2,7-Disubstituted-Pyrrolotriazine Kinase Inhibitors with an Unusually

ARTICLE pubs.acs.org/crt

2,7-Disubstituted-Pyrrolotriazine Kinase Inhibitors with an Unusually High Degree of Reactive Metabolite Formation Kevin J. Wells-Knecht,*,† Gregory R. Ott,‡ Mangeng Cheng,‡ Gregory J. Wells,‡ Henry J. Breslin,‡ Diane E. Gingrich,‡ Linda Weinberg,‡ Eugen F. Mesaros,‡ Zeqi Huang,‡ Mehran Yazdanian,† Mark A. Ator,‡ Lisa D. Aimone,‡ Kelli Zeigler,‡ and Bruce D. Dorsey‡ †

Worldwide Development and ‡Worldwide Discovery Research, Cephalon Inc., 145 Brandywine Parkway, West Chester, Pennsylvania 19380, United States ABSTRACT: There are numerous published studies establishing a link between reactive metabolite formation and toxicity of various drugs. Although the correlation between idiosyncratic reactions and reactive metabolite formation is not 1:1, the association between the two is such that many pharmaceutical companies now monitor for reactive metabolites as a standard part of drug candidate testing and selection. The most common method involves in vitro human microsomal incubations in the presence of a thiol trapping agent, such as glutathione (GSH), followed by LC/MS analysis. In this study, we describe several 2,7-disubstituted-pyrrolotriazine analogues that are extremely potent reactive metabolite precursors. Utilizing a UPLC/ UV/MS method, unprecedented levels of GSH adducts were measured that are 5 10 times higher than previously reported for high reactive metabolite-forming compounds such as clozapine and troglitazone.

’ INTRODUCTION There are many documented cases of drug biotransformation forming reactive metabolites that elicit adverse reactions and toxicity arising from modification to biomolecules (e.g., proteins, DNA, etc.),1 5 oxidative stress,6,7 and induction of immunemediated8 and inflammatory responses.9 Reactive metabolite formation has played a role in the termination of compounds during clinical drug development, FDA-imposed black-box warnings, and withdrawal of drugs from the market.1 A recent in-depth review by Stepan et al. noted that reactive metabolite formation played a causative role in 69% of drugs associated with toxicity.10 Clearly, it is in the interest of the pharmaceutical industry to identify drugs that form reactive metabolites at an early stage to develop synthetic strategies to eliminate or minimize this type of metabolism.11,12 The degree of reactivity can vary widely. For example, some reactive acyl glucuronides and epoxides have been detected in biological samples after in vivo dosing of the parent compound.13,14 However, the majority of reactive metabolites are r 2011 American Chemical Society

short-lived and can only be detected with trapping reagents. The most common method involves incubating the compound with human liver microsomes in the presence of glutathione (GSH) and using LC/MS techniques to detect the presence of GSH adducts.15 Unfortunately, the relevance of simple yes/no results afforded by this technique are often confusing to project team members, especially in oncology programs where a certain degree of toxicity is tolerated. Gan et al. utilized dansyl-GSH (dGSH) in their incubations to quantitate the formation of GSH adducts using in-line UV and fluorescence detectors to the mass spectrometer.16 When the technique was applied to 50 marketed drugs, there was an association observed between those known to exhibit drug-induced toxicity and the presence of GSH adducts.17 The correlation was even more substantial when the relative amounts of GSH adducts and dose were taken into consideration. Other studies have also noted the connection between dose Received: July 25, 2011 Published: October 24, 2011 1994

dx.doi.org/10.1021/tx200304r | Chem. Res. Toxicol. 2011, 24, 1994–2003

Chemical Research in Toxicology and drug-induced toxicity.18,19 While this information substantiates a link between daily exposure to reactive metabolites and drug-induced toxicology for marketed compounds, it still leaves some uncertainty regarding the interpretation of reactive metabolite results during lead optimization, when potential dosing regimens in humans are far from being established. During investigations to evaluate novel 2,7-disubstituted pyrrolo[2,1-f][1,2,4]triazines analogues as kinases inhibitors, it was noted that several compounds had unusual mass spectral characteristics. These findings raised suspicions about their potential ability to form reactive metabolites. When these compounds were incubated with microsomes in the presence of GSH, the degree of GSH adducts formed was unexpectedly high. We instituted a reactive metabolite screen using the same incubation conditions as Gan et al. with the exception of using unlabeled GSH as the trapping reagent. Reactive metabolites are then detected and measured by ultra performance liquid chromatography (UPLC)/UV/MS. The levels GSH adducts formed by these pyrrolotriazine analogues were much higher than ever described for compounds known to form reactive metabolites. In this study, we describe the unusual properties of these analogues, along with the identification and measurement of GSH adducts, and then compare these results to other kinase inhibitors.

