Metabolically labile cannabinoid esters: A 'soft drug' approach for the development of cannabinoid-based therapeutic drugs

Article abstract of DOI:10.1016/j.bmcl.2007.06.055

Biphenylic ester derivatives, designed by using a 'soft-drug' approach, proved to possess good binding properties toward cannabinoid CB₁ and CB₂ receptors and, at the same time, their metabolically labile ester portion would promote a rapid systemic inactivation. This may constitute a possible solution to the psychotropic side effects encountered when cannabinoids are therapeutically employed as local analgesic or antiglaucoma agents.

Full text of DOI:10.1016/j.bmcl.2007.06.055

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4878–4881
Metabolically labile cannabinoid esters: A ‘soft drug’ approach
for the development of cannabinoid-based therapeutic drugs
F. Minutolo,a,* M. G. Cascio,b I. Carboni,a T. Bisogno,b G. Prota,a S. Bertini,a
M. Digiacomo,a M. Bifulco,c V. Di Marzob and M. Macchiaa
a
`
Dipartimento di Scienze Farmaceutiche, Universita di Pisa, Via Bonanno 6, 56126 Pisa, Italy
bIstituto di Chimica Biomolecolare, Consiglio Nazionale delle Ricerche, Via Campi Flegrei 34, 80078 Pozzuoli (NA), Italy
c
`
Dipartimento di Scienze Farmaceutiche, Universita di Salerno, Via Ponte Don Melillo, 84084 Fisciano (SA), Italy
Received 16 May 2007; revised 11 June 2007; accepted 12 June 2007
Available online 20 June 2007
Abstract—Biphenylic ester derivatives, designed by using a ‘soft-drug’ approach, proved to possess good binding properties toward
cannabinoid CB1 and CB2 receptors and, at the same time, their metabolically labile ester portion would promote a rapid systemic
inactivation. This may constitute a possible solution to the psychotropic side effects encountered when cannabinoids are therapeu-
tically employed as local analgesic or antiglaucoma agents.
ꢀ 2007 Elsevier Ltd. All rights reserved.
The medical applications of Cannabis sativa derivatives
have been known for many centuries, but their approved
use was commonly discontinued in the early 1900s due
to the well-known psychotropic side effects. Nowadays,
at least two receptor subtypes, CB1 and CB2, have been
recognized within the mammalian cannabinergic system.
CB1 is mainly found in central and peripheral neurons,
whereas CB2 exists primarily in immune cells.1 It has
been shown that stimulation not only of CB1, but also
of CB2 receptors causes analgesic effects.2 Moreover,
stimulation of the ocular CB1 receptor produces a reduc-
tion of intraocular pressure (IOP), indicating a route by
which new anti-glaucoma drugs may be developed.3
Therefore, molecules able to stimulate CB-receptors
have the potential to be used locally as analgesic or
anti-glaucoma agents. The main problem in the thera-
peutic use of CB-agonists is the difficulty in limiting
the occurrence of central psychotropic effects. A possible
solution is given by the ‘soft drug approach’.4
has led to the development of safer marketed drugs since
they are endowed with the appropriate local activity,
whereas they are devoid of systemic side effects.5 The
soft drug approach would, therefore, be particularly
suitable for the development of cannabinoids to be used
as locally active analgesic or anti-glaucoma agents, with-
out causing the unwanted central effects. These consider-
ations have already inspired some excellent examples of
‘soft-cannabinoids’ possessing CB1-mediated IOP lower-
ing properties, to be developed as safe anti-glaucoma
agents.6
We focused our attention on structurally simple biphe-
nylic derivatives that had already been reported to be
good CB-ligands7a,b and potent CB-agonists.7c In par-
ticular, compound 1 (Fig. 1) showed Kis of 79 and
2 nM, on CB1 and CB2, respectively.7b
Soft drugs are bioactive molecules deliberately designed
to go through a rapid, efficient, and predictable meta-
bolic inactivation after having exerted their effects at
the site of action.4 In many cases the use of soft drugs
(H,Me)
Me
HO
(H,Me)
O
O
OH
(Me,H)
Me
O
(H,Me)
2-8
Alk
C5H11
O
1
Keywords: Cannabinoids; CB1; CB2; Soft-drugs; Retrometabolic.
