Kilogram-scale synthesis of an inhaled corticosteroid

Article abstract of DOI:10.1021/op200257g

The development and implementation of a safe and scalable process for the manufacture of corticosteroid PF-4714224 (1) is described. Initial routes used to synthesise analogues from this series directly from fluocinolone acetonide (2) were unsuitable for large-scale use. Key aspects of the route are the efficient and simple method for the preparation of the steroid tetraol (6), acetal formation by reaction of the tetraol with a bisulphite adduct (12), and isolation of the product by sequential recrystallisations.

Full text of DOI:10.1021/op200257g

Communication
pubs.acs.org/OPRD
Kilogram-Scale Synthesis of an Inhaled Corticosteroid†
Craig J. Knight,‡ David S. Millan,§ Ian B. Moses,‡ Aurelie A. Robin,‡ and Matthew D. Selby*,§,⊥
§
‡Research API and Worldwide Medicinal Chemistry, Pfizer Global Research and Development, Sandwich, Kent CT13 9NJ, U.K.
S
* Supporting Information
chloric acid (7:3 ration of cHCl:H2O, at ambient temperature)
ABSTRACT: The development and implementation of a
safe and scalable process for the manufacture of cortico-
steroid PF-4714224 (1) is described. Initial routes used to
synthesise analogues from this series directly from
fluocinolone acetonide (2) were unsuitable for large-scale
use. Key aspects of the route are the efficient and simple
method for the preparation of the steroid tetraol (6),
acetal formation by reaction of the tetraol with a bisulphite
adduct (12), and isolation of the product by sequential
recrystallisations.
gave the tetraol in good yield (82%) but progress was
hampered by the apparent instability of the tetraol to acidic
reaction conditions and the need to very closely monitor the
reaction progress in order to avoid degradation of the product.
Simple acidic conditions were investigated to directly hydrolyse
the acetonide 2 to the tetraol 6. A range of aqueous acids
(including H2SO4, HCl, and HBF4) and the use of n-butanol, 2-
methyltetrahydrofuran and isopropyl alcohol as cosolvents were
investigated in an effort to achieve a homogeneous reaction
mixture with the potential for a direct-drop isolation. The
apparent instability to acidic reaction conditions was observed
during most of the hydrolysis reactions screened; for example,
the highest conversion to product observed during the
screening process was only 55% when using conc. H2SO4 (10
equiv) in 50% aqueous isopropyl alcohol.
Efficient conversion to product was achieved by heteroge-
neous acetonide hydrolysis in 48% aqueous HBF4;5 under these
conditions both starting material and product were poorly
soluble, but conversion proceeded smoothly, and the desired
tetraol was provided without the need for exclusion of oxygen.
Dilution with water upon completion of the reaction and
collection of the product by filtration followed by washing with
ethyl acetate gave clean tetraol 6 in good yield (Scheme 3).
Further optimisation of this step allowed reaction volumes to
be driven down to 5 mL per gram of feed acetal. In contrast to
the hydrochloric acid conditions previously employed, this
method gave a very stable reaction mixture with no significant
degradation even when workup was delayed for several days.
Optimisation of the acetal-forming reaction between
fluocinolone tetraol 6 and aldehyde 3 (used as a representative
example of this series) identified the use of 3 or more
equivalents of trifluoromethanesulphonic (triflic) acid, in the
presence of dried magnesium sulphate, in dioxane, as improved
conditions for this step (Scheme 4).6 Weaker acids such as
trifluoroacetic or sulphuric acid failed to give product under any
conditions investigated; conversion to product was only
observed when using very strong acids. Additionally, the
presence of a drying agent such as dried sodium sulphate or
magnesium sulphate was found to be essential; the use of triflic
acid in the absence of drying agent gave very low conversion
and a very messy reaction profile. Whilst dioxane is not the
ideal solvent for large-scale chemistry, due to safety concerns,
these reaction conditions were workable for multigram-scale
preparation of 1 and analogues of 1. The use of sodium or
magnesium sulphate as a solid drying agent additionally served
to maintain a mobile reaction mixture, replacing sand and
INTRODUCTION
■
Inhaled corticosteroids, such as fluticasone propionate, are
widely used as treatments for asthma.1 As part of a project to
identify agents with an improved therapeutic index, a series of
acetals of fluocinolone were investigated of which PF-4714224
(1)2 was selected for preclinical development requiring a
kilogram-scale supply of material.
The medicinal chemistry route to early analogues proceeded
directly from commercially available fluocinolone acetonide (2)
by transacetalisation, catalysed by perchloric acid3 (Scheme 1).
