Evaluation of aspirin metabolites as inhibitors of hypoxia-inducible factor hydroxylases

Article abstract of DOI:10.1039/b814440k

Known and potential aspirin metabolites were evaluated as inhibitors of oxygen-sensing hypoxia-inducible transcription factor (HIF) hydroxylases; some of the metabolites were found to stabilise HIF-α in cells. The Royal Society of Chemistry 2008.

Full text of DOI:10.1039/b814440k

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Evaluation of aspirin metabolites as inhibitors of hypoxia-inducible factor
hydroxylasesw
Benoit M. Lienard,a Ana Conejo-Garcıa,a Ineke Stolze,b Christoph Loenarz,a
´
a
Neil J. Oldham,z Peter J. Ratcliffeb and Christopher J. Schofield*a
Received (in Cambridge, UK) 19th August 2008, Accepted 9th October 2008
First published as an Advance Article on the web 5th November 2008
DOI: 10.1039/b814440k
Known and potential aspirin metabolites were evaluated as
inhibitors of oxygen-sensing hypoxia-inducible transcription
factor (HIF) hydroxylases; some of the metabolites were found
to stabilise HIF-a in cells.
The a,b-hypoxia-inducible transcription factor (HIF) regulates
the transcription of a gene array whose products work to
counteract the effects of hypoxia and establish oxygen home-
ostasis in animals. Prolyl-4-hydroxylation signals for HIF-a
degradation, and asparaginyl-3-hydroxylation reduces the
interaction between HIF-a and a transcriptional co-activator.
In the presence of oxygen, the hydroxylation of human HIF-a
is catalysed by 2-oxoglutarate- (2OG, 1; Fig. 1) dependent
Pro-hydroxylases (PHD1–3) and an Asn-hydroxylase (FIH).
The reduction of HIF hydroxylase activity in hypoxia causes
the accumulation of transcriptionally active HIF-a, so enabling
the hypoxic response. PHD2 has been shown to be the most
important HIF hydroxylase for down-regulating the hypoxic
response during normoxia in healthy cells.1–3
During our efforts to identify HIF hydroxylase inhibitors,
we considered known and potential metabolites of acetyl
salicylic acid (aspirin, 2; Fig. 1). As well as being an anti-
inflammatory, 2 is used for the prevention of heart disease and
cerebral thrombosis.4 Metabolites of 2 and salicylic acid
(3; Fig. 1) may also have biological activity; 2,3-dihydroxy-
benzoic acid (4; Fig. 1),
a
known siderophore,5 and
Fig. 1 Potential inhibitors and related compounds used in the study.
2,5-dihydroxybenzoic acid (5; Fig. 1) and the potential meta-
bolite 2,6-dihydroxybenzoic acid (6; Fig. 1) are proposed to
impact on a variety of biochemical systems.6–8
C-P4H inhibitor.9 3 and 12 are present in the human diet, and
can be metabolised into 4–6 under inflammatory conditions
(ESI, Fig. S1w).11 Aspirin metabolite 4 occurs in humans via
tryptophan metabolism.12
Hippuric acid (7; Fig. 1) and salicyluric acid (8; Fig. 1) are
known major metabolites of aspirin. Related glycine deriva-
tives of hydroxyaryl compounds, such as 9–11 (Fig. 1), are
reported to be collagen prolyl-4-hydroxylase (C-P4H) and
PHD2 inhibitors.9,10 Benzoic acid (12; Fig. 1) can be meta-
bolised via the benzoate degradation/hydroxylation pathway
into 3,4-dihydroxybenzoic acid (13; ESI, Fig. S1w), which is a
Derivatives of 2 and 3 were synthesised (see ESIw) to test the
possibility that they might be HIF hydroxylase inhibitors
(8, 10 and 14–26; Fig. 1). 10 and 24 were of interest, as 10 is
a metabolite of depsipeptide antibiotic pristinamycin P1,13,14
and 24 is a metabolite of anti-inflammatory compounds
mesalazine and sulfasalazine.
a The Department of Chemistry and the Oxford Centre for Integrative
Systems Biology, Chemistry Research Laboratory, University of
Oxford, Mansfield Road, Oxford, United Kingdom OX1 3TA.
