![[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}
|
|
|
|
|
||
Preclinical Candidate Optimization (D.Z., V.R., R.D., M.Z., L.K., J.M.), Discovery Chemistry (M.O., J.K.R.), Analytical Research and Development, Pharmaceutical Research Institute (R.G.), Bristol-Myers Squibb, Princeton, New Jersey
(Received December 16, 2002; accepted March 20, 2003)
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
-amino group of a lysine
residue. Microsomal incubations of [14C]BMS-204352 resulted in low
levels of covalent binding of radioactivity to proteins. This in vitro
covalent binding required cytochrome P450-reductase cofactor NADPH and was
attenuated by glutathione. P4503A inhibitors ketoconazole and troleadomycin
selectively prevented the covalent binding in vitro. Based on these
observations, a two-step bioactivation process for the protein covalent
binding of BMS-204352 was postulated: 1) P4503A-mediated
O-demethylation leading to spontaneous release of HF and the
formation of an ortho-quinone methide reactive metabolite and 2)
nucleophilic addition of the -amino group of protein lysine residue(s)
in protein to form des-fluoro des-O-methyl BMS-204352 lysine
adduct.
). This compound has potential to prevent and treat ischemic
stroke. Stroke is a major cause of death and long-term disability, affecting
more than 700,000 people in the US annually
(Williams et al., 1999). Acute
ischemic stroke is the most common form, producing pathologically fatal levels
of intracellular calcium (Ca2+) in neurons at risk.
Maxi-K channels are large-conductance voltage- and
Ca2+-activated K+ channel proteins
(Chang et al., 1997). The
fluoro-oxyindole BMS-204352 is a potent maxi-K channel opener and provides
significant levels of cortical neuroprotection in rat models of stroke by
augmenting an endogenous mechanism for regulating Ca2+
entry and membrane potential to protect the neurons
(Cheney et al., 2001;
Gribkoff et al., 2001).
|
Positron emission tomography imaging showed that F-18-labeled BMS-204352 is
rapidly and widely distributed in rats; the brain uptake was rapid and
occurred at high levels (Kiesewetter et
al., 2002). The compound had a brain/plasma exposure ratio of 9:1
following an intravenous bolus dose of 5 mg/kg in rats
(Hewawasam et al., 2002).
BMS-204352 was 99.5% protein bound in rat serum
(Krishna et al., 2001).
Following an intraarterial infusion or oral dose of [14C]BMS-204352
(6 mg/kg) (Krishna et al.,
2002), the elimination of total radioactivity in plasma was very
slow compared with that of the parent compound. Radioactivity was primarily
excreted in the feces (more than 85% of administered dose over a 7-day
collection period). The comparable recoveries of the radioactive dose in urine
(5.9% for intraarterial and 4.5% for oral dose) suggested that the oral
absorption of BMS-204352 was nearly complete. However, the absolute oral
bioavailability of BMS-204352 was 55% in rats. Collectively, these data
suggested that a C-14-derived radioactive material in plasma had a much longer
t1/2 than the parent compound and that the compound
exhibited the first-pass metabolism in rats.
Within 1 h following i.v. dosing of [14C]BMS-204352 to rats, more than 85% of the radioactivity in plasma was found to be unextractable by organic solvents. In this study, we characterized the covalently bound radioactive material by acid hydrolysis of rat plasma proteins, HPLC isolation, analysis by LC/MS and NMR, and organic synthesis of a metabolite analog. We also propose a P450-catalyzed bioactivation mechanism for the covalent binding through a novel ortho-quinone methide intermediate.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
).
des-Fluoro BMS-204352 lysine adduct (BMS-349821, 98% pure) was synthesized as
described below. The structure of BMS-204352 is shown in
Fig. 1. Ammonium acetate and
polyethylene glycol-300 were purchased from Aldrich Chemical Co. (Milwaukee,
WI). Sodium hydroxide, hydrochloric acid (36.5–38%), sodium phosphate
(dibasic), silica gel TLC plates (250 nm), and silica gel were purchased from
EM Scientific (Gibbstown, NJ). N-
-Boc-L-lysine
methyl ester was obtained from Bachem Biosciences (King of Prussia, PA). P450
inhibitors and pooled microsomes were purchased from BD Gentest (Woburn, MA).
Ecolite liquid scintillation cocktail was purchased from ICN Biomedicals, Inc.
(Costa Mesa, CA). For solid phase extraction, Oasis HLB C18
cartridge columns (1 and 10 ml) were obtained from Waters Corporation
(Milford, MA). Acetonitrile-d3 (D 99.8%) and
DMSO-d6 were obtained from Cambridge Isotope Laboratories,
Inc. (Andover, MA). All other chemicals used were of reagent grade or
better.
Synthesis of des-Fluoro BMS-204352 Lysine Adduct BMS-349821.