’ EXPERIMENTAL PROCEDURES Materials. All reagents and solvents used in this study were of HPLC grade or of the highest available purity. Diclofenac, clozapine, acetaminophen, ascorbic acid, and GSH were obtained from Sigma-Aldrich (St. Louis, MO). Troglitazone and nifedipine were obtained from Calbiochem (San Diego, CA). Sorafenib was obtained from American Custom Chemicals (San Diego, CA). Imatinib and nilotinib were obtained from Toronto Research Chemicals (North York, Ontario, Canada). CP-690,550 was obtained from Thesis Chemistry (Cambridge, ON, Canada). Human liver microsomes were obtained from Xenotech LLC (Lenexa, KS). P450 baculosomes used for reaction phenotyping were purchased from Invitrogen (Madison, WI). Baculosomes are microsomes prepared from insect cells infected with a recombinant baculovirus containing a human P450 isozyme and a rabbit NADPH-P450 reductase. Both microsomes and baculosomes were stored at 80 °C until use. The synthesis of pyrrolotriazine analogues used in this study has been described previously.20 All other tested compounds were synthesized using reported procedures.21 26 Microsome Incubations. Test compounds were initially dissolved in dimethyl sulfoxide (DMSO) at a concentration of 15 mM. The incubations were performed in duplicate using human liver microsomes (1 mg/mL protein) in 96-well plates, with 50 mM potassium phosphate buffer containing 5 mM MgCl2 and 1 mM GSH (pH 7.4, 500 μL total volume). Test compounds were preincubated with the microsomes for 5 min at 37 °C at 50 μM (0.3% DMSO final concentration). A time zero control aliquot (100 μL) was removed, added to 100 μL of icecold acetonitrile (ACN) containing 0.1% formic acid, vortexed, and placed on ice. NADPH (1 mM final concentration) was then added to initiate the reaction, which was placed in a 37 °C water bath. Aliquots (100 μL) were removed at 30 and 60 min, and the reaction was quenched by the addition of an equal volume of ice-cold ACN containing 0.1% formic acid, vortexed, and placed on ice. The samples were centrifuged at 2061g for 15 min at 5 °C. A 150 μL aliquot was transferred to another 96-well plate, and then, 300 μL of 0.1% formic acid in water was added and mixed with the aliquot. The plate was sealed and placed in an autosampler maintained at 5 °C pending UPLC/UV/MS analysis. Reaction Phenotyping with cDNA-Expressed Cytochrome P450 Enzymes. P450 baculosome incubations were carried out using

