An improved process for the preparation of pimavanserin tartrate

Article abstract of DOI:10.1177/1747519819873643

A practical synthetic route to pimavanserin tartrate, in which the target compound was obtained with 99.84% purity and in 46% total yield via a 5-step synthesis starting from 4-hydroxybenzaldehyde and (4-fluorophenyl)methanamine, is reported. The main advantages of the route include inexpensive starting materials, mild reaction conditions and an acceptable overall yield.

Full text of DOI:10.1177/1747519819873643

873643CHL0010.1177/1747519819873643Journal of Chemical ResearchWu et al.
research-article2019
Research Paper
Journal of Chemical Research
1–6
© The Author(s) 2019
Article reuse guidelines:
sagepub.com/journals-permissions
An improved process for the preparation
of pimavanserin tartrate
https://doi.org/10.1177/1747519819873643
DOI: 10.1177/1747519819873643
journals.sagepub.com/home/chl
Caijiao Wu*, Qifan Zhou*, Dake Song, Hui Li, Changshun Bao,
Xuelong Liu, Xuefei Bao and Guoliang Chen
Abstract
A practical synthetic route to pimavanserin tartrate, in which the target compound was obtained with 99.84% purity
and in 46% total yield via a 5-step synthesis starting from 4-hydroxybenzaldehyde and (4-fluorophenyl)methanamine,
is reported. The main advantages of the route include inexpensive starting materials, mild reaction conditions and an
acceptable overall yield.
Keywords
optimization, Parkinson’s disease, pimavanserin tartrate, synthesis, 5-HT2A receptors
Date received: 31 March 2019; accepted: 13 August 2019
However, unsafe N,N,N’,N’-tetramethylnaphthalene-1,8-
diamine and diphenyl phosphorazidate were used in this
route. To avoid the application of toxic phosgene, the
groups of Wang and Cheng reported another two novel
Introduction
Pimavanserin has been administered as its tartrate salt (ACP-
103, P) (Figure 1) and was used as a novel inverse agonist of
5-HT2A receptors. It was independently developed by Acadia
Pharmaceuticals and approved by the US Food and Drug
Administration (FDA) in 2016 for the treatment of psychosis
and other neuropsychiatric indications.1,2
Key Laboratory of Structure-Based Drug Design & Discovery, Ministry
of Education, Shenyang Pharmaceutical University, Shenyang, P.R. China
The reported routes to 1 are depicted in Scheme 1.
*These authors contributed equally to this work.
Initially, Acadia Pharmaceuticals disclosed a method for its
synthesis in the early stage of drug discovery (Route A).3 In
this method, 4-hydroxybenzaldehyde was used as a starting
material in 6-step process to provide 1 with a total yield of
11%. The main drawbacks of this approach were that the
use of toxic phosgene and H2/Pd-C as a reducing reagent
under high pressure, which seriously hampered its large-
scale production. Gant and Sarshar4 developed another
method (Route B) to provide 1 from methyl-2-(4-hydroxy-
phenyl) acetate in five steps with a total yield of 21%.
Corresponding authors:
Xuefei Bao, Key Laboratory of Structure-Based Drug Design &
Discovery, Ministry of Education, Shenyang Pharmaceutical University,
Shenyang 110016, P.R. China.
Email: bxf551@163.com
Guoliang Chen, Key Laboratory of Structure-Based Drug Design &
Discovery, Ministry of Education, Shenyang Pharmaceutical University,
Shenyang 110016, P.R. China.
Email: spucgl@163.com