BMS-225113, a chloro-analog of BMS-204352, was synthesized as described
previously (Kiesewetter et al.,
2002) and used as the stating material. des-Fluoro BMS-204352
lysine adduct was synthesized in three steps as shown in
Fig. 2 and described below.
|
des-Fluoro BMS-204352 Boc-L-lysine methyl ester. To a solution
of BMS-225113 (221.3 mg, 0.59 mmol) dissolved in CH3CN (3 ml) was
added N--Boc-L-lysine methyl ester (274.3 mg, 0.61
mmol) in CH3CN (2 ml) followed by addition of solid
K2CO3 (197.8 mg, 1.41 mmol). After the reaction was
stirred at 90°C under a blanket of argon for 1 h, TLC (ethyl
acetate/hexanes 3:7) showed the starting material BMS-225113 was consumed. The
reaction was cooled to room temperature, filtered, and the solid washed with 2
x 2 ml CH3CN. The filtrate was evaporated under vacuum to
afford 380.5 mg (yield 108%) of crude solid, which was used without further
purification. TLC (ethyl acetate/hexanes 3:7): one spot at
Rf 0.1 by UV.
des-Fluoro BMS-204352 Boc-L-lysine adduct. To a solution of the above methyl ester (370 mg) dissolved in tetrahydrofuran (5 ml) was added 1 M LiOH (2.5 ml, 2.5 mmol). The resulting solution was stirred under argon at 60°C. After 1 h, TLC (CHCl3/MeOH/HOAc 8:2:0.1) showed the starting ester was consumed. The reaction was cooled to room temperature, and CH2Cl2 (20 ml) was added. The solution was extracted with 2 M HCl (10 ml), 2 x 10 ml H2O, and brine (10 ml). The organic phase was dried over MgSO4, filtered and evaporated under vacuum to afford 331.2 mg (yield 96%) of the product. TLC (CHCl3/MeOH/HOAc, 8:2:0.1): one spot at Rf 0.6 by UV. MS: 586.2 (M + H)+; 584.2 (M – H)–.
des-Fluoro BMS-204352 Lysine Adduct (BMS-349821). A solution of the Boc-protected adduct (331.2 mg, 0.57 mmol) was dissolved in trifluoroacetic acid (5 ml) and stirred at room temperature. After 1 h, TLC (CHCl3/MeOH/HOAc, 8:2:0.1) showed disappearance of the starting material. The solvent was evaporated under a stream of nitrogen. The resulting residue was evaporated twice from CHCl3/MeOH (9:1) (10 ml), and dried under high vaccuum to afford 385.5 mg (96%, trifluoroacetic acid salt) of a foam. The crude product was chromatographed on silica (30 g) eluting with CHCl3/MeOH/HOAc (74:25:1). Fractions containing the desired product were combined and evaporated under vacuum to afford 252.8 mg of a residue. This residue was dissolved in MeOH (5 ml), boiled with activated carbon (200 mg), and filtered through a bed of celite. The celite was washed with MeOH (30 ml), and the combined filtrate was evaporated under vacuum to afford a glass-like material. This solid was evaporated from Et2O (10 ml) to afford 233.5 mg (69%) of the product as a solid. TLC (CHCl3/MeOH/HOAc, 8:2:0.1): one spot at Rf 0.1 by UV. MS: 486.2 (M + H)+; 484.1 (M – H)–. 1H NMR (d6-DMSO) 10.77 (br, 1H, H1), 7.91 (d, 1H, J = 3 Hz, H23), 7.33 (dd, 1H, J = 8.5, 3 Hz, H21), 7.20 (dd, 1H, J = 7.5, 1.5 Hz, H8), 7.09 (d, 1H, J = 1.5 Hz, H6), 7.01 (d, 1H, J = 7.5 Hz, H9), 6.93 (d, 1H, J = 8.5 Hz, H20), 3.70 (br t, 1H, J = 6.0 Hz, H16), 3.44 (s, 3H, H25), 2.35 (m, 1H, H12a), 2.04 (m, 1H, H12b), 1.71–1.67 (m, 2H), 1.41 (m, 2H), 1.37–1.30 (m, 2H); 13C NMR (d6-DMSO) 179.21 (C2), 171.88 (C17), 155.80 (C19), 144.68 (C5), 136.17 (C4), 132.08 (C18), 129.73 (C7), 129.15 (C21), 128.21 (C23), 125.20 (C9), 125.23 (C22), 124.71 (C24), 119.05 (C8), 114.72 (C20), 105.78 (C6), 66.12 (C3), 56.36 (C25), 52.76 (C16), 42.13 (C12), 30.44 (C15), 29.69 (C13), 22.68 (C14); 19F NMR (d6-DMSO) –62.
