Aromatic hydroxylation of salicylic acid and aspirin by human cytochromes P450

Article abstract of DOI:10.1016/j.ejps.2015.03.015

Aspirin (acetylsalicylic acid) is a well-known and widely-used analgesic. It is rapidly deacetylated to salicylic acid, which forms two hippuric acids - salicyluric acid and gentisuric acid - and two glucuronides. The oxidation of aspirin and salicylic acid has been reported with human liver microsomes, but data on individual cytochromes P450 involved in oxidation is lacking. In this study we monitored oxidation of these compounds by human liver microsomes and cytochrome P450 (P450) using UPLC with fluorescence detection. Microsomal oxidation of salicylic acid was much faster than aspirin. The two oxidation products were 2,5-dihydroxybenzoic acid (gentisic acid, documented by its UV and mass spectrum) and 2,3-dihydroxybenzoic acid. Formation of neither product was inhibited by desferrioxamine, suggesting a lack of contribution of oxygen radicals under these conditions. Although more liphophilic, aspirin was oxidized less efficiently, primarily to the 2,5-dihydroxy product. Recombinant human P450s 2C8, 2C9, 2C19, 2D6, 2E1, and 3A4 all catalyzed the 5-hydroxylation of salicylic acid. Inhibitor studies with human liver microsomes indicated that all six of the previously mentioned P450s could contribute to both the 5- and 3-hydroxylation of salicylic acid and that P450s 2A6 and 2B6 have contributions to 5-hydroxylation. Inhibitor studies indicated that the major human P450 involved in both 3- and 5-hydroxylation of salicylic acid is P450 2E1.

Full text of DOI:10.1016/j.ejps.2015.03.015

European Journal of Pharmaceutical Sciences 73 (2015) 49–56
Contents lists available at ScienceDirect
European Journal of Pharmaceutical Sciences
journal homepage: www.elsevier.com/locate/ejps
Aromatic hydroxylation of salicylic acid and aspirin by human
cytochromes P450
1
1
⇑
´
Mirza Bojic , Carl A. Sedgeman , Leslie D. Nagy, F. Peter Guengerich
Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, TN 37232-0146, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Aspirin (acetylsalicylic acid) is a well-known and widely-used analgesic. It is rapidly deacetylated to sali-
cylic acid, which forms two hippuric acids—salicyluric acid and gentisuric acid—and two glucuronides.
The oxidation of aspirin and salicylic acid has been reported with human liver microsomes, but data
on individual cytochromes P450 involved in oxidation is lacking. In this study we monitored oxidation
of these compounds by human liver microsomes and cytochrome P450 (P450) using UPLC with fluores-
cence detection. Microsomal oxidation of salicylic acid was much faster than aspirin. The two oxidation
products were 2,5-dihydroxybenzoic acid (gentisic acid, documented by its UV and mass spectrum) and
2,3-dihydroxybenzoic acid. Formation of neither product was inhibited by desferrioxamine, suggesting a
lack of contribution of oxygen radicals under these conditions. Although more liphophilic, aspirin was
oxidized less efficiently, primarily to the 2,5-dihydroxy product. Recombinant human P450s 2C8, 2C9,
2C19, 2D6, 2E1, and 3A4 all catalyzed the 5-hydroxylation of salicylic acid. Inhibitor studies with human
liver microsomes indicated that all six of the previously mentioned P450s could contribute to both the 5-
and 3-hydroxylation of salicylic acid and that P450s 2A6 and 2B6 have contributions to 5-hydroxylation.
Inhibitor studies indicated that the major human P450 involved in both 3- and 5-hydroxylation of sali-
cylic acid is P450 2E1.
Received 23 December 2014
Received in revised form 28 January 2015
Accepted 23 March 2015
Available online 31 March 2015
Keywords:
Aspirin (acetylsalicylic acid)
Salicylic acid
Cytochromes P450
Aromatic hydroxylation
Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction
The major route of metabolism is conjugation with glycine to
form salicyluric acid, accounting for 20–65% of the products
Aspirin (acetylsalicylic acid) is a commonly used analgesic and
can be used to treat arthritis as well as inflammation. Aspirin is
also used as an antithrombotic to reduce the risk of heart attacks.
Although Charles Frederich von Gerhardt synthesized aspirin in
1853 (Gerhardt, 1853) and Felix Hoffman developed it as a drug
in 1897 (Chemical Heritage Foundation, 2014), our knowledge of
how aspirin is metabolized in humans is still not completely
understood. The acetyl group is quickly hydrolyzed (enzymatically
and non-enzymatically) after oral ingestion to form salicylic acid in
the body (Trnavsky and Zachar, 1975). Humans further metabolize
salicylic acid in a number of different ways (Fig. 1), or the com-
pound can be directly excreted (1–31%) (Hutt et al., 1986).
(Hutt et al., 1986). The other major metabolites result from
conjugation to glucuronides (ester and ether, 1–42%) (Hutt et al.,
1986). Glucuronidation involves a wide array of uridine 50-diphos-
phoglucuronosyltransferases (UGTs), including the 1A1, 1A3, 1A4,
1A6, 1A7, 1A8, 1A9, 1A10, 2B4, 2B7, 2B15, and 2B17 enzymes
(Kuehl et al., 2006). Aspirin metabolism can vary, depending on
the UGTs in different animals. For instance, cats show increased
toxicity to aspirin due to a lack of UGT and have a prolonged
half-life of aspirin (22 h) compared to dogs (7.5 h) (Parton et al.,
2000). The half-life of aspirin has been reported to vary from 1 to
38 h in various animals, being ꢀ6 h in humans (Trnavsky and
Zachar, 1975).