E-mail: christopher.schofield@chemistry.oxford.ac.uk;
Fax: +44 (0) 1865 275674; Tel: +44 (0) 1865 275625
b Henry Wellcome Building for Molecular Physiology, University of
Oxford, Oxford, United Kingdom OX3 7BN
Using a non-denaturing electrospray ionisation mass spectro-
metry (ESI-MS) assay (Fig. 2), 3–6, 8, 10 and 14–27, but not
their corresponding esters or compounds without the aromatic
a-hydroxyl group, were found to bind to tPHD2–Fe (with
different affinities) and to compete for binding with 2OG;
aspirin itself was not observed to bind in this assay (tPHD2 is
a truncated form of PHD2 that can be efficiently produced in a
soluble form).15 In the presence of a 19-residue peptide sub-
strate representing the HIF-1a C-terminal oxygen-dependent
w Electronic supplementary information (ESI) available: Experimental
details for synthetic procedures and biological assays are described.
See DOI: 10.1039/b814440k
z Current address: School of Chemistry, University of Nottingham,
University Park, Nottingham, NG7 2RD
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 6393–6395 | 6393

View Article Online
Fig. 3 HIF-1a stabilisation monitored by immunoblot staining SDS-
PAGE analysis. (a) The effect of incubating 28–33 with Hep3B cell
lines on HIF-1a levels after 1, 6 and 18 h. (b) The dose-dependence of
the activity of 28 (from left to right: 0.00625, 0.0125, 0.025, 0.05, 0.1,
0.2 mM) in Hep3B, U-2OS and MCF7 cell lines (18 h incubation);
levels of the HIF target gene CA9 were monitored as a measure of
transcriptional activation. (c) The effect of varying the 2OG concen-
tration (or DMSO as a control) on 28- and 33-mediated HIF-1a
stabilisation.
Fig. 2 Deconvoluted ESI-MS spectra showing the results of the
incubation of tPHD2–Fe with an equimolar amount of 8, 10, 14–18 or
24 after o5 min incubation at 23 1C and sample cone voltage = 80 V.
only very low levels of activity (IC50 4 1 mM) in assays with
isolated tPHD2; the observed low activity may have been due
to partial ester hydrolysis under the incubation conditions. 28
and 32 induced HIF-1a stabilisation in human Hep3B cell
lines in a concentration-dependent manner, whereas 29–31
had no effect (Fig. 3a). These results are consistent with the
inhibition of tPHD2 by 109 and 14 observed in in vitro
experiments. Interestingly, 28 was more potent in cells than
32 or dimethyloxalylglycine (33, Fig. 1) in inducing HIF-1a
stabilisation, and maintained a notably high level of HIF-1a,
even after 18 h when used at 100 mM. Up-regulation of HIF
target genes was observed for 28 above 50 mM in different cell
lines (Fig. 3b), consistent with the determined IC50 value for 14
(the acid analog of 28). The effect of variation in 2OG
concentration on the activity of 28 and 33 is notable because
2OG was observed to reduce the activity of 33, but not 28
(Fig. 3c). Although the reason for this difference is unclear, the
in vivo mechanism of action may be more complex than simple
enzyme inhibition. Despite not being identified as a PHD
inhibitor in vitro, 30 caused HIF-1a accumulation (ESI,
Fig. S4dw). It cannot be ruled out that the compounds were
metabolised to other inhibitors,11,12 that there were other
inhibition modes, or that transport effects account for differ-
ences in potency.
degradation domain, which includes a prolyl hydroxylation site
(CODD, residues 556–574), only 10 and 16 were observed to
form significant amounts of a tPHD2–Fe–CODD–inhibitor
complex (ESI, Fig. S2w). Overall, the MS results were highly
reproducible (except for 16, which appeared to undergo partial
reaction) and indicate that: (i) the glycine side-chain was not
required for binding of the tested compounds to the tPHD2–Fe
complex, but that its presence increased the affinity, (ii) com-
pounds without the a-hydroxyl and glycinyl groups did not
form complexes, (iii) the hydroxyl group stabilised complex
formation (compare 8 and 14–17 in Fig. 2), and (iv) the
mesalazine metabolite, 24, could form
a complex with
tPHD2–Fe (Fig. 2h).
