Discovery of neuroprotective agents that inhibit human prolyl hydroxylase PHD2

Article abstract of DOI:10.1016/j.bmc.2021.116115

Prolyl hydroxylase (PHD) enzymes play a critical role in the cellular responses to hypoxia through their regulation of the hypoxia inducible factor α (HIF-α) transcription factors. PHD inhibitors show promise for the treatment of diseases including anaemia, cardiovascular disease and stroke. In this work, a pharmacophore-based virtual high throughput screen was used to identify novel potential inhibitors of human PHD2. Two moderately potent new inhibitors were discovered, with IC₅₀ values of 4 μM and 23 μM respectively. Cell-based studies demonstrate that these compounds exhibit protective activity in neuroblastoma cells, suggesting that they have the potential to be developed into clinically useful neuroprotective agents.

Full text of DOI:10.1016/j.bmc.2021.116115

Bioorg. Med. Chem. 38 (2021) 116115
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Discovery of neuroprotective agents that inhibit human prolyl
hydroxylase PHD2
a
b
a
c
Nicole L. Richardson , Laura J. O’Malley , Daniel Weissberger , Anthony Tumber ,
c
a
b,
*
a,*
Christopher J. Schofield , Renate Griffith , Nicole M. Jones , Luke Hunter
a
School of Chemistry, University of New South Wales (UNSW), Sydney, Australia
b
School of Medical Sciences, University of New South Wales (UNSW), Sydney, Australia
c
Chemistry Research Laboratory, Department of Chemistry and the Ineos Oxford Institute for Antimicrobial Resistance, 12, Mansfield Road, Department of Chemistry,
University of Oxford, OX1 3TA, United Kingdom
A R T I C L E I N F O
A B S T R A C T
Keywords:
Prolyl hydroxylase (PHD) enzymes play a critical role in the cellular responses to hypoxia through their regu-
Prolyl hydroxylase
Hypoxia inducible factor
Neuroprotection
Virtual screen
Oxidative stress
Docking
lation of the hypoxia inducible factor
α
(HIF- ) transcription factors. PHD inhibitors show promise for the
α
treatment of diseases including anaemia, cardiovascular disease and stroke. In this work, a pharmacophore-based
virtual high throughput screen was used to identify novel potential inhibitors of human PHD2. Two moderately
potent new inhibitors were discovered, with IC50 values of 4
μ
M and 23 M respectively. Cell-based studies
μ
demonstrate that these compounds exhibit protective activity in neuroblastoma cells, suggesting that they have
the potential to be developed into clinically useful neuroprotective agents.
1
. Introduction
to enhanced expression of various hypoxia-inducible genes involved in
ameliorating the response to hypoxia.1
1-14
In humans there are three
Exposure to mild hypoxia can give a variety of benefits in health and
HIF-
α
isoforms (HIF-1
α
–3
α
), which have different and at least to some
medicine. Some athletes subject themselves to intermittent hypoxia to
extent, context dependent roles in upregulating expression of gene sets.
1
enhance their endurance and performance. In the medical field, expo-
HIF target genes include vascular endothelial growth factor, which
promotes angiogenesis and neurogenesis;1
5-17
sure to mild hypoxia is reported to have positive effects in diseases
ranging from chronic heart and lung diseases to iron-deficiency and
anaemia.2 Neuroprotective effects have also been demonstrated:
intermittent hypoxia can improve walking function in patients suffering
erythropoietin, which
stimulates red blood cell production and reduces blood–brain barrier
,3
18,19
leakage;
pyruvate dehydrogenase kinase-1, which reduces mito-
chondrial oxygen consumption;20 and glucose transporter-1, which
4
21
from an incomplete spinal cord injury; it can alleviate short term
returns the cell toward normal metabolism and glucose transport.
memory deficits in patients with mild cognitive impairments;5 and
hypoxic postconditioning can improve function in animal models of
Under normoxic conditions, HIF-
The degradation process begins with the hydroxylation of two proline-
residues located in oxygen dependent degradation domains of HIF-1
and 2 . The presence of these hydroxyl groups promotes binding of HIF-
to the von Hippel-Lindau protein, which is part of an ubiquitin E3
α
is rapidly degraded in the cell.22
6
stroke.
α
Many of these protective effects may involve the chronic hypoxic
response,7,8 a biochemical system in which the transcription factors
known as hypoxia inducible factors (HIFs) play a key role. HIF is an
α
1α
2
3-26
ligase complex, leading to proteasomal degradation of HIF-1
hydroxylation of HIF- is catalyzed by prolyl hydroxylase (PHD) en-
zymes (Fig. 1a).23 The PHDs are dependent on a ferrous iron cofactor
α.
The
α
,β-heterodimer, comprising an oxygen-sensitive alpha subunit (HIF-
α
)
α
that under normoxic conditions localizes in the cytoplasm and a
2
7
constitutively-expressed beta subunit (HIF-β) that localizes in the nu-
and the cosubstrates 2-oxoglutarate, and molecular oxygen (Fig. 1b).
9
,10
cleus.
Under hypoxic conditions, HIF-
α
accumulates, leading to its
In the proposed PHD mechanism, molecular oxygen coordinates directly
to the single Fe(II) ion in the PHD active site, and it ultimately provides
the oxygen atom of the hydroxyl group that becomes attached to Pro564
translocation into the nucleus, where it dimerizes with HIF-β. The het-
erodimeric
,β -HIF complex binds to hypoxia response elements leading
α
*
Corresponding authors.
E-mail addresses: n.jones@unsw.edu.au (N.M. Jones), l.hunter@unsw.edu.au (L. Hunter).
https://doi.org/10.1016/j.bmc.2021.116115
Received 29 January 2021; Accepted 13 March 2021
Available online 24 March 2021
0
968-0896/© 2021 Elsevier Ltd. All rights reserved.

N.L. Richardson et al.