Two minor metabolites of salicylic acid are 2,3-dihydroxyben-
zoic acid (2,3-DHBA) (Crabtree et al., 1958; Grootveld and
Halliwell, 1986) and 2,5-dihydroxybenzoic acid (2,5-DHBA) (Hutt
et al., 1986; Reidl, 1983). These products can be attributed to oxi-
dation by cytochrome P450 (P450) enzymes, as well as non-enzy-
matic Fenton-type reactions. Further conjugation of 2,5-DHBA with
glycine yields gentisuric acid (Wilson et al., 1978) (which in prin-
ciple could also occur by hydroxylation of salicyluric acid)
(Fig. 1). The evidence for P450 oxidation may also be supported
⇑
Corresponding author at: Department of Biochemistry, Vanderbilt University
School of Medicine, 638 Robinson Research Building, 2200 Pierce Avenue, Nashville,
TN 37232-0146, USA. Tel.: +1 615 322 2261; fax: +1 615 343 0704.
´
E-mail addresses: mbojic@pharma.hr (M. Bojic), carl.a.sedgeman@vanderbilt.edu
(C.A. Sedgeman), cellavporte@gmail.com (L.D. Nagy), f.guengerich@vanderbilt.edu
(F.P. Guengerich).
1
These authors contributed equally to the work.
http://dx.doi.org/10.1016/j.ejps.2015.03.015
0928-0987/Ó 2015 Elsevier B.V. All rights reserved.

´
M. Bojic et al. / European Journal of Pharmaceutical Sciences 73 (2015) 49–56
50
Fig. 1. Aspirin metabolism. Aspirin is readily hydrolyzed to salicylic acid, both enzymatically and non-enzymatically. Salicylic acid undergoes conjugation reactions
generating the major metabolites salicyluric acid (apparently catalyzed by an acyl-CoA N-acyltransferase) and glucuronides (gluc) (catalyzed by UGTs). Products of oxidation
of salicylic acid have been attributed to catalytic activity of P450s as well as non-enzymatic Fenton-type reactions (2,3-DHBA). Gentisic acid (2,5-DHBA) can also be
conjugated by an acyl-CoA N-acyltransferase to form gentisuric acid (Wilson et al., 1978). The conjugation reaction shown with the broken line is hypothetical.
by studies on aspirin sensitivity due to polymorphisms in P450
2C9, an enzyme that has been proposed to oxidize aspirin. One
study reported that the variant rs4918758 had significant associa-
tion with aspirin-intolerant urticaria (Palikhe et al., 2011). Attack
by hydroxyl radicals has been reported to form the 2,3-DHBA pro-
duct and is speculated to form some of the 2,5-DHBA product as
well (Grootveld and Halliwell, 1986, 1988; Coudray et al., 1995;
Ghiselli et al., 1992; Ingelman-Sundberg et al., 1991).
NADPH-generating system in a 37 °C reciprocal shaking water
bath. Reactions were quenched using 6 M NaOH to bring the pH
to 13. The aspirin samples stood for 15 min to deacetylate the
acetylsalicylic acid. HClO4 (6 M) was added to precipitate protein
and to acidify (pH 1) the products for extraction with (C2H5)2O
(aspirin) or CH2Cl2 (salicylic acid). Extraction was done twice with
3 ml of (C2H5)2O or CH2Cl2, followed with centrifugation (3000g,
10 min) to separate the organic and aqueous layers.
The objective of this work was to determine the rates of aro-
matic hydroxylation and contributions of individual P450 enzymes
involved in metabolism of aspirin and salicylic acid.
The organic layer was concentrated under a stream of nitrogen,
and the residue was dissolved in 30 ll of CH3CN-H2O (1:9, v/v) for
LC analysis. 2,3-DHBA and 2,5-DHBA were separated by UPLC on a
Waters Acquity octadecylsilane (C18) column (2.1 mm ꢁ 100 mm,
1.7 lm) and quantified using UV (photodiode array) and fluores-
2. Materials and methods
cence detectors (Acquity, Waters, Milford, MA). CH3CN-H2O
(5:95, with 0.1% CH3CO2H, v/v) was used as mobile phase A and
CH3CN-H2O (40:60, with 0.1% CH3CO2H (v/v)) as mobile phase B.
The following gradient program was used with a flow rate of
2.1. Chemicals and enzymes
Acetylsalicylic acid, salicylic acid, and 2,3- and 2,5-DHBA were
obtained from SigmaAldrich (St. Louis, MO). Chelex 100Ò resin
was from Bio-Rad (Hercules, CA). Ten individual human liver sam-
ples from a stock (Schadt et al., 2008) in our laboratory were
pooled (equal weights) to prepare human liver microsomes
(Guengerich, 2014). Escherichia coli recombinant human P450s
2C8 (Shimada et al., 2001), 2C9 (Sandhu et al., 1993), 2C19
(Komatsu et al., 2000), 2D6 (Gillam et al., 1995; Hanna et al.,
2001), 2E1 (Gillam et al., 1994), and 3A4 (Gillam et al., 1993;
Hosea et al., 2000) were prepared and purified as described in
the indicated references. E. coli recombinant rat NADPH-P450
reductase (Hanna et al., 1998) and E. coli recombinant human cyto-
chrome b5 (Guengerich, 2005; Shimada et al., 1986) were prepared
as described.
300 ll/min: starting point of 100% (v/v) A, held at 100% to 2 min
and raised linearly to 70% B (v/v) at 13 min. The column was re-
equilibrated to 100% A at 15 min and held for 2 min more.
