Synthesis and antiproliferative activity of alkylphosphocholines

Article abstract of DOI:10.1016/j.chemphyslip.2003.08.002

Alkylphosphocholines (APC) with one or more methylene groups in the alkyl chain replaced by oxygen atoms or carbonyl groups, or both have been assembled modularly using ω-diols as central building blocks. Out of 25 new compounds of this kind, 11 were evaluated for their antiproliferative activity on four cell lines and compared with miltefosine to evaluate their hemolytic activity (HA) and cytotoxicity on non-tumoral cells (MT2), used as markers of adverse effects. Compound 13 was more active on cancer cell lines than on non-tumoral cells and the data were similar for MTT and thymidine incorporation assays. It had less HA than miltefosine. Compound 13 could therefore be a candidate for the preparation of compounds with higher cytotoxicity on cancer cells and lower general toxicity.

Full text of DOI:10.1016/j.chemphyslip.2003.08.002

Chemistry and Physics of Lipids 126 (2003) 201–210
Synthesis and antiproliferative activity of alkylphosphocholines
Mandy Agrestaa, Paola D’Arrigoa, Ezio Fasolia, Daniele Losib,
Giuseppe Pedrocchi-Fantonia, Simona Rivab, Stefano Servia,∗, Davide Tessaroa
a
Dipartimento di Chimica, Materiali, Ingegneria Chimica “Giulio Natta”, Politecnico di Milano and CNR, Istituto di Chimica del
Riconoscimento Molecolare “Adolfo Quilico”, Via Mancinelli 7, 20131 Milano, Italy
b
Biosearch Italia, Via Roberto Lepetit 34, 21040 Gerenzano, VA, Italy
Received 19 April 2002; received in revised form 31 July 2003; accepted 13 August 2003
Abstract
Alkylphosphocholines (APC) with one or more methylene groups in the alkyl chain replaced by oxygen atoms or carbonyl
groups, or both have been assembled modularly using ␻-diols as central building blocks. Out of 25 new compounds of this kind,
11 were evaluated for their antiproliferative activity on four cell lines and compared with miltefosine to evaluate their hemolytic
activity (HA) and cytotoxicity on non-tumoral cells (MT2), used as markers of adverse effects. Compound 13 was more active
on cancer cell lines than on non-tumoral cells and the data were similar for MTT and thymidine incorporation assays. It had less
HA than miltefosine. Compound 13 could therefore be a candidate for the preparation of compounds with higher cytotoxicity
on cancer cells and lower general toxicity.
© 2003 Elsevier Ireland Ltd. All rights reserved.
Keywords: Alkylphosphocholines; Miltefosine; Antitumor activity
1. Introduction
et al., 1980). Alkylphosphocholines (APC) are the
largest group of phosphocholine derivatives inves-
tigated for their biomedical applications since their
antitumoral activity has been recognized (Eibl et al.,
1992; Kotting et al., 1992; Brachwitz and Vollgraf,
1995): they constitute a new class of antineoplastic
agents interfering with membrane signal transduction
pathways.
The clinical development of the lead compound,
hexadecylphosphocholine (see Fig. 1: miltefosine)
(Unger and Eibl, 1990), for oral administration in
patients with solid tumours was prevented by its
dose-limiting gastrointestinal toxicity. Miltefosine is
used by topical application, the intravenous route be-
ing excluded because of the high hemolytic activity
(HA).
Phosphatidylcholines (PCs) and derivatives have
been subjected to considerable synthetic and bio-medical
investigation for their use as enzyme inhibitors (Cox
et al., 1979; Chandrakumar and Hajdu, 1982; Yuan
et al., 1987; Magolda and Johnson, 1985), platelet
activating factors (Ohno et al., 1985; Grivilsdalsky
and Bittman, 1989; Terashita et al., 1983; Nakamura
et al., 1990), CNS and cardiac agents (Puricelli, 1989),
and antifungal, antimicrobial (Bierer et al., 1995;
Tsushima et al., 1982) and antitumor agents (Runge
∗
Corresponding author. Tel.: +39-02-23993047;
fax: +39-02-23993080.
E-mail address: stefano.servi@polimi.it (S. Servi).
0009-3084/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.chemphyslip.2003.08.002

202
M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
2. Biological activity
O
O P O
+
N
-
O
It has been proposed that APC act in vivo as
inhibitors of mitogenic signalling (Grunicke et al.,
1996), or that their antiproliferative activity is due
to the inhibition of phosphatidylinositol specific
phospholipase C. APC, being phospholipid (PL) ana-
logues, show antitumor activity because they interfere
with enzymes acting on or regulated by phospho-
lipids. Protein kinase C (PKC) can be activated by
phosphatidic acid (PA) generated by phospholipase D
(PLD)-catalysed hydrolysis of PL. Hydrolysis of PC,
catalysed by PC-specific phospholipase C (PCPLC),
generates diacylglycerol (DAG), a potent activator of
PKC. APCs, competing with the natural substrates,
may interfere with PKC activation, inactivating cell
growth.
miltefosine
perifosine
1
O
O
P O
+
N
-
O
2
Fig. 1. APC with antitumor activity.
