Identification of a dual δ or antagonist/μ or agonist as a potential therapeutic for diarrhea-predominant Irritable Bowel Syndrome (IBS-d)

Article abstract of DOI:10.1016/j.bmcl.2012.05.042

A small set of acyclic analogs 5 were prepared to explore their structure-activity relationships (SARs) relative to heterocyclic core, opioid receptor (OR) agonists 4. Compound 5l was found to have very favorable OR binding affinities at the δ and μ ORs (r K_i δ = 1.3 nM; r K_i μ = 0.9 nM; h K_i μ = 1.7 nM), with less affinity for the κ OR (gp K_i κ = 55 nM). The OR functional profile for 5l varied from the previously described dual δ/μ OR agonists 4, with 5l being a potent, mixed dual δ OR antagonist/μ OR agonist [δ IC₅₀ = 89 nM (HVD); μ EC₅₀ = 1 nM (GPI); κ EC₅₀ = 1.6 μM (GPC)]. Compound 5l has progressed through a clinical Phase II Proof of Concept study on 800 patients suffering from diarrhea-predominant Irritable Bowel Syndrome (IBS-d). This Phase II study was recently completed successfully, with 5l demonstrating statistically significant efficacy over placebo.

Full text of DOI:10.1016/j.bmcl.2012.05.042

Bioorganic & Medicinal Chemistry Letters 22 (2012) 4869–4872
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification of a dual d OR antagonist/
l OR agonist as a potential
therapeutic for diarrhea-predominant Irritable Bowel Syndrome (IBS-d)
,
⇑
Henry J. Breslin , Craig J. Diamond , Robert W. Kavash , Chaozhong Cai, Alexey B. Dyatkin ,
Tamara A. Miskowski , Sui-Po Zhang, Paul R. Wade , Pamela J. Hornby, Wei He
Janssen Research and Development, L.L.C. (formerly known as Johnson and Johnson Pharmaceutical Research and Development, L.L.C.), Welsh and McKean Roads, PO Box 776,
Spring House, PA 19477-0776, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A small set of acyclic analogs 5 were prepared to explore their structure–activity relationships (SARs) rel-
ative to heterocyclic core, opioid receptor (OR) agonists 4. Compound 5l was found to have very favorable
Received 16 March 2012
Revised 4 May 2012
Accepted 8 May 2012
Available online 24 May 2012
OR binding affinities at the d and
for the OR (gp Ki = 55 nM). The OR functional profile for 5l varied from the previously described dual
OR agonists 4, with 5l being a potent, mixed dual d OR antagonist/ OR agonist [d IC50 = 89 nM (HVD);
EC50 = 1 nM (GPI); EC50 = 1.6 M (GPC)]. Compound 5l has progressed through a clinical Phase II
l ORs (r Ki d = 1.3 nM; r Ki l = 0.9 nM; h Ki l = 1.7 nM), with less affinity
j
j
d/
l
l
l
j
l
Keywords:
Mixed d OR antagonist/l OR agonist
Opioid receptor
Irritable Bowel Syndrome
IBS-d
Proof of Concept study on 800 patients suffering from diarrhea-predominant Irritable Bowel Syndrome
(IBS-d). This Phase II study was recently completed successfully, with 5l demonstrating statistically sig-
nificant efficacy over placebo.
Ó 2012 Elsevier Ltd. All rights reserved.
Compounds that modulate opioid receptors (ORs) have long
been accepted as a standard of care for pain management (e.g.,
morphine, 1, and fentanyl, 2), and more recently have become
appreciated as therapeutics for their gastrointestinal (GI) motility
modulation (e.g., loperamide, 3) (Fig. 1).1,2 The ORs are categorized
as OR ligands. Key experimental findings are discussed below for
some ‘B’ acyclic analogs 5. These analogs 5 are summarized in
Table 1, along with their respective d and
An initial set of acyclic analogs (5a–5d) had modest OR binding
affinities, considerably poorer than parent cyclic analogs 4a–c (
Ki d: >200-fold; Ki : 20–250-fold). For analogs such as 5a–5d,
l OR binding affinities.
D
into three major subclasses, d,
l
, and
j
, with their classifications
D
l
based on well defined pharmacological profiles.2
where the amide R0 is H, it is well established that the expected ste-
reochemical preference around the 2° amide bond is more predom-
inately pseudo trans (Fig. 4) relative to analogous 3° amides.
