Conformationally-restricted analogues and partition coefficients of the 5-HT3 serotonin receptor ligands meta-Chlorophenylbiguanide (mCPBG) and meta-Chlorophenylguanidine (mCPG)

Article abstract of DOI:10.1016/S0960-894X(03)00044-1

The present investigation examined two features of arylbiguanide and arylguanidine 5-HT₃ ligands: conformation and partition coefficients. Several conformationally-constrained analogues of mCPBG (2) and mCPG (11; K_i=32 nM) were prepared and of these only 2-amino-5-chloro-3,4-dihydroquinazoline (14; K_i=34 nM) retained high affinity. The partition coefficient of compound 11 (LogP_app=-0.64) was less than that of its corresponding arylbiguanide 2 (LogP_app=-0.38). The quinazoline structure may represent a pharmacologically-active conformation of these agents, and the arylbiguanides were found more lipid soluble than their arylguanidine counterparts at physiological pH.

Full text of DOI:10.1016/S0960-894X(03)00044-1

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1119–1123
Conformationally-Restricted Analogues and Partition
Coefficients of the 5-HT3 Serotonin Receptor Ligands
meta-Chlorophenylbiguanide (mCPBG) and
meta-Chlorophenylguanidine (mCPG)
Ashraf A. Rahman,a Maha Khalifa Daoud,a Ma•gorzata Dukat,a
Katharine Herrick-Davis,b Anil Purohit,b Milt Teitler,b
Antonia Taveres do Amaral,c Alberto Malvezzic and Richard A. Glennona,*
aDepartment of Medicinal Chemistry, School of Pharmacy, Medical College of Virginia Campus,
Virginia Commonwealth University, Richmond, VA 23298-0540, USA
bCenter for Neuropharmacology and Neuroscience, Albany Medical College, Albany, NY 12208, USA
cInstituto de Quimica, Universidade de Sa˜o Paulo, Sa˜o Paulo, Brazil
Received 30 September 2002; accepted 26 December 2002
Abstract—The present investigation examined two features of arylbiguanide and arylguanidine 5-HT3 ligands: conformation and
partition coefficients. Several conformationally-constrained analogues of mCPBG( 2) and mCPG( 11; Ki=32 nM) were prepared
and of these only 2-amino-5-chloro-3,4-dihydroquinazoline (14; Ki=34 nM) retained high affinity. The partition coefficient of
compound 11 (LogPapp=À0.64) was less than that of its corresponding arylbiguanide 2 (LogPapp=À0.38). The quinazoline struc-
ture may represent a pharmacologically-active conformation of these agents, and the arylbiguanides were found more lipid soluble
than their arylguanidine counterparts at physiological pH.
# 2003 Elsevier Science Ltd. All rights reserved.
Serotonin (5-HT) receptors are divided into seven major
families (5-HT1–5-HT7) and all are G-protein coupled
except for 5-HT3 receptors; the latter belong to the ion
channel superfamily of receptors.1À3 5-HT3 receptor
antagonists have found therapeutic application in the
treatment of emesis,2,3 but much less is known about the
potential therapeutic utility of 5-HT3 ligands with ago-
nist activity. One of the first 5-HT3 ligands identified
was phenylbiguanide (1). Although it was widely used as
an agonist in preclinical studies for several years, its
affinity for 5-HT3 receptors was only in the low micro-
molar range.2 meta-Chlorophenylbiguanide (mCPBG;
2)2,4 represented a significant advance in that it is an
agonist that binds at 5-HT3 receptors with higher affinity
(Ki <50 nM). It has also been demonstrated that the
chlorophenyl substituent of 2 can be replaced by a
2-naphthyl group (i.e., 3) with retention of affinity.5,6
Compared in the same investigation, the Ki values deter-
mined for 1–3 were 1200, 18, and 14 nM, respectively.7
The biguanide moiety common to 1–3 is con-
formationally flexible. In order to identify possible bio-
logically-preferred conformations that might account
for the binding of such compounds at 5-HT3 receptors,
we synthesized and examined several conformationally
restricted analogues (see Fig. 1). Another issue addressed
was the lipid solubility of the arylbiguanides. Evidence
*Corresponding author. Tel.: +1-804-828-8487; fax: +1-804-828-
7404; e-mail: glennon@hsc.vcu.edu
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00044-1

1120
A. A. Rahman et al. / Bioorg. Med. Chem. Lett. 13(2003) 1119–1123
Figure 1. Arylbiguanides and arylguanidines analogues examined in this study.
suggests that 1 displays poor brain penetration; hence, we
determined the partition coefficients of several derivatives
for comparison with their arylguanidine counterparts.
binding of arylbiguanides at 5-HT3 receptors. For
example, compound 8 (Ki=1,460 nM), a pyrimidine
analogue of 6, lacks one of the ring nitrogen atoms yet
binds with an affinity similar to that of unsubstituted
derivative 1.