In order to obtain compounds with better antiprolif-
erative action and less hemolysis (Berger et al., 1990;
Brachwitz et al., 1996, 1997) or compounds able to
influence apoptosis, a number of structural analogues
have been prepared (Matzke et al., 2000, 2001). New
APC have been tested in culture systems of human
tumor cells (Matzke et al., 2001).
If this is the function of APCs, the activity could
be improved better in structurally related compounds
with known ability as substrates of PL hydrolyzing
enzymes.
We recently examined PL analogues as substrates
for bacterial PLDs (Bossi et al., 2001) and found
that among simple phosphodiesters, the hydrolysis rate
could be correlated with substrate polarity. Methylene
group substitution with an oxygen atom, would be a
Perifosine (Fig. 1), the heterocyclic analogue of
miltefosine, is a promising candidate among the nu-
merous compounds obtained by modification of either
the polar head or alkyl chains (Matzke et al., 2000;
Traiser et al., 1998). Here we report the synthesis
and properties of a series of alkylphosphocholines,
in which one or more methylene groups in the alkyl
chain have been substituted with oxygen atoms or
carbonyl groups, or both.
HO
HO
PhCH2Cl
O-Na+
+
n O
Ph
n
4
3
NaH
THF
Br
m
O
O
n O
n OH
Ph
m
m
6
5
POCl3, Et3N
Choline Tos, Py
H2O
O
P
+
N
O
n O
O
m
-
O
7
m = 1, 3, 5, 7, 9, n = 1, 2, 3, 7, 9
Scheme 1.

M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
203
simple way to improve the polarity of APCs, making
them more competitive with the natural PLs.
All solvents and reagents were of analytical grade.
Pyridine was refluxed over, distilled from and stored
over KOH pellets. Ethanol-free chloroform was dried
and distilled over CaCl2. N,N-Dimethylformamide
was dried over molecular sieves. Phosphorus oxy-
chloride and triethylamine were distilled before
use.
3. Synthesis of new oxa-APC
Compounds of type 7 in Scheme 1 were assembled
modularly using type 3 ␻-diols as central building
blocks. Thus, the type 3 diols with n + 1carbon atoms
were protected at one end as benzyl ethers, elongated
with alkyl bromides of m + 1carbon atoms to give
compounds of type 5, which were deprotected by hy-
drogenolysis to the primary alcohols 6. These were
then transformed into the corresponding PC through
the procedure outlined in Scheme 1. The compounds
obtained were tested for their HA properties and only
those with reduced HA were considered as candi-
dates for further evaluation. Hemolysis was tested as
described in the experimental part (Mosmann, 1983;
Sims and Page, 1984; Alley et al., 1988).
Seven compounds (8–14) out of the 25 synthe-
sised were selected as having equal or lower HA than
miltefosine. To this set of substances we added four
additional compounds (15–18) with significant struc-
tural differences. The ether linkage in the previous
set of compounds was replaced by an oligoethyleneg-
lycol chain, an ester linkage or an –OH group, which
should help clarify the role of the ethereal oxygen
atom in the previous set of lipids (Fig. 2). Finally,
the compounds were tested for their cytotoxic ac-
tivity on four cell lines (Mosmann, 1983; Sims and
Page, 1984; Alley et al., 1988) using two different
assays.
4.1. ω-Hydroxybenzylethers (4)
NaH (1.1 mol) was suspended in DMF (330 ml) and
the diol (1 mol) was added dropwise. After stirring at
50 ◦C for 10 h, the mixture was allowed to cool at room
temperature. Benzylchoride (1 mol) was then added
dropwise and the reaction mixture was left under stir-
ring at room temperature overnight. After addition of
a 5% HCl solution, the mixture was extracted with
ethyl acetate or hexane. The extract was washed with
water, dried and concentrated. The residue was chro-
matographed on silica gel to give the desired product
(yield 60–70%).
4.2. ω-Benzyloxyalkyl alkylethers (5)
␻-Hydroxybenzylether 4 (1 mol) was added drop-
wise to a suspension of NaH (1.1 mol) in THF
(660 ml). After stirring at 50 ◦C for 4 h, the reaction
mixture was allowed to cool at room temperature. The
alkylbromide (1 mol) was added slowly dropwise in
THF. The mixture was stirred for 12 h, acidified with
HCl 1N and evaporated to dryness. The residue was
then purified by chromatography on silica gel giving
the desired product (yield 70–75%).
