Discovery of orally efficacious RORγt inverse agonists, part 1: Identification of novel phenylglycinamides as lead scaffolds

Article abstract of DOI:10.1016/j.bmc.2017.12.006

A series of novel phenylglycinamides as retinoic acid receptor-related orphan receptor-gamma t (RORγt) inverse agonists were discovered through optimization of a high-throughput screen hit 1. (R)-N-(2-((3,5-Difluoro-4-(trimethylsilyl)phenyl) amino)-1-(4-m

Full text of DOI:10.1016/j.bmc.2017.12.006

Accepted Manuscript
Discovery of orally efficacious RORγt inverse agonists, Part 1: Identification
of novel phenylglycinamides as lead scaffolds
Junya Shirai, Yoshihide Tomata, Mitsunori Kono, Atsuko Ochida, Yoshiyuki
Fukase, Ayumu Sato, Shinichi Masada, Tetsuji Kawamoto, Kazuko Yonemori,
Ryoukichi Koyama, Hideyuki Nakagawa, Masaharu Nakayama, Keiko Uga,
Akira Shibata, Keiko Koga, Toshitake Okui, Mikio Shirasaki, Robert Skene,
BiChing Sang, Isaac Hoffman, Wes Lane, Yasushi Fujitani, Masashi Yamasaki,
Satoshi Yamamoto
PII:
S0968-0896(17)31989-2
https://doi.org/10.1016/j.bmc.2017.12.006
BMC 14111
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
11 October 2017
30 November 2017
3 December 2017
Please cite this article as: Shirai, J., Tomata, Y., Kono, M., Ochida, A., Fukase, Y., Sato, A., Masada, S., Kawamoto,
T., Yonemori, K., Koyama, R., Nakagawa, H., Nakayama, M., Uga, K., Shibata, A., Koga, K., Okui, T., Shirasaki,
M., Skene, R., Sang, B., Hoffman, I., Lane, W., Fujitani, Y., Yamasaki, M., Yamamoto, S., Discovery of orally
efficacious RORγt inverse agonists, Part 1: Identification of novel phenylglycinamides as lead scaffolds, Bioorganic
&
Medicinal Chemistry (2017), doi: https://doi.org/10.1016/j.bmc.2017.12.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Discovery of orally efficacious RORγt inverse agonists, Part 1: Identification of
novel phenylglycinamides as lead scaffolds
a,*
a
a
a
Junya Shirai , Yoshihide Tomata , Mitsunori Kono , Atsuko Ochida , Yoshiyuki
a,c
a,d
a
a,d
a
Fukase , Ayumu Sato , Shinichi Masada , Tetsuji Kawamoto , Kazuko Yonemori ,
a
a
a
a,d
Ryoukichi Koyama , Hideyuki Nakagawa , Masaharu Nakayama , Keiko Uga , Akira
a
a
a
a,d
b
Shibata , Keiko Koga , Toshitake Okui , Mikio Shirasaki , Robert Skene , BiChing
b
b
b
a
a,d
Sang , Isaac Hoffman , Wes Lane , Yasushi Fujitani , Masashi Yamasaki , and Satoshi
a
Yamamoto
a
Pharmaceutical Research Division, Takeda Pharmaceutical Co., Ltd., Shonan
Research Center, 26-1, Muraoka-Higashi 2-chome, Fujisawa, Kanagawa 251-8555,
Japan
b
Takeda California, 10410 Science Center Drive, San Diego, California 92121, United
States
Abstract
A series of novel phenylglycinamides as retinoic acid receptor-related orphan
receptor-gamma t (RORγt) inverse agonists were discovered through optimization of a
high-throughput screen hit 1.
amino)-1-(4-methoxyphenyl)-2-oxoethyl)-3-hydroxy-N-methylisoxazole-5-carboxamide
22) was identified as one of the best of these compounds. It displayed higher subtype
(R)-N-(2-((3,5-Difluoro-4-(trimethylsilyl)phenyl)
(
selectivity and specificity over other nuclear receptors and demonstrated in vivo
potency to suppress the transcriptional activity of RORγt in a mouse PD
(
pharmacodynamic) model upon oral administration.
Key words: Retinoic acid receptor-related orphan receptor-gamma t (RORγt), Inverse
agonist, Autoimmune disease, Th17, IL-17, Phenylglycinamide
*
Corresponding author at: Cardurion Pharmaceuticals K.K. c/o Shonan Research Center,
Takeda Pharmaceutical Company Ltd. 26-1, Muraoka-Higashi 2-chome Fujisawa,
Kanagawa 251-8555, Japan
1

