Novel selective glucocorticoid receptor agonists (SEGRAs) with a covalent warhead for long-lasting inhibition

Article abstract of DOI:10.1016/j.bmcl.2016.08.091

The synthesis and in vitro properties of six analogues of the selective glucocorticoid receptor (GR) agonist GSK866, bearing a warhead for covalent linkage to the glucocorticoid receptor, is described.

Full text of DOI:10.1016/j.bmcl.2016.08.091

Bioorganic & Medicinal Chemistry Letters xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel selective glucocorticoid receptor agonists (SEGRAs)
with a covalent warhead for long-lasting inhibition
Oksana Ryabtsova a, Jurgen Joossens a, Pieter Van Der Veken a, Wim Vanden Berghe b, Koen Augustyns a,
Hans De Winter a,
⇑
a Laboratory of Medicinal Chemistry, Department of Pharmaceutical Sciences, Faculty of Pharmaceutical, Biomedical and Veterinary Sciences,
Campus Drie Eiken, Building A, Universiteitsplein 1, 2610 Antwerp, Belgium
b Laboratory of Protein Chemistry, Proteomics and Epigenetic Signaling, Department of Biomedical Sciences, Faculty of Pharmaceutical,
Biomedical and Veterinary Sciences, Campus Drie Eiken, Building A, Universiteitsplein 1, 2610 Antwerp, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
The synthesis and in vitro properties of six analogues of the selective glucocorticoid receptor (GR) agonist
GSK866, bearing a warhead for covalent linkage to the glucocorticoid receptor, is described.
Ó 2016 Elsevier Ltd. All rights reserved.
Received 2 August 2016
Revised 26 August 2016
Accepted 29 August 2016
Available online xxxx
Keywords:
Covalent inhibition
Glucocorticoid receptor
Segra
GSK866
Inflammation
Glucocorticosteroids are an established class of medicines that
are still successfully used to treat inflammatory and auto-immune
diseases. However, the main disadvantage of these compounds,
especially at high dosing levels and long-term usage, is the ten-
dency to induce skin atrophy, tachyphylaxis and the occurrence
of a rebound effect upon discontinuation of treatment.1 For this
purpose, the search towards more selective glucocorticoid receptor
agonists (SEGRAs) is still an active research domain, with GSK866
and mapracorat being the prototypes of this class of compounds
(Fig. 1).2 Both glucocorticosteroids and selective glucocorticoid
receptor agonists show a pharmacological effect by binding to
and activating the glucocorticoid receptor (GR). However, in con-
trast to the glucocorticosteroids – which pharmacological effects
arises from the activation of both a transactivation and transre-
pression pathway – SEGRAs selectively activate the GR and exert
their pharmacological effect by activation of only the transrepres-
sion pathway.
could have a long-lasting effect on the GR but without the potential
systemic side-effects that common SEGRA compounds still have.
Having a covalent linkage between the agonist and the GR might
potentially reduce systemic release of the compound with less sys-
temic side-effects as a result. This proof-of-concept study builds
further upon the ideas that emerged from the recent development
of covalent-binding kinase inhibitors,3,4 with the underlying idea
to develop long-lasting GR agonists with applicability in the anti-
inflammatory domain.
The crystal structure of the ligand-binding domain of the GR has
been solved in complex with GSK866 (pdb-code 3E7C),5 and this
structure shows the location of a cysteine residue (Cys643) in close
proximity of the dichorophenyl group of GSK866 (Fig. 2). Com-
plexes of the GR with other SEGRA compounds are available from
the PDB (Fig. 1),6 but we have selected GSK866 as a synthetic start-
ing point for the design of covalent analogues due to synthetic fea-
sibility reasons for the introduction of the covalent warhead, and
also because of the fact that molecular graphics analyses were
favorable in terms of this compound to be used as a starting tem-
plate structure. The close presence of such cysteine residue has
prompted us to investigate the possibility of modifying GSK866
by introducing a reactive warhead into the structure, with the pos-
sibility of forming a covalent linkage between protein structure
and ligand upon binding. Supported by molecular graphics-based
Over the past years, we have been involved in the synthesis of
close analogues of GSK866 with the intention to search for cova-
lent inhibitors of the GR. The rationale behind the covalent binding
approach is to identify GR agonists for topical application and that
⇑
Corresponding author.
E-mail address: hans.dewinter@uantwerpen.be (H. De Winter).
http://dx.doi.org/10.1016/j.bmcl.2016.08.091
0960-894X/Ó 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Ryabtsova, O.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.08.091