HPLC. HPLC was performed on a Shimadzu Class VP system equipped with two pumps (model LC-10AT; Shimadzu, Kyoto, Japan), an autoinjector (SIL 10AD), and a diode array detector (SPD-M10A). A Zorbax RX C18 column (4.6 mm x 250 mm, 5 µm) equipped with a guard column was used. All HPLC analyses were performed by enclosing the column in an Eppendorf CH-30 column heater maintained at 35°C. The mobile phase flow rate was 1 ml/min. The retention time of the parent compound was confirmed before each run by radioactivity or UV detection. A gradient of two solvent systems, A and B, was used. Solvent A was water containing 15 mM ammonium acetate, pH 4.3. Solvent B consisted of acetonitrile/solvent A (95:5, v/v). Solvent B was started at 25% and then increased in a linear manner as follows: 50% (35 min), 80% (40 min), and then maintained at 80% for 2 min before it was decreased to 25% in 3 min.
Radioactivity Detection. Radioactivity in biological samples was
determined by mixing aliquots of the samples with 15 ml of Ecolite cocktail
and counting with a Packard Tri-Carb 2250 liquid scintillation analyzer
(PerkinElmer Life Sciences, Boston, MA). An on-line ß-Ram radioactivity
detector (0.5-ml liquid cell; IN/US Systems Inc., Tampa, FL) was also used for
HPLC radioactivity profiling. In addition, HPLC effluent was collected at 25-s
intervals after sample injection into 96-deep-well Lumaplates with a Gilson
model 202 fraction collector (Gilson Medical Electronics, Middleton, WI). The
effluent in the plates was dried with a Speed-Vac (Savant Instruments,
Holbrook, NY), and the plates were counted for 10 min per well with a TopCount
scintillation analyzer (PerkinElmer Life Sciences)
(Zhu et al., 2000).
LC/MS Analysis. LC/MS analysis was performed on a Finnigan LCQ mass spectrometer with an electrospray ionization source Thermo Finnigan MAT (San Jose, CA). Samples were analyzed in the negative ion mode. Half of the HPLC eluent was directed to the Finnigan LCQ mass spectrometer through a divert valve set to divert the flow to waste from 0 to 10 min. From 10 min until the end of the HPLC run, the eluent flow was directed to the mass spectrometer. The capillary temperature used for analysis was between 220 and 230°C. The nitrogen gas flow rate, spray current, and voltages were adjusted to give maximum sensitivity. The other half of the HPLC eluent was directed to either the on-line detector or a fraction collector. The collision energy was 20 to 35%.
Exact mass measurements of the parent and product ions were obtained on a Micromass Q-TOF Ultima mass spectrometer that was equipped with lock-spray under positive ion mode. The m/z 556.2771 and 278.1141 of an infused 20 ng/µl leucine enkephalin solution were used as lock masses. The Q-Tof was tuned to 18,000 resolution at half peak height using an insulin tuning solution. The experimentally obtained masses of product ions (326.0199 and 147.1138 Da) agreed to their respective calculated values with an error less than 1 mDa.
NMR. NMR experiments were performed on a Bruker 600 system (Newark, DE) equipped with a 3-mm probe and operating at 599.87 MHz. 1D 1H, 2D correlation spectroscopy (proton-proton coupling), 2D 1H-13C heteronuclear multiple quantum coherence spectroscopy and heteronuclear multiple-bond correlation spectroscopy (long range nuclear coupling, and bond-bond coupling), 13C, and 19F NMR experiments were performed. All chemical shifts are reported in ppm relative to tetramethylsilane or trichlorofluoromethane.
Sample Collections. After male Sprague-Dawley rats were administered single 2 mg/kg intravenous doses of [14C]BMS-204352 (21.5 µCi/mg), plasma samples (0.25, 1, 4, 8, 24, 72, and 168 h) were obtained from three rats for each time point. For isolation of a large quantity of the metabolite, plasma was collected from nine rats 6 h after single 5 mg/kg intravenous doses of [14C]BMS-204352 (9.5 µCi/mg). Plasma samples were prepared in tubes containing EDTA and centrifugation (10 min, 1300g at 4°C).
Extraction of Radioactivity from Plasma. Pooled plasma samples (0.2 ml) were mixed with 1 ml of methanol. After vortexing, the protein precipitates were pelleted by centrifugation at 3500 rpm in an IEC Centra-7R refrigerated centrifuge (IEC, Needham Heights, MA) or 15 min at 10°C. Similarly, separate portions of plasma were extracted with methanol after mixing with potassium hydroxide, potassium carbonate, and urea at 0.1, 0.5, and 6 M, respectively. Portions (0.2 ml) of the plasma samples were also extracted with 0.5-ml size C18 cartridges, which were eluted with 1 ml of methanol, 2 ml of water, and then 1 ml of methanol. Control plasma samples were made by spiking [14C]BMS-204352 in rat plasma (equivalent to the radioactivity level in rat plasma) and extracted with methanol or C18 cartridges. The control plasma samples were also incubated at 37°C for 0, 1, 3, and 5 h before extraction. Extraction with methanol, acetonitrile, or ethyl acetate was repeated four times.
