In vitro metabolism of 17-(dimethylaminoethylamino)-17- demethoxygeldanamycin in human liver microsomes

Article abstract of DOI:10.1124/dmd.110.036418

The objective of this study was to investigate the oxidative metabolism pathways of 17-(dimethylaminoethylamino)-17-demethoxy-geldanamycin (17-DMAG), a geldanamycin (GA) derivative and 90-kDa heat shock protein inhibitor. In vitro metabolic profiles of 17-DMAG were examined by using pooled human liver microsomes (HLMs) and recombinant CYP450 isozymes in the presence or absence of reduced GSH. In addition to 17-DMAG hydroquinone and 19-glutathionyl 17-DMAG, several oxidative metabolites of 17-DMAG were detected and characterized by liquid chromatography-tandem mass spectrometry. Different from previously reported primary biotransformations of GA and GA derivatives, 17-DMAG was not metabolized primarily through the reduction of benzoquinone and GSH conjugation in HLMs. In contrast, the primary biotransformations of 17-DMAG in HLMs were hydroxylation and demethylation on its side chains. The most abundant metabolite was produced by demethylation from the methoxyl at position 12. The reaction phenotyping study showed that CYP3A4 and 3A5 were the major cytochrome P450 isozymes involved in the oxidative metabolism of 17-DMAG, whereas CYP2C8, 2D6, 2A6, 2C19, and 1A2 made minor contributions to the formation of metabolites. On the basis of the identified metabolite profiles, the biotransformation pathways for 17-DMAG in HLMs were proposed. Copyright

Full text of DOI:10.1124/dmd.110.036418

0090-9556/11/3904-627–635$20.00
DRUG METABOLISM AND DISPOSITION
Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics
DMD 39:627–635, 2011
Vol. 39, No. 4
36418/3672095
Printed in U.S.A.
In Vitro Metabolism of 17-(Dimethylaminoethylamino)-17-
demethoxygeldanamycin in Human Liver Microsomes
Nan Zheng, Peng Zou, Shaomeng Wang, and Duxin Sun
Department of Pharmaceutical Sciences, College of Pharmacy, University of Michigan, Ann Arbor, Michigan (N.Z., P.Z., D.S.);
and Departments of Internal Medicine and Medicinal Chemistry and Comprehensive Cancer Center, University of Michigan,
Ann Arbor, Michigan (S.W.)
Received September 23, 2010; accepted December 22, 2010
ABSTRACT:
The objective of this study was to investigate the oxidative metab- DMAG was not metabolized primarily through the reduction of
olism pathways of 17-(dimethylaminoethylamino)-17-demethoxy- benzoquinone and GSH conjugation in HLMs. In contrast, the pri-
geldanamycin (17-DMAG), a geldanamycin (GA) derivative and 90- mary biotransformations of 17-DMAG in HLMs were hydroxylation
kDa heat shock protein inhibitor. In vitro metabolic profiles of and demethylation on its side chains. The most abundant metab-
17-DMAG were examined by using pooled human liver microsomes
(HLMs) and recombinant CYP450 isozymes in the presence or 12. The reaction phenotyping study showed that CYP3A4 and 3A5
absence of reduced GSH. In addition to 17-DMAG hydroquinone were the major cytochrome P450 isozymes involved in the oxida-
olite was produced by demethylation from the methoxyl at position
and 19-glutathionyl 17-DMAG, several oxidative metabolites of tive metabolism of 17-DMAG, whereas CYP2C8, 2D6, 2A6, 2C19,
17-DMAG were detected and characterized by liquid chromatog- and 1A2 made minor contributions to the formation of metabolites.
raphy-tandem mass spectrometry. Different from previously re- On the basis of the identified metabolite profiles, the biotransfor-
ported primary biotransformations of GA and GA derivatives, 17- mation pathways for 17-DMAG in HLMs were proposed.
Introduction
et al., 2007). Although GA and 17-AAG are structurally similar (see
Fig. 1), their metabolite profiles in liver microsomes are different
(Lang et al., 2007). GA is primarily (40–73%) reduced into geldana-
mycin hydroquinone (GAH2) (Lang et al., 2007; Guo et al., 2008).
During exposure to oxygen, GAH2 slowly reverts to GA. In the
presence of reduced GSH, more than 50% of GA is rapidly converted
into 19-glutathionyl geldanamycin hydroquinone (Cysyk et al., 2006;
Lang et al., 2007). No significant amount of oxidative metabolites of
GA in the incubations with human liver microsomes (HLMs) has been
detected (Lang et al., 2007). The metabolic pathways of 17-AAG in
liver microsomes are controversial. Guo et al. (2008) reported that
quinone/hydroquinone conversion was the primary metabolism mode
of 17-AAG and 17-DMAG in microsomal preparation. In the presence
of reduced GSH, 15% of 17-AAG was conjugated with GSH after
incubation in liver microsomes for 24 h. However, Lang et al. (2007)
observed that only 2% of 17-AAG was reduced into hydroquinone in
HLMs, and no significant amount of 19-GSH conjugate of 17-AAG
was detected in HLMs in the presence of 5 mM GSH. Furthermore,
they found that, different from GA, 17-AAG in HLMs primarily
underwent oxidative metabolism on the 17-allylamino side chain to
form 17-aminogeldanamycin (17-AG) (see Fig. 1) and 17-(2Ј,3Ј-
dihydroxypropylamino)-geldanamycin, which was consistent with a
previous study (Egorin et al., 1998).
The 90-kDa heat shock protein (Hsp90) is a molecular chaperone to
mediate the folding, activation, and assembly of many oncogenic
client proteins, which stimulate cancer cell growth (McIlwrath et al.,
1996). Geldanamycin (GA) is an Hsp90 inhibitor that binds to Hsp90
and disrupts the interaction between Hsp90 and its client proteins (An
et al., 1997). This disruption depletes the oncogenic proteins and
results in antitumor activity. To develop potent antitumor agents, a
number of GA derivatives have been synthesized and characterized
biologically. Among GA derivatives, 17-(allylamino)-17-demethoxy-
geldanamycin (17-AAG) and 17-(dimethylaminoethylamino)-17-de-
methoxygeldanamycin (17-DMAG) have been introduced into clini-
cal trials (Glaze et al., 2005).
Both GA and 17-AAG are known to undergo extensive metabolism
(Egorin et al., 1998; Musser et al., 2003; Guo et al., 2005, 2006; Lang
This work was supported in part by the National Institutes of Health National
Cancer Institute [Grants R01-CA120023, R21-CA143474]; and the University of
Michigan Cancer Center Research [Grant Munn].
N.Z. and P.Z. contributed equally to this work.
Article, publication date, and citation information can be found at
http://dmd.aspetjournals.org.
doi:10.1124/dmd.110.036418.
