Discovery Tactics To Mitigate Toxicity Risks Due to ... - ACS Publications

May 27, 2010 - The quest for effective treatment options for osteoporosis continues to spark interest because of the widespread prevalence of this dis...
0 downloads 0 Views 427KB Size
Chem. Res. Toxicol. 2010, 23, 1115–1126

1115

Discovery Tactics To Mitigate Toxicity Risks Due to Reactive Metabolite Formation with 2-(2-Hydroxyaryl)-5(trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-one Derivatives, Potent Calcium-Sensing Receptor Antagonists and Clinical Candidate(s) for the Treatment of Osteoporosis Amit S. Kalgutkar,*,† David A. Griffith,‡ Tim Ryder,† Hao Sun,† Zhuang Miao,† Jonathan N. Bauman,† Mary T. Didiuk,‡ Kosea S. Frederick,† Sabrina X. Zhao,† Chandra Prakash,† John R. Soglia,† Scott W. Bagley,‡ Bruce M. Bechle,‡ Ryan M. Kelley,‡ Kenneth Dirico,‡ Michael Zawistoski,‡ Jianke Li,‡ Robert Oliver,‡ Angel Guzman-Perez,‡ Kevin K. C. Liu,‡ Daniel P. Walker,‡ John W. Benbow,‡ and Joel Morris‡ Pharmacokinetics, Dynamics and Metabolism Department and Department of Medicinal Chemistry, Pfizer Global Research and DeVelopment, Groton, Connecticut 06340 ReceiVed April 12, 2010

The synthesis and structure-activity relationship studies on 5-trifluoromethylpyrido[4,3-d]pyrimidin-4(3H)-ones as antagonists of the human calcium receptor (CaSR) have been recently disclosed [Didiuk et al. (2009) Bioorg. Med. Chem. Lett. 19, 4555-4559). On the basis of its pharmacology and disposition attributes, (R)-2-(2-hydroxyphenyl)-3-(1-phenylpropan-2-yl)-5-(trifluoromethyl)pyrido[4,3d]pyrimidin-4(3H)-one (1) was considered for rapid advancement to first-in-human (FIH) trials to mitigate uncertainty surrounding the pharmacokinetic/pharmacodynamic (PK/PD) predictions for a short-acting bone anabolic agent. During the course of metabolic profiling, however, glutathione (GSH) conjugates of 1 were detected in human liver microsomes in an NADPH-dependent fashion. Characterization of the GSH conjugate structures allowed insight(s) into the bioactivation pathway, which involved CYP3A4mediated phenol ring oxidation to the catechol, followed by further oxidation to the electrophilic orthoquinone species. While the reactive metabolite (RM) liability raised concerns around the likelihood of a potential toxicological outcome, a more immediate program goal was establishing confidence in human PK predictions in the FIH study. Furthermore, the availability of a clinical biomarker (serum parathyroid hormone) meant that PD could be assessed side by side with PK, an ideal scenario for a relatively unprecedented pharmacologic target. Consequently, progressing 1 into the clinic was given a high priority, provided the compound demonstrated an adequate safety profile to support FIH studies. Despite forming identical RMs in rat liver microsomes, no clinical or histopathological signs prototypical of target organ toxicity were observed with 1 in in vivo safety assessments in rats. Compound 1 was also devoid of metabolism-based mutagenicity in in vitro (e.g., Salmonella Ames) and in vivo assessments (micronuclei induction in bone marrow) in rats. Likewise, metabolism-based studies (e.g., evaluation of detoxicating routes of clearance and exhaustive PK/PD studies in animals to prospectively predict the likelihood of a low human efficacious dose) were also conducted, which mitigated the risks of idiosyncratic toxicity to a large degree. In parallel, medicinal chemistry efforts were initiated to identify additional compounds with a complementary range of human PK predictions, which would maximize the likelihood of achieving the desired PD effect in the clinic. The back-up strategy also incorporated an overarching goal of reducing/ eliminating reactive metabolite formation observed with 1. Herein, the collective findings from our discovery efforts in the CaSR program, which include the incorporation of appropriate derisking steps when dealing with RM issues are summarized. Introduction The quest for effective treatment options for osteoporosis continues to spark interest because of the widespread prevalence of this disease. Osteoporosis affects a substantial proportion of * To whom correspondence should be addressed. Tel: 860-715-2433. E-mail: [email protected]. † Pharmacokinetics, Dynamics and Metabolism Department. ‡ Department of Medicinal Chemistry.

the elderly population and causes notable morbidity, deterioration in quality of life, and mortality due to associated fragility fractures (1). While currently available antiresorptive agents (e.g., bisphosphonates, calcitonin, estrogen, and selective estrogen receptor modulators) prevent further bone loss, they cause relatively small increases in bone formation. The ability to stimulate bone growth and thereby replace bone already lost to the disease could mark a significant advancement in osteoporosis therapy. A first step in this direction is the recent FDA approval

10.1021/tx100137n  2010 American Chemical Society Published on Web 05/27/2010

1116

Chem. Res. Toxicol., Vol. 23, No. 6, 2010

of recombinant human parathyroid hormone (PTH1-34),1 teriparatide (Forteo), to treat severe osteoporosis in the United States (2). Daily subcutaneous administration of teriparatide has been associated with increases in bone mineral density and overall reduction in fracture rates in the osteoporotic population. However, given the usual difficulties associated with biotherapeutics (e.g., administration by injection and high cost of therapy), a viable alternative is the identification of an orally active small molecule agent that would stimulate secretion of endogenously stored PTH from the parathyroid glands via blockade of the calcium-sensing receptor (CaSR). CaSR is a G-protein-coupled, seven-pass transmembrane molecule present in the thyroid gland, whose function is to coordinate calcium homeostasis by regulating the release of PTH (3, 4). Low serum calcium concentrations trigger PTH release from the parathyroid glands, whereas an increase in calcium suppresses the release of this hormone (5, 6). By antagonizing CaSR, small-molecule agents can replicate a hypocalcemic state and stimulate PTH release (5, 6). From a drug discovery perspective, the identification of orally active and clinically viable CaSR antagonists poses significant challenges for two reasons. First, CaSR antagonists must stimulate the release of sufficient PTH for efficacy. Second, the profile of PTH stimulation following CaSR antagonism must be rapid and transient for bone anabolism since sustained activation causes prolonged PTH secretion and a catabolic state, such as hyperthyroidism (5-9). To achieve such a pharmacodynamic (PD) profile in vivo, the pharmacokinetic (PK) attributes of a CaSR antagonist must comply with a fast onset of action and a short PK halflife (t1/2) of approximately 1 h (10-12). Extensive in-house PK/ PD modeling (13, 14) has indicated that to mimic a PTH profile similar to teriparatide, an orally active CaSR antagonist must demonstrate in vitro CaSR binding potency (IC50) e 60 nM, rapid oral absorption, projected human blood clearance (CLb), and steady state distribution volume (Vdss) of 10 mL/min/kg and 1 L/kg, respectively, which would yield the desired t1/2 of 1 h. Toward the pursuit of novel and proprietary chemotypes, our laboratories recently disclosed a series of 5-(trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-one derivatives as potent and orally active CaSR antagonists (13). Within members of the class, (R)2-(2-hydroxyphenyl)-3-(1-phenylpropan-2-yl)-5-(trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-one 1 (Figure 1) was generally compliant with the stringent human PK/PD requirements, making it an attractive drug candidate for first-in-human (FIH) studies. However, in vitro metabolism studies in human liver microsomes (HLM) indicated that the phenol group in 12 underwent NADPH-dependent bioactivation to a reactive ortho1 Abbreviations: PTH, parathyroid hormone; CaSR, calcium-sensing receptor; PD, pharmacodynamic; PK, pharmacokinetic; t1/2, half-life; CLb, blood clearance; Vdss, steady state distribution volume; 1, (R)-2-(2hydroxyphenyl)-3-(1-phenylpropan-2-yl)-5-(trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-one; FIH, first-in-human; HLM, human liver microsomes; GSH, reduced glutathione; RM, reactive metabolite; GSH-EE, ethyl ester of reduced glutathione; UDPGA, uridine-5′-diphosphoglucuronic acid; SAM, S-adenosyl methionine; CYP, cytochrome P450; COMT, catechol-O-methyl transferase; 11, (R)-3-(1-(3,4-difluorophenyl)propan-2-yl)-2-(3-hydroxypyridin-2-yl)-5-(trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-one; LC-MS/ MS, liquid chromatography tandem mass spectrometry; SRM, selected reaction monitoring; CID, collision-induced dissociation; tR, retention time; SAR, structure-activity relationship; RLM, rat liver microsomes; IADRs, idiosyncratic adverse drug reactions. 2 Crystallographic data for 1 has been deposited with the Cambridge Crystallographic Data Center as supplementary publication no. CCDC 771675. Copies of the data can be obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, United Kingdom [Fax: +44-(0)1223-336033 or e-mail: [email protected]].

