Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Glucuronidation is a major pathway in human drug metabolism. The resulting glucuronic acid metabolites are generally considered to be pharmacologically inactive and rapidly excreted. A wealth of data, however, points to the fact that the generalized assumption of inactivity and rapid excretion is incorrect (for review see Kroemer and Klotz, 1992). First, glucuronides of drugs may accumulate during chronic therapy (Walle et al., 1979; Sutfin et al., 1988; Fromm et al., 1995). The increased hydrophilicity of drug glucuronides makes renal elimination a common route of clearance, so glucuronides accumulate in particular in patients with reduced kidney function, who have an impaired tubular secretion of glucuronides (Lieberman et al., 1985; Eriksson et al., 1989). For example, Fromm and coworkers (1995) observed stereoselective accumulation of the glucuronides of the antiarrhythmic propafenone (dose-corrected steady-state concentrations 2783 nmol/ml/mol dose and 7340 nmol/ml/mol dose for S-propafenone glucuronide and R-propafenone glucuronide, respectively, which is 15 to 20 times higher than in healthy volunteers) in four patients with renal failure. In addition, some glucuronic acid metabolites have direct pharmacological activity (for review see Kroemer and Klotz, 1992), the most prominent example being the analgesic effects that follow the administration of morphine-6-glucuronide (Osborne et al., 1988). Besides such direct activity, recruitment of the parent compound from accumulating glucuronides has been suggested for several drugs, such as clofibric acid, nonsteroidal anti-inflammatory drugs and lorazepam (Meffin et al., 1983; Brater, 1988; Herman et al., 1989). Although some data indicate involvement of esterases in the hydrolysis of clofibric acid glucuronides, the enzyme that catalyzes cleavage has not been definitively identified.

Cleavage of glucuronides of drugs can be catalyzed by the enzyme -Gluc, thereby modulating drug disposition and action in two different ways. Cleavage of inactive glucuronides may liberate the active parent compound and increase or prolong net drug effect. The opposite scenario arises from the action of -Gluc on active drug glucuronides, which would reduce drug effects. Thus the activity of this enzyme can modulate drug disposition, and hence drug action, via the cleavage of active or inactive glucuronides. Although the potential role of the enzyme in human drug metabolism has not been evaluated in a systematic manner, its physiological function has been characterized in detail.

The physiological role of the acid hydrolase -glucuronidase (EC3.2.1.31) is proteoglycan degradation in lysosomes. Genetic deficiency of the enzyme leads to a lysosomal storage disease known as mucopolysaccharidosis type VII (Sly et al., 1973). The tetrameric glycoprotein is composed of identical subunits with a molecular weight of 77 kD (Brot et al., 1978). After glycosylation and C-terminal processing within the endoplasmic reticulum and the Golgi complex, a portion of the synthesized -Gluc is directed to lysosomes via the mannose 6-phosphate receptor (Erickson and Blobel, 1983; Kornfeld, 1992; Shipley et al., 1993). Another portion is retained within the endoplasmic reticulum by association with the esterase egasyn (Tomino and Paigen, 1975; Medda and Swank, 1985). At both sites, -Gluc seems to be active in rodents (Brunelle and Verbeeck, 1993; Whiting et al., 1993), whereas dual localization and activity has not yet been shown in humans.

Several attempts have been made to utilize directly the activity of human -Gluc for bioactivation of nontoxic compounds (Henle et al., 1988; Bosslet et al., 1994). The reduced toxicity of drug glucuronides has been used as a takeoff point for administration of drug glucuronides as prodrugs, from which the active moiety is released because of the action of -Gluc. One example for such a prodrug is Dox-S-G, which shows considerably less systemic toxicity than Dox. Cytotoxic action of Dox requires cleavage of the glucuronide by -Gluc (Bosslet et al., 1994). The enzyme necessary for cleavage is administered as a fusion protein consisting of a humanized antibody directed against a tumor-specific surface antigen, e.g., the carcinoembryonic antigen (CEA) coupled to a -Gluc moiety (Bosslet et al., 1992). As an alternative to this "antibody-directed enzyme prodrug therapy" (ADEPT), administration of the prodrug alone (e.g., 8-hydroxyquinoline glucuronide; Henle et al., 1988) may result in enhanced tumor selectivity in view of the fact that a high activity of -Gluc has been reported for many tumors (Fishman and Anlyan, 1947).

