Clinical Relevance of Drug Metabolism

CLINICAL RELEVANCE OF DRUG METABOLISM

The dose and frequency of administration required to achieve effective therapeutic blood and tissue levels vary in different patients because of individual differences in drug distribution and rates of drug metabolism and elimination. These differences are determined by genetic factors and nongenetic variables, such as age, sex, liver size, liver function, circadian rhythm, body tem-perature, and nutritional and environmental factors such as con-comitant exposure to inducers or inhibitors of drug metabolism. The discussion that follows summarizes the most important of these variables.

Individual Differences

Individual differences in metabolic rate depend on the nature of the drug itself. Thus, within the same population, steady-state plasma levels may reflect a 30-fold variation in the metabolism of one drug and only a two-fold variation in the metabolism of another.

Genetic Factors

Genetic factors that influence enzyme levels account for some of these differences, giving rise to “genetic polymorphisms” in drug metabolism. The first examples of drugs found to be subject to genetic polymorphisms were the muscle relaxant succinylcholine, the anti-tuberculosis drug isoniazid, and the anticoagulant warfarin. A true genetic polymorphism is defined as the occurrence of a variant allele of a gene at a population frequency of ≥ 1%, result-ing in altered expression or functional activity of the gene product, or both. Well-defined and clinically relevant genetic polymor-phisms in both phase I and phase II drug-metabolizing enzymes exist that result in altered efficacy of drug therapy or adverse drug reactions (ADRs). The latter frequently necessitate dose adjust-ment (Table 4–4), a consideration particularly crucial for drugs with low therapeutic indices.

A. Phase I Enzyme Polymorphisms

Genetically determined defects in the phase I oxidative metabo-lism of several drugs have been reported (Table 4–4). These defects are often transmitted as autosomal recessive traits and may be expressed at any one of the multiple metabolic transformations that a chemical might undergo. Human liver P450s 3A4, 2C9, 2D6, 2C19, 1A2, and 2B6 are responsible for about 75% of all clinically relevant phase I drug metabolism (Figure 4–4), and thus for about 60% of all physiologic drug biotransformation and elimination. Thus, genetic polymorphisms of these enzymes, by significantly influencing phase I drug metabolism, can alter their pharmacokinetics and the magnitude or the duration of drug response and associated events.

Three P450 genetic polymorphisms have been particularly well characterized, affording some insight into possible underlying molecular mechanisms, and are clinically noteworthy, as they require therapeutic dosage adjustment. The first is the debrisoquin-sparteine oxidation type of polymorphism, which apparentlyoccurs in 3–10% of Caucasians and is inherited as an autosomal recessive trait. In affected individuals, the CYP2D6-dependent oxidations of debrisoquin and other drugs (Table 4–2; Figure 4–6) are impaired. These defects in oxidative drug metabolism areprobably co-inherited. The precise molecular basis for the defect appears to be faulty expression of the P450 protein due to either defective mRNA splicing or protein folding, resulting in little or no isoform-catalyzed drug metabolism and thereby conferring a poor metabolizer (PM) phenotype. This PM phenotype corre-lates with a higher risk of relapse in patients with breast cancer treated with tamoxifen, an anti-cancer drug that relies on its CYP2D6-dependent metabolic activation to endoxifen for its efficacy. More recently, however, another polymorphic genotype has been reported that results in ultrarapid metabolism of rele-vant drugs due to the presence of CYP2D6 allelic variants with up to 13 gene copies in tandem. 

This ultrarapid metabolizer (UM) genotype is most common in Ethiopians and Saudi Arabians, populations that display it in up to one third of individuals. As a result, these subjects require two-fold to three-fold higher daily doses of nortriptyline (an antidepressant and a CYP2D6 substrate) to achieve therapeutic plasma levels. The poor responsiveness to antidepressant therapy of the UM phenotype also clinically cor-relates with a higher incidence of suicides relative to that of deaths due to natural causes in this patient population. Conversely, in these UM populations the prodrug codeine (another CYP2D6 substrate) is metabolized much faster to morphine, often resulting in undesirable adverse effects of morphine, such as abdominal pain. Indeed, intake of high doses of codeine by a mother of the ultrarapid metabolizer type was held responsible for the morphine-induced death of her breast-fed infant.