E-mail address: junya.shirai@cardurion.com (J. Shirai)
c
Present address: Tri-Institutional Therapeutics Discovery Institute, Inc., 413 East 69th
Street, New York, NY 10021, U.S.A.
d
Present address: Axcelead Drug Discovery Partners, Inc., 26-1, Muraoka-Higashi
2
-chome, Fujisawa, Kanagawa 251-0012, Japan
2

1
. Introduction
The modulation of a nuclear receptor (NR) is of pharmaceutical interest because NRs
are deeply involved in the regulation of cell life and death processes such as
1
homeostasis and metabolism.
The retinoic acid receptor-related orphan receptor
(
ROR) is a subclass of NRs, and the three subtypes are known as ROR, ROR and
2
ROR. In particular, RORγt (the second isoform of , and highly restricted to the
thymus) is a key transcription factor and a master regulator of differentiation and
+
activation of human Th17 (T helper 17) cells, which are a subset of CD4 T cells,
3
essential in the immune response. The Th17 cell itself produces proinflammatory
cytokines (IL-17A, IL-17F, etc.) that play crucial roles in the immunopathology of
human autoimmune diseases such as inflammatory bowel disease (IBD), psoriasis,
multiple sclerosis (MS), rheumatoid arthritis (RA), chronic obstructive pulmonary
4
disease (COPD), and asthma. Therefore, inhibiting transcription of RORγt could be a
promising new therapeutic option, replacing existing immunosuppressive agents for the
treatment autoimmune diseases.
RORγt is classified as an orphan receptor because its endogenous ligands are still
unknown. Some small molecules have been shown to bind to RORγt (Figure 1).
Cholesterol and cholesterol derivatives show affinity for RORγt and act as agonists to
5
enhance gene transcription. In contrast, digoxin, a steroid glycoside purified from the
foxglove plant, inhibits RORγt transcriptional activity, suppresses murine Th17 cell
6
differentiation, and reduces the severity of autoimmune disease in mice. Although
digoxin seems to have the attractive biological attribute of suppressing the
transcriptional activity of RORγt, it also has Na+/K+ ATPase inhibitory activity and is
used clinically only for control of arrhythmia due to its rather narrow safety window.
Thus, synthetic small molecules (selective RORγt inverse agonists) are highly expected
to selectively inhibit the transcriptional activity of RORγt and be effective against
Th17-driven autoimmune diseases at clinically safe doses.
3

Figure 1. Chemical structure of reported RORγt ligands (agonist/inverse agonist)
Under these circumstances, several small molecules have been disclosed recently in
7
public as potential RORγt inhibitors (inverse agonists) in vitro and in vivo. Some of
them are entering the development phase of clinical study, including VTP-43742
8
(
autoimmune disease and psoriasis in phase 2) from Vitae Pharmaceuticals Inc,
9
JNJ-3534 (psoriasis in phase 1) from Phenex Pharmaceuticals AG/Janssen Biotech Inc,
ARN-6039 (multiple sclerosis and psoriasis in phase 1) from Arrien Pharmaceuticals
1
0
LLC, AZD-0284 (psoriasis and respiratory disease in phase 1) from AstraZeneca plc,
and JTE-451 (allergy, autoimmune disease and psoriasis in phase 1) from Japan Tabaco
11
Inc/Orphagen Pharmaceuticals (only some of the structures have been disclosed).
Since we believed significant unmet needs still exist in the treatment of Th17-driven
autoimmune diseases, we initiated a program to explore novel RORγt inverse agonists.
Herein, we report the discovery of a series of novel phenylglycinamides as potent,
orally active, selective RORγt inverse agonists with properties confirmed
pharmacologically in an in vivo model.
A high-throughput screening (HTS) of our compound library was performed using a
binding assay that measured the amount of fluorescence-labeled cholesterol
(
BODIPY-cholesterol) displacement from the cholesterol-depleted RORγt protein (apo
1
2
RORγt) in the presence of test compounds.
A phenylglycinamide 1 (a purchased
compound, racemic) was identified as a moderate binder (IC50 = 200 nM) by the
BODIPY-cholesterol
displacement
assay,
and
also
displayed
3
[
(
H ]25-hydroxycholesterol displacement from apo RORγt (93% inhibition at 10 μM)
2
1
3
Figure 2).
This compound was then tested in a reporter gene assay to confirm its
4