Results and Discussion
It has been suggested that arylbiguanides might form an
intramolecular hydrogen bond to afford structures such
as 9.11 Several dihydrotriazines were prepared (i.e., 10)12
to mimic the general structure of 9. Both 10a and 10b
(Ki >10,000 nM) lacked 5-HT3 receptor affinity.
Conformationally-constrained analogues
Radioligand binding data9 are summarized in Table 1.
The first compound examined was the triazole 48 in
which the biguanide moiety was constrained in a five-
membered ring system. Compound 4 (Ki >10,000 nM)
was found to lack affinity for 5-HT3 receptors relative to
its ring-open counterpart 2 (Ki=18 nM). Compounds
1–3 were also conformationally constrained in a six-
membered ring (i.e., triazines 5–7).10 None of these ana-
logues displayed measurable affinity (Ki >10,000 nM in
each case) for 5-HT3 receptors. It is true that in addition
to fixing the conformation of the compounds the basi-
city of 4–7 has likely been altered relative to that of their
parents; however, at this time it is not known what role
the basicity of the various nitrogen atoms plays in the
We have previously found that the biguanide portion of
the arylbiguanides can be abbreviated to a guanidine
moiety with retention of 5-HT3 receptor affinity.5,7 For
example, 11 (meta-chlorophenylguanidine, mCPG;
Ki=32 nM; Table 1) is the guanidine analogue of 2. It
was thought that this smaller structure might aid in
minimizing the conformational possibilities of the flex-
ible chain. The first conformationally restricted analo-
gue of 11 to be considered was benzimidazole 12.
Morain et al.13 have previously shown that benzimida-
zole 13 binds at 5-HT3 receptors only in the low micro-
molar range. However, because replacement of the
chloro group of 2 with a –CF3 group dramatically
reduces its affinity, it is not known if the low affinity of
13 is attributable to its restricted conformation or to the
Table 1. Radioligand binding data for compounds examined
Compd
5-HT3 affinity
Ki, nM (ÆSEM)a
4
5
6
7
>10,000
>10,000
>10,000
>10,000
1460 (290)
>10,000
>10,000
32 (6)
8
10a
10b
11
12
14
725 (75)
34 (8)
aSEM not provided where Ki values are >10,000 nM.

A. A. Rahman et al. / Bioorg. Med. Chem. Lett. 13(2003) 1119–1123
1121
presence of the trifluoromethyl substituent. Hence, 1214
was prepared and examined to address this issue. Ben-
zimidazole 12 (Ki=725 nM) displayed 20-fold lower
affinity for 5-HT3 receptors than 11; yet, its affinity was
higher than that of any of the conformationally-con-
strained analogues examined up to this point.
the inability of 1 to establish stimulus control of beha-
vior might now be explained by its low lipophilicity and
inability to penetrate the blood–brain barrier. However,
2 has been used as a training drug in drug discrimina-
tion studies using rats and its stimulus effects appeared
to be centrally mediated.16 Likewise, 2 substituted for
arylguanidine 11 in rats trained to discriminate 11 from
vehicle; here too, it was shown that the stimulus effect is
likely centrally mediated.6 In contrast, it might be noted
that the 4-chloro analogue of 2 (i.e., 3,4-dichloro-
phenylbiguanide), which binds with higher affinity than
2 and which is presumably more lipophilic than 2, failed
to serve as a discriminative stimulus.17 Thus, while cer-
tain of these agents might penetrate the blood–brain
barrier only with difficulty, there is also evidence indi-
cating that some have central actions.
The centroid-to-amine distance of the flexible 11 is cal-
culated to range from 3.7 to 4.9 A.7 That for 12 (cen-
troid to NH2) is calculated to be 4.5 A. Because the
centroid-to-amine distance of 14 is calculated to be 4.9 A,
it was thought that 14 might better mimic the more exten-
ded form of 11. Compound 14 (Ki=34 nM), the ring-
expanded analogue of 12, was examined and found to bind
with enhanced affinity, and with an affinity nearly identical
to that of 11 itself. Of the various conformationaly-con-
strained analogues examined, only quinazoline 14 was
found to bind with low nanomolar affinity.