The second well-studied genetic drug polymorphism involves the stereoselective aromatic (4)-hydroxylation of the anticonvul-sant mephenytoin, catalyzed by CYP2C19. This polymorphism, which is also inherited as an autosomal recessive trait, occurs in 3–5% of Caucasians and 18–23% of Japanese populations. It is genetically independent of the debrisoquin-sparteine polymor-phism. In normal “extensive metabolizers” (EMs) (S)-mephenytoin is extensively hydroxylated by CYP2C19 at the 4 position of the phenyl ring before its glucuronidation and rapidexcretion in the urine, whereas (R)-mephenytoin is slowly N-demethylated to nirvanol, an active metabolite. PMs however,appear to totally lack the stereospecific (S)-mephenytoin hydroxy-lase activity, so both (S)- and (R)-mephenytoin enantiomers are N-demethylated to nirvanol, which accumulates in much higherconcentrations. Thus, PMs of mephenytoin show signs of pro-found sedation and ataxia after doses of the drug that are well tolerated by normal metabolizers. Two defective CYP2C19 variant alleles (CYP2C19∗2 and CYP2C19∗3), the latter predominant in Asians, are responsible for the PM genotype. The molecular bases include splicing defects resulting in a truncated, nonfunctional protein. CYP2C19 is responsible for the metabolism of various clinically relevant drugs (Table 4–4). Thus, it is clinically impor-tant to recognize that the safety of each of these drugs may be severely reduced in persons with the PM phenotype. On the other hand, the PM phenotype can notably increase the therapeutic efficacy of omeprazole, a proton-pump inhibitor, in gastric ulcer and gastroesophageal reflux diseases.

Another CYP2C19 variant allele (CYP2C19∗17) exists that is associated with increased transcription and thus higher CYP2C19 expression and even higher functional activity than that of the wild type CYP2C19-carrying EMs. Individuals carrying this CYP2C19∗17 allele exhibit higher metabolic activation of pro-drugs such as the breast cancer drug tamoxifen, the antimalarial chlorproguanil, and the antiplatelet drug clopidogrel. The former event is associated with a lower risk of breast cancer relapse, and the latter event with an increased risk of bleeding. Carriers of the CYP2C19∗17 allele are also known to enhance the metabolism and thus the elimination of drugs such as the antidepressants esci-talopram and imipramine, as well as the antifungal voriconazole. This consequently impairs the therapeutic efficacy of these drugs, thus requiring clinical dosage adjustments.

The third relatively well-characterized genetic polymorphism is that of CYP2C9. Two well-characterized variants of this enzyme exist, each with amino acid mutations that result in altered metab-olism. The CYP2C9∗2 allele encodes an Arg144Cys mutation, exhibiting impaired functional interactions with POR. The other allelic variant, CYP2C9∗3, encodes an enzyme with an Ile359Leu mutation that has lowered affinity for many substrates. For example, individuals displaying the CYP2C9∗3 phenotype have greatly reduced tolerance for the anticoagulant warfarin. The war-farin clearance in CYP2C9∗3-homozygous individuals is about 10% of normal values, and these people have a much lower toler-ance for the drug than those who are homozygous for the normal wild type allele. These individuals also have a much higher risk of adverse effects with warfarin (eg, bleeding) and with other CYP2C9 substrates such as phenytoin, losartan, tolbutamide, and some nonsteroidal anti-inflammatory drugs (Table 4–4).

Allelic variants of CYP3A4 have also been reported, but their contribution to its well-known interindividual variability in drug metabolism apparently is limited. On the other hand, the expres-sion of CYP3A5, another human liver isoform, is markedly poly-morphic, ranging from 0% to 100% of the total hepatic CYP3A content. This CYP3A5 protein polymorphism is now known to result from a single nucleotide polymorphism (SNP) within intron 3, which enables normally spliced CYP3A5 transcripts in 5% of Caucasians, 29% of Japanese, 27% of Chinese, 30% of Koreans, and 73% of African Americans. Thus, it can significantly contribute to interindividual differences in the metabolism of preferential CYP3A5 substrates such as midazolam. Two other CYP3A5 allelic variants that result in a PM phenotype are also known.

Polymorphisms in the CYP2A6 gene have also been recently characterized, and their prevalence is apparently racially linked. CYP2A6 is responsible for nicotine oxidation, and tobacco smok-ers with low CYP2A6 activity consume less and have a lower incidence of lung cancer. CYP2A6 1B allelic variants associated with faster rates of nicotine metabolism have been recently discov-ered. It remains to be determined whether patients with these faster variants will fall into the converse paradigm of increased smoking behavior and lung cancer incidence.

