Investigations into the mechanisms of pyridine ring cleavage in vismodegibs

Article abstract of DOI:10.1124/dmd.113.055715

Vismodegib (Erivedge, GDC-0449) is a first-in-class, orally administered small-molecule Hedgehog pathway inhibitor that is approved for the treatment of advanced basal cell carcinoma. Previously, we reported results from preclinical and clinical radiolabe

Full text of DOI:10.1124/dmd.113.055715

Supplemental Material can be found at:
http://dmd.aspetjournals.org/content/suppl/2014/01/03/dmd.113.055715.DC1.html
1521-009X/42/3/343–351$25.00
http://dx.doi.org/10.1124/dmd.113.055715
DRUG METABOLISM AND DISPOSITION
Drug Metab Dispos 42:343–351, March 2014
Copyright ª 2014 by The American Society for Pharmacology and Experimental Therapeutics
Investigations into the Mechanisms of Pyridine Ring
Cleavage in Vismodegibs
S. Cyrus Khojasteh, Qin Yue,1 Shuguang Ma, Georgette Castanedo, Jacob Z Chen,
Joseph Lyssikatos, Teresa Mulder, Ryan Takahashi, Justin Ly, Kirsten Messick, Wei Jia,
Lichuan Liu, Cornelis E. C. A. Hop, and Harvey Wong
Department of Drug Metabolism and Pharmacokinetics (S.C.K., Q.Y., S.M., J.Z.C., T.M., R.T., J.L., C.E.C.A., H.W.), Department of
Discovery Chemistry (G.C., J.L.), gRED Non Clinical Operations (K.M.), Small Molecule Pharmaceutical Sciences (W.J.), and Small
Molecule Clinical Pharmacology (L.L.), Genentech Inc., South San Francisco, California
Received November 6, 2013; accepted December 11, 2013
ABSTRACT
Vismodegib (Erivedge, GDC-0449) is a first-in-class, orally adminis- studies. The use of stable-labeled (13C2,15N)vismodegib on the
tered small-molecule Hedgehog pathway inhibitor that is approved pyridine ring exhibited that the loss of carbon observed in both M9
for the treatment of advanced basal cell carcinoma. Previously, we and M13 was from the C-6 position of pyridine. Interestingly, the
reported results from preclinical and clinical radiolabeled mass source of the nitrogen atom in the amide of M9 was from the
balance studies in which we determined that metabolism is the main
pyridine. Evidence for the formation of aldehyde intermediates was
route of vismodegib elimination. The metabolites of vismodegib are observed using trapping agents as well as 18O-water. Finally, we
primarily the result of oxidation followed by glucuronidation. The conclude that cytochrome P450 is involved in the formation of M9,
focus of the current work is to probe the mechanisms of formation of M13, and M18 and that M3 (the major mono-oxidative metabolite) is
three pyridine ring-cleaved metabolites of vismodegib, mainly M9, not the precursor for the formation of these cleaved products; rather,
M13, and M18, using in vitro, ex vivo liver perfusion and in vivo rat M18 is the primary cleaved metabolite.
Introduction
and conjugative metabolic reactions on the basis of metabolite profile
analyses. The amount of unchanged drug present in urine and bile
accounted for less than 1% in rats and less than 5% in dogs. In humans,
the percentage of absorbed drug could be estimated only on the basis of
absolute bioavailability studies and human mass balance studies, and it
was clear that hepatic elimination predominates for vismodegib.
The major primary metabolites of vismodegib were oxidative, with the
oxidation taking place mainly on the pyridine and chlorophenyl rings,
followed by glucuronide, sulfonate, or, to a lesser extent, glucose con-
jugation. The major metabolites were qualitatively similar in all species.
The phase 2 conjugates were observed in the urine and bile samples from
rats and dogs and were not detected in feces, which is consistent with gut
bacteria hydrolysis of these conjugates to their respective oxidative me-
tabolites. None of these metabolites was circulating in plasma at significant
levels (,1% of total drug-related material).
Vismodegib is an orally active, small molecule that is among the
new generation of targeted therapies for oncology. This drug targets
the Hedgehog signaling pathway that is activated in certain types of
cancer and is currently approved for treatment of metastatic or locally
advanced basal cell carcinoma (Scales and de Sauvage, 2009). Further
investigations are under way for other indications. Vismodegib was
optimized for suitable potency; selectivity; and absorption, distribu-
tion, metabolism, and excretion properties (Robarge et al., 2009;
Wong et al., 2009; Castanedo et al., 2010). Its unique absorption, dis-
tribution, metabolism, and excretion properties contribute to vismodegib’s
prolonged circulating half-life (Wong et al., 2010; Giannetti et al.,
2011; Graham et al., 2011a, 2012). Here we report on studies to determine
the mechanism of formation of three ring-opened metabolites formed
as a result of metabolism of the pyridine moiety.
In addition to simple oxidative products on the pyridine and
chlorophenyl rings to form phenol-type metabolites, three pyridine
ring cleavage metabolites were found in excreta (Yue et al., 2011)
(Fig. 1). The literature precedence for such a reaction is limited. These
metabolites were designated M9, M13, and M18 and were charac-
We previously reported the identification and characterization of
vismodegib’s metabolites (Wong et al., 2009, 2010; Graham et al.,
2011b; Yue et al., 2011). After a single dose of 14C-vismodegib, the
radioactivity was predominately excreted in feces (80%–90% in rat, dog,
and healthy human subjects), partly as unabsorbed material. Urine
accounted for less than 5% of the dose in all species tested. The same
terized using high-resolution mass spectrometry (MS) and, in the case
1
studies showed that absorbed vismodegib underwent extensive oxidative of M13, also by H NMR (Yue et al., 2011). Different amounts of
these metabolites were found in the excreta of each species. In the rat,
M9 was 4.4% of the administered dose, and M18 and M13 were
barely detectable (,1%), whereas in dog and human, M13 was the
major metabolite, making up 10% and 3% of the dose, respectively.