4.3. ω-Hydroxyalkyl-alkylethers (6)
4. Materials and methods
The di-ethers 5 in ethanol (1–1.5 M) were treated
with Pd/C (5%) and the mixture was hydrogenated at
room temperature (r.t.) for several hours. The mixture
was then filtered and the filtrate was evaporated to
dryness. Products 6 were used as such.
1H NMR analysis was run on a Varian EMX
250 MHz instrument. H chemical shifts are reported
1
in δ units (ppm) relative to tetramethylsilane (TMS)
as internal standard and all spectra were recorded
in CDCl3/CH3OH (2/1) unless otherwise indicated.
Silica gel 60 F254 plates (Merck) were used for ana-
lytical TLC in solvent systems of hexane/ethylacetate
(8/2), CHCl3/CH3OH/H2O in different ratios. Detec-
tion was with UV light or by spraying with a 4%
solution of phosphomolybdic acid in 10% H2SO4,
followed by heating.
4.4. Alkyloxophosphocholines (7)
A solution of ␻-hydroxyalkyl-alkylethers 6 (1 mol)
in CHCl3 was added dropwise to a solution of POCl3
(1.3 mol) and TEA (1.3 mol) in CHCl3, keeping the
temperature around 0 ◦C with external cooling. The

204
M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
O
+
O P O
8
O
N
O
-
O
+
O P O
O
9
N
O
-
O
+
O P O
O
10
N
O
-
O
+
O P O
11
O
N
O
-
O
+
O P O
12
13
O
N
O
-
O
+
O P O
O
N
O
-
O
+
O
O P O
14
17
N
O
-
O
O
+
O P O
O
N
O
-
O
O
+
O P O
18
16
O
N
O
-
O
+
O
O
O P O
O
O
O
N
O
-
O
O
O
+
O P O
15
HO
N
O
-
Fig. 2. oxa-APC with HA lower or equal to miltefosine.

M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
205
mixture was allowed to reach r.t. and stirred for 3 h,
4.9. Compound 12
then cooled to 10 ◦C. A warm solution (50 ◦C) of
choline tosylate (1.5 mol) in pyridine (8 mol) was then
added slowly and the reaction mixture was left un-
der stirring for two days at r.t. After addition of water
(15 mol), the mixture was concentrated. The residue
was treated with CH2Cl2/toluene 1:1, filtered and the
filtrate evaporated to dryness. The residue was dis-
solved in THF/H2O, 9:1, and chromatographed on
Amberlite MB3 ionic exchange resin. The product
was purified by chromatography on silica gel (yield
50–65%).
Analysis calcd. for C23H50NPO5 (451) H 11.16, C
61.17, N 3.10; found H 11.11, C 61.24, N 3.17. Rf
0.4 (CHCl3/CH3OH/H2O 65/25/4); [M+H] 452.6; 1H
NMR δ: 0.93 (t, 3H); 1.36 (m, 22H); 1.64 (m, 6H␤);
3.28 (s, 9H); 3.47 (t, 4H␣); 3.67 (m, 2Hcol); 3.93 (q,
2H␣1); 4.3 (m, 2Hcol).
4.10. Compound 13
Analysis calcd. for C23H50NPO5 (451) H 11.16, C
61.17, N 3.10; found H 11.18, C 61.09, N 3.07. Rf
0.33 (CHCl3/CH3OH/H2O 65/25/4); [M + H] 452.6;
1H NMR δ: 0.91 (t, 3H); 1.3 (m, 22H); 1.55 (m, 4H␤);
1.68 (m, 2H␤); 3.26 (s, 9H); 3.41 (t, 4H␣); 3.8 (m,
2Hcol); 4.05 (q, 2H␣1); 4.4 (m, 2Hcol).
4.5. Compound 8
Analysis calcd. for C21H46NPO5 (423) H 10.95, C
59.55, N 3.31 O; found H 11.04, C 59.34, N 3.22. Rf
0.4 (CHCl3/CH3OH/H2O 65/25/4); [M + H] 424.4;
1H NMR δ: 0.88 (t, 3H); 1.3 (m, 18H); 1.57 (m, 6H␤);
3.22 (s, 9H); 3.41 (t, 4H␣); 3.59 (m, 2Hcol); 3.85 (q,
2H␣1); 4.22 (m, 2Hco1).
4.11. Compound 14
Analysis calcd. for C18H40NPO5 (381) H 10.57, C
56.67, N 3.67; found H 10.44, C 56.71, N 3.69. Rf
0.2 (CHCl3/CH3OH/H2O 65/25/4); [M+H] 382.4; 1H
NMR δ: 0.98 (t, 3H); 1.25 (m, 14H); 1.57 (m, 2H␤);
1.9 (m, 2H␤); 3.25 (s, 9H); 3.40 (t, 4H␣); 3.55 (m,
2Hcol); 3.83 (q, 2H␣1); 4.3 (m, 2Hcol).
