![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Center for Neuropharmacology and Neuroscience, Albany Medical College, Albany, New York (R.S., J.Y., J.W.N., L.B.H.); and Curragh Chemistries, Cleveland, Ohio (J.G.P.)
(Received July 26, 2007; Accepted December 18, 2007)
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
). In laboratory animals and/or humans, however, cimetidine can produce significant additional effects including mental confusion, seizures, antinociception, and adverse drug interactions (Goodman et al., 2001; Cannon et al., 2004). The antinociceptive and seizure-producing effects of cimetidine are actions on the brain that are not mediated by the H2 receptor (Cannon et al., 2004). Adverse drug interactions with cimetidine have been attributed to its well known ability to act as a low-potency inhibitor of the cytochrome P450 (P450) superfamily (Sorkin and Darvey, 1983).
|
). Although early studies concluded that this binding represented brain H2 receptors, subsequent work showed that this conclusion was incorrect. For example, imidazole-containing H2 ligands have high affinity for the brain 3HCIM-binding site, but nonimidazole H2 ligands lack such activity (Rising et al., 1980; Smith et al., 1980; Warrander et al., 1983). Furthermore, Warrander et al. (1983) reported the existence of imidazole-containing compounds with high affinity for the 3HCIM-binding site but lacking affinity for the H2 receptor. Thus, cimetidine has a high-affinity, biological target in the brain that is distinct from the H2 receptor and remains to be identified. Liver, kidney, and gastric mucosa have also been reported to contain a 3HCIM-binding site (Rising et al., 1980; Chansel et al., 1982; Reilly et al., 1983; Warrander et al., 1983).
Because several of the side effects of cimetidine are not mediated by H2 receptors, the identity of the 3HCIM-binding site is of biological interest. For example, we recently suggested that the 3HCIM-binding site might represent an antinociceptive target. In support of this idea, CC12 (Fig. 1), a new derivative of cimetidine, showed high affinity for 3HCIM-binding sites and behaved in vivo as a competitive antagonist of several antinociceptive agents (Hough et al., 2007). In the present study, we have characterized the pharmacological nature of 3HCIM binding in brain homogenates and tested the hypothesis that this binding site is a member of the P450 superfamily.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Synthesis of CC11 and CC12. For synthesis of CC11, (1H-imidazol-4-yl)-methanol (3.44 g, 0.035 mol) and benzyl mercaptan (1.25 g, 0.010 mol) were dissolved in glacial acetic acid (12 ml) and heated at reflux under N2 for 4 days. The reaction was cooled to room temperature, and acetic acid was evaporated in vacuo. The residue was dissolved in dichloromethane, and the pH was adjusted to 9–10 with aqueous 10% sodium hydroxide solution. The dichloromethane layer was separated, the aqueous layer was extracted twice with dichloromethane, and the dichloromethane extracts were combined, dried over MgSO4, filtered, and evaporated to give 2.24 g of crude product, a brown viscous oil. Purification by flash silica chromatography using CH2Cl2-methanol (95:5) gave 1.5 g of a colorless viscous oil that became a waxy solid upon standing. The free base (100 mg) was dissolved in 4 ml of 4 N HCl in dioxane and stirred overnight at room temperature. Organic volatile compounds were evaporated in vacuo and dried on high vacuum for 1 h at 50°C, yielding CC11 [4(5)-(benzylthiomethyl)-1H-imidazole hydrochloride] (Fig. 1). Nuclear magnetic resonance (300 MHz, dimethyl sulfoxide-d6) 
(9.05, s, 1H), (7.55,s, 1H), (7.37–7.20, m, 5H), (3.78, s, 4H). CC12 [4(5)-((4-iodobenzyl)thiomethyl)-1H-imidazole] (Fig. 1) was synthesized as described recently (Hough et al., 2007).