1Current affiliation: Novartis Institutes for BioMedical Research, 4560 Horton
St, Emeryville, CA 94608
S.C.K., Q.Y., and S.M. contributed equally to this work.
M18 constituted 4.5% and 2% of the dose in dogs and humans,
respectively, and M9 was not detected in either species.
dx.doi.org/10.1124/dmd.113.055715.
s This article has supplemental material available at dmd.aspetjournals.org.
ABBREVIATIONS: ACN, acetonitrile; ABT, 1-aminobenzyltriazole; FA, formic acid; MS, mass spectrometry; TBA, thiobarbituric acid.
343

344
Khojasteh et al.
Fig. 2. Chemical structure of (13C2,15N)vismodegib.
obtained from Santa Cruz Biotechnology (Dallas, TX). Sprague-Dawley rat (Lot
62547), cynomolgus monkey (Lot 61370), and human (Lot 38289) liver
microsomes were obtained from BD Biosciences (San Jose, CA). Vismodegib,
(
13C2,15N)vismodegib, N-oxide of vismodegib, and M3 (Supplemental Method)
were synthesized at Genentech, Inc. (South San Francisco, CA).
In Vitro Studies
Liver Microsomal Incubations. Vismodegib (5 mM), N-oxide of vismode-
gib (5 mM), or M3 (5 mM), was incubated in liver microsomes (1.0 mg/ml),
NADPH (1 mM), and MgCl2 (2 mM) in potassium phosphate buffer (100 mM,
pH 7.4) with or without ABT (1 mM). Vismodegib (5 mM) was also incubated in
cynomolgus monkey liver microsomes (1.0 mg/ml), NADPH (1 mM), and
MgCl2 (2 mM) in potassium phosphate buffer (100 mM, pH 7.4) with
methoxylamine (5 mM) or thiobarbituric acid (5 mM) or incubation mixture
Fig. 1. Chemical structures of vismodegib, M3, M9, M13, and M18.
To probe the mechanisms for formation of these pyridine ring cleavage containing approximately 50:50 18O:16O-water. The liver microsomes were
warmed at 37°C for 5 minutes before reaction initiation by addition of the test
article and cofactor mixture. The incubation lasted 60 minutes at 37°C and was
quenched with three volumes of cold acetonitrile. Samples were centrifuged at
3000g for 5 minutes, and the supernatant was analyzed as described as follows.
Hydrolysis of M18. An aliquot of a rat bile sample was extracted with ethyl
acetate, and the organic phase was removed and dried as previously reported
(Yue et al., 2011). The sample was reconstituted in 18O-water with or without
sulfuric acid (0.5 N). The mixture was incubated for 1 hour at 50°C and
extracted with ethyl acetate. The organic fraction was dried and reconstituted in
the mobile phase before analysis.
HepatoPac Assays. Vismodegib (10 mM) was incubated with cocultured
HepatoPac as described previously by Wang et al. (2010) for 7 days.
Micropatterned cocultures were generated from Sprague-Dawley rat, beagle
dog, cynomolgus monkey, and human hepatocytes. Cultured hepatocytes for all
species except dog were cryopreserved, whereas dog hepatocytes were freshly
isolated. At the end of the incubation, cell viabilities were assessed using the
Cell Titer-Glo Luminescent cell viability assay (Promega, Madison, WI)
according an established protocol (Crouch et al., 1993). The culture medium
was removed from the well and analyzed as described as follows.
products, several studies were carried out, including in vitro studies using
liver microsomes and long-lived hepatocyte cultures. Except in the
monkey, the very low intrinsic clearance in rat, dog, and human liver
microsomes posed a significant challenge for effectively studying the
formation of these metabolites in vitro; therefore, we experimented with
various systems. These studies included ex vivo rat liver perfusion and
in vivo studies under various conditions, including pretreatment with
1-aminobenzotriazole (ABT; a broad cytochrome P450 inactivator).
Stable-labeled vismodegib with two 13C atoms and one 15N atom on the
pyridine [(13C2,15N)vismodegib] (Fig. 2) was synthesized to examine the
source of the atoms that constitute the cleaved side chain. In addition, we
investigated whether the other oxidative products are the precursors for
the formation of these cleaved products.
Materials and Methods
Chemicals
High-performance liquid chromatography grade acetonitrile (ACN), ammo-
nium formate, ammonium hydroxide, diethyl ether, ethyl acetate, formic acid
(FA), methanol, and water were purchased from either Mallinckrodt Baker, Inc.