ARTICLE

10 μM substrate and 50 nM P450 content. Other conditions were similar to the microsome incubations previously described. UPLC/UV/MS Method. Samples were injected (15 μL) onto a Waters Acquity UPLC system using a Waters HSS T3 column (2.1 mm 100 mm, 1.8 μm) designed for improved retention of polar metabolites. The reactive metabolite assay is used for a wide array of compounds so generic conditions were utilized for the analysis. The mobile phases were 0.1% formic acid in water (A) and 0.1% formic acid in ACN (B). The gradient was 2% B for 0.5 min, then ramped to 100% B at 10.3 min using convex curve 7. The column was held at 100% B for 1.2 min, then returned to 2% B in 0.2 min, and held for 2.3 min. The flow rate was 0.5 mL/min, and the column temperature was maintained at 40 °C. Flow from the UPLC column was directed to a Waters Acquity photodiode array (PDA) detector scanning 220 450 nm at 20 spectra/ s. A divert valve (VIVI Valco Instrument, Houston, TX) was used after the PDA detector to direct flow to waste during the first 2 min to minimize contamination of the mass spectrometer source. Samples were first analyzed with the mass spectrometer (AB Sciex Q-Trap 4000, Foster City, CA) operated in negative ion electrospray (ESI) mode collecting 272 precursor ion spectra to detect GSH adducts.27 The scanned mass range was typically kept below 250 Da to maintain adequate sensitivity with upper and lower limits sufficient for detecting GSH adducts to parent and possible metabolites. Samples were injected a second time with the mass spectrometer operated in positive ion mode collecting full-scan spectra (enhanced MS). This was done to (a) match retention times of parent compound to UV peaks, (b) detect metabolites, (c) confirm the mass of GSH adducts obtained from the precursor ion analysis, and (d) confirm that metabolites are not coeluting with GSH adducts. Generic parameters used by the mass spectrometer provided good sensitivity and minimal in-source fragmentation of GSH adducts across a diverse range of test compounds. Conditions for precursor ion scans utilized an ionspray voltage = 4.5 kV, declustering potential = 50 V, and a collision energy ramp of 25 to 35 V. The enhanced MS scans used an ionspray voltage = 4.75 kV and a declustering potential = 75 V. Both methods utilized gas 1 and 2 = 50 psi and turbo gas temperature = 500 °C. UV spectra were obtained for each test compound, and chromatograms were constructed based on a suitable wavelength range in case there was a shift in UV maxima for GSH adducts. The % turnover of test compound was calculated from a comparison of peak areas obtained from 0 and 30 min samples. The % GSH adduct was calculated as a percentage of the GSH adduct peak area at 30 min relative to the test compound peak area at 0 min. High-Resolution Mass Spectrometry. High-resolution mass spectrometry was obtained on a Waters Synapt G2 Q-TOF mass spectrometer using positive ion electrospray and the same UPLC conditions as described above. The mass spectrometer parameters were capillary voltage = 3.1 kV, sampling cone = 35 V, desolvation temperature = 350 °C, desolvation gas flow = 1000 L/h, source temperature = 150 °C, and cone gas flow = 50 L/h. Mass spectra were obtained with sequential low and high collision energy (MSE) at 10 and 30 V, respectively, using leucine enkephalin (m/z = 556.2771) as a lock mass standard. In Vivo Rat Study. Bile duct-cannulated male CD rats were purchased from Charles River Laboratories (Wilmington, MA). Six rats (304 374 g) were given a 30 or 100 mg/kg oral dose of compound 2 formulated in PEG 400, and bile was collected during 0 2 and 2 4 h postdose. Urine samples were also collected at various times from 20 min to 1.5 h. Formic acid was added to an aliquot of bile samples (0.5% final concentration) to adjust the pH to approximately 3.0 and stabilize potential acyl glucuronide metabolites against acyl migration. Samples were stored at 20 °C pending analysis, at which time aliquots were diluted 1/10 (bile) or 1/3 (urine) with 0.1% formic acid, vortexed, and then centrifuged for 15 min at 3200 rpm at 4 °C. The samples were 1995

dx.doi.org/10.1021/tx200304r |Chem. Res. Toxicol. 2011, 24, 1994–2003

Chemical Research in Toxicology

Figure 1. MS spectra of 1 infused alone (A) showing the semiquinonediimine radical ion (m/z 514) and the quinonediimine ion (m/z 513) formed by the electrospray process. Infusion with ascorbic acid (B) significantly reduced the formation of these oxidation products. analyzed by multiple reaction monitoring (MRM) using transitions for metabolites previously identified from microsomal studies. Full scan [enhanced mass spectrometry (EMS)] and enhanced product ion (EPI) spectra were also obtained to detect and structurally characterize new metabolites. For the detection of glucuronide and GSH conjugates, 176 neutral loss (positive ion) and 272 precursor (negative ion) modes were performed, respectively.

’ RESULTS AND DISCUSSION Mass Spectrometry of Pyrrolotriazine Analogues. Figure 1A shows the mass spectrum obtained by infusion of compound 1 to optimize mass spectrometry parameters for an upcoming metabolite identification study in microsomes. The spectrum obtained is representative for many of the pyrrolotriazine compounds described in this study. The expected [M + H]+ ion was 515; however, there was also a prominent signal at m/z 514, which suggested the presence of a M+• radical ion caused by oxidation from the electrospray source. The weak signal at m/z 513 likely represents an additional one-electron oxidation to form the quinonediimine. When ascorbic acid was infused with compound 1, the mass spectrum in Figure 1B showed a significant reduction

ARTICLE

Figure 2. Proposed scheme depicting the formation of semiquinonediimine radical (m/z = 514) and quinonediimine (m/z = 513) ions during electrospray ionization of compound 1. The formation of the quinonediimine could potentially form by additional one-electron oxidation semiquinonediimine radical or via disproportionation of two molecules of the semiquinonediimine radical. The addition of ascorbic acid allows for one-electron reduction of the semiquinonediimine and quinonediimine species.