2
O. Ryabtsova et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
Figure 1. The structure of GSK866 and two other SEGRA analogues of which the crystal structure of the complex with the GR ligand-binding domain is available from the PDB
(entry codes are given in parentheses). The structure of mapracorat is also shown.
qualitative modeling, six such close analogues of GSK866 were
synthesized and their biochemical properties were evaluated. The
structures are shown in Figure 3.
Compound 4a was synthesized by coupling 11 with the com-
mercially available 2,6-dichloro-3-nitrobenzoic acid (19) to yield
20, followed by reduction to the aniline 21 and nucleophilic addi-
tion–elimination with chloroacetyl chloride to yield the final pro-
duct 4a (Scheme 4). Compound 4b was prepared by coupling
structure 11 with 24 (from a mixture of 24 and 25); compound
24 was synthesized from the commercially available 19, reduction
to the corresponding aniline 22, and conversion into 24 by means
of nucleophilic addition–elimination reaction with chloroacetyl
chloride (generating also 25 resulting from an electrophilic addi-
tion side-reaction) (Scheme 5).
The main building block 11 was synthesized according
Scheme 1, starting from the commercially available 1,3-bis(benzy-
loxy)propan-2-ol (6), followed by a Corey-Kim oxidation to the
corresponding ketone 7.7 Conversion of the ketone into the
a-tri-
fluoromethyl-
a-tosyloxymethyl epoxide 8 was done according
the method as described by Keeling and coworkers.8 Nucleophilic
substitution, initially with N-benzylethanamine and subsequently
with ammonia in methanol, yielded compound 9.9 Coupling of 9
with 13 finally yielded 10, which was then debenzylated into 11
by means of hydrogenolysis.5
Synthesis of 5 was accomplished in three steps, starting from 11
by coupling the secondary amine to the commercially available
N-Z-(R)-(+)-pipecolinic acid to yield 26, removal of the benzyl car-
bamate group by means of Pd/C-catalyzed hydrogenolysis, and
Synthesis of the target compound 2 (Scheme 2) was achieved by
coupling 11 with 2,5-dioxopyrrolidin-1-yl 3-(2,5-dioxo-2,5-dihy-
dro-1H-pyrrol-1-yl)propanoate (16). This latter was prepared from
the reaction between furan-1,5-dione (14) and 3-aminopropanoic
acid (15).10
finally
a
nucleophilic addition–elimination reaction with
chloroacetylchloride to yield the target compound 5 in moderate
yields (Scheme 6).
Compound 3a was prepared from 11 through coupling with
N-Z-(D)-proline,11 deprotected by means of Pd/C-catalyzed
hydrogenolysis, and finally converted into the target structure after
nucleophilic addition–elimination with chloroacetyl chloride.
Treatment of 18 with acryloylchloride yielded compound 3b
(Scheme 3).
Cellular inhibitory activity of the GSK866 analogues was deter-
mined using a human GR receptor radioligand binding assay using
1.5 nM [3H]-dexamethasone as radioligand.12 IM-9 cells were used
as the source of the soluble GR. Assays were performed at Cerep.13
The results are given in Table 1. Only compounds 4a and 4b
displayed significant specific binding to the receptor, suggesting
Please cite this article in press as: Ryabtsova, O.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.08.091