Tissue Preparation. Sprague-Dawley rats were euthanized with an overdose of pentobarbital or CO2, and brains were rapidly removed. In some cases frozen brains were purchased (Taconic Farms, Germantown, NY). Crude whole brain fractions were prepared by homogenizing (Polytron) in 10 volumes of homogenate buffer (100 mM Tris-HCl and 0.5 mM EDTA, pH 7.4) and centrifugation (26,000g for 15 min). Supernatant fractions were removed, and the pellets were resuspended in the same volume of buffer with a glass-Teflon homogenizer. Resuspensions were recentrifuged, supernatants were discarded, and pellets were stored at -80°C. On the day of assay, pellets were resuspended in assay buffer (100 mM Tris-HCl, pH 7.4) and centrifuged at 26,000g for 10 min, supernatants were discarded, and pellets were again resuspended in assay buffer and analyzed for 3HCIM-binding activity. For some experiments, whole brains were homogenized in sucrose buffer (250 mM sucrose, 50 mM Tris-HCl, and 4 mM EDTA, pH 7.4) with a glass-Teflon homogenizer and spun at 100,000g for 1 h. Supernatants were discarded, and pellets were stored at -80°C. On the day of the assay, pellets were resuspended in assay buffer and assayed for 3HCIM-binding activity.
3HCIM Radioligand Binding. 3HCIM-binding experiments were performed following Smith et al. (1980). Resuspended crude membrane pellets (360–470 µg of rat brain protein) or 2 pmol of recombinant P450s (rP450s) (5.4–14.3 µg of protein, Supersomes; BD Biosciences, Woburn, MA) were incubated in a total volume of 0.1 ml containing 50 nM 3HCIM (22–25 Ci/mmol; GE Healthcare Biosciences Corp., Piscataway, NJ), appropriate concentrations of competing ligand, and assay buffer for 60 min on ice, in polystyrene test tubes. Burimamide (30 µM) or cimetidine (10 µM) was used to evaluate nonspecific binding. After incubation, samples were rapidly filtered through GF/B filters and rinsed three times with 1.5 ml of ice-cold assay buffer. Filters were placed in 5 ml of Ecoscint scintillation fluid and counted in a scintillation counter. Protein content was determined using the bicinchoninic acid method (Pierce Chemical, Rockford, IL). For competition studies, percent specific binding was calculated using the following formula: (drug - nonspecific)/(total - nonspecific) x 100, where drug indicates the amount of binding in the presence of competing ligand. For antibody studies, anti-CYP2C8/9/18/19 antibody (0.1–0.5 µg of protein, clone 1-68-11; Calbiochem, La Jolla, CA) and rat brain homogenate (resuspended 100,000g pellet, 308 µg) were preincubated for 20 min at 37°C in a volume of 60 µl, chilled (5 min), and assayed as described for 3HCIM binding in a final volume of 0.1 ml.
Quantification of Recombinant P450 Activities. Activities of CYP2B6, CYP2C19, and CYP2C11 were quantified by measurement of 9-anthraldehyde oxidation, similar to the methods of Marini et al. (2003). Briefly, recombinant CYP2B6 (rCYP2B6) (2 pmol, 26 µg of protein; BD Biosciences) was preincubated for 15 min at 37°C with 9-anthraldehyde (5 µM final) and inhibitors in 100 mM potassium phosphate buffer, pH 7.4, at a volume of 766 µl. The reaction was initiated by the addition of an NADPH-regenerating system (NADPH-RS) (0.4 mM NADP+, 0.85 mM glucose 6-phosphate, 0.1 U/ml glucose-6-phosphate dehydrogenase, and 0.85 mM MgCl2; BD Biosciences) and incubated for 60 min at 37°C in a final volume of 1 ml. The reaction was stopped by the addition of NaOH (1 ml, 0.5 N). Ethyl acetate (4 ml) was then added, and the mixture was shaken for 5 min and centrifuged at 1300g for 5 min. HCl (1 ml, 0.5 N) was added to an aliquot of the resulting aqueous phase (1.5 ml) and extracted with ethyl acetate as before. Fluorescence of this final organic phase was measured with a Perkin-Elmer spectrofluorimeter at 255 nm (excitation) and 458 nm (emission) wavelengths and 9-anthracene carboxylic acid levels were calibrated from a standard curve. Measurement of rCYP2C19 (2 pmol, 9.4 µg of protein) activity was performed identically to that of rCYP2B6 except that the preincubation was at room temperature, the 9-anthraldehyde concentration was 25 µM, NADPH-RS was approximately 3-fold higher in concentration, inhibitors were added at the start of the reaction, and incubation time was 45 min. Measurement of rCYP2C11 (0.5–2 pmol, 1.15–4.6 µg of protein) activity was performed identically to that of rCYP2B6, except that higher concentrations of substrate (9-anthraldehyde, 100 µM) and NADPH-RS (2-fold) were used, no preincubation was performed, and incubation time was 45 min (see Figs. 7 and 8). Recombinant CYP19A1 activity was quantified by a high-throughput screening method using dibenzylfluorescein (0.2 µM) as substrate, exactly as described by Stresser et al. (2000). Pilot P450 inhibitor studies (Table 2, not in brackets) were performed by commercial assays (MDS Pharma Services, Taipei, Taiwan; see www.mdsps.com).
