Synthesis of allyloxycarbonyloxymethyl-5-fluorouracil and copolymerizations with N-vinylpyrrolidinone

Article abstract of DOI:10.1039/B002092N

Poly(N-vinylpyrrolidinone) (PNVP) bearing 5-fluorouracil (5-FU) moieties was synthesised via a three-step method. Firstly, 5-FU reacted with formaldehyde to form a mixture of mono- and disubstituted hydroxymethyl-5-FUs. The mixture was then treated with allyl chloroformate to afford allyloxycarbonyloxymethyl-5-FUs. This reaction showed site-specificity: the hydroxymethyl goup at the N-I position readily reacted with chloroformate whereas the N-3 hydroxymethyl group partially decomposed into formaldehyde and the amide group. 1-Allyloxycarbonyloxymethyl-5-FU (4) and NVP were copolymerized in dioxane using azobisisobutyronitrile as an initiator. The monomer reactivity ratios, r₄= 0.32 and r(NVP)= 0.97, were evaluated by both linear and non-linear methods.

Full text of DOI:10.1039/B002092N

View Article Online / Journal Homepage / Table of Contents for this issue
J O U R N A L O F
Synthesis of allyloxycarbonyloxymethyl-5-¯uorouracil and
copolymerizations with N-vinylpyrrolidinone
Zuifang Liu,a Nigel Fullwoodb and Steve Rimmer*a
aThe Polymer Centre, Lancaster University, Lancaster, UK LA1 4YA.
E-mail: S.Rimmer@Lancaster.ac.uk
C H E M I S T R Y
bBiological Sciences, Lancaster University, Lancaster, UK
Received 15th March 2000, Accepted 17th May 2000
Published on the Web 3rd July 2000
Poly(N-vinylpyrrolidinone) (PNVP) bearing 5-¯uorouracil (5-FU) moieties was synthesised via a three-step
method. Firstly, 5-FU reacted with formaldehyde to form a mixture of mono- and disubstituted
hydroxymethyl-5-FUs. The mixture was then treated with allyl chloroformate to afford
allyloxycarbonyloxymethyl-5-FUs. This reaction showed site-speci®city: the hydroxymethyl goup at the N-1
position readily reacted with chloroformate whereas the N-3 hydroxymethyl group partially decomposed into
formaldehyde and the amide group. 1-Allyloxycarbonyloxymethyl-5-FU (4) and NVP were copolymerized in
dioxane using azobisisobutyronitrile as an initiator. The monomer reactivity ratios, r4~0.32 and rNVP~0.97,
were evaluated by both linear and non-linear methods.
synthetic route involves the preparation of allyloxycarbonyl-
oxymethyl-5-¯uorouracils.
Introduction
It has been reported that 5-¯uorouracil (5-FU), a widely used
antimetabolite, is effective in inhibiting the progression of
disease such as proliferative vitreoretinopathy (PVR), which
remains a leading cause of failure in retinal detachment
surgery,1 and various cancers.2 Several polymeric routes to
controlled administration of 5-FU have now been examined.
These include encapsulation in synthetic or natural hydro-
gels,3±8 followed by diffusive release; encapsulation in degrad-
able polymer devices from which release occurs as the polymer
degrades;9±20 conjugation to soluble peptides, which may then
be attached to other polymers;21±25 copolymerization of 5-FU
monomers with other water soluble monomers26±35 or attach-
ment of reactive 5-FU compounds to polymers.36 Release from
the three latter methodologies involves degradation of a group
linking the polymer to 5-FU. In the case of the peptide linked
species this degradation occurs through enzymatic attack on
the peptide, which can also function as a targeting group,
whereas in non-peptide degradable systems release relies on
hydrolysis. Several 5-FU functional monomers have been
reported. Thus, Cho et al. have reported the synthesis and
polymerization of 1,2,3,6-tetrahydrophthaloyl, acrylic or
itaconyl monomers with 5FU functionality.27±30,32±34 Berger
et al. have shown that the action of 5-FU in PVR can be
enhanced by coupling to dexamethasone using a degradable
carbonate linking group.1 The coupled molecule was not active
but as the carbonate group hydrolysed, releasing only carbon
dioxide as the by-product,1 the two active drugs were released.
