Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Digest
Breakthroughs in neuroactive steroid drug discovery
Maria-Jesus Blanco ⇑, Daniel La, Quinn Coughlin, Caitlin A. Newman, Andrew M. Griffin,
Boyd L. Harrison, Francesco G. Salituro
Sage Therapeutics, Inc., 215 First Street, Cambridge, MA 02142, USA
a r t i c l e
i n f o
Article history:
Received 7 October 2017
Revised 26 November 2017
Accepted 27 November 2017
Available online 2 December 2017
Keywords:
Neurosteroid
NAS
GABAA
NMDA
Synaptic
Extrasynaptic
Clinical candidate
a b s t r a c t
Endogenous and synthetic neuroactive steroids (NASs) or neurosteroids are effective modulators of multiple signaling pathways including receptors for the c-aminobutyric acid A (GABAA) and glutamate, in
particular N-methyl-D-aspartate (NMDA). These receptors are the major inhibitory and excitatory neurotransmitters in the central nervous system (CNS), and there is growing evidence suggesting that dysregulation of neurosteroid production plays a role in numerous neurological disorders. The significant unmet
medical need for treatment of CNS disorders has increased the interest for these types of compounds. In
this review, we highlight recent progress in the clinical development of NAS drug candidates, in addition
to preclinical breakthroughs in the identification of novel NASs, mainly for GABAA and NMDA receptor
modulation.
Ó 2017 Elsevier Ltd. All rights reserved.
Neuroactive steroids (NASs) or neurosteroids are among the
most potent and effective modulators of neuronal excitability.
The term ‘‘neurosteroid” was first mentioned by Etienne Baulieu
in 19800 s and initially referred to endogenous steroids synthesized
in the brain and central nervous system (CNS) from cholesterol.1
Neurosteroids have been shown to impact CNS function primarily
through allosteric modulation of the GABA (c-aminobutyric acid)A
receptor (GABAAR) or the N-methyl-D-aspartic acid class of glutamate receptors (NMDAR). However, at high concentrations neurosteroids have been shown to act on other receptor systems like
nicotinic acetylcholine, serotonin 5-HT3, and sigma1 receptors.
The term neurosteroid has been expanded to include synthetic
and naturally-derived analogs that have CNS actions similar to
endogenous neurosteroids. Increasing evidence2,3 indicates that
dysregulation of neurosteroid production plays a role in the pathophysiology of stress and stress-related psychiatric disorders,
including mood and anxiety disorders. In addition to agonist or
antagonist modes of action at different receptors, these receptors
could be either positively (PAM) or negatively (NAM) modulated
by NAS compounds at allosteric sites. For example, allopregnanolone (1) is a PAM at the GABAAR (Fig. 1).
Other endogenous NASs, such as 24(S) hydroxycholesterol (2)
are PAMs at NMDA receptors.4 Such mechanisms are providing
novel approaches to treat CNS disorders, and the steroid field is
⇑ Corresponding author.
E-mail address: Maria-Jesus.Blanco@sagerx.com (M.-J. Blanco).
https://doi.org/10.1016/j.bmcl.2017.11.043
0960-894X/Ó 2017 Elsevier Ltd. All rights reserved.
showing a significant resurgence with multiple compounds
advancing to clinical studies. Clinical trials currently underway
are assessing the efficacy of various NASs for the treatment of
diverse CNS disorders such as epilepsy,5 super refractory status
epilepticus (SRSE), Fragile X, traumatic brain injury and
Alzheimer’s disease.3 In this BOMCL Digest, we will provide a brief
perspective on recently disclosed GABAAR and NMDAR NASs that
have either advanced to clinical studies or have been described
preclinically within the last 3–5 years.
Clinical NASs. Ten NAS compounds have reached clinical development status since the 19700 s (Table 1), however many have since
been withdrawn. For example, an intravenous anesthetic combining two NASs, alphaxolone (3) and alphadolone in a cremophor
vehicle was withdrawn from the market in 1984 due to issues with
anaphylaxis. Later in 2001, it was determined that Cremophor EL
was responsible for these anaphylactic reactions in humans.7 Currently, alphaxolone alone is under additional human studies in a
sulfobutyl ether-b-cyclodextrin formulation (‘‘Phaxan”) as an
intravenous anesthetic.8
Minaxolone (CCI-12923, 4), a GABAAR PAM, was developed as a
water-soluble anesthetic NAS. It was withdrawn before registration due to toxicity observed in long-term studies in rats.9 Early
emphasis on water-soluble NASs led to the discovery of ORG20599 (5), a potent GABAAR PAM. Unacceptable clinical profile,
however led to withdrawal from further clinical development.10
Marinus Pharmaceuticals has been developing ganaxolone
(GX, 6), a 3b methyl derivative of allopregnanolone for focal-onset
62
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
Fig. 1. Structures of Allopregnanolone (1) and 24 (S) hydroxycholesterol (2). Generic steroid nomenclature used in this manuscript. aCarbon numbering and ring designations
of the steroid core.6 Substituents above the plane of the paper are described as b and are shown as a solid line; those below the plane are described as a and are shown by a
broken line. Carbons at positions 4, 5 and/or 6 may be saturated or unsaturated.
seizures in adults and in children with epilepsy. Focal-onset seizures are manifestations of abnormal epileptic firing of brain cells.
It is estimated that 65 million people worldwide are living with
some form of epilepsy.24 A disclosure in 2016 announced the discontinuation of phase 3 clinical studies for adult focal onset seizures and advancing 6 in status epilepticus and pediatric orphan
indications.11
GABAAR antagonist sepranolone12 (7, Table 1) is currently in
clinical studies for the treatment of premenstrual dysphoric disorder (PMDD). This disease is a severe, debilitating form of premenstrual distress comprising emotional, physical symptoms and
functional impairment. PMDD affects 3–8% of women in fertile
ages.25 A phase 2b trial in PMDD patients was expected to begin
in the second half of 2017 after an initial phase 1/2 study demonstrated a statistically significant difference between active treatment and placebo.
