Discovery, structure-activity relationship study and biological evaluation of 2-thioureidothiophene-3-carboxylates as a novel class of C-X-C chemokine receptor 2 (CXCR2) antagonists

Article abstract of DOI:10.1016/j.ejmech.2020.112387

The C-X-C motif ligand 8 and C-X-C chemokine receptor 2 (CXCL8-CXCR2) axis is involved in pathogenesis of various diseases including inflammation and cancers. Various CXCR2 antagonists are under development for several diseases. Our previous high-throughput cell-based assay specific for CXCR2 has identified a pyrimidine-based compound CX797 acting on CXCR2 down-stream signaling. A lead optimization campaign through scaffold-hopping strategy led to a series of 2-thioureidothiophene-3-carboxylates (TUTP) as novel CXCR2 antagonists. Structure-activity relationship study of TUTPs led to the identification of compound 52 that significantly inhibited CXCR2-mediated β-arrestin recruitment signaling (IC₅₀ = 1.1±0.01 μM) with negligible effect on CXCL8-mediated cAMP signaling and calcium flux. Similar to the known CXCR2 antagonist SB265610, compound 52 inhibited CXCL8-CXCR2 induced phosphorylation of ERK1/2. TUTP compounds also inhibited CXCL8-mediated cell migration and showed synergy with doxorubicin in ovarian cancer cells, thereby supporting TUTPs as promising compounds for cancer treatment.

Full text of DOI:10.1016/j.ejmech.2020.112387

Journal Pre-proof
Discovery, structure-activity relationship study and biological evaluation of 2-
thioureidothiophene-3-carboxylates as a novel class of C-X-C chemokine receptor 2
(CXCR2) antagonists
Ding Xue, Wenmin Chen, Nouri Neamati
PII:
S0223-5234(20)30357-3
DOI:
https://doi.org/10.1016/j.ejmech.2020.112387
EJMECH 112387
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 10 February 2020
Revised Date: 9 April 2020
Accepted Date: 23 April 2020
Please cite this article as: D. Xue, W. Chen, N. Neamati, Discovery, structure-activity relationship study
and biological evaluation of 2-thioureidothiophene-3-carboxylates as a novel class of C-X-C chemokine
receptor 2 (CXCR2) antagonists, European Journal of Medicinal Chemistry, https://doi.org/10.1016/
j.ejmech.2020.112387.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©
2020 Published by Elsevier Masson SAS.

Discovery, Structure-Activity Relationship Study and
Biological Evaluation of 2-Thioureidothiophene-3-
carboxylates as a Novel Class of C-X-C Chemokine
Receptor 2 (CXCR2) Antagonists
†
,
†,
*,†
Ding Xue, Wenmin Chen, ∇ and Nouri Neamati
∇
†
Department of Medicinal Chemistry, College of Pharmacy, Rogel Cancer Center, University of
Michigan, North Campus Research Complex, 1600 Huron Parkway, Ann Arbor, Michigan,
4
8109, United States.
ABSTRACT
The C-X-C motif ligand 8 and C-X-C chemokine receptor 2 (CXCL8-CXCR2) axis is involved
in pathogenesis of various diseases including inflammation and cancers. Various CXCR2
antagonists are under development for several diseases. Our previous high-throughput cell-based
assay specific for CXCR2 has identified a pyrimidine-based compound CX797 acting on
CXCR2 down-stream signaling. A lead optimization campaign through scaffold-hopping
strategy led to a series of 2-thioureidothiophene-3-carboxylates (TUTP) as novel CXCR2
antagonists. Structure-activity relationship study of TUTPs led to the identification of compound
5
2 that significantly inhibited CXCR2-mediated β-arrestin recruitment signaling (IC50
=
1
.1±0.01 µM) with negligible effect on CXCL8-mediated cAMP signaling and calcium flux.
Similar to the known CXCR2 antagonist SB265610, compound 52 inhibited CXCL8-CXCR2
induced phosphorylation of ERK1/2. TUTP compounds also inhibited CXCL8-mediated cell
1

migration and showed synergy with doxorubicin in ovarian cancer cells, thereby supporting
TUTPs as promising compounds for cancer treatment.
KEYWORDS
CXCR2; Thioureidothiophene; β-arrestin; GPCR; CXCL8; cAMP;
INTRODUCTION
CXCR2 is a member of the seven transmembrane G-protein-coupled receptor (GPCR) family
expressed on immune cells, endothelial cells, epithelial cells, as well as various types of cancer
1
cells. CXCR2 mediates its downstream biological effects upon binding with high affinity to
+
CXCL8 as well as other ELR (Glu4-Leu5-Arg6)-containing chemokines including CXCL1, 2, 3,
2
5
, 6 and 7.
The CXCL8-CXCR2 axis is involved in the pathogenesis of various diseases including
inflammation and cancers. As the first and most intensively studied chemokine, CXCL8 plays an
essential role in the recruitment and migration of neutrophils during inflammatory process, thus
the modulation of this pathway is beneficial in the treatment of various diseases including
3
chronic obstructive pulmonary disorder (COPD), asthma, cystic fibrosis and arthritis. The
secretion of CXCL8 and its activation of CXCR2 are also increasingly recognized as key factors
for tumor growth, invasion, angiogenesis and metastasis. CXCR2 signaling is important for the
microenvironment of select tumor types and has a regulatory effect on cancer stem cells
2
, 4
proliferation and self-renewal. While overexpression of CXCL8 is associated with several late
stage diseases, CXCR2 expression level is correlated with prognosis and therapeutic outcome of
5
6
7
patients with colorectal cancer, bladder cancer, hepatocellular carcinoma and laryngeal
2

