Calcium acamprosate (N-acetylhomotaurine calcium salt, 1) is a well-established drug for the treatment of alcohol dependence. Acamprosate
is thought to stabilize the chemical balance in the brain that would otherwise be
disrupted by alcohol withdrawal.[1] Until it became a generic in the US, calcium acamprosate (under the brand name Campral)
was manufactured and marketed in the US by Forest Laboratories, and by Merck KGaA
in the rest of the world.[2] Legal in Europe since 1989, the FDA approved the drug in July 2004.
Preparation processes for 1 are based on Merck’s original patent[3] that implies the acetylation of homotaurine (2), followed by in situ formation of the calcium salt (Scheme [1a]).[3]
[4]
[5] The chemical procedure performs very well (up to 90% yield), but the final crystallization
drops the overall yield to 30%.
Scheme 1 Current and proposed routes to calcium acamprosate (1)
Preparation of homotaurine, however, presents remarkable drawbacks. A first approach,
starting from 3-aminopropanol (4), involves two reaction steps and the use of HCl(g) is required to obtain the reactive
chloramine 3 which is transformed to homotaurine (2) by sulfonylation.[6] In another approach, starting from 1,3-propanesultone (5), the use of ammonia gas or aqueous solution is required and, under these conditions,
homotaurine (2) is obtained together with the isomeric sulfonamide as byproduct.[7] As a consequence, purification of 2 is mandatory prior to its transformation to 1.
We felt, therefore, that there was ample room for improvement, and we sought to develop
an alternative, more straightforward procedure that would obviate the need of the
intermediate preparation of homotaurine (2) and would allow for the direct precipitation and filtration of the approved crystalline
form of 1. After careful evaluation of several alternatives, we decided to concentrate our
efforts on a previously unknown route, involving the opening of commercially available
1,3-propanesultone (5) with acetamide in the presence of a strong base and isolation of the calcium salt
1 from the reaction medium after in situ cation exchange (Scheme [1b]). This approach has resulted in the development of a direct, one-pot, and operationally
simple procedure for the synthesis of 1, with industrial applicability. Moreover, it also avoids the use of reactive chloramines.
Although this procedure uses the relatively toxic sultone 5, this compound, due to its high reactivity, was expected to react totally under the
set reaction conditions and subsequent workup and, as a consequence, would be not
detected in the isolated calcium acamprosate.
Although there are no direct precedents for the nucleophilic ring opening of γ-sultone
5 in the literature, we hoped that suitable conditions would be found based on previous
reports of the selective N-alkylation of acetamide with primary alkyl bromides[8] or mesylates[9]
[10] that called for the use of relatively strong bases in aprotic solvents (KOH/Al2O3/dioxane[8] or NaH/DMF,[9]
[10] respectively).
Taking this into account, we set out to find the optimal reaction conditions for the
sultone opening step, with regard to the base, solvent, and temperature, by 1H NMR monitoring of the reaction crude (Scheme [2]). Due to the nature of the compounds involved in the process, 1H NMR spectroscopy was the technique allowing a better assessment of the reaction
outcome. All experiments were run with 0.5–1.0 g of 5.
After an extensive exploration of the reaction variables using NaH as base (Table
[1]), we reached the following conclusions: Treatment of 5 with 5 equivalents of acetamide and 5 equivalents of NaH in anhydrous N,N-dimethylformamide (DMF) (14 mL/g of 5) at 0 °C and warming to room temperature in a 2-hour period (entry 1) led to complete
conversion of sultone 5, with the formation of sodium acamprosate as the major product according to 1H NMR analysis. Use of less equivalents of acetamide/NaH (entry 2), preformation of
sodium acetamide (entry 3), or extended reaction time (entry 4) resulted either in
incomplete reaction or in the production of more complex crudes. Tetrahydrofuran and
2-methyltetrahydrofuran were less suitable solvents than DMF (entries 5 and 6).
