Abstract
All current 3-hydroxy-3-methylglutaryl-CoA reductase inhibitors [simvastatin (SV), lovastatin (LV), atorvastatin, pravastatin, fluvastatin, and cerivastatin] are believed to undergo an atypical β-oxidation of the dihydroxy heptanoic or heptanoic acid side chain. Metabolites, which are shortened by two- and/or four-carbon units consistent with β-oxidation products, have been reported exclusively in rodents following LV and SV administration and across species (rodents, dogs, and humans) following the other statins. In this study, in vitro formation of a β-oxidation product of simvastatin hydroxy acid (SVA) and its intermediates in mouse livers is described. Incubation of SVA with mouse liver preparations fortified with CoASH and ATP led to formation of SV and two major products (P1 and P2). Based on mass spectrometry (MS), tandem mass spectrometry, and/or NMR spectral characteristics, P1 was an α,β-unsaturated metabolite, formed by dehydration of thed,d-dihydroxy heptanoic acid side chain, whereas P2 was probably the l,d-dihydroxy acid isomer of SVA, formed by stereospecific hydration of P1. When NAD+ was also included in the incubation mixture, there were two additional metabolites with the MS and/or NMR characteristics consistent with a two-carbon shortened product (P3) and its dehydrated derivative (P4). In a complete incubation system with all cofactors (ATP, CoASH, NAD+, and NADPH) present, there was an additional product with MS spectra and liquid chromatography retention time identical to the β-oxidized, unsubstituted pentanoic acid metabolite (P5) detected in rats and mice following simvastatin administration. The involvement of CoASH and NAD+ and the presence of the four metabolic intermediates suggest that SVA (and presumably the other statins) is a substrate for the β-oxidation enzyme complex in mice. Additionally, the present finding of CoASH-dependent formation of SV substantiates a mechanism proposed previously for the in vivo lactonization of statin hydroxy acids.
3-Hydroxy-3-methylglutaryl-CoA reductase inhibitors, or “statins”, are used widely for the treatment of hypercholesterolemia and hypertriglyceridemia (Mauro, 1993). Except for simvastatin (SV1) and lovastatin (LV), all current statins are administered as the pharmacologically active hydroxy acid form. SV and LV are inactive lactones, which upon conversion to their respective hydroxy acids (SVA and LVA), are potent competitive inhibitors of 3-hydroxy-3-methylglutaryl-CoA reductase (Duggan and Vickers, 1990). Most of the statins have been shown to undergo extensive metabolism in both animals and humans (Vickers et al., 1990a; Everett et al., 1991; Dain et al., 1993; Halpin et al., 1993; Cheng et al., 1994; Le Couteur et al., 1996; Prueksaritanont et al., 1997; Black et al., 1999). Biotransformation of the statins exhibits noticeable species differences and some qualitative similarities. In the case of SV and LV, esterase-dependent hydrolysis to SVA or LVA in plasma was very rapid in rodents but not in humans and dogs (Vickers et al. 1990b; Draganov et al., 2000). One of the biotransformation pathways reported for all statins is oxidation at the dihydroxy heptanoic or heptanoic acid side chain, a structural feature common to all statins. Pentanoic acid derivatives of SVA or LVA, corresponding to the loss of a two-carbon unit from the dihydroxy heptanoic acid side chain, have been reported to occur exclusively in rodents following SV or LV administration (Vickers et al., 1990b;Halpin et al., 1993). Metabolites shortened by two- or four-carbon units, resulting in pentanoic derivatives or propanoic products, respectively, have been observed in vivo for atorvastatin and pravastatin (both contain the same dihydroxy heptanoic side chain) primarily in rodents and minimally in dogs and in humans (Arai et al., 1988; Everett et al., 1991; Le Couteur et al., 1996; Black et al., 1998, 1999). Analogous metabolites also have been described in animals and humans for cerivastatin and fluvastatin, both of which contain the dihydroxy heptanoic acid moiety (Dain et al., 1993; Boberg et al., 1998). These metabolites have been referred to as β-oxidation products because they resemble those observed in the β-oxidation of fatty acids, which is characterized by stepwise oxidation of the carbon chain, two carbons for each cycle.