Kalgutkar et al.

Figure 1. CaSR antagonist 1 as a potential clinical candidate for the treatment of osteoporosis.

quinone species, which was trapped with reduced glutathione (GSH). Although the prospects of toxicity (both short- and longterm) due to reactive metabolite (RM) liability in 1 was concerning, a more immediate concern was the uncertainty around achieving the narrow window of human PK required to achieve teriparatide-like PD effects. Because of the availability of a serum biomarker (PTH) as an early indicator of clinical target engagement, a decision was made to rapidly progress 1 into an FIH study to obtain preliminary proof-of-mechanism for this target. Appropriate studies (metabolism/safety) were conducted to demonstrate a favorable safety profile to support the FIH study. In parallel, additional compounds with a complementary range of human CLb predictions (5-8 mL/min/ kg) were sought that would maximize the likelihood of achieving the target CLb and t1/2. The overall program strategy was to progress this cohort of compounds into FIH in as short a time as possible to select the best candidate for advanced development on the basis of safety and an optimal PK/PD response. Identification of compounds with little to no RM liability was a major laboratory objective in back-up compound selection. The collective findings from our efforts including our risk mitigation strategies in dealing with RM issues at candidate nomination stage are summarized herein.

Materials and Methods Materials. Test compounds utilized in the metabolism studies were synthesized at Pfizer (13). All compounds had purities >97% upon assessment by HPLC. All solvents were of analytical grade or higher. All other reagents including NADPH, uridine-5′diphosphoglucuronic acid (UDPGA), S-adenosyl methionine (SAM), GSH, reduced glutathione ethyl ester (GSH-EE, site of esterification is on the glycine carboxylic acid), alamethicin, and butyl boronic acid were obtained from Sigma Aldrich (St. Louis, MO). HLM fractions pooled from 53 individual male and female donors were purchased from BD Bioscience (Woburn, MA). A pooled human liver S-9 (15 human livers) fraction was obtained from In Vitro Technologies, Inc. (Baltimore, MD). In Vitro Metabolism Studies. Stock solutions of test compounds were prepared in DMSO, and the final concentration of DMSO in the incubation media was 0.2% (v/v). Incubations were carried out at 37 °C for 60 min in a shaking water bath. The incubation volume was 1 mL and consisted of the following: 0.1 M potassium phosphate buffer (pH 7.4) containing MgCl2 (3.3 mM), HLM (final protein concentration ) 1 mg/mL), NADPH (1.3 mM), and test compound (10 µM). RM formation was assessed via the addition of GSH or GSH-EE. The final concentration of the individual trapping agents in the incubation mixture was 1 mM. Incubations that lacked either NADPH or thiol trapping agents served as negative controls, and reactions were terminated by the addition of ice-cold acetonitrile (3 mL). Glucuronidation of 1 and (R)-3(1-(3,4-difluorophenyl)propan-2-yl)-2-(3-hydroxypyridin-2-yl)-5(trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-one (11, the backup candidate to 1) was also examined in HLM, which were treated with UDPGA (2 mM) and alamethicin (10 µg/mg). Incubations were conducted in the absence and presence of NADPH (1.3 mM) and GSH (1 mM) or GSH-EE (1 mM). The potential role of cytochrome P450 (CYP) 3A4 in the oxidative metabolism of 1 and

Metabolite Formation with Phenolic CaSR Antagonists 11 was examined in NADPH-, GSH-, and UDPGA-supplemented HLM in the presence of the CYP3A4-selective inhibitor ketoconazole (1 µM). Methylation by catechol-O-methyltransferase (COMT) was examined for 1 (10 µM) and its purified catechol metabolite M6 (10 µM) in human liver S-9 fraction (1 mg/mL) in the presence of NADPH (1.3 mM) and SAM (0.5 mM). Paroxetine (20 µM) was used a positive control in the human liver S-9 incubations to test for O-methylation (15). Following the addition of acetonitrile, incubation mixtures were centrifuged (3000g, 15 min), and the supernatants were dried using an evaporative centrifuge set at 37 °C. The residue was reconstituted with mobile phase and analyzed for metabolite formation by liquid chromatography tandem mass spectrometry (LC-MS/MS). Metabolite Identification by Orbitrap LC-MS/MS. The HPLC system consisted of an Accela quaternary solvent delivery pump, an Acella autoinjector, and a Surveyor photodiode array detector (Thermo Electron Corp., Waltham, MA). Chromatography was performed on a Phenomenex Hydro RP column (4.6 mm ×150 mm, 4 µ) (Phenomenex, Torrance, CA). The mobile phase was composed of 5 mM ammonium formate buffer (pH 3.0) (solvent A) and acetonitrile (solvent B). The flow rate was 1.0 mL/min. The LC gradient started at 10% B for 5 min, was ramped linearly to 90% B over 35 min, was held at 90% B over 10 min, was returned to the initial condition over 1.0 min, and was allowed to equilibrate for 4.0 min. Postcolumn flow passed through the diode array detector to provide UV (λ ) 254 nm) detection prior to being split to the mass spectrometer at a rate of 100 µL/min. The LC system was interfaced to a Thermo Orbitrap mass spectrometer operating in positive ion electrospray mode. Xcalibur software version 2.0 was used to control the LC/MS system. Full scan data were collected at 15000 resolution. Data-dependent product ion scans of the two most intense ions found in the full scan were obtained at 15000 mass resolution. The dynamic exclusion function was used with a 1 min exclusion duration after three successive product ion scans with an early exclusion if the precursor ion falls below a signal-to-noise of 20. Metabolites were identified by comparing t ) 0 samples to t ) 60 min samples (with or without cofactors), and structural information was generated from collisioninduced dissociation (CID) spectra of the corresponding protonated molecular ions. Isolation and Structural Characterization of M6, the Major Metabolite of 1 in HLM. M6, the major metabolite of 1, was biosynthesized via a scale-up of a NADPH-supplemented HLM incubation of 1 (100 µM) in a total volume of 50 mL. The incubation was conducted at 37 °C in a shaking water bath for 60 min. Air bubbles were introduced into the incubation mixture every 5 min to provide adequate oxygenation. Following the addition of cold acetonitrile (20 mL) at the end of the incubation, the solution was centrifuged at 3000g for 15 min and evaporated to dryness under a stream of nitrogen gas. The residue was reconstituted with a mixture of water:acetonitrile (50:50), and an aliquot (1 mL) was injected onto a HPLC system. The system consisted of a HP-1050 solvent delivery system, a HP-1050 membrane degasser, an HP1050 autoinjector (Hewlett-Packard, Palo Alto, CA), and a Thermo Separations spectromonitor 3200 UV (Thermo Fisher Scientific, Waltham, MA). The monitoring UV wavelength was λ 254 nm. Chromatography was performed on a Gemini (Phenomenex) C-18 column (4.6 mm × 150 mm, 5 µ). The mobile phase was composed of 5 mM ammonium formate (pH 3.0) (solvent A) and acetonitrile (solvent B). The flow rate was 8.0 mL/min. The LC gradient started at 30% B, was ramped linearly to 95% B over 32 min, was returned to the initial condition over 1.0 min, and was allowed to equilibrate for 7.0 min. The HPLC effluent fractions containing M6 from eight injections were combined and evaporated to dryness. To assess whether M6 was a catechol derivative, a solution of purified M6 (∼1 mg) in acetonitrile (5 mL) was treated with butylboronic acid (1 mL of a 25 mg/mL suspension in acetonitrile). After the reaction was stirred for 15 min at 37 °C, an aliquot (0.5 mL) was mixed with mobile phase and directly infused on a Finnigan LTQ spectrometer as well as on the Thermo Orbitrap to obtain MS2/

