The expanding role of gene-based prescribing for phase II drug-metabolizing enzymes

Background

The human NAT gene locus comprises two genes, NAT1 and NAT2, that encode 2 enzymes (NAT1 and NAT2), responsible for the transfer of acetyl groups (acetylation) to aromatic amines and hydrazines, as well as to heterocyclic amine groups present in many drugs and carcinogens.^[4] These enzymes are responsible for the acetylation of a variety of drugs, but the highly polymorphic NAT2 is known for its role in the N-acetylation of several antituberculosis agents including isoniazid and pyrazinamide. It is also responsible for acetylating sulfonamide-containing antibiotics (sulfamethazine and sulfamethoxazole) and the antihypertensive agent, hydralazine. The correlation between NAT2 genotype and acetylation phenotype is a classic in the field of pharmacogenomics and has been extensively reviewed.^[5-9]

Results from early experiments with isoniazid identified a bimodal distribution in urinary acetylated metabolites and led to the assignment of “slow” and “rapid” acetylator phenotypes. These studies also showed that not all N-acetylated drugs exhibited this polymorphism, and for those that did, NAT2 was the enzyme responsible. Later studies suggested that the acetylator phenotype was more trimodal than bimodal incorporating the “intermediate” or “heterozygous rapid” acetylator phenotype.^[10,11]

Genetic variation and testing

The reference allele for the NAT2 gene is the NAT2*4 allele. Phenotypic expression (acetylation) of this allele is described as rapid, intermediate, and slow for individuals with two, one, or zero copies of the *4 allele. The various haplotypes identified for NAT2 are available online.^[12] The most common single-nucleotide polymorphisms (SNPs) and nucleotide changes associated with acetylator phenotypes are rs1801279 (191G > A), rs1041983 (282C > T), rs1801280 (341T > C), rs1799929 (481C > T), rs1799930 (590G > A), rs1208 (803A > G), and rs1799931 (857G > A). Varying combinations of these yield the characterized NAT2*5, *6, *7, and *14 (slow acetylator) and the NAT2*11, *12, and *13 (rapid acetylator) haplotypes. Each of these numbered alleles have several clusters of these SNPs assigned and are designated with an alphabet suffix for differentiation (e.g., the *5 haplogroup has the NAT2*5A to *5V haplotypes). The 7 SNPs above serve as the basis for NAT2 genotyping, although there is an extremely wide variation of allele frequencies between and within various ethnic groups.^[13-15]

The marked variation in allele frequencies^[14,16] has generated differing opinions on the usefulness of the seven SNP panels. In 2011, a tag SNP (rs1495741) was identified from genome-wide studies.^[17,18] For economies of scale in genotyping, especially for resource-constrained areas, efficient SNP selection was advocated based on computational strategies^[19] with 2-, 3-, and 4-SNP panels proposed and tested.^[20] The ability of the tag SNP to infer acetylator phenotype has been compared with the 2, 3, 4, and 7 SNP panels; the 4-SNP panel is suggested to maximize sensitivity and specificity, particularly in populations of non-European ancestry in whom inference of NAT2 phenotypes from SNP panels has been problematic.^[21-24]

Background

COMTs are responsible for the O-methylation of biogenic catecholamines such as the neurotransmitters dopamine, epinephrine, and norepinephrine, as well as catechol-containing drugs such as the antihypertensive drug, methyldopa, and antiparkinsonian drug, L-dopa. It also methylates other endogenous catechol compounds such as the catechol estrogens formed in vivo from estrone and 17β-estradiol.^[25] The pharmacogenetics of COMT has been reviewed by Weinshilboum et al.^[25] and the distribution of genotypes differs by ethnicity. With respect to genotype– phenotype correlations, functional polymorphisms leading to high and low activity forms of the soluble enzyme have been identified; these can be phenotyped by measuring erythrocyte COMT activity.^[26]