functional activity (i.e., its ability to suppress the transcriptional activity of RORγt in
human Jurkat cells). We found that 1 actually decreased the transcriptional activity of
RORγt (IC50 = 1,600 nM). Although the functional activity of 1 was weak, subtype
selectivity against ROR and ROR was seen in the reporter gene assays (> 15-fold for
both subtypes). Therefore, we were confident that phenylglycinamide 1 was a genuine
hit. Sufficient subtype selectivity at a starting point could be advantageous for us to
conduct optimization toward identification of selective inverse agonists that cause
efficient immunosuppression with pharmacological safety profiles.
The phenylglycinamide 1 had a dipeptide structure, which we thought could help
establish a structure-activity relationship (SAR) quickly by making use of peptide
synthesis chemistry (i.e., amidation, hydrolysis, protection, and deprotection). And
also 1 had the shape of a three-pronged trident; thus we had enough spatial SAR
information to attempt improvement of its binding to the protein by changing the ends
and the central core, even if we had no information on the structure of the complex with
RORγt protein. In our initial SAR study for 1, we divided the molecule into three
parts (region A, B, and C) and investigated the stereochemical requirements at the
central carbon. Through these efforts, we examined whether one of the best
compounds for in vivo testing could suppress the transcriptional activity of RORγt.
Figure 2. Chemical structure of a lead compound 1 obtained from HTS campaign, its
biological profiles and SAR study plan
5

2
2
. Results and Discussion
.1. Initial SAR study of a phenylglycinamide lead 1
The compounds synthesized were initially evaluated using in vitro binding and
reporter gene assays similar to those used in the HTS campaign. In the binding assay
cell-free system), BODIPY-cholesterol combined with time-resolved fluorescence
(
resonance energy transfer (TR-FRET) technology was used for evaluation of the affinity
of the test compounds for human RORγt. In the cell-based reporter gene assay, the
ability of the test compounds to affect the transcriptional activity of RORγt was
assessed in human Jurkat cells by measuring luciferase activity. For key compounds,
the selectivity of ROR subtypes ( and ) was determined by the reporter gene assay in
human Jurkat cells, and specificity of other subclass NRs was confirmed by Invitrogen
®
SelectScreen Cell-based Nuclear Receptor Profiling Service using 20 kinds of
receptors in antagonist mode. Cassette dosing pharmacokinetic (PK) studies in mice
were carried out for evaluation of oral availability of the selected compounds by
measuring AUC0-8h values at 1 mg/kg.
First, we modified the left-side (region A) of the lead 1 compound, through changing
the isopropyl group on the phenyl ring (Table 1). We found that bulkiness at this
position clearly affected both the binding and reporter gene activity (1–3) and that
adding a large lipophilic substituent such as tert-butyl had the most favorable impact.
Introducing a chlorine atom adjacent to the tert-butyl resulted in acceptable binding and
reporter gene activity (4). Hence we were much interested in a bicyclic system with a
large substituent, and we examined dimethylindan 5 and confirmed that it had the same
strong affinity and functional activity as observed in 4. Furthermore, we were curious
about effects of increasing bulkiness at the para-position on activity and replaced the
tert-butyl group of 3 with a trimethylsilyl group, leading to the discovery of 6 with
1
4
equipotent binding activity and significant enhancement of the functional activity.
The increase of lipophilicity from tert-butyl to trimethylsilyl group could contribute this
functional activity enhancement. Hashimoto’s group recently reported similarly that
silicon-containing compounds showed increased ROR-inhibitory activity compared to
1
5
their original carbon compound counterparts.
From these results, we concluded that
1
the substituent R on the phenyl group is critical for in vitro activity, and that high
6

lipophilicity and large bulkiness would be desirable for strong binding and reporter gene
activity.
Table 1. SAR at left-side
b
c
Binding
Reporter gene
IC50 (nM)
a
Structure
Compound
No.
1
IC50 (nM)
R
d
d
[
95%CI]
[95%CI]
2
00
1,600
1
[
160–250]
[1,200–2,200]
(
lead compound)
6
,600
2
3
>30,000
[
1,500–7,900]
1
00
270
[
[
98–104]
[210–330]
7
5
120
4
5
45–130]
[84–170]
8
4
190
[
64–110]
[140–250]
7

