The recent crystal structures of CC chemokine receptors 2 and 9 (CCR2 and CCR9) have provided structural evidence for an allosteric, intracellular binding site. The high conservation of residues involved in this site suggests its presence in most chemokine receptors, including the close homologue CCR1. By using [3H]CCR2-RA-[ R], a high-affinity, CCR2 intracellular ligand, we report an intracellular binding site in CCR1, where this radioligand also binds with high affinity. In addition, we report the synthesis and biological characterization of a series of pyrrolone derivatives for CCR1 and CCR2, which allowed us to identify several high-affinity intracellular ligands, including selective and potential multitarget antagonists. Evaluation of selected compounds in a functional [35S]GTPγS assay revealed that they act as inverse agonists in CCR1, providing a new manner of pharmacological modulation. Thus, this intracellular binding site enables the design of selective and multitarget inhibitors as a novel therapeutic approach.
The recent crystal structures of CC chemokine receptors 2 and 9 (CCR2 and CCR9) have provided structural evidence for an allosteric, intracellular binding site. The high conservation of residues involved in this site suggests its presence in most chemokine receptors, including the close homologue CCR1. By using [3H]CCR2-RA-[ R], a high-affinity, CCR2 intracellular ligand, we report an intracellular binding site in CCR1, where this radioligand also binds with high affinity. In addition, we report the synthesis and biological characterization of a series of pyrrolone derivatives for CCR1 and CCR2, which allowed us to identify several high-affinity intracellular ligands, including selective and potential multitarget antagonists. Evaluation of selected compounds in a functional [35S]GTPγS assay revealed that they act as inverse agonists in CCR1, providing a new manner of pharmacological modulation. Thus, this intracellular binding site enables the design of selective and multitarget inhibitors as a novel therapeutic approach.
Chemokines are chemotactic
cytokines that control the migration
and positioning of immune cells during physiological and pathological
conditions by interacting with more than 20 different chemokine receptors.[1] Chemokine receptors mainly belong to the class
A of G protein-coupled receptors (GPCRs) and can be divided into four
different subtypes, namely C, CC, CXC, and CX3C, according to the
pattern of specific cysteine residues in their major endogenous chemokines.[2] To exert their function, chemokines bind at the
extracellular side of their receptors in a binding mechanism involving
the N-terminal domain, extracellular loops, and the upper half of
the transmembrane bundle.[3,4] After activation, most
chemokine receptors signal through heterotrimeric G proteins, mainly
Gi/o class, and β-arrestins.[2] CC chemokine receptors 1 (CCR1) and 2 (CCR2) are two of the 10 members
of the CC subtype of chemokine receptors. CCR1 and CCR2 are expressed
in a variety of immune cells, such as monocytes, dendritic cells,
and T helper type-1 (TH1) cells, from where they regulate
diverse inflammatory and homeostatic functions.[5] Multiple chemokines activate these two receptors, including
CCL3, CCL5, and CCL8 in the case of CCR1, and CCL2, CCL7, and CCL8
in the case of CCR2.[2]Dysregulation
of CCR1, CCR2, and their ligands has been linked
to several inflammatory and immune diseases,[6,7] which
has resulted in many drug discovery efforts to develop small molecules
that target these receptors.[8,9] Several lines of evidence
support a role for both CCR1 and CCR2 in the pathogenesis of diseases
such as rheumatoid arthritis (RA) and multiple sclerosis (MS): increased
expression of both receptors and their ligands in disease models and
patients,[10,11] protective effect of genetic knockout of
CCR1 or CCR2 in disease models,[12,13] and positive preclinical
studies with chemokine-neutralizing monoclonal antibodies or small-molecule
inhibitors of CCR1 or CCR2.[14−16] Yet, only few clinical studies
have shown promising results,[17,18] while most of the drugs
developed so far have failed in clinical trials due to lack of efficacy.[8,9] In this regard, the development of multitarget drugs has been proposed
as a strategy to overcome the lack of efficacy. Multitarget drugs
are designed to specifically act on more than one drug target, which
might be necessary in highly heterogeneous diseases, such as RA and
MS, where more than one chemokine receptor is involved.[19] The design of dual antagonists has been previously
undertaken for CCR1/CCR3,[20] CCR2/CCR5,[21] CCR5/CXCR4,[22] and
CXCR1/CXCR2;[23] however, no CCR1/CCR2 dual
antagonists have so far been reported.Recently, the crystal
structures of CCR2[24] and CCR9[25] have revealed a novel allosteric
binding site for small molecules in chemokine receptors. Both CCR2-RA-[R] in CCR2 and vercirnon in CCR9 bind in a pocket located
in the intracellular surface of the receptors, partially overlapping
with the binding site for G proteins and β-arrestins.[24,25] These intracellular ligands can inhibit the receptors in a noncompetitive
and insurmountable manner with regard to chemokine binding, as demonstrated
previously in CCR2.[26] This might result
in higher efficacy even in the presence of a high local concentration
of chemokines during a disease state. Together with the potential
advantages of allosteric modulators of chemokine receptors, this intracellular
binding site seems to be quite conserved among chemokine receptors,
which suggests the presence of homologous pockets in other receptors
such as CCR1.[27] This conservation might
provide an opportunity for the design of both selective and dual-targeting
inhibitors of CCR1 and CCR2 as a novel approach to treat inflammatory
and immune diseases.For CCR2, several compounds belonging to
different scaffolds have
already been reported to bind to this intracellular binding site,
including pyrrolone derivatives such as CCR2-RA-[R], sulfonamide derivatives, and 2-mercapto imidazoles.[26,28] When tested for selectivity, some of these compounds also displayed
a moderate activity on CCR1,[29−31] suggesting that they might also
bind to CCR1. Thus, we selected the pyrrolone scaffold to explore
a potential intracellular binding site in CCR1. In our current study,
we report the synthesis and the biological evaluation of novel and
previously patented pyrrolone derivatives[32,33] at both CCR1 and CCR2 in order to determine their selectivity and
structure–affinity relationships (SAR) for both receptors.
Finally, compounds were tested in a [35S]GTPγS binding
assay in order to determine their functional effects in CCR1 and CCR2.
Overall, our results provide evidence that CCR1 can also be targeted
with intracellular allosteric modulators and that this binding site
can be used for the design of multitarget compounds.
Results and Discussion
Synthesis
of Pyrrolone Derivatives
The racemic pyrrolones
(6–24, 26–46) depicted in Scheme were synthesized via a one-pot three-component condensation
reaction, starting from the commercially available substituted aldehydes 1a–l, anilines 2a–q, and ethyl 2,4-dioxo-butanoates 3a–i in acetic acid[33] (6–23, 26–46)
or THF[29] (24). The ethyl 2,4-dioxo-butanoates
(3b–d,f,i), which were not commercially available, were prepared by a Claisen
condensation starting from the methyl ketones (4b–d,f,i) and diethyl oxalate 5.[34] Pyrrolone 25 was
prepared via a transesterification of 24 by the use of p-toluenesulfonic acid in 2-propanol.
Scheme 1
Synthesis Route of
Pyrrolones 6–48, with Different R1, R2, and R3 Substituents
Reagents and conditions: (a)
acetic acid, reflux for 2–4 h or THF, rt, overnight; (b) Na,
EtOH, 0–20 °C, overnight; (c) p-toluenesulfonic
acid, 2-propanol, reflux, 48 h.
Synthesis Route of
Pyrrolones 6–48, with Different R1, R2, and R3 Substituents
Reagents and conditions: (a)
acetic acid, reflux for 2–4 h or THF, rt, overnight; (b) Na,
EtOH, 0–20 °C, overnight; (c) p-toluenesulfonic
acid, 2-propanol, reflux, 48 h.
