The in vitro metabolism of phospho-sulindac amide, a novel potential anticancer agent

Article abstract of DOI:10.1016/j.bcp.2014.07.007

Phospho-sulindac amide (PSA) is a novel potential anti-cancer and anti-inflammatory agent. Here we report the metabolism of PSA in vitro. PSA was rapidly hydroxylated at its butane-phosphate moiety to form two di-hydroxyl-PSA and four mono-hydroxyl-PSA metabolites in mouse and human liver microsomes. PSA also can be oxidized or reduced at its sulindac moiety to form PSA sulfone and PSA sulfide, respectively. PSA was mono-hydroxylated and cleared more rapidly in mouse liver microsomes than in human liver microsomes. Of eight major human cytochrome P450s (CYPs), CYP3A4 and CYP2D6 exclusively catalyzed the hydroxylation and sulfoxidation reactions of PSA, respectively. We also examined the metabolism of PSA by three major human flavin monooxygenases (FMOs). FMO1, FMO3 and FMO5 were all capable of catalyzing the sulfoxidation (but not hydroxylation) of PSA, with FMO1 being by far the most active isoform. PSA was predominantly sulfoxidized in human kidney microsomes because FMO1 is the dominant isoform in human kidney. PSA (versus sulindac) is a preferred substrate of both CYPs and FMOs, likely because of its greater lipophilicity and masked-COOH group. Ketoconazole (a CYP3A4 inhibitor) and alkaline pH strongly inhibited the hydroxylation of PSA, but moderately suppressed its sulfoxidation in liver microsomes. Together, our results establish the metabolic pathways of PSA, identify the major enzymes mediating its biotransformations and reveal significant inter-species and inter-tissue differences in its metabolism.

Full text of DOI:10.1016/j.bcp.2014.07.007

Biochemical Pharmacology 91 (2014) 249–255
Contents lists available at ScienceDirect
Biochemical Pharmacology
journal homepage: www.elsevier.com/locate/biochempharm
The in vitro metabolism of phospho-sulindac amide, a novel potential
anticancer agent
Gang Xie a, Ka-Wing Cheng a, Liqun Huang a, Basil Rigas a,b,
*
a Division of Cancer Prevention, Department of Medicine, Stony Brook University, Stony Brook, NY 11794, USA
b Medicon Pharmaceuticals, Inc., Stony Brook, NY 11790, USA
A R T I C L E I N F O
A B S T R A C T
Article history:
Phospho-sulindac amide (PSA) is a novel potential anti-cancer and anti-inflammatory agent. Here we
report the metabolism of PSA in vitro. PSA was rapidly hydroxylated at its butane-phosphate moiety to
form two di-hydroxyl-PSA and four mono-hydroxyl-PSA metabolites in mouse and human liver
microsomes. PSA also can be oxidized or reduced at its sulindac moiety to form PSA sulfone and PSA
sulfide, respectively. PSA was mono-hydroxylated and cleared more rapidly in mouse liver microsomes
than in human liver microsomes. Of eight major human cytochrome P450s (CYPs), CYP3A4 and CYP2D6
exclusively catalyzed the hydroxylation and sulfoxidation reactions of PSA, respectively. We also
examined the metabolism of PSA by three major human flavin monooxygenases (FMOs). FMO1, FMO3
and FMO5 were all capable of catalyzing the sulfoxidation (but not hydroxylation) of PSA, with FMO1
being by far the most active isoform. PSA was predominantly sulfoxidized in human kidney microsomes
because FMO1 is the dominant isoform in human kidney. PSA (versus sulindac) is a preferred substrate of
both CYPs and FMOs, likely because of its greater lipophilicity and masked–COOH group. Ketoconazole (a
CYP3A4 inhibitor) and alkaline pH strongly inhibited the hydroxylation of PSA, but moderately
suppressed its sulfoxidation in liver microsomes. Together, our results establish the metabolic pathways
of PSA, identify the major enzymes mediating its biotransformations and reveal significant inter-species
and inter-tissue differences in its metabolism.
Received 16 May 2014
Accepted 11 July 2014
Available online 18 July 2014
Keywords:
Phospho-sulindac amide
Cytochrome P450
Liver microsomes
Hydroxylation
Sulfoxidation
ß 2014 Elsevier Inc. All rights reserved.
