Novel naftopidil derivatives containing methyl phenylacetate and their blocking effects on α1D/1A-adrenoreceptor subtypes

Article abstract of DOI:10.1016/j.bmcl.2018.01.068

α₁-Adrenoceptor (α₁-AR) antagonists are considered to be the most effective monotherapy agents for lower urinary tract symptoms associated with benign prostatic hyperplasia (LUTS/BPH). In this study, we synthesized compounds 2–17, which are novel piperazine derivatives that contain methyl phenylacetate. We then evaluated the vasodilatory activities of these compounds. Among them, we found that compounds 2, 7, 12, which contain 2-OCH₃, 2-CH₃ or 2, 5-CH₃, respectively, exhibited potent α₁-blocking activity similar to protype drug naftopidil (1). The antagonistic effects of 2, 7, and 12 on the (?)-noradrenaline-induced contractile response of isolated rat prostatic vas deferens (α_1A), spleen (α_1B) and thoracic aorta (α_1D) were further characterized to assess the sub receptor selectivity. Compared with naftopidil (1) and terazosin, compound 12 showed the most desirable α_1D/1A subtype selectivity, especially improved α_1A subtype selectivity, and the ratios pA₂ (α_1D)/pA₂ (α_1B) and pA₂ (α_1A)/pA₂ (α_1B) were 17.0- and 19.5-fold, respectively, indicating less cardiovascular side effects when used to treat LUTS/BPH. Finally, we investigated the chiral pharmacology of 12. We found, however, that the activity of enantiomers (R)-12 and (S)-12 are not significantly different from that of rac-12.

Full text of DOI:10.1016/j.bmcl.2018.01.068

Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel naftopidil derivatives containing methyl phenylacetate and their
blocking effects on a1D/1A-adrenoreceptor subtypes
b
b
a
a
c
a,c
Jun-Jun Huang a, , Zhi-Han Zhang , Fei He , Xia-Wen Liu , Xing-Jie Xu , Li-Jun Dai , Qi-Meng Liu
,
⇑
Mu Yuan a,
⇑
a Guangdong Provincial Key Laboratory of Molecular Target & Clinical Pharmacology, School of Pharmaceutical Sciences and the Fifth Affiliated Hospital, Guangzhou
Medical University, Guangdong 511436, PR China
b Guangdong Province Key Laboratory of Microbial Signals and Disease Control, Department of Plant Pathology, South China Agricultural University, Guangzhou 510642, PR China
c Laboratory Animal Center, Guangzhou Medical University, Guangzhou, Guangdong 511436, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
a
1-Adrenoceptor (a1-AR) antagonists are considered to be the most effective monotherapy agents for
Received 20 November 2017
Revised 30 January 2018
Accepted 31 January 2018
Available online xxxx
lower urinary tract symptoms associated with benign prostatic hyperplasia (LUTS/BPH). In this study,
we synthesized compounds 2–17, which are novel piperazine derivatives that contain methyl phenylac-
etate. We then evaluated the vasodilatory activities of these compounds. Among them, we found that
compounds 2, 7, 12, which contain 2-OCH3, 2-CH3 or 2, 5-CH3, respectively, exhibited potent
activity similar to protype drug naftopidil (1). The antagonistic effects of 2, 7, and 12 on the (À)-
noradrenaline-induced contractile response of isolated rat prostatic vas deferens ( 1A), spleen ( 1B) and
thoracic aorta ( 1D) were further characterized to assess the sub receptor selectivity. Compared with
naftopidil (1) and terazosin, compound 12 showed the most desirable a1D/1A subtype selectivity, espe-
cially improved a1A subtype selectivity, and the ratios pA2 1D)/pA2 1B) and pA2 1A)/pA2 1B) were
a1-blocking
Keywords:
a
a
a
1-Adrenoreceptor
a
Benign prostatic hyperplasia
Piperazine derivatives
Functional assays
(
a
(a
(
a
(a
17.0- and 19.5-fold, respectively, indicating less cardiovascular side effects when used to treat LUTS/
BPH. Finally, we investigated the chiral pharmacology of 12. We found, however, that the activity of
enantiomers (R)-12 and (S)-12 are not significantly different from that of rac-12.