8
squamous cell carcinoma. Upregulation of CXCR2 in resistant cancer cell lines also indicates its
9
-11
involvement in the development of resistance to chemotherapies.
The multiple effects of
CXCL8 signaling within the tumor pathology suggest that the targeting of CXCL8-CXCR2 may
have important implications to halt cancer progression and to sensitize tumor cells to
chemotherapies.
Given the importance of CXCR2 signaling pathway in inflammatory disorders and cancers, the
1
2-16
antagonists of CXCR2 have received increasing attention over the past twenty years.
Since
1
7
the discovery of the first small-molecule CXCR2 antagonists in 1998, a number of CXCR2
antagonists have been developed and at least nine of them have advanced to clinical stage for
COPD, asthma, arthritis and cancer treatments (Figure. 1). Diarylureas SB656933 and danirixin
1
8-20
were tested in Phase I and II clinical trial respectively for the treatment of COPD.
The
bioisosteric replacement of the urea led to the development of another series of CXCR2
antagonists with a squaramide core. Navarixin from this class exhibited significant anti-tumor
2
1, 22
activity in melanoma and colon cancer mouse models,
and it is now under Phase II clinical
2
3
trial. Several other antagonists with distinct structure motifs are also under clinical
development for the treatment of cancer and inflammatory diseases. Repertaxin is tested as
single agent or in combination for the treatment of breast cancer or symptoms associated with
24
organ transplantations.
AZD8309, AZD5069 and AZD4721 are a series of 2-
(benzylthio)pyrimidines that have been testing for nasal inflammation, COPD, asthma, and
2
5
prostate cancer. SX-682 is another close analog with a distinct boronic acid group that enhance
T cell activation and antitumor immunity and is now tested in combination with pembrolizumab
2
6, 27
in Phase I/II trial for metastatic melanoma.
3

Since the crystal structure of CXCR2 has not yet been solved, ligand-based drug design and
high-throughput screening are the most feasible approaches for discovery of CXCR2 antagonists.
We previously discovered a novel pyrimidine-based compound CX797 as a potent CXCR2
signaling modulator through a high-throughput screening of our in-house compound library.
CX797 inhibits CXCL8 mediated cAMP signaling (IC = 7.79 ± 0.15 µM), receptor degradation
5
0
2
8
as well as cell migration, while stimulates CXCL8-mediated β-arrestin recruitment. CX797 has
a planar polycyclic structure and relatively high lipophilicity, which could be detrimental to its
drug-like properties like metabolic stability, bioavailability, and hERG binding induced
2
9, 30
toxicity.
Using the scaffold-hopping strategy, we opened the fused polycyclic ring system of
CX797 to produce a series of compounds with thiourea and carboxylate moieties (Figure 2).
These compounds, namely 2-thioureidothiophene-3-carboxylates, proved to be a novel class of
CXCR2 antagonists and exhibited a distinct profile in the CXCR2 inhibition assay. An extensive
SAR study was carried out and the detailed mechanism of action regarding the CXCL8/CXCR2
signaling is presented. Its inhibition of CXCL8-mediated cell migration and the synergistic effect
with doxorubicin further warrants its development for treating cancers.
4

Figure 1. Structures of representative CXCR2 antagonists
Figure 2. Modification of CX797 using a scaffold-hopping approach
RESULTS AND DISCUSSION
Chemistry
As depicted in Scheme 1, the synthesis of the 2-thioureidothiophene-3-carboxylates and
carboxamides derivatives started with a one-pot Gewald reaction employing cyclic ketones, 2-
cyanoacetates/2-cyanoacetamide, sulfur and morpholine as starting materials, leading to the
5

generation of 2-aminothiophene-3-carboxylates/carboxamide 2a-l. Compared to the standard
3
1-33
heated conditions in MeOH, EtOH, or DMF at 50-80℃ ,
neat condition under room
temperature provided improved yields of 2 (average yields 65% versus 31%). Compound 2 was
then treated with the appropriate isothiocyanates to afford 11, 26-28, 30-32, 36-40, 42 and 43.
Compound 29 was obtained through hydrolysis of the tert-butyl ester of 31 by TFA. To facilitate
the modification on R of the final compounds, 2 was converted into isothiocyanate derivative 3
4
that was reacted with various anilines/benzyl amines to yield 12-25, 44-50 and 52-56. To
synthesize compound 41, intermediate 2l was selectively protected with Fmoc on the piperidine
moiety and then reacted with phenyl isothiocyanate, followed by the deprotection of piperidine
in DMF. Amides 33 and 35 were synthesized by a different scheme starting from the Boc
protection of the 2-amino group of 2a to give 5, of which the ethyl ester was hydrolyzed in the
presence of NaOH under 80℃ to give 6. Compound 6 was then coupled with methyl amine or
benzylamine in the presence of HATU and DIEA. The resulting compound 7a and 7b were
deprotected and subsequently reacted with phenyl isothiocyanate to give the final compounds 33
and 35, respectively (Scheme 1).
a
Scheme 1. Synthesis of 2-Thioureidothiophene-3-carboxylate Derivatives
6