ABBREVIATIONS: Hsp90, 90-kDa heat shock protein; 17-AAG, 17-(allylamino)-17-demethoxygeldanamycin; 17-AAG-SG, 19-glutathionyl 17-
(allylamino)-17-demethoxygeldanamycin; 17-AAGH2, 17-(allylamino)-17-demethoxygeldanamycin hydroquinone; 17-AG, 17-aminogeldanamycin;
17-DMAG, 17-(dimethylaminoethylamino)-17-demethoxy geldanamycin; 17-DMAG-SG, 19-glutathionyl 17-(dimethylaminoethylamino)-17-
demethoxy geldanamycin; 17-DMAGH2, 17-(dimethylaminoethylamino)-17-demethoxygeldanamycin hydroquinone; EPI, enhanced product
ion; GA, geldanamycin; GAH2, geldanamycin hydroquinone; HLM, human liver microsome; HPLC, high-performance liquid chromatography;
LC-MS/MS, liquid chromatography-tandem mass spectrometry; Mr, protonated molecular ion; MRM, multiple reaction monitoring; MS, mass
spectrometry; P450, cytochrome P450.
627

628
ZHENG ET AL.
values for the formation of M1 and M3 from 17-DMAG, various concen-
17-DMAG is much more metabolically stable than 17-AAG be-
trations of 17-AAG (0.5–100 ␮M) were incubated in 0.5 mg/ml HLMs for
30 min in triplicate. The formation rates of the metabolites were fitted with
the Michaelis-Menten equation using WinNonlin (Pharsight, Mountain
View, CA).
Incubations with HLMs for Metabolite Identification. 17-DMAG was
incubated with HLMs in 0.6 ml of phosphate buffer (0.1 M) at 37°C. The final
concentrations of microsomes, ␤-NADPH, phosphate buffer, and MgCl2 were
the same as described in stability assay. The initial concentration of 17-DMAG
was 10 ␮M. Two different negative controls were prepared simultaneously by
cause of the limited oxidative metabolism on 17-dimethylaminoeth-
ylamino side chain (Glaze et al., 2005). Compared with 17-AAG,
17-DMAG exhibits a longer terminal half-life of 16 to 19 h (Hwang
et al., 2006; Moreno-Farre et al., 2006) (4 h for 17-AAG) and a lower
total clearance of 7.4 to 17.7 l/h (Hwang et al., 2006; Moreno-Farre et
al., 2006) (36 l/h for 17-AAG) in humans. Although the preclinical
(Egorin et al., 2002) and clinical (Glaze et al., 2005; Goetz et al.,
2005) pharmacokinetics of 17-DMAG have been investigated, to our
knowledge, the biotransformation information of 17-DMAG is still using boiled microsomes (100°C for 5 min) or spiking 10 ␮M 17-DMAG after
protein precipitation. Both samples and negative controls were incubated for 2
h, and the reaction was terminated with 1.2 ml of ice-cold acetonitrile to
precipitate proteins. In addition, to evaluate GSH conjugation, 17-DMAG
(10 ␮M) was incubated with reduced GSH (5 mM) and ␤-NADPH (1 mM)
in 0.1 M phosphate buffer (containing MgCl2) in the presence or absence
of HLMs (1 mg/ml) at 37°C for 2 h. After protein precipitation, samples
and negative controls were centrifuged at 14,000 rpm ϫ 5 min for LC-MS
analysis.
limited and controversial. Reduction of quinone was proposed to be
the primary metabolism of 17-DMAG in liver microsomes, and 17-
DMAG was observed to undergo more rapid GSH conjugation than
17-AAG (Guo et al., 2008). However, these findings cannot explain
the less in vivo metabolism of 17-DMAG than that of 17-AAG in
animals and humans (Musser et al., 2003; Hwang et al., 2006).
Biotransformation of GA and its derivatives is related to their
antitumor activity and toxicity. For example, the reduction of benzo-
quinone ansamycins into hydroquinone ansamycins enhanced Hsp90
LC-MS/MS. An Agilent 1200 HPLC system (Agilent Technologies, Santa
Clara, CA) was used for separation. The processed samples were injected on
inhibition (Guo et al., 2006; Lang et al., 2007), whereas GSH conju- a Zobarx SB-C18 column (2.1 ϫ 50 mm, 3.5 ␮m) (Agilent Technologies).
Mobile phase A, consisting of 0.1% formic acid and 10 mM ammonium
formate in water, and mobile phase B, consisting of 0.1% formic acid in
acetonitrile, were used for a linear gradient elution as follows: 10 to 90% B in
10 min, hold 90% B for 3 min, return to 10% B in 0.1 min, and hold 10% B
for 3.9 min to equilibrate the column. The flow rate was 0.4 ml/min. Mass
spectrometric detection was performed on a QTRAP 3200 mass spectrometer
(Applied Biosystems/MDS Sciex, Foster City, CA) equipped with an electro-
spray ionization source. 17-DMAG was detected under positive ionization
mode, whereas 17-AAG and GA were detected under negative ionization mode.
The multiple reaction monitoring (MRM) ion transition was m/z 617 3 58 for
gation of benzoquinone ansamycins was correlated with their hepatic
toxicity (Guo et al., 2008). Hence, it is important to elucidate the
major biotransformation pathways of 17-DMAG in liver microsomes
for discovery of more stable, potent, and less toxic GA analogs.
In this study, we investigated the biotransformation pathways of
17-DMAG in HLMs and especially focused on quinone-hydroquinone
conversion and GSH conjugation. The relative percentages of major
metabolites of 17-DMAG were estimated by normalizing their peak
areas. The major metabolites in incubations were tentatively charac-
terized using liquid chromatography-tandem mass spectrometry (LC- 17-DMAG, m/z 559 3 516 for GA, and m/z 584 3 541 for 17-AAG (Smith et al.,
2004). Ionization source temperature was set at 650°C. Curtain gas, gas 1, and gas
2 were set at 30, 50, and 50 arbitrary units, respectively. Ion spray voltage was set
at 5500 V for 17-DMAG and Ϫ4500 V for GA and 17-AAG. A neutral loss of 43
Da was observed, likely because of the loss of N-methylenemethanamine from 17
side chain. Hence, the MRM ion transitions at m/z 561 3 518, 864 3 821, and
866 3 823 were used to detect GAH2, 19-glutathionyl GA, and 19-glutathionyl
geldanamycin hydroquinone, respectively. MRM ion transitions at m/z 586 3
543, 889 3 846, and 891 3 848 were used to detect 17-(allylamino)-17-
demethoxygeldanamycin hydroquinone (17-AAGH2), 19-glutathionyl 17-(allyl-
amino)-17-demethoxygeldanamycin (17-AAG-SG), and 19-glutathionyl 17-(allyl-
amino)-17-demethoxygeldanamycin hydroquinone, respectively. Meanwhile, the
MS/MS), and the cytochrome P450 (P450) enzymes responsible for
the formation of the metabolites were identified. On the basis of the
identified metabolite profiles, the biotransformation pathways for
17-DMAG in HLMs were proposed.