Bioorganic & Medicinal Chemistry 38 (2021) 116115
of HIF-
α
.28 The oxygen dependence of PHDs is the key feature that en-
2. Results and discussion
ables the hypoxic response to be activated under the appropriate
conditions.29
We pursued a pharmacophore-based virtual screening approach,41,42
There are normally three mammalian isoforms of the PHDs, but
PHD2 is the most highly conserved. PHD2 is localized mainly in the
cytoplasm and its expression is induced by hypoxia in a HIF dependent
manner. Suppression of PHD2 by small interfering RNA upregulates HIF-
because pharmacophore-based screening has been shown to be more
efficient and effective than docking-based screening.4
3,44
A literature
search was conducted for all crystal structures of PHD2 with an inhibitor
bound, and nine pharmacophores were generated, which encode for the
interactions between the inhibitor and the protein, from these crystal
structures using the program LigandScout.44 A virtual screen was then
performed using compound libraries from SPECS and InterBioScreen,
first including all features of the pharmacophore, then progressively
lowering the number of matched features to increase the number of hits.
By this means, 104 candidates that matched at least two pharmaco-
phores were identified. We then scanned these 104 candidates using
1
in normoxia.31 Complete deletion of PHD2 is fatal to mice, but con-
α
ditional deletion leads to increased levels of vascular endothelial growth
factor and erythropoietin.23
There has been significant interest in identifying small-molecule
inhibitors of PHD2, with a view to therapeutic applications of such in-
hibitors through their ability to upregulate the hypoxia response.2
Most inhibition studies of PHD2 have focused on the iron cofactor, with
some early inhibitors either sequestering iron or competing with 2-oxo-
,3
FAFDrugs4 (Free ADME-Tox Filtering Tool).45 This tool identifies so-
3
46
glutarate for binding to PHD2. For example, the iron chelator defer-
called pan-assay interference compounds (PAINs)
and assigns a
oxamine32 has been investigated as
a
potential neuroprotective
toxicity risk rating to each candidate. We excluded all candidates that
contained PAINs or had a risk rating higher than “low,” thereby reducing
the number of candidates from 104 to 72. The next step was to exclude
candidates that had physicochemical properties unfavorable for
blood–brain barrier permeability (e.g. molecular weight above 400 Da;
number of hydrogen bond donors and acceptors greater than 3 and 7
respectively).39 We exercised some judgement in this regard, because we
assumed that there would be opportunities to subsequently improve the
physicochemical properties of candidates that were slightly outside the
optimal parameters. For example, we retained some candidates that
contained carboxylic acid groups because we reasoned that they might
subsequently be replaced with bioisosteres.38 Such considerations of
possible blood–brain barrier permeability narrowed the list of candi-
dates from 72 to 42. We then docked each of these 42 candidates into a
crystal structure of PHD2 (PDB: 4KBZ) using the Gold software and
assessed the docked poses using ChemScore. Any compound with a
docking score inferior to –30 was excluded. Amongst the remaining
candidates were several families of closely related analogs, and we
excluded the higher molecular weight members of such families because
our priority was to discover scaffold diversity. Finally, we performed a
literature search of each remaining candidate to ensure that any future
medicinal chemistry development would be unencumbered by intel-
lectual property constraints. This process delivered a final set of 18
virtual hit compounds (Table 1).
treatment for intracerebral haemorrhage as well as for neurodegenera-
3
3,34
tive conditions such as Parkinson’s and Alzheimer’s diseases.
However, simple iron chelators have poor selectivity because iron is also
3
important in many other cellular processes. Direct active site binders of
PHD2 that compete with 2-oxoglutarate35 are approved and are un-
dergoing preclinical and clinical testing for diseases such as anaemia,
inflammatory diseases and heart disease.3
6,37
Such compounds bind to
the active site both through chelation of iron and by competing with 2-
oxoglutarate.2 The majority of known inhibitors were not designed
with blood–brain barrier permeation as a priority and contain groups
that replace the salt bridge interaction between 2-oxoglutarate and
Arg383 (Fig. 1b).38 These groups usually contain a negative charge
which is problematic for blood–brain barrier permeation, rendering
such compounds inapplicable to the treatment of neurological
diseases.3
,35
9,40
The aim of this work was to identify novel PHD2 inhibitors that could
serve as lead compounds for the treatment of neurological diseases and
injuries such as stroke. Herein, we describe a pharmacophore-based
virtual high throughput screening approach for the discovery of such
inhibitors. We also report the results of validation experiments to
measure the virtual hits’ PHD2 inhibitory potency and their effects on
cell viability and neuroprotection.
To enable the virtual hits to be validated, compounds 1–17 were
Fig. 1. HIF-
green), with N-terminus and C-terminus labelled as N and C respectively (PDB: 5L9B). Atoms colored by element (C in grey, N in blue, O in red). (b) The active site
of PHD2 contains a metal ion (pink; Mn substituting for Fe(II) in the crystal structure) co-ordinated by two histidines, an aspartate, and in a bidendate manner by 2-
α
binding by human PHD2. (a) View from a crystal structure of PHD2 (grey) bound to the C-terminal oxygen dependent degradation domain of HIF-1
α
3
0
(
oxoglutarate (yellow). The proposed site of O
respectively.
2
coordination is indicated by an arrow. Salt bridges and hydrogen bonds are shown by orange and green dashed lines
2

N.L. Richardson et al.
Bioorganic & Medicinal Chemistry 38 (2021) 116115
Table 1
Table 1 (continued)
Structures, docking scores, physiochemical properties, and PHD2 inhibitory
activities of the virtual hit compounds. Physicochemical properties are color-
a–e
coded based on their adherence to BBB permeability guidelines.
a
MW green if < 400.
b
ClogP green if 1.5–2.7, yellow if 0.5–3.7, orange if outside this range.
HBD green if 3 or less.
c
d
e
HBA green if 7 or less, yellow if 9 or less.
tPSA green if < 75, yellow if 75–100, orange if > 100. ND: not determined.
3

N.L. Richardson et al.