(Phillipsburg, NJ) or EM Science (Gibbstown, NJ). Ammonium acetate
(analytical grade), 1-aminobenzyltriazole (ABT), methoxylamine, thiobarbituric
Ex Vivo Studies
Rat Liver Perfusion. Rats (n = 2) were pretreated i.p. 2 hours before dosing of
the test article with or without ABT at 100 mg/kg, using 50 mg/ml solution (dose
acid (TBA), nicotinamide adenine dinucleotide phosphate (reduced form; volume 2 ml/kg). Vismodegib or M3 (each at 20 mM) was perfused separately
NADPH), bacitracin, neomycin sulfate, and streptomycin sulfate were purchased with isolated rat livers for 3 hours. The isolated liver perfusion was performed
from Sigma Chemical Co. (St. Louis, MO). 18O-Water (97% isotopic purity) was as described previously (Mehvar and Reynolds, 1995). Briefly, adult male

Pyridine Ring Cleavage of Vismodegib
Sprague-Dawley rats were anesthetized by i.p. injection of a ketamine/xylazine
345
TABLE 1
cocktail and, after opening the abdominal cavity, the hepatic portal vein was
catheterized with a 14-gauge i.v. catheter to serve as the inlet. After cutting the
diaphragm, another i.v. catheter, inserted into the thoracic inferior vena cava,
served as the outlet. The liver was then excised and transferred to a temperature-
controlled perfusion tray. The isolated livers were perfused using a commercial
perfusion apparatus MX Perfuser II from MX International (Aurora, CO) in
a recirculating manner. The perfusate consisted of a Krebs-Henseleit bicarbonate
buffer (pH 7.35–7.45) containing glucose (5.0 g/l), bovine serum albumin (4%,
w/v), and bovine red blood cells (20%, v/v) delivered at a flow rate of 12 ml/min.
The perfusate was oxygenated with an O2:CO2 (95:5) mixture for at least 30
minutes before entering the liver. Perfusate and bile samples were collected, and
the liver was harvested. Liver samples were homogenized with a 4-fold volume of
water, and aliquots of liver homogenates and the perfusate were subjected to liquid-
liquid extraction with ethyl acetate. The extracts were dried down and reconstituted
in 30% ACN in water followed by centrifugation at 3000g for 2 minutes. Bile
samples were centrifuged at 3000g at 4°C for 10 minutes to remove undissolved
solids. The viability of the liver was confirmed through the overall macroscopic
appearance of the liver, the color of the inlet and outlet perfusate, and the extent of
perfusate and bile flow. Samples were analyzed as described as follows.
Generation of the cleaved metabolites M9, M13, and M18, from incubations
of vismodegib with liver microsomes, hepatocytes, and HepatoPac
from various species
Matrix
Species
Rat
Dog
Monkey
Cofactor
M9
M13
M18
Liver microsomes
+N
+N
-N
+N
+N+ABT
+N
NA
NA
NA
NA
2
2
2
+
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
+
2
2
2
2
+
Human
Rat
Dog
Monkey
Human
HepatoPac
+
2, No formation of metabolite; +, formation of metabolite; ABT, 1-aminobenzotriazole; +N,
with NADPH; 2N, without NADPH; NA, not applicable.
cleaved products were detected. On the other hand, in monkey liver
microsomes, M9 and M18 were detected, but M13 was not. Formation
of these metabolites was NADPH-dependent and decreased by more
than 90% after ABT preincubation, suggesting that cytochrome P450
enzymes are involved in their formation. We also examined whether
other metabolites could be the precursor for the formation of the open-
ring metabolite. For this reason, N-oxide vismodegib and M3 were
incubated in cross-species liver microsomes but did not result in the
formation any of the pyridine-cleaved metabolites.
In Vivo Studies
Stable-Labeled Vismodegib In Vivo Metabolites. One rat was orally
administered a single dose of stable-labeled vismodegib at a target dose of
50 mg/kg formulated in 0.5% methylcellulose and 0.2% Tween 80 in reverse
osmosis water (pH 3.0). The animal was treated as described previously (Wang
et al., 2010). Feces and urine were collected from metabolic cages on dry ice at
0–8, 8–24, and 24–48 hours after the dose and stored at 270°C before analysis.
In monkey liver microsomes, to test the formation of aldehyde in-
termediate, aldehyde trapping agents (methoxylamine and TBA) were
Liquid Chromatography-Mass Spectometry Analysis
The liquid chromatography–mass spectrometry system consisted of an Accela coincubated with vismodegib. With methoxylamine, one trapped metabolite
UPLC (Thermo Fisher Scientific, San Jose, CA) and an LTQ-Orbitrap Velos
mass spectrometer (Thermo Fisher Scientific). Separation was achieved on
a Hypersil Gold C18 column (100 Â 2.1 mm, 1.9 mm; Thermo Fisher Scientific)
at a flow rate of 0.4 ml/min. A binary mobile phase consisted of water with 0.1%
formic acid (A) and acetonitrile with 0.1% formic acid (B). The high-
performance liquid chromatography gradient was initiated at 5% B for 1 minute,
increased to 20% B in 11 minutes, followed by a gradient to 40% B over
9 minutes, and then changed to 60% over 2.9 minutes. The gradient was rapidly
ramped to 95% B in 0.1 minute and remained at 95% B over 1.8 minutes. The
gradient was then returned to the initial condition of 5% B in 0.2 minute and
was detected with a protonated molecular ion at m/z 511.06021, indicating
that its empirical formula is C21H21Cl2N4O5S (mass accuracy, 20.41 ppm).
