The inhibition of the nuclear receptor retinoic-acid-receptor-related orphan receptor γt (RORγt) is a promising strategy in the treatment of autoimmune diseases. RORγt features an allosteric binding site within its ligand-binding domain that provides an opportunity to overcome drawbacks associated with orthosteric modulators. Recently, trisubstituted isoxazoles were identified as a novel class of allosteric RORγt inverse agonists. This chemotype offers new opportunities for optimization into selective and efficacious allosteric drug-like molecules. Here, we explore the structure-activity relationship profile of the isoxazole series utilizing a combination of structure-based design, X-ray crystallography, and biochemical assays. The initial lead isoxazole (FM26) was optimized, resulting in compounds with a ∼10-fold increase in potency (low nM), significant cellular activity, promising pharmacokinetic properties, and a good selectivity profile over the peroxisome-proliferated-activated receptor γ and the farnesoid X receptor. We envisage that this work will serve as a platform for the accelerated development of isoxazoles and other novel chemotypes for the effective allosteric targeting of RORγt.
The inhibition of the nuclear receptor retinoic-acid-receptor-related orphan receptor γt (RORγt) is a promising strategy in the treatment of autoimmune diseases. RORγt features an allosteric binding site within its ligand-binding domain that provides an opportunity to overcome drawbacks associated with orthosteric modulators. Recently, trisubstituted isoxazoles were identified as a novel class of allosteric RORγt inverse agonists. This chemotype offers new opportunities for optimization into selective and efficacious allosteric drug-like molecules. Here, we explore the structure-activity relationship profile of the isoxazole series utilizing a combination of structure-based design, X-ray crystallography, and biochemical assays. The initial lead isoxazole (FM26) was optimized, resulting in compounds with a ∼10-fold increase in potency (low nM), significant cellular activity, promising pharmacokinetic properties, and a good selectivity profile over the peroxisome-proliferated-activated receptor γ and the farnesoid X receptor. We envisage that this work will serve as a platform for the accelerated development of isoxazoles and other novel chemotypes for the effective allosteric targeting of RORγt.
Nuclear receptors (NRs) are a family of ligand-dependent transcription
factors and attractive drug targets because of their central role
in several regulatory processes in the human body.[1−4] Within the NR family, the retinoic-acid-receptor-related
orphan receptor γt (RORγt) has received increased attention
because of its essential role in the immune system.[5] RORγt is a key regulator in the differentiation of
naïve CD4+ T cells into T helper 17 (Th17)
cells, and the production of the pro-inflammatory cytokine IL-17a.[5−7] Elevated IL-17a levels are associated with the development of autoimmune
diseases, including psoriasis, multiple sclerosis, and rheumatoid
arthritis.[8−11] Disrupting the Th17/IL-17a pathway shows high potential for the
treatment of autoimmune disorders, which has already been validated
by the clinical success of anti-IL-17a monoclonal antibodies.[8,12,13] The inhibition of RORγt
could pose an attractive alternative strategy to decrease IL-17a production
in the treatment of autoimmune diseases.[12] Many research groups have shown significant interest in the identification
of small-molecule RORγt inhibitors (or, more specifically, inverse
agonists), with several synthetic compounds progressing into clinical
trials.[14−18]With the exception of RTA-1701 (Figure C),[18,19] all of the RORγt
inverse agonists that have entered clinical trials thus far likely
target the highly conserved ligand-binding pocket, termed the orthosteric-binding
site, within the ligand-binding domain (LBD) of RORγt (Figure A).[5,14] While orthosteric targeting has been highly successful, novel molecular
modalities with alternative modes of action, such as allosteric modulators,
offer an interesting alternative for targeting NRs and could circumvent
issues related to promiscuity between NRs.[20−23] Recently, an allosteric binding
site was identified in the LBD of RORγt (formed by helices 3,
4, 11, and 12), at a location that is topographically distinct to
the orthosteric site (Figure A).[24] The indazole MRL-871 (1) (Figure C) was identified as a prototypical allosteric ligand for
RORγt, acting as a potent inverse agonist and decreasing coactivator
binding with a similar efficacy as orthosteric inverse agonists.[24,25] The allosteric pocket of RORγt is unique within the NR family
and thought not to be the target of endogenous ligands.[24,26] Therapeutic compounds targeting the allosteric-binding pocket potentially
have a higher selectivity profile and do not act in competition, but
in synergy, with endogenous agonists.[20,21,27] Therefore, these compounds would be greatly beneficial
for drug discovery and chemical biology applications.
Figure 1
(A) Co-crystal structure
of the RORγt LBD in complex with
allosteric ligand FM26 (shown as orange sticks) (PDB: 6SAL),[28] where the allosteric site (orange circle) and the orthosteric
site (blue circle) are indicated; (B) enlarged view of FM26 (shown as orange sticks) in the allosteric pocket, where H-bond
interactions are indicated with orange dashes; (C) chemical structures
of allosteric RORγt ligands FM26 (2), MRL-871 (1), and RTA-1701; and (D) exploration of the hit-to-lead SAR profile of FM26.
(A) Co-crystal structure
of the RORγt LBD in complex with
allosteric ligand FM26 (shown as orange sticks) (PDB: 6SAL),[28] where the allosteric site (orange circle) and the orthosteric
site (blue circle) are indicated; (B) enlarged view of FM26 (shown as orange sticks) in the allosteric pocket, where H-bond
interactions are indicated with orange dashes; (C) chemical structures
of allosteric RORγt ligands FM26 (2), MRL-871 (1), and RTA-1701; and (D) exploration of the hit-to-lead SAR profile of FM26.Despite the high potential of
allosteric inverse agonists for RORγt,
the number of examples have remained limited to compounds based on
the indazole core of 1 or similar chemotypes, and the
previously mentioned allosteric ligand RTA-1701 (which
binds to RORγt via covalent attachment to Cys476
of helix 11 and partly overlaps with the allosteric site).[19,24,25,29−34] Follow-up studies improved the pharmacokinetic (PK) and selectivity
profile of 1 and its derivatives. However, off-target
effects toward other NRs, most notably toward the peroxisome proliferated-activated
receptor γ (PPARγ), and challenges regarding cytotoxicity
and metabolic stability have been observed.[24,29−32] The scarcity and lack of scaffold diversity of allosteric RORγt
ligands, and the limitations regarding specificity and drug-like properties,
necessitates further exploration of the allosteric pocket. Specifically,
establishing structure–activity relationship (SAR) profiles
for allosteric chemotypes distinct from the indazole-analogous ligand
classes is of great importance.Recently, we used an in silico pharmacophore search
approach to identify a novel class of RORγt allosteric inverse
agonists, featuring a trisubstituted isoxazole core.[28,35] SAR studies around this novel chemotype resulted in the discovery
of FM26 (2) (Figure C), which shows sub-micromolar potency as
a RORγt inverse agonist and significant inhibition of cellular
IL-17a expression levels. Initial SAR studies of 2 identified
two key pharmacophore features, where the 2,6-disubstituted phenyl
ring at the C-3 position and a C-4benzoic acid moiety were shown
to be optimal (Figure D).[24,25,28,30−33] Less knowledge has been garnered on the linker at
the C-4 position and the C-5 substituent. Regarding the C-4 linker,
an amine linker (as is used for 2) was found to be preferred
for potency over some more rigid linkers.[28] Additionally, the co-crystal structure (Figure B) showed that the presence of a hydrogen
bond donating N-heterocycle at the C-5 isoxazole
position (as is the case for the pyrrole in 2) significantly
increased the potency toward RORγt by the formation of an additional
polar interaction with the backbone carbonyls of residues Leu353 and
Lys354.[28] To identify compounds suitable
for focused lead optimization, crucial explorations around the C-4
and C-5 positions are needed to further explore the SAR, improve the
potency of the isoxazole series, increase the RORγt specificity,
and direct PK optimization.Utilizing X-ray crystallography
to guide focused library development,
herein, we report on the synthesis of allosteric RORγt inverse
agonists diversified at the C-4 and C-5 isoxazole positions (Figure D). The structure-based
SAR study was specifically focused on improved RORγt activity,
mitigating off-target PPARγ activity, and a first exploration
of PK profiles.
Results and Discussion
In Silico Docking Studies
Guide a SAR Study at the Isoxazole C-4 Position
The initial
investigation started with expanding the SAR data around the C-4 position
of the isoxazole series guided by in silico docking
studies. A virtual library of derivatives around 2, containing
different C-4 linkers and benzoic acid substituents, was designed
with a focus on diversity and synthetic feasibility (see Table S1). This library was docked into the allosteric
site of RORγt (PDB: 4YPQ(24)), using Glide, the molecular
docking tool in Schrödinger (12.3, 2020-1).[36,37] For each ligand, the docking pose was evaluated and given a “glide
score”, which is an empirical scoring function that approximates
the ligand binding free energy (the more negative the values, the
higher the expected potency) but will not always translate into experimental
IC50 values.[36,37]Table shows the compounds that resulted in the
most promising glide scores, indicating that linkers responsible for
an increased clog P of the compounds
[i.e., ether (3), thioether (4), methylated
amine (5), and alkene (6 and 7)] would be beneficial for affinity. Additionally, the data suggest
that a fluoro substitution at the ortho position
of the benzoic acid moiety would be well tolerated (8 and 9).
Table 1
SAR Studies around
the C-4 Isoxazole
Positiona
cLogP
values, glide scores, TR-FRET
IC50 values (μM) from coactivator recruitment and
AlexaFluor-MRL-871 displacement assays, ΔTm values (°C) from thermal shift assays, and fold decrease
in IL-17a mRNA expression levels relative to DMSO reverse transcriptase
PCR (RT-PCR). Abbreviations: n.d., not determined. TR-FRET and TSA
data are recorded in triplicate; values are representative of ≥3
repeated experiments. RT-PCR data are recorded in triplicate; values
are representative of ≥2 repeated experiments. cLogP values
were predicted using MarvinSketch (20.10).
cLogP
values, glide scores, TR-FRET
IC50 values (μM) from coactivator recruitment and
AlexaFluor-MRL-871 displacement assays, ΔTm values (°C) from thermal shift assays, and fold decrease
in IL-17a mRNA expression levels relative to DMSO reverse transcriptase
PCR (RT-PCR). Abbreviations: n.d., not determined. TR-FRET and TSA
data are recorded in triplicate; values are representative of ≥3
repeated experiments. RT-PCR data are recorded in triplicate; values
are representative of ≥2 repeated experiments. cLogP values
were predicted using MarvinSketch (20.10).