|
|
|
|
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
) and zolantidine did not inhibit 3HCIM binding at H2-receptor relevant concentrations (IC50 values >30 µM, also not shown).
|
). To investigate whether the 3HCIM-binding site resembles a P450-like protein, the effects of the nonselective P450 inhibitors metyrapone and cyanide were determined on 3HCIM binding (Fig. 3). Both drugs produced concentration-dependent inhibition [pIC50 values of -4.68 ± 0.04 (IC50 = 20.8 µM) and -2.65 ± 0.05 (IC50 = 2.2 mM), for metyrapone and cyanide, respectively].
|
A series of selective inhibitors of various P450 isoforms were chosen to further characterize 3HCIM binding (see Table 1 for inhibitors, P450 targets, and Ki values). Figure 4 shows the effects of these compounds on the 3HCIM-binding site. Fluvoxamine (an inhibitor of CYP1A2), thioTEPA (CYP2B6), quercetin (CYP2C8), sulfaphenazole (CYP2C9), quinidine (CYP2D6), diethyldithiocarbamate (CYP2E1), and DL-aminoglutethimide (CYP19A1) all produced less than 50% inhibition of 3HCIM binding at 100 µM (Fig. 4), a concentration well above all of the reported P450-inhibitory potencies (Table 1). In contrast, ketoconazole (CYP3A4) and tranylcypromine (CYP2C19) both completely displaced 3HCIM binding with pIC50 values of 4.96 ± 0.15 (IC50 = 11.1 µM) and 4.72 ± 0.06 (IC50 = 18.9 µM) and Hill coefficients of -0.8 ± 0.2 and -0.8 ± 0.1, respectively (Fig. 4). The potency of ketoconazole as an inhibitor of 3HCIM binding (calculated Ki = 6.4 µM) (Table 1) is greater than 300-fold lower than the published affinity of this drug as a CYP3A4 inhibitor (Ki = 0.02 µM) (Table 1). Tranylcypromine, however, competed with 3HCIM binding at a concentration comparable with its potency as an inhibitor of CYP2C19 activity (Ki = 10.8 versus 8 µM, respectively) (Table 1). However, ticlopidine (another inhibitor of CYP2C19) had no effect on 3HCIM binding (Fig. 4).
|
|
High-Affinity Ligands for 3HCIM-Binding Proteins. The search for high-affinity ligands led to the study of the cimetidine analogs CC11 (Warrander et al., 1983) and CC12 (Hough et al., 2007) (Fig. 1). In agreement with earlier results (Warrander et al., 1983), CC11 produced complete, concentration-dependent inhibition of 3HCIM binding, yielding a pIC50 of 8.05 ± 0.03 (IC50 = 9.0 nM) and a Hill coefficient of -1.4 ± 0.1 (Fig. 5). CC12 was also shown to be a potent inhibitor of 3HCIM binding, producing a pIC50 of 7.78 ± 0.11 (IC50 = 16.5 nM) and a Hill coefficient of -1.1 ± 0.2 (Fig. 5).
|
Due to the unexpectedly wide range of activities of CC12 and increased availability of commercial P450 testing, CC12 was subjected to preliminary screening on seven additional P450 human isoforms (Table 2). A single concentration of CC12 (200 nM) inhibited the activities of these enzymes as well (33–100%). Thus, CC12 exhibited some affinity for all 12 P450 isoforms.
Because of the apparently high affinity of CC12 for CYP2B6, CYP2C19, and CYP19A1 in these preliminary studies, additional detailed experiments were performed to validate these findings. CC12 produced concentration-dependent inhibition of CYP2B6, CYP2C19, and CYP19A1 activities, yielding pIC50 values of 7.93 ± 0.08 (IC50 = 11.7 nM), 7.29 ± 0.09 (IC50 = 51.4 nM), and 6.85 ± 0.06 (IC50 = 140.7 nM), respectively (Fig. 6).
