Highly potent and selective zwitterionic agonists of the δ-opioid receptor. Part 1

Article abstract of DOI:10.1016/j.bmcl.2005.10.102

A series of zwitterionic δ-opioid agonists, with targeted physicochemistry, as a strategy to limit potential for CNS exposure, were prepared. These agents were found to possess exquisite potency and selectivity over mu and κ-opiate activity. Furthermore,

Full text of DOI:10.1016/j.bmcl.2005.10.102

Bioorganic & Medicinal Chemistry Letters 16 (2006) 905–910
Highly potent and selective zwitterionic agonists of the
d-opioid receptor. Part 1
Donald S. Middleton,a,* Graham N. Maw,a Clare Challenger,a Alan Jessiman,a
Patrick S. Johnson,a William A. Million,a Carly L. Nichols,a
Jenny A. Pricea and Michael Trevethickb
aDepartment of Discovery Chemistry, Pfizer Global Research and Development, Sandwich, Kent, CT13 9NJ, UK
bDepartment of Discovery Biology, Pfizer Global Research and Development, Sandwich, Kent, CT13 9NJ, UK
Received 3 October 2005; revised 31 October 2005; accepted 31 October 2005
Available online 15 November 2005
Abstract—A series of zwitterionic d-opioid agonists, with targeted physicochemistry, as a strategy to limit potential for CNS expo-
sure, were prepared. These agents were found to possess exquisite potency and selectivity over mu and j-opiate activity. Further-
more, analogue 3a was found to display restricted CNS exposure, as evidenced by its inactivity in a rodent hyperlocomotion assay of
central opiate activity. Dog pharmacokinetic studies on 3a indicated encouraging oral bioavailability.
Ó 2005 Elsevier Ltd. All rights reserved.
p
Irritable bowel syndrome (IBS) is a highly common
functional bowel disorder, which significantly impacts
on the quality of life for those who suffer its effects.
While several mechanistic approaches to the treatment
of IBS have been identified, their efficacy is often limited
and new therapies are still required to treat this debili-
tating condition.1
O
( )
z
Et2N
N
HO2C
A
X
( )
X
m
N
N
N
N
(2) X = -OMe, -OH, -F
p = 1, m = 1. Z = 2H or O
(3) X = -OMe, -OH, -F
p = 2, m = 1. Z = 2H or O
A = alkyl or phenyl
(1a) X = -OMe. SNC-80
In this first communication, we wish to report our preli-
minary findings targeting selective agonism of peripheral
delta (d)-opioid receptors, as a strategy to both treat IBS
and, by excluding these agents from the CNS, minimise
the potential for centrally mediated adverse events asso-
ciated with compounds possessing opiate activity. These
studies, to the best of our knowledge, have resulted in
the identification of the most potent and selective d-opi-
oid agonists2 yet reported. Furthermore, the physico-
chemical design strategy outlined below has also
successfully delivered agents which are CNS excluded,
yet retain the potential to be orally bioavailable, as as-
sessed by dog pharmacokinetic studies.
δ (pIC50) =8.7, µ (pIC50) = <5
(1b) X = -OH. BW373U86
δ (pIC50) = 9.2
Figure 1.
1990s. The optimisation of these intriguing non-peptidic
compounds into drugs, however, remains a significant
and enduring challenge.3
Central to our own strategy was a plan to exploit the
attractive primary pharmacology observed in agonists
such as 1a and 1b and to design compounds which pos-
sessed physicochemical properties consistent with the
potential for good oral absorption, but which were
likely to possess more limited potential for CNS
penetration. While there have been many reports in
the literature defining properties which afford good oral
absorption,4 properties, which define good passive
Small, non-peptidic agonists of the d-opioid receptor
1a3a and 1b,3b Figure 1, have been known since the early
Keywords: d-Opioid; Agonist.