2OG turnover assays, without and with pre-incubation,
revealed that 10 and 14 (B95% at 1 mM) inhibited tPHD2
under both conditions (ESI, Fig. S3w). In contrast, only 14
(B90% at 1 mM) displayed significant inhibition under the
standard assay conditions with FIH. IC50 values for tPHD2
inhibition were determined for the reported9 inhibitors, 10
(113 ꢁ 1.2 mM) and 14 (27.4 ꢁ 1.3 mM), using a fluorescence-
based assay (ESI, Fig. S8w).16
To test the ability of the aspirin derivatives to inhibit HIF
hydroxylases in cells, an ethyl ester derivative, 28, of 14 that
was predicted to be cell-penetrating was prepared, as were
other arylglycinate esters (29–32; Fig. 1). Ester 28 displayed
It was considered that if 14 was an aspirin metabolite, it
might be possible to observe some HIF-1a accumulation by
incubating the Hep3B cell lines with aspirin. HIF-1a
ꢀc
This journal is The Royal Society of Chemistry 2008
6394 | Chem. Commun., 2008, 6393–6395

View Article Online
accumulation was not observed when Hep3B cell lines were
incubated with 0.1–5 mM of 34, 35, 2 or 3 (ESI, Fig. S4a–cw),
suggesting that 3 (or 34) and 2 were not significantly metabo-
lised into 14 or other species that caused HIF-1a stabilisation
under the analysed conditions. In contrast, 35 stabilised
HIF-1a under these conditions (ESI, Fig. S4cw).
hydroxylase inhibition. However, the possibility that this
occurs should be considered, especially for pharmaceuticals
like aspirin that are taken on a long term basis or in high
doses, and whose physiological effects are mediated via in-
completely understood mechanisms.22 It is possible that the
prophylactic properties of aspirin are partly related to the
ability of its metabolites to impact on the HIF system, though
it is important to state that the present study provides no
direct evidence for this proposal.
We then carried out preliminary investigations into whether
aspirin could be metabolised into a HIF hydroxylase inhibitor
in a human. Aspirin (1 g, single dose) was ingested on two
occasions by one of us (B. M. L.), with urine analyses by
LC-MS (ESI, Fig. S5–S7w). Together with experiments in
which synthetic derivatives were mixed with the biological
samples and high resolution MS analyses (o10 ppm mass
accuracy), the results revealed that: (i) 3 was only observed in
urine after aspirin ingestion (ESI, Fig. S6c and dw), (ii) a
significant amount of 7 was present in the urine (ESI, Fig. S6e
and fw) and (iii) 8, reported earlier to be a major aspirin
metabolite,17 was observed only after aspirin intake (ESI,
Fig. S6g and hw). Molecules with a mass of 211 Da were
consistently present in the urine almost exclusively after
aspirin ingestion (ESI, Fig. S7b and cw; a low level of a peak
corresponding to 16 was observed without aspirin ingestion),
suggesting the formation of dihydroxybenzoglycine deriva-
tives (e.g. 14–17) via aspirin metabolism. One of these peaks
(retention time 2.21 min; ESI, Fig. S7cw) was assigned to 14 on
the basis of a comparison with standards, accurate mass and
fragmentation data (ESI, Fig. S7c and dw). Peaks (ESI,
Fig. S7cw) were also assigned to 16 (2.60 min) and 15
(3.18 min) (MS accuracy o15 ppm). 17 was not observed,
consistent with prior work.8 Although these analyses are
preliminary, and the estimated concentration of 14 in the
urine in the two analyses varied (B0.05 mM and B0.5 mM),
they provide evidence for the metabolism of aspirin into 14.
Several endogenous or diet-/drug-derived small organic
molecules have been proposed to act as HIF hydroxylase
inhibitors, including succinate, fumarate, hydralazine (used as
a vasodilator) and flavonoids.15,18–20 Our data reveal the
possibility that aspirin metabolites affect the HIF system via
hydroxylase inhibition. Interestingly, aspirin has recently been
shown to specifically up-regulate the expression of uPAR
(urokinase-type plasminogen activator receptor),21 which is also
strongly induced in a HIF-1a-dependent manner by hypoxia
and the generic 2OG oxygenase inhibitor, 33 (see ESIw).