*
Figure 1. Structural derivation of metabolically labile esters 2–8 (see
Table 1) from biphenylic CB-ligand 1.
Corresponding author. Tel.: +39 050 2219557; fax: +39 050
2219605; e-mail: minutolo@farm.unipi.it
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.06.055

F. Minutolo et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4878–4881
4879
We were inspired by this class of biphenyl cannabimi-
metic derivatives because their simple molecular struc-
tures guarantee greater synthetic accessibility
compared to classical cannabinoids, and, most impor-
tantly, they generally display high potencies of action,
which have been considered as a useful feature for their
application in the treatment of pain and glaucoma.7c
R
R
O
O
+ Alk-OH
OH
a
O
O
Alk
O
O
O
active molecule
inactive metabolic products
Figure 2. Predicted metabolic hydrolytic inactivation of ester-type soft
cannabinoids.
Therefore, we envisaged the possibility of inserting a
metabolically labile ester group within the linear ali-
phatic side chain of biphenyl derivative 1, in an attempt
to develop soft-cannabinoid analogues (2–8, Table 1).
Compound 2 is the direct ester analogue of parent ali-
phatic CB-ligand 1, obtained by simply replacing the
gem-dimethyl-substituted ethylene portion of 1 with an
ester –C(O)O– group. The linear five-carbon-aliphatic
side chain (n-C5H11) of the ester group in 2 was chosen
in order to maintain an overall seven-carbon-length as
in 1. Moreover, we also wanted to verify the effect on
CB1/CB2 affinity due to small structural modifications
on several molecular portions of the biphenylic core
originally present in 1. Changes in the methyl and/or
free OH aryl substituents, as well as in the length of
the linear aliphatic ester chain, were obtained in the ser-
ies of new compounds 3–8.
OMe
OMe
Br
Br
1) SOCl2
OH
OR4
MeO
MeO
2) R4-OH
THF
O
O
13a: R4
13b: R4
=
=
n
n
-C5H11
-C4H9
12
13c: R4 = CH3
R1
OMe
BBr3
(2 or 6 eq)
R1
ArB(OH)2
Pd(PPh3)4
2,3
5-8
OR4
MeO
CH2Cl2
-78 C to RT
-C5H11
= n-C4H9
O
n
°
4: R1 = CH3; R4
=
aq K3PO4
dioxane
80 °C
14a: R1 = CH3; R4
14b: R1 = CH3; R4 = CH3
14c: R1 = H; R4
-C5H11
=
n
Scheme 1. Synthesis of esters 2–8 from commercially available 12.
These newly designed ester-type soft cannabinoids
should act only locally, as long as they preserve their
structural integrity. However, as soon as they are sys-
tematically absorbed, they are predicted to undergo a
rapid hydrolytic inactivation that leads to inactive car-
boxylic acids and aliphatic alcohols, as the metabolic
products (Fig. 2).
the appropriate alcohol in THF, afforded esters 13a–c.
Subsequent Pd-catalyzed cross-coupling reaction with
the properly substituted arylboronic acid in the presence
of Pd(PPh3)4 as the catalyst yielded biphenyl derivatives
4 and 14a–c. In spite of the fact that the bromo-aryl
group in intermediates 13a–c is rather hindered because
of the simultaneous presence of two ortho-methoxy sub-
stituents, the yields of this step were quite satisfactory,
ranging between 64% and 75%. Final progressive
demethylation steps of the aryl-methoxy groups were
carried out by finely modulating the reaction conditions.
Esters 2–8 were prepared as outlined in Scheme 1 from
commercially available 4-bromo-3,5-dimethoxybenzoic
acid 12.