Whilst suitable for rapid analogue provision on small scale, this
method was unsuitable for larger-scale use for a number of
reasons; the addition of sand was required to prevent the
products from forming intractable gums under reaction
conditions, the use of perchloric acid, a powerful oxidizing
agent, represented an unacceptable safety hazard on scale, and
the isolated yield was low. In addition, considerable difficulties
were presented in purifying a poorly soluble, crude reaction
product containing both epimers of the product and the
acetonide starting material 2.
DISCUSSION
■
Reaction of aldehydes with the steroid tetraol fluocinolone (6)
was found to proceed smoothly and to give a cleaner reaction
profile (Scheme 2) than reaction with the acetonide 2, giving
the product as a mixture of α- and β-epimers. The epimers
could be separated by chromatography, and any residual tetraol
was more readily separated by chromatography than the
acetonide.
The reported route to the tetraol 6 uses formic acid to
solvolyse the acetonide 2 to a mixture of formate esters which
are then cleaved by methanolic potassium hydroxide4 (Scheme
3). However, the need for strict exclusion of oxygen from the
second step (to avoid oxidative cleavage of the hydroxyl-ketone
moiety to a carboxylate), led the process team to explore
alternatives to this method. Hydrolysis using dilute hydro-
Received: September 18, 2011
Published: March 14, 2012
© 2012 American Chemical Society
697
dx.doi.org/10.1021/op200257g | Org. Process Res. Dev. 2012, 16, 697−703

Organic Process Research & Development
Communication
a
Scheme 1. Transacetalisation route to 4
a
Reagents and conditions: HClO4, PhMe, sand, ambient temperature, 18 h, 44%.
a
Scheme 2. Preparation of 4 by reaction of aldehyde and tetraol
a
Reagents and conditions: HClO4, DCM, ambient temperature, 36%.
a
Scheme 3. Routes to steroid tetraol 6
molecule, to identify features of the route likely to cause
difficulties and safety concerns upon scale-up.
PF-4714224 (1) was identified as the project lead and
subsequently as a development candidate. The aldehyde
fragment (10) required for the preparation of 1 was initially
prepared by the homogeneous reaction of commercially
available 4-fluorobenzaldehyde with the known 3-methylthio-
thiophenol (8),7 itself prepared by the reported alkylation of
1,3-benzenedithiol (7) with dimethyl sulphate (Scheme 5).
Acetal formation to give 1 by reaction of 6 and aldehyde 10
using the project conditions and reaction in dioxane in the
presence of triflic acid and magnesium sulphate gave the
product in 18% yield following an unoptimised isolation
procedure (Scheme 6).
Further optimisation of the acetal formation was carried out
by automated screening8 of a range of solvents and of reagent
stoichiometries, using the aldehyde 10 and tetraol 6 (Table 1).
Acetonitrile was identified as a suitable reaction solvent, being
less hazardous than dioxane and giving equivalent results.
Substituting dimethoxyethane for acetonitrile gave similar
results but a range of other solvents screened (including
tetrahydrofuran, 2-methyltetrahydrofuran, propionitrile, N-
methylpyrrolidinone and dimethylsulphoxide) gave either
poorer conversion or a more complicated reaction profile. A
small excess of aldehyde was found to be required for full
conversion of the tetraol to product, no improvement in
a
Reagents and conditions: (a) (i) 60% HCO2H(aq), reflux; (ii) KOH,
MeOH, ambient temperature, 63% (over two steps); (b) 48%
HBF4 (aq), ambient temperature, 94%.
avoiding the scouring action on laboratory glassware observed
when sand was used. With the use of these reaction conditions,
a number of analogues of 1 were prepared;2 in each case the
crude reaction mixture contained both epimers with α-epimer
predominating.
Efforts were made to move towards chemical routes suitable
for large-scale use prior to candidate nomination arising from
close collaboration between the medicinal chemistry and
research-API (process chemistry) groups. In light of the need
to progress the project rapidly to clinical assessment, work was
carried out, prior to the identification of the candidate
a
Scheme 4. Improved conditions for preparation of 4 by reaction of aldehyde and tetraol
a
Reagents and conditions: CF3SO3H, dioxane, MgSO4, ambient temperature, 46%.
698
dx.doi.org/10.1021/op200257g | Org. Process Res. Dev. 2012, 16, 697−703

Organic Process Research & Development
Communication
a
Scheme 5. Preparation of aldehyde 10, medicinal chemistry route
a
Reagents and conditions: (a) Me2SO4, H2O, 0 °C (used without isolation); (b) K2CO3, MeCN, ambient temperature, 86%.
a
Scheme 6. Preparation of 1 by medicinal chemistry route
a
Reagents and conditions: CF3SO3H, dioxane, MgSO4, ambient temperature, 1 isolated in 18% yield (note, crude reaction product was a mixture of
1:11 in ration of approximately 87:13).