3HCIM-Binding Potential of Recombinant Candidate P450s. Although low concentrations of CC12 bind to the 3HCIM-binding site and also inhibit the activities of CYP2B6, CYP2C19, and CYP19A1, only the selective CYP2C19 inhibitor, tranylcypromine, competed with 3HCIM binding at relevant concentrations. Even though results with ticlopidine (another 2C19 inhibitor) (Table 1) failed to substantiate the relevance of CYP2C19, rat homologs of the human CYP2C19 were sought. A National Center for Biotechnology Information BLAST search of the human 2C19 amino acid sequence was performed against the rat genome. This BLAST search identified two rat proteins (CYP2C11 and CYP2C6) as homologs of human CYP2C19. Thus, experiments were performed to directly test the hypothesis that CYP2C11 or CYP2C6 accounts for the specific binding of 3HCIM in the rat brain.
When 3HCIM was incubated with sf9 microsomes containing rCYP2C11 (Fig. 7, left), no specific binding was measured. Companion studies were performed to ensure that the recombinant enzyme was viable and was properly filtered in the radioligand-binding assays (Fig. 7, right). To ensure that optimal binding conditions for the recombinant enzymes were met, variations in assay conditions, including adding higher enzyme amounts (4 pmol) or coincubating the enzyme (2–6 pmol/tube) with brain homogenate during radioligand-binding assays, were used and had no effect on binding activity. Furthermore, preincubation of the enzyme with 3HCIM and an NADPH-RS failed to produce a reliable specific radioligand binding signal. Additional studies similar to those in Fig. 7 showed that eight human P450s (2C19, 2A6, 2B6, 2C8, 2C9, 2C18, 3A5, and 19A1) and two other rat P450s (2C6 and 2B1, rat homologs of CYP2C19 and CYP2B6, respectively) did not exhibit specific binding of 3HCIM (data not shown, see Table 2).
A CYP2C8/9/18/19 antibody was also used to further evaluate the similarities between 3HCIM binding in the rat brain and CYP2C19. As shown (Fig. 8), antibody treatment caused concentration-dependent inhibition of rCYP2C11 activity but had no effect on rat brain 3HCIM binding.
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
), 52.2 nM (Rising et al., 1980), and 40 nM (Devoto et al., 1980)]. Additional KD values from 15.7 to 400 nM (as reviewed in Warrander et al., 1983) have been reported. As discussed by Warrander et al. (1983), variations in assay conditions (including incubation temperature and time, buffering conditions, concentrations of radioligand studied, type of drug added to evaluate nonspecific binding, and species) may contribute to this variability. Such variations in conditions may also affect the measurement of receptor density. The Bmax value currently reported (0.94 pmol/mg of protein) (Fig. 2) is consistent with several other laboratory results [0.68 (Rising et al., 1980) and 1.3 pmol/mg (calculated from Burkard, 1978)]. However, other reported values vary from 0.23 to 3.9 pmol/mg (Warrander et al., 1983). Compared with many receptors for biogenic amines in the brain, this density is high, yet higher levels have been found for the cannabinoid CB1 receptor in the brain (1.85 pmol/mg) (Devane et al., 1988). The high density of 3HCIM-binding sites (much greater than H2 receptor expression levels) (Gajtkowski et al., 1983) reinforces the conclusion that these sites are not the H2 receptor. The present findings, showing that pharmacologically active concentrations of ranitidine and zolantidine (both nonimidazole-containing H2 receptor antagonists) fail to inhibit 3HCIM binding, are similar to earlier results (Smith et al., 1980; Chansel et al., 1982; Gajtkowski et al., 1983; Warrander et al., 1983), further supporting the conclusion that the binding is not to the histamine H2 receptor. Furthermore, other histamine receptors or histamine-binding sites seem to be excluded by the finding that very high concentrations of histamine are required to inhibit 3HCIM binding (IC50 values were 0.8 mM in unpublished data from this laboratory and >1 mM from Smith et al., 1980).
Cimetidine is a well known inhibitor of several human P450 isoforms, although affinities are low (Ki values = 14–140 µM) (Table 1). These Ki values are much higher than the dissociation constant of cimetidine for the 3HCIM-binding site in the rat brain (KD = 67 nM) (Fig. 2). Despite this discrepancy, it seems possible that 3HCIM could bind to a P450 at concentrations lower than those required to inhibit enzyme function. The Ki values for cimetidine on human P450s are also much higher than therapeutic plasma levels of cimetidine, even though these levels can modify drug metabolism (Madeira et al., 2004). This finding suggests the possibility that cimetidine may bind to additional P450 isoforms.