Characterization
of [3H]-CCR2-RA-[R] Binding on CCR1 and
CCR2
[3H]-CCR2-RA-[R] is the
(R)-isomer of [3H]-CCR2-RA,
a high-affinity radioligand previously characterized in our group
for CCR2.[26] To avoid a possible effect
of the lower-affinity isomer, we used the tritium-labeled (R)-isomer in the present study. As expected, [3H]-CCR2-RA-[R] binds with high affinity to osteosarcoma
(U2OS) cells stably expressing CCR2b (U2OS-CCR2) as shown by saturation
experiments (KD of 6.3 nM and Bmax of 2.6 pmol/mg, Supporting Information, Figure S1 and Table S1). Kinetic characterization
showed that [3H]-CCR2-RA-[R] associates
and dissociates in a biphasic manner (Supporting Information, Table S1), consistent with the previously reported
[3H]-CCR2-RA kinetics.[26] We
had reported that [3H]-CCR2-RA binds with low affinity
to CCR5 (KD of 100 nM),[28] suggesting that CCR2-RA-[R] is a nonselective
antagonist that can bind several chemokine receptors. In this regard,
CCR1 is a close homologue of CCR2, with 61% amino acid similarity
and 47% identity; furthermore, this amino acid similarity is >90%
when only considering the amino acid residues involved in the intracellular
binding site of CCR2-RA-[R] in CCR2[24] (Supporting Information, Figure S2). This prompted us to investigate the binding of [3H]-CCR2-RA-[R] in membrane preparations from U2OS cells stably expressing
CCR1 (U2OS-CCR1). [3H]-CCR2-RA-[R] homologous
displacement assays on U2OS-CCR1 yielded a KD of 13.5 nM and a Bmax of 6.1
pmol/mg (Figure a,
Supporting Information, Table S1), suggesting
the presence of an intracellular site in CCR1 and making it a suitable
tool to study such binding pocket. Binding of [3H]-CCR2-RA-[R] to U2OS-CCR1 was also assessed in kinetic experiments
at 25 °C. These experiments showed that [3H]-CCR2-RA-[R] associates and dissociates in a biphasic manner, similar
to our findings in CCR2, but the association and dissociation rates
were significantly higher in CCR1 than in CCR2 (Supporting Information, Figure S1 and Table S1).
Figure 1
(a) Homologous displacement
curves of 3, 6, and 12 nM [3H]-CCR2-RA-[R] specific binding by increasing concentrations
of CCR2-RA-[R] in U2OS-CCR1 at 25 °C. (b) Displacement
curves of 6 nM [3H]-CCR2-RA-[R] specific
binding by increasing concentrations of SD-24, JNJ-27141491, and BX471
in U2OS-CCR1 at 25 °C. BX471 significantly enhanced the binding
of [3H]-CCR2-RA-[R] up to 120%. Statistical
significance between binding in absence (100%) and presence of 10
μM BX471 (116 ± 2%) was determined using an unpaired, two-tailed
Student’s t-test with Welch’s correction.
(c,d) Displacement curves of 6 nM [3H]-CCR2-RA-[R] specific binding by compounds 39, 41, 43, and 45 (b) in U2OS-CCR1
or (c) in U2OS-CCR2 at 25 °C. In the case of U2OS-CCR2, compound 45 did not displace more than 50% of [3H]-CCR2-RA-[R], thus only single-point data at 1 μM is shown.
The dashed blue line corresponds to the nonlinear regression fit for
compound 45 by GraphPad Prism 7.0. Data shown are mean
± SEM of at least three experiments performed in duplicate.
(a) Homologous displacement
curves of 3, 6, and 12 nM [3H]-CCR2-RA-[R] specific binding by increasing concentrations
of CCR2-RA-[R] in U2OS-CCR1 at 25 °C. (b) Displacement
curves of 6 nM [3H]-CCR2-RA-[R] specific
binding by increasing concentrations of SD-24, JNJ-27141491, and BX471
in U2OS-CCR1 at 25 °C. BX471 significantly enhanced the binding
of [3H]-CCR2-RA-[R] up to 120%. Statistical
significance between binding in absence (100%) and presence of 10
μM BX471 (116 ± 2%) was determined using an unpaired, two-tailed
Student’s t-test with Welch’s correction.
(c,d) Displacement curves of 6 nM [3H]-CCR2-RA-[R] specific binding by compounds 39, 41, 43, and 45 (b) in U2OS-CCR1
or (c) in U2OS-CCR2 at 25 °C. In the case of U2OS-CCR2, compound 45 did not displace more than 50% of [3H]-CCR2-RA-[R], thus only single-point data at 1 μM is shown.
The dashed blue line corresponds to the nonlinear regression fit for
compound 45 by GraphPad Prism 7.0. Data shown are mean
± SEM of at least three experiments performed in duplicate.Overall, these findings allowed
us to set up a [3H]-CCR2-RA-[R] competitive
displacement assay on both U2OS-CCR1 and
U2OS-CCR2 to determine the binding affinity (Ki) of unlabeled compounds. Using this assay, we first determined
the ability of known ligands to displace this radioligand from CCR1,
i.e., the CCR2 intracellular ligands SD-24 and JNJ-27141491[26,28] and the CCR1 orthosteric antagonist BX471[35] (Figure b). SD-24
and JNJ-27141491 fully displaced [3H]-CCR2-RA-[R] from CCR1 in a concentration-dependent manner, indicating
that these compounds bind at the same binding site as CCR2-RA-[R]. SD-24 displaced the radioligand with a pKi of 7.45 ± 0.05 (Ki =
36 nM), while JNJ-27141491 displaced [3H]-CCR2-RA-[R] with a pKi of 6.9 ±
0.06 (Ki = 138 nM), consistent with previously
reported activities in CCR1.[30,31] To rule out that these
compounds bind at the orthosteric binding site of CCR1, we also investigated
the effect of BX471 in [3H]-CCR2-RA-[R] binding. As expected, BX471 was not able to displace the radioligand
(Figure b); on the
contrary, BX471 significantly enhanced the binding of [3H]-CCR2-RA-[R] by approximately 20% (116 ±
2% in the presence of 10 μM BX471), in a similar manner as previously
reported with CCR2 orthosteric antagonists.[24,26] This allosteric enhancement is consistent with two different binding
sites in CCR1: the orthosteric binding site where BX471 binds and
an intracellular pocket for CCR2-RA-[R], SD-24, and
JNJ-27141491.This [3H]-CCR2-RA-[R] assay was also
used to determine the affinity of the synthesized pyrrolone derivatives.
All pyrrolone derivatives 6–46 were
first tested at a single concentration of 1 μM in both U2OS-CCR1
and U2OS-CCR2 (Tables –3). Compounds which displaced more
than 50% of [3H]-CCR2-RA-[R] binding were
further evaluated in this assay using at least six different concentrations
of unlabeled compound in order to determine their binding affinity
for the corresponding receptor subtypes (Figure c,d and Tables –3). Finally,
we selected four compounds (39, 41, 43 and 45) to be tested in a functional [35S]GTPγS binding assay (Figure ). The potency (pIC50) of these
compounds was determined in the presence of an EC80 concentration
of CCL3 (8 nM) or CCL2 (20 nM) in U2OS-CCR1 or U2OS-CCR2 membranes,
respectively.
Table 1
Binding Affinities
of Compounds 6–26 on Human CCR1 and
Human CCR2
pKi ± SEM (Ki, nM)a or
displacement at 1 μM (%)b
compd
R1
R3
CCR1
CCR2
6
c-hexyl
Me
7.26 ± 0.04 (56)
7.10 ± 0.03 (81)
7
c-heptyl
Me
7.26 ± 0.03 (56)
7.02 ± 0.06 (96)
8
c-octyl
Me
7.24 ± 0.01 (57)
6.79 ± 0.09 (170)
9
Ph
Me
6.79 ± 0.04 (162)
39% (38, 40)
10
4-Me Ph
Me
6.71 ± 0.06 (198)
36% (42, 31)
11
4-OMe Ph
Me
6.27 ± 0.01 (541)
5% (5, 5)
12
4-Cl Ph
Me
7.17 ± 0.01 (67)
6.70 ± 0.08 (207)
13
4-Br Ph
Me
7.07 ± 0.07 (87)
6.67 ± 0.03 (214)
14
3-Me Ph
Me
47% (51, 44)
11% (14, 8)
15
3-OMe Ph
Me
28% (34, 22)
0% (3, −3)
16
3-Cl Ph
Me
6.70 ± 0.01 (198)
19% (25, 14)
17
3-Br Ph
Me
6.74 ± 0.02 (181)
19% (20, 18)
18
c-hexyl
Et
7.52 ± 0.01 (30)
6.99 ± 0.06 (104)
19
c-hexyl
Pr
7.54 ± 0.04 (29)
6.86 ± 0.10 (144)
20
c-hexyl
Bu
7.50 ± 0.004 (31)
6.81 ± 0.05 (158)
21
c-hexyl
I-Pr
7.39 ± 0.06 (42)
6.50 ± 0.05 (316)
22
c-hexyl
c-Pr
7.74 ± 0.08 (19)
6.80 ± 0.05 (160)
23
c-hexyl
t-Bu
7.66 ± 0.05 (22)
6.81 ± 0.07 (158)
24
c-hexyl
OEt
6.70 ± 0.01 (200)
31% (36, 26)
25
c-hexyl
OiPr
36% (45, 26)
6% (10, 1)
26
c-hexyl
–Ph
7.11 ± 0.01 (77)
37% (45, 30)
pKi and Ki (nM) values obtained from [3H]-CCR2-RA-[R] binding assays on U2OS membranes stably expressing human
CCR1 or human CCR2. Values are means ± standard error of the
mean (SEM) of at least three independent experiments performed in
duplicate.