Bioorganic & Medicinal Chemistry 38 (2021) 116115
(a)
γ-aminobutyric acid,
toluene, 140 °C
H
H
O
O
O
70%
N
CO H
2
H
H
O
O
1
9
18
(
b)
NH2
(
i) carbonyl diimidazole,
S-methylisothiouronium
sulfate, NaOH, H2O,
H N
2
THF
H N NH
N
NH
O
2
(
ii) H NNH .H O
2 2 2
HO C
F
F
50 °C
F
2
O
1
4%
Cl
Cl
Cl
(
3 steps)
2
0
21
24
22
H
N
H
H O, ∆
68%
2
N
N
N
F
F
N
N
HN
HN
Cl
Cl
O
O
2
Boc-β-Ala, HBTU,
H
Et N, DMF
3
N
N
BocHN
N
N
N
F
N
33%
N
N
N
N
H N
2
NHBoc
Cl
N
O
O
23
nOe
N
N
N
N
F
F
N
N
H N
H N
2
2
Cl
Cl
2
25
Scheme 1. Synthesis of potential PHD2 inhibitors. (a) Synthesis of compound 18; (b) Attempted synthesis of compound 2.
purchased and compound 18 was synthesized following a literature
the putative precursor 23, but further progress was thwarted when the
acylation of 23 occurred not at the exocylic primary amino group, but
4
7
procedure (Scheme 1a). The purity and identity of each virtual hit
compound was checked by NMR analysis. All compounds were found to
be highly purified and of the correct structure, except for 2 which
seemed to have become degraded and/or to be of the wrong constitu-
tion. No reported synthetic methods for 2 are available, so we attempted
to synthesize it via a new route (Scheme 1b). We successfully obtained
rather on the triazole ring (Scheme 1b); there is some precedent for the
regiochemistry of such a transformation.4
8,49,50
We speculate that the
commercial material that was supplied as 2 might actually have been a
different structural isomer (e.g. 2′, Scheme 1a). Further, it is known that
ring-acylated 1,2,4-triazoles are unstable,50 and this could explain the
Fig. 2. Docked poses of 8 (green) and 16 (orange) in the active site of PHD2. See Fig. 1 for colors. Additionally, grey, yellow, green and dark and light purple dashed
lines depict metal acceptor, pi-sulfer, hydrogen bond, pi-pi stacked and pi-alkyl interactions respectively.
4

N.L. Richardson et al.
Bioorganic & Medicinal Chemistry 38 (2021) 116115
extraneous signals that we observed in the NMR spectra of 2/2′. We did
not pursue further methods for synthesizing 2, so this candidate was
excluded from subsequent biological studies.
permeability, and while the physicochemical properties of compounds 8
and 16 are not ideal in this regard, there is scope for their future opti-
mization through medicinal chemistry. The preliminary findings
described herein could assist in the future development of treatments for
neurological diseases and injuries such as stroke.
The virtual hit compounds 1 and 3–18 were subjected to a reported
mass spectrometry-based biochemical assay to determine their ability to
inhibit the hydroxylation of a peptide representing the C-terminal oxy-
gen dependent degradation domain of HIF-1
α
as catalyzed by the cata-
4. Experimental
5
1
lytic domain of PHD2 (aa. 181–426). Two of the virtual hits (8 and 16)
exhibited significant inhibition of PHD2, with IC50 values of 4.09 µM and
4.1. Programs used
2
6.03 µM, respectively (PHD final assay concentration: 150 nM)
(
Table 1). Our earlier docking results had predicted that both 8 and 16
would inhibit PHD2 by coordinating the iron cofactor and occupying the
-oxoglutarate binding site including by interacting with Arg383, which
Discovery Studio Client v17.2.0.16349 (Dassault Syst `e mes Biovia
Corp, San Diego, CA, USA); Gold Version 5.3.0 (Cambridge Crystallo-
graphic Data Centre); Maestro Version 11.3.016 (Schr o¨ dinger, LLC, New
York, NY, USA); LigandScout Version 4.1.1 (Inte:Ligand, Vienna,
Austria)
2
forms a salt bridge with the C-5 carboxylate of 2-oxoglutarate within the
PHD2 active site (Fig. 2).
The cellular effects of 8 and 16 were explored in SH-SY5Y neuro-
blastoma cells (Fig. 3). The MTT cell viability assay was first used to
investigate possible toxicity (Fig. 3a). With this assay compound 8 was
found to be essentially non-toxic at all concentrations employed, while
4.2. Generating pharmacophores
The crystal structures with PDB codes 4BQW52, 4BQX52, 4BQY52
,
4KBZ, 2HBT, 2HBU, 2G1953 and 2G1M were opened in LigandScout.
Pharmacophores of each were generated by encoding the electronic and
steric interactions between the bound ligand and protein as features.
53
1
6 reduced cell viability only at the highest applied concentration of
1
00 µM. The toxic concentration of 16 is substantially higher than the
IC50 value of this compound (i.e. 26.03 µM, Table 1). Next, the neuro-
protective activities of 8 and 16 were evaluated by administering each
compound in various concentrations to the cell culture then exposing the
4.3. Virtual high throughput screen
cells to oxidative stress (50 µM H
2 2
O ). Manual counting of phase
contrast microscopy images revealed that pre-treatment with 8 or 16
significantly reduced the rate of cell death, in a concentration-
dependent fashion (Figure 3b).
Digital compound libraries were acquired from SPECS (https
://www.specs.net/) and InterBioScreen (https://www.ibscreen.com/).
Screening was performed in LigandScout by comparing the structures of
the virtual compounds with the generated pharmacophores and deter-
mining how many features are matched in each structure. Checking for
exclusion volumes was disabled, where exclusion volumes are the po-
sitions in the active site that are sterically occupied by the macromo-
lecular environment, and all features were considered in the first screen
before sequentially decreasing the number of matched features to in-
crease the number of hits. The hits generated were analysed for multiple
pharmacophore matches before subjecting to a PAINS screen (https:
//fafdrugs4.rpbs.univ-paris-diderot.fr/).
3
. Conclusion
We have identified two moderately potent PHD2 inhibitors through a
pharmacophore-based virtual high throughput screening approach.
Compounds 8 and 16 both exhibit concentration-dependent protective
activity in neuroblastoma cells when administered as pre-treatments to
an oxidative stress insult. The mechanism of neuroprotection by com-
pounds 8 and 16 likely involves increasing levels of HIF-1 , thereby
α
upregulating the hypoxia response, although this has yet to be demon-
strated. We designed our virtual screen to prioritize blood–brain barrier
Fig. 3. Effects of 8 and 16 in neuroblastoma
cells. (a) MTT assays reveal that 8 and 16 are
essentially non-toxic in SH-SY5Y cells at
concentrations below 100
μ
M; red dashed
line indicates 100% viability control. (b)
Compounds 8 and 16 both exhibit dose-
dependent neuroprotective activity in SH-
SY5Y cells when administed as
treatment 2 h before H 50
a
pre-
2
O
2
μ
m treat-
ment; Data expressed as mean ± SEM, from 4
separate cultures. Data analysed by one-way
ANOVA (Dunnett’s multiple comparisons).