Chem. Res. Toxicol., Vol. 23, No. 6, 2010 1117 MS3 and full scan accurate mass spectra of M6 and its butylboronate derivative (16). Analysis of GSH Conjugates of Compounds 1 and 11 in HLM Fortified with GSH and UDPGA. A mixture containing HLM (protein concentration ) 1 mg/mL) in phosphate buffer (100 mM, pH 7.4, at 37 °C), MgCl2 (3.3 mM), and alamethicin (10 µg/ mg protein) was placed on ice for 15 min. Compounds 1 (10 µM) or 11 (10 µM) in DMSO were added to this mixture to obtain a final incubation volume of 100 µL and preincubated at 37 °C for 3 min. The first set of reactions (n ) 3) was initiated by the addition of both NADPH (1.3 mM) and GSH (1 mM). A second group of reactions (n ) 3) were initiated by the addition of NADPH, GSH, and UDPGA (3 mM). Blank incubations were also performed without GSH or UDPGA. After 60 min, the reactions were stopped by the addition of 200 µL of acetonitrile containing nefazodone as an internal standard. The relative levels of GSH conjugates formed in the absence or presence of UDPGA were determined by monitoring the presence of trapped GSH adducts by LC/MS. Quenched incubates were centrifuged at 3000g for 10 min. Supernatants were transferred, and aliquots of 10 µL were analyzed by LC/MS. Injections were made directly onto an HPLC column (Gemini C18, 50 mm × 4.6 mm, Phenomenex) coupled to an API 4000 Q-Trap mass spectrometer (Applied Biosystems/MDS SCIEX, Ontario, Canada), equipped with a TurboIonSpray source. The mass spectrometer was operated in the selected reaction monitoring (SRM) mode (SRM for compound 1: 747 f 618; SRM for compound 11: 784 f 655; in both cases, the SRM reflected the loss of 129 Da from the molecular weight of GSH conjugate) for semiquantitative analyses. The metabolites were separated on the HPLC column using a gradient solvent system consisting of acetonitrile and 0.1% formic acid with the flow rate set at 0.4 mL/ min. The initial conditions consisted of a mixture of acetonitrile and 0.1% formic acid (80:20 v/v). The percentage of acetonitrile was increased linearly from 20 to 100% in 1 min. After an additional 0.5 min at 100% acetonitrile, the column was re-equilibrated with the initial mobile phase for 1 min before the next injection. The percent of GSH conjugate formed in each sample was estimated from peak area ratios of the analyte to internal standard, and the values were normalized at 60 min to those obtained at 0 min. Quantum Chemical Calculations. The catechol metabolites of 1 and 11 and their corresponding ortho-quinone structures were built, and Gaussian 03 was used to calculate the effect of electronics on the two electron oxidation of the catechol the ortho-quinone species. The activation energies were estimated using the DFT/ B3LYP method with the 6-31G* basis set selected. The calculation was accomplished on Pfizer portable batch system Linux GRID in parallel. Single Crystal X-ray Analysis of 1. A representative crystal was surveyed, and a 0.90 Å data set (maximum sin Θ/λ ) 0.56) was collected on a Bruker APEX II/R diffractometer. Atomic scattering factors were taken from the International Tables for Crystallography (17). All crystallographic calculations were facilitated by the SHELXTL system (18). All diffractometer data were collected at room temperature. A trial structure obtained by direct methods was refined routinely. Hydrogen positions were calculated wherever possible. The methyl hydrogens were located by difference Fourier techniques and then idealized. The hydrogen atom attached to the phenolic oxygen was located by difference Fourier techniques and allowed to refine. The shifts calculated in the final cycles of least-squares refinement were all less than 0.1 of the corresponding standard deviations. The final R index was 3.16%. A final difference Fourier revealed no missing or misplaced electron density. The refined structure was plotted using the SHELXTL plotting package.

Results Metabolite Profiling of 1 in HLM. Figure 2 depicts the extracted ion chromatogram of an incubation mixture of 1 in NADPH- and GSH-supplemented HLM in the absence (panel A) and presence of UDPGA (panel B). In the absence of UDPGA, four metabolites (labeled M1, M2, M5, and M6) of 1

1118

Chem. Res. Toxicol., Vol. 23, No. 6, 2010

Kalgutkar et al.

Figure 2. Extracted ion chromatogram of 1 upon incubation with HLM supplemented with NADPH and GSH in the absence (A) or presence (B) of UDPGA. In panel B, metabolites M3 and M4 are glucuronide conjugates.

Table 1. LC-MS Data for 1 and Its Metabolites (M1-M6) in HLMa compound

tR (min)b

observed [M + H]+

calculated (mDa)c

M1 M2 M3 M4 M5

20.0 20.5 22.5 23.1 26.3

747.2055 747.2055 618.1694 602.1740 442.1373

747.2060 (-0.5) 747.2060 (-0.5) 618.1699 (-0.5) 602.1750 (-1.0) 442.1378 (-0.5)

M6

28.5

442.1373

442.1378 (-0.6)

1

31.0

426.1424

426.1423 (0.1)

LC-MSn fragments, m/z MS2 on m/z same as M1 MS2 on m/z MS2 on m/z MS2 on m/z MS3 on m/z MS4 on m/z MS2 on m/z MS3 on m/z MS2 on m/z MS3 on m/z

747: 618, 500, 356 618: 500, 442, 324, 304, 276, 248 602: 426, 308, 288, 260, 240 442: 324 442f324: 304 442f324f304: 286, 276 442: 324 442f324: 304, 276 426: 308 426f308: 288, 268, 260, 240, 220

a

Proposed MS fragment assignments are shown in Figure 3. b Under HPLC conditions listed in the Materials and Methods. c Mass difference in mDa between the observed and the calculated mass.

were detected at the retention times (tR) listed in Table 1. Besides these products, inclusion of UDPGA revealed the presence of two additional metabolites, M3 and M4 (see Table 1). Metabolites M1-M3, M5, and M6 were not detected when NADPH was omitted from the microsomal incubations, suggesting that CYP isozymes catalyzed the rate-limiting step in their formation (data not shown). The formation of M1-M3, M5, and M6 was also abolished upon inclusion of the selective CYP3A4 inhibitor ketoconazole, suggesting a predominant role for the major human isozyme in their formation. Structural Characterization of the Metabolites of 1. The mass spectral fragmentation patterns of 1 and its metabolites M1-M6 are summarized in Table 1, and proposed assignments of observed fragments are displayed in Figure 3. Metabolites M5 and M6 possessed a molecular ion at 442 (M + H)+, an addition of 16 Da to the molecular weight of 1), which suggested that they were monohydroxylated derivatives of 1. The presence of fragment ions at m/z 324, 304, and 276 in the CID spectra of M5 and M6 (addition of 16 Da to the fragment ions at m/z 308, 288, and 260 in the CID spectrum of 1) ruled out the 1-phenylpropan-2-yl substituent as a site of monohydroxylation. To examine whether M6, the major metabolite of 1, is a catechol derivative, M6 was biosynthesized in HLM, isolated, purified,

and reacted with butyl boronic acid, a reagent used to selectively derivatize vicinal diols and catechols (16, 19). An aliquot of the crude reaction mixture was mixed with the HPLC mobile phase and directly infused on the mass spectrometer to obtain spectral information for the putative butylboronate derivative. As seen in Figure 4, the observed molecular weight of 508 (M + H)+ and MS2/MS3 data were consistent with the formation of the butylboronate derivative of M6, unambiguously confirming that M6 was a catechol derivative. Because M5 was formed in very minute quantities in microsomal incubations, scale up and purification of this metabolite for structure elucidation were not possible. Furthermore, besides the characteristic loss of HF and CO, no other fragment ions were discerned in the CID spectrum of M5 that could allow insight into the site of hydroxylation on the 2-(2-hydroxyphenyl)-5-(trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-one scaffold. A tentative structure for M5, which involves phenol ring hydroxylation in a similar fashion as M6, is shown in Figure 3. Metabolites M3 and M4 possessed masses [(M + H)+] of 618 and 602, respectively, suggesting that they were glucuronides of a monohydroxylated derivative of 1 and the parent compound, respectively. Tentative structures of the glucuronide conjugates M3 and M4, which are consistent with the CID

Metabolite Formation with Phenolic CaSR Antagonists

Chem. Res. Toxicol., Vol. 23, No. 6, 2010 1119

Figure 3. Proposed assignments of the MS fragmentation pattern for 1 and its metabolites.