Compounds with LogP values in the 1.5–2.5 range will,
in general, readily enter the brain; in fact, some com-
pounds with significantly lower LogP values can pene-
trate the blood–brain barrier and display central
actions.19 An optimum value for LogP is most crucial
for agents with low receptor affinity.19 Nevertheless, the
LogP values obtained for 2 and 11 are certainly not
inconsistent with the concept that they might penetrate
the blood–brain barrier only with difficulty.
Partition coefficients
Intuitively, it might have been thought that the aryl-
guanidines, lacking the additional amidine moiety of the
arylbiguanides, would be more lipid soluble than their
corresponding arylbiguanides. This was not found to be
the case. Three pairs of arylbiguanides and arylguani-
dines were examined (Table 2) and in each case the aryl-
biguanides were found to be more lipid soluble than their
corresponding arylguanidines at physiological pH. In
both cases, the unsubstituted compounds (1 and 16) were
the most soluble in aqueous media; the monochloro
derivatives 2 and 11 were more lipid soluble, and the tri-
chloro derivatives 15 and 17 were the most lipid soluble.
Summary
The present investigation provides evidence that quina-
zoline analogue 14, the first member of a novel series of
5-HT3 ligands, likely represents a biologically active
conformation of the arylguanidines. Also, the arylbi-
guanides were found to possess greater lipid solubility
than their corresponding arylguanidines; as expected,
introduction of multiple chloro groups resulted in the
most lipophilic derivatives. Because both the arylbigua-
nides and aryl guanidines displayed low lipid solubility,
future studies will focus on the synthesis of quinazoline
analogues of these agents and, in particular, on multi
chloro-substituted analogues. This study, then, demon-
strates how the future construction of novel, brain-
penetrant 5-HT3 ligands with enhanced lipid solubility
might be achieved. Such studies are now underway in
our laboratories.
There is relatively limited literature information on the
lipid solubility of arylbiguanides and their ability to pene-
trate the blood–brain barrier. In an ex vivo study, Bachy
et al.15 found that 1 and 2 displayed poor brain penetra-
tion in mice as determined by measuring their ability to
displace [3H]granisetron 30min following systemic
administration. A number of years ago we unsuccessfully
attempted to train rats to discriminate 1 from vehicle in
a two-lever operant procedure (unpublished findings);
Table 2. LogPapp of several arylbiguanides and arylguanidines18
Experimental
Melting points (uncorrected) were obtained with a
Thomas Hoover apparatus. 1H NMR spectra were
recorded with a Varian EM-390 spectrometer, and peak
positions are given in parts per million (d) downfield
from tetramethylsilane as internal standard. Micro-
analyses were performed by Atlantic Microlab (GA,
USA) for the indicated elements, and the results are
within 0.4% of theory. Reactions and product mixtures
were routinely monitored by thin-layer chromatography
on silica gel precoated F254 Merck plates. Several of the
compounds used in the present study were known and
were prepared according to literature procedures; the
literature citations are provided in the text.
Compd
R
LogPapp (SEM)
1
2
15
H
3-Cl
3,4,5-Cl3
À1.17 (0.04)
À0.38 (0.02)
1.47 (0.01)
16
11
17
H
3-Cl
3,4,5-Cl3
À1.32 (0.18)
À0.64 (0.08)
1.16 (0.02)

1122
A. A. Rahman et al. / Bioorg. Med. Chem. Lett. 13(2003) 1119–1123
2-Amino-4-(3-chlorophenyl)aminopyrimidine hydrochloride
(8). A mixture of 2-amino-4-chloropyrimidine (2.0 g,
20 mmol), 3-chloroaniline (1.6 mL, 20 mmol) and concd
HCl (0.1 mL) in H2O (15 mL) was heated at reflux for
2 h. The clear dark solution was warmed with charcoal
powder and filtered hot. The filtrate was allowed to
cool to room temperature, and made strongly alkaline
by the addition of 10 N NaOH. The precipitated pro-
duct was collected and recrystallized from benzene to
afford 3.2 g (94%) of product (mp 126–128 ꢀC) which
was converted to its HCl salt: mp 191–194 ꢀC; IR
References and Notes
1. Glennon, R. A.; Dukat, M.; Westkaemper, R.B. In Wat-
son, S. J., Ed.; Psychopharmacology, A Generation of Progress;
Lippincott-Raven: New York, 1998 (CD ROM Version).