R3
X
O
R3
X
O
e
R1
R2
R1
R2
NCS
NH R4
NH
S
S
S
3
a-c
1
2-25, 44-50, 52-56
d
R3
O
R3
X
O
O
X
OH
NH
O
For 31
c
R1
R2
a
R1
R2
b
R1
NH R4
NH2
R2
S
NH
S
S
NH
S
S
2
a-l
1
1, 26-28, 30-32, 36-40, 42-43
29
For 2l
f
i
For 2a
R5
NH
O
R5
NH
O
O
O
O
O
OH
j
k
l, m
O
NH
S
NH
Boc
NH
Boc
NH
Boc
S
NH
NH2
S
S
S
N
S
Fmoc
7
7
a R5 = Me
b R5 = Bn
33 R5 = Me
35 R5 = Bn
4
5
6
g, h
t
2
2
2
a -R1....R2- = -CH(CH3)CH2CH2CH2- R3 = Et X = O
2i -R1....R - = -CH OC(CH ) CH -
R3 = Bu X = O
2
2
3
2
2
i
t
b -R1....R2- = -CH(CH3)CH2CH2CH2- R3 = Pr X = O
2j -R1....R - = -CH CH N(CH )CH -
R3 = Bu X = O
2
2
2
3
2
t
t
c -R1....R2- = -CH(CH3)CH2CH2CH2- R3 = Bu X = O
2k -R1....R - = -CH CH N(CH Ph)CH - R3 = Bu X = O
O
2
2
2
2
2
t
O
2d -R1....R2- = -CH(CH3)CH2CH2CH2- R3 = H X = NH 2l -R1....R2- = -CH2CH2NHCH2-
R3 = Bu X = O
t
2
e -R1....R2- = -CH2CH2CH2CH2-
R3 = Bu X = O
3a -R1....R2- = -CH(CH3)CH2CH2CH2- R3 = Et X = O
t
t
NH
S
2f -R1....R - = -CH(CH )CH OCH -
R3 = Bu X = O
3b -R1....R - = -CH(CH )CH CH CH - R3 = Bu X = O
2
3
2
2
2
3
2
2
2
HN
S
NH
t
t
2
2
g -R1....R2- = -CH2CH2OCH2-
h -R1....R2- = -CH2CH2CH2O-
R3 = Bu X = O
3c -R1....R2- = -CH2CH2OCH2-
R3 = Bu X = O
t
R3 = Bu X = O
4
1
1
1-28 -R1....R2- = -CH(CH3)CH2CH2CH2- R3 = Et X = O
3
0-32 -R1....R2- = -CH(CH3)CH2CH2CH2- R4 = Ph
1
1
1
1
1
1
1
1
1
1
2
3
4
5
6
7
8
9
R4 = Ph
R4 = 4-F-Ph
20 R4 = 4-NH2-Ph
21 R4 = 4-OH-Ph
22 R4 = 2-OH-Ph
23 R4 = 2-F-Ph
24 R4 = 3-OMe-Ph
25 R4 = Bn
R4 = 4-OMe-Ph
R4 = 4-OCF3-Ph
R4 = 4-NMe2-Ph
i
3
3
0 R3 = Pr X = O
t
1 R3 = Bu X = O
t
32 R3 = H
X = NH
R4 = 4- Bu-Ph
R4 = 2,2-difluorobenzo[d][1,3]dioxole-5-yl 26 R4 = pyrazine
R4 = 3,4-di-OMe-Ph
R4 = 4-CF3-Ph
27 R4 = benzoyl
28 R4 = -C(=O)OEt
t
t
t
3
6-40, 42, 43 R3 = Bu R4 = Ph X = O 44-50 -R1....R2- = -CH(CH3)CH2CH2CH2- R3 = Bu X = O
52-56 -R1....R2- = -CH2CH2OCH2- R3 = Bu X = O
3
6 -R1....R2- = -CH2CH2CH2CH2-
7 -R1....R2- = -CH(CH3)CH2OCH2-
8 -R1....R2- = -CH2CH2OCH2-
44 R4 = 3-OMe-Ph
45 R4 = 4-OCF3-Ph
46 R4 = 4-O Bu-Ph
47 R4 = 4-OPh-Ph
48 R4 = 4-(4-methylpiperazin-1-yl)-Ph
49 R4 = 3,4-di-OMe-Ph
52 R4 = 4-F-Ph
3
3
3
4
4
4
53 R4 = 4-OCF -Ph
3
t
54 R4 = 3,4-di-OMe-Ph
55 R4 = (S)-1-phenylethan-1-yl
56 R4 = (R)-1-phenylethan-1-yl
9 -R1....R2- = -CH2CH2CH2O-
0 -R1....R2- = -CH2OC(CH3)2CH2-
2 -R1....R2- = -CH2CH2N(CH3)CH2-
3 -R1....R2- = -CH2CH2N(CH2Ph)CH2-
50 R4 = 2,2-difluorobenzo[d][1,3]dioxole-5-yl
a
Reagents and conditions: (a) 2-cyanoacetates or 2-cyanoacetamide, sulfur, morpholine, rt; (b)
isothiocyanates, pyridine, 60 ℃ ; (c) TFA, DCM, rt; (d) thiophosgene, CaCO , DCM, H O, rt; e)
3
2
amines, pyridine, 60 ℃ ; (f) Fmoc-Cl, Na
2
CO
3
, dioxane, H
2
O, rt; (g) phenyl isothiocyanate,
pyridine, 60 ℃ ; (h) piperidine, DMF, rt; (i) (Boc)
2
O, DMAP, dioxane, rt; (j) NaOH, THF,
7

2
MeOH, H O, 80 ℃ ; (k) amines, HATU, DIEA, DMF, rt; (l) HCl, dioxane, rt; (m) phenyl
isothiocyanate, pyridine, 60 ℃ .
A different scheme was applied to the synthesis of amides 34 and 51 (Scheme 2). 2-cyano-N-
phenylacetamide 8 was obtained by the reaction of 2-cyanoacetic acid with aniline in the
presence of EDCI and HOBt. The Gewald reaction as described above was not applicable to
3
4
amide 8, and thus a two-step procedure involving the olefin intermediate 9 was employed.
Compound 8 was treated with 2-methylcyclohexanone or tetrahydro-4H-pyran-4-one in the
presence of NH OAc in AcOH and toluene at 80 ℃ , providing 9a (mixture of Z- and E- isomers)
4
and 9b, respectively. 9a/9b was subjected to the cyclization condition to give 10a/10b in the
presence of sulfur and morpholine. For 9a, only the Z-isomer was reacted. 10a/10b was then
heated with phenyl isothiocyanate in pyridine to give 34 and 51.
a
Scheme 2. Synthesis of compound 34 and 51
a
3
Reagents and conditions: (a) aniline, EDCI, HOBt, Et N, DMF, rt; (b) 2-methylcyclohexanone
(
for 9a) or tetrahydro-4H-pyran-4-one (for 9b), NH OAc, AcOH, toluene, 80 ℃ ; (c) sulfur,
4
o
morpholine, EtOH, 60 C; (d) phenyl isothiocyanate (for 34) or 4-fluorophenyl isothiocyanate
(for 51), pyridine, 50 ℃ .
8