Optimized Synthesis Route Allowed the Efficient
Synthesis of C-4 Isoxazole Derivatives
Isoxazoles 2–9 (Table ) were synthesized
(Scheme ) in order
to biochemically evaluate the predictions from the in silico docking experiments. Common intermediate 14 was synthesized via a 1,3-dipolar cycloaddition using a nitrile oxide and
alkynyl bromide, as described previously (Scheme S1).[28] In order to prepare derivatives 2–9, it was important to obtain core intermediate 16 with a N-Boc-protected pyrrole moiety,
instead of the free pyrrole, as was the case in the original synthesis
route for 2 (Scheme S2).[28] This was achieved via a Suzuki
cross-coupling reaction, employing a dppf instead of a tetrakis palladium
catalyst (used in previous research), which proved to be essential
for maintaining the Boc-protected pyrrole 15 in an acceptable
yield (Scheme ). DIBAL
was then used to selectively reduce the ethyl ester, yielding 16 without a concomitant loss of the Boc group (Scheme ), which was previously observed
with LiAlH4 (Scheme S1).[28]
Scheme 1
Synthesis Route for Different Trisubstituted
Isoxazoles (C-4 Library)
Reagents and conditions: (a) N-Boc-pyrrole-B(pin), Pd(dppf)Cl2, Cs2CO3, and DME, 85 °C, 8 h, 50%; (b) DIBAL and CH2Cl2, −78 °C, 3 h, 80%; (c) (i) (MeSO2)2O, Et3N, and CH2Cl2, 0 °C → rt, 3 h, (ii) aniline or N-methylaniline, rt, 3 h, 36% (17) and 52% (18); (d) (i) DMP and CH2Cl2, rt, 3 h, 44%, (ii)
methyl 4-amino benzoate, MeOH, and AcOH, reflux, 24 h, 42%, (iii)
NaBH4 and EtOH, reflux, 3 h, 17% (19); (e)
LiOH, EtOH, and H2O, 95 °C, 3 h, 29–81%; (f)
methyl 4-hydroxybenzoate or methyl 2-fluoro-4-hydroxybenzoate, DIAD,
PPh3, Et3N, and THF, reflux, 3 h, 55% (20), 25% (21); (g) methyl 4-mercaptobenzoate,
DIAD, PPh3, Et3N, and THF, reflux, 3 h, 27%;
(h) DMP and CH2Cl2, rt, 3 h, quant.; (i) (4-(methoxycarbonyl)benzyl)triphenylphosphonium
(see Experimental Section for the synthetic
procedure of the triphenylphosphonium), LiHMDS, and THF, −78
°C → rt, 24 h, 7% (24), 27% (25).
Synthesis Route for Different Trisubstituted
Isoxazoles (C-4 Library)
Reagents and conditions: (a) N-Boc-pyrrole-B(pin), Pd(dppf)Cl2, Cs2CO3, and DME, 85 °C, 8 h, 50%; (b) DIBAL and CH2Cl2, −78 °C, 3 h, 80%; (c) (i) (MeSO2)2O, Et3N, and CH2Cl2, 0 °C → rt, 3 h, (ii) aniline or N-methylaniline, rt, 3 h, 36% (17) and 52% (18); (d) (i) DMP and CH2Cl2, rt, 3 h, 44%, (ii)
methyl 4-amino benzoate, MeOH, and AcOH, reflux, 24 h, 42%, (iii)
NaBH4 and EtOH, reflux, 3 h, 17% (19); (e)
LiOH, EtOH, and H2O, 95 °C, 3 h, 29–81%; (f)
methyl 4-hydroxybenzoate or methyl 2-fluoro-4-hydroxybenzoate, DIAD,
PPh3, Et3N, and THF, reflux, 3 h, 55% (20), 25% (21); (g) methyl 4-mercaptobenzoate,
DIAD, PPh3, Et3N, and THF, reflux, 3 h, 27%;
(h) DMP and CH2Cl2, rt, 3 h, quant.; (i) (4-(methoxycarbonyl)benzyl)triphenylphosphonium
(see Experimental Section for the synthetic
procedure of the triphenylphosphonium), LiHMDS, and THF, −78
°C → rt, 24 h, 7% (24), 27% (25).The primary alcohol of 16 was used as a functional
handle for the derivatization of the isoxazole C-4 position. Critical
to the synthesis of this series of isoxazoles was the optimization
of the reductive amination step, which provided sub-optimal yields
in the synthesis of 2 (Scheme S2).[28] In order to improve the total yield
of the synthesis, the reductive amination was substituted for a nucleophilic
substitution reaction. Final compounds 2 and 5 were accessed in an efficient manner by the mesylation of alcohol 16 with methane sulfonic anhydride (monitored by NMR), the in situ addition of the substituted aniline, followed by
the hydrolysis of the benzoate ester. In contrast, 8 (fluoro
substituent) was synthesized via the original reductive
amination route (Scheme ).Compounds with an ether and thioether linkage (3, 9, and 4) were synthesized from core
intermediate 16via a Mitsunobu reaction
with hydroxy-
or mercaptobenzoate, respectively, furnishing the products in acceptable
yields (Scheme ).
For the cis and trans alkene linker
(6 and 7), the alcohol of compound 16 was oxidized to the aldehyde, which was subsequently used
in a Wittig reaction, obtaining a 3:1 mixture of the cis/trans isomers (Scheme ). These were separated via preparative high-performance liquid chromatography/ultraviolet (HPLC-UV)
to afford 24 and 25, and followed by ester
hydrolysis, which afforded final compounds 6 (trans) and 7 (cis) (for which
the stereochemistry was determined based on the relative 1H NMR J-coupling values between the alkene protons).
Biochemical Assays and X-ray Crystallography
Reveal a Positive Correlation between the C-4 Linker Lipophilicity
and Potency
The potency of the C-4 modified compounds for
RORγt was investigated in a time-resolved FRET (TR-FRET) coactivator
recruitment assay (Table , Figure A,
see Figure S3A for a schematic representation
of the assay setup).[38] RORγt is constitutively
active, which means that it shows a basic level of transcriptional
activity, thus allowing the partial recruitment of coactivators in
the absence of an agonist.[39] Reference
compounds 1 and 2 showed potent inhibition
of coactivator binding, with IC50 values comparable to
previous studies.[24,28] Interestingly, compound 3 (FM156) containing an ether linkage demonstrated
a 9-fold increase in potency compared to 2 with an amine
linker (IC50 of 31 ± 3 nM vs 270 ± 20 nM, respectively),
resulting in an IC50 value in the same range as indazole 1. The trans-alkene linker (6, FM260) resulted in a low nM IC50 value
as well (IC50 of 20 ± 2 nM), in contrast to the cis-alkene linker (7) which was 25 times less
potent (IC50 of 490 ± 30 nM), as predicted by the
docking scores. In contrast, a thioether linkage (4)
resulted in a significantly decreased potency compared to 2 (IC50 of 6.6 ± 0.8 μM), whereas a methylated
amine linker (5) resulted in a loss of activity (IC50 > 100 μM). The introduction of a fluorine substituent
at the ortho position of the benzoic acid moiety
proved detrimental to coactivator inhibition with a 7–13 fold
decrease in potency compared to the compound without a substituent
(8 vs 2 and 9 vs 3), which was also the case for an isoxazole analogue in a previous
SAR study.[28]
Figure 2
Biochemical analysis
and X-ray crystallography data for isoxazole
compounds with C-4 modifications. (A) Dose–response curves
from the TR-FRET coactivator recruitment assay for 1, 2, 3, and 6; (B) dose–response
curves from the ligand displacement TR-FRET assay using the AlexaFluor
MRL-871 probe for 1, 2, 3,
and 6; (C) melting temperatures (ΔTm in °C), as measured in TSA for 1, 2, 3, and 6. Data were recorded
in ≥3 independent experiments, each recorded in triplicate
(one representative dataset shown). Error bars represent the SD of
the mean; (D,G) tertiary co-crystal structure of RORγt in complex
with 3 (PDB: 7NPC) (D) and 9 (PDB: 7NP5) (G) (stick representation).
The final 2Fo–Fc electron density map of compounds is shown
as an isomesh contoured at 1σ; (E,H) overlay of the co-crystal
structure of RORγt in complex with FM26 (2) (PDB: 6SAL) and RORγt bound to 3 (E), and an overlay of
RORγt with 3 and 9 (H); and (F,I)
enlarged view of the allosteric pocket of RORγt showing the
interactions between 3 (F)/9 (I) and the
protein.
Biochemical analysis
and X-ray crystallography data for isoxazole
compounds with C-4 modifications. (A) Dose–response curves
from the TR-FRET coactivator recruitment assay for 1, 2, 3, and 6; (B) dose–response
curves from the ligand displacement TR-FRET assay using the AlexaFluor
MRL-871 probe for 1, 2, 3,
and 6; (C) melting temperatures (ΔTm in °C), as measured in TSA for 1, 2, 3, and 6. Data were recorded
in ≥3 independent experiments, each recorded in triplicate
(one representative dataset shown). Error bars represent the SD of
the mean; (D,G) tertiary co-crystal structure of RORγt in complex
with 3 (PDB: 7NPC) (D) and 9 (PDB: 7NP5) (G) (stick representation).