Electrophoresis of Radioactivity in Plasma Proteins. Proteins from 0.25- to 72-h plasma were precipitated with methanol and washed twice with methanol. Portions of the plasma proteins (0.3 mg) were dissolved by boiling in 20 µl of an electrophoresis sample buffer containing (w/v) 4% SDS, 20% glycerol, 10% ß-mercaptoethanol, and 0.1% bromophenol blue in 125 mM Tris buffer (pH 6.0) and loaded on a 10% polyacrylamide stacking gel. The gel was run for 70 min at 160 V on a Bio-Rad Mini-Protean II system (Bio-Rad, Hercules, CA), stained with Comassie blue R solution for 1 h, and destained in water/methanol/acetic acid (65:25:10, v/v/v) for 3 h. After drying onto a piece of filter paper, the gel was scanned for 24 h on a Packard Electronic Autoradiography, Instantimager.
Dialysis of Radioactivity in Plasma Proteins. A portion of the methanol-precipitated and washed 1-h plasma proteins (30 mg) was dissolved in 2 ml of 4 M urea. Control plasma was spiked with an equivalent level of [14C]BMS-204352 and dissolved in 4 M urea. The plasma protein and control samples were dialyzed in membrane tubes with a molecular mass cutoff of 3,500 Da. The membrane tubes were purchased from American Scientific Products (McGaw Park, IL). The samples were dialyzed in 1 liter of 35 mM phosphate buffer containing 4 M urea, pH 7.4, for 24 h followed by additional 40 h in an additional 2 liters of buffer. Aliquots of samples from the dialysis bags were taken for determination of radioactivity.
Acid Hydrolysis of Rat Plasma Proteins for Metabolite Isolation. Plasma was pooled from nine rats 6 h after a single 5 mg/kg intravenous dose of [14C]BMS-204352 (9.5 µCi/mg). The 6-h time point was selected to obtain maximal concentrations of protein bound radioactivity. A 28-ml plasma sample was extracted with 60 ml of methanol and washed twice with 60 ml of methanol. The protein pellet was dissolved in 31 ml of 6 M HCl. The mixture was incubated in a 110°C oil bath for 24 h in a sealed tube. Potassium phosphate was added to the hydrolysate to a final concentration of 200 mM and the pH of the solution was adjusted to 5.0 with 12 M NaOH. The solution was passed through four 1-ml C18 cartridges. The radioactive material was eluted with methanol and concentrated under a stream of nitrogen. The concentrated sample was separated by HPLC on a semipreparative Zorbax C18, 9.4 x 250 mm 5 µ column (10 injections). The effluent was collected as fractions. A 50 µl portion of each fraction was counted for radioactivity. The fractions that corresponded to the major radioactive peak were pooled and repurified on a Supelco ABZ+, 4.6 x 250 mm column (6 injections). For the semipreparative Zorbax column, solvent B was held at 1% for 2 min, increased to 99% over 31 min, then held at 99% for 5 min at 4 ml/min. For the Supelco ABZ column, solvent B was held at 30% for 2 min and then increased to 50% over 20 min at 1 ml/min. The isolated sample was analyzed by MS and NMR.
Acid Hydrolysis of BMS-349821 and Rat Plasma Proteins for Comparison. Pooled (1, 4, 12 h) plasma was extracted by addition of 4 ml of acetonitrile to 2 ml of plasma, vortexed for 1 min, and centrifuged at 3000 rpm for 10 min. The extraction was repeated three times. The protein residues were dried under a nitrogen stream in a 15-ml pyrex tube. HCl (6 N, 8 ml) was added to the protein pellet. The solution was refluxed for 24 h at 110°C. Two separate hydrolysis controls were incubations of [14C]BMS-204352 (50 µg, 9.5 µCi/mg) and BMS-349821 (1 mg, the des-fluoro lysine adduct) under similar conditions with 6 N HCl in the presence of proteins from 6 ml of control plasma. The hydrolysate was neutralized by adding a 6 N NaOH solution containing 1 M sodium phosphate to adjust to pH 7.5 and passed through a 10 cc C18 cartridge. The radioactive material was eluted with methanol and concentrated under a stream of nitrogen. The residue was redissolved in 250 µl of 70% acetonitrile in 15 mM ammonium acetate and centrifuged at 3000 rpm for 10 min before injecting onto the HPLC column. The radioactive fractions were pooled, evaporated, dissolved in 70% acetonitrile, and analyzed by LCQ with negative electrospray ionization.