SAR study
Upon stimulation with CXCL8, the plasma membrane CXCR2 is phosphorylated and β-arrestins
are recruited to their C terminus to cause the internalization of CXCR2, which mediates the
3
5-37
activation of downstream signaling pathways including ERK1/2, JNK1, and p38 MAPK.
Homologous CXCR4 trans-membrane protein triggers similar cascades as stimulated by stromal
3
8
cell-derived factor 1 (SDF-1). Tango assays based on beta-lactamase reporter provide a highly
3
9
selective readout of target receptor activation and subsequent β-arrestin recruitment. We tested
all our compounds in both CXCR2 and CXCR4 Tango assays. A good selectivity for CXCR2
inhibition over CXCR4 was desirable as it indicates a more specific mechanism of action
towards CXCR2, and could efficiently exclude experimental artifacts arising from the
interferences with the assay.
Using a scaffold-hopping strategy, we first obtained compound 11. Compound 11 inhibited
CXCL8-mediated β-arrestin recruitment with an IC50 of 6.4 ± 2.5 ꢀM in CXCR2 Tango assay
(Table 1). Encouraged by this result, we set out to investigate the SAR through the modification
of the phenylthiourea moiety (Table 1). Most of the compounds were active in CXCR2 Tango
assay and showed selectivity over CXCR4. Hydrophobic substituents at either para- meta- or
ortho- positions were well tolerable (12-19, 23 and 24), and compound 17 with a 2,2-
difluorobenzo[d][1,3]dioxole moiety showed the best potency with an IC50 of 1.9 ± 0.5ꢀM.
Hydrophilic groups like hydroxyl and amino groups (20-22) showed detrimental effect on their
potencies, indicating phenylthiourea moiety is probably positioned in a hydrophobic cavity
within its binding pocket. Replacement of the phenyl group with a benzyl (25) or an
ethoxycarbonyl group (28) led to a slight decrease in potency while a pyrazine (26) or a benzoyl
(27) group was not tolerated at this position.
9

Table 1. CXCR2/4 inhibition of TUTPs with modifications on the thiourea moiety
a
a
CXCR2
CXCR4
Selectivity of
b
Compound
R
cLogP
IC50 (ꢀM)
IC50 (ꢀM)
CXCR2/CXCR4
11
12
13
14
6.4±2.5
> 10
>1.6
5.5
5.05
5.25
5.00
6.10
2.1±0.7
4.4±2.0
2.2±1.3
11.7±1.3
11.5±6.5
12.7±5.5
2.6
5.7
15
16
3.9±0.7
7.0±4.6
> 10
> 10
>2.6
>1.4
5.22
6.88
1
7
8
1.9±0.5
7.7±0.6
26.7
>10
14
6.52
4.62
1
>1.3
1
0

19
20
21
22
23
2.0±0.3
> 30
14.6±5.1
> 10
7.2
NA
6.07
3.83
4.39
4.39
4.80
> 30
11.7±2.7
> 10
<0.4
NA
> 30
7.5±2.3
12.0±2.8
>1.6
2
4
5
3.0±1.0
6.7 ±2.5
8.9±1.9
> 20
2.9
>3
5.00
5.53
2
26
> 10
> 10
NA
3.36
2
7
8
> 10
>10
>10
NA
5.44
5.18
2
8.7±1.3
>1.1
a
CXCR2/4 inhibition was determined by Tango assay and IC50 values are the mean±SD of at
b
least three independent experiments. cLog P values were calculated using ChemBioDraw
Professional 16.
1
1

We then investigated the effect of bioisosteric replacements of the ethyl ester moiety at position
of the tetrahydrobenzo[b]thiophene core (Table 2). The corresponding carboxylic acid showed
3
a similar profile with the ethyl ester (29), while bulkier esters like isopropyl (30) or tert-butyl
ester groups (31) were beneficial to the CXCR2 inhibition activity. A similar trend was observed
within the amide analogs. Primary amide (32) or methyl amide (33) were not favored, which led
to a complete loss of potency, whereas the phenyl (34) or benzyl (35) amides were active. Taken
together, it is proposed that the carboxyl moiety is inserted into a hydrophobic cavity that
requires a substantial volume.
Table 2. CXCR2/4 inhibition of TUTPs with modifications on the carboxyl moiety
CXCR2
CXCR4
Selectivity of
b
Compound
R
cLogP
a
a
IC50 (ꢀM)
IC50 (ꢀM)
CXCR2/CXCR4
2
3
9
0
6.7±2.2
2.7±1.2
> 10
>1.5
3.4
4.79
5.36
9.3±1.1
3
3
3
1
2
3
2.5±1.4
>30
11.9±0.4
>10
4.9
NA
NA
5.76
3.61
3.84
> 10
>10
3
4
2.7±1.4
>10
>3.7
5.62
1
2

3
5
2.5±0.5
>30
>11.8
5.80
a
CXCR2/4 inhibition was determined by Tango assay and IC values are the mean ± SD of at
5
0
b
least three independent experiments. cLog P values were calculated using ChemBioDraw
Professional 16.
The preliminary SAR study suggested that hydrophobic groups were generally favored on either
the phenyl thiourea or the carboxyl moieties, such as compounds 30, 31, 34, and 35. However,
these modification leads to cLogP value > 5, that can likely compromise aqueous solubility and
3
0
simultaneously increases the risk of non-specific toxicity. In view of this hypothesis, we
proposed that the introduction of a hetero-atom into the hexane ring could possibly serve as a
“lever” to balance the increase in cLogP by other hydrophobic moieties and thus be beneficial to
overall physicochemical properties. As expected, dihydrothienopyran analogs 37 and 38 turned
out to have desirable cLogP values and retained similar CXCR2 inhibition as compared to
compound 31. Moving the oxygen to other positions seemed not favorable as exemplified by
compounds 39 and 40. Replacement of the hexane with piperidines (41-43) diminished the
activity of this series of compounds.
Table 3. CXCR2/4 inhibition of TUTPs with modifications on the tetrahydrobenzothiophene
core
1
3

