A new method for the synthesis of 5-fluorouracil prodrugs

Article abstract of DOI:10.1055/s-1999-2938

Bromo-tris(dimethylamino)-phosphonium-hexafluorophosphate (BROP) is a particularly suitable reagent for synthesising some 5-fluorouracil prodrugs in a one step reaction. Several acids, especially retinoic acid, were bound to the N₁ of 5-fluorou

Full text of DOI:10.1055/s-1999-2938

LETTER
1829
A New Method for the Synthesis of 5-Fluorouracil Prodrugs
Pascale Jolimaître a, Christiane André-Barres, Myriam Malet-Martino a*, Robert Martino a, Isabelle Rico-Lattes
Laboratoire des IMRCP (Groupe de RMN Biomédicalea), UMR CNRS 5623, Université Paul Sabatier, 118 route de Narbonne,
31062 Toulouse, France
E-mail: rmnbio@ramses.ups-tlse.fr
Received 12 August 1999
Abstract: Bromo-tris(dimethylamino)-phosphonium-hexafluoro-
phosphate (BROP) is a particularly suitable reagent for synthesising
some 5-fluorouracil prodrugs in a one step reaction. Several acids,
especially retinoic acid, were bound to the N1 of 5-fluorouracil, with
yields of about 50% after purification.
Key words: N1-retinoyl-5-fluorouracil, BROP, artificial nucleo-
sides, prodrug
Proliferative vitreoretinopathy (PVR) is characterised by
a rapid and uncontrolled proliferation of ocular cells with-
in the vitreous body of the eye, frequently resulting in ret-
inal detachment.1 Silicone oil is often used to treat
Structures of 5-fluorouracil and prodrugs synthesised.
mechanical retinal detachment, but in the case of PVR, a
Figure 1
reproliferation is common.2 The injection of an antitumor
agent would stop the proliferation and the silicone oil
could be both a vehicle for this agent and effective against
We first used a known method for the preparation of N1-
acyl-FU compounds with a saturated alkyl chain5a
(Scheme 1). The acid chloride was prepared by a known
method, using oxalyl chloride, then added to a solution of
FU potassium salt 6 in acetonitrile.6 The corresponding
N1-acyl-FU was purified by chromatography. All com-
pounds were characterised with 1H NMR, 13C NMR, mass
spectrometry and elemental microanalysis.7 Since the N1
position of FU is more reactive than the N3 position to-
wards acyl chlorides, the preferred formation of N1-alkyl
or N1-alkenyl carbonyl derivatives was expected. Never-
the retinal detachment. Another difficulty is the length of
the treatment; to be effective the antitumor agent must be
present in the vitreous body for at least 6 weeks. Pharma-
cological studies have shown that 5-fluorouracil (FU) 1
inhibits the growth of rabbit fibroblasts both in vitro and
in vivo as well as being capable of reducing experimental
intraocular proliferation in rabbits.3 However, FU is insol-
uble in silicone oil and is eliminated very rapidly follow-
ing intravitreal injection. So, in patients, repeated
injections would be necessary, which is uncomfortable.
Other investigations showed that retinoic acid (RA) inhib-
its retinal cell proliferation. The antimetabolite effect of
RA in a rabbit model of PVR has already been proved.4
1
theless, the N1 substitution was established from the H
NMR chemical shift of H6, which is ≈8.2-8.3 ppm
(CDCl3) for N1-substituted derivatives and ≈7.2-7.3 ppm
for N3-substituted derivatives.5 Compounds 2-4 were ob-
tained (Table 1), but we were unable to get retinoyl-FU 5
by this method.
We thus chose to prepare a prodrug containing both FU
and RA, i.e. retinoyl-FU 5. This lipophilic compound
should slowly liberate the two active compounds, thus
maintaining a therapeutic intraocular level of FU and RA
in the vitreous body for several days or weeks.
Before preparing 5 from RA, which is difficult to handle
because of its rapid isomerisation in the presence of light
and temperature and its sensitivity to water traces, three
model compounds were synthesised:
– N1-lauroyl-FU 2: a prodrug with a long alkyl chain,
– N1-oleyl-FU 3: a prodrug with a long unsaturated alkyl
chain,
– N1-octenoyl-FU 4: a prodrug with a long unsaturated
alkyl chain conjugated with the carbonyl group (Figure 1).