1
00
35
6
[
47–220]
[24–50]
a
b
c
d
All of the compounds are racemic.
Displacement assay of BODIPY-cholesterol with apo RORγt
Inhibition of RORγt transcriptional activity in human Jurkat cells
9
5% confidence intervals (CI) in duplicate (n = 2)
Next, we moved onto an SAR investigation of the central core scaffold (region B).
Since we found that the trimethylsilyl compound 6 was unstable under strong acidic
conditions (i.e., 4 M HCl in EtOAc), we thought difficulties might be encountered
during some chemical reactions (i.e., hydrolysis, deprotection) to obtain other
trimethylsilyl derivatives. Thus, we used the tert-butyl compound 3 as a reference
compound for this purpose. We inserted a carbon atom at the central methine of 3 and
modified the methyl on the right amide nitrogen as well (Table 2). The carbon
insertion clearly decreased both the affinity and the functional activity (3 vs 7, 8 and 9),
which implied that the relative position of each end of the trident structure in space
might be important for RORγt affinity. In contrast, we were interested to find that
similar activities in 10 and 11 could be achieved by modifying the methyl group on the
amide nitrogen of 3. These transformations may alter thermodynamically stable
structures around the right amide moiety (E/Z rotational isomerization) and may affect
in vitro activity. Therefore, we supposed that the pocket of RORγt protein might
loosely recognize the right amide moiety of phenylglycinamide compounds. On the
other hand, we observed a strict limitation on left-side amide modification. Compound
1
2 having a methyl group on the nitrogen of the left-side amide was significantly less
active than compound 3. In addition, introduction of a methyl group to the 2'-position
on the central phenyl ring reduced activity relative to that of 3 (3 vs 13). These results
indicated compounds 12 and 13 had a less stable conformation than 3 presumably due
to steric repulsion caused by the additional methyl group. On the basis of the SAR
around the central core, we supposed that the RORγt protein recognizes the spatial
8

structure of the phenylglycinamides strictly and this information could enable us to
generate a new scaffold from that of phenylglycinamides. Further investigations
1
6
regarding scaffold generation will be reported in the upcoming publication.
Table 2. Effects by central core changes
b
c
Binding
Reporter gene
IC50 (nM)
a
Compound
No.
Core
IC50 (nM)
d
d
[
95%CI]
[95%CI]
1
00
270
3
7
8
[
97–104]
[210–330]
>10,000
>30,000
>30,000
7
,200
[
4,200–12,000]
9

9
>10,000
>30,000
9
9
210
1
0
[
78–120]
[150–280]
2
30
370
11
[
130–410]
[250–540]
1
1
2
3
>10,000
>30,000
5
,700
21,000
[
4,200–7,700]
[11,000–41,000]
a
b
c
d
All of the compounds are racemic.
Displacement assay of BODIPY-cholesterol with apo RORγt
Inhibition of RORγt transcriptional activity in human Jurkat cells
9
5% confidence intervals (CI) in duplicate (n = 2)
1
0

Figure 3. Summary of SAR from initial modification of the phenylglycinamide
lead 1
With basic SAR on the phenylglycinamides for the left and the center core moieties
in hand (Figure 3), we proceeded to optimize the pyridone ring (region C) to improve
both the compound’s in vitro potency and PK profile and ultimately show its in vivo
potency in a PD model. Extensive explorations of the pyridone ring of 3 and 6 by
high-throughput organic synthesis resulted in discovery of 5-membered heterocyclic
rings such as furyl and isoxazolyl rings with acceptable binding and functional ability
(
14, 15, 17, 18). We confirmed that the TMS-anilides had consistently better binding
and functional activity than the tert-butyl-anilides in three types of the heterocycles (3
vs 6, 14 vs 17 and 15 vs 18). Since the isoxazolyl compounds (15 and 18) seemed to
have better activity compared with the furyl ones (14 and 17), a tautomeric hydroxyl
group was introduced onto the isoxazolyl ring like the pyridone ring. A slight activity
enhancement was observed by this modification. Interestingly, PK study in mice
revealed that the plasma exposure AUC0-8h was very large for compound 19 (1.59 g･
h/mL at 1 mg/kg), but not so high for compound 6 (AUC0-8h: 0.086 g･h/mL at 1
mg/kg). These results encouraged us to investigate the hydroxyisoxazoles for further
optimization.
11

Table 3. SAR at right-side in comparison with t-Bu- and TMS-anilide analogs
b
c
Binding
Reporter gene
a
Compound
No.
Structure
IC (nM)
IC (nM)
5
0
50
d
d
[
[
[
95%CI]
[95%CI]
1
00
97–104]
00
47–230]
00
180–480]
60
100–240]
70
120–240]
10
65–180]
270
[210–330]
35
3
6
1
[24–50]
1,100
3
1
1
1
1
4
7
5
8
[
[
[
[730–1,600]
190
1
[58–640]
870
1
[590–1,300]
73
1
[
[22–240]
1
2

