausanne

2Department of Pharmaceutical Sciences, University of Geneva, University of Lausanne; Switzerland

3Institute of Microbiology, University Hospital Center, University of Lausanne, Switzerland

Expert Opin. Drug Metab. Toxicol. (2011) 7(5):609-62

Introduction

Important advances in the anti-retroviral therapy and development of new target molecules have been achieved within the last 20 years. Besides therapeutic effectiveness, drug tolerability is an important issue for HIV lifelong treatment. Given the high prevalence of toxicity in HIV therapy ¹, efforts must be pursued to improve not only antiretroviral efficacy but also the safety and tolerability of the treatments. Those elements represent, in addition to treatment simplification, an important element to promote adherence. Nowadays, treatment modifications occurring during the management of HIV infection are more often triggered by adverse effects than by efficacy issues ².

Patient management could benefit from approaches integrating pharmacogenetics and applied clinical pharmacokinetics to better understand and predict drug exposure, toxicity and clinical response. Most investigations have focused on the parent drug solely, considering it as the best pharmacokinetic marker of antiretroviral drug exposure, and in case of higher levels, of toxicity. However, drug metabolites resulting from complex genetic and environmental influences can also contribute to treatment outcomes. For most drugs, metabolism is a detoxification process, but it can constitute in certain cases the rate-limiting step in drug toxicity. For some drugs in other therapeutic areas, the accumulation of reactive metabolites due to reduced elimination have been implicated in the occurrence of tissue toxicity, carcinogenicity, teratogenicity, and/or immune-mediated injuries ³. Unfortunately, integration of this aspect with pharmacogenetics and pharmacokinetics has attracted little attention in the field of antiretroviral therapy. Distinct metabolite profiles could modulate toxicity, tolerability and outcome of antiretroviral therapy.

This review focuses on current knowledge of the metabolic pathways of Protease Inhibitors, Non Nucleoside Reverse Transcriptase Inhibitors and Integrase Inhibitors and the implication and mechanism of parent and metabolite toxicity. Current knowledge of the main characteristics of their pharmacokinetics along the pharmacogenetic-attributed variation in drug exposure and potentially toxicity will be outlined. Finally, the use of metabolites as markers of metabolic capacity will be addressed. The metabolic pathways of the reviewed antiretroviral drugs, their metabolites, as well as their genetic variants and the respective clinical consequences are summarized in table 1.

Non Nucleoside Reverse Transcriptase Inhibitors (NNRTIs)

NEVIRAPINE

Nevirapine (NVP) was the first NNRTI approved by the U.S. Food and Drug Administration. It is used either in combination with other antiretroviral agents for the treatment of HIV-1 infection or in a single prophylactic dose to prevent mother to child transmission. Due to its low cost, NVP is one of the most widely prescribed antiretroviral drugs in resource-limited countries. NVP use is limited by its FDA Black Box warning for idiosyncratic liver toxicity and skin hypersensitivity.

After oral administration, NVP is rapidly and readily absorbed with an absolute bioavailability of around 93%. Peak plasma concentrations are attained within approximately 4 hours following a single 200 mg dose. Extent of absorption does not appear to be altered by concomitant anti-acid administration or food intake. The molecule is highly lipophilic, crosses the placenta, is excreted into breast milk and is about 60% bound to plasma proteins ⁴. NVP induces hepatic cytochromes (CYP) CYP3A4 and CYP2B6 mediating its own degradation, therefore auto-induction increases NVP apparent oral clearance by approximately 1.5- to 2-fold, and decreases the terminal phase half-life, from 45 hours following a single dose intake to 25-30 hours under continuous treatment. Induction is complete within 28 days ^{5, 6} and steady state NVP plasma concentration is then reached ⁷.