Mechanistically, the β-oxidation of fatty acids comprises CoASH-dependent activation of the carboxyl group, dehydrogenation followed by stereospecific hydration to give al-β-hydroxy derivative, NAD+-dependent dehydrogenation, and thiolytic cleavage of the α,β-bond (Nelson and Cox, 2000). Since all of the statins have a d-β-hydroxy configuration, an epimerization to the l-configuration is needed for the β-oxidation cycle to occur. Statins that form propanoic or propanoic acid metabolites (loss of the four-carbon unit) are believed to undergo two cycles of β-oxidation (Boberg et al., 1998). The mechanisms for the formation of unsubstituted pentanoic acid products of statins, including LVA or SVA, have been proposed to occur following a cycle of fatty acid oxidation yielding a d-β-hydroxy pentanoic acid derivative, followed by fatty acid biosynthetic processes (Halpin et al., 1993; Boberg et al., 1998). The biosynthesis requires thed-configuration of the hydroxyl moiety and involves dehydration of the remaining d-hydroxyl group, followed by hydrogenation to form the unsubstituted pentanoic acid metabolites (Halpin et al., 1993). Although theoretically conceivable, evidence supporting these proposals is lacking. To date, there have been no reports concerning CoASH and NAD+ involvement or demonstrating the presence of key anticipated intermediates leading to the chain shortened products of these statins. Therefore, the present studies were undertaken, using an in vitro approach and SVA as a model substrate, to illustrate CoASH and NAD+involvement in the β-oxidation process of statins, and to provide evidence for β-oxidation intermediates for the unsubstituted pentanoic acid derivative of SVA in mice.
Experimental Procedures
Materials.
SV, SVA, and [14C]SVA with specific activity of 50 μCi/μmol (Fig. 1) were synthesized at Merck Research Laboratories (Rahway, NJ). Triton X-100, CoASH, ATP, NAD+, and NADPH were obtained from Sigma (St. Louis, MO). All other reagents were of analytical or HPLC grade.
Proposed pathway for the β-oxidation of SVA.
∗, indicates the position of the 14C label.
Animals.
Ten male CF-1 mice (∼25–40 g) were obtained from Charles River Laboratories (Wilmington, MA). Following cervical dislocation, livers were quickly removed, weighed, and washed with 1.15% potassium chloride. The livers were homogenized in 4 volumes of ice-cold 0.05 M Tris buffer, pH 7.4, and 1.15% potassium chloride. Liver S2 was prepared following centrifugation of the homogenate at 2000gfor 15 min at 8°C to remove cell debris. Protein concentrations were measured by the method of Lowry et al. (1951).
In Vitro Metabolism of SVA.
A typical incubation mixture, in a final volume of 0.2 ml, contained 0.6 mg of liver S2 preincubated with 30 μl of 2% Triton X-100 for 15 min, 20 mM MgCl2, 6 mM ATP, 0.5 mM CoASH, and 0.05 M Tris buffer, pH 8.0. The preincubation step with Triton X-100 was found to be necessary for high-enzyme activity. Unless otherwise specified, the reaction was started by the addition of [14C]SVA (prepared by mixing [14C]SVA with nonradiolabeled SVA to achieve an isotopic ratio of 14C/12C close to 1:1) following a 3-min preincubation at 37°C, and the reaction was incubated for up to 60 min. Experiments also were conducted in the presence of the cofactors NAD+(2 mM) and/or NADPH (1 mM), in addition to CoASH and ATP. Control experiments included incubation mixtures with one or more components missing. The reaction was terminated at appropriate time intervals by the addition of 0.2 ml of acetonitrile (ACN). The samples were centrifuged, and the supernatants were analyzed immediately by HPLC or LC/MS, as described below.
For the purpose of isolation and purification of metabolites of SVA (P1 and P3), large scale incubations of SVA (60 μM; 20 × 1-ml incubations) were carried out with mouse liver S2 (3 mg/ml) and CoASH and ATP (for P1), or CoASH, ATP, and NAD+ (for P3) for 60 min. The metabolites were first isolated by using the HPLC conditions described below. The isolated fractions, after being dried in a lyophilizer, were subjected to further repurification, using a linear gradient from 15 to 65% ACN in water containing 0.1% formic acid. The purified metabolites were characterized by MS and NMR spectroscopy.