1. Introduction
and Morrow, 2001). This toxicity results from the inhibition of
cyclooxygenase (COX) and the subsequent prostaglandin synthe-
Inflammation plays critical roles at various stages of tumor
development, including tumor initiation, promotion, malignant
conversion, invasion, and metastasis [1]. These findings provide a
foundation and rationale for the use of anti-inflammatory drugs in
cancer prevention and therapy. For example, sulindac, a widely-
used nonsteroidal anti-inflammatory drug, prevents colon cancer
in patients with sporadic adenomas or familial adenomatous
polyposis [2].
sis; the –COOH group of sulindac is critical for its binding to COX
[4]. Therefore, we recently modified sulindac at its–COOH group to
develop a sulindac derivative, phospho-sulindac (PS) (Fig. 1). PS
exhibited more potent anti-cancer efficacy and markedly reduced
toxicity than conventional sulindac in animal models [5]. However,
the carboxylester bond of PS is labile, resulting in the extensive
hydrolysis of PS to sulindac in vitro and in vivo [6]. We previously
demonstrated that protecting PS against its hydrolysis enhanced
its anti-cancer efficacy [7]. Thus, the extensive hydrolysis of PS in
vivo may be detrimental to its efficacy and safety. This
consideration prompted us to modify PS at its ester bond to
generate phospho-sulindac amide (PSA; Fig. 1) to abrogate its
undesired hydrolysis.
Sulindac, however, has a broad range of side effects, including
gastrointestinal, central nervous system, skin rash and pruritus,
and transient elevations of hepatic enzymes in plasma [3] (Roberts
Drug metabolism plays a key role in defining the efficacy and
toxicity of drugs. While the assessment of PSA as a novel
anticancer/anti-inflammatory agent is still ongoing, herein, we
report the major metabolic pathways of PSA, showing that PSA is
extensively hydroxylated and sulfoxidized but not hydrolyzed by
microsomes from various species and tissues. These results are
Abbreviations: CYP, cytochrome P450; FMO, flavin monooxygenase; HL, Mhuman
liver microsomes; ML, Mmouse liver microsomes; PS, phospho-sulindac; PSA,
phospho-sulindac amide.
*
Corresponding author at: Stony Brook University, Division of Cancer Prevention,
HSC, T17-080; Stony Brook, NY 11794-8173, USA. Tel.: +1 631 444 9538;
fax: +1 631 444 9553.
E-mail address: basil.rigas@stonybrook.edu (B. Rigas).
http://dx.doi.org/10.1016/j.bcp.2014.07.007
0006-2952/ß 2014 Elsevier Inc. All rights reserved.

250
G. Xie et al. / Biochemical Pharmacology 91 (2014) 249–255
Sulindac
Phospho-sulindac (PS)
Phospho-sulindac amide (PSA)
Fig. 1. Structures of sulindac, phospho-sulindac and phospho-sulindac amide.
essential for a better understanding of the pharmacology of PSA in
animal and human studies.
subjected to HPLC analyses. The HPLC fractions corresponding to
each metabolite of PSA were collected and analyzed by mass
spectrometry.
2. Materials and methods
2.5. The metabolism of PA by rat hepatocytes
2.1. Reagents
Cryopreserved rat hepatocytes were thawed and incubated
following the manufacturer’s protocol. Briefly, hepatocytes were
incubated with PSA in 24-well tissue culture plates at a density of
2.5 Â 105 cells/well in 5% CO2 at 37 8C. At the designated time-
points, the cells were mixed with 2-fold volume of acetonitrile to
stop the reaction. After centrifugation, the supernatants were
analyzed by HPLC.
PSA was provided by Medicon Pharmaceuticals, Inc. (Stony
Brook, NY). Sulindac, sulindac sulfone, sulindac sulfide, dithio-
threitol, methimazole, and CH3CN of HPLC grade were purchased
from Sigma-Aldrich (St. Louis, MO). Quinidine and ketoconazole
were purchased from Toronto Research Chemicals (North York,
ON, Canada). Mouse and human liver microsomes, rat liver cytosol,
recombinant human CYPs (CYP1A2, 2A6, 2B6, 2C9, 2C19, 2D6, 2E1
and 3A4), FMOs (FMO1, FMO3 and FMO5), NADPH regenerating
solution, and cryopreserved rat hepatocytes were purchased from
BD Biosciences (San Jose, CA). Human intestine, kidney and lung
microsomes were purchased from XenoTech LLC (Lenexa, KS).
2.6. The metabolism of PSA by rat liver cytosol
PSA (50
mM) was pre-incubated at 37 8C for 5 min with 10 mM
dithiothreitol in 0.1 M Tris buffer (pH 7.4). The reaction was
initiated by the addition of liver cytosol (protein concentration
2 mg/ml) and samples were maintained at 37 8C for various time
periods. At the end of each of the incubations, 0.1 ml aliquots were
mixed with 0.2 ml of CH3CN, vortexed and then centrifuged for
10 min at 13,000 Â g. The supernatants were analyzed by HPLC.