Table 1. Optimisation of acetal formation (reaction of
aldehyde 10 with tetraol 6)
reaction rate or profile was observed when increasing this
beyond 1.3 equiv. The quantity of triflic acid used was varied
between 1 and 5 equiv. When using less than 3 equiv of acid the
reaction slowed significantly, taking 4−5 h to reach maximum
conversion, but the quantity of product produced did not
match that generated when a larger excess was used. The
quantity of magnesium sulphate used was found to be less
critical with 3 or 5 equiv performing equally well. Given the low
cost and hazards associated with this reagent and the ease of its
removal during workup, no further variation of this quantity
was explored. When applied on gram-scale, the optimized
reaction conditions employing 3 equiv of trifilic acid, 5 equiv of
magnesium sulphate, and 1.3 equiv of aldehyde 10 relative to
tetraol 2, in acetonitrile at ambient temperature, yielded the
product 1 in 32% yield after 1 h reaction time.
aldehyde
(equiv)
MgSO4
(equiv)
TfOH
(equiv)
temp
(°C)
conversion
(%)
a
solvent
1.0
1.2
1.5
1.2
1.2
1.2
1.2
1.2
1.2
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
3
3
3
5
5
5
5
5
5
5
5
5
1
2
3
4
5
5
20
20
20
20
20
20
20
60
20
98
100
99
28
52
66
95
58
100
a
Reactions were sampled after 2 h. Reaction screening was carried out
on 20-mg scale and monitored by HPLC.
Throughout this project, the acetal formations, using a range
of aldehydes, proceeded to give the product as a mixture of
Scheme 7. Final route to 1
699
dx.doi.org/10.1021/op200257g | Org. Process Res. Dev. 2012, 16, 697−703

Organic Process Research & Development
Communication
epimers, with the desired α-epimer predominating. The
optimised conditions for the preparation of 1 gave a crude
87:13 mixture of isomers (α:β),9 a ratio which was not
improved upon despite considerable variation of reaction
conditions. The use of dimethoxyethane or dioxane in place of
acetonitrile gave almost identical ratios of epimeric products
when compared by HPLC analysis of the crude reaction
mixtures.
Extending the reaction time beyond 70 min led to
degradation of the product under reaction conditions. Thus
prompt analysis of the reaction mixture and response to the
results were crucial to success upon scale-up. Rapid degradation
of the reaction mixture was also observed when the reaction
was heated; thus, it was vital to maintain the reaction vessel
temperature below 25 °C during the addition of triflic acid.
Attempts to equilibrate the isolated β-epimer (11) to a mixture
of α- and β-epimers by resubmitting crude β-epimer to the
reaction conditions gave significant degradation and very little
observed conversion to the desired product.
crystallisation from ethyl acetate yielded material sufficient for
early, medicinal chemistry, requirements. The solvent volumes
required for workup (∼40 mL per gram), however, made this
impractical for use on large scale. Initial attempts to extract the
product, following water quench of the acetonitrile reaction
mixture and using methyl-THF, were hampered by the
formation of emulsions and the requirement for large extraction
volumes to give phase separations. Measurement of solubility in
a range of organic solvents showed an unexpected trend
towards higher solubility in the heavier ester solvents (Table 2).
a
Table 2. Solubility of 1 in a range of solvents
solvent
solubility (g/L)
solvent
solubility (g/L)
Me-THF
MeCN
49.4
5.4
MeOAc
EtOAc
3.1
2.4
2-butanone
toluene
5.4
i-PrOAc
n-BuOAc
6.7
14.1
>100
a
Solubility measured by HPLC at ambient temperature.
Telescoping the alkylation and arylation steps gave the
aldehyde 10 as a crude solution in ethyl acetate. A simple,
homogeneous procedure was employed, using potassium
hydroxide and dimethyl sulphate to alkylate benzene-1,3-thiol
7, followed by extraction into tert-butylmethylether. During the
development of this process it was found that too slow a rate of
addition of base led to greater levels of bis-methylation; thus, it
is important to keep the addition time to a minimum.