Percent of [3H]-CCR2-RA-[R] displacement by 1 μM
compound. Values represent
the mean of two independent experiments performed in duplicate.
Table 3
Binding Affinities of Compounds 43–46 on Human CCR1 and Human CCR2
pKi ± SEM
(Ki, nM)a or
displacement at 1 μM (%)b
compd
R1
R3
CCR1
CCR2
43
c-hexyl
c-propyl
8.27 ± 0.02 (5)
7.82 ± 0.04 (15)
44
c-hexyl
Ph
7.56 ± 0.04 (28)
7.18 ± 0.03 (66)
45
3-Br Ph
c-propyl
7.30 ± 0.01 (50)
45% (49, 42)
46
3-Br Ph
Me
7.19 ± 0.02 (65)
6.67 ± 0.01 (216)
pKi and Ki (nM) values
obtained from [3H]-CCR2-RA-[R] binding
assays on U2OS membranes stably expressing human CCR1 or
human CCR2. Values are means ± standard error of the mean (SEM)
of at least three independent experiments performed in duplicate.
Percent of [3H]-CCR2-RA-[R] displacement by 1 μM compound. Values represent
the mean of two independent experiments performed in duplicate.
Figure 3
(a) [35S]GTPγS binding upon stimulation of U2OS-CCR1
and U2OS-CCR2 by increasing concentrations of CCL3 and CCL2, respectively.
In both cases, the response was corrected by subtracting the basal
activity (approximately 8000 dpm for both CCR1 and CCR2). (b) Inhibition
of CCL3-induced [35S]GTPγS binding by compounds 39, 41, 43, and 45 in
U2OS-CCR1. (c) Inhibition of CCL2-induced [35S]GTPγS
binding by compounds 39, 41, 43, and 45 in U2OS-CCR2. The level of basal activity in
U2OS-CCR1 and U2OS-CCR2 is indicated by a dashed line. In all cases,
data shown are mean ± SEM of at least three experiments performed
in duplicate.
Docking of CCR2-RA-[R] in
CCR1 and CCR2
To better understand the binding mode of CCR2-RA-[R] in both humanCCR1 and CCR2b, we docked this compound
into models
of both receptors (Figure ). In the case of CCR2, homology modeling
was used to model the CCR2 residues between Ser2265x62 and
Lys2406x32, which correspond to the M2 muscarinic acetylcholine
receptor sequence in the CCR2b crystal structure (PDB 5T1A).[24] These residues were modeled because this region is in close
proximity to the CCR2-RA-[R] binding site. As expected
from the sequence alignment (Supporting Information, Figure S2), CCR2-RA-[R] was predicted to
bind to CCR1 in an overlapping binding site as the one reported in
the crystal structure of CCR2,[24] in a solvent-exposed
intracellular pocket found between the intracellular ends of transmembrane
segments 1–3, 6, 7, and helix 8 (Figure ). The vinylogous carboxylic acid functionality
makes similar interactions in CCR1 as in CCR2: the hydroxyl and the
two carbonyl groups are involved in hydrogen-bond interactions with
the side chain of Arg1313x50, and the backbone of Arg3078x49 and Phe3088x50 (Figure ). A similar hydrophobic subpocket is also
observed around the cyclohexyl moiety, which interacts with Ala6x33, Val/Leu6x36, Ile6x37, and Ile6x40. Interestingly, Val2446x36 in CCR2 is replaced
by the bigger Leu2406x36 in CCR1, which pushes the ligand
down against Arg1313x50, resulting in a slightly different
binding orientation of CCR2-RA-[R] in this receptor
(Figure ). In addition,
the exchange of Lys3118x49 in CCR2 by Arg3078x49 in CCR1 might also contribute to the stabilization of this slightly
altered binding pose. This difference in orientation could result
in CCR1 selectivity, as this orientation seems to open up the subpockets
in the proximity of the cyclohexyl and the acetyl group of CCR2-RA-[R] in CCR1, allowing the introduction of bigger and more
lipophilic substituents at these positions.
Figure 2
Proposed binding mode
of compound CCR2-RA-[R]
in the homology models of CCR1 and CCR2, based on the crystal structure
of CCR2 (PDB 5T1A).[24] For CCR1, representative residues
are shown as green “sticks” and for CCR2 as orange “sticks”.
In all cases, oxygen and nitrogen atoms are represented in red and
blue, respectively, and hydrogen bonds with dashed yellow lines. Residues
are numbered based on the corresponding residue numbers and with structure-based
Ballesteros–Weinstein numbers in superscript.[37]
Proposed binding mode
of compound CCR2-RA-[R]
in the homology models of CCR1 and CCR2, based on the crystal structure
of CCR2 (PDB 5T1A).[24] For CCR1, representative residues
are shown as green “sticks” and for CCR2 as orange “sticks”.
In all cases, oxygen and nitrogen atoms are represented in red and
blue, respectively, and hydrogen bonds with dashed yellow lines. Residues
are numbered based on the corresponding residue numbers and with structure-based
Ballesteros–Weinstein numbers in superscript.[37]
Structure–Affinity
Relationships (SAR)
Modifications Replacing the Cyclohexyl Group
(R1, Table )
Several
pyrrolone derivatives have been previously evaluated at CCR2,[29,32,33,36] resulting in the identification of CCR2-RA-[R]
as a hit compound for further development,[29] but characterization of these compounds in CCR1 is mostly missing.
Compound 6, previously reported and characterized in
CCR2 by Zou et al. (2007),[36] was selected
as our starting point for the analysis of SAR in both CCR1 and CCR2.
In our assay, compound 6 showed an affinity of 81 nM
for CCR2 and a slightly higher affinity of 56 nM for CCR1 (Table ). To note, the binding affinities reported previously for
these pyrrolone derivatives were obtained with a 125I-CCL2
binding assay,[29,36] resulting in lower affinities
compared with our [3H]-CCR2-RA-[R] binding
assay, as previously observed in our group.[26] For our SAR study, we first examined different C5 substituents of
the pyrrolone core (R1), as shown in Table . In line with previous studies,[29] we found that increasing the size of the cycloalkyl
group from cyclohexyl (6) to cycloheptyl (7) or cyclooctyl (8) resulted in a decrease in binding
affinity for CCR2; however, the affinity for CCR1 was retained, indicating
that bulkier groups are better tolerated in CCR1 than in CCR2 and
providing an avenue for selectivity on CCR1 over CCR2. Previous studies
showed that decreasing the size of the cycloalkyl group was also detrimental
for CCR2,[29] so we decided not to explore
smaller ring sizes.pKi and Ki (nM) values obtained from [3H]-CCR2-RA-[R] binding assays on U2OS membranes stably expressing humanCCR1 or humanCCR2. Values are means ± standard error of the
mean (SEM) of at least three independent experiments performed in
duplicate.Percent of [3H]-CCR2-RA-[R] displacement by 1 μM
compound. Values represent
the mean of two independent experiments performed in duplicate.Substitution of the cycloalkyl group
by a phenyl group (9) led to a great loss of CCR2 affinity
(39% displacement at 1 μM),
consistent with previously reported values showing a decreased affinity
for an almost similar pair of compounds.[36] Yet this substitution only led to a 3-fold decrease in CCR1 affinity
(Ki of 162 nM), thus showing much higher
selectivity for CCR1. Next, we explored the effect of N-aryl modifications in both affinity and selectivity (compounds 10–17), specifically the effect of para
and meta substituents. In general, N-aryl groups
on the R1 position resulted in increased selectivity toward
CCR1, as most compounds did not displace more than 36% [3H]-CCR2-RA-[R] binding in CCR2 at a concentration
of 1 μM. Only compounds 12 and 13,
with halogen substitutions in para position (Cl and Br, respectively),
regained CCR2 affinity (12, 207 nM; 13,
214 nM). Furthermore, para-substituted derivatives displayed significantly
higher affinities compared with their meta-substituted analogues.In the case of CCR1, introduction of a para-methyl
moiety (10) resulted in a slight decrease in affinity
compared with the unsubstituted 9; in contrast, the meta-substituted
analogue (14) showed less than 50% displacement at 1
μM. Introduction of an electron-donating substituent (methoxy, 11 and 15) was not well tolerated in any position,
as it led to an approximately 3-fold decrease in affinity when placed
in para position (11, 541 nM) and a near complete loss
of affinity when placed in meta position (15, 28% displacement
at 1 μM). Halogen substituents in para position were also more
favored in the case of CCR1, yielding higher affinities compared with
the unsubstituted 9 and regardless of the halogen used
(67 nM for R1 = 4-Cl phenyl (12), p < 0.0001 to 9; 87 nM for R1 = 4-Br phenyl (13); p = 0.0002 to 9). However, selectivity for CCR1 was notably reduced considering
that these compounds displayed binding affinities of around 200 nM
in CCR2. Although moving the halogens to the meta position (16 and 17) decreased the affinities more than
2-fold compared with their para analogues, selectivity for CCR1 was
restored as these compounds showed less than 20% displacement of [3H]-CCR2-RA-[R] binding in CCR2. Together,
the results for compounds 6–17 indicate
that in CCR1 aliphatic groups yield higher affinities, while aromatic
groups yield lower affinities but improved selectivity over CCR2.