The existence of several cimetidine-P450 interactions seems to support the hypothesis that the brain 3HCIM-binding site is a P450, even though this has never been tested. Inhibition of the binding by cyanide and metyrapone (Fig. 3) (discussed further below) is also consistent with (but not proof of) a P450-binding mechanism. Because of the lack of selectivity of these inhibitors and the multiplicity of P450 isoforms (currently 88 P450 isoforms are known in the rat) (http://drnelson.utmem.edu/CytochromeP450.html), two additional approaches were taken to identify candidate 3HCIM-binding P450 isoforms. The first approach was to study the effects of selective P450 inhibitors on 3HCIM binding (Fig. 4; Table 1). By initially focusing on P450 isoforms known to be inhibited by cimetidine (Table 1), it was predicted that if the 3HCIM-binding site resembles any of these P450 isoforms, then selective inhibitors of one or more of these enzymes should compete with 3HCIM binding at concentrations equivalent to (or less than) those reported to inhibit P450 activities. The findings that pharmacologically active concentrations of fluvoxamine, sulfaphenazole, quinidine, and diethyldithiocarbamate all failed to inhibit 3HCIM binding (Fig. 4; Table 1) suggest that the 3HCIM-binding site does not resemble human CYP1A2, 2C9, 2D6, or 2E1, respectively. Moreover, the binding site does not resemble CYP3A4, as the 3HCIM-binding constant of ketoconazole (Ki = 6.4 µM) (Fig. 4; Table 1) is much larger than the comparable value on CYP3A4 enzyme activity (Ki = 0.02 µM). In contrast, the affinity of tranylcypromine for the binding site (Ki = 10.8 µM) is virtually identical with its activity at CYP2C19 (Ki = 8 µM) (Fig. 4; Table 1), suggesting that the 3HCIM-binding site may resemble a CYP2C19-like protein. This similarity was not sustained by results with other CYP2C19 inhibitors, as ticlopidine was inactive on the binding site (Fig. 4; Table 1). Because ticlopidine has been suggested to be a mechanism-based suicide inhibitor of CYP2C19 (Ha-Duong et al., 2001), additional experiments (i.e., preincubation of brain membranes with ticlopidine at 37°C with NADPH-RS) ensured that this drug did not require metabolic activation to inhibit 3HCIM binding (data not shown).
Because pharmacological screening failed to clearly implicate a 3HCIM-binding candidate P450, other targets for tranylcypromine and ketoconazole were considered. Although these drugs are widely used as isoform-selective inhibitors (Chauret et al., 1997; Erickson et al., 1999), both have documented actions at additional P450 isoforms. In addition to its activity on CYP2C19, tranylcypromine has been reported to inhibit CYP2B6 and CYP19A1 (Stresser, 2004), but failure of thioTEPA (an inhibitor of 2B6) and DL-aminoglutethimide (an inhibitor of 19A1) to displace 3HCIM binding (Fig. 4; Table 1) indicates that the binding site does not resemble these enzymes. Tranylcypromine is also a potent inhibitor of CYP2A6 (IC50 = 0.05 µM) (Draper et al., 1997), but the binding site seems to not resemble this isoform because the affinity of tranylcypromine for 3HCIM binding is much lower (IC50 = 10.8 µM) (Table 1). Ketoconazole is a well known inhibitor of CYP3A4, but this drug has also been reported to inhibit CYP2C8 at concentrations (IC50 = 4.4 µM) (Stresser, 2004) similar to its observed affinity for 3HCIM binding (IC50 = 11.1 µM) (Fig. 4; Table 1), suggesting the possible relevance of this isoform. However, the inability of quercetin (another 2C8 inhibitor) to influence 3HCIM binding (Fig. 4; Table 1) seems to exclude a role for CYP2C8-like activity. Thus, despite these documented additional actions of tranylcypromine and ketoconazole, no clear 3HCIM-binding P450 candidate emerged from the pharmacological screening.
High-affinity ligands were synthesized and tested on 3HCIM-binding and P450 activities as a second approach toward identifying 3HCIM-binding proteins. Warrander et al. (1983) showed that the 3HCIM-binding site was not the H2 receptor by demonstrating a series of cimetidine analogs that competed with 3HCIM binding but lacked H2 receptor activity. One compound (herein named CC11) from that study had high affinity for the binding site (Ki = 5 nM) (Warrander et al., 1983), but the synthesis, rationale, and possible relevance to P450s were not given. CC11 was synthesized and tested in this study presently to confirm the high affinity of this drug for 3HCIM binding (Ki = 5.2 nM) (Fig. 5). Although not mentioned by Warrander et al. (1983), earlier literature reported CC11 to be a weak inhibitor of aldrin epoxidation (a P450-mediated reaction) in rat liver microsomes (Wilkinson et al., 1972).