Analytical Procedures for SVA and Metabolites.
An HPLC method previously described for SV and its metabolites was used for quantification purposes, with minor modifications (Prueksaritanont et al., 1997). The samples were chromatographed on a Waters C18-Symmetry column (Milford, MA) (150 × 4.6 mm, 5 μm), preceded by a C8 guard column with a linear gradient of ACN and 5 mM formic acid (30–70% ACN in 25 min). The eluate was monitored by UV absorption spectra set at 240 nm and by an on-line IN/US β-ram radioactivity detector (IN/US Systems, Tampa, FL).
For the purpose of metabolite identification, the LC system consisted of an Hewlett Packard 1100 binary pump and a PerkinElmer 200 Autosampler (PerkinElmer Instruments, Norwalk, CT), maintained at 5°C, coupled to a Finnigan MAT LCQ ion trap mass spectrometer (Thermo Finnigan, San Jose, CA) and a Packard 500TR flow scintillation analyzer (Meriden, CT). HPLC separation was carried out either on an Eclipse XDB-C18 or Betasil C18 column (Keystone Scientific, Inc., Bellefonte, PA) (4.6 × 250 mm, 5 μm). The mobile phase consisted of 0.1% aqueous formic acid (solvent A) and ACN (solvent B) and was delivered at a constant flow rate of 1 ml/min. The initial mobile phase consisted of 16% B, which remained unchanged for 6 min, and increased linearly to 60% over 39 min, then to 90% over 5 min, and held for 10 min. Mass spectral analyses were performed using electrospray ionization in the negative ion mode. The electrospray ionization spray voltage was 2.5 kV, and the heated capillary temperature was 160°C. Metabolites were identified based on LC-radio-MS and/or LC-MS/MS analysis. Detection of metabolites was assisted by radioactive peak detection and the visual recognition of dual peaks, which resulted from ∼1:1 ratio of14C/12C of the [14C]SVA substrate in full scan MS spectra and molecular ion identification.
NMR spectra were obtained in CD3OD (99.96% deuterium content; Isotec Inc., Miamisburg, OH) at 3°C on a Varian Inova 500 MHz spectrometer (Varian Inc., Palo Alto, CA) equipped with a 3-mm inverse detection probe (MIDG-3; Nalorac Corp., Martinez, CA). 1H chemical shifts (in ppm) are relative to the solvent CD2HOD signal, which is set at 3.31 ppm.
Results and Discussion
Figure 2, A and B, illustrates typical HPLC chromatograms derived from incubates of mouse liver S2 with SVA in the presence of all of the cofactors (ATP, CoASH, NAD+, and NADPH). The formation of these metabolites (five major metabolites designated as P1 to P5 and few minor peaks) was dependent on the presence of both cofactors and liver protein. When incubations were carried out in the presence of only the cofactors ATP and CoASH, two major radioactive peaks (P1 and P2) and one minor peak were apparent. P2 exhibited a UV absorption spectrum similar to SVA, whereas the minor peak had a UV spectrum and HPLC retention time identical to SV. Formation of both P1 and P2 increased more or less in parallel with increased incubation time, whereas that of SV appeared to increase during the first 30 min (∼2% of initial concentration) and then to plateau (Fig.3A). Except for a small amount of SV (<0.6% of initial concentration), metabolites were not detectable in control incubations without either liver protein, ATP, or CoASH. The results suggested that formation of P1 and P2 was mediated by CoASH- and ATP-dependent enzyme(s), whereas SV was formed by both enzymatic (by CoASH- and ATP-dependent enzymes) and chemical processes. The latter provides evidence supporting a mechanism proposed previously for the lactonization of the hydroxy acid forms of statins observed in vivo (Duggan and Vickers, 1990).
Representative HPLC radioactivity (A) and UV (B) profiles of SVA metabolites in a mouse liver S2 incubate.
Incubations were carried out at 37°C for 40 min using mouse liver S2 (3 mg/ml) and [14C]SVA (50 μM) with 0.5 mM CoASH, 6 mM ATP, 2 mM NAD+, and 1 mM NADPH.
Formation of SVA metabolites as a function of time following incubation with mouse liver S2.