In Vitro Experiments. The cofactor requirements, the influence of GSH, and the effect of P450 inhibitors on covalent binding were evaluated using human liver microsomes. [14C]BMS-204352 (25 µM) was incubated with microsomal proteins (2 mg/ml), NADPH (1.25 mM), and bovine serum albumin (8 mg/ml) in 0.5 ml of 60 mM phosphate buffer, pH 7.4. The incubations were done in 15-ml glass test tubes at 37°C in a shaking water bath. Each assay was started by addition of the microsomes followed by incubation for 2 to 5 h at 37°C. A control incubation without NADPH was done with each set of experiments. In some experiments, GSH (5 mM) was also included in the incubation mixtures for attenuating covalent binding activity. In the inhibition experiments, inhibitors (2.5–5 µl, dissolved in methanol or acetonitrile) were added to 0.5 ml of incubation mixtures containing 1.4 mg of microsomal proteins and 25 µM [14C]BMS-204352 in 60 mM phosphate buffer, pH 7.4. Suicide inhibitors furafylline, orphenadrine, troleadomycin, and 1-aminobenzotrazole were preincubated with microsomes in the presence of NADPH for 10 min before the substrate was added. The reactions were initiated by adding the substrate, continued for 60 min at 37°C, and terminated by protein precipitation with methanol. After centrifugation, the protein pellet was saved, thoroughly washed with methanol, acetonitrile, or ethyl acetate for three times, digested with 1 M KOH for 24 h, and neutralized with 1M HCl, and analyzed for radioactivity by scintillation counting.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
), the percentage of the radioactive dose covalently bound to
plasma proteins was calculated to be about 5.8% at 4 h after i.v. dosing in
rats. In contrast to these ex vivo samples, however, control rat plasma
incubated with [14C]BMS-204352 gave quantitative recovery of
radioactivity from extraction by methanol. The spiked plasma samples showed
high percentages (87–99%) of extractable radioactivity even after
incubation for up to 5 h at 37°C.
|
Figure 3 shows an image of the radioactivity bound to plasma proteins of rats dosed with [14C]BMS-204352 after the proteins were separated on an SDS-polyacrylamide gel by an electric field. About 80% of the radioactivity comigrated with albumin at 68 kDa and 20% comigrated with other protein components.
|
Figure 4 shows the results of dialysis of unextractable radioactivity in plasma proteins from rats dosed with [14C]BMS-204352. In this experiment, substances with molecular mass of less than 3,500 Da would diffuse out of the dialysis membrane bag. After 24 and 64 h dialysis, respectively, 93 and 95% of the radioactivity in methanol-washed plasma proteins remained inside the dialysis bags, compared with 42 and 12% of the radioactivity for spiked rat plasma proteins.
|
Isolation and Structural Confirmation of des-Fluoro
Des-O-Methyl BMS-204352 Lysine Adduct. From the acid hydrolysis,
83.5% of the radioactivity in plasma proteins was recovered after initial
C18 cartridge clean-up. HPLC analysis showed that more than 90% of
the radioactivity was eluted in one radioactive peak. After a two-step
purification procedure, about 24 µg of this major radioactive component was
isolated from acid hydrolysis of plasma proteins from the nine rats dosed with
[14C]BMS-204352. LC/MS/MS analysis showed that this isolate
contained a single compound with a molecular ion at [M –
1]– of 470 (in negative mode). MS/MS analysis gave a major
fragmentation ion at m/z 324, suggesting neutral loss of
lysine (146 Da). High resolution MS analyses (in positive mode) showed a
molecular ion [M + 1]+ at m/z 472.1262
(–0.0029 Da difference from calculated 472.1291 of
C21H22N3O4F3Cl). High
resolution MS/MS (Fig. 5)
showed fragment ions at m/z 326.0199 (+0.0003 Da difference
from calculated 326.0196 of
C15H8NO2F3Cl) and at
m/z 147.1138 (+0.0004 Da difference from calculated 147.1134
of C6H15N2O2, lysine).
One-dimensional 1H NMR of the isolate in CD3OD showed
six aromatic protons. In DMSO, the isolate showed a broad peak (1H)
at 5.3 ppm (a -methine proton of an amino acid). Two-dimensional
(correlation spectroscopy) 1H NMR supported a lysine structure.
Proton signals were observed at 2.0 (2H), 2.1 (2H), 3.2
(2H), and 5.3 ppm (1H) in the region of 2.0 to 5.5 ppm. The
couplings between protons at 3.2 ppm (a CH2 next to amine nitrogen)
and at 2.1 ppm (a CH2 in the chain), between protons at 2.1 and 2.0
ppm, and between protons at 2.0 ppm and at 5.3 ppm (a -methine proton)
were observed. The isolate had 
max values at 220, 250, and
293 nm in its UV spectrum, similar to those of the parent drug. Fluorine NMR
of the isolate showed loss of the fluorine from the ring carbon (–160
ppm to –180 ppm versus CFCl3) but retention of the three
fluorine atoms in the trifluoromethyl group (–83 ppm and –70 ppm
versus CFCl3) of BMS-204352. Based on these MS, 1H, and
19F NMR results, the 24-h acid hydrolysis product of the
BMS-204352-rat plasma protein adduct was identified as the lysine conjugate of
desfluoro des-O-methyl BMS-204352.