4.6. Compound 9
Analysis calcd. for C20H44NPO5 (409) H 10.83, C
58.66, N 3.42; found H 10.78, C 58.72, N 3.34. Rf
0.39 (CHCl3/CH3OH/H2O 65/25/4); [M + H] 410.4;
1H NMR δ: 0.88 (t, 3H); 1.27 (m, 16H); 1.57 (m, 2H␤);
1.68 (m, 4H␤); 3.22 (s, 9H); 3.42 (m, 4H␣); 3.62 (m,
2Hcol); 3.92 (quart., 2H␣1); 4.26 (m, 2Hcol).
4.12. 8-Benzyloxyoctane-1-phosphocholine
This compound was prepared from 1,8-octanediol-
monobenzylether according to the method previously
described for the synthesis of alkyloxophospho-
cholines. Rf 0.3 (CHCl3/CH3OH/H2O 65/25/4);
4.7. Compound 10
1
Analysis calcd. for C26H56NPO5 (493) H 11.43, C
63.25, N 2.84; found H 11.50, C 63.35, N 2.83. Rf
0.5 (CHCl3/CH3OH/H2O 65/25/4); [M+H] 494.5; 1H
NMR δ: 0.9 (t, 3H); 1.3 (m, 30H); 1.6 (m, 6H␤); 3.2
(s, 9H); 3.42 (t, 4H␣); 3.61 (m, 2Hcol); 3.87 (quart.,
2H␣1); 4.25 (m, 2Hcol).
[M + H] 402; H NMR δ: 1.32 (m, 8H); 1.61 (m,
4H␤); 3.2 (s, 9H); 3.48 (t, 2H␣); 3.57 (m, 2Hcol); 3.85
(quart., 2H␣1); 4.44 (s, 2HBn); 7.34 (m, 5H).
4.13. Compound 15
8-Benzyloxyoctane-1-phosphocholine in ethanol
(1–1.5 M) was treated with Pd/C (5%) and the mixture
was hydrogenated at r.t. for several hours. The mixture
was then filtered and the filtrate was evaporated to
dryness giving the desired product. Analysis calcd. for
C13H30NPO5 (311) H 9.71 C 50.15, N 4.50; found H
9.63, C 50.22, N 4.44. Rf 0.19 (CHCl3/CH3OH/H2O
4.8. Compound 11
Analysis calcd. for C17H38NPO5 (367) H 10.42, C
55.57, N 3.81; found H 10.51, C 55.60, N 3.77. Rf
0.47 (CHCl3/CH3OH/H2O 65/25/4); [M + H] 368.3;
1H NMR (in D2O) δ: 1.16 (t, 3H); 1.33 (m, 12H);
1.62 (m, 4H␤); 3.24 (s, 9H); 3.54 (m, 4H␣); 3.73 (m,
2Hcol); 3.93 (m, 2H␣1); 4.32 (m, 2Hcol).
1
65/25/4); [M + H] 312.3; H NMR δ: 1.32 (m, 8H);
1.53 (m, 2H␤); 1.62 (m, 2H␤);3.22 (s, 9H); 3.57 (t,

206
M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
2H␣); 3.61 (m, 2Hcol); 3.76 (quart., 2H␣1); 4.23 (m,
2Hcol).
5. Biological assay
The in vitro activity of oxa-alkylphosphocholines
(oxa-APC) was determined on four cell lines
(Mosmann, 1983; Sims and Page, 1984; Alley et al.,
1988): (a) Jurkat (human T-cell leukaemia); (b)
K562 (human chronic myelogenous leukemia); (c)
Hela (human cervix epitheloid carcinoma); (d) MT2
(murine macrophages, non-tumoral cells). Cell lines
(a), (b), (c) were from Istituto Zooprofilattico di Bres-
cia and MT2 was from Milan University. All cells
were kept in liquid nitrogen in 10% DMSO in serum.
Culture medium was RPMI 1640 supplemented with
10% fetal calf serum, Penicillin 100 U/ml and Strep-
tomycin 100 ␮g/ml. The cell culture medium and
4.14. Compound 17
Palmitic acid (1.6 mol), DCC (1.06 mol) and
4-dimethylaminopyridine (0.1 mol) were added to a
solution of compound 15 in CHCl3 and the mixture
was stirred at r.t. for 8 h. The solid formed was fil-
tered off and washed with CHCl3. The filtrate and
the CHCl3 solution were evaporated to dryness. The
residue was then purified by chromatography on silica
gel giving the desired product (yield 70%). Analy-
sis calcd. for C29H60NPO6 (549) H 11.00, C 63.36,
N 2.55; found H 11.19, C 63.41, N 2.58. Rf 0.5
1
3
(CHCl3/CH3OH/H2O 65/25/4); [M + H] 550.8; H
reagents were from Irvine. H thymidine was from
NMR δ: 0.88 (t, 3H); 1.1 (m, 34H); 1.62 (m, 4H␤);
2.35 (t, 2H); 3.21 (s, 9H); 3.6 (m, 2Hcol); 3.85 (quart.,
2H␣1); 4.05 (t, 2H␣); 4.23 (m, 2Hcol).