2.3. Mass spectrometry analysis
The incubation products were quenched and extracted as
described above. Prior to evaporating the (C2H5)2O under a stream
of nitrogen, the reaction products were reacted with diazomethane
(prepared by alkaline treatment of DiazaldÒ(SigmaAldrich) in 2-(2-
ethoxyethoxy)ethanol) to esterify the carboxylic acids to the
corresponding methyl esters. The diazomethane solution was
added until a yellow color persisted in the solution. The (C2H5)2O
was evaporated under a stream of nitrogen, and the dried product
2.2. Human liver microsome incubations
was dissolved in 30 ll of a mixture of CH3CN-H2O (1:9, v/v). The
products were analyzed using LC–MS by introduction into a
Waters Acquity UPLC system (Waters, Milford, MA) interfaced to
a Thermo-Finnigan LTQ mass spectrometer (Thermo Scientific
Corp., San Jose, CA). Chromatographic separation was achieved
Incubations were conducted at 37 °C in 0.1 or 1.0 ml incubation
mixtures containing 5
buffer (pH 7.4), an NADPH-generating system (10 mM glucose 6-
phosphate, 0.5 mg/ml NADP+, and 2
g/ml yeast glucose 6-phos-
lM P450 in 50 mM potassium phosphate
l
with a Macherey–Nagel HILIC column (4.6 mm ꢁ 100 mm, 5
lm).
phate dehydrogenase) (Guengerich, 2014), and the reaction sub-
strate (0.5 mM aspirin/salicylic acid if not otherwise stated).
Greater volume of incubations was used for aspirin (1.0 ml)
because lower rates of hydroxylation were observed when com-
pared to salicylic acid (incubation volume of 0.1 ml). Incubations
were performed generally for 30 min with or without the
The LC conditions used were an isocratic solution of 97%
CH3CN:3% 5 mM NH4CH3CO2 in H2O (v/v) at 0.3 ml/min. Mass
spectrometry analysis was performed using an atmospheric pres-
sure chemical ionization source (negative ion) with a full scan
range of m/z 50–400. The source current was set at 10
lA, the
capillary temperature at 300 °C, and the vaporizer temperature at

´
51
M. Bojic et al. / European Journal of Pharmaceutical Sciences 73 (2015) 49–56
350 °C. The sheath gas flow rate was 30 and the auxiliary gas flow
3. Results
rate was 5. MS fragmentation analysis was performed with a colli-
sion energy of 35%, 2 m/z isolation width, activation Q of 0.25, and
an activation time of 30 ms. Data was acquired using a Finnigan
Xcalibur software package.
3.1. Detection and separation of oxidized metabolites
Previous studies have utilized direct HPLC electrochemical mea-
surements (Coudray et al., 1995; Ghiselli et al., 1992; Grootveld
and Halliwell, 1986, 1988; Reidl, 1983) for the separation and
analysis of the oxidation metabolites of aspirin. Chromatography
of the products is problematic due to their hydrophilicity.
Derivatization methods to dansylate or to form an acetonide
proved to be unsuccessful. Esterification (with diazomethane)
facilitated separation of the standards, but the assays were not sen-
sitive enough for quantitation (data not shown).
2.4. P450 incubations
Incubations were done at 37 °C in 0.10 ml incubation mixtures
containing 50 mM potassium phosphate buffer (pH 7.4), an
NADPH-generating system (10 mM glucose 6-phosphate, 0.5 mM
NADP+, and 2
lg/ml yeast glucose 6-phosphate dehydrogenase)
(Guengerich, 2014), and 5 mM substrate (salicylic acid or aspirin).
Each P450 enzyme system used in the incubations included
2,5-DHBA has an absorbance maximum at 328 nm (Fig. 2A). This
wavelength was used for fluorescence excitation, and the maxi-
mum fluorescence emission was recorded at 422 nm (Fig. 2C).
2,3-DHBA has an absorbance maximum at 313 nm (Fig. 2B) and a
maximum fluorescence emission at 432 nm (Fig. 2D). A UPLC sep-
aration method developed by Sucharitakul et al. (2013) was used
to separate 2,3-DHBA from 2,5-DHBA using an increasing gradient
of CH3CN in H2O (with 0.1% HCO2H, v/v). Using this protocol, the
tR for 2,5-DHBA was 4.1 min (Fig. 2E), and 2,3-DHBA eluted at
5.9 min (Fig. 2F). The differences in tR and spectral properties
allowed us to quantify the formation of both 2,5-DHBA and 2,3-
DHBA.
0.25
recombinant rat NADPH-P450 reductase, 0.5
nant human cytochrome b5, and a lipid mixture (L-
leoyl-sn-glycero-3-phosphocholine, L- -1,2-dilauroyl-sn-glycero-
3-phosphocholine, and bovine brain phosphatidylserine in a ratio
of 1:1:1 (w/w/w), 40 g/ml) (with 0.5 mM sodium cholate only
l
M of each E. coli recombinant human P450, 0.5
M E. coli recombi-
-1,2-dio-
lM E. coli
l
a
a
l
for P450 3A4) (Imaoka et al., 1992). Incubations proceeded for
30 min with or without the NADPH-generating system.
Incubation samples were quenched, extracted, and analyzed as
shown above in the human liver microsomal incubations.
2.5. Inhibition studies
3.2. Dependence of product formation on NADPH
Salicylic acid (0.5 mM) was incubated with pooled human liver
microsomes in the presence of selective P450 inhibitors.
Naphthoflavone (for P450 1A2, 1 M), methoxalen (for P450 2A6,
M), ticlopidin (for P450 2B6, 5 M), quercetin (for P450 2C8,
M), sulfaphenazole (for P450 2C9, 5 M), fluconazole (for
P450 2C19, 10 M), ketoconazole (for P450 3A4, 2 M), quinidine
(for P450 2D6, 2 M), and 4-methylpyrazole (for P450 2E1,
100 M) (Correia, 2005) were individually added at inhibiting con-
centrations to the incubation mixture by adding 0.1 l of the inhibi-
a-
Product formation from salicylic acid was not detected with
microsomes for 2,3-DHBA or 2,5-DHBA in the absence of NADPH
(Fig. 3A and C). With NADPH, both 2,5-DHBA and 2,3-DHBA were
detected by fluorescence (Fig. 3B). In the aspirin incubation, 2,5-
DHBA was detected but 2,3-DHBA was not (after the cleavage of
the acetyl group in the post-incubation analysis) (Fig. 3D).