Modifications to the Acetyl Group (R3, Table )
Previous modifications
to the vinylogous carboxylic acid functionality in CCR2 showed detrimental
effects in binding affinity.[29,36] Indeed, mutagenesis
and structural studies have shown crucial interactions of the hydroxyl
and the two carbonyl groups with Glu3108x48, Lys3118x49, and Phe3128x50 (residues according to structure-based
Ballesteros–Weinstein numbering[37]) in CCR2.[24,28] Sequence alignment of CCR1 and
CCR2 (Supporting Information, Figure S2) and our docking study (Figure ) suggest similar interactions in CCR1, as only position
8.49 differs (arginine in CCR1 and lysine in CCR2). Therefore, we
decided to keep the vinylogous carboxylic acid moiety and explore
different modifications to the acetyl group at the R3 position
(Table ). A gradual
increase in the length of the alkyl chain from a methyl group (6) to a butyl group (18–20) resulted in
a ∼2-fold increase in CCR1 affinity (30 nM for R3 = ethyl (18), p = 0.0004 against 6; 29 nM for R3 = propyl (19), p = 0.0002 against 6; and 31 nM for R3 = butyl (20), p = 0.0010 against 6). In contrast, for CCR2, we observed a similar or a slight
decrease in affinity. Introduction of a bulkier isopropyl group led
to a decrease in affinity in both receptors, with a more drastic effect
in CCR2 affinity. Replacing the isopropyl group with cyclopropyl (22) or tert-butyl (23) restored
the affinity in CCR2 to values similar to compound 20 (22, 160 nM; 23, 158 nM); in CCR1, these
modifications further improved the binding affinity to approximately
20 nM, yielding compounds with the highest affinity and selectivity
observed in these series of R1 and R3 modifications
(22, 19 nM; 23, 22 nM). These results suggest
a larger hydrophobic subpocket in CCR1, able to accommodate larger
and branched alkyl chains.We also explored the effect of adding
heteroatoms (oxygen in this case) between the carbonyl and an ethyl
or isopropyl group (24 and 25, respectively).
Overall, this led to a drastic drop in affinity for both receptors.
This detrimental effect was most pronounced in compound 25, which displaced less than 40% of [3H]-CCR2-RA-[R] binding in CCR1 and less than 10% in CCR2. The transformation
of the ketone into an ester might decrease the electron density on
the carbonyl oxygen as well as the acidity of the adjacent protons,
thus weakening or disrupting key hydrogen bonding interactions with
Lys8x49 in CCR2[24,28] or Arg8x49 in CCR1. The need of an acidic function for intracellular antagonists
has also been reported in a study with N-benzylindole-2-carboxylic
acids, where the authors found a correlation between higher acidity
and higher CCR2 affinity.[38] Finally, replacing
the methyl group in R3 with a phenyl group (26) had no effect on CCR1 affinity, while it only displaced 37% of
[3H]-CCR2-RA-[R] binding in CCR2. Altogether,
these findings indicate that bigger, more lipophilic groups in R3 are better tolerated in CCR1, while in CCR2 methyl is preferred.
Modifications to the Phenyl Ring (R2, Table )
In addition, we explored
different N-aryl modifications in the phenyl ring
(R2, Table ), starting with modifications in para position.
Removing the methyl group in 6 yielded compound 27, with an unsubstituted phenyl group, which displaced less
than 50% of the radioligand in both receptors. Increasing the size
of the alkyl group from methyl (6) to ethyl (28) caused a 3-fold decrease in CCR1 affinity, while the affinity in
CCR2 was maintained (28, 168 nM in CCR1 versus 66 nM
in CCR2). Adding an electron-donating methoxy group was unfavorable
for both receptors, as affinities dropped to 260 nM in CCR1 and 217
nM in CCR2. In contrast, an electron-withdrawing substituent (trifluoromethyl, 32) restored the affinity to 92 nM in CCR2, similar to our
starting compound 6 and to 144 nM in CCR1. The substitution
of the para-methyl group with halogens yielded derivatives
with improved binding affinities in both receptors (30 and 31) but no gain in selectivity. Substitution with
a chlorine (30) or bromine atom (31) led
to a 4.5-fold increase in CCR2 affinity compared with 6, with Ki values around 20 nM regardless
of the halogen. In the case of CCR1, the bromine atom (31) led to a 2-fold increase compared with 6 (31, 24 nM), while the smaller chlorine atom did not affect the affinity
much (30, 40 nM). Although not synthesized in our study,
Dasse et al. (2007)[29] showed that the para-fluoro analogue performed worse in CCR2 than other para-halogen derivatives. In this regard, from fluoro to
chloro there is an important increase in polarity (σ), lipophilicity
(π), and size, whereas from chloro to bromo only lipophilicity
and size increase.[39,40] Taken together, these results
suggest that lipophilicity and size of the halogen might be more important
in CCR1 than in CCR2, while electronegativity or polarity could play
a bigger role in CCR2.
Table 2
Binding Affinities
of Compounds 6, 27–42 on Human CCR1 and
Human CCR2
pKi ± SEM (Ki, nM)a or
displacement at 1 μM (%)b
compd
R2
CCR1
CCR2
27
H
42% (41, 42)
45% (44, 45)
6
4-Me
7.26 ± 0.04 (56)
7.10 ± 0.03 (81)
28
4-Et
6.78 ± 0.02 (168)
7.19 ± 0.05 (66)
29
4-OMe
6.60 ± 0.07 (260)
6.67 ± 0.05 (217)
30
4-Cl
7.41 ± 0.05 (40)
7.73 ± 0.08 (19)
31
4-Br
7.62 ± 0.05 (24)
7.80 ± 0.12 (17)
32
4-CF3
6.86 ± 0.08 (144)
7.04 ± 0.02 (92)
33
3-Me
6.31 ± 0.07 (500)
6.58 ± 0.05 (265)
34
3-F
44% (45, 42)
47% (48, 47)
35
3-Cl
6.28 ± 0.08 (541)
6.62 ± 0.02 (239)
36
3-CF3
25% (23, 27)
6.54 ± 0.11 (305)
37
2-F, 4-Me
7.56 ± 0.10 (29)
7.44 ± 0.05 (37)
38 (CCR2-RA)
2-F, 4-Cl
7.82 ± 0.06 (15)
8.00 ± 0.09 (11)
39
2-F, 4-Br
7.98 ± 0.04 (11)
8.25 ± 0.02 (6)
40
3,4-diMe
7.37 ± 0.03 (43)
7.75 ± 0.02 (18)
41
3-Me, 4-Cl
7.51 ± 0.01 (31)
8.09 ± 0.08 (9)
42
3-F, 4-Me
7.32 ± 0.07 (49)
7.24 ± 0.02 (57)
pKi and Ki (nM) values obtained
from [3H]-CCR2-RA-[R] binding assays on
U2OS membranes stably expressing human
CCR1 or human CCR2. Values are means ± standard error of the
mean (SEM) of at least three independent experiments performed in
duplicate.