2. Gozlan, H. 5-HT3 Receptors. In Olivier B, van Wijngaar-
den I, Soudin, W., Eds.; Serotonin Receptors and Their
Ligands; Elsevier: Amsterdam, 1997; p 221.
3. King, F. D., Jones, B. J., Sanger, G. J., Eds. 5-Hydroxy-
tryptamine-3Receptor Antagonists . CRC: Boca Raton,
1994.
4. Kilpatrick, G. J.; Butler, A.; Burridge, J.; Oxford, W. Eur.
J. Pharmacol. 1990, 182, 193.
1
(KBr, cmÀ1): 3340, 1560. H NMR (DMSO-d6) d 6.22
(d, 1H, Pyr–H), 6.90 (br s, 2H, D2O-exchangeable),
7.11 (d, 1H, ArH), 7.40 (m, 1H, ArH), 7.80 (d, 1H,
ArH), 7.90 (s, 1H, ArH), 8.05 (d, 1H, Pyr–H), 9.40 (s,
1H, NH, D2O-exchangeable). Anal. (C10H9ClN4) C, H,
N.
5. Dukat, M.; Abdel-Rahman, A. A.; Ismaiel, A. M.; Ingher,
S.; Teitler, M.; Gyermek, L.; Glennon, R. A. J. Med. Chem.
1996, 39, 4017.
6. Dukat, M.; Young, R.; Darmani, N. N.; Ahmed, B.; Glen-
non, R. A. Psychopharmacology (Berl) 2000, 150, 200.
7. Dukat, M.; Choi, Y.; Teitler, M.; Du Pre, A.; Herrick-
Davis, K.; Smith, C.; Glennon, R. A. Bioorg. Med. Chem.
Lett. 2001, 11, 1599.
2-Amino-7-chloro-3,4-dihydroquinazoline hydrochloride
(14). Following the general procedure of Grosso et al.20
a suspension of S-methylisothiouronium sulphate (3.6 g,
18 mmol) and Na2CO3 (2.14 g, 20 mmol) in dry dioxane
was heated until solution was complete. Chloroisatoic
anhydride (3.56 g, 13 mmol) was added to the reaction
mixture; the white suspension, which became yellow,
was heated at reflux for further 20 h. The reaction
mixture was allowed to cool at room temperature,
poured into H2O (60 mL), and allowed to stir for
15 min. The product was recovered by suction filtration
and the filter cake was dried under vacuum at 60 ꢀC for
24 h. The product was recrystallized from dioxane to
yield 1.50 g (40%) of the desired quinazolinone: mp
>300 ꢀC.
8. Synthesized as reported by Blank, B.; Nichols, D. M.; Vai-
dya, P. D. J. Med. Chem. 1972, 15, 694.
9. The radioligand binding assays were performed in triplicate
as previously described7 using NG108-15 cells which express
the 5-HT3 receptor. [3H]GR65630 (New England Nuclear) was
used as radioligand and 1 mM tropisetron (ICS 205-930) was
used to define nonspecific binding; a 64% specific binding sig-
nal was produced using 1 nM [3H]GR65630 (84.2 Ci/mmol).
Assay tubes were incubated for 30 min at room temperature;
suspensions were filtered on presoaked Schleicher & Schuell
#32 glass fiber filters and washed with 10 mL of ice-cold buf-
fer. The filters were counted by a Beckman 3801 liquid scin-
tillation counter in 5 mL of aqueous counting scintillant
(Ecoscint; National Diagnostic). Data from binding assays
were plotted as log concentration versus percent inhibition
and analyzed by nonlinear least squares techniques in which
100% maximal inhibition was assumed at high test compound
concentrations. IC50 values obtained from such data treatment
were used to calculate apparent inhibition constants from the
following equation: Ki=IC50/1+([c]/KD) where [c] is the con-
centration of radioligand employed in the binding assay, and
KD is its receptor dissociation constant (KD=0.7 nM) for
[3H]GR65630.
