Novel carbamate cholinesterase inhibitors that release biologically active amines following enzyme inhibition

Article abstract of DOI:10.1016/j.bmcl.2009.04.089

Conjugation of the phenol derived from rivastigmine with amphetamines gave access to novel carbamate cholinesterase inhibitors. All compounds possessed increased affinity and selectivity for AChE compared to rivastigmine and were orally bioavailable. Comp

Full text of DOI:10.1016/j.bmcl.2009.04.089

Bioorganic & Medicinal Chemistry Letters 19 (2009) 3243–3246
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel carbamate cholinesterase inhibitors that release biologically active
amines following enzyme inhibition
*
Jeroen C. Verheijen , Kjesten A. Wiig, Shoucheng Du, Stacie L. Connors, Ashley N. Martin, Jennifer P. Ferreira,
Vladimir I. Slepnev, Ulrike Kochendörfer
Sention Inc., 1 Richmond Square, Providence, RI 02906, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Conjugation of the phenol derived from rivastigmine with amphetamines gave access to novel carbamate
cholinesterase inhibitors. All compounds possessed increased affinity and selectivity for AChE compared
to rivastigmine and were orally bioavailable. Compound 4a, incorporating d-amphetamine, caused signif-
icant inhibition of cholinesterase in vivo at doses that were well tolerated. Release of amphetamine from
4a was demonstrated following in vitro and in vivo inhibition of cholinesterase. Compound 4a was also
effective in alleviating scopolamine induced amnesia in a rat passive avoidance model.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 24 March 2009
Revised 17 April 2009
Accepted 21 April 2009
Available online 24 April 2009
Keywords:
Acetylcholinesterase
Cholinesterase
Butyrylcholinesterase
Carbamates
Amphetamine
Alzheimer’s disease
Memory enhancement
Cholinesterase inhibitors (ChEIs) of the carbamate type (‘stig-
mines’) have been known for decades. Towards the end of the
19th century, physostigmine found medicinal use in the treatment
of glaucoma. More recently, physostigmine has been explored for
treatment of Alzheimer’s disease.1 However, the severity of the
side effects associated with high doses of physostigmine has
spurred the search for other carbamate cholinesterase inhibitors
that are safer and better tolerated. Examples of such stigmines
are the approved AD drug rivastigmine2 (marketed as ExelonÒ)
and the experimental AD drug phenserine3 (Fig. 1). Although these
second generation cholinesterase inhibitors are better tolerated
than physostigmine, their application is still hampered by limited
efficacy and a narrow therapeutic window.4
stigmines could provide both cholinesterase inhibition and actions
at additional relevant targets in a single molecule, potentially lead-
ing to increased efficacy and tolerability compared to known cho-
linesterase inhibitors.
Attempts to provide ChEIs with additional pharmacology have
been published.6–10 In contrast to our model, the published ap-
proaches required significant structure–activity modifications to
retain both ChE inhibition and the secondary pharmacology within
a single molecule. We have developed a general approach towards
ChE inhibitors with dual activity. By relying on the release of a
known amine with beneficial pharmacological properties, we ex-
pect to gain access to molecules incorporating a variety of activities
with minimal structural optimization.
Stigmines inhibit AChE by transferring their carbamoyl group to
a serine residue in the active site5 (semi-irreversible inhibition, see
Fig. 2). The covalently bound carbamate is slowly hydrolyzed to
reconstitute the active enzyme. During this process, a carbamic
acid is released that in turn dissociates into carbon dioxide and
an amine. In known stigmines, this amine is a small molecule that
is considered pharmacologically inactive. We hypothesized that
this mechanism could be leveraged to release a biologically active
amine during the process of carbamic acid dissociation. Thus, these
In this Letter, we present the synthesis and in vitro and in vivo
inhibitory activity of a number of novel carbamate ChEIs. Rates of
decarbamylation of AChE following inhibition by the novel conju-
gates are determined. Amine release following inhibition of cholin-
esterase is demonstrated in vitro as well as in vivo. Finally,
behavioral and cognitive effects are described.
