Discovery of a Potent and Selective Steroidal Glucocorticoid Receptor Antagonist (ORIC-101)

Article abstract of DOI:10.1021/acs.jmedchem.8b00743

The glucocorticoid receptor (GR) has been linked to therapy resistance across a wide range of cancer types. Preclinical data suggest that antagonists of this nuclear receptor may enhance the activity of anticancer therapy. The first-generation GR antagonist mifepristone is currently undergoing clinical evaluation in various oncology settings. Structure-based modification of mifepristone led to the discovery of ORIC-101 (28), a highly potent steroidal GR antagonist with reduced androgen receptor (AR) agonistic activity amenable for dosing in androgen receptor positive tumors and with improved CYP2C8 and CYP2C9 inhibition profile to minimize drug-drug interaction potential. Unlike mifepristone, 28 could be codosed with chemotherapeutic agents readily metabolized by CYP2C8 such as paclitaxel. Furthermore, 28 demonstrated in vivo antitumor activity by enhancing response to chemotherapy in the GR⁺ OVCAR5 ovarian cancer xenograft model. Clinical evaluation of safety and therapeutic potential of 28 is underway.

Full text of DOI:10.1021/acs.jmedchem.8b00743

Subscriber access provided by MT ROYAL COLLEGE
Article
Discovery of a Potent and Selective Steroidal
Glucocorticoid Receptor Antagonist (ORIC-101)
Yosup Rew, Xiaohui Du, John Eksterowicz, Haiying Zhou, Nadine Jahchan, Liusheng
Zhu, Xuelei Yan, Hiroyuki Kawai, Lawrence R. McGee, Julio C. Medina, Tom Huang,
Chelsea Chen, Tatiana Zavorotinskaya, Dena Sutimantanapi, Joanna Waszczuk,
Erica Jackson, Elizabeth Huang, Qiuping Ye, Valeria R. Fantin, and Daqing Sun
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00743 • Publication Date (Web): 09 Aug 2018
Downloaded from http://pubs.acs.org on August 9, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of a Potent and Selective Steroidal
Glucocorticoid Receptor Antagonist (ORICꢀ101)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Yosup Rew,* Xiaohui Du, John Eksterowicz, Haiying Zhou, Nadine Jahchan, Liusheng Zhu,
Xuelei Yan, Hiroyuki Kawai, Lawrence R. McGee, Julio C. Medina, Tom Huang, Chelsea Chen,
Tatiana Zavorotinskaya, Dena Sutimantanapi, Joanna Waszczuk, Erica Jackson, Elizabeth
Huang, Qiuping Ye, Valeria R. Fantin, Daqing Sun*
ORIC Pharmaceuticals, 240 E Grand Avenue, Fl2, South San Francisco, CA 94080, USA.
KEYWORDS: Glucocorticoid receptor (GR), antagonist, androgen receptor (AR), selectivity.
ABSTRACT: The glucocorticoid receptor (GR) has been linked to therapy resistance across a
wide range of cancer types. Preclinical data suggest that antagonists of this nuclear receptor may
enhance the activity of antiꢀcancer therapy. The firstꢀgeneration GR antagonist mifepristone is
currently undergoing clinical evaluation in various oncology settings. Structureꢀbased
modification of mifepristone led to the discovery of ORICꢀ101 (28), a highly potent steroidal GR
antagonist with reduced androgen receptor (AR) agonistic activity amenable for dosing in
androgen receptor positive tumors, and with improved CYP2C8 and CYP2C9 inhibition profile
to minimize drugꢀdrug interaction potential. Unlike mifepristone, 28 could be coꢀdosed with
chemotherapeutic agents readily metabolized by CYP2C8 such as paclitaxel. Furthermore, 28
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
+
demonstrated in vivo antitumor activity by enhancing response to chemotherapy in the GR
OVCAR5 ovarian cancer xenograft model. Clinical evaluation of safety and therapeutic
potential of 28 is underway.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment
2