Both for release of parent compounds from glucuronides and for cleavage of prodrugs such as Dox-S-G, variability in expression of -Gluc is a modulating factor. A systematic evaluation of interindividual variability in expression of human -Gluc and its consequences for drug metabolism in the human has not been reported. We therefore investigated variability in expression and function of -Gluc in human liver and kidney samples by using the following approaches: 1) Interindividual variability of function of -Gluc was assessed by cleavage of the model compound MUG in human liver (n = 30) and human kidney (n = 18). 2) Protein expression was assessed by immunoblotting of the same samples. 3) Detailed enzyme kinetics were described for MUG and Dox-S-G. 4) The inhibition constant was estimated for various other glucuronides of drugs. Using these techniques, we describe a wide interindividual variability in expression and function of -Gluc in liver and kidney, which is due to variations in protein content.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Tissue samples and chemicals. Human liver and kidney samples were obtained as surgical waste during partial hepatectomy or nephrectomy, respectively. Kidney samples were derived from cortical tissue. After surgery, samples were immediately frozen in liquid nitrogen and subsequently stored at 80°C. The mean age ± S.D. of the patients was 53.1 ± 16.5 and 62.4 ± 12.9 years for hepatectomy (15 male, 15 female; body weight ± S.D., 64.6 ± 13 kg) and nephrectomy (10 male, 8 female; body weight ± S.D., 74.7 ± 14.7 kg), respectively.

Preparation of tissue homogenates and enzymatic reactions. Frozen human liver or kidney samples (300 mg) were homogenized in 3 ml of 20 mM Tris-HCl, pH 7.4, at 4°C using an Ultra Turrax homogenizer (Bachhofer, Reutlingen, Germany) for 3 × 30 s at full speed. Protein contents were determined according to the method of Lowry et al. (1951). Incubation mixtures contained 2.25 or 1.13 µg of protein (for MUG and Dox-S-G cleavage assays, respectively) in 50 µl of assay buffer (200 mM sodium acetate, pH 5; 10 mM EDTA; 0.01% [w/v] bovine serum albumin, 0.1% [v/v] Triton X-100, 2.5 mM MUG). For enzyme kinetic experiments, increasing amounts of MUG (156 µM-5 mM) or Dox-S-G (6.25 µM-200 µM) were incubated. In inhibition experiments, the respective compounds were preincubated with the homogenates for 5 min in assay buffer, followed by the addition of MUG. All incubations were carried out at 37°C for 30 min to 2 h as duplicate or triplicate determinations with deviations of the mean below 10%. The enzymatic reaction was stopped by adding 150 µl of 200 mM sodium carbonate. Subsequently, 4 µl of 5 mM 9-CMA in dimethyl sulfoxide was added as an internal standard for analysis of MUG cleavage. Before analysis, the mixtures were centrifuged for 5 min at 13,000 rpm to separate from residual particles and precipitated protein. In order to exclude nonspecific binding, tissue homogenates were incubated with the specific -Gluc inhibitor saccharolactone at a final concentration of 1 mM.

HPLC analysis of cleaved -Gluc substrates. The liquid chromatographic system (Shimadzu, Duisburg, Germany) consisted of a LC-9A pump unit, a SIL-9A auto injector with a 100-µl loop, a RF 530 fluorescence detector (excitation at 355 nm, emission at 460 nm) and a C-R6A integrator.