Genetic variation and testing

The first COMT polymorphism, described by Lachman et al.,^[26] was the functional polymorphism rs4680 322/472 G > A (Val108/158Met substitution in the soluble or membrane-bound forms of the enzyme, respectively). The major G (or Val) allele is expressed as the high COMT activity allele and the A (or Met) as the low activity allele. There are ethnic differences in prevalence; the frequency of the A allele ranges from 29% to 51% in Asian and European populations.^[26] While rs4680 is a key COMT variant, 3 other SNPs, rs6269 A > G, rs4633 C > T, and rs4818 C > G, found in a tightly linked haploblock have been combined with it to yield high, intermediate, and low activity COMT haplotypes.^[27] There is currently no defined star (*) allele nomenclature for COMT. In addition to the 4 SNPs mentioned, other studies have included rs737865 A > G, rs9332377 C > T, and rs165599 G > A SNPs.^[28,29] To date, rs4680 remains the only functional COMT non-synonymous coding SNP, although a single association study identified another functional non-synonymous SNP rs6267 G > T (Ala72Ser substitution in the membrane bound form).^[30] Other studies have reported functional haplotypes that include rs4680 and other synonymous coding SNPs (rs4633, rs4818).^[31,32] Many of these COMT SNPs are represented on large-scale genome and exome-wide genotyping chips and rs4680 remains the most commonly referenced COMT SNP in commercial genotyping panels.^[33]

Background

TPMT catalyzes the S-methylation of thiopurine drugs such as azathioprine (AZA), 6-mercaptopurine (6-MP), and thioguanine (TG), as well as other heterocyclic sulfur-containing compounds.^[34] TG and 6-MP are widely used for the treatment of lymphoid (6-MP) and myeloid malignancies,^[35,36] and AZA and 6-MP are used to treat non-malignant conditions such as inflammatory bowel disease, systemic lupus erythematosus, and rheumatoid disease. All three compounds are prodrugs that are activated to the same TG nucleotide (TGN) active metabolites; TPMT tempers this activation process by yielding inactive methyl 6-MP and methyl TG compounds.^[37]

Thiopurines have a class effect of cytotoxicity, with a side effect profile that includes severe myelosuppression. The clinical implication is that individuals with low TPMT activity have higher blood levels of active TGN metabolites, and thus greater myelosuppression. The risk of developing life-threatening myelosuppression has been shown to be greatest in individuals with very low red cell TPMT activity treated with “standard” doses of thiopurines.^[38] This relationship between TPMT activity and TGN levels was demonstrated in key studies reporting both trimodality^[39] and a correlation between TPMT activity and TGN levels.^[40,41] The pharmacogenetics of TPMT has been extensively studied and validated and a concise summary of the enzyme is available.^[42] This summary along with information regarding its genetic architecture and relevant clinical annotations is also found on the Pharmacogenomics Knowledge Base (PharmGKB) website.^[43,44]

Genetic variation and testing

TPMT genotypes are grouped into predicted phenotypes of normal or high activity, intermediate activity, and low or deficient activity based on the number of functional or variant alleles present. The reference allele for TPMT is the *1 or the high activity allele. Variant alleles representing low or deficient activity include *2 (rs1800462 C > G), *3A (rs1142345 T > C and rs1800460 C > T), *3B (rs1800460), *3C (rs1142345), and *4 (rs1800584 C > T). The *3A is a combination haplotype of the *3B and C alleles and is the most common TPMT variation in Caucasians (frequency of approximately 5%).^[45] The *3C occurs more frequently in East Asian and African American populations, exemplifying the finding of major ethnic differences in allele frequency. TPMT phenotype was initially obtained from erythrocyte TPMT activity;^[9,39] however, as with other DMEs, genotyping is now typically used to predict phenotype. There are over 40 variant alleles reported on PharmGKB and the TPMT allele nomenclature websites.^[44,46] There are commercial and in-house Clinical Laboratory Improvement Amendments approved assays for TPMT genotyping with extremely high call rates, concordance, and accuracy.