CC12 (the para-iodo derivative of CC11) (Fig. 1) was described for the first time in 2007 (Hough et al., 2007), and the activity of this drug on P450 enzymes has not been previously studied. Originally synthesized to investigate the biological relevance of 3HCIM binding, the compound is a potent ligand for the binding site (Ki = 9.5 nM) (as in Fig. 5) and also behaves in vivo as an antagonist of several analgesic drugs including morphine and the cimetidine derivative improgan (Hough et al., 2007). The mechanisms accounting for this latter activity are not known (Hough et al., 2007). CC12 was found to be a moderately potent H3 antagonist (Ki = 50 nM), but the drug lacks activity in vitro on a variety of other central nervous system receptors (Hough et al., 2007). Chemically, CC12 belongs to the family of phenylimidazole P450 inhibitors that commonly interact with the heme moiety of P450, yielding a "type II" difference spectrum (Correia and Ortiz de Montellano, 2005). It is remarkable that within this family, the nature of the aromatic heteroatom and the structure of the side chain still permit a wide variety of potencies and selectivities for various isoforms. Drugs from this class include metyrapone, ketoconazole, and letrozole (Correia and Ortiz de Montellano, 2005). However, even when compared with many of these other compounds, CC12 is a very high-potency P450 inhibitor (on CYP2B6, IC50 = 11.7 nM, equivalent to Ki = 5.4 nM) (for Km value, see Stresser et al. 2000). Other compounds reported to act on CYP2B isoforms include miconazole (Ki = 50 nM) (Zhang et al., 2002), sulconazole (Ki = 40 nM) (Zhang et al., 2002), and 4-(4-chlorophenyl)imidazole (IC50 = 70 nM) (Spatzenegger et al., 2001). Additional studies are needed to assess the potency, selectivity, and utility of CC12 as a tool for P450 research.
Because CC12 has high affinity for the 3HCIM-binding site (Ki = 9.5 nM) (Fig. 5), this drug should be a potent inhibitor of any P450 isoform capable of 3HCIM binding. Although CC12 showed inhibitory activity on a total of 12 P450s (estimated IC50 values = 11.7–494 nM) (Table 2; Fig. 6), this drug showed high affinity for only three isoforms: CYP2B6, CYP2C19, and CYP19A1 (IC50 values = 11.7, 51.4, and 140.7 nM, respectively) (Fig. 6). The lack of 3HCIM-binding affinity by inhibitors of CYP2B6 and CYP19A1 (Fig. 4; Table 1) suggests little relevance of these isoforms to the binding site. However, two inhibitors of CYP2C19 (CC12 and tranylcypromine) also inhibited 3HCIM binding, whereas another inhibitor (ticlopidine) did not, suggesting the possible significance of CYP2C19-like proteins.
The lack of 3HCIM-binding activity in rCYP2C11 or rCYP2C6 indicates that these proteins (which are highly homologous to human CYP2C19) do not account for 3HCIM binding in the rat brain. Furthermore, the findings that rCYP2A6, 2B6, 2C8, 2C9, 2C18, 2C19, 3A5, and 19A1, as well as the 2B6 rat homolog, rCYP2B1, all failed to bind 3HCIM (Table 2) suggest that these enzymes do not resemble 3HCIM-binding proteins.
The conclusion that the rat brain 3HCIM-binding site does not resemble any of the P450s presently tested assumes that these methods would have detected 3HCIM binding to the rP450s. Calculations show that the amount of enzymes added to the binding assay (2 pmol) should have yielded a specific binding signal of 0.85 pmol (43% of the P450s should have bound according to a KD of 67 nM) (Fig. 2). Such a result would have been easily detectable, because it corresponds to a signal approximately twice that of the Bmax value for rat brain homogenate (Fig. 2). It was also necessary to show that the P450 proteins, supplied in insect microsomal fractions, could be filtered in the radioligand binding assay. This result was confirmed by the lack of enzyme activity in the filtrate of the binding assay (Fig. 7, right).