NVP undergoes significant oxidative metabolism followed by glucuronidation to water soluble conjugates which are primarily cleared in the urine. Renal excretion of the unchanged parent drug is insignificant ⁶. 2 and 12-hydroxynevirapine (OH-NVP) formation is predominantly mediated by CYP3A4, while CYP2B6 is involved in the formation of 3- and 8-OH-NVP ⁸. Secondary oxidation of 12-OH-NVP yields carboxy-nevirapine ⁶.

NVP has been associated with a significant incidence of hepatotoxicity (3%) and cutaneous adverse reactions (9%) with occasional life threatening side effects such as Stevens-Johnson syndrome or toxic epidermal necrolysis transition syndrome (0.3%) ⁹. The onset of these reactions occurs usually during the first 6 weeks of treatment ¹⁰, therefore, close clinical and laboratory monitoring is strictly required to quickly identify life threatening adverse events. Rash and liver toxicity have been reported to be more frequently observed in female with a low body mass index ⁹.The causal element of NVP idiosyncratic toxicity is not determined and multiple mechanistic pathways have been proposed to explain this adverse reaction. It has been advocated that drug toxicity could be generated through reactive intermediates which produce cellular damage by alkylating human DNA, thus resulting in hypersensitivity reactions ^{11, 12}. Furthermore direct immune mediated mechanisms have also

been suggested since NVP is contraindicated in patients with high CD4+ count ⁴. Prediction of increased risk of NVP-associated hypersensitivity ¹³ and hepatotoxicity ¹⁴ has been proposed with the use of the human leucocyte antigen (HLA) class II allele HLADRB*0101.

HLA-B*3505 is another candidate that could possibly be implicated in NVP induced skin rash in Asian patients ¹⁵.

Several in vitro and in vivo approaches have been used to elucidate the involvement of NVP metabolites in the idiosyncratic drug reactions. CYP3A activity was used as a marker for NVP toxicity, demonstrating that rats pretreated with CYP3A inducers such as dexamethasone or NVP exhibited increased hepatotoxicity compared to naïve animals, thus implying the role of metabolites in liver injury ¹⁶. Skin rash and hepato-toxicity were observed in brown rats secondary to the formation of the antigen quinine methide; This metabolite is thought to be generated either via the sulfoconjuguaison of 12-OH-NVP directly in the skin or by cytochromes mediated formation in the liver ¹⁷. NVP incubation with NADPH-supplemented microsomes together with glutathione, a tripeptide found in all mammalian tissues and used as trapping agent for reactive intermediates ¹⁸, revealed the P450-mediated formation of a glutathione-mediated sulfhydryl conjugate that could initiate hepatotoxicity ¹⁹. Using a synthetic model, Antunes et al. ^20-22 successfully demonstrated that phase II esterification of 12-OH-NVP to electrophilic derivatives yielded covalent DNA adducts that could potentially play a role in the hepatocarcinogenicity of NVP in vivo. Recently, Srivastava et al ²³ provided the first evidence of NVP bioactivation in human patients by isolating and quantifying NVP mercaptopurates in urine, an approach that can be particularly useful for the assessment of human exposure to environmental and biogenic toxic compounds.

The current available knowledge is not conclusive regarding the association between higher NVP or NVP metabolites exposure and toxicity. A study in 49 case-control pairs did not find a relationship between NVP induced toxicities and higher blood levels of NVP or 12-OH-NVP in patients at week 4 of treatment ²⁴, even though higher blood levels of the 12-OH NVP were found in a sub-group at higher risk of skin rash, that is, women and patients co-treated with prednisone. The role of genetic polymorphisms in NVP-induced toxicity was evaluated in two studies. The first reported that a genetic polymorphism of CYP2B6 at position G516T is associated with greater NVP plasma exposure, and might therefore lead to toxicity ^25-27. The second study showed that variation in the MDR1 position 3435C>T allele was associated to a decreased risk of NVP hepatotoxicity ^{28, 29}. Recent observations suggest that CYP3A5 6986A>G and MDR1 3435C>T could be correlated to elevation in liver transaminases ³⁰.