The product ion mass spectrum of m/z 511 contains fragment ions at m/z
480, 479, 449, and 448 (Fig. 3A; Supplemental Fig. 1). The fragment ions
at m/z 480 and 479 resulted from loss of methoxyl radical and methanol
from the protonated molecular ion, respectively. The fragment ions at m/z
449 and 448 were further loss of methoxyl radical and methanol from the
fragment ion at m/z 480, which indicated two methoxylamine molecules
reacted with two aldehydes. The MS3 spectrum of m/z 449 resulted in
equilibrated at 5% B for 4 minutes before the next injection. The total run time fragment ions at m/z 431, 422, 414, 395, 370, 232, and 217 (Fig. 3A;
was 30 minutes. A longer gradient was used to achieve better separation of
methoxylamine or thiobarbituric acid trapped adducts from the incubation of
vismodegib with cynomolgus monkey liver microsomes. The gradient was
initiated at 5% B for 2 minutes, increased to 25% B in 38 minutes, followed by
a gradient to 45% B over 10.5 minutes. The gradient was rapidly ramped to 95%
B in 0.5 minute and remained at 95% B over 4.8 minutes. The gradient was then
returned to the initial condition of 5% B in 0.2 minute and equilibrated at 5% B
for 4 minutes before the next injection. The total run time was 60 minutes.
The LTQ-Orbitrap mass spectrometer was operated in the positive ion mode
with electrospray ionization source. The ion source conditions were as follows:
Supplemental Fig. 1A). The characteristic fragment ion at m/z 217 was the
same as unchanged vismodegib, suggesting that 2-chloro-4-methylsulfonyl
benzene moiety remained unchanged. The fragment ion at m/z 370 resulted
from loss of methylsulfonyl benzene radical (78.93538 Da). Mass
accuracies for all fragment ions were within 6 5 ppm of the proposed
structures, and the mechanism of its formation is shown in Supplemental
Fig. 2. TBA reacted with an a,b-unsaturated aldehyde intermediate to form
a cyclic conjugate indicative of a ring-opened intermediate (Fig. 3B;
Supplemental Fig. 1B). It had a protonated molecular ion at m/z 578.99528,
spray voltage, 4.0 kV; sheath and auxiliary gas (N2) flow, 60 and 10 (arbitrary which is consistent with a molecular formula of C23H16Cl2N4O6S2 (mass
units), respectively; capillary temperature, 300°C; and source heater temper-
ature, 150°C. The scan event cycle consisted of a full-scan mass spectrum at
a resolving power of 30,000 and the corresponding data-dependent tandem MS
(MS/MS) and multistage mass spectrometry (MS3) scans acquired at a resolving
power of 7500 with collision energy of 30% and an isolation mass window
of 2 Da. To separate the isotope peaks of 37Cl- from 18O-, the full-scan mass
spectra were acquired at a mass resolving power of 100,000. Data were pro-
cessed using Xcalibur software 2.2 SP1 (Thermo Fisher Scientific).
accuracy 21.43 ppm). The product ion mass spectrum of m/z 579 produced
a fragment ion at m/z 551 by loss of a carbon monoxide. The MS3
spectrum of m/z 551 resulted in fragment ions at m/z 534, 475, and 449
by further loss of ammonia, thiourea, and 4-thioxo-1,3-diazetidin-2-one,
respectively. Mass accuracies for all fragment ions were within 64 ppm
of the proposed structures and the mechanism of its formation is shown
in Supplemental Fig. 2.
Vismodegib was also incubated with monkey liver microsomes in
50:50 of 16O-water:18O-water, which resulted in incorporation of one
18O atom into the carboxylic acid of M18. This was evident by the
Results
The results from vismodegib incubations with liver microsomes were doublet peaks at m/z 471.01725 and 473.02141 with approximately
the formation of three pyridine ring–opened metabolites (M9, M13, and equal intensity (Supplemental Fig. 3). Similarly, the presence of
M18; Table 1). In liver microsomes from rat, dog, and human, no doublet peaks at m/z 443.02237 and 445.02659 suggested that one

346
Khojasteh et al.
Fig. 3. Interpretation of fragment ions of (A)
methoxylamine-trapped conjugates at molecu-
lar ion at m/z 511 and (B) thiobarbituric acid
(TBA) trapped conjugate at m/z 579. Vismode-
gib (5 mM) was incubated with monkey liver
microsomes (1 mg/ml) fortified with NADPH
and methoxylamine (5 mM) or TBA (5 mM).
18O- atom was incorporated into M9. Using high-resolution MS at dissociation. MA eluted at 17.6 minutes (whereas M9 eluted at 16.6
a mass resolving power of 100,000, the contribution of 37Cl- from minutes) and had a protonated molecular ion at m/z 443.0225,
18O- in M9 and M18 were separated (Supplemental Fig. 3).
indicating that its empirical formula is C18H17O5N2Cl2S (21.06 ppm;
Vismodegib was incubated for 7 days in HepatoPac cocultures. Fig. 5). The product ion scan at m/z 443 from MA produced fragment
During this period, no changes in cell viability of the cocultures were ions at m/z 425.99814 (loss of NH3), 425.01404 (loss of H2O),
observed. Supernatants were profiled, and the metabolites formed 397.01899 (a loss of formic acid), and 385.01910 (loss of oxiran-2-
were matched to in vivo samples on the basis of accurate masses and one) (Fig. 6). The MS3 spectrum of m/z 425 gave characteristic
retention times. M18 was detected in cocultures containing human and fragment ions at m/z 216.97319 and 154.99118, which suggests that
monkey hepatocytes but not in rat or dog cocultures. Formation of this the 2-chloro-4-methylsulfonyl benzene moiety remained unchanged
metabolite was greatest in monkeys (normalized to 100%) and was and that the modification occurred on the pyridine-chloroaniline
36% in humans on the basis of the MS signal.