Following solvent swap into acetonitrile, arylation was effected
by reaction with fluorobenzaldehyde in the presence of
tetramethyl guanidine. Sparging of the reaction solution with
nitrogen prior to the arylation step was found to be vital to
avoid formation of the symmetrical disulphide impurity which
was readily formed by the thiol under basic conditions. The
aldehyde 9 had initially been isolated as an oil; however, this
was found to be vulnerable to oxidation upon storage. To
circumvent this and to allow the isolation of a convenient, solid
intermediate, the bisulphite adduct (12) was prepared directly
from the crude solution of aldehyde. Following aqueous
workup and solvent exchange into acetonitrile, treatment with
aqueous sodium metabisulphite gave the solid bisulphite adduct
which was collected by filtration and washed with water and
acetonitrile (Scheme 7).
Fortuitously, the bisulphite adduct 12 was found to react
cleanly with fluocinolone tetraol 6 without any need to convert
to the aldehyde.10 The optimal reaction profile was achieved by
premixing the tetraol with magnesium sulphate for several
hours prior to the addition of bisulphite adduct and
trifluoromethyl sulphonic acid. This was believed to be due
to the slow dissolution of the tetraol under reaction conditions;
premixing the reactants allowed the very granular tetraol to be
broken down into smaller particles, allowing rapid dissolution
as the reaction progressed. Attempts to improve the ratio of
isomers by carrying out the acetal formation at lower
temperature gave a more complex mixture of products, possibly
due to the low solubility of the tetraol. Applying the previously
optimized reaction conditions (3 equiv of triflic acid, 5 equiv of
magnesium sulphate, and 1.3 equiv of bisulphite adduct, in
acetonitrile at ambient temperature) gave product 1 in 37%
yield, with a ratio of epimers in the crude reaction mixture
identical to that observed when employing the aldehyde (α:β,
87:13).
The use of n-butyl acetate to extract the crude product from an
aqueous workup allowed the use of moderate total volumes and
gave clean phase splits. Thus, following complete reaction as
determined by HPLC, the reaction mixture was diluted with n-
butyl acetate and washed with water, sodium bicarbonate
solution, and finally water before filtration and partial
concentration.
Early batches of 1 recrystallised from ethyl acetate or
isopropyl alcohol were found to be stable solvates resistant to
vacuum drying but recrystallisation from aqueous ethanol gave
nonsolvated material; this was found to be a mixture of
polymorphs but, nevertheless, was suitable for precandidate
studies. Material recrystallised from acetonitrile could be dried
in vacuo or at elevated temperature (>125 °C) to give a
homogeneous form with a melting point of 190 °C. This form
was subsequently found to be a desolvated solvate but was
considered suitable for development purposes, possessing
suitable physical properties for inhaled dosing following
micronisiation by jet milling. Harsh conditions (jet mill at 7.5
bar pressure) were required to reduce the particle size to that
required for inhalation use, but the resulting API was stable at
70 °C and 75% relative humidity without any observed increase
in particle size from agglomeration or recrystallisation.
Following the identification of PF4714224 1 as a develop-
ment candidate, an isolation process was developed, avoiding
chromatography and making use of the inherently high
crystallinity of the steroid product. The crude n-butyl acetate
solution of product was partially concentrated and triturated
with 2-butanone to give crude product, purging it of residual
tetraol and unreacted aldehyde. The removal of n-butyl acetate
by codistillation with 2-butanone was found to be rather
inefficient, giving an azeotrope of 8:1 2-butanone:n-butyl
acetate. Hence, the concentration of the crude n-butyl acetate
extract by distillation was crucial to the successful isolation of
the product. Recrystallisation of the crude solid from ethanol
was found to remove the undesired β-epimer (reducing levels
of β-epimer to <2%), affording a granular solid, which could be
subsequently recrystallised from acetontrile to give the product
in the desired crystalline form. Recrystallisation of the crude
product resulting from 2-butanone trituration using acetonitrile
failed to purge the β-epimer. The precipitation of the crude
product following the addition of MEK was found to occur
Isolation of product from the acetal formation by aqueous
workup using ethyl acetate, followed by chromatography and
700
dx.doi.org/10.1021/op200257g | Org. Process Res. Dev. 2012, 16, 697−703

Organic Process Research & Development
Communication
reliably without seeding, as were the intermediate and final
recrystallisations.
(6α,11β,16α)-6,9-Difluoro-11,16,17,21-tetrahydroxy-
pregna-1,4-diene-3,20-dione (Fluocinolone Tetraol) (2).
A suspension of fluocinolone acetonide (5.39 kg; 11.91 mol) in
tetrafluoroboric acid (48% aq, 37.2 kg) was stirred at 22 °C for
2 h. The suspension was then diluted with water (27 kg) and
stirred at 22 °C for 1 h before the solid was collected by
filtration, washing with water (15 kg then 30 kg) and ethyl
acetate (24.3 kg) to give the product as white solid (4.64 kg,
94.5%) after drying under vacuum at 45 °C.