16 EFAVIRENZ

Efavirenz (EFV) is among the preferred regimens for first line therapy in HIV treatment guidelines because it exhibits excellent virological and immunological response ^{31, 32}. Once daily dosage promotes adherence, which is a crucial element since first generation NNRTI have a low barrier to resistance that rapidly develops after virologic failure, and is mainly attributed to a single mutation in the reverse transcriptase ^32-34.

EFV is readily absorbed with a time to peak plasma concentration reached within 3 to 5 hours. Bioavailability is increased upon administration with high caloric/fat meal, where a small increase in area under the concentration curve (AUC) (17-22 %) and in maximal concentration (39%-51%) was observed compared to fasted conditions. EFV is approximately 99.5-99.75% bound to plasma proteins, predominantly to albumin. EFV is a dose-dependent inducer of CYP2B6, CYP3A4 and P-gp ^35-38 and is a CYP2B6 reversible inhibitor as well ³⁹. Due to auto-induction, terminal elimination half-life drops from 52-76 hours after single dose to 40-55 hours after multiple dosing with 200-400 mg daily for 10 days.

EFV is extensively metabolized primarily by hepatic CYP2B6 with partial involvement of CYP3A4 and CYP2A6 ⁴⁰. CYP2B6 mediated conversion to 8-OH-EFV accounts for 77 to 92% of overall EFV clearance, whereas only 7 to 23% are eliminated as 7-OH-EFV essentially via CYP2A6. The metabolite 8,14-dihydroxyefavirenz is also detected in vitro and in vivo and is formed either by direct 14-hydroxylation of EFV or the sulfate of 8-OH-EFV ^{40, 41}. Hydroxylated EFV metabolites undergo further conjugation with glucuronic acid forming N-glucuronide-EFV. EFV is excreted in urine (14 – 34%) mainly as metabolites and in feces (16 – 61%) mainly as parent compound. Recently, EFV metabolite profiling was used as a novel tool to enable the quantification of the metabolic capacity of the primary and secondary metabolic pathways and consequently their importance in drug detoxification ⁴⁰. It was used as well to identify the behavior of accessory metabolic pathways in case of decrease or loss of function of the major metabolizing enzyme ⁴².

EFV is characterized by its frequent induction (25 -70 %) of central nervous system (CNS) side effects. Early CNS symptoms include dizziness, depression, anxiety, irritability, headache and sleep disturbance ^{31, 32, 43} which occur at therapy initiation and usually resolve spontaneously after 2 to 4 weeks of treatment. Symptoms may however persist longer and lead to at least 4 to 10% of treatment discontinuation in some patients ^44-46. It has been suggested that neuropsychological symptoms are correlated to plasma EFV concentrations

47, with plasma levels higher than 4 mg/L being associated with higher frequency of CNS side effects ^{48, 49}; a dosage reduction in such patient was shown to prevent discontinuation due to

toxicity ⁵⁰. However, CNS toxicity is not uniformly found at higher EFV levels ⁵¹ and might therefore be attributed to variations in EFV metabolite concentrations or in CNS penetration.

The relationship between metabolites and toxicity in vivo has not been studied. EFV metabolite-induced nephrotoxicity was solely detected in rats and involved glutathione adducts, proven responsible for renal tubular cell necrosis ^{52, 53, 54}.

Since EFV major metabolizing enzyme CYP2B6 and minor pathways CYP2A6 and CYP3A4/3A5 are highly polymorphic, the influence of CYP2B6 genetic variants on EFV exposure and consequently on EFV induced CNS side effects was extensively studied.

