Lipophilic methotrexate conjugates with antitumor activity

Article abstract of DOI:10.1016/S0928-0987(00)00062-2

Lipophilic methotrexate (MTX)-lipoamino acid conjugates coupled with amide or ester linkages (1a-1r) were synthesised. The inhibitory activity of the conjugates was evaluated on bovine liver DHFR. The in vitro growth inhibitory effect against MTX-sensitive human lymphoblastoid CCRF-CEM cells and an MTX-resistant sub-line (CEM/MTX), which displays defective intracellular transport of MTX, was determined under short-term and continuous (120-h incubation) exposure conditions. The α, γ, or α,γ amide conjugates showed different activity in inhibiting the growth of parent cells. CEM/MTX cells were much less susceptible than CCRF-CEM cells to inhibition by α or α,γ-substituted lipoamino acid conjugates, whereas both cell lines were almost equally sensitive to the MTX-γ conjugates. Although less potent than MTX, they could partially circumvent the impaired transport system. These findings confirm that lipophilic MTX conjugates may be good lead compounds on the drug development for the treatment of some MTX-resistant tumors. Ester-type conjugates displayed an interesting activity against parent CCRF-CEM cells, although they were less potent against the transport-resistant sub-line. Stability studies on these molecules indicated that they are not degraded into MTX in the culture medium, thus suggesting that they are not able to over-cross cell resistance despite of their lipophilicity. Copyright (C) 2000 Elsevier Science B.V.

Full text of DOI:10.1016/S0928-0987(00)00062-2

European Journal of Pharmaceutical Sciences 10 (2000) 237–245
www.elsevier.nl/locate/ejps
Lipophilic methotrexate conjugates with antitumor activity
Rosario Pignatelloa,b , , Giuseppina Spampinato , Valeria Sorrenti , Claudia Di Giacomo ,
Luisa Vicaric, John J. McGuiree, Cynthia A. Russelle, Giovanni Puglisib, Istvan Toth1,a
aDepartment of Pharmaceutical and Biological Chemistry, The School of Pharmacy, University of London, 29/39 Brunswick Square,
c
d
d
*
WC1 N 1AX London, UK
b
`
Dipartimento di Scienze Farmaceutiche, Universita di Catania, v.le A. Doria 6, 95125 Catania, Italy
c
`
III Cattedra di Patologia Generale, Universita di Catania, via Androne 86, 95124 Catania, Italy
d
`
Istituto di Chimica Biologica, Universita di Catania, v.le A. Doria 6, 95125 Catania, Italy
eRoswell Park Cancer Institute, Grace Cancer Drug Center, Elm and Carlton Streets, Buffalo, NY 14263, USA
Received 9 September 1999; received in revised form 20 December 1999; accepted 22 December 1999
Abstract
Lipophilic methotrexate (MTX)–lipoamino acid conjugates coupled with amide or ester linkages (1a–1r) were synthesised. The
inhibitory activity of the conjugates was evaluated on bovine liver DHFR. The in vitro growth inhibitory effect against MTX-sensitive
human lymphoblastoid CCRF-CEM cells and an MTX-resistant sub-line (CEM/MTX), which displays defective intracellular transport of
MTX, was determined under short-term and continuous (120-h incubation) exposure conditions. The a, g, or a,g amide conjugates
showed different activity in inhibiting the growth of parent cells. CEM/MTX cells were much less susceptible than CCRF-CEM cells to
inhibition by a or a,g-substituted lipoamino acid conjugates, whereas both cell lines were almost equally sensitive to the MTX-g
conjugates. Although less potent than MTX, they could partially circumvent the impaired transport system. These findings confirm that
lipophilic MTX conjugates may be good lead compounds on the drug development for the treatment of some MTX-resistant tumors.
Ester-type conjugates displayed an interesting activity against parent CCRF-CEM cells, although they were less potent against the
transport-resistant sub-line. Stability studies on these molecules indicated that they are not degraded into MTX in the culture medium, thus
suggesting that they are not able to over-cross cell resistance despite of their lipophilicity.
reserved.
2000 Elsevier Science B.V. All rights
Keywords: Methotrexate conjugates; Lipoamino acids; Antitumor agents; Lipophilicity; CCRF-CEM cells; CEM/MTX cells
1. Introduction
active carrier (RFC) used by the cells for the uptake of
physiologically circulating reduced folate cofactors and of
the antifolate drugs (Henderson, 1986).
We have previously reported the synthesis and bio-
logical evaluation of methotrexate (MTX) lipophilic amide
derivatives. Thus, MTX has been derivatised with amino-
alkanes (Pignatello et al., 1996, 1999) and lipoamino acids
(LAAs) (Pignatello et al., 1998), in order to achieve
compounds able to penetrate passively into tumor cells.
With such drugs, structurally related to the ’non-classical’
antifolates, the possibility arises of over-crossing one of
the forms of resistance commonly shown by many tumor
cell lines toward MTX. Such a ‘transport resistance’ is due
to a defective efficiency of the low-affinity, high-capacity
As many literature reports suggest, it is reasonable that
an overall reduction in the polarity of MTX molecule gives
a key for enhancing passive internalization of the cytotoxic
agent into transport defective cell clones (Kinsky and
Loader, 1987; Rahman and Cchabra, 1988; Westerhof,
1995). The modification of one or both the carboxyl
groups, present in the glutamate moiety of the MTX
molecule (defined as a region of ‘bulk tolerance’), ap-
peared as an ideal target, since the binding of MTX to its
target enzymes [dihydrofolate reductase (DHFR) and
thymidilate synthetase (TS)], is mainly linked to the
diaminopyrimidine moiety of the pteridine ring (Sirotnak
et al., 1979). For instance, Rosowsky et al. (1984) found
that increasing chain length in a series of alkyl monoesters
of MTX had a divergent effect upon their inhibitory
activity against DHFR (which is reduced) and CEM cells,
*Corresponding author. Tel.: 139-95-738-4021; fax: 139-95-222-239.
E-mail address: pignatel@mbox.unict.it (R. Pignatello)
1Present address: School of Pharmacy, University of Queensland, Bris-
bane 4072, Australia.
0928-0987/00/$ – see front matter
PII: S0928-0987(00)00062-2
2000 Elsevier Science B.V. All rights reserved.

238
R. Pignatello et al. / European Journal of Pharmaceutical Sciences 10 (2000) 237–245
against which the more lipophilic derivatives displayed the
maximum activity.
evaluated in vitro, along with their inhibitory activity
against bovine liver DHFR.