The final 2Fo–Fc electron density map of compounds is shown
as an isomesh contoured at 1σ; (E,H) overlay of the co-crystal
structure of RORγt in complex with FM26 (2) (PDB: 6SAL) and RORγt bound to 3 (E), and an overlay of
RORγt with 3 and 9 (H); and (F,I)
enlarged view of the allosteric pocket of RORγt showing the
interactions between 3 (F)/9 (I) and the
protein.To prove an allosteric mode of
action, the compounds were also
tested in two other TR-FRET assay formats. First, a previously described
AlexaFluor 647-labeled MRL-871 probe[24] (Figure S1) was used, which upon binding to RORγt
shows FRET pairing with an anti-His terbium cryptate donor on the
protein (see Figure S3B).[24] All compounds showed displacement of the AlexaFluor-MRL
allosteric probe, proving an allosteric binding mode (Table , Figure B), with the IC50 values correlating
with the IC50 values from the TR-FRET coactivator recruitment
assay (Table ). In
particular, ether 3 and trans-alkene 6 demonstrated efficient displacement of the probe, with IC50 values of 12 ± 1 and 7.4 ± 0.9 nM, respectively,
significantly increased as compared to 2 (IC50 value of 100 ± 10 nM).The most potent ligands 3 and 6 were
also measured in a competitive TR-FRET coactivator recruitment assay,
where the compounds were titrated in the presence of cholesterol,
which is an orthosteric agonist (activator) for RORγt[40] (Figure S2, see Figure S3C for a schematic representation of
the assay setup). The binding curves for compounds 3 and 6 show that the IC50 values did not decrease upon
increasing concentrations of cholesterol, demonstrating that their
binding mode is independent to that of cholesterol, supporting their
allosteric-binding mode. In fact, the presence of cholesterol slightly
enhanced the potency of both 3 and 6, suggesting
a cooperative behavior between orthosteric and allosteric ligand binding,
as was observed previously.[27,28]A thermal shift
assay (TSA) was performed as an orthogonal assay
to investigate the effect of the compounds on the thermal denaturation
of the RORγt protein (Table , Figure C). Ligand binding typically improves the thermal stability of a
protein by stabilizing the protein fold, as indicated by the change
in the melting temperature ΔTm.[41−43] The C-4 isoxazole derivatives 3–9 showed a thermal
stabilization effect according to their potency, as observed in the
TR-FRET assays. The most potent compounds 3 and 6 showed particularly high ΔTm values (4.9 and 6.4 °C resp.). These values are improved compared
to that for 2, which provides further indication that 3 and 6 have a high-binding affinity for RORγt.Crystallization studies provided the co-crystal structures of the
RORγt LBD in complex with compounds 3 (ether linker)
and 9 (fluoro substituent), with resolutions of 1.46
and 1.55 Å, respectively (Tables S4 and S5, Figure S6). The co-crystal structures of RORγt with 3 and 9 showed clear electron density for the
compounds in the allosteric binding site between helices 3, 4, 11,
and 12 (Figure D,G).
In Figure E, an overlay
of RORγt in complex with 2 and with 3 is shown, demonstrating a common binding pose, with the C-3, C-4,
and C-5 isoxazole substituents anchored at the same position. The
pyrrole moiety forms a hydrogen bond interaction with the main chain
carbonyls of residues Leu353 and Lys354 (Figure F). Additionally, the carboxylic acid moiety
forms hydrogen bond interactions with the side chain of Gln329 and
backbone nitrogens of Phe498 and Ala497, as is also the case for 1 and 2. Interestingly, the electron density
for 3 and 2 would allow two different conformations
of the C-4 linker,[28] but the preferred
conformation for the ether linker in 3 is opposite to
the preferred conformation for the amine linker in 2 (Figure E). For compound 9, the electron density and binding mode are highly similar
to 3, with the same conformation for the ether linker
present (Figure G–I).
The ortho-fluoro substituent on the benzoic acid
ring does not significantly influence the conformation of the compound,
except for the disubstituted phenyl ring at the isoxazole C-3 position,
which is normally fixed at the same position but is now slightly shifted
(Figure H).The increased potency of ether 3 and alkene 6 could be the result of increased hydrophobic effects of
these more lipophilic compounds toward the hydrophobic allosteric-binding
pocket (see the protein–ligand interaction plot in Figure S5). The different preferred conformations
of the ether linker for 3, as seen in the co-crystal
structure, might also be contributing to the potency. In contrast,
the compound with a thioether linkage (4) shows a significantly
lower potency, which might be caused by a slight change in the bond
angle and length of the linker. The docking pose shows that the isoxazole
core and pyrrole moiety are clearly shifted compared to the compound
with ether linker (3) (Figure S4A). The drop in potency for 5, featuring a methylated
amine linker, likely results from the restricted rotation of the linker.
The electron density for the isoxazoles shows that two conformations
of the linker are present, but when the linker is more rotationally
restricted, as is the case for 5, only one of these linker
conformations will most likely be present, potentially resulting in
a higher entropic penalty and in turn a decrease in the binding affinity.
The introduction of an ortho-fluoro substituent at
the benzoic acid moiety leads to a slight decrease in potency (8 vs 2 and 9 vs 3).
The co-crystal structure in complex with 9 shows that
the fluoro substituent is placed in the same plane as the carboxylic
acid moiety (Figure I). This unfavorable conformation is believed to be caused in order
to both fit the fluoro substituent in the pocket and also allow hydrogen
bond interactions between the carboxylic acid and the protein, which
could lead to charge repulsion between the two moieties and a decreased
potency.
SAR Studies Show the Necessity of a H-Bond
Donor Moiety at the Isoxazole C-5 Position
The SAR was further
explored at the isoxazole C-5 position to investigate the hydrogen
bonding character of the pyrrole moiety of 2. Compound 3 with the ether linkage was chosen as a core scaffold, based
on the combination of its high potency, availability of the co-crystal
structure, and synthetic feasibility. A focused library of C-5 derivatives
(compounds 10–13, Table ) was designed and synthesized (the full
library used for docking is shown in Table S2). Specifically, the effect of the nitrogen position (10), pKa (11) (pKa = 10.17 for 11 vs 14.99 for 3) and substitution (12) was investigated, as well as
the hydrophobic space around the pyrrole (13).
Table 2
SAR Studies around the C-5 Isoxazole
Positiona
Glide
scores, TR-FRET IC50 values (μM) from the coactivator
recruitment and AlexaFluor-MRL-871
displacement assays, ΔTm values
(°C) from TSA, and fold decrease in IL-17a expression levels
relative to DMSO (RT-PCR) are shown. Abbreviations: n.d., not determined.
TR-FRET and TSA data are recorded in triplicate; values are representative
of ≥3 repeated experiments. RT-PCR data are recorded in triplicate;
values are representative of ≥2 repeated experiments.
Glide
scores, TR-FRET IC50 values (μM) from the coactivator
recruitment and AlexaFluor-MRL-871
displacement assays, ΔTm values
(°C) from TSA, and fold decrease in IL-17a expression levels
relative to DMSO (RT-PCR) are shown. Abbreviations: n.d., not determined.
TR-FRET and TSA data are recorded in triplicate; values are representative
of ≥3 repeated experiments. RT-PCR data are recorded in triplicate;
values are representative of ≥2 repeated experiments.Compounds 10–13 were synthesized
using an analogous synthesis route as shown previously (Scheme ). Because the C-5 moiety is
incorporated early in the synthesis, it was not possible to diversify
from a late-stage intermediate, as for the C-4 modifications. Bromide 14 was subjected to a Suzuki reaction with the associated
pinacol ester, containing a Boc-, THP-, or methyl-protected heterocycle,
followed by ester hydrolysis with DIBAL to yield 30–33 (Scheme ). For the N-methyl pyrrole, the original conditions were used (Scheme S2) (the tetrakis palladium catalyst in
the Suzuki reaction and LiAlH4 for reduction) because no
labile-protecting group was used in this synthesis route. Next, a
Mitsunobu reaction with methyl 4-hydroxybenzoate and ester hydrolysis
combined with heterocycle deprotection were conducted to afford the
final compounds 10–13.
Scheme 2
Synthesis Route for
Different Trisubstituted Isoxazoles (C-5 Library)
Reagents
and conditions: (a) N-Boc-pyrrole-B(pin), N-THP-pyrazole-B(pin)
or 5-methyl N-Boc-pyrrole-B(pin) (see Experimental Section for the synthetic procedure of this pinacol
ester), Pd(dppf)Cl2, Cs2CO3, and
DME, 85 °C, 8 h, 49% (26), 34% (27),
and 24% (29); (b) N-methyl-pyrrole-B(pin),
Pd(PPh3)4, Na2CO3, DME,
and H2O, 85 °C, 8 h, 55% (28); (c) DIBAL
and CH2Cl2, −78 °C, 3 h, 69% (30), 71% (31), and 58% (33); (d)
LiAlH4 and THF, 0 °C → rt, 2 h, 76% (32); (e) (i) methyl 4-hydroxybenzoate, DIAD, PPh3, and THF, reflux, 3 h, 21%, (ii) LiOH, EtOH, and H2O,
95 °C, 3 h, 41% (10); (f) (i) methyl 4-hydroxybenzoate,
DIAD, PPh3, Et3N, and THF, reflux, 3 h, 44–45%,
(ii) LiOH, EtOH, and H2O, 95 °C, 3 h, 86% (12), 83% (13); (g) (i) methyl 4-hydroxybenzoate, DIAD,
PPh3, Et3N, and THF, reflux, 3 h, 49%, (ii)
TFA and CH2Cl2, 40 °C, 2 h, 87%, (iii)
LiOH, EtOH, and H2O, 95 °C, 3 h, 89% (11).
Synthesis Route for
Different Trisubstituted Isoxazoles (C-5 Library)
Reagents
and conditions: (a) N-Boc-pyrrole-B(pin), N-THP-pyrazole-B(pin)
or 5-methyl N-Boc-pyrrole-B(pin) (see Experimental Section for the synthetic procedure of this pinacol
ester), Pd(dppf)Cl2, Cs2CO3, and
DME, 85 °C, 8 h, 49% (26), 34% (27),
and 24% (29); (b) N-methyl-pyrrole-B(pin),
Pd(PPh3)4, Na2CO3, DME,
and H2O, 85 °C, 8 h, 55% (28); (c) DIBAL
and CH2Cl2, −78 °C, 3 h, 69% (30), 71% (31), and 58% (33); (d)
LiAlH4 and THF, 0 °C → rt, 2 h, 76% (32); (e) (i) methyl 4-hydroxybenzoate, DIAD, PPh3, and THF, reflux, 3 h, 21%, (ii) LiOH, EtOH, and H2O,
95 °C, 3 h, 41% (10); (f) (i) methyl 4-hydroxybenzoate,
DIAD, PPh3, Et3N, and THF, reflux, 3 h, 44–45%,
(ii) LiOH, EtOH, and H2O, 95 °C, 3 h, 86% (12), 83% (13); (g) (i) methyl 4-hydroxybenzoate, DIAD,
PPh3, Et3N, and THF, reflux, 3 h, 49%, (ii)
TFA and CH2Cl2, 40 °C, 2 h, 87%, (iii)
LiOH, EtOH, and H2O, 95 °C, 3 h, 89% (11).To assess the SAR around the C-5 position,
analogues 10–13 were evaluated in the TR-FRET
coactivator recruitment assay (Figure A, Table ). A nitrogen shift (2- instead
of 3-position) in the pyrrole ring (10) resulted in a
4.5-fold decrease in potency (IC50 = 140 ± 10 nM)
compared to 3. Pyrazole 11 (FM257) showed a slight drop in potency, with an IC50 value
of 110 ± 10 nM. For compound 12 (containing a methylated
pyrrole) and compound 13 (methylation of the pyrrole
at the 5-position), a lower activity was observed with an IC50 value of 3.3 ± 0.3 and 2.9 ± 0.2 μM, respectively.
