Bing-Kou Tang

The use of caffeine for enzyme assays: A critical appraisal

Clinical Pharmacology and Therapeutics (1993) 53, 503–514; doi:10.1038/clpt.1993.63

Use of caffeine metabolite ratios to explore CYP1A2 and xanthine oxidase activities

Caffeine was used as a metabolic probe to screen healthy subjects for their activities of two enzymes, deduced to be CYP1A2 (an inducible cytochrome P450) and xanthine oxidase. A longitudinal study revealed modest effects of caffeine dose, ethanol intake, and time‐of‐day on the CYP1A2 index, without any effect on the xanthine oxidase index. The coefficients of intraindividual variation not accounted for were 5.0% for the xanthine oxidase and 17.2% for the CYP1A2 index. In a population study, both indexes showed a log normal distribution, with CYP1A2 values of most subjects covering a 6.3‐fold range but only a 1.7‐fold range with xanthine oxidase. The CYP1A2 index was 33% decreased in women who used oral contraceptives and substantially increased in cigarette smokers. Neither the CYP1A2 nor the xanthine oxidase index differed between volunteers of Chinese and European extraction. Four of 178 subjects showed unexplained low xanthine oxidase values (i.e., values several standard deviations below the mean).

Caffeine as a metabolic probe: Exploration of the enzyme‐inducing effect of cigarette smoking

It has been realized recently that the primary metabolism of caffeine in humans is catalyzed by P‐450IA2 and that the rate of caffeine metabolism can be estimated from a metabolic ratio in a single urine sample. A population of 178 students including 19 smokers were subjected to this caffeine test to establish their P‐450IA2 index. Both stated numbers of cigarettes smoked per day and urinary cotinine levels as a confirmatory measure correlated significantly with enzyme activity showing dose‐effect relationships (r = 0.62 and 0.89, respectively). Nevertheless, more nonsmokers than smokers had the highest enzyme indexes, suggesting that dietary elements or other factors may determine P‐450IA2 activities in populations. Because P‐450IA2 is a monooxygenase that may be confined to the liver, caffeine reveals directly the Ah‐receptor‐dependent enzyme induction only in the liver, but it may also be a signal of induction elsewhere.

Prominence of slow acetylator phenotype among patients with sulfonamide hypersensitivity reactions

Delayed hypersensitivity reactions are among the most severe adverse effects of the sulfonamides in current clinical use. These reactions appear to occur because of differences in the metabolism and detoxification of reactive metabolites of the sulfonamides. N‐Acetylation is a major metabolic pathway for the sulfonamides. Slow acetylation phenotype might be a risk factor for the development of these reactions. We determined the acetylation phenotype of 21 patients who had suffered hypersensitivity reactions to the sulfonamides. There were 11 females and 10 males in the group, with a mean age of 15 years (age range, 1.8 to 50 years). Their acetylator phenotype was determined by determining the ratio of urinary caffeine metabolites (1‐methylxanthine to 5‐amino‐6‐formylmethyluracil after an oral dose of 50 mg caffeine). Nineteen (90%) of the patients were slow acetylators compared to a 55% incidence of slow acetylators in a race‐matched control population (p < 0.008). This suggests that a slow acetylator phenotype is a risk factor for the development of sulfonamide hypersensitivity reactions and provides further support for the role of imbalances in genetically determined pathways of metabolism and detoxification of the sulfonamides in the pathogenesis of these reactions.

Caffeine as a metabolic probe: validation of its use for acetylator phenotyping.

