Sterol-modified phospholipids: Cholesterol and phospholipid chimeras with improved biomembrane properties

Article abstract of DOI:10.1021/ja8065557

We synthesized a family of sterol-modified glycerophospholipids (SML) in which the sn-1 or sn-2 position is covalently attached to cholesterol and the alternative position contains an aliphatic chain. The SML were used to explore how anchoring cholesterol to a phospholipid affects cholesterol behavior in a bilayer. Notably, cholesterol in the SML retains the membrane condensing properties of free cholesterol regardless of the chemistry or position of its attachment to the glycerol moiety of the phospholipid. SMLs by themselves formed liposomes upon hydration and in mixtures between an SML and diacylglycerophospholipids (C14 to C18 chain length) the thermotropic phase transition is eliminated at the SML equivalent of about 30 mol % free cholesterol. Osmotic-induced contents leakage from SML (C14-C18) liposomes depends upon the linkage and position of cholesterol but in general is similar to that observed in 3/2 diacylphosphatidylcholine/cholesterol (mole ratio) liposomes. SML liposomes are exceptionally resistant to contents release in the presence of serum at 37°C. This is probably due to the fact that SML exchange between bilayers is more than 100 fold less than the exchange rate of free cholesterol in the same conditions. Importantly, SML liposomes containing doxorubicin are as effective in treating the murine C26 colon carcinoma as Doxil, a commercial liposome doxorubicin formulation. SMLs stabilize bilayers but do not exchange and hence provide a new tool for biophysical studies on membranes. They may improve liposomal drug delivery in organs predisposed to the extraction of free cholesterol from bilayers, such as the skin, lung, or blood.

Full text of DOI:10.1021/ja8065557

Published on Web 10/25/2008
Sterol-Modified Phospholipids: Cholesterol and Phospholipid
Chimeras with Improved Biomembrane Properties
Zhaohua Huang and Francis C. Szoka, Jr.*
Departments of Pharmaceutical Chemistry and Biopharmaceutical Sciences, School of
Pharmacy, UniVersity of California at San Francisco, San Francisco, California 94143-0912
Received August 18, 2008; E-mail: Szoka@cgl.ucsf.edu
Abstract: We synthesized a family of sterol-modified glycerophospholipids (SML) in which the sn-1 or
sn-2 position is covalently attached to cholesterol and the alternative position contains an aliphatic chain.
The SML were used to explore how anchoring cholesterol to a phospholipid affects cholesterol behavior in
a bilayer. Notably, cholesterol in the SML retains the membrane condensing properties of free cholesterol
regardless of the chemistry or position of its attachment to the glycerol moiety of the phospholipid. SMLs
by themselves formed liposomes upon hydration and in mixtures between an SML and diacylglycerophos-
pholipids (C14 to C18 chain length) the thermotropic phase transition is eliminated at the SML equivalent
of about 30 mol % free cholesterol. Osmotic-induced contents leakage from SML (C14-C18) liposomes
depends upon the linkage and position of cholesterol but in general is similar to that observed in 3/2
diacylphosphatidylcholine/cholesterol (mole ratio) liposomes. SML liposomes are exceptionally resistant
to contents release in the presence of serum at 37 °C. This is probably due to the fact that SML exchange
between bilayers is more than 100 fold less than the exchange rate of free cholesterol in the same conditions.
Importantly, SML liposomes containing doxorubicin are as effective in treating the murine C26 colon
carcinoma as Doxil, a commercial liposome doxorubicin formulation. SMLs stabilize bilayers but do not
exchange and hence provide a new tool for biophysical studies on membranes. They may improve liposomal
drug delivery in organs predisposed to the extraction of free cholesterol from bilayers, such as the skin,
lung, or blood.