Western blotting and densitometry. Western blot analysis was performed as described previously (Sperker et al., 1991). Briefly, 50 µg of tissue homogenate was subjected to 8% sodium dodecyl sulfate (SDS) gels. The blot was incubated with a hybridoma supernatant containing the monoclonal anti--Gluc antibody 2156/215 (diluted 1:250) for 1 h at room temperature. Densitometric analysis was done with an Elscript 400 densitometer (Hirschmann, München, Germany) at a wavelength of 546 nm. Band intensities were expressed in AU, and specific -Gluc contents were expressed as AU/mg protein.

Statistical analysis. Calculation of SD values and linear regression analysis were done with GraphPad Prism (GraphPad Software, Inc., San Diego, CA) or SlideWrite 3.0 software (Advanced Graphics Software, Inc., Carlsbad, CA). The kinetics of aglycon formation, V_max, K_m and the Hill coefficient n were calculated using the Michaelis-Menten equation: or the Hill equation: with the GraphPad Prism Software. EC₅₀ values were calculated by the equation EC₅₀ = K^1/n. V_max was expressed in nmol/mg/h, specific enzyme activity at a concentration of 2.5 mM MUG in µmol/mg/h and K_m in µM or mM. K_i was determined graphically from Dixon plots and expressed in µM or mM. Statistical distribution analysis and the Kolmogorov-Smirnov test for goodness of fit were performed with Statgraphics (Manugistics, Inc., Rockville, MD).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Specific -Gluc activities and protein levels in liver and kidney homogenates. Addition of the specific inhibitor saccharolactone (1 mM) completely abolished enzyme activity in both liver and kidney, thereby excluding nonspecific cleavage. The specific -Gluc activities measured in human liver samples ranged from 0.32 to 1.85 µmol/mg/h, with a mean value of 0.87 ± 0.34 µmol/mg/h. Kidney samples had a lower activity, which ranged from 0.07 to 1.00 µmol/mg/h, with a mean value of 0.39 ± 0.21 µmol/mg/h. We observed a log normal distribution in liver and kidney, with median values of 0.785 and 0.362 µmol/mg/h, respectively (fig. 1).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Distribution of specific -Gluc activities in 30 liver (panel A) and 18 kidney (panel B) tissue samples. Each bar represents the number of samples revealing the respective range of specific activity. Homogenates were diluted to 50 µg protein/ml in assay buffer containing 2.5 mM MUG and incubated at 37°C for 30 min. After termination of the reaction by addition of sodium carbonate, liberated MU was detected by HPLC analysis. Each sample was measured as duplicate determination.

View larger version (38K): [in this window] [in a new window]   Fig. 2.   Specific -Gluc contents and correlation to the respective specific activities in liver (panel A) and kidney (panel B) homogenates. Representative Western blots with 17 liver and 17 kidney homogenates are shown. First, 50 µg of homogenate protein was loaded on 8% SDS gels. Then it was transferred to nitrocellulose and incubated with the monoclonal antibody 2156/215. The intensities of the respective total spots were quantified densitometrically for determination of the specific -Gluc contents. Linear regression analyses of the -Gluc contents of liver (n = 30) and kidney (n = 18) samples were performed with the respective specific activities (panel A: r = 0.8, P < .001; panel B: r = 0.71, P < .05).

Enzymatic characterization of liver and kidney -Gluc. To assess the enzymatic characteristics in different tissues, enzyme kinetics were investigated with MUG and Dox-S-G. Using three different liver and kidney samples, we determined K_m and V_max values. With MUG, Michaelis-Menten-like kinetics were obtained. No difference between liver and kidney samples was observed with respect to K_m (table 1), whereas the V_max values were closely correlated with enzyme contents. In contrast to the observation with MUG, cleavage of Dox-S-G showed a sigmoidal velocity curve (fig. 3) as described by the Hill equation for allosteric enzymes. These data indicate a cooperative substrate binding pattern of Dox-S-G to -Gluc; the respective Hill coefficients n are displayed in table 1. The EC₅₀ values obtained for affinity of Dox-S-G indicate an affinity toward -Gluc two orders of magnitude higher than that of MUG. In contrast, maximum rates of formation correlated well with those determined for MUG (r = 0.98, P < .001).