l
2
50
l
l
l
l
l
l
l
l
l
tor (dissolved in CH3CN with the exception of 4-methylpyrazole, in
H2O). Inhibition assays were done in triplicate with control CH3CN
and H2O blanks. Concentration-dependent inhibition of P450 2E1
was examined by incubating salicylic acid in the presence of 0, 10,
20, 50, and 100 lM concentrations of 4-methylpyrazole with human
liver microsomes. Statistical analysis was done using a t-test and a
3.3. Confirmation of 2,5-DHBA by mass spectrometry
Aspirin was incubated in the presence of human liver micro-
somes and NADPH, and the hydrolyzed products were extracted
into (C2H5)2O and reacted with diazomethane. Products were
injected onto an LC-MS system for mass analysis. Standard methyl
2,5-DHBA (tR 2.29 min) (Fig. 4A) yielded ions at m/z 135, 152, and
one-way ANOVA in Prism 6 (GraphPad Software, San Diego, CA).
C 400000
2.0
A
E
1.5
1.0
0.5
0.0
300000
200000
100000
0
200
250
300
400
500
600
700
Wavelength (nm)
Wavelength (nm)
2.5
2.0
1.5
1.0
0.5
0.0
D
20000
15000
10000
5000
0
B
F
200
250
300
400
500
600
700
Wavelength (nm)
Wavelength (nm)
Fig. 2. Separation and detection of oxidized products of aspirin and salicylic acid. 2,5-DHBA showed absorbance maxima at 239 and 328 nm (A). Maximum fluorescence
emission was observed at kexcitation 328 nm and kemmission 422 nm (C); the 422 nm emission was used for UPLC detection (tR 4.1 min, E). 2,3-DHBA showed absorbance maxima
at 246 and 313 nm (B). Maximum fluorescence emission was observed at kexcitation 313 nm and kemmission 432 nm (D); the 432 nm emission was used for UPLC detection (tR
4.1 min, F).

´
M. Bojic et al. / European Journal of Pharmaceutical Sciences 73 (2015) 49–56
52
A
B
C
D
− NADPH
+ NADPH
− NADPH
+ NADPH
2,3-DHBA
2,5-DHBA
2,5-DHBA
Fig. 3. NADPH-dependent formation of 2,5- and 2,3-dihydroxybenzoic acid (DHBA). To determine P450 dependent metabolism, salicylic acid and aspirin were incubated in
presence and absence of NADPH. Salicylic acid (1 mM final concentration) was incubated with human liver microsomes (0.5 nmol P450), and NADPH-dependent generation of
2,5- and 2,3-DHBA was observed (A,B) indicating P450-mediated oxidation of salicylic acid. When aspirin (1.5 mM final concentration) was incubated with human liver
microsomes under these conditions, NADPH-dependent formation of 5-hydroxy-2-acetylsalicylic acid (detected as 2,5-DHBA after hydrolysis) was observed (D). No product
formation was observed without NADPH (C) indicating P450 dependent hydroxylation of aspirin.
167 (Fig. 4B), with the aspirin incubation product (tR 2.19 min)
(Fig. 4C) also yielded ions at m/z 135, 152, and 167 (Fig. 4D).
was the major product formed, compared to 2,5-DHBA. With
aspirin only 2,5-DHBA was found (after deacetylation). The maxi-
mum rate of 5-hydroxylation was approximately five times lower
with aspirin than salicylic acid (data not shown).
3.4. Effect of desferrioxamine
Assays with the substrates, salicylic acid and aspirin, were per-
formed in the presence and absence of the iron chelator desfer-
rioxamine (using Chelex 100Ò-treated buffers) (Fig. 5) to address
the contribution of oxygen radicals in non-enzymatic formation
of the observed products. The presence of the chelating agent did
not inhibit but instead increased the formation of both 2,3- and
2,5-DHBA.
3.7. Catalytic activity of individual cytochrome P450s
Recombinant P450 enzymes were incubated with salicylic acid
(Fig. 8). Because of the limit of detection (more sensitive for 2,5-
DHBA, Fig. 2) in these assays, only 2,5-DHBA was quantified.
P450 2C9 showed the highest product formation of 2,5-DHBA, fol-
lowed by 2D6, 2C8, 3A4, 2E1, and 2C19 in respective order. With
aspirin as the substrate, concentrations of the products were too
low to quantitate.
3.5. Time dependence
In a kinetic assay was with salicylic acid, a 30-min time point
was selected for further studies for the two products (Fig. 6).
3.8. Inhibition studies
In order to determine which P450 enzymes are responsible for
the metabolism in liver microsomes, inhibition studies were per-
formed (Fig. 9A and B). Selective inhibitors for P450s 1A2, 2A6,
2B6, 2C8, 2C9, 2C19, 2D6, 3A4, and 2E1 were individually added
to reaction incubation mixtures, either in CH3CN or H2O (only for
3.6. Rates of oxidation
Formation of both 2,3- and 2,5-DHBA from salicylic acid was
concentration-dependent (Fig. 7). With salicylic acid, 2,3-DHBA

´
53
M. Bojic et al. / European Journal of Pharmaceutical Sciences 73 (2015) 49–56
150000
500000
400000
300000
200000
100000
0
B
A
C
t
R 2.29
135
100000
50000
0
152
167
60
80 100 120 140 160 180 200
0
2
4
6
10000
8000
6000
4000
2000
0
2000
1500
1000
500
0
135
D
t
R 2.19
167
152
0
2
4
6
60
80 100 120 140 160 180 200
m/z
t , min
Fig. 4. Confirmation of 5-hydroxylation of aspirin by mass spectrometry. To confirm that aspirin can be directly oxidized, the hydroxylated product was hydrolyzed with
perchloric acid, neutralized, and methylated with diazomethane. The UPLC trace (A) and mass spectrum (B) of standard 2,5-DHBA corresponded to the hydrolyzed and
subsequently esterified aspirin product (C and D, respectively) after incubation with human liver microsomes (0.7 nmol P450).