Ó 2018 Published by Elsevier Ltd.
Lower urinary tract symptoms (LUTS) suggestive of benign pro-
static hyperplasia (BPH) are highly prevalent in aging men.1,2 The
American and European Urological Associations recommend the
of GABA and glycine via a receptor other than an a1 adrenocep-
tor.10 And NAF is also effective for reduced the spinal cord
injury-related tissue remodeling in the bladder.11
use of
for the first-line medical treatment for patients with LUTS/BPH.3
1-AR is linked to Gq/11 protein involving phospholipase
activation, and further divided into a1A 1B, and
1D-subtypes.4
a
1-AR antagonists as prostatic smooth muscle relaxants
However, NAF has some disadvantages, such as low subtype
selectivity, high required dosage, and low bioavailability.12 N-het-
ero arylpiperazine derivatives have been structurally modified to
improve their affinity and selectivity.6,13,14 Research of NAF and
similar a1 subtype blockers has been ongoing for years in our lab.
The present study is a continuation of our previous work on the
synthesis of heterocyclic compounds containing acetophenone.15
a
C
,
a
a
As the three subtypes having distinct pharmacology and tissue
expression5, recent studies have shown that treatment with antag-
onists with better selectivities for a1A-AR and a1D-AR over a1B-AR
may efficiently alleviate BPH symptoms without causing signifi-
The proposed ligand-based pharmacophore model for a1-AR antag-
cant cardiovascular side effects.6,7
onists is mainly consisted of positive ionisable (PI), hydrophobic
features (HY) and hydrogen bond acceptor (HBA).16–18 Our docking
study of two NAF enantiomers towards a1D showed that the right
arylpiperazine part was placed on the entrance of hydrophobic
pocket, and the left naphthalene moiety entered into a deep
hydrophobic region.19 We changed naphthalene to methyl pheny-
lacetate in order to improve the hydrophobic property and thus
novel derivatives 2–17 (Scheme 1) were designed and synthesized.
All new compounds were subjected to preliminary pharmacologi-
cal screening for their vasodilatory effects. Finally, a functional
Naftopidil (NAF, Fig. 1) is a N-aryl piperazine derivative and a
specific
a1D-AR antagonist. It has been marketed in Japan since
1999 for the treatment of LUTS/BPH and proven to be more effec-
tive than tamsulosin, a major a1A blocker, in treating storage symp-
toms.8,9 Recent reports indicate that NAF modulates sensory
transmission in the substantia gelatinosa by enhancing the release
⇑
Corresponding authors.
E-mail address: huangjunjun1985@sina.com (J.-J. Huang).
https://doi.org/10.1016/j.bmcl.2018.01.068
0960-894X/Ó 2018 Published by Elsevier Ltd.
Please cite this article in press as: Huang J.-J., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.01.068

2
J.-J. Huang et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
O
O
O
N
H
O
O
N
N
N
O
S
N
H
OH
N
O
O
O
N
O
O
NH2
naftopidil (NAF, 1)
tamsulosin
terazosin
Fig. 1. Marketed drugs for treating LUTS/BPH.
HN
b
O
N
a
N
6
3
O
R1
+
R1 OH
R1 OH
N
Cl
O
+
5
R2
O
O
R2
2
2-17
4
ꢀ
2: R2=2-OCH3
3: R2=3-OCH3
4: R2=4-OCH3
8: R2=3- CH3
9: R2=4- CH3
14: R2=3,5-CH3
R1=
15: R2=4-CH3-5-Cl
16: R2=2-Cl
O
10: R2=2,3- CH3
5: R2=2, 4-OCH3-5-Cl 11: R2=2,4- CH3
17: R2=4-Br
6: R2=3-OCH2CH3
7: R2=2-CH3
12: R2=2,5- CH3
13: R2=3,4-CH3
Scheme 1. Synthetic route of target compounds 2–17. Reagents and conditions: a) NaOH, 0 ꢀ 60 °C; b) 1-(substituted phenyl)-piperazine, 2-propanol, reflux.