Page 3 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
Glucocorticoids (GCs) such as cortisol are steroid hormones secreted by the adrenal gland in a
circadian and stressꢀassociated manner to regulate metabolism, cell growth, apoptosis and
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
differentiation, inflammation, mood and cognitive function. GCs signal through the
glucocorticoid receptor (GR), a member of the superfamily of nuclear receptors expressed across
a wide variety of tissues. Upon ligandꢀbinding, GR translocates to the nucleus where it triggers
transcription of a spectrum of genes that mediate its pleiotropic biological effects. Dysregulation
of GR signaling has been associated with a variety of pathological states including Cushing’s
2
syndrome, psychotic depression, obesity and cancer. Therefore, there has been considerable
interest to target this nuclear receptor for therapeutic purposes.
Several lines of evidence point to GR as a potential mechanism of therapy resistance in cancers
3
of epithelial origin. Elevated levels of GR protein have been correlated with poor prognosis in
-
4
5
ESR1 breast cancer and endometrial cancer patients, as well as to poor outcome in castrationꢀ
6
resistant prostate cancer (CRPC) patients treated with enzalutamide. In addition, cumulative
preclinical data over the past decade have functionally linked GR activation with resistance to a
variety of chemotherapeutic agents across an array of solid tumors that include ovarian cancer,
tripleꢀnegative breast cancer (TNBC), prostate cancer, pancreatic cancer, nonꢀsmall cell lung
7ꢀ13
cancer, urological cancers, etc.
At the molecular level, GR drives resistance to therapy at
least in part, by driving transcriptional activation of mediators of cell survival, antiꢀapoptosis,
9
, 14ꢀ17
cell adhesion and invasion.
Collectively, these studies have established that genetic
ablation or pharmacologic inhibition of GR reverses the proꢀsurvival effects mediated by this
6
ꢀ9, 14, 18
nuclear receptor and enhances chemotherapy and antiꢀandrogen activity in cancer models.
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
These data have also provided rationale to evaluate the efficacy of GR antagonists in
combination with chemotherapeutic agents as well as enzalutamide in the clinic.
1
9
Furthermore, it is well established that GR is expressed in a variety of immune cells.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0ꢀ21
Endogenous and synthetic glucocorticoids are potent regulators of inflammation.
Glucocorticoid binding to GR activates or represses transcription of a variety of proꢀ and antiꢀ
inflammatory genes, and the functional consequences of glucocorticoid action differ by cell type.
GR activation has been shown to promote expansion of regulatory T cells and myeloid derived
22ꢀ24
25
suppressor cells,
stimulation of Th2 differentiation, as well as suppression of activated
+
26
CD8 T cells, and most of these effects were reversed by the GR antagonist mifepristone (also
known as RU486). Importantly, promotion of immunosuppressive cells and T cell exhaustion
27
have been implicated in cancer immune escape as well poor response to immune therapy.
Furthermore, disruption in cortisol secretion has been correlated with poor prognosis and
immunosuppression across a variety of advanced solid tumors including breast, ovarian and lung
2
8ꢀ30
cancer.
In addition, extraꢀadrenal production of cortisol by colorectal cancer, squamous skin
+
cancer and melanoma cells has been associated with decreased CD8 T cell proliferation and
3
1ꢀ32
activity,
suggesting that aberrant activation of GR could contribute to local tumor
immunosuppression. Altogether these observations point to the immunomodulatory potential of
GR, and provide a rational basis for the clinical evaluation of GR antagonists in combination
with emerging immunotherapies.
The marketed drug mifepristone (RUꢀ486) is a steroidal GR antagonist, discovered in the
33
1
980’s as a potent progesterone receptor (PR) antagonist. Due to its potent GR inhibitory
activity, mifepristone has also been evaluated for conditions associated with alterations in GR
ACS Paragon Plus Environment
4

Page 5 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
34
activity. Mifepristone was approved in 2011 for the treatment of Cushing’s syndrome, and is
currently undergoing clinical testing in various solid tumors.
Compared to recently discovered synthetic nonꢀsteroidal GR antagonists and allosteric
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
35ꢀ40
inhibitors,
mifepristone (1) stands out for its potency. However, mifepristone is a partial AR
agonist, and its agonistic activity is sufficient to stimulate the proliferation of CRPC LNCaP/ARꢀ
luc (LNAR) cells both in vitro and in vivo. In addition, mifepristone has also been shown to
+
41
induce AR target gene expression in AR TNBC MDAꢀMBꢀ453 cells. Besides prostate cancer,
previous studies have shown the oncogenic roles of AR and androgen signaling in TNBC and
4
2ꢀ45
ovarian cancer.
In fact, the therapeutic potential of the antiꢀandrogen enzalutamide is
+
45ꢀ47
currently being clinically evaluated in AR TNBC and ovarian cancer patients.
Clearly,
partial AR agonism activity is an unwanted feature for any GR antagonist with potential clinical
application in ARꢀdriven cancers. We envisioned that the undesirable AR agonism of
mifepristone could be eliminated while retaining its GR antagonistic potency by structureꢀbased
modification of mifepristone. We report here the results of our structureꢀactivity relationship
(
SAR) studies, which led to the discovery of ORICꢀ101 (28) (Figure 1), a highly potent GR
antagonist with significantly reduced AR agonism and improved Cytochrome P450 (CYP)
inhibition profiles. These features of 28 make this novel compound an attractive combination
+
partner to antiꢀandrogens and paclitaxel in AR cancers including CRPC, as well as a subset of
1
4, 44, 46, 48
TNBC and ovarian cancers.
In addition, the favorable CYP inhibition profile of 28
makes this compound combinable with therapeutic agents that are metabolized by CYP2C8 and
CYP2C9.
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
b
Figure 1. Mifepristone (1) and ORICꢀ101 (28). IC in GR luciferase antagonist assay. EC
5
0
50
[
Emax]* in AR luciferase agonist assay. * % mifepristone.
ACS Paragon Plus Environment
6

