CYP3A

Enzymes of the CYP3A subfamily are the most abundantly expressed isoforms in the liver, comprising 25–40% of total hepatic P450 content⁶ and thought to account for most of the monooxygenase activity present in the gastrointestinal tract.⁷ Expression of CYP3A in the enterocyte of the small intestine and the liver accounts for the marked first-pass effect often observed following oral administration of its substrates. Because of its broad substrate specificity, CYP3A is involved in the metabolism of over 50% of currently used therapeutic agents, including most PIs and NNRTIs.⁸ In addition to being substrates for CYP3A, most PIs are inhibitors of CYP3A4 whereas NNRTIs induce CYP3A4.

CYP3A5-mediated metabolism has wide interindividual variability reflective of both environmental factors such as drug interactions involving inhibition and induction^9,10 and genetic components.¹¹ Not surprisingly, a number of laboratories have looked for single nucleotide polymorphisms (SNPs) in the coding and noncoding regions of both CYP3A4 and CYP3A5. Most of these appear to be uncommon (frequency <1–2%) and/or distributed to specific populations.¹² A 2392A>G transition in the 5′-regulatory region of CYP3A4 (i.e., CYP3A4*1B) occurs more frequently, particularly among those of African descent.¹³ With CYP3A5, an A>G substitution within intron 3 (i.e., CYP3A5*3) a variant forming a truncated inactive protein is noted to be present in most subjects.¹⁴ A rarer G>A transition in exon 7 (i.e., CYP3A5*6) also results in a splicing defect. It has been suggested that these variants may contribute to the observed interindividual variability in CYP3A activity. Several studies have been performed in order to evaluate the effect of genetic variations in CYP3A4 on the pharmacokinetics of nelfinavir,^15–17 efavirenz,^15–18 and saquinavir^19,20 (Table 29-1). These studies indicate that CYP3A4*1B, CYP3A5*3, and CYP3A5*6 have no influence on nelfinavir and efavirenz pharmacokinetics. CYP3A5*3 polymorphism has been associated with the urinary metabolic ratio of saquinavir to its hydroxy metabolites in healthy individuals. The mean metabolic ratio was significantly lower and the clearance was higher in subjects with active CYP3A5 compared to individuals with a functionally inactive enzyme.^19,20

Table 29-1 Inherited Differences in the Metabolism of Antiretroviral drugs

† See http://www.imm.ki.se/CYP alleles/for detail description of CYP alleles,¹³⁵ http://www.hiv-pharmacogenomics.org for updated information for genetic influences on antiretroviral drugs.¹³³

Summarized from references 15–20, 28–31, and 127

CYP2C19

CYP2C19 is important in the metabolism of the PI nelfinavir. Several loss of function genetic variants exist, two variants (CYP2C19*2 and CYP2C19*3) accounting for more than 95% of poor metabolizer phenotypes. Genotypic differences are observed among populations since 2–3% of Caucasians and 4% of Africans have the poor metabolizer phenotype versus 10–25% of Asians.²¹ A recent study demonstrated that nelfinavir exposure was significantly higher in treatment-naive HIV patients with GA or AA genotypes at position 681 in CYP2C19 (i.e., CYP2C19*2) compared to the wild type.¹⁷ This same study showed a trend toward decreased virologic failure in subjects with the GA genotype (Table 29-1).

CYP2D6

The CYP2D6 genetic polymorphism was originally discovered as a result of differences in the pharmacokinetics and therapeutic effects of drugs metabolized by this enzyme such as codeine, dextromethorphan, metoprolol, nortriptyline, and debrisoquin.²² Approximately 5–10% of Caucasian subjects were found to have a deficiency in their ability to oxidize the antihypertensive drug debrisoquin.²³ Subjects with poor metabolism tend to be homozygous for mutations in CYP2D6 associated with marked loss or complete absence of enzymatic activity. However, it should also be noted that some subjects have multiple copies of CYP2D6 resulting in an extensive metabolizer phenotype.²⁴ A study published by Fellay et al¹⁵ analyzed the association between CYP2D6*3, *4, and *6, representing the most frequent inactivating mutations in Caucasians leading to a poor metabolizer phenotype, and plasma levels of nelfinavir and efavirenz in treatment-naive HIV-infected patients. Patients homozygous or heterozygous for a loss of function CYP2D6 allele had a higher median plasma concentration of both nelfinavir and efavirenz when compared to patients with a CYP2D6 extensive meta-bolizer genotype (Table 29-1).