CYP2B6 allelic variation (CYP2B6*6, *11,*15, *18, *27, *28, *29), associated with loss or decreased function or expression accounted for most of the inter-individual differences in EFV clearance and yielded higher EFV exposure⁵⁵. An association has been reported between CYP2B6 polymorphism and early occurrence and higher frequency of CNS side effects with a resulting treatment discontinuation and increased emergence of virologic failure ⁵⁶. ^{27, 57-59}. Some evidence is in favor of the importance of CYP2A6 and CYP3A4/A5 accessory pathways in EFV disposition, independently from CYP2B6, or in case of loss of function of this pathway 27, 40, 42, 60. CYP3A4*1B, CYP3A4_rs4646437 and CYP3A5*3 as well as CYP2A6 allelic variants (*2, *4, *1H, *1J, *5, *7, *9, *10, *12, *13, *15, *17, *19, *34) explained a small but relevant part of the inter-patient variability in drug exposure and might also affect the metabolic profile of this drug ^{27, 60, 61}. A genetic association study showed that individual with CYP2B6, CY2A6 and CYP3A4 loss of function genotypes experienced higher discontinuation rates due to toxic events that individuals with only CY2B6 polymorphism.

These results suggest that EFV genotyping might represent a useful tool in predicting toxicity and consequently in treatment individualization Others did not find an effect of CYP3A polymorphism on EFV exposure without assessing the correlation between plasma EFV and CNS side effects 27, 60, 62-64. The role of MDR1 in EFV kinetic and toxicity profiles appears to be controversial as well 27, 59, 60, 62, 64, 65. Whether variations in EFV metabolite levels induced by genetic polymorphisms are related to toxicity remains to be established.

ETRAVIRINE

Etravirine (ETV) is the first member of the second generation NNRTI; it has a potent activity against wild type and mutant HIV-1 strains and is characterized by an increased resistance barrier. It is prescribed as a second line therapy in treatment experienced patient with HIV-1 mutations and resistance to either EFV or NVP, since ETV retains antiviral activity in presence of several common NNRTI resistance associated mutation ⁶⁶. ETV is neither recommended for treatment naïve patients nor for pediatric patients ^67-70.

ETV absolute bioavailability is not yet known. Maximal concentrations are reached within 2.5 to 4 hours. ETV absorption is not affected by agents increasing gastric pH; on the other hand, an increase in AUC by approximately 50 % was reported following the administration of a meal. ETV is highly bound to plasma proteins (99.9%), principally to albumin and to alpha-1-acid glycoprotein (AAG) ⁶⁹ .

ETV undergoes hepatic metabolism by CYP3A4, CYP2C9 and CYP2C19 and is a CYP3A4 inducer and an inhibitor of CYP2C9 and CYP2C19 and P-gp ⁷¹. Methylhydroxylation of the dimethylbenzonitrile to form mono or dihydroxy–ETV accounts for the majority of ETV metabolism whereas hydroxylation of the dimethylbenzonitrile without the methyl groups plays only a minor role. Glucuronide conjugates of these metabolites are also detected.

Elimination occurs in a large portion in feces (85%) and bile (11%) while only 1% is recovered in urine. ETV has a long elimination half-life of 30 to 40 hours. Although once and twice daily dosing results in comparable exposure ⁷², evidence for once daily dosing is considered not yet sufficient ^{69, 73}. ETV metabolites have minimal anti-HIV activity compared to the parent drug ⁷² and are mainly excreted in urine ⁷³.

The most commonly reported side effect is a mild to moderate rash, which occurs in 10% of patients at the beginning of the treatment and resolves spontaneously within 1 to 2 weeks on continued ETV. Post-marketing experience revealed rare cases of severe Stevens-Johnson including hepatic failure, life threatening conditions and even fatalities ^{70, 73, 74}. So far, no evidence of causal relationship between the occurrence of ETV side effects and plasma concentrations of parent drug, toxic metabolites has been identified to date.

Pharmacogenetics of ETV metabolizing enzymes has been recently analyzed ⁷⁵ suggesting that carriers of CYP2C9*3 AND CYP2C19*2 had 44% and 20% lower ETV clearance respectively. The clinical influence of these genetic polymorphisms was not clear as it was confounded by the opposite effect of the co-administered medications on ETV disposition.