Initial efforts were directed towards synthesis of rivastigmine-
based conjugates of amphetamine isomers. Rivastigmine was cho-
sen as starting point for the novel compounds because it has pro-
ven clinical utility in the management of Alzheimer’s disease and
because the corresponding phenol is chemically well character-
ized, accessible and stable. Amphetamines were selected due to
their attractive and well characterized pharmacological properties.
* Corresponding author at present address: Wyeth Research, 401 N. Middletown
Rd, Pearl River, NY 10965, USA.
E-mail address: verheij@wyeth.com (J.C. Verheijen).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.04.089

3244
J. C. Verheijen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3243–3246
H
H
O
N
O
N
O
N
N
N
H
N
O
O
N
O
N
H
Phenserine
Physostigmine
Rivastigmine
Figure 1. Representative examples of carbamate cholinesterases (‘stigmines’).
X
OH
O
X
R2
R1
O
N
R2
H
N
R1
O
B
B
H
H
O
H
H
O
O
B
B
Slow Hydrolysis
R2
CO2 + HNR1R2
R1
HO
N
B
H
O
H
H
B
O
B
H
H
O
B
Figure 2. Schematic mechanism of action for cholinesterase inhibition by carbamates.
Amphetamines are neurotransmitter uptake inhibitors11 with
stimulant properties12,13 which could be beneficial in the treat-
ment of (geriatric) depression and fatigue, two symptoms that of-
ten accompany Alzheimer’s disease. In addition, d-amphetamine is
a well known cognitive enhancer14–19 and it has recently been
shown that l-isomers of the amphetamines are potent cognitive
enhancers possessing only limited stimulant properties.20
Synthesis of the target compounds was accomplished as de-
picted in Scheme 1. Thus, (À)-3-hydroxyphenylethyldimethyl-
amine21 1 was reacted with carbonyl diimidazole in dry ethyl
acetate to form the activated imidazolide. Addition of the amphet-
amine isomers (2a–d) gave access to the target carbamates 3a–d in
good yields (40–60% after purification by column chromatogra-
phy). Free base carbamates 3a–d were converted into the corre-
sponding hydrochloride salts 4a–d by addition of hydrogen
chloride in diethyl ether followed by removal of the solvents. The
identity and purity of the resulting white solids was firmly estab-
lished by 1H NMR and HPLC analysis.
BuChE (human, purified from serum). Cholinesterase activity was
determined spectrophotometrically by a modified Ellman proce-
dure.22 The results of these experiments are provided in Table 1
and demonstrate that the novel carbamates were all more potent
inhibitors of acetylcholinesterase than the parent structure
(rivastigmine).
Interestingly, all novel carbamates had less affinity for butyr-
ylcholinesterase than rivastigmine. Consequently, all four novel
inhibitors were more selective AChE inhibitors than rivastigmine.
The increase in AChE selectivity was especially pronounced for
compounds 4a (incorporating d-amphetamine, which was 225-fold
more selective for AChE than rivastigmine) and 4d (incorporating l-
methamphetamine, which was 1200-fold more selective than
rivastigmine).
No clear structure–activity rules can be established from the
data in Table 1. For example, the compound incorporating the d-
isomer of amphetamine (4a) is 13-fold more selective than that
incorporating the l-isomer (4b), whereas incorporation of the d-
isomer of methamphetamine (4c) results in a compound with
20-fold less selectivity than in case of incorporation of the l-isomer
Having carbamates 4a–d in hand, attention was focused on
their inhibitory potential against AChE (human recombinant) and
R1
ii
i
OH
O
N
N
N
H
N
R2
2
O
R3
R
R1
1
R3
3a-d
2a-d
a. R1 = H, R2 = Me, R3 = H
b. R1 = H, R2 = H, R3 = Me
c. R1 = Me, R2 = Me, R3 = H
d. R1 = Me, R2 = H, R3 = Me
R1
N
O
N
HCl.
2
O
R3
R
4a-d
Scheme 1. Reagents and conditions: (i) carbonyldiimidazole, then 2a–d, 40–60%; (ii) HCl, Et2O, quant.