a
1-AR subtype selectivity study of selected compounds (2, 7, and
pointingly low vasodilatory activity. Out of all derived compounds,
6 with 3-OCH2CH3 showed the worst 1-blocking activity. Its poor
12) was performed. Results showed that methyl phenylacetate
derivative with 2, 5-CH3 group (12) was a moderately selective
a1D/1A blocker with lower antagonist potency against rat spleen
a
performance could be attributed to large steric hindrance. Para-
substituted derivatives (4, 9, 15, and 17) also exhibited poor phar-
macological effects. Among the disubstituted derivatives 11–14, 2,
5-disubstituted compounds exhibited preferable activity second
only to 2-substituted compounds. For example, the activity of com-
pound 12 was more potent than those of other disubstituted
derivatives. Among compounds 7–14, which were derived through
one or more methyl substitution, 2-CH3 (7) and 2, 5-CH3 (12)
exhibited most desirable vasodilatory activity.
a1B-AR than NAF (pA2 of 6.12 for 12 vas 6.75 for NAF).
To begin with, compounds 2–17 were synthesized through the
routes outlined in Scheme 1. Substituted phenols were converted
into epoxide through nucleophilic substitution with epichlorohy-
drin, followed by substitution with piperazine in 2-propanol to
yield the target compounds 2–17.15 The derivatives were charac-
terized on the basis of HRESI-MS, 1H NMR and 13C NMR spectral
data, which were fully consistent with their depicted structures.
Compounds 2–17 were then preliminarily screened at concen-
Among all methyl phenylacetate-derived arylpiperazines, 2, 7,
and 12 substituted with 2-OCH3, 2-CH3 or 2, 5-CH3, producing
trations of 0.1, 1, and 10
thoracic aortic rings precontracted with (À)-noradrenaline (NE)
where b- and 2-ARs were inhibited. NAF was used as a reference
l
M for their vasodilatory activity on rabbit
relaxation response more than 60% at 1.0
aortic rings, exhibited the best vasodilatory activity and a degree
of effect compared with 1. Thus, the 1-subtype blocking activity
lM on rabbit thoracic
a
a
compound. The percentage response produced by various com-
of these derivatives was further studied.
pounds is shown Table 1. NAF (1) was initially used as an antihy-
It has been suggested that contractions of rat vas deferens and
spleen to exogenous NE are mediated predominantly by a1A and
a1B adrenoceptor, whereas contractions to NE in rat aorta are
mediated predominantly by a1D adrenoceptor.21,22 Thus, contrac-
tions of these three tissues to NE have been used as a functional
pertensive drug given its strong vasodilatory activity (1
lM, 86.9
3.3%; 10
l
M, 97.8 1.5%).20 The replacement of methyl pheny-
lacetate resulted in compounds with slightly weaker activity than
prototype 1, and some compounds, such as 4 and 6, exhibited weak
performances. Different substituent groups, including alkoxy, alkyl
and halogen, were designed on the phenylpiperazine moiety to
verify the role of different groups. Meanwhile, substitution at the
meta-, para- and ortho-positions conferred differing levels of phar-
macological activity to the derivatives. In compounds 2–4, ortho-
substitution (2) exhibited the best a1-blocking activity, as well as
compounds 7 and 16 (relaxation response >90%). Substitution on
position 3-alone (3, 6, and 8) resulted in compounds with disap-
model for a1A
of compounds 2, 7, and 12 on Sprague-Dawley rat prostatic vas
deferens ( 1A), spleen ( 1B), and thoracic aorta ( 1D) were charac-
, a1B,
a1D adrenoceptors.23 The antagonistic effects
a
a
a
terized to assess the sub receptor selectivity of the compounds
(Table 2). NAF and terazosin were used as reference compounds.
All three compounds appeared to be competitive antagonists at
each sub receptor, because the slopes of the Schild plots were
not significantly different from unity on the basis of parallel shifts
Please cite this article in press as: Huang J.-J., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.01.068

J.-J. Huang et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
3
Table 1
Percentage relaxation response of compounds 2–17 at 0.1, 1 and 10
lM in rabbit thoracic aorta, previously contracted with NE (3 lM).