Percent of [3H]-CCR2-RA-[R] displacement by 1 μM
compound. Values represent
the mean of two independent experiments performed in duplicate.
pKi and Ki (nM) values obtained
from [3H]-CCR2-RA-[R] binding assays on
U2OS membranes stably expressing humanCCR1 or humanCCR2. Values are means ± standard error of the
mean (SEM) of at least three independent experiments performed in
duplicate.Percent of [3H]-CCR2-RA-[R] displacement by 1 μM
compound. Values represent
the mean of two independent experiments performed in duplicate.Moving the substituents from the
para to the meta position resulted
in poor affinities for both receptors compared with their para-substituted
analogues. In CCR1, the meta-methyl (33) and meta-chlorine (35) groups led
to a 9-fold and 13-fold decrease in affinity, respectively; in CCR2,
the affinities decreased 3-fold and 13-fold after the same substitutions.
The addition of a trifluoromethyl group in meta position (36) also led to a 3-fold decrease in CCR2 affinity compared with its
para-substituted analogue 32. In CCR1, 36 only displaced 25% of [3H]-CCR2-RA-[R] binding at a concentration of 1 μM, displaying the highest
selectivity toward CCR2 in these series of modifications. Also detrimental
was the addition of a fluorine group in meta position (34), which led to less than 50% displacement of [3H]-CCR2-RA-[R] binding in both receptors. Overall, substituents in the
para position were more favored in both receptors, especially halogen
substituents, yet none of the compounds displayed selectivity toward
CCR1. Similarly as reported by Dasse et al. (2007),[29] attempts to introduce different substituents in the ortho
position were unsuccessful, thus we continued to explore different
combinations of phenyl substituents.As part of our SAR analysis,
we synthesized compound 38 (also referred as CCR2-RA),
which corresponds to the racemic mixture
of the radioligand [3H]-CCR2-RA-[R] used
in this study. This compound displayed an affinity of 15 nM in CCR1
and 11 nM in CCR2, similar to the KD values
obtained in homologous displacement or saturation assays (Supporting
Information, Table S1). Replacing the para-chloro group in 38 with a methyl moiety
(37), while keeping the ortho-fluorine
group, led to an expected decrease in affinity for both receptors,
as compound 6 with a methyl group in para position performed
worse than 30 with a chlorine atom in the same position.
When the para substituent was replaced with a bromine atom (39), the affinity was restored to 11 nM in CCR1 and 6 nM in
CCR2. Subsequent combinations of meta and para substituents (40–42) generated compounds with decreased
CCR1 affinities compared with 38, as expected from the
data on the monosubstituted meta analogues. Compound 41 displayed a slightly higher selectivity for CCR2 (9 nM in CCR2 versus
31 nM in CCR1). Overall, disubstituted derivatives performed better
than the monosubstituted compounds in both receptors; however, no
clear trend in selectivity was observed in these series.In
an attempt to improve both affinity and selectivity for CCR1,
we decided to combine some of the best features observed at R1, R2, and R3 positions: a disubstituted
phenyl ring with an ortho-fluoro and para-bromo moieties for R2 in order to retain the high affinity
of 39, a cyclopropyl group or an unsubstituted phenyl
ring at R3 (22 and 26) to gain
selectivity, and a meta-bromo phenyl ring at R1 (17) to further improve selectivity for CCR1.
These combinations resulted in four final compounds shown in Table (43–46). To maintain a high
affinity for CCR1, we kept the 2-fluoro-4-bromophenyl group at R2 constant, and we combined it with different R1 and R3 substituents. The combination with a cyclopropyl
group at R3 position (43) led to the highest
CCR1 affinity in our study (Ki of 5 nM),
but selectivity over CCR2 was reduced compared with 22 (3-fold versus 8-fold). Replacing the cyclopropyl group at R3 by a phenyl group (44) decreased the affinity
for CCR1 by more than 5-fold compared with 43. Compound 43, somewhat unexpectedly, bound to CCR2 with an affinity
of 66 nM, more than 15-fold better than 26. Replacing
the cyclohexyl group at R1 (43) by a 3-bromo-phenyl
group (45) resulted in an improved selectivity over CCR2,
as this compound did not displace more than 50% of [3H]-CCR2-RA-[R] binding at 1 μM, whereas it showed an affinity
of 50 nM in CCR1. Finally, replacing the cyclopropyl with a methyl
group at R3 (46) maintained the affinity for
CCR1 and restored the affinity for CCR2 (65 nM in CCR1 and 216 nM
in CCR2), with a concomitant loss of selectivity.pKi and Ki (nM) values
obtained from [3H]-CCR2-RA-[R] binding
assays on U2OS membranes stably expressing humanCCR1 or
humanCCR2. Values are means ± standard error of the mean (SEM)
of at least three independent experiments performed in duplicate.Percent of [3H]-CCR2-RA-[R] displacement by 1 μM compound. Values represent
the mean of two independent experiments performed in duplicate.
Functional Characterization
of Selected Compounds
Following
the SAR analysis, four compounds (39, 41, 43, and 45) were selected for further
characterization in a G protein-dependent functional assay in order
to assess their inhibitory potencies (pIC50) in both CCR1
and CCR2. The four compounds were selected based on their affinity
and selectivity profile: compounds 43 and 39, with the highest affinity for either CCR1 or CCR2, respectively,
compound 41, with higher selectivity toward CCR2, and
compound 45, with higher selectivity toward CCR1. As
a functional assay, we used a previously reported [35S]GTPγS
binding assay on U2OS-CCR2 membranes, which had been applied in the
functional characterization of several allosteric and orthosteric
CCR2 ligands.[26] Similarly as reported by
Zweemer et al. (2013),[26] CCL2 stimulated
[35S]GTPγS binding in a concentration-dependent manner,
displaying a potency of 5 nM in CCR2 (pEC50 = 8.3 ±
0.09, Figure a). Using the same assay conditions, we characterized
the G protein activation of CCL3 in U2OS-CCR1 membranes. In this assay,
CCL3 induced [35S]GTPγS binding in CCR1 with a higher
potency than CCL2 in CCR2 (1.3 nM, pEC50 = 8.9 ± 0.06)
and with a higher maximum effect (Emax) (Figure a). It
should be noted that the potency of CCL3 in our study is lower than
previously reported,[41] which might be related
to the differences in cell line and/or assay conditions.(a) [35S]GTPγS binding upon stimulation of U2OS-CCR1
and U2OS-CCR2 by increasing concentrations of CCL3 and CCL2, respectively.
In both cases, the response was corrected by subtracting the basal
activity (approximately 8000 dpm for both CCR1 and CCR2). (b) Inhibition
of CCL3-induced [35S]GTPγS binding by compounds 39, 41, 43, and 45 in
U2OS-CCR1. (c) Inhibition of CCL2-induced [35S]GTPγS
binding by compounds 39, 41, 43, and 45 in U2OS-CCR2. The level of basal activity in
U2OS-CCR1 and U2OS-CCR2 is indicated by a dashed line. In all cases,
data shown are mean ± SEM of at least three experiments performed
in duplicate.For the antagonist assays,
we used a submaximal EC80 concentration of CCL3 (8 nM)
and CCL2 (20 nM) in CCR1 or CCR2, respectively,
in order to evoke 80% stimulation of [35S]GTPγS binding.
Although all compounds were able to inhibit CCL3- or CCL2-induced
G protein activation, their potencies (IC50) ranged between
30 nM to 8 μM (Table and Figure b,c). In CCR2, the potency of the compounds increased
in the same order observed for affinity (Figure c, 45 < 43 < 41 < 39). In CCR1, 39 displayed
the highest potency (590 nM), followed by 43 (950 nM),
contrary to their binding affinity (Figure b, 43 > 39).
In
addition, the moderate selectivity observed in the binding assays
was lost in this functional assay: except for 45, all
compounds were more potent inhibitors of CCR2 than CCR1, as their
potencies were 3-fold (43), 19-fold (39),
or 48-fold (41) lower in CCR1. Upon comparison of potencies
in the [35S]GTPγS assay and the affinities in the
[3H]-CCR2-RA-[R] binding assay, we observed
that all compounds displayed between 5 and 10-fold difference between
assays in CCR2 (Tables –4), in agreement with previous characterization
of CCR2-RA-[R] on this receptor.[26] In contrast, all compounds displayed at least a 50-fold
difference between assays when tested on CCR1. Such lack of correlation
between apparent potencies and binding affinities in CCR1 might be
dependent on the assay conditions used, G protein concentrations,
or the chemokine used in this study; thus, further studies are warranted
to fully characterize these ligands for their selectivity.