*
Corresponding author. Tel.: +1304 648393; fax: +44 1304 651987;
e-mail: don.s.middleton@pfizer.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.10.102

906
D. S. Middleton et al. / Bioorg. Med. Chem. Lett. 16 (2006) 905–910
diffusion across the blood–brain barrier (BBB), appear
less clearly characterised, though they are generally
regarded to be more stringent.5 We recognised at the
outset, however, that our target permeability profile of
good oral exposure and CNS exclusion was extremely
challenging and that any physicochemical window to
achieve this was likely to be narrow at best.
on 3a have shown this series to possess encouraging oral
bio availability (ꢀ20%).
All compounds 2 and 3 were prepared by the general
method shown in Scheme 1. Reaction of (À)-(2R,5S)-
1-benzyl-2,5-dimethylpiperazine 48 and the appropriate
aldehyde 5a or 5b in the presence of benzotriazole in tol-
uene, under Dean–Stark conditions, produced the inter-
mediate imine. These underwent diastereoselective aryl
Grignard addition using the method of Katritsky,9 to
give intermediates 6a or 6b as the major isomeric prod-
ucts isolated. Conversion to the target compounds was
then accomplished using the general methods previously
described.9 BOC-deprotection using HCl(g) in CH2Cl2,
(to give 7a and 7b), followed by acylation or alkylation
of the secondary resulting amine, with the appropriate
ester-containing reagent and subsequent ester hydrolysis
using 2 N NaOH in dioxan/H2O, gave the respective
acids. Phenolic compounds 2b, 2d, 2g, 2h, 2i, 2j, 3b,
3d, 3g and 3i were then accessed, from their anisole pre-
cursors, using BBr3 in CH2C12 at À78 °C.
Our target was to identify an orally bioavailable, potent d-
opioid receptor agonist (pIC50 ꢀ9) with high selectivity
over mu (l) and kappa (j) sub-types (P500-fold) which
showed minimal CNS exposure, as measured in a mouse
hyperlocomotion model of central opiate activity.
Our design strategy (2 and 3), Figure 1, focused on zwit-
terionic compounds, which were precedented to exhibit
limited CNS penetration.6 Within this class, we then tar-
geted compounds with a moderately positive log D (ꢀ2)
and relatively high MWt (500–550), which we felt
occupied the most likely physicochemical window for
achieving reasonable oral (GI) absorption with minimal
BBB penetration potential.5 As an in vitro measure of
likely permeability, we sought compounds with moder-
ate flux in a Caco-2 screen7 (apical-basolateral (A-B)
rate ꢀ5%/h), indicative in our assay of permeability
likely to be consistent with GI absorption, though not
considered high enough to indicate significant CNS
penetration.
Aldehyde 5a was prepared as shown in Scheme 2. BOC-
protection of dipropargylamine gave 8, which was treat-
ed with propargyl alcohol and Wilkinsonꢀs catalyst in
EtOH at 0 °C, to give alcohol 9 in 62% yield over 2
steps. Oxidation of the alcohol to the aldehyde 5a was
then achieved in 51% yield using palladium acetate in
DMF.
Execution of this strategy has led to the discovery of
exquisitely potent and selective zwitterionic d-opioid
agonists which were found in a rodent hyperlocomotion
model of CNS activity, to be highly peripherally selec-
tive. Furthermore, initial dog pharmacokinetic studies
Aldehyde 5b was prepared as shown in Scheme 3. Cyc-
lisation of imino-acetal 10 using BF3ÆOEt2 gave isoquin-
oline 11 in good yield (72%). Demethylation using 48%
HBr at reflux, followed by selective reduction of the
p
( )
p
( )
H
N
(a), (b)
boc
N
boc
N
+
X
N
( )
( )
m
N
N
m
O
Ph
Ph
(4)
(5a) p = 1, m = 1
(5b) p = 2, m = 1.
(6a) p = 1, m = 1
(6b) p = 2, m = 1.
(c)
p
( )
p
( )
z
(d), (e), (f)
N
H
N
HO2C
A
X
X
( )
( )
m
m
N
N
N
N
Ph
Ph
(7a) p = 1, m = 1
(7b) p = 2, m = 1.