Amersham (TRK61) and test plates were 96-well
cell culture clusters tissue culture treated (Costar cat.
N◦3596).
4.15. 12-Hydroxylauric acid methyl ester
5.1. Hemolytic activity
A solution of 12-hydroxylauric acid in ethanol
(0.5 M) was treated with an ethereal solution of di-
azomethane. The reaction was complete after about
30 min. The product was recovered by evaporation to
dryness, with quantitative yields.
Petri plates containing ram blood/agar were spot-
ted with 30 ␮l of oxa-APC solution at concentrations
from 2 mM to 16 ␮M and incubated at 37 ◦C for 24 h.
Results were expressed as the minimal oxa-APC con-
centration with HA; identified by the clear pink spot
on the dark red surface of the agar plate.
4.16. Compound 18
5.2. Antiproliferative activity: [3H] thymidine
uptake
12-Hydroxylauric acid methyl ester was phospho-
rylated according to the method previously described
for the synthesis of alkyloxophosphocholines. Analy-
sis calcd. for C18H38NPO6 (395) H 9.68, C 54.67, N
3.54; found H 9.65, C 54.67, N 3.66. Rf 0.29 (CHCl3/
Antiproliferative activity of the various compounds
in the different cell lines was calculated by measur-
ing the incorporation of radioactive thymidine into
cellular DNA. Cells were continuously cultured in
complete medium at 37 ◦C and 5% CO2 and split
at confluence using trypsin. Cells for testing were
seeded in microtiter plates at a density of 2×104 cells
per well; 180 ␮l per well of complete medium plus
20 ␮l of an oxa-APC solution were transferred into
the well (concentrations from 100 ␮M to 0.78 ␮M
obtained by serial dilution). Eight wells per plate
without oxa-APC solution were used as positive con-
trols. Plates were incubated at 37 ◦C and 5% CO2 for
48 h. Each test was repeated twice.
1
CH3OH/H2O 65/25/4); [M + H] 396.3; H NMR δ:
1.31 (m, 16H); 1.64 (m, 2H); 2.34 (t, 2H); 3.25 (s,
9H); 3.68 (s, 3H); 3.9 (m, 2H␣ e 2H␣); 4.35 (m, 2Hcol).
4.17. Compound 16
Polyethylene glycol 350 monomethyl ether was
used as substrate for the phosphorylation according
to the method previously described for the synthesis
of alkyloxophosphocholines. Rf 0.4 (CHCl3/CH3OH/
H2O 65/25/4); [M + H] 506; 1H NMR δ: 3.21 (s,
9H); 3.42 (s, 3H); 3.68 (m, 28H); 4.06 (m, 2H); 4.3
(m, 2H).
After incubation the supernatant was removed and
cells were pulsed for 4 h with 3H thymidine 0.1 ␮Ci in

M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
207
100 ␮l per well of medium without serum. Cells were
5.4. Oxa-APC hydrolysis mediated by PLD from
Streptomyces sp. PMF
then trypsinized and harvested on a glass fiber filter
with a semi-automated cell harvester (Wallac). The in-
corporated radioactivity was measured as counts per
minutes (CPM) using a Betaplate scintillation counter
(Wallac). CPM were captured, recorded and processed
using an Excel macro and transformed to the percent-
age inhibition of thymidine uptake (cytotoxicity) us-
ing the following formula:
Fifty mg of PC were dissolved in 5 ml of 0.1 M
acetate buffer, 0.1 M in Ca2+ at pH 5.6. The solu-
tion was mixed with 0.5 mg of crude PLD (6 U/mg,
3 U). The mixture was kept at 36 ◦C for 24 h and
samples were drawn and quenched (heating at 100 ◦C
for 30 s) every 30 min. The solution was assayed for
the choline released in the hydrolysis using a quanti-
tative enzymatic assay, as described in the literature
(Shimbo et al., 1989).
% inhibition
ꢀ
ꢁ
CPM compound − CPM blank
CPM control − CPM blank
= 100 −
× 100
5.5. Competitive PLD catalysed hydrolysis of
compound 13 and PC
where CPM compound: CPM in wells with oxa-APC;
CPM blank: average CPM in wells without cells; CPM
control: average CPM in wells without oxa-APC.
The concentration inhibiting thymidine uptake by
50% (IC50) was then obtained by interpolation of the
inhibition curves.
Twenty-five mg of phosphocholine and 25 mg of PC
were suspended in 5 ml of 0.1 M acetate buffer, 0.1 M
in Ca2+ at pH 5.6 and dispersed by sonication for
5 min. The emulsion was treated with 0.5 mg of crude
PLD (6 U/mg, 3 U). The mixture was kept at 36 ◦C for
2 h and heated at 100 ◦C for 1 min, then analyzed for
the choline released in the hydrolysis using the usual
quantitative enzymatic assay and for the amount of PA
formed. PA was determined by HPLC as previously
described (D’Arrigo et al., 1996).
