Manufacture of (5Z,8Z,11Z,13E)(15S)-15-hydroxyeicosa-5,8,11,13-tetraenoic acid sodium salt for clinical trials

Article abstract of DOI:10.1021/op100244n

A robust synthesis of (5Z,8Z,11Z,13E)(15S)-15-hydroxyeicosa-5,8,11,13- tetraenoic acid (15(S)-HETE) sodium salt was established, utilising a biooxidation process. Treatment of arachidonic acid with soybean lipoxidase in 0.1 M sodium tetraborate buffer under oxygen pressure resulted in formation of the hydroperoxide, 15(S)-HPETE. Addition of sodium borohydride to the reaction mixture reduced the hydroperoxide to 15(S)-HETE, which was then purified by column chromatography. 15(S)-HETE sodium salt was prepared by treatment of an ethanol solution of HETE with aqueous sodium hydrogen carbonate. Multiple 10-g batches of 15(S)-HETE sodium salt with >98% enantiomeric excess and >98% chemical purity were prepared to support clinical trials.

Full text of DOI:10.1021/op100244n

Organic Process Research & Development 2011, 15, 301–304
Manufacture of (5Z,8Z,11Z,13E)(15S)-15-Hydroxyeicosa-5,8,11,13-tetraenoic Acid
Sodium Salt for Clinical Trials
Raymond E. Conrow,‡ Paul Harrison,†,§ Mark Jackson,*,† Shaun Jones,† Christel Kronig,† Ian C. Lennon,†,⊥ and
Shaun Simmonds†
Dr. Reddy’s Custom Pharmaceutical SerVices, Chirotech Technology Limited, Unit 162 Cambridge Science Park, Milton Road,
Cambridge CB4 0GH, U.K., and Alcon Laboratories, 6201 South Freeway, Fort Worth, Texas 76134, United States
Abstract:
reaction mixture at the start, allowing in situ reduction of the
hydroperoxide and access to 15(S)-HETE directly. The major
problem of this process is the low volume efficiency (0.5 g/L).
The oxidation can be achieved much more rapidly and with
greater volume efficiency if it is carried out under increased
oxygen pressure, typically 35 psi.4 However, in this procedure
the isolation of the hydroperoxide is very dilute (<1 g/L). The
hydroperoxide was reduced using triphenylphosphine, and
15(S)-HETE was isolated as the methyl ester derivative after
treatment with diazomethane. From our literature assessment
we decided that a biooxidation process would be best, but none
of these methods could be used directly without further
development. Therefore, we set out to devise a modified
biooxidation process to produce batches of 15(S)-HETE sodium
salt.
A robust synthesis of (5Z,8Z,11Z,13E)(15S)-15-hydroxyeicosa-
5,8,11,13-tetraenoic acid (15(S)-HETE) sodium salt was established,
utilising a biooxidation process. Treatment of arachidonic acid with
soybean lipoxidase in 0.1 M sodium tetraborate buffer under
oxygen pressure resulted in formation of the hydroperoxide, 15(S)-
HPETE. Addition of sodium borohydride to the reaction mixture
reduced the hydroperoxide to 15(S)-HETE, which was then
purified by column chromatography. 15(S)-HETE sodium salt was
prepared by treatment of an ethanol solution of HETE with
aqueous sodium hydrogen carbonate. Multiple 10-g batches of
15(S)-HETE sodium salt with >98% enantiomeric excess and
>98% chemical purity were prepared to support clinical trials.
Introduction
Results and Discussion
(5Z,8Z,11Z,13E)(15S)-15-Hydroxyeicosa-5,8,11,13-tetraeno-
ic acid (15(S)-HETE, Icomucret) has been in clinical trials for
the treatment of dry-eye syndrome.1 Chirotech was required to
develop a robust synthesis and manufacture several 10 g batches
of 15(S)-HETE sodium salt to support these clinical trials. Due
to the high potency of this compound, only small volumes of
the active ingredient were required. The sodium salt was
preferred to the acid due to improved stability in solution.
Our initial literature survey showed that 15(S)-HETE can
be prepared by a variety of synthetic organic routes. For
example, Nicolaou has reported a synthesis where the correct
stereochemistry is introduced by the coupling of a single
enantiomer vinyl bromide with a terminal acetylene.2 The
overall yield is a very respectable 25% but due to the number
of steps, 15(S)-HETE is most conveniently prepared biosyn-
thetically from arachidonic acid. In the procedure described by
Baldwin, arachidonic acid was treated with soybean lipoxidase
under one atmosphere of oxygen.3 The intermediate hydroper-
oxide was reduced by treatment with sodium borohydride in
ethanol. Alternatively, sodium borohydride can be added to the
We decided to follow the biooxidation process of Iacazio,4
using 35 psi oxygen pressure. After the biooxidation we
investigated methods to improve the reduction of 15(S)-HPETE
and isolation of 15(S)-HETE. Scheme 1 shows the biooxidation
of arachidonic acid and preparation of 15(S)-HETE sodium salt.