|
Figure 6 shows comparative HPLC-UV or radiochromatographic profiles of the acid hydrolysis of the synthetic lysine adduct BMS-349821 (top panel), of [14C]BMS-204352 (middle panel), and of plasma proteins from rats dosed with [14C]BMS-204352 (bottom panel). Acid hydrolysis of BMS-349821 (top panel) produced a major peak 1 (eluted at 19.6 min) and a minor peak 2 (eluted at 21.5 min). LC/MS analysis showed that the peak 1 had a molecular ion [M – 1]– at m/z 484, consistent with the intact structure of BMS-349821 and that the minor peak had a molecular ion [M – 1]– at m/z 470 and a prominent MS2 fragment ion at m/z 324 (loss of lysine), suggesting a structure of demethylated product of BMS-349821, namely, desfluoro des-O-methyl BMS-204352 lysine adduct. Acid hydrolysis of [14C]BMS-204352 (middle panel) produced one major peak 3 at 42 min with complete disappearance of BMS-204352 at a retention time of 43.5 min. LC/MS analysis showed that the peak 3 at 42 min had a molecular ion at m/z 342 and major MS2/MS3 fragment ions at m/z 325, 314, and 298, which was consistent with des-fluoro des-O-methyl hydroxy BMS-204352 (with an F-replaced by -OH). Acid hydrolysis of plasma protein of rats dosed with [14C] BMS-204352 produced a single prominent radioactive peak 2 at 21.5 min (Fig. 6, bottom panel). No other significant radioactive peaks were observed from the acid hydrolysis of rat plasma protein. LC/MS analysis showed that this peak 2 at 21.5 min also had a molecular ion at m/z 470 and a prominent MS2 fragment ion at m/z 324, suggesting a common structure to the minor acid hydrolysis product of BMS-349821. Based on the retention time, LC/MS properties, and the NMR data of rat plasma isolate, the peak 2 was identified as des-fluoro des-O-methyl BMS-204352 lysine adduct.
|
P4503A-Mediated Protein Covalent Binding in Vitro. Table 2 shows protein bound (unextractable) radioactivity from incubations of [14C]BMS-204352 with human liver microsomes in the presence and absence of NADPH. Radioactivity covalently bound to proteins in microsomal incubations was determined by extensive extractions with methanol and acetonitrile of unbound label from microsomal incubations followed by scintillation counting of residual C-14. The percentage of unextractable radioactivity in the incubation mixtures increased 14-fold (2 h) to 25-fold (5 h) compared with incubations run for the same period without NADPH. GSH only attenuated the protein covalent binding by about 30%. Table 3 shows the effect of various P450 inhibitors on protein covalent binding of radioactivity for in vitro incubations of [14C]BMS-204352 with human liver microsomes. Aminobenzotriazole, a nonselective P450 suicide inhibitor, reduced binding to near background levels. Selective inhibitors of P450 isozymes 1A2, 2A6, 2C19, 2C9, 2D6 slightly inhibited the protein covalent binding. Although 2B6 inhibitor orphenadrine and 2E1 inhibitor 4-methylpyrazole caused significant inhibitions, the isozyme 3A inhibitors, ketoconazole and troleadomycin, almost completely inhibited protein covalent binding. In summary, the covalent binding of radioactivity in vitro required NADPH and was selectively inhibited by P4503A inhibitors, suggesting a bioactivation mechanism by P4503A. A mechanism for covalent binding of BMS-204352 is proposed in Fig. 7.
|
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Since the structure of the BMS-204352 covalently bound protein adduct could
not be determined directly, the covalently bound protein sample was treated
with acid to hydrolyze the amino acid backbone to its component amino acids.
The hydrolysis product was identified as a lysine conjugate of des-fluoro
des-O-methyl BMS-204352. Hence, the BMS-204352-protein adduct simply
represents the net replacement of the ring fluorine atom of
des-O-methyl BMS-204352 with the -amino group of a lysine
residue. Although BMS-204352 was hydrolyzed to the des-O-methyl
product (des-fluoro des-O-methyl hydroxy BMS-204352), the
O-demethylation in the protein adduct (des-fluoro
des-O-methyl BMS-204352 lysine adduct) was not likely caused during
acid hydrolysis. The reason is that only <5% of synthetic des-fluoro
BMS-204352 lysine adduct was O-demethylated to des-fluoro
des-O-methyl BMS-204352 lysine adduct under the same acid hydrolysis
conditions. Therefore, the O-demethylation probably proceeded the
adduct formation between lysine residue and an activated BMS-204352 species. A
proposed mechanism for the metabolic activation of BMS-204352 is shown in
Fig. 7 to account for these
observations. In this mechanism, BMS-204352, a
3-(2-methoxyphenyl)-3-fluoro-2-oxindole, is oxidized by P4503A to lead to
O-demethylation. A nonenzymatic elimination of hydrogen fluoride from
the formed phenol leads to formation of the ortho-quinone methide
metabolite. The resulting electrophilic reactive metabolite undergoes a
nonenzymatic Michael addition reaction with the -amino group of lysine
residue(s) in proteins to form a covalently bound adduct.