Bruschettini SRL is developing tauroursodeoxycholic acid [8,
TUDCA, the taurine conjugate of ursodeoxycholic acid (UDCA)]
for the treatment of amyotrophic lateral sclerosis (ALS).13 ALS is
a progressive CNS disease where nerve cells in the brain and spinal
cord that control voluntary movement gradually deteriorate, causing loss of muscle function and paralysis. Several studies have
demonstrated that TUDCA imparts anti-apoptotic activity in a
number of neurodegenerative diseases, including ALS, Alzheimer’s
disease, Parkinson’s disease, and Huntington’s disease.26 It has
been indicated that UDCA modifies the function of the bile salt
export pump (BSEP, ABCB11) however underlying mechanisms
remain unknown. On February 27th 2017, orphan designation
was granted by the European Commission for TUDCA for the treatment of ALS.
The Shanghai Innovative Research Center of Traditional Chinese
Medicine (SIRC-TCM) is developing S-111 (9, Yuxintine, or 20(S)protopanaxadiol, Table 1), an active ginseng intestinal metabolite
and an inhibitor of serotonin and norepinephrine uptake, for the
treatment of depression, including major depressive disorder
(MDD). MDD is a widely distributed medical condition that
includes abnormalities of mood, appetite, sleep, cognition and psychomotor activity.27 Preclinical antidepressant-like activity of
orally administered S-111 was measured in various animal models
of depression and demonstrated antidepressant-like activity with
similar potency to fluoxetine.28 The latest report in 2015, indicated
the drug was in phase 2 development.14,15
Sage Therapeutics has recently disclosed phase 2 results for its
first generation NAS, SAGE-547 (10, Brexanolone).29 A parenteral,
continuous infusion formulation of SAGE-547, was in phase 3 clinical trials for the treatment of SRSE, a life-threatening condition in
which the brain is in a state of persistent seizure that fails to
respond to standard treatments.18,30,31 The study did not meet
the primary endpoint,32 comparing success in weaning of thirdline agents and resolution of potentially life-threatening status
epilepticus with brexanolone vs. placebo when added to standard-of-care. In addition, SAGE-547 has completed a phase 2 clinical trial in severe post-partum depression (PPD)29,33 and an
exploratory study in essential tremor (ET).34 Simultaneous with
the development program in SRSE, SAGE-547 was studied in a
phase 3 program in moderate and severe post-partum depression.29 There is considerable preclinical research supporting the
potential for GABAAR modulation imparting benefits in a number
of mood disorders. For post-partum depression in particular, there
is evidence sustaining the potential utility of NASs, such as allopregnanolone, in depressive mood disorders through modulation
of synaptic and extrasynaptic GABAARs.35 This disease, with no
approved drugs to date, is estimated to affect between 10 and
20% of women in the United States after childbirth.36 In early
November 2017, Sage Therapeutics announced positive top-line
results for phase 3 studies in moderate and severe post-partum
depression. Brexanolone achieved the primary endpoint in both
trials with a mean reduction from baseline in the Hamilton Rating
Scale for Depression (HAM-D) total score compared to placebo at
60 h.37
SAGE-547 is a potent GABAAR PAM, active at both synaptic and
extrasynaptic GABAARs, and is ideally suited for parenteral administration, due to its high intrinsic clearance and low volume of distribution, yielding a fast on/fast off pharmacokinetic profile.38 In
searching for next generation NASs, Sage Therapeutics has developed a molecule with robust pharmacological PAM activity at
GABAAR but with low intrinsic clearance, high oral bioavailability
and potential for once daily dosing aiming to reach larger populations of patients. To this end, Sage Therapeutics reported the discovery of SAGE-217 (11, Table 1),38,39 a clinical candidate which
has now completed single and multiple ascending doses (SAD,
MAD) phase 1 clinical trials and has progressed into phase 2 clinical trials for the treatment of GABAARs mediated movement disorders such as essential tremor (ET) and Parkinson’s disease as well
as in mood disorders such as PPD and MDD.
In April 2017, Sage Therapeutics announced the advancement of
SAGE-718 (12, Table 1) into phase 1 clinical studies.23 SAGE-718 is
a novel, oral, first-in-class oxysterol-based PAM of the NMDAR.
Positive modulation of NMDARs has potential benefit in the treatment of a range of neurological disorders associated with a variety
of cognitive, neurological and behavioral symptoms. SAGE-718
also has potential in the treatment of CNS disorders associated
with a high prevalence of anti-NMDA antibodies or reduced levels
of endogenous 24(S)-hydroxycholesterol (2). In preclinical studies,
SAGE-718 improved social behavior in an animal model of NMDA
hypofunction, and ameliorated both behavioral and electrophysiological deficits in a model of compromised cholesterol regulation.23
In general, development of NASs faces several challenges as
seen from previous examples. Many of the issues are related to formulation40 as those compounds tend to have physicochemical
properties outside of the traditional small molecule drug-like properties (Table 1). Attempts were made to reduce lipophilicity of the
compounds and increase aqueous solubility (4, 5), however the
compounds led to unacceptable margins of safety. It is important
to highlight that the receptors are membrane bound. Chisari41 postulated that NASs might require a membranous route of access to
63
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
Table 1
Progression of NASs in clinical development.
a
Compd
number
Compd
name
3
Structure
Company/
organization
Current Status
Ref.
Physicochemical
properties (MW/
cLogP/PSA)a
Alphaxolone &
alphadolone (Althesin)
Alphaxolone (Phaxan)
Glaxo Research
Group Ltd
(Althesin)
Drawbridge
Pharmaceuticals
(Phaxan).
Withdrawn from the market (1984).
Alfaxolone still used for veterinary
medicine. Alphaxalone alone (Phaxan) is
currently in human studies as anesthetic
formulated in sulphobutyl-ether bcyclodextrin.
7,8
332.5/3.7/54.4
4
Minaxolone
(CCI-12923)
Glaxo Research
Group Ltd
Withdrawn before registration (1979)
9
405.6/3.1/49.8
5
ORG-20599
Organon
International
Halted further clinical studies (2001)
10
438.1/4.3/49.8
6
Ganaxolone (GX)
Marinus
Pharmaceuticals
Phase 3 for adult focal-onset seizures
discontinued (6/2016). Phase 2 studies
ongoing for epilepsy indications (status
epilepticus and pediatric orphan
indications).