Materials and Methods
Materials. GA, 17-AAG, and 17-DMAG were purchased from LC Lab-
oratories (Woburn, MA). Reduced GSH, reduced ␤-NADPH, MgCl2, 0.1 M
phosphate buffer, formic acid ␣-naphthoflavone, tranylcypromine, querce-
tin, sulfaphenazole, ticlopidine, diethyldithiocarbamate, quinidine, and ke-
toconazole were supplied by Sigma-Aldrich (St. Louis, MO). MI-63 was tertiary amine of 17-DMAG facilitated its protonation and resulted in a high MS
obtained from Prof. S. Wang (Department of Medicinal Chemistry, Uni- sensitivity under positive ionization. As a result, MRM ion transitions at m/z
versity of Michigan, Ann Arbor, MI). High-performance liquid chroma- 619 3 58, 922 3 58, and 924 3 58 were used to detect 17-(dimethylami-
tography (HPLC)-grade acetonitrile was purchased from Thermo Fisher noethylamino)-17-demethoxygeldanamycin hydroquinone (17-DMAGH2),
Scientific (Waltham, MA). HPLC water was purified using a MilliQ water 19-glutathionyl 17-(dimethylaminoethylamino)-17-demethoxy geldanamy-
system (Millipore Corporation, Billerica, MA). Pooled HLMs (20 mg/ml), cin (17-DMAG-SG), and 19-glutathionyl 17-(dimethylaminoethylamino)-
purified Escherichia coli-expressed recombinant human P450 enzymes
coexpressed with human cytochrome b5 (1 nmol/vial), and E. coli-ex-
17-demethoxygeldanamycin hydroquinone, respectively.
Screening and Characterization of Metabolites. GA, 17-AAG, and 17-
pressed control microsomes were obtained from XenoTech, LLC DMAG were infused into a mass spectrometer to obtain their MS, MS2, and
(Lenexa, KS).
Metabolic Stability Assay. GA, 17-AAG, or 17-DMAG was incubated spectra, the structures of fragment ions of protonated 17-DMAG were pro-
with HLMs in the absence or presence of reduced GSH at 37°C. The enzymes posed tentatively. MRM ion transitions 617 3 58, 617.3 3 159, and 617 3
were activated by reduced ␤-NADPH. The incubation solution was diluted 524 were used to generate 240 additional MRM ion transitions for metabolites
with 0.1 M phosphate buffer (containing MgCl2) to 0.6 ml. The final concen- screening using Metabolite ID software (Applied Biosystems), including 40
trations of drug, microsomes, ␤-NADPH, phosphate buffer, and MgCl2 were 5 common biotransformation processes. To detect all the metabolites, full scan
MS3 spectra. On the basis of the similarities and difference among their mass
␮M, 1 mg/ml, 1 mM, 0.1 M, and 3.3 mM, respectively. For GSH conjugation
assays, the final concentration of reduce GSH was 5 mM (Lang et al., 2007).
An aliquot of 40 ␮l of mixture was collected at 1, 5, 10, 15, 30, 45, 60, and 120
min, and then proteins were precipitated with 120 ␮l of ice-cold acetonitrile
and precursor scan also were conducted. Only the components detected in the
sample and absent in all the control samples were regarded as possible
metabolites. To characterize the possible metabolites, both the sample and
controls were injected on the LC-MS for EPI and MS3 scans to obtain their
containing an internal standard MI-63 (100 ng/ml). The samples were centri- MS2 and MS3 spectra. On the basis of the MS2, MS3 spectra of the metabolites,
fuged at 14,000 rpm ϫ 5 min, and 10 ␮l of supernatant was injected into
and the proposed structures of 17-DMAG fragment ions, the metabolites of
LC-MS. To determine the apparent Km values for the formation of 17-AG from 17-DMAG were characterized.
17-AAG, various concentrations of 17-AAG (1–200 ␮M) were incubated in
Inhibition of 17-DMAG Oxidative Metabolism by Selective P450 Inhib-
0.25 mg/ml HLMs for 15 min in triplicate. Likewise, to determine the Km itors. The incubation mixtures, containing 10 ␮l of pooled HLMs (20 mg

IN VITRO METABOLISM OF 17-DMAG IN HLMs
629
Results
Relative Percentages of Metabolites in Incubations. The appar-
ent Km values for the formation of 17-AG from 17-AAG and M1 and
M3 from 17-DMAG in HLMs were determined as 29.1, 6.61, and 7.83
␮M, respectively, which were higher than the initial concentrations of
17-AAG and 17-DMAG (5 ␮M) in stability assays. The relative
percentages of GA and 17-AAG and their corresponding hydroqui-
none and 19-glutathionyl conjugates determined in stability assays
were shown in Fig. 2. The values were expressed as percentages
normalized by peak area ratio of GA or 17-AAG and internal standard
at 1 min of incubation. GA exhibited high metabolic stability in HLMs
in the absence of reduced GSH (Fig. 2A). Seventy-three percent of
GA was not metabolized after 2 h of incubation in 1 mg/ml HLMs. A
total of 6.4 to 9.3% of GAH2 was obtained during the course of
incubation. A similar amount of GAH2 was also detected in the
incubation with boiled HLMs (negative control), suggesting that the
formation of GAH2 was not dependent on CYP450 enzymes. In
contrast, the presence of 5 mM reduced GSH resulted in the rapid
metabolism of GA during the first 5 min of incubation. Only 75.3 and
46.6% of GA remained after 5- and 120-min incubation, respectively.
Consistently, the reduced GSH resulted in a quick onset of the
formation of GAH2 in HLMs, and 85.7% of GAH2 formed 1 min after
the addition of 5 mM reduced GSH. In addition, the formation of
GAH2 was also observed in the incubation of GA and GSH in the
absence of HLMs, and the peak area ratio of GAH2 and GA was 1.93
after 2 h of incubation. No significant amount of GSH conjugates
were detected in all the above incubations using negative MRM ion
transitions m/z 864 3 821 and 866 3 823.
FIG. 1. Structures of GA, 17-AAG, 17-DMAG, and 17-AG.