500
25
400
2,3-DHBA
2,5-DHBA
2,3-DHBA
2,5-DHBA
20
15
10
5
300
200
100
0
0
20
40
60
80
100
120
140
0
Time, min
-
+
-
+
Desferrioxamine
Fig. 6. Time dependence of salicylic acid oxidation. If the hydroxylation of salicylic
acid is metabolism dependent it should show time dependence as well. To confirm
this incubations were performed for 15, 30, 60, and 120 min. The incubation
reaction was done with 0.5 nmol of microsomal P450. A 30 min incubation time
was selected for further kinetic and inhibition assays.
Fig. 5. Effect of desferrioxamine on P450 oxidation of salicylic acid using human
liver microsomes. To determine if hydroxylation of salicylic acid is result of Fenton-
type reactions, incubations were performed in presence and absence of 100 lM
desferrioxamine with 0.5 nmol microsomal P450, incubated for 30 min. Under
these conditions reactive oxygen species production should be inhibited due to
chelation of iron with desferrioxamine. However, no inhibition in production of 2,3
and 2,5- DHBA was observed, indicating P450-dependent enzymatic hydroxylation
of salicylic acid.
6
4
2,3-DHBA
4-methylpyrazole, for P450 2E1). In these assays, it was found that
P450s 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 3A4, and 2E1 all had signifi-
cant contribution (p < 0.05) to the 5-hydroxylation reaction
(Fig. 9A). For 3-hydroxylation, P450s 2C8, 2C9, 2C19, 2D6, and
2E1 had significant contribution (p < 0.05) to the enzymatic activ-
ity (Fig. 9B). When aspirin was used as substrate, there were also
decreases in the oxidation rates in the presence of inhibitors,
although none of the individual enzymes reached statistical signifi-
cance (p < 0.05) (Fig. 9C).
2
2,5-DHBA
0
0
2
4
6
8
Salicylicacid (mM)
For P450 2E1, which showed a major role in Fig. 9A, the concen-
tration dependence was examined for the selective inhibitor 4-
methylpyrazole. 4-Methylpyrazole decreased the formation of
both 2,3- and 2,5-DHBA in human liver microsomal incubations
(Fig. 10).
Fig. 7. Rates of microsomal oxidation of salicylic acid in human liver microsomes
(0.5 nmol P450). Data points were fit to linear equation (linear regression
analysis). Due to limits of solubility of salicylic acid enzyme saturation (Vmax
cannot be achieved and consequently Km determined. Under these conditions the
slope is an upper limit of catalytic efficiency (Vmax/Km), ꢂ0.50 minꢃ1 Mꢃ1 for 2,5-
DHBA, 0.71 minꢃ1 Mꢃ1 for 2,3-DHBA, based on total P450).
a
)

´
54
M. Bojic et al. / European Journal of Pharmaceutical Sciences 73 (2015) 49–56
40
30
20
10
0
6
4
2
2,3-DHBA
2,5-DHBA
0
0
50
100
3A4
2E1
2C8
2D6
2C9
2C19
4-Methylpyrazole (µM)
P450
Fig. 10. P450 2E1 inhibitor titration. The contribution of P450 2E1 to the oxidation
of salicylic acid (0.5 mM) to 2,5- and 2,3-dihydroxybenzoic acid (DHBA) was
measured using different concentrations of the selective P450 2E1 inhibitor 4-
Fig. 8. Rates of 2,5-DHBA production by individual recombinant P450s. Salicylic
acid (0.5 mM) was incubated with 25 pmol of each P450 enzyme for 30 min. Low
hydroxylation rates were observed (11–30 pmol/nmol/min).
methylpyrazole (0, 1, 10, 20, 50, 100 lM) in incubations with human liver
microsomes (0.5 nmol P450, different pool of microsomes than used in Fig. 9). As
production of dihydroxybenzoic acids was dependent on the inhibitor concentra-
tion, hydroxylation of salicylic acid can be attributed to metabolic activity of P450
2E1.
4. Discussion
Previous studies involved HPLC for separation of 2,3- and 2,5-
DHBA using sodium citrate-acetate buffers and CH3OH, with elec-
trochemical detection of the catechol and hydroquinone (Coudray
et al., 1995; Ghiselli et al., 1992; Grootveld and Halliwell, 1986,
1988). Electrochemical detectors are no longer generally used,
and their use is restricted to a few electrochemically active mole-
cules. We utilized acidic H2O/CH3CN mixtures, based on other pre-
vious literature (Sucharitakul et al., 2013). CH3CN was used instead
of CH3OH to limit UV background absorbance during analysis.
Major differences in the fluorescence of the products allowed fur-
ther distinction and quantitation of these compounds.
In human liver microsomes fortified with NADPH, salicylic acid
reproducibly generated 2,3- and 2,5-DHBA but acetylsalicylic acid
generated only 2,5-DHBA (Fig. 3). The results vary with some pre-
vious results with rat and rabbit liver microsomes, which catalyzed
the formation of 2,5- but not 2,3-DHBA (Ingelman-Sundberg et al.,
1991). The differences may result from variance in the P450s in the
different species. Our limits of detection were much higher for 2,3-
DHBA (30 nM) than 2,5-DHBA (3 nM). This could have lead to the
inability to quantify 2,3-DHBA in earlier studies although it is not
clear how the sensitivity compares with an electrochemical
detector.