Page 7 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
With the goal of developing a potent, steroidal GR antagonist with reduced AR agonism as
compared to 1, an analysis of the binding mode of mifepristone to AR and GR was initiated. The
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
9ꢀ50
protein coꢀcrystal structure of 1 bound to GR, but not to AR, has been solved (Figure 2).
Since the two nuclear receptors are closely related, a homology model of AR was constructed
based on the GR structure (pdb: 3H52). Figure 2A shows a view of the mifepristone (1)/GR
binding site illustrating the key hydrogen bond interactions of the steroid carbonyl with Arg611
and Gln570. Outside of the interaction with these two polar residues, the majority of binding site
residues are hydrophobic in nature and form favorable van der Waals interactions with the
predominately hydrophobic ligand. Figure 2B shows the complementary fit between the ligand
and the binding site, with 1 almost completely buried in the steroid binding pocket. The helices
that contain residues that interact with mifepristone and the NCoR peptide are also labeled.
The agonistic and antagonistic functions of GR and AR ultimately depend on the protein
5
1ꢀ53
conformation,
in particular, the positions of helices 11 and 12 and the ability or inability of
54
the receptor to bind coꢀactivators or coꢀrepressors. To that end, structural hypotheses were
developed along three distinct vectors with the goal of altering the GR and AR structures and
thus affecting the agonist and antagonist profiles of the ligands. Using the mifepristone scaffold
as a foundation, modification of the C11 4ꢀdimethylaminophenyl group provided a vector to alter
the position of helix 12 and also to directly challenge coꢀpeptide binding. Modification of the
C17 hydroxyl and introduction of C16 substituents afforded the opportunity to disrupt the
55
positions of both helices 11 and 12 directly, and modification of the C17 1ꢀpropynyl group
provided a cascading structural effect where disruption of helix 7 in turn altered the position of
helices 11 and 12.
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(A)
(B)
NCoR
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
H12
H3
H11
H5
H7
Figure 2. (A) 1 (gray) bound to GR (pdb: 3H52). GR active site residues shown in blue. (B)
GR protein surface with helices labeled and NCoR peptide shown in purple.
Initial investigation of 1 began with replacement of the 4ꢀdimethylaminophenyl group at the
5
6ꢀ58
C11 position in the steroid core with various substituents (Table 1, compounds 2ꢀ6).
To
measure antagonist and agonist activities of these compounds on GR and AR, luciferase reporter
59ꢀ60
assays were employed.
Substitution of the 4ꢀdimethylaminophenyl group with a small alkyl
group, such as an ethyl, switched the GR activity, converting antagonism to agonism (2 vs. 1).
Increase of the substituent’s size to cyclohexyl or unsubstituted phenyl group restored GR
antagonism and attenuated GR agonism (3 and 4 vs. 2). These results imply that a phenyl or a
similarꢀsized cyclohexyl moiety at the C11 position is an important feature for GR antagonism.
Moving the dimethylamino group from paraꢀposition to metaꢀposition reduced AR agonism
dramatically (5.8% of 1), but decreased GR antagonism by 20ꢀfold as well (5 vs. 1). The 4ꢀ
ACS Paragon Plus Environment
8

Page 9 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
isopropylphenyl derivative elicited improvement in GR antagonism (IC = 1.93 nM for 6).
5
0
However, it also induced more AR agonism (200% of 1).
Attention was then focused on the modification of the 1ꢀpropynyl group at the C17 position
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
6, 61
(Table 1, compounds 7ꢀ11).
Changing the 1ꢀpropynyl group to a saturated nꢀalkyl group
amplified AR agonism while diminishing GR antagonism (7 vs 1), suggesting that a CꢀC triple
bond at the C17 position is crucial for minimized AR agonism and enhanced GR antagonism.
The introduction of an electron withdrawing group by replacing the 1ꢀpropynyl group with a
3
,3,3ꢀtrifluoroꢀ1ꢀpropynyl group maintained the same level of AR agonism (8 vs 1).
Interestingly, gradually increasing alkyl substituent size in the 1ꢀpropynyl group led to the 3,3ꢀ
dimethylꢀ1ꢀbutynyl derivative 11, which possessed substantially reduced partial agonism in the
AR luciferase assay (13% of 1). However, 11 also exhibited a 4.5ꢀfold decrease in GR
antagonism potency compared to 1 (compounds 9ꢀ11 vs. 1).
Table 1. Exploration of substituents at C11 and C17 positions in mifepristone
GR luc. antagonism
IC50 (nM)
GR luc. agonism
AR luc. agonism
compound
R1
R2
EC50 (nM) [Emax]a
EC50 (nM) [Emax]b
1
3
.3±0.6
>2,500 [3.1±1.4]
11.9 ± 5.6 [100]
(mifepristone)
2
3
>5,000
57±35 [36±18]
>2,500 [47]
>2,500 [26]
18.1±6.3
>2,500 [9.3±0.6]
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
5
14.2±1.5
64±20
17.4±1.5 [7.0±3.4]
>2,500 [27]
>2,500 [5.8]
>2,500 [1.5±0.5]
>2,500 [3.1±2.0]
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
1.93±0.20
13±3 [198±14]
1.5 [210]
7
32±13
>2,500 [1.4±0.6]
>2,500 [2.4±1.2]
>2,500 [2.1±1.1]
8
9
11.2±2.8
8.5±1.7
5
4 [117]
0 [101]
3
10
11
26±12
>2,500 [1.8±0.9]
>2,500 [1.4±0.9]
>
2,500 [50]
14.6±4.5
>
2,500 [13.4±5.2]
a
Potency and Emax data with SD are reported as the average of at least two determinations. % dexamethasone.
b
%
mifepristone.
As part of the lead optimization process, a coꢀcrystal structure of 11 bound to human GR was
determined at a resolution of 3.05 Å (Figure 3). The ketone moiety of 11 makes key hydrogen
bond interactions with Arg611 and Gln570, and overall the ligand sits in a similar position as
that of mifepristone in the binding site. There is however, a significant difference in the protein
structure in the region interacting with the C17 substituent. As can be observed in the
mifepristone coꢀcrystal structure, the 1ꢀpropynyl group sits in a hydrophobic region between
helices 3, 6 and 7 (Figure 2). The orientation of these helices in turn has an effect on the position
ACS Paragon Plus Environment
10