5
7
540
[320–930]
48
1
1
6
9
[
47–69]
6
0
[
34–110]
[17–130]
a
b
c
d
All of the compounds are racemic.
Displacement assay of BODIPY-cholesterol with apo RORγt
Inhibition of RORγt transcriptional activity in human Jurkat cells
9
5% confidence intervals (CI) in duplicate (n = 2)
To identify a compound with potent in vivo agonist activity, we need to enhance
functional activity. We introduced two fluorine atoms adjacent to the TMS group
because lipophilicity and bulkiness in region A would favor greater pharmacological
activity as described previously.
Eventually, we saw a slightly increased
pharmacological activity in 20 with high AUC0-8h value like that of 19 (Table 4).
Finally, we examined each optical enantiomer of 20 and identified 22 with
(
R)-configuration as a desired stereoisomer (eutomer). Compound 22 was found to be
highly selective for ROR subtypes (>192-fold for  and ) and very specific for other
®
17
NRs according to the results of Invitrogen SelectScreen (at least >250-fold).
Compound 22 also displayed acceptable PK profiles, thus further studies were carried
out to see the inhibitory effects of RORγt on transcriptional activity in vivo.
1
3

Table 4. TMS-anilides with a hydroxyisoxazole
Reporter
a
c
Binding
Mouse PK
b
a
gene
Compound
No.
Stereo-
Structure
IC50 (nM)
IC50 (nM)
AUC0-8h
chemistry
d
d
[
95%CI]
[95%CI]
(μg·h/mL)
6
0
48
19
racemic
1.59
[
34–110]
[17–130]
5
7
17
2
2
2
0
1
2
racemic
1.51
[
30–110]
[12–26]
3
,400
4
20
e
(S)
NT
[
1,300–9,00
[
290–600]
0
]
1
8
13
(R)
0.37
[
9.7–32]
[10–16]
a
b
c
d
e
Displacement assay of BODIPY-cholesterol with apo RORγt
Inhibition of RORγt transcriptional activity in human Jurkat cells
The values shown as means of three determinations (1 mg/kg, po)
9
5% confidence intervals (CI) in duplicate (n = 2)
Not tested
1
4

2
.2. In vivo RORγt inhibition (PD study)
With 22 (an orally available, in vitro potent and selective RORγt inverse agonist) in
hand, we used a mouse PD model to assess its potency to inhibit the transcriptional
activity of RORγt in vivo. IL-23 injection in the ear of mice increases IL-17a gene
1
8
expression.
PD model mice were given 22 orally at a dose range of 10 to 100 mg/kg
and changes in their IL-17a gene expression response to IL-23 injection in the ear were
monitored by real-time RT-PCR. We found that 22 significantly suppressed the
increase of the gene expression at 100 mg/kg compared with the vehicle control (Figure
4
). We also measured PK profiles in mice at the effective dose of 100 mg/kg to make
sure there was sufficient plasma exposure. Table 5 and Figure 5 show
pharmacokinetic parameters and a plasma drug concentration versus time curve, which
suggested that the compound concentration in the plasma was certainly high enough to
inhibit completely transcriptional activity of RORγt in the reporter gene assay over 8
hours. We assumed that the distribution of the compound between the plasma and
tissues (i.e., the ear) would be important to understand the result of this PD study. We
observed that the level of the compound in the ear was almost proportional to its plasma
concentration, increased dose-dependently and time-dependently to at least the level
needed for complete inhibition, and remained at that concentration for 8 hours in the
reporter gene assay (data not shown). Taking into consideration the biological results
and the PK profiles, we conclude that enough 22 reached the ear to suppress
IL-23-induced IL-17a gene expression enhancement through inhibiting the
transcriptional activity of RORγt. Therefore, our phenylglycinamides could become
an orally available and potentially therapeutic agent for treating Th17-driven
autoimmune diseases through RORγt inhibition, and also a valuable tool to understand
RORγt biology.
1
5

7
1% inhibition
Figure 4. In vivo inhibitory effect of compound 22 on IL-23-induced IL-17a gene
expression in mice. ##p<0.01 (Aspin-welch test) compared with PBS-treated group (N:
normal group), p<0.025 (one-tailed Williams' test) compared to the IL-23-injected and
vehicle-administered group (V: vehicle group).
1
6