2
Journal of Chemical Research 00(0)
reaction results. When 9 was added to a mixture of 6 in
DCM, impurity A-2 (Figure 2) was readily formed. If 6 was
added to a mixture of 9 in DCM, A-2 was not detected.
On this basis, the effects of the amount of BTC, the solvent
and the temperature were investigated. As shown in Table 2,
the reaction temperature of the first step had a significant
influence on the yield. It was found that slow, dropwise addi-
o
tion of a mixture of 5 and Et3N at -10 C to the solution of
BTC (1.5 equiv with respect to phosgene) in DCM was opti-
mum. After the reaction was complete, the precipitates were
removed and the filtrate was added dropwise to a solution of
9. The reaction mixture was stirred for 3 h at room tempera-
ture and the total yield of the two steps was 70% (entry 8).
An attempt to increase the purity and yield by changing
the recrystallization solvents proved successful. It was
found that a yield of 70% could be obtained by recrystalli-
zation from ethyl acetate (Table 3).
Finally, salt formation process was carried out in ethyl
acetate, when ethyl acetate was used as solvent, the reac-
tion mixture was stirred for 1 h at 50 oC, an acceptable yield
of 89% could be obtained.
Figure 1. Chemical structure of pimavanserin tartrate (1).
synthetic routes by using 4-hydroxybenzylamine (Route C,
in <16% total yield)5 and 4-fluorobenzylamine (Route D, in
<52% total yield),6 respectively, as starting materials.
However, the reagents phenyl chloroformate and benzyl
chloroformate have certain toxicity and corrosiveness.
Moreover, 4-hydroxybenzylamine and 4-fluoroben-
zylamine are relatively expensive.
Accordingly, we discuss our attempts to develop a safe
and scalable process for the synthesis of 1 and seek to define
a more efficient and cost-effective route that is appropriate
for more sustainable long-term supply of the drug (Scheme
2). The new synthesis started from commercially available
p-hydroxybenzaldehyde (2), which was converted into ether
4viareactionwith1-chloro-2-methylpropane. Subsequently,
ether 4 was transformed into amine 5 via reductive amina-
tion reaction using Raney-nickel. Reaction of amine 5 and
with bis(trichloromethyl)carbonate (BTC) yielded isocy-
anate 6, which upon treatment with fluoro compound 9 fur-
nished pimavanserin. Finally, pimavanserin was reacted
with tartaric acid to afford pimavanserin tartrate (1).
Conclusion
In summary, we have provided an alternative method for
the production of pimavanserin tartrate. The modified route
avoids several drawbacks of the original synthesis such as
removal of highly toxic and expensive reagents. The
method also reduces the cost and improves the yield. The
procedure involves in five linear steps with an overall yield
of 46% and a 99.84% high-performance liquid chromatog-
raphy (HPLC) purity of the target compound.
Experimental
Results and discussion
General
During the process for the assembly of 5, we found that a
small quantity of an impurity, a light oil, that was identified
5a formed (Table 1). Efforts to improve the conversion rate
and to reduce the yield of 5a by controlling the temperature
and amount of catalyst were unsuccessful. As 5a did not
affect the next synthetic step, the reduced product was directly
used in the next step without any further purification.
The main methods to synthesize isocyanates are as fol-
low: amines reacting with phosgene,3 using carboxylic acids
in Curtius rearrangements in the presence of azide,4 and uti-
lizing N,N’-carbonyldiimidazole (CDI) as an activated
amide.7 Unfortunately, CDI readily formed the self-conden-
sation byproduct A-1 (Figure 2). Consequently, we chose
BTC (bis(trichloromethyl)carbonate) to synthesize isocy-
anates in order to reduce the generation of impurity A-1.8
Interestingly, it was found that the order of addition had
a significant influence on the reaction. When BTC was
added to 5 in dichloromethane (DCM), a large quantity of
A-1 was observed. It is speculated that this was due to 5
being present in excess compared to BTC in the reaction
mixture. This problem was solved by adding 5 to a mixture
of BTC in DCM in a dose-controlled manner. During fur-
ther study on the synthesis of pimavanserin, it was found
that the order of addition had a significant influence on the
Reagents and solvents were obtained from commercial sup-
pliers and used without further purification. Melting points
were determined with a Yanaco apparatus and were uncor-
1
rected. H and 13C nuclear magnetic resonance (NMR)
spectra were recorded using a Bruker ARX-400 apparatus
(Billerica, MA, USA) with tetramethylsilane (TMS) as an
internal standard. Electrospray ionization mass spectrome-
try (ESI-MS) data were obtained using an
AlliAnce+QuattromicroTM Series liquid chromatography-
tandem quadrupole mass spectrometer. The reactions were
monitored by thin-layer chromatography (TLC; HG/
T2354-92, GF254), and compounds were visualized on
TLC with ultraviolet (UV) light. HPLC analyses were per-
formed on a Shimadzu LC-20, column, Shimadzu
Inertsil-SP C18 (5 µm; 250 mm × 4.6 mm).
Preparation of 4-isobutoxybenzaldehyde (4): Potassium
carbonate (69.0 g, 500 mmol) and potassium iodide (6.6 g,
40 mmol), 4-hydroxybenzaldehyde (2) (24.4 g, 200 mmol)
and isobutyl chloride (3) (36.8 g, 400 mmol) were added
to dimethylformamide (DMF) (140 mL). After being
stirred at reflux for 18 h, the mixture was cooled to ambi-
ent temperature filtered under vacuum. The filtrate was
concentrated in vacuo to remove DMF. The residue was