Page 11 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of helices 11 and 12. The resulting overall conformation of GR is favorable for binding of the
coꢀrepressor peptide, NCoR. This is consistent with the antagonistic nature of mifepristone
towards GR. The replacement of the alkynyl methyl with a tꢀbutyl group in 11 results in
significant changes in the shape and size of the binding site and the overall conformation of GR.
Figure 3 illustrates a significant increase in the size and depth of this part of the binding pocket
that is necessary to accommodate 11. Consequently, significant changes occur in the
orientations of helices 6 and 7, resulting in a shift in the positions of helices 11 and 12 as well.
In fact, the helix 12 position prevents the binding of coꢀrepressor or coꢀactivator peptides
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
2
(
Supplemental Figure S4). While mifepristone does not show significant agonism towards GR
(Emax relative to dexamethasone of 3.1%), 11 further reduces the relative Emax by more than
0%.
5
(A)
(B)
H12
H3
H11
H5
H7
Figure 3. (A) 11 (dark gray) bound to GR (pdb: 6DXK). GR active site residues shown in olive.
B) GR protein surface with helices labeled.
(
ACS Paragon Plus Environment
1
1

Journal of Medicinal Chemistry
Page 12 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
As mentioned, the addition of tꢀbutyl to the alkyne has a significant effect on AR selectivity. A
more than 7ꢀfold reduction in AR agonism (Emax) is observed for 11 versus mifepristone. To
date, efforts to obtain a coꢀcrystal structure of AR with 11 have not been successful. However,
given the close structural relationship between AR and GR, a similar structural hypothesis can be
developed through molecular modeling to explain the reduced AR agonism of 11. The
orientation of 11 in the binding site of AR is expected to be similar to the one observed for GR
and thus a similar disruption of the positions of helices 6, 7, 11 and 12 is expected. These large
conformational changes in turn disrupt the coꢀpeptide binding site further reducing the AR
agonism for 11 versus mifepristone.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A single crystal small molecule xꢀray structure of 11 was also obtained (Figure 4). This
structure allowed for confirmation of the absolute stereochemistry of all of the chiral centers in
the molecule and was also used to evaluate the coꢀcrystal structure binding mode. The steroid
scaffold is rigid and overall 11 has few rotatable bonds. Nevertheless, structural distortion of the
ligand is still possible and could result in a high energy binding conformation. The ligand
conformation of 11 from the GR protein coꢀcrystal structure is shown in B, while C shows the
overlay of the two and illustrates that the conformations are almost identical. There is no
significant change in ligand conformation upon binding to GR and thus no additional strain is
introduced into the molecule.
ACS Paragon Plus Environment
12

Page 13 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
A)
(B)
(C)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Comparison of small molecule single crystal xꢀray structure of 11 (A, orange) with the
GR coꢀcrystal xꢀray structure 11 (B, dark gray, pdb: 6DXK). The overlay of the two structures is
shown in C.
Having identified the 3,3ꢀdimethylꢀ1ꢀbutynyl group as the optimized 1ꢀalkynyl group at the
C17 position with respect to AR selectivity, attention was turned to substituent modification of
the aryl ring at the C11 position (Table 2). Replacing the 4ꢀdimethylamino group in 11 by small
alkyl groups was tolerated in terms of GR antagonism potency, but those derivatives displayed
greater AR agonism as the size of the alkyl group increased (12, 13, and 14). Replacement of the
4
ꢀdimethlyamino group with a methoxy or methyloxetanyl group was also well tolerated for GR
antagonism (15 and 25). Shifting the dimethylamino substituent from the paraꢀposition to metaꢀ
position led to a drop in GR antagonism potency (16) as observed in the previous SAR with
compound 5. Additional substitutions at the metaꢀand orthoꢀpositions were then evaluated while
preserving the dimethylamino group at the para position. The 2,5ꢀdifluorophenyl analog 18
showed GR antagonism potency similar to that of 11, whereas potency of all other analogs was
diminished (17, 19, and 20). The optimization of the N,Nꢀdialkyl group within 11 was attempted
to improve GR antagonism potency and to address the inherent metabolic instability of the N,Nꢀ
dimethylamino group. Compound 21, which exchanged one Nꢀmethyl group with an isopropyl
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
group, was lower in GR antagonism activity compared to 11 (IC = 43 nM for 21 vs. IC = 14
5
0
50
nM for 11) but improved AR agonism selectivity (Emax = 8.2% of 1 for 21 vs. Emax = 13% of 1
for 11). Introduction of various cyclic tertiary amines was also tolerated with regard to GR
antagonist potency (22, 23, and 24). Overall, no significant improvement in GR antagonist
potency was observed from the aryl group modification at the C11 position.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2. Exploration of aryl substituents at C11 position in compound 11
GR luc. antagonism
IC50 (nM)
GR luc. agonism
AR luc. agonism
compound
Ar
EC50 (nM) [Emax]a
EC50 (nM) [Emax]b
1
1
14.6±4.5
>2,500 [1.4±0.9]
>
2,500 [13.4±5.2]
1
1
2
3
16.1±7.8
17.5±9.3
>2,500 [1.5±0.8]
>2,500 [1.5±0.7]
>2,500 [16±10]
>2,500 [20±13]
1
4
16.0±7.0
>2,500 [1.5±0.8]
>2,500 [25]
1
1
5
6
16.6±3.5
205±112
>2,500 [1.1±0.6]
>2,500 [1.3±0.6]
>2,500 [9.8±4.3]
>2,500 [7.5]
1
7
43±19
>2,500 [1.3±0.6]
>2,500 [8.8±7.7]
ACS Paragon Plus Environment
1
4