J. C. Verheijen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3243–3246
3245
Table 1
and 180 min following oral dosing of rivastigmine to assess time-
dependence. As there was no consistent difference in the level of
inhibition achieved between these times (Bonferroni t-test), which
is in accord with reports of its long duration of pseudo-irreversible
inhibition,23 plasma and brain levels of cholinesterase inhibition of
all other compounds were determined at a single time point
(60 min). All compounds resulted in significant inhibition of cho-
linesterase following oral dosing in both plasma and brain, demon-
strating the oral bioavailability and blood–brain barrier
penetration of compounds 4a–d.
Inhibitory concentrations (IC50’s) of cholinesterases 4a–d and the parent structure
rivastigmine against recombinant human acetylcholinesterase and purified human
butyrylcholinesterase (plasma)
Compd
Amine
rhAChE
IC50 (nM)
hBuChE
IC50 (nM)
Selectivityb
a
a
Rivastigmine
N/A
2615
508
404
131
302
179
7322
416
334
24400
1:15
14:1
1:1
3:1
81:1
4a
4b
4c
4d
d-Amphetamine
l-Amphetamine
d-Methamphetamine
l-Methamphetamine
The data in Table 2 reveal that there is no clear correlation be-
tween plasma or brain ChE inhibition and maximum tolerated
dose. Compound 4a, incorporating d-amphetamine, resulted in sig-
nificant inhibition of both brain (72%) and plasma (59%) ChE with-
out causing severe side effects (as assessed by salivation,
lacrimation and overt behavior). On the other hand, significant side
effects were observed following a dose of 32 mg/kg of compound
4b, associated with limited inhibition of brain (25%) and plasma
(26%) ChE. Although the increase in MTD for 4a–d could be ex-
plained by poor oral bioavailability, the high level of brain ChE
inhibition without severe side effects following dosing with 4a
suggests that there are other factors that influence the MTD as
well.
Next, the release of amphetamine from inhibited cholinesterase
was investigated indirectly by monitoring reconstitution of en-
zyme activity as described above. Recombinant human cholines-
terase was incubated with an excess of carbamates 4a–d,
resulting in >80% inhibition of enzyme activity. The enzyme was
separated from small organic molecules (carbamate and possible
degradation products) by size exclusion chromatography over a
Sephadex column. Purified, inhibited enzyme was incubated in
phosphate buffer at 37 °C, aliquots were drawn at various time
points and the enzymatic activity was determined as described
above. The data were plotted as the natural logarithm of the per-
cent inhibition versus time according to a first-order kinetics mod-
el. The decarbamylation rate constant k was determined as the
a
Values represent the mean of two or three independent experiments, each
performed in triplicate. Standard deviations were typically within 10% of the IC50
value.
b
Selectivity (IC50 hBuChE/IC50 rhAChE).
(4d). The addition of a methyl substituent on the d-amphetamine
backbone results in a fourfold loss of selectivity (compare 4a and
4c), whereas the same substitution results in a 70-fold increase
in selectivity in case of l-amphetamines 4b and 4d. The observed
IC50 reflects a complicated interplay between affinity for the en-
zyme, rate of enzyme carbamylation and rate of reconstitution of
enzymatic activity. Although more detailed studies are required
to fully explain the observed SAR, it is clear that replacement of
the methylethylamine moiety in rivastigmine with pharmacologi-
cally active amines can be tolerated.
Having established the in vitro inhibitory potential of the novel
stigmines, we next evaluated whether these conjugates were effec-
tive in vivo inhibitors of ChE. Initially, the Maximum Tolerated
Dose (MTD) was determined for each compound. The MTD was
the dose at which clear, yet reversible and non-life threatening
cholinergic effects were observed. The data in Table 2 clearly dem-
onstrate that the novel ChEIs were better tolerated than their par-
ent structure rivastigmine. Especially in case of d-amphetamine
derivative 4a, side-effects were very mild. At the highest dose
tested (100 mg/kg), only moderate cholinergic effects were ob-
served (mild fasciculations, some salivation). For comparison, riv-
astigmine (5 mg/kg, po) resulted in increased diarrhea, and
caused seizures and severe fasciculation, salivation and lacrima-
tion. It should also be noted that even after the highest dose of
4a, corresponding to 41 mg/kg d-amphetamine, no stimulant ef-
fects (e.g., locomotor stimulation) were observed, suggesting that
initial release of d-amphetamine following oral dosing of 4a is low.
