Tacrine is not an ideal probe drug for measuring CYP1A2 activity in vivo

Abstract

Aims

The aim of the present study was to examine the CYP1A2 substrate tacrine as a possible alternative to caffeine for assessing CYP1A2 activity in vivo.

Methods

Eighteen, healthy, nonsmoking men participated. Each volunteer was tested by caffeine (200 mg orally), and caffeine metabolic ratios were calculated. Subsequently, on two occasions, separated by at least 4 weeks, each volunteer was tested with tacrine (40 mg orally). The apparent oral clearance, partial clearances and different metabolic ratios of tacrine were determined.

Results

The median oral clearances of tacrine in the two study periods were 1893 l h⁻¹ (range: 736–3098) and 1890 l h⁻¹ (range: 438–4175), respectively. The interindividual coefficient of variation was 42% and 49%, respectively. The intraindividual coefficients of variation ranged from 0.28% to 64% (median: 13%). In both study periods, the oral clearance of tacrine correlated with the caffeine urinary metabolic ratio. However, only modest magnitudes of correlation were observed (r_s: 0.64–0.66, P < 0.01). No tacrine metabolic ratio correlating with the oral clearance of tacrine was found.

Conclusion

The applicability of tacrine as a probe drug for measuring CYP1A2 activity in vivo appears limited.

Introduction

Cytochrome P4501A2 enzyme (CYP1A2) constitutes approximately 15% of the human liver cytochrome P450 content [1]. This hepatic enzyme plays a major role in the metabolism of a range of drugs (e.g. clozapine [2], theophylline [3], imipramine [4], propranolol [5]), and has also been shown to activate various procarcinogens [6].

The CYP1A2 activity exhibits substantial intra-and interindividual variation [7, 8]. Therefore, assessment of CYP1A2 activity in patients would be of theoretical value before treatment with CYP1A2 substrates, in particular those drugs having a low therapeutic index, and for which clinical dose titration is not feasible (e.g. clozapine, theophylline).

At present, caffeine is the model drug of choice for measuring CYP1A2 activity. However, caffeine may not be the ideal probe drug for assessing CYP1A2 activity. The clinical use of the caffeine test to predict the CYP1A2-mediated metabolism of drugs in the individual patient is hampered by the need for pretest methylxanthine abstinence (caffeine metabolism exhibits dose-dependency [9]). Further, caffeine has a very complex metabolism, and several different metabolic ratios in plasma, saliva and urine have been suggested in the literature [10].

The anticholinesterase tacrine (1,2,3,4-tetrahydro-9-aminoacridine; THA) is extensively metabolized to monohydroxy-, dihydroxy- and phenol glucuronide metabolites [11]. In vitro studies suggest that the formation of the monohydroxymetabolites (1-hydroxytacrine (1-OH-THA), 2-hydroxytacrine (2-OH-THA) and 4-hydroxytacrine (4-OH-THA)) involves CYP1A2 [12–14]. In one study [12], a correlation between the formation of the major metabolite, 1-OH-THA and the CYP1A2 content in liver microsomal samples was found. Evidence for a role of CYP1A2 in tacrine metabolism has also emerged from two recent in vivo studies showing an approximately 85% reduction in the apparent oral clearance of tacrine with concomitant fluvoxamine treatment [15, 16]. Fluvoxamine is a potent inhibitor of CYP1A2, both in vitro [17] and in vivo [18].

We therefore decided to investigate whether tacrine would be a useful alternative probe drug for measuring CYP1A2 in vivo.

Methods

Eighteen healthy, nonsmoking men, aged between 20 and 29 years (median: 24 years), participated in the study. All subjects completed a thorough medical examination, including ECG, clinical chemical and haematological screening, before taking any medications. None of the subjects had any history of regular drug intake, excessive alcohol consumption or disorders of the liver, GI tract or kidneys known to interfere with drug absorption, disposition, metabolism or excretion.