Conversely, there is a higher percentage of pseudo cis configuration
expected for 3° amides such as 4, which is the observed energy-min-
imized modeling configuration of 4a.3 This pseudo cis/trans amide
conjecture for analogs 4 and 5 is a potential contributing factor for
the noted relative loss of OR binding affinities for 5a–5d. Further
supportive evidence for this pseudo cis/trans amide supposition is
that the additionally R00 sterically encumbered 5c (R0 = i-Pr; Ki
We previously reported on a series of phenyl imidazoles 4 that
demonstrated potent dual d/
l OR agonist activities, both in vitro
and in vivo (Fig. 2).3,4 Initially described was the discovery of 4a,
a potent d OR agonist whose OR activity was directly correlated
with modulation of GI motility in vivo.3 Subsequently reported
was the improvement of 4a’s overall OR activities by exploring
variations of its Tyr moiety (arbitrarily labeled ‘A’ segment), ulti-
mately yielding the more potent compounds 4b–e at both the d
and
l
ORs.4
Following are described secondary structure–activity relation-
ship (SAR) endeavors evaluating OR activities relative to alterations
of the central tetrahydroisoquinoline (Tic) and piperidine ‘B’ seg-
ments of analogs 4a–e. Recognizing that this SAR exploration could
prove laboriously slow if various core substituted heterocycles had
to be synthesized in lieu of 4a–e’s Tics and piperidines, we chose to
prepare more readily accessible acyclic analogs akin to 5 (Fig. 3)
with the hope they would prove equivalent, or hopefully better,
d = 5198 nM; Ki
l
= 121 nM) is considerably less active than 5b
NCH3
O
H
HO
N(CH3)2
HO
O
N
N
N
O
OH
2
3
CH3
1
Cl
Fentanyl
Loperamide
(Imodium)
Morphine
⇑
Corresponding author.
E-mail address: HBreslin60@gmail.com (H.J. Breslin).
(Duragesic)
Not currently employed within the Janssen organization.
Figure 1.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.05.042

4870
H. J. Breslin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4869–4872
closely resemble those for its more closely related cyclic Tic ana-
logs 4a, 4b, and 4d, whose relative d/ OR binding affinities gener-
ally favored better binding affinity at the d OR.
Ar
l
{
B
N
R
N
Having identified 5i with exceedingly desirable OR binding
N
H
affinities (Ki d = 1.5 nM; Ki
l = 0.03 nM), we subsequently replaced
O
A
{
its DMT moiety with a DMT bioisostere, 4-(aminocarbonyl)-2,6-di-
NH2
X
R
methyl-Phe,4 to give compound 5j. Analog 5j (Ki d = 12 nM; Ki
4
l
= 0.3 nM) possessed about 10-fold weaker binding affinities rela-
tive to 5i at both the d and ORs. Tighter binding at the d and
ORs was revived by exploring various substitutions on the Bn
group of 5i, as exemplified by 5k (Ki d = 0.5 nM; Ki = 1.0 nM)
and 5l (Ki d = 1.3 nM; Ki = 0.9 nM).
l
l
OR Binding
OR Functional
r Ki
r Ki
Ar
R
X
δ EC50
μ EC50
δ (nM)
0.9
0.1
μ (nM)
l
l
4a
Fused Ph
Fused Ph
---
Fused Ph
---
H
OH
OH
OH
CONH2
CONH2
55
0.3
19
0.9
37
2445
27
2
The compounds with low nanomolar binding affinities at the d
4b
4c
4d
4e
CH3
CH3
CH3
CH3
OR were next evaluated for d OR functional agonist activities, as
measured by a cell membrane-based [35S]GTP S assay6 (Table 1).
c
1.9
0.06
14
0.05
1.4
0.13
Compound 5h (Ki d = 1.4 nM; d EC50 = 103 nM) showed modest d
OR agonist functional activity, despite very promising d OR binding
affinity. Compounds 5i (Ki d = 1.5 nM; d EC50 = 20 nM) and 5j (Ki
d = 12 nM; d EC50 = 35 nM) had similar d OR agonist functional
activities, even though their OR binding affinities were ꢀ10-fold
different. The comparable OR functional results for 5i and 5j shares
additional credence to the previously reported conclusion that the
4-(aminocarbonyl)-2,6-dimethyl-Phe group is a good bioisostere
for the DMT moiety.4 Compounds 5k and 5l exhibited no d OR ago-
nist functional activities at the maximum testing concentration of
22
135
161
9
Figure 2.