Table 4
Functional Characterization of Compounds 37, 39, 41, and 43 in
U2OS-CCR1 and U2OS-CCR2 Using a [35S]GTPyS Binding Assay
inhibition
of [35S]GTPyS bindinga
CCR1b
CCR2c
compd
pIC50 ± SEM (IC50, μM)
Hill slope
pIC50 ±
SEM (IC50, μM)
Hill
slope
39
6.26 ± 0.10 (0.59)***
–0.62 ± 0.05**
7.57 ± 0.08 (0.03)
–0.94 ± 0.18
41
5.73 ± 0.09 (1.94)***
–0.72 ± 0.08*
7.47 ± 0.10 (0.04)
–0.88 ± 0.13
43
6.03 ± 0.04 (0.95)
–0.73 ± 0.02*
6.54 ± 0.16 (0.33)
–0.80 ± 0.13
45
5.07 ± 0.05 (8.64)
–0.93 ± 0.01
5.06 ± 0.05 (8.77)
–1.20 ± 0.08
All values
are means ± SEM
of at least three independent experiments performed in duplicate.
Unpaired t-test analysis with Welch’s correction
was performed to analyze differences in pIC50 values between
receptors, with differences noted as ***, p <
0.001. One-Way ANOVA with Dunnett’s posthoc test was performed
to compare pseudo-Hill slopes against compound 45, which
showed a pseudo-Hill slope of approximately unity in both receptors,
with significant differences displayed as *, p <
0.05; **, p < 0.01.
Inhibition of CCL3-induced [35S]GTPyS binding
in U2OS membranes stably expressing human
CCR1. A concentration of 8 nM CCL3 was used in the assays to evoke
an 80% response.
Inhibition
of CCL2-induced [35S]GTPyS binding in U2OS membranes stably
expressing human
CCR2. A concentration of 20 nM CCL2 was used in the assays to evoke
an 80% response.
All values
are means ± SEM
of at least three independent experiments performed in duplicate.
Unpaired t-test analysis with Welch’s correction
was performed to analyze differences in pIC50 values between
receptors, with differences noted as ***, p <
0.001. One-Way ANOVA with Dunnett’s posthoc test was performed
to compare pseudo-Hill slopes against compound 45, which
showed a pseudo-Hill slope of approximately unity in both receptors,
with significant differences displayed as *, p <
0.05; **, p < 0.01.Inhibition of CCL3-induced [35S]GTPyS binding
in U2OS membranes stably expressing humanCCR1. A concentration of 8 nM CCL3 was used in the assays to evoke
an 80% response.Inhibition
of CCL2-induced [35S]GTPyS binding in U2OS membranes stably
expressing humanCCR2. A concentration of 20 nM CCL2 was used in the assays to evoke
an 80% response.In CCR1,
all compounds behaved as inverse agonists, as they all
significantly decreased the basal activity of CCR1 at the highest
concentration tested (Supporting Information, Figure S3a). In this regard, it was previously demonstrated
that CCR1 exhibits constitutive activity leading to ligand-independent
G protein-activation, β-arrestin recruitment, and receptor internalization,[42] which points to the development of inverse agonists
as a potential therapeutic option for inflammatory diseases. Yet,
only BX-471[35] has been reported to act
as inverse agonist in CCR1.[42] This prompted
us to further characterize these compounds as inverse agonists in
CCR1 by measuring their inhibitory potency in absence of the agonist
CCL3 (Supporting Information, Figure S3b and Table S2). Compounds 39 and 41 were more
potent inverse agonists than antagonists, displaying a 3-fold and
almost 10-fold higher potency, respectively, as inverse agonists.
As such, their potencies as inverse agonists were more comparable
to their binding affinities (Table and Supporting Information, Table S2). In contrast, 43 and 45 showed
similar potencies when measured in the absence or presence of CCL3
and thus displayed more than 130-fold difference between functional
and binding assays (Table and Supporting Information, Table S2). Interestingly, both compounds 43 and 45 have a cyclopropyl in the R3 position while 39 and 41 have a methyl group (Tables and 3), which suggests
that this larger group might be responsible for the difference in
their efficacy and functional profile. Moreover, most compounds displayed
pseudo-Hill slopes of less than unity in CCR1 when tested in the presence
or absence of CCL3 (Table and Supporting Information, Table S2), indicative of a more complex mechanism of inhibition, combining
negative allosteric modulation and inverse agonism.[43] Of note, the basal levels of constitutive activity in the
[35S]GTPγS assay are very dependent on the assay
conditions used, such as GDP concentrations. Yet, at a single concentration
(100 μM) tested, all compounds consistently decreased the basal
activity in CCR1 after varying GDP concentrations. For instance, compound 41 decreased basal activity by 22% (1 μM GDP), 26% (10
μM GDP), and 25% (20 μM GDP) (data not shown). To the
best of our knowledge, these compounds represent the first intracellular
ligands with demonstrated inverse agonism in CCR1. Both 45 and 43 decreased the basal activity of CCR2 to a similar
or smaller level than in CCR1 (45, maximal decrease of
58%; 43, maximal decrease of 27%), indicative of inverse
agonism (Supporting Information, Figure S3a). However, no constitutive activity has been reported for CCR2,
with only one constitutively active mutant (CAM) described so far.[44] In fact, Gilliland et al. (2013) showed that
CCR2 was not able to induce ligand-independent cell migration or to
constitutively associate with β-arrestin, pointing to a lack
of constitutive activity.[42] Moreover, several
classes of orthosteric and allosteric CCR2 ligands did not show evidence
of inverse agonism when previously tested in a similar [35S]GTPγS binding assay.[26] Thus, the
inverse agonism observed in this study might be the consequence of
the expression level, ligand concentration, and/or assay conditions
employed, so further research is warranted to investigate ligand-independent
signaling in CCR2.
Conclusions
In this study, we have
characterized [3H]-CCR2-RA-[R], a high-affinity
intracellular antagonist previously
described for CCR2,[26] in both CCR1 and
CCR2, which allowed us to conclude that this radioligand binds to
CCR1 with a similar high affinity. By characterizing this radioligand
in CCR1, we have provided evidence that CCR1 possesses an intracellular
binding site that can be used for the design of noncompetitive compounds.
In addition, this intracellular radioligand allowed us to explore
the SAR of a series of pyrrolone derivatives in both CCR1 and CCR2.
Although some of these derivatives had been previously described for
CCR2, their characterization in CCR1 had not been reported. With the
SAR analysis we learned that introduction of bulkier and more lipophilic
groups at R1 and R3 positions was better tolerated
in CCR1, allowing us to obtain better selectivity for this receptor.
The high conservation between the intracellular pockets of CCR1 and
CCR2 prevented us from finding high selectivity in these series of
compounds but allowed us to find several potential dual-target antagonists.
Finally, characterization of four selected compounds in a functional
assay allowed us to determine their functional effects as antagonists
in CCR2 and inverse agonists in the constitutively active CCR1, which
opens up a novel avenue to modulate these receptors in inflammatory
diseases. In addition, this highly conserved binding site might allow
the design of both selective and multitarget inhibitors for chemokine
receptors beyond CCR1 and CCR2.
Experimental
Section
Chemistry: General Methods
All solvents and reagents
were purchased from commercial sources and were of analytical grade.