Scheme 2 Initial exploration of the sultone opening step
Table 1 Sultone Opening Using Sodium Hydride as Basea
|
Entry
|
Base
|
Acetamide
|
Solvent
|
Time (h)
|
Crude analysisb (NMR)
|
|
1
|
NaH (5 equiv)
|
5 equiv
|
DMF
|
2
|
CaAC major
|
|
2
|
NaH (1.8 equiv)
|
2 equiv
|
DMF
|
2
|
complex crude (CaAC minor)
|
|
3
|
NaH (1.8 equiv)
|
2 equiv, preformation
|
DMF
|
2
|
complex crude (CaAC minor)
|
|
4
|
NaH (1.1 equiv)
|
1.2 equiv
|
DMF
|
overnight
|
incomplete reaction
|
|
5
|
NaH (5 equiv)
|
5 equiv
|
THF
|
6
|
complex crude (CaAC minor)
|
|
6
|
NaH (1.8 equiv)
|
2 equiv
|
2-MeTHF
|
overnight
|
complex crude (CaAC minor)
|
a Experiments carried out on 0.5–1.0 g scale.
b CaAC: calcium acamprosate.
Other bases were also tested (Table [2]). The use of 2 equivalents of acetamide and 1.8 equivalents of KOtBu (entry 7), but also NaOtBu (entry 8), in anhydrous DMF (16 mL/g of 5, 0 °C to r.t., 2 h) led to results similar to those of entry 1. Preformation of potassium
acetamide (r.t., 1 h) had a beneficial effect on the purity of the reaction crude.
LDA or sBuLi in DMF can also provide complete conversion of 5, albeit with more complex reaction crudes (entries 9 and 10, respectively). All attempts
with nBu2Mg (entries 11 and 12) or iPrMgCl (entry 13) as the base failed to produce pure reaction crudes under the conditions
necessary for complete sultone conversion. When weaker bases (K2CO3, DIPEA, NaOMe) were used, sultone ring opening did not take place (data not shown).
Table 2 Sultone Opening Using Other Basesa
|
Entry
|
Base
|
Acetamide
|
Solvent
|
Time (h)
|
Crude analysisb (NMR)
|
|
1
|
NaH (5 equiv)
|
5 equiv
|
DMF
|
2
|
CaAC major
|
|
7
|
KOtBu (1.8 equiv)
|
2 equiv, preformation
|
DMF
|
2
|
CaAC major
|
|
8
|
NaOtBu (1.8 equiv)
|
2 equiv, preformation
|
DMF
|
2
|
CaAC major
|
|
9
|
LDA (1.8 equiv)
|
2 equiv
|
DMF
|
2
|
complex crude
|
|
10
|
sBuLi (1.8 equiv)
|
2 equiv
|
DMF
|
2
|
complex crude
|
|
11
|
nBu2Mg (1.8 equiv)
|
2 equiv
|
THF
|
5
|
incomplete reaction
|
|
12
|
nBu2Mg (1.8 equiv)
|
2 equiv
|
DMF
|
5
|
complex crude
|
|
13
|
iPrMgCl (1.8 equiv)
|
2 equiv, preformation
|
THF
|
overnight
|
incomplete reaction
|
a Experiments carried out on 0.5–1.0 g scale.
b CaAC: calcium acamprosate.