(2) and (3)
Scheme 1. Synthesis of 5-[(R)-[(2S,5R)-4-benzyl-2,5-dimethylpiperazinyl](aryl)methyl]isoindoline (2) and 7-(R)-[(2S,5R)-4-benzyl-2,5-dimethylpip-
erazinyl](aryl)methyl]-1,2,3,4-tetrahydroisoquinolines 3. Reagents and conditions: (a) benzotriazole, toluene, Dean–Stark conditions, 3 h; (b) add to
aryl magnesium bromide, THF, À20 °C–rt, 1 h; (c) HCl(g), CH2Cl2, 0 °C; (d) acylation or alkylation; (e) ester hydrolysis; (f) BBr3, CH2Cl2, À78 °C
(for 2b, 2d, 2g, 2h, 2i, 2j, 3b, 3d, 3g and 3i).

D. S. Middleton et al. / Bioorg. Med. Chem. Lett. 16 (2006) 905–910
907
O
O
(c)
(b)
(a)
N
N
N
N
H
O
O
O
O
O
OH
(8)
(9)
(5a)
Scheme 2. Synthesis of tert-butyl 5-formyl-1,3-dihydro-2H-isoindole-2-carboxylate 5a. Reagents and conditions: (a) Di-tert-butyl-dicarbonate, NEt3,
CH2Cl2, 0 °C–rt; (b) propargyl alcohol, Wilkinsonꢀs catalyst, EtOH, 0 °C–rt; (62% over two steps); (c) Pd(OAc)2, 4-iodo-toluene, tetraethyl
ammonium chloride, NaHCO3, DMF, 100 °C, 2 h (51%).
OEt
(b), (c)
(a)
OEt
NH
N
HO
MeO
N
MeO
(11)
(12)
(d), (e)
(10)
O
(f)
(g), (h)
N
O
N
O
N
O
F3C S O
O
EtO2C
O
O
O
O
(5b)
(14)
(13)
Scheme 3. Synthesis of tert-butyl 7-formyl-3,4-dihydro-2(1H)-isoquinolinecarboxylate (5b). Reagents and conditions: (a) BF3ÆOEt2, (CF3CO)2O,
0 °C–rt (72%); (b) 48% HBr, reflux (64%); (c) glacial acetic acid, PtO2, H2, 40 p.s.i., 16 h; (d) di-tert butyl carbonate, H2O/THF (2:5), NEt3, rt, 16 h
(91%); (e) N-phenylbis(trifluoromethanesulfonimide), CH2Cl2, rt, 48 h (76%); (f) ethyl acrylate, Pd(OAc)2, (o-Tol)3P, NEt3, CH3CN, reflux, 16 h
(44%); (g) OsO4, N-methyl morpholine N-oxide, acetone/H2O (5:1), 16 h, rt (53%); (h) sodium periodate, Et2O/H2O (4:3), rt (98%).
isoquinoline heteroaromatic ring using H2 and PtO2 as
catalyst, furnished 12. N-BOC-protection and conver-
sion of the phenolic hydroxyl to the triflate then gave
13, which was readily converted to the a,b-unsaturated
ester 14 using ethyl acrylate and Pd(OAc)2. 14 was then
converted to the target aldehyde 5b in two steps using
OsO4 and periodate cleavage of the resulting diol.
less lipophilic (log D 2.1). Encouragingly, acetic acid
analogue 16, which possessed a still lower log D (1.1)
than 15, showed reduced permeability, Caco-2 (6%/h),
and this was also reflected in a 10-fold increase in the
dose required for observation of threshold effects in
the CNS hyperlocomotion model. That differentiated
profiles could be observed in the hyperlocomotion
model in this low MWt series was highly encouraging.