Wu et al.
3
Scheme 1. Synthetic routes toward pimavanserin tartrate or pimavanserin.
dissolved in EtOAC (100 mL) and H2O (100 mL). The and the filtrate was concentrated in vacuo to afford 4 as a
layers were separated, and the aqueous layer extracted yellow oil (34.3 g, 96%). Crude 4 was used directly in the
with EtOAC (100 mL). The organic phase was collected, next step. A sample of 4 was analyzed as follows. ESI-MS
washed with brine (100 mL) and then dried over anhy- m/z: 179.1 [M+H]+.1H NMR (400 MHz, CDCl3): δ (ppm)
drous sodium sulfate. The sodium sulfate was filtered off 9.88 (s, 1H), 7.82 (d, J = 8.7 Hz, 2H), 6.99 (d, J = 8.7 Hz,

4
Journal of Chemical Research 00(0)
Scheme 2. Our an alternative synthetic route to pimavanserin tartrate (1).
Table 1. Optimization of the reduction reactiona.
Entry
Ni (%)b
Temperature (oC)
Yieldc (%)
5 . . . 5a
1
2
30
4
5
6
7
8
5
5
30
50
30
50
30
50
50
50
26 . . . 0
43 . . . 1
35 . . . 1
46 . . . 1
37 . . . 1
80 . . . 1
81 . . . 2
81 . . . 2
10
10
15
15
20
25
aStandard conditions: 4 (200 mmol), 10% w/v solution of ammonia in methanol (150 mL).
bBased on the quality of 4.
cIsolated yield after chromatography.
2H), 3.80 (d, J = 6.5 Hz, 2H), 2.15–2.08 (m, 1H), 1.04 (d, 6.91–6.86 (m, 4H), 4.61 (s, 4H), 3.72 (d, J = 6.6 Hz, 4H),
J = 6.7 Hz, 6H).
2.08 (dt, J = 13.3, 6.7 Hz, 2H), 1.02 (d, J = 6.7 Hz, 12H).
Preparation of (4-isobutoxyphenyl) methanamine (5):
Preparation of N-(4-fluorobenzyl)-1-methylpiperidin-4-
4-isobutoxybenzaldehyde (4) (28 g, 200 mmol), Raney- amine (9): (4- fluorophenyl)methanamine (7) (15.0 g, 120
nickel (4.2 g, 15%, 47% w/w) and a 10% w/v solution of mmol), 1-methylpiperidin-4-one (8) (13.6 g, 120 mmol),
ammonia in methanol (150 mL) were stirred at 50–55 °C for and zinc chloride (3.3 g, 24 mmol) were added to methanol
8 h at a hydrogen pressure of 2.02 kg/cm2. The reaction mix- (200 mL). The mixture was cooled below 0 °C, and sodium
ture was cooled to ambient temperature and isolated by cyanoborohydride (11.3 g, 180 mmol) was slowly added.
vacuum filtration. The filtrate was concentrated in vacuo to The mixture was stirred at 40 °C for 4 h, then cooled to
afford 5 as a light yellow oil (22.6 g, 80%). ESI-MS m/z: ambient temperature and concentrated in vacuo to remove
1
180.1 [M+H]+. H NMR (400 MHz, CDCl3): δ (ppm) methanol. The residue was dissolved in DCM (300 mL),
7.23–7.19 (m, 2H), 6.89–6.84 (m, 2H), 3.80 (s, 2H), 3.71 (d, washed with 10% sodium hydroxide (450 mL) and brine
J = 6.6 Hz, 2H), 2.12–2.02 (m, 1H), 1.02 (d, J = 6.7 Hz, (200 mL), and then dried over anhydrous sodium sulfate.
6H). Impurity 5a (yield: 1.5%). ESI MS m/z: 342.2 [M+H]+. The sodium sulfate was filtered off and the filtrate was con-
1H NMR (400 MHz, CDCl3): δ (ppm) 7.30–7.26 (m, 4H), centrated in vacuo to 9 as a yellow oil (25.9 g, 97.1%).