Following on from this work we report here a synthesis of a
polymerizable carbonate derivative of 5-FU and detail
copolymerization behaviour with NVP. PNVP has been
selected as the major component of the polymer backbone
because of its excellent biocompatibility when used as vitreous
substitute or when implanted in the vitreous body.1,37,38 The
Experimental
Materials
5-FU (Fluka) was used as supplied. N-Vinylpyrrolidinone
(NVP) (Aldrich) was puri®ed by distillation under reduced
pressure prior to polymerization. Azobisisobutyronitrile
(AIBN) (Merck) was recrystallised from methanol. Triethyla-
mine (TEA) and solvents were puri®ed and dried by standard
procedures.39 Distilled and deionised water was used through-
out. Other chemicals were used as received from Aldrich.
Monomer synthesis
5-FU (1.3 g, 10 mmol) and 37 wt% formaldehyde in aqueous
solution (1.14 g, 14 mmol) were added to water (10 g) in a
100 mL round-bottomed ¯ask which was immersed in a
60¡1 ³C oil bath. The reaction was conducted under
magnetic agitation for 6 hours. The resultant solution was
concentrated using a rotary evaporator and then dried in a
vacuum oven at 60 ³C for 48 hours. The obtained oily
product was dissolved in dry acetonitrile (30 mL) in a two-
necked round-bottomed ¯ask containing TEA (1.37 g,
13.5 mmol). To this solution, allyl chloroformate (1.54 g,
12.4 mmol) was added dropwise during about 5 minutes
under dry nitrogen at room temperature. A white precipitate
formed as soon as the chloroformate was added. After
stirring at room temperature for 3 hours, the mixture was
®ltered and acetonitrile was removed under reduced
pressure. The resultant was then dissolved in dichloro-
methane (30 mL) and washed three times with HCl
(1.0 mol dm23), saturated NaHCO3aq. (twice) and water
(twice). The organic layer was dried over sodium sulfate,
DOI: 10.1039/b002092n
J. Mater. Chem., 2000, 10, 1771±1775
This journal is # The Royal Society of Chemistry 2000
1771

View Article Online
®ltered and ®nally concentrated by using a rotary evapora-
tor. Column chromatography on silica gel (silica gel 60,
Fluka) using a mixture of methanol and dichloromethane
(1 : 20 v/v) afforded two fractions. Evaporation of the
solvent gave 0.68 g (19% yield, based on charged 5-FU) of
an oily compound and 0.83 g (34% yield) of a wax-like
compound for the ®rst and second fraction, respectively. H
NMR spectrometry revealed that the former contained
mainly 1,3-bis(allyloxycarbonyloxy-
and the residual monomer levels were calculated using a standard
calibration curve. Pure linear polymer samples, for NMR
spectroscopy, were prepared by double reprecipitation from
methanol into diethyl ether.
Analysis
1
1
200 MHz H nuclear magnetic resonance (NMR) spectra were
recorded in dimethyl sulfoxide-d6 (DMSO-d6) with tetra-
methylsilane (TMS) as an internal standard using a Varian
NMR 200 spectrometer (Oxford, UK). Polymerization was
followed by measuring the residual monomer concentrations
using HPLC (column: Waters C8 3.96150 mm, at 30 ³C; eluent:
30 vol% methanolz70 vol% 10 mM ammonium acetate aqueous
solution; ¯ow rate: 1.0 mL min21; detector: UV234 nm and
UV262 nm for NVP and monomer 4, respectively). The molecular
weights of copolymers of 4 and NVP with different monomer
feed ratios were determined by size exclusion chromatography
(SEC) (column: PSS HEMA 86300 nm, at 70 ³C; eluent 0.1%
(w/v) ammonium acetate in dimethylacetamide; ¯ow rate:
1.0 mL min21; refractive index detector). The calibration curves
were obtained using poly(ethylene oxide) standards.
methyl)-5-¯uorouracil (6), and the latter 1-allyloxycarbonyl-
oxymethyl-5-¯uorouracil (4).
1
Analytical data for 4: H NMR (DMSO-d6) d~4.54 (m, O-
CH2-CHL), 5.32 (m, LCH2), 5.62 (s, N-CH2-O), 5.95 (m, -
CHL), 8.15 (d, CH, 5-FU), 12.03 (s, NH, 5-FU); 13C NMR
(DMSO-d6) d~68.4, 73.0, 118.7, 129.0, 129.7, 131.7, 137.2,
141.7, 149.2, 153.3, 157.1, 157.6; FT-IR/cm21~3200, 3050,
1700 (broad), 1490, 1400, 1350, 1250, 1150, 1100, 950, 800; m/
z~244 (ESIz); mp 82.2 ³C (DSC).