CYP2B6

CYP2B6 was thought to be expressed at low levels and only in a fraction of human liver samples. However, CYP2B6 has gained more attention since it was shown to be significantly involved in the metabolic activation and inactivation of a number of clinically important drugs such as efavirenz.²⁵ A large interindividual variability in the hepatic content of CYP2B6 has been reported²⁶ as well as functional differences between genetic variants.²⁷ A number of studies have investigated for a possible role of genetic differences in CYP2B6 as the basis for toxicity to the NNRTI efavirenz. The relationship between efavirenz pharmacokinetics, central nervous system (CNS) side effects, weight, race, and virologic response was investigated in 154 HIV-infected individuals randomized to receive efavirenz-containing regimen.¹⁸ The authors found that a 516G>T polymorphism in exon 4 of CYP2B6 (i.e., CYP2B6*6) was more common in African-Americans compared with Caucasians. Furthermore, efavirenz AUC was three times higher in patients with the 516 TT genotype compared to GG homozygotes or GT heterozygotes. CYP2B6 516 TT genotype was also significantly associated with increased CNS side effects during the first week of efavirenz therapy. No association was observed between the allelic variants and virologic response. A number of studies have confirmed a role for the 516G>T polymorphism for efavirenz^28–31 and nevirapine (Table 29-1).²⁸

Induction of Drug Metabolizing Enzymes

In addition to genetic variations, expressed levels of CYP enzymes and drug transporters can be modulated by activation of certain nuclear receptors such as the Pregnane X Receptor (PXR).³² In organs such as the liver, these adopted orphan receptors mediate transcriptional activation of target genes by their ability to translocate into the nucleus once activated by ligands, which include a number of HIV drugs. In the nucleus, activated receptors then heterodimerize with the 9-cis retinoic acid receptor (RXR) and bind to response elements of target genes which then initiate target gene transcription.³³ PXR has been shown to be a key factor responsible for CYP3A4, 2B6, and 2C9 transcriptional activation^34–37 as well as for the efflux transporter MDR1.³⁸ Of importance, it is now known that most PIs and certain NNRTIs act as PXR ligands leading to the induction of a number of phase I and II metabolic pathways. Therefore, it is not surprising that the combination of several PIs with or without NNRTIs often result in complex drug interactions. Unlike CYP enzymes, genetic variations in nuclear receptors such as PXR are rare.

MDR1 (ABCB1)

Due to its broad substrate specificity, P-glycoprotein (P-gp), the gene product of MDR1, has received the greatest attention in terms of identification and characterization of SNPs. P-gp functions as a transmembrane efflux pump thereby moving drugs from the intracellular to extracellular domain; it can also act on drug molecules entrapped within the cell membrane lipid bilayer. ATP hydrolysis provides the energy for active drug transport, enabling the transporter to function against steep concentration gradients.