Page 15 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
8
12.4±5.8
48±17
>2,500 [1.3±0.7]
>2,500 [1.0±0.5]
>2,500 [11±8.4]
>2,500 [5.7]
19
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0
28±16
43±21
>2,500 [1.2±0.7]
>2,500 [1.7±0.8]
>2,500 [17.0±1.7]
>2,500 [8.2±4.4]
21
2
2
3
13.3±2.7
19.5±5.1
>2,500 [1.5±0.7]
>2,500 [1.2±0.8]
>2,500 [18.7±3.2]
>2,500 [13.3±4.1]
2
2
4
5
16.7±1.0
9.6±3.4
>2,500 [1.8±1.1]
>2,500 [2.2±1.2]
>2,500 [23±10]
2
>2,500 [11.0±4.3]
a
Potency and Emax data with SD are reported as the average of at least two determinations. % dexamethasone.
b
% mifepristone.
The Xꢀray coꢀcrystal structure of 11 bound to GR protein (Figure 3) shows that the A ring
carbonyl oxygen forms direct hydrogen bonds to the guanidinium of Arg611 and to the γꢀamide
of Gln570. The Aꢀring carbonyl oxygen in compound 11 is α,βꢀγ,δꢀdiene carbonyl oxygen with
extended conjugation, which is known to be less electronegative than α,βꢀunsaturated carbonyl
oxygen. Therefore, we hypothesized that reducing the C9 CꢀC double bond in the B ring could
enhance hydrogen bond acceptor ability of the A ring carbonyl oxygen. Consequently, an effort
was made to explore modification of the steroid core in 11 (Table 3).
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 3. Modification of the steroid core of 11
GR luc. antagonism GR luc. agonism
AR luc. agonism
compound
Structure
IC50 (nM)
EC50 (nM) [Emax]a
EC50 (nM) [Emax]b
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
3
.3±0.6
>2,500 [3.1±1.4]
11.9±5.6 [100]
(mifepristone)
1
1
14.6±4.5
5.6±2.4
14.6±7.0
>2,500 [1.4±0.9]
>2,500 [2.2±1.0]
>2,500 [1.7±0.9]
>2,500 [13.4±5.2]
224±97 [22±11]
2
6
2
7
>
2,500 [10.6±4.3]
28
5
.6±2.6
>2,500 [2.3±1.1]
>2,500 [11.0±5.8]
(ORICꢀ101)
a
Potency and Emax data with SD are reported as the average of at least two determinations. % dexamethasone.
b
%
mifepristone.
Of significance was the observation that 26, the corresponding C9 CꢀC double bond reduced
analog of 11, had a marked 3ꢀfold improvement in GR antagonist potency (IC = 5.6 nM for 26
5
0
vs. IC = 14 nM for 11). However, 26 showed a slight increase in AR agonist activity (Emax =
5
0
2
2% of 1) compared to 11 (Emax = 13% of 1). This increase is not negligible. For example, the
comparable partial agonistic activity of bicalutamide (Emax = 23% of 1) compromises its
ACS Paragon Plus Environment
16

Page 17 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
therapeutic efficacy as an antiꢀandrogen relative to full antagonists such as enzalutamide (Figure
5
) and apalutamide.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Relative AR agonism of 11 and 28 compared to 1, enzalutamide, and bicalutamide.
The 3,3ꢀdimethylꢀ1ꢀbutynyl group in 26 was further modified by replacing one methyl of the tꢀ
butyl group with a methoxy moiety. This was attempted to reduce AR agonism and improve
solubility, but it also resulted in lower GR antagonism (27). On the other hand, similar to
previous observations, the replacement of one Nꢀmethyl group in 26 with an isopropyl group
provided 28 with significantly reduced AR agonism (Emax = 11% of 1), which is comparable to
AR agonism of enzalutamide (Emax = 8.7% of 1) (Figure 5). Overall, compound 28 is a weak
full AR antagonist with IC of 484 ± 110 nM in the AR luciferase agonism assay. Furthermore,
50
compound 28 is equipotent to 26 and only 2ꢀfold less potent than mifepristone in the GR
luciferase antagonism assay (IC = 5.6 nM).
5
0
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Induction of endogenous AR target genes by synthetic androgen (R1881) and 1 in
LNAR cells. Endogenous AR target genes, PSA and TMPRSS2, are induced by R1881 and 1 in
LNAR cells while those genes are not induced by 28. Enz (enzalutamide) inhibits R1881ꢀ
induced expression but does not fully inhibit the expression induced by 1.
CLDN8
CYPGF8
SEC14L2
28
1
2
8
1
28
1
+
Figure 7. Stimulation of AR activity in AR TNBC cells by 1. Compound 1, but not 28,
+
stimulates AR target gene expression in the AR MDAꢀMBꢀ453 TNBC cell line.
ACS Paragon Plus Environment
1
8

