Pharmacogenomics and Dermatological Therapeutics FREE

Advances in the understanding of the structure and function of the human genome offer an opportunity to reexamine some old questions about the interplay between genetics and individual reactions to drugs. Pharmacogenetics examines questions related to individual variations in drug efficacy as well as drug toxicity. Why is it that a drug affects one individual in one way, while it may lead to a more blunted or a more dramatic therapeutic effect in another individual, or cause a serious untoward reaction in another? Why is it that the same drug can affect a person in one way on one occasion, but in a different way at another point in time? Although exogenous factors such as concomitant viral infection and multiple drug interactions may account for some variation in drug response, genetic variability in drug metabolism, drug clearance, or end-organ effect, accounts for much of the variation that previously was considered to be "idiopathic."^1- 2

Although adverse reactions to drugs remain a leading cause of morbidity and mortality, at present, we lack effective tools to identify individuals with elevated risk for most toxic effects. In several important cases, however, the underlying explanation for an "idiopathic" event has been elucidated. We will briefly discuss 5 such examples; the fact that only a small number of the hundreds of unexplained drug reactions are well worked out illustrates how much work is yet to be accomplished in this field. The examples, however, demonstrate the potential power of pharmacogenetics to unravel many of these mysteries and to help us to anticipate and avoid many of the toxic drug effects that we see today. Some of the pharmacogenetic conditions with relevence to dermatology are listed below.³

An example of a drug toxicity mystery that was solved is the case of primaquine-induced hemolytic anemia. It had been long recognized that a small subset of patients, particularly African American or Caribbean American males, developed hemolytic anemia during treatment with primaquine. Affected individuals were found to have low levels of functional activity of the enzyme G6PD. The cause for the faulty enzyme was found to lie in a defect in the gene coding for G6PD: a single base substitution (asparagine to aspartic acid) accounted for the defective enzyme that produced this toxic effect. The G6PD gene has subsequently been found to have more than 400 variants. A variation that occurs in more than 1% of the studied population is designated a "polymorphism." Because G6PD is involved in the metabolism of several other drugs, such as hydroxychloroquine and dapsone, the administration of these drugs to susceptible individuals also entails the risk of drug-induced hemolytic anemia.

The next example of a genetic polymorphism with clinical relevance is that of the acetylation polymorphism. In the 1950s, a high variation in individual rates of excretion of isoniazid was observed among people being treated for tuberculosis.⁴ Following a single oral dose of isoniazid, a bimodal pattern of plasma isoniazid levels was demonstrated, leading to the concept of rapid and slow eliminators of this drug. The concept that this variation may have had a genetic basis arose from the observation that monozygotic and dizygotic twins had a high concordance rate for excretion rates. Further investigation revealed that the enzyme responsible for the metabolism of isoniazid was N-acetyltransferase (NAT). This enzyme is central in the metabolism of a wide variety of drugs, all of which contain an arylamine or hydrazine group. The genetic basis for variability in the action of this enzyme results from polymorphisms at the NAT2 gene locus. Fifteen variant alleles for NAT2 have been identified. Several of the alleles have been associated with the rapid acetylator phenotype (NAT2*4, NAT2*12, and NAT2*13), while others have been associated with slow acetylation (NAT2*5, NAT2*6, NAT2*7, and NAT2*14S).⁵

Extensive investigation of the clinical differences between slow and rapid acetylators has since taken place. As expected, individuals who are rapid acetylators excrete the target drugs rapidly and therefore experience higher than expected rates of failure of therapy with isoniazid. Similarly, rapid acetylators require higher doses of hydralazine for blood pressure control. Individuals who are slow acetylators are more likely to develop a host of toxic side effects from medications, including drug-induced lupus from procainamide hydrochloride and hydralazine; neuropathy from isoniazid; and even toxic epidermal necrolysis from sulfonamides.⁶ The acetylation polymorphism does not completely explain these reactions, however, because more than 50% of white individuals have the slow acetylator phenotype.

Cytochrome P-450 (CYP) enzymes (hepatic and extrahepatic) are central to the metabolism of many classes of drugs. Most drugs are small lipophilic compounds that require conversion to a more polar state to be excreted. This conversion is generally accomplished through 2 steps: oxidation by phase 1 enzymes such as the cytochromes, then conjugation of polar side groups by phase 2 enzymes such as acetylases, sulfatases, or glucuronidases. Genetic polymorphisms may occur in all CYP enzymes; however, most of the clinically important polymorphisms occur in 5 isozymes: CYP2A6, CYP2C9, CYP2C19, CYP2D6, and CYP3A4, the latter 2 being the most important drug metabolizing enzymes in humans. Numerous alleles have been identified in association with each of these. Six percent of whites lack the CPY2D6 enzyme altogether due to the presence of 2 null alleles. This makes these individuals susceptible to more adverse reactions as well as alteration in efficacy of these drugs, which include, among others, amitryptyline hydrochloride, clotrimazole, codeine, imipramine hydrochloride, ketoconazole, loratadine, methoxsalen, ondansetron hydrochloride, perphenazine, and sulfasalazine.

Azathioprine is an immunosuppressive compound used in the treatment of rheumatologic disease and organ transplantation. Although the medication is usually well tolerated, a subset of patients who receive azathioprine develop severe, life-threatening leukopenia. Thiopurine methyltransferase (TPMT) is the enzyme responsible for the metabolism of azathioprine. Thiopurine methyltransferase is polymorphic in humans, with a triphasic distribution: 89% of people have normal enzyme levels, 11% have intermediate activity, but 0.3% of people have virtually no TPMT activity at all. Studies of individuals with low erythrocyte TPMT levels demonstrate that these individuals are at high risk for severe bone marrow suppression during azathioprine therapy. Levels of TPMT can be obtained in some commercial laboratories; however, the test is not yet widely available.