With these results in hand, we shifted our attention to the isolation of calcium acamprosate
(1) from the reaction mixture, after cation exchange by addition of CaCl2 at a controlled pH. We selected the conditions of entry 1 for sultone opening for
the optimization of the process. Firstly, we found that when the sultone opening was
performed on a 5 g scale, the equivalents of acetamide and of NaH could be reduced
to 2.0 and 1.8, respectively. After careful addition of water to the cold suspension
to eliminate the excess sodium acetamide and acidification to pH 6, CaCl2 (1.1 equiv) was added in one portion; cation exchange was complete after 1 hour of
stirring at room temperature (a change of the appearance of the suspended solid was
observed). At this point, all of our attempts to precipitate selectively 1 from the reaction mixture were unsuccessful. We were delighted to find, however,
that dilution with 2-propanol (IPA) (16 mL/g 5) allowed us to selectively precipitate NaCl, with a weight content of 1 lower than 2% (most of the theoretical expected amount of NaCl was recovered). This
solid was filtered out and the resulting solution was concentrated under vacuum. The
resulting residue, containing 1, less than 10% weight of acetamide, and 50–60% of solvent (DMF) was treated with
IPA (33 mL/g 5), and seeding with 1 allowed the isolation of 86% of the theoretical amount of crude 1 as a readily filterable solid in the commercial crystalline form [by X-ray powder
diffraction (XRPD); see the Supporting Information]. The sequence of sultone opening
(using 2.0 equiv acetamide and 1.8 equiv NaH) followed by the optimized purification
conditions proved to be reproducible on a 5–10 g scale, affording 70–86% of the theoretical
amount of 1 with a purity range of 65–78% (by quantitative 1H NMR, maleic acid as internal standard) (Scheme [3], Procedure A).
Scheme 3 Calcium acamprosate developed isolation method
While Procedure A provided a robust and a quite reproducible method for obtaining
good yields of crude calcium acamprosate (1), in the commercial crystalline form, there were some concurrent drawbacks that we
still wanted to also avoid to increase the consistency of the procedure:
a) The hazards associated with the use of NaH;
b) The need to add Celite in order to clarify the solution and to remove the mineral
oil prior to the addition of CaCl2;
c) The purity of crude calcium acamprosate (1) was only moderate, and some NaCl was still detected by XRPD.
To that end, we applied the cation-exchange and precipitation conditions from Procedure
A to the entry 7 reaction conditions (sultone opening with KOtBu, Table [2]). We ascertained that NaH could indeed be advantageously replaced by KOtBu, with some methodological modifications:
a) DMF turned out to be the best solvent also for the isolation step, and best results
were obtained with slightly more diluted reaction mixtures (final DMF volume 10 mL/g
5 instead of 8 mL/g 5).
b) NaOtBu was found to be clearly inferior to KOtBu, for the isolation step, and although complete sultone opening can be achieved
with both reagents, NaOtBu use affects both the purity (50–59% vs >95% by quantitative 1H NMR) and the crystallinity of the final product. Moreover, precipitation of NaCl
was not complete, whereas KCl can be almost quantitatively removed by filtration.
After concentration of the mother liquors, addition of IPA, and seeding, crude calcium
acamprosate (1) was obtained as a readily filterable solid in around 74–77% yield, and with a 1H NMR purity higher than 95%. A KCl content between 0.5% and 1.0% (w/w) was also estimated,
by XRPD.
c) The intermediate treatment with Celite (to remove mineral oil from NaH) is avoided
with the use of KOtBu.
This process (Scheme [3], Procedure B) has been successfully run with 25 g of sultone 5, and constitutes a robust, more readily industrializable (with respect to the use
of NaH), and reproducible procedure for the preparation of 1. Table [3] summarizes the main advantages of Procedure B with respect to the initial Procedure
A.
Table 3 Comparison of Procedure A and Procedure B
|
Procedure A
|
Procedure B
|
|
Base
|
NaH
|
KOtBu
|
|
Oil removal
|
required
|
–
|
|
Quenching
|
H2 release, exothermic
|
–
|
|
Salt removal
|
85–87% NaCl recovered
|
90–98% KCl recovered
|
|
Recovery of 1
|
70–86%
|
74–77%
|
|
Purity of 1
|
65–78% by qNMR
|
95–99% by qNMR
|
To achieve a higher calcium acamprosate quality, a further purification step was required
and therefore preliminarily investigated (Table [4]). Crude 1 was conveniently recrystallized by solubilization in water (1 mL/g crude 1, 50–60 °C) followed by addition of an antisolvent such as MeOH (9 mL/g crude 1, stirring overnight at 0–5 °C, and filtration under vacuum; entry 1). The solid was
subsequently washed with cold MeOH/H2O (9:1, 1 mL/g crude 1) and dried overnight at 100 °C under vacuum. In this way, crystalline calcium acamprosate
(1) (>95% purity by quantitative 1H NMR, 98% purity by HPLC, containing less than 0.5% KCl according to XRPD) was obtained
in 73% yield (entry 1). Moreover, in this case, according to the developed HPLC analysis
conditions, acetamide was not detected. Other purification conditions were also tested
(Table [4]): crystallization time (entry 2), temperature (entry 3), different antisolvents
(entries 4–6). However, the final quality of calcium acamprosate was not improved
with respect to that of entry 1.