Among the previously described lipophilic MTX a,g-
bis-conjugates (Pignatello et al., 1996, 1998, 1999), some
derivatives showed interesting in vitro growth inhibitory
activity, compared to free MTX, against a human leukemic
T cell lymphoblastoid CCRF-CEM line. Furthermore, most
derivatives showed an inhibitory activity against bovine
liver DHFR, despite the absence of a free carboxylic group
in the a-position of glutamate moiety, whose importance
in the binding of folates to DHFR is known (Sirotnak et
al., 1979).
2. Experimental procedures
2.1. Apparatus
IR spectra (not reported) were recorded in KBr disks on
a Perkin–Elmer 1600 series FTIR spectrophotometer; the
data were in agreement with the structure of the com-
pounds. 1H-NMR were obtained in CDCl3 or in 1:1 (v/v)
¨
However, when an MTX transport-resistant cell sub-line
(CEM/MTX cells) was used, which shows an impaired
CDCl3 /MeOD with a Bruker AM500 instrument at 500
MHz; chemical shifts are reported in ppm downfield from
TMS as the internal reference. Mass spectra were recorded
with a VG Analytical ZAB-SE instrument, using fast-atom
bombardment (FAB) ionization. A 20-kV Cs1 ion bom-
bardment was used, with 2 ml of appropriate matrix (either
3-nitrobenzyl alcohol or thioglycerol/glycerol/TFA); a
NaI methanol solution was added to produce nitrated
species when no protonated molecular ions were observed.
TLC was performed on Merck F2541366 silica gel
aluminum backed plates; spots were detected by UV light,
exposure to iodine vapors or pouring into a 10% sulfuric
acid ethanol solution. Column chromatography was carried
out on Merck dry silica gel (230–400 mesh), with the
RFC system activity and
a reduced MTX uptake
(Rosowsky et al., 1980), lipophilic disubstituted deriva-
tives maintained a good cell growth inhibitory activity with
respect to the parent drug. Such findings confirm the
possibility of bypassing the cell defective RFC transport
system (Pignatello et al., 1996, 1998, 1999).
We describe here the synthesis and biological evaluation
of further a- and g-mono-substituted and a,g-disubstituted
lipoamino acid conjugates of MTX coupled with an ester,
amide or ester-amide linkage (Table 1). The growth
inhibitory activity of the conjugates against sensitive and
resistant CCRF-CEM lymphoblastoid cell sublines was
Table 1
Structure of mono- and disubstituted conjugates of MTX with lipoamino acids
Compound
R1
R2
1a
1b
1c
1d
1e
1f
1g
1h
1i
1j
1k
1l
1m
1n
1o
1p
1r
OH
NHCH[CONH2](CH2)13CH3
OH
NHCH[COOCH3](CH2)7CH3
NHCH[COOCH3](CH2)11CH3
OCH3
OCH3
OH
OH
NHCH[COOCH3](CH2)7CH3
NHCH[COOCH3](CH2)11CH3
OCH2CONHCH[COOC2H5](CH2)9CH3
OCH2CONHCH[COOC2H5](CH2)13CH3
OH
NHCH[CONH2](CH2)13CH3
NHCH[COOCH3](CH2)7CH3
NHCH[COOCH3](CH2)11CH3
NHCH[COOCH3](CH2)7CH3
NHCH[COOCH3](CH2)11CH3
NHCH[COOCH3](CH2)7CH3
NHCH[COOCH3](CH2)11CH3
OH
OH
OCH2CONHCH[COOC2H5](CH2)9CH3
OCH2CONHCH[COOC2H5](CH2)13CH3
OCH2CONHCH[COOC2H5](CH2)9CH3
OCH2CONHCH[COOC2H5](CH2)13CH3
OH
OH
OCH2CONHCH[COOC2H5](CH2)9CH3
OCH2CONHCH[COOC2H5](CH2)13CH3
OCH[COOC2H5](CH2)13CH3
OH
OCH[COOC2H5](CH2)13CH3

R. Pignatello et al. / European Journal of Pharmaceutical Sciences 10 (2000) 237–245
239
eluent systems reported below. Elemental analysis was
carried out on a Carlo Erba mod. 1106 analyzer; samples
were kept in vacuo for 24 h over P2O5 before analysis;
values found are within60.4% of theoretical ones.
HBTU, and 42 ml of DIEA. The activation of APA was
completed after 5 min. The activated APA was then
injected into the system and allowed to recirculate for 2 h
(with continuous absorbance check at 288 nm) until the
acylation cycle was completed. The resin was finally
washed with dry DMF (2320 ml) and then dry dichloro-
methane (2320 ml). The conjugate was then cleaved from
the resin by a 90-min treatment with 95% aqueous
trifluoroacetic acid (TFA). The crude product was precipi-
tated with ice-cooled diethyl ether and purified by semi-
preparative HPLC (see Section 2.4).
2.2. Chemicals
MTX (free acid; purity .98%, HPLC) was gifted by
Cyanamid (Catania, Italy). MTX-g-ethyl ester (Piper et al.,
1982) was a generous gift of Dr. J. R. Piper (Southern
Research Institute, Birmingham, AL, USA). 2-Aminoal-
kanoic acids (LAAs) were synthesised from their corre-
sponding 1-bromoalkanes and diethylacetamido malonate
(Gibbons et al., 1990). Bovine liver DHFR (specific
activity 7.8 U/mg protein), dihydrofolic acid and amino-
pteroic acid emihydrochloride (APA) were from Aldrich–
Sigma Chimica, (Milan, Italy). Solid phase synthesis
procedures were carried out using Novabiochem (Nottin-
gham, UK) products. N,N-dimethylformamide (DMF) was
distilled and dried over molecular sieves before use; all
other solvents and reactants were reagent grade or better
and did not need further purification.
2.3.2. Reaction of MTX mono-ester with lipoamino acids
The coupling reaction between MTX g-methyl ester
(Piper et al., 1982) and methyl a-aminodecanoate hydro-
chloride (Gibbons et al., 1990) was carried out as de-
scribed previously (Pignatello et al., 1998). Treatment of
the reaction crude product by column chromatography on
dry silica gel with increasing polarity eluent mixtures
(from pure dichloromethane to 7:3, v/v, dichloromethane–
methanol), gave the following compounds (FAB-MS peaks
and yield given): 1c (Pignatello et al., 1998) (m/z 822;
35%); MTX g-methoxy-a-(methyl 2-aminodecanoate) 1e
(m/z 653, 652; 24%); a mixture of MTX a-(methyl 2-
aminodecanoate) and g-(methyl 2-aminodecanoate) (1g, 1i)
(m/z 624). The latter mixture was subjected to a further
chromatographic separation with 7:3 (v/v) dichlorome-
thane–methanol, allowing the collection of the two iso-
mers. According to IR data and literature reports
(Rosowsky et al., 1978), the faster eluting compound was
identified as the monoconjugate having a free g-COOH
group (i.e., the a-substituted conjugate 1g). However, the
overall yield of these two products was very low (below
10%).