O
N
6
3
OH
O
N
5
R2
4
2
O
Comp.
R2
% Relaxation Response SEM (n = 5)
0.1
lM
1
l
M
10 lM
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1
2-OCH3
3-OCH3
4-OCH3
2,4-OCH3-5-Cl
3-OCH2CH3
2-CH3
27.6 4.8
13.0 3.1
6.3 1.4
8.2 2.9
6.1 0.8
27.3 5.4
14.8 1.1
15.3 2.5
12.5 2.0
13.7 1.9
24.3 4.5
8.9 1.3
5.8 0.5
71.1 3.5
23.7 2.0
7.0 3.7
100.0 6.6
39.8 2.8
15.8 2.9
46.1 9.0
14.9 1.6
99.3 9.5
63.5 2.8
56.7 6.2
81.7 8.8
92.7 4.4
100.0 6.3
70.2 5.8
61.4 2.1
38.7 5.1
98.3 3.6
43.5 2.4
97.8 1.5
20.4 5.0
10.2 2.5
62.3 7.4
36.9 3.2
28.7 3.4
43.3 6.8
43.3 5.0
66.2 6.9
28.4 3.2
18.6 1.4
25.9 3.9
58.7 1.9
16.6 4.2
86.9 3.3
3-CH3
4-CH3
2,3-CH3
2,4-CH3
2,5-CH3
3,4-CH3
3,5-CH3
4-CH3-5-Cl
2-Cl
19.5 4.8
29.0 2.4
7.9 0.9
4-Br
–
25.1 1.4
Table 2
Functional antagonistic potency of 2, 7, and 12 together with its enantiomers, expressed as pA2, at
spleen ( 1B) and thoracic aorta ( 1D).
a
1-adrenoceptor subtypes of SD rat isolated tissues: prostatic vas deferens (
Selectivity ratio
a1A),
a
a
Compound
pA2a(slope)
a1A
a1B
a1D
a1D
/a1A
a1D
/
a1B
a1A/a1B
2
6.83 0.07
(1.09 0.03)
6.31 0.09
(1.17 0.15)
7.41 0.21
(1.06 0.02)
7.20 0.28
(1.00 0.04)
7.52 0.06
(1.20 0.18)
7.48 0.07
(1.00 0.11)
7.90 0.15
(1.13 0.07)
6.22 0.07
(1.03 0.15)
6.49 0.14
(0.86 0.07)
6.12 0.34
(0.84 0.10)
6.35 0.22
(0.92 0.10)
6.57 0.21
(1.06 0.07)
6.75 0.11
(1.22 0.14)
8.59 0.08
(0.99 0.12)
7.11 0.03
(1.22 0.20)
7.70 0.12
(1.12 0.04)
7.35 0.14
(1.14 0.05)
7.42 0.18
(0.99 0.12)
7.81 0.25
(1.15 0.02)
7.93 0.11
(1.10 0.04)
8.83 0.17
(1.06 0.19)
1.91
24.55
0.87
1.66
1.95
2.82
8.51
7.76
4.07
0.66
19.50
7.08
8.91
5.37
0.20
7
16.22
17.00
11.75
17.38
15.14
1.74
12
(R)-12
(S)-12
1b
terazosinc
a
b
c
pA2 values, expressed as mean S.E.M. of three different concentrations, each tested at least four times.
The corresponding pKi values were 8.43 0.06, 7.70 0.02, 8.92 0.00, obtained in binding experiment.8
The reference data was 8.04, 8.60, 8.65.24
to concentration-response curves. In addition, treatment with
increasing concentrations of these compounds did not decrease
the maximum agonist-induced contractile response (Fig. 2).