O. Ryabtsova et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
3
pivotal importance of the dichlorophenyl ring in target recognition,
an observation which is in agreement with the reported SAR
around the GSK866 series.9
In order to investigate whether compounds 4a and 4b bind to
the GR in an irreversible manner, a wash-out experiment was set
up using dexamethasone mesylate and cortisol. Dexamethasone
mesylate has been reported as an irreversibly binding covalent
ligand of the receptor and served as a positive control for covalent
modification.14 Compared to the corresponding C-21 mesylates of
cortisol and deacylcortivazol, dexamethasone mesylate has been
reported to be the most useful derivative due to its favorable bal-
ance of high receptor affinity and predominantly irreversible
antiglucocorticoid activity.15,16 Cortisol served as a positive control
of a non-covalent binder. The protocol was based on the method as
described by Simons and Thompson.17 After initial incubation of
the compounds with the receptor and subsequent removal of
unbound compounds, the ability of the receptor to bind [3H]-dex-
amethasone added in excess was assessed. The IM-9 cell-line
cytosolic preparations (240 lg/assay) were pre-incubated for six
hours with 4a, 4b, dexamethasone mesylate and cortisol at a con-
centration chosen as to occupy approximately 80% of the receptors.
For 4a, 4b and dexamethasone mesylate the concentration tested
was 100 nM, and for cortisol this was 60 nM. After incubation, free
compounds were removed by activated charcoal followed by cen-
trifugation. Subsequently, an excess (3 nM) of [3H]-dexamethasone
was added a second incubation was performed. Binding capacity of
[3H]-dexamethasone was quantified at several incubation times
(1 h, 2 h, 3 h, 4 h, 16 h, 24 h). The results are given in Table 2 (‘Ini-
tial experiments’).
The kinetics of the reaction were considered to be rapid due to
the fact that after one hour the maximum binding was already
attained (data not shown). For cortisol, binding to the GR was
totally recovered, confirming that cortisol binds reversibly to the
receptor. However, although the fact that compounds 4a and 4b
were designed as irreversible binders to the GR, we could not
confirm this covalent linkage in the present experiments as the
Figure 2. Close-up in the ligand-binding pocket of the ligand-binding domain of the
glucocorticoid receptor.5 Shown with purple-colored carbon atoms is the dichlor-
ophenyl ring of GSK866, with the yellow dotted line indicating the distance
between this ring and the thiol group of Cys643
.
Figure 3. Design of novel GSK866 (1) analogues with the dichlorophenyl substituent replaced by reactive groups to form a covalent bond with the thiol group of Cys643 of the
ligand-binding domain of the GR (2–5). Proposed docking poses of 4a and 4b in the receptor are shown in Figure S1 (Supplementary Material).
Please cite this article in press as: Ryabtsova, O.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.08.091

4
O. Ryabtsova et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
Scheme 1. Synthesis of 11. Reagents and conditions: (a) NCS, Me2S, Et3N, toluene; (b) TMSCF3, Bu4NF, THF; (c) H2, Pd/,; EtOH; (d) TsCl, pyridine; (e) K2CO3, DCM; (f)
PhCH2NHEt, dioxane, N2, followed by NH3 in MeOH; (g) 13, HATU, DMF; (h) H2, EtOH, Pd/C; (i) LiOH, EtOH/H2O, reflux, then acidification.
Scheme 2. Synthesis of 2. Reagents and conditions: (a) 16; DMF, rt; (b) DMF, rt; (c) DCC, NHS.
Scheme 3. Synthesis of 3a and 3b. Reagents and conditions: (a) N-Z-(D)-proline, HATU, DIPEA, DMF, rt; (b) Pd/C, H2, MeOH, rt; (c) chloroacetylchloride, DIPEA, DCM; (d)
acryloylchloride, DIPEA, DCM.
recovery was complete for 4a and significant for 4b (although not
complete: 74% vs 21%). Strangely, we could not confirm the pub-
lished results of dexamethasone mesylate as being a covalent bin-
der (according to reference 10, the pre-incubation with
dexamethasone mesylate reduced the ability of the receptor by
75% to bind [3H]-dexamethasone), as in our hands the recovery
Please cite this article in press as: Ryabtsova, O.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.08.091

O. Ryabtsova et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
5
Scheme 4. Synthesis of 4a. Reagents and conditions: (a) first SOCl2, benzene, reflux 1 h, then DCM; (b) Zn, NH4Cl/THF; (c) chloroacetylchloride; DCM, NaHCO3.
Scheme 5. Synthesis of 4b. Reagents and conditions: (a) 24, first SOCl2, benzene, DrySyn 100 °C, then 1 equiv NEt3, DCM, À50 °C, overnight; (b) Zn, NH4Cl/THF; (c)
acryloylchloride, DCM, then 0.5 M NaOH, then 1 M HCl; (d) SOCl2, benzene.
Scheme 6. Synthesis of 5. Reagents and conditions: (a) N-Z-(R)-(+)-pipecolinic acid, HATU, DIPEA, DMF, rt, overnight; (b) Pd/C, H2, MeOH, rt, overnight; (c)
chloroacetylchloride, DIPEA, DCM.
of dexamethasone mesylate was almost complete. Potential causes
could have been (1) the presence of b-mercaptoethanol in the incu-
bation buffer solutions used by Cerep, (2) a decomposed form of
dexamethasone mesylate in which the mesylate group would be
cleaved prior to performing the experiments, or (3) differences in
the protein sequence of the GR depending on the species from
which the protein has been extracted. To exclude the first two
possibilities, all experiments were repeated using buffer solutions
without b-mercaptoethanol, and also using a new batch of dexam-
ethasone mesylate (the original batch was supplied by ARC,18 the
second batch was synthesized from dexamethasone according a
method as described by Simons and coworkers19). Results are
shown in the second part of Table 2 (‘Replication experiments’)
and confirm the earlier findings. The two different batches of
Please cite this article in press as: Ryabtsova, O.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.08.091