Page 19 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
We also compared the AR agonism of compounds 1 and 28 in the LNAR cells, which exhibit
high AR, but low GR, and low PR expression, as well as in the MDAꢀMBꢀ453 cells with high
AR, high GR, but no PR expression by RTꢀqPCR (Figures 6 and 7). The results showed that 1,
but not 28, is able to stimulate endogenous AR target gene expression in both LNAR and MDAꢀ
MBꢀ453 cell lines at about 100 nM. These expression levels are similar to those achievable by
0.1 nM R1881, the potent synthetic androgen in the same cell lines. On the other hand, 1 ꢀM
enzalutamide completely reverted gene expression induced by R1881, but not 1, in the LNAR
cell line (Figure 6). Notably, endogenous AR target gene induction by compound 28 was
minimal in both cell lines (Figures 6 and 7). These data suggest that compound 28 exhibits no
AR agonism and might have an advantage over 1 in treating ARꢀdriven cancers.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The single crystal small molecule xꢀray structure of 28 was solved. Figure 8 shows the
structure of 28 as compared to the protein coꢀcrystal structure of 11. Since the overlay shows
that the overall shape of both compounds is very similar, 28 would be expected to have a very
similar binding mode to that of 11. This is consistent with the binding mode of 28, as predicted
via docking calculation with GR.
(A)
(Β)
(C)
Figure 8. Comparison of the small molecule single crystal xꢀray structure of 28 (A, blue) with
the GR coꢀcrystal xꢀray structure 11 (B, dark gray, pdb: 6DXK). The overlay of the two
structures is shown in C.
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
To identify a molecule suitable for preclinical development, several selected compounds were
evaluated for their rat pharmacokinetic (PK) profile (Table 4). The results showed that 13 and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3 had low clearance, but poor oral bioavailability (CL = 0.41 and 0.28 L/kg/h; %F = 0 and
.7%, respectively). Compounds 11 and 28 had moderate clearance (CL =1.4 and 1.9 L/kg/h,
0
respectively) and acceptable oral bioavailability (%F = 13 and 24%, respectively). The rest of
the compounds 21, 25, and 27 showed high CL (CL = 3.0˗5.0 L/kg/h) although 27 exhibited
good oral bioavailability (%F = 54%).
a
Table 4. Rat pharmacokinetic data for selected compounds
iv (0.5 mg/kg)b
po (5.0 mg/kg)c
F (%) Cmax (ꢀg/L)
compound
CL
AUC
V
ss (L/kg)
t1/2 (iv, h)
(
L/kg/h)
(ꢀg—h/L)
1
3.7
4.0
0.41
4.0
0.28
3.0
5.0
4.1
4.9
1.7
4.7
1.5
6.7
7.2
4.6
1.6
1.9
3.4
1.5
3.6
4.4
2.4
4.4
6.4
37
0
2
4
85
471
0
1
1
1
7
2
3
0
21
23
0.7
7.2
54
24
5
9
290
167
2
3
21
25
59
123
544
635
2
7
295
149
2
8
1.9
ACS Paragon Plus Environment
2
0

Page 21 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
n = 3 mice for both iv and po, iv time points: one preꢀdose and 10 postꢀdose, po time points: one preꢀdose and
b
c
8 postꢀdose; formulated in 10% DMSO, 70% PEG 400, and 20% water; formulated in 5% DMSO, 95% 0.2%
Tween 80 in 0.25% CMC.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Given the GR antagonist potency, low AR agonism, and favorable rat PK profile, compound
2
8 was selected for further evaluation. The preclinical pharmacokinetics of 28 were evaluated in
mouse and minipig following singleꢀdose intravenous (iv) or oral administration. The molecule
was found to have low clearance and reasonable oral bioavailability in mouse and minipig (Table
6
3
5
).
Table 5. Pharmacokinetic data for 28 in mouse and minipig
iv (0.5 mg/kg)a
CL (L/kg/h)
1.63
po (5 mg/kg for mouse and 10 mg/kg for minipig)b
Species
V
ss (L/kg)
t
1/2 (h)
F (%)
21
Cmax (ꢀg/L)
AUC (ꢀg—h/L)
Mouse
5.01
7.8
4.83
12.2
2
74
644
Minipig
1.1
20
1
53
1721
a
Formulations: 28 was in 5% DMA, 10% EtOH, 40% PEG400, and 45% D5W for mouse iv study; and in 5%
b
DMA, 10% EtOH, 40% PEG 400, 43.8% D5W, pH 5 for minipig iv study; 28 was in 0.25% CMC, 0.2% Tween
8
0, and 5% DMSO for mouse oral study, and 28 HCl salt was in 2% HPMC and 0.2% Tween 80 for minipigꢀoral
study.
Species differences in the unbound fraction of compound 28 were minimal (< 2ꢀ3 fold). The
unbound fraction (0.14) of compound 28 is 10ꢀfold higher than that of compound 1 in human
plasma. As compound 1 is reported to have high potential for inhibition of CYP2C8, we next
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
examined the liability of 11 and 28 in the CYP inhibition assay. It was found that both 11 and 28
have low potential of CYP2C8 and CYP2C9 inhibition compared to 1 although the IC values
5
0
of 11 and 28 for CYP3A4 are lower than that of 1 (Table 6). In general, the replacement of the
C17 alkynyl methyl in 1 with a tꢀbutyl group improves the CYP inhibition profile, especially
against CYP2C8 and CYP2C9.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 6. CYP inhibition of compounds 1, 11, and 28
IC50 (ꢀM)
P450
Substrate
Isoform
1
11
28
3A4
2C8
2C9
Midazolam
9.5
2.9
>10
>10
1.6
Amodiaquine 1.5
Diclofenac 4.9
>10
>10
Furthermore, the in vivo combination PK studies of 1 or 28 with paclitaxel, a CYP2C8
substrate, in female mice demonstrated that 28 showed less potential for CYP2C8 driven drugꢀ
drug interaction compared to 1 (Figure 9). The results showed that adding 1 to paclitaxel
resulted in higher paclitaxel plasma exposure. Unlike 1, 28 did not result in the increase in
paclitaxel concentration in plasma.
ACS Paragon Plus Environment
22