2.3. Synthesis of products
2.3.1. 1a and 1b
A NovaSyn TGR resin (Rink amide linker) (substitution
ratio: 0.2 mmol/g) (200 mg) was swollen at room
temperature in dry DMF for 60 min and placed in a
sintered glass peptide synthesis reaction vessel of a
NovaSyn Gem apparatus (Novabiochem). 2-[N-1-(4,4,-di-
methyl-2,6-dioxocyclohex-1-ylidene)ethyl]-aminohexade-
canoic acid (Kellam et al., 1998) (48.8 mg, 0.12 mmol) in
2 ml of DMF was coupled to the resin through the
assistance of 2-(1 H-benzotriazole-1-yl)-1,1,3,3-tetra-
methyluronium hexafluorophosphate (HBTU) (45.5 mg,
0.12 mmol) and in the presence of N-hydroxybenzotriazole
(HOBt) (10 mg, 0.06 mmol) and diisopropylethylamine
(DIEA) (42 ml, 0.24 mmol). The reaction mixture was
recycled until a 98% substitution was achieved (about 20
min, ninhydrin test). The resin was then washed with dry
DMF for 5 min and the protecting group was removed by
3-min treatment with a 2% hydrazine hydrate solution in
DMF; the deprotection was monitored at 288 nm. After a
5-min wash with DMF, the N-a-Fmoc-L-glutamic acid a
or g-t-butyl ester (51 mg, 0.12 mmol) was injected into the
synthesizer along with 10 mg of HOBt, 45.5 mg of HBTU,
and 42 ml of DIEA in 2 ml of dry DMF. The conjugation
was allowed to proceed and the continuous flow reaction
was monitored at 288 nm until a plateau was achieved,
determining the end of the acylation phase (about 30 min).
After the Fmoc group was removed by treatment with 20%
piperidine in DMF, the resin was washed for 5 min with
DMF. One equivalent of APA (41 mg) was dissolved in 2
ml of DMF and added with 10 mg of HOBt, 45.5 mg of
The same reaction was carried out with the methyl
a-aminotetradecanoate hydrochloride (Gibbons et al.,
1990), giving the corresponding derivatives 1d, 1f, 1h and
1j with the same order of elution (m/z 920, 709, and 680,
respectively; yields: 23%, 20%,,5%,,5%, respectively).
2.3.3. a-Bromoethanoyl-v-methoxy[imino-1(decyl-2-oxo-
1,2-ethanediyl)]
The compound was obtained by a modified procedure of
that described by Wood et al. (1992). Bromoacetic acid
(140 mg, 0.1 mmol), methyl 2-aminododecanoate hydro-
chloride (325 mg, 0.1 mmol) (Gibbons et al., 1990),
triethylamine (100 mg, 1 mmol), HOBt (155 mg, 1 mmol),
and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hy-
drochloride (EDAC) (200 mg, 1.1 mmol) were dissolved
in 5 ml dichloromethane and allowed to react at 08C for 2
h and then overnight at room temperature. The mixture
was diluted with 10 ml dichloromethane and extracted as
described above for compound 1c. Crystallization of the
residue from acetonitrile gave the title compound with a
70% yield. Analysis for C15H28BrNO3 (350.30): % calc.

240
R. Pignatello et al. / European Journal of Pharmaceutical Sciences 10 (2000) 237–245
(found)5C 51.43 (51.71), H 8.06 (8.29), N 4.00 (4.21).
FAB-MS (m/z, %): 374, 372 [M-H1Na]1 (8), 352, 350
[M]1 (19, 22), 330, 328 (12, 32), 308, 306 (34, 100), 248,
246 (23, 77).
dipotassium salt and methyl 2-bromoacetamido-hexade-
canoate: yield: 63%; FAB-MS (m/z, %): 1150 (10) [M1
2Na]1, 1128 (59) [M1Na]1, 1106 (33) [M11]1, 634
1
(23), 583 (34), 555 (100), 527 (54); H-NMR (CDCl3):
The corresponding higher homologous a-bromo-
ethanoyl-v-methoxy[imino-1(tetradecyl-2-oxo-1,2-eth-
anediyl)] was obtained using the procedure described
above, by the reaction of bromoacetic acid with methyl
a-aminohexadecanoate hydrochloride (Gibbons et al.,
1990). Analysis for C19H36BrNO3 (406.41): % calc.
(found)5C 56.15 (56.26), H 8.93 (9.07), N 3.45 (3.26).
Yield: 52%. FAB-MS (m/z, %): 430, 428 [M-H1Na]1
(28), 408, 406 [M]1 (57, 64), 364, 362 (14, 34), 348, 346
(58, 64), 226, 224 (100, 27), 154 (67), 137 (97), 109, 107
(52, 63). 1H-NMR (CDCl3): d57.05, 6.90 (1 H, 2d, NH;
J59 Hz), 4.58 (1 H, m, a-CH), 4.07, 3.89 (2 H, 2 s,
CH2Br), 3.76 (3 H, s, COOCH3), 1.90–1.85, 1.75–1.70 (2
H, 2 m, b-CH2), 1.28–1.25 (24 H, s, 12 CH2), 0.88 (3 H,
t, CH3; J513 Hz).
d58.60 (1 H, bs, H-7), 7.70, 6.75 (4 H, 2d, aromatic H;
J56.2 Hz), 7.40 (2 H, d, NH2; J56 Hz), 5.29 (2 H, s,
CH), 4.76 (2 H, s, 9-CH2), 4.62 (2 H, m, 2 a-CH),
3.80–3.65 (6 H, m, 2 OCH3), 3.49–3.40 (4 H, m,
COOCH2CO), 3.19 (3 H, s, NCH3), 1.85–1.65 (4 H, m,
bCHg2CH2CO), 1.43 (4 H, s, 2 CH2), 1.24 (48 H, s, 24
CH2), 0.89 (6 H, t, 2 CH3; J510 Hz); HPLC: RT 10.48
min.
Minor products (yield,10%) were 1n and 1p; FAB-
MS: m/z 780 [M11]1, 803 [M1Na]1); HPLC: RT57.9
min.