From an analogue design perspective, however, work-
ing within such a narrow MWt window, unfortunately,
proved impractical. Interestingly, we noted that biphe-
nyl analogue 17 possessed a similar moderate Caco-2
profile (A-B, 3%/h) to 16, (A-B, 6%/h), despite possess-
ing similar log D to a highly fluxed propionic acid ana-
logue 15, (A-B, 23%/h). This suggested to us that
targeting higher MWt compounds (500–550) with mod-
erate log D (ꢀ2) might allow us to access compounds
with the permeability profile we sought, but in a series
with greater synthetic scope for analogue design, to as-
sist optimisation of the overall profile.
Functional d-opioid activity was initially demonstrated
using a mouse isolated electrically stimulated vas defer-
ens assay.10 Functional selectivity at l and j-opioid
receptors was determined using the electrically stimu-
lated guinea pig isolated myenteric plexus preparation.11
Activity at the human d-opioid receptor expressed in
CHO cells was determined either by radioligand binding
or functionally using cytosensor technology developed
at Pfizer. The mouse hyperlocomotion assay was per-
formed by the method reported previously.12
Initial evaluation of a range of zwitterionic com-
pounds8 in the mouse hyperlocomotion model, Table
1, in an attempt to define the property space which
could confer limited CNS exposure, yielded some
intriguing SAR. Lipophilic amine, SNC-80 (1a), dem-
onstrated pronounced CNS effects (ED200 400 lg/kg),
indicative of its ability to rapidly penetrate the BBB
and consistent with its moderate MWt (449), low H-
bond donor count and relatively high log D (3.3). Sim-
ple propionic acid analogue 15 also showed effects at
low dose, with threshold effects seen at 100 lg/kg, con-
sistent with its also very high Caco-2 flux permeability
(A-B, 23%/h), comparable to SNC-80, despite being
Our initial work in this area13 had indicated that while
encouraging, variable, permeability could be observed
in such zwitterionic compounds, no clear SAR could
be discerned. The design of compounds of types 2 and
3 was, therefore, also driven by a pragmatic desire to
work in a readily synthetically accessible template,
allowing easy and sequential modification of the acid
moiety as a mechanism to tune MWt and log D to
quickly allow systematic SAR development. The sec-
ondary amine in both the isoindoline or tetrahydroiso-
quinoline moieties provided an ideal location to effect

908
D. S. Middleton et al. / Bioorg. Med. Chem. Lett. 16 (2006) 905–910
Table 1. Functional agonist activity against d (mouse vas deferens) and mouse hyperlocomotion data for zwitterionic d-opioid agonists SNC-80, 15,
16 and 17
R
OH
N
N
Compounda
R
d-pIC50
MWt
Log D
Caco-2 (%/h)
Mouse locomotion (iv)c
b
SNC-80
15
16
—
–(CH2)2–CO2H
–CH2–CO2H
8.7
9.9
9.8
449
459
445
3.3
2.1
1.1
23 ER = 1
23 ER = 1
6 ER = 2
400 lg/kg
100 lg/kg
1000 lg/kg
CO2H
17
10.7
507
2.3
3 ER > 5
NT
.
NT, not tested. ER, efflux ratio (ratio of apical-basolateral (A-B):basolateral-apical (B-A)).
a All compounds are full agonists and single homochiral diastereomers.
b Mouse vas deferens.
c Threshold dose at which effects were observed.
these changes. Initial analogues prepared are shown in
Table 2.