2.2. HPLC-UV analysis
The HPLC system consisted of
a Waters Alliance 2695
Separations Module equipped with a Waters 2998 photodiode
array detector (328 nm) and a Thermo Hypersil BDS C18 column
(150 Â 4.6 mm, particle size 3
m
m). The mobile phase consisted of
2.7. The metabolism of PSA by human CYP and FMO isoforms
a gradient between aqueous phase [Trifluoroacetic acid, CH3CN,
H2O (0.1:4.9:95, v/v/v)] and CH3CN at a flow rate of 1 ml/min at
30 8C. We applied gradient elution from 0% to 100% CH3CN in
15 min, and it was maintained at 100% CH3CN for 5 min.
PSA were pre-incubated at 37 8C for 5 min with an NADPH-
regenerating solution (1.3 mM NADP, 3.3 mM D-glucose 6-
phosphate, 3.3 mM MgCl2, and 0.4 U/ml glucose-6-phosphate
dehydrogenase) in 0.1 M potassium phosphate buffer (pH 7.4).
The reaction was initiated by the addition of individual recombi-
nant human CYP isoforms (25 pmol/ml) or human FMO isoforms
(0.125 mg protein/ml) in a total volume of 1 ml and samples were
maintained at 37 8C for various time periods. At each designated
time point, an aliquot was mixed with 2-fold volume of CH3CN,
vortexed, and then centrifuged for 10 min at 13,000 Â g. The
supernatants were subjected to HPLC analysis.
2.3. LC-MS/MS analysis
The LC-MS/MS system consisted of a Thermo TSQ Quantum
Access (Thermo-Fisher) electrospray ionization triple quadrupole
mass spectrometer coupled to an Ultimate 3000 HPLC system
(Dionex Corporation, Sunnyvale, CA). Chromatographic separa-
tions were achieved using a Luna C18 column (150 Â 2 mm) and a
mobile phase consisting of a gradient from 10% to 95% CH3CN.
3. Results
2.4. The metabolism of PSA in mouse and human microsomes
3.1. The metabolism of PSA in mouse and human liver microsomes
PSA was pre-incubated at 37 8C for 5 min with an NADPH-
regenerating solution (1.3 mM NADP, 3.3 mM D-glucose 6-
phosphate, 3.3 mM MgCl2, and 0.4 U/ml glucose-6-phosphate
dehydrogenase) in 0.1 M potassium phosphate buffer (pH 7.4).
The reaction was initiated by the addition of mouse or human liver
microsomes (protein concentration 0.5 mg/ml) or human intes-
tine, kidney, liver or lung microsomes (protein concentration
0.25 mg/ml) and samples were maintained at 37 8C for various
time periods. At each of the designated time-points, 0.1-ml
aliquots were mixed with 0.2 ml of CH3CN, vortexed, and then
centrifuged for 10 min at 13,000 Â g. The supernatants were
We examined the metabolism of PSA in mouse liver micro-
somes (MLM) and human liver microsomes (HLM). We identified
four mono-hydroxyl metabolites (henceforce referred to as mono-
hydroxyl-PSA) and two di-hydroxyl metabolites using LC-MS/MS.
The sodium adduct ions of the four mono-hydroxyl-PSA were
observed at m/z 602.3, 602.4, 602.5 and 602.3, respectively. The
sodium adduct ions of the two di-hydroxyl-PSA were observed at
m/z 618.2 and 618.3, respectively. To determine the position of the
hydroxyl groups of the metabolites, we treated these metabolites
at pH 13 and 70 8C for 3 h. These metabolites were hydrolyzed to

G. Xie et al. / Biochemical Pharmacology 91 (2014) 249–255
251
Fig. 2. Identification of PSA sulfone using LC-MS/MS analysis. (A) Top, MS spectrum of protonated PSA sulfone. (B) Bottom, MS2 spectrum of the above ion at m/z 580.2.
Mouse
Human
give sulindac, suggesting that their hydroxyl groups are not located
on the sulindac moiety, but on the butane-phosphate moiety. The
protonated PSA sulfone was observed at m/z 580.2 (Fig. 2A), which
60
60
was further fragmented to generated ion at m/z 426.06 (Fig. 2B).
40
40
20
0
The identification of PSA sulfone was also supported by our
observation that PSA was hydrolyzed to give sulindac sulfone at pH
13. We did not detect sulindac in the liver microsomes, indicating
that the amide bond of PSA is rather stable in the liver microsomes.