(R0 = Me;Ki d = 708 nM;Ki
= 13 nM). Not surprisingly, also observed from this preliminary set
of analogs was improved OR binding affinities for the 2,6-di-Me-
phenyl substituted Tyr (DMT) analog 5b (R = Me; Ki d = 708 nM; Ki
= 17 nM) relative to Tyr analog 5a (R = H; Ki d = 5660 nM; Ki
= 1260 nM). These improved OR affinity trends for DMT analogs
l
= 17 nM)or5d(R0 = Bn;Ki d = 255 nM;Ki
l
10 l cS binding stimulation of
M, and actually inhibited [35S]GTP
l
l
the d OR agonist, SNC 80. Noteworthy, the inclusion of a meta car-
boxy moiety on the phenyl ring of R0 was the added structural com-
monality for both of these compounds that resulted in loss of all d
OR agonist functional activity, despite that both analogs main-
tained very favorable d OR binding affinities. Based on the interest-
ing d OR functional profiles for 5k and 5l, both compounds were
are consistent with our previously reported work, 4 and further illus-
trate the utility of the DMT discovery by Lazarus and co-workers.5
Next evaluated was the SAR for varied R0 and R00 alkyl substitu-
tions (5e–i), where R and X were held constant as Me and OH,
respectively (i.e., with DMT as the ‘A’ substituent). Pleasingly, OR
binding affinities for the R0 alkyl substituted acyclic analogs 5 were
more reflective overall of the favorable OR binding affinities for
parent cyclic structures 4b and 4c than had been observed for
the R0 = H analogs. For example, relatively consistent binding affin-
subsequently evaluated for
5l (
EC50 = 1 nM) proved ꢀ60-fold more potent as a
relative to 5k ( EC50 = 61 nM), where the OR functional activity
was also determined by a GTP
S assay.6
Because of 5l’s interesting preliminary d OR functional result, in
conjunction with its promising OR functional activity, it was
l
OR functional activities. Compound
l
l
OR agonist
l
l
c
ities for the
Ki d = 15 nM; Ki
d = 1.9 nM; Ki = 0.05 nM), although there was still a comparative
l
OR were seen for acyclic analog 5g (R0 = Me; R00 = Me;
= 0.1 nM) and cyclic parent analog 4c (Ki
l
l
more extensively profiled biologically. Foremost, 5l was found to
have potent d OR antagonist activity (d IC50 = 89 nM), based on a
hamster vas deferens (HVD) tissue assay.7 Identifying a dual d OR
l
loss (ꢀ7 fold) of binding affinity at the d OR for 5g. Weighing the
relative activities of 5g to 5e (R0 = Me; R00 = H; Ki d = 26 nM; Ki
antagonist/
l OR agonist compound was viewed as a potentially
l
= 0.3 nM) suggested a slight enhancement in OR binding affini-
ties by having a methyl group as R00. A more prominent example
where the OR binding affinities were enhanced by the R00 equals
Me substituent was observed when comparing 5h (R0 = i-Pr;
favorable finding, based on the reports of attenuated dependence
liability for such dual acting ligands8 as well as possible analgesic
advantages.9 In contrast to the ꢀ1 nM d and
l
OR binding affinities
= 55 nM) and in a
rather weak agonist
M).10 Compound 5l, as its dihydrochloride salt, also
R00 = Me; Ki d = 1.4 nM; Ki
d = 15 nM; Ki
affinities for acyclic analog 5h was that they mirrored those of a
cyclic relative, 4c (Ki d = 1.9 nM; Ki = 0.05 nM). The benzyl (Bn)
analog 5i also showed very favorable OR binding affinities (R0 = Bn;
R00 = Me; Ki d = 1.5 nM; Ki
= 0.03 nM). Somewhat unexpected, the
relative d/ OR binding affinities for benzyl analog 5i did not
l
= 0.03 nM) with 5f (R0 = i-Pr; R00 = H; Ki
for 5l, its affinity to the
j OR was lower (gp Ki j
guinea pig colon tissue assay was
a
l
= 0.1 nM). Most encouraging about the binding
(EC50 = 1.6
l
showed favorable pharmaceutical properties (Table 2 lists some
key results), and had very positive outcomes in a battery of
ex vivo and in vivo GI experiments, as recently determined.11
Based on 5l’s overall compelling in vitro, ex vivo, and in vivo OR
l
l
l
Break one of the red
Ar
highlighted bonds and
subsequently vary
{
B
R
substituents R' and R"
R"
O
5
R'
N
N
R
N
N
N
H
N
H
O
NH2
NH2
X
R
X
R
4
Figure 3.