CXCR2
CXCR4
Selectivity of
b
Compound
cLogP
IC (ꢀM)
IC (ꢀM)
CXCR2/CXCR4
5
0
50
3
3
6
7
1.8 ±0.3
1.7±1.3
>10
>15
>5.6
>8.8
5.24
3.66
3
3
8
9
2.3±2.0
4.1±1.7
>30
>10
>13.2
>2.4
3.14
4.25
4
0
>10
>10
NA
4.18
4
4
1
2
8.7±2.6
>10
>10
>10
>1.1
NA
3.12
3.57
4
3
> 10
>10
NA
5.56
a
CXCR2/4 inhibition was determined by Tango assay and IC50 values are the mean±SD of at
b
least three independent experiments. cLog P values were calculated using ChemBioDraw
Professional 16.
After exploring the favorable modifications on the core and substitutions individually, we then
sought to investigate if the combination could be more beneficial to their potency and selectivity.
Surprisingly, the tert-butyl ester group seemed not to be compatible with bulky substituents on
the phenyl group when a tetrahydrobenzothiophene core was retained (45-50). Only compound
4
4 with a 3-methoxygroup showed a moderate potency. This incompatibility was also observed
1
4

when a dihydrothienopyran core structure was incorporated along with a phenyl amide moiety
51), which led to a complete loss of its activity. Gratifyingly, the dihydrothienopyran core was
(
compatible with the tert-butyl ester group (52-56) where all compounds exhibited moderate to
good potencies while retaining a good selectivity over CXCR4. Compound 52 was identified as
the most potent CXCR2 antagonist (IC50 =1.1µM) and it exhibited a 27-folds selectivity over
CXCR4.
Table 4. CXCR2/4 inhibition of TUTPs with modifications on multiple sites
CXCR2
CXCR4
Selectivity of
b
Compound
R3
R4
cLogP
5.70
6.81
6.94
7.86
IC50 (ꢀM) IC50 (ꢀM) CXCR2/CXCR4
t
4
4
4
4
4
5
6
7
O Bu 2.7±2.1
>10
>10
>10
>10
>3.7
>0.9
>1.5
NA
t
O Bu 10.6±3.6
t
O Bu 6.8±2.5
t
O Bu
>10
>10
t
4
8
O Bu
>10
NA
5.66
1
5

t
4
5
9
0
O Bu 5.7±2.0
>10
>10
>1.8
NA
5.33
7.22
t
O Bu
>10
>10
5
5
5
5
5
5
1
2
3
4
5
6
NHPh
>10
>30
>30
>10
>30
>30
NA
3.21
3.34
4.19
2.71
3.93
t
O Bu 1.1±0.01
>26.7
>18.4
>2.9
t
O Bu 1.6±0.6
t
O Bu 3.4±0.2
t
O Bu 2.4±1.2
>12.7
>18.7
t
O Bu 1.6±0.3
3.93
-
SB265610
-
-
-
0.15±0.02
a
CXCR2/4 inhibition was determined by Tango assay and IC50 values are the mean±SD of at
b
least three independent experiments. cLog P values were calculated using ChemBioDraw
Professional 16.
Different from the lead compound CX797 but similar to reported CXCR2 antagonist
4
0
SB265610, compound 52 selectively inhibited CXCL8-meditated β-arrestin recruitment (IC =
5
0
1
.1± 0.01 µM, Figure 3). Compound 52 inhibited CXCL8-meditated recruitment of β-arrestin by
87% at 10 µM, while CX797 cause an increase of ~117% of this signaling at the same
~
1
6

concentration. Both compounds had negligible effects on β-arrestin recruitment upon ligand
stimulation with SDF-1. (Figure 3B). This result indicates that TUTPs represent a novel class of
CXCR2 modulators with different mechanism of action on CXCR2 as compared with CX797.
Figure 3. Compound 52 inhibited CXCL8-mediated β-arrestin recruitment, similar to the known
CXCR2 antagonist SB265610 while different from CX797. (A) Dose-response curved of
TM
compound 52 and SB265610 in the CXCR2 Tango assay. (B) CX797 increased CXCL8-
mediated β-arrestin recruitment while compound 52 inhibited CXCL8-mediated β-arrestin
recruitment at the concentration of 10 µM. Both compounds had negligible effects on β-arrestin
recruitment upon ligand stimulation with SDF-1. CXCR2-bla U2OS cells were pre-treated with
indicated concentrations of compounds for 30 min. Cells were then stimulated with CXCL8 for 5
h. Cells were loaded with a β-lactamase substrate for 2 h and the amount of cleaved and
uncleaved substrate was measured using the Clario Star plate reader (excitation at 405 nm and
emissions at 460 and 535 nm).
Compounds 37, 52 and 56 are selective over other GPCRs.
1
7

Compounds 37, 52 and 56 were tested in a GPCR screening assay containing 50 different
GPCRs (Supplementary Table S1 & S2). Compound 37 showed good selectivity over all the
tested GPCRs (<50% inhibition at 10 ꢀM or Ki > 10 ꢀM in secondary binding assay) except
binding to peripheral benzodiazepine receptor (PBR) at a Ki of 3.6 ꢀM (Table 5). Compounds 52
and 56 exhibited excellent selectivity over all the GPCRs tested, warranting their further
development.
Table 5. GPCR inhibition screening results of compounds 37, 52 and 56
CXCR2
CXCR4
GPCRs with over 50% inhibition
(Ki in secondary conformational binding assay)
PBR 63.1% (4.0 ± 0.9 ꢀM)
Compound
a
IC (ꢀM) IC (ꢀM)
50
50
3
7
1.7±1.3
>15
5-HT2B 58.0% (>10 ꢀM)
Alpha2A 55.2% (>10 ꢀM)
5
5
2
6
1.1±0.01
1.6±0.3
>30
>30
None
PBR 64.01% (>10 ꢀM)
Alpha2A 59.01% (>10 ꢀM)
a
Compounds were screened in primary radioligand binding assays at a single concentration (10 µM).
Compounds showing a minimum of 50% inhibition at 10 µM are tagged for secondary
radioligand binding assays to determine equilibrium binding affinity at specific targets.
TUTP Compounds Do Not Interfere with CXCL8-mediated cAMP signaling and calcium
mobilization
1
8