As shown in Fig. 3, PSA was rapidly hydroxylated and
sulfoxidized to yield mono-hydroxyl-PSA, di-hydroxyl-PSA and
PSA sulfone with the former being the major metabolite in liver
microsomes. PSA generated more mono-hydroxyl-PSA in MLM
than in HLM, resulting in the more rapid clearance of PSA in MLM
than in HLM (Fig. 3).
20
0
0
2
4
0
2
4
6
6
4
2
0
4
2
0
3.2. The metabolism of PSA in human liver, intestine, kidney and lung
microsomes
0
2
4
0
2
4
We compared the metabolic behavior of PSA in human liver,
intestine, kidney and lung microsomes. PSA was extensively
metabolized in human liver and kidney microsomes, and to a lesser
extent in intestine microsomes, but not in lung microsomes
(Fig. 4). While mono-hydroxyl-PSA was the major metabolite in
human liver and intestine microsomes, PSA sulfone was dominant
in kidney microsomes.
Time, h
Fig. 3. Kinetics of PSA metabolism in mouse and human liver microsomes. Time course
of the levels of PSA and its metabolites in PSA-treated liver microsomes. PSA
(50
mM) was incubated with mouse or human liver microsomes (protein
concentration 0.5 mg/ml) at 37 8C for up to 3 h. PSA and its metabolites were
extracted at the designated time points and assayed using HPLC.

252
G. Xie et al. / Biochemical Pharmacology 91 (2014) 249–255
Mono-Hydroxyl-PSA
Kidney
Lung
Liver
Di-hydroxyl-PSA
PSA sulfone
Intestine
60
40
20
8
8
6
4
2
0
CYP3A4
CYP2C9
6
4
2
0
0
0
0
2
2
4
6
6
8
8
30
0
2
4
6
0
2
4
6
20
10
0
8
6
4
2
0
8
6
4
2
0
CYP2D6
CYP2C19
4
30
20
10
0
0
0
2
4
6
0
0
2
4
6
6
8
6
4
2
0
8
6
4
2
0
CYP2B6
CYP2E1
CYP1A2
CYP2A6
0
2
4
6
8
Time, h
2
4
6
2
4
Fig. 4. Kinetics of PSA metabolism in human liver, intestine, kidney, and lung
microsomes. Time courses of the levels of PSA and its metabolites in PSA-treated
Time, h
human microsomes. PSA (50
mM) was incubated individually with human liver,
intestine, kidney and lung microsomes (protein concentration 0.25 mg/ml) at 37 8C
for up to 4 h. PSA and its metabolites were extracted at the designated time points
and assayed as described in Methods.
Fig. 5. Kinetics of the metabolism of PSA by human CYPs. PSA (50
mM) was incubated
withindividualhumanCYPisoformsand NADPH-regeneratingsolutionat37 8Cforup
to 6 h. The metabolites were extracted at the designated time points and assayed by
HPLC.
3.3. The metabolism of PSA and sulindac by human CYPs
Cytochrome P450s (CYPs) account for ꢀ75% of drug metabolism
[8]. Of the eight major human CYPs, CYP3A4 uniquely catalyzes the
hydroxylation of PSA to form mono-hydroxyl-PSA and di-hydrox-
yl-PSA (Fig. 5). CYP2D6 is by far the most active isoform to catalyze
the sulfoxidation of PSA at its sulindac moiety to form PSA sulfone.
CYP3A4, 2C9 and 2C19 exhibited minimal activity in the
sulfoxidation of PSA. In contrast, CYP1A2, 2A6, 2B6 and 2E1are
inactive towards PSA (Fig. 5). In comparison, all the CYPs tested
failed to appreciably sulfoxidize sulindac under the same
experimental condition (data not shown), indicating that PSA
(versus sulindac) is a preferred substrate of CYPs.
FMO1
FMO3
FMO5
50
40
30
20
10
0
3.4. The metabolism of PSA and sulindac by human FMOs
We next evaluated the metabolism of PSA by three major
human flavin monooxygenases (FMOs). FMO1, FMO3 and FMO5
were all capable of catalyzing the sulfoxidation of PSA with FMO1
being by far the most active (Fig. 6). Twenty min after the initiation
of the reactions, FMO1 generated 21.4 and 6.5 times higher level of
PSA sulfone than FMO3 and FMO5, respectively. All the FMOs
sulfoxidized PSA far more rapidly than sulindac (Fig. 6), indicating
that PSA (versus sulindac) is a preferred substrate of FMOs. We did
not detect hydroxyl-PSA in the reaction mixtures, indicating that
FMOs do not catalyze the hydroxylation of PSA.