6.2, respectively). Conversion of each of these anisoles
to their respective phenols, 2b (d, pIC50 9.3) and 2d (d,
pIC50 9.0) however, significantly increased potency in
each case. Attempts to replace the phenolic hydroxy
group in 2d, with either a meta-fluoro 2e or hydrogen
Isoindolines 2a and 2c were found to possess moderate
activity against the d-opioid receptor (pIC50 7.6 and
Table 2. Functional agonist activity against d (mouse vas deferens) and l and j (guinea pig myenteric plexus) receptors for 5-[(R)-[(2S,5R)-4-benzyl-
2,5-dimethylpiperazinyl](aryl)methyl] isoindoline 2a–j
R N
X
N
N
b
c
c
Compounda
R
X
d-Agonist activity (pIC50
)
l-Agonist activity (pIC50
)
j-Agonist activity (pIC50)
2a
2b
2c
2d
2e
2f
–CH2–CO2H
–CH2–CO2H
–OMe
–OH
–OMe
–OH
–F
7.6
9.3
6.2
9.0
7.6
7.2
NT
6.9
<5
NT
5.9
<5
–(CH2)2–CO2H
–(CH2)2–CO2H
–(CH2)2–CO2H
–(CH2)2–CO2H
6.2
NT
NT
5.1
NT
NT
–H
CO2H
O
2g
–OH
–OH
9.5
7.3
5.4
CO2H
2h
8.5
NT
NT
HO2C
O
2i
2j
–OH
–OH
8.8
8.1
7.0
5.5
5.9
6.4
O
HO2C
a All compounds are full agonists and single homochiral diastereomers.
b Mouse vas deferens.
c Guinea pig myenteric plexus.

D. S. Middleton et al. / Bioorg. Med. Chem. Lett. 16 (2006) 905–910
909
2f, resulted in a significant drop in potency. A series of
acid-containing benzamide analogues 2g, 2i and 2j were
also prepared to explore the role of the acid orientation
in conferring potency in this series, with the ortho-acid
analogue 2g found to be highly potent against the d-opi-
oid receptor (pIC50 9.5) and >100-fold selective over l
and j sub-types.
While tetrahydroisoquinoline series 3 afforded excellent
pharmacological profiles, our goal of combining this
pharmacology with moderate Caco-2 permeability in a
CNS-excluded agent was found to be extremely chal-
lenging. For example, we discovered that all phenolic
analogues screened in this series 3b, 3d, 3g and 3i, while
extremely potent and possessing our target moderate,
positive, lipophilicity, showed very poor permeability
characteristics (Caco-2 (A-B) <1%/h), effectively elimi-
nating them from consideration for further progression.
Indeed, within this series, we discovered that permeabil-
ity hinged on the presence or absence of this single, phe-
nolic, H-bond donor (e.g., 3d (OH, <1%/h) vs 3c (OMe,
11%/h)). Attempts to temper the deleterious effects of
the phenolic OH, by replacing it with a fluoro group,
3e unfortunately gave compounds with too rapid perme-
ability (Caco-2 >30%/h).
The ortho-acid benzyl analogue 2h was found be
approximately 10-fold less potent than its amide ana-
logue 2g.
Work undertaken in the homologous tetrahydroisoquin-
oline series 3 showed that even greater potency and
selectivity could be achieved, Table 3. For example, ace-
tic acid analogue 3a (pIC50 8.7) was found to be approx-
imately 10-fold more potent than the equivalent
isoindoline 2a (pIC50 7.6), with excellent human transla-
tional pharmacology (human d binding pIC50 8.3). As in
the isoindoline series, conversion of the anisole in 3a to
the phenol was found to result in a significant increase in
potency (d, pIC50 10.4), with excellent selectivity over l
(pIC50 6.5) and j (pIC50 6.0) receptors. Propionic acid
analogue 3d (d, pIC50 11.3) was found to be approxi-
mately 10-fold more potent than its lower homologue
3b, again demonstrating excellent selectivity over l
and j. Attempts to replace the phenolic hydroxy moiety
found in 3d with a meta-fluoro group, 3e, resulted in a
precipitous drop in activity (pIC50 7.0). Interestingly
and to our surprise, amides 3g (d, pIC50 12.0) and 3i
(pIC50 11.8) demonstrated exceptional levels of potency
and selectivity.
In the case of exquisitely potent phenol 3g (d, pIC50
ꢀ12, Caco-2 A-B <1%/h), we hoped that replacing the
phenolic –OH with –OMe 3f or –F 3h, while predicted
to result in a drop in potency, would still yield potent
compounds and with reasonable flux. Disappointingly,
and somewhat to our surprise, 3h and 3f both showed
a massive loss in potency relative to 3g, which precluded
their progression. A more equitable balance between
target potency and permeability was, however, found
in the anisole analogues 3a and 3c.