Introduction
in turn dynamically influences membrane protein organization
and ultimately cell physiology.1,10 These profound effects on
Cholesterol is an essential component of animal cell mem-
branes and contributes to bilayer stability and bilayer fluidity
and promotes the liquid condensed state in lipid mixtures
containing unsaturated and saturated diacyl chains.1-3 When
the cholesterol content of the bilayer is about 30 mol %, the
thermotropic phase transition of synthetic diacylphospholipids
is eliminated,4,5 the membrane permeability of hydrophilic
molecules is reduced,6,7 and the insertion of hydrophobic
compounds and proteins into the bilayer inhibited.8 Above 50
mol %, the cholesterol phase separates from the membrane into
cholesterol-rich domains that can lead to the formation of
cholesterol crystals.9 Thus, the mole ratio of cholesterol in
biomembranes critically regulates their solvent properties, which
the physical characteristics of membranes require that the
cholesterol membrane concentration be carefully regulated by
a feedback system of proteins that monitor excess free choles-
terol in cell membranes and ultimately modulate cholesterol
synthesis and uptake at the transcriptional level.11,12 Needless
to say, the importance of cholesterol in health and disease has
generated an enormous literature in the century since Windaus
reported that atheromatous lesions had 6 times as much
cholesterol as a normal arterial wall leading to his pioneering
studies on cholesterol biosynthesis and metabolism.13
In mammals, cholesterol is found as a free molecule, modified
at the 3 position by sulfate or esterified at the 3 position to
fatty acids. Neither cholesterol nor cholesterol esters can form
bilayer membranes by themselves, whereas cholesterol deriva-
tives with a hydrophilic headgroup at the 3 position such as a
sulfate, phosphate, phosphatidylcholine, ethylene oxide, or
hemisuccinate assemble into membranes as bilayer vesicles
(liposomes).14-19 Liposomes containing a high mole percentage
of cholesterol are generally more stable and less leaky than those
(1) McIntosh, T. J.; Simon, S. A. Annu. ReV. Biophys. Biomol. Struct.
2006, 35, 177–198.
(2) Sugahara, M.; Uragami, M.; Yan, X.; Regen, S. L. J. Am. Chem. Soc.
2001, 123, 7939–7940.
(3) McConnell, H. M.; Vrljic, M. Annu. ReV. Biophys. Biomol. Struct.
2003, 32, 469–492.
(4) Mabrey, S.; Mateo, P. L.; Sturtevant, J. M. Biochemistry 1978, 17,
2464–8.
(5) McMullen, T. P.; Lewis, R. N.; McElhaney, R. N. Biochemistry 1993,
32, 516–22.
(10) Anderson, R. G. Trends Cell. Biol. 2003, 13, 534–9.
(11) Brown, M. S.; Goldstein, J. L. Science 1986, 232, 34–47.
(12) Goldstein, J. L.; DeBose-Boyd, R. A.; Brown, M. S. Cell 2006, 124,
35–46.
(6) Demel, R. A.; Bruckdorfer, K. R.; van Deenen, L. L. Biochim. Biophys.
Acta 1972, 255, 321–30.
(7) Papahadjopoulos, D.; Nir, S.; Oki, S. Biochim. Biophys. Acta 1972,
266, 561–83.
(13) Windaus, A. In Nobel Lectures Chemistry 1922-1941; Elsevier
Science Publishing Co.: New York, 1996; pp 105-121.
(14) Brockerhoff, H.; Ramsammy, L. S. Biochim. Biophys. Acta 1982, 691,
227–232.
(8) Haines, T. H. Prog. Lipid Res. 2001, 40, 299–324.
(9) Bach, D.; Wachtel, E. Biochim. Biophys. Acta 2003, 1610, 187–197.
9
15702 J. AM. CHEM. SOC. 2008, 130, 15702–15712
10.1021/ja8065557 CCC: $40.75
2008 American Chemical Society

Sterol-Modified Phospholipids
A R T I C L E S
Table 1. Summary of Sterol-Modified Phospholipids (SMLs)
that include: ester, carbonate, carbamate, and ether linkages.
These linkages were selected because the manner in which
cholesterol and the aliphatic chain are linked to the glycerol
backbone may significantly affect the properties of SMLs in
bilayers and their fate in vivo. We adopted synthetic strategies
that have been applied to the synthesis of glycerophospholipids25
for the synthesis of SMLs. Semisynthetic approaches used
commercially available precursors, such as glycerophospho-
choline (Scheme 1) and lyso-lipids (Schemes 5 and 6), which
were employed to prepare stereospecific molecules. For SMLs
containing an ether linkage (Schemes 2-4), a longer synthetic
route is required to obtain the specific 1,2-disubstituted glycerols.