1
Analytical data for 6: H NMR (DMSO-d6) d~4.64 (m, O-
CH2-CHL), 5.32 (m, LCH2), 5.69 (s, N-CH2-O, N-1 position),
5.85 (s, N-CH2-O, N-3 position), 5.95 (m, -CHL), 8.31 (d, CH,
5-FU).
Results and discussion
Polymerization
Preparation and characterization of monomers
1 Preparation of linear copolymers. A series of polymerizations
were performed with various monomer feed ratios of 4 and
NVP. In a typical polymerization, AIBN (0.016 g, 0.98 mmol)
and a mixture of NVP (as de®ned by the designated feed) and 4
(10 mmol, 2.16 g) in dioxane (20 mL) were charged to a two-
necked round-bottomed ¯ask equipped with a magnetic stirrer
bar, nitrogen inlet and re¯ux condenser. The solution was
deoxygenated by bubbling with nitrogen at room temperature for
1 hour. Polymerization was then started by immersing the ¯ask in
an oil bath at 60¡1 ³C. Samples (0.2 mL) were taken as the
polymerization progressed and these were added to sample tubes
containing 1.5 mL propan-2-ol (HPLC grade). The sample tubes
were shaken for 20 minutes and then centrifuged to separate the
precipitated polymer from the solution. The supernatants were
®nally analysed by high pressure liquid chromatography (HPLC)
5-FU can react with formaldehyde (in stoichiometric excess) to
give 3 in quantitative yield.40±42 In the current work, reaction 1 in
Scheme 1wasconductedindiluteaqueoussolutionusingaratioof
1.0 mole 5-FU to 1.4 mole formaldehyde and afforded a mixture
of 1-hydroxymethyl-5-¯uorouracil (1), 3-hydroxymethyl-5-¯uoro-
uracil(2)and1,3-bis(hydroxymethyl)-5-¯uorouracil(3). Fig. 1A
showsthe1HNMRspectrumandtheassignmentsoftheproduct
of reaction 1 after 6 hours. It is evident that the product was a
mixtureof1, 2and3. Theproductcompositionsweredetermined
1
by H NMR and the results after 3 and 6 hours are given in
Table 1. As expected, the amount of 1 was higher than that of 2
throughout the reaction, whilst the amount of 3 increased as the
reaction progressed. Finally, a mixture comprising 31.7 mol% 1,
21.1 mol% 2 and 32.7 mol% 3 was obtained.
Scheme 1
1772
J. Mater. Chem., 2000, 10, 1771±1775

View Article Online
Fig. 1 1H NMR spectra: A, products of reaction 1, 6 hours; B, product of reaction 2, after work-up.
After drying, this mixture was treated with allyl chlorofor-
1
mate (reaction 2) in the presence of TEA. H NMR spectro-
methanol is more easily removed than the N-1 methanol group.
On preparing 5-FU derivatives from 1,3-bis(hydroxymethyl)-5-
FU and some chloroformates, Nagase et al.43 reported that the
reaction occurred exclusively on the N-1 position at the early
stage and the remaining chloroformate then reacted with the
N-3 hydroxymethyl group to give disubstituted compounds. In
the current study, the amount of allyl chloroformate
(12.4 mmol) was larger than the total amount of hydroxy-
methyl groups (11.8 mmol, calculated from the composition
result). However, only partial conversion of 3 to 6 was
observed. Furthermore, no signal arising from the proton on
the hydroxymethyl group at the N-3 position was detectable
(see Fig. 1B). These results suggest that, under the reaction
conditions, the hydroxymethyl group at the N-1 position could
be activated to react with allyl chloroformate. The hydro-
xymethyl group at the N-3 position, on the other hand, reacted
with chloroformate by decomposition into formaldehyde and
NH group. This is further supported by the fact that only a very
small amount of 5 was formed from about 2.1 mmol of 2.
scopy of the product revealed that the amount of N-1
substituted monomer, 4 (56.3 mol%), in the reaction mixture
was signi®cantly larger than that of the disubstituted monomer,
6 (31.5 mol%), and only a small amount of the N-3 substituted
monomer, 5 (1.0 mol%), was detected (Fig. 1B). By examining
the NMR spectra in Fig. 1, it is clear that the product
distribution was signi®cantly different from the product
distribution predicted by reference to the starting mixture,
containing 1, 2 and 3. This implies different behaviours of the
hydroxymethyl groups at the N-1 and N-3 positions; that is,
demethylolation was a function of the substitution pattern.