Figure 4. Mass spectral analysis for M6 and its butylboronate derivative (A) following infusion of a reaction mixture of purified M6 with butyl boronic acid into the mass spectrometer. Panels B and C depict the MS2 and MS3 ionization of ions at m/z 508 (M + H)+ and m/z 390, respectively.

1120

Chem. Res. Toxicol., Vol. 23, No. 6, 2010

fragment ions, are provided in Figure 3. Metabolites M1 and M2 possessed identical masses [(M + H)+ ) 747] and CID spectra. Their molecular weights indicated that they were isomeric GSH derivatives obtained via conjugation of the thiol nucleophile to a monohydroxylated metabolite of 1. The fragment ion at m/z 618 in the CID spectra of M1 and M2 is derived from the characteristic loss of the pyroglutamic acid component of GSH (20). The origin of the ion at m/z 500 in the CID spectrum of M1 and M2 can be explained via the combined loss of the phenylpropan-2-yl and pyroglutamate components. The ion at m/z 500 indicates that the 1-phenylpropan-2-yl functionality is unmodified in both M1 and M2. Likewise, the occurrence of the fragment ion at m/z 356 in the CID spectra of M1 and M2 is consistent with the presence of an aromatic thioether motif in these metabolites (19). On the basis of these observations, proposed structures for regioisomeric GSH conjugates M1 and M2, which are consistent with their respective exact masses and fragmentation patterns, are shown in Figure 3. Attempts were made to scale-up the GSH conjugate for NMR studies; however, these conjugates proved to be unstable during the course of isolation and purification. Finally, catechol O-methylation was not observed upon incubation of 1 or purified M6 with human liver S-9 fractions fortified with NADPH and SAM. Under the current experimental conditions, the antidepressant paroxetine demonstrated the formation of the isomeric guaiacol metabolites that are derived from a CYP-mediated 1,3-benzdioxole ring scission followed by O-methylation of the catechol intermediate (15). Strategies to Eliminate/Reduce RM Formation Liability in 1. We were aware that strategies to eliminate RM formation via gross structural modifications on the phenol motif (e.g., phenol ring replacement with five-membered heterocyclic bioisosteres, incorporation of ortho and/or para phenol ring substituents, and/or replacement of phenolic OH with functional groups of complementary polarity) were out of scope considering that such alterations generated compounds with weaker CaSR antagonist activity and/or low predicted human microsomal clearance (13, 21). Six-membered heterocyclic rings such as pyridine are routinely used in medicinal chemistry as phenyl ring isosteres. In the present situation, the pyridine ring also seemed attractive from a metabolism standpoint, since it is more electron-deficient than the phenyl ring and can potentially attenuate catechol oxidation to quinone as was recently demonstrated by Samuel et al. (22). Consequently, we focused on the targeted synthesis of a series of C-2 azaphenol derivatives and evaluated their CaSR antagonist activity, predicted HLM CLb, and RM formation potential. In many instances, the corresponding C-2 phenol variants were available for direct comparison of these properties. RM formation in NADPHsupplemented HLM was examined using our high-throughput RM screen, which uses GSH-EE as a trapping agent (23). Putative GSH-EE conjugates of electrophilic quinone intermediates were detected by SRM of the anticipated molecular ions MH+ to (MH - 129)+ transitions, where the 129 Da ion represents the neutral loss of the pyroglutamate moiety from GSH-EE. In addition, constant neutral loss scanning of 129 Da (loss of pyroglutamic acid from GSH-EE) technique was used to detect potential GSH-EE conjugates other than those derived from conjugation to quinone intermediates. Peak areas of the quinone/GSH-EE conjugates were used as an approximate measure of bioactivation potential for individual compounds, assuming that ionization efficiencies within the chemical series would be comparable under the SRM conditions.

Kalgutkar et al.

As seen in Table 2, the formation of GSH-EE conjugates was observed for C-2 phenols as well as C-2 azaphenols. The observed molecular weight (M + H)+ of the GSH-EE conjugates for the test compounds indicated that they were derived from conjugation of the thiol nucleophile with the corresponding quinone species. Because GSH-EE was used as a trapping agent instead of GSH, the molecular ion (M + H)+ of the conjugates increased by 28 mass units [for example, the GSH and GSHEE conjugates of 1 possessed (M + H)+ values of 747 and 775 Da, respectively]. Comparison of the peak areas of GSH-EE conjugates, however, indicated that azaphenols were less susceptible to RM formation than the corresponding phenols. Constant neutral loss analysis of 129 Da did not reveal the presence of any unique GSH conjugate(s) for test compounds. Isosteric exchange of the phenol ring in 1 with the corresponding azaphenol moiety resulted in 5, which demonstrated significant reduction in GSH-EE conjugate formation, while retaining the CaSR pharmacology and HLM CLb attributes of 1. Of considerable interest, however, was azaphenol 11, which was 3-5-fold more potent as a CaSR antagonist than compounds 1 and 5 with a predicted human CLb (7.4 mL/min/kg) in a complementary range relative to 1 (12 mL/min/kg). In addition, the extent of GSH-EE conjugate formation with 11 was significantly diminished when compared with the corresponding peak area for the GSH-EE conjugate of 1 (peak area of GSH-EE conjugate for 1 ) 166810 ( 19400, peak area of GSH-EE conjugate for 11 ) 3000 ( 858). On the basis of these collective findings and the in vivo PD effects (transient PTH stimulation) following oral administration to rats, compound 11 was selected as a back-up candidate to 1. It is noteworthy to point out that our initial enthusiasm in compound 12, based upon its potency and complete lack of GSH-EE conjugate formation, diminished due to its predicted low human CLb (∼2.0 mL/min/kg) in HLM. While this finding was somewhat disappointing, we were pleasantly surprised to learn that replacement of the phenolic OH group on the azaphenol 2 with CF3 (compound 13) or Cl (compound 14) afforded analogues that retained in vitro CaSR potency while demonstrating human CLb values complementary to compounds 1 and 11 (Figure 5) (13). Furthermore, no GSH-EE conjugates of 13 and 14 were observed in incubations with NADPHsupplemented HLM. The somewhat superior CaSR potency of 13 led to its further consideration as the third back-up candidate. Metabolite Profiling of 11 in HLM. Figure 6 depicts the extracted ion chromatogram of an incubation mixture of 11 in NADPH- and GSH-supplemented HLM in the absence (panel A) and presence of UDPGA (panel B). In absence of UDPGA, three metabolites (labeled M7, M11, and M12) were detected at the tR listed in Table 3. Inclusion of UDPGA revealed the presence of three additional metabolites: M8, M9, and M10. Metabolites M7-M9, M11, and M12 were not detected when NADPH was omitted from the microsomal incubations or when ketoconazole was added to the HLM incubations, suggesting that CYP3A4 catalyzed the rate-limiting step in their formation. Structural Characterization of the Metabolites of 11. The mass spectral fragmentation patterns of 11 and its metabolites M7-M12 are summarized in Table 3, and proposed assignments of observed fragments are displayed in Figure 7. The masses of M11 and M12 [(M + H)+ ) 479] indicated that they were monohydroxylated derivatives of 11. The presence of fragment ion at m/z 155 ruled out the difluorophenylpropan-2-yl group as the site of hydroxylation in M11 and M12. A tentative structural assignment for M12, the major oxidative metabolite of 11, is shown in Figure 7. We propose M12 to be the