The study was approved by the regional ethics committee and the Danish National Board of Health. The volunteers consented to participate on the basis of both verbal and written information.

Study procedure

The study consisted of prestudy visits and two study periods (I and II) separated by a period of at least 4 weeks.

At the prestudy visits the volunteers were phenotyped with regard to sparteine, caffeine and mephenytoin oxidation.

When performing the caffeine test, the subjects abstained from ingesting methylxanthine-containing beverages, foods and medications for at least 24 h before and 6 h after intake of caffeine. Each subject ingested 200 mg caffeine (Nycomed DAK, Denmark). Six hours later a 10 ml blood sample was drawn, and a spot urine sample was given.

In the two study periods abstinence from all methylxanthine-containg beverages, foods and medications was instituted from one day before taking the first study dose and until the last blood sample was drawn. Further, no medications or alcoholic beverages were allowed. The volunteers fasted from midnight before taking the study medication.

The two study periods were identical. At 08.00 h the volunteers took a single oral dose of 40 mg tacrine (two capsules of Cognex® 20 mg, Parke-Davis Scandinavia AB, Sweden) and fasted for 1 h. At this time a standardized meal was served. Blood samples were drawn at 0, 0.5, 1, 2, 3, 4, 5, 6, 8, 10 and 24 h, and urine was collected for 0–6 h, 6–14 h and 14–24 h postdose. In addition a 10 ml spot urine was collected at 6 h.

All blood samples, except at 24 h, were drawn via an intravenous catheter (Venflon®, BOC Ohmeda AB, Sweden). The sample at 24 h was drawn by regular vein puncture. The blood was drawn in heparinized tubes (Venoject®, Terumo, Belgium). After centrifugation, the plasma was separated and frozen at −20° C until analysis. The urine volumes were recorded and 10 ml urine from each sample was kept at −20° C until analysis.

Analytical methods

Caffeine assay

Urine samples were analysed for 5- acetylamino-6-formylamino-3-methyluracil (AFMU), 1-methyluric acid (1MU), 1-methylxanthine (1MX) and 1,7-dimethyluric acid (17DMU) by a previously published h.p.l.c. method [19]. The limit of detection was 1–2 μm for all four metabolites. The accuracy was within ±3%. Plasma samples were analysed for caffeine (137TMX) and 1,7-dimethylxanthine (17DMX) by another h.p.l.c. method [20]. For this method the limit of detection was 0.2 μm for 17DMX and 0.6 μm for caffeine. The accuracy was within ±13%.

Tacrine assay

Plasma and urine samples were analysed for tacrine, 1-hydroxy-, 2-hydroxy-and 4-hydroxytacrine by an h.p.l.c. assay with fluorescence detection. The method was developed in our laboratory and has been described in detail elsewhere [21]. In brief, the method is based on a simple one-step liquid-liquid extraction with ethylacetate followed by isocratic, reversed phase high performance liquid chromatography and fluorescence detection. The limit of detection in plasma was 0.5 nm for 2-hydroxy-and 4-hydroxytacrine and 2 nm for tacrine and 1-hydroxytacrine. In urine the limit of detection was 60 nm for 2-hydroxy-and 4-hydroxytacrine, 30 nm for 1-hydroxytacrine and 80 nm for tacrine. The limit of quantification in plasma was 2.5 nm for 2-hydroxy-and 4-hydroxytacrine, 10 nm for 1-hydroxytacrine and 2 nm for tacrine. In urine the limit of quantification was 120 nm for all components. The overall mean recoveries ranged from 84 to 105% in plasma and 64–100% in urine for all four compounds. The accuracy was within ±7% for plasma and ±9% for urine.