in the m/z 514 and 513 ions. This supports the identity of those ions as the semiquinonediimine radical and quinonediimine as outlined in Figure 2, which were later confirmed by high-resolution mass spectrometry (514.2688, 0.8 ppm error; 513.2610, 0.8 ppm error). While oxidation of compounds to radical cations during electrospray ionization has been described for some compounds,28 the process usually requires high concentrations of analyte and use of nonnucleophilic solvents such as dry ACN and dichloromethane. Interestingly, oxidation of dopamine during electrospray analysis has also been described,29 and this compound is known to form reactive metabolites in vivo that are implicated with its toxicity.30 32 When a similar radical ion was observed for another pyrrolotriazine compound undergoing a metabolism study, incubations were added that included 1 mM GSH. The results shown in Figure 3 illustrate the extensive amounts of GSH adduct formed in these incubations. GSH adducts, when present, typically appear as small peaks on a UV chromatogram; however, the conditions used for these incubations resulted in complete conversion of compound 2 to a GSH adduct of the parent compound in monkey and human microsomes. No GSH adduct was observed in the absence of NADPH when incubated up to 3 h nor was there a decrease in compound 2 (data not shown), which suggests that autoxidation is not responsible for generating the reactive metabolite. The LC/MS spectra obtained later by 1996



ARTICLE

Figure 3. UV chromatograms (300 400 nm) from microsomal incubations containing GSH taken at 0 (solid line) and 60 min (red dotted line) showing the disappearance of 2 and the formation of the GSH adduct.

high-resolution mass spectrometry (Figures 4 and 5) confirmed the presence of the GSH adduct. At the time these results were obtained, there was no in vitro reactive metabolite assay in our early discovery process. On the basis of the unexpected results with compound 2, we instituted an assay that provided meaningful data to medicinal chemists so they could develop synthetic strategies to eliminate or minimize potential reactive metabolites. With a LC/MS-based assay, signal intensities can easily vary 10-fold between similar compounds because ionization efficiencies are affected by basicity, polarity, hydrophobicity, and molecular size.33 In addition, matrix effects from components in the sample and the percentage of organic mobile phase present during the chromatographic run can further affect the LC/MS peak area intensities. For this reason, many companies using a LC/MS-based assay report a simple yes/no result for the presence of GSH adducts.34 A quantitative method for reactive metabolites was described by Gan et al., which used dGSH and fluorescent peak detection.17 Our reactive metabolite assay was a modification of this method that used unlabeled GSH and UPLC/UV/MS to detect and measure the amounts of GSH adduct(s) formed relative to the initial amount of parent compound. The incubation conditions were analogous to those outlined by Gan et al.,17 and we obtained similar results for several reference compounds (Figure 6 and Table 1). The fate of a compound is not judged solely on the results of the reactive metabolite assay but rather in the context of other numerous tests gauging the developability [e.g., pharmacokinetics, cytochrome P450 (CYP) inhibition, solubility, etc.]. Metabolic factors also need to be taken into consideration which may limit reactive metabolite formation in vivo.35,36

Figure 4. Accurate mass LC/MS spectra of the GSH adduct to compound 2 formed during microsomal incubations obtained with low (A) and high collision energy (B).

While it was recognized that there are no current tests that predict drug-induced toxicity with certainty, the minimization of in vitro reactive metabolite formation was considered to be a desirable goal to decrease the probability that such toxicity might be associated with a given molecule. Ideally, this screening approach, combined with a mechanistic understanding of the oxidative/metabolic process, would be used by medicinal chemists as a guide to eliminate or minimize the reactive metabolite liability and, thus, the potential for idiosyncratic drug toxicities. Even though some reactive metabolites are not effectively trapped by GSH because they are considered “hard electrophiles” (e.g., iminium ions, aldehydes), the assay could be adapted to utilize cyanide37 or a bifunctional trapping agent to potentially cover a broader range of reactive metabolites.38 The data in Table 1 show the GSH adduct levels obtained for numerous kinase inhibitors, several of which contain similar structural motifs to the pyrrolotriazine kinase inhibitors. Some of these are currently on the market against various kinase targets (Figure 7). With the exception of dasatinib, all of the marketed kinase inhibitors tested in this study had no detectable GSH adducts. The amount of dasatinib GSH adduct measured was similar to diclofenac. The formation of dasatinib reactive metabolites has been previously reported39 and implicated in mechanism-dependent inactivation of CYP3A4; however, it is unknown whether reactive metabolites are responsible for reported hepatitis40 or lupus41 adverse events. 1997



ARTICLE

Figure 5. Proposed fragmentation pattern for the GSH adduct to compound 2.