The conversion of 1,2-disubstituted glycerol to the corresponding
phosphocholine was accomplished by the commonly used steps:
phosphorylation with phosphorus oxychloride, followed by the
coupling with choline tetraphenyl borate using 2,4,6-triisopro-
pylbenzysulfonyl chloride as the condensing reagent. This is a
practical and reliable way to convert 1,2-diradyl glycerol to
phosphocholine in our laboratory and providing an overall yield
of above 70%. It should be noted that some SMLs are
diastereomers when racemic starting material was used. The
packing of diastereomers in the membrane may be quite different
from the optical pure SMLs.
In the synthesis of 1a-c (Scheme 1), the yield-limiting step
was the introduction of alkylcarbamoyl chain at the C2 position
through the reaction of isocyanate and the 2-hydroxy group.
The low yield (ca. 30%) of this reaction may be due to the side
reaction of isocyanate and the poor solubility of 7. The presence
of the carbamate bond at the C2 position may favor the
formation of strong hydrogen bonding in the water-oil surface
region and alter the properties of 1a-c. The synthesis of 2a-d
started with the selective alkylation of the C2 position of 1,3-
protected glycerol (Scheme 2) followed by the removal of the
protective group through acid hydrolysis. The reaction of 11a-d
and cholesteryl chloroformate with controlled stoichiometry led
mainly to the desired product 12a-d with slight 1,3-disubstitued
side product which was easily removed by high performance
flash chromatography (HPFC). The conversion of 12a-d to
2a-d was achieved by the common procedure. Cholesteryl
glycerol (15), the key intermediate of 3a-d, was obtained by
the treatment of cholesteryl tosylate with excessive soketal at
90 °C followed by acid hydrolysis (Scheme 3). Our effort to
repeat the previously reported method26 ended only with side
product from the elimination reaction of cholesteryl tosylate due
to the harsh conditions: the presence of strong base and high
reaction temperature (120 °C). After the protection of the C3
hydroxyl group with trityl group, the acyl group was selectively
coupled to the C2 position. The removal of the trityl group from
17a-d at 0 °C with boron trifluoride etherate led to 18a-d
with no detectable isomers as judged by TLC. Nevertheless,
they were immediately converted to 3a-d using the standard
procedure to avoid acyl migration. The synthesis of 4a-d
(Scheme 4) was similar to that of 3a-d except that the position
of cholesterol and aliphatic chain was switched. The synthesis
of 21a-d was carried out under higher temperature because of
the stability of the aliphatic chain. After selective attachment
of cholesterol at the C2 position through a carbonate linkage,
the trityl group was removed and the 1,2-diradyl glycerols were
transformed to 4a-d. The synthesis of 5a-d (Scheme 5) was
achieved by direct coupling of cholesteryl chloroformate and
lipids
R1
R2
method
1a-c
2a-d
3a-d
Chol-OC(O)a
Chol-OC(O)
Chol
CnH2n+1NHC(O)b
route 1
route 2
route 3
c
CnH2n+1 or C18H35
CnH2n+1C(O)d or
C17H33C(O)
e
4a-d
5a-d
CnH2n+1 or C18H35
Chol-OC(O)
Chol-OC(O)
route 4
route 5
CnH2n+1C(O)f or
C17H33C(O)
6a-d
CnH2n+1C(O)for
C17H33C(O)
Chol-OC(O)CH2
CH2C(O)
route 6
a Chol-OH ) cholesterol. b n ) 18, 16, 14. c n ) 18, 16, 14. d n )
17, 15, 13. e n ) 18, 16, 14. f n ) 17, 15, 13.
without cholesterol and are widely used in the formulation of
chemotherapeutic drugs.20 However, when liposomes composed
of free cholesterol and phospholipids are placed in biological
fluids in contact with biomembranes and proteins, free choles-
terol rapidly transfers from the liposome into the biomem-
branes.21-23 This transfer of free cholesterol from the liposome
results in the decrease in liposome stability and the subsequent
loss of encapsulated contents from the liposome.
We were interested in learning the consequences on bilayer
properties of an amphiphilic cholesterol analog that could form
bilayers but was unable to rapidly transfer between membranes.