10. Synthesized as HCl salts: Laszlo, S.; Laszlo, S.; Otto, C.
Acta Pharm. Hung. 1961, 31, 163. Shah, M. H.; Deliwala,
C. V.; Sheth, U. K. J. Med. Chem. 1968, 11, 1167.
11. Shapiro, S. L.; Parrino, V. A.; Rogow, E.; Freedman, L.
J. Am. Chem. Soc. 1959, 81, 3725.
12. Compounds 10a and 10b were obtained as their di- and
mono-HCl salts respectively according to Modest, E. J. J. Org.
Chem. 1956, 21, 1. Colebrook, L. D.; Giles, H. G.; Rosowsky,
A.; Bentz, W. E.; Fehlner, J. R. Can. J. Chem. 1976, 54, 3757.
13. Morain, P.; Abraham, C.; Portevin, B.; DeNanteuil, G.
Mol. Pharmacol. 1994, 46, 732.
14. Prepared as the acetate salt as reported by Leonard, N. J.;
Curtin, D. Y.; Beck, K. M. J. Am. Chem. Soc. 1947, 69, 2459.
15. Bachy, A.; Heaulme, M.; Giudice, A.; Michaud, J. C.;
Lefevre, I. A.; Souilhac, J.; Manara, L.; Emerit, M. B.;
Gozlan, H.; Hamon, M.; Keane, P. E.; Soubrie, P.; Le Fur, G.
Eur. J. Pharmacol. 1993, 237, 299.
The quinazolinone (0.75 g, 3.82 mmol) was added to
1 M BH3–THF (12.25 mL) and the resulting solution
was heated at reflux under N2 for 2 h. The borate com-
plex and excess reagent were hydrolyzed by the drop-
wise addition of 6 N HCl (3 mL), and the acidic solution
was basified with 6 N NaOH (12 mL). The mixture was
concentrated and the residue was extracted with hot
chloroform (3 Â 10 mL). The combined extracts were
concentrated under vacuum to yield 0.15 g (21%) of a
white solid: mp 230 ꢀC. The solid in absolute EtOH
(10 mL) was titrated to pH ꢁ3 with 37% aqueous HCl.
Solvent was removed under vacuum and the residue was
dissolved in absolute EtOH (10 mL) and evaporated to
dryness under vacuum; the process was repeated three
more times. The residue was recrystallized from abso-
lute EtOH/anhydrous Et2O to yield 0.07 g (38%) of 14:
mp 240–242 ꢀC; IR (KBr, cmÀ1): 3180, 3078, 1620. H
1
NMR (DMSO-d6) d 4.72 (s, 2H, CH2Ar), 6.98–7.01 (d,
1H, ArH), 7.76 (s, 1H, ArH), 8.66 (bs, 2H, NH, D2O-
exchangeable), 11.12 (s, 1H, NH, D2O-exchangeable).
Anal. (C8H8ClN.3 HCl) C, H, N.
16. De La Garza, R.; Callahan, P. M.; Cunningham, K. A.
Pharmacol. Biochem. Behav. 1996, 54, 533.
17. Olivier, B.; Broersen, L. M.; Slangen, J. L. Prog. Neuro-
Psychopharm. Biol. Psychiat. 2002, 26, 463.
Acknowledgements
Both A.R. and M.K.D. were recipients of Egyptian
Channel Program scholarships. We would like to thank
Dr. Ola El-Sayed for her assistance with the synthesis of
compound 4.
18. Partition coefficients: LogPapp was determined in an aqu-
eous 1-octanol buffer system by the shake-flask method. All
experiments were performed at pH=7.4 (ionic strength of
0.10 M NaCl). Room temperature was controlled and set to

A. A. Rahman et al. / Bioorg. Med. Chem. Lett. 13(2003) 1119–1123
1123
23Æ1.0 ꢀC. Concentrations of compound in the aqueous
phase were measured via spectrophotometer and the results
represent the means of duplicate determinations run in tripli-
cate. See do Amaral et al.21 and references therein for greater
experimental detail. Compounds were prepared as previously
described.7
19. Hansch, C.; Bjorkroth, J. P.; Leo, A. J. Pharm. Sci. 1987,
76, 663.
20. Grosso, J. A.; Nichols, D. E.; Nichols, M. B.; Yim,
G . K. W.J. Med. Chem. 1980, 23, 1261.
21. do Amaral, A. T.; Oliveira, A. C.; Neidlein, R. M.; Gallacci,
M.; Caprara, L.; Miyazaki, Y. Eur. J. Med. Chem. 1997, 32, 433.