Scheme 1
Synlett 1999, No. 11, 1829–1831 ISSN 0936-5214 © Thieme Stuttgart · New York

1830
P. Jolimaître et al.
LETTER
To synthesise this product, we tried without success some
classical coupling agents, such as DCC in the presence
of HOBt or 2-mercaptothiazoline. However, bromo-
tris(dimethylamino)-phosphonium-hexafluorophosphate
(BROP) 7, which has been described for the coupling of
N-methyl-aminoacids8 and disubstituted amines, allowed
us to get compounds 2-5 (Table 1), according to the mech-
anism described in scheme 2. The reaction of the carbox-
ylic acid and FU 1 in the presence of BROP and
triethylamine in DMF gave the desired products 2-5.6
BROP was added last to avoid as far as possible the for-
mation of anhydride. Since this agent is very reactive, we
made the coupling at 0°C to prevent the formation of the
N3 substituted prodrugs.
References and Notes
(1) a) Glaser, B.M.; Cardin, A.; Biscoe, B. Ophthalmology 1987,
94, 327. b) Machmer, R. Invest. Ophthalmol. Vis. Sci. 1988,
29, 1771.
(2) Lucke, K.; Laqua, H. Silicone oil in the treatment of
complicated retinal detachment. Techniques, results and
complications; Springer: Berlin, 1990.
(3) a) Stern, W.; Geoffrey, L.P.; Erickson, P.; Guerin, G.;
Anderson, D.; Fisher, S.; O’Donnell, J. Am. J. Ophthalmol.
1983, 96, 33. b) Blumenkranz, M.S.; Claflin, A.; Hajek, A.S.
Arch. Ophthalmol. 1984, 102, 598.
(4) Manzanas, I.; Marin, J.; Refojo, M.F. Invest. Ophthalmol. Vis.
Sci. 1990, 31 (suppl), 24.
(5) a) Beall, H.D.; Prankerd, R.J.; Sloan, K.B. Drug Dev. and Ind.
Pharm. 1996, 22, 85. b) Beall, H.D.; Prankerd, R.J.; Todaro,
L.J.; Sloan, K.B. Pharm. Res. 1993, 10, 905. c) Beall, H.D.;
Prankerd, R.J.; Sloan, K.B. Drug Dev. and Ind. Pharm. 1997,
23, 517.
(6) Typical experimental procedure for both methods:
Method 1: in a first step, reaction of a suitable acid (1.5 mmol)
with oxalyl chloride (7.5 mmol) in dry toluene (20 ml) for 1
hour led to the desired acid chloride. At the same time, stirring
FU (1.02 mmol) with potassium hydroxyde (1 mmol) in
methanol (5 ml) for 30 minutes gave the potassium salt of FU.
In a second step, the acid chloride was added dropwise to a
solution, refrigerated at 0°C, containing the white residue of
FU potassium salt in acetonitrile (10 ml). The mixture was
then stirred at room temperature for 4 hours.
Method 2: the carboxylic acid (1 mmol) and FU (1.5 mmol)
were dissolved in anhydrous DMF (35 ml), and stirred at 0°C.
Triethylamine (3 mmol) and BROP (1.5 mmol) were added.
The mixture was then stirred at 0°C for 2 hours.
Scheme 2
(7) a) N1-lauroyl-5-fluorouracil 2: the different salts, FU and
anhydride were separated on a small dry silica gel
The purification was the most difficult step, because of the
lack of stability of compounds 2-5 towards silica acidity
and their different solubility in solvents. Table 1 shows
that the two methods led to compounds 2-4 in comparable
yields.
chromatography column. The remaining acid was removed by
dissolving it in cyclohexane. 1H NMR (400 MHz, CDCl3):
0.85 (3H, t, J 6.8 Hz, CH3), 1.23 (16H, m, (CH2)8-CH3), 1.69
(2H, app quin, J 7.3 Hz, CH2-(CH2)8-CH3), 3.09 (2H, t, J 7.3
Hz, O=C-CH2), 8.27 (1H, d, 3JH-F 6.6 Hz, H6), 9.2 (1H, s, NH).