11
332.5/5.0/37.3
7
Sepranolone
Swedish Asarina
Pharma
Phase 2, premenstrual syndrome (GABAA
Antagonist); iv
12
318.5/4.5/37.3
8
Tauroursodeoxycholic
acid
Bruschettini SRL
Phase 2, Motor neuron disease (ALS); oral
13
499.7/2.1/123.9
9
S-111 (Yuxintine)
Shanghai
Innovative
Research Center
of Traditional
Chinese Medicine
Phase 2, Major depressive disorder (MDD);
oral
14,15
460.4/6.8/60.7
10
SAGE-547
(Brexanolone)
Sage Therapeutics
Achieved primary endpoints for two Phase 3
studies in postpartum depression (PPD, 11/
2017); Phase 2 for Essential tremor (ET).
Discontinued Phase 3 for super refractory
status epilepticus (SRSE, 9/2017).
16–18
318.5/4.5/37.3
11
SAGE-217
Sage Therapeutics
Phase 2, Severe PPD, MDD, ET, Parkinson
disease
19–22
409.6/4.3/76.7
12
SAGE-718
Sage Therapeutics
Phase 1 (4/2017)
23
Undisclosed
Physico-chemical properties calculated using ChemDraw 16.0 software. PerkinElmer Informatics Inc.
transmembrane-domain binding sites within GABAAR. As more
compounds move into development, there is an opportunity to
understand the translation from preclinical to clinical studies
including target engagement,42 dose, route of administration,
duration, and safety.
Preclinical Neurosteroids. Moving into the second part of the
manuscript, the focus will be on summarizing preclinical reports
on GABAAR and NMDAR modulators. Each section will include a
brief introduction on the receptor.
Preclinical GABAAR NAS PAMs. The GABAAR is a pentameric ion
channel surrounding a central chloride ion-selective channel gated
by GABA (Fig. 2). Functional receptors consist of five subunits
selected from 19 known receptor subunits: a1–6, b1–3, c1–3, d,
e, h,p, and q1–3. This specific assembly results in a complex
64
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
Fig. 3. Oxygen bridge NAS analogs described by Coirini et al. including IC50
inhibition values of [35S]-TBPS binding to male rat synaptosomes.
Fig. 2. Schematic representation of the GABAAR.
heterogeneity of GABAAR subtypes, and each subtype has a distinct
physiological and pharmacological profile. Distinct receptor subtypes have a specific regional and cellular expression pattern. For
example, GABAARs that contain a1–3, b1-3, and c2 subunits are
mainly synaptic, whereas a4–6 and d-containing receptors are
mainly peri- or extrasynaptically located. Allopregnanolone and
other NASs are potent allosteric modulators of both synaptic and
extrasynaptic GABAARs. Hence, NASs can enhance synaptic phasic
and extrasynaptic tonic inhibition. The resulting chloride current
conductance generates a form of shunting inhibition that controls
network excitability, seizures, and behavior.43
Structurally, NASs have an ABCD steroid core and follows the
steroid nomenclature depicted in Fig. 1. NASs have the potential
to differentiate from classical GABAAR modulators in several ways.
First, NASs differentiate from benzodiazepines by targeting different populations of GABAARs. NASs have been suggested to putatively bind to three or four distinct sites on the GABAAR (Fig. 2).
However, benzodiazepines bind to an allosteric site distinct from
Compound
1
a
the GABA-binding site, at the interface of the a and c subunits. This
limits their ability to potentiate synaptic GABA currents to receptor
assemblies that contain a c subunit. Conversely, NASs potentiate
the GABAAR by binding to residues within the a subunit and so
modulate receptors independently of their subunit composition.
Therefore, unlike benzodiazepines, NASs are capable of targeting
extrasynaptic GABAAR that include the d subunit in addition to
synaptic c-containing receptors. As such, NASs may exhibit a therapeutic advantage44 over benzodiazepines that has led to renewed
medicinal chemistry efforts to explore novel analogs. For example,
Coirini and colleagues45 recently reported the synthesis of allopregnanolone and pregnanolone analogs with an intramolecular
oxygen bridge between A and B rings (Fig. 3). These conformationally constrained analogs were hypothesized, according to the
authors, to possess a more favorable spatial arrangement for binding. The compounds (13–16) were evaluated by [35S]-TBPS binding
competition assay and shown to have similar activity in this assay
relative to allopregnanolone (1, IC50 = 86 nM reported). Compound
14 was further evaluated in cerebral cortex and hippocampus cultures subjected to hypoxia to evaluate neuroprotective activity.
Compound 14 was shown to prevent the increase of glial fibrillary
acidic protein as well as neurofilament (NF160/200) decrease in
hippocampus cultures subjected to ischemia. Although no DMPK
[35S]-TBPS binding competition assay
IC50 (nM)
80.0 + 14.3
17
255.3 + 17.9
18
625.0 + 125.0
19
5000.0 + 100.0
IC50 records the steroid concentration producing a half-maximal inhibition of the [35S]-TBPS binding
competition assay.
Fig. 4. Epoxide neurosteroids reported by Kasal and coworkers.
65
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
Activity/property
GX (6)
UCI-50027 (20)
0.2
1.2
0.1
0.2
Solubility
20 mM in 45% HP CD
52 mM in 20% HP CD
Rat %F
No data reported
1 2 2L GABAAR EC50
(oocytes, M)
2 1 2L GABAAR EC50
(oocytes, M)
b
In vivo activity (mouse, po)
EPM MED= 20 mg/kg
Preliminary safety (Mouse)
aReported
77a
c
b
PTZ
EPM MED < 0.3 mg/kg
ED50 = 23 mg/kg
PTZc ED50 = 6 mg/kg
TI as anxiolytic (EPM) = 3.3
TI as anxiolytic (EPM) > 127
TI as anticonvulsant = 3
TI as anticonvulsant ~ 6
dosing 3 mg/kg p.o. and 1 mg/kg i.v. bEPM= elevated plus maze (EPM) model for
anxiety; cPTZ = pentylenetetrazole-induced seizures assay
Fig. 5. Comparison of preclinical profile of UCI-50027 (20) vs. GX (6).
Fig. 7. Nor-19-pregnanolone analog SGE-516.