The blocking activities of the test compounds on the three a1
subtypes were less effective than that of NAF (1). The pA2 values
azosin showed no subtype selectivity at all. Introduction of methyl
moiety as in 7 and 12 gave rise to the antagonistic activity of a1D
Compared with 1 and terazosin, 12 exhibited higher selectivity for
1D and 1A, with selectivity ratios of approximately 17- and 19.5-
fold. Given its low potency for 1B, 12 may be an effective agent for
.
a
a
a
of compounds 2 and 7 for
are considerably lower than that of 1 (pA2 = 7.48). Similar to 1, the
selectivity of all compounds for 1D were more potent than that for
1A. However, the compounds had different subtype selectivity. As
we can see, the selectivity ratios pA2 ( 1D)/pA2 ( 1B) and pA2 ( 1A)/
pA2 ( 1B) of 1 were approximately 15- and 5-fold, respectively. Ter-
a
1A were 6.83 and 6.31 (<7); these values
improving symptoms related to bladder filling in BPH.
As 12 has one chiral carbon center, we also synthesized two
enantiomers of 12, namely, (R)-12 and (S)-12 (Scheme 2). Optically
active compounds were prepared with optically active (2R)- or
(2S)-glycidyltosylate in DMF to obtain chiral epoxide25, followed
by substitution with piperazine in 2-propanol to provide the target
a
a
a
a
a
a
Please cite this article in press as: Huang J.-J., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.01.068

4
J.-J. Huang et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
prostatic vas deferens
spleen
(A)
(B)
120
120
100
80
60
40
20
0
100
80
60
40
20
0
12 [
μ
M]
0
0.3
1
3
12 [μM]
0
1
3
10
-12
-10
-8
-6
-4
-2
-12
-10
-8
-6
-4
-2
log [NE], M
log [NE], M
thoracic aorta
(D)
2.5
(C)
120
α1A
α1B
α1D
100
80
60
40
20
0
2.0
1.5
1.0
0.5
0.0
12 [μM]
0
0.3
1
3
-7.0
-6.5
-6.0
-5.5
-5.0
-4.5
-12
-10
-8
-6
-4
-2
log [compound 12], M
log [NE], M
Fig. 2. Concentration-response curves of NE-induced contractions in rat isolated prostatic vas deferens (A), spleen (B), and thoracic aorta (C) in compound 12 treatment at
different concentrations. Data are the mean SEM of five to eight separate experiments. (D) Schild plots for the antagonism to NE response through compound 12 in rat
isolated rat prostatic vas deferens (
a
1A), spleen (
a
1B) and thoracic aorta (
a
1D), whose slopes were not significantly different from unity.
(R)
O
O
N
O
(R)
(S)
OH
N
(R)
OH
TsO
O
O
a
O
b
(R)-12
O
O
HN
O
N
(S)
a
O
O
N
O
O
OH
N
(S)
TsO
O
O
O
(S)-12
O
O
Scheme 2. Enantiomeric synthesis of compound 12. Reagents and conditions: a) DMF, NaH, rt; b) 2-propanol, reflux.
enantiomers (R)-12 and (S)-12 (Scheme 2). These enantiomers did
not exhibit substantially different blocking activities or sub recep-
tor selectivity, as evidenced by the results of functional tests
(Table 2). (S)-12 exhibited slightly better blocking activity than
the (R)-enantiomer and its racemate without increased subtype
however, both (R)-NAF and (S)-NAF displayed similar affinity7, as
well as compound 12 in current paper.
Compared with the previous acetophenone15, carbazole and
benzotriazole17,26 derivatives, methyl phenylacetate series did
not display a better antagonistic effect on a1 subtypes, as their
pA2 values were all less than eight. It is worth noting that methyl
phenylacetate derivative 12 has highly improved the subtype
selectivity. Similar blocking effect on the
among 12 and two enantiomers, however, the pA2
of 12 exhibited higher than the two enantiomers, as the low affin-
ity on 1B subtype (pA2 of 6.12 for 12). Chiral pharmacological dif-
ferences were observable in some structures of NAF derivatives15
a1A subtype was observed
(a1A)/pA2
(a1B)
selectivity on the
for the 1A than for the
type selectivity was reduced when using the enantiomers of 12.
a1A, that is, 12 exhibited 19.5-fold higher potency
a
a
a
1B. Moreover, this effect of higher 1A sub-
a
,
Please cite this article in press as: Huang J.-J., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.01.068

J.-J. Huang et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
5
In conclusion, we report the synthesis, structural characteriza-
tion, and in vitro vasodilatory effect of novel piperazine derivatives.