0
2
4
30
20
10
0
3.5. Effect of ketoconazole and quinidine on the metabolism of PSA
0
2
4
We evaluated the effect of ketoconazole (CYP3A4 inhibitor) and
quinidine (CYP2D6 inhibitor) on the metabolism of PSA in liver
microsomes. Ketoconazole strongly inhibited the hydroxylation of
PSA in MLM and HLM in a concentration-dependent manner
(Fig. 7). Ketoconazole also significantly inhibited the sulfoxidation
of PSA in HLM, but unexpectedly enhanced its sulfoxidation in
Time, h
Fig. 6. Kinetics of the oxidation of PS versus sulindac by FMOs. PSA or sulindac (each at
50 M) was incubated with individual human FMOs (FMO1, FMO3 and FMO5) and
m
NADPH-regenerating solution at 37 8C for up to 20 h. The oxidized products were
extracted at the designated time points and analyzed by HPLC. Results were
expressed as percent of input drug.
MLM. On the other hand, quinidine (up to 100
mM) had no

G. Xie et al. / Biochemical Pharmacology 91 (2014) 249–255
253
pH7.4
pH9.5
100
80
60
40
20
0
Mouse
120
Human
100
80
60
40
20
0
0
1
10
100
140
120
100
80
MLM
HLM
160
120
80
40
0
60
40
20
0
0
1
10
100
Ketoconazole, μM
Fig. 7. Effect of ketoconazole on the metabolism of PSA in liver microsomes. PSA
(50 M) was incubated with or without ketoconazole in mouse or human liver
MLM
HLM
HKM
m
microsomes) at 37 8C for 1 h. PSA metabolites were extracted and assayed using
HPLC. Presented are the percentages of control at the indicated ketoconazole
concentrations.
Fig. 8. Effect of alkaline pH on the metabolism of PSA in microsomes. (A) PSA (50
was incubated in liver microsomes at pH 7.4 or pH 9.5 at 37 8C for 1 h. PSA
metabolites were extracted and assayed using HPLC. (B) PSA (50 M) was
mM)
m
incubated in microsomes at pH 7.4 or pH 9.5 at 37 8C for 1 h. PSA metabolites were
extracted and assayed using HPLC.
appreciable effect on the hydroxylation and sulfoxidation of PSA in
MLM and HLM, suggesting that CYP2D6 plays a negligible role in
the metabolism of PSA in liver microsomes.
mono-hydroxyl-PSA by rat hepatocytes. One hour after the
incubation of PSA (35
mM) with rat hepatocytes, 51.5%, 6.8% and
3.6. Effect of alkaline pH and methimazole on the metabolism of PSA
1.9% of PSA were converted to mono-hydroxyl-PSA, di-hydroxyl-
PSA and PSA sulfone, respectively. These results are consistent with
the extensive hydroxylation of PSA in liver microsomes (Fig. 3).
The activities of CYPs (but not FMOs) could be abrogated at
alkaline pH (Cashman, 2005). We observed that the human CYPs
were completely inactive at pH 9.5, whereas the activities of FMOs
were essentially not affected by this alkaline pH. We examined the
metabolism of PSA at pH 9.5 versus pH 7.4. The alkaline pH strongly
inhibited the hydroxylation of PSAin MLM and HLMby 95% and 99%,
respectively (Fig. 8A), confirming that the hydroxylation of PSA in
the liver microsomes is catalyzed by CYPs, rather than FMOs.
As is the case with ketoconazole, the alkaline pH produced
opposite effects on PSA sulfoxidation in MLM and HLM (Fig. 8B).
Ketoconazole and alkaline pH unexpectedly enhanced PSA
sulfoxidation in MLM because they strongly inhibited PSA
hydroxylation, resulting in more substrate available for PSA
3.8. The metabolism of PSA in rat liver cytosol
Since conventional sulindac can be reduced to sulindac sulfide
in rat liver cytosol [9], we treated PSA with rat liver cytosol at 37 8C
in the absence of NADPH. Two hours after the initiation of the
reaction, 32% of PSA was reduced to form PSA sulfide. PSA sulfide
was identified using LC-MS/MS analysis, and its sodium adduct ion
was observed at m/z 570.3.
4. Discussion
sulfoxidation. Indeed, ketoconazole (10
mM) did inhibit the
sulfoxidation of PSA by 21% in MLM at pH 9.5. The strong
sulfoxidation of PSA in human kidney microsomes was essentially
not affected by the alkaline pH (Fig. 8B), indicating that this
reaction in kidney is catalyzed by FMOs but not by CYPs.