*
***p < 0.0001 vs. 100% viability control
MEM); #p < 0.05, ##p < 0.01, ###p <
.001, ####p < 0.0001 vs. vehicle treat-
(
0
ment; red dashed line indicates control (no
pre-treatment).
5

N.L. Richardson et al.
Bioorganic & Medicinal Chemistry 38 (2021) 116115
4
.4. General minimisation procedure
(t, J = 7.2 Hz, 2H), 1.89 (tt, J = 7.2, 7.4 Hz, 2H), 1.51 (dt, J = 9.9, 1.7
Hz, 1H), 1.20 (d, J = 9.9 Hz, 1H); data in accordance with literature
values.54
Energy minimisations were performed in Discovery Studio (DS) after
applying the CHARMm forcefield. Default settings were used except the
maximum steps were changed to 10,000. Minimizations were consid-
Synthesis of 21: Carbonyldiimidazole (1.48 g, 9.12 mmol) was
added to a solution of 3-chloro-4-fluorobenzoic acid (1.22 g, 7.01 mmol)
in THF (10 mL) and the mixture was stirred at rt for 3 h. The mixture was
added dropwise to 50–60% hydrazine hydrate (2 mL, 20.6 mmol) and
the mixture was stirred overnight. The solvent was evaporated and the
crude product was carried on to the next step without further
purification.
ꢀ
1
ered converged with gradient tolerance (0.1000000 kcal/mol Å
satisfied.
)
4
.5. Crystal structure preparation
Crystal structures were prepared by first downloading the pdb file
Synthesis of 22: S-Methylisothiourea hemisulfate (194 mg, 1.39
mmol) was added to a solution of intermediate 21 (259 mg) in aqueous
NaOH (1% w/v, 8 mL). The mixture was stirred for 48 h at rt, then for a
from the Protein Data Bank and deleting the bonds associated with the
metal cation. The corrected file was then imported into DS and water
molecules were removed and hydrogens added to account for any
missing in the crystal structure. A three-part minimisation was per-
formed with different components of the crystal structure constrained at
each part. The first minimisation was performed on the hydrogens with
all atoms fixed excluding hydrogen. The second was performed on the
side chains with the backbone and bound ligand fixed. The final mini-
misation was performed on all amino acids outside of the active site,
where the active site is a defined sphere around the ligand that en-
compasses the interacting amino acid residues. The ligand was then
removed from the structure and the receptor was ready for docking.
◦
◦
further 2 h at 50 C. The mixture was cooled to 0 C, filtered and the
precipitate washed with water and dried. The precipitate was purified
by flash chromatography (5–30% MeOH/DCM) to yield a grey solid
◦
ꢀ 1
)
(45.5 mg, 14% over 2 steps); m.p. 155–156 C; IR (neat) vmax (cm
–
–
1
3489 (NH
2
), 1649 (C O); H NMR (600 MHz, DMSO‑d ) δ 10.15 (br s,
6
1H), 8.11 (dd, J = 7.6, 1.8 Hz, 1H), 7.90 (ddd, J = 8.6, 5.0, 1.8 Hz, 1H),
1
3
1
7.30 (dd, J = 8.6, 8.9 Hz, 1H), 7.04 (br s, 2H), 6.77 (br s, 2H); C{ H}
NMR (150 MHz, DMSO‑d ) δ 159.3, 158.0, 156.4, 152.9, 128.8, 127.3
6
(d, J = 7.1 Hz), 118.4 (d, J = 17.3 Hz), 115.7 (d, J = 20.5 Hz); HRMS
+
+
(ESI, +ve) C
H
8 9
ClFN
4
O
[M + H ] requires m/z 231.0443, found
2
31.0440.
4
.6. Docking
Synthesis of 23: A solution of 22 (29.5 mg, 0.128 mmol) in water (4
mL) was heated at reflux for 24 h. After cooling to rt, the product was
Docking was performed using GOLD docking software through DS.
extracted with EtOAc (3 × 5 mL), dried and the solvent evaporated to
◦
The crystal structure ligand was docked into the crystal structure to
determine efficiency of the docking method, with the method that had
the closest pose and interactions to the original carried forward. The
fitness function ChemScore and the default settings were used except the
number of dockings was changed to 100, detect cavity was false and
early termination was false. The flexibility parameters were changed
from the default to increase the flexibility: explore ring conformations
was changed to flip ring corners, flip ring amide bonds was set to true,
flip planar R-NR1R2 set to flip all, flipping pyramidal nitrogens set to
true, intramolecular hydrogen bonds set to true, protonated carboxylic
acids set to flip and fix rotatable bonds set to none. Clustering of the 100
poses was performed and the top poses in large clusters at 2 Å were
considered and the interactions with the protein analysed.
yield a white solid (18.5 mg, 68%); m.p. 242–243 C; IR (neat)
ꢀ
1
v
max (cm ) 3427 (NH), 3185 (NH); 1H NMR (600 MHz, DMSO‑d
6
) δ
12.20 (br s, 1H), 7.94 (dd, J = 6.1, 1.8 Hz, 1H), 7.84 (ddd, J = 8.8, 6.1,
1
3
1
1.8 Hz, 1H), 7.45 (t, J = 8.8 Hz, 1H), 6.15 (br s, 2H); C{ H} NMR (150
MHz, DMSO‑d
) δ 158.4, 158.0, 156.8, 130.6, 127.5, 126.3 (d, J = 7.1
6
Hz), 120.1 (d, J = 18.6 Hz), 117.6 (d, J = 20.4 Hz); HRMS (ESI, +ve)
+
+
C
8
7
H ClFN
4
[M + H ] requires m/z 213.0338, found 213.0332.