To establish the in vivo cholinesterase inhibition properties of
the novel conjugates, rats were dosed orally with saline solutions
of the test compounds at the MTD (or at the highest dose tested
in case of 4a and 4c). Following administration, rats were sacrificed
and blood and brain samples were collected, processed and ChE
activity was determined spectrophotometrically as described
above. Initially, cholinesterase inhibition was quantified at 30, 60
slope of this line. The half-life t1/2 is directly related to
(k = Àln(0.500)/t1/2).
k
The results of the decarbamylation experiments are presented
in Table 3. Reconstitution of activity was significantly faster for
carbamates of primary amines (4a and 4b) than for carbamates
of secondary amines (rivastigmine, 4c and 4d). Hydrolysis of d-
amphetamine from the enzyme was especially rapid. A rapid
decarbamylation rate raises concerns about insufficient duration
of cholinesterase inhibition. However, following oral doses of 16–
64 mg/kg of 4a, the in vivo cholinesterase inhibition at 3 h post-
dosing was only 10–20% less than at 1 h post-dosing.
Based on the data presented above, compound 4a was selected
for further characterization. Compound 4a shows potent and selec-
tive inhibition of AChE in vitro and is a potent inhibitor of brain
ChE following oral dosing. More interestingly, doses of 4a associ-
ated with high levels of ChE inhibition are very well tolerated.
The relatively rapid rate of reconstitution of ChE activity following
inhibition by 4a suggests that relatively high levels of amine may
be released following doses of this compound.
Table 2
Inhibition of total brain cholinesterase and total plasma cholinesterase 60 min
following oral dosing of compounds 4a–d and rivastigmine at the maximum tolerated
dose or highest dose tested
Compd
Amine
MTD
(mg/kg)
Plasma
Brain
ChEIa (%)
ChEIa (%)
Table 3
Rivastigmineb
N/A
5
>100
32
>64
10
39
59
26
50
45
56
72
25
50
64
Decarbamylation rates (k) and decarbamylation half-lives (t1/2) following inhibition of
rhAChE by compounds 4a–d and rivastigmine
4a
4b
4c
4d
d-Amphetamine
l-Amphetamine
d-Methamphetamine
l-Methamphetamine
Compd
Amine
k (hÀ1
)
t1/2 (h)
Rivastigmine
N/A
0.007
0.558
0.057
0.003
0.011
>24
1.2
12
>24
>24
a
4a
4b
4c
4d
d-Amphetamine
l-Amphetamine
d-Methamphetamine
l-Methamphetamine
Values represent the mean of n = 4 animals per group. Standard deviations were
typically within 10% of the determined cholinesterase activity.
b
Rivastigmine data is the mean of three measurements between 30 and 180 min
following dosing.

3246
J. C. Verheijen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3243–3246
peripheral cholinergic effects are antagonized by the release of
300
250
200
150
100
50
Scopolamine (0.2 mg/kg)
adrenergic agents, the increased tolerability may be a reflection
of in vivo pharmacological effects of released d-amphetamine.
However, further research is warranted to confirm this hypothesis.
Acknowledgments
The authors are grateful to Dr. Kazumi Shiosaki and Dr. Randall
Carpenter for helpful advice.
*
*
References and notes
1. Thal, L. J.; Ferguson, J. M.; Mintzer, J.; Raskin, A.; Targum, S. D. Neurology 1999,
52, 1146.
2. For a recent review on rivastigmine, see: Williams, B. R.; Nazarians, A.; Gill, M.
A. Clin. Ther. 2003, 25, 1634.
3. Greig, N. H.; Sambamurti, K.; Yu, Q. S.; Brossi, A.; Bruinsma, G.; Lahiri, K. Alz.
Res. 2005, 2, 281.