We also evaluated the effect of methimazole (FMO inhibitor) on
the metabolism of PSA in microsomes. Methimazole (5 mM)
completely inhibited the sulfoxidation of PSA in human kidney
microsomes, confirming that this reaction is catalyzed exclusively
by FMOs rather than CYPs. On the other hand, methimazole
partially inhibited the sulfoxidation of PSA in MLM and HLM by
48% and 73%, respectively.
This work establishes the metabolic pathways of PSA. As
illustrated in Fig. 9, PSA undergoes the following metabolic
reactions: (1) mono-hydroxylation at its butane-phosphate moiety
to form mono-hydroxyl-PSA; (2) di-hydroxylation at its butane-
phosphate moiety to form di-hydroxyl-PSA; (3) sulfoxidation at its
sulindac moiety to form PSA sulfone; (4) reduction at its sulindac
moiety to form PSA sulfide. The amide bond of PSA is rather stable
since we did not detect sulindac in PSA-treated liver microsomes.
CYPs play a major role in the metabolism of PSA by catalyzing
its hydroxylation reactions. Of the eight major human CYPs,
CYP3A4 uniquely hydroxylates PSA to generate four mono-
hydroxyl metabolites and two di-hydroxyl metabolites. Similarly,
CYP3A4 uniquely catalyzes the hydroxylation of phospho-ibupro-
fen [10] and phospho-tyrosol-indomethacin [11]. Substrate-
binding to CYP3A4 induces a large increase in the volume of its
active site [12], likely accounting for its preference for bulky
3.7. The metabolism of PSA by rat hepatocytes
We examined the metabolism of PSA by cryopreserved
rat hepatocytes. PSA was primarily metabolized to form

254
G. Xie et al. / Biochemical Pharmacology 91 (2014) 249–255
Fig. 9. Overall metabolic pathways of PSA in vitro. PSA undergoes hydroxylation reactions at its butane-phosphate moiety to form mono-hydroxyl-PSA and di-hydroxyl-PSA,
which are uniquely catalyzed by CYP3A4. PSA also can be oxidized and reduced at its sulindac moiety to form PSA sulfone and PSA sulfide, respectively.
substrates like phospho-compounds. Indeed, CYP3A4 accounts for
45–60% of the metabolism of all drugs currently used [13].
On the other hand, among the eight major human CYPs, CYP2D6
is the most active in catalyzing the sulfoxidation of PSA. CYP2D6
substrate binding is uniquely governed by an ion-pair interaction
between a positively charged nitrogen atom and a negatively
charged Asp301 residue [14]. The partially positively charged sulfur
and phosphorus atoms of PSA may interact with the Asp301 residue
of CYP2D6, thereby facilitating its binding and favorable orienta-
tion in the active site. As is the case with phospho-tyrosol-
indomethacin [11], PSA was hydroxylated and sulfoxidized by
CYP3A4 and CYP2D6, respectively. Together, these findings
demonstrate not only the important roles of CYP3A4 and 2D6 in
the metabolism of phospho-compounds, but also the unique CYP-
catalyzed reaction specificity within a given substrate.
We demonstrated that PSA (versus sulindac) is a preferred
substrate of CYPs. Substrate lipophilicity is highly correlated with
their binding affinity with CYPs [15], and PSA is more lipophilic
than sulindac as evidenced by its greater LogP value (4.0 versus
2.3). Moreover, human CYPs generally prefer neutral or basic
substrates. Thus, the phospho-modification of sulindac at its–
COOH group enhances its lipophilicity and masks its acid group,
thus promoting its binding to CYPs.
compounds containing a negative charge (e.g., sulindac) are
generally poor substrates of FMOs [16]. Second, PSA is much
larger than sulindac, and the rate of FMO-catalyzed reactions
increases with larger size of the substrate [17].
FMO1 catalyzes the sulfoxidation of PSA far more rapidly than
FMO3 and FMO5. Similar observations have been reported for other
relatively large compounds, including four thioether compounds
[18], lorcaserin [19] and arecoline [20]. One exception is trimethy-
lamine, a rather small compound whose sulfoxidation reaction is
predominantly catalyzed by FMO3 [21]. These observations are not
surprising since FMO3 (versus FMO1) has a smaller substrate
binding channel and thus prefers smaller substrates [22].
We observed significant inter-tissue differences in the metab-
olism of PSA by human microsomes. PSA was significantly
hydroxylated in human liver and intestinal microsomes but
essentially not in lung or kidney microsomes, likely because
CYP3A4 (the isoform that exclusively catalyzes the hydroxylation
of PSA) is abundant only in human liver and intestine [23]. On the
other hand, PSA was sulfoxidized predominantly in human kidney
microsomes because FMO1 (the only isoform that efficiently
catalyzes the sulfoxidation of PSA) is predominantly expressed in
human kidney [24]. In comparison, PSA was minimally sulfox-
idized in HLM likely because human liver expresses mainly FMO3
and FMO5, but not FMO1 [25]. Human lung microsomes were
essentially inactive towards PSA, and FMO2 is the predominant
isoform in human lung [25]. Therefore, FMO2 should have little or
no activity towards PSA.