Table 4 Purification Conditionsa
|
Entry
|
CaAC (weight, HPLCb purity)
|
Antisolventc
|
Temp, time
|
CaAC (1) obtained (weight, HPLCb purity)
|
Yield (%)d
|
|
1
|
3 g, 90%
|
MeOH (9 V)
|
0–5 °C, overnight
|
1.98 g, 98% (acetamide not detected)
|
73
|
|
2
|
0.5 g, 84%
|
MeOH (5 V)
|
0–5 °C, 2 h
|
0.17 g, 95% (0.1% acetamide)
|
40
|
|
3
|
0.5 g, 85%
|
MeOH (5 V)
|
r.t., overnight
|
0.25 g, 96% (0.2% acetamide)
|
59
|
|
4
|
1 g, 78%
|
EtOH (9 V)
|
r.t., overnight
|
0.71 g, 88% (0.2% acetamide)
|
80
|
|
5
|
1 g, 89%
|
MeOH (9 V), MEK (0.5 V)
|
r.t., overnight
|
0.68 g, 97% (0.2% acetamide)
|
76
|
|
6
|
3 g, 90%
|
MeOH (9 V), iBuOAc (0.5 V)
|
r.t., overnight
|
1.74 g, 97% (0.1% acetamide)
|
64
|
a All experiments were carried out using 1 V of hot water to dissolve calcium acamprosate
(CaAC).
b Gemini NX C18, 250 × 4.6 mm, 3 μm, 10 mL potassium dihydrogen phosphate, pH 3/acetonitrile.
Given values corresponding to area%.
c V = mL of solvent per gram of crude material.
d Yield corrected considering purity of starting CaAC.
In summary, we have developed a new, direct, robust procedure suitable for the preparation
of calcium acamprosate (1) on industrial scale. This one-pot methodology is based on the nucleophilic opening
of the available 1,3-propanesultone (5) by potassium acetamidate, followed by cation exchange with Ca2+ ion at controlled pH and selective precipitation of 1 as a readily filterable solid in the commercial crystalline form (74–77% yield, >95%
purity). This new procedure compares favorably with previously described methods both
in terms of number of reaction steps (one-step reaction) and workup procedures (very
simple and suitable for scale-up). Moreover, it also avoids the use of reactive chloramines
and although the relatively toxic sultone 5 is used, due to its high reactivity, sultone 5 is not detected in the isolated calcium acamprosate. Crude 1 can be further purified by recrystallization from water/methanol.
Preparation of Calcium Acamprosate (1) from Sultone 5 and Sodium Acetamide (from Acetamide/Sodium
Hydride); Procedure A
To an ice-cooled solution of acetamide (4.84 g, 82 mmol) in anhydrous DMF (30 mL)
under inert atmosphere, NaH (60% w/w in mineral oil; 2.95 g, 73.7 mmol) was added
in one portion. The mixture was stirred for 1 h at r.t. before cooling (ice bath).
A solution of 1,3-propanesultone (5; 5.0 g, 40.9 mmol) in anhydrous DMF (10 mL) was added dropwise to the cooled mixture.
The mixture was stirred for 2 h at r.t. H2O (4 mL) was carefully added dropwise to the ice-cooled mixture which was stirred
overnight at r.t. Then, the pH was adjusted with 6 M aq HCl (9.3 mL, pH 6.1). Celite
(250 mg) was added and the mixture was stirred for 30 min at r.t. before filtering
(Büchner, Whatman paper, vacuum). The filtrate was washed with cyclohexane (2 × 5
mL) and the aqueous layer was concentrated to dryness (Rotavapor, 45 °C, 25 mbar).