Synthesis of 25: Boc-β-alanine (97.8 mg, 0.517 mmol), HBTU (268
mg, 0.707 mmol) and Et
3
N (50 L, 0.36 mmol) were added to a solution
μ
of 23 (54.6 mg, 0.257 mmol) in DMF (3 mL). The mixture was stirred at
rt for 2.5 h then quenched with ice-cold water (3 mL). The mixture was
extracted with EtOAc (3 × 4 mL) and the organic layer was washed with
1 M NaOH (2 × 5 mL) and brine (3 × 5 mL), dried, and the solvent was
evaporated. The crude product was purified by flash chromatography
(2% MeOH/DCM) to yield a white solid (32.7 mg, 33%) [regiochemistry
4
.7. Synthetic reagents and instrumentation:
5
0
◦
–
–
assigned by analogy with literature ]; m.p. 162 C; IR (neat) vmax
ꢀ 1
–
Synthetic reagents were from Sigma-Aldrich, Combi-Blocks or
(cm ) 3470 (NH), 3383 (NH), 3133 (NH), 1712 (C
O), 1704 (C O);
–
1
Enamine and were used without further purification. Reactions were
monitored by thin layer chromatography performed on plates contain-
ing Merck aluminium-backed silica gel 60 F254 (0.2 mm), and visual-
H NMR (400 MHz, CDCl
3
) δ 8.10 (dd, J = 7.2, 2.2 Hz, 1H), 7.91 (ddd, J
= 8.7, 4.6, 2.2 Hz, 1H), 7.19 (t, J = 8.7 Hz, 1H), 6.33 (br s, 2H), 5.04 (br
s, 1H), 3.58 (dt, J = 6.0, 5.8 Hz, 2H), 3.29 (t, J = 5.8 Hz, 2H), 1.43 (s,
9H); 13C{ H} NMR (150 MHz, CDCl
1
) δ 160.3, 158.8, 158.6, 157.2,
isation was achieved by KMnO
4
stain or UV light. Flash column
m silica
3
chromatography was performed with SiliaFlash® P60 40–63
μ
129.5, 127.3 (d, J = 4.0 Hz), 127.0 (d, J = 7.5 Hz), 121.6 (d, J = 18.0
gel. NMR spectra were recorded at 298 K on Bruker Avance III 300, 400
and 600 MHz instruments. IR spectra were recorded on a Cary 630 FTIR
spectrophotometer with a single-bounce diamond ATR accessory. HRMS
data were recorded on an Orbtitrap LTQ XL ion trap mass spectrometer
in positive ion mode using an electrospray ionisation (ESI) source.
Melting points were determined using an SRS MPA100 OptiMelt melting
point apparatus.
Hz), 117.0, 116.9, 36.3, 35.5, 31.1, 28.5; HRMS (ESI, +ve)
+
C
16
H19ClFN
5
O
3
+ [M + H ] requires m/z 384.1233, found 384.1232.
4.9. Solid phase extraction coupled to mass spectrometry based assays:
Inhibition of the catalytic domain of PHD2 (tPHD2181-426) was
measured by a reported procedure.51 Using the C-terminal oxygenase
dependent domain peptide substrate (CODD) DLDLEMLA-
PYIPMDDDFQL (with a C-terminal amide) and appearance of the hy-
droxylated peptide product in assay buffer (50 mM Tris.Cl pH 7.5, 50
mM NaCl). Titrations of compounds for IC50 determinations (3-fold and
11-point IC50 curves) were performed using an ECHO 550 acoustic
dispenser (Labcyte) and dry dispensed into 384-well polypropylene
assay plates. The final assay concentration of DMSO was kept constant at
0.5% (v/v). tPHD2181-426 was at a concentration of 300 nM (2x final
assay concentration) in the assay buffer and substrate was prepared in
4
.8. Synthesis of VH18
γ-Aminobutyric acid (103 mg, 1.00 mmol) was added to a solution of
1
9 (187 mg, 1.14 mmol) in dry toluene (8 mL), and the resulting mixture
◦
was heated at 140 C for 24 h. After cooling to rt, the mixture was
washed with 1 M HCl (3 × 10 mL) and brine (10 mL) and dried with
4
MgSO . The solvent was evaporated to yield a yellow-white solid (173
◦
1
mg, 70%); m.p. 116.8–117.0 C; H NMR (400 MHz, CDCl
3
) δ 6.28 (t, J
=
1.7 Hz, 2H), 3.54 (t, J = 7.2 Hz, 2H), 3.27 (m, 2H), 2.68 (s, 2H), 2.37
assay buffer (20
μM ferrous iron sulfate, 200
μ
M L-ascorbic acid, 10 M
μ
6

N.L. Richardson et al.
Bioorganic & Medicinal Chemistry 38 (2021) 116115
CODD and 20
across each 384-well assay plate. tPHD2181-426 was equilibrated with
compounds for 15 min and the enzyme reaction initiated by 25
dispense of substrate; incubations were for 15 min. Reaction was
μ
M 2-oxoglutarate); 25
μ
L of tPHD2181-426 was dispensed
(Specs, Bleiswijkseweg, Zoetermeer, Netherlands). Cell treatments were
made up in the following solvents: 100% DMSO stock solution. Relevant
DMSO parallel control dilutions were made to account for any toxic or
protective effects of DMSO alone. Compounds 8 and 16 alone were
μL
v
quenched by 10% ( /v) formic acid (5
to a RapidFire RF365 sampling robot (Agilent) connected to an Agilent
550 quadrupole-time-of-flight (Q- TOF) mass spectrometer, aspirated
μL). Assay plates were transferred
tested for toxicity (1–100
treatment with compound 8 (3
against 50 M H was assessed using phase-contrast microscopy and
counting dead cells.
μ
M each) in MTT assay. The effect of 2 h pre-
μ
M, 10 M)
μ
M) or compound 16 (10–50
μ
6
μ
2 2
O
under vacuum, then loaded onto a C4 solid phase extraction (SPE)
v
cartridge. The C4 SPE cartridge was washed with 0.1% ( /v) formic acid
in water to remove non-volatile buffer salts. The peptide was then eluted
Declaration of Competing Interest
v
from the SPE with 85% acetonitrile, 15% water containing 0.1% ( /v)
formic acid into the mass spectrometer. Peptide charge states were
monitored in the positive ion mode and peak area data integrated using
RapidFire Integrator software (Agilent). The % conversion of the CODD
to the + 16 hydroxylated product was calculated using:
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
%
conversion = 100 × hydroxylated / (hydroxylated + non-hydroxylated
Acknowledgements
peptide)
NR thanks the Australian Government for an Australian Postgraduate
Award and the Mark Wainwright Analytical Centre at UNSW. CJS and
AT thank the Wellcome Trust and Cancer Research UK for financial
support. The authors thank William Figg Jr. for performing the expres-
sion and purification of the PHD2 used in this study. This research was
funded in whole, or in part, by the Wellcome Trust [Grant number
106244/Z/14/Z]. For the purpose of open access, the author has applied
a CC BY public copyright licence to any Author Accepted Manuscript
version arising from this submission.