Figure 6. Structures of reference compounds tested in this study that are known to form reactive metabolites.

The remaining compounds shown in Figure 8 are (or were) in various stages of research and/or development at pharmaceutical companies and designed to target various kinases, including the pyrrolotriazine analogues from Cephalon. As shown in Table 1, there was a wide variation in the quantity of GSH adducts formed for these compounds. Adducts for three of these (NVP-TAE226, NVP-TAE684, and TG101209) were produced at levels similar to or exceeding those observed with troglitazone. It should be noted the Janus kinase (JAK2) inhibitor, TG101209, has been replaced in development by SAR302503 (previously known as TG101348), which does not produce detectable GSH adducts. While reactive metabolite formation was fairly high for these compounds, we were amazed by the amplified GSH levels for the pyrrolotriazine analogues containing the same putative quinonediimine-forming structural motifs. To our knowledge, there have been no reports of GSH adduct formation of this magnitude by metabolic activation of a compound in vitro. The high levels

do not appear to be related to enhanced UV extinction coefficients for the adducts. The UV spectra of parent compounds and their adducts are similar, and we can account for a loss in parent peak area as GSH adduct(s) and metabolites. As shown in Figure 3, the complete loss of compound 2 chromatographic peak in monkey and human microsomal incubations led to a GSH adduct peak with similar intensity. The peak areas of compound 2 and its GSH adduct in prepared samples are unchanged after 14 h; therefore, the results are not biased by differences in their stabilities. For all of the pyrrolotriazine compounds listed in Table 1, with the exception of compound 5, the major GSH adduct formed was to parent compound. The GSH adduct to compound 4 with a mass of 714 results from replacement of the methylpiperazine ring with a hydroxyl group. This type of product likely results from deamination of the quinonediimine to a quinoneimine following metabolic oxidation42,43 with subsequent addition of GSH. The presence of this minor adduct (∼4% of total) is evidence that metabolic activation occurs at the 1,4-phenylenediamine ring in the structure. A similar GSH adduct was also observed for TG101209 (mass = 732), representing ∼4% of the total adducts formed. The metabolic activation of compound 5 first requires C-oxidation at the morpholine ring, yielding a p-aminophenol structure that can be further oxidized to a quinoneimine with subsequent addition of GSH. The GSH adducts thus arise from oxidative O-dealkylation (813), followed by reduction of the N-ethanal group (815) or oxidative N-dealkylation (771). These results suggest the pyrrolotriazine ring is somehow greatly enhancing reactive metabolite formation. No further experiments were conducted to measure oxidation potentials; however, detection of the semiquinonediimine radical in the mass spectrometer (Figure 1) suggests that these analogues have a high oxidation potential. Another possibility is that the pyrrolotriazine ring produces a three-dimensional structure that strongly favors oxidation within the P450 active site. The results of a P450 reaction-phenotyping study, shown in Figure 9, 1998



ARTICLE

Figure 7. Structures of marketed kinase inhibitors tested in this study.

Table 1. GSH Adducts Formed from Pyrrolotriazine Kinase Inhibitors as Compared to Other Kinase Inhibitors and Reference Compoundsa drug

% substrate turnover

adduct mass (Da)

proposed adduct composition

% GSH adduct(s)

456

M + GSH-2H

0.7b

diclofenac

69

582

M + GSH + O-HCl

0.9c

clozapine

26

631

M + GSH-2H

6.4d

troglitazone

22

746

M + GSH-2H

13.8e

sorafenib (Nexavar)

28

ND

imatinib (Gleevec)

15

ND

nilotinib (Tasigna) dasatinib (Sprycel)

14 24

acetaminophen

PF2341066

6.4

M + GSH-2H + O

5

SAR302503 (TG101348)

10

INCB18424

42

PF-562271

808

7.6

ND 1.0 ND ND

629

M+GSH+H2O

1.0

812

M + GSH-2H (two peaks)