To examine whether M18 was the primary metabolite leading to indicates that MA has an aliphatic carboxylic acid moiety.
formation of M13 and M9, a bile sample from a rat dosed with Stable-labeled vismodegib was synthesized by incorporating two
moiety. The loss of formic acid from m/z 443 to form m/z 397
vismodegib was incubated under acidic conditions. The ethyl acetate 13C atoms at the C-2 and C-6 positions plus one 15N atom on the
extract from rat bile, which contained M18 and M13 among other pyridine (Fig. 2). Metabolites excreted in feces from intact rats dosed
metabolites of vismodegib, was dried and incubated in 18O-water in the orally with this labeled vismodegib were compared with those excreted
presence of 0.5 N sulfuric acid for 1 hour at 50°C. Under these from rats dosed with vismodegib in previous studies. Table 2
conditions, one 18O atom was incorporated into the carboxylic acid of summarizes the findings from this study. As expected, the molecular
M13 and no changes to M9 (Fig. 4). For M13 formed from the ion of labeled vismodegib is 3 Da higher than that of nonlabeled
incorporation of one 18O atom, we observed the generation of m/z vismodegib. The fragment ions from labeled vismodegib also include
463.03660 ([M + NH4]+) for addition of 18O versus m/z 461.03250 for the stable labels. M9 is the amide metabolite, and the interpretation of
the addition of 16O. Using high-resolution MS, we were also able to its collision-induced dissociation is shown in Fig. 7. The difference of 2
separate the contribution of 35Cl/37Cl from 16O/18O. For M13 generated Da between M9 and stable-labeled M9 is indicative of a loss of one of
from M18, two atoms of 18O atoms were incorporated, resulting in the the labeled atoms from the original parent. The neutral loss of ammonia
molecular ion at m/z 465.04084. Of the two 18O atoms incorporated into (NH3) from M9 and its stable-labeled analog lead to the formation of
M13, one was at the carboxylic acid and one other incorporated into the m/z 426 and 427, respectively. This suggests that the nitrogen atom on
ketone (Fig. 4). With high-resolution MS, the mass difference showing the amide was the same nitrogen atom originally located on the pyridine
the incorporation of two 18O atoms confirmed the formation of M13 ring. The other fragment ions confirmed that the labeled carbon retained
from M18. In summary, under these experimental conditions, M18 did on the molecule was at the C-2 position, suggesting that the labeled
lead to the formation of M13 but not M9.
carbon at the C-6 position was lost. Metabolite M13 is a carboxylic acid
To understand the role of the liver in the formation of M9, M13, and metabolite, and there is a 1 Da difference between M13 and its stable-
M18, rat liver perfusion studies were conducted. Vismodegib was labeled analog. Based on the fragment ion interpretation, only one label
continuously perfused to the isolated rat liver for 3 hours. Detectable levels was retained at the C-2 position. M18 has a molecular ion consistent
of M18, but not M13 or M9, plus other oxidative metabolites of with all three stable-labeled atoms being retained in the compound. The
vismodegib that were reported previously were found in the perfusate and summary of these findings is described in Fig. 1.
bile samples. Interestingly, another cleaved metabolite, MA, which had not
been previous detected, was discovered. Formation of these metabolites
(M18 and MA) decreased after the pretreatment of rats with ABT (Fig. 5).
Discussion
MA has the same elemental composition as M9, but it has a different
In previous in vivo studies, vismodegib was shown to form three
retention time and fragment ion spectrum under collision-induced pyridine ring-opened metabolites M9, M13, and M18 (Graham et al.,

Pyridine Ring Cleavage of Vismodegib
347
Fig. 4. Molecular ion of M13 ([M+NH4]+)
from M18 and M13 present in a rat bile sample
after incubation with H218O under acidic
conditions.
2011b; Yue et al., 2011). These metabolites were excreted mainly in understand the mechanisms of their formation as well as P450
feces, with the greatest amount in the dog (15%), followed by similar involved in this biotransformation.
amounts in the excreta of human and rat (;5%). M18 and M13 were
The biotransformation of a pyridine moiety resulting in a ring-opened
detected in all species, but M9 was detected only in rats. Here we metabolite is rarely reported. Pyridine is an aromatic heterocycle that is
investigated the formation of these ring-opened metabolites to slightly basic in nature and is commonly found in key biologic systems
Fig. 5. Total ion chromatogram for vismodegib
and ring-cleaved metabolites from rat perfusate
in the (A) absence and (B) presence of ABT.

348
Khojasteh et al.
Fig. 6. Product ion spectra for metabolite MA
at m/z 443 with (A) MS2 and (B) MS3.
such as electron carriers (NADH and NADPH) and nucleotides, as well
Vismodegib provided an opportunity to further understand the
as in many pharmaceutical drugs. Pyridine moiety, compared with mechanism of pyridine-ring cleavage; however, the compound posed
phenyl rings, is attractive to pharmaceutical scientists because it imparts a particular challenge for mechanistic studies. In vitro studies had
increased solubility and metabolic stability against cytochrome P450 limited utility because of the compound’s low in vitro hepatic
enzymes (St. Jean and Fotsch, 2012). This stability is due mainly to the clearance in relevant species (rats, dogs, and humans); accordingly,
electronegativity of nitrogen that makes the ring electron deficient and, these liver microsome studies generated small amount of metabolites.