NAS
Protection against A 42 toxicity
on adult neural stem cells
1
+
21
++
22
++
Table 2
In vitro GABAA pharmacology and DMPK profile for SGE-516.49
Fig. 6. O-allyl neurosteroid analogs described by Karout et al.48
a1b2 c2
EC50 (nM)/Emax
(%)
a4b3d
EC50 (nM)/Emax
(%)
Rat Oral%
Fa
Brain:plasma
ratioa
125/663
240/579
27
1.8
a
data or other supporting data was provided, the authors stated that
these oxygen bridge analogs might offer promise in other therapeutic applications.
Kasal et. al. have explored the hypothesis that a 2a,3a-epoxy
ring could replace the 3a-hydroxyl group,46 synthesizing several
analogs (Fig. 4).
Epoxides 17 and 18 displayed somewhat weaker inhibition relative to 1, while 19 displayed significantly lower activity than 1.
Additionally, the corresponding diol formed by epoxide opening,
2b,3a-dihydroxy-5a-pregnan-20-one, did not inhibit [35S]-TBPS
binding. This result demonstrated that the activity of epoxide 17
is due to the compound itself and not to its hydrolytic product.
Compound 17 exhibited moderate anticonvulsant and short-lasting activity when tested in motor seizures induced by pentylenetetrazol (PTZ) at 20 mg/kg. The authors highlighted that although the
epoxide could be an acceptable substitution for 3a-hydroxy group,
in vivo results showed that the epoxide analogs are highly
metabolized.
Hogenkamp47 and colleagues have studied potential novel bioisostere replacements of the 17b-acetyl side chain to improve the
pharmacological profile of GX (6, Table 1). Isoxazoles are known
In vivo PK parameters following iv (5 mg/kg) and oral (20 mg/kg). Brain:plasma
ratios are obtained by single point at 30 min post iv. Oral bioavailability and
brain:plasma ratios were calculated from mean plasma concentrations from two
rats.
bioisosteres for ketones, and in this case, it was postulated that
an isoxazole would maintain activity at GABAARs while improving
the DMPK profile of GX (6). This work led to the discovery of
3-[3a-hydroxy-3b-methyl-5a-androstan-17b-yl]-5-(hydroxymethyl)isoxazole also known as UCI-50027 (20, Fig. 5). Like 6, UCI-50027
(20) was shown to have activity as GABAA PAM when tested in
Xenopus oocytes expressing a1b2c2L and a2b1c2L GABAARs
(Fig. 5). While less potent than 6 in vitro, UCI-50027 is more potent
in vivo both as an anxiolytic and as an anticonvulsant. The minimum effective dose (MED) for UCI-50027 is 0.3 mg/kg when
administered orally in the mouse elevated plus maze (EPM) paradigm, while GX has an MED of 20 mg/kg. Similarly, UCI-50027 is
more potent than GX in anticonvulsant models. The ED50 for UCI50027 is 6 mg/kg when administered orally in mouse against
PTZ-induced seizures assay, while GX has an oral ED50 of 23 mg/
kg. The authors hypothesized that the subtype receptor selectivity
could explain the differences in the anxiolytic profile between the
66
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
Table 3
Activity of SGE-516 in mouse PTZ-induced seizure assay and PK parameters.50
a
Dose (mg/kg, po)
% clonic seizure
% tonic seizure
% death
Plasma concentration (ng/mL)a
Brain concentration (ng/g)a
vehicle
3
10
30
100
90
20
0
80
40
0
0
80
30
0
0
NA
87
703
1689
NA
87
773
1955
All samples collected 30 min after administration.
Table 4
Correlation between NASs modulation of tonic current and anticonvulsant profiles in
the 6-Hz seizure model in mice.57
Compound
NAS-mediated potentiation of tonic
currents by 1 lM GABAa
6-Hz Test ED50
(mg/kg)
1
24
25
26
6
3
4.36
2.24
0.76
3.52
3.47
1.98
4.2
7.7
>100
5.0
1.5
8.8
a
Values derived from fold potentiation of mean, normalized tonic current
response to 1 lM GABA. Tonic current responses were recorded from voltageclamped (65 mV) DGGCs from wild type female mice. All mean values are representative of four to eight cells per neurosteroid and concentration.
two compounds (and it was not linked to brain penetration profile). Pharmacokinetic (PK) studies indicated that UCI-50027
showed larger exposures than GX when dosed orally in rats. When
UCI-50027 was dosed orally at 3 mg/kg in rats, the maximum
plasma levels were 200 ng/mL while a 10 mg/kg p.o. dose of GX
led to plasma levels of 37 ng/mL. The authors indicated that the
observed superior in vivo potency for UCI-50027 (vs. GX) was
due to an increased oral bioavailability, compensating for the
lower in vitro potency of UCI-50027.
Karout48 and coworkers tested a set of synthetic derivatives of
allopregnanolone to potentially identify NASs with increased
efficacy, in particular their neuroprotective activity against
the toxic effects of amyloid beta 42 peptide (Ab42), a major component of amyloid plaques. The authors focused their research on
3-O-allyl analogs (Fig. 6) to increase metabolic stability vs allopregnanolone (1).
Compound 1 or its analogs were added to neurosphere cultures
1 h before Ab42 was applied for 2 days. Fluorometric measurement
of caspase-3/7 activity was used to determine the extent of apoptotic cell death. Under these conditions, the Ab42-induced increase
in fluorescence was significantly reduced (40%) in the presence of
1. The 3-O-allyl analogs, irrespective of orientation of the substituent, efficiently protected against Ab42 toxicity, reducing apoptotic activity to basal levels. The authors postulated that since both
isomers showed equal neuroprotective effects, the mechanism of
action could be off target rather than mediated through the
GABAAR.