13C NMR (100 MHz, CDCl3): 14.1 (s, CH3), 22.6, 24.4, 28.9,
29.28, 29.29, 29.4, 29.54, 29.56, 31.9, 39.0 (10s, (CH2)10-
CH3), 121.8 (d, 2JC-F 36.7 Hz, C6), 141.3 (d, 1JC-F 244.0 Hz,
C5), 147.8 (s, C2), 156.8 (d, 2JC-F 28.4 Hz, C4), 172.0 (s, CH2-
C=O). MS (FAB+, MNBA) m/z 313 [MH]+, 625 [2MH]+, 131
[FU+H]+, 183 (loss of one FU molecule). Anal. for
C16H25N2O3F: C, 61.51; H, 8.07; N, 8.97. Found: C, 61.46; H,
7.95; N, 8.98. White solid: mp 92°C.
b) N1-oleyl-5-fluorouracil 3: since it is very unstable on a
silica gel column, the compound was purified by a liquid/
liquid extraction in heptane/ acetonitrile. 1H NMR (400 MHz,
CDCl3): 0.88 (3H, t, J 6.6 Hz, CH3), 1.29 (20H, m, (CH2)4-
CH2-CH=CH-(CH2)6-CH3), 1.72 (2H, app quin, J 7.0 Hz,
CH2-CH2-C=O), 2.02 (4H, m, CH2-CH=CH-CH2), 3.12 (2H,
t, J 7.3 Hz, CH2-C=O), 5.35 (2H, m, CH=CH), 8.29 (1H, d,
3JH-F 6.6 Hz, H6), 9.2 (1H, s, NH).13C NMR (100 MHz,
CDCl3): 14.1 (s, CH3), 22.7, 24.4, 27.17, 27.23, 28.9, 29.1,
29.25, 29.3 (2C), 29.5, 29.7, 29.8, 31.9 (12s, (CH2)5-CH2-
CH=CH-CH2-(CH2)6-CH3), 39.1 (s, CH2-C=O), 121.8 (d,
2JC-F 36.6 Hz, C6), 130.1 and 129.7 (2s, CH=CH), 141.3 (d,
1JC-F 244.0 Hz, C5), 147.9 (s, C2), 157.0 (d, 2JC-F 28.4 Hz, C4),
172.1 (s, CH2-C=O). MS (FAB+, MNBA) m/z 395 [MH]+,
417 [MNa]+, 265 (loss of one FU molecule). Viscous oil.
c) N1-octenoyl-5-fluorouracil 4: the different salts, FU and
anhydride were separated on a small dry silica gel
Table 1 Yields obtained with both methods described in schemes 1
and 2 after purification of the compounds.
Method 1: carboxylic acid (1.5 eq), oxalyl chloride (7.5 eq), dry to-
luene, RT, 1 h; FU (1.02 eq), KOH (1 eq), MeOH, RT, 30 min; acid
chloride formed, potassium salt of FU, acetonitrile, 0°C, 4 h.
Method 2: carboxylic acid (1 eq), FU (1.5 eq), BROP (1.5 eq), NEt3
(3 eq), dry DMF, 0°C, 2 h.
The moderate yields can be explained by the low stability
of the prodrugs. The N1-CO bond hydrolyses5a very easily
in water and in basic medium, so this makes the reaction
and the purification difficult. The reactions have to be
done under nitrogen, all solvents must be anhydrous, the
silica gel for chromatography has to be dried and the prod-
ucts have to be stored under nitrogen. However, only the
BROP method led to compound 5. Moreover, the one step
reaction is shorter and easier to run. BROP is thus an ex-
cellent reagent for coupling poorly reactive amines or
amides and acids.