therefore, less prone to electron abstraction. Other reports on The monkey was the only species that generated an appreciable level
metabolism of pyridine include N-oxidation, N-methylation (Damani of the ring-opened metabolites; however, the monkey was not one of
et al., 1986), and N-glucronidation (Lin, 1999; Wiener et al., 2004). On the species studied for metabolism and preclinical toxicity because of
the other hand, few cases of ring opening of pyridine have been the very high clearance. The mechanistic studies in monkey liver
reported, including microbial ring opening of pyridine derivatives microsomes still provided the opportunity to further examine the
(Kaiser et al., 1996), the formation of muconaldehyde from benzene mechanism of formation of the ring open metabolites. Aldehyde
mediated by cytochrome P450 (Latriano et al., 1986), and the formation trapping agents methoxylamine and TBA confirmed the presence of
of E-2-(acetaminomethylene) succinate from a vitamin B6 degradant two aldehyde moieties in an intermediate formed in the process of
(McCulloch et al., 2009). It is noteworthy that there are reports, albeit generating the open-ring metabolites (Fig. 3). 18O-Water demonstrated
rare, on other six-membered aromatic heterocyclic ring-opening bio- that after the formation of ring-opened intermediate, there is an
transformation, such as pyrimidine ring–cleaved metabolites of PF- opportunity for exchange with water, most likely by aldehyde. The
00734200, a dipeptidyl peptidase inhibitor (Sharma et al., 2012).
ring-opened metabolites were also generated by coculture hepatocytes,
which can be maintained for a long incubation period. Even in this
case, only small amounts of the M18 metabolite was generated in the
human and monkey. We, therefore, complemented the in vitro data
with studies using stable-labeled compound as well as ex vivo liver
perfusion and in vivo studies.
TABLE 2
Major collision-induced product ions of vismodegib and the stable-labeled
vismodegib and their pyridine cleaved metabolites in feces of rats dosed orally with
50 mg/kg
The use of stable-labeled (13C2,15N)vismodegib (Fig. 2) allowed us
to understand the source of the atoms in the ring-opened metabolites.
Metabolites from the stable-labeled compound in the rat excreta
confirmed that, as expected, M18 retained all three stable-labeled
atoms, and M13 retained only one at the C-2 position. Surprisingly,
M9 retained 13C atom at the C-2 position as well as the 15N at the
primary amine. This means that the C-6 carbon (Fig. 1) was lost,
followed by rearrangement of the pyridine nitrogen to form a terminal
amide.
Analyte
[M + H]+
Major Fragment Ionsa
m/z
m/z
Vismodegib
Labeled vismodegib
M9
Labeled M9
M13
Labeled M13
M18
Labeled M18
421
424
443
445
444
445
471
474
342 (100), 313, 306, 217, 188, 155, 139
345 (100), 316, 309, 217, 191, 155, 139
426 (100), 398, 380, 217, 155
427 (100), 399, 381, 217, 155
426 (100), 398, 380, 362, 301, 217, 155
427 (100), 399, 381, 217, 155
453 (100), 443, 426, 425, 397
456 (100), 445, 427, 399
To understand the relationship between the three ring-opened
metabolites, extracts from rat bile samples were treated with a low
concentration of acid in the presence of 16O- or 18O-water. Under
aEach base peak is labeled as 100 in parentheses.

Pyridine Ring Cleavage of Vismodegib
349
Fig. 7. Product ion spectra and interpretations
of (A) M9 and (B) stable-labeled M9 formed
from the metabolism of vismodegib and its
stable-labeled analog, respectively. The frag-
ment ions displayed in the structures are based
on theoretical masses.
these conditions, M18 converted to M13, but no or little conversion to could lead to ring-opened intermediate Ib. It is out of the scope of this
M9 was observed. These results suggest that M18 conversion could be manuscript to probe in this first step of oxidation for the formation of
acid-catalyzed to M13, whereas the same was not true for the Ib, but one could also consider that cytochrome P450 enzymes act as
formation of M9 from M18. This finding was also consistent with the a peroxidase (Vaz et al., 1998) and insert two oxygen atoms into the
previous in vivo observations. If M9 were formed only by chemical double bond between C-5 and C-6 to form a dioxetane Ia’ intermediate
degradation, then we would expect to detect M9 whenever M18 and (Fig. 8, pathway A). This is similar to the previously proposed
M13 are detected; however, M9 is detected only in rats and not in mechanism, but not proven, for the formation of muconaldehyde from
humans or dogs despite M18 and M13 being present in measurable benzene oxidation by P450 (Latriano et al., 1986). The dioxetane
quantities.
intermediate (Ia’) could rearrange to form the intermediate Ib (Fig. 8,
Because of vismodegib’s low rat hepatic intrinsic clearance, rat liver pathway B). This intermediate could then rearrange by hydrogen
perfusion studies were used to study the formation of the ring-opened cation abstraction to form a ketene intermediate Ie (pathway C).
metabolites. In addition to finding detectable levels of M18 in the Alternatively, the Ib aldehyde could hydrate to form the Ic
perfusate and bile samples, a new metabolite (MA) was identified that intermediate (pathway D), which could rearrange by hydrogen
had not been previously observed in in vivo or in vitro studies. The role abstraction and eventually lead to formation of M18 (pathway F).
of P450 enzymes was established from pretreatment of ABT in a liver Methoxylamine and TBA confirmed the presence of Ib and 18O-water
perfusion study. This pretreatment resulted in significantly decrease exchange confirmed intermediates Ic and Ie. M18 could undergo
levels of M18 and MA, which is consistent with cytochrome P450 hydrolysis by loss of formic acid, leading to an amine and the
enzymes being involved in the formation of these metabolites.