6
O. Ryabtsova et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
Table 1
dexamethasone mesylate produce the same results, in the sense
that a total recovery of GR binding was observed without blockage
of the receptor by the compound. In addition, b-mercaptoethanol
does not seem to influence the reversibility of dexamethasone
mesylate as well as compounds 4a and 4b.
Specific binding measured as % inhibition of control. All experiments were performed
in duplicate
% inhibition of control
1
l
M ligand concentration
0.1 lM ligand concentration
The dexamethasone mesylate binding experiments performed
by Simons and coworkers16,17 were performed on the glucocorti-
coid receptor extracted from rat hematoma tissue culture cells,
while the present experiments are performed on human IM-9
cell-line cytosolic preparations. In the work done by Simons,
degradation experiments identified Cys656 as the target for
18
3b
27
4a
4b
5
À20%
À23%
À12%
92%
96%
À29%
78%
74%
Table 2
Specific binding measured as % inhibition of control, before and after wash-out at given time points. All experiments were performed in duplicate. ‘Dex-mes’ stands for
dexamethasone mesylate. For the replication experiments, buffer solutions were devoid of b-mercaptoethanol and a new batch of dexamethasone mesylate was used
% inhibition of control
Before wash-out
1 h
2 h
3 h
4 h
16 h
24 h
Average (1–24 h)
Initial experiments
Cortisol
Dex-mes
4a
69%
93%
78%
74%
3%
7%
6%
3%
4%
7%
À1%
À14%
À28%
6%
5%
13%
1%
12%
23%
2%
5%
6%
À4%
À16%
À19%
À11%
4b
21%
17%
24%
31%
%
21%
Replication experiments
Cortisol
Dex-mes
4a
69%
93%
78%
74%
À16%
3%
À15%
À11%
À8%
À23%
À12%
16%
46%
37%
À5%
À8%
13%
8%
À2%
À10%
9%
5%
À14%
À2%
3%
À1%
À6%
0%
À10%
4b
À14%
À33%
À22%
Figure 4. Comparison of the binding of dexamethasone mesylate (‘Dex-mes’, green-colored carbon atoms) and GSK866 (purple-colored carbon atoms) projected into the
crystal structure of the GSK866. Also shown is the binding of CHAPS21 (cyan-colored carbon atoms) as observed in the structure of the dexamethasone complex. The three
cysteine residues which are located in the vicinity of the ligand-binding pocket are highlighted by yellow spheres. Finally, Gly638 is indicated by a purple sphere. In the
sequence of the human and rat GR, this glycine is replaced by a cysteine. The docking pose of dexamethasone mesylate was generated starting from the crystal structure of
dexamethasone in complex with the ligand-binding domain of the GR receptor,20 and manually adding the mesylate sidechain whilst keeping the conformation of this
sidechain in a low energy conformation.
Please cite this article in press as: Ryabtsova, O.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.08.091