protein/ml), were preincubated in 0.1 M phosphate buffer (pH 7.4) in the
presence of reduced ␤-NADPH (1 mM), MgCl2 (3.3 mM), and chemical
inhibitor for 3 min at 37°C in a shaking water bath. Parallel control incubations
were conducted in the absence of chemical inhibitors. The reactions were
initiated by adding 17-DMAG (final concentration, 1 ␮M). The reaction
mixtures (final volume 0.2 ml) were incubated for 20 min at 37°C. The P450
isoform-selective inhibitors used were ␣-naphthoflavone for 1A2 (10 ␮M),
tranylcypromine for 2A6 (1 ␮M), ticlopidine (10 ␮M) for 2B6 and 2C19,
quercetin (1 ␮M) for 2C8, sulfaphenazole (1 ␮M) for 2C9, quinidine (1 ␮M)
for 2D6, diethyldithiocarbamate (10 ␮M) for 2E1, and ketoconazole (1 ␮M)
for 3A (Kim et al., 2003; Lee et al., 2006). In positive control experiments, 10
␮M ␣-naphthoflavone inhibited the formation of acetaminophen from phen-
acetin (20 ␮M); 1 ␮M tranylcypromine inhibited the formation of 7-hydroxy-
coumarin from coumarin (5 ␮M); 10 ␮M ticlopidine inhibited the formation of
hydroxybupropion from bupropion (50 ␮M) and formation of hydroxyompra-
zole from omprazole (50 ␮M); 1 ␮M quercetin inhibited the formation of
6␣-hydroxypaclitaxel from paclitaxel (10 ␮M); 1 ␮M sulfaphenazole inhibited
hydroxytolbutamide formation from tolbutamide (100 ␮M); 1 ␮M quinidine
inhibited dextrorphan formation from dextromethorphan (5 ␮M); 10 ␮M
diethyldithiocarbamate inhibited the formation of hydroxychlorzoxazone from
chlorzoxazone (50 M); and 1 ␮M ketoconazole inhibited 1Ј-hydroxymidazo-
lam formation from midazolam (5 ␮M) (Lee et al., 2006). The reactions were
terminated by adding 400 ␮l of ice-cold acetonitrile containing 100 ng/ml
MI-63. The samples were vortexed for 30 s and centrifuged at 14,000 rpm for
5 min. The supernatant was analyzed by LC-MS/MS to monitor formation rate
of 17-DMAG metabolites. The MRM ion transitions for determination of M1,
M2, M3, M4, and M6 were 633 3 175, 633 3 322, 603 3 510, 603 3 524,
and 665 3 157, respectively. The percentages of inhibition were calculated by
the ratio of the amounts of metabolites formed with and without the specific
inhibitor.
Incubation with Recombinant CYP450 Enzymes. The incubation mix-
tures (200 ␮l), containing 20 pmol of recombinant CYP450 enzyme, 0.2 mg of
reduced ␤-NADPH, and 3.3 mM MgCl2, were preincubated in 0.1 M phos-
phate buffer (pH 7.4) at 37°C in a shaking water bath. E. coli-expressed control
microsomes without cDNA of human P450 were used as a negative control.
The reactions were initiated by adding 17-DMAG (final concentration 1 ␮M).
After 30-min incubation, the reactions were terminated by adding 400 ␮l of
ice-cold acetonitrile containing 100 ng/ml MI-63. LC-MS/MS was used to
monitor formation rate of 17-DMAG metabolites. The formation rates of the
metabolites in each CYP450 isozyme were expressed as the percentages
relative to their formation rates in CYP3A4.
FIG. 2. Relative percentages of GA and GAH2 (A); 17-GA and 17-GAH2 (B); and
17-DMAG (C) and their metabolites in the incubations with HLMs in the presence
or absence of 5 mM reduced GSH. The values were expressed as percentages
normalized by peak area ratio of parent compound and internal standard at 1 min of
incubation. ء, presence of 5 mM GSH.

630
ZHENG ET AL.
In contrast to GA, 17-AAG was extensively metabolized in 1 was assigned to C1–10 as shown in Fig. 3B (Lang et al., 2007). On the
mg/ml HLMs and totally disappeared 45 min after the beginning of MS3 spectra, the ion at m/z 187 was further fragmented to generate
incubation (Fig. 2B). Less than 1% of 17-AAGH2 was detected during ions at m/z 159, 131, and 117, and their structures were tentatively
the first 30 min of incubation in HLMs, and no 17-AAGH2 was proposed in Fig. 3B.
detected after incubation for 1 h. The presence of reduced GSH did
In addition to 17-DMAGH2 and 17-DMAG-SG, 12 metabolites of
not significantly change the metabolism rate of 17-AAG in HLMs but 17-DMAG were detected in the HLM incubation but not in negative
resulted in a quicker onset of 17-AAGH2 formation. Furthermore, controls by MS full scan and MRM scan. Three metabolites had a
3.04% of 17-AAGH2 was obtained when 5 ␮M 17-AAG was incu- protonated molecular ion (Mr) of 665, five had an Mr of 633, one had
bated with 5 mM reduced GSH for 2 h in the absence of HLMs. GSH an Mr of 627, one had an Mr of 615, and two had an Mr of 603. The
conjugates of 17-AAG were not detected over the course of incuba- major metabolites were the two compounds with an Mr of 603, five
tions using the negative MRM ion transitions m/z 889 3 846 and compounds with an Mr of 633, and one compound with an Mr of 665.
891 3 848.
The metabolites with Mr of 603 and 633 were also observed previ-
Figure 2C showed the relative percentages of 17-DMAG and its ously in the bile of rats administered with 17-DMAG, although their
metabolites during the course of incubation. The characterization of structures were not identified (Egorin et al., 2002).
17-DMAG metabolites M1 to M7 is discussed in the following
Five components were detected in the LC-MS/MS chromatogram
section. Compared with 17-AAG, 17-DMAG exhibited improved (m/z 633) of HLM incubation extract but not the negative controls
metabolic stability. Twenty-four percent of 17-DMAG remained in (Fig. 4A). The Mr (m/z 633) of each metabolites showed a mass shift
HLMs after 2 h of incubation. Only a trace amount (0.24%) of of 16 Da compared with protonated 17-DMAG, indicating the addi-
17-DMAGH2 was obtained after incubation for 2 h. Reduced GSH tion of one oxygen. The MS/MS spectrum (Fig. 4B) of the component
showed no effect on the metabolism of 17-DMAG and formation of eluted at 8.21 min (M1) showed the product ions at m/z 58 and 72,
17-DMAGH2 in HLMs. The maximum percentage (0.33%) of 17- suggesting that the oxidation did not occur at the 17 side chain. The
DMAG-SG during the incubation was detected at 60 min, whereas product ions of M1 at m/z 203, 175, 147, and 133 exhibited a mass
19-glutathionyl 17-(dimethylaminoethylamino)-17-demethoxygel- shift of 16 Da compared with the corresponding product ions of
danamycin hydroquinone was not detected during the incubation. In 17-DMAG, indicating that one hydroxyl group generated on C1–10
addition, 1.16% of 17-DMAG-SG was obtained when 17-DMAG was Furthermore, MS3 spectrum of protonated M1 (m/z 617 3 203)
incubated with reduced GSH for 2 h in the absence of HLMs. showed no neutral loss of water from the ion at m/z 203, indicating
.
Identification of 17-DMAG Metabolites. To assign the major that there was no hydrogen on the carbon next to the hydroxylated
fragment ions of protonated 17-DMAG, the MS/MS and MS3 spectra carbon. Hence, the hydroxyl of M1 was tentatively assigned to C23.
of GA, 17-AAG, and 17-DMAG were obtained. Figure 3A showed Product ions at m/z 187 and 159 were detected in the MS/MS spec-
the MS/MS spectrum of protonated 17-DMAG. The product ions at trum (Fig. 4C) of the component eluted at 9.04 min (M2), suggesting
m/z 58 and 72 of 17-DMAG, which were not detected in the MS/MS that the hydroxylation did not occur at C1–10 moiety. The detection of
spectra of GA and 17-AAG, were assigned to the N,N-dimethyl- a product ion at m/z 70 instead of 72 suggested the oxidation on
ethanamine and trimethylamine moieties of 17-dimethylaminoethyl- N,N-dimethylethanamine, whereas the presence of a product ion at
amino side chain (Fig. 3B). Both product ions at m/z 187 and 159 were m/z 58 indicated that the trimethylamine moiety was intact. Hence, the
detected in MS/MS spectra of GA, 17-AAG, and 17DMAG. On the hydroxyl was tentatively assigned to the ␣-carbon. The product ion at
basis of the accurate mass measurement, the product ion at m/z 187 m/z 70 was proposed to be generated from ␣-hydroxylated 17-DMAG
FIG. 3. A, MS/MS spectrum and MS3 spectrum (m/z 617 3 187)
of protonated 17-DMAG. B, the proposed fragmentation pathways
of protonated 17-DMAG.