12
10
A
B
10
8
8
6
*
6
*
4
*
*
4
*
**
*
2
0
*
2
*
*
*
*
0
Inhibited P450
Inhibited P450
2.0
1.5
1.0
0.5
0.0
C
Inhibited P450
Fig. 9. Analysis of P450s responsible for oxidation of salicylic acid (A, B) and aspirin (C) using selective P450 inhibitors on human liver microsomes (see Materials and
Methods). Statistically significant changes (p < 0.05, ⁄) in the rate of salicylic acid (0.5 mM) oxidation using 0.5 nmol human liver microsomes were observed when P450s 2A6,
2B6, 2C8, 2C9, 2C19, 2D6, 3A4, and 2E1 were inhibited. Solvent controls were performed using CH3CN (used for P450s 1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, and 3A4) and H2O
(used for P450 2E1) to examine solvent inhibition.

´
55
M. Bojic et al. / European Journal of Pharmaceutical Sciences 73 (2015) 49–56
Product formation was time-dependent (Fig. 6) and concentra-
tion-dependent (Fig. 7). Mass spectrometry was used to confirm
the formation of 2,5-DHBA in aspirin incubation. Both hydrox-
ylations of salicylic acid were not saturating (Fig. 7), and the results
can be presented as estimates of catalytic efficiency (Vmax/Km),
ꢂ0.50 minꢃ1 Mꢃ1 for 2,5-DHBA, 0.71 minꢃ1 Mꢃ1 for 2,3-DHBA)
but Vmax and Km cannot be calculated individually. This result
may be attributed to multiple P450 enzymes having a role in the
oxidation, e.g. (Figs. 8 and 9), or simply reflect the kinetic behavior
of one or more P450s with the substrate. Although we originally
hypothesized that aspirin would be a better substrate than salicylic
acid (due to consideration of the hydrophilicity of salicylic acid),
the opposite was found to be the case. The oxidation of salicylic
acid is probably more relevant, in light of the known rapid hydroly-
sis of aspirin (Trnavsky and Zachar, 1975).
Incubations done in the presence of the iron chelator desfer-
rioxamine showed that the formation of 2,5-DHBA was catalyzed
by P450 instead of oxygen radicals formed in solution (Fig. 5).
Inhibition assays with selective P450 inhibitors showed roles for
multiple P450 enzymes, i.e. 2C8, 2C9, 2C19, 2D6, 3A4, and 2E1
(Fig. 9). These results (for 2,5-DHBA) were confirmed by perform-
ing incubations in the presence of recombinant P450 enzymes
(Fig. 8).
However, another report showed a similar association with other
non-steroidal anti-inflammatory drugs (Carbonell et al., 2010),
suggesting that the issue may be pharmacodynamic (and possibly
the result of linked genes, not CYP2C9). Our evidence (Figs. 7 and 8)
indicates that P450 2C9 is not dominant, being only one of several
P450s involved (Figs. 8 and 9) similar to the promiscuity observed
with UGTs in the glucuronidation reactions (Kuehl et al., 2006). The
evidence favors P450 2E1 as having the largest single role among
P450s (Figs. 9 and 10), which is not surprising in light of the small
size of the active site of that P450 (Porubsky et al., 2008).
5. Conclusion
We found that human liver microsomes catalyzed the forma-
tion of 2,3- and 2,5-DHBA from salicylic acid and 2,5-DHBA from
aspirin. This formation was NADPH-, time-, and concentration-de-
pendent and was confirmed by mass spectrometry. Formation of
these metabolites was shown not to be due to oxygen radicals
(Fig. 5). Individual P450 incubations and inhibitor assays were used
to determine the enzymes that have activity in these reactions,
including P450s 2C8, 2C9, 2C19, 2D6, 3A4, and 2E1. The con-
tribution of multiple P450s and UGTs (Kuehl et al., 2006) in meta-
bolism is consistent with a lack of issues regarding enzyme
polymorphisms with this widely used drug.
The low in vitro rates of formation of 2,3- and 2,5-DHBA can
explain why the oxidized products are minor metabolites of sali-
cylic acid. The formation of low levels of products may be impor-
tant, in the context of aspirin sensitivity in people with
polymorphisms in P450 2C9 (vide infra).
Disclosures
The authors declare no conflicts of interest. This work was
financially supported by National Institutes of Health grants R37
CA090426 and T32 GM065086.
Although 3-hydroxylation of salicylic acid has been suggested
to be a marker of hydroxyl radicals (Grootveld and Halliwell,
1986), our results with desferrioxamine (Fig. 5) did not show a
decrease in activity. This conclusion is in contrast to a report with
rabbit and rat liver microsomes (Ingelman-Sundberg et al., 1991).
We are unsure what the differences are, but we used Chelex
100Ò-treated buffers in these experiments, and we routinely pre-
pare liver microsomes in the presence of EDTA and butylated
hydroxytoluene to prevent lipid peroxidation artifacts
(Guengerich, 2014; van der Hoeven and Coon, 1974). We have
not specifically prepared microsomes according to the procedure
of Ingelman-Sundberg et al. (Eliasson et al., 1988; Ingelman-
Sundberg and Johansson, 1984; Ingelman-Sundberg et al., 1991)
(lower EDTA concentration, no butylated hydroxytoluene
(Eliasson et al., 1988)), nor have we examined these animal species
in this work.
Although P450 oxidation is a minor pathway of metabolism
(Fig. 1) and low-dose aspirin is often used in cardioprotective dose,
it should be considered that aspirin is a high-dose nonsteroidal
anti-inflammatory drug (typical recommended dose of 2700 to
6000 mg/day for humans) (Beers and Berkow, 1999) and that the
oxidation products (2,3- and 2,5-DHBA) are o- and p-dihydrox-
yphenolic compounds (i.e., a catechol and hydroquinone, respec-
tively) and are capable of oxidation to quinones, potentially toxic
chemicals. However, we are unaware of any evidence that this is
an issue in any toxicity related to aspirin use.