The first small-scale experiments were carried out in a 50-
mL pressure vessel equipped with a glass liner, allowing the
input of 1 mmol of substrate in 10 mL of solvent. The following
variables were investigated to ensure a robust process, capable
of manufacturing several 10-g batches to the same high quality.
Temperature. Temperature was found to be a critical factor,
with more side products observed at higher temperature (10
°C) and a slight drop in enantiomeric excess observed. To ensure
high enantiomeric excess the reaction mixture was cooled to
0-3 °C before commencing the oxidation and maintained below
5 °C throughout the reaction.
Concentration. An attempt was made to increase the
volume efficiency of the reaction, but the reaction at 50 g/L
was slower and gave less pure product. At 20 g/L a relatively
clean reaction was observed. Further dilution would be expected
to give an even cleaner reaction; however, 20 g/L was chosen
so we could manufacture 10-g batches in a 2-L pressure vessel.
This still represented a significant improvement on what had
been achieved previously.
* pjackson@drreddys.com.
† Dr. Reddy’s Custom Pharmaceutical Services.
‡ Alcon Laboratories.
§ Current address: Nice-Pak UK, Aber Park, Flint CH6 5EX, U.K.
⊥ Current address: Chiral Quest Corp. Trinity House, Cambridge Business
Park, Cowley Road, Cambridge CB4 0WZ, U.K.
(1) Schneider, L. W.; Conrow, R. E.; Gamache, D. A.; Pasquine, T.; Yanni,
J. M.; Bhagat, H. G. Chem. Abstr. 2001, 134, 366731. U.S. Patent
6,429,227 B1, 2002.
Reaction Time. The yield was lower when the reaction
mixture was left under oxygen pressure for an extended period.
Polar impurities resulting from further reaction of the hydro-
(2) Nicolaou, K. C.; Ladduwahetty, T.; Elisseou, E. M. J. Chem. Soc., Chem.
Commun. 1985, 1580–1581.
(3) Baldwin, J. E.; Davies, D. I.; Hughes, L.; Gutteridge, N. J. A. J. Chem.
Soc., Perkin Trans. 1 1979, 115–121.
(4) Martini, D.; Buono, G.; Iacazio, G. J. Org. Chem. 1996, 61, 9062–
9064.
10.1021/op100244n  2011 American Chemical Society
Published on Web 12/02/2010
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
301

Scheme 1. Biooxidation of arachidonic acid
peroxide were observed. In order to maximize the yield, it was
necessary to achieve a rapid reaction and a release of the
pressure as soon as the reaction was complete (oxygen uptake
ceased). However, it was important for the reaction to reach
completion since residual arachidonic acid made chromato-
graphic purification of the product difficult, whereas any polar
impurities were easily removed by silica gel chromatography.
As the reaction was scaled up, vigorous stirring was found to
be important in promoting a fast reaction and ensuring all
arachidonic acid was consumed. Carrying out the reaction with
13 wt % soybean lipoxidase at 35 psi oxygen enabled complete
conversion in approximately 25 min. Further increase in
pressure did not have a significant effect.
15(S)-HETE was obtained in approximately 50% yield after
silica gel chromatography, eluting with MTBE/heptanes. The
sodium salt was prepared by treatment with aqueous sodium
bicarbonate in ethanol followed by evaporation to dryness to
give the product as a white powder. To improve stability this
was conveniently stored as a 1% wt/v solution in ethanol. Due
to concerns over stability of the product towards exposure to
light and heat, initial reactions were carried out in amber
glassware, and rotary evaporation was limited to 20 °C.
However, a series of stressing experiments indicated that
processing could be carried out in standard laboratory glassware
and evaporation temperatures raised to 30 °C without detriment
to the product. This facilitated manufacture of the clinical
batches.
Reduction of 15(S)-HPETE. Various ways of carrying out
the reduction of 15(S)-HPETE were investigated. The in situ
reduction with sodium borohydride is not suitable in a pressure
reactor even though this appears to be the method of choice
when carrying out the reaction under one atmosphere of oxygen.
Workup of the hydroperoxide and then treatment with sodium
borohydride in ethanol as described by Baldwin was successful,
but there is a potential risk associated with concentrating the
unstable hydroperoxide, especially as the reaction is scaled up.