Overall, 3a was selected for further evaluation, as it pos-
sessed a profile closest to our target pharmacology (d,
pIC50 8.7, >500-fold selective over l and j) and moder-
Table 3. Functional agonist activity against d (mouse vas deferens), l and j (guinea pig myenteric plexus) receptors and Caco-2 permeability data for
7-(R)-[(2S,5R)-4-benzyl-2,5-dimethylpiperazinyl](aryl)methyl]-l,2,3,4-tetrahydroisoquinolines 3a–3i
N
R
X
N
N
Compounda
R
X
d-Agonist l-Agonist j-Agonist MWt Log D
activity activity activity (pH 7.4) (%/h)
Caco-2 (A-B) Mouse hyperlocomotion
(iv, lg/kg)
b
c
c
(pIC50
8.7
)
(pIC50
)
(pIC50)
SNC-80 1a
—
—
<5
5.3
6.5
<5
6
5.9
5.9
6.0
6
449
514
500
528
514
516
542
528
530
514
3.3e
2.5
2.1
2.3
2.1
2.1e
1.3
0.4
1.0e
1.6
23 ER = 1
5 ER = 4
<1
400 (ED200)
No effect at 10,000
3a
3b
3c
3d
3e
3f
–CH2–CO2H
–CH2–CO2H
–(CH2)2–CO2H
–(CH2)2–CO2H
–(CH2)2–CO2H
–OMe
–OH
–OMe
–OH
–F
8.7d
10.4
7.9
NT
NT
NT
NT
NT
NT
NT
NT
11 ER = 1.6
<1
11.3
7.0
5.3
NT
NT
6.4
NT
6.4
NT
NT
6.7
NT
6.7
>30 ER = 1
NT
<1
–CO–CH2–CO2H –OMe
–CO–CH2–CO2H –OH
–CO–CH2–CO2H –F
7.8
3g
3h
3i
ꢀ12
7.7
11.8
14 ER = 0.7
<1
–CO–CO2H
–OH
NT, not tested. ER, efflux ratio (ratio of apical-basolateral (A-B):basolateral-apical (B-A)).
a All compounds are full agonists and single homochiral diastereomers.
b Mouse vas deferens.
c Guinea pig myenteric plexus.
d (human) d-opioid binding (pIC50 = 8.3).
e Calculated log D (pH 7.4).

910
D. S. Middleton et al. / Bioorg. Med. Chem. Lett. 16 (2006) 905–910
ate Caco-2 flux (A-B, 5%/h). Evaluation of this analogue
in the mouse hyperlocomotion model of CNS exposure
showed the compound to be very poorly centrally pene-
trant with no effects seen at 10,000 lg/kg, relative to (1a,
ED200 400 lg/kg) and no convulsant or seizure behav-
iour observed.
Sciences Department for analytical and spectroscopic
data.
References and notes
1. Farhadi, A.; Bruninga, K.; Fields, J.; Keshavarzian, A.
Exp. Opin. Investig. Drugs 2001, 10, 1211.
2. Maw, G. N.; Middleton, D. S. Ann. Rep. Med. Chem.
2002, 37, 159.
Based on this overall profile, compound 3a was pro-
gressed to dog pharmacokinetic studies to assess
whether the physicochemical design parameters, which
had successfully prevented CNS exposure, were compat-
ible with achieving oral bioavailability. Encouragingly,
3a demonstrated moderate clearance (17 ml/min/kg)
and Vd (5.4 L/kg), with 20% oral bioavailability, indica-
tive of significant, though incomplete, oral absorption
(ca. 30–50%) in this study.
3. (a) Calderon, S. N.; Rothman, R. B.; Porreca, F.; Flippen-
Anderson, J. L.; McNutt, R. W.; Xu, H.; Smith, L. E.;
Bilsky, E. J.; Davis, P.; Rice, K. C. J. Med. Chem. 1994,
37, 2125; (b) Chang, K. J.; Rigdon, G. C.; Howard, J. L.;
McNutt, R. W. J. Pharm. Exp. Ther. 1993, 267, 852.