Page 23 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 9. In vivo combination pharmacokinetics of 1 and 28 with paclitaxel in mice.
To evaluate the selectivity of compound 28 for GR over PR, both 28 and 1 were tested in the
PR luciferase reporter assay. The results showed that the inhibitory activity of 28 was
substantially lower than that of 1 (IC = 21 ± 10 nM for 28; IC = 0.4 ± 0.2 nM for 1). The
5
0
50
PR/GR ratio of 28 is 3.8 while that of 1 is 0.12. Compound 28 does not show any agonism in the
PR luciferase reporter assay. Unlike the estrogen receptor (ER) ligand 17βꢀestradiol and the ER
antagonist tamoxifen, 28 does not physically bind to ER up to 1 ꢀM (Supplemental Figure S1).
Similarly, 28 neither induces nor inhibits aldosteroneꢀinduced interaction between MR and its
coactivator SRC at the concentration of 5 ꢀM (Supplemental Figures S2 and S3), suggesting 28
is not a mineralocorticoid receptor (MR) agonist or antagonist.
To compare the activity of 1 and 28 on GR across multiple species, expression of endogenous
GR target genes was evaluated using human, rat, dog, monkey, and minipig peripheral blood
mononuclear cells (PBMCs) treated ex vivo with dexamethasone in the presence or absence of 1
or 28. Figure 10 shows an example of quantitative polymerase chain reaction (RTꢀqPCR) results
measuring the doseꢀdependent expression levels of two established GR target genes, GILZ and
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
64
FKBP5 in human PBMCs treated with dexamethasone and 28. Treatment with 28 significantly
inhibited ligandꢀinduced target gene expression with potencies comparable to 1 in all species
tested (Table 7).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1.2
GILZ
FKBP5
0.9
0.6
0.3
0.0
0.1
1
10
100
1000
Concentration (nM)
Figure 10. GR target gene expression in PBMCs. Representative RTꢀqPCR measuring FKBP5
and GILZ in PBMCs from a healthy human donor treated with dexamethasone at 30 nM and 28
at dose ranging concentrations for 6 hours.
Table 7. Crossꢀspecies evaluation of the effect of 1 and 28 on endogenous target gene expression
in PBMCs
species
human
GILZ
17.8
Rat
dog
monkey
minipig
Target Gene
IC50 of 1 (nM)
FKBP5 GILZ FKBP5
GILZ FKBP5
GILZ FKBP5
GILZ FKBP5
14.8
19.5
NDa
59.4
NDa
45.6
32.6
34.3
20.2
24.5
13.9
14.9
21.9
11.1
25.7
181
19.7
39.3
IC50 of 28 (nM) 24.9
a
not determined.
ACS Paragon Plus Environment
24

Page 25 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Next, we evaluated whether 28 could inhibit endogenous GR target gene expression and
+
compared its relative potency with 1 in cancer cells. To this end, the GR human ovarian cancer
cell line OVCARꢀ5, which has no detectable AR and low PR expression levels, was stimulated
with dexamethasone in the presence or absence of 28 or 1 and mRNA levels of GR target genes
were measured by RTꢀqPCR. Compound 28 reduced expression of FKBP5 and GILZ in a doseꢀ
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
dependent manner (IC = 17.2 and 21.2 nM, respectively) with similar potency as 1 (IC = 10.9
50
50
and 13.0 nM, respectively) (Figure 11).
Figure 11. Inhibitory effects of 28 and 1 on GRꢀmediated target gene expression in ovarian
cancer OVCARꢀ5 cells.
The ability of 28 to inhibit GR transcriptional activity was further evaluated in vivo. The
6
5
major murine glucocorticoid, corticosterone, is a weak agonist of human GR. Therefore, when
human cancer cells are grown as tumor xenografts in mice, it is necessary to provide an
exogenous ligand such as cortisol or the synthetic ligand dexamethasone to effectively activate
human GR. Expression of the GR target gene FKBP5 was analyzed in OVCAR5 tumors isolated
from animals 6 hours after administration of vehicle or dexamethasone (0.5 mg/kg results in a
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
plasma glucocorticoid concentration equivalent to a human cortisol dose of 0.100 to 0.467 ꢀM
corrected for potency) in the presence or absence of a single oral dose of 28 at 3 different
concentrations. In the presence of 28 at 25, 50 or 75 mg/kg, dexamethasoneꢀinduced FKBP5
expression was reduced in a dose dependent manner (Figure 12A). Expression of FKBP5 in the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
5 mg/kg group was not significantly different than expression in the vehicle group. The
unbound plasma concentration of 28 in circulating mouse plasma ranged from 0.02 to 0.12 ꢀM
across the 3 groups (Figure 12B). The plasma exposures of dexamethasone were similar in all
animals that received the dose of glucocorticoid (data not shown). These data suggest that 28 is
effective at inhibiting dexamethasoneꢀinduced GR target gene expression in OVCARꢀ5
xenograft tumors.
(A)
(B)
Figure 12. Effect of 28 on dexamethasoneꢀinduced target gene expression in OVCAR5
xenograft tumors. (A) mRNA levels of the GR target gene, FKBP5, in OVCAR5 tumors relative
to vehicle 6 hours after treatment (n = 4 mice per group). Dexamethasone (Dex) was
administered by intraperitoneal injection at 0.5 mg/kg and a single dose of 28 was administrated
orally by gavage at 25, 50, and 75 mg/kg. (B) Bar graph showing the mean unbound plasma
ACS Paragon Plus Environment
2
6