Additionally, the TR-FRET AlexaFluor-MRL assay was performed, validating
an allosteric binding mode for all compounds (Figure B, Table ). Analysis of the IC50 values for this
series showed that these were in line with the IC50 values
from the TR-FRET coactivator assay (Table ). In the TSA, compounds 10 and 11 induced thermal stabilization with ΔTm values of 2.8 and 2.7 °C respectively, which is
lower than for 3 (ΔTm = 4.9 °C) but slightly higher than for 2 (ΔTm = 2.4 °C). Compounds 12 and 13 did not show any response (Figure C, Table ).
Figure 3
Biochemical analysis and X-ray crystallography data for
isoxazole
compounds with C-5 modifications. (A) Dose–response curves
from the TR-FRET coactivator recruitment assay for 2, 3, 10, and 11; (B) dose–response
curves from the ligand displacement TR-FRET assay using the AlexaFluor
MRL-871 probe for 2, 3, 10,
and 11; (C) melting temperatures (ΔTm in °C), as measured in TSA for 2, 3, 10, and 11. Data were recorded
in ≥3 independent experiments, each recorded in triplicate
(one representative dataset shown). Error bars represent the SD of
the mean. (D,G) Tertiary co-crystal structure of RORγt in complex
with 10 (D) (PDB: 7NEC) and 11 (G) (PDB: 7NP6) (stick representation).
The final 2Fo–Fc electron density map of the compounds is shown
as an isomesh contoured at 1σ; (E,H) overlay of the crystal
structure of RORγt in complex with 3 and RORγt
bound to 10 (E) or 11 (H); and (F,I) enlarged
view of the allosteric pocket of RORγt showing the interactions
between 10 (F)/11 (I) and the protein.
Biochemical analysis and X-ray crystallography data for
isoxazole
compounds with C-5 modifications. (A) Dose–response curves
from the TR-FRET coactivator recruitment assay for 2, 3, 10, and 11; (B) dose–response
curves from the ligand displacement TR-FRET assay using the AlexaFluor
MRL-871 probe for 2, 3, 10,
and 11; (C) melting temperatures (ΔTm in °C), as measured in TSA for 2, 3, 10, and 11. Data were recorded
in ≥3 independent experiments, each recorded in triplicate
(one representative dataset shown). Error bars represent the SD of
the mean. (D,G) Tertiary co-crystal structure of RORγt in complex
with 10 (D) (PDB: 7NEC) and 11 (G) (PDB: 7NP6) (stick representation).
The final 2Fo–Fc electron density map of the compounds is shown
as an isomesh contoured at 1σ; (E,H) overlay of the crystal
structure of RORγt in complex with 3 and RORγt
bound to 10 (E) or 11 (H); and (F,I) enlarged
view of the allosteric pocket of RORγt showing the interactions
between 10 (F)/11 (I) and the protein.Crystallization studies provided the co-crystal
structures of the
RORγt LBD in complex with C-5-modified compounds 10 (2-pyrrole) and 11 (pyrazole), with resolutions of
1.95 and 1.84 Å, respectively (Tables S6 and S7, Figure S6). The co-crystal structure of RORγt
with 10 reveals a clear ligand electron density in the
allosteric site with the ether linker in the same preferred conformation
as for 3 (Figure D,E). The 2-pyrrole at the C-5 position, with the nitrogen
at a different position, does not establish a direct hydrogen bond
interaction with the protein, as expected, potentially explaining
the lower potency of the compound for RORγt. Interestingly,
the NH of the pyrrole substituent of 10 forms an alternative
hydrogen bond network via a water molecule in the
binding pocket at a distance of 3.2 Å from the NH of the pyrrole
(Figure F) (this water
molecule appears to be present in all crystal structures with an allosteric
ligand, and thus has a structural role). For pyrazole 11, the binding mode is highly similar to 3, again establishing
a hydrogen bond interaction between the NH of the pyrazole and the
backbone carbonyls of RORγt (Figure G–I).These results show that
small changes at the pyrrole C-5 moiety
lead to relevant changes in the potency of the compounds. The 4.5-fold
decrease in potency for the compound with a changed position of the
nitrogen in the pyrrole ring (10) (compared to 3) is most probably due to the loss of the characteristic
H-bond interaction between the pyrrole ring and the backbone of the
protein. The additional H-bond interaction with a water molecule explains
why 10 is still active on RORγt with an IC50 value < 150 nM, despite the lack of the characteristic
C-5 H-bond interaction. For the pyrazole (11, FM257), the 3.5-fold lower potency indicates that the lower pKa of the pyrazole is not beneficial for H-bond formation,
although other factors might also be involved, for example, the existence
of the pyrazole in two tautomeric forms for which only one can establish
the hydrogen bond interaction with the protein. The much lower potency
for compound 12 (methylated pyrrole) suggests that the
space around the pyrrole is limited, which is also supported by the
low potency of the methyl substitution of the pyrrole at the 5-positon
(13). The docking score predicts compound 12 to be highly active, based on the docking pose where the pyrrole
ring is rotated and the methyl substituent could point toward a small
cavity (Figure S4B). Together, based on
the biochemical data, the methylated pyrrole does not appear to have
enough space for this rotation in the allosteric pocket. Combined,
these data show the need for an H-bond donor at the right position
of the heterocycle at the C-5 position of the isoxazole scaffold,
while additional substituents at the ring appear to be too bulky to
be tolerated.
The New Isoxazole Series
Show the Inhibition
of IL-17a Expression in EL4 Cells
A selection of the novel
isoxazole compounds was tested in a quantitative RT-PCR assay to investigate
their functional effect. EL4 cells were treated with 10 μM of
the compound or dimethyl sulfoxide (DMSO), after which the IL-17a
mRNA levels were measured (Figure ). Hit compound 2 showed a 9.3-fold reduction,
in accordance with our previous reports.[28] The optimized compounds 3 and 6 (IC50 values of 31 and 20 nM, respectively) showed a 15- and 16-fold
reduction of IL-17a levels compared to DMSO, which is in line with
their high biochemical potency and indicates good cellular uptake
and activity. Compounds 10 and 11 induced
a 12- and 6.6-fold decrease in IL-17a mRNA levels, correlating with
their slightly lower biochemical potency compared to 3 and 6. Compound 12 only resulted in a
2.4-fold decrease compared to DMSO, as expected, given its significantly
lower potency in the TR-FRET coactivator recruitment assay.
Figure 4
IL-17a mRNA
expression levels in EL4 cells treated with ligands 1, 2, 3, 6, 10, 11, and 12 (10 μM, 24 h) or DMSO
and fold decrease of the IL-17a expression relative to DMSO. The level
of IL-17a expression was normalized to that of GAPDH expression. All
data are expressed as mean ± s.d. (n = 3). The
relative gene expression was calculated by the 2–ΔΔt (Livak) method using DMSO control as
a calibrator. Statistical analysis was performed using a one-way analysis
of variance (ANOVA) compared against the DMSO control following Dunnett’s post hoc test; ***P < 0.001 and ****P < 0.0001.
IL-17a mRNA
expression levels in EL4 cells treated with ligands 1, 2, 3, 6, 10, 11, and 12 (10 μM, 24 h) or DMSO
and fold decrease of the IL-17a expression relative to DMSO. The level
of IL-17a expression was normalized to that of GAPDH expression. All
data are expressed as mean ± s.d. (n = 3). The
relative gene expression was calculated by the 2–ΔΔt (Livak) method using DMSO control as
a calibrator. Statistical analysis was performed using a one-way analysis
of variance (ANOVA) compared against the DMSO control following Dunnett’s post hoc test; ***P < 0.001 and ****P < 0.0001.
The Isoxazole
Series Shows an Improved Selectivity
Profile for RORγt
Previous studies have shown that 1 is selective for RORγt over other NRs, except for
PPARγ on which it shows a significant cross-reactivity to the
orthosteric binding site, as was also supported by the co-crystal
structure.[24,44] Isoxazole 2 already
showed a higher selectivity for RORγt over PPARγ than 1,[28] so we were interested to establish
if the novel isoxazole compounds further improved this promising selectivity
profile.In a TR-FRET coactivator recruitment assay (Table ), 1 showed
PPARγ agonism with an EC50 value of 0.34 ± 0.02
μM, while the EC50 value for 2 was >20-fold
higher (EC50 = 8.2 ± 0.3 μM).[24,28] All novel isoxazole ligands developed in this study were at least
five times less active on PPARγ than 1 (EC50 values > 1.7 μM), except for compounds 12 and 13 (both bearing a methyl substituent at the pyrrole
moiety), which showed a higher cross-reactivity on PPARγ (EC50 values of 1.2 ± 0.1 and 0.79 ± 0.10 μM resp.).
This might be due to the more bulky substitution pattern around the
pyrrole C-5 substituent.
Table 3
EC50 Values
Observed in
the TR-FRET Coactivator Recruitment Assay, IC50 Values
from the TR-FRET Competition Assay with PPARγ, and ΔTm Values (°C) from TSA Assaysa
fold-selective for
RORγt over PPARγ
compound
EC50 (μM) PPARγ
IC50 (μM) PPARγ (competition
rosiglitazone)
ΔTm (°C)
PPARγ
TR-FRET
TSA
1
0.34 ± 0.02
6.3 ± 0.7
2.5 ± 0.2
26
3.1
2
8.2 ± 0.3
81 ± 7
0.3 ± 0.0
30
8.0
3
3.1 ± 0.1
63 ± 8
0.6 ± 0.0
100
8.1
4
18 ± 1
>100
0.7 ± 0.0
2.7
0.5
5
>100
>100
0.0 ± 0.0
6
2.5 ± 0.1
64 ± 3
1.1 ± 0.2
125
5.8
7
7.5 ± 0.3
>100
0.1 ± 0.0
15
7.0
8
15 ± 1
>100
0.1 ± 0.0
1.6
7.0
9
5.8 ± 0.2
67 ± 6
0.4 ± 0.0
67
7.5
10
1.7 ± 0.1
29 ± 4
1.2 ± 0.0
7.1
2.3
11
3.6 ± 0.1
79 ± 11
0.7 ± 0.0
33
3.9
12
1.2 ± 0.1
20 ± 2
1.5 ± 0.0
0.36
<0.1
13
0.79 ± 0.10
19 ± 3
1.5 ± 0.0
0.27
<0.1
TR-FRET and TSA data are recorded
in triplicate; values are representative of ≥3 repeated experiments.