Wu et al.
5
Figure 2. Chemical structures of A-1 and A-2.
Table 2. Optimization of the amount of BTC, the solvent and the temperature on the yield of pimavanserina.
Entry
BTC (equiv)
Solvent
Temperature 1b (oC) Temperaturec 2 (oC) Yield (%)
1
2
3
4
5
6
7
8
9
1.2
1.5
1.5
1.5
1.5
1.5
1.8
01.5
1.5
1.5
THF
THF
THF
THF
THF
THF
THF
DCM
DCM
DCM
25
25
0
–10
–10
–20
–10
–10
–20
–10
25
25
25
25
50
50
50
25
20
41
51
69
68
66
68
70
69
69
25
reflux
10
BTC: bis(trichloromethyl)carbonate; DCM:dichloromethane;THF:tetrahydrofuran.
aStandard conditions: 5 (100 mmol), Et3N (290 mmol).
bTemperature for the reaction of 5 and BTC.
cTemperature for the reaction of 6 and 9.
Table 3. Effect of the solvent on the yield of recrystallizationa.
(29.8 g, 290 mmol) in DCM (80 mL) while maintaining the
temperature below 10 °C. When the addition was complete
(0.5 h), the reaction mixture was isolated by vacuum filtra-
tion. The filtrate 6 was added dropwise to a mixture of
N-(4-fluorobenzyl)-1-methylpiperidin-4-amine (9) (21.8 g,
100 mmol) and DCM (80 mL) over 2.5 h. The mixture was
then stirred at 25 °C for 3 h. The reaction mixture was
extracted with H2O (720 mL) and the combined aqueous
layer was extracted with DCM (240 mL). The organic
phase was collected, washed with brine (300 mL) and then
dried over anhydrous sodium sulfate. The sodium sulfate
was filtered off, and the filtrate was concentrated in vacuo
to afford a brown oil (40.0 g). The crude material was
recrystallized form EtOAC (160 mL) to afford pimavan-
serin as a beige solid (29.4 g, 70%). Recrystallization pro-
cess: crude pimavanserin was dissolved by heating in
EtOAC and cooled for 30 min. The mixture was then fil-
tered to obtain pure pimavanserin. A sample of purified
aStandard conditions: Pimavanserin (93.6 mmol).
ESI-MS m/z: 223.1 [M+H]+. 1H NMR (400 MHz, CDCl3):
δ (ppm) 7.31–7.26 (m, 2H), 7.03–6.97 (m, 2H), 3.78 (s,
2H), 2.85–2.80 (m, 2H), 2.51–2.46 (m, 1H), 2.27 (s, 3H),
2.03–1.97 (m, 2H), 1.93–1.87 (m, 2H), 1.50–1.41 (m, 2H).
Preparation of Pimavanserin tartrate (1): A solution of
BTC (14.6 g, 150 mmol) in DCM (80 mL) was stirred at
−10 °C, followed by the dropwise addition of (4-isobutoxy-
phenyl) methanamine (5) (17.6 g, 100 mmol) and Et3N
o
pimavanserin was analyzed as follows. M.p. 119–122 C
(Lit.,9 m.p.124 oC). ESI-MS m/z: 428.3 [M+H]+. 1H NMR
(600 MHz, CDCl3): δ (ppm) 7.18 (dd, J = 8.4, 5.3 Hz, 2H,
ArH), 6.99 (m, 4H, ArH), 6.78 (d, J = 8.6 Hz, 2H, ArH),
4.50 (t, J = 5.5 Hz, 1H, NH), 4.34 (s, 2H, CH2), 4.33-4.31