Sage Therapeutics has also reported on a series of nor-19 pregnanolone bearing a heterocycle substituent at C21. These efforts
resulted in the identification of SGE-516,49 a PAM of both gamma
and delta- containing GABAAR (Fig. 7, Table 2). The broad GABAAR
activity differentiates NASs like SGE-516 from benzodiazepines, a
class of anticonvulsants which have been shown in vitro to selectively target gamma-subunit containing GABAARs. In addition,
SGE-516 has PK properties suitable for chronic oral dosing
(Table 2). Compound SGE-516 has been profiled in numerous
in vivo preclinical models of seizure activity. In mice, SGE-516 protected against acute seizures in the PTZ-induced chemo-convulsant seizure model (ED50 = 2.0 mg/kg, ip) and the 6-Hz
psychomotor seizure model (ED50 = 3.8 mg/kg). In addition, SGE516 demonstrated anticonvulsant activity in the mouse corneal
electrical kindling assay, a model thought to be relevant for chronic
epilepsy. After oral administration, SGE-516 increased the latency
to clonic and tonic seizures in a dose-dependent fashion (ED50 =
2.82 mg/kg) and prevented death after PTZ administration
(Table 3).50
For the last several years, there has been an increasing interest
in extrasynaptic GABAAR activity and its translation into potentially differentiated pharmacology through tonic modulation not
previously studied for neuronal excitability.51–54 The importance
of this tonic regulation is reflected in the potential therapeutic
Fig. 8. Reddy’s description of functional activation of extrasynaptic d GABAARs.57
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
67
Fig. 12. Formula used to estimate IC50 by Kudova et al.
Fig. 9. Nor-19-pregnanolone analog SGE-872.
Table 5
In vitro ephys GABAA pharmacology of SGE-872.49.
a1b2c2
EC50 (nM)/Emax (%)
a4b3d
EC50 (nM)/Emax (%)
>3000/>744
178/858
Fig. 10. Representatives of the two classes of NAS negative modulators of NMDARs.
opportunities of extrasynaptic GABAARs as anesthetics, sleep-promoting drugs, and alcohol addiction disorders. NASs have shown
greater sensitivity to extrasynaptic GABAARs55 in comparison to
other compounds such as benzodiazepines, leading to their importance in stress-, ovarian cycle-, and pregnancy-related mood disorders. Additionally, disruptions in network dynamics associated
with schizophrenia, epilepsy, and Parkinson’s disease might
involve alterations in the tonic GABAA receptors-mediated conductance.56 These potential therapeutic opportunities have increased
the relevance to identify suitable NAS tools (see Table 4).
In 2016, Reddy and colleagues reported some key structural features for the identification of a NAS pharmacophore to modulate
extrasynaptic GABAAR–mediated tonic inhibition in murine dentate gyrus granule cells (DGGCs) in the hippocampus using well
known natural and synthetic NASs (Fig. 8).57 The pharmacophore
map summarizes the correlation between NAS modulation of tonic
current and anticonvulsant profiles in the 6-Hz seizure model. The
authors built this correlation after performing experiments in
granule cells from d-knockout mice showing that NAS potentiation
of tonic currents was completely (95%) diminished. Furthermore,
the authors confirmed the NAS sensitivity of dGABAARs at the systems level using anticonvulsant 6-Hz seizure mice model. As a
result, the authors highlighted that the 3a-OH (24) substitution
was crucial for extrasynaptic receptor functional activity, since
the 3b-OH epimer (25) was inactive in activating tonic currents.
Allopregnanolone (1) exhibited the highest potency and maximal
efficacy in promoting tonic currents. Modifications at the C20
position of NAS, (tetrahydrodeoxycorticosterone, THDOC, 26)
Fig. 13. Various pregnane derivatives investigated for inhibitory activity.
significantly changed the transduction kinetics of tonic current
activation. It is important to emphasize the moderate correlation
between an ex-vivo slice assay in DGGCs and an in vivo assay
(6-Hz seizure model) with compounds that might have different
PK profiles.
The NAS-mediated increase in phosphorylation and surface
levels of extrasynaptic GABAARs seen with THDOC (26) leads to
an increase in tonic current. This is an additional mechanism by
which NASs can increase the inhibitory tone in the brain in addition to the known allosteric effect. Modgil and collaborators58 evaluated various NASs (1, 6 and SGE-516) on their ability to increase
tonic current by promoting GABAAR phosphorylation and membrane trafficking dependent on protein kinase C (PKC) activity. A
sustained increase in tonic current was observed following exposure to 1, or SGE-516 however no increase in tonic current was
observed with exposure to 6. In agreement with the observations
of an increased tonic current, 1 and SGE-516 increased the phosphorylation and surface expression of the b3 subunit-containing
GABAARs. The authors postulate that those results open an interesting area of potential therapeutic targets aimed at modulating
the trafficking of a particularly important subset of GABAARs.
In 2015, Sage Therapeutics reported the identification of SGE872,49 a 5-N azabenzopyrazole NAS with a strong preference for
a4b3d GABAA vs. a1b2c receptors (>17-fold, Fig. 9, Table 5) as a
potential tool molecule to investigate extrasynaptic pharmacology.
Preclinical NMDAR NASs. In addition of GABAAR modulators,
there have been recent advancements within the context of
NMDARs. While there has been significant activity outside of
steroids,59 the last section of this review will focus specifically on
the NAS genre of modulators. The NMDARs are non-selective
cation channels, that play an important role in excitatory synaptic
transmission and several forms of synaptic plasticity.60 NMDARs
belong to the ionotropic glutamate receptor family but unlike most
receptors, NMDARs contain a magnesium ion which blocks the
channel and requires dissociation prior to activation. The ion is
removed through activation of co-localized AMPA receptors which
allows the influx of sodium ions thereby increasing internal positive charge resulting in electrostatic repulsion liberating the magnesium ion. Upon binding of glutamate and glycine or D-serine to
Fig. 11. Key compounds demonstrating the effect of C17 substitution.
68
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
Fig. 14. Amide-based steroidal inhibitors.
Fig. 15. SAR of NAS-like perhydrophenanthrene analogs.
the ligand binding domain (LBD) of NMDAR, tertiary changes
results in the passage of calcium ions into the postsynaptic neuron.
The influx of calcium leads to downstream cellular cascades which
strengthens signaling through increasing expression of signaling
receptors and triggering development of more synapses in the
same pathway. The signaling pathway, therefore, is strengthened
only when sufficiently stimulated; this is the basic principle
behind long-term potentiation (LTP) and a hypothesized biological
basis of learning and memory. Given NMDARs central role in
synaptic physiology, excessive activation has been implicated in
excitotoxicity, central to the damage of neurological disorders such
as stroke, traumatic brain injury (TBI), neurodegeneration and
depression. In addition, NMDAR hypofunction has been associated
with schizophrenia, Alzheimer’s diseases and dementia.61
The NMDAR is a heterotetramer comprised of a dimer-ofdimers arising from two N1 and two N2(A–D) or N3(A, B) subunits.