Besides β-arrestin recruitment, G-protein signaling is another main cascade involved in CXCL8-
CXCR2 axis. G-protein activation inhibits adenylyl cyclase resulting in decreased cAMP
2
+
production and induces rapid intracellular Ca mobilization released from the endoplasmic
reticulum (ER). We measured the level of those second messengers (cAMP and calcium) upon
5
2 and 56 treatment.
2
93T-CXCR2-GFP-p22F cells was engineered to express the cAMP sensitive firefly luciferase
4
1
which served as readout of the cAMP level. Forskolin activates the enzyme adenylyl cyclase
and increases intracellular levels of cAMP as shown in Figure 4A, while the addition of CXCL8
inhibit the stimulation of cAMP signaling. The treatment of SB265610 and CX797 blocked the
inhibition of cAMP production caused by CXCL8 treatment, while TUTP compound 52 and 56
did not.
As shown in Figure 4B, CXCL8 rapidly induces calcium mobilization in 293T-CXCR2-GFP
cells, reaching a peak within 30 s. Pre-treatment with 15 ꢀM SB265610 significantly reduced
CXCL8-mediated calcium mobilization. However, co-treatment with TUTP compound 52 or
CX797 had negligible effect on calcium flux. This observation demonstrates that although TUTP
compounds inhibit CXCL8-mediated β-arrestin recruitment, they don’t show significant effect
on cAMP signaling and calcium mobilization, suggesting that TUTP compounds have a biased
impact on CXCR2 downstream signaling.
1
9

Figure 4. TUTP compounds did not inhibit CXCL8-mediated cAMP signaling and calcium
mobilization. (A) CX797 and SB265610 inhibited CXCL8-mediated, forskolin-stimulated cAMP
signaling while TUTP compounds 52 and 56 did not. Cells were pretreated with compounds at
indicated concentrations for 10 min prior to stimulation with 50 nM of CXCL8. Cells were then
stimulated with 50 ꢀM of forskolin until max signal was reached (~80 min). (B) 52 and CX797
did not inhibit CXCL8 induced calcium mobilization in CXCR2-overexpressing 293T-CXCR2-
GFP cells. The cells were pretreated with compounds before the addition of CXCL8 (300 nM)
through injector. SB265610 inhibited CXCL8-stimulated calcium internalization while 52 and
CX797 did not.
TUTP Compounds Inhibit CXCL8-CXCR2 mediated ERK1/2 Phosphorylation
After the stimulation of CXCR2 by CXCL8, both β-arrestin recruitment and G-protein-mediated
signaling result in phosphorylation of ERK1/2, which promotes chemotaxis. Other growth and
stress proteins such as P38, JNK, cJUN, and AKT are also stimulated to facilitate cell survival
2
, 35-37
and proliferation.
The impact of TUTP compounds on ERK1/2 phosphorylation and other
downstream signaling was investigated. We observed that CXCL8 treatment induced
2
0

phosphorylation of ERK1/2 in T293-CXCR2-GFP cells in a dose-dependent manner (Figure 5A).
Using 60 nM of CXCL8 to stimulate phosphorylation of ERK1/2, pretreatment with SB265610
(10 µM) or TUTP compound 52 and 56 blocked CXCL8-induced ERK1/2 phosphorylation
(
Figure 5B). In contrast, CX797 stimulated ERK1/2 phosphorylation. The increased
phosphorylation of P38, JNK, cJUN, and AKT was also observed post-CXCL8 treatment that
could not be reversed by SB265610 or TUTP compounds treatment (Figure 5C).
Figure 5. The influences of TUTPs on the CXCL8-CXCR2 down-stream cell signaling pathway.
(A) Dose-response of CXCL8 stimulated phosphorylation of ERK1/2. 60 nM of CXCL8 is used
in the subsequent experiments. (B) SB265610 (SB), TUTP compounds 52 and 56 inhibited the
phosphorylation of ERK1/2 while CX797 enhanced the phosphorylation of ERK1/2. (C) 60 nM
CXCL8 treatment lead to the enhanced phosphorylation of P38, JNK, cJUN, and AKT, which
were not inhibited by SB265610 or TUTP compound 52.
TUTP compound inhibited the cell migration in the wound healing assay
2
1

CXCL8-CXCR2 mediates the chemotaxis of neutrophils and endothelial cells during infection
4
2
43
and angiogenesis. CXCL8 also induce cancer cell migration and invasion. Thus, we used a
wound healing assay to assess the effects of selected TUTP compounds on CXCL8-induced cell
migration (Figure 6). Treatment of U2OS cells with CXCL8 induced cell migration and
promoted wound healing as compared to the unstimulated control group. Pretreatment with 52,
5
6 or SB265610 prior to CXCL8 stimulation inhibited CXCL8-mediated wound closure.
Figure 6. TUTP compounds and SB265610 inhibited CXCL8-mediated cell migration. (A)
CXCR2-bla U2OS were seeded in 96-well plates at 35, 000 cells per well in DMEM
supplemented with 1% FBS overnight. A single wound was induced in the monolayer with a
sterile pipette tip and washed with PBS. Cells were treated with TUTP compounds (10 µM) or
SB265610 (10 µM) for 20 min and then stimulated with CXCL8 (60 nM) for 24 h. Control wells
were not stimulated with CXCL8. +CXCL8 wells were treated with CXCL8 (60 nM).
Representative images of two independent experiments performed in duplicate are shown here.
Combination of TUTP Compounds and Doxorubicin is Synergistic in
Ovarian Cancer Cells
2
2