O. Ryabtsova et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
7
dexamethasone mesylate covalent binding.16 In the crystal struc-
References and notes
ture of the human GR with GSK866,5 this cysteine has been
mutated to a Gly638 and is located in close proximity to the dichlor-
ophenyl ring of GSP866 and to the C-21 atom of the dexametha-
sone ligand (Fig. 4).20 Interestingly, the crystal structure of the
GR/dexamethasone complex reveals binding of second steroid-like
molecule (CHAPS, see Fig. 4), also located in very close proximity of
this Cys656, indicating that alternative binding modes are observ-
able in the case of the GR receptor. It might therefore be possible
that the discrepancy with the work of Simons and coworkers16,17
could be attributed to undocumented differences in experimental
conditions, thereby exploring alternative binding modes of which
the details are not yet understood.
1. Du Vivier, A.; Stoughton, R. B. Arch. Dermatol. 1975, 111, 581.
2. De Bosscher, K.; Beck, I. M.; Ratman, D.; Berghe, W. V.; Libert, C. Trends
Pharmacol. Sci. 2016, 37, 4.
3. Liu, Q.; Sabnis, Y.; Zhao, Z.; Zhang, T.; Buhrlage, S. J.; Jones, L. H.; Gray, N. S.
Chem. Biol. 2014, 20, 146.
4. Garuti, L.; Roberti, M.; Bottegoni, G. Curr. Med. Chem. 2011, 18, 2981.
5. Madauss, K. P.; Bledsoe, R. K.; Mclay, I.; Stewart, E. L.; Uings, I. J.; Weingarten,
G.; Williams, S. P. Bioorg. Med. Chem. Lett. 2008, 18, 6097.
6. Biggadike, K.; Bledsoe, R. K.; Coe, D. M.; Cooper, T. W.; House, D.; Iannone, M.
A.; Macdonald, S. J.; Madauss, K. P.; McLay, I. M.; Shipley, T. J.; Taylor, S. J.; Tran,
T. B.; Uings, I. J.; Weller, V.; Williams, S. P. Proc. Natl. Acad. Sci. U.S.A. 2009, 106,
18114.
7. Corey, E. J.; Kim, C. U. J. Am. Chem. Soc. 1972, 94, 7586.
8. Keeling, S. P.; Campbell, I. B.; Coe, D. M.; Cooper, T. W. J.; Hardy, G. W.; Jack, T.
I.; Jones, H. T.; Needham, D.; Shipley, T. J.; Skone, P. A.; Sutton, P. W.;
Weingarten, G. A.; Macdonald, S. J. F. Tetrahedron Lett. 2008, 49, 5101.
9. Barnett, H. A.; Coe, D. M.; Cooper, T. W. J.; Jack, T. I.; Jones, H. T.; Macdonald, S. J.
F.; McLay, I. M.; Rayner, N.; Sasse, R. Z.; Shipley, T. J.; Skone, P. A.; Somers, G. I.;
Taylor, S.; Uings, I. J.; Woolven, J. M.; Weingarten, G. G. Bioorg. Med. Chem. Lett.
2009, 19, 158.
The current experiments cannot exclude, nor confirm, that dex-
amethasone mesylate can bind covalently to another cysteine resi-
due not necessary located in the ligand-binding pocket. In
preliminary and unpublished mass spectroscopy experiments, we
identified dexamethasone mesylate to bind covalently to Cys622
,
10. Song, H. Y.; Ngai, M. H.; Song, Z. Y.; MacAry, P. A.; Hobley, J.; Lear, M. J. Org.
Biomol. Chem. 2009, 7, 3400.
11. Falkiewicz, B. Nucl. Acids Symp. Series 1999, 42, 153.
12. Clark, A. F.; Lane, D.; Wilson, K.; Miggans, S. T.; McCartney, M. D. Invest.
Ophtalmol. Vis. Sci. 1996, 37, 805.
13. Catalog reference 0469
14. Eisen, H. J.; Schleenbaker, R. E.; Simons, S. S., Jr. J. Biol. Chem. 1981, 256, 12920.
15. Simons, S. S., Jr; Miler, P. A. J. Steroid Biochem. 1986, 24, 25.
16. Simons, S. S., Jr; Pumphrey, J. G.; Rudikoff, S.; Eisen, H. J. J. Biol. Chem. 1987, 262,
9676.
17. Simons, S. S., Jr.; Thompson, E. B. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 3541.
18. Ref. ARCD 0206, batch 120301.
but further research around the exact mechanism is needed to
understand and confirm these findings.22 Initial docking models
of dexamethasone mesylate in the ligand-binding domain of the
GR do not reveal a close contact between the mesylate group and
Cys622 or Cys643, respectively, but rather a potential contact with
Cys736 (see Fig. S2 of the Supplementary Material). However, more
refined docking studies as well as elongated molecular dynamics
simulations will be essential to validate these hypotheses.
19. Simons, S. S., Jr; Pons, M.; Johnson, D. F. J. Org. Chem. 1980, 45, 3084.
20. Edman, K.; Hosseini, A.; Bjursell, M. K.; Aagaard, A.; Wissler, L.; Gunnarsson, A.;
Kaminski, T.; Kohler, C.; Backstrom, S.; Jensen, T. J.; Cavallin, A.; Karlsson, U.;
Nilsson, E.; Lecina, D.; Takahashi, R.; Grebner, C.; Geschwindner, S.; Lepisto, M.;
Hogner, A. C.; Guallar, V. Structure 2015, 23, 2280.
21. 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate
22. Molecular graphics analysis of the crystal structure of the GR binding domain
in complex with GSK866 confirms the location of Cys622 in proximity of the
GSK866 ligand. However, while Cys643 is positioned in close proximity of the
2,6-dichlorophenyl sidechain of GSK866 (see Fig. 2), Cys622 is located along the
4-fluorophenylpyrazole side of GSK866 with the thiol group pointing outwards
Acknowledgements
This work was supported by a IOF-POC grant from the Univer-
sity of Antwerp and an O&O project from the Flemish IWT.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2016.08.091.
into the solvent.
A close-up on GSK866 and both cysteines is shown in
Figure S3 of the Supplementary Material.
Please cite this article in press as: Ryabtsova, O.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.08.091