Background

The GSTs are a diverse group of enzymes responsible for conjugating reduced glutathione to electrophilic centers present on a variety of substrates, including endogenous and exogenous compounds. GSTs are implicated in drug metabolism as well as detoxification (and sometimes activation) of procarcinogenic agents.^[2] They catalyze glutathione-conjugated inactivation of the anticancer agents, etoposide and busulfan, and the platinum-derived compounds; cisplatin, oxaliplatin, and carboplatin, in addition to drugs such as isoniazid, rifampicin, and pyrazinamide.

GSTs belong to a superfamily of soluble enzymes that have been further subdivided into different classes – alpha, delta, kappa, mu, omega, pi, theta, zeta, and microsomal proteins. Unlike many of the genes encoding other DMEs, each GST member has a unique chromosomal location.^[47] Four of these, GST alpha (GSTA), GST theta (GSTT), GST mu (GSTM), and GST pi (GSTP), encoded by the GSTA1, GSTT1, GSTM1, and GSTP1 genes, respectively, are implicated in phase II metabolism and are the most studied with respect to their pharmacogenetics.

Genetic variation and testing

GSTA1 is a major hepatic GST enzyme,^[48] but has few genetic variations and those present are either of no known relevance or of minimal clinical importance.^[49] GSTM1, GSTP1, and GSTT1, on the other hand, are more polymorphic, with GSTM1 and GSTT1 possessing common insertion/deletion (InDel) variants.^[48]

In GSTM1, there is a common deletion expressed as a null allele or a non-null allele for the insertion variant, and a promoter SNP, rs3754446A > C, occurring with different ethnic frequencies (ranging from 1% to 6% in those of African ancestry to 67% in Asians). In GSTT1, there is also a deletion polymorphism characterized as the null or GST*0 or GST1 negative allele as well as rs1007888C > T and rs4630> A SNPs.^[50-53] GSTP1 has no known deletion polymorphisms. Rather, there are two common non-synonymous SNPs, rs1695G > A (Ile105Val) and rs1138272C > T (Ala114Val), that appear to lower enzyme activity.^[54,55]

All common GSTT1, GSTM1, and GSTP1 polymorphisms are represented on many commercial genotyping panels and are part of core ADME gene lists.

Background

Human SULTs are responsible for the sulfate conjugation of many important endogenous compounds and drugs. Substrate specificity differs for each SULT enzyme family of which the best characterized are SULT1A, 1B, 1C, 2A, 2B, and 4A families.^[54] The SULT1A family is composed of SULT1A1, 1A2, and 1A3 enzymes that catalyze the sulfate conjugation of phenolic compounds, including simple phenols like nitrophenol, as well as catecholamines. Well-known substrates include drugs such as acetaminophen, dopamine, and minoxidil. Of all the SULT enzymes, SULT1E1 has the highest substrate affinity and is responsible for the sulfation of both natural and synthetic estrogens, while other steroid hormones are sulfated mostly by the SULT1B and 2B families.^[56-58] Genotype–phenotype correlation for SULT activity was initially conducted using biochemical assays for SULT1A1 and subsequently SULT1A2 enzymes.^[59,60] These studies showed over 50-fold variation in phenotypic (biochemical) activity in human platelet SULT activity that correlated with low- and high-activity genotypes.

Genetic variation and testing

The SULT isoforms currently implicated in human phase II drug metabolism are SULT1A1, 1A2, 1A3, and 1E1, all encoded by highly polymorphic genes. SULT1A1 is the most abundant and most important. This isoform has four haplotypes assigned – *1 the reference allele, *2 (rs9282861 G > A), *3 (rs1801030 G > A), and *5 (rs28374453 T > C). In addition, three SNPs in the coding region of the gene, SULT1A1, have copy number polymorphisms (X2 to X5, reflecting 2–5 copies of the gene) and two promoter SNPs rs3760091 G > C and rs750155 G > A that are of functional importance.^[61]

SULT1A2 has three haplotypes assigned, *1 the reference allele, *2 (with two variants rs1136703 A > T, C, G and rs1059491 T > G), and *3 (rs10797300 C > G). All the above-listed SULT1A SNPs are non-synonymous coding polymorphisms that cause amino acid changes.