IC50 data were determined from non-linear regression plots using
GraphPad prism 6.0. The level of +16 (methionine residue oxidation) as
observed in the no enzyme control was ≤5%. All data were normalized
to a no enzyme control.
4
.10. Cell culture
Human-derived undifferentiated neuroblastoma cells (SH-SY5Y,
9
4030304, Sigma Aldrich, St Louis, MO, USA) were chosen as the in vitro
model based on their capacity to undergo oxidative stress among other
Appendix A. Supplementary material
cell death mechanisms5
5,56
and were cultured as per Lizarme et al.
55
◦
with modifications. Cells were maintained at 37 C, 5% CO
2
in a hu-
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.bmc.2021.116115.
midified incubator using media containing 43.5% Dulbecco’s Modified
Eagle’s Medium (DMEM, 11960044), 43.5% Ham’s nutrient mixture
F12 (31765035), 10% Fetal Bovine Serum (FBS, P9423), 1% antibiotic/
antimycotic (15240096), 1% sodium pyruvate (11360070) and 1%
glutaMAX supplement (35050061). Cells were passaged with 0.05%
trypsin/EDTA solution (25399962); harvested cells were recombined
with suspension cells and centrifuged at 700 rpm for 4 min at room
temperature. Cells were resuspended in complete media at the desired
passage dilution. All cell culture reagents were obtained from Life
Technologies (Mulgrave, VIC, Australia) with the exception of FBS and
Minimum Essential Medium Eagle (MEM, M2279) being supplied by
Sigma Aldrich (St Louis, MO, USA). All treatments and vehicle were
made up in MEM from stocks.
References
1
2
Wilber RL. Application of altitude/hypoxic training by elite athletes. Med. Sci. Sport.
Exerc. 2007;39:1610–1624. https://doi.org/10.1249/mss.0b013e3180de49e6.
Kim S, Yang E. Recent advances in developing inhibitors for hypoxia-inducible factor
prolyl hydroxylases and their therapeutic implications. Molecules. 2015;20:
20551–20568. https://doi.org/10.3390/molecules201119717.
3
4
Fraisl P, Aragon
´
es J, Carmeliet P. Inhibition of oxygen sensors as a therapeutic
strategy for ischaemic and inflammatory disease. Nat Rev Drug Discovery. 2009;8(2):
1
39–152. https://doi.org/10.1038/nrd2761.
Navarrete-Opazo A, Alcayaga J, Sepúlveda O, Rojas E, Astudillo C. Repetitive
intermittent hypoxia and locomotor training enhances walking function in
incomplete spinal cord injury subjects: a randomized, triple-blind, placebo-
controlled clinical trial. J Neurotrauma. 2017;34(9):1803–1812. https://doi.org/
1
0.1089/neu.2016.4478.
5
6
Wang H et al. Intermittent hypoxia training for treating mild cognitive impairment: a
pilot study. Am J Alzheimers Dis Other Demen 35, (2020) https://doi.org/10.1177/
4
.11. Cell viability assay
1
533317519896725.
Cells were harvested when 85–95% confluent and seeded at a density
Nguyen HL, Ruhoff AM, Fath T, Jones NM. Hypoxic postconditioning enhances
functional recovery following endothelin-1 induced middle cerebral artery occlusion
in conscious rats. Exp Neurol. 2018;306:177–189. https://doi.org/10.1016/j.
expneurol.2018.05.018.
3
of 3 × 10 cells/mL, in complete media in Costar 96-well plates (3599,
Corning Inc, Corning, NY, USA) for the MTT cell viability assay. Cells
were treated and cell viability was determined using MTT assay 24 h
after insult treatment, as per Lizarme et al,55 with modifications. MTT
7
8
9
Lee P, Chandel NS, Simon MC. Cellular adaptation to hypoxia through hypoxia
inducible factors and beyond. Nat Rev Mol Cell Biol. 2020;21(5):268–283. https://
doi.org/10.1038/s41580-020-0227-y.
The Nobel Prize in Physiology or Medicine 2019 - Advanced information -
NobelPrize.org. https://www.nobelprize.org/prizes/medicine/2019/advanced-
information/.
Wang GL, Jiang BH, Rue EA, Semenza GL. Hypoxia-inducible factor 1 is a basic-helix-
loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci USA.
1995;92(12):5510–5514. https://doi.org/10.1073/pnas.92.12.5510.
(
0.1 mg/mL; M5655, Sigma Aldrich, St Louis, MO, USA) was incubated
◦
with cells for 3 h at 37 C, 5% CO
2
in a humidified incubator. Media was
removed and the formazan product dissolved in dimethyl sulfoxide
DMSO, D8418, Sigma Aldrich, St Louis, MO, USA). Plates were shaken
for 30 min and read at 570 nm using FLUOstar optima microplate reader
BMG Labtech, Ortenberg, Germany). Data was normalised against
00% viability control (MEM as vehicle) and 0% viability control
0.25% triton-X 100, 234729, Sigma Aldrich, St Louis, MO, USA).
(
(
10 Wang GL, Semenza GL. Purification and characterization of hypoxia-inducible factor.
J Biol Chem. 1995;270:1230–1237. https://doi.org/10.1074/jbc.270.3.1230.
1
1
1
1
2
Ke Q, Costa M. Hypoxia-inducible factor-1 (HIF-1). Mol Pharmacol. 2006;70(5):
(
1
469–1480. https://doi.org/10.1124/mol.106.027029.
Semenza GL, Nejfelt MK, Chi SM, Antonarakis SE. Hypoxia-inducible nuclear factors
bind to an enhancer element located 3 to the human erythropoietin gene. Proc Natl
′
4
.12. Cell injury and protective treatments
(7722841, ThermoFisher, Wal-
Acad Sci USA. 1991;88(13):5680–5684. https://doi.org/10.1073/pnas.88.13.5680.