Protease Inhibitors (PIs)

LOPINAVIR

Lopinavir (LPV) is used in adults and children ^{76, 77} and remains a commonly prescribed PI in pregnancy ⁶⁸. It is co-formulated with low dose RTV as a booster, thus increasing its plasma AUC 77- to 100-fold by inhibiting P-gp and first pass inactivation ^78-80.

After multiple dosing, time to peak concentration occurred after approximately 4 hours.

Administration of LPV/RTV with a moderate to high fat meal improved bioavailability, with a

more relevant effect for the liquid formulation. At steady state, LPV is approximately 98-99%

bound to both AAG and albumin with a higher affinity for AAG ⁸¹. At a dosing of 400 mg + RTV 100mg twice daily, the apparent elimination half life ranges from 4 to 6 hours ^{81, 82}. LPV is both a substrate and inhibitor of CYP3A family.

LPV is metabolized by CYP3A4 and CYP3A5 to 3 major metabolites, 4-OXO-LPV and 2 epimeric 4-OH-LPV with other minor metabolites. Even though 4-OXO-LPV and the 2 OH-epimers are the main metabolites, their total measured radioactivity in plasma is relatively small and they exhibit only minor HIV protease inhibition capacity ^{83, 84}. LPV derivatives are primarily excreted in the feces with only a minor role of renal excretion (10.4%). Nearly 20% of the dose is excreted unchanged in the feces, while renal elimination of unchanged compound accounts for less than 3% after multiple dosing.

Moderate to severe diarrhea, nausea and vomiting as well as severe (grade 3 to 4) laboratory abnormalities, such as elevated cholesterol, triglycerides and alanine aminotransferase (ALT) and aspartate aminotransferase (AST) are reported in >2% of patients treated with LPV/RTV 81, 82, 85, 86. LPV exposure has been suggested to be directly correlated to drug tolerability and therapeutic efficacy: a target trough concentration of 1 mg/L⁶⁸ is required to achieve viral suppression and a threshold of 4 mg/L is required when five or fewer LPV baseline resistance mutations are present ⁸⁷. A higher incidence of nausea and diarrhea were observed in patients taking once daily 800/200 mg LPV/RTV compared to twice daily 400/100 mg after 3 months of treatment ⁸¹. Further results suggest that high RTV boosting leads to higher LPV exposure and is associated with more frequent nausea and vomiting ⁸⁸.

The impact of genetic polymorphisms on LPP/RTV elimination was recently investigated ^{57, 89}. Genetic variations in solute carrier SLCO1B1*4 induced a marked increase in LPV clearance compared to carriers of the reference allele, thus reducing drug exposure. Another variant of this transporter SLCO1B1*5 that is characterized by impaired drug transport activity, induced a small decrease in LPV clearance. Genetic variation in the ATP-Binding Cassette subfamily C member 2 (ABCC2) genes encoding the Multidrug resistance-related protein MRP2 was associated with impaired LPV elimination as well related to a decreased biliary efflux of the drug. It is worth noting that despite the inhibitory action of the associated low dose RTV, the effect of CYP3A allelic variants on LPV clearance remained significantly relevant. Genetic variations in the efflux pump P-gp encoded by the Multidrug resistance 1 gene (MDR1) MDR1 3435C>T and 2677G>T polymorphism did not influence LPV plasma levels ⁹⁰. Interestingly, ultra rapid metabolizers of CYP2D6, a cytochrome partially involved in RTV disposition, were found to have lower RTV and LPV concentrations⁹¹. Although variations in

drug exposure consequent to genetic polymorphisms were identified for this drug, the relationship between LPV, metabolites and toxicity profile has not been investigated so far.

ATAZANAVIR

Excellent bioavailability, favorable lipid profile and gastrointestinal tolerability in addition to low pill burden, are the advantages that ensured for atazanavir (ATV) a relevant role in HIV pharmacotherapy. It is the preferred PI for the treatment of HIV+ naïve patient ⁹². It is administered in both adults and children, most frequently with low dose RTV as a booster.

Dans le document Population pharmacokinetics and pharmacogenetics of antiretroviral drugs (Page 24-56)