6
Journal of Chemical Research 00(0)
(m, 1H, CH), 4.27 (d, J = 5.4 Hz, 2H, CH2), 3.68 (d, J =
6.6 Hz, 2H, CH2), 2.88-2.85 (m, 2H, CH2), 2.26 (s, 3H,
CH3), 2.09-2.05 (m, 3H, CH and CH2), 1.73-1.65 (m, 4H,
CH2), 1.01 (d, J = 6.7 Hz, 6H, CH). 13C NMR (400 MHz,
DMSO-d6): δ (ppm) 162.54, 160.14, 158.03, 157.92,
137.55, 133.57, 128.85, 128.78, 128.61, 115.29, 115.08,
114.53, 74.24, 55.38, 52.85, 46.29, 44.43, 43.59, 30.45,
28.19, 19.52. HRMS (ESI) calcd for C25H34FN3O2
[M+H]+: 428.2635, found: 428.2728. Impurity A-1:
CH3).13C NMR (600 MHz, DMSO-d6): δ (ppm) 174.78,
161.36 (d, J = 240 Hz), 157.93, 137.26, 133.43, 128.80
(d, J = 8.0 Hz), 128.61, 115.24 (d, J =21 Hz), 114.54,
74.24, 72.23, 54.48, 51.79, 44.68, 44.55, 43.57, 29.08,
28.17, 19.52. Pimavanserin tartrate 1 with 99.84% (rela-
tive area) of HPLC purity (eluent, acetonitrile/water =
55/45; flow rate = 1 mL/min; temperature 40 oC; wave-
length 231 nm. The HPLC analysis data are reported in
relative area % and was not adjusted to weight %).
1
ESI-MS m/z: 385.4 [M+H]+, 407.4 [M+Na]+. H NMR
(400 MHz, CDCl3): δ(ppm) 7.14 (d, J = 8.5 Hz, 4H), 6.81
(d, J = 8.6 Hz, 4H), 4.69 (s, 2H), 4.24 (d, J = 5.3 Hz, 4H),
3.67 (d, J = 6.6 Hz, 4H), 2.11–2.01 (m, 2H), 1.01 (d, J =
6.7 Hz, 12H). Impurity A-2: ESI-MS m/z: 633.4 [M+H]+.
1H NMR (400 MHz, CDCl3): δ(ppm) 7.17–7.15 (m, 2H),
7.07–7.03 (m, 4H), 6.87–6.77 (m, 6H), 5.96 (t, J = 5.2 Hz,
1H), 4.70 (s, 2H), 4.29 (s, 2H), 4.28 (d, J=5.4 Hz, 2H), 3.70
(d, J = 6.6 Hz, 2H), 3.68 (d, J = 6.6 Hz, 2H), 3.65–3.63
(m, 1H), 2.77 (d, J=11.2 Hz, 2H), 2.23 (s, 3H), 2.11–2.03
(m, 2H), 1.86 (d, J=11.4 Hz 2H), 1.73–1.64 (m, 2H), 1.28
(d, J=10.8 Hz, 2H), 1.02 (d, J = 6.7 Hz, 12H).
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
Funding
The author(s) received no financial support for the research,
authorship, and/or publication of this article.
ORCID iD
Pimavanserin (20 g, 46.8 mmol) in EtOAC (200 mL)
was stirred at 50 °C. Then L-(+)-tartaric acid (3.6 g,
24.0 mmol) was added to the mixture which was stirred
Guoliang Chen
https://orcid.org/0000-0003-3674-5226
Supplemental material
o
at 50 C for 1 h. The mixture was cooled to −15 °C,
1H spectra, MS spectra, HPLC traces: This material can be found
via the “Supplementary Content” section of this article’s web
page.
stirred for 0.5 h and then was filtered. The residue was
washed with cold EtOAC (40 mL) and dried at 45–55 °C
under vacuum to afford 1 as a white solid (22.4 g, 95%,
based on pimavanserin). Crude 1 was recrystallized from
ethanol (60 mL) to afford pimavanserin tartrate (1) as a
white solid (20.8 g, 89%, based on pimavanserin).
Recrystallization process: crude pimavanserin tartrate
was dissolved by heating in ethanol and then cooled to
ambient temperature and isolated by vacuum filtration to
obtain pure pimavanserin tartrate. M.p. 129–131 oC
(Lit.,9 m.p.133–135 oC). ESI-MS m/z: 428.3 [M+H]+.1H
NMR (400 MHz, DMSO-d6): δ (ppm) 7.24 (dd, J = 8.3,
5.5 Hz, 4H, ArH), 7.14–7.09 (m, 8H, ArH), 6.93 (t, J =
5.7 Hz, 2H, NH), 6.84 (d, J = 8.2 Hz, 4H, ArH), 4.41 (s,
4H, CH2), 4.18 (d, J = 5.6 Hz, 4H, CH2), 4.07-4.04 (m,
2H, CH), 4.01 (m, 2H, CH), 3.70 (d, J = 6.6 Hz, 4H,
CH2), 2.99-2.96 (m, 4H, CH2), 2.36-2.30 (m, 10H, CH2
and CH3), 2.01-1.96 (m, 2H, CH), 1.73-1.66 (m, 4H,
CH2), 1.52–1.48 (m, 4H, CH2), 0.97(d, J = 6.7 Hz, 12H,
References
1. Roberts C. Curr Opin Investig Drugs 2006; 7: 653–660.
2. Meltzer HY, Elkis H, Vanover K, et al. Neuropsychophar-
macology 2010; 35: 881–892.
3. Mikkel T, Nathalic S, Bo-ragnar T, et al. CN102153505 A,
China, 2008.
4. Gant TG and Sarshar S. US20080280886 Al, USA, 2008.
5. Wang SJ, Wang HP and Zhao LS. CN104961672 A, China,
2015.
6. Cheng G. CN105481757 A, China, 2016.
7. Neelakandan K, Manikandan H, Santosha N, et al. Org
Process Res Dev 2013; 17: 981–984.
8. Liu DZ, Tian Z, Yan ZH, et al. Bioorg Med Chem 2013; 21:
2960–2967.
9. Mikkel T, Nathalic S, Bo-ragnar T, et al. CN101035759 A,
China, 2007.