2.3.5. Ethyl 2-bromohexadecanoate
2-Bromohexadecanoic acid (Fluka, Sigma–Aldrich Srl,
Milan, Italy) was dissolved in a saturated solution of
hydrochloric acid in absolute ethanol, under stirring for 1 h
at room temperature. The crude product was purified by
flash column chromatography (dichloromethane/methanol,
10:1, v/v) and the ester was obtained as oil in a 74% yield
(FAB-MS: 366, 364 [M111]).
2.3.4. Compound 1k
MTX was suspended in methanol and the required
amount of a methanol–water (1:1, v/v) KOH solution was
added. After 10 min, the resulting yellow solution was
filtered, evaporated in vacuo and the residue freeze-dried to
give the MTX dipotassium salt. The latter (15 mg, 0.028
mmol) was dissolved in 5 ml methanol and 18-crown-6
ether (25 mg, 0.095 mmol) was added. After stirring at
room temperature for 30 min, the solvent was evaporated,
the residue treated twice with 10 ml benzene and dried in
vacuo. The salt (15 mg, 0.028 mmol) was then dissolved
in 2 ml anhydrous DMSO and methyl 2-bromoacetamido-
dodecanoate (23 mg, 0.057 mmol) was added. After
stirring for 24 h at room temperature, the solvent was
removed under high vacuum and the residue treated with a
small volume of water. The crude compound was purified
by column chromatography using 10:1 (v/v) dichlorome-
thane/methanol mixture.
2.3.6. Compound 1r
MTX dipotassium salt (11 mg, 0.02 mmol) was treated
with 15 mg 18-crown-6 ether in methanol. Ethyl 2-bromo-
hexanoate (15 mg, 0.04 mmol) dissolved in 1 ml dry DMF
was added and the mixture was left under stirring at room
temperature for 16 h. After evaporation under high vac-
uum, the residue was treated with 2 ml of cold water and
extracted with diethyl ether (3310 ml). Removal of the
solvent under vacuum gave a crude product which was
further purified by flash column chromatography, using
hexane (to remove the unreacted starting bromoester) and
then dichloromethane–methanol (8:2, v/v) as the eluent.
Compound 1r was collected with an overall 64% yield.
TLC, Rf: 0.52 (dichloromethane–methanol, 10:1, v/v);
FAB-MS (m/z, %): 1042 (24) [M1Na]1, 1020 (100)
[M11]1, 858 (22), 647 (41), 590 (43), 441 (77). 1H-NMR
(CDCl3): d58.65 (1 H, bs, H-7), 8.15 (1 H, m,
CONHGlu), 7.70, 6.70 (4 H, 2 m, aromatic H), 7.36 (2 H,
s, NH2), 4.80 (2 H, s, 9-CH2), 4.25 (4 H, q, 2 CH2), 4.23
(2 H, m, 2 a-CH), 3.20 (3 H, s, NCH3), 1.79–1.75 (4 H,
m, bCHg2CH2CO), 1.43 (4 H, s, 2 CH2), 1.30–1.25 (54 H,
5s, 24 CH212 COOCH2CH3), 0.88 (6 H, t, 2 CH3;
J513.5 Hz).
Yield: 48%; FAB-MS (m/z, %): 1032 (9) [M1K]1, 993
(100) [M11]1, 724 (63), 494 (23); 1H-NMR (CDCl3):
d58.63 (1 H, s, H-7), 7.70, 6.75 (4 H, 2s, aromatic H),
7.37 (2 H, m, NH2), 5.40 (2 H, bs, CH), 4.75 (2 H, s,
9-CH2), 4.62 (2 H, m, 2 a-CH), 3.71–3.64 (6 H, m, 2
OCH3), 3.19 (3 H, s, NCH3), 1.84–1.80 (4 H, bm,
bCHg2CH2CO), 1.26–1.22 (36 H, s, 18 CH2), 0.88 (6 H, t,
2 CH3; J56.6. Hz).
Compound 1o was obtained as a minor product (yield:
12%). MS m/z (%)5746 (3) [M1Na]1, 724 (10) [M1]1,
1
338 (66), 303 (100), 287 (94); H-NMR (CDCl3): d58.57
(1 H, s, H-7), 7.66, 6.83 (4 H, 2d, aromatic H; J58.5 Hz
and 8.7 Hz), 4.78 (2 H, s, 9-CH2), 4.47 (2 H, m, 2 a-CH),
3.81–3.53 (3 H, m, OCH3), 3.20 (3 H, s, NCH3), 1.75-
1.60 (4 H, bm, bCH2gCH2CO), 1.20 (18 H, s, 9 CH2), 0.85
(3 H, t, CH3; J56.7 Hz); HPLC, RT: 6.6 min.
2.4. Conjugate analysis and purification
The purity of the prepared compounds was assessed by
means of RP-HPLC, using a Waters 616/600S twin pump
system equipped with a 717plus automatic sampler and a
486 tunable UV detector monitored by a Millennium
Compound 1l was synthesized using the same procedure
described for compound 1k, by the reaction of MTX

R. Pignatello et al. / European Journal of Pharmaceutical Sciences 10 (2000) 237–245
241
software package. A Vydac C4 column (4.63150 mm)
linked to a Beckman Ultrasphere C8 pre-column was used,
at a constant flow-rate of 1.2 ml min21. Wavelength and
sensitivity were set at 310 nm and 1.0 AUFS, respectively.
Mobile phase consisted of 0.1% TFA (v/v) (A) and 90%
v/v acetonitrile/0.09% TFA (B); solvents were filtered
through a 0.2-mm polyamide filter and degassed by a
helium flow before use. The solvent gradient ranged
linearly from 0 to 100% solvent B in 10 min and then
decreased to 0% solvent B in 5 min. Test samples were
dissolved in 20% acetic acid and filtered through a 0.45-
mm polyamide filter before injection. Analysis of mixtures
of each compound with MTX always showed a good
separation of the two peaks under the above conditions.
Purification of crude reaction mixtures was also afforded
by semi-preparative RP-HPLC on a Vydac C6 column
(83150 mm). Separation was achieved with the same
solvents as above, using a linear gradient between 0 and
60% (solvent B) in 180 min, staying at these conditions for
30 min, and then decreasing steadily to 0% solvent B in 30
min, at a constant flow-rate of 6 ml min21. The gradient
was effected by two microprocessor-controlled Gilson 302
single-piston pumps. Eluates were detected by a Holoch-
rome UV–Vis detector at 300 nm and chromatograms were
recorded using a LKB 2210 single-channel recorder. The
identity of conjugates in the suitable collected fractions
was confirmed by FAB-mass spectrometry.
rpm. The supernatant was collected and diluted twice with
distilled water, and the mixture was injected in the HPLC
column and tested by the procedure described above.