chromatography column. The remaining acid was removed by
dissolving it in cyclohexane. 1H NMR (400 MHz, CDCl3):
0.90 (3H, t, J 6.6 Hz, CH3), 1.33 (4H, m, (CH2)2-CH3), 1.52
Synlett 1999, No. 11, 1829–1831 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
A New Method for the Synthesis of 5-Fluorouracil Prodrugs
1831
(2H, app quin, J7.2 Hz, CH2-CH2-CH=), 2.34 (2H, app q, J 7.2
(1H, d, 3JH-F 6.8 Hz, H6), 8.46 (1H, s, NH). 13C NMR (100
MHz, CDCl3): 12.0 (s, C19’), 14.6 (s, C20’), 18.1 (s, C3’), 20.8 (s,
C18’), 28.0 (2C, s, C16’, C17’), 32.2 (s, C2’), 33.3 (s, C1’), 38.6 (s,
C4’), 117.4 (s, C14’), 122.1 (d, 2JC-F 36.8 Hz, C6), 128.4 (s, C10’
or C12’), 129.1 (s, C7’), 129.7 (s, C5’), 133.4 (s, C11’), 133.9 (s,
C12’ or C10’), 136.0 (s, C8’), 136.6 (s, C6’ or C9’), 140.1 (d, 1JC-F
242.1 Hz, C5), 141.2 (s, C9’ or C6’), 146.7 (s, C2), 155.5 (d,
2JC-F 27.8 Hz, C4), 158.5 (s, C13’), 161.9 (s, C15’). MS (FAB+,
MNBA) m/z 435 [MNa]+, 131 [FU+H]+, 283 (loss of one FU
molecule). Orange solid: mp 151°C.
Hz, CH2-CH=), 7.08 (1H, dt, 3JH-Htrans 15.2 Hz, 4JH-H 1.5 Hz,
=CH-C=O), 7.33 (1H, td, J 6.8 Hz, 3JH-Htrans 15.2 Hz, CH2-
CH=), 8.23 (1H, d, 3JH-F 6.6 Hz, H6), 8.7 (1H, s, NH). 13
C
NMR (100 MHz, CDCl3): 13.9 (s, CH3), 22.4, 27.5, 31.3, 33.1
(4s, (CH2)4-CH3), 121.8 (s, =CH-C=O), 122.5 (d, 2JC-F 36.5
Hz, C6), 141.3 (d, 1JC-F 243.1 Hz, C5), 147.8 (s, C2), 156.0 (s,
CH2-CH=), 156.5 (d, 2JC-F 28.4 Hz, C4), 164.0 (s, =CH-C=O).
MS (FAB+, MNBA): m/z 255 [MH]+ (→ MS/MS 131, 125),
509 [2M+H]+, 531 [2M+Na]+. Anal. for C12H15N2O3F: C,
56.68; H, 5.95; N, 11.02. Found: C, 56.83; H, 6.04; N, 10.98.
White solid: mp 77°C.
(8) Coste, J.; Dufour, M.J.; Pantaloni, A.; Castro, B. Tetrahedron
Lett. 1990, 31, 669.
d) N1-retinoyl-5-fluorouracil 5: the different salts, FU and
anhydride were separated in the dark, on a small dry silica gel
chromatography column. The remaining acid was removed by
dissolving it in cyclohexane. 1H NMR (400 MHz, CDCl3):
0.97 (6H, s, CH3 16', CH3 17'), 1.40 (2H, m, H2’), 1.55 (2H, m,
H3’), 1.64 (3H, s, CH3 18'), 1.95 (2H, s, H4’), 1.97 (3H, s, CH3
19'), 2.35 (3H, s, CH3 20'), 6.10 (1H, d, 3JH-Htrans 15.9 Hz, H8’),
6.13 (1H, d, J 11.3 Hz, H10’ or H12’), 6.30 (1H, d, 3JH-Htrans 15.9
Hz, H7’), 6.37 (1H, d, 3JH-Htrans 15.0 Hz, H12’ or H10’), 6.78 (1H,
s, H14’), 7.13 (1H, dd, J 11.3 Hz, 3JH-Htrans 15.0 Hz, H11’), 8.14
(9) Using another coupling reagent, bromo-tris-pyrrolidino-
phosphonium-hexafluorophosphate, can prevent the
formation of HMPA, which is very toxic. The by-product
formed is tris-pyrrolidinophosphoramide and the yields are
close to those obtained with BROP.
Article Identifier:
1437-2096,E;1999,0,11,1829,1831,ftx,en;G20999ST.pdf
Synlett 1999, No. 11, 1829–1831 ISSN 0936-5214 © Thieme Stuttgart · New York