View this table: [in this window] [in a new window]   TABLE 1 Enzyme kinetics of liver and kidney -Gluc using Dox-S-G and MUG as substrates

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Enzyme kinetics with different -Gluc substrates performed with liver and kidney homogenates. A) Substrate dependence of MU formation after incubation of liver and kidney homogenates (50 µg protein/ml, 30 min, 37°C) with serial 2-fold dilutions of MUG (0.156-5 mM), respectively. Data were fitted using the Michaelis-Menten equation (goodness of fit: r = 0.992 for liver and r = 0.997 for kidney). B) Substrate dependence of Dox formation after incubation of the respective homogenates (25 µg protein/ml, 1 h, 37°C) with serial 2-fold dilutions of Dox-S-G (6.25-200 µM). Data were fitted using the Hill equation (goodness of fit: r = 0.999 for liver and r = 0.991 for kidney).

Inhibition experiments with other glucuronides. We tested the potency of various drug glucuronides to inhibit cleavage of MUG. The compounds and their respective K_m, K_i or EC₅₀ values are displayed in table 2. The specific inhibitor of -Gluc, saccharolactone, was included as a positive control (K_i 5 µM). Dox-S-G was the compound with the lowest inhibition constant (60 and 50 µM for liver and kidney, respectively), followed by the glucuronidated steroid glycyrrhizin (470 and 570 µM, respectively). Estradiol 3-glucuronide (0.9 and 1.2 mM, respectively) and paracetamol glucuronide (1.6 and 2 mM, respectively) were less potent inhibitors of MUG clearance.

View this table: [in this window] [in a new window]   TABLE 2 Comparison of Km, EC50 and Ki values of different -Gluc substrates and inhibitors as measured with liver and kidney homogenates

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In this paper, we describe a pronounced interindividual variability of -Gluc activity in human liver and kidney. These data raise the question of which mechanisms underlie the variability observed. The close correlation between the protein content of the individual samples and the turnover rates suggests that this variability is due to different steady-state levels of -Gluc protein. Different levels of -Gluc expression may be caused by transcriptional induction and gene dose effects (Swank et al., 1978; Chabas et al., 1991) or other regulation events.

Another possibility is that the existence of various forms of the enzyme causes interindividual variability. Such isoenzymes of -Gluc have been suggested to exist in mice, in which experimental evidence pointed to the possibility of several enzymes with different thermolability and charge (Paigen, 1989). In order to investigate this possibility, we determined interindividual and intertissue differences in enzyme characteristics for liver and kidney. Assessed on the basis of cleavage of two substrates, MUG and Dox-S-G, enzyme characteristics are similar both interindividually and in different tissue types. Moreover, Western blots showed similar banding patterns.

The wide interindividual variation in expression of -Gluc can lead to variable cleavage of glucuronides. This is of particular interest in the kidney, in which we and others (Corrales-Hernandez et al., 1988) observed a wider interindividual variability as compared with liver. A high activity of renal -Gluc, in combination with the above-described tubular secretion of glucuronides, could result in a local release of the aglycon, which in the case of Dox could then result in local cytotoxicity.

An interesting observation is that of different enzyme kinetics for the two -Gluc substrates investigated. MUG showed a typical hyperbolic Michaelis-Menten curve, which is in agreement with data reported previously by other groups (Szasz, 1967; Fishman et al., 1967). In contrast, cleavage of Dox from the respective glucuronide was best described by a sigmoidal curve following the Hill equation. Because the enzyme is known to exist as a dimer-composed tetramer (Gehrmann et al., 1994), it is likely that cooperative interactions exist between the different subunits. The Hill coefficient n indicates the number of subunits in an oligomeric enzyme. We observed an n value between 2 and 3, which suggests either an incomplete cooperativity or cooperative interactions taking place between subunits in dimers. The question arises why -Gluc shows cooperativity with Dox-S-G, whereas we and other groups observed a hyperbolic velocity curve with MUG. One can speculate that the higher affinity of Dox-S-G compared with MUG leads to stronger conformational changes and hence interactions of the subunits.