Jiang B-H, Rue E, Wang GL, Roe R, Semenza GL. Dimerization, DNA binding, and
transactivation properties of hypoxia- inducible factor 1. J Biol Chem. 1996;271(30):
1
1
3
4
Cells were treated with 50 µM H
2 2
O
1
7771–17778. https://doi.org/10.1074/jbc.271.30.17771.
tham, MA, USA), to induce oxidative stress. Cell viability was assessed
Carrero P, Okamoto K, Coumailleau P, O’Brien S, Tanaka H, Poellinger L. Redox-
by MTT assay. SH-SY5Y cells were treated with compounds 8 or 16
regulated recruitment of the transcriptional coactivators CREB-binding protein and
7

N.L. Richardson et al.
Bioorganic & Medicinal Chemistry 38 (2021) 116115
SRC-1 to hypoxia-inducible factor 1
org/10.1128/MCB.20.1.402-415.2000.
α
. Mol Cell Biol. 2000;20(1):402–415. https://doi.
35 Yeh T-L, Leissing T, Abboud MI, et al. Molecular and cellular mechanisms of HIF
prolyl hydroxylase inhibitors in clinical trials. Chem Sci. 2017;8(11):7651–7668.
https://doi.org/10.1039/C7SC02103H.
1
1
1
5
6
7
Ruiz de Almodovar C, Lambrechts D, Mazzone M, Carmeliet P. Role and therapeutic
potential of VEGF in the nervous system. Physiol Rev. 2009;89(2):607–648. https://
doi.org/10.1152/physrev.00031.2008.
Lange C, Storkebaum E, de Almod o´ var CR, Dewerchin M, Carmeliet P. Vascular
endothelial growth factor: a neurovascular target in neurological diseases. Nat Rev
Neurol. 2016;12(8):439–454. https://doi.org/10.1038/nrneurol.2016.88.
Zechariah A, ElAli A, Doeppner TR, et al. Vascular endothelial growth factor
promotes pericyte coverage of brain capillaries, improves cerebral blood flow during
subsequent focal cerebral ischemia, and preserves the metabolic penumbra. Stroke.
36 Eltzschig HK, Bratton DL, Colgan SP. Targeting hypoxia signalling for the treatment
of ischaemic and inflammatory diseases. Nat Rev Drug Discovery. 2014;13(11):
852–869. https://doi.org/10.1038/nrd4422.
37 Locatelli F, Fishbane S, Block GA, Macdougall IC. Targeting hypoxia-inducible
factors for the treatment of Anemia in chronic kidney disease patients. Am J Nephrol.
2017;45(3):187–199. https://doi.org/10.1159/000455166.
38 Chan MC, Atasoylu O, Hodson E, et al. Potent and selective triazole-based inhibitors
of the hypoxia-inducible factor prolyl-hydroxylases with activity in the murine brain.
PLoS ONE. 2015;10(7):e0132004. https://doi.org/10.1371/journal.pone.0132004.
39 Pajouhesh H, Lenz GR. Medicinal chemical properties of successful central nervous
system drugs. NeuroRx. 2005;2(4):541–553. https://doi.org/10.1602/
neurorx.2.4.541.
2
013;44(6):1690–1697. https://doi.org/10.1161/STROKEAHA.111.000240.
1
1
2
2
8
9
0
1
Heeschen C, Aicher A, Lehmann R, et al. Erythropoietin is a potent physiologic
stimulus for endothelial progenitor cell mobilization. Blood. 2003;102(4):
1
340–1346. https://doi.org/10.1182/blood-2003-01-0223.
Liu K, Sun T, Wang P, Liu Y-H, Zhang L-W, Xue Y-X. Effects of erythropoietin on
40 Mikitsh JL, Chacko AM. Pathways for small molecule delivery to the central nervous
system across the blood-brain barrier. Perspect Medicin Chem. 2014;6:11–24. https://
doi.org/10.4137/PMc.s13384.
41 Liao C, Sitzmann M, Pugliese A, Nicklaus MC. Software and resources for
computational medicinal chemistry. Future Med Chem. 2011;3(8):1057–1085.
https://doi.org/10.4155/fmc.11.63.
42 Wolber G, Langer T. LigandScout: 3-D pharmacophores derived from protein-bound
ligands and their use as virtual screening filters. J Chem Inf Model. 2005;45:160–169.
https://doi.org/10.1021/ci049885e.
43 Ambure P, Kar S, Roy K. Pharmacophore mapping-based virtual screening followed
by molecular docking studies in search of potential acetylcholinesterase inhibitors as
anti-Alzheimer’s agents. BioSystems. 2014;116:10–20. https://doi.org/10.1016/j.
biosystems.2013.12.002.
44 Chen Z, Li H-L, Zhang Q-J, et al. Pharmacophore-based virtual screening versus
docking-based virtual screening: a benchmark comparison against eight targets. Acta
Pharmacol Sin. 2009;30(12):1694–1708. https://doi.org/10.1038/aps.2009.159.
45 Lagorce D, Sperandio O, Baell JB, Miteva MA, Villoutreix BO. FAF-Drugs3: a web
server for compound property calculation and chemical library design. Web Serv issue
Publ online. 2015;43(W1):W200–W207. https://doi.org/10.1093/nar/gkv353.
46 Whitty A. Growing PAINS in academic drug discovery. Future Med Chem. 2011;3(7):
797–801. https://doi.org/10.4155/fmc.11.44.
blood-brain barrier tight junctions in ischemia-reperfusion rats. J Mol Neurosci. 2013;
4
9(2):369–379. https://doi.org/10.1007/s12031-012-9883-5.
Papandreou I, Cairns RA, Fontana L, Lim AL, Denko NC. HIF-1 mediates adaptation
to hypoxia by actively downregulating mitochondrial oxygen consumption. Cell
Metab. 2006;3(3):187–197. https://doi.org/10.1016/j.cmet.2006.01.012.
Vemula S, Roder KE, Yang T, Bhat GJ, Thekkumkara TJ, Abbruscato TJ. A functional
role for sodium-dependent glucose transport across the blood-brain barrier during
oxygen glucose deprivation. J Pharmacol Exp Ther. 2009;328(2):487–495. https://
doi.org/10.1124/jpet.108.146589.
2
2
Li L, Saliba P, Reischl S, Marti HH, Kunze R. Neuronal deficiency of HIF prolyl 4-
hydroxylase 2 in mice improves ischemic stroke recovery in an HIF dependent
manner. Neurobiol Dis. 2016;91:221–235. https://doi.org/10.1016/j.
nbd.2016.03.018.