Inflammatory chemokines play an important role in the angiogenesis and progression of ovarian
cancer. The interactions between chemokines produced by the ovarian cancer cells and
chemokine receptors expressed by endothelial cells and the tumor microenvironment promote
tumor growth by stimulating angiogenesis and increasing migration, invasion, and cell
4
4, 45
proliferation.
CXCR2 inhibition combined with sorafenib improved antitumor and
4
6
antiangiogenic response in preclinical models of ovarian cancer. We investigated the
combination of selected TUTP compounds with several anticancer drugs in ovarian cancer cell
lines. Results showed that compound 52, 56 or 37 was synergistic with doxorubicin in SKOV3
cells (Figure 7). Doxorubicin did not show significant cytotoxicity in SKOV3 cells at the
concentration up to 38 nM. However, combination with TUTP compounds (8.9 µM, 13.3 µM
and 20 µM) resulted in significant inhibition of colony formation. Compound 56 exerted the best
synergistic effect in the colony formation assay. This result indicates that TUTP compounds hold
the potential to aid standard-of-care cancer chemotherapies.
2
3

Figure 7. The combination of TUTP compounds (56, 52 or 37) with doxorubicin exerted
synergistic anticancer activity against SKOV3 cells in colony formation assay.
CONCLUSION
Our previous study has identified a pyrimidine-based compound CX797 as a modulator of
CXCR2/CXCL8 signaling, which inhibited CXCL8-mediated cAMP signaling while specifically
up-regulated CXCL8-mediated
β-arrestin recruitment (Figure 8). In this study, structure
modification of CX797 by a scaffold-hopping approach resulted in the discovery of a series of
TUTP compound as novel CXCR2 antagonists. An extensive SAR study was carried out and the
most potent compound 52 obtained inhibited β-arrestin recruitment with an IC of 1.1 µM in
50
Tango assay. In addition, select TUTP compounds showed good selectivity for CXCR2
inhibition over CXCR4 and other GPCRs, further implicating they are specific CXCR2
antagonists.
Interestingly, TUTP compound 52 inhibited CXCL8/CXCR2 β-arrestin coupling cascade while
does not interfere with the G-protein signaling mediated by second messengers cAMP or calcium.
Similar to CXCR2 antagonist SB265610, TUTP 52 also inhibited phosphorylation of ERK1/2
induced by CXCR2 activation (Figure 8). Because of the complexity and flexibility of the
transmembrane protein, multiple antagonist allosteric binding sites are reported to exist in the
4
7
transmembrane regions and C-terminal domains. Meanwhile, different regions of CXCR2 such
as C/N-terminus, transmembrane domain and loop specifically contribute to ligand binding and
specificity, receptor desensitization and endocytosis, receptor phosphorylation, and G-protein
2
coupling. Therefore, antagonists with different binding sites on CXCR2 could behave
differently in modulating the down-stream signaling of CXCR2/CXCL8 axis. It is most likely
2
4

that TUTP compounds act as biased ligands and their interaction with CXCR2 does not shut
down the entire receptor. Some other biased ligands that selectively alter GPCR-mediated
36, 48-51
signaling have been reported in GPCRs including CXCR2, CXCR4 and CXCR7.
We also demonstrated that TUTP compounds inhibited cancer cell migration induced by CXCL-
, which is an important driver of tumor metastasis. TUTP compounds also exhibited synergistic
8
effect with doxorubicin in ovarian cancer cells at non cytotoxic concentrations, which may be
advantageous for aiding standard-of-care chemotherapies. Additionally, further development and
characterization of TUTP compounds will offer important insights into CXCR2 signaling and
function, facilitating the design and application of CXCR2 inhibitors.
Figure 8. Schematic illustration of the proposed mechanism of action of CX797, TUTP
compound 52 and SB265610.
EXPERIMENTAL SECTION
2
5

Chemistry. Unless otherwise specified, all reactions were carried out in oven-dried glassware
with magnetic stirring. All commercial reagents and anhydrous solvents were purchased and
used without purification, unless specified. Column chromatography was performed on silica gel
(200−300 mesh). Preparative HPLC was performed on Shimadzu LC20-AT system, using
Kinetex® 5 µm, XB-C18 100 Å, 150 x 21.2 mm column at room temperature. HPLC gradient
method utilized was 10% to 95% MeCN in H O with 0.05% trifluoroacetic acid over 25 minutes
2
with a 8 mL/min flow rate. NMR spectra were recorded on a Bruker 300 or 400MHz NMR
1
spectrometer. Chemical shifts for proton magnetic resonance spectra ( H NMR) and carbon
1
3
magnetic resonance spectra ( C NMR) are quoted in parts per million (ppm) referenced to the
appropriate solvent peak or 0.0 ppm for tetramethylsilane (TMS). The following abbreviations
are used to describe the peak-splitting patterns when appropriate: br, broad; s, singlet; d, doublet;
t, triplet; q, quartet; m, multiplet; and dd, doublet of doublets. Coupling constants, J, are reported
in hertz (Hz). All compounds were
liquid chromatography system using Kinetex® 2.6 µm, XB-C18 100 Å, 75 x 4.6 mm column at
room temperature. HPLC gradient method utilized was 10% to 95% MeCN in H O with 0.1%
≥95% purity as determined by Shimadzu LC-2030C 3D
2
formic acid over 15 minutes with a 0.80 mL/min flow rate. Mass spectra were recorded using
ESI ion source on a Shimadzu LCMS-2020 system.
Ethyl 2-amino-4-methyl-4,5,6,7-tetrahydrobenzo[b]thiophene-3-carboxylate (2a). A mixture of
2
0
-methylcyclohexanone (33.6 g, 0.3 mol), ethyl 2-cyanoacetate (33.9 g, 0.3 mol), sulfur (9.6 g,
.3 mol) and morpholine (26.1 g, 0.3 mol) were stirred at room temperature for 48 h. The
mixture was directly purified with silica chromatography (5% EtOAc in hexane) to give a pale
1
yellow solid (30.1 g, 42%). H NMR (400 MHz, CDCl
3
)
δ
4.37–4.19 (m, 2H), 3.30–3.19 (m,
2
6