TR-FRET and TSA data are recorded
in triplicate; values are representative of ≥3 repeated experiments.In order to further explore
the binding site of the isoxazole compounds
on PPARγ, a TR-FRET assay was performed whereby the competition
between the isoxazole ligands and a known orthosteric PPARγ
agonist, rosiglitazone, was measured (Table ).[28,45] All compounds showed
competition with rosiglitazone at high concentrations, indicating
binding to the orthosteric site of PPARγ. Interestingly, at
least a 3-fold lower competition with rosiglitazone was observed for
the isoxazole compounds compared to 1 (i.e., at least
a 3-fold increase of the IC50 values compared to 1), with most analogues demonstrating IC50 values
> 50 μM. Overall, a similar trend was observed for the IC50 values in this competition assay as for the EC50 values in the previously mentioned TR-FRET coactivator recruitment
assay (Table ). Again,
compounds 12 and 13 appeared to be less
selective for RORγt than the more potent RORγt-binding
isoxazoles (IC50 values of 20 ± 2 and 19 ± 3
μM resp.). Compound 10 showed some cross-reactivity
with PPARγ as well, although with an IC50 value >
20 μM.Next, the compounds were evaluated in a PPARγ
TSA (Table ). Compounds 10, 12, and 13 showed the highest
thermal stabilization effect (ΔTm of 1.2 and 1.5 °C), in line with the TR-FRET data, and also 6 appeared to show a significant stabilization effect (ΔTm of 1.1 °C). Compound 1 again
showed the most significant activity on PPARγ, with a ΔTm of 2.5 °C. The selectivity of the compounds
for RORγt over PPARγ was calculated (Table ) by comparing the data for
PPARγ and RORγt for both the TR-FRET coactivator recruitment
assay and TSA (see Tables –3). Interestingly, compounds 3 and 6, which are highly potent RORγt
inverse agonists (IC50 values < 35 nM), show a 100-
and 125-fold selectivity for RORγt over PPARγ, respectively,
based on the TR-FRET results, which is a significant improvement compared
to 2 (30-fold). Contrastingly, 12 and 13 containing a methyl substituent at the pyrrole moiety show
higher cross-reactivity toward PPARγ, in the same range as for 1. Together, the cross-reactivity of the isoxazoles on PPARγ
can be modulated by only small changes in the pyrrole substitution
pattern. Gratifyingly, the most potent compounds for RORγt (3 and 6) also feature the highest selectivity
for RORγt over PPARγ.Lastly, the isoxazole compounds
were tested on the farnesoid X
receptor (FXR) to exclude cross-reactivity because trisubstitutedisoxazole ligands, such as GW4064, have been reported
as potent FXR agonists.[46,47] The most potent RORγt
isoxazole ligands and GW4064 were evaluated in a fluorescence
polarization (FP) assay. GW4064 showed a typical agonistic
dose–response curve for FXR, whereas the novel isoxazole ligands
did not elicit an agonistic response on FXR (Figure S7A). Additionally, the EC50 value of GW4064 was not significantly affected by the presence of isoxazole 3 (Figure S7B), thus also excluding
an antagonistic behavior of 3 on FXR.
Absorption, Distribution, Metabolism, and
Excretion Profile
Previous studies on the absorption, distribution,
metabolism, and excretion (ADME) parameters of the isoxazoles showed
that these were promising but not optimal yet for compound 2, most probably governed by the pyrrole moiety and the amine linker,
which are both likely to be prone to oxidation.[28,48] The ADME profiling of the novel isoxazoles (Table ) demonstrates that compounds 3, 6, 9, and 10 in general
show comparable properties to hit compound 2, with promising
values for the kinetic solubility and (passive) membrane permeability,
especially because the compounds contain a carboxylic acid moiety
which generally leads to lower permeability values. The metabolic
stability in human liver microsomes showed that the clearance values
in phase I metabolism are rather high, similar to 2.
Glucuronidation assays (as the most important phase II pathway) showed
promising improved values (% remain), especially for compound 10. The plasma stability (except for 6) was slightly
lower than for 1, and relatively high levels of blood
plasma protein binding were measured. Contrastingly, compound 11, containing a pyrazole instead of a pyrrole substituent
at the C-5 position, showed a rather different ADME profile. It demonstrated
a promising stability in phase I metabolism and was highly stable
in plasma, although the phase II stability (glucuronidation) and permeability
were lower than for the other isoxazoles.
Table 4
ADME Properties
for Compounds 1, 2, 3, 6, 9, 10, and 11
microsomal
stability
compound
kinetic solubility
(μM)
PAMPA (% flux)
phase 1 (Clint, μL/min/mg)
glucuronidation in vitro (% remaining)
plasma stability (% remaining)
plasma
protein
binding (% bound)
1
423
29.7
0.4
37.7
95.5
99.7
2
460
36.1
31.2
73.0
79.1
99.9
3
436
40.9
57.8
30.7
74.2
99.7
6
392
40.3
69.0
58.3
59.5
99.9
9
445
37.3
77.0
54.6
97.1
99.8
10
421
50.5
68.0
68.9
89.5
99.3
11
476
23.0
8.8
39.1
100
98.8
The open-source program
OSIRIS Property Explorer[49] was used to
predict the drug likeness and toxicity of the
compounds (Table S3) to have a first indication
on these aspects. All isoxazole compounds (except 5)
were predicted to have no side effects (i.e., mutagenic, tumorigenic,
irritant, and reproductive effects), whereas 1 was predicted
to have a potentially high risk of reproductive toxicity. Additionally,
all isoxazole ligands showed higher drug scores than 1, with pyrazole 11 showing the best properties, again
highlighting its potential for further development.Combined,
the change of an amine to an ether/alkene linker (2 vs 3 and 6) does not significantly
affect the ADME(T) properties (except for the glucuronidation of 3), while the change of a pyrrole to a pyrazole moiety (11) results in a differentiated, promising profile. The C-4/C-5
isoxazole modifications thus provide valuable entry points for optimization
of the ADME properties.
Conclusions
The
NR RORγt has an important regulatory role in the immune
system, and inhibition of RORγt via allosteric
inverse agonists could be a promising strategy in the treatment of
autoimmune disorders. Apart from indazoles, with 1 (MRL-871) as a prominent example, there has been little chemical
diversity of allosteric RORγt modulators. Trisubstituted isoxazoles
constitute a distinct novel chemotype in this matter, with 2 (FM26) as an exemplary potent hit compound.Here,
we further optimized this isoxazole series, specifically
focusing on the linker at the C-4 position and the heterocycle at
the C-5 position to deliver lead compounds for further optimization.
The optimization of the linker at the C-4 position revealed a clear
correlation between potency and the lipophilicity and flexibility
of the linker, with an ether or alkene linker resulting in the highest
potency. The co-crystal structure of RORγt in complex with ether
compound 3 revealed the role of hydrophobic interactions
and conformation of the linker at the C-4 position. The most potent
compounds 3 (FM156) and 6 (FM260) were highly selective for RORγt over PPARγ,
which is a valuable improvement compared to the exemplary indazole 1.The focused library of C-5 isoxazoles revealed several
routes for
chemical exploration. Changing the position of the hydrogen bond donor via the repositioning of the pyrrole ring (10) led only to a 3-fold lower potency, with the formation of another
hydrogen bond interaction network via a water molecule
in the RORγt pocket. Changing the pyrrole to a pyrazole (with
a lower pKa) in compound 11 (FM257) was also tolerated and resulted in a promising
PK profile, opening up routes for affinity and ADME optimization.In conclusion, by probing the effect of the linker at the C-4 position
and the importance of the polar character of the C-5 moiety, this
study has led to valuable leads as allosteric inverse agonists for
RORγt with improved potency, increased selectivity against PPARγ
and FXR, and promising initial ADME properties. The results not only
provide new insights into the SAR for this specific isoxazole class
of allosteric RORγt ligands but can also most probably similarly
be translated to other allosteric RORγt chemotypes. Overall,
the trisubstituted isoxazole class of allosteric RORγt ligands
shows high potential for chemical biology approaches as well as for
future development in drug discovery programs against autoimmune diseases.
Experimental Section
In Silico Studies
Molecular
Docking Studies
The receptor–ligand
complex structure (PDB code: 4YPQ) was prepared using the Protein Preparation Wizard
within Maestro (version 12.3 (2020-1), Schrödinger LLC, New
York, NY, USA) (default parameters). A receptor grid centered on the
bound ligand was generated using the Protein Preparation Wizard. All
parameters were kept as the default. The examined ligands were drawn
in ChemDraw. Energy-minimized 3D ligand conformations for each molecule
to be investigated were generated using the Ligand Preparation Wizard
within Maestro (default parameters). Ligands were docked using Glide
in standard precision mode with flexible ligand sampling. The predicted
binding modes of all ligands were ranked according to their glide
score.
In Silico Drug Likeness
and Toxicity Predictions
OSIRIS Property Explorer utilizes
the database of traded drugs and commercially available compounds
(Fluka), assumable as nondrug-like data set, to assess the occurrence
frequency of each fragment in the individual structure.[49] The program was used to estimate the risks of
side effects, such as mutagenic, tumorigenic, irritant, and reproductive
effects, as well as drug likeness and overall drug score (by combining
the outcome of clog P, log S (solubility), MW, toxicity risks, and drug likeness).
The drug score is a measure of the compound’s potential to
meet the criteria of a possible drug candidate.
General Chemistry
All nonaqueous
reactions were performed under an argon atmosphere unless otherwise
stated. Water-sensitive reactions were performed in oven-dried glassware,
cooled under argon before use. All solvents were supplied by Biosolve
or Sigma-Aldrich and used without further purification. Dry solvents
were obtained from a MBRAUN solvent purification system (MB-SPS-800).