0
Saline
0
8
12
16
24
Dose 4a (mg/kg)
Figure 3. Step-through latencies in a rat passive avoidance model. Rats received a
dose of 0.2 mg/kg scopolamine sc 40 min before training and doses of 0, 8, 12, 16 or
24 mg of 4a ip 30 min before training.
4. Benzi, G.; Moretti, A. Eur. J. Pharmacol. 1998, 346, 1.
5. Bar-On, P.; Millar, C. B.; Harel, M.; Dvir, H.; Enz, A.; Sussman, J. L.; Silman, I.
Biochemistry 2002, 41, 3555.
6. Brühlmann, C.; Ooms, F.; Carrupt, P.-A.; Testa, B.; Catto, M.; Leonetti, F.;
Altomare, C.; Carotti, A. J. Med. Chem. 2001, 44, 3195.
7. McKenna, M. T.; Proctor, G. R.; Young, L. C.; Harvey, A. L. J. Med. Chem. 1997, 40,
3516.
8. Toda, N.; Tago, K.; Marumoto, S.; Takami, K.; Ori, M.; Yamada, N.; Koyama, K.;
Naruto, S.; Abe, K.; Yamazaki, R.; Hara, T.; Aoyagi, A.; Abe, Y.; Kaneko, T.; Kogen,
H. Bioorg. Med. Chem. 2003, 11, 1935.
9. Kogen, H.; Toda, N.; Tago, K.; Marumoto, S.; Takami, K.; Ori, M.; Yamada, N.;
Koyama, K.; Naruto, S.; Abe, K.; Yamazaki, R.; Hara, T.; Aoyagi, A.; Abe, Y.;
Kaneko, T. Org. Lett. 2002, 4, 3359.
10. Sterling, J.; Herzig, Y.; Goren, T.; Finkelstein, N.; Lerner, D.; Goldenberg, W.;
Miskolczi, I.; Molnar, S.; Rantal, F.; Tamas, T.; Toth, G.; Zagyva, A.; Zekany, A.;
Lavian, G.; Gross, A.; Friedman, R.; Razin, M.; Huang, W.; Krais, B.; Chorev, M.;
Youdim, M. B.; Weinstock, M. J. Med. Chem. 2002, 45, 5260.
In order to unequivocally demonstrate that enzyme reconstitu-
tion results in release of amphetamine, aliquots from the decarb-
amylation experiments with 4a were analyzed for levels of
amphetamine using an LC/MS/MS method.24 It was demonstrated
that aliquots drawn at t = 0 did not contain amphetamine, whereas
aliquots drawn 4 h later (corresponding to 80% reconstitution) con-
tained 53 ng/mL of d-amphetamine. Similarly, plasma collected
from rats 1 h following oral doses (64 mg/kg) of 4a contained sig-
nificant amounts of amphetamine (17 ng/mL).
11. Creese, I.; Iversen, S. D. Brain Res. 1975, 83, 419.
Finally, in order to determine whether compound 4a had any
memory enhancing properties, we investigated its effects in a sco-
polamine model of passive avoidance.25 Prior to training, rats were
injected with scopolamine hydrobromide (0.2 mg/kg, sc, 40 min
before training) and with either saline or compound 4a (8, 12, 16
or 24 mg/kg, ip, 30 min prior to training). The animals received a
retention test 24 h following training. The retention test was iden-
tical to training except that no foot-shock or drug was delivered.
Latency to enter the dark chamber was recorded. As shown in Fig-
ure 3, pre-training administration of scopolamine produced a ro-
bust and statistically significant amnesia (p < 0.0001). This
amnesia could be alleviated by doses of 8 and 12 mg/kg of com-
pound 4a (p < 0.01). These results clearly demonstrate that com-
pound 4a is able to improve mnemonic performance in rats.
Moreover, the effectiveness of 4a in this model was comparable
to the effect of rivastigmine. Interestingly, the approved AD drug
donepezil was less effective in this model than 4a, and only
showed a trend toward reversal of amnesia (p = 0.09).