2.6. DHFR inhibition assay
The drug inhibitory effect on DHFR activity was
measured by following NADPH oxidation at l5340 nm
(Pignatello et al., 1996). The assay mixture contained 1 M
sodium acetate buffer (pH 6.0), 0.6 M KCl, 50 mM
NADPH, 50 ml DMSO (control) or the same volume of a
DMSO solution of test compounds, to a final concentration
ranging between 1025 and 10211 M, and 0.02 units of
bovine liver DHFR, in a final volume of 3 ml. After
enzyme addition, the mixture was incubated at 238C for 5
min; reaction was started by adding 33 mM dihydrofolic
acid. The change in absorbance at l5340 nm (Dabs340
)
was followed for ten min, during which the activity was
linear with respect to time. Results are reported as percent
inhibition of the enzymatic activity vs. control (Fig. 1);
each value is the mean6S.D. of six experiments.
2.7. In vitro cell cytotoxicity tests
2.7.1. Cell growth inhibition assay
The human T-lymphoblastic leukemia cell line CCRF-
CEM (Foley et al., 1965) and an MTX-resistant (transport-
defective) cell line (CEM/MTX) (Rosowsky et al., 1980;
Westerhof et al., 1995) were grown at 378C, in a 5% CO2
2.5. Stability evaluation of MTX conjugates
The stability of some compounds, selected as prototypes
of the series of prepared conjugates, was evaluated both in
the same culture medium used for the cell growth inhib-
itory assays (see below) and in FCS, in order to test their
susceptibility toward the hydrolysis into free MTX.
For the first experiment MTX, MTX-diethyl ester
(Rosowsky, 1973) or conjugates 1c, 1k, and 1o were
dissolved in 100 ml DMSO at 1 mM concentration and the
solution added to 2 ml of complete RPMI 1640 culture
medium, in separate screw-capped vials. At time 0 and
after 12, 24 and 48 h of incubation at 378C, 250 ml of
medium were withdrawn and added with 500 ml of
chloroform. Samples were vigorously stirred and then
centrifuged at 108C, for 10 min at 5000 rpm. The organic
layer was discarded and samples of the aqueous phase
were spotted on TLC RP-18 glass sheets (Merck F254),
together with a pure (untreated) MTX specimen. After
elution with a pH 8.0 phosphate buffer, the compounds
were visualized by UV light at 254 nm.
To assay the hydrolytic stability of conjugates in a
biological medium, compounds 1c and 1k (1 mg) were
dissolved in 200 ml DMSO and the solution added to 1.8
ml of fresh FCS. Samples were vortexed and incubated at
378C (618C) in a water bath. At the decided time
intervals, 50 ml of the solution was withdrawn, added with
450 ml of ice-cooled methanol and centrifuged at 5000
Fig. 1. Bovine liver DHFR inhibition by MTX conjugates.

242
R. Pignatello et al. / European Journal of Pharmaceutical Sciences 10 (2000) 237–245
atmosphere, in RPMI 1640 (HyClone) supplemented with
10% fetal calf serum, glutamine (2 mM) and antibiotics.
Cells were plated in plastic 24-well plates at a con-
centration of 13105 cells per ml per well and test
compounds were concomitantly added in a concentration
range of 0.01–10 mM. Dilutions were made using the
culture medium from 1 mM stock solutions of the drugs in
anhydrous DMSO. Controls were treated with an equiva-
lent amount of solvent, diluted as above, to assess the
absence of toxicity due to the solvent.
After 72 h of incubation, viable cells were counted with
a haemocytometer by trypan blue exclusion assay. EC50
(drug concentration reducing viability to 50% that of
untreated control) were determined visually from plots of
relative growth (percent of untreated control) versus the
logarithm of drug concentration and are reported in Table
2. The experiments were repeated three times in duplicate.
3. Results
3.1. Chemistry
An isomeric pair of g- and a-substituted monoconju-
gates of MTX with 2-aminohexadecanoic acid 1a and 1b
were prepared using standard solid-phase peptide synthesis
methods (SPPS) (Merrifield, 1963). An N-acetyldimedone
(Dde)-protected 2-aminohexadecanoic acid (Kellam et al.,
1998) was linked to a TGR (Rink amide linker) resin
through HBTU assistance. After removal of the protecting
group with 2% hydrazine hydrate, N-a-Fmoc-L-glutamic
acid a- or g-tert.-butyl ester was added, followed by
cleavage with 20% piperidine. The deprotected compound
was then reacted with 2,4-diaminopteroic acid
hemihydrochloride (APA). Removal of the final compound
from resin by treatment with 95% aqueous TFA gave the
expected derivatives 1a,b with a very high purity (.95%,
HPLC). The major advantages of this new method, com-
pared with those described in the literature, were (i) the
easy Dde deprotection and (ii) the use of non-toxic HBTU
as the activating agent.
Monosubstituted conjugates of MTX 1c–j were also
obtained as side products from the reaction between MTX-
g-methyl ester and LAAs. Separation of the of a- and
g-substituted isomer (lower Rf) was accomplished by
column chromatography, by taking advantage from the
little difference in the pKa of the two carboxyl functions
(Rosowsky et al., 1978; Rahman and Cchabra, 1988). The
structure elucidation of the compounds was accomplished
by using modern physicochemical methods functions
(Rosowsky et al., 1978; Antonjuk et al., 1984a; Rahman
and Cchabra, 1988). It is noteworthy that a partial hy-
drolysis of MTX-g-methyl ester occurred in the reaction
medium, resulting in a free g-COOH group, which reacted
with the LAA residues present in the reaction mixture
(Rosowsky et al., 1978).
2.7.2. Continuous drug exposure test
The CCRF-CEM cell line and its MTX-resistant sub-line
CEM/MTX were cultured as described in RPMI 1640
medium (GIBCO) supplemented with 10% horse serum
(McCloskey et al., 1991); growth inhibition by continuous
drug exposure for 120 h was measured as described
(McCloskey et al., 1991). Both cell lines (generation
times: CCRF-CEM, 20.460.1 h; CEM/MTX, 19.960.6 h)
grew logarithmically for 120 h if seeded at 104 cells per ml
at tzero. The 120-h exposure time increases the sensitivity
of the growth inhibition assay by allowing the maximum
number of logarithmic cell doublings to occur. MTX was
dissolved in water; compounds 1a, 1b and 1j were
dissolved in DMSO. Solvent controls were included in
each experiment and only data generated at DMSO con-
centrations that were themselves nontoxic (#0.23%
DMSO for both cell lines) were considered valid. Both cell
lines were verified to be negative for Mycoplasma con-
tamination using the GenProbe test kit. EC50 values were
determined as reported above (Table 2).