Pharmacokinetic analysis

Caffeine

Based on the analysis of each of the urine samples, a caffeine metabolic ratio (CMRurine) was calculated according to Campbell [22]:

1

In plasma, a caffeine metabolic ratio (CMRplasma) calculated according to Fuhr & Rost [23]:

2

Tacrine

All pharmacokinetic parameters were calculated using the pharmacokinetic software PK Solutions 2.0 (Summit Research Services, Ohio, USA). The calculations were based on noncompartmental analysis. Thus, by means of a curve stripping technique, the concentration-time curves were described by biexponential functions in all cases, and parameter estimates based on the linear terminal part of the log-concentration curve were calculated. In the calculation of area under the concentration-time curve (AUC) use was made of the trapezoidal rule with extrapolation to infinity using the elimination rate constant and last observed concentration. Apparent oral clearance of tacrine was calculated as:

3

assuming complete absorption of tacrine.

The metabolic clearances for the formation of 1-, 2-and 4-hydroxytacrine were calculated according to:

4

where Ae_M(0→t) is the amount of metabolite (i.e. 1-, 2-or 4-hydroxytacrine) excreted in the urine during 0–24 h. AUC(0→t) is the corresponding AUC of tacrine. The use of equation 4 is based on the assumption that there is no further metabolism, that excretion of the metabolites is exclusively renal and that the rate of appearance of the metabolites in the urine is equal to their rate of formation.

In addition, the following tacrine metabolic ratios in urine and plasma were calculated:

5

6

7

Statistical analysis

All statistical analyses were carried out using the statistical software StatView version 4.5 for Macintosh (Abacus Concepts Inc., California, USA). Values are presented as medians and ranges. To describe the variation, coefficients of variation have been calculated for the tacrine clearance. Spearman’s rank correlation test was used for analysing the correlation between the various metabolic ratios and clearances.

P values <0.05 were considered statistically significant.

Results

All 18 volunteers completed the full study. Compliance was considered excellent.

With regard to cytochrome P4502D6 (CYP2D6) phenotype, all 18 volunteers, except number 15, were extensive metabolisers. All 18 volunteers were extensive metabolisers of mephenytoin (cytochrome P4502C19).

The median apparent oral clearances of tacrine in the two study periods (Table 1) were 1893 l h⁻¹ (range: 736–3098) and 1890 l h⁻¹ (range: 438–4175). A statistically significant correlation was found between clearances in the two study periods (Spearman rank correlation test; r_s = 0.58, P < 0.02). The interindividual coefficient of variation was 42% and 49% in the two study periods, respectively. The intraindividual coefficients of variation ranged from 0.28% to 64% (median value: 13%).

Table 1.

Apparent oral clearance of tacrine in study periods I and II.

The only caffeine metabolic ratio consistently correlating with any of total or partial clearances of tacrine was (AFMU+1MU+1MX)/17DMU in urine (Table 2). This metabolic ratio exhibited a statistically significant correlation with total apparent oral clearance of tacrine (r_s = 0.64 (period 1), r_s = 0.66 (period 2); P < 0.01). Also, the 1-OH-THA/THA ratio in plasma 4 h postdose was correlated to the formation clearance of 1-OH-THA. All other listed caffeine or tacrine metabolic ratios showed either no statistically significant correlations with the different clearances of tacrine or conflicting results.

Table 2.

The correlations between caffeine and tacrine metabolic ratios and oral and partial clearances of tacrine in urine and plasma samples from 18 healthy volunteers in the two study periods (Period I/Period II).

Likewise, no consistent statistically significant correlations between the different metabolic ratios of tacrine in plasma and urine, and the caffeine metabolic ratios were found (Table 3).

Table 3.

The correlations between tacrine metabolic ratios and caffeine metabolic ratios in urine and plasma samples from 18 healthy volunteers in the two study periods (Period I/Period II).