Analyses of the structures of other pharmaceuticals, includ-
ing those of mesalazine, sulfasalazine and pristinamycin,
suggest that they could give rise to metabolites that are HIF
hydroxylase inhibitors. There is no evidence as yet that the
concentrations of the relevant metabolites will reach high
enough levels to induce physiological effects, nor whether the
observed HIF-a induction in cells (Fig. 3) is solely due to
We thank the European Commission (A. C.-G., Marie
Curie Programme MEIF-CT-2003-500525), the Rhodes Trust
(C. L.), the Wellcome Trust and the BBSRC for support.
Notes and references
1 K. Hirota and G. L. Semenza, Biochem. Biophys. Res. Commun.,
2005, 338, 610.
2 C. J. Schofield and P. J. Ratcliffe, Biochem. Biophys. Res.
Commun., 2005, 338, 617.
3 W. G. Kaelin, Jr, Biochem. Biophys. Res. Commun., 2005, 338, 627.
4 G. Weissmann, Sci. Am., 1991, 264, 58.
5 M. A. Parent, B. H. Bellaire, E. A. Murphy, R. M. Roop,
P. H. Elzer and C. L. Baldwin, Microb. Pathog., 2002, 32, 239.
6 D. R. Haynes, P. F. Wright, S. J. Gadd, M. W. Whitehouse and
B. Vernon-Roberts, Agents Actions, 1993, 39, 49.
7 A. L. Sagone, in Oxygen and Oxy-Radicals in Chemistry and
Biology, ed. M. A. J. Rodgers and E. L. Powers, Academic Press,
New York, 1981, pp. 719–724.
8 M. W. Whitehouse and P. D. Dean, Biochem. Pharmacol., 1965,
14, 557.
9 M. Hirsila, P. Koivunen, V. Gunzler, K. I. Kivirikko and
J. Myllyharju, J. Biol. Chem., 2003, 278, 30772.
10 M. A. McDonough, V. Li, E. Flashman, R. Chowdhury, C. Mohr,
B. M. R. Lienard, J. Zondlo, N. J. Oldham, I. J. Clifton, J. Lewis,
L. A. McNeill, R. J. M. Kurzeja, K. S. Hewitson, E. Yang,
S. Jordan, R. S. Syed and C. J. Schofield, Proc. Natl. Acad. Sci.
U. S. A., 2006, 13, 9814.
11 G. C. Tremblay and I. A. Qureshi, Pharmacol. Ther., 1993, 60, 63.
12 KEGG
PATHWAY
Database-tryptophan
metabolism,
http://www.genome.ad.jp/kegg/pathway/map/map00380.html,
Kanehisa Laboratories, Japan.
13 F. Jolles, B. Terlain and J. P. Thomas, Nature, 1965, 207, 199.
14 G. Jolles, B. Terlain and J. P. Thomas, Therapie, 1965, 20, 1471.
15 K. S. Hewitson, B. M. Lienard, M. A. McDonough, I. J. Clifton,
D. Butler, A. S. Soares, N. J. Oldham, L. A. McNeill and
C. J. Schofield, J. Biol. Chem., 2007, 282, 3293.
16 J. H. Dao, R. J. M. Kurzeja, J. M. Morachis, H. Veith, J. Lewis,
V. Yu, C. M. Tegley and P. Tagari, Anal. Biochem., 2008, DOI:
10.1016/j.ab.2008.09.052.
17 G. Levy and L. P. Amsel, Biochem. Pharmacol., 1966, 15, 1033.
18 H. J. Knowles, Y. M. Tian, D. R. Mole and A. L. Harris, Circ.
Res., 2004, 95, 162.
19 A. Triantafyllou, P. Liakos, A. Tsakalof, G. Chachami,
E. Paraskeva, P. A. Molyvdas, E. Georgatsou, G. Simos and
S. Bonanou, Free Radical Res., 2007, 41, 342.
20 W. J. Wilson and L. Poellinger, Biochem. Biophys. Res. Commun.,
2002, 293, 446.
21 S. Jamaluddin, Biochem. Biophys. Res. Commun., 2006, 348, 618.
22 C. Patrono, B. Coller, G. A. FitzGerald, J. Hirsh and G. Roth,
Chest, 2004, 126, 234.
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 6393–6395 | 6395