P-gp was first isolated from colchicine-resistant Chinese hamster ovary cells.⁴⁰ Subsequently, the gene coding for P-gp (MDR1) was identified because of its over-expression in tumor cells associated with an acquired cross-resistance to multiple cytotoxic anticancer agents.⁴¹ This drug transporter was also recognized to be expressed in many normal tissues.⁴² P-gp is found on the canalicular domain of hepatocytes, the apical surface of proximal tubular cells in the kidney, the brush border surface of enterocytes, the epithelium of the brain choroid plexus as well as the luminal surface of blood capillaries in the brain, the placenta, the ovaries, the testes and, of relevance to HIV therapy, in CD56+, CD8+, and CD4+ lymphocytes.^42–44 The function and tissue expression pattern for P-gp suggest that this transporter acts as a protective barrier to keep toxins out through excretion of substrate compounds into bile, urine, and the intestinal lumen and thereby limit the accumulation of such toxins or xenobiotics in organs such as the brain, gonads, and bone marrow, or in the fetus. P-gp is co-localized with the drug metabolizing enzyme CYP3A4 in the small intestine and liver suggesting that the extent of interplay between P-gp and CYP3A4 likely account for the net absorption and elimination of shared substrate drugs. Expression of P-gp at the level of the blood–brain barrier has shown to be critical to limiting the CNS entry of many drugs. Supportive evidence was obtained from animal models. A mouse strain naturally deficient in mdr1a, demonstrated marked sensitivity to the neurotoxic effects of ivermectin, an antiparasiticide substrate of P-gp.^45,46 The absence of P-gp in the blood–brain barrier of such mice resulted in a more than 80-fold higher brain accumulation of ivermectin.⁴⁵ Additionally, P-gp can alter oral drug absorption⁴⁷ or enhance their biliary and renal excretion.⁴⁸ P-gp has been shown to be particularly relevant to HIV therapy as PIs have been shown to be substrates of this transporter.⁴⁹

Screening of the entire MDR1 coding region identified a synonymous SNP in exon 26 (i.e., 3435C>T) associated with altered protein expression although the SNP does not change the encoded amino acid (Ile).⁵⁰ In the same study P-gp expression in duodenal biopsy samples among healthy Caucasians with the homozygous T allele (variant) was noted to be decreased when compared to those with the C allele (common). Subjects with the variant allele were also shown to have increased digoxin plasma concentrations after oral administration, suggesting greater drug absorption in individuals with low intestinal P-gp levels. The first study exploring associations between SNPs in MDR1 and antiretroviral pharmacokinetic parameters was performed by Fellay et al.¹⁵ The authors examined the influence 3435C>T on the median concentrations of nelfinavir and efavirenz in treatment-naive HIV-infected patients. Patients with the variant allele were shown to have lower nelfinavir and efavirenz levels compared to the wild type. Subsequent studies attempting to define associations between this same polymorphism in exon 26 and another in exon 21 (i.e., 2677G>T/A) and the pharmacokinetics of several PIs and efavirenz have resulted in conflicting and controversial findings (Table 29-2).^{16–18,29,51–53}

Table 29-2 Inherited Differences in the Transport of Antiretroviral Drugs

† See http://www.hiv-pharmacogenomics.org for updated information for genetic influences on antiretroviral drugs.¹³³

Summarized from references 15–17, 51–54, 56, 57, 128, 129.

In addition to drug levels, SNPs in MDR1 may also alter the physiological protective role of P-gp and therefore influence disease.⁴⁹ Fellay et al found a relationship between expression of P-gp in peripheral blood mononuclear cells (PBMCs) of HIV-infected patients and CD4 lymphocyte response to treatment.¹⁵ Patients with the 3435T allele in exon 26 had a significantly greater rise in CD4 cells count, 6 months after starting ART. It was hypothesized that this benefit associated with the T allele could result from an enhanced HIV PI penetration into CD4 cells. Three additional studies have reported a better virologic outcome associated with the 3435T allele.^16,17,54 However, other studies reported no virologic or immunologic effects associated with the study allele (Table 29-2).^55–57 While the exact mechanism by which SNPs in MDR1 affect the disposition or efficacy of HIV drugs needs to be clarified, there is a growing acceptance of the fact that in many cases SNPs in MDR1 do predict response to certain HIV drugs. It should be remembered however, while P-gp is the best characterized ABC transporter, other transporters also have the potential to affect the disposition of HIV drugs (Table 29-2).