Additional genetic polymorphisms in drug metabolism (eg, CYP2B6) that are inherited independently from those alreadydescribed are being discovered. For instance, a 20- to 250-fold variation in interindividual CYP2B6 expression partly due to genetic polymorphisms has been reported. This may have a sig-nificant impact on the metabolism of several clinically relevant drugs such as cyclophosphamide, methadone, efavirenz, selegiline, and propofol. Studies of theophylline metabolism in monozygotic and dizygotic twins that included pedigree analysis of various families have revealed that a distinct polymorphism may exist for this drug and may be inherited as a recessive genetic trait. Genetic drug metabolism polymorphisms also appear to occur for amin-opyrine and carbocysteine oxidations. Regularly updated informa-tion on human P450 polymorphisms is available at http://www.imm.ki.se/CYPalleles/.

Although genetic polymorphisms in drug oxidations often involve specific P450 enzymes, such genetic variations can also occur in other enzymes. Recently, genetic polymorphisms in POR, the essential P450 electron donor, have been reported. In particular, an allelic variant (at a 28% frequency) encoding a POR A503V mutation has been reported to result in impaired CYP17-dependent sex steroid synthesis and impaired CYP3A4- and CYP2D6-dependent drug metabolism in vitro. Its involvement in clinically relevant drug metabolism, while predictable, remains to be established. Descriptions of a polymorphism in the oxidation of trimethylamine, believed to be metabolized largely by the flavinmonooxygenase (Ziegler’s enzyme), result in the “fish-odorsyndrome” in slow metabolizers, thus suggesting that genetic vari-ants of other non–P450-dependent oxidative enzymes may also contribute to such polymorphisms.

B. Phase II Enzyme Polymorphisms

Succinylcholine is metabolized only half as rapidly in persons with genetically determined deficiency in pseudocholinesterase (now generally referred to as butyrylcholinesterase [BCHE]) as in per-sons with normally functioning enzyme. Different mutations, inherited as autosomal recessive traits, account for the enzyme deficiency. Deficient individuals treated with succinylcholine as a surgical muscle relaxant may become susceptible to prolonged respiratory paralysis (succinylcholine apnea). Similar pharmacoge-netic differences are seen in the acetylation of isoniazid. The defectin slow acetylators (of isoniazid and similar amines) appears to be caused by the synthesis of less of the NAT2 enzyme rather than of an abnormal form of it. Inherited as an autosomal recessive trait, the slow acetylator phenotype occurs in about 50% of blacks and whites in the USA, more frequently in Europeans living in high northern latitudes, and much less commonly in Asians and Inuits (Eskimos). The slow acetylator phenotype is also associated with a higher incidence of isoniazid-induced peripheral neuritis, drug-induced autoimmune disorders, and bicyclic aromatic amine-induced bladder cancer.

A clinically important polymorphism of the TPMT (thiopurine S-methyltransferase) gene is encountered in Europeans (frequency,1:300), resulting in a rapidly degraded mutant enzyme and conse-quently deficient S-methylation of aromatic and heterocyclic sulfhydryl compounds including the anti-cancer thiopurine drugs 6-mercaptopurine, thioguanine, and azathioprine, required for their detoxification. Patients inheriting this polymorphism as an autosomal recessive trait are at high risk of thiopurine drug-in-duced fatal hematopoietic toxicity.

Genetic polymorphisms in the expression of other phase II enzymes (UGTs and GSTs) also occur. Thus, UGT polymor-phisms (UGT1A1∗28) are associated with hyperbilirubinemic diseases (Gilbert’s syndrome) as well as toxic side effects due to impaired drug conjugation and/or elimination (eg, the anticancer drug irinotecan). Similarly, genetic polymorphisms (GSTM1) in GST (mu1 isoform) expression can lead to significant adverse effects and toxicities of drugs dependent on its GSH conjugation for elimination.