IN VITRO METABOLISM OF 17-DMAG IN HLMs
631
FIG. 4. A, LC-MS/MS chromatogram (m/z 633) of HLM incuba-
tion extract. Five metabolites of 17-DMAG formed by the addition
of one oxygen; B, MS/MS spectrum of the metabolite eluted at 8.21
min (M1); C, MS/MS spectrum of the metabolite eluted at 9.04
min (M2).
by cleaving the bond between nitrogen and ␣-carbon, followed by from C12, the loss of methanol from C6 position was more energeti-
dehydration. Similar ␣-hydroxylation on a 17 side chain also was cally favored because it would increase the amount of conjugation in
observed in the metabolism of 17-AAG in HLMs (Lang et al., 2007). the ansamycin ring. Hence, the product ion of M3 at m/z 571 was
The metabolite eluted at 7.14 min (Fig. 4A) showed a similar generated by the loss of methanol from C6, and M3 was produced by
profile of product ions as that of M2, except the presence of the the demethylation from the methoxyl at C12 of 17-DMAG (Fig. 5C).
metabolite’s product ion was at m/z 72 instead of m/z 70. The major Likewise, the product ions of M4 at m/z 585 and 524 were generated
product ions of this metabolite were ions at m/z 159 and 187 instead by the loss of water from C6 and subsequent loss of carbamic acid
of ions at m/z 322 and 304. The metabolite was probably formed by from C7 (Fig. 5D). M4 was generated by the demethylation from the
the hydroxylation of C25 or the nitrogen adjacent to C17. The metab- methoxyl at C6. The component eluted at 8.08 min was also detected
olites eluted at 6.71 and 7.53 min exhibited the same MS/MS pattern, in the negative controls, indicating that it was not a metabolite of
and the major product ions were ions at m/z 572, 540, 306, 274, 246, 17-DMAG.
192, and 177, suggesting the oxidation might occur at C1–10 moiety,
One metabolite eluted at 8.4 min (M5) exhibited a Mr at m/z 615
C22, C24, or two methoxyl groups. The components eluted at 9.65 and (parent Ϫ 2), indicating a loss of H2O from hydroxylized 17-DMAG
10.14 min might not be metabolites because they were also detected (Fig. 6A). The detection of m/z 187 suggested that the metabolism did
in the negative controls. A complete separation of the peaks at 7.53 not occur at C1–10 position. The ions at m/z 72 and 58 were not
and 7.80 min on the chromatogram was not obtained. The structure of detected in the MS/MS spectrum of M5, suggesting that the metabo-
the metabolite at 7.80 min was not identifiable because of the low lism was likely to occur on N,N-dimethylethanamine moiety. The base
intensities of its product ions.
product ion peak of protonated M5 at m/z 304 was also detected in the
Two metabolites with the identical Mr at m/z 603 (parent Ϫ 14) MS/MS spectrum of M2 (Fig. 4C). Hence, M5 is proposed to be
were eluted at 7.57 min (M3) and 7.13 min (M4). Both metabolites derived from M2 by the loss of H2O, which resulted in the formation
were likely formed by demethylation (Fig. 5A). The two metabolites of a carbon-nitrogen double bond and increased amount of conjuga-
had similar MS/MS fragmentation pattern, except for the detection of tion in the system (Fig. 6A).
product ions of M3 at m/z 571 and 510 and product ions of M4 at m/z
It is worth noting that the product ion at m/z 306 was detected in the
585 and 524 (Fig. 5, B and D). The ions at m/z 571 and 510 were MS/MS spectra of M1 (Fig. 4B), M3 (Fig. 5B), and M4 (Fig. 5C). Its
derived from protonated M3 by the loss of methanol and subsequent counterpart ion at m/z 322 in MS/MS spectrum of M2 (Fig. 4C)
loss of carbamic acid from C7. Compared with the loss of methanol showed a mass shift of 16 Da. Another counterpart ion at m/z 304 in

632
ZHENG ET AL.
FIG. 5. A, LC-MS/MS chromatogram (m/z
603) of HLM incubation extract. Two me-
tabolites of 17-DMAG formed by demeth-
ylation; B, MS/MS spectrum and MS3
spectrum (m/z 603 3 306) of the metabo-
lite eluted at 7.57 min (M3); C, the proposed
formation and fragmentation pathways of the
product ion at m/z 306; D, MS/MS spectrum
of the metabolite eluted at 7.13 min (M4).
MS/MS of M5 (Fig. 6A) showed a mass shift of 2 Da. Hence, the ion two oxygen atoms were added to C11–20 moiety. The presence of the
at m/z 306 was possibly generated by the loss of methanol (M1 and ion at m/z 58 suggested that two N-methyl groups were not oxidized.
M4) or H2O (M3) from C12 and subsequent cleavage of the amide Two hydroxyl groups were tentatively assigned to the ␣-carbon of 17
bond in the ring and the bond between C10 and C11 (Fig. 4C). MS3 side chain and C25, respectively. The product ions of M6 at m/z 185
spectrum of M3 (m/z 633 3 306) showed that the ion at m/z 306 and 157 (Fig. 6B) exhibited a mass shift of 2 Da from their counterpart
generated its fragment ions at m/z 274, 246, 229, and 218 (Fig. 5B) by ions of 17-DMAG at m/z 187 and 159, respectively, indicating a third
the loss or sequential losses of CH3OH, CO, and NH3 (Fig. 5C). The hydroxyl group located at C1–10 moiety of M6, and this hydroxyl was
product ion of M2 at m/z 322 (Fig. 4C) and the product ion of M5 at ready to be deprived from protonated M6 through dehydration. Hence,
m/z 304 (Fig. 6A) were produced through similar fragmentation.
the third hydroxyl was tentatively assigned to C24 because the forma-
The metabolite eluted at 6.8 min (M6) showed an Mr at m/z 665 tion of a double bond between C24 and C10 would increase the amount
(parent ϩ 48), indicating the addition of the three oxygen to 17- of conjugation and stabilize the product ions of M6. The relative
DMAG. The detection of ion at m/z 338 (Fig. 6B), a counterpart of the amounts of the six metabolites M1 to M6 in HLM incubation during
ion at m/z 306 of M3 and M4 (a mass shift of 32 Da), indicated that 1 to 120 min were shown in Fig. 2C. The relative amounts of

IN VITRO METABOLISM OF 17-DMAG IN HLMs
633
FIG. 6. MS/MS spectrum of the metabolite eluted at 8.40 min (M5)
(A) and the metabolite eluted at 6.80 min (M6) (B).