Water was purified by a Millipore purification train. Deuterated solvents
were obtained from Cambridge Isotope Laboratories. Solvents were removed in vacuo using a Büchi rotary evaporator and a diaphragm
pump. Commercially available starting materials were obtained from
Sigma-Aldrich, TCI Chemicals and Fluorochem. Proton (1H)
NMR (400 MHz), carbon (13C) NMR (100 MHz), and 2D NMR (400
MHz) were recorded on a Bruker Avance 400 MHz NMR spectrometer. Proton
spectra are referenced to tetramethyl silane (TMS). Carbon spectra
are referenced to TMS or the solvent peak of the deuterated spectrum.
NMR spectra are reported as follows: chemical shift (δ) in parts
per million (ppm), multiplicity (s = singlet, d = doublet, t = triplet,
q = quartet, m = multiplet, dd = doublet of doublet, td = triplet
of doublets, app. = apparent), coupling constant (J) in Hertz (Hz) (if applicable), and integration (proton spectra
only). Peak assignments are based on additional 2D NMR techniques
(COSY, HMBC, and HSQC). Analytical liquid chromatography coupled with
mass spectrometry (LC–MS) was performed on a C4 Jupiter SuC4300A
150 × 2.0 mm column using ultrapure water with 0.1% formic acid
(FA) and acetonitrile with 0.1% FA, in general with a gradient of
5–100% acetonitrile for over 10 min, connected to a Thermo
Fisher LCQ fleet ion-trap mass spectrometer. The purity of the samples
was assessed using a UV detector at 254 nm. Unless otherwise stated
all final compounds were >95% pure as judged by HPLC. High-resolution
mass spectra (HRMS) were recorded using a Waters ACQUITY UPLC I-Class
LC system coupled to a Xevo G2 quadrupole time-of-flight (Q-TOF) mass
spectrometer. Preparative HP-LC was performed on a Gemini S4 110A
150 × 21.20 mm column using ultrapure water with 0.1% FA and
acetonitrile with 0.1% FA with various gradients (mentioned for each
compound specifically). Column chromatography was either performed
manually using silica gel (60–200 μm particle size, 60
AÅ) or using an automated Grace Reveleris X2 chromatograph with
pre-packed silica columns supplied by Buchi/Grace (40 μm particle
size). The reaction progress was monitored by thin-layer liquid chromatography
(TLC) using Merck TLC silica gel 60 F254 plates. The visualization
of the plates was achieved using an ultraviolet lamp (λmax = 254 nm).
General Procedure for
Suzuki Coupling
Under an inert atmosphere, pinacol boronate
(2.0 equiv), Cs2CO3 (2.0 equiv), and Pd(dppf)Cl2 (0.1
equiv) were added to a solution of bromide 14 (1.0 equiv)
in degassed 1,2-dimethoxyethane (DME) (0.1 M). The reaction mixture
was heated at 85 °C for 8 h, cooled to room temperature, diluted
with H2O, and extracted with EtOAc (3×). The combined
organic phase was washed with brine, dried over MgSO4,
filtered, and concentrated in vacuo. The crude product
was purified by flash column chromatography using the specified eluent
to afford the desired product.
General
Procedure for the Reduction of Esters
to Alcohols
Under an inert atmosphere, DIBAL (1.23 g/mL in
cyclohexane, 15.0 equiv) was added dropwise to a solution of ester
(1.0 equiv) in anhydrous CH2Cl2 (0.1 M) at −78
°C. The reaction mixture was followed by TLC analysis. Upon complete
consumption of the starting material, the reaction mixture was warmed
to room temperature, quenched by the addition of a saturated Rochelle
salt (KNaC4H4O6·4H2O) solution and stirred vigorously for 60 min. Subsequently, the
mixture was extracted with CH2Cl2 (3×)
and separated. The combined organic phase was washed with brine, dried
over MgSO4, filtered, and concentrated in vacuo to afford the title compound which was purified as described.
General Procedure for Mesylation and Substitution
Under an inert atmosphere, triethylamine (3.0 equiv) was added
to a solution of alcohol compound 16 (1.0 equiv) in anhydrous
CH2Cl2 (0.1 M). Methanesulfonic anhydride (1.5
equiv) was added to the flask, and the reaction mixture was heated
at 40 °C. The reaction was monitored using NMR analysis. Subsequently,
aniline (5.0 equiv) was added, and the reaction mixture was stirred
for 1 h, cooled to room temperature, and CH2Cl2 was removed in vacuo. The crude product was purified
by flash column chromatography using the specified eluent.
General Procedure for Ester Hydrolysis and
N-Boc Deprotection
LiOH·H2O (10.0 equiv)
was added to a suspension of ester (1.0 equiv) in a 4:1 mixture of
EtOH/H2O (0.05 M). The reaction mixture was heated to 95
°C followed by TLC analysis. Upon complete consumption of the
starting material, EtOH was removed in vacuo, and
the resulting aqueous mixture was acidified to pH 3 using 10% v/v
aqueous HCl and extracted with a 9:1 mixture of CH2Cl2/MeOH (5×). The combined organic phase was dried (MgSO4), filtered, and concentrated in vacuo to
furnish the final compound which was purified as described.
General Procedure for Mitsunobu Coupling
Under an inert
atmosphere, triphenylphosphine (2.0 equiv) and DIAD
(2.0 equiv) were dissolved in anhydrous tetrahydrofuran (THF) (0.05
M) and stirred for 15 min at 0 °C. Subsequently, the alcohol
compound (1.0 equiv), the benzoate (1.1 equiv), and triethylamine
(1.0 equiv) were added, and the reaction mixture was heated at 80
°C for 3 h, cooled to room temperature, and THF was removed in vacuo. The crude product was purified by flash column
chromatography using the specified eluent.
According to the general procedure for Mitsunobu
coupling, alcohol 31 (0.100 g, 0.234 mmol) was reacted
with methyl 4-hydroxybenzoate (0.039 g, 0.257 mmol). The crude product
was purified by column chromatography, eluting with a gradient of
10–25% EtOAc in n-heptane, to furnish the
ether compound (65.0 mg, 49%). The ether compound (0.045 g, 0.080
mmol), containing a THP-protected pyrazole, was diluted in a mixture
of TFA/CH2Cl2 (50/50) (0.05 M), and the reaction
mixture was stirred at 40 °C for 2 h. Afterward, it was cooled
to room temperature, and the solvent was removed in vacuo. The crude product was purified by column chromatography, eluting
with 25–40% EtOAc in n-heptane, to furnish
the pyrazole product (28.0 mg, 88%). The resulting product was subjected
to ester hydrolysis according to the general procedure for ester hydrolysis
and N-Boc deprotection. The crude product was purified
via preparative HPLC (gradient of 62–67% acetonitrile in H2O) to furnish carboxylic acid 11 (26.0 mg, 89%)
as a white solid. 1H NMR (400 MHz, acetone-d6): δ (ppm) 8.24 (2H, s, pyrazole H-3 and H-5),
7.92 (4H, m, ArH-3, ArH-5 and benzoate C-2), 7.80 (1H, app. t, J = 7.9 Hz, ArH-4), 6.94 (2H, d, J = 8.1
Hz, benzoate H-3), 5.12 (2H, s, CH2O); 13C NMR (100 MHz, acetone-d6):
δ (ppm) 167.2 (CO2H), 164.6 (C-5),
162.9 (benzoate C-4), 159.9 (C-3), 137.1 (ArC-2), 134.4 (ArC-3), 133.6
(pyrazoleC-3 and C-5), 133.0 (q, J = 30.4 Hz, ArC-6),
132.8 (ArC-4), 132.4 (benzoate C-2), 127.4 (ArC-1), 126.2 (q, J = 5.0 Hz, ArC-5), 124.3 (benzoate C-1), 122.7 (q, J = 274.3 Hz, CF3), 115.1 (benzoateC-3), 110.2 (C-4), 109.6 (pyrrole C-4). LC–MS (ESI): calcd
for C21H13ClF3N3O4 [M + H]+, 464.05; observed, 464.08 (Rt = 4.45 min). HRMS (ESI): calcd for C21H13ClF3N3O4 [M + H]+, 464.0625; observed, 464.0610.
Alcohol compound 16 (0.256
g, 0.580 mmol) was dissolved in anhydrous CH2Cl2 (8 mL). To this was added Dess–Martin periodinane (0.369
mg, 0.870 mmol), and the reaction mixture was stirred at room temperature
for 3 h. The reaction mixture was quenched by the addition of 10%
aqueous Na2S2O3 solution and extracted
with CH2Cl2 (3×). The combined organic
phase was washed with saturated aqueous NaHCO3 and H2O, dried over MgSO4, filtered, and concentrated in vacuo. The crude product was purified by column chromatography,
eluting with a gradient of 20–30% EtOAc in n-heptane, to furnish the aldehyde (112.0 mg, 44%). Methyl 4-amino-2-fluorobenzoate
(0.052 g, 0.310 mmol) was added to a solution of the aldehyde (0.112
g, 0.250 mmol) and AcOH (1.43 μL, 0.025 mmol) in MeOH (2 mL).
The reaction mixture was heated at the reflux for 24 h and then concentrated in vacuo. The intermediate imine was isolated by flash column
chromatography, eluting with a gradient of 10–30% EtOAc in n-heptane (62.3 mg, 42%), and then dissolved in EtOH (1
mL), cooled to 0 °C (ice), and treated with NaBH4 (0.020
mg, 0.526 mmol). The reaction mixture was stirred at 86 °C for
4 h. The solvent was removed in vacuo, and the mixture
was dissolved in CH2Cl2 and washed with water.
The aqueous phase was further extracted with CH2Cl2 (3×), dried (MgSO4), filtered, and concentrated in vacuo. The crude product was purified by flash column
chromatography, eluting with a gradient of 30–40% EtOAc in n-heptane, to furnish compound 19 (10.5 mg,
17%). 1H NMR (400 MHz, DMSO-d6): δ 11.4 (1H, s, pyrrole-NH), 7.89 (1H, d, J = 8.0 Hz, ArH-5), 7.84 (1H, d, J = 7.5
Hz, ArH-3), 7.72 (1H, app. t, J = 8.0 Hz, ArH-4),
7.47 (1H, app. t, J = 8.7 Hz, benzoate H-2), 7.40
(1H, m, pyrrole H-2), 6.99 (1H, m, pyrrole H-5), 6.76 (1H, app. t, J = 4.6 Hz, CH2NH), 6.55 (1H,
m, pyrrole H-4), 6.25 (1H, dd, J = 8.9, 1.9 Hz, benzoate
H-3a), 6.10 (1H, d, J = 14.6 Hz, benzoate
H-3b), 4.17 (2H, m, CH2NH).