Metabolite Formation with Phenolic CaSR Antagonists

Chem. Res. Toxicol., Vol. 23, No. 6, 2010 1121

Table 2. SAR Analysis of CaSR Antagonism, HLM CLb, and RM Formation Potential with 5-(Trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-ones Containing C-2 Phenol and C-2 Azaphenol Substituents

a Procedures for measuring CaSR antagonist activity have been described in detail (13). IC50 values are means of three individual experiments. Human CLb values (average of two independent determinations) were estimated from the intrinsic clearance date from HLM using the well-stirred model (incorporating unbound fraction in plasma/microsomes) as previously described by Obach (24). c Peak areas as estimated for GSH-EE conjugates were derived from the catechol f quinone bioactivation pathway. d Peak areas of GSH-EE conjugates are mean values obtained from n ) 2-4 independent experiments on test compound (10 µM) in HLM fortified with NADPH (1.3 mM) and GSH-EE (1 mM). e Not detected. b

corresponding catechol metabolite based on analogous findings with the structural analogue 1. Likewise, M11 is speculated to be a regioisomer of M12. Metabolites M8 and M9 possessed identical masses of ∼655, suggesting that they were obtained from glucuronidation of a monohydroxylated derivative of 11. In contrast, metabolite M10 revealed a mass of 639 [(M + H)+],

which was consistent with glucuronidation of phenolic OH group in 11. Metabolite M7 possessed a protonated mass [(M + H)+] of 784, suggesting that it originated from conjugation of GSH to a monohydroxylated metabolite of 11. The fragment ions m/z 709 and m/z 655 were derived from the characteristic loss of the glycine and pyroglutamic acid components of GSH

1122

Chem. Res. Toxicol., Vol. 23, No. 6, 2010

Figure 5. Nonphenolic 5-(trifluoromethyl)pyrido[4,3-d]pyrimidin4(3H)-ones devoid of RM formation.

(20). Furthermore, the ion at m/z 501 indicated that the difluorophenylpropan-2-yl functionality in M7 remained unaltered. The proposed structure of M7 that is consistent with the observed molecular weight and fragmentation pattern is shown in Figure 7. Effect of Glucuronidation on RM Formation in Compounds 1 and 11. The impact of glucuronidation on the CYP3A4-mediated bioactivation of 1 and 11 was assessed by comparing the amount of GSH conjugates (M1/M2 for 1 and M7 for 11) formed in NADPH and GSH-supplemented HLM incubations of 1 or 11 (10 µM, each) in the presence or absence of UDPGA. Figure 8 shows the effect of glucuronidation on GSH conjugate formation with 1 (panel A) and 11 (panel B). Incubation of 1 or 11 with NADPH and GSH-supplemented HLM containing UDPGA for 60 min resulted in approximately 75% formation of M1/M2 and 35% formation of M7 relative to incubations lacking UDPGA. This suggested that the addition of UDPGA changed the levels of GSH conjugate formed in the microsomal incubations for both compounds. Quantum Chemical Calculations. The electronic energies of the catechol and ortho-quinone forms of 1 and 11 and the relative transition energy (∆E ) [quinone + H2O] - [catechol + 1/2O2]) were calculated using density functional theory as previously described for the atypical antipsychotic remoxipride (25). The relative activation energy for both compounds is positive, which indicated that the reaction to form ortho-quinone from catechol is not spontaneous but rather enzyme catalyzed (Table 4). In addition, the difference in relative transition energy (∆∆E ) ∆E11 - ∆E1) was used to elucidate the thermodynamic change of the reaction, which indicated that the catechol metabolite of 11 is more difficult to oxidize to the ortho-quinone form because of a higher energy barrier (10.8 kcal/moL).

Discussion Although the primary discovery objective in the CaSR antagonist program was to identify a compound that would possess the highest probability of achieving the stringent PK profile needed for anabolic bone efficacy, similar to teriparatide, historical experience within Pfizer suggested that this would be a high risk endeavor. Because of the paradoxical effects of PTH on bone, a slower than anticipated clearance would increase the risk of observing catabolic effects due to sustained PTH elevation. In contrast, a more rapidly cleared compound would significantly increase efficacious dose requirements. Currently available methods (e.g., liver microsomes, hepatocytes, and/or allometric scaling) for predicting human PK parameters suggest that a prediction accuracy better than 2-fold is difficult to achieve (26). Adding to the risk in a similar manner is the lack of methodology to assess potential variability in clinical PK a priori. In addition, intersubject variability in the PD response (PTH elevation) remains unknown. Given the high overall

Kalgutkar et al.

confidence in mechanism (clinical efficacy of teriparatide), uncertainty around human PK predictions and concomitant PD response was identified as the greatest program risk. To achieve our laboratory objectives, structure-activity relationship (SAR) studies were conducted to simultaneously optimize for CaSR potency and human CLb using HLM (13). Our initial efforts resulted in the identification of 1, which fulfilled the necessary PD (potent in vitro CaSR antagonism resulting in rapid and transient PTH stimulation in rats upon oral administration) and PK (projected human CLb ) 12 mL/ min/kg, Vdss ) 1.1 L/kg, t1/2 ) 1.1 h, and oral F ) 15-40%) criteria. Confidence in the human CLb estimate using HLM was fairly high since 1 demonstrated (1) a good correlation ((2fold) between in vivo rat CLb and the clearance from rat liver microsomes (RLM); (2) substantial decrease in in vivo CLb and an increase in t1/2 in rats pretreated with CYP inactivator 1-aminobenzotriazole, indicating that CYP isozymes were principally involved in elimination; and (3) an in vitro metabolic profile in RLM, which was virtually identical to the one observed in human (oxidative metabolism significantly dominated parent glucuronidation). Collectively, these observations significantly increased the level of confidence in the predicted CLb of 1 using HLM, on the basis of which it was decided to progress 1 into an FIH study to test the accuracy of our human PK predictions and corresponding PD effects. However, the finding that 1 formed GSH conjugates in HLM posed a hurdle with regards to compound progression to FIH and in larger clinical trials. The formation of GSH conjugate(s), an indicator of RM formation, was of particular concern due to the cumulative evidence supporting the hypothesis that RMs can be involved in direct organ toxicity as well as in idiosyncratic adverse drug reactions (IADRs), leading to suspension from clinical trials, black box warnings, or, in some cases, withdrawal from the market (27-29). Mass spectral characterization of the GSH conjugates of 1 suggested that they were derived from the conjugation of the thiol nucleophile to an electrophilic quinone intermediate, a hypothesis that is supported by the finding that the major metabolite of 1 in HLM, that is, M6, was a catechol derivative. Straightforward solutions to eliminate RM formation were not conducive with the SAR for CaSR antagonism. In light of the rigid SAR requirements for primary pharmacology, a conscious decision was made to progress 1 toward clinical evaluation, while ensuring that appropriate safety tests were conducted to derisk the prospect of both short- and long-term toxicity, especially due to RM formation. In parallel, medicinal chemistry efforts were initiated to identify back-up compounds with equipotent or superior CaSR potency, complementary human CLb range, and a key objective of reducing or eliminating RM formation. These attributes were largely achieved with compounds 11 and 13, respectively. From a structure-metabolism standpoint, our findings that the excessive levels of GSH conjugate(s) formation through phenol bioactivation in 1 could be lowered by replacing the phenyl group with a pyridyl group is in line with previous literature observations. For instance, Hartz et al. demonstrated that the formation of electrophilic quinone-imine species from the sequential metabolism of the 4-difluoromethoxy-2,6-dichloroaniline scaffold in corticotrophin-releasing factor 1 receptor antagonists could be significantly attenuated by isosteric replacement of the phenyl ring with a pyridyl ring (30). Similarly, colleagues from Merck revealed that NADPH-dependent covalent binding of electrophilic ortho-benzoquinones to HLM could be reduced by exchanging the phenyl ring with a pyridine

Metabolite Formation with Phenolic CaSR Antagonists

Chem. Res. Toxicol., Vol. 23, No. 6, 2010 1123

Figure 6. Extracted ion chromatogram of 11 upon incubation with HLM supplemented with NADPH and GSH in the absence (A) or presence (B) of UDPGA. In panel B, metabolites M8, M9, and M10 are glucuronide conjugates.