To the resulting oil was added anhydrous CaCl2 (5 g, 45 mmol). The mixture was stirred for 1 h at r.t. and IPA (82 mL) was added
dropwise while cooling with ice. The mixture was stirred for 1 h at 0–5 °C and filtered
(vacuum, filter pore 2). The filtrate was concentrated to an oil and IPA (165 mL)
was added dropwise. The solution was seeded twice with pure calcium acamprosate before
solid was formed. Then, the suspension was stirred overnight at r.t. The mixture was
filtered (vacuum, filter pore 2) and the solid was washed with IPA (10 mL). The wet
solid was dried (100 °C, vacuum, overnight) to obtain crude calcium acamprosate (1); yield: 7.1 g (86%). Quantitative 1H NMR (maleic acid as internal standard) indicated a purity of 70% for this product.
Preparation of Calcium Acamprosate (1) from Sultone (5) and Potassium Acetamidate
(from Acetamide/Potassium tert-Butoxide); Procedure B
To an ice-cold solution of acetamide (24.18 g, 409 mmol) in anhydrous DMF (200 mL)
under inert atmosphere, KOtBu (41.3 g, 368 mmol) was added in one portion (no exotherm was observed). The mixture
was stirred for 1 h at r.t. before cooling (ice bath). A solution of 1,3-propanesultone
(5; 25.0 g, 205 mmol) in anhydrous DMF (50 mL) was added dropwise over 15 min to the
cooled mixture. The resultant dense mixture was vigorously stirred for 4 h at r.t.
H2O (20 mL) and 6 M aq HCl (20 mL) were added to the ice-cooled mixture which was stirred
overnight at r.t. Then, the pH was adjusted with 6 M aq HCl (14.5 mL, pH >3 to <6)
and H2O (8 mL) was added to achieve a final DMF/H2O ratio of 4:1. Anhydrous CaCl2 (25.0 g, 225 mmol) was added and, after stirring for 1 h at r.t., IPA (825 mL) was
added dropwise over 20 min while cooling (ice bath). The mixture was stirred for 1
h at 0–5 °C and filtered (vacuum, filter pore 2). The filtrate was concentrated in
vacuo (distillation at 40–50 °C, 20–24 mbar for 1 h) to an oil before the dropwise
addition of IPA (825 mL) at r.t. over 20 min. The resulting solution was seeded twice
with pure calcium acamprosate before solid was formed. Then, the suspension was stirred
overnight at r.t. The mixture was filtered (vacuum, filter pore 2) and the solid was
washed with IPA (50 mL). The wet solid was dried (100 °C, vacuum, overnight) to obtain
crude calcium acamprosate (1); yield: 31.73 g (77%). Quantitative 1H NMR (maleic acid as internal standard) indicated a purity of >95% for this product.
Recrystallization of Calcium Acamprosate (1)
A stirred aqueous suspension of crude calcium acamprosate (1; 3.0 g, 7.49 mmol, 90% HPLC purity) in H2O (3 mL) was heated to 50–60 °C until a clear solution was obtained. At this point,
MeOH (27 mL) was added dropwise, upon which the formation of an opalescent solution
was observed. After overnight stirring at 0–5 °C (previous seeding with pure 1 is sometimes needed to induce crystallization), the precipitate was collected by
filtration (vacuum, filter pore 2), washed with cold MeOH/H2O (9:1, 3 mL/g), and dried overnight at 100 °C under vacuum to afford calcium acamprosate
(1) of 98% HPLC purity; yield: 1.98 g (66%).
1H NMR (400 MHz, D2O): δ = 3.26 (t, J = 6.8 Hz, 2 H), 2.89 (t, J = 8.0 Hz, 2 H), 1.96 (s, 3 H), 1.90 (m, 2 H).