Treatment of 12 with refluxing thionyl chloride, fol-
lowed by reaction of the resulting acid chloride with
Table 1. Binding affinity of esters 2–8 and carboxylic acids 9–11
R1
R2
O
R1
O
R4
O
R3
2-11
O
Compound
R1
R2
R3
R4
hCB1 Kia (lM)
hCB2 Kia (lM)
2
3
4
5
6
7
8
CH3
CH3
CH3
H
H
H
n-C5H11
n-C5H11
n-C5H11
n-C5H11
n-C4H9
CH3
2.0
1.4
1.1
CH3
CH3
H
H
0.27
>10
4.0
CH3
H
>10
>10
2.6
CH3
CH3
CH3
H
H
3.2
4.5
H
CH3
H
>10
>10
H
CH3
6.4
9
10
11
CH3
CH3
CH3
H
H
H
H
H
>10
>10
>10
>10
>10
>10
CH3
CH3
H
CH3
a Data represent mean values for at least three separate experiments performed in duplicate and are expressed as Ki (lM). hCB1, human CB1 receptor;
hCB2, human CB2 receptor. Standard errors are not shown for the sake of clarity and were never higher than 10% of the means.

4880
F. Minutolo et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4878–4881
In details, when 4 was treated with only 2 equiv of boron
tribromide at ꢀ78 ꢁC, and then the reaction mixture was
slowly allowed to reach rt, mono-demethylated product
38 could be isolated by stopping the reaction immedi-
ately after TLC showed the consumption of starting
material. On the other hand, exhaustive demethylation
of 4 was achieved by using 6 equiv of BBr3 and monitor-
ing by TLC the consumption of both starting material
and partially demethylated intermediate 3, thus obtain-
ing, di-hydroxy-substituted ester 2.9 The same proce-
dures were applied to the mono- and di-demethylation
of methyl ester 14b, which, respectively, afforded com-
pounds 8 and 7. Only completely demethylated products
5 and 6 were instead obtained by BBr3 treatment of pre-
cursors 14c and 14a, respectively.
6 min at 4 ꢁC. The supernatant (750 lg) and compound
3 (100 lg) were incubated in 50 mM Tris–HCl buffer,
pH 7. The organic phase, after extraction of the incuba-
tion mixture with two volumes of CHCl3/CH3OH (2:1
by vol), was lyophilized under vacuum and then frac-
tionated by TLC (20 · 20 cm) on silica gel F254 (Merck)
by using AcOEt/EtPt/AcOH (60:40:1) as the eluting
system.
Compound 2, the ester derivative possessing the closest
structural resemblances to CB-ligand 1, showed interest-
ing affinity for both CB1 and CB2 receptors, with Ki val-
ues of 2.0 and 1.1 lM, respectively. The presence of a
single methoxy-group in the place of one phenol OH,
as in compound 3, even enhanced the binding affinity
of the resulting molecule, especially for the CB2-subtype
(Ki = 0.27 lM). On the contrary, the replacement of
both phenol—with two MeO—groups as in 4 caused a
complete loss of binding affinity for both receptor sub-
types. The importance of the methyl-aryl substituents
(R1) was checked by assaying compound 5, bearing an
unsubstituted phenyl-ring: in this case the resulting mol-
ecule showed no affinity for CB1, whereas it still main-
tained a certain level of affinity (Ki = 4.0 lM) for CB2.
These data show that the presence of two methyl-groups
(R1) in the 30 and 50 position of the phenyl-substituents
is fundamental for CB1-binding and is also beneficial for
an efficient CB2-affinity.
We also synthesized carboxylic acids 9–11, which are the
predicted metabolic products of the newly designed soft-
cannabinoids (Fig. 2), in order to test whether or not they
preserved any (unwanted) affinity for the CB-receptors.
To this purpose, we started their synthesis (Scheme 2)
from methyl ester 14b, prepared as previously shown
in Scheme 1, which was hydrolyzed by treatment with
potassium tert-butoxide in a diethylether/water mixture,
producing dimethylated carboxylic acid 11.10 Also in
this case, it was possible to selectively obtain mono-
and di-hydroxy-substituted products (10 and 9, respec-
tively),11 by tuning the BBr3-promoted demethylation
conditions as seen before.