C. The Role of Pharmacogenetic Testing in Clinically Safe & Effective Drug Therapy

Despite our improved understanding of the molecular basis of pharmacogenetic defects in drug-metabolizing enzymes, their impact on drug therapy and ADRs, and the availability of vali-dated pharmacogenetic biomarkers to identify patients at risk, this clinically relevant information has not been effectively translated to patient care. Thus, the much-heralded potential for personal-ized medicine, except in a few instances of drugs with a relatively low therapeutic index (eg, warfarin), has remained largely unreal-ized. This is so even though 98% of US physicians are apparently aware that such genetic information may significantly influence therapy. This is partly due to the lack of adequate training in translating this knowledge to medical practice, and partly due to the logistics of genetic testing and the issue of cost-effectiveness. Severe ADRs are known to contribute to 100,000 annual US deaths, about 7% of all hospital admissions, and an increased aver-age length of hospital stay. Genotype information could greatly enhance safe and efficacious clinical therapy through dose adjust-ment or alternative drug therapy, thereby curbing much of the rising ADR incidence and its associated costs.

Diet & Environmental Factors

Diet and environmental factors contribute to individual variations in drug metabolism. Charcoal-broiled foods and cruciferous vegetables are known to induce CYP1A enzymes, whereas grapefruit juice is known to inhibit the CYP3A metabolism of co-administered drug substrates (Table 4–2). Cigarette smokers metabolize some drugs more rapidly than nonsmokers because of enzyme induction (see previous section). Industrial workers exposed to some pesticides metabolize certain drugs more rapidly than unexposed individu-als. Such differences make it difficult to determine effective and safe doses of drugs that have narrow therapeutic indices.

Age & Sex

Increased susceptibility to the pharmacologic or toxic activity of drugs has been reported in very young and very old patients com-pared with young adults. Although this may reflect differences in absorption, distribution, and elimina-tion, differences in drug metabolism also play a role. Slower metabolism could be due to reduced activity of metabolic enzymes or reduced availability of essential endogenous cofactors.

Sex-dependent variations in drug metabolism have been well documented in rats but not in other rodents. Young adult male rats metabolize drugs much faster than mature female rats or pre-pubertal male rats. These differences in drug metabolism have been clearly associated with androgenic hormones. Clinical reports suggest that similar sex-dependent differences in drug metabolism also exist in humans for ethanol, propranolol, some benzodiaz-epines, estrogens, and salicylates.

Drug-Drug Interactions during Metabolism

Many substrates, by virtue of their relatively high lipophilicity, are not only retained at the active site of the enzyme but remain non-specifically bound to the lipid endoplasmic reticulum membrane. In this state, they may induce microsomal enzymes, particularly after repeated use. Acutely, depending on the residual drug levels at the active site, they also may competitively inhibit metabolism of a simultaneously administered drug.

Enzyme-inducing drugs include various sedative-hypnotics, antipsychotics, anticonvulsants, the antitubercular drug rifampin, and insecticides (Table 4–5). Patients who routinely ingest barbi-turates, other sedative-hypnotics, or certain antipsychotic drugs may require considerably higher doses of warfarin to maintain a therapeutic effect. On the other hand, discontinuance of the seda-tive inducer may result in reduced metabolism of the anticoagu-lant and bleeding—a toxic effect of the ensuing enhanced plasma levels of the anticoagulant. Similar interactions have been observed in individuals receiving various combinations of drug regimens such as rifampin, antipsychotics, or sedatives with contraceptive agents, sedatives with anticonvulsant drugs, and even alcohol with hypoglycemic drugs (tolbutamide).

It must also be noted that an inducer may enhance not only the metabolism of other drugs but also its own metabolism. Thus, continued use of some drugs may result in a pharmacokinetic type of tolerance—progressively reduced therapeutic effectiveness due to enhancement of their own metabolism.