In conclusion, we have shown that bifunctional cholinesterase
inhibitors can be prepared through combination of a pharmacolog-
ically active amine and the phenol moiety of a known cholinester-
ase inhibitor in a single molecule. The resulting hybrid molecules
retain their ability to inhibit cholinesterase, both in vitro and
in vivo, and as demonstrated for 4a release pharmacologically ac-
tive amines following decarbamylation of the inhibited enzyme.
The high level of brain and plasma cholinesterase inhibition in
absence of severe side effects following oral doses of 4a is unprec-
edented and suggests that novel bifunctional cholinesterase inhib-
itors may have a greater therapeutic window than currently
known cholinesterase inhibitors. Since it can be expected that
12. Wise, R. A.; Rompre, P. P. Ann. Rev. Psychol. 1989, 40, 191.
13. Mason, S. T. Neurosci. Biobehav. Rev. 1983, 7, 325.
14. M’Harzi, M.; Willig, F.; Costa, J. C.; Delacour, J. Physiol. Behav. 1988, 42, 575.
15. Sara, S. J.; Deweer, B. Behav. Neural Biol. 1982, 36, 146.
16. Quartermain, D. J.; Martin, E.; Jung, H. Physiol. Behav. 1988, 43, 239.
17. Soetens, E.; Hooge, R. D.; Hueting, J. D. Neurosci. Lett. 1993, 161, 9.
18. Soetens, E.; Casaer, S.; D’Hooge, R.; Hueting, J. E. Psychopharmacology 1995, 119,
155.
19. Brown, R. W.; Bardo, M. T.; Mace, D. D.; Phillips, S. B.; Kraemer, P. J. Behav. Brain
Res. 2000, 114, 135.
20. Wiig, K. A.; Whitlock, J. R.; Epstein, M. H.; Carpenter, R. L.; Bear, M . F. Neurobiol.
Learn. Mem. 2009. doi:10.1016/j.nlm.2009.02.001.
21. (À)-3-Hydroxyphenylethyldimethylamine was accessible by acid hydrolysis
(concentrated HCl, 85 °C) of rivastigmine.
22. ChE activity was measured by the spectrophotometric method of Ellmann et al.
(Ellman, G. L.; Courtney, K. D.; Andres, V. Jr.; Feather-Stone, R. M. Biochem.
Pharmacol. 1961, 7, 88) with some modifications. The reaction was monitored
by following the change in absorbance spectrophotometrically, at 405 nm in a
Polarstar (BMG) spectrophotometer. The standard assay was carried out as
follows:Human recombinant Acetylcholinesterase (rhAChE) and human
butyrylcholinesterase purified from human serum were purchased from
Sigma Chemical Co.; St Louis, MO. ATCh (Sigma A-5751) (0.3 mM) was used
as substrate. Compounds were preincubated with the enzyme for 30 min at
25 °C in 10 mM Tris buffer pH 8.0 before addition of substrate. The final DTNB
(Sigma D-8130) concentration was 0.33 mM. Assays were performed in
triplicate at three independent measurements. Hydrolysis was monitored by
measuring the optical density at 405 nm for 5 min plotted against time. Initial
rates are calculated from the slope of the linear portion of the graph. Activities
were compared to the activity of the control and expressed as percent
inhibition. The data for a range of at least five [I] at a constant [S] were plotted
as a Dixon Plot. The IC50 values were determined from the opposite value of the
x-intercept at the given substrate concentration. Inhibitors were tested at
concentrations of 10 nM–100
observed initial IC50, the appropriate range for follow-up experiments was
selected from 10 to 1000 nM, 100 to 10,000 nM or 1 to 100 M.
lM in an initial experiment. Based on the
l
23. Enz, A.; Boddeke, H.; Gray, J.; Spiegel, R. Ann. NY Acad. Sci. 1991, 640, 272.
24. Assays were performed by MedTox Laboratories, St. Paul, MN.
25. Bejar, C.; Wang, R.-H.; Weinstock, M. Eur. J. Pharmacol. 1999, 383, 231.