The conclusion that the brain binding site does not resemble any of the currently tested P450s could also depend on specific differences in rat brain and insect cell preparations. For example, the recombinant enzyme systems currently studied are coexpressed with high levels of cytochrome P450 reductase and cytochrome b5. If these proteins prevented 3HCIM binding, then coincubation of the insect microsomes with brain homogenates should have decreased 3HCIM binding in the homogenates. Alternatively, if the recombinant enzymes (in the presence of an NADPH-RS) did not contain the necessary components and/or cofactors to bind 3HCIM (as might be present in brain homogenates), then coincubation of the homogenate with the recombinant enzymes should have increased binding activity. Because neither of these results was obtained, it is concluded that differences in the brain and insect cell preparations are not relevant.
Other potential differences between the rat brain homogenate and the rP450s that might affect 3HCIM binding include post-translational modifications and potential allelic variations of P450s. Although the potential significance of these differences is difficult to assess, the present use of a CYP2C8/9/18/19 antibody known to inhibit CYP2C11 activity (Park et al., 1989) and that is immunoreactive with 2C6 (Waxman et al., 1987) provided another approach. The finding that this antibody (at concentrations that inhibited rCYP2C11 activity) (Fig. 8) did not alter 3HCIM-binding activity in rat brain fractions reinforces the conclusion that the 3HCIM-binding site is not CYP2C6 or CYP2C11.
The potency of cimetidine as a P450 inhibitor has been reported to be increased after preincubation of the drug with rat liver microsomes and an NADPH-RS (Levine et al., 1998). Under these conditions, it has been suggested that the higher potency of cimetidine can be explained by the formation of a metabolite-intermediate complex with P450 isoforms (Levine et al., 1998). If this hypothesis is correct, then one reason why 3HCIM binding to rP450s was not detected could be that conditions for forming this complex were present in rat brain homogenates but not in insect cell microsomes. Accordingly, mimicking these preincubation conditions should have resulted in an increase in 3HCIM binding of rat homogenates, thereby permitting the detection of specific 3HCIM binding to rP450s. The finding that the NADPH-RS preincubation had no effect on 3HCIM binding suggests that metabolic activation is not required for 3HCIM binding in either the brain or in the recombinant systems.
The present results show that the 3HCIM-binding site in rat brain is not CYP2C11, CYP2C6, or CYP2B1, but other P450 isoforms, and even non-P450 proteins remain to be tested. The present search for 3HCIM-binding P450s in the rat has depended heavily on testing the commercially available human isoforms. Rat P450s with high (>70%) homology to CYP2C19 (CYP2C6 and CYP2C11) lacked 3HCIM-binding activity. However, three additional rat P450s (CYP2C7, CYP2C24, and CYP2C79) also have high homologies to CYP2C19 (http://drnelson.utmem.edu/CytochromeP450.html) and remain viable candidates for 3HCIM-binding activity. None of these are commercially available, and one of these (CYP2C79) has only been described in silico. It is not known whether the CYP2C8/9/18/19 antibody used in the present study (Fig. 8) cross-reacts with these additional 2C isoforms. These isoforms remain to be cloned and tested for 3HCIM-binding activity.
Another limitation of the present study rests on the assumption that rat P450s with high homologies to a human P450 will share ligand-binding profiles. However, P450s are considered to be rather promiscuous, in that multiple enzymes from different families and subfamilies, can have overlapping substrate/inhibitor profiles (Rendic, 2002). For example, mutations of only three amino acid residues can alter CYP2C19 substrate selectivity to more closely resemble that of CYP2C9 (Jung et al., 1998). Thus, it is possible that the only homologies relevant to 3HCIM and CC12 binding are those in the substrate and inhibitor recognition sites. Subtle rat-human differences might also explain why the rat-binding protein shows 2C19-like activity with some, but not all, 2C19 inhibitors. Because crystal structures and homology models of several P450s with inhibitors have been described (Otyepka et al., 2007), it may be possible to use this approach in future searches for 3HCIM-binding P450s.
The classic demonstration of the existence of a P450 in tissues or of a drug-P450 interaction is the carbon monoxide (CO) difference spectrum, in which dithionite-reduced microsomes combine with CO to yield the characteristic 450-nm spectrum (Guengerich, 1991). Thus, the ideal experiment to demonstrate high-affinity binding of cimetidine to brain P450s would use cimetidine-induced alterations in the CO difference spectra from brain. Unfortunately, brain P450s are not in high enough abundance to yield a reproducible CO difference spectrum (X. Ding, personal communication).