After work-up and column separation, the two major
components, 4 and 6, were isolated in 34% and 19% yield
(based on charged 5-FU), respectively. It is noteworthy that the
amount of isolated 6 (1.9 mmol) is signi®cantly smaller than the
amount predicted by the amount of 3 (ca. 3.3 mmol) present in
the starting material. In contrast, the amount of 4 (3.4 mmol) is
slightly larger than the amount expected based on the amount
of 1 (3.2 mmol) present in the starting material. These results
suggest that 1, 2 and 3 undergo demethylolation in competition
with allyloxyformylation so that a fraction of 3 undergoes
demethylolation at N-3 followed by allyloxyformylation of the
remaining hydroxymethyl group at N-1. Thus reaction of 3
generated both 4 and 6. Similarly, 1 could react readily to form
4 but a fraction of 1 underwent demethylolation. Since 5 was
produced only in low yield it is reasonable to assume the N-3-
Copolymerization of monomer 4 with NVP
Isolated monomer 4 was copolymerized with NVP at various
feed ratios (reaction 3 in Scheme 1). Fig. 2 shows the variations
of the conversions of both 4 and NVP with time for a typical
polymerization. It is evident that 4 polymerized at a slightly
slower rate than NVP. A typical 1H NMR spectrum for one of
the copolymers is shown in Fig. 3. The characteristic signal
corresponding to the proton on the 5-FU residue at around
J. Mater. Chem., 2000, 10, 1771±1775
1773

View Article Online
Table 1 Product compositions for reaction 1 in Scheme 1
Product composition (mol%)
Reaction time/h
1
2
3
3
6
32.1
31.7
22.5
21.1
17.7
32.7
Fig. 3 1H NMR spectrum of copolymer prepared under the conditions
shown in Fig. 2.
Fig. 2 Conversion of monomer versus polymerization time
[4 : NVP~44 : 56 (mol/mol)].
8.15 ppm is clearly seen. However, the initial attempt to
1
determine the copolymer compositions by using the H NMR
spectroscopy was not successful because of the wide and
incompletely resolved peaks arising from the protons on the
NVP repeat unit. To evaluate the monomer reactivity ratios,
the copolymer compositions were determined from the residual
monomer levels measured by HPLC. The composition of one
®nal copolymer prepared using about 44 mol% of 4 in the feed
was also determined by elemental analysis (performed by
C.H.N Analysis Ltd, Leicester, UK). The results obtained by
the two methods are in good agreement (the fractions of
monomer 4 in the copolymer are 0.35 and 0.33 determined by
HPLC and by elemental analysis, respectively). In order to
satisfy the well-known differential equation relating the
copolymer composition to the monomer composition, data
with monomer conversions lower than 10% were used for the
evaluation of the monomer reactivity ratios. The compositions
of the monomer feed and of the corresponding copolymers are
given in Table 2. A pair of reactivity ratios, r4~0.31¡0.06 and
Fig. 4 95% con®dence ellipse for the reactivity ratios determined by the
non-linear least-squares method.
r
NVP~0.97¡0.02, was obtained by using the Kelen±TuÈdoÄs
linear regression method.44 These data were then used as initial
approximate values for the non-linear least-square Tidwell±
Mortimer method,45 which gave r4~0.32 and rNVP~0.97. The
95% con®dence ellipse is shown in Fig. 4. Molecular weights of
the copolymers were determined by SEC and plotted against
the fraction of monomer 4 in the monomer feed in Fig. 5. It can
be seen that as expected an increase in the fraction of 4 resulted
in a decrease in the molecular weight.
Fig. 5 Variation of copolymer molecular weight with fraction of 4 in
the monomer feed. The monomer conversions are presented in Table 2.