The effect of the length of the aliphatic ester chain was
also investigated in n-butyl (6) and methyl (7 and 8) es-
ters. When compared to n-pentyl ester 2, the corre-
sponding n-butyl ester 6 showed a nearly unmodified
affinity for CB1, whereas its binding for CB2 was more
negatively affected (Ki = 3.2 lM). Methyl ester 7 showed
a reduction of affinity for both receptor subtypes, which
led to a substantial loss of affinity for CB1, still main-
taining some binding on CB2 (Ki = 4.5 lM). In this lat-
ter case, the transformation of a single phenol OH of 7
into a methoxy-group, as in 8, did not cause the same
binding improvement observed with their longer-chain
analogues (cf. 2 and 3); in fact, the affinity of 8 generally
was even worse than that of 7. Therefore, the most
appropriate length of the linear aliphatic ester side chain
turned out to be that of five carbon atoms (R4 =
n-C5H11), as in 2 and 3. As a matter of fact, an alkyl
chain of this extent, together with the ester linker, covers
an approximately same spatial occupation as the seven-
carbon fully-aliphatic side chain in parent CB-ligand 1.
Compounds 2–11 were tested for their affinity for hu-
man recombinant CB1 and CB2 receptors using previ-
ously described procedures and employing HEK cells
transfected with either the hCB1 or hCB2 receptor and
[3H]-(ꢀ)-cis-3-[2-hydroxy-4-(1,1-dimethylheptyl)
phe-
([3H]CP-
nyl]trans-4-(3-hydroxypropyl)cyclo-hexanol
55,940; Kd = 0.18 and 0.31 nM for CB1 and CB2
receptor, respectively) as the high affinity ligand, as
described by the manufacturer (Perkin-Elmer, Italy).12
Displacement curves were generated by incubating
drugs with [3H]-CP-55,940 (0.14 and 0.084 nM for CB1
and CB2 binding assay, respectively). In all cases, Ki val-
ues were calculated by applying the Cheng–Prusoff
equation to the IC50 values (obtained by GraphPad)
for the displacement of the bound radioligand by
increasing concentrations of the test compounds. The
resulting Ki values are reported in Table 1. In addition,
in order to evaluate the metabolic transformations of
the compound 3, rat liver was homogenated in 50 mM
Tris–HCl buffer, pH 7, and centrifuged at 800g, for
Mostimportantly, carboxylicacids9–11,whichconstitute
the predicted metabolic products of their ester precursors,
were completely devoid of any affinity for either receptor
subtype. This inability to bind the CB-receptors by 9–11
is a fundamental requirement for the ester derivatives to
be considered as possible soft-cannabinoid agents.
Me
OMe
t
-BuOK
Me
11
Et2O/H2O
0 °C to RT
OMe
MeO
14b
O
Finally, preliminary assays on metabolic transforma-
tions of the highest affinity compound 3 performed in
rat liver homogenates confirmed that about 40% of this
molecule is hydrolyzed after only 30 min of incubation
at 37 ꢁC. Although these first in vitro metabolic data
are not sufficient to completely exclude the possible
BBr3
BBr3
(6 eq)
(2 eq)
CH2Cl2
-78 °C to RT
9
10
Scheme 2. Synthesis of carboxylic acids 9–11 from ester 14b.

F. Minutolo et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4878–4881
4881
2. Walker, J. M.; Huang, S. M. Pharmacol. Ther. 2002, 95,
127.
3. Ja¨rvinen, T.; Pate, D. W.; Laine, K. Pharmacol. Ther.
2002, 95, 203.
occurrence of CNS side effects, they still constitute a
starting basis for more in-depth evaluations. Further
studies are currently under way to fully characterize
the functional and metabolic behavior of compounds 2
and 3 and the kinetics of their hydrolytic inactivation.
4. (a) Bodor, N.; Buchwald, P. AAPS J. 2005, 7, E820; (b)
Bodor, N.; Buchwald, P. Med. Res. Rev. 2000, 20, 58.
5. Bodor, N.; Buchwald, P. Mol. Biotechnol. 2004, 26, 123.
6. (a) Buchwald, A.; Derendorf, H.; Ji, F.; Nagaraja, N. Y.;
Wu, W. M.; Bodor, N. Pharmazie 2002, 57, 108; (b)
Buchwald, A.; Browne, C. E.; Wu, W. M.; Ji, F.; Bodor,
N. Pharmazie 2000, 55, 196.
7. (a) Martin, B.R.; Razdan, R.K. U.S. Patent 159,449, 2005;
Chem. Abstr. 2005, 143, 153534; (b) Gareau, Y.; Dufresne,
C.; Gallant, M.; Rochette, C.; Sawyer, N.; Slipetz, D. M.;
Tremblay, N.; Weech, P. K.; Metters, K. M.; Labelle, M.