Aspirin has also been reported to induce P450s. A low dose
(50 mg/day for 14 days) was reported to increase P450 2C19 and,
to a lesser extent, P450 3A4, as judged by in vivo assays (mepheny-
toin and midazolam oxidations) (Chen et al., 2003). Thus, in con-
sidering Figs. 8 and 9, aspirin might be expected to increase its
own metabolism.
We found that P450 2C9 had catalytic activity in oxidizing sali-
cylic acid (Figs. 8 and 9) and possibly aspirin itself (Fig. 9C). A
major role for P450 2C9 in aspirin metabolism has been proposed
(Agundez et al., 2009; Palikhe et al., 2011). However, this conclu-
sion appears to be based only on association studies of side effects
(urticaria) of aspirin and some CYP2C9 alleles (Palikhe et al., 2011).
Acknowledgement
We thank K. Trisler for assistance in preparation of the
manuscript.
References
Agundez, J.A., Martinez, C., Perez-Sala, D., Carballo, M., Torres, M.J., Garcia-Martin,
E., 2009. Pharmacogenomics in aspirin intolerance. Curr. Drug Metab. 10, 998–
1008.
Beers, M.H., Berkow, R. (Eds.), 1999. The Merck Manual. Merck Research
Laboratories, Whitehouse Station, NJ, p. 1364.
Carbonell, N., Verstuyft, C., Massard, J., Letierce, A., Cellier, C., Deforges, L., Saliba, F.,
Delchier, J.C., Becquemont, L., 2010. CYP2C9⁄3 loss-of-function allele is
associated with acute upper gastrointestinal bleeding related to the use of
NSAIDs other than aspirin. Clin. Pharmacol. Ther. 87, 693–698.
Chemical Heritage Foundation, 2014. <http://www.chemheritage.org/discover/
online-resources/chemistry-in-history/themes/pharmaceuticals/relieving-
symptoms/hoffmann.aspx>.
Chen, X.P., Tan, Z.R., Huang, S.L., Huang, Z., Ou-Yang, D.S., Zhou, H.H., 2003. Isozyme-
specific induction of low-dose aspirin on cytochrome P450 in healthy subjects.
Clin. Pharmacol. Ther. 73, 264–271.
Correia, M.A., 2005. Inhibition of cytochrome P450 enzymes. In: Ortiz de
Montellano, P.R. (Ed.), Cytochrome P450: Structure, Mechanism, and
Biochemistry. Kluwer Academic, New York, pp. 247–322.
Coudray, C., Talla, M., Martin, S., Fatome, M., Favier, A., 1995. High-performance
liquid
chromatography-electrochemical
determination
of
salicylate
hydroxylation products as an in vivo marker of oxidative stress. Anal.
Biochem. 227, 101–111.
Crabtree, R.E., Data, J.B., Christian, J.E., 1958. A study of the oxidative metabolism of
acetylsalicylic, salicylic, and gentisic acid in fevered animals. J. Am. Pharmaceut.
Assoc. 47, 602–605.
Eliasson, E., Johansson, I., Ingelman-Sundberg, M., 1988. Ligand-dependent
maintenance of ethanol-inducible cytochrome P-450 in primary rat
hepatocyte cell cultures. Biochem. Biophys. Res. Commun. 150, 436–443.
Gerhardt, C.F., 1853. Untersuchungen über die wasserfreien organischen Säuren.
Justus Liebigs Ann. Chem. 87, 149–179.
Ghiselli, A., Laurenti, O., De Mattia, G., Maiani, G., Ferro-Luzzi, A., 1992. Salicylate
hydroxylation as an early marker of in vivo oxidative stress in diabetic patients.
Free Radical Biol. Med. 13, 621–626.
Gillam, E.M., Baba, T., Kim, B.R., Ohmori, S., Guengerich, F.P., 1993. Expression of
modified human cytochrome P450 3A4 in Escherichia coli and purification and
reconstitution of the enzyme. Arch. Biochem. Biophys. 305, 123–131.

´
M. Bojic et al. / European Journal of Pharmaceutical Sciences 73 (2015) 49–56
56
Gillam, E.M., Guo, Z., Guengerich, F.P., 1994. Expression of modified human
cytochrome P450 2E1 in Escherichia coli, purification, and spectral and
catalytic properties. Arch. Biochem. Biophys. 312, 59–66.
Palikhe, N.S., Kim, S.H., Nam, Y.H., Ye, Y.M., Park, H.S., 2011. Polymorphisms of
aspirin-metabolizing enzymes CYP2C9, NAT2 and UGT1A6 in aspirin-intolerant
urticaria. Allergy Asthma Immunol. Res. 3, 273–276.
Gillam, E.M., Guo, Z., Martin, M.V., Jenkins, C.M., Guengerich, F.P., 1995. Expression
of cytochrome P450 2D6 in Escherichia coli, purification, and spectral and
catalytic characterization. Arch. Biochem. Biophys. 319, 540–550.
Grootveld, M., Halliwell, B., 1986. Aromatic hydroxylation as a potential measure of
hydroxyl-radical formation in vivo. Identification of hydroxylated derivatives of
salicylate in human body fluids. Biochem. J. 237, 499–504.
Grootveld, M., Halliwell, B., 1988. 2,3-Dihydroxybenzoic acid is a product of human
aspirin metabolism. Biochem. Pharmacol. 37, 271–280.
Guengerich, F.P., 2005. Reduction of cytochrome b5 by NADPH-cytochrome P450
reductase. Arch. Biochem. Biophys. 440, 204–211.
Parton, K., Balmer, T.V., Boyle, J., Whittem, T., MacHon, R., 2000. The
pharmacokinetics and effects of intravenously administered carprofen and
salicylate on gastrointestinal mucosa and selected biochemical measurements
in healthy cats. J. Vet. Pharmacol. Ther. 23, 73–79.
Porubsky, P.R., Meneely, K.M., Scott, E.E., 2008. Structures of human cytochrome P-
450 2E1. Insights into the binding of inhibitors and both small molecular weight
and fatty acid substrates. J. Biol. Chem. 283, 33698–33707.