Demineralized water is simply referred to as H2O, as was
used in all cases unless stated otherwise (i.e., brine). 1H NMR spectra were recorded on a Bruker AV 400 liquid spectrometer
(1H NMR, 400 MHz) at ambient temperature. Chemical shifts
are reported in parts per million (ppm), are designated by δ,
and are downfield to the internal standard tetramethylsilane (TMS)
in CDCl3. Coupling constants are reported in Hz and are
designated as J. As a representative example of the
obtained 1H NMR spectra, Supporting Information, Figure S4 shows the 1H NMR spectrum
of compound 43. Analytical purity of the final compounds
was determined by high pressure liquid chromatography (HPLC) with
a Phenomenex Gemini 3 × C18 110A column (50 mm × 4.6 mm,
3 μm), measuring UV absorbance at 254 nm. Sample preparation
and HPLC method was, unless stated otherwise, as follows: 0.3–0.8
mg of compound was dissolved in 1 mL of a 1:1:1 mixture of CH3CN/H2O/tBuOH and eluted from the column within
15 min, with a three component system of H2O/CH3CN/1% TFA in H2O, decreasing polarity of the solvent mixture
in time from 80/10/10 to 0/90/10. All compounds showed a single peak
at the designated retention time and are at least 95% pure. Liquid
chromatography–mass spectrometry (LC–MS) analyses were
performed using Thermo Finnigan Surveyor–LCQ Advantage Max
LC–MS system and a Gemini C18 Phenomenex column (50 mm ×
4.6 mm, 3 μm). The elution method was set up as follows: 1–4
min isocratic system of H2O/CH3CN/1% TFA in
H2O, 80:10:10, from the fourth minute, a gradient was applied
from 80:10:10 to 0:90:10 within 9 min, followed by 1 min of equilibration
at 0:90:10 and 1 min at 80:10:10. Thin-layer chromatography (TLC)
was routinely performed to monitor the progress of reactions, using
aluminum coated Merck silica gel F254 plates. Purification by column
chromatography was achieved by use of Grace Davison Davisil silica
column material (LC60A 30–200 μm). Yields and reaction
conditions were not optimized. Additionally, all compounds were screened
using FAF-Drugs4[45,46] in order to detect potential
pan-assay interference compounds (PAINS). None of the compounds was
identified as PAINS after application of three different filters based
on Baell et al.[47]
General Procedure for the
Synthesis of Compounds 6–23, 26–46.[33]
The respective aldehyde 1a–l (1.0
equiv), aniline 2a–q (1.0 equiv),
and ethyl 2,4-dioxo-butanoate analogue 3a–i (1.0 equiv) were dissolved in acetic
acid (2.5 mL/mmol) and heated at 95 °C for 2–4 h under
a nitrogen atmosphere. Upon completion of the reaction (TLC 1/7 EtOAct/petroleum
ether), acetic acid was removed under reduced pressure, the residue
was triturated with Et2O and stirred for 30 min, after
which the pure product was collected by filtration.
Sodium 1,4-diethoxy-1,4-dioxobut-2-en-2-olate
(1.25 g, 6.00 mmol) was dissolved in 25 mL of H2O, and
25 mL of Et2O was added. The mixture was acidified to pH
2 with 6 M HCl (aq) and was extracted with Et2O from the
aqueous phase, dried over MgSO4, and concentrated in vacuo,
yielding 1.05 g, 4.97 mmol, 83% diethyl 2-oxosuccinate as a yellow
oil.[52]Diethyl 2-oxosuccinate 3h (1.05 g, 4.97 mmol, 1.12 equiv) was added to a mixture
of cyclohexane carboxaldehyde 1a (534 μL, 4.42
mmol, 1.00 equiv) and 4-methylaniline 2b (474 mg, 4.42
mmol, 1.00 equiv) in 10 mL of dry THF and stirred at room temperature
overnight. The reaction mixture was concentrated in vacuo, Et2O was added, and the white precipitate was collected by filtration.
Yield: 400 mg, 26%, white solid. 1H NMR (400 MHz, CDCl3): δ 11.24 (s, 1H), 7.40 (d, J = 8.4
Hz, 2H), 7.24 (d, J = 8.4 Hz, 2H), 5.00 (d, J = 1.8 Hz, 1H), 4.32–4.13 (m, 2H), 2.32 (s, 3H),
1.85–1.76 (m, 1H), 1.66–1.59 (m, 1H), 1.56–1.44
(m, 3H), 1.32 (d, J = 12.0 Hz, 1H), 1.26 (t, J = 7.2 Hz, 3H), 1.06–0.75 (m, 4H), 0.63–0.53
(m, 1H) ppm. MS [ESI + H]+: 344.07.
[3H]-CCR2-RA-[R]
(specific activity 59.6 Ci mmol–1),
corresponding to the (R)-isomer of compound 38 ([3H]-(R)-4-acetyl-1-(4-chloro-2-fluorophenyl)-5-cyclohexyl-3-hydroxy-1,5-dihydro-2H-pyrrol-2-one)), was custom-labeled by Vitrax (Placentia,
CA). [35S]GTPγS (guanosine 5′-O-(3-[35S]thio)triphosphate), with a specific activity
of 1250 Ci mmol–1, was purchased from PerkinElmer
(Waltham, MA). CCR2-RA-[R], SD-24, and JNJ-27141491
were synthesized as described previously.[30,36,54] BX471 was purchased from Cayman Chemical
(Ann Arbor, MI, USA). Chemokine ligands CCL2 and CCL3 were purchased
from PeproTech (Rocky Hill, NJ). Bovine serum albumin (BSA, fraction
V) was purchased from Sigma (St. Louis, MO, USA). Bicinchoninic acid
(BCA) and BCA protein assay reagent were purchased from Pierce Chemical
Company (Rockford, IL, USA). Tango CCR1-bla and Tango
CCR2-bla osteosarcoma (U2OS) cells stably expressing
the humanCCR1 or humanCCR2b (U2OS-CCR1 or U2OS-CCR2, respectively)
were obtained from Invitrogen (Carlsbad, CA). All other chemicals
were obtained from standard commercial sources.
Cell Culture
and Membrane Preparation
U2OS-CCR1 and
U2OS-CCR2 were grown in a humidified atmosphere at 37 °C and
5% CO2 in McCoy’s 5A medium supplemented with 10%
fetal calf serum, 2 mM glutamine, 0.1 mM nonessential amino acids
(NEAAs), 25 mM HEPES, 1 mM sodium pyruvate, 100 IU/ml penicillin,
100 μg/mL streptomycin, 100 μg/mL G418, 50 μg/mL
hygromycin, and 125 μg/mL zeocin (200 μg/mL zeocin for
U2OS-CCR1). Cells were subcultured twice a week at a ratio of 1:3
to 1:8 on 10 cm Ø plates by trypsinization. For membrane preparation
cells were subcultured on 15 cm Ø plates using dialyzed fetal
calf serum. Membranes from U2OS-CCR1 and U2OS-CCR2 cells were prepared
as described previously.[26] Briefly, cells
were detached from confluent 15 cm Ø plates by scraping them
into 5 mL of phosphate-buffered saline (PBS), collected, and centrifuged
for 5 min at 3000 rpm (700g). The pellets were resuspended
in ice-cold 50 mM Tris-HCl buffer, pH 7.4, supplemented with 5 mM
MgCl2, and homogenized with an Ultra Turrax homogenizer
(IKA-Werke GmbH & Co. KG, Staufen, Germany). Membranes were separated
from the cytosolic fraction by several centrifugation steps in an
Optima LE-80 K ultracentrifuge (Beckman Coulter, Inc., Fullerton,
CA) at 31000 rpm for 20 min at 4 °C. Finally, the membrane pellets
were resuspended in 50 mM Tris-HCl buffer supplemented with 5 mM MgCl2, pH 7.4, divided into aliquots of 100 μL and 250 μL,
and stored at −80 °C. Membrane protein concentrations
were measured using a BCA protein determination with BSA as a standard.[55]
[3H]-CCR2-RA-[R] Binding Assays
[3H]-CCR2-RA-[R] (homologous) displacement
assays in U2OS-CCR1 and U2OS-CCR2 were performed in a 100 μL
reaction volume containing assay buffer (50 mM Tris-HCl, 5 mM MgCl2, and 0.1% CHAPS, pH 7.4), 6 nM [3H]-CCR2-RA-[R], 8–15 μg of membrane protein, and the competing
ligand. Homologous displacement assays were carried out with three
different concentrations of [3H]-CCR2-RA-[R], namely 3, 6, and 12 nM. In all cases, at least six concentrations
of competing ligand were used and the reaction mixture was incubated
for 120 min at 25 °C. Nonspecific binding was determined in the
presence of 10 μM CCR2-RA-[R]. Total radioligand
binding did not exceed 10% of the amount added to prevent ligand depletion.
[3H]-CCR2-RA-[R] saturation binding assays
in U2OS-CCR2 were also performed in a 100 μL reaction volume
containing assay buffer, [3H]-CCR2-RA-[R] in 12 different concentrations ranging from 0.05 to 70 nM, and
15 μg of membrane protein. Nonspecific binding was determined
in the presence of 10 μM JNJ-27141491 at four different concentrations
of radioligand, namely 0.1, 0.4, 2.5, and 20 nM. In association assays,
U2OS-CCR1 and U2OS-CCR2 membrane preparations were added to the reaction
mix at different time points of incubation, ranging from 1 to 180
min incubation; in dissociation assays, membranes were first incubated
with 6 nM [3H]-CCR2-RA-[R] for 90 min,
and dissociation was initiated by the addition of 10 μM CCR2-RA-[R] at different time points, up to 150 min for CCR1 and
180 min for CCR2. For all experiments, incubations were terminated
by dilution with ice-cold wash buffer (50 mM Tris-HCl buffer supplemented
with 5 mM MgCl2 and 0.05% CHAPS, pH 7.4). Separation of
bound from free radioligand was performed by rapid filtration through
a 96-well GF/B filter plate using a PerkinElmer Filtermate harvester
(PerkinElmer, Groningen, The Netherlands). Filters were washed 10
times with ice-cold wash buffer. Then 25 μL of Microscint scintillation
cocktail (PerkinElmer, Groningen, The Netherlands) was added to each
well and the filter-bound radioactivity was determined by scintillation
spectrometry using the P-E 2450 Microbeta[2] scintillation plate counter (PerkinElmer, Groningen, The Netherlands).