MTX–LAAs conjugates with an ester linkage 1k–p
were obtained by crown-ether assisted condensation. First
Table 2
In vitro growth inhibitory activity (EC50, nM) of MTX and conjugates against MTX-sensitive CCRF-CEM cells and the transport-resistant CEM/MTX
sub-line
Drug
CCFR-CEM
(72-h)
CEM/MTX
(72-h)
Relative resistancea
fold
CCRF-CEM
(120-h)
CEM/MTX
(120-h)
Relative resistancea
fold
1a
1b
1j
1k
1o
1p
MTX
385
900
85
35
60
390
.5000
.5000
1900
10 000
570
1.01
n.d.
n.d.
54.3
167
2.1
310
1800b
.10 000
.1000
–
–
–
2200
5.8
n.d.
n.d.
–
c
5800b
.1000
–
–
–
–
270
30
–
4300
143
14
157d
a Relative resistance is the ratio of EC50 (CEM/MTX) to EC50 (CCRF-CEM).
b In a second experiment, the EC50 was not reached but the trend in the data allowed extrapolation to an approximate EC50 which was consistent with the
value presented. For 1a and CEM/MTX, the extrapolated EC50 was 1300 nM; for 1b and CCRF-CEM cells, the extrapolated EC50 was 6000 nM.
c n.d.5value cannot be determined.
d In separate experiments, relative resistance of CEM/MTX sub-line toward MTX ranged between 100- and 150-fold.

R. Pignatello et al. / European Journal of Pharmaceutical Sciences 10 (2000) 237–245
243
dipotassium salt of MTX was reacted with crown-16-ether
and then with the methyl ester of the desired LAA
(Gibbons et al., 1990) in DMSO or dichloromethane. The
purification of 1k–p was achieved by column chromatog-
raphy. A similar reaction was used to obtain the long-chain
aliphatic MTX a,g-bis(ester) 1r.
Qualitative stability studies were carried out on selected
terms, i.e., an a,g-bis(amide) conjugate (1c) and two ester
derivatives (1k and 1o), by testing the appearance of free
MTX, under the same conditions used for the cell out-
growth inhibition assays. Neither compound 1c or the ester
derivatives 1k and 1o were converted into MTX after 48 h
of incubation in the culture medium, whereas MTX diethyl
ester already was partially hydrolyzed into both the
monoethyl ester and the free drug after 24 h of incubation.
MTX did not display any degradation sign after 48 h of
incubation at 378C.
After incubation in FCS the results are divergent for the
amide and ester-type conjugates. Amide conjugate 1c was
almost stable after 5 days of incubation, giving only traces
of free MTX (,3%, based on the HPLC peak area). On
the contrary, the ester 1k was rapidly degraded: after 15
min a 20% MTX was detected in the incubation mixture
(retention time: 3.85 min), and its concentration remained
almost constant up to 8 h incubation. In the meantime, the
disappearance of the starting compound (retention time:
8.83 min) was associated by increasing levels of two
degradation products (retention times: 6.30 and 6.90 min),
corresponding to the isomeric pair of monosubstituted
conjugates 1m and 1o, as confirmed by the FAB-MS
analysis on the collected HPLC fraction. Their combined
concentration in the assay medium went from 22% after 15
min of incubation to 55% after 8 h.
Fig. 2. Growth inhibitory activity of 1a and MTX against sensitive and
resistant CCRF-CEM cells.
relative resistance) gave divergent results. In fact, only
g-substituted monoconjugates (1a and 1p) maintained a
similar potency against both cell lines, whereas MTX was
about 150 times less active against the resistant sub-line
than towards the parental line. In Fig. 2, the concentration–
activity profile is enlightened for compound 1a and MTX.
Compounds 1a, 1b and 1j were compared to MTX as
inhibitors of MTX-sensitive CCRF-CEM human leukemia
cells growth in horse serum (instead of FCS, as above) also
during continuous exposure in vitro (Table 2). Compounds
1a and 1b were 22-fold and 414-fold less potent than MTX
as growth inhibitors of CCRF-CEM cells. The well-char-
acterized MTX-transport-defective CEM/MTX sub-line
was 157-fold resistant to MTX, but only slightly cross-
resistant to 1a (1b did not affect cell growth at 10 mM, the
highest concentration testable).
By comparing the results from the two experiments,
compound 1a displayed the same order of cytotoxicity
both in the 72-h and 120-h incubation assays (relative
resistance of 1.01 and 5.8, respectively). The a-substituted
isomer 1b showed a lower activity against sensitive cells
and was almost inactive against the resistant CEM/MTX
sub-line.
3.2. Biochemical screening
Some of the prepared conjugates were tested against
DHFR as representative terms of the different series of
lipophilic MTX conjugates. All of them showed an inhib-
itory activity against bovine liver DHFR (Fig. 1). Com-
pounds 1b and 1g, substituted on the a-carboxyl function
were less potent than MTX. However, the g mono-amide
conjugates 1a and 1j, and the a,g-bis(ester) conjugate 1k
showed a similar or greater inhibitory activity than the
parent drug.
3.3. In vitro tumor cell cytotoxicity assay
To perform these assays selected compounds were also
chosen, belonging to the three series of lipophilic MTX
conjugates. Against the MTX-sensitive (parental) CCRF-
CEM cell line, compounds 1a, 1b, 1j and 1p were less
active than MTX, but the activity of ester conjugates 1k
and 1o was comparable with that one of the parent drug
(Table 2). However, data relative to the activity exerted
against the transport-deficient resistant cells (i.e., the
4. Discussion
In general, the inhibitory activity of the described
lipophilic MTX conjugates was lower than the activity of
the parent drug against sensitive cells. However, some

244
R. Pignatello et al. / European Journal of Pharmaceutical Sciences 10 (2000) 237–245
conjugates showed the same cytotoxicity to CEM/MTX
cells, which were primarily resistant to MTX by virtue of
an impaired cellular drug uptake. Similar findings have
been reported for other lipophilic antifolates and MTX
derivatives (Rosowsky et al., 1980; Piper et al., 1982;
Rosowsky et al., 1984; Kinsky and Loader, 1987; Rahman
and Cchabra, 1988; Pignatello et al., 1996, 1998, 1999).