generation of MA. M13 could be formed by further hydrolysis of M18
To determine whether M18 might be formed via other metabolites, or MA (pathway H or I). The formation of M9 could possibly be
M3 and N-oxide vismodegib were synthesized and dosed in rats or enzyme mediated because this metabolite is detected only in the rat,
incubated in vitro. M3 was a candidate to be the precursor to ring- despite M18 and M13 being present in measurable quantities in human
opened metabolites since this was the major oxidative metabolite and dog. We speculate that MA is the precursor for the formation of
detected in all species. N-oxide vismodegib, although not detected as M9 since the amine has the ability to be more nucleophilic. This
a metabolite, could potentially be converted to the ring-opened reaction could proceed in two possible ways. One is a nucleophilic
metabolites. Neither of these compounds generated any of the ring- attack of amine on C-5 to form 2-pyrroline. There are two
opened metabolites. In the case of M3, we detected further oxidized considerations that could make this mechanism less likely. One is
metabolites that were not detected to a significant level in rat mass that the carboxylic acid is not usually susceptive to nucleophilic attack
balance studies.
by amine under normal conditions; second, there is a chance that
We also considered the involvement of gut microflora in the 2-pyrroline loses water and forms stable pyrrole, which was not
formation of these metabolites since pyridine ring-opening bio- detected. We therefore propose that the carboxylic acid could be
transformation has been reported (Kaiser et al., 1996). Preliminary activated similarly to the amino acid conjugation reactions (Knights
studies with pretreatment of antibiotics did not result in diminished et al., 2007). Once the carboxylic acid is activated (2ROX; Fig. 8,
levels of the ring-opened metabolites. In addition, when vismodegib pathway J), amine could react with C-5 and eventually be converted
was incubated with rat feces under anaerobic conditions, the drug to M9.
concentration remained stable and no formation of any metabolite was
observed (data not shown).
In summary, because of the limitations of in vitro tools, alternative
methods, such as the use of stable-labeled vismodegib, in vivo and ex
On the basis of the results of these studies, we propose a mechanism vivo studies, were used to investigate the vismodegib mechanism of
in which M18 is the primary metabolite and the other cleaved ring cleavage in human and other preclinical species. M18 is the
products, M13 and M9, are formed from M18 via MA. The formation primary ring-opened metabolite formed in the liver through cyto-
of M18 might occur through cytochrome P450–mediated epoxidation chrome P450 biotransformation. The subsequent hydrolysis of M18
(Fig. 8, pathway A). The epoxide intermediate (Ia) could be leads to formation of MA, which could be the precursor for both M13
hydrolyzed to a dihydrodiol intermediate and further oxidized that and M9. Under slightly acidic conditions, M13 is formed from M18;

350
Khojasteh et al.
Fig. 8. Proposed metabolic cleavage of vismodegib by cytochrome P450. Reactions with methoxylamine (CH3ONH2) and TBA are most likely with Ib intermediate. Â is to
denote either hydroxyl or the modification of carboxylic acid to an activated form, perhaps similar to amino acid conjugation process via acetyl-CoA to form thioester.
Authorship Contributions
Participated in research design: Khojasteh, Yue, Ma, Chen, Liu, Takahashi,
Hop, Wong.
however, M9 was not formed from M18 at significant levels, sug-
gesting that the transformation to M9 may be enzymatic in nature.
Conducted experiments: Yue, Ma, Castanedo, Chen, Mulder, Takahashi, Ly,
Messick, Jia.
Acknowledgments
The authors acknowledge the editorial contribution of Ronitte Libedinsky.
Contributed new reagents or analytic tools: Castanedo, Chen, Takahashi.

Pyridine Ring Cleavage of Vismodegib
351
Mehvar R and Reynolds J (1995) Input rate-dependent stereoselective pharmacokinetics. Ex-
perimental evidence in verapamil-infused isolated rat livers. Drug Metab Dispos 23:637–641.
Robarge KD, Brunton SA, Castanedo GM, Cui Y, Dina MS, Goldsmith R, Gould SE, Guichert O,
Gunzner JL, and Halladay J, et al. (2009) GDC-0449-a potent inhibitor of the hedgehog
pathway. Bioorg Med Chem Lett 19:5576–5581.
Performed data analysis: Khojasteh, Yue, Ma, Lyssikatos, Takahashi.
Wrote or contributed to the writing of the manuscript: Khojasteh, Yue, Ma.
References
Scales SJ and de Sauvage FJ (2009) Mechanisms of Hedgehog pathway activation in cancer and
implications for therapy. Trends Pharmacol Sci 30:303–312.
Castanedo GM, Wang S, Robarge KD, Blackwood E, Burdick D, Chang C, Dijkgraaf GJ, Gould
S, Gunzner J, and Guichert O, et al. (2010) Second generation 2-pyridyl biphenyl amide
inhibitors of the hedgehog pathway. Bioorg Med Chem Lett 20:6748–6753.
Crouch SP, Kozlowski R, Slater KJ, and Fletcher J (1993) The use of ATP bioluminescence as
a measure of cell proliferation and cytotoxicity. J Immunol Methods 160:81–88.
Damani LA, Shaker MS, Crooks PA, Godin CS, and Nwosu C (1986) N-methylation and qua-
ternization of pyridine in vitro by rabbit lung, liver and kidney N-methyltransferases: an
S-adenosyl-L-methionine-dependent reaction. Xenobiotica 16:645–650.
Giannetti AM, Wong H, Dijkgraaf GJ, Dueber EC, Ortwine DF, Bravo BJ, Gould SE, Plise EG,
Lum BL, and Malhi V, et al. (2011) Identification, characterization, and implications of
species-dependent plasma protein binding for the oral Hedgehog pathway inhibitor vismodegib
(GDC-0449). J Med Chem 54:2592–2601.