The use of two caffeine metabolite ratios for acetylator phenotyping was validated by demonstrating concordance with two sulfamethazine tests in 178 unrelated healthy subjects. The caffeine metabolites used for this purpose were 5‐acetylamino‐6‐amino‐3‐methyluracil (AAMU), 1‐methylxanthine (1X), and 1‐methylurate (1U). The ratio AAMU/(AAMU + 1X + 1U), referred to as molar ratio or N‐acetyltransferase, was compared with the ratio AAMU/1X. The results indicated that, for screening purposes, the acetylator phenotype can be determined by analysis of a 6‐hour urine sample after a cup of coffee or strong tea or a can of caffeine‐containing soft drink. The ratio AAMU/1X is the ratio of choice for the study of subjects in whom variability of xanthine oxidase can be neglected; use of the ratio AAMU/(AAMU + 1X + 1U) appears appropriate for special purposes. Gender, ethnic origin, habitual or moderate consumption of coffee, tea, soft drinks, or ethanol, or cigarette smoking have little if any effect on the caffeine tests for acetylator phenotyping.

CYP1A2 activity as measured by a caffeine test predicts clozapine and active metabolite norclozapine steady-state concentration in patients with schizophrenia

Clozapine is an atypical antipsychotic drug and displays efficacy in 30% to 60% of patients with schizophrenia who do not respond to traditional antipsychotics. A clozapine concentration greater than 1,150 nmol/L increases the probability of antipsychotic efficacy. However, plasma clozapine concentration can vary more than 45-fold during long-term treatment. The aim of this study was to assess the contribution of CYP1A2 to variability in steady-state concentration of clozapine and its active metabolite norclozapine. Patients with schizo-phrenia or schizoaffective disorder were prospectively monitored during clozapine treatment (N = 18). The in vivo CYP1A2 activity was measured using the caffeine metabolic ratio (CMR) in overnight urine. Trough plasma samples were drawn after at least 5 days of treatment with a constant regimen of clozapine. A significant negative association was found between the CMR and the dose-corrected clozapine (r s = −0.87, p < 0.01) and norclozapine (r s = −0.76, p < 0.01) concentrations. Nonsmokers displayed a higher clozapine (3.2-fold) and norclozapine (2.3-fold) concentration than smokers (p < 0.05). Furthermore, there was marked person-to-person variation in CYP1A2 activity during multiple-dose clozapine treatment (coefficient of variation = 60%). Age, weight, serum creatinine, and grapefruit juice consumption did not significantly contribute to variability in clozapine and norclozapine concentration (p > 0.05). In conclusion, CYP1A2 is one of the important contributors to disposition of clozapine during multiple-dose treatment. Although further in vitro experiments are necessary, the precise metabolic pathways catalyzed by CYP1A2 seem to be subsequent to the formation of norclozapine, hitherto less recognized quantitatively important alternate disposition routes, or both. From a clinical perspective, an environmentally induced or constitutively high CYP1A2 expression can lead to a decrease in steady-state concentration of clozapine as well as its active metabolite norclozapine. Thus, interindividual variability in CYP1A2 activity may potentially explain treatment resistance to clozapine in some patients. CYP1A2 phenotyping with a simple caffeine test may contribute to individualization of clozapine dosage and differentiate between treat ment noncompliance and high CYP1A2 activity.

Variable activation of lovastatin by hydrolytic enzymes in human plasma and liver

Lovastatin, widely used to lower cholesterol, is a pro-drug that requires metabolic activation through hydrolysis by carboxyesterases. There appear to be at least three distinct esterases in humans capable of catalysing this reaction, one in plasma and two in the liver.The rate of lovastatin hydroxy acid formation was measured as 15.8 pmol · ml−1 · min−1 in plasma, 2.13 pmol · mg−1 protein · min−1 in hepatic microsomes and 0.92 pmol · mg−1 protein · min−1 in cytosol. The data suggest that on average the three esterases together are capable of activating about 220 nmol (90 μg) lovastatin per minute per person, to which the esterases of plasma, liver microsomes and liver cytosol contribute approximately 18, 15 and 67%, respectively.All three esterases showed evidence of inter-individual variability. In one of 17 livers, both cytosolic and microsomal esterase activity was completely missing, while two other liver specimens lacked one esterase.Such variability must be expected to influence the therapeutic efficacy of the drug, and they might be related to its occasional toxicity.