Table 5 and Figure 5. Pharmacokinetic Parameters in Plasma of Compound 22 orally
administered to mice at 100 mg/kg
a
PK parameter of compound 22 in mice upon 100 mg/kg, po
Cmax
Tmax
(h)
AUC0-24h
(
μg/mL)
(
μg·h/mL)
In plasma
9.12
1.0
47.4
a
The values shown as means of three determinations
1
7

2
.3 Crystal structure of a eutomer of 19 with C-terminal truncated RORγt
During medicinal chemistry optimization, we were able to obtain a crystal of
C-terminal truncated apo RORγt protein. We tried ligand-soaking experiments to
increase our understanding of the interactions between phenylglycinamide derivatives
and the protein, and thereby facilitate creation of a new scaffold from this series in the
future. One of the complex structures, 19, a eutomer that binds to RORγt-LBD, is
1
9
shown in Figure 6.
Three key interactions were detected in the crystal: 1) two amide
groups formed hydrogen bonds with the backbones of Phe377 and Glu379; 2) two
phenyl groups appeared to interact with Met365 through CH/π contacts; and 3) the
trimethylsilylphenyl group extended toward the hydrophobic residues, Phe388, Ile397,
and Ile400. From our SAR study, we recognized the importance of the central
methoxyphenyl group, the strict modification of the left amide NH and the significant
impacts on in vitro potency enhancement by increasing lipophilicity and bulkiness on
the left-side of the molecule. The crystal structure seems to be consistent with the
SAR information, and the key interactions tell us how to facilitate our medicinal
chemistry optimization of RORγt-specific ligands going forward.
1
8

Figure 6. The X-ray of a complex structure of the human C-terminal truncated
RORγt-LBD (gray) with a eutomer of compound 19 (green)
1
9

2
.4 Chemistry
The target compounds were synthesized by utilizing basic reactions in peptide
synthesis chemistry, which include amidation, esterification, hydrolysis, and
protection/de-protection. There were three targets: the left aniline, the central core,
and the right carboxylic acid (Figure 2). Most of the left aniline moieties were
prepared by known procedures, and some special aniline derivatives were prepared as
below. As an example, 3,5-difluoro-4-trimethylsilylaniline 25 was prepared by
introduction of a TMS group to 3,5-difluoronitrobenzene 23 under basic conditions
followed by reduction of the nitro group by standard hydrogenation (Scheme 1).
Scheme 1. Synthesis of an intermediates 25. Reagents and conditions: (a) TMSCl,
NaHMDS, THF, –78 °C, 80%; (b) H (1 atm), 10% Pd-C (50% wet), MeOH, 82%.
2
The central core of the phenylglycine skeleton was constructed by the Petasis
2
0
reaction.
The phenylglycinamides 2–6 were synthesized as described in Scheme 2,
where N-allyl phenylglycine 27 was prepared with three components in the first step,
followed by sequential introduction of the right and left-side moieties with
accompanying protection and de-protection of both ends. In the preparation of 12
2
1
having a methyl on the left amide, 4-(tert-butyl)-N-methyl-aniline was used for
coupling with carboxylic acid 31. This coupling needed higher temperature (60 °C) to
complete the reaction compared with the synthesis of 2–6 (room temperature).
2
0

Scheme 2. Synthesis of phenylglycinamides 2–6 and 12. Reagents and conditions: (a)
glyoxylic acid monohydrate, N-methylallylamine, MeCN, 60 oC, 92%; (b)
trimethylsilyldiazomethane, MeOH, 96%; (c) 1,3-dimethylbarbituric acid, Pd(PPh
3
4
) ,
2
1

THF, 88%; (d) 6-hydroxynicotinic acid, DIEA, HATU, DMF, 60%; (e) 1M NaOH,
MeOH, THF, acidic workup, 78%; (f) anilines, DIEA, HATU, DMF, 11–72%; (g)
o
4
-(tert-butyl)-N-methylaniline, DIEA, HATU, DMF, 60 C, 67%.
Other phenylglycinamides 13–20 having a variety of right-side moieties were
synthesized in similar ways (Scheme 3). The left-side aniline was firstly introduced to
the central core, followed by removal of the allyl group and coupling with the right-side
carboxylic acid. The compound possessing a methyl at the 2'-position of the central
phenyl ring was also prepared by this method.
Scheme 3. Synthesis of phenylglycinamides 13–20. Reagents and conditions: (a)
2
2