If the 3HCIM-binding protein in brain is not a P450, then inhibition of this binding by cyanide (Fig. 3) suggests the importance of other heme-containing proteins. Cyanide is a well known inhibitor of P450s with affinities similar to values found in the present study (IC50 values = 1–2 mM) (Kitada et al., 1977). However, this reagent also inactivates other iron-containing (especially heme-containing) non-P450 proteins (Dixon and Webb, 1964). A query of the rat proteome (UniProt Knowledgebase) using the Integr8 Web Portal (http://www.ebi.ac.uk/integr8/EBI-Integr8-HomePage.do) with the gene ontology parameter "heme-binding" (GO:0020037) returns 115 unique heme-binding proteins. Approximately one-half of these are P450s. Notable among the remaining half are several electron transport proteins (e.g., cytochrome-related), oxygen transporters (hemoglobin, myoglobin, cytoglobin, and neuroglobin), and other oxidases and peroxidases. The list also includes specialized enzymes such as soluble guanylate cyclase and three forms of nitric-oxide synthase (NOS). Some of these hemoproteins can be excluded as potential 3HCIM-binding targets because they are soluble and would not have been present in the membrane fractions used in the present study. Two other metalloenzymes (endothelial NOS and neural NOS) are unlikely to bind 3HCIM because CC12 lacked activity on these enzymes (Hough et al., 2007). In contrast to the case with cyanide, a literature review of metyrapone shows that its experimental use has been limited to P450 inhibition (Ignarro et al., 1985; Masubuchi et al., 1998). Thus, inhibition of 3HCIM binding by metyrapone suggests that the binding site may be a rat P450 isoform that was not tested in the present study. Recently we used the high affinity of CO for heme in unpublished pilot studies and found that 3HCIM binding in brain homogenates was reduced by approximately 30% after CO treatment under reducing (dithionite) conditions. These findings, along with cyanide results, suggest the relevance of heme-containing proteins (P450 or non-P450).
The present studies suggest that the 3HCIM-binding site in brain is a heme-containing protein, but its identity as a P450 could not be confirmed. This protein needs to be identified for several reasons, including the biological activity of inhibitors of this binding (e.g., CC12), the pharmacological activities of cimetidine that are not mediated by the H2 receptor and the well known (but poorly understood) nature of cimetidine-P450 drug interactions. Furthermore, the affinity of cimetidine for this binding site is more than 10-fold higher than the affinity for any other known protein (including the H2 receptor) (Gajtkowski et al., 1983), arguing for its potential importance. Several approaches may be needed to identify this site, including cloning and testing of additional rat P450 isoforms and non-P450 hemoproteins, modeling ligand-P450 active sites or protein purification.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
ABBREVIATIONS: P450, cytochrome P450; 3HCIM, [3H]cimetidine; TEPA, N,N',N''-triethylene phosphoramide; CC11, 4(5)-(benzylthiomethyl)-1H-imidazole hydrochloride; CC12, 4(5)-((4-iodobenzyl)thiomethyl)-1H-imidazole; rP450, recombinant cytochrome P450; NADPH-RS, NADPH regenerating system; CO, carbon monoxide; NOS, nitric-oxide synthase.
Address correspondence to: Dr. Lindsay B. Hough, Center for Neuropharmacology and Neuroscience, Albany Medical College MC-136, 47 New Scotland Ave, Albany, NY 12208. E-mail: houghl{at}mail.amc.edu
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
, 3HCIM binding: resuspended 100,000g pellets (308 µg of protein) from rat brain were preincubated with the antibody (micrograms of IgM, ordinate) in 0.1 M Tris-HCl, pH 7.4, for 20 min at 37°C in a volume of 60 µl. After preincubation, 3HCIM, unlabeled cimetidine (to evaluate nonspecific binding), and buffer were added to a final volume of 100 µl and specific binding was measured as in 
, 2C11 activity: recombinant CYP2C11-containing sf9 microsomes (2 pmol, 4.6 µg of protein of P450) in 0.1 M potassium phosphate buffer, pH 7.4, were preincubated with antibody for a total of 20 min in a final volume of 60 µl. After preincubation, an NADPH-RS and buffer were added to a final volume of 1 ml and the 9-anthraldehyde oxidation assay commenced as described. Control (0 µg of IgM) CYP2C11 activity was 4.34 pmol/(min x mg of protein). For both data sets, data points represent the mean fractional inhibition of activity ± S.E.M. of triplicate determinations.