An alternative test for acetylator phenotyping with caffeine

Previously published methods allow the determination of the genetically controlled acetylator status using caffeine as a test drug, based on the urinary excretion of a ring‐opened metabolite of caffeine, an acetylated uracil (5‐acetylamino‐6‐formylamino‐3‐methyluracil). 5‐Acetylamino‐6‐formylamino‐3‐methyluracil is labile but can be converted into a stable, deformylated product referred to as 5‐acetylamino‐6‐amino‐3‐methyluracil, which has recently been shown to be quantifiable by exclusion chromatography. The first part of the present article represents a longitudinal study of three subjects to assess the intraindividual variability of those caffeine metabolite ratios that are of potential interest for the determination of acetylator phenotypes. Effects of single and multiple doses, as well as of different periods of urine collection, were tested. A ratio relating the excretion of 5‐acetylamino‐6‐amino‐3‐methyluracil to that of all products of the 7‐demethylation pathway of paraxanthine proved to be highly reproducible, particularly after collection of overnight urine after coffee consumption during the day. This ratio showed complete concordance with the plasma index for sulfamethazine acetylation. The second part of this article showed the use of this ratio in a population study. It allowed a good separation of slow and fast acetylators and probably also a separation of homozygous and heterozygous fast acetylators.

A method for studying drug metabolism in populations: racial differences in amobarbital metabolism.

The two main metabolites of amobarbital excreted in urine are 3′‐hydroxyamobarbital (C‐OH) and 1‐(β‐D‐glucopyranosyl) amobarbital (N‐glu). When testing the metabolite ratio in small single samples of urine, it was found that the urine in a Caucasian population contained about one‐third glucose conjugation and two‐thirds hydroxylation product, while an Oriental population excreted both metabolites in equal proportion. Attempts to learn the causes for the different metabolite ratios led to an investigation of metabolite concentrations in urine. The sums of the 2 metabolite concentrations were the same in both populations. The average urinary concentration of C‐OH was greater in Caucasians than in Orientals, no matter how the data were expressed; the reverse was true for the N‐glu metabolite. C‐OH data was scattered more widely among Orientals than Caucasians; this might indicate bimodality of the distribution curves. There also was a trend toward more N‐glu metabolite in urine of females than of males. Measuring the metabolite/creatinine ratios narrowed the distribution range of the data, particularly after correction for sex difference in creatinine, but population differences were not changed. Expected relationships between metabolite content of urine, sampling times, and plasma half‐life (t½) were established by calculation. A Caucasian female with no capacity for N‐glucosidation was found during the first part of this population survey.16 An Oriental male with only trace capacity for amobarbital hydroxylation was found in the second part.

The science of pharmacological variability: An essay

1 introduced the concept of an LD 50, the dose killing 50% of a group of animals. Many ingenious observers before him thought of drugs, their effects, and their use for therapy. However, the special aspect of a science is that it has a component that guides the measurements, the quantification, of a set of observations. Quantification was often irregular and random before Trevan. For example, toxicity was measured in terms of a “minimal toxic dose,” a quantity that would tend to vary drastically from case to case. The LD 50 is basically a statistical measurement, but an important aspect of it is its assertion that all drug responses differ from case to case, from person to person, or from one animal or tissue to the next. The LD 50, or its modern functional counterpart, the ED 50, means recognition that variability is a factor that pertains to all drugs under all circumstances. It was about 40 years ago when it became firmly established that variability of drug response can have a genetic cause. 2,3 Since that time, the subspecialty of pharmacogenetics put on record dozens of examples that show determination of an altered drug response by genetic variation of a particular protein and, more recently, by proving the effects of mutant genes. 4 The variable proteins initially investigated were most often drug-metabolizing enzymes. Well-known examples are variants of the P450 cytochromes such as CYP2D6, responsible for the metabolism of debrisoquin (INN, debrisoquine), dextromethorphan, and at least three dozen other drugs. 5 At the present time, advanced techniques are also showing genetic variability of drug targets, for instance, drug receptors. 6 Examples are studies of the variability of the different forms of dopamine, serotonin, or other G-protein–associated receptors. 7