SAFETY CONSIDERATIONS
■
The very high potency of the product 1 and, consequently, the
very low predicted dose size (<100 μg) led to concerns about
worker safety in handling these products; glucocorticosteroid
agonists are, as a class, known to be potential teratogens and to
display dose-limiting systemic side effects when administered
by inhalation or orally. Early assessment of the potential
hazards presented by 1 led to the assignment of an occupational
exposure limit of <1 μg/m3. As a consequence, special
precautions were taken to minimise exposure to the API; the
use of over sleeves and disposable lab coats, worn over regular
lab coats and gloves, coupled with a rigorous approach to
occupational hygiene was deemed sufficient to ensure worker
safety during gram-scale work. When planning the 100-g and
kilogram-scale campaigns, it was considered necessary to also
minimise the operational complexity of the isolation procedure
and to handle intermediate solids as solvent-damp filter cakes
to reduce exposure to airborne powders. Both the ethanol and
acetonitrile recrystallization steps were found to be tolerant of a
significant degree of solvent carry-over from the previous
isolations. Deliberate spiking of each of these recrystallisations
with 5% v/v of 2-butanone or ethanol, respectively, gave no
significant loss of yield, purity, or crystalline form.
1H NMR (400 MHz, DMSO-d6) δ: 0. 82 (s, 3H), 1.46 (s,
3H), 1.31−1.51 (m, 3H), 1.77−1.87 (m, 1H), 2.08−2.33 (m,
3H), 2.37−2.46 (m, 1H), 4.08 (dd, J = 12 and 2 Hz, 1H),
4.09−4.16 (m, 1H),4.48 (dd, J = 12 and 3 Hz, 1H), 4.63 (br,
1H), 4.75 (dd, J = Hz, J = Hz, 1H), 5.35 (d, J = 5 Hz, 1H),
5.51−5.69 (m, 1H), 6.08 (s, 1H), 6.26 (dd, J = 10 and 2 Hz,
1H), 7.24 (d, J = 10 Hz, 1H). 19F NMR (400 MHz, DMSO-d6)
δ: −164.55 (dd J = 9 and 32 Hz), −186.39 (dd, J = 49 and 16
Hz). LRMS (ESI): m/z 411 [M − H]−.
3-(Methylthio)benzenethiol (8). A solution of benzene-
1,3-dithiol (1.6 kg; 11 mol) in 2-methyltetrahydrofuran (10.1
kg) was treated with dimethylsulphate (1.42 kg; 11.3 mol),
followed by a line wash of 2-methyltetrahydrofuran (0.5 kg)
and cooling to 2 °C. The mixture was treated with a solution of
2 M sodium hydroxide (1.35 kg 40% in 5.8 kg water), over 25
min (addition time should not exceed 40 min), whilst
maintaining the vessel temperature below 15 °C. A line wash
of 2-methyltetrahydrofuran (0.6 kg) was added, and the
mixture was heated to 50 °C for 3 h before cooling to 20
°C. The mixture was diluted with tert-butylmethylether (11.8
kg) and stirred for 20 min, and the layers were separated. The
organic layer was treated with 2 M sodium hydroxide solution
(7.15 kg) and stirred for 10 min; the layers were separated, and
the organic phase was treated as waste. The combined aqueous
layers were treated with 6 M HCl (17.6 kg), whilst maintaining
the temperature below 25 °C, and extracted with tert-
butylmethylether (two portions of 5.9 kg). The two tert-
butylmethylether extracts were combined and washed with
water (16 kg) to give a solution of the product (1.33 kg, 11%
solution in tert-butylmethylether; 75%). The product was found
by HPLC to be identical to material prepared by the method of
Rumpf.5 1H NMR (400 MHz, CDCl3) δ: 2.47 (s, 3H), 3.45 (s,
H), 7.03 (m, 2H), 7.12−7.16 (m, 2H).