Since the lack of drug transport represents one important
form of resistance to MTX in both human and murine cell
lines (Sirotnak, 1980; Westerhof, 1995), the alterations of
the transport kinetics have a potential clinical value.
Comparison of results relative to DHFR activity and
tumor cell growth inhibition by the conjugates, confirmed
that the presence of a free carboxylic group in the a
position of MTX glutamate moiety, is important for the
binding to the target enzyme (Antonjuk et al., 1984b,
1989); as expected, the g-mono substituted compounds 1a,
1j and 1k inhibited DHFR at the same level than MTX.
The g- and a-substituted mono-amide conjugates 1a and
1b exhibited a very different affinity for DHFR and in
vitro inhibition potency against CCRF-CEM cell growth.
The g-substituted isomer 1a showed a higher inhibitory
activity against DHFR than the corresponding a-substi-
tuted derivative 1b, with a higher IC50 value (0.034 nM)
than MTX itself (0.37 nM) (Fig. 1). Similarly, compound
1a was 20–50 times more potent than 1b against sensitive
CCRF-CEM cells and resistant (CEM/MTX) sub-line
(Table 2).
Conjugate 1a showed an activity profile close to that one
of MTX. However, the lipophilic conjugate retained its
activity against the MTX-uptake defective CEM/MTX cell
sub-line, showing a relative resistance of 1.02 and 5.8,
after 72-h and 120-h drug exposure respectively (Table 2
and Fig. 2.). As expected, MTX was much less potent in
both assays. CEM/MTX cells become less sensitive to
compound 1a with exposure time. Its EC50 value was in
fact 393 nM after 72-h drug exposure and 1800 nM after
120-h exposure. For compound 1b differences were less
marked, while MTX displayed similar EC50 values in both
the culture conditions. Several interpretations can be given
to these results: conjugate 1a may slowly be degraded into
MTX in the long-term (120-h) culture assay, and a time-
dependent release of free MTX may occur in the external
medium, leading to a progressive reduction of cell growth
inhibition. It is noteworthy that we used different culture
media (FCS or horse serum) in the two in vitro assays, and
this could induce a different hydrolytic activity upon the
MTX conjugates.
number of logarithmic growth doublings to occur during
the assay period. The longer exposure period also allows
delayed growth inhibition to be detected, if it occurs. The
greater dynamic range allows more closely spaced drug
concentrations to be tested (typically spacing of, for
example, 10 nM, 20 nM, 50 nM, 100 nM, 200 nM, etc.).
This spacing allows a more accurate interpolation of the
EC50 value than the wider ten- to 100-fold spacing used in
the 72-h viability studies (Fig. 2). It should also be noted
that the 120-h assay measures growth (based on particle
counts), while the shorter assay measures actual cell
viability (based on trypan blue dye exclusion). Thus the
results of the two assays may not coincide if cells are
killed (not viable), but do not lyse (particle intact).
Compound 1j (a g-mono-substituted amide conjugate)
showed a higher affinity than MTX for DHFR (IC50,0.01
nM); it also showed an activity close to MTX against
sensitive CCRF-CEM cells, but was inactive against MTX-
resistant cells (EC50 .1000 nM; data not shown). It would
then seem that it was not able to take advantage of its
lipophilic character to penetrate passively into the trans-
port-resistant cell lines.
An important consideration about the lipophilic conju-
gates of MTX is that their biological activity can either be
the result of a direct effect on cell enzymes, or of being
mediated by the intracellular release of free MTX, that will
ultimately be responsible for blocking cell outgrowth. For
this reason, we prepared further compounds where MTX
and the LAA residue were linked through an ester bond
(compound 1r), or an ester–amide bond (compounds 1k–
1p). The hydrolysis rate of an ester in the biological
environment is normally faster than the amide bonds
(Bundgaard, 1985), so we expected that these compounds
would act better as MTX prodrugs.
In our experiments, the biological activity of compounds
1k and 1o was very close to that of MTX against both the
parent and resistant cell lines (72-h incubation time) (Table
2). By considering that these ester conjugates were not
converted into MTX when incubated in the culture
medium (see Results), it seems that they are unable to enter
resistant cells via passive routes, despite their lipophilic
character. However, the gradual hydrolysis into the corre-
sponding monosubstituted esters and then free MTX that
we observed after incubation of some ester conjugates in
fetal calf serum, made the conjugate 1k a potential prodrug
for the slow systemic release of MTX.
Alternatively, by considering that these conjugates could
not be subjected to intracellular polyglutamylation
(McGuire, 1998), a steady-state equilibrium between drug
influx and efflux could occur, lowering the effective
concentration of the inhibitor inside the tumor cells.
In any case, it must be considered that the 120-h
exposure time increases the sensitivity and dynamic range
of the growth inhibition assay by allowing the maximum
5. Conclusions
We have confirmed that lipophilic antifolates, like
MTX–LAA conjugates can be used as lead compounds for
the development of drugs for MTX-resistant tumors.
Lipophilic derivatization of MTX glutamate allowed the
passive entry of the drug into tumor cells. The activity of
the amide conjugates of MTX 1a and 1p, which showed an

R. Pignatello et al. / European Journal of Pharmaceutical Sciences 10 (2000) 237–245
245
McCloskey, D.E., McGuire, J.J., Russell, C.A., Rowan, B.G., Bertino,
J.R., Pizzorno, G. et al., 1991. Decreased folylpolyglutamate synthet-
ase activity as a mechanism of methotrexate resistance in CCRF-CEM
human leukemia sublines. J. Biol. Chem. 266, 6181–6187.
McGuire, J.J., 1998. Antifolate polyglutamylation in preclinical and
clinical antifolate resistance. In: Jackman, A.L. (Ed.), Anticancer Drug
Development Guide: Antifolate Drugs in Cancer Therapy, Humana
Press, Totowa, pp. 339–363.
Merrifield, R.B., 1963. Solid phase synthesis of a tetrapeptide. J. Am.
Chem. Soc. 85, 2149–2154.
Pignatello, R., Sorrenti, V., Spampinato, G., Pecora, T., Panico, A., Di
Giacomo, C. et al., 1996. Synthesis and preliminary in vitro screening
of lipophilic a,g-bis(amides) as potential prodrugs of methotrexate.
Anti-Cancer Drug Design 11, 253–264.
Pignatello, R., Spampinato, G., Sorrenti, V., Vicari, L., Di Giacomo, C.,
Vanella, A. et al., 1999. Aliphatic a,g-bis(amides) of methotrexate.