5.3. Cell viability: MTT tetrazolium dye assay
The tetrazolium dye assay is based on the abil-
ity of living cells to form MTT-formazan, a dark
blue product, whose concentration is measured from
the absorbance. Cells plated at a density of 5 × 104
cells/100 ␮l in 96-well microplates were incubated
for one night at 37 ◦C. Then 100 ␮l of an oxa-APC
solution were added to each microplate and cells were
allowed to grow for another 48 h at 37 ◦C. After this
50 ␮l of a MTT phosphate buffered solution was added
to each well and cultures were incubated at 37 ◦C for
4 h. The supernatant was removed from the wells by
slow aspiration and replaced with DMSO (100 ␮l per
plate) to solubilize the MTT tetrazolium dye. Optical
densities (OD) were read at λ = 595 nm, recorded
and processed using a customized Excel macro. OD
values were transformed into percentage inhibition of
viability (cytotoxicity) using the following formula:
6. Results and discussion
6.1. Hemolytic activity and cytotoxicity
Eleven APC compounds were tested with mil-
tefosine to compare their HA and cytotoxicity on
non-tumoral cells (MT2), used as markers of adverse
effects. Table 1 sets onto the lowest concentrations
causing hemolysis (HA) and IC50, determined with the
MTT and thymidine incorporation assays. The com-
pounds can be classified in two categories: compounds
13, 15, 16 and 18 had lower HA and cytotoxicity
than miltefosine 1, and the others caused at least one
adverse effect more intense than miltefosine 1. Both
groups were tested for cytotoxicity on cancer cell lines.
Tables 2 and 3 show the IC50 of the four selected com-
pounds against tumor cell lines in the MTT and thymi-
dine uptake assays. Only compound 13 showed cyto-
toxic effects comparable with miltefosine 1, slightly
higher against Jurkat cells. The others were not active.
% inhibition
ꢀ
ꢁ
OD compound − OD blank
OD control − OD blank
= 100 −
× 100
where OD compound: OD in wells with oxa-APC; OD
blank: average OD in wells without cells; OD control:
average OD in wells without oxa-APC.

208
M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
Table 4
Table 1
Relative rates of PLD catalysed hydrolysis (PC = 1)
Compound
Minimal concentrations (␮M) causing hemolysis and IC50 deter-
mined with the MTT and thymidine incorporation assay in MT2
cells for a series of oxa-APC
Relative rates
PC comp
PC
13
15
16
18
1
–
0.9
0.7
0.6
0.7
0.4
0.7
0.8
0
Compound
Hemolytic
activity
Cytotoxicity
Concentration
(␮M)
MTT on
MT2 cells
IC50 (␮M)
Thymidine
uptake on MT2
cells: IC50 (␮M)
Relative rates of PLD-catalysed hydrolysis of oxa-APCs compared
to PC, taken as 1. PC comp: hydrolysis rate of PC in the presence
of a equimolar amount of oxa-APC (see Section 2).
Miltefosine 1
8
9
400
2000
2000
400
>2000
400
2000
>2000
>2000
>2000
400
>100
>100
2
23
8
<0.8
12
79
93
>100
<0.8
>100
>100
>100
>100
10
11
12
13
14
15
16
17
18
30
12
6.2. APC hydrolysis by PLD
>100
>100
<0.8
>100
>100
>100
>100
In order to check whether oxa-APC were better cy-
totoxic agents on account of their ability to inhibit
PLD, we measured the relative rate of hydrolysis of
all oxa-APCs, as described in the experimental sec-
tion for compound 13. All compounds were substrates
for PLD, as expected (Bossi et al., 2001). For com-
pounds 13, 15, 16, 18 we compared their hydrolysis
to that with PC (taken as 1) under similar conditions
(Table 4). The table also reports the effects of equimo-
lar concentrations of oxa-APC on PC hydrolysis in
competitive experiments (PC comp).
>2000
Hemolytic activity was tested in ram blood-agar plates spotted
with 30 ␮l of oxa-APC solution and incubated at 37 ◦C for 24 h. A
high concentration indicates low hemolytic activity. Cytotoxicity
was measured in non-tumoral cells (MT2) with the MTT and
thymidine uptake assay. Data are reported as IC50 (␮M); high
values indicate low cytotoxicity.