Compounds 2–17 were designed from naftopidil (1), an 1-adreno-
ceptor subtype antagonist used for the treatment of LUTS/BPH.
Compound 12 is a moderate therapeutic agent for BPH given its
References
1. Mobley D, Feibus A, Baum N. Postgrad Med. 2015;127:301–307.
2. Berry SJ, Coffey DS, Walsh PC, et al. J Urol. 1984;132:474–479.
3. Van Asseldonk B, Barkin J, Elterman DS. Can J Urol. 2015;22:7–17.
4. Zhong H, Minneman KP. Eur J Pharmacol. 1999;375:261–276.
5. Hennenberg M, Stief CG, Gratzke C. Neurourol Urodyn. 2014;33:1074–1085.
6. Huang JJ, He F, Huang MY, et al. Eur J Med Chem. 2015;96:83–91.
7. Huang JJ, Cai Y, Huang MY, et al. Euro J Pharmacol. 2016;791:473–481.
8. Takei R, Ikegaki I, Shibata K, et al. Jpn J Pharmacol. 1999;79:447–454.
9. Oh-oka H. Urology. 2008;72:1051–1055.
a
combined antagonistic effects against a1A and
a1D. Moreover, no
obvious differences among stereo configurations of 12 exist.
Rac-12 exhibits higher a1A subtype selectivity than the two
enantiomers. Our results show that the methyl phenylacetate
ligand is a versatile class of compounds that provide a new starting
point for the development of ligands with high selectivity for a1
subtypes.
10. Uta D, Xie DJ, Hattori T, et al. J Neurol Sci. 2017;380:205–211.
11. Kadekawa K, Majima T, Kawamorita N, et al. Neurourol Urodyn. 2016;9999:1–8.
12. Liu XW, Zhang XF, Huang JJ, et al. J Pharm Biomed. 2015;112:147–154.
13. Sorbi C, Franchini S, Tait A, et al. ChemMedChem. 2009;4:393–399.
14. Némethy A, Vavrinec P, Vavrincová-Yaghi D, et al. Arch Pharm. 2017;350.
https://doi.org/10.1002/ardp.201600394. n/a.
15. Huang JJ, Huang YJ, Zhu L, et al. Pharmazie. 2014;69:578–584.
16. Abou-Seri SM, Abouzid K, Abou El Ella DA. Eur J Med Chem. 2011;46:647–658.
17. Xu W, Huang JJ, Shao BH, et al. Bioorg Med Chem. 2016;24:5565–5572.
18. MacDougall IJA, Griffith R. J Mol Graph Model. 2006;25:146–157.
19. Xu W, Huang JJ, Jiang RW, et al. Acta Pharmaceut Sin B. 2017;7:496–501.
20. Himmel HM. Cardiovasc Drug Rev. 1994;12:32–47.
21. Aboud R, Shafii M, Docherty JR. Br J Pharmacol. 1993;109:80–87.
22. Buckner SA, Oheim KW, Morse PA, et al. Eur J Pharmacol. 1996;297:241–248.
23. Docherty JR. Cell Mol Life Sci. 2010;67:405–417.
24. Hancock AA, Buckner SA, Brune ME, et al.
2002;300:478–486.
Conflicts of interest
There are no conflicts of interest to declare.
Acknowledgements
This work was supported by the Natural Science of Guangdong
Province (B17016005002), the Education Administration Research
Foundation of Guangzhou City (1201630368), and Guangdong
Science and Technology Plan Project (2013B060300023).
J Pharmacol Exp Ther.
25. Klunder JM, Ko SY, Sharpless KB. J Org Chem. 1986;51:3710–3712.
26. Huang JJ, Zhu L, Huang YJ, et al. Lat Am J Pharm. 2014;33:273–277.
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.bmcl.2018.01.068.
Please cite this article in press as: Huang J.-J., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.01.068