Table 3. LC-MS Data for 11 and Its Metabolites (M7-M12) in HLMa compound

tR (min)b

observed [M + H]+

calculated (mDa)c

M7 M8 M9 M10 M11 M12 11

19.8 20.3 21.3 22.3 24.9 25.9 29.6

784.1807 655.1447 655.1449 639.1497 479.1130 479.1133 463.1184

784.1819 (-1.2) 655.1458 (-1.1) 655.1458 (-0.9) 639.1509 (-1.2) 479.1137 (-0.7) 479.1137 (-0.4) 463.1188 (-0.4)

LC-MSn fragments, m/z MS2 MS2 MS2 MS2 MS2 MS2 MS2

on on on on on on on

m/z m/z m/z m/z m/z m/z m/z

784: 655: 655: 639: 479: 479: 463:

766, 479, 479, 463, 461, 325, 309,

709, 325 325 309, 325, 305, 289,

655, 501, 357 289 305, 155 155 155

a

Proposed MS fragment assignments are shown in Figure 3. b Under HPLC conditions listed in the Materials and Methods. c Mass difference in mDa between the observed and the calculated mass.

Table 4. Calculated Energies for the Formation of Reactive ortho-Quinone Intermediates from Catechols Using the DFT B3LYP/6-31G* Method compound

1

11

catechol (Eha) quinone intermediate (Eh) ∆E (kcal/mol)b ∆∆E ) ∆E11 - ∆E1 (kcal/mol)

-1576.74 -1575.50 212.10 10.8

-1791.24 -1789.99 222.90

a The atomic unit of energy hartree (Eh) ) 627.509391 kcal/mol ) 2625.5 kJ/mol. b ∆E is defined from the reaction: catechol + 1/2O2 ) quinone + H2O.

ring (22). The experimental findings on the diminished GSH conjugate formation with 11 is also supported by the results of our quantum chemical calculations, which suggest that the two electron oxidation sequence of catechol f ortho-benzoquinone is energetically less favored for the aza-catechol metabolite of 11 (relative to the catechol metabolite of 1) presumably due to resonance stabilization of the phenolic OH group with the pyridine nitrogen atom. In the case of 13, the back-up to 11, removal of the phenolic OH group completely abrogated RM formation. From a retrospective standpoint, while the exercise to reduce/eliminate RM formation (1 f 11 f 13) seems trivial, the lower predicted human CLb of 11 and 13 would have precluded further interest in these compounds had they emerged prior to the identification of 1. The observation that 1 and 11 formed GSH conjugates in NADPH-supplemented RLM and/or ariclor-1254-induced rat liver S-9 fraction, identical to the ones observed in HLM incubations, suggested that the rat was an appropriate species

to test oVert (or direct) toxicity due to RM formation. Because of the strong ties between RM formation and mutagenicity (31, 32), compounds 1 and 11 were first evaluated in genotoxicity screens. Despite bioactivation in the rat, the two compounds were devoid of metabolism-dependent (ariclor-1254-induced rat liver S-9/NADPH) genotoxicity in the Salmonella Ames test, the in vitro micronucleus, and in vitro cytogenetic assays. Lack of in vitro genotoxic response was also confirmed in in vivo 5 day rat safety studies where micronuclei formation in rat bone marrow was not observed after administration of 1 and 11 at oral doses of 10, 50, and 400 mg/kg. From a direct target organ toxicity perspective, no clinical signs or histopathological changes (e.g., elevations in liver enzymes indicative of hepatotoxicity) were noted with 1 and 11 in the rat toleration studies. Overall, these observations suggested that in an acute setting the toxicological profiles of 1 and 11 did not pose a significant risk of toxicity in human. Suffice to say, no adverse findings were noted with 13 in these toxicological assessments. Confident that the safety information gathered thus far would be sufficient to support short-term FIH studies to derisk human PK/PD, we focused our attention on studies that could mitigate the risk(s) of coVert toxicological concerns (e.g., IADRs) due to RM formation with 1 and 11, especially since the concordance between toxicity in animals and humans as it relates to IADRs is very low (33). Many drugs form RMs and/or covalently bind to hepatic tissue in vitro, but only a fraction thereof cause toxicity in vivo (34-36). Possible reason(s) for this anomaly include the existence of competing detoxication pathways and/ or a low daily dosing regimen (34-36). For instance, the phenol

1124

Chem. Res. Toxicol., Vol. 23, No. 6, 2010

Kalgutkar et al.

Figure 7. Proposed assignments of the MS fragmentation pattern for 11 and its metabolites.

Figure 8. Relative percentage of GSH conjugates (M1/M2 for 1 and M7 for 11) formed after incubation of 1 (A) or 11 (B) with NADPHand GSH-supplemented HLM in the presence of UDPGA. The amount of GSH conjugates formed in the presence of UDPGA (9) is normalized to amounts in incubations containing NADPH and GSH but without UDPGA (0). Data presented here are an average of three independent experiments.

motif in raloxifene is metabolized in HLM by CYP3A4 to electrophilic quinone species, which reacts covalently with GSH or microsomal protein (34, 37, 38). However, competing high

first-pass phenol glucuronidation restricts RM formation with the drug in vivo, an attribute that potentially explains the lack of IADRs with raloxifene (39). IADRs are also very rare with the antidepressant paroxetine despite CYP2D6-mediated methylenedioxyphenyl ring scission to the corresponding catechol, which is then oxidized to GSH- and protein-reactive orthoquinone intermediates in HLM (15). Likely explanation(s) for the discrepancy with paroxetine include O-methylation of the catechol as a competing detoxication pathway and the low daily paroxetine dose of 20 mg. Lending further credence to low daily dose theory as a mitigating factor for IADR risks is the example of remoxipride, which was withdrawn from the market due to several cases of aplastic anemia. GSH- and human neutrophilreactive electrophilic quinone metabolites of remoxipride have been characterized, which provide the circumstantial link with observed toxicity (25). It is interesting to note that the daily dose of remoxipride ranges from 300 to 600 mg with side effects more frequent at doses over 300 mg (40). In our situation, phenol glucuronidation in the starting material (1 or 11) was a minor component of overall metabolism. Inclusion of UGT cofactor UDPGA in HLM incubations, however, reduced the levels of GSH conjugates of 1 and 11, suggesting that phase II glucuronidation of the catechol metabolites (rather than parent) competed more efficiently with the bioactivation process. Because glucuronidation represents a low affinity, high capacity metabolic pathway, we anticipate that phase II conjugation will not be saturated in humans and will offer an alternative route of detoxication for the catechol metabolites of 1 and 11. To probe whether lack of significant glucuronidation in parent phenols was the result of a stable internal H-bond between the phenolic OH and the quinazolinone nitrogen, the X-ray structure of 1 was solved. The measured distance between the phenolic hydrogen and the quinazolinone nitrogen atom (N10) in the X-ray structure of 1 is ∼3.73 Å (Figure 9), which precludes the formation of a noncovalent sixmembered transition state. However, what remains unclear at the present time is the likelihood of the formation of such a transition state following binding of 1 or 11 in the UGT active site. PK/PD studies were also undertaken on 1 and 11 to predict

Metabolite Formation with Phenolic CaSR Antagonists

Chem. Res. Toxicol., Vol. 23, No. 6, 2010 1125

References

Figure 9. ORTEP diagram of compound 1. The measured distance between the phenolic (O26) hydrogen and the quinazolinone nitrogen atom (N10) in the X-ray structure of 1 is ∼3.73 Å.