o
glyoxylic acid monohydrate, N-methylallylamine, MeCN, 60 C, 99%; (b)
-(tert-butyl)aniline or 4-(trimethylsilyl)aniline or 25, TEA or DIEA, HATU or T3P,
DMF, 57–62%; (c) thiosalicylic acid, Pd (dba) , THF, quant. or 1,3-dimethylbarbituric
acid, Pd(PPh , THF, 66–100%.; (d) carboxylic acids, TEA, HATU, DMF, 19–64%.
4
2
3
)
3 4
Three types of the compounds 7–9 were synthesized by inserting a carbon atom
around the central methine of the phenylglycinamide 3 using the corresponding known
amino acid derivative as starting material (Scheme 4–6). For preparation of
acetyltyrosinamide 7, hydrolysis of methyl ester 32, followed by coupling with
4
-(tert-butyl)aniline to the left-side moiety, removal of the N-Boc group, and
condensation with 6-hydroxynicotinic acid gave the target compound (Scheme 4).
o
Scheme 4. Synthesis of 7. Reagents and conditions: (a) NaH, MeI, THF, 0 C then 1M
NaOH, acidic workup, 67%; (b) 4-(tert-butyl)aniline, TEA, HATU, DMF, 89%; (c) 4M
HCl in EtOAc, EtOAc, basic workup, quant.; (d) 6-hydroxynicotinic acid, TEA, HATU,
DMF, 42%.
For synthesis of compound 8, attachment of the left-side moiety to 36 was conducted
first, followed by de-protection to provide free amine 38, introduction of a methyl group
2
3

to the free amine by three steps to give 39, and coupling with the right-side carboxylic
acid (Scheme 5).
Scheme 5. Synthesis of 8. Reagents and conditions: (a) 4-(tert-butyl)aniline, TEA,
o
HATU, DMF, quant.; (b) hydrazine monohydrate, EtOH, 80 C, 77%; (c) TFAA, TEA,
o
THF, 5 C, 93%; (d) MeI, K
2
CO , DMF, 82%; (e) 2M NaOH, EtOH, 82% (crude); (f)
3
6
-hydroxynicotinic acid, TEA, HATU, DMF, 53%.
Starting with a known N-protected amino acid 40, compound 9 was also synthesized
by N-methylation and sequential installation at both ends (Scheme 6).
2
4

Scheme 6. Synthesis of 9. Reagents and conditions: (a) NaH, MeI, acidic workup,
0%; (b) 4-(tert-butyl)aniline, TEA, HATU, DMF, 77%; (c) 4M HCl in EtOAc, EtOAc,
9
basic workup, 99%; (d) 6-hydroxynicotinic acid, TEA, HATU, DMF, 65%.
To change substituents on the nitrogen of the right-side amide, compound 10 without
alkyl substituents on the right-side amide was synthesized by coupling of
6
-hydroxynicotinic acid with 46, which was prepared from 44 by aniline introduction
and removal of the Boc group. Compound 11 having an n-propyl group was obtained
by reductive alkylation of 46 and then coupling to the right-side moiety (Scheme 7).
2
5

Scheme 7. Synthesis of 10 and 11. Reagents and conditions: (a) 4-(tert-butyl)aniline,
TEA, HATU, DMF, quant.; (b) 4M HCl in EtOAc, EtOAc, basic workup, 67%; (c)
6
-hydroxynicotinic acid, TEA, HATU, DMF, 81%; (d) propionaldehyde, NaBH(OAc)
3
,
THF, 83%; (e) 6-hydroxynicotinic acid, TEA, HATU, DMF, 27%.
2
6

As shown in Scheme 8, racemic compound 20 was separated into chiral
stereoisomers (including 21 [distomer] with shorter retention time and 22 [eutomer]
with longer retention time) by preparative HPLC using a CHIRALPAK IA column.
Their absolute stereochemistry was determined as follows. A chiral intermediate 60
was prepared from a racemic 54 in 3 steps. Boc-protection of the racemic 54 followed
by optical resolution with chiral HPLC afforded (+)- or (–)-59, which was treated under
Boc-deprotection conditions to give the chiral 60.
Acylation of 60 with
p-bromobenzoylchloride provided 61 whose absolute stereochemistry was determined
as (R)-configuration by single X-ray crystallography (Figure 7). Condensation of 60
with the right-side carboxylic acid yielded a product with no loss of optical purity. The
product was identical to 22 on chiral HPLC analysis. These experiments suggested
that compound 22 should have an (R)-configuration and be a RORγt inverse agonist
eutomer.
Scheme 8. Preparation of chiral phenylglycinamides 21 and 22, and their absolute
stereochemistry determination. Reagents and conditions: (a) preparative HPLC by chirai
HPLC, >99%ee for 21 and 22; (b) Boc O, TEA, THF, 99%; (c) preparative chiral HPLC,
2
2
7