Incubations were carried out in duplicate at 37°C using mouse liver S2 (3 mg/ml) and [14C]SVA (50 μM) with 0.5 mM CoASH, and 6 mM ATP (A); 0.5 mM CoASH, 6 mM ATP, and 2 mM NAD+(B); and 0.5 mM CoASH, 6 mM ATP, 2 mM NAD+, and 1 mM NADPH (C).
The mass spectrum of P1 showed a prominent [M − H]− ion cluster at m/z417 and 419, 18 mass units less than that of the [14C]/[12C]SVA substrate mixture (m/z 435/437), suggesting possible dehydration of SVA. NMR studies of the isolated P1 confirmed that it was a dehydrated product of SVA with a double bond between the α- and β-positions relative to the carboxyl carbon of the dihydroxy heptanoic acid side chain (Fig. 1). The key features of the NMR spectrum of this metabolite were the appearance of two new signals at 5.88 ppm (doublet with J = 15.1 Hz) and 6.74 ppm (broad) (Table 1) and a cross peak between the two resonances in the two-dimentional total correlation spectroscopy spectrum (Summers et al., 1986). The chemical shifts and the scalar-coupling constant confirmed a double bond across carbons 2 and 3 (α and β to the carboxyl carbon, respectively) and that these protons were trans to each other (Fig. 1).
Identifiable 1H chemical shifts of the parent SVA and its metabolites P1 and P3 in CD3 OD at 3°C
The identification of P2 was based on MS, MS/MS, and UV absorption spectra. P2 showed a significant [M − H]− ion cluster at m/z435 and 437, similar to SVA (Fig. 4, A–D). The ion cluster at 481 and 483 detected in the MS spectrum of P2 (Fig. 4C) or SVA (Fig. 4A) corresponded to a formic acid adduct ofm/z 435 and 437. In the negative ionization mode, MS/MS spectra of SVA (Fig. 4B) and P2 (Fig. 4D) also were similar; the ion 435 mainly underwent neutral losses of 116 and 104, yielding two characteristic product ions of m/z 319 andm/z 215, respectively. The loss of 116 was attributed to a loss of the dimethyl-oxobutoxy moiety, whereas that of 104 corresponded to the loss of the β-hydroxy-butanoic acid side chain. P2 showed a UV absorption spectrum identical to SVA. Based on these observations, P2 was assigned as an isomer of SVA. Since SVA is ad-β-hydroxy isomer, P2 was probably al-β-hydroxy isomer of SVA (Fig. 1), arising from rehydration of the α,β-double bond of P1 (Fig. 1). This conclusion corresponds with the well characterized fatty acid oxidation reactions, which, after activation with CoASH, involve dehydrogenation at the α and β carbons with respect to the carboxy carbon to give atrans-Δ2-enoyl CoA, followed by stereospecific hydration of the unsaturated acyl CoA to give l-β-hydroxy isomer because the subsequent enzyme in the fatty acid oxidation cycle is specific for the l-β-hydroxyacyl CoA (Nelson and Cox, 2000).
LC-MS (A) and LC-MS/MS (B) spectra of SVA and LC-MS (C) and LC-MS/MS (D) spectra of P2.
When NAD+ was also included in the incubation mixture, there were two additional metabolites (P3 and P4). Formation of P1 and P2 in this mixture (Fig. 3B) was decreased significantly compared with the incubations with only CoASH and ATP present (Fig.3A). P3 and P4 were not observed in the absence of liver protein or in the presence of liver protein and NAD+, but without ATP and CoASH. These results suggested that P3 and P4 were NAD+-dependent, enzyme-mediated products of either P1 or P2. SV was barely detectable in this system. P3 showed a major [M − H]− cluster atm/z 391 and 393, corresponding to the loss of an acetyl group. The MS/MS spectrum showed an intense ion atm/z 275, suggesting that this metabolite contained the unchanged dimethyl-oxobutoxy moiety. NMR studies (Table1) supported that P3 was a substituted pentanoic acid analog of SVA (Fig. 1). 1H NMR spectrum of P3 indicated the presence of a new peak at 3.89 ppm. The following connectivity of 2.41—3.89—∼1.60—1.25, 1.41—1.72 ppm, corresponding to 4—5—6—7 protons was observed in the 2D total correlation spectroscopy spectrum, consistent with a loss of two carbons in the heptanoic acid side chain. The proposed pathway from P2 to P3 shown in Fig. 1 is in complete agreement with the next two steps in the fatty acid oxidation cycle: 1) an NAD+-dependent dehydrogenation of thel-β-hydroxyacyl-CoA to form a β-ketoacyl product; and 2) a thiolytic cleavage by thiolase resulting in a loss of acetyl-CoA.