There are two unusual aspects regarding to the above proposed mechanism
involving the ortho-quinone methide covalently bound to plasma
protein (mostly albumin) through a lysine adduct. First, the
ortho-quinone methide reactive intermediate generated by P4503A in
rat livers or microsomal incubation was apparently stable enough to escape
from the enzyme active site and covalently bound to albumin molecules being
processed at endoplasmic reticulum membranes. Second, the hard nucleophile,
-amino group of lysine residue(s) and not the soft nucleophile, the
sulfhydryl group of cysteine residue(s) involved in the adduct formation.
Quinone methides are electrophilic reactive intermediates with positive charge
density centered mainly at the exocyclic methylene carbon
(Turner, 1964). Both the
formation and the rate/type of subsequent nucleophilic additions of
para-quinone methides of alkylated phenols can be affected by
delocalization of the positive charge density through extended
-conjugation at the exocyclic methylene carbon
(Bolton et al., 1995) or
through hydrogen binding interactions at the carbonyl and spatial interactions
(Filar and Winstein, 1960;
Bolton et al., 1992). Extended
-conjugation at the para-position of 4-alkylcatechols enhanced
the rates of isomerization of an ortho-quinone to
para-quinone methide through increasing the stability of quinone
methides (Iverson et al.,
1995). Eugenol para-quinone methides were stabilized
dramatically by extended -conjugation
(Bolton et al., 1995). The
electrophilic reactivity of para-quinone methide of butylated
hydroxytoluene increased as increased stabilization through hydrogen binding
interactions with solvents (Bolton et al.,
1992). The relative stability of the ortho-quinone
methide in this study might be due to the extended conjugation to an aromatic
ring system and a carbonyl group at the nucleophilic attack site. Our
preliminary results indicated that BMS-204352 was also a time-dependent
inactivator of P4503A in microsomal incubations (data will be published
separately). It is also important to point out that treating BMS-204352 with
BCl3 led to formation of an uncharacterized black polymer material,
indicating that the phenol precursor formed in tetrahydrofuran or methylene
chloride was not stable and polymerized. The sulfhydryl group in cysteine or
glutathione was the common nucleophile to form adducts with
para-quinone methides or imine methides
(Mizutani et al., 1983;
Nakagawa et al., 1983;
Tajima et al., 1983;
Nocerini et al., 1985;
Thompson et al., 1990). Bolton
and colleagues used amino acid and peptide models to study influence of
para-quinone methide reactivity on the alkylation of thiol and amino
groups in protein (Bolton et al.,
1997). Their results indicated that peptide alkylation should
occur in the order of cysteine thiol > N-terminal amino >
N--lysine based on chemical reactivity. Actually,
para-quinone methide from 2-tert-utyl-4,6-dimethylphenol was
at least 10-fold more reactive to form adduct with cysteine than with
-amino group of lysine (McCracken et
al., 1997). However, the para-quinone methide alkylated
hemoglobin (containing no cysteine residues) mainly at
N-lysine, suggesting that other factors such as the relative
concentrations of competing nucleophilic groups and their accessibility
determined by protein secondary and tertiary structures might be also
important for protein covalent binding
(Bolton et al., 1997). In
addition, solvolysis rates of the para-quinone methide by the hard
nucleophile, water, increased as increased stabilization of the methide
positive charge through hydrogen binding interactions
(Bolton et al., 1992). Rat
albumin contains 32 cysteine residues and 52 lysine residues
(Peters and Reed, 1980).
However, the ortho-quinone methide in this study formed adducts
exclusively with lysine, which might be explained by the relative stability of
this extended methide. GSH only modestly attenuated the protein covalent
binding, further demonstrating the strong preference of the hard nucleophile,
-amino group over the soft nucleophile, sulfhydryl group for adduct
formation. There might be specific binding or motifs in proteins directing the
ortho-quinone methide to a lysine-containing pocket of plasma
proteins before the nucleophilic addition occurs to form the lysine adduct.
Overall, the stabilization of the ortho-quinone methide through
extended conjugation would make the reactive intermediate possible to escape
from the active site of P450 enzyme(s) and to react with the hard nucleophile,
-amino group of lysine residue(s) in proteins.
para-Alkylated phenol such as butyl hydroxytoluene and eugenol,
and 3-methylindole undergo oxidation through a direct abstraction of a
hydrogen atom from the exocyclic methylene group
(Mizutani et al., 1983,
Thompson et al., 1990;
Skiles and Yost, 1996).