Conclusion
Table 2 Monomer conversion and compositions of the feed and the
copolymers
We report the synthesis of a new 5-FU functional monomer
and its copolymerization with NVP. The allyl carbonate
monomer can be successfully copolymerized with NVP over
a wide composition range. The reactivity ratios imply that
copolymerization would lead to polymers that are slightly
rich in NVP at low conversion and rich in the new monomer
towards the end of the polymerization. However, the values
(r4~0.32, rNVP~0.97) were found to be suf®ciently close to
minimize the production of homopolymer during the
polymerization. Although a substantial decrease in molecu-
fa
Fb
Conversion of 4 (%)
Conversion of NVP (%)
0.15
0.20
0.29
0.44
0.60
0.14
0.18
0.25
0.36
0.47
6.8
7.8
6.2
6.7
5.5
7.3
8.8
7.8
9.4
9.1
aMolar fraction of 4 in monomer feed. bMolar fraction of 4 in copo-
lymer.
1774
J. Mater. Chem., 2000, 10, 1771±1775

View Article Online
18 K. Ciftci, A. A. Hincal, H. S. Kas, M. T. Ercan and S. Ruacan,
Eur. Pharm. Sci., 1994, 1, 249.
19 T. Moritera, Y. Ogura, Y. Honda, R. Wada, S. H. Hyon and
Y. Ikada, Invest. Ophthalmol. Visual Sci., 1991, 32, 1785.
20 J. Heller, Y. F. Maa, P. Wuthrich, S. Y. Ng and R. Duncan,
J. Controlled Release, 1991, 16, 3.
21 M. Nichifor, E. H. Schacht and L. W. Seymour, J. Controlled
Release, 1997, 48, 165.
22 K. S. Kim, D. S. Lim and S. H. Cho, Korea Polym. J., 1996, 4, 16.
23 M. Nichifor, E. H. Schacht and L. W. Seymour, J. Controlled
Release, 1996, 39, 79.
lar weight occurred as the amount of 4 in the feed increased,
high molecular weight polymers were formed at all
compositions so that transfer to monomer does not appear
to be a problem in these copolymerizations. 5-FU is a useful
therapeutic agent that requires quite high local concentration
to be effective so that studies of local sustained release from
these and similar polymers is ongoing and will be reported in
due course.
24 D. Putnam and J. Kopecek, Bioconjugate Chem., 1995, 6, 483.
25 M. Nichifor, E. H. Schacht, L. W. Seymour, D. Anderson and
M. Shoaibi, J. Bioact. Compat. Polym., 1997, 12, 265.
26 S. M. Lee, W. M. Choi, C. S. Ha and W. J. Cho, J. Polym. Sci.
Part A: Polym. Chem., 1999, 37, 2619.
Acknowledgements
The authors wish to thank EPSRC, UK (GR/M36526) for the
®nancial support to this project.
27 J. G. Park, C. S. Ha and W. J. Cho, J. Polym. Sci. Part A: Polym.
Chem., 1999, 37, 2113.
28 J. G. Park, C. S. Ha and W. J. Cho, J. Polym. Sci. Part A: Polym.
Chem., 1999, 37, 1589.
29 J. G. Park, S. H. Kim, C. S. Ha and W. J. Cho, J. Polym. Sci. Part
A: Polym. Chem., 1998, 36, 2985.
30 W. M. Choi, I. D. Chung, N. J. Lee, Y. W. Lee, C. S. Ha and
W. J. Cho, J. Polym. Sci. Part A: Polym. Chem., 1998, 36, 2177.
31 A. Kishida, H. Goto, K. Murakami, K. Kakinoki, M. Akashi and
T. Endo, J. Bioact. Compat. Polym., 1998, 13, 222.
32 J. G. Park, W. M. Choi, N. J. Lee, C. S. Ha and W. J. Cho,
J. Polym. Sci. Part A: Polym. Chem., 1998, 36, 1625.
33 W. M. Choi, I. D. Jung, N. J. Lee, S. H. Kim, C. S. Ha and
W. J. Cho, Polym. Adv. Technol., 1997, 8, 701.
34 N. J. Lee, C. S. Ha and W. J. Cho, J. Macromol. Sci., Pure Appl.
Chem., 1992, 29, 161.
35 T. Akiyama, Y. Takesue, M. Kumegawa, H. Nishimoto and
S. Ozaki, Bull. Chem. Soc. Jpn., 1991, 64, 2266.
References
1
A. S. Berger, C. K. Cheng, P. A. Pearson, P. A. Shton,
P. A. Crooks, T. Cynkowski, G. Cynkowska and G. J. Jaffe,
Invest. Ophthalmol. Visual Sci., 1996, 37, 2318.