Metabolite P4 was detected at m/z 373 and 375, 18 mass units less than P3, suggesting a possible loss of water. Unfortunately, due to the relatively low abundance of P4, the MS/MS and NMR spectra could not be obtained. Nevertheless, the proposal that P4 was a dehydrated product of P3 (Fig. 1) is substantiated by a subsequent finding that P5 was formed with a concomitant decline in P3 and P4 formation when NADPH was added to the incubation mixture (Fig.3C). P5 was detected at m/z 375 and 377, which was 2 mass units more than P4, consistent with hydrogenation of P4 proposed earlier (Halpin et al., 1993). This metabolite P5 had an MS and a retention time identical to the unsubstituted pentanoic acid derivative of SVA (structure confirmed by NMR) detected in mouse liver (Vickers et al., 1990b) and in rat plasma (data not shown) following simvastatin administration.
The proposed sequence leading to the formation of the unsubstituted pentanoic acid metabolite shown in Fig. 1 was based on the evidence presented here combined with the knowledge of the well characterized fatty acid oxidation and biosynthesis pathways. The findings of CoASH and NAD+ involvement, together with the SVA epimer formation, suggest that SVA is a substrate for the β-oxidation enzyme complex. The finding of the dehydrated product P1 from the d-d-dihydroxy SVA in the presence of CoASH and ATP suggests an additional step analogous to the dehydration in the fatty acid biosynthesis pathway before the oxidation process. Previously, only formation of P4, but not P1, was proposed to occur (Halpin et al., 1993). The dehydration of P3, giving rise to P4 and the hydrogenation of P4 to yield P5, was suggested earlier, although this was postulated without supporting evidences and was simply based on the knowledge of the biosynthesis pathways of fatty acids (Halpin et al., 1993). It is noteworthy that, although anticipated, we did not observe thioester conjugates of these SVA intermediates. The reason for this remains unclear, since attempts were made to minimize possible hydrolysis of thioester conjugates by rapid sample work-up and immediate sample analysis. Inclusion ofp-chloromercuribenzoic acid, a thiol blocking agent used to inhibit acyl-CoA hydrolase (Yamada et al., 1996), also did not result in detectable levels of the anticipated thioesters. Interestingly, consistent with the previous in vivo observations (Duggan and Vickers, 1990), species differences in the formation of these in vitro metabolites were also observed in our preliminary studies (i.e., the CoASH-dependent metabolites P1 and P2 were detectable only with liver preparations from mice and rats, but not dogs and humans).
Thus, the present in vitro study demonstrated, for the first time, the pathways for the formation of the pentanoic acid metabolite of SVA in mice. To our knowledge, this is also the first study to demonstrate the β-oxidation process of statins in vitro. Possibly, all of the statins, which share the common d,d-dihydroxy acid side chain, would form β-oxidation products by a similar sequence, at least initially, to the one proposed here for SVA. The present study also substantiates the proposal that the in vivo statin lactonization is mediated, at least in part, by CoASH-dependent enzyme(s).
Acknowledgments
We thank Drs. A. Jones and C. Raab, Greg Gatto, and Nathan Yu for synthesis and purification of [14C]SVA and J. Brunner for mouse liver preparations.
Footnotes
- Abbreviations used are::
- SV
- simvastatin
- LV
- lovastatin
- SVA
- hydroxy acid form of simvastatin
- LVA
- hydroxy acid form of lovastatin
- HPLC
- high-pressure liquid chromatography
- ACN
- acetonitrile
- LC
- liquid chromatography
- MS
- mass spectrometry
- MS/MS
- tandem mass spectrometry
- Received March 5, 2001.
- Accepted June 14, 2001.
- The American Society for Pharmacology and Experimental Therapeutics