Page 27 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
concentrations of 28 in the 3 treated groups, 6 hours after treatment with Dex at 0.5 mg/kg (n = 4
mice per group). Significance of effects on FKBP5 expression was determined by oneꢀWay
ANOVA using Dunnet’s test to correct for multiple comparisons. The pꢀvalues compared to the
vehicle group are 0.0004 (***) for the Dexamethasone group, 0.0036 (**) for the Dex + 28 at 25
mg/kg group, and not significant (ns) for the Dex + 28 at 50 mg/kg group (p = 0.1111) and the
Dex + 28 at 75 mg/kg group (p = 0.7506).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
An in vivo study was carried out to assess the efficacy of 28 in combination with gemcitabine
and carboplatin in the OVCAR5 ovarian cancer xenograft model (Figure 13). As previously
mentioned, corticosterone is the prevailing circulating glucocorticoid in rodents and is a weak
activator of human GR. Therefore, to achieve proper activation of GR in the human xenograft
transplant model, cortisol was administered chronically in the drinking water at 100 mg/L before
tumor implantation and for the duration of the study in the groups indicated. The resulting
plasma cortisol levels are consistent with reported mean plasma cortisol levels in cancer patients
6
6ꢀ67
throughout the day, which range from 0.100 to 0.496 ꢀM.
Daily treatment with 28 did not
affect the growth of the tumors grown under chronic cortisol. Gemcitabine and carboplatin
induced tumor growth inhibition in groups with or without 28 compared to vehicle alone in the
presence of cortisol. However, addition of 28 significantly enhanced the antitumor effect of
chemotherapy. Tumor volumes on Day 21 in mice receiving gemcitabine + carboplatin +
cortisol were significantly larger than the tumor volumes in mice receiving gemcitabine +
carboplatin + cortisol + 28. These data indicated that 28 was effective in reversing the GR
mediated resistance to chemotherapy driven by cortisol in vivo.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 66
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
*
Figure 13. Effects of 28 on gemcitabine and carboplatin response in OVCAR5 xenograft tumor
in cortisolꢀtreated mice. Tumor growth curves for mice with established OVCAR5 xenografts
grown in the presence of cortisol at 100 mg/L (n = 10 mice per group). Animals were dosed with
Gemcitabine 100 mg/kg and Carboplatin 60 mg/kg every 3 days for 4 doses. 28 was dosed at 75
mg/kg twice a day by oral gavage for the duration of the study. Data are displayed as mean ±
SEM. The star (*) in both panels indicates statistical significance on day 21 between the 2 groups
indicated, with p < 0.05.
CHEMISTRY
Compounds 2–23 were prepared in a straightforward manner, through the route outlined in
6
1
Scheme 1. Their synthesis was initiated by incorporation of C17 substituent in the
commercially available ethylene deltanone 29. Grignard reaction of ketone 29 with
alkynylmagnesium bromide or alkynyllithium provided 17βꢀhydroxy ketal 30 exclusively.
Treatment of 30 with 30% hydrogen peroxide and 1,1,1,3,3,3ꢀhexafluoropropanꢀ2ꢀone trihydrate
in the presence of sodium hydrogen phosphate buffer provided epoxide 31 as a major
ACS Paragon Plus Environment
28

Page 29 of 66
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
diastereomer (dr > 3:1). Copperꢀcatalyzed addition of an alkyl or aryl Grignard reagent to the
vinylic epoxide 31, followed by hydrolysis with 70% acetic acid at 55 °C gave compounds 2–23
in moderate to high yields.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 1
a
2
Reagents and conditions: (a) for 30a: CH C≡CMgBr, THF, 0 °C to 25 °C; for 30b–30f: R Li,
3
1
˗78 °C to 25°C, 24–87%; (b) 35% H O , CF COCF , Na HPO , DCM, 42–94%; (c) R MgBr,
2
2
3
3
2
4
CuI, THF, 0 °C to 25°C, 59–83%; (d) 70% AcOH, 55 °C, 1–2 h, 46–96%.
ACS Paragon Plus Environment
29