H. J. Breslin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4869–4872
4871
Table 1
All final compounds 5 with respective d and
l rat OR in vitro binding affinities. The binding assays may be associated with a margin of error between 10–20%. Compounds 5
identified with d and
l
OR functional activities were screened in an OR [35S]GTP
c
S in vitro functional assay; NT = not tested
R"
R'
N
R
N
N
H
O
NH2
R
X
5
R
X
R0
R00
OR binding
r Ki
OR functional
d EC50 (nM) l EC50 (nM)
r Ki d (nM)
l
(nM)
5a
5b
5c
5d
5e
5f
5g
5h
5i
H
OH
OH
OH
OH
OH
OH
OH
OH
H
H
H
H
Me
i-Pr
Me
i-Pr
Bn
Bn
Me
Me
i-Pr
Bn
H
5660
708
5198
255
26
15
15
1.4
1.5
1260
17
121
13
0.3
0.1
NT
NT
NT
NT
NT
NT
NT
103
20
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
Me
Me
Me
Me
0.1
0.03
0.03
0.3
OH
CONH2
5j
12
35
O
O
HO
5k
5l
CH3
CONH2
CONH2
Me
Me
0.5
1.3
1.0
0.9
>10,000
>10,000
61
OMe
HO
CH3
1.0
R"
N
HO
HO
N
N
N
H
R"
R
R
X
NH2
NH2
N
R
R
HN
X
Psuedo trans tautomer more
preferred for secondary amides
relative to tertiary amides
Pseudo cis tautomer less
preferred for secondary amides
relative to tertiary amides
Figure 4.
results, in addition to the local activity of 5l in the GI tract and its
low overall oral bioavailability,11 this compound was further ad-
vanced through drug development and recently completed a Phase
II Proof of Concept clinical trial successfully in over 800 patients
with diarrhea-predominant Irritable Bowel Syndrome (IBS-d).12
The preparation of analog 5l (Scheme 1) exemplifies a standard
route used to synthesize acyclic compounds 5. The commercially
Table 2
Pharmaceutical properties for 5l
Compound 5l
Solubility (pH 7.4)a
>1 mg/mL
t1/2 = 150 min
Metabolic stability (HLM)a
b
hERG IC50
P-450 (3A4) IC50
>10
>20
l
l
M
M
c
available starting materials Cbz-N-protected
L
-alanine (6) and ami-
nobenzophenone (7) were coupled via a standard amidation reac-
tion to give 8 in reasonable yield. Intermediate 8 was then treated
with ammonium acetate under cyclization/dehydration conditions
to cleanly generate imidazole 9. The Cbz protecting group of 9 was
a
Assays performed at Absorption Systems following their standard protocols,
HLM = human liver microsomes.13
Inhibition by 5l of [3H]-astemizole binding to the hERG-encoded K+ channel.14
b
c
P450 (isoform 3A4) inhibition in HLM by 5l.15

4872
H. J. Breslin et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4869–4872
H2N
HCl
O
Me
O
Me
O
O
N
7
·
O
N
H
O
N
H
O
N-CBZ-L-Ala
HN
N
H
EDAC, HOBt
NMM, CH2Cl2
NH4OAc, AcOH
reflux, 7h, 95%
9
6
8
83%
Me
Me
OH
O
O
O
MeO
Me
Me
H2N
HN
Me
O
O
O
O
MeO
O
t-Bu
H
11
O
Me
13
N
N
H2N
N
H
N
EDAC, HOBt
DMF, 56%
NaBH4, MeOH
70%
N
H2, 10%Pd/C
MeOH, 100%
H
H
12
10
Me
Me
Me
O
O
O
O
O
O
EtOAc, THF
HCl
71% over
two steps
HO
MeO
HO
LiOH, H2O
THF, MeOH
Me
Me
Me
N
N
N
Me
N
Me
N
Me
N
N
H
N
H
N
H
O
O
O
O
O
HN
Me
H2N
H2N
HN
NH2
Me
H2N
Me
·
2HCl
O
O
O
O
O
15
5l
14
t-Bu
t-Bu
Scheme 1.
2. Fries, D. S. In Principles of Medicinal Chemistry; Foye, W. O., Lemke, T. L.,
Williams, D. A., Eds., 4th ed.; Willimams and Wilkins: Baltimore, MD, 1995; pp
247–269.