There are four domains, the amino terminal domain (ATD), the
ligand binding domain (LBD), the transmembrane domain (TMD)
and the cytoplasmic carboxy terminal domain (CTD). The TMD
contains a series of four a-helices, M1-4, which undergo structural
changes in response to conformational movement upon agonist
binding to the LBD therefore coupling ligand association and channel gating.62 NASs that modulate NMDARs can be categorized in
two general classes, the oxysterols, which arise from cholestane
hydroxylation4 and pregnane derivatives that bear a charge group
at C3 such as a sulfate, hemisuccinate, glutamate, etc. Representative examples, 24 (S)-hydroxycholesterol (2, 24HC) and pregnenolone sulfate (28, PS, pregn-5-en-3b-ol-20-one 3b-sulfate),
are shown in Fig. 10. Modulators from both classes can act as either
PAMs or NAMs. While the exact binding sites of these NASs are not
fully understood, chimera experiments that swapped domains
between rat NMDAR and GluK2 kainate receptors revealed that 2
requires the TMD and 28 requires the TMD and LBD of the NMDAR
for potentiation.63 The authors suggest that while more extensive
mechanistic studies are required to elucidate the exact site of binding for the NAS modulators, these regions could contain the site of
binding.
Given the strong association between NMDAR hyperactivity
and neurological disorders, NASs inspired by pregnanolone sulfate
(PAS, 29, Fig. 11), an endogenous inhibitor of NMDAR, have been
investigated.64 The amphiphilic nature of this class of NASs gives
rise to complications stemming from the equilibrium between single vs aggregate forms in solution. The authors, therefore assert
that the formula below, which assumes 100% inhibition at saturating inhibitor concentrations, accurately estimates IC50. The formula
employed is shown in Fig. 12, where [compound] is the steroid
concentration, II is the relative degree of inhibition and h (the
apparent Hill coefficient) is fixed at 1.2.
From the SAR, dramatic changes in potency were observed with
modification at the C17 position. Improvements in potency were
observed up to 273-fold (31) over 29 (Fig. 11) and generally tolerant to halogenation and unsaturation. As a tool molecule, compound 30 (0.60 mM) demonstrated activity against all isoforms of
GluN2(A–D). In addition, 30 is active when evaluating NMDA,
AMPA and GABA receptor response in hippocampal neurons. The
authors also presented a strong correlation between experimental
DGexp values and calculated logP and logD. Given this relationship,
the authors concluded that the potency within this class could be
dictated by protein affinity as well as changes in membrane concentration proximal to the receptor. Nevertheless, the same group
suggested that the site of action of 29 is the extracellular vestibule
of the ion channel pore65 and is consistent with the chimera work
described above.63
In addition to sulfate NAS inhibitors, charged ester derivatives
such as those in Fig. 13, have also been shown to selectively inhibit
tonic versus phasic NMDAR signaling.66 The authors indicated that
within the context of neurological disease treatment, a tonic inhibitor may offer a greater therapeutic index due to the association
between synaptic/phasic inhibition and psychomimetic symptoms.
This investigation revealed that as the chain length increases, so
does potency and tonic selectivity. The steady-state potencies ranged from 10–60 lM on GluN1/GluN2B receptors where the general trend was greater potency with increased chain length [PAS
(29) was measured at 23 lM]. Ultimately, the hemipimelate analog
(PA-hPim, 37) was determined to have a potency of 7.4–12 mM for
tonic inhibition of native NMDARs. In contrast, this compound did
not inhibit synaptically activated NMDARs. Compound 37 was further tested in several in vivo models evaluating psychomimetic
symptoms such as the locomotor activity, passive avoidance task
and the Y-maze spontaneous alteration task assays. Hyperlocomotion was not observed with 1 or 10 mg/kg of 37. In addition, as
opposed to control MK-801, it had no effect in the step-through
passive avoidance test or on spontaneous alteration in the Y-maze.
When advanced to a intrahippocampal NMDA infusion study, 37
reversed learning deficits. In totality, the authors suggested that
this body of work emphasized the potential of NASs as novel
therapeutics.
Influenced by this study as well as those with 29, analogs of
pregnanolone glutamate (PAG, 38, Fig. 14)67 were synthesized,
attempting to improve the pharmacokinetic profiles.68 With the
four compounds illustrated in Fig. 14, improvements in potency
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
Fig. 16. Structure of tetrol.
Fig. 17. Structures of NMDAR PAMs SGE-201 and SGE-301.
were seen, in particular with compounds 41 and 42. The permeability, as measured in the Caco-2 assay, was as low as 0.22 ±
0.060 106 cm/s and did not exceed 5.85 ± 0.060 106 cm/s
with recoveries that were typically <50%. Compounds 39–42 were
further assessed in a spontaneous locomotor activity test and all
compounds had significant effects on distance travelled except
for compound 41. The authors suggested that the reduced effect
in relation to the other analogs was due to lower CNS penetration
stemming from poor Caco-2 permeability. Compound brain concentrations were not provided to support these assumptions.
Advances in sulfate and glutamate-NAS research has prompted
further investigation of this scaffold through modifications of the
tetracyclic core. Deconstruction of the D-ring was examined by
the synthesis of perhydrophenanthrenes (compounds 43–46) as
illustrated in Fig. 15 which highlights the most potent compounds
reported.69 The potency values observed, which were measured in
HEK293 cells transfected with rat GluN1 and GluN2B subunits,
were similar to pregnanolone sulfate (29, 24.6 mM, Fig. 11). In general, methyl ether analogs (43–44, R2 = CH2OCH3 in Fig. 15) were
not as potent as the unsubstituted analogs (45–46, R2 = H). Further
in vitro evaluation of 45 revealed that it was not subtype selective
(GluN2A-D) and less potent against AMPA receptor responses (83
mM) in hippocampal neurons. Conversely, 45 was more potent and
effective in inhibiting GABAAR responses (3.1 lM). The authors
concluded, therefore, that the neuroid D-ring is critical for
NMDAR/GABAR selectivity and provides potential avenues for
optimization.