17-DMAG metabolites at 60 min were M3 (2.19%) Ͼ M1 (1.48%) Ͼ were summarized in Fig. 7. The formation of M1, M2, M3, M4, and
M6 (1.40%) Ͼ M2 (1.38%) Ͼ M4 (0.98%) Ͼ M5 (0.52%) Ͼ
17-DMAGH2 (0.19%).
M6 in HLMs was inhibited dramatically by a CYP3A inhibitor.
Ketoconazole (1 ␮M) resulted in 75 to 88% inhibition on the forma-
tion of the five metabolites. The other inhibitors did not cause re-
markable inhibition on the formation of metabolites in HLMs. Results
from positive control experiments confirmed that the incubations were
performed under optimum conditions.
Oxidative Metabolism of 17-DMAG by Recombinant P450
Isoenzymes. 17-DMAG (1 ␮M) was incubated with recombinant
human P450 isozymes, and the relative formation rates (percentage of
that of CYP3A4) of M1, M2, M3, M4, and M6 were shown in Fig. 8.
Among the CYP450 isoforms tested, CYP3A4 produced the highest
formation rates of M1, M2, M3, M4, and M6. CYP3A5 showed a
lower activity for the metabolism of 17-DMAG relative to CYP3A4.
CYP2C8, 2D6, 2A6, 2C19, and 1A2 produced M6 at moderate rates
compared with CYP3A4 and 3A5. Other CYP450 enzymes did not
metabolize 17-DMAG at an appreciable rate. The incubation of 17-
DMAG with E. coli-expressed control microsomes did not show the
formation of the metabolites.
One minor metabolite eluted at 9.2 min exhibited an Mr at m/z 627,
a mass shift of 10 Da from the counterpart ion of 17-DMAG. The
major product ion of this metabolite included ions at m/z 566, 538, and
331. The product ion at m/z 566 was derived from the Mr by the loss
of carbamic acid. A subsequent neutral loss of 28 Da instead of 32 Da
(the loss of methanol C12 or C6) from m/z 566 was detected and gave
rise to m/z 538, suggesting the possible loss of an ethylene or carbon
monoxide. Because of the low concentration of this metabolite in
HLM incubation, the MS3 spectra of its product ions were not ob-
tainable. Additional study is required to definitively identify the
structure of the metabolite.
Inhibition of 17-DMAG Oxidative Metabolism by Selective
P450 Inhibitors. Identification of the P450 isozymes responsible for
oxidative metabolism of 17-DMAG was performed using both P450-
selective inhibitors. The results of experiments using specific P450
inhibitors to prevent the formation of 17-DMAG metabolites in HLMs
FIG. 8. Relative formation rates of M1, M2, M3, M4, and M6 by recombinant
human CYP450 isozymes. The values were expressed as percentages normalized by
the formation rate in CYP3A4 incubation.
FIG. 7. Effect of selective CYP450 inhibitors on the oxidative metabolism of
17-DMAG by HLMs.

634
ZHENG ET AL.
Lang et al., 2007), no significant amount of GSH conjugates of 17-AAG
Discussion
was detected. Likewise, the GSH conjugates of GA were not detected in
the presence and absence of HLMs.
The molecular modeling analysis showed that the reduction of
quinone into hydroquinone facilitated the formation of two more
hydrogen bonds between GAH2 and Hsp90, resulting in a tighter
binding between GAH2 and Hsp90 (Lang et al., 2007). GSH plays an
important role in the detoxification of reactive drugs and metabolites
formed by hepatic drug-metabolizing enzymes (Satoh, 1995). GA and
its analogs were observed to react chemically (i.e., nonenzymatically)
with GSH to form GSH conjugates at 19-position, and the GSH
conjugation could be important for the toxicity of GA and its analogs
(Cysyk et al., 2006). One objective of this study was to detect and
compare the quinone-hydroquinone conversion and GSH conjugation
among GA, 17-AAG, and 17-DMAG in HLM incubations. Previous
study (Lang et al., 2007) and our preliminary study showed that only
trace amounts of hydroquinone metabolites and 19-glutathionyl me-
tabolites of 17-AAG or 17-DMAG formed in the HLM incubation. To
generate enough metabolites, a substrate concentration of 5 ␮M and a
HLM concentration of 1 mg/ml were selected for metabolic stability
assays. The substrate concentration was lower than the Km values for
the formation of major metabolites of 17-AAG and 17-DMAG. To
evaluate the formation of GSH conjugates of GA derivatives in
HLMs, GSH in reduced form was added to the HLM incubations to
reach a final concentration of 5 mM as described previously (Lang et
al., 2007; Guo et al., 2008). It was also reported (Guo et al., 2008) that
less than 15% of 17-AAG-SG formed even after 24 h of incubation in
the presence of 5 mM reduced GSH, indicating a slow GSH conju-
gation rate. Hence, the total incubation time of metabolic stability
assays was chosen as 2 h to acquire as high an amount of GSH
conjugates as possible while maintaining reasonable enzyme activity.
A total of 6.4 to 9.3% of GAH2 was detected in HLM incubation in the
absence of GSH, much lower than the amount (40–73%) detected in a
published report (Lang et al., 2007). The inconsistency might be partially
caused by oxygen exposure during sample preparation as some GAH2
was oxidized into GA. The reduced GSH stimulated the reduction of GA,
and 78 to 88% of GAH2 was detected, which was consistent with the
amount of GAH2 obtained in HLM incubation under hypoxic conditions
(Lang et al., 2007). The reduction of GA was not dependent on CYP450
enzymes because the incubation of GA with GSH in the absence of
HLMs produced more GAH2 than the incubation in the presence of
HLMs. In contrast, reduced GSH did not enhance the formation
of 17-AAGH2 and 17-DMAGH2 in HLMs. Less than 1% of 17-AAGH2
and 17-DMAGH2 formed in HLM incubations. This result agreed well
with a previous study (Lang et al., 2007) in which only 2% of 17-AAGH2
were detected after incubation in HLMs for 60 min. The reason for
different formation rates of GAH2 and 17-AAGH2 in HLMs has been
discussed: a 17-allylamino group of 17-AAG is a stronger electron-
donating substituent than the 17-methoxy of GA, leading to a more
negative shift in one-electron redox potential for 17-AAG. As a result,
17-AAG is not as easily reduced as GA (Lang et al., 2007). Likewise,
17-dimethylaminoethylamino is also a stronger electron-donating sub-
stituent than the 17-methoxy, resulting in the slower formation of 17-
DMAGH2 than GAH2. With the tight binding between GAH2 and
Hsp90, the tendency to get reduced might result in a higher nonselective
toxicity of GA than that of 17-AAG and 17-DMAG.
Quinone/hydroquinone cycling was proposed to be the primary
metabolism pathway of GA and its analogs (Guo et al., 2008).