Conversely, simultaneous administration of two or more drugs may result in impaired elimination of the more slowly metabolized drug and prolongation or potentiation of its pharmacologic effects (Table 4–6). Both competitive substrate inhibition and irreversible substrate-mediated enzyme inactivation may augment plasma drug levels and lead to toxic effects from drugs with narrow thera-peutic indices. Indeed, such acute interactions of terfenadine (a second-generation antihistamine) with a CYP3A4 substrate-inhibitor (ketoconazole, erythromycin, or grapefruit juice) resulted in fatal cardiac arrhythmias (torsades de pointe) requiring its with-drawal from the market. Similar drug-drug interactions with CYP3A4 substrate-inhibitors (such as the antibiotics erythromy-cin and clarithromycin, the antidepressant nefazodone, the anti-fungals itraconazole and ketoconazole, and the HIV protease inhibitors indinavir and ritonavir), and consequent cardiotoxicity led to withdrawal or restricted use of the 5-HT₄ agonist, cisapride. Similarly, allopurinol both prolongs the duration and enhances the chemotherapeutic and toxic actions of mercaptopurine by competitive inhibition of xanthine oxidase. Consequently, to avoid bone marrow toxicity, the dose of mercaptopurine must be reduced in patients receiving allopurinol. Cimetidine, a drug used in the treatment of peptic ulcer, has been shown to potentiate the pharmacologic actions of anticoagulants and sedatives. The metabolism of the sedative chlordiazepoxide has been shown to be inhibited by 63% after a single dose of cimetidine; such effects are reversed within 48 hours after withdrawal of cimetidine.Impaired metabolism may also result if a simultaneously administered drug irreversibly inactivates a common metabolizing enzyme. These inhibitors, in the course of their metabolism by cytochrome P450, inactivate the enzyme and result in impairment of their own metabolism and that of other cosubstrates. This is indeed the case of the furanocoumarins in grapefruit juice that inactivate CYP3A4 in the intestinal mucosa and consequently enhance its proteolytic degradation. 

This impairment of their intes-tinal first-pass CYP3A4-dependent metabolism significantly enhances the bioavailability of drugs, such as felodipine, nifedipine, terfenadine, verapamil, ethinylestradiol, saquinavir, and cyclosporine A, and is associated with clinically relevant drug-drug and food-drug interactions.Recovery from these interactions is dependent on CYP3A4 resynthesis and thus may be slow.

Interactions between Drugs & Endogenous Compounds

Some drugs require conjugation with endogenous substrates such as GSH, glucuronic acid, or sulfate for their inactivation. Consequently, different drugs may compete for the same endoge-nous substrates, and the faster-reacting drug may effectively deplete endogenous substrate levels and impair the metabolism of the slower-reacting drug. If the latter has a steep dose-response curve or a narrow margin of safety, potentiation of its therapeutic and toxic effects may result.

Diseases Affecting Drug Metabolism

Acute or chronic diseases that affect liver architecture or function markedly affect hepatic metabolism of some drugs. Such condi-tions include alcoholic hepatitis, active or inactive alcoholic cirrhosis, hemochromatosis, chronic active hepatitis, biliary cir-rhosis, and acute viral or drug-induced hepatitis. Depending on their severity, these conditions may significantly impair hepatic drug-metabolizing enzymes, particularly microsomal oxidases, and thereby markedly affect drug elimination. For example, the half-lives of chlordiazepoxide and diazepam in patients with liver cirrhosis or acute viral hepatitis are greatly increased, with a cor-responding increase in their effects. Consequently, these drugs may cause coma in patients with liver disease when given in ordi-nary doses.Some drugs are metabolized so readily that even marked reduction in liver function does not significantly prolong their action. However, cardiac disease, by limiting blood flow to the liver, may impair disposition of those drugs whose metabolism is flow-limited (Table 4–7). These drugs are so readily metabolized by the liver that hepatic clearance is essentially equal to liver blood flow. Pulmonary disease may also affect drug metabolism, as indi-cated by the impaired hydrolysis of procainamide and procaine in patients with chronic respiratory insufficiency and the increased half-life of antipyrine (a P450 functional probe) in patients with lung cancer. The impaired enzyme activity or defective formation of enzymes associated with heavy metal poisoning or porphyria also results in reduced hepatic drug metabolism.Although the effects of endocrine dysfunction on drug metab-olism have been well explored in experimental animal models, corresponding data for humans with endocrine disorders are scanty. Thyroid dysfunction has been associated with altered metabolism of some drugs and of some endogenous compounds as well. Hypothyroidism increases the half-life of antipyrine, digoxin, methimazole, and some β blockers, whereas hyperthy-roidism has the opposite effect. A few clinical studies in diabetic patients indicate no apparent impairment of drug metabolism, although impairment has been noted in diabetic rats. Malfunctions of the pituitary, adrenal cortex, and gonads markedly reduce hepatic drug metabolism in rats. On the basis of these findings, it may be supposed that such disorders could significantly affect drug metabolism in humans. However, until sufficient evidence is obtained from clinical studies in patients, such extrapolations must be considered tentative.Finally, the release of inflammatory mediators, cytokines, and nitric oxide associated with bacterial or viral infections, cancer, or inflammation are known to impair drug metabolism by inactivat-ing P450s and enhancing their degradation.