While marine steroid 24-methylenecholestane-3b,5a,6b,19tetrol70 (tetrol, 47, Fig. 16) has previously been characterized as
an anti-cancer agent, more recently, it has been determined to be
an NMDA negative modulator.71 Tetrol inhibits NMDA-induced
calcium concentration (IC50 = 7.8 mM) and NMDA current in cortical neurons (IC50 = 10.3 mM). Tetrol was synthesized in 10 steps
from hyodeoxycholic acid (HDCA) which enabled further evaluation in models of neuronal injury. For instance, 10 mM of tetrol
increased the number of surviving cerebellar granule neurons
when exposed to 200 mM glutamate, a concentration that kills neurons in the control. Staining experiments employing fluorescin
diacetate revealed that improvements in viability were observed
at concentration as low as 2.5 mM. The authors also demonstrated
that 47 significantly reduce the infarction volume by roughly 30%
in a rat permanent focal cerebral ischemia model induced by middle cerebral artery occlusion (MCAO).
In addition to the negative modulators described above, considerable progress has been made towards identifying PAM NASs at
69
the NMDAR. Through a screen of endogenous oxysterols and
cholesterol metabolites, Paul et al.4 reported the identification of
2 (Fig. 1) as a PAM with an estimated EC50 of 1.2 mM in hippocampal neurons. In addition, two synthetic analogs, SGE-201 and SGE301 (48–49, Fig. 17), were shown to have PAM activity and the
authors suggested that while the binding site of 2 is disparate from
PS (28, Fig. 10), it is likely shared with these structural homologous
analogs. Compounds 2 and 48 at 1 mM each potentiated NMDAR
EPSCs which were pharmacologically isolated by NBQX and gabazine, agents that antagonize the AMPA and GABA receptors respectively, thereby eliminating confounding signaling. Compounds 2
and 48 enhanced long-term potentiation (LTP) in hippocampal
slices and reversed synaptic plasticity deficits induced by ketamine
and also reverses MK-801 induced deficits in spontaneous alterations in a mouse Y-maze test. Additionally, 49 rescues social
interaction deficits and novel object recognition in PCP treated rats.
As an extension of this work, Sage Therapeutics has recently
advanced the first NMDAR PAM, SAGE-718, into phase 1 clinical
trials.23
In summary, there has been tremendous interest in the development of novel NASs during the past 5 years. There has been a
continuous flow of new NASs entering clinical studies with some
showing promising positive results in phase 2 and 3. Although
we have emphasized the importance of GABAA and NMDA receptor
modulators, opportunities exist to explore the effects of NASs on
alternative intracellular targets and other signaling pathways
intersecting with actions at the membrane with GABA, glutamate
or other ion channels. NASs have demonstrated specific interactions with multiple CNS targets and represent an attractive starting
point for novel therapeutics.
Acknowledgements
We thank Dr. A. Robichaud and Dr. J. Doherty for valuable discussions during the elaboration of this review.
References
1. Baulieu EE. Steroid hormone regulation of the brain. In: Fuxe K, Gustafsson J-Å,
Wetterberg L, eds. Steroid hormones in the brain: several mechanisms,
34. Stockholm: Pergamon Press Ltd; 1981:3–4.
2. Zorumski CF, Paul SM, Izumi Y, Covey DF, Mennerick S. Neurosci Biobehav Rev.
2013;37:109.
3. Reddy DS, Estes WA. Trends Pharmacol Sci. 2016;37:543.
4. Paul SM, Doherty JJ, Robichaud AJ, et al. J Neurosci. 2013;33:17290.
5. Serrano, E., Kanner, A. F1000Prime Rep. 2015, 7.
6. Panico R, Powell WH, Richer JC. IUPAC commission on the nomenclature of organic
chemistry: a guide to IUPAC nomenclature of organic compounds, vol. 8. Oxford
UK: Blackwell Scientific Publications Ltd.; 1994. 71.
7. Clarke RS, Dundee JW, Carson IW. Proc R Soc Med. 1973;66:49.
8. Monagle J, Siu L, Worrell J, Goodchild CS, Serrao JM. Anesth Analg.
2015;121:914.
9. McNeill HG, Clarke RS, Dundee JW, Briggs LP. Anaesthesia. 1981;36:592.
10. Drug Profile: ORG 20599. <http://adisinsight.springer.com/drugs/800003651>;
2001 Accessed 16.08.17.
11. Caperelli,
L.
<http://ir.marinuspharma.com/releasedetail.cfm?releaseid=
975400>; Published 2016 Accessed 16.08.17.
12. Bixo M, Ekberg K, Poromaa IS, et al. Psychoneuroendocrinology. 2017;80:46.
13. Elia AE, Lalli S, Monsurro MR, et al. Eur J Neurol. 2016;23:45.
14. Ling J, Yu Y, Zhu J, et al. J Chromatogr B Anal Technol Biomed Life Sci.
2016;1031:214.
15. Comparison of Yuxintine with Placebo in Treatment of MDD.
<https://clinicaltrials.gov/ct2/show/NCT02395263?term=Yuxintine&cntry1=ES
%3ACN&rank=1>; 2015 Accessed 16.08.17.
16. A Study to Evaluate SAGE-547 in Patients with Moderate Postpartum
Depression.
<https://clinicaltrials.gov/ct2/show/NCT02942017>;
2016
Accessed 16.08.17.
17. A Study to Evaluate SAGE-547 in Patients with Severe Postpartum Depression.
<https://clinicaltrials.gov/ct2/show/NCT02614547>; 2015 Accessed 16.08.17.
18. A Study with SAGE-547 for Super-Refractory Status Epilepticus.
<https://clinicaltrials.gov/ct2/show/NCT02477618>; 2015 Accessed 16.08.17.
19. A Study to Evaluate SAGE-217 in Subjects with Moderate to Severe Major
Depressive Disorder. <https://clinicaltrials.gov/ct2/show/NCT03000530>; 2016
Accessed 16.08.17.
70
M.-J. Blanco et al. / Bioorganic & Medicinal Chemistry Letters 28 (2018) 61–70
20. A Study to Evaluate SAGE-217 in Subjects with Parkinson’s Disease.
<https://clinicaltrials.gov/ct2/show/NCT03000569>; 2016 Accessed 16.08.17.
21. A Study to Evaluate SAGE-217 in Subjects with Essential Tremor.
<https://clinicaltrials.gov/ct2/show/NCT02978781>; 2016 Accessed 16.08.17.