Use of triphenylphosphine instead of sodium borohydride
resulted in slower reaction. This made it difficult to reliably
determine the end point of the reaction since a faint positive
test for peroxide could be observed after 2-3 h, whereas in
the case of the borohydride reduction a negative peroxide test
was achieved within 30 min. In addition, isolation of the product
from phosphorus impurities was more troublesome. This method
appears more suitable for preparation of the corresponding ester,
methyl 15(S)-HETE since the chromatographic separation is
less difficult. We found that reduction of the hydroperoxide in
aqueous solution directly after the biooxidation gave the most
convenient procedure. Use of an excess of sodium borohydride
(3.7 equiv), as described in the original in situ method, was
successful, but considerable foaming occurred during workup.
By using one equivalent of sodium borohydride a significantly
easier workup with reduced foaming was developed. Conve-
niently, the reaction could be followed by TLC (MTBE/heptane,
1:1, containing 0.25% acetic acid) and was found to be complete
in 60-90 min. Another benefit of this procedure is that fewer
nonpolar impurities were observed during chromatography.
Identification of Process-Related Impurities. Samples of
15(S)-HETE sodium salt were analysed by liquid chromatog-
raphy (HPLC) for related substances (see Experimental Section,
Table 1, for the HPLC method). Figure 1 shows relative
retention times of potential impurities and lists the amount
present in each of the first five batches manufactured. These
were identified by comparison with appropriate standards. (Note
15-KETE was detected at 260 nm). The main impurities in
batches of 15(S)-HETE produced by this process were found
to be the trans-isomer, triene, and 11-HETE. The 15(R)-isomer
was quantified using chiral SFC (see Experimental Section for
the method).
To conclude, we have demonstrated a robust and reproduc-
ible synthesis of 15(S)-HETE sodium salt that was used to
produce multiple batches of high-quality material, with purity
and enantiomeric excess >98%, to support clinical trials.
Experimental Section
General Procedures. 1H NMR spectra were recorded at 400
MHz (Bruker DPX 400). 13C NMR spectra were recorded at
100 MHz. Chemical shifts (δ) are quoted in ppm, and coupling
constants (J) are given in hertz (Hz). Analytical thin layer
chromatography was performed on Merck silica gel precoated
plates and visualised using ceric ammonium molybdate. Soy-
bean lipoxidase was obtained from Fluka and had an activity
of 150,000 U/mg (where 1U corresponds to the amount of
enzyme that causes an increase in absorbance at 234 nm of
0.001 min -1 at 25 °C and pH 9.0 on linoleic acid substrate).
302
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Figure 1. Impurity profile for five batches of 15(S)-HETE.
(5Z,8Z,11Z,13E)(15S)-15-Hydroxyeicosa-5,8,11,13-tetraeno-
ic Acid. Sodium tetraborate decahydrate (41.32 g, 108 mmol)
and water (1065 mL) were charged to the 2-L vessel. The
mixture was stirred until a clear solution was obtained. The
solution was cooled to 0-3 °C. Arachidonic acid (21.55 g, 70.8
mmol) and soybean lipoxidase (2.83 g) were charged to the
vessel. The vessel was closed, and the mixture was stirred
vigorously. The vessel was charged to 35 psi with oxygen
(vented and refilled three times to obtain an atmosphere of pure
oxygen). The mixture was stirred until uptake of oxygen ceased
(pressure was recharged to 35 psi every 3-5 min and temper-
ature was maintained below 5 °C). Disappearance of arachidonic
acid was confirmed by TLC. The vessel was flushed with
nitrogen and then charged with sodium borohydride (2.70 g,
71.4 mmol). The mixture was stirred at 0-5 °C until TLC
indicated disappearance of the hydroperoxide. The solution was
acidified to pH 3 by charging 2 N sulfuric acid (185 mL). The
vessel contents were transferred to a separating funnel and
extracted with ice-cold MTBE (720 mL and 2 × 360 mL). The
combined organic phases were carefully decanted, and the
residual emulsion was filtered through a pad of Kieselguhr (28.6
g). The product was washed through with MTBE (300 mL).
The combined MTBE solutions were washed with water (2 ×
360 mL) and brine (360 mL). The organic solution was dried
over sodium sulfate (100 g) and then filtered. The filter cake
was washed with MTBE (250 mL). The MTBE solution was
evaporated under reduced pressure (20 °C, final pressure 7
mbar) to give the crude product (28.46 g). The crude product
was applied to a silica column (714 g) and eluted with MTBE/
heptane (1:1 v/v). Fractions containing pure 15(S)-HETE were
combined and concentrated under reduced pressure (20 °C, final
pressure 0.4 mbar). The vacuum was released under nitrogen
to give 15(S)-HETE (11.79 g, 36.8 mmol, 52%) as a colourless
oil.
1H NMR (400 MHz, CD3OD) δ ppm 6.44 (1 H, dd, J 15,
11), 5.89 (1 H, t, J 11), 5.56 (1 H, dd, J 15.5, 7), 5.35-5.20 (5
H, m), 4.00 (1 H, q, J 6), 2.93-2.86 (2 H, m), 2.77-2.73 (2
H, m), 2.20 (2 H, t, J 7.5), 2.06-2.01 (2 H, m), 1.57 (2 H,
quintet, J 7.5), 1.45-1.17 (9 H, m), and 0.81 (3 H, t, J 7).