FMOs catalyze the sulfoxidation of PSA but not its hydroxyl-
ation at its C–H bonds, since FMOs act on substrates containing soft
nucleophiles such as nitrogen and sulfur. FMOs oxidize PSA far
more rapidly than sulindac, likely for two reasons. First,

G. Xie et al. / Biochemical Pharmacology 91 (2014) 249–255
255
[7] Wong CC, Cheng KW, Xie G, Zhou D, Zhu CH, Constantinides PP, et al.
Carboxylesterases 1 and 2 hydrolyze phospho-nonsteroidal anti-inflammato-
ry drugs: relevance to their pharmacological activity. J Pharmacol Exp Ther
2012;340:422–32.
[8] Sweeney BP, Bromilow J. Liver enzyme induction and inhibition: implications
for anaesthesia. Anaesthesia 2006;61:159–77.
[9] Ratnayake JH, Hanna PE, Anders MW, Duggan DE. Sulfoxide reduction, In vitro
reduction of sulindac by rat hepatic cytosolic enzymes. Drug Metab Dispos:
Biol Fate Chem 1981;9:85–7.
[10] Xie G, Sun Y, Nie T, Mackenzie GG, Huang L, Kopelovich L, et al. Phospho-
ibuprofen (MDC-917) is a novel agent against colon cancer: efficacy, metabo-
We previously examined and documented the metabolism of PS,
whose amide form is PSA (Fig. 1). As is the case with PSA, CYP2D6 is
by farthe most active isoform catalyzing the sulfoxidation of PS [26],
suggesting that the amide bond of PSA is not critical for the CYP-
mediated sulfoxidation of PSA. However, we observed significant
difference in FMO-catalyzed metabolism of PS and PSA. While FMO1
catalyzes the sulfoxidation of PS and PSA at similar rates, FMO3 and
FMO5 are much more active towards PS than PSA. This observation
suggests that the amide bond of PSA plays an important role in FMO-
mediated sulfoxidation of PSA.
The metabolic fate of PSA has important implications for its
safety and efficacy. The CYP3A4-mediated hydroxylation of PSA
leads to its detoxification and elevated water-solubility, thus
facilitating its excretion from the body. The reduction of PSA on its
sulindac moiety to form PSA sulfide may enhance its bioactivity
since PS sulfide (versus PS) has been shown to be more active in
inhibiting cancer growth in mice [27].
lism, and pharmacokinetics in mouse models.
2011;337:876–86.
J Pharmacol Exp Ther
[11] Xie G, Zhou D, Cheng KW, Wong CC, Rigas B. Comparative in vitro metabolism
of phospho-tyrosol-indomethacin by mice, rats and humans. Biochem Phar-
macol 2013;85:1195–202.
[12] Ekroos M, Sjogren T. Structural basis for ligand promiscuity in cytochrome
P450 3A4. Proc Natl Acad Sci U S A 2006;103:13682–87.
[13] Hustert E, Zibat A, Presecan-Siedel E, Eiselt R, Mueller R, Fuss C, et al. Natural
protein variants of pregnane X receptor with altered transactivation activity
toward CYP3A4. Drug Metab Dispos: Biol Fate Chem 2001;29:1454–9.
[14] Ellis SW, Hayhurst GP, Smith G, Lightfoot T, Wong MM, Simula AP, et al.
Evidence that aspartic acid 301 is a critical substrate-contact residue in the
active site of cytochrome P450 2D6. J Biol. Chem. 1995;270:29055–58.
[15] Lewis DF. On the recognition of mammalian microsomal cytochrome
P450 substrates and their characteristics: towards the prediction of human
p450 substrate specificity and metabolism. Biochem Pharmacol 2000;60:
293–306.
In summary, PSA undergoes significant biotransformations in
vitro, generating an array of metabolites. These results will greatly
help our understanding of the pharmacological effects of PSA in
animal and human studies.
[16] Krueger SK, Williams DE. Mammalian flavin-containing monooxygenases:
structure/function, genetic polymorphisms and role in drug metabolism.
Pharmacol Ther 2005;106:357–87.
Conflicts of interest
[17] Nagata T, Williams DE, Ziegler DM. Substrate specificities of rabbit lung and
porcine liver flavin-containing monooxygenases: differences due to substrate
size. Chem Res Toxicol 1990;3:372–6.