NADPH-dependent redox cycling rates of GA analogs were deter-
mined by measuring the oxygen consumption rates: 17-DMAG Ͼ
GA Ͼ 17-AAG; however, the GSH conjugation rates are as follows:
GA Ͼ 17-DMAG Ͼ 17-AAG. However, these findings were incon-
sistent with the rapid in vitro and in vivo metabolism of 17-AAG
(Egorin et al., 2001; Musser et al., 2003).
Our study showed that benzoquinone reduction and GSH conjuga-
tion were not the primary metabolism pathways of 17-DMAG in
HLMs. This observation was supported by in vivo metabolism data of
17-DMAG in rats (Egorin et al., 2002). In the study, 11 compounds
with UV spectra similar to that of 17-DMAG were detected in rat bile
after intravenous administration of 17-DMAG. Although the struc-
tures of the 11 compounds were not identified, the metabolism did not
involve the alterations to the benzoquinone moiety because their UV
spectra were not altered compared with 17-DMAG. Among the 11
metabolites, 4 had Mr of 633 and 2 had Mr of 603, which were also
detected in our HLM incubations.
The results of metabolite identification revealed that the biotrans-
formation of 17-DMAG by CYP450 enzymes differed from that of
17-AAG. The extensive oxidative metabolism of 17-AAG on the
17-allylamino side chain resulted in the formation of 17-AG as well as
epoxide and diol metabolites (Egorin et al., 1998). The hydroxylation
on C22 and demethylation from methoxy moieties were detected as the
minor biotransformation pathways of 17-AAG in HLMs (Lang et al.,
2007). In contrast, the hydroxylation on C23 (M1) and demethylation
from methoxy moieties (M3 and M4) were the primary biotransfor-
mation of 17-DMAG in HLMs (Fig. 8). Similar to 17-AAG, the
hydroxylation of the adjacent carbon to the nitrogen generated a stable
carbinolamine intermediate (M2). Subsequently, M2 underwent the
loss of H2O and formed an imine metabolite (M5). In general, carbi-
nolamines are known as unstable intermediates in N-demethylation.
Most carbinolamines tend to form imine intermediates by the loss of
water or decompose into dealkylated amine by the loss of aldehyde or
ketone. However, stable carbinolamines and imines exist in aqueous
solutions under some specific circumstances. For example, carbino-
lamines and imines can be stabilized by adjacent electron-withdraw-
ing groups (Upthagrove and Nelson, 2001). Furthermore, a stable
carbinolamine, hydroxymethylpentamethylmelamine, was detected as
a metabolite in mouse plasma (Abikhalil et al., 1986) and rat liver
microsomes (Ames et al., 1983) because the adjacent substituents
delocalize the lone pair nitrogen electrons and stabilize the carbino-
lamine. In addition, stable carbinolamines have been isolated or
detected as metabolites to a number of drugs including benzamides
(Huizing et al., 1980), carbamates (Dorough and Casida, 1964), pro-
carbazine (Weinkam and Shiba, 1978), N-methylcarbazole (Ebner et
al., 1991), medazepam (Schwartz and Kolis, 1972), verapamil (Walles
et al., 2002), pyrimidinyl-pyridopyrazines (Prakash and Soliman,
1997), and ecabapide (Fujimaki et al., 1995). Quinone is known as an
electron-withdrawing group by resonance (Hiramatsu et al., 1983).
The presence of the quinone adjacent to the nitrogen in M2, M5, and
M6 of 17-DMAG allows the delocalization of the lone pair nitrogen
electrons and stabilizes the metabolites.
GA, 17-AAG, and 17-DMAG were reported to react chemically with
GSH to form GSH conjugates at 19-position (Cysyk et al., 2006). How-
ever, in the current study, only 0.33% of 17-DMAG-SG was formed in
the incubation of 17-DMAG with GSH in the presence of HLMs, and
1.16% of 17-DMAG-SG was formed in the absence of HLMs. The
17-AAG was a known substrate of CYP3A4 (Lang et al., 2007). In the
current study, the reaction phenotyping experiments using recombinant
human P450s and selective P450 chemical inhibitors revealed that
presence of proteins in the incubation might decrease GSH conjugation. CYP3A4 and CYP3A5 were the major P450 isozymes responsible for the
Consistent with previous in vitro and in vivo studies (Egorin et al., 1998; oxidative metabolism of 17-DMAG in HLMs. In addition, CYP2C8,

IN VITRO METABOLISM OF 17-DMAG IN HLMs
635
FIG. 9. Summary of in vitro oxidative me-
tabolism pathways of 17-DMAD in HLMs.
Goetz MP, Toft D, Reid J, Ames M, Stensgard B, Safgren S, Adjei AA, Sloan J, Atherton P,
Vasile V, et al. (2005) Phase I trial of 17-allylamino-17-demethoxygeldanamycin in patients
with advanced cancer. J Clin Oncol 23:1078–1087.
Guo W, Reigan P, Siegel D, and Ross D (2008) Enzymatic reduction and glutathione conjugation
of benzoquinone ansamycin heat shock protein 90 inhibitors: relevance for toxicity and
mechanism of action. Drug Metab Dispos 36:2050–2057.
2D6, 2A6, 2C19, and 1A2 were observed to contribute to the formation
of M6. On the basis of the identified metabolite profiles, the in vitro
oxidative biotransformation pathways for 17-DMAG in HLMs were
proposed (Fig. 9).
Guo W, Reigan P, Siegel D, Zirrolli J, Gustafson D, and Ross D (2005) Formation of 17-allylamino-
demethoxygeldanamycin (17-AAG) hydroquinone by NAD(P)H:quinone oxidoreductase 1: role of
17-AAG hydroquinone in heat shock protein 90 inhibition. Cancer Res 65:10006–10015.
Guo W, Reigan P, Siegel D, Zirrolli J, Gustafson D, and Ross D (2006) The bioreduction of a
series of benzoquinone ansamycins by NAD(P)H:quinone oxidoreductase 1 to more potent
heat shock protein 90 inhibitors, the hydroquinone ansamycins. Mol Pharmacol 70:1194–
1203.
Hiramatsu M, Shiozaki K, Fujinami T, and Sakai S (1983) Preparation and electrochemical
properties of palladium(0) complexes coordinated by quinones and 1,5-cyclooctadiene. J
Organomet Chem 246:203–211.
Huizing G, Segura J, and Beckett AH (1980) On the mechanism of metabolic N-dealkylation.
Isolation of a relatively stable carbinolamine. J Pharm Pharmacol 32:650–651.
Hwang K, Scripture CD, Gutierrez M, Kummar S, Figg WD, and Sparreboom A (2006)
Determination of the heat shock protein 90 inhibitor 17-dimethylaminoethylamino-17-
demethoxygeldanamycin in plasma by liquid chromatography-electrospray mass spectrome-
try. J Chromatogr B Analyt Technol Biomed Life Sci 830:35–40.