Table 2
6.3. Activity versus toxicity
IC50 of four selected compounds tested against tumor cell lines
using the MTT assay
The biological results indicate that compound 13 is
more active on cancer cell lines than on non-tumoral
cells, with similar data for both MTT and thymidine
incorporation. It had less HA than miltefosine. These
results are promising and suggest that compound 13
might be worthwhile candidate for the design of com-
pounds with fewer adverse effects. Other compounds,
like 14, a lower isomeric homologue of 13, had low
HA but high toxicity towards MT2 cells. Surprisingly,
compounds 15–18 had no cytotoxic activity, and com-
pounds 9–11 and 14 showed higher cytotoxic activity
on non-tumoral cells than cancer cells. Thus, com-
pounds 8, 12 and 13 form a homogeneous structural
group with behavior similar to miltefosine but with
much higher cytotoxic activity on cancer cells than
on normal cells (MT2). Compound 13 is also the best
substrate for PLD, although there was no real correla-
tion between PLD hydrolysis and activity. In fact, 13
attenuated PC hydrolysis slightly more than expected
from the hydrolysis rate of the two compounds and
Compound
Hela
K562
Jurkat
Miltefosine 1
26
>100
>100
>100
>100
>100
33
29
>100
>100
>100
13
15
16
18
99
>100
>100
>100
The data are reported as concentrations (␮M) inducing 50% cy-
totoxicity.
Table 3
IC50 of four selected compounds tested against tumor cell lines
in the [3H] thymidine incorporation assay
Compound
Hela
K562
Jurkat
Miltefosine 1
6
62
68
>100
>100
>100
33
18
>100
>100
>100
13
15
16
18
30
>100
>100
>100
The data are reported as concentrations inducing 50% cytotoxicity.

M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
209
bonds as substrates for phospholipase C. Chem. Phys. Lipids
25, 369–380.
Eibl, H., Hilgard P., Unger C. (Eds.), 1992. Alkylphosphocholines:
New Drugs in Cancer Therapy. Progress in Experimental
Tumor Research, vol. 34. Karger, Basel.
D’Arrigo, P., de Ferra, L., Piergianni, V., Selva, S., Servi, S., Strini,
A., 1996. Preparative transformations of natural phospholipids
catalysed by phospholipase D from Streptomyces. J. Chem.
Soc. Perkin Trans. I, pp. 2651–2656.
Grivilsdalsky, P.N., Bittman, R., 1989. An efficient stereocontrolled
route of both enantiomers of platelet activating factor
and analogues with long-chain esters at C2: saturated
and unsaturated ether glycerolipids by opening of glycidyl
arenesulfonates. J. Org. Chem. 54, 4643–4648.
Grunicke, H., Maly, K., Überall, F., Schubert, C., Kindler, E.,
Stekar, J., Brachwitz, H., 1996. Cellular signalling as a target
in cancer chemotherapy. Phospholipid analogues as inhibitors
of mitogenic signal transduction. Adv. Enzyme Regulat. 36,
385–407.
Kotting, J., Bergen, M.P., Unger, C., Eibl, H., 1992. Alkylphospho-
cholines: influence of structural variations on biodistribution
at antineoplastically active concentrations. Cancer Chemother.
Pharmacol. 85–102.
slightly more if compared to the other compounds of
the set which, however, had no cytotoxicity. The sig-
nificance of the findings on PLD hydrolysis is also re-
duced by the fact that all the oxa-APC are substrates
for PLD irrespective of their cytotoxicity. It is difficult
to evaluate cooperative effects since it is well known
that structural variation may dramatically change the
phase properties of phospholipid analogues and con-
sequently the observed hydrolysis rates (Servi, 1998).
The data presented reveal the difficulty on iden-
tifying well defined structure–activity correlation in
the group of compounds studied. Yet minor structural
changes can cause a distinct modulation in biologi-
cal properties, measured with in vitro assays. Com-
pound 13 emerges as a new oxa-APC with interesting
properties.
References
Magolda, R.L., Johnson, P.R., 1985. A new efficient and versatile
synthesis of alkylphosphorylcholines. Tetrahedron Lett. 26,
1167–1170.
Matzke, A., Massing, U., Krug, H.F., 2000. Synthetic phospho-
lipids as antitumor agents: a new concept. Nachr.-Forschungs-
zent. Karlsruhe 32, 105–113.
Alley, M.C., Scudiero, D.A., Monke, A., Hursey, M.L., Czerwinski,
M.J., Firre, D.L., Abbott, B.J., Mayo, J.G., Shoemaker, R.H.,
Boyd, M.R., 1988. Feasibility of drug screening with panels
of human tumors cell lines using a microculture tetrazolium
assay. Cancer Res. 48, 589–601.
Berger, M.R., Richter, H., Seelig, M.H., Eibl, H., Schmaehl, D.,
1990. New cytostatics—more activity and less toxicity. Cancer
Treat. Rev. 17, 143–154.
Matzke, A., Massing, U., Krug, H.F., 2001. Killing tumour cells
by alkylphosphocholines: evidence for involvement of CD95.
Eur. J. Cell Biol. 80, 1–10.