the likelihood of a low predicted human dose as a mitigating factor for IADRs. A preclinical PK/PD relationship between systemic exposure of 1 or 11 and plasma PTH was characterized in rats using a precursor pool model commonly applied to pharmacological mechanisms that result in the stimulation of hormone release into plasma (14). Assuming a 1:1 translation of quantitative pharmacology from rat to human, the model predicted low daily doses of 1 and 11 to 45 and 10 mg, respectively. On the basis of the low predicted human efficacious dose, we speculate that the level of ortho-quinone species formed in humans after administration of 1 and 11 at these low doses may be readily sequestered by the liver’s pool of GSH. Satisfied with the level of diligence undertaken to mitigate the risk of oVert as well as coVert toxicity, we turned our attention to the prototypic ADME concern of drug-drug interaction potential due to RM formation. The role of RMs in mechanism-based inactivation of CYP isozymes, which catalyzes their formation, is a well-established phenomenon (41). Compounds 1 and 11 did not cause time- and concentrationdependent inactivation of human CYP3A4 despite the isozyme’s role in their metabolism to RMs, suggesting that PK interactions due to CYP3A4 inactivation would be minimal. In summary, multiple CaSR compounds were identified for exploratory FIH studies to resolve the uncertainty surrounding human PK/PD predictions. Phenol 1, which possessed the highest likelihood of achieving the targeted PK, was also found to possess RM liability. Despite undergoing bioactivation, 1 was devoid of overt in Vitro and in ViVo metabolism-dependent toxicity in standard preclinical toxicity screens, findings that qualified the further advancement of 1 into FIH. The existence of a competing detoxication metabolic pathway and a predicted low daily dosing regimen further increased the level of comfort around the safety profile of 1. An understanding of the mechanism leading to RM formation with 1 allowed the implementation of medicinal chemistry tactics to reduce/ eliminate metabolic activation in back-up compounds. Our studies illustrate potential strategies, which can be adopted in the drug discovery environment when dealing with RM-positive drug candidates for unprecedented pharmacological targets. Supporting Information Available: CID spectra of compounds 1, 11, and their metabolites. This material is available free of charge via the Internet at http://pubs.acs.org.

(1) Berry, S. D., Kiel, D. P., Donaldson, M. G., Cummings, S. R., Kanis, J. A., Johansson, H., and Samelson, E. J. (2010) Application of the national osteoporosis foundation guidelines to postmenopausal women and men: The Framingham osteoporosis study. Osteoporosis Int. 21, 53–60. (2) Blick, S. K., Dhillon, S., and Keam, S. J. (2008) Teriparatide: A review of its use in osteoporosis. Drugs 68, 2709–2737. (3) Garrett, J. E., Capuano, I. V., Hammerland, L. G., Hung, B. C., Brown, E. M., Hebert, S. C., Nemeth, E. F., and Fuller, F. (1995) Molecular cloning and functional expression of human parathyroid calcium receptor cDNAs. J. Biol. Chem. 270, 12919–12925. (4) Brown, E. M., and Macleod, R. J. (2001) Extracellular calcium sensing and extracellular calcium signaling. Physiol. ReV. 81, 239–297. (5) Arey, B. J., Seethala, R., Ma, Z., Fura, A., Morin, J., Swartz, J., Vyas, V., Yang, W., Dickson, J. K., Jr., and Feyen, J. H. (2005) A novel calcium-sensing receptor antagonist transiently stimulates parathyroid hormone secretion in vivo. Endocrinology 146, 2015–2022. (6) Nemeth, E. F., Delmar, E. G., Heaton, W. L., Miller, M. A., Lambert, L. D., Conklin, R. L., Gowen, M., Gleason, J. G., Bhatnagar, P. K., and Fox, J. (2001) Calcilytic compounds: Potent and selective Ca2+ receptor antagonists that stimulate secretion of parathyroid hormone. J. Pharmacol. Exp. Ther. 299, 323–331. (7) Gowen, M., Stroup, G. B., Dodds, R. A., James, I. E., Votta, B. J., Smith, B. R., Bhatnagar, P. K., Lago, A. M., Callahan, J. F., Delmar, E. G., Miller, M. A., Nemeth, E. F., and Fox, J. (2000) Antagonizing the parathyroid calcium receptor stimulates parathyroid hormone secretion and bone formation in osteopenic rats. J. Clin. InVest. 105, 1595–1604. (8) Miller, M. A., Chin, J., Miller, S. C., and Fox, J. (1998) Disparate effects of mild, moderate, and severe secondary hyperparathyroidism on cancellous and cortical bone in rats with chronic renal insufficiency. Bone 23, 257–266. (9) Deal, C. (2009) Potential new drug targets for osteoporosis. Nature Clin. Pract. Rheumatol. 5, 20–27. (10) Yang, W., Wang, Y., Roberge, J. Y., Ma, Z., Liu, Y., Lawrence, R. M., Rotella, D. P., Seethala, R., Feyen, J. H., and Dickson, J. K., Jr. (2005) Discovery and structure-activity relationships of 2-benzylpyrrolidinesubstituted aryloxypropanols as calcium-sensing receptor antagonists. Bioorg. Med. Chem. Lett. 15, 1225–1228. (11) Balan, G., Bauman, J., Bhattacharya, S., Castrodad, M., Healy, D. R., Herr, M., Humphries, P., Jennings, S., Kalgutkar, A. S., Kapinos, B., Khot, V., Lazarra, K., Li, M., Li, Y., Neagu, C., Oliver, R., Piotrowski, D. W., Price, D., Qi, H., Simmons, H. A., Southers, J., Wei, L., Zhang, Y., and Paralkar, V. M. (2009) The discovery of novel calcium sensing receptor negative allosteric modulators. Bioorg. Med. Chem. Lett. 19, 3328–3332. (12) Marquis, R. W., Lago, A. M., Callahan, J. F., Rahman, A., Dong, X., Stroup, G. B., Hoffman, S., Gowen, M., DelMar, E. G., Van Wagenen, B. C., Logan, S., Shimizu, S., Fox, J., Nemeth, E. F., Roethke, T., Smith, B. R., Ward, K. W., and Bhatnagar, P. (2009) Antagonists of the calcium receptor. 2. Amino alcohol-based parathyroid hormone secretagogues. J. Med. Chem. 52, 6599–6605. (13) Didiuk, M. T., Griffith, D. A., Benbow, J. W., Liu, K. K., Walker, D. P., Bi, F. C., Morris, J., Guzman-Perez, A., Gao, H., Bechle, B. M., Kelley, R. M., Yang, X., Dirico, K., Ahmed, S., Hungerford, W., DiBrinno, J., Zawistoski, M. P., Bagley, S. W., Li, J., Zeng, Y., Santucci, S., Oliver, R., Corbett, M., Olson, T., Chen, C., Li, M., Paralkar, V. M., Riccardi, K. A., Healy, D. R., Kalgutkar, A. S., Maurer, T. S., Nguyen, H. T., and Frederick, K. S. (2009) Short-acting 5-(trifluoromethyl)pyrido[4,3-d]pyrimidin-4(3H)-one derivatives as orally-active calcium-sensing receptor antagonists. Bioorg. Med. Chem. Lett. 19, 4555–4559. (14) Abraham, A. K., Kalgutkar, A. S., Gao, X., Li, M., Healy, D. R., Petersen, D. N., Griffith, D. A., Mager, D. E., and Maurer, T. (2010) Pharmacodynamic model of parathyroid hormone modulation by a negative allosteric modulator of the calcium-sensing receptor. J. Pharmacol. Exp. Ther., in press. (15) Zhao, S. X., Dalvie, D. K., Kelly, J. M., Soglia, J. R., Frederick, K. S., Smith, E. B., Obach, R. S., and Kalgutkar, A. S. (2007) NADPHdependent covalent binding of [3H]-paroxetine to human liver microsomes and S-9 fractions: Identification of an electrophilic quinone metabolite of paroxetine. Chem. Res. Toxicol. 20, 1649–1657. (16) Giachetti, C., Zanolo, G., Assandri, A., and Poletti, P. (1989) Determination of cyclic butylboronate esters of some 1,2- and 2,3diols in plasma by high-resolution gas chromatography/mass spectrometry. Biomed. EnViron. Mass Spectrom. 18, 592–597. (17) Still, W. C., Tempczyk, A., Hawley, R. C., and Hendrickson, T. (1990) Semianalytical treatment of solvation for molecular mechanics and dynamics. J. Am. Chem. Soc. 112, 6127–6129.