Influence of chain length on in vitro activity against sensitive and
resistant tumor cells. Pharm. Pharmacol. Commun. 5, 299–305.
Pignatello, R., Jansen, G., Kathmann, I., Puglisi, G., Toth, I., 1998.
Lipoamino acid conjugates of methotrexate with antitumor activity. J.
Pharm. Sci. 87, 367–371.
Piper, J.R., Montgomery, J.A., Sirotnak, F.M., Chello, P.L., 1982.
Syntheses of a- and g-substituted amides, peptides, and esters of
methotrexate and their evaluation as inhibitors of folate metabolism. J.
Med. Chem. 25, 182–187.
intrinsic potency against cell growth, was not due to the
release of free MTX inside the cells.
The therapeutic association of MTX with one of its
lipophilic derivatives may prevent the development of
transport resistance to the drug, which has often been
observed during clinical use of MTX. Moreover, the
described lipophilic MTX conjugates would not be ex-
pected to form g-polyglutamates once they enter a cell,
unlike the parent drug. This distinctive property has
potential therapeutic implications for the treatment of
certain MTX-resistant tumors whose resistance may be
associated with a lower than normal capacity to form
g-polyglutamates, in comparison with highly proliferative
tissues such as intestinal mucosa or marrow (McCloskey et
al., 1991; McGuire, 1998).
Acknowledgements
The present work was supported by an Italian
M.U.R.S.T., C.N.R., and A.I.R.C. grants, and by an award
from the Roswell Park Alliance Foundation (JJM).
Rahman, L.K.A., Cchabra, S.R., 1988. The chemistry of methotrexate and
its analogues. Med. Res. Rev. 8, 95–156.
Rosowsky, A., 1973. Methotrexate analogs. 2. A facile method of
preparation of lipophilic derivatives of methotrexate and 39,59-dich-
loromethotrexate by direct esterification. J. Med. Chem. 16, 1190–
1193.
References
Rosowsky, A., Beardsley, G.P., Ensminger, W.D., Lazarus, H., Yu, C.-S.,
1978. Methotrexate Analogues. 11. Unambiguous chemical synthesis
and in vitro biological evaluation of a- and g-monoesters as potential
prodrugs. J. Med. Chem. 21, 380–386.
Rosowsky, A., Lazarus, H., Yuan, G.C., Beltz, W.R., Mangini, L.,
Abelson, H.T. et al., 1980. III. Effects of methotrexate esters and other
lipophilic antifolates on methotrexate-resistant human leukemia lym-
phoblasts. Biochem. Pharmacol. 29, 648–652.
Rosowsky, A., Forsch, R.A., Yu, C.S., Lazarus, H., Beardsley, G.P., 1984.
Methotrexate Analogues. 21. Divergent influence of alkyl chain length
on the dihydrofolate reductase affinity and cytotoxicity of methotrex-
ate monoesters. J. Med. Chem. 27, 605–609.
Sirotnak, F.M., 1980. Correlates of folate analog transport, phar-
macokinetics and selective antitumor action. Pharmacol. Ther. 8,
71–103.
Sirotnak, F.M., Chello, P.L., Piper, J.R., Montgomery, J.A., DeGraw, J.I.,
1979. In: Kisliuk, R.L., Brown, G.M. (Eds.), Chemistry and Biology of
Pteridines, Elsevier/North-Holland, New York, pp. 597–608.
Westerhof, G.R. (1995) Folate and antifolate transport in tumour cells,
Ph.D. Thesis. Febodruk, Enschede, Leiden.
Westerhof, G.R., Schornagel, J.H., Kathmann, I., Jackman, A.L.,
Rosowsky, A., Forsch, R.A. et al., 1995. Carrier and receptor-mediated
transport of folate antagonists targeting folate-dependent enzymes,
correlates of molecular-structure and biological activity. Molec. Phar-
macol. 48, 459–471.
Antonjuk, D.J., Boadle, D.K., Cheung, H.T.A., Tran, T.Q., 1984a.
Synthesis of monoamides of methotrexate from L-glutamic acid
monoamide t.-butyl esters. J. Chem. Soc. Perkin Trans. 1, 1989–2003.
Antonjuk, D.J., Birdsall, B., Cheung, H.T.A., Clore, G.M., Feeney, J.,
Gronenborn, A. et al., 1984b. A 1H NMR study of the role of the
glutamate moiety in the binding of methotrexate to Lactobacillus casei
dihydrofolate reductase. Br. J. Pharm. 81, 309–315.
Antonjuk, D.J., Boadle, D.K., Cheung, H.T.A., Friedlander, M.L., Gre-
gory, P.M., Tattersall, M.H.N., 1989. a-Monoamides of methotrexate
as potential prodrugs. Arzeinem.-Forsch./Drug Research 39, 12–15.
Bundgaard, H., 1985. Design of prodrugs: bioreversible derivatives for
various functional groups and chemical entities. In: Bundgaard, H.
(Ed.), Design of Prodrugs, Elsevier, Amsterdam–New York–Oxford,
pp. 1–92.
Foley, G.F., Lazarus, H., Farber, S., Uzman, B.G., Boone, B.A., McCar-
thy, R.E., 1965. Continuous culture of lymphoblasts from peripheral
blood of a child with acute leukemia. Cancer 18, 522–529.
Gibbons, W.A., Hughes, R.A., Charalambous, M., Christodoulou, M.,
Szeto, A., Aulabaugh, A.E. et al., 1990. Synthesis, resolution and
structural elucidation of fatty amino acids and their homo- and hetero-
oligomers. Liebigs Ann. Chem., 1175–1183.
Henderson, G.B., Blakley, R.L., 1986. In: Whitehead, V.M. (Ed.), Folates
and Pterins, Vol. 3, John Wiley, New York, pp. 207–250.
Kellam, B., Drouillat, B., Dekany, G., Starr, M.S., Toth, I., 1998.
Synthesis and in vitro evaluation of lipoamino acid and carbohydrate
modified enkephalins as potential antinociceptive agents. Int. J. Pharm.
161, 55–64.
Wood, I.P., Toth, I., Holley, J.L., Gibbons, W.A., 1992. Lipidic peptides.
XIII: synthesis, structure elucidation and in vitro toxicity assessment
of chlorambucil conjugates with lipidic acids, lipidic amino acids and
their oligomers. Int. J. Pharm. 87, 141–147.
Kinsky, S.C., Loader, J.E., 1987. Circumvention of the methotrexate
transport system by methotrexate-phosphatidylethanolamine deriva-
tives: effect of fatty acid chain length. Bioch. Biophys. Acta 921,
96–103.