Bierer, D.E., Dubenko, L.G., Litvak, J., Gerber, R.E., Chu, J., Thai,
D.L., Tempesta, M.S., Truong, T.V., 1995. Synthesis of (2R)-
and (2S)-␤-d-glucopyranose-1-yl-2-O-palmitoyl-3-O-phospha-
tidylcholinylglycerol: a new antifungal phosphocholine. J. Org.
Chem. 60, 7646–7653.
Bossi, L., D’Arrigo, P., Pedrocchi-Fantoni, G., Mele, A., Servi, S.,
Leiros, I., 2001. The substrate requirements of phospholipase
D. J. Mol. Cat. B: Enzymatic 11, 433–438.
Brachwitz, H., Lachmann, U., Thomas, Y., Bergmann, J.,
Berdel, W.E., Langen, P., 1996. Synthesis and antiproliferative
activity of cytidine-5ꢀ-alkylphosphates and structurally related
compounds. Chem. Phys. Lipids 83, 77–85.
Brachwitz, H., Ölke, M., Bergmann, J., Langen, P., 1997. Alkyl-
phospho-l-serine analogues: synthesis of cytostatically active
alkylphosphono derivatives. Bioorg. Med. Chem. Lett. 7 (13),
1739–1742.
Mosmann, T., 1983. Rapid colorimetric assay for cellular growth
and survival: application to proliferation and cytotoxicity
assays. J. Immunol. Methods 65, 55–63.
Nakamura, N., Miyazaki, H., Ohkawa, N., Oshima, T., Koike, H.,
1990. An efficient synthesis of platelet-activating factor (PAF)
via 1-O-alkyl-2-O-(3-isoxazolyl)-SN-glycero-3-phosphocho-
line, a new PAF agonist. Utilization of the 3-isoxazolyloxy
group as a protected hydroxyl. Tetrahedron Lett. 31, 699–702.
Ohno, M., Fujita, K., Nakai, H., Kobayashi, S., Inoue, K., Nojima,
S., 1985. An enantioselective synthesis of platelet-activating
factors, their enantiomers, and their analogs from d- and
l-tartaric acids. Chem. Pharm. Bull. 33, 572–582.
Puricelli, L., 1989. EP 0335396.
Runge, M.H., Andersen, R., Pfleiderer, A., Munder, P.G., 1980.
Destruction of human solid tumors by alkyl lysophospholipids.
J. Natl. Cancer Inst. 64, 1301–1321.
Brachwitz, H., Vollgraf, C., 1995. Analogs of alkyl lysophos-
pholipids: chemistry, effects on the molecular level and their
consequences for normal and malignant cells. Pharmacol.
Therap. 66 (1), 39–82.
Chandrakumar, N.S., Hajdu, J., 1982. Synthesis of enzyme-
inhibitory phospholipid analogs. Stereospecific synthesis of
2-amidophosphatidylcholines and related derivatives. J. Org.
Chem. 47, 2144–2147.
Shimbo, K., Yano, H., Miyamoto, Y., 1989. Two Streptomyces
strains that produce phospholipase D with transphosphatidyl-
ation activity. Agric. Biol. Chem. 53, 3083–3087.
Servi, S., 1998. Phospholipases as synthetic catalysts. In: Fessner,
W.-D. (Ed.), Topics in Current Chemistry. Biocatalysis:
From Discovery to Application, vol. 200. Springer, Berlin,
pp. 127–158.
Sims, T.J., Page, R.C., 1984. An improved method for assessing
the incorporation of labelled precursors into DNA by human
mononuclear cells. J. Immunol. Methods 67, 255–269.
Cox, J.W., Snyder, W.R., Horrocks, L.A., 1979. Synthesis of
choline and ethanolamine phospholipids with thiophosphoester

210
M. Agresta et al. / Chemistry and Physics of Lipids 126 (2003) 201–210
Terashita, Z., Tsushima, S., Yoshioka, Y., Nomura, H., Inada, Y.,
Nishikawa, K., 1983. CV-3988-a specific antagonist of platelet
activating factor. Life Sci. 32, 1975–1982.
Traiser, M., Reichert, S., Voss, A., 1998. Current development
status of the second generation alkylphosphocholine analogue
perifosine. Drugs Today 34, 67–71.
of alkyl lysophospholipids. Chem. Pharm. Bull. 80, 3260–
3270.
Unger, C., Eibl, H., 1990. Hexadecylphosphocholine: preclinical
and the first clinical results of a new antitumor drug. Lipids
17, 143–154.
Yuan, W., Berman, R.J., Gelb, M.H., 1987. Synthesis and
evaluation of phospholipid analogues as inhibitors of cobra
venom phospholipase A2. J. Am. Chem. Soc. 109, 8071–8081.
Tsushima, S., Yoshioka, Y., Tanida, S., Nomura, H., Nojima,
S., Hosumi, M., 1982. Syntheses and antimicrobial activities