1126

Chem. Res. Toxicol., Vol. 23, No. 6, 2010

(18) International Union of Crystallography (1974) International Tables for X-Ray Crystallography, Vol. IV, pp 55, 99, 149, Kynoch Press, Birmingham. (19) Kirsch, N. H., and Stan, H.-J. (1994) Gas chromatographic-mass spectrometric determination of chlorinated cis-1,2-dihydroxycyclohexadienes and chlorocatechols as their boronates. J. Chromatogr. A 684, 277–287. (20) Baillie, T. A., and Davis, M. R. (1993) Mass spectrometry in the analysis of glutathione conjugates. Biol. Mass Spectrom. 22, 319– 325. (21) Shcherbakova, I., Balandrin, M. F., Fox, J., Ghatak, A., Heaton, W. L., and Conklin, R. L. (2005) 3H-Quinazolin-4-ones as a new calcilytic template for the potential treatment of osteoporosis. Bioorg. Med. Chem. Lett. 15, 1557–1560. (22) Samuel, K., Yin, W., Stearns, R. A., Tang, Y. S., Chaudhary, A. G., Jewell, J. P., Jr., Lin, L. S., Hagmann, W. K., Evans, D. C., and Kumar, S. (2003) Addressing the metabolic activation potential of new leads in drug discovery: A case study using ion trap mass spectrometry and tritium labeling techniques. J. Mass Spectrom. 38, 211–221. (23) Soglia, J. R., Harriman, S. P., Zhao, S., Barberia, J., Cole, M. J., Boyd, J. G., and Contillo, L. G. (2004) The development of a higher throughput reactive intermediate screening assay incorporating microbore liquid chromatography-micro-electrospray ionization-tandem mass spectrometry and glutathione ethyl ester as an in vitro conjugating agent. J. Pharm. Biomed. Anal. 36, 105–116. (24) Obach, R. S. (1999) Prediction of human clearance of twenty-nine drugs from hepatic microsomal intrinsic clearance data: An examination of in vitro half-life approach and nonspecific binding to microsomes. Drug Metab. Dispos. 27, 1350–1359. (25) Erve, J. C. L., Svensson, M. A., von Euler-Chelpin, H., and KlassonWehler, E. (2004) Characterization of glutathione conjugates of the remoxipride hydroquinone metabolite NCQ-344 formed in vitro and detection following oxidation by human neutrophils. Chem. Res. Toxicol. 17, 564–571. (26) Hosea, N. A., Collard, W. T., Cole, S., Maurer, T. S., Fang, R. X., Jones, H., Kakar, S. M., Nakai, Y., Smith, B. J., Webster, R., and Beaumont, K. (2009) Prediction of human pharmacokinetics from preclinical information: Comparative accuracy of quantitative prediction approaches. J. Clin. Pharmacol. 49, 513–533. (27) Ju, C., and Uetrecht, J. P. (2002) Mechanism of idiosyncratic drug reactions: Reactive metabolite formation, protein binding and the regulation of the immune system. Curr. Drug Metab. 3, 367–377. (28) Evans, D. C., Watt, A. P., Nicoll-Griffith, D. A., and Baillie, T. A. (2005) Drug-protein adducts: an industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem. Res. Toxicol. 17, 3–16. (29) Liebler, D. C., and Guengerich, F. P. (2005) Elucidating mechanisms of drug-induced toxicity. Nat. ReV. Drug DiscoVery 4, 410–420. (30) Hartz, R. A., Ahuja, V. T., Zhuo, X., Mattson, R. J., Denhart, D. J., Deskus, J. A., Vrudhula, V. M., Pan, S., Ditta, J. L., Shu, Y. Z., Grace, J. E., Lentz, K. A., Lelas, S., Li, Y. W., Molski, T. F., Krishnananthan, S., Wong, H., Qian-Cutrone, J., Schartman, R., Denton, R., Lodge, N. J., Zaczek, R., Macor, J. E., and Bronson, J. J. (2009) A strategy to minimize reactive metabolite formation: Discovery of (S)-4-(1-

Kalgutkar et al.

(31) (32) (33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

cyclopropyl-2-methoxyethyl)-6-[6-(difluoromethoxy)-2,5-dimethylpyridin-3-ylamino]-5-oxo-4,5-dihydropyrazine-2-carbonitrile as a potent, orally bioavailable corticotropin-releasing factor-1 receptor antagonist. J. Med. Chem. 52, 7653–7668. Marnett, L. J. (1999) Chemistry and biology of DNA damage by malondialdehyde. IARC Sci. Publ. 150, 17–27. Dobo, K. L., Obach, R. S., Luffer-Atlas, D., and Bercu, J. P. (2009) A strategy for the risk assessment of human genotoxic metabolites. Chem. Res. Toxicol. 22, 348–356. Olson, H., Betton, G., Robinson, D., Thomas, K., Monro, K., Kolaja, G., Lilly, P., Sanders, J., Sipes, G., Bracken, W., Dorato, M., Van Deun, K., Smith, P., Berger, B., and Heller, A. (2000) Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul. Toxicol. Pharmacol. 32, 56–67. Obach, R. S., Kalgutkar, A. S., Soglia, J. R., and Zhao, S. X. (2008) Can in vitro metabolism-dependent covalent binding data in liver microsomes distinguish hepatotoxic from nonhepatotoxic drugs? An analysis of 18 drugs with consideration of intrinsic clearance and daily dose. Chem. Res. Toxicol. 21, 1814–1822. Bauman, J. N., Kelly, J. M., Tripathy, S., Zhao, S. X., Lam, W. W., Kalgutkar, A. S., and Obach, R. S. (2009) Can in vitro metabolismdependent covalent binding data distinguish hepatotoxic from nonhepatotoxic drugs? An analysis using human hepatocytes and liver S-9 fraction. Chem. Res. Toxicol. 22, 332–340. Obach, R. S., Kalgutkar, A. S., Ryder, T. F., and Walker, G. S. (2008) In vitro metabolism and covalent binding of enol-carboxamide derivatives and anti-inflammmatory agents sudoxicam and meloxicam: insights into the hepatotoxicity of sudoxicam. Chem. Res. Toxicol. 21, 1890–1899. Chen, Q., Ngui, J. S., Doss, G. A., Wang, R. W., Cai, X., DiNinno, F. P., Blizzard, T. A., Hammond, M. L., Stearns, R. A., Evans, D. C., Baillie, T. A., and Tang, W. (2002) cytochrome P450 3A4-mediated bioactivation of raloxifene: irreversible enzyme inhibition and thiol adduct formation. Chem. Res. Toxicol. 15, 907–914. Yu, L., Liu, H., Li, W., Zhang, F., Luckie, C., van Breemen, R. B., Thatcher, G. R., and Bolton, J. L. (2004) Oxidation of raloxifene to quinoids: Potential toxic pathways via a diquinone methide and o-quinones. Chem. Res. Toxicol. 17, 879–888. Dalvie, D., Kang, P., Zientek, M., Xiang, C., Zhou, S., and Obach, R. S. (2008) Effect of intestinal glucuronidation in limiting hepatic exposure and bioactivation of raloxifene in humans and rats. Chem. Res. Toxicol. 21, 2260–2271. Lapierre, Y. D., Ancill, R., Awad, G., Bakish, D., Beaudry, P., Bloom, D., Chandrasena, R., Das, M., Durand, C., and Elliott, D. (1992) A dose-finding study with remoxipride in the acute treatment of schizophrenic patients. J. Psychiatry Neurosci. 17, 134–145. Kalgutkar, A. S., Obach, R. S., and Maurer, T. S. (2007) Mechanismbased inactivation of cytochrome P450 enzymes: Chemical mechanisms, structure-activity relationships and relationship to clinical drugdrug interactions and idiosyncratic adverse drug reactions. Curr. Drug Metab. 8, 407–447.

TX100137N