22. A Study to Evaluate SAGE-217 in Subjects with Severe Postpartum Depression.
<https://clinicaltrials.gov/ct2/show/NCT02978326>; 2016 Accessed 16.08.17.
23. Cox, P. <http://investor.sagerx.com/news-releases/news-release-details/sagetherapeutics-announces-initiation-phase-1-development-and>; 2017 Accessed
16.08.17.
24. Thurman DJ, Beghi E, Begley CE, et al. Epilepsia. 2011;52:2.
25. Lanza di Scalea T, Pearlstein T. Psychiatr Clin North Am. 2017;40:201.
26. Vang S, Longley K, Steer CJ, Low WC. Glob Adv Heal Med. 2014;3:58.
27. Fava M, Kendler KS. Neuron. 2000;28:335.
28. Xu C, Teng J, Chen W, et al. Prog Neuro-Psychopharmacol Biol Psychiatry.
2010;34:1402.
29. Kanes SJ, Colquhoun H, Doherty J, et al. Hum Psychopharmacol Clin Exp.
2017;32:e2576.
30. Vaitkevicius H, Husain AM, Rosenthal ES, et al. Ann Clin Transl Neurol.
2017;4:411.
31. Bialer M, Johannessen SI, Levy RH, Perucca E, Tomson T, White HS. Epilepsy Res.
2015;111:85.
32. Cox, P. <http://www.businesswire.com/news/home/20170912005509/en/
Sage-Therapeutics-Reports-Top-Line-Results-Phase-3>;
2017
Accessed
21.11.17.
33. Kanes S, Colquhoun H, Gunduz-Bruce H, et al. Lancet. 2017;390:480.
34. Ellenbogen A, Raines S, Kanes S. Neurology. 2016;86:P4–297.
35. Maguire J, Mody I. Neuron. 2008;59:207.
36. Gavin NI, Gaynes BN, Lohr KN, Meltzer-Brody S, Gartlehner G, Swinson T. Obstet
Gynecol. 2005;106:1071.
37. Cox, L. <http://investor.sagerx.com/news-releases/news-release-details/sagetherapeutics-announces-brexanolone-achieves-primary>;
2017
Accessed
21.11.17.
38. Martinez Botella G, Salituro FG, Harrison BL, et al. J Med Chem. 2017;60:7810.
39. Robichaud, A. J. Abstr Pap 252nd ACS Natl Meet Expo, Philadelphia. 2016.
40. Irwin RW, Solinsky CM, Brinton RD. Front Cell Neurosci. 2014;8.
41. Chisari M, Eisenman LN, Covey DF, Mennerick S, Zorumski CF. Trends Neurosci.
2010;33:299.
42. Durham TB, Blanco M-J. Bioorg Med Chem Lett. 2015;25:998.
43. Carver CM, Reddy DS. Psychopharmacology. 2013;230:151.
44. Martinez Botella G, Ackley MA, Salituro FG, Doherty JJ. Annu Rep Med Chem.
2014;49:27.
45. Rey M, Kruse MS, Alvarez LD, et al. Exp Neurol. 2013;249:49.
46. Kasal A, Buděšínský M, Mareš P, et al. Steroids. 2016;105:12.
47. Hogenkamp DJ, Tran MB, Yoshimura RF, Johnstone TB, Kanner R, Gee KW.
Psychopharmacology. 2014;231:3517.
48. Karout M, Miesch M, Geoffroy P, et al. J Neurochem. 2016;139:782.
49. Martinez Botella G, Salituro FG, Harrison BL, et al. J Med Chem. 2015;58:3500.
50. Hammond RS, Althaus AL, Ackley MA, et al. J Epilepsy Res. 2017;134:16.
51. Carlson SL, Bohnsack PJ, Patel V, Morrow LA. J Pharmacol Exp Ther.
2016;356:148.
52. Clossen BL, Reddy DS. J Neurosci Res. 2017;95:1906.
53. Errington AC. Extrasynaptic GABAA receptors. In: Errington AC, Di Giovanni G,
Crunelli V, eds. New York: Springer; 2014:1–14.
54. Fritschy J, Panzanelli P. Extrasynaptic GABAA Receptors. In: Errington A, Di
Giovanni G, Crunelli V, eds. New York: Springer; 2014:15–32.
55. Abramian AM, Comenencia-Ortiz E, Modgil A, et al. Proc Natl Acad Sci USA.
2014;111:7132.
56. Kim YS, Yoon B. Exp Neurobiol. 2017;26:122.
57. Carver C, Reddy DS. J Pharmacol Exp Ther. 2016;357:188.
58. Modgil A, Parakala ML, Ackley MA, Doherty JJ, Moss SJ, Davies PA.
Neuropharmacology. 2017;113:314.
59. Strong KL, Jing Y, Prosser AR, Traynelis SF, Liotta DC. Expert Opin Ther Pat.
2014;24:1349.
60. Paoletti P, Bellone C, Zhou Q. Nat Rev Neurosci. 2013;14:383.
61. Zorumski CF, Izumi Y. Neurosci Biobehav Rev. 2012;36:989.
62. Traynelis SF, Wollmuth LP, Mcbain CJ, et al. Pharmacol Rev. 2014;62:405.
63. Wilding TJ, Lopez MN, Huettner XE. J Neurosci. 2016;36:8815.
64. Kudova E, Chodounska H, Slavikova B, et al. J Med Chem. 2015;58:5950.
65. Vyklicky V, Krausova B, Cerny J, et al. Sci Rep. 2015;5:10935.
66. Vyklicky V, Smejkalova T, Krausova B, et al. J Neurosci. 2016;36:2161.
67. Holubova K, Nekovarova T, Pistovcakova J, Sulcova A, Stuchlík A, Vales K. Front
Behav Neurosci. 2014;8:130.
68. Adla SK, Hubalkova P, Krausova B, et al. Steroids. 2017;117:52.
69. Slavikova B, Chodounska H, Nekardova M, et al. J Med Chem. 2016;59:4724.
70. Duh C-Y, Wang S-K, Chu M-J, Sheu J-H. J Nat Prod. 1998;61:1022.
71. Leng T, Liu A, Wang Y, Chen X, Zhou S, Li Q. Steroids. 2016;105:96.