3. Breslin, H. J.; Miskowski, T. A.; Rafferty, H. M.; Coutinho, S. V.; Palmer, J. M.;
Wallace, N. H.; Schneider, C. R.; Kimball, E. S.; Ahang, S.-P.; Li, J.; Colburn, R. W.;
Stone, D. J.; Martinez, R. P.; He, W. J. Med. Chem. 2004, 47, 5009.
4. Breslin, H. J.; Cai, C.; Miskowski, T. A.; Coutinho, S. V.; Zhang, S.-P.; Hornby, P.;
He, W. Bioorg. Med. Chem. Lett. 2006, 16, 2505.
5. Bryant, S. D.; Jinsmaa, Y.; Salvadori, S.; Okada, Y.; Lazarus, L. H. Biopolymers
2003, 71, 86.
6. Purchased CHO-hg cell membrane was used for the GTPcS OR functional
assays. Complete experimental details for these assays are described in Ref. 3.
7. Complete experimental details for hamster vas deferens tissue assay:
McKnight, A. T.; Corbett, A. D.; Marcoli, M.; Kosterlitz, H. W.
Neuropharmacology 1985, 24, 1011.
8. (a) Schiller, P. W. Life Sci. 2010, 86, 598; (b) Schiller, P. W.; Weltrowska, G.;
Berezowska, I.; Nguyen, T. M.-D.; Wilkes, B. C.; Lemieux, C.; Chung, N. N.
Biopolymers 2000, 51, 411.
readily removed by Pd catalyzed hydrogenolysis to quantitatively
give primary amine 10. A sodium borohydride reductive amination
of benzaldehyde 11 with 10 gave secondary amine 12, which was
subsequently coupled with acid 1316 providing a 56% isolated re-
turn of 14. The ester moiety of 14 was easily hydrolyzed with lith-
ium hydroxide to give acid 15, which was then treated with
hydrogen chloride to remove the Boc protecting group with simul-
taneous precipitation of desired final product 5l straightaway as its
dihydrochloride salt.
In summary, appropriately substituted acyclic analogs 5, de-
rived by opening the heterocyclic cores of parent structures 4
(Fig. 3), had favorable d and
l OR binding affinities and were con-
sistent with the binding affinities previously reported for com-
pounds 4. However, dependent on the R0 substituents of 5, the
compounds varied as d OR agonists (E.g., 5h, 5i, 5j; Table 1) or d
OR antagonists (E.g., 5k, 5l). From the set of acyclic analogs 5 eval-
uated, compound 5l was identified as a compound of particular
9. Dietis, N.; Guerrini, R.; Calo, G.; Salvadori, S.; Rowbotham, D. J.; Lambert, D. G.
Br. J. Anaesth. 2009, 103, 38.
10. Complete experimental details for
described in Ref. 11.
j OR binding and functional assays are
11. Wade, P.R.; Palmer, J.M.; McKenney, S.; Kenigs, V.; Chevalier, K.; Moore, B.A.;
Mabus, J.R.; Saunders, P.; Wallace, N.H.; Schneider, C.R.; Kimball, E.; Breslin,
H.J.; He, W.; Hornby, P.J. Br. J. Pharmacol., in press.
12. Clinical Study: NCT01130272.
13. http://www.absorption.com/site/.
14. Cardiovascular long QT syndrome was assessed by 5l inhibition of [3H]-
astemizole binding to the hERG-encoded channel; internal assay as described
by Chiu, P. J.; Marcoe, K. F.; Bounds, S. E.; Lin, C. H.; Feng, J. J.; Lin, A.; Cheng, F.
C.; Crumb, W. J.; Mitchell, R. J. Pharmacol. Sci. 2004, 95, 311.
15. Experimental in vitro inhibition (IC50) by 5l of cytochrome P450 3A4 using 6b-
hydroxylation of substrate testosterone in HLM; based on Wang, R. W.;
Newton, D. J.; Liu, N.; Atkins, W. M.; Lu, A. Y. Drug Metab. Dispos. 2000, 28, 360.
16. Preparation of 13 previously reported: Cai, C.; Breslin, H. J.; He, W. Tetrahedron
2005, 61, 6836.
interest, demonstrating potent dual d OR antagonist/
l OR agonist
activities as well as possessing a favorable overall biological profile.
Compound 5l is currently advancing through clinical studies, re-
cently completing a large Phase II Proof of Concept study success-
fully where it showed statistically significant efficacy for IBS-d
patients.
References and notes
1. Aldrich, J. V. Burger’s Medicinal Chemistry and Drug Discovery In Therapeutic
Agents; John Wiley Son, 1996; Vol. 3, pp 321–441. 5th ed.