LC–MS (ESI): calcd for C23H16ClF4N3O3 [M + H]+, 494.08; observed,
494.25 (Rt = 7.14 min).
Alcohol
compound 16 (0.150 g, 0.407 mmol) was dissolved in anhydrous
CH2Cl2 (5 mL). To this was added Dess–Martin
periodinane (0.216 mg, 0.610 mmol), and the reaction mixture was stirred
at room temperature for 3 h. The reaction mixture was quenched by
the addition of 10% aqueous Na2S2O3 solution and extracted with CH2Cl2 (3×).
The combined organic phase was washed with saturated aqueous NaHCO3 and H2O, dried over MgSO4, filtered,
and concentrated in vacuo. The crude product was
purified by column chromatography, eluting with a gradient of 15–30%
EtOAc in n-heptane, to furnish 21 (150
mg, 100%) as colorless oil. 1H NMR (400 MHz, CDCl3): δ (ppm) 9.67 (1H, s, CHO), 8.54 (1H, s,
pyrrole H-2), 7.81–7.73 (2H, m, ArH-3 and ArH-5), 7.63 (1H,
app. t, J = 8.0 Hz, ArH-4), 7.43–7.38 (1H,
m, pyrrole H-5), 6.97–6.92 (1H, m, pyrrole H-4), 1.66 (9H,
s, C(CH3)3). LC–MS (ESI):
calcd for C20H16ClF3N2O4 [M + H]+, 441.08; observed, 440.92 (Rt = 5.87 min).
tert-Butyl (E) or (Z)-3-(3-(2-Chloro-6-(trifluoromethyl)phenyl)-4-(4-(methoxy
carbonyl)styryl)isoxazol-5-yl)-1H-pyrrole-1-carboxylate
(24 and 25)
Under an inert atmosphere,
triphenylphosphine (0.890 g, 3.39 mmol) and methyl 4-(bromomethyl)benzoate
(0.519 g, 2.27 mmol) were dissolved in acetonitrile (20 mL). The reaction
mixture was stirred at 82 °C for 3 h, when TLC indicated the
full conversion of the starting material. The reaction mixture was
cooled to room temperature and precipitated in toluene (150 mL). The
solid was filtered, collected, and dried under reduced pressure to
yield (4-(methoxycarbonyl)benzyl) triphenylphosphonium as a crystalline
white solid (930 mg, 100%). The resulting product (0.152 g, 0.369
mmol) was dissolved in anhydrous THF (2 mL) under an inert atmosphere.
The solution was cooled to −78 °C, LiHMDS (0.4 mL, 1 M
in hexane) was added and the solution was stirred at −78 °C
for 2 h. Aldehyde 23 (0.125 g, 0.284 mmol), dissolved
in anhydrous THF (2 mL), was added dropwise. The reaction mixture
was slowly warmed to room temperature and stirred for 24 h. The reaction
was quenched with H2O (10 mL) and extracted with EtOAc
(3×). The combined organic phase was washed with H2O and brine, dried over MgSO4, filtrated, and concentrated in vacuo. The crude product was purified by column chromatography,
eluting with a gradient of 5–20% EtOAc in n-heptane, to furnish a mixture of cis/trans isomers (3:1) (70 mg).
The mixture of cis/trans isomers was separated via preparative HPLC
(gradient of 80–100% acetonitrile in H2O) to furnish 24 (trans isomer, 12.0 mg, 7%) and 25 (cis isomer,
44.0 mg, 27%), both as white solids. Trans isomer (24): 1H NMR (400 MHz, CDCl3): δ 7.92 (2H,
d, J = 7.9 Hz, benzoate H-2), 7.80 (3H, m, ArH-3,
ArH-5 and pyrrole H-2), 7.63 (1 H, app. t, J = 7.8
Hz, ArH-4), 7.39 (1H, m, pyrrole H-5), 7.24 (2H, d, J = 7.9 Hz, benzoate H-3), 7.03 (1H, d, J = 16.5
Hz, C4-HC=CH), 6.73 (1H, m, pyrrole H-4),
6.17 (1H, d, J = 16.4 Hz, benzoate C4-HC=CH), 3.90 (3H, s, OCH3), 1.64
(9H, s, (C(CH3)3); 13C NMR (100 MHz, CDCl3): δ (ppm) 166.7 (COOCH3), 162.7 (C-5), 157.1 (C-3), 148.1 (NCO2), 141.3 (benzoate C-4), 136.7 (ArC-2), 133.2 (ArC-3),
132.2 (q, J = 31.5 Hz, ArC-6), 131.1 (ArC-4), 129.9
(benzoate C-2), 129.8 (benzoate C4-HC=CH),
129.2 (benzoate C-1), 127.6 (ArC-1), 126.0 (benzoate C-3), 125.1 (q, J = 5.0 Hz, ArC-5), 121.5 (pyrrole C-5), 121.4 (q, J = 274.5 Hz, CF3), 119.7 (pyrroleC-2), 118.0 (C4-HC=CH), 114.6 (pyrrole C-3),
112.6 (C-4), 110.4 (pyrrole C-4), 85.1 (C(CH3)3), 52.1 (OCH3), 27.9
(C(CH3)3). LC–MS (ESI):
calcd for C29H24ClF3IN2O5 [M + H]+, 573.13; observed, 573.00 (Rt = 6.44 min).Cis isomer (25): 1H NMR (400 MHz, CDCl3): δ 7.77 (2H,
d, J = 8.0 Hz, benzoate H-2), 7.68 (2H, m, ArH-3
and ArH-5), 7.53 (1 H, app. t, J = 8.0 Hz, ArH-4),
7.38 (1H, br s, pyrrole H-2), 7.28 (2H, d, J = 8.0
Hz, benzoate H-3), 7.08 (1H, br s, pyrrole H-5), 6.62 (1H, d, J = 12.1 Hz, C4-HC=CH), 6.50 (1H,
br s, pyrrole H-4), 6.18 (1H, d, J = 12.1 Hz, benzoate
C4-HC=CH), 3.87 (3H, s, OCH3), 1.57 (9H, s, (C(CH3)3); 13C NMR (100 MHz, CDCl3): δ
(ppm) 166.7 (COOCH3), 161.3 (C-5), 159.0
(C-3), 148.1 (NCO2), 140.8 (benzoate C-4),
136.6 (ArC-2), 133.9 (C4-HC=CH), 133.1 (ArC-3),
132.0 (q, J = 31.5 Hz, ArC-6), 130.8 (ArC-4), 129.5
(benzoate C-2), 129.1 (benzoate C-1), 128.2 (benzoate C-3), 126.8
(ArC-1), 124.9 (q, J = 5.0 Hz, ArC-5), 121.5 (q, J = 274.5 Hz, CF3), 120.6 (pyrroleC-5), 119.7 (pyrrole C-2), 118.2 (benzoate C4-HC=CH),
114.6 (pyrrole C-3), 110.8 (C-4), 110.0 (pyrrole C-4), 84.7 (C(CH3)3), 52.1 (OCH3), 27.9 (C(CH3)3). LC–MS (ESI): calcd for C29H24ClF3IN2O5 [M + H]+, 573.13; observed,
573.00 (Rt = 6.26 min).
Under
an inert atmosphere, a Schlenk tube was charged with tert-butyl 2,4-dimethyl-1H-pyrrole-1-carboxylate
(1.10 g, 6.05 mmol). Two separate flasks under argon were charged
with [Ir(OMe)(COD)]2 (0.060 g, 0.091 mmol) and dtbpy (0.049
g, 0.182 mmol). HBPin (1.16 g, 9.08 mmol) was added to the [Ir(OMe)(COD)]2 containing flask. Anhydrous THF (18 mL) was added to the
dtbpy-containing flask, and the solution was mixed with the [Ir(OMe)(COD)]2 and HBPin mixture. The resulting solution was transferred
to the Schlenk flask, and the reaction mixture was stirred at 60 °C
for 6 h. Afterward, the reaction mixture was cooled to room temperature
and THF was removed in vacuo. The crude product was
purified by flash column chromatography, eluting with a gradient of
0–5% EtOAc in n-heptane, to furnish tert-butyl 2-methyl-4-(4,4,5-trimethyl-1,3,2-dioxaborolan-2-yl)-1H-pyrrole-1-carboxylate (1.30 g, 70%).[50]Subsequently, according to the general procedure
for Suzuki coupling, bromide 14 (1.09 g, 2.82 mmol) was
reacted with the synthesized pinacol boronate (1.30 g, 4.23 mmol).
The crude product was purified by flash column chromatography, eluting
with a gradient of 30–60% CH2Cl2 in n-heptane, to furnish 29 (0.325 g, 24%) as
colorless oil. 1H NMR (400 MHz, CDCl3): δ
(ppm) 8.39 (1H, s, pyrrole H-2), 7.74–7.65 (2H, m, ArH-3 and
ArH-5), 7.54 (1H, app. t, J = 8.0 Hz, ArH-4), 6.70
(1H, s, pyrrole H-4), 3.61 (3H, s, CO2CH3), 2.51 (3H, s, NCCH3), 1.64
(9H, s, C(CH3)3); 13C NMR (100 MHz, CDCl3): δ (ppm) 169.0 (C-5), 161.6
(CO2CH3), 158.9 (C-3), 148.9
(NCO2), 136.2 (ArC-2), 132.8 (pyrroleC-5), 132.6 (ArC-3), 131.3 (q, J = 31.4 Hz, ArC-6),
130.5 (ArC-4), 127.7 (ArC-1), 124.9 (pyrrole C-2), 124.5 (q, J = 4.9 Hz, ArC-5), 121.6 (q, J = 274.4
Hz, CF3), 111.6 (pyrrole C-3), 110.9 (pyrroleC-4), 106.9 (C-4), 84.7 (C(CH3)3), 51.6 (CO2CH3), 28.0 (C(CH3)3), 15.3 (NCCH3). LC–MS (ESI): calcd for C22H20ClF3N2O5 [M + H]+, 485.10; observed, 485.00 (Rt = 6.22
min).
Ester 29 (0.400 g, 0.825 mmol) was treated according
to the general procedure for the reduction of esters to alcohols.