D. Putnam and J. Kopecek, Adv. Polym. Sci., 1995, 122, 55.
O. Garcia, M. D. Blanco, J. A. Martin and J. M. Teijon, Eur.
Polym. J., 2000, 36, 111.
2
3
4
5
6
7
8
9
Y. Luo, Y. Aso and S. Yoshioka, Chem. Pharm. Bull., 1999, 47,
579.
O. Garcia, M. D. Blanco, C. Gomez and J. M. Teijon, Polym.
Bull., 1997, 38, 55.
T. D. Blanco, O. Garcia, R. Olmo, J. M. Teijon and I. Katime,
J. Chromatogr. B: Biomed. Appl., 1996, 680, 243.
X. M. Li, W. M. Shen, C. J. Liu, S. I. Nishimoto and T. Kagiya,
Radiat. Phys. Chem., 1991, 38, 377.
C. Zinutti, F. Kedzierewicz, M. Hoffman and P. Maincent,
J. Microencapsulation, 1994, 11, 555.
36 M. Nichifor, V. Coessens and E. H. Schacht, J. Bioact. Compat.
Polym., 1995, 10, 199.
Y. H. E. Lin and R. C. Vasavada, J. Microencapsulation, 2000, 17,
1.
37 M.
J.
Bruining,
P.
S.
Edelbroek
Hoogendoorn,
10 S. Einmahl, M. Zignani, E. Varesio, J. Heller, J. L. Veuthey,
C. Tabatabay and R. Gurny, Int. J. Pharm., 1999, 185, 189.
11 M. B. Sintzel, J. Heller, S. Y. Ng, C. Tabatabay,
K. SchwachAbdellaoui and R. Gurny, J. Controlled Release,
1998, 55, 213.
12 T. H. Zhou, H. Lewis, R. E. Foster and S. P. Schwendeman,
J. Controlled Release, 1998, 55, 281.
13 G. A. Peyman, D. H. Yang, B. Khoobehi, M. H. Rahimy and
S. Y. Chin, Ophthalmic Surgery and Lasers, 1996, 27, 384.
14 L. K. Chiu, W. J. Chiu and Y. L. Cheng, Int. J. Pharm., 1995, 126,
169.
H. G. T. Blaauwgeers, C. M. Mooy, F. H. Hendrikse and
L. H. Koole, J. Biomed. Mater. Res., 1999, 47, 189.
38 Y. Hong, T. V. Chirila, S. Vijayasekaran, W. Shen, X. Lou and
P. D. Dalton, J. Biomed. Mater. Res., 1998, 41, 650.
39 W. L. F. Armarego and D. D. Perrin, Puri®cation of Laboratory
Chemicals, 4th edn., Butterworth-Heinemann, Bath, UK,
1996.
40 S. Ahmad, S. Ozaki, T. Nagase, M. Iigo, R. Tokuzen and
A. Hoshi, Chem. Pharm. Bull., 1987, 35, 4137.
41 Y. Ohya, H. Kobayashi and T. Ouchi, React. Polym., 1991, 15,
153.
15 G. E. Trope, Y. L. Cheng, H. Sheardown, G. S. Liu, I. A. Menon,
J. G. Heathcote, D. S. Rootman, W. J. Chiu and L. Gould, Can.
J. Ophthalmol., 1994, 29, 263.
42 J. C. Maurizis and B. Pucci, Macromol. Chem. Phys., 1999, 200,
1351.
43 T. Nagase, K. Shiraishi, Y. Yamada and S. Ozaki, Heterocycles,
1988, 27, 1155.
44 T. Kelen and F. TuÈdoÄs, J. Macromol. Sci. Chem., 1975, A9, 1.
45 P. W. Tidwell and G. A. Mortimer, J. Polym. Sci., 1965, A3,
369.
16 S. F. Bernatchez, A. Merkli, T. L. Minh, C. Tabatabay,
J. M. Anderson and R. Gurny, J. Biomed. Mater. Res., 1994,
28, 1037.
17 L. W. Seymour, R. Duncan, J. Duffy, S. Y. Ng and J. Heller,
J. Controlled Release, 1994, 21, 201.
J. Mater. Chem., 2000, 10, 1771±1775
1775