Bioorg. Med. Chem. Lett. 1996, 6, 189; (c) Makriyannis,
A.; Lai, X.-Z.; Lu, D. PCT Int. Appl. WO 017920, 2004;
Chem. Abstr 2004, 140, 217824.
8. Compound 3 purified by column chromatography (Hex/
EtOAc); white solid, mp 62 ꢁC; 1H NMR (CDCl3,
200 MHz) d 0.90–0.97 (m, 3H), 1.36–1.47 (m, 4H), 1.71–
1.85 (m, 2H), 2.32 (s, 6H), 3.75 (s, 3H), 4.30 (t, 2H,
J = 6.6 Hz), 6.92 (s, 2H), 6.95 (s, 1H), 7.20 (d, 1H,
J = 1.5 Hz), 7.30 (d, 1H, J = 1.5 Hz), 8.11 (br s, 1H); MS-
EI (70 eV) m/e 342 (M+).
The generally lower binding levels found with all the es-
ter analogues of CB-ligand 1 could be largely expected,
since the original bulky and lipophilic gem-dimethyl-
ethyl portion of 1 was replaced by the polar ester group,
needed to provide a metabolically cleavable site. Never-
theless, the real challenge of this work was to keep some
sufficient receptor affinity for compounds that were de-
signed to: (i) be administered locally; (ii) lose their CB-
affinity when they get into the systemic circulation.
Therefore, Ki values reaching low- or sub-micromolar
levels, such as those shown by 2 and 3, may be good en-
ough to achieve these goals. Moreover, this ‘soft-can-
nabinoid’ approach does not even require a particular
receptor-subtype selectivity, since the selective periphe-
ral action, devoid of central effects, would probably be
supported by the programmed rapid metabolic inactiva-
tion of these derivatives.
9. Compound 2 purified by column chromatography (Hex/
EtOAc); white solid, mp 93 ꢁC; 1H NMR (CDCl3,
200 MHz) d 0.89–0.96 (m, 3H), 1.39–1.48 (m, 4H), 1.68–
1.79 (m, 2H), 2.31 (s, 6H), 4.26 (t, 2H, J = 6.5 Hz), 6.95 (s,
1H), 6.99 (s, 2H), 7.18 (s, 2H), 8.08 (br s, 2H); MS-EI
(70 eV) m/e 328 (M+).
Further studies will be devoted to a deeper pharmaco-
logical and metabolic characterization of the highest
affinity esters 2 and 3, as well as to the development of
other biphenylic soft-CB-ligands, by moving the ester
portion along the aliphatic side chain, in an attempt to
verify the possibility of obtaining molecules possessing
even better binding properties and/or more suitable met-
abolic behaviors.
10. Compound 11: white solid, mp 190 ꢁC; 1H NMR (CDCl3,
200 MHz) d 2.36 (s, 6H), 3.81 (s, 6H), 6.94 (s, 1H), 6.98 (s,
2H), 7.41 (s, 2H).
1
11. (a) Compound 9: white solid, mp 206–208 ꢁC; H NMR
(acetone-d6, 200 MHz) d 2.31 (s, 6H), 6.93 (s, 1H), 6.98 (s,
2H), 7.21 (s, 2H), 8.01 (br s, 2H). (b) Compound 10: white
1
solid, mp 198 ꢁC; H NMR (CDCl3, 200 MHz) d 2.30 (s,
References and notes
6H), 3.76 (s, 3H), 6.93 (m, 3H), 7.22 (d, 1H, J = 1.3 Hz),
7.33 (d, 1H, J = 1.3 Hz), 8.09 (br s, 2H).
1. Pertwee, R. G.; Ross, R. A. Prostaglandins, Leukotrienes
Essent. Fatty Acids 2002, 66, 101.
12. Brizzi, A.; Brizzi, V.; Cascio, M. G.; Bisogno, T.; Siriani,
R.; Di Marzo, V. J. Med. Chem. 2005, 48, 7343.