SULT1E1, encoding the only member in the 1E family, has four haplotypes assigned with the reference allele designated as *1 and 3 other polymorphic alleles differing in single nucleotide variations from the reference designated as *2 (rs11569705 C > A), *3 (rs34547148 G > A), and *4 (rs11569712 G > T). As with the SULT1A genes, these SNPs are non-synonymous coding polymorphisms. The SULT1E1 variants occur at very low frequencies, ranging from 0.5 to 0.9% for the *2 and *4 alleles, and there is no well-defined minor allele frequency for the *3 variant. With respect to genotyping tests, only SULT1A1 polymorphisms are represented (with varying degrees of SNP and copy number coverage) on some commercial genotyping platforms.

Background

The superfamily of enzymes responsible for the conjugation of a glucuronic acid moiety onto suitable acceptors on a wide range of endogenous and exogenous substrate compounds are known as UDP-glucuronosyltransferases or uridine diphosphoglucuronosyl transferases (UGTs). The products of this conjugation are called glucuronides. Glucuronidation is an efficient metabolizing process contributing to over 30% of all phase II drug metabolism.^[62] In addition to playing a role in the metabolic elimination of oral drugs, glucuronidation also serves a protective function through detoxification and elimination of carcinogens. Human UGTs are found in two main families – UGT1 and UGT2.

UGT1

The UGT1 gene family is composed of 12 UGT1A genes on a large and complex locus. The complexity occurs because each UGT1A gene has a unique exon 1 spliced to four common conserved exons (2–5).^[63] Each gene is named UGT1A1 to UGT1A12 based on the proximity of the unique first exon to the shared ones. Although 12 distinct gene products exist in humans, four of these, UGT1A2, UGT1A11, UGT1A12, and UGT1A13, are pseudogenes. Genetic variations spanning the range of polymorphism mechanisms have been reported for all UGT1A active gene products and many have been characterized and studied in detail.^[64-66] The pharmacogenetics of UGTs and their role in health, disease, and cancer have been extensively reviewed.^[67-90]

UGT2

Enzymes in the UGT2 family fall into either 2A or 2B subfamilies. UGT2A members are mostly olfactory enzymes implicated in the detoxification of airborne xenobiotics and toxins. UGT2B enzymes, on the other hand, are responsible for the glucuronidation of both drugs and endogenous steroid hormones. Of the human UGT2 enzymes, UGT2B7, 2B15, and 2B17 are the most studied, although UGT2B7 seems to be the isoform with the greatest contribution to the glucuronidation of clinically useful entities.^[72,91-94]

NATs

This class of enzymes is involved in the metabolism of many drugs and possesses genetic variants mostly implicated in clinical studies of acetylator phenotypes and isoniazid effects and toxicity. Other NAT pharmacogenetic areas of interest include adverse events such as skin reactions with sulfonamides and drug-induced lupus with hydralazine [Table 1].

Close

Figure 1:

Simplified flowchart of criteria involved in determining clinical utility of any identified drug-metabolizing enzyme variants with the potential for altering drug response in humans.

Export to PPT

Isoniazid remains a first-line drug for the treatment of latent and active tuberculosis, but problems include hepatic injury and treatment failure. Both problems have been associated with genetic differences in acetylation capacity, but the evidence is mixed with some studies reporting conclusive associations and others little or no association. Various meta-analyses of hepatic injury associated with antituberculosis drugs attributed an odds ratio between 1.59 and 4.7 to the slow acetylator genotype and risk of developing hepatic injury.^[102-106] This association was more robust in individuals of Asian, Middle Eastern and Brazilian ancestry, with no significant associations for Caucasians, and very limited data in the African population.