Sodium Hydroxyl(4-{[3-(methylthio)phenyl]thio}-
phenyl)methanesulphonate (12). A solution of 3-(methyl-
thio)-benzenethiol 7 (1.33 kg; 8.51 mol) in TBME (11.80 kg)
was charged and distilled to a volume of 9 L, maintaining the
temperature below 90 °C to remove TBME, and acetonitrile
(10.5 kg) was added. The solution was distilled at atmospheric
pressure to give a volume of 9 L, and acetonitrile (10.5 kg) was
added. The solution was distilled a third time to give a final
volume of 13.5 L; this solution was cooled to 10 °C and
sparged with nitrogen for 1 h. This solution was treated with 4-
fluorobenzaldehyde (1.06 kg; 8.55 mol) followed by a line wash
of acetonitrile (0.79 kg). 1,1′,3′3′-Tetramethyl guanidine was
added slowly (1.08 kg; 9.3 mol) with stirring, followed by a line
wash of acetonitrile (0.79 kg). The reaction mixture was heated
to 50 °C for 16 h, then cooled to 20 °C and treated with ethyl
acetate (12 kg) followed by 2 M HCl (6.68 kg) and a line wash
of ethyl acetate (0.6 kg). The mixture was stirred for 20 min,
and then the layers were separated. The organic phase was
washed successively with 1 M sodium bicarbonate solution
(14.42 kg) and brine (1.73 kg NaCl in 6.7 kg water). The
organic solution was treated with acetonitrile (3.9 kg) and
A batch of 2.56 kg of PF4714224 1 was prepared in 43.5%
yield and 97.9% purity by the method outlined in Scheme 7.
CONCLUSION
■
A concise synthesis of PF-4714224 1 was achieved in four linear
steps from commercially available starting materials (Scheme
7). Fluocinolone tetraol 6 was prepared by an operationally
simple HBF4 hydrolysis of fluocinolone acetonide 2, in a
significant improvement over the method reported in the
literature. The aldehyde 10 was prepared by monomethylation
of benzene-1,3-dithiol and subsequent reaction with fluoro-
benzaldehyde. The bisulphite adduct 12 was found to be a
convenient, solid surrogate for the aldehyde 10 in this route.
The inherent crystallinity of 1 was used to isolate the product
from the undesired β-epimer 11, which was inevitably formed
as a byproduct of the acetal formation. The potential hazards
represented by this exquisitely potent steroid were managed by
careful design of the purification and isolation procedure to
minimise exposure to dry solids by deliberately handling
intermediate filter cakes whilst solvent-damp.
EXPERIMENTAL SECTION
■
1H NMR spectra were recorded on a Bruker 400 MHz NMR
spectrometer. HPLC analyses were performed using a reverse
phase technique. LC/MS analysis was performed using the
following system; Hewlett-Packard 1100 with SB C18 3.0 mm
×50 mm, 1.8 μm particles; mobile phase consisting of solvent
A, 0.05% TFA in water, solvent B, 0.05% TFA in acetonitrile; 0
min = 5% solvent B; 3.5 min = 100% solvent B; 4.5 min =
100% solvent B; 4.6 min = 5% solvent B; run time 5 min;
column temperature 50 °C; 225 nm; with Waters Micromass
ZQ 2000/4000 mass detector. Combustion analyses were
performed by Warwick Analytical Service, University of
Warwick Science Park, The Venture Centre, Sir William
Lyons Road, Coventry CV4 7EZ, U.K. Experimental data are
provided for the largest batch produced.
701
dx.doi.org/10.1021/op200257g | Org. Process Res. Dev. 2012, 16, 697−703

Organic Process Research & Development
Communication
distilled at atmospheric pressure to 10-L volume, diluted with
acetonitrile (11.8 kg), and concentrated by distillation at
atmospheric pressure to a volume of 10 L. The resulting
solution was diluted with acetonitrile (11.8 kg) and
concentrated by distillation at atmospheric pressure to a
volume of 15 L before cooling to 22 °C (solvent composition
found to be approximately 0.95% ethyl acetate). The solution
was treated with sodium metabisulphite (1.88 kg) in water (15
kg) and stirred at 22 °C for 48 h before the solid was isolated
by filtration. The crude solid was washed with water (two
portions of 12.8 kg) and acetonitrile (two portions of 10.1 kg)
to give the product bisulphite adduct as a white solid (2.18 kg;
70%) after drying under vacuum at 50 °C.