Safrole [1-allyl-3,4-(methylenedioxy)-benzene] undergoes
O-demethylation to catechol (1-allyl-3,4-dihydroxybenzene), which can
be oxidized to an ortho-quinone followed by isomerization to form a
para-quinone methide (Bolton et
al., 1994). Tamoxifen undergoes the following two steps of P450
oxidation: para-hydroxylation on one benzene ring or
O-demethylation at a different ring to generate
para-alkylated phenols followed by P450-mediated proton abstraction
to form the para-quinone methides
(Dehal and Kupfer, 1999;
Fan and Bolton, 2001). In this
study, we identified a novel ortho-quinone methide, which was formed
by a P450-mediated O-demethylation followed by a nonenzymatic
elimination of HF. Another ortho-quinone methide was formed through a
quinone reduction to ortho-alkylated phenol followed by a distal
elimination of cyanide in the quinone anticancer antibiotics saframycins A and
C, which rapidly led to an intramolecular ring close to form a hydroxyquinone
immonium ion (Lown et al.,
1982). Other examples of reactive electrophilic metabolites that
covalently bind to proteins include ortho-quinone from catechols
(Dehal and Kupfer, 1999),
epoxides (Sahali-Sahly et al.,
1996; Yamamoto et al.,
2002), iminoquinone (Shen et
al., 1999; Lai et al.,
2000), and acylglucuronides
(Bolze et al., 2002). The
lysine residue in protein can form covalent binding with carbonyl groups
through Schiff base formation (Braun et
al., 1995), with ,ß-unsaturated ketones through
Michael addition reactions (Szweda et al.,
1993), or with epoxides through nucleophilic substitution
reactions (Sujatha et al.,
2001). BMS-204352 lysine adduct identified in this study
represents another example for the lysine residue forming a covalent bond
through nucleophilic addition reaction.
There are reversible- and irreversible-types of protein covalent binding.
Covalent binding through disulfide bonds of cysteine residues of proteins with
sulfhydryl-containing drugs such as captopril and omapatrilat are
reversible-type covalent binding
(Ramaswamy et al., 2001). The
drug material(s) can be released from the covalent bound adduct(s) through
replacement by free thiols in glutathione or cysteine. Covalent binding
through nucleophilic addition reactions of amino acids such as cysteine or
lysine residues represent irreversible-type covalent binding. Their further
metabolism may depend on the metabolism of targeted proteins. It is noteworthy
that the half-life of radioactivity in rat plasma following
[14C]BMS-204352 is similar to that of albumin in plasma [i.e., 48 h
(Reed et al., 1988)].
Protein covalent binding may lead to such consequences as the production of
autoantibodies, inflammation, and cancer. Some of these reactive metabolites
are related to hepatic toxicities (Pumford
and Halmes, 1997). The formation of methide reactive intermediates
has been linked to hepatotoxicity, pulmonary toxicity, carcinogenicity, and
tumor promotion (Yost, 1989;
Thompson et al., 2001;
Monks and Jones, 2002).
Protein covalent binding has been observed with drugs that are generally
considered safe (captopril, aspirin, valproic acid), as well as drugs that
have been associated with hepatic impairment (acetaminophen overdose) and
idiosyncratic or allergic reactions (diflunisal and zomepirac)
(Uetrecht, 1992;
Pumford and Halmes, 1997).
Although the biological consequences of this P450 activation pathway of
BMS-204352 are unknown, no toxicological findings directly attributable to its
protein covalent binding have been observed with the compound in rats.
In summary, within an hour of dosing rats with [14C]BMS-204352,
over two-thirds of the drug related material (radioactivity) in plasma is
covalently bound to plasma proteins. The covalently bound adduct of BMS-204352
in rat plasma results from O-demethylation and replacement of the
ring-fluorine atom of [14C]BMS-204352 by -amino group of a
lysine residue in proteins. Metabolic activation of BMS-204352 appears to
proceed through an ortho-quinone methide, which is formed by
P450-mediated O-demethylation and spontaneous loss of hydrogen
fluoride. The toxicological consequences of the protein covalent binding
observed with BMS-204352 are presently unknown.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Address correspondence to: Dr. Donglu Zhang, Department of Preclinical Candidate Optimization, P.O. BOX 4000, Bristol-Myers Squibb, Princeton, NJ 08543. Email: Donglu.Zhang{at}BMS.com
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
-conjugation
dramatically stabilizes the quinone methide formed from eugenol.
Chem-Biol Interact 95:
279–290.[CrossRef][Medline]
This article has been cited by other articles:
|
|
 |
|
  D. Zhang, R. Krishna, L. Wang, J. Zeng, J. Mitroka, R. Dai, N. Narasimhan, R. A. Reeves, N. R. Srinivas, and L. J. Klunk METABOLISM, PHARMACOKINETICS, AND PROTEIN COVALENT BINDING OF RADIOLABELED MAXIPOST (BMS-204352) IN HUMANS Drug Metab. Dispos., January 1, 2005; 33(1): 83 - 93. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Zhang, W. Zhao, V. A. Roongta, J. G. Mitroka, L. J. Klunk, and M. Zhu AMIDE N-GLUCURONIDATION OF MAXIPOST CATALYZED BY UDP-GLUCURONOSYLTRANSFERASE 2B7 IN HUMANS Drug Metab. Dispos., May 1, 2004; 32(5): 545 - 551. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