Graham RA, Hop CE, Borin MT, Lum BL, Colburn D, Chang I, Shin YG, Malhi V, Low JA,
and Dresser MJ (2012) Single and multiple dose intravenous and oral pharmacokinetics of the
hedgehog pathway inhibitor vismodegib in healthy female subjects. Br J Clin Pharmacol 74:
788–796.
Graham RA, Lum BL, Cheeti S, Jin JY, Jorga K, Von Hoff DD, Rudin CM, Reddy JC, Low JA,
and Lorusso PM (2011a) Pharmacokinetics of hedgehog pathway inhibitor vismodegib (GDC-
0449) in patients with locally advanced or metastatic solid tumors: the role of alpha-1-acid
glycoprotein binding. Clin Cancer Res 17:2512–2520.
Graham RA, Lum BL, Morrison G, Chang I, Jorga K, Dean B, Shin YG, Yue Q, Mulder T,
and Malhi V, et al. (2011b) A single dose mass balance study of the Hedgehog pathway
inhibitor vismodegib (GDC-0449) in humans using accelerator mass spectrometry. Drug Metab
Dispos 39:1460–1467.
Kaiser JP, Feng Y, and Bollag JM (1996) Microbial metabolism of pyridine, quinoline, acridine,
and their derivatives under aerobic and anaerobic conditions. Microbiol Rev 60:483–498.
Knights KM, Sykes MJ, and Miners JO (2007) Amino acid conjugation: contribution to the
metabolism and toxicity of xenobiotic carboxylic acids. Expert Opin Drug Metab Toxicol 3:
159–168.
Sharma R, Sun H, Piotrowski DW, Ryder TF, Doran SD, Dai H, and Prakash C (2012) Me-
tabolism, excretion, and pharmacokinetics of ((3,3-difluoropyrrolidin-1-yl)((2S,4S)-4-(4-
(pyrimidin-2-yl)piperazin-1-yl)pyrrolidin-2-yl)methanone, a dipeptidyl peptidase inhibitor, in
rat, dog and human. Drug Metab Dispos 40:2143–2161.
St. Jean DJ, Jr and Fotsch C (2012) Mitigating heterocycle metabolism in drug discovery. J Med
Chem 55:6002–6020.
Vaz AD, McGinnity DF, and Coon MJ (1998) Epoxidation of olefins by cytochrome P450:
evidence from site-specific mutagenesis for hydroperoxo-iron as an electrophilic oxidant. Proc
Natl Acad Sci USA 95:3555–3560.
Wang WW, Khetani SR, Krzyzewski S, Duignan DB, and Obach RS (2010) Assessment of
a micropatterned hepatocyte coculture system to generate major human excretory and circu-
lating drug metabolites. Drug Metab Dispos 38:1900–1905.
Wiener D, Doerge DR, Fang JL, Upadhyaya P, and Lazarus P (2004) Characterization of
N-glucuronidation of the lung carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL)
in human liver: importance of UDP-glucuronosyltransferase 1A4. Drug Metab Dispos 32:72–79.
Wong H, Chen JZ, Chou B, Halladay JS, Kenny JR, La H, Marsters JC, Jr, Plise E, Rudewicz PJ,
and Robarge K, et al. (2009) Preclinical assessment of the absorption, distribution, metabolism
and excretion of GDC-0449 (2-chloro-N-(4-chloro-3-(pyridin-2-yl)phenyl)-4-(methylsulfonyl)
benzamide), an orally bioavailable systemic Hedgehog signalling pathway inhibitor. Xen-
obiotica 39:850–861.
Wong H, Theil FP, Cui Y, Marsters JC, Jr, Khojasteh SC, Vernillet L, La H, Song X, Wang H,
and Morinello EJ, et al. (2010) Interplay of dissolution, solubility, and nonsink permeation
determines the oral absorption of the Hedgehog pathway inhibitor GDC-0449 in dogs: an
investigation using preclinical studies and physiologically based pharmacokinetic modeling.
Drug Metab Dispos 38:1029–1038.
Yue Q, Chen YH, Mulder T, Deese A, Takahashi R, Rudewicz PJ, Reynolds M, Solon E, Hop
CE, and Wong H, et al. (2011) Absorption, distribution, metabolism, and excretion of
[¹⁴C]GDC-0449 (vismodegib), an orally active hedgehog pathway inhibitor, in rats and dogs:
a unique metabolic pathway via pyridine ring opening. Drug Metab Dispos 39:952–965.
Latriano L, Goldstein BD, and Witz G (1986) Formation of muconaldehyde, an open-ring me-
tabolite of benzene, in mouse liver microsomes: an additional pathway for toxic metabolites.
Proc Natl Acad Sci USA 83:8356–8360.
Lin JH (1999) Role of pharmacokinetics in the discovery and development of indinavir. Adv Drug
Deliv Rev 39:33–49.
McCulloch KM, Mukherjee T, Begley TP, and Ealick SE (2009) Structure of the PLP degradative
enzyme 2-methyl-3-hydroxypyridine-5-carboxylic acid oxygenase from Mesorhizobium loti
MAFF303099 and its mechanistic implications. Biochemistry 48:4139–4149.
Address correspondence to: Dr. Cyrus Khojasteh, Drug Metabolism and
Pharmacokinetics, Genentech, Inc., 1 DNA Way, MS 412a, South San Francisco,
CA 94080. E-mail: pars@gene.com