was cooled to 20 °C over 6 h and the resulting solid collected
by filtration. The solid was then washed with acetonitrile (two
portions of 5.5 kg) and dried in vacuum at 50 °C to give pure
acetal 1 as a white solid (2.56 kg; 46%). Mp 189.9 °C; 1H NMR
(400 MHz, DMSO-d6) δ: 0.85 (s, 3H), 1.43−1.53 (m, 1H),
1.48 (s, 3H), 1.65−1.71 (m, 3H), 1.97−2.06 (m, 1H), 2. 14−
2.31 (m, 2H), 2.41 (s, 3H), 2.55−2.67 (m, 1H), 4.16−4.22 (m,
2H), 4.52 (dd, J = 6 and 19 Hz, 1H), 4. 95 (d, J = 5 Hz, 1H),
5.06 (t, J = 6 Hz, 1H), 5.48 (s, 1H), 5.50−5.51(m, 1H), 5.54−
5.79 (m, 1H), 6.10 (s, 1H), 6.27 (dd, J = 10 and 2 Hz, 1H),
7.06 (dt, J = 8 and 2 Hz, 1H), 7.18−7.20 (m, 2H), 7.23−7.26
(m, 1H), 7.27−7.29 (m, 1H), 7.32 (d, J = 8 Hz, 2H), 7.42 (d, J
= 8 Hz, 2H). 13C NMR (100 MHz, DMSO-d6) δ: 14.18, 16.19,
22.25, 30.90, 32.12, 33.57, 35.58, 42.69, 45.17, 47.91, 65.77,
70.09, 81.04, 86.50, 97.44, 99.50, 102.09, 119.24, 124.90,
127.36, 127.57, 127.65, 128.72, 129.79, 129.88, 134.81, 134.92,
136.93, 140.03, 151.39, 162.60, 184.54, 208.66. LRMS (ESI):
m/z 655 [M + H]+. 19F NMR (400 MHz, DMSO-d6) δ:
−164.88 (dd, J = 49 and 15 Hz), −186.50 (dd, J = 28 and 9
Hz). HRMS ESI positive; m/z 655.2002 (C25H37F2O6S2,
theoretical 655.199962, −1.20), 677.1809 (C35H36F2NO6S2,
theoretical 677.181906, 0.70). Anal. Calcd For C35H36F2S2O6:
C, 64.2%; H, 5.53%; N, 0.00%. Found: C, 64.15%: H, 5.53%; N,
0.04%.
1H NMR (400 MHz, DMSO-d6) δ: 2.41 (s, 3H), 4.97 (d, J =
6 Hz, 1H), 5.90 (d, J = 6 Hz, 1H), 6.98 (m, 1H), 7.12 (m, 2H),
7.26 (m, 3H), 7.46 (d, J = 10 Hz, 2H). 13C NMR (100 MHz,
DMSO-d6) δ: 15.25, 85.19, 124.98, 126.65, 126.91, 129.71,
130.47, 131.17, 132.16, 137.59, 140.35, 140.54. HRMS (ESI)
m/z 283.0226 (C14H12NaOS2, theoretical 283.022726, −1.60),
261.0407 (C14H13OS2, theoretical 261.040781, −1.80).
(4aS,4bR,5S,6aS,6bS,8R,9aR,10aS,10bS,12S)-4b,12-Di-
fluoro-6b-glycoloyl-5-hydroxy-4a,6adimethyl-8-(4-{[3-
(methylthio)phenyl]thio}phenyl)-4a,4b,5,6,6a,6-
b,9a,10,10a,10b,11,12-dodecahydro-2H-naphtho-
[2′,1′:4,5]indeno[1,2-d][1,3]dioxol-2-one (1). A suspension
of magnesium sulphate (3.7 kg; 30.8 mol) and fluocinolone
tetraol 6 (3.5 kg; 8.5 mol) in acetonitrile (26.7 kg) was stirred
at 18 °C for 12 h. Bisulphite adduct 12 (3.63 kg; 9.96 mol) was
charged to the vessel followed by trifluoromethanesulphonic
acid (6.37 kg; 42.4 mol) whilst maintaining the vessel
temperature below 25 °C (critical process parameter), and
the mixture was stirred for 70 min at 22 °C. n-Butyl acetate
(20.8 kg) followed by water (35 kg) was added, and the mixture
was stirred for 40 min. The aqueous phase was removed, and
the organic phase was washed with water (35 kg) followed by
sodium bicarbonate solution (two portions of 1.57 kg in 17.5 kg
water) and water (17.5 kg) before being passed through an
inline ceramic filter which was subsequently washed with n-
butyl acetate (6.2 kg). The combined organic solution was
concentrated by distillation at reduced pressure (−0.940 barg,
pot temperature of 29.5 °C) to a volume of approximately 10 L,
cooled to 20 °C, treated with 2-butanone (24.7 kg), and aged
for 11 h. The crude solid was collected by filtration (visual
inspection to check for presence of solid precipitate 2 h
prior to filtration stage is a critical process parameter) and
washed with 2-butanone (two portions of 8.45 kg); whilst still
damp the solid was transferred to a reaction vessel which had
been charged with ethanol (55.2 kg); the resulting suspension
was heated to reflux until a clear solution had formed. The
solution was cooled to 20 °C over 20 min and stirred at this
temperature for 2 h before collection of the resulting solid by
filtration. The solid was then washed with ethanol (two
portions of 13.8 kg) and acetonitrile (5.3 kg). The solid was
transferred, whilst still damp, to a reaction vessel which had
been charged with acetonitrile (24.8 kg) and the resulting
solution concentrated by distillation at reduced pressure
(−0.925 g, pot temperature 30 °C) to a volume of
approximately 11 L. The solution was treated with acetonitrile
(8.5 kg) and heated to reflux, and two further portions of
acetonitrile (two portions of 1.73 kg) were added whilst
maintaining the mixture at reflux, until a clear solution was
obtained (the visual observation of a clear solution at this
stage is a critical process parameter). The resulting solution
ASSOCIATED CONTENT
■
S
* Supporting Information
Proton and carbon NMR data and assignments of 1 and epimer
11. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Telephone: +44 0 1753 534 655. E-mail: matt.selby@ucb.