3.0

CP-690,550

12

617


5.1

NVP-TAE226

28

773

M + GSH-2H

12.8

789 704

M + GSH-2H + O M + GSH-2H C4H7N

NVP-TAE684

12

918

M + GSH-2H

13.7

TG101209

27

814

M + GSH-2H

20.0

732

M + GSH-2H-C5H10N2 + O

1

25

819

M + GSH-2H

23.7

3

37

779

M + GSH-2H

39.7

4

68

796

M+GSH-2H

49.6

59

714 815

M+GSH-2H-C5H10N2 + O M + GSH-2H + H2O

51.0

813

M + GSH-2H + O

5

771

M + GSH-2H-C2H2

2

76

870

M + GSH-2H

61.0

6

80

907


81.0

a

The adduct mass is the nonionized form. ND, not detected. b Gan et al. reported 0.5%.16 c Gan et al. reported 1.516 and 0.52%.17 d Gan et al. reported 9.816 and 3.15%.17 e Gan et al. reported 12.516 and 3.9%.17

indicate that 3A4 is the major enzyme responsible for producing the reactive metabolite with some minor contribution by 2C9. It is interesting to note that compound 2 appeared to have excellent stability in our metabolic screen (t1/2 > 40 min, human microsomes, 0.5 μM initial concentration). Likewise, when compound 2 was incubated at 10 μM with human microsomes for metabolite identification studies, there was only 17% metabolism after 1 h, which primarily occurred at the piperazine ring resulting in N-dealkylation, N-oxidation, and C-oxidation. Yet,

when the same incubation included 1 mM GSH, nearly 100% of compound 2 was converted to a GSH adduct after 1 h (Figure 3). This does not appear to result from mechanism-based inhibition. A single-point time-dependent inhibition study using nifedipine as a 3A4 probe substrate revealed that 15% irreversible inhibition of 3A4 occurred after preincubating with compound 2 at 10 μM. A possible explanation for this paradox, outlined in Figure 10, may involve oxidation of compound 2 to a reactive quinonediimine intermediate as the major metabolic route, which is then 1999



ARTICLE

Figure 8. Structures of various kinase inhibitors tested in this study. The structure of compound 1 is shown in Figures 1 and 2.

converted back to the parent compound by successive oneelectron reductions via P450 reductase or two-electron reduction via NAD(P)H:quinone oxidoreductase 1 (NQO1; DT-diaphorase).44,45 The inclusion of GSH would disrupt the cycle, diverting the pathway toward a GSH adduct end product. Although 1,4-benzoquinone compounds have been shown to undergo multiple GSH additions with increasing degrees of nephrotoxicity,46,47 only one GSH molecule was added to compound 2 (Figure 3). The levels of the GSH adduct remained unchanged even when the microsomal incubation was carried out to 3 h, which suggests that the GSH adduct is not a suitable substrate for further redox reactions or microsomal metabolism. There are many examples of quinone compounds involved in this type of redox cycling pathway,48,49 although the results are often complex when trying to assign the contributions of redox cycling and protein modifications to toxicity. While it was outside the scope of this paper, possible ways to test whether these pyrrolotriazine analogues can participate in redox cycling may

involve monitoring for increased reactive oxygen species, such as superoxide and superoxide,50,51 or markers of increase oxidative stress in cells.52 Evaluating the effect on compound half-lives in microsome incubations after adding NQO1 inhibitors or P450 reductase may also be used. If this scenario outlined in Figure 10 occurred in vivo, it could potentially lead to increased oxidative stress via formation of reactive oxygen species and decreased GSH levels.6,7,53 However, it should be noted that mice dosed with compound 2 at 55 mg/ kg/day bid for 12 days did not exhibit any overt toxicity or body weight loss. Perhaps the number of mice used in this study was not sufficient to highlight an idiosyncratic response or there may be other metabolic factors in mice to limit damage from reactive metabolites.54,55 On the basis of the high amounts of GSH adducts obtained in vitro in all species tested, there was speculation whether this would also be observed in vivo. Therefore, a subsequent study was conducted in bile duct cannulated rats to assess the presence 2000



ARTICLE

of GSH adducts of compound 2 and/or metabolites. The LC/ UV chromatogram, shown in Figure 11, is a representative bile sample for all six rats dosed with compound 2. On the basis of UV peak areas, it is evident that GSH conjugates make up a large percentage (∼75%) of compound 2 and metabolites present in rat bile. The most abundant GSH adduct, GSH-1, is the same adduct found in microsomal incubations. The amount of compound 2 was

2,7-Disubstituted-Pyrrolotriazine Kinase Inhibitors with an Unusually

Recommend Documents