Kim KA, Kim MJ, Park JY, Shon JH, Yoon YR, Lee SS, Liu KH, Chun JH, Hyun MH, and Shin
JG (2003) Stereoselective metabolism of lansoprazole by human liver cytochrome P450
enzymes. Drug Metab Dispos 31:1227–1234.
Lang W, Caldwell GW, Li J, Leo GC, Jones WJ, and Masucci JA (2007) Biotransformation of
geldanamycin and 17-allylamino-17-demethoxygeldanamycin by human liver microsomes:
reductive versus oxidative metabolism and implications. Drug Metab Dispos 35:21–29.
Lee HK, Moon JK, Chang CH, Choi H, Park HW, Park BS, Lee HS, Hwang EC, Lee YD, Liu
KH, et al. (2006) Stereoselective metabolism of endosulfan by human liver microsomes and
human cytochrome P450 isoforms. Drug Metab Dispos 34:1090–1095.
McIlwrath AJ, Brunton VG, and Brown R (1996) Cell-cycle arrest and p53 accumulation induced
by geldanamycin in human ovarian tumour cells. Cancer Chemother Pharmacol 37:423–428.
Moreno-Farre J, Asad Y, Pacey S, Workman P, and Raynaud FI (2006) Development and validation of a
liquid chromatography/tandem mass spectrometry method for the determination of the novel anticancer
agent 17-DMAG in human plasma. Rapid Commun Mass Spectrom 20:2845–2850.
Musser SM, Egorin MJ, Zuhowski EG, Hamburger DR, Parise RA, Covey JM, White KD, and
Eiseman JL (2003) Biliary excretion of 17-(allylamino)-17-demethoxygeldanamycin (NSC
330507) and metabolites by Fischer 344 rats. Cancer Chemother Pharmacol 52:139–146.
Prakash C and Soliman V (1997) Metabolism and excretion of a novel antianxiety drug
candidate, CP-93,393, in Long Evans rats. Differentiation of regioisomeric glucuronides by
LC/MS/MS. Drug Metab Dispos 25:1288–1297.
Satoh K (1995) The high non-enzymatic conjugation rates of some glutathione S-transferase
(GST) substrates at high glutathione concentrations. Carcinogenesis 16:869–874.
Schwartz MA and Kolis SJ (1972) Pathways of medazepam metabolism in the dog and rat.
J Pharmacol Exp Ther 180:180–188.
Smith NF, Hayes A, Nutley BP, Raynaud FI, and Workman P (2004) Evaluation of the cassette
dosing approach for assessing the pharmacokinetics of geldanamycin analogues in mice.
Cancer Chemother Pharmacol 54:475–486.
In conclusion, this study tentatively characterized six oxidative metab-
olites of 17-DMAG in HLMs and revealed that the primary biotransfor-
mations of 17-DMAG in HLMs were hydroxylation and demethylation
on its side chains. CYP3A4 and CYP3A5 were identified as the major
P450 isozymes responsible for the oxidative metabolism of 17-DMAG in
HLMs. The metabolic data generated in the current study have important
implications for the discovery of more metabolically stable and less toxic
GA analogs for the treatment of solid tumors.
Authorship Contributions
Participated in research design: Wang.
Conducted experiments: Sun.
Contributed new reagents or analytic tools: Wang.
Performed data analysis: Sun, Zheng, and Zou.
Wrote or contributed to the writing of the manuscript: Zheng and Zou.
References
Abikhalil F, Dubois J, Hanocq M, and Atassi G (1986) A new algorithm for computing the
parameters of linear compartment models in pharmacokinetics. Eur J Drug Metab Pharma-
cokinet 11:51–59.
Ames MM, Sanders ME, and Tiede WS (1983) Role of N-methylolpentamethylmelamine in the
metabolic activation of hexamethylmelamine. Cancer Res 43:500–504.
An WG, Schnur RC, Neckers L, and Blagosklonny MV (1997) Depletion of p185erbB2, Raf-1
and mutant p53 proteins by geldanamycin derivatives correlates with antiproliferative activity.
Cancer Chemother Pharmacol 40:60–64.
Cysyk RL, Parker RJ, Barchi JJ Jr, Steeg PS, Hartman NR, and Strong JM (2006) Reaction of
geldanamycin and C17-substituted analogues with glutathione: product identifications and
pharmacological implications. Chem Res Toxicol 19:376–381.
Dorough HW and Casida JE (1964) Nature of certain carbamate metabolites of insecticide sevin.
J Agric Food Chem 12:294–304.
Ebner T, Meese CO, and Eichelbaum M (1991) Regioselectivity and stereoselectivity of the
metabolism of the chiral quinolizidine alkaloids sparteine and pachycarpine in the rat.
Xenobiotica 21:847–857.
Egorin MJ, Lagattuta TF, Hamburger DR, Covey JM, White KD, Musser SM, and Eiseman JL
(2002) Pharmacokinetics, tissue distribution, and metabolism of 17-(dimethylaminoethyl-
amino)-17-demethoxygeldanamycin (NSC 707545) in CD2F1 mice and Fischer 344 rats.
Cancer Chemother Pharmacol 49:7–19.
Egorin MJ, Rosen DM, Wolff JH, Callery PS, Musser SM, and Eiseman JL (1998) Metabolism
of 17-(allylamino)-17-demethoxygeldanamycin (NSC 330507) by murine and human hepatic
preparations. Cancer Res 58:2385–2396.
Egorin MJ, Zuhowski EG, Rosen DM, Sentz DL, Covey JM, and Eiseman JL (2001) Plasma
pharmacokinetics and tissue distribution of 17-(allylamino)-17-demethoxygeldanamycin
(NSC 330507) in CD2F1 mice1. Cancer Chemother Pharmacol 47:291–302.
Fujimaki Y, Hosokami T, and Ono K (1995) Identification of urinary metabolites of ecabapide
in rat. Xenobiotica 25:501–510.
Upthagrove AL and Nelson WL (2001) Carbinolamines, imines, and oxazolidines from fluorinated
propranolol analogs. (19)F NMR and mass spectral characterization and evidence for formation as
intermediates in cytochrome P450-catalyzed N-dealkylation. Drug Metab Dispos 29:1114–1122.
Walles M, Thum T, Levsen K, and Borlak J (2002) Verapamil: new insight into the molecular
mechanism of drug oxidation in the human heart. J Chromatogr A 970:117–130.
Weinkam RJ and Shiba DA (1978) Metabolic activation of procarbazine. Life Sci 22:937–945.
Address correspondence to: Dr. Duxin Sun, Department of Pharmaceutical
Sciences, College of Pharmacy, University of Michigan, 428 Church St., Ann
Arbor, MI 48109. E-mail: duxins@umich.edu
Glaze ER, Lambert AL, Smith AC, Page JG, Johnson WD, McCormick DL, Brown AP, Levine
BS, Covey JM, Egorin MJ, et al. (2005) Preclinical toxicity of a geldanamycin analog,
17-(dimethylaminoethylamino)-17-demethoxygeldanamycin (17-DMAG), in rats and dogs:
potential clinical relevance. Cancer Chemother Pharmacol 56:637–647.