5
0% (>99%ee); (d) TFA then basic workup, 98%; (e) p-bromobenzoylchoride, TEA,
THF, 91%; (f) 3-hydroxyisoxazole-5-carboxylic acid, HATU, DIEA, DMF, 93%
99%ee).
(
Figure 7. ORTEP of 61, thermal ellipsoids are drawn at 50% probability.
3
. Conclusions
We discovered a series of novel phenylglycinamides with RORγt inverse agonistic
activity by optimization of an HTS hit 1 as lead compound. Although we had no
structural information about RORγt protein at the beginning of this study, we were able
to obtain a highly potent, orally active, and selective inverse agonist 22 through
examination of the three parts of 1 by using peptide synthesis chemistry. Compound
2
2 suppressed the in vivo transcriptional activity of RORγt in mice, and we believe 22
could become an oral therapeutic agent for treating Th17-driven autoimmune diseases
and a valuable tool to understand RORγt biology. We also succeeded in identifying the
types of interactions of the RORγt complex crystal structure with compounds in this
series, and will be able to utilize this information for further medicinal chemistry
development. Further investigation of compound 22 continues to increase its potency
and these results will be reported in due course.
2
8

4
4
. Experimental
.1. Chemistry
1
H NMR spectra were recorded on a Bruker Ultra Shield-300 (300 MHz)
spectrometer. Chemical shifts are given in δ values (ppm) using tetramethylsilane as
an internal standard. Abbreviations are used as follows: s = singlet, d = doublet, t =
triplet, q = quartet, m = multiplet, dd = doublets of doublet, br = broad, br.s. = broad
singlet. Coupling constants (J values) are given in hertz (Hz).
LC−MS analysis was performed on a Shimadzu LC-20AD separations module or
Agilent 1200 series, operating in ESI (+ or −) ionization mode. Analytes were eluted
using a linear gradient of 0.05% TFA containing H
O/CH CN or 5 mM ammonium acetate containing H
containing H O/CH CN or 0.01% heptafluorobutyric acid/1.0% isopropyl alcohol
containing H O/CH CN mobile phase.
2
O/CH
3
CN or 0.01% TFA containing
H
2
3
2
O/CH
3
CN or 10 mM NH HCO
4
3
2
3
2
3
Reactions were followed by TLC on Silica gel 60 F 254 precoated TLC plates (E.
Merck) or NH TLC plates (Fuji Silysia Chemical Ltd.). Purification by column
chromatography was performed on an SiO
or an NH-SiO column (Chromatorex® NH, Fuji Silysia Chemical Ltd.) or a Diol-SiO
column (Chromatorex® DIOL, Fuji Silysia Chemical Ltd.) using the indicated eluents.
2
column (Purif-Pack, SHOKO SCIENTIFIC)
2
2
Purification by preparative HPLC was performed on a Waters 2525 separations
module (L-column2 ODS (20 × 150 mm ID), CERI, Japan); MS spectra were recorded
using a Waters ZQ2000 with electrospray ionization or on a GILSON system, equipped
with a Lcolumn2 ODS (20 × 150 mm ID, CERI, Japan) or on a Waters Deltaprep 300
system or on a Shimadzu 10A VP system. Samples were eluted using a linear gradient
of 0.1% TFA in H
2
O/CH
3
CN or 10 mM NH
4
HCO
3
in H
2
O/CH
3
CN, or H
2
O/CH CN.
3
The purity of all test compounds was assessed by analytical HPLC. The data were
collected by a HPLC with Corona CAD (Charged Aerosol Detector), Nano quantity
analyte detector (NQAD), or photo diode array detector. The column was a Capcell Pak
C18AQ (50 mm × 3.0 mm ID, Shiseido, Japan) or L-column 2 ODS (30 mm × 2.0 mm
ID, CERI, Japan) with a temperature of 50 °C and a flow rate of 0.5 mL/min. Mobile
phases A and B under a neutral condition were a mixture of 50 mM Ammonium acetate,
H
2
O, and CH
3
CN (1:8:1, v/v/v) and a mixture of 50 mM ammonium acetate and
CH CN (1:9, v/v), respectively. The ratio of mobile phase B was increased linearly
3
2
9