Takeda K, Cowan A, Fong G-H. Essential role for prolyl hydroxylase domain protein 2
in oxygen homeostasis of the adult vascular system. Circulation. 2007;116(7):
23
24
25
26
27
28
29
7
74–781. https://doi.org/10.1161/CIRCULATIONAHA.107.701516.
Jaakkola P, et al. Targeting of HIF- to the von Hippel-Lindau ubiquitylation
α
complex by O2-regulated prolyl hydroxylation. Science (80-.). 2001;292:468–472.
https://doi.org/10.1126/science.1059796.
Ivan M, et al. HIF
Implications for O2 sensing. Science (80-.). 2001;292:464–468. https://doi.org/
0.1126/science.1059817.
α
targeted for VHL-mediated destruction by proline hydroxylation:
47 Best M, Degen A, Baalmann M, Schmidt TT, Wombacher R. Two-step protein labeling
by using lipoic acid ligase with norbornene substrates and subsequent inverse-
electron demand diels-alder reaction. ChemBioChem. 2015;16(8):1158–1162.
https://doi.org/10.1002/cbic.v16.810.1002/cbic.201500042.
48 Boraei ATA, El Ashry ESH, Duerkop A. Regioselectivity of the alkylation of S-
substituted 1,2,4-triazoles with dihaloalkanes. Chem Cent J. 2016;10:1–13. https://
doi.org/10.1186/s13065-016-0165-0.
49 Fidler ZN, Shibanova EF, Makerov PV, et al. Acyl derivatives of 3-amino-1,2,4-
triazole. Chem Heterocycl Compd. 1980;16(10):1079–1083. https://doi.org/10.1007/
BF00496616.
50 D z˙ ygiel A, Masiukiewicz E, Rzeszotarska B. Acetylation of 5-amino-1H-[1,2,4]tria-
zole revisited. J Agric Food Chem. 2002;50(6):1383–1388. https://doi.org/10.1021/
jf010734a.
1
Masson N, Willam C, Maxwell PH, Pugh CW, Ratcliffe PJ. <Masson-2001-
Independent function.pdf>. 20, 5197–5206 (2001) https://doi.org/10.1093/emboj/
2
0.18.5197.
Rabinowitz MH. Inhibition of hypoxia-inducible factor prolyl hydroxylase domain
oxygen sensors: tricking the body into mounting orchestrated survival and repair
responses. (2013) doi:10.1021/jm400386j https://doi.org/10.1021/jm400386j.
McNeill LA, Hewitson KS, Gleadle JM, et al. The use of dioxygen by HIF prolyl
hydroxylase (PHD1). Bioorganic Med Chem Lett. 2002;12(12):1547–1550. https://doi.
org/10.1016/S0960-894X(02)00219-6.
Flashman E, Hoffart LM, Hamed RB, Bollinger Jr JM, Krebs C, Schofield CJ. Evidence
for the slow reaction of hypoxia-inducible factor prolyl hydroxylase 2 with oxygen.
FEBS J. 2010;277(19):4089–4099. https://doi.org/10.1111/j.1742-
51 Holt-Martyn JP, Chowdhury R, Tumber A, et al. Structure-activity relationship and
crystallographic studies on 4-hydroxypyrimidine HIF prolyl hydroxylase domain
inhibitors. ChemMedChem. 2020;15(3):270–273. https://doi.org/10.1002/cmdc.
v15.310.1002/cmdc.201900557.
4
658.2010.07804.x.
3
3
0
1
Chowdhury R, Leung IKH, Tian Y-M, et al. Structural basis for oxygen degradation
domain selectivity of the HIF prolyl hydroxylases. Nat Commun. 2016;7(1). https://
doi.org/10.1038/ncomms12673.
52 Chowdhury R, Candela-Lena JI, Chan MC, et al. Selective small molecule probes for
the hypoxia inducible factor (HIF) prolyl hydroxylases. ACS Chem Biol. 2013;8(7):
1488–1496. https://doi.org/10.1021/cb400088q.
Karuppagounder SS, Ratan RR. Hypoxia-inducible factor prolyl hydroxylase
inhibition: Robust new target or another big bust for stroke therapeutics. J Cereb
Blood Flow Metab. 2012;32(7):1347–1361. https://doi.org/10.1038/jcbfm.2012.28.
Naser M, et al. A review on iron chelators in treatment of iron overload syndromes.
Int J Hematol Stem Cell Res. 2016;10:239–247.
53 McDonough MA, Li V, Flashman E, et al. Cellular oxygen sensing: crystal structure of
hypoxia-inducible factor prolyl hydroxylase (PHD2). Proc Natl Acad Sci USA. 2006;
103(26):9814–9819. https://doi.org/10.1073/pnas.0601283103.
3
3
2
3
Guo C, Hao L-J, Yang Z-H, et al. Deferoxamine-mediated up-regulation of HIF-1
α
54 Liu J, Burts AO, Li Y, et al. ‘Brush-first’ method for the parallel synthesis of
photocleavable, nitroxide-labeled poly(ethylene glycol) star polymers. J Am Chem
Soc. 2012;134(39):16337–16344. https://doi.org/10.1021/ja3067176.
55 Lizarme Y, Wangsahardja J, Marcolin GM, Morris JC, Jones NM, Hunter L. Synthesis
and neuroprotective activity of dictyoquinazol A and analogues. Bioorganic Med
Chem. 2016;24(7):1480–1487. https://doi.org/10.1016/j.bmc.2016.02.016.
56 Zille M, et al. Ferroptosis in neurons and cancer cells is similar but differentially
regulated by histone deacetylase inhibitors. eNeuro. 2019;6:263–281. https://doi.
org/10.1523/ENEURO.0263-18.2019.
prevents dopaminergic neuronal death via the activation of MAPK family proteins in
MPTP-treated mice. Exp Neurol. 2016;280:13–23. https://doi.org/10.1016/j.
expneurol.2016.03.016.
3
4
Guo C, et al. Intranasal deferoxamine attenuates synapse loss via up-regulating the
P38/HIF-1
α
pathway on the brain of APP/PS1 transgenic mice. Front Aging Neurosci.
2
015;7. https://doi.org/10.3389/fnagi.2015.00104.
8