The crude product was purified by flash column chromatography, eluting
with a gradient of 10–25% EtOAc in n-heptane,
to furnish alcohol 33 (0.220 g, 58%) as colorless oil. 1H NMR (400 MHz, CDCl3): δ (ppm) 7.83 (1H,
s, pyrrole H-2), 7.77–7.58 (2H, m, ArH-3 and ArH-5), 7.58 (1H,
app. t, J = 8.0 Hz, ArH-4), 6.48 (1H, s, pyrrole
H-4), 4.42 (2H, br s, CH2OH), 2.50 (3H,
s, NCCH3), 1.63 (9H, s, C(CH3)3); 13C NMR (100 MHz, CDCl3): δ (ppm) 164.8 (C-5), 158.7 (C-3), 149.1 (NCO2), 136.5 (ArC-2), 133.4 (pyrrole C-5), 133.1
(ArC-3), 132.2 (q, J = 31.4 Hz, ArC-6), 131.0 (ArC-4),
126.8 (ArC-1), 124.9 (q, J = 4.9 Hz, ArC-5), 121.5
(q, J = 274.4 Hz, CF3), 120.7 (pyrrole C-2), 112.9 (C-4), 112.3 (pyrrole C-3), 109.9 (pyrroleC-4), 84.6 (C(CH3)3), 53.9
(CH2OH), 28.0 (C(CH3)3), 15.4 (NCCH3).
LC–MS (ESI): calcd for C21H20ClF3N2O4 [M + H]+, 457.11; observed,
456.92 (Rt = 5.53 min).
Biophysical Assays
RORγt-LBD Expression
and Purification
His-tagged RORγt LBD was expressed
and purified as described
in previous studies.[28]
TR-FRET Coactivator Recruitment Assay
Assays were conducted
using 100 nM N-terminal
biotinylated SRC-1 box2 coactivator peptide (Biotin-N-PSSHSSLTARHKILHRLLQEGSPSD-CONH2) and 20 nM His6-RORγt-LBD in buffer containing
10 mM HEPES (pH 7.5), 150 mM NaCl, 5 mM dithiothreitol (DTT), 0.1%
bovine serum albumin (BSA) (w/v), and 0.1 mM CHAPS. A terbium-labeled
anti-His antibody (CisBio Bioassays, 61HISTLA) and D2-labeled streptavidin
(CisBio Bioassays, 610SADLA) were used at the concentrations recommended
by the supplier. Compounds (dissolved in DMSO) were titrated using
a 2× dilution series in Corning white low volume, low binding,
384-well plates at a final volume of 10 μL. The final DMSO concentration
was 1% v/v throughout. The plate was directly measured and centrifuged
before reading (excitation = 340 nm; emission = 665 and 620 nm) on
a Tecan infinite F500 plate reader using the parameters recommended
by CisBio Bioassays. The data were analyzed with GraphPad Prism 7.0
software. The dose–response curve was fitted represented bywhere y is the FRET ratio
((acceptor/donor) × 1000), A1 is
the bottom asymptote, A2 is the top asymptote, p is the Hill slope, x is the ligand concentration
in μM, and x0 is the IC50 value in μM. Where dose–response curves did not reach
a bottom asymptote, this was fixed at the value of the negative control.
The data were recorded in triplicate in three independent experiments
(one representative data set shown). Error bars represent the SD of
the mean.
Competition TR-FRET Coactivator
Recruitment
Assay
Competition assays were performed in an analogous fashion
to that described above, only in the presence of fixed concentrations
of cholesterol: 0 μM (DMSO), 0.25 μM, and 1.0 μM
such that the final concentration of DMSO remained at 1.2% v/v.
TR-FRET AlexaFluor-MRL-871 Recruitment Assay
Assays were conducted using 100 nM AlexaFluor647-labeled MRL-871
and 20 nM His6-RORγt-LBD in buffer as described above.
A terbium-labeled anti-His antibody (CisBio Bioassays, 61HISTLA) was
used at the concentrations recommended by the supplier. The assay
was carried out in the same manner as described above.
Thermal Shift Assay
TSA were performed
using 40 μL samples containing 5 μM RORγt-LBD, 10
μM compound, and 2.5x SYPRO Orange (Sigma) in buffer containing
150 mM NaCl, 10 mM HEPES (pH 7.5), and 1% DMSO. Hard-Shell 96-well
PCR plates (low profile, thin wall, skirted, green/white #hsp9645)
were used. The samples were heated from 25 to 80 °C at a rate
of 0.5 °C per 5 s in a CFX96 touch real-time PCR detection system
(Bio-Rad). After each increment, Sypro Orange intensity was measured
using the plate read option in Bio-Rad CFX Manager 3.1. Melting temperatures
were determined by Bio-Rad CFX Manager 3.1 software in negative mode.
The data were recorded in triplicate in three independent experiments
(one representative data set shown). Error bars represent the SD of
the mean.
Quantitative IL-17a mRNA
RT-PCR Assay
Cell culture, RT-PCR experiments, and data analysis
were performed
as described in previous studies.[28] Statistical
analysis was performed using a one-way ANOVA, comparing against the
DMSO control following the Dunnett’s post hoc test (GraphPad Prism 7.0 software). A P value <
0.05 was considered statistically significant. (Data recorded in triplicate
from two independent experiments. Data shown are representative of
two independent experiments. Error bars represent the SD of the mean.)
Protein X-ray Crystallography
RORγt-LBD
Expression and Purification
(Used for Crystallography)
His-tagged RORγt LBD containing
a C455H mutation was expressed and purified as described in previous
studies.[28]
X-ray
Crystallography
X-ray crystallography
experiments and data analysis were performed as described in previous
studies.[28] Diffraction data were collected
at the P11 beamline of the PETRA III facility at DESY (Hamburg, Germany),
the ID30B beamline of ESRF (Grenola, France), and the i04 beamline
of Diamond Light Source (Didcot, UK). The structures of RORγtC455H
in complex with 3, 9, 10, and 11 were deposited in the Protein Data Bank (PDB) under codes 7NPC, 7NP5, 7NEC, and 7NP6.
TR-FRET competition assays were performed in
an analogous fashion to that described above, only using 10 nM His6-PPARγ LBD instead of 20 nM His6-RORγt
LBD and 200 nM N-terminal biotinylated PGC1a coactivator
peptide (Biotin-N-GTPPPQEAEEPSLLKLLLAPANTQ-CONH2) instead
of 100 nM SRC-1 box 2 coactivator peptide.
Competition
TR-FRET Coactivator Recruitment
Assays on PPARγ
TR-FRET competition assays were performed
in an analogous fashion to that described above, only using 100 nM
His6-PPARγ LBD instead of 20 nM His6-RORγt
LBD. The assay was performed in the presence of 1 μM rosiglitazone,
in order to initially activate PPARγ.
TSA
on PPARγ
TSA were performed
in an analogous fashion to that described above, only using PPARγ
LBD instead of RORγt LBD.
FP
Assays on FXR
Assays were conducted
using 100 nM N-terminal FITC-labeled SRC-1 box2 coactivator
peptide (FITC-N-PSSHSSLTARHKILHRLLQEGSPSD-CONH2) and 1
μM GST-tagged FXR-LBD in buffer containing 10 mM HEPES (pH 7.5),
150 mM NaCl, 5 mM DTT, 0.1% BSA (w/v), and 0.1 mM CHAPS. For the competition
assay, the assay was performed in the presence of 0, 1, 10, or 50
μM of ligand 3. Compounds (dissolved in DMSO) were
titrated using a 2× dilution series in Corning black round-bottom,
low-volume, low-binding, 384-well plates at a final volume of 10 μL.
The final DMSO concentration was 1% v/v throughout. The plate was
directly measured (excitation = 485 nm; emission = 535 nm) on a Tecan
infinite F500 plate reader. The data were analyzed with GraphPad Prism
7.0 software. The dose–response curve was fitted in the same
way as described for the TR-FRET data on RORγt. The data were
recorded in triplicate in three independent experiments (one representative
data set shown). Error bars represent the SD of the mean.
ADME Experiments
Kinetic
Solubility
Aqueous solubility
of compounds was determined by the spectrophotometrical measurement
of the kinetic solubility of a 500 μM compound solution in HEPES
buffer pH 7.4 compared to a solution containing 50% of the organic
solvent acetonitrile. To this end, saturated samples of the compounds
in buffer were prepared starting from DMSOstocks, and samples were
shaken for 90 min at room temperature. The precipitated material was
removed by filtration and samples were further diluted with the same
volume of acetonitrile. Absorbance spectra (250–500 nm) were
recorded and relative solubility was calculated in comparison to the
acetonitrile–buffer solution of the compounds using absorbance
ratios.
Parallel Artificial Membrane Permeability
Assay
Permeability through artificial membranes [parallel
artificial membrane permeability assay (PAMPA)] was performed at an
initial concentration of 500 μM of the compound in the donor
compartment. After an incubation period of 20 h, the absorption of
the receiver wells was measured by spectrophotometry and permeation
was calculated by the normalization of the compound flux across a
blank filter.
Microsomal Stability—Phase
I
Metabolic stability under oxidative conditions was measured
using
NADPH-supplemented human liver microsomes. Compound depletion was
analyzed by liquid chromatography with tandem mass spectrometry (LC–MS/MS)
at a concentration of 1 μM over a time up to 50 min at 37 °C.
Based on compound half-life t1/2, in vitro
intrinsic clearance Clint was calculated. , with V = assay volume
and mg = amount of microsomal protein.
Microsomal
Stability—In Vitro Glucuronidation
Metabolic stability under conjugative conditions
was measured in the glucuronidation assay by the LC–MS-based
determination of % remaining of selected compounds. Prior to the assay,
human liver microsomes were activated using alamethicin and further
supplemented with 5 mM of the cofactor UDPGA. 5 μM compound
was added to the reaction mixture, and it was incubated for 1 h at
37 °C.
Plasma Stability
Plasma stability
was measured by the LC–MS-based determination of % remaining
of selected compounds at a concentration of 5 μM after incubation
in 50% plasma in phosphate-buffered saline (obtained from humans)
for 1 h at 37 °C.
Plasma Protein Binding
Plasma protein
binding was determined by equilibrium dialysis. Plasma containing
5 μM of test compound was allowed to equilibrate with the buffer
compartment for 6 h at 37 °C. Compound concentrations on both
sides of the semipermeable membrane were analyzed by LC–MS/MS
and % bound was calculated.
Figure 1
Scheme 1
Figure 2
Scheme 2
Figure 3
Figure 4