[35S]GTPγS Binding Assays
[35S]GTPγS binding assays were performed as described previously.[26] Briefly, binding assays were performed in a
100 μL reaction volume containing assay buffer (50 mM Tris-HCl,
5 mM MgCl2, 100 mM NaCl, 1 mM EDTA, and 0.05% BSA, pH 7.4),
10 μM GDP, 10 μg of saponin, and 10 μg of membrane,
either U2OS-CCR1 or U2OS-CCR2. To determine the EC50 value
of CCL2 and CCL3, the membrane mixture was preincubated with increasing
concentrations of chemokine for 30 min at 25 °C. To determine
the IC50 values of the ligands, the membrane mixture was
preincubated with increasing concentrations of the ligand of interest
in the absence or presence of a fixed concentration of CCL2 (20 nM)
or CCL3 (8 nM). Basal activity was determined in the absence of any
ligand or chemokine. Finally, the mixture was incubated for another
90 min at 25 °C after the addition of 0.3 nM [35S]GTPγS
in all cases. For all experiments, incubations were terminated by
dilution with ice-cold 50 mM Tris-HCl, 5 mM MgCl2 buffer.
Separation of bound from free [35S]GTPγS was performed
as described under “[3H]-CCR2-RA-[R] binding assays”.
Data Analysis
All experiments were
analyzed using GraphPad
Prism 7.0 (GraphPad Software Inc., San Diego, CA, USA). The KD and Bmax values
of [3H]-CCR2-RA-[R] in U2OS-CCR2 were
calculated from saturation experiments by fitting the data to the
equation Bound = (Bmax*[L])/([L] + KD), where Bmax is
the maximum number of binding sites and KD is the concentration required to reach half-maximum binding at equilibrium
conditions. In the case of U2OS-CCR1 membranes, the KD and Bmax values were calculated
from homologous binding experiments by nonlinear regression analysis,
using the “One Site–Homologous” model that assumes
that unlabeled and labeled CCR2-RA-[R] have identical
affinities. The (p)IC50 values of unlabeled ligands from
[3H]-CCR2-RA-[R] binding assays were obtained
by nonlinear regression analysis of the displacement curves and further
converted into (p)Ki values using the
Cheng–Prusoff equation.[56] The (p)IC50 or (p)EC50 values from [35S]GTPγS
curves were also obtained by nonlinear regression. The observed association
rate constants (kobs,fast; kobs,slow) were calculated by fitting the data to a two-phase
exponential association function; similarly, dissociation rate constants
(koff,fast; koff,slow) were calculated using a two-phase exponential decay function. All
values obtained are means ± standard error of the mean (SEM)
of at least three independent experiments performed in duplicate,
unless stated otherwise. Differences in kinetic rates and pIC50 values between receptors or between assay formats (in the
absence or presence of chemokine) were analyzed using an unpaired,
two-tailed Student’s t-test with Welch’s
correction; differences in pKi values
between compounds, in maximal [35S]GTPγS inhibition
against basal activity or in pseudo-Hill slopes from [35S]GTPγS inhibition curves against compound 45,
which showed a pseudo-Hill slope of approximately unity, were analyzed
using a one-way ANOVA with Dunnett’s posthoc test. Significant
differences are displayed as *, p < 0.05; **, p < 0.01; ***, p < 0.001; and ****, p < 0.0001.
Computational Receptor Modeling and Docking
All modeling
was performed in the Schrodinger suite,[57]Figure b,c were
made in a later version[58] that includes
the interaction and orientation of residues (e.g., backbone, side
chain). As a starting point for the structure-based studies we used
the recently published crystal structure of CCR2b in complex with
both BMS-681 and CCR2-RA-[R] (PDB 5T1A).[24] We replaced the sequence (CCR2b: sequence between L2265x62 and R2406x32) of the M2 muscarinic acetylcholine
receptor, close to the intracellular binding site, by the CCR2b sequence
using homology modeling[59−61] and CCR5 as template (PDB 4MBS).[62] A homology model of CCR1 was constructed on the basis of
this CCR2b model. For both models, the knowledge-based scoring function
was used. For the ligand docked, the pKa of the hydroxyl hydrogen was calculated to be 4.5 using Jaguar;[63,64] therefore, the negatively charged protonation state was used. Compound
CCR2-RA-[R] was docked in both models using Induced
Fit docking.[65,66] Visualizations were created using
PyMOL;[67] residues within 5 Å of the
ligand and facing the binding site are shown.
Authors: Natalia V Ortiz Zacarías; Eelke B Lenselink; Adriaan P IJzerman; Tracy M Handel; Laura H Heitman Journal: Trends Pharmacol Sci Date: 2018-04-10 Impact factor: 14.819
Authors: Annelien J M Zweemer; Julia Bunnik; Margo Veenhuizen; Fabiana Miraglia; Eelke B Lenselink; Maris Vilums; Henk de Vries; Arthur Gibert; Stefanie Thiele; Mette M Rosenkilde; Adriaan P IJzerman; Laura H Heitman Journal: Mol Pharmacol Date: 2014-07-14 Impact factor: 4.436
Authors: Anna Junker; Artur K Kokornaczyk; Annelien J M Zweemer; Bastian Frehland; Dirk Schepmann; Junichiro Yamaguchi; Kenichiro Itami; Andreas Faust; Sven Hermann; Stefan Wagner; Michael Schäfers; Michael Koch; Christina Weiss; Laura H Heitman; Klaus Kopka; Bernhard Wünsch Journal: Org Biomol Chem Date: 2015-02-28 Impact factor: 3.876
Authors: I Sabroe; M J Peck; B J Van Keulen; A Jorritsma; G Simmons; P R Clapham; T J Williams; J E Pease Journal: J Biol Chem Date: 2000-08-25 Impact factor: 5.157
Authors: P K Smith; R I Krohn; G T Hermanson; A K Mallia; F H Gartner; M D Provenzano; E K Fujimoto; N M Goeke; B J Olson; D C Klenk Journal: Anal Biochem Date: 1985-10 Impact factor: 3.365
Authors: M Amat; C F Benjamim; L M Williams; N Prats; E Terricabras; J Beleta; S L Kunkel; N Godessart Journal: Br J Pharmacol Date: 2006-10-03 Impact factor: 8.739
Authors: Vignir Isberg; Chris de Graaf; Andrea Bortolato; Vadim Cherezov; Vsevolod Katritch; Fiona H Marshall; Stefan Mordalski; Jean-Philippe Pin; Raymond C Stevens; Gerrit Vriend; David E Gloriam Journal: Trends Pharmacol Sci Date: 2014-12-22 Impact factor: 14.819
Authors: Yi Zheng; Ling Qin; Natalia V Ortiz Zacarías; Henk de Vries; Gye Won Han; Martin Gustavsson; Marta Dabros; Chunxia Zhao; Robert J Cherney; Percy Carter; Dean Stamos; Ruben Abagyan; Vadim Cherezov; Raymond C Stevens; Adriaan P IJzerman; Laura H Heitman; Andrew Tebben; Irina Kufareva; Tracy M Handel Journal: Nature Date: 2016-12-07 Impact factor: 49.962
Authors: Natalia V Ortiz Zacarías; Kirti K Chahal; Tereza Šimková; Cas van der Horst; Yi Zheng; Asuka Inoue; Emy Theunissen; Lloyd Mallee; Daan van der Es; Julien Louvel; Adriaan P IJzerman; Tracy M Handel; Irina Kufareva; Laura H Heitman Journal: J Med Chem Date: 2021-02-18 Impact factor: 7.446
Authors: Natalia V Ortiz Zacarías; Jacobus P D van Veldhoven; Lisa S den Hollander; Burak Dogan; Joseph Openy; Ya-Yun Hsiao; Eelke B Lenselink; Laura H Heitman; Adriaan P IJzerman Journal: J Med Chem Date: 2019-12-13 Impact factor: 7.446
Scheme 1
Figure 1
Figure 3
Figure 2