Although most patients treated with phenytoin and other aromatic anticonvulsant drugs for seizures tolerate these medications well, a small subset of patients develop a catastrophic hypersensitivity syndrome consisting of fever, rash, facial swelling, lymphadenopathy, hepatitis, nephritis, pneumonitis, and eosinophilia.⁷ The pathophysiology of the hypersensitivity syndrome has not been fully elucidated. Phenytoin, carbamazepine, and phenobarbital are hydroxylated by the P-450 system. It has been hypothesized that this process may generate reactive intermediates such as arene oxides, which can bind to cellular components, possibly triggering apoptosis or a secondary immune response. The putative reactive metabolite has not yet been directly measured or detected; however, evidence suggests that it can be detoxified by the enzyme epoxide hydrolase. The lymphocyte toxicity assay is an in vitro drug metabolite toxicity system that measures survival of lymphocytes when incubated with the drug in question and murine hepatic microsomal enzymes. The murine enzymes are thought to convert the drug into its metabolites, some of which may be toxic. Lymphocytes from healthy individuals are able to survive in this environment; however, lymphocytes from patients with the hypersensitivity syndrome die earlier in the assay. It is presumed that the lymphocytes of healthy individuals are able to detoxify the metabolites produced by the murine microsomes (with the enzyme epoxide hydrolase), whereas the lymphocytes from patients with the hypersensitivity syndrome are unable to perform this detoxification, and therefore become cellular targets of the toxic compounds and do not survive. Evidence is strong for a genetic component to these reactions; affected individuals have markedly decreased lymphocyte survival time in the lymphocyte toxicity assay, and their first-degree relatives have a greater than expected chance of diminished lymphocyte survival.

The genetic underpinning of these reactions is still not known, however. In an elegant study, a comparison of the structure and sequence of the microsomal epoxide hydrolase gene (mEH) was made between patients with the anticonvulsant hypersensitivity syndrome and healthy controls. While several distinct point mutations were detected, none correlated with the anticonvulsant toxicity phenotype.⁸ From this information we may conclude either that mEH is a false lead, and the specific gene defect continues to elude us, or that the explanation for the hypersensitivity reaction mystery is more complex than a single gene defect.

So, where does this leave us? Better understanding of the genetic control of drug metabolism will help us to identify patients at risk for drug toxicity. It also will give us the power to select pharmacologic agents that are best suited to a patient's particular set of drug metabolizing enzymes.

At this point, however, we face a series of challenges. First, genetic information related to drug reactions and drug efficacy currently is sparse and only applicable to a small handful of drugs. The logic behind the TPMT azathioprine story is impeccable and we now have the tools to anticipate a serious side effect, and protect at-risk patients from a previously "idiopathic" reaction. Why, however, can we not do the same for hundreds of other reactions that occur in only a subset of treated patients, such as drug-induced lupus from minocycline, or toxic epidermal necrolysis from allopurinol, or contact dermatitis from neomycin, or cutaneous atrophy from topical corticosteroids? The possibilities are endless, but much work is ahead to capitalize on the potential that our genetic understanding may offer.

A second challenge is that most of the available pharmacogenetic information has still not crossed over and become integrated into clinical practice. The TPMT assay is not widely commercially available. If we can check G6PD levels before starting dapsone therapy, we should be able to check TPMT levels before starting azathioprine therapy. Once the clinician is aware that it is possible to identify at-risk patients, it can be maddening that the test cannot be performed by the local laboratory. Once the test becomes more widely available, guidelines for dosing of the drug in people with normal TPMT levels will also need to be revised.

A third challenge will be to face the numerous nongenotypic factors that affect an individual's drug metabolism. We are already aware of situations in which an individual's acetylation phenotype differs from that which would be expected, based upon the patient's alleles for the NAT2 gene. Individuals infected with the human immunodeficiency virus (HIV) have, for example, been shown to have a greater chance of having a slow-acetylator phenotype than HIV-negative controls. While this finding has been both confirmed⁹ and refuted,¹⁰ it offers a possible window of explanation into why HIV-positive patients have such high rates of drug eruptions due to sulfonamides. The coadministration of other medications can also profoundly affect drug metabolism. Terbinafine and diphenhydramine both inhibit the activity of the CYP2D6 gene, which alters the host's ability to metabolize many other drugs metabolized by this enzyme system. Itraconazole, as a potent inhibitor of CYP3A4, can lead a similar alteration in the metabolism of its group of drugs. The medications can therefore essentially override the individual's genotype for drug metabolism, and alter the individual's pharmacologic phenotype.

A fourth challenge will be to expand from a pharmacogenetic mind-set, namely, a focus on monogenic defects that result in relatively easy-to-measure phenotypes (such as hemolytic anemia from dapsone use in G6PD-deficient individuals), to a pharmacogenomic mind-set—one that allows us to look for explanations to more complex traits related to drug toxicity and response.^11- 12 We may use sulfonamide-induced toxic epidermal necrolysis as an example. While it has been demonstrated that patients with the slow acetylator phenotype are at elevated risk for this devastating reaction, the preselection of individuals at highest risk is still not possible with acetylator status alone, because more than 50% of some populations (whites, for example) are slow acetylators. Some other factors must be involved: perhaps a unique allele for one of the cytochrome enzymes, or perhaps another gene involved in the regulation of sulfonamide metabolism. High throughput genomic techniques will allow a search among thousands of possible single nucleotide polymorphisms that may be associated with a given reaction.

Accepted for publication August 21, 2001.

Corresponding author and reprints: Mark H. Lowitt, MD, Department of Dermatology, University of Maryland School of Medicine, 405 W Redwood St, Sixth Floor, Baltimore, MD 21201.