The authors have nothing to disclose except for BR, who has an
equity position in Medicon, Pharmaceuticals, Inc.
[18] Usmani KA, Karoly ED, Hodgson E, Rose RL. In vitro sulfoxidation of thioether
compounds by human cytochrome P450 and flavin-containing monooxygen-
ase isoforms with particular reference to the CYP2 C subfamily. Drug Metab
Dispos: Biol Fate Chem 2004;32:333–9.
Acknowledgements
[19] Usmani KA, Chen WG, Sadeque AJ. Identification of human cytochrome P450
and flavin-containing monooxygenase enzymes involved in the metabolism of
lorcaserin, a novel selective human 5-hydroxytryptamine 2 C agonist. Drug
Metab Dispos: Biol Fate Chem 2012;40:761–71.
[20] Giri S, Krausz KW, Idle JR, Gonzalez FJ. The metabolomics of (+/-)-arecoline 1-
oxide in the mouse and its formation by human flavin-containing monoox-
ygenases. Biochem Pharmacol 2007;73:561–73.
This work was supported by the National Institute of Health
Grant R01CA139454 and the shared instrumentation grant, NIH/
NCRR 1 S10 RR023680-1. We thank R. Rieger and T. Koller, Stony
Brook University, for their expert LC-MS/MS analysis of our
samples.
[21] Lang DH, Yeung CK, Peter RM, Ibarra C, Gasser R, Itagaki K, et al. Isoform
specificity of trimethylamine N-oxygenation by human flavin-containing
monooxygenase (FMO) and P450 enzymes: selective catalysis by FMO3.
Biochem Pharmacol 1998;56:1005–12.
[22] Lomri N, Yang Z, Cashman JR. Regio- and stereoselective oxygenations by adult
human liver flavin-containing monooxygenase 3. Comparison with forms 1
and 2. Chem Res Toxicol 1993;6:800–7.
[23] Martignoni M, Groothuis GM, de Kanter R. Species differences between mouse,
rat, dog, monkey and human CYP-mediated drug metabolism, inhibition and
induction. Exp Opin Drug Metab Toxicol 2006;2:875–94.
[24] Yeung CK, Lang DH, Thummel KE, Rettie AE. Immunoquantitation of FMO1 in
human liver, kidney, and intestine. Drug Metab Dispos: Biol Fate Chem
2000;28:1107–11.
[25] Zhang J, Cashman JR. Quantitative analysis of FMO gene mRNA levels in human
tissues. Drug Metab Dispos: Biol Fate Chem 2006;34:19–26.
[26] Xie G, Wong CC, Cheng KW, Huang L, Constantinides PP, Rigas B. Regioselective
oxidation of phospho-NSAIDs by human cytochrome P450 and flavin mono-
oxygenase isoforms: implications for their pharmacokinetic properties and
safety. Brit J Pharmacol 2012;167:222–32.
[27] Huang L, Zhu C, Sun Y, Xie G, Mackenzie GG, Qiao G, et al. Phospho-sulindac
(OXT-922) inhibits the growth of human colon cancer cell lines: a redox/
polyamine-dependent effect. Carcinogenesis 2010;31:1982–90.
References
[1] Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. Cell
2010;140:883–99.
[2] DiS.Ario JA, Alberts DS, Tietze CC, Khullar SK, Bohman VD, Larsen BR, et al.
Sulindac induces regression and prevents progression of sporadic colorectal
adenomas. Gastroenterology 1997;112:A555–.
[3] Roberts II J, Morrow J. Analgesic-antipyretic and antiinflammatory agents and
drugs employed in the treatment gout. In: Hardman JG, Limbird LE, editors.
Goodman and Gilman’s the pharmacological basis of therapeutics. New York:
McGraw-Hill; 2001. p. 687–731.
[4] Piazza GA, Keeton AB, Tinsley HN, Gary BD, Whitt JD, Mathew B, et al. A novel
sulindac derivative that does not inhibit cyclooxygenases but potently inhibits
colon tumor cell growth and induces apoptosis with antitumor activity. Cancer
Prev Res 2009;2:572–80.
[5] Mackenzie GG, Sun Y, Huang L, Xie G, Ouyang N, Gupta RC, et al. Phospho-
sulindac (OXT-328), a novel sulindac derivative, is safe and effective in colon
cancer prevention in mice. Gastroenterology 2010;139:1320–32.
[6] Xie G, Nie T, Mackenzie GG, Sun Y, Huang L, Ouyang N, et al. The metabolism
and pharmacokinetics of phospho-sulindac (OXT-328) and the effect of
difluoromethylornithine. Brit J Pharmacol 2012;165:2152–66.