13C NMR (100 MHz, CD3OD) δ 182.07, 142.60, 135.22,
134.78, 134.47, 134.20, 134.02, 133.33, 130.86, 77.98, 43.07,
38.93, 37.64, 32.20, 31.63, 31.19, 30.95, 30.61, 28.37, and
19.09.
(5Z,8Z,11Z,13E)(15S)-15-Hydroxyeicosa-5,8,11,13-tetraeno-
ic Acid, Sodium Salt. 15(S)-HETE (11.79 g, 36.8 mmol) was
dissolved in absolute ethanol (352 mL). Sodium hydrogen
carbonate (3.08 g, 36.7 mmol) was dissolved in water (79 mL)
and added to the ethanolic 15(S)-HETE solution. The mixture
was stirred until carbon dioxide evolution ceased. The solution
was filtered through a Whatman No. 3 filter paper. The reaction
flask was rinsed, and the filter was washed through with absolute
ethanol (100 mL). The solution was evaporated under reduced
pressure (20 °C, final pressure 0.8 mbar). The vacuum was
released under nitrogen. 15(S)-HETE sodium salt was dissolved
in absolute ethanol (390 mL). The solution was evaporated
under reduced pressure (20 °C, final pressure 0.8 mbar). The
vacuum was released under nitrogen. The solution and evapora-
tion sequence was repeated twice more. 15(S)-HETE sodium
salt was dried under high vacuum (20 °C, 0.7 mbar) until weight
loss was less than 100 mg/h. 15(S)-HETE sodium salt (12.4 g,
36.1 mmol, 98%) was obtained as a white solid. The sodium
salt was dissolved in absolute ethanol (1227 mL) and filtered
through a Whatman No. 3 filter paper. The solution was
transferred to an amber bottle, flushed with nitrogen, and sealed
with a Sure/Seal cap. The product was stored at -20 °C.
ee 98.6% (Chiralpak AD, 90/10 CO2/MeOH, 3.0 mL/min,
3000 psi, 35 °C, 220 nm, retention times R 4.6 min, S 7.7 min).
Purity 98.5% (Luna phenyl hexyl column, acetonitrile/
ammonium acetate/methanol gradient).
ee 98.6% (Chiralpak AD, 90/10 CO2/MeOH, 3.0 mL/min,
3000 psi, 35 °C, 220 nm, retention times R 4.6 min, S 7.7 min).
[R]25 +13.3° (c ) 1.0, EtOH).
D
[R]25 +23.5° (c ) 1.0, EtOH).
νmax (Nujol) 3283, 1563, 1455, 1418, and 1377 cm-1.
1H NMR (400 MHz, CD3OD) δ 6.56 (1 H, dd, J 15, 11),
6.01 (1 H, t, J 11), 5.66 (1 H, dd, J 15, 7), 5.45-5.32 (5 H, m),
D
νmax (film) 3500-2500 (br), 1707, 1457, 1410, and 1237
cm-1.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
303

Table 1. HPLC Method for Related Substances in
15(S)-HETE
4.11 (1 H, q, J 6), 3.04-2.93 (2 H, m), 2.87 (2 H, t, J 6), 2.20
(2 H, t, J 7.5), 2.15 (2 H, q, J 7.5), 1.68 (2 H, quintet, J 7.5),
1.55-1.25 (9 H, m), and 0.93 (3 H, t, J 7).
ThermoFinnigan P4000 pump
UV6000 photodiode array
AS3000 autosampler
13C NMR (100 MHz, CD3OD) δ ppm 183.23, 138.38, 131.38,
131.00, 130.16, 129.78, 129.44, 128.98, 126.63, 73.70, 39.28,
38.82, 33.40, 28.72, 28.23, 27.37, 26.97, 26.71, 24.12, 14.83.
See Table 1:
column
LUNA phenylhexyl, 150 mm × 3.0 mm, 3 µm
temperature 40 °C
eluent A
eluent B
acetonitrile/0.1 M ammonium acetate/methanol
(270:650:80)
acetonitrile/0.1 M ammonium acetate/methanol
(695:100:205)
0.6 mL/min
Acknowledgment
flow rate
detection
injection
retention
We thank Evan P. Kyba and Mark DuPriest of Alcon
Laboratories for support of this work through a collaborative
research program.
235 nm (200-300 nm scanning)
10 µL
15-HETE typically 30-35 min
pump events
time (min)
eluent A%
eluent B%
0
100
0
60
100
0
60.1
70
0
100
70.1
100
0
80
100
0
Received for review September 6, 2010.
OP100244N
0
100
304
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development