Discussion

Caffeine has been established as the model drug of choice for measuring CYP1A2 activity. By correlating caffeine clearance to measures of CYP1A2 content in human liver samples, Fuhr et al. [24] were able to show that CYP1A2 activity is the major determinant of caffeine clearance. The metabolic ratio most closely resembling caffeine clearance is probably 17DX/137TMX in plasma or saliva 5–7 h after administration of caffeine [23]. For population studies a urinary measure is often preferred [25]. The urinary metabolic ratio suggested by Campbell et al. [22] has been widely used. Some authors [26] have found that, due to variation in renal clearance, urinary metabolic ratios may not be valid measures of CYP1A2 at all. However, the urinary metabolic ratio used here seems to be superior in terms of sensitivity to variability in urine flow [27].

The predictive value of the caffeine test has been the subject of a few studies. While no correlation was found between the caffeine test and theophylline clearance [20], the caffeine test has been shown to correlate with clozapine clearance [2]. Likewise, the caffeine breath test has been shown to correlate with the logarithm of the steady-state plasma tacrine concentration [28]. While the present study was in progress, Fontana et al. [29] published a study on the utility of caffeine based probes of CYP1A2 activity in predicting the pharmacokinetics of tacrine in 19 patients with Alzheimer’s disease. The caffeine breath test and two urinary caffeine metabolic ratios were shown to correlate with tacrine oral clearance. However, the magnitude of the correlations did not seem to be clinically meaningful.

The present study is explorative, and a large number of tests of correlation have been performed. No adjustments of the significance level of the multiple tests were made. On this basis, the results must be interpreted cautiously.

A correlation between the apparent oral clearance of tacrine and the urinary caffeine metabolic ratio was found. In fact, the apparent oral clearance of tacrine also correlated with the plasma caffeine metabolic ratio, but only in the second study period. These results corroborate the findings of Fontana et al. [29], and further support a role of CYP1A2 in tacrine metabolism. Although statistically significant, the low degree of correlation between tacrine clearance and measures of CYP1A2 activity may be explained by the known intraindividual variation in CYP1A2 over time. To reduce variation in the sample, all participants in this study were nonsmoking men. However, apart from gender and smoking [30], a number of other factors (dietary components [31], e.g. cruciferous vegetables, charcoal broiled or smoked food, heavy exercising [32], coffee and alcohol intake [33], paracetamol use [33], etc.) have been shown to modify CYP1A2 activity. Further, being a highly extracted drug, the first-pass metabolism, and consequently, the total tacrine clearance, is dependent not only on enzyme activity, but also on hepatic blood flow [34], which may also vary in each individual.

Unfortunately, no relevant tacrine metabolic ratio, correlating with the apparent oral clearance of tacrine, was found. An explanation for this may be that the metabolic pathways studied here only account for a minor proportion of the total tacrine metabolism.

The metabolic ratio 1-OH-THA/THA appeared to correlate with the formation clearance of 1-OH-THA. As mentioned above, an in vitro study have shown that CYP1A2 content was correlated with 1-hydroxylation of tacrine [12]. As no correlation was found between caffeine metabolic ratios and 1-hydroxylation of tacrine in the present study, we were not able to corroborate this finding in vivo.

In the present study, the interindividual variation in the apparent oral clearance of tacrine was markedly greater than the intraindividual variation. This finding is much in line with studies on inter and intraindividual variation of caffeine metabolism [7, 8]. Kalow et al. [35] have advocated that comparisons of inter and intraindividual variation can substitute for twin studies to determine the heritability of drug responses. Theoretically, if a given drug response is much influenced by genetic variation, the differences between individuals will be larger than the time-to-time differences within individuals. One major exception being situations where the environmental effects vary much less for the individual than for the group. Tacrine metabolism, like caffeine metabolism, seems to be more influenced by genetic variation than environmental influences.

In conclusion, the oral clearance of tacrine was found to correlate with one measure of CYP1A2 activity. However, the magnitude of this correlation was modest, and no tacrine metabolic ratio correlating with either oral clearance of tacrine or measures of CYP1A2 activity was found. Therefore, the applicability of tacrine as a probe drug for measuring CYP1A2 activity in vivo appears limited.