Reidl, U., 1983. Determination of acetylsalicylic acid and metabolites in biological
fluids by high-performance liquid chromatography. J. Chromatogr. 272, 325–
331.
Guengerich, F.P., 2014. Analysis and characterization of enzymes and nucleic acids
relevant to toxicology. In: Hayes, A.W., Kruger, C.L. (Eds.), Hayes’ Principles and
Methods of Toxicology, sixth ed. CRC Press-Taylor & Francis Boca Raton, FL, pp.
1905–1964.
Hanna, I.H., Teiber, J.F., Kokones, K.L., Hollenberg, P.F., 1998. Role of the alanine at
position 363 of cytochrome P450 2B2 in influencing the NADPH- and
hydroperoxide-supported activities. Arch. Biochem. Biophys. 350, 324–332.
Hanna, I.H., Kim, M.S., Guengerich, F.P., 2001. Heterologous expression of
cytochrome P450 2D6 mutants, electron transfer, and catalysis of bufuralol
hydroxylation: The role of aspartate 301 in structural integrity. Arch. Biochem.
Biophys. 393, 255–261.
Hosea, N.A., Miller, G.P., Guengerich, F.P., 2000. Elucidation of distinct ligand
binding sites for cytochrome P450 3A4. Biochemistry 39, 5929–5939.
Hutt, A.J., Caldwell, J., Smith, R.L., 1986. The metabolism of aspirin in man: a
population study. Xenobiotica 16, 239–249.
Imaoka, S., Imai, Y., Shimada, T., Funae, Y., 1992. Role of phospholipids in
reconstituted cytochrome P450 3A forms and mechanism of their activation
of catalytic activity. Biochemistry 31, 6063–6069.
Sandhu, P., Baba, T., Guengerich, F.P., 1993. Expression of modified cytochrome
P450 2C10 (2C9) in Escherichia coli, purification, and reconstitution of catalytic
activity. Arch. Biochem. Biophys. 306, 443–450.
Schadt, E.E., Molony, C., Chudin, E., Hao, K., Yang, X., Lum, P.Y., Kasarskis, A., Zhang,
B., Wang, S., Suver, C., Zhu, J., Millstein, J., Sieberts, S., Lamb, J., GuhaThakurta, D.,
Derry, J., Storey, J.D., Avila-Campillo, I., Kruger, M.J., Johnson, J.M., Rohl, C.A., van
Nas, A., Mehrabian, M., Drake, T.A., Lusis, A.J., Smith, R.C., Guengerich, F.P.,
Strom, S.C., Schuetz, E., Rushmore, T.H., Ulrich, R., 2008. Mapping the genetic
architecture of gene expression in human liver. PLoS Biol. 6, e107.
Shimada, T., Misono, K.S., Guengerich, F.P., 1986. Human liver microsomal
cytochrome P-450 mephenytoin 4-hydroxylase,
polymorphism in oxidative drug metabolism.
a
prototype of genetic
Purification and
characterization of two similar forms involved in the reaction. J. Biol. Chem.
261, 909–921.
Shimada, T., Oda, Y., Gillam, E.M., Guengerich, F.P., Inoue, K., 2001. Metabolic
activation of polycyclic aromatic hydrocarbons and other procarcinogens by
cytochromes P450 1A1 and P450 1B1 allelic variants and other human
cytochromes P450 in Salmonella typhimurium NM2009. Drug Metab. Dispos.
29, 1176–1182.
Sucharitakul, J., Tongsook, C., Pakotiprapha, D., van Berkel, W.J., Chaiyen, P., 2013.
The reaction kinetics of 3-hydroxybenzoate 6-hydroxylase from Rhodococcus
jostii RHA1 provide an understanding of the para-hydroxylation enzyme
catalytic cycle. J. Biol. Chem. 288, 35210–35221.
Ingelman-Sundberg, M., Johansson, I., 1984. Mechanisms of hydroxyl radical
formation and ethanol oxidation by ethanol-inducible and other forms of
rabbit liver microsomal cytochromes P-450. J. Biol. Chem. 259, 6447–6458.
Ingelman-Sundberg, M., Kaur, H., Terelius, Y., Persson, J.O., Halliwell, B., 1991.
Hydroxylation of salicylate by microsomal fractions and cytochrome P-450.
Lack of production of 2,3-dihydroxybenzoate unless hydroxyl radical formation
is permitted. Biochem. J. 276, 753–757.
Trnavsky, K., Zachar, M., 1975. Correlation of serum aspirin esterase activity and
half-life of salicylic acid. Agents Actions 5, 549–552.
Komatsu, T., Yamazaki, H., Asahi, S., Gillam, E.M., Guengerich, F.P., Nakajima, M.,
Yokoi, T., 2000. Formation of a dihydroxy metabolite of phenytoin in human
liver microsomes/cytosol: roles of cytochromes P450 2C9, 2C19, and 3A4. Drug
Metab. Dispos. 28, 1361–1368.
van der Hoeven, T.A., Coon, M.J., 1974. Preparation and properties of partially
purified cytochrome P-450 and reduced nicotinamide adenine dinucleotide
phosphate-cytochrome P-450 reductase from rabbit liver microsomes. J. Biol.
Chem. 249, 6302–6310.
Kuehl, G.E., Bigler, J., Potter, J.D., Lampe, J.W., 2006. Glucuronidation of the aspirin
metabolite salicylic acid by expressed UDP-glucuronosyltransferases and
human liver microsomes. Drug Metab. Dispos. 34, 199–202.
Wilson, J.T., Howell, R.L., Holladay, M.W., Brilis, G.M., Chrastil, J., Watson, J.T., Taber,
D.F., 1978. Gentisuric acid: metabolic formation in animals and identification as
a metabolite of aspirin in man. Clin. Pharmacol. Ther. 23, 635–643.