As concerning isoniazid-associated hepatic injury, treatment failure for tuberculosis could have broader clinical and economic consequences since second-line drug regimens are usually more expensive and toxic. It has been hypothesized that rapid acetylators, due to extremely rapid inactivation of the parent isoniazid, do not attain adequate blood levels for optimal therapeutic response. The concept of using pharmacogenetics-guided isoniazid dosing to treat pulmonary tuberculosis using the WHO-recommended 6-month four-drug regimen was tested in a prospective randomized clinical trial.^[107] Based on the results of PK dose-ranging studies, the isoniazid genotype-guided dose for slow and rapid acetylators was 0.5 and 1.5 times the standard dose used for intermediate acetylators, respectively. Despite the small sample size for slow (n = 7) and rapid (n = 44) acetylators in the pharmacogenetics guided arm and n = 9 and 48 in the standard treatment arm, the results were interesting. Hepatic injury occurred in 7 of 9 slow acetylators (78%) receiving standard dosing and in none of the seven who received pharmacogenetic-guided doses. In rapid acetylators, there was a significantly lower incidence of treatment failure in the test arm (15%) compared to standard treatment (38%). This study along with others modeled around it^[108] illustrates the potential for NAT2 genotype guided dosing, at least in the context of isoniazid.

Recommendation

Although there is no clear consensus regarding the clinical utility of NAT2 genotyping, the potential of NAT2 genotype-guided regimens for isoniazid-based therapies bears consideration for use in treatment. This genotype-guided consideration may be most relevant for populations with a high TB burden and a prevalence of slow acetylators. The WHO over a decade ago recommended an increase in standard dosing for isoniazid from 5 to 10 mg/kg body weight for pediatric patients (Rapid advice: Treatment of Tuberculosis in Children, WHO, 2010). This higher dosing most likely favors increased efficacy in rapid acetylators while posing a greater risk of hepatotoxicity for slow acetylators. As a result, the potential utility of genotyping remains important, and more clinical studies clearly defining the relationship between acetylator status and isoniazid efficacy or toxicity will be helpful.

COMT

For COMT, the most relevant associations include those for exogenous levodopa used in the treatment of Parkinson’s disease, and entacapone, the direct COMT inhibitor, used as an adjunct to boost effects of levodopa. There are studies on COMT genotypes and response to levodopa with some finding no association, and others showing the association between increased levodopa doses and greater functional activity COMT haplotypes.^[26,109,110] A recent meta-analysis confirmed an association between the well-studied rs4680 and levodopa-induced dyskinesia.^[111] The association between entacapone effects and COMT genotypes has also been inconsistent.^[112,113]

Recommendations

Although a recent review on Parkinson’s disease called for personalized treatments relying on pharmacogenetic procedures to optimize therapeutics in its management,^[114] there is not enough evidence for using COMT pharmacogenomics profiles for gene-based prescribing.

Recommendations

Despite the numerous studies cited on pharmacogenomic studies of GSTs and pharmacotherapy with cytotoxic agents such as busulfan and platinum-containing agents, there is inadequate evidence to suggest that genotyping for GST polymorphisms would alter current standards of therapy.

Recommendations

At present, there is inadequate evidence to support the clinical utility of SULT genotyping.

Recommendations

The TA repeat polymorphism (UGT1A1*28 in the Caucasian population) was one of the earliest pharmacogenomic biomarkers approved by the United States FDA for reference on a drug label.^[151] The agency-approved label for one irinotecan product has the following stated recommendation – “when administered in combination with other agents or as a single agent, a reduction in the starting dose by at least one level should be considered for patients known to be homozygous for the UGT1A1*28 allele” with no precise dose reduction provided. Another product, however, has a dose specification with a caveat to consider dose modifications based on individual tolerance to treatment (50 mg/m² administered by intravenous infusion over 90 min, with an increased dose to 70 mg/m² as tolerated in subsequent cycles).^[152]