4. Lipinski, C. A. Drug Discovery Today 2004, 1, 337.
5. Many methods have been evaluated to establish physico-
chemical trends for good passive CNS penetration These
include targeting low MWt (<450), low polar surface area
In summary, a series of zwitterionic d-opioid agonists
were designed with relatively high MWt (500–550) and
moderately positive log D (ꢀ2) as strategy to minimise
potential for CNS penetration, yet retain potential for
oral (GI) absorption. These compounds were found to
possess excellent and full functional agonist potency
againstthed-opioidreceptorandveryhighselectivityover
l and j-opioid sub-types. In addition, analogue 3a was
found to possess moderate permeability in the Caco-2 as-
say and poor CNS penetration properties, as evidenced by
its inactivity in a mouse hyperlocomotion model of opiate
activity up to 10 mg/kg. Dog pharmacokinetic studies on
3a showed it to be orally bioavailable (20%), indicating
that despite being non-BBB penetrant, it was still signifi-
cantly orally absorbed in the GI tract (30–50%).
2
˚
(<80 A ), low ionisation potential, low H-bond count, low
affinity for P-gp or efflux mechanism and high lipophilic-
ity, however, given the complex multi-factorial nature of
CNS permeability, no strongly predictive correlation has
been reported. See: (a) Mahar Doan, K. M.; Humphreys,
J. E.; Webster, L. O.; Wring, S. A.; Shampine, L. J.;
Serabjit-Singh, C. J.; Adkinson, K. D.; Polli, J. W. J.
Pharmacol. Exp. Ther. 2002, 303, 1039; (b) van de
Waterbeemd, H.; Smith, D. A.; Beaumont, K.; Walker,
D. J. Med. Chem. 2001, 44, 1313, and references cited
therein.
6. For an example of an orally bioavailable zwitterionic
compound that does not significantly access the CNS, see:
Pagliara, A.; Testa, B.; Carrupt, P.-A.; Jolliet, P.; Morin,
C.; Morin, D.; Urien, S.; Tillement, J.-P.; Rihoux, J.-P. J.
Med. Chem. 1998, 41, 853. The basis for such observations
may be rationalised, in part, by the enthalpic cost of
desolvating a charged residue to enable partitioning into
lipid.
Further work, describing the successful optimisation of
this approach as a strategy for designing orally bioavail-
able, CNS excluded, compounds, will be the subject of a
future communication from these laboratories.
7. Artursson, P.; Palm, K.; Luthman, K. Adv. Drug Delivery
Rev. 1996, 22, 67.
8. Compounds as delta opioid agonists. Maw, G. N.;
Middleton, D. S. U.S. Patent 6,200,978.
9. Bishop, M. J.; McNutt, R. W. Bioorg. Med. Chem. Lett.
1995, 5, 1311, and references cited therein.
Acknowledgments
10. Hughes, J.; Kosterlitz, H. W.; Leslie, F. M. Br. J.
Pharmacol. 1975, 53, 371.
The authors thank Dr. Jem Gale, Dr. Aileen McHarg
for helpful discussions. We also acknowledge the able
assistance of Mrs. Pam Mclntyre, Mr Simon Tickner
and Mr Tony Hawcock in providing screening data,
Dr. Joanne Bennett for permeability studies and
Dr. Susan Cole for pharmacokinetic studies. We also
thank the staff of the Structural and Separation
11. Pencheva, N.; Pospisek, J.; Hauzerova, L.; Barth, T.;
Milanov, P. Br. J. Pharmacol. 1999, 128, 569.
12. Brocco, M.; Dekeyne, A.; Veiga, S.; Girardon, S.; Millan,
M. Pharmacol. Biochem. Behav. 2002, 71, 667.
13. Unpublished results. These data will be reported
elsewhere.