com.
Present Address
⊥UCB, 216 Bath Road, Slough SL1 4EN, United Kingdom.
Notes
The authors declare no competing financial interest.
†The work described was conducted within Research-API and
Worldwide Medicinal Chemistry at Pfizer Global Research and
Developments Laboratories, Sandwich.
ACKNOWLEDGMENTS
■
We thank Carol Bains, Jonathan Fray, Steven Fussell, Christian
Ngono, and James R. Sayer for contributions to the chemistry
and Asayuki Kamatani, Matthew Morland, and Paul Glynn for
gram-lab and kilo-lab support. We also thank Michael
Hawksworth for process safety work, Edouard Guillabert and
Toby A. Payne-Cook for material sciences work, and Derek
Bradley, Catriona Thom, Kerry Paradowski, Simone Gabriele,
and Emma Stubbs for analytical support.
REFERENCES
(1) Barnes, P. J. Pharmaceuticals 2010, 3, 514.
■
(2) (a) Glossop, P. A.; Millan, D. S.; Price, D. A. WO Patent 2009/
069032 A1, 2009. (b) Millan, D. S.; Ballard, S. A.; Chunn, S;
Duybowski, J. A.; Fulton, C. K.; Glossop, P. A.; Guillabert, E.; Hewson,
C. A.; Jones, R. M.; Lamb, D. J.; Napier, C. M.; Payne-Cook, T. A.;
Renery, E. R.; Selby, M. D.; Tutt, M. F.; Yeadon, M. Bioorg. Med.
Chem. Lett. 2011, 21, 5826.
702
dx.doi.org/10.1021/op200257g | Org. Process Res. Dev. 2012, 16, 697−703

Organic Process Research & Development
Communication
(3) Jakupovic, E.; Sourander, R. T.; Stenhede, J. U.; Svensson, R. P.
European Patent 1988/262108 A1, 1998. Parkinson, N. C.; Van Sickle,
A. P.; WO Patent 1996/9616978 A1, 1996.
(4) Mills, J. S.; Bowers, A.; Djerassi, C.; Ringold, H. J. J. Am. Chem.
Soc. 1960, 82, 3399.
(5) Rainer, A.; Dax, K.; Plenschko, R.; Stiltz, A. E. Carbohydr. Res.
1985, 137, 282.
(6) Morrell, A. I. In New Synthetic Technologies in Medicinal
Chemistry; Farrant, E., Ed.; Royal Society of Chemistry Publishing,
Cambridge; 2012; p 42.
(7) Rumpf, P. Bull. Soc. Chim. Fr. 1940, 7, 632.
(8) Reaction of aldehyde (8) with fluocinolone acetonide (2) under
the optimized acetal formation conditions (MeCN, TfOH, MgSO4,
ambient temperature) gave a mixture of product and unreacted
acetonide which proved to be a more challenging mixture from which
to isolate the desired product than that resulting from reaction with
the tetraol (6).
(9) The structures of the α- and β-epimers were assigned following
1
full H and 13C NMR assignment. See Supporting Information. For a
full discussion of the characteristic NMR spectra of a closely related
pair of epimers, see Thalen, A Acta Pharm. Suec. 1987, 24, 97.
(10) The direct, perchloric acid-catalyzed reaction of a related steroid
with the bisulphite adduct of an aldehyde to yield a steroid acetal has
been previously reported: Phull, M. S.; Rao, D. R.; Kankan, R. N.
World Patent 2008/2008035066 A2, 2008.
703
dx.doi.org/10.1021/op200257g | Org. Process Res. Dev. 2012, 16, 697−703