1
H), 2.52–2.45 (m, 2H), 1.94–1.70 (m, 3H), 1.67–1.58 (m, 1H), 1.35 (t, J = 7.1 Hz, 3H), 1.15 (d,
+
J
= 6.8 Hz, 3H). LC-MS (ESI) m/z 240.1 [M + H] .
Isopropyl 2-amino-4-methyl-4,5,6,7-tetrahydrobenzo[b]thiophene-3-carboxylate (2b). Using a
similar procedure as described for 2a with 2-methylcyclohexanone (560 mg, 5.0 mmol),
isopropyl 2-cyanoacetate (635 mg, 5.0 mmol), sulfur (160 mg, 5.0 mmol) and morpholine (435
mg, 5.0 mmol) and purified with silica chromatography (20% EtOAc in hexane). Light yellow
1
solid (806 mg, 64%). H NMR (300 MHz, CDCl
3
)
δ
5.22 (p,
J
= 6.2 Hz, 1H), 3.30 (d,
J = 12.4
Hz, 1H), 2.53 (s, 2H), 1.98–1.61 (m, 4H), 1.36 (t,
J
= 6.6 Hz, 6H), 1.18 (d,
J
= 6.8 Hz, 3H).
tert-Butyl 2-amino-4-methyl-4,5,6,7-tetrahydrobenzo[b]thiophene-3-carboxylate (2c). Using a
similar procedure as described for 2a with 2-methylcyclohexanone (794 mg, 7.1 mmol), tert-
butyl 2-cyanoacetate (1.0 g, 7.1 mmol), sulfur (227 mg, 7.1 mmol) and morpholine (617 mg, 7.1
mmol), and purified with silica chromatography (10% EtOAc in hexane). White solid (1.38 g,
1
7
1
2
3%). H NMR (300 MHz, CDCl )
δ
3.25 (t,
J
= 6.3 Hz, 1H), 2.51 (s, 2H), 1.94–1.61 (m, 5H),
3
.58 (s, 8H), 1.18 (d,
J = 6.8 Hz, 3H).
-Amino-4-methyl-4,5,6,7-tetrahydrobenzo[b]thiophene-3-carboxamide (2d). Using a similar
procedure as described for 2a with 2-methylcyclohexanone (667 mg, 6.0 mmol), 2-
cyanoacetamide (500mg, 6.0 mmol), sulfur (190 mg, 6.0 mmol) and morpholine (518 mg, 6.0
mmol), and purified with silica chromatography (10% MeOH in DCM). Light yellow solid (213
1
mg, 17%). H NMR (300 MHz, CDCl )
δ
3.08 (t,
J
= 6.6 Hz, 1H), 2.51 (t, J = 5.9 Hz, 2H), 1.97–
3
+
1
.56 (m, 4H), 1.17 (d,
J
= 6.9 Hz, 3H). LC-MS (ESI) m/z 211.1 [M + H] .
tert-Butyl 2-amino-4,5,6,7-tetrahydrobenzo[b]thiophene-3-carboxylate (2e). Using a similar
procedure as described for 2a with cyclohexanone (500 mg, 5.1 mmol), tert-butyl 2-cyanoacetate
(719 mg, 5.1 mmol), sulfur (163 mg, 5.1 mmol) and morpholine (444 mg, 5.1 mmol), and
2
7

1
purified with silica chromatography (10% EtOAc in hexane). White solid (1.0 g, 78%). H NMR
300 MHz, CDCl3) 2.72–2.65 (m, 2H), 2.55–2.48 (m, 2H), 1.82–1.73 (m, 4H), 1.56 (s, 9H).
tert-Butyl 2-amino-4-methyl-5,7-dihydro-4H-thieno[2,3-c]pyran-3-carboxylate (2f). Using a
similar procedure as described for 2a with 3-methyldihydro-2 -pyran-4(3 )-one (150 mg, 1.32
(
δ
H
H
mmol), tert-butyl 2-cyanoacetate (186 mg, 1.32 mmol), sulfur (42 mg, 1.32 mmol) and
morpholine (115 mg, 1.32 mmol), and purified with silica chromatography (20% EtOAc in
1
hexane). Light yellow solid (286 mg, 81%). H NMR (300 MHz, CDCl
3
)
δ
4.66–4.52 (m, 2H),
= 6.8 Hz, 3H).
tert-Butyl 2-amino-5,7-dihydro-4H-thieno[2,3-c]pyran-3-carboxylate (2g). Using a similar
procedure as described for 2a with dihydro-2 -pyran-4(3 )-one (500 mg, 5.0 mmol), tert-butyl
-cyanoacetate (705 mg, 5.0 mmol), sulfur (160 mg, 5.0 mmol) and morpholine (435 mg, 5.0
3
.90–3.74 (m, 2H), 3.04 (d, J = 7.4 Hz, 1H), 1.58 (s, 9H), 1.31 (d, J
H
H
2
mmol and purified with silica chromatography (20% EtOAc in hexane). White solid (1.18 g,
1
9
2
2%). H NMR (300 MHz, CDCl )
δ
4.56 (d,
J
= 2.1 Hz, 2H), 3.91 (t,
J
= 5.6 Hz, 2H), 2.85–
3
.75 (m, 2H), 1.55 (s, 9H).
tert-Butyl 6-amino-3,4-dihydro-2H-thieno[2,3-b]pyran-5-carboxylate (2h). Using a similar
procedure as described for 2a with dihydro-2 -pyran-3(4 )-one (300 mg, 3.0 mmol), tert-butyl
-cyanoacetate (423 mg, 3.0 mmol), sulfur (96 mg, 3.0 mmol) and morpholine (261 mg, 3.0
H
H
2
mmol), and purified with silica chromatography (20% EtOAc in hexane). Light yellow solid
1
(
425 mg, 56%). H NMR (300 MHz, CDCl )
δ
4.20–4.11 (m, 2H), 2.68 (t,
J
= 6.5 Hz, 2H), 1.99
3
(dt, J = 10.8, 6.2 Hz, 2H), 1.55 (s, 9H).
tert-Butyl 2-amino-6,6-dimethyl-6,7-dihydro-4H-thieno[3,2-c]pyran-3-carboxylate (2i). Using a
similar procedure as described for 2a with 6,6-dimethyldihydro-2 -pyran-3(4 )-one (300 mg,
.34 mmol), tert-butyl 2-cyanoacetate (330 mg, 2.34 mmol), sulfur (75 mg, 2.34 mmol) and
H
H
2
2
8