Allosteric effects have been described for other drug-metabolizing enzymes (Johnson and Schwab, 1984) and may have implications for therapy with Dox-S-G. Assuming that our in vitro data are predictive for the in vivo situation, a small change in substrate concentration will lead to a marked increase in rate of release of Dox from the glucuronide. Further increase in the concentration will not enhance the formation rate in a proportional manner, and concentrations below a threshhold will not produce significant amounts of the active product. This "off-on switch" in enzyme kinetics based on the cooperative binding of Dox-S-G to -Gluc will require an exact achievement, during therapy, of target concentrations that are in the linear range of the substrate vs. velocity relationship. The cooperative binding may, in combination with the enhanced -Gluc expression in the tumor, lead to a further increase in tumor selectivity of Dox-S-G.

In terms of affinity to -Gluc, we observed a wide variation between substrates. Glucuronides of several model compounds were used to inhibit cleavage of MUG. It is known that the K_i value of a substance is in the same range as the K_m value, so we used K_i for MUG cleavage as a measure for substrate affinity toward -Gluc. Dox-S-G, which has the glucuronic acid moiety bound distant from the anthracycline residue via a synthetic spacer, showed the highest affinity to the enzyme, with the K_i value being in the same range as the respective EC₅₀ value. Glycyrrhizin, a naturally occurring compound that has antiviral, steroid-like and interferon-inducing activities (Kumagai et al., 1957; Pompei et al., 1979; Abe et al., 1982), contains two glucuronosyl moieties linked to a steroid. This substance has an affinity about 10 times lower than that of Dox-S-G. The putative K_m value for deglucuronidation of glycyrrhizin, however, is in the range of concentrations observed in humans (Kanaoka et al., 1986), and in vivo deglucuronidation has to be expected. In fact, deglucuronidation of glycyrrhizin was reported in a patient with pseudo-aldosteronism after i.v. administration of large doses of this drug (Kanaoka et al., 1986). Similar phenomena may be expected for both paracetamol glucuronide and estradiol 3-glucuronide, because both compounds show relatively high affinity to human -Gluc.

A pivotal question that arises when we address a potential -Gluc-mediated metabolism in the human is whether the activity of -Gluc is restricted to the lysosomal compartment or may occur at extralysosomal sites. If activity of -Gluc takes place exclusively in lysosomes, then cleavage of drug glucuronides would require targeted uptake into this compartment, which appears to be rather unlikely. There are, however, several lines of evidence arguing for an extralysosomal activity of -Gluc in the human. First, considerable elevations of serum -Gluc have been reported in various disease processes (Ohta et al., 1992; Camisa et al., 1988). For example, serum -Gluc concentrations are increased more than 16-fold in patients with AIDS (Saha et al., 1991).

Second, in rodents, -Gluc was shown to be present in both lysosomes and the endoplasmic reticulum. The enzyme is retained in the endoplasmic reticulum by means of the esterase egasyn. Several experiments indicate a functional role of -Gluc present in the endoplasmic reticulum. For example, cleavage of bilirubin glucuronides was significantly reduced in a mouse strain with low microsomal -Gluc activity (Whiting et al., 1993). Consequently, addition of the -Gluc inhibitor saccharolactone to rat and human liver microsomes leads to an apparent increase in the formation of glucuronic acid conjugates, because deconjugation by -Gluc is blocked (Brunelle and Verbeeck, 1993; Brunelle and Verbeeck, 1996). The functional role of -Gluc in the human endoplasmic reticulum, however, remains unclear.

In summary, our paper describes a high variability in the expression of -Gluc in human liver and kidney. Therefore, cleavage of drug glucuronides that accumulate during chronic therapy or are used as prodrugs can show a wide interindividual variability in humans, which might result in variable response to drugs.