So we have designed and synthesized a new category of
chimeric sterol-modified phospholipids (SMLs) consisting of
both cholesterol and an aliphatic chain covalently linked to the
glycerol backbone of phosphatidylcholine (Table 1). We
hypothesize that the covalent hybridization of cholesterol and
a phospholipid will confine cholesterol in the lipid bilayer and
improve the bilayer cohesion properties.24 In this paper, we
report the synthesis of SMLs and their biophysical properties
and demonstrate their potential as an anticancer drug carrier.
Results and Discussion
Chemistry of SMLs. The idea of a covalent chimera between
a sterol and a phospholipid opens the way to a myriad of
structures given the various combinations of sterol, aliphatic
chain length, aliphatic chain type, substitution position, linkage
chemistry, stereochemistry, and headgroup. To test the concept,
we focused on lipids bearing a cholesterol, an aliphatic chain
with length between C-14 and C-18 and phosphocholine as
the headgroup. These constituents were selected because of their
abundance and importance in mammalian membranes. A series
of six SMLs (Table 1, and Schemes 1-6) were synthesized
(15) Demel, R. A.; Lala, A. K.; Kumari, S. N.; Vandeenen, L. L. M.
Biochim. Biophys. Acta 1984, 771, 142–150.
(16) Patel, K. R.; Li, M. P.; Schuh, J. R.; Baldeschwieler, J. D. Biochim.
Biophys. Acta 1984, 797, 20–26.
(17) Lai, M. Z.; Duzgunes, N.; Szoka, F. C. Biochemistry 1985, 24, 1646–
1653.
(18) Epand, R. M.; Bottega, R.; Robinson, K. Chem. Phys. Lipids 1990,
55, 49–53.
(19) Gotoh, M.; Ribeiro, N.; Michels, B.; Elhabiri, M.; Albrecht-Gary,
A. M.; Yamashita, J.; Hato, M.; Ourisson, G.; Nakatani, Y. Chem.
BiodiVers. 2006, 3, 198–209.
(20) Torchilin, V. P. Nat. ReV. Drug DiscoVery 2005, 4, 145–60.
(21) Phillips, M. C.; Johnson, W. J.; Rothblat, G. H. Biochim. Biophys.
Acta 1987, 906, 223–76.
(22) Hamilton, J. A. Curr. Opin. Lipidol. 2003, 14, 263–271.
(23) Kan, C. C.; Yan, J.; Bittman, R. Biochemistry 1992, 31, 1866–74.
(24) Needham, D.; Evans, E. Biochemistry 1988, 27, 8261–8269.
(25) Paltauf, F.; Hermetter, A. Prog. Lipid Res. 1994, 33, 239–328.
(26) MacKellar, C.; Graham, D.; Will, D. W.; Burgess, S.; Brown, T.
Nucleic Acids Res. 1992, 20, 3411–7.
9
J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008 15703

A R T I C L E S
Huang and Szoka
Scheme 1. Synthesis of 1a-ca
a Reagents and conditions: (A) trityl chloride (1.03 equiv.), ZnCl2 (0.95 equiv.), DMF, 4 °C, 10 h; (B) alkyl isocyanate (1.0 equiv.), DMSO, 100 °C, 24 h;
(C) TFA in CHCl3 (16.7%), r.t., 4 h; (D) cholesteryl chloroformate (5 equiv.), DIPEA (5.2 equiv.), CHCl3, r.t., 16 h.
Scheme 2. Synthesis of 2a-da
a Reagents and conditions: (A) (1) NaH (1.2 equiv.), toluene, r.t., 30 min; (2) iodoalkane (1.25 equiv.), reflux, overnight; (B) HCl (conc.) in MeOH
(10%), reflux, 5 h; (C) cholesteryl chloroformate (1.05 equiv.), DIPEA (1.4 equiv.), DMAP (0.5 equiv.), CHCl3, 0 °C, 0.5 h, and then r.t., overnight; (D)
POCl3 (1.1 equiv.), pyridine (2 equiv.), THF, 0 °C, 2-3 h; (E) choline tetraphenyl borate (2 equiv.), TPS (2.5 equiv.), pyridine, 70 °C, 1 h, then r.t., 3 h.
Scheme 3. Synthesis of 3a-da
a Reagents and conditions: (A) cholesteryl tosylate, toluene, 90 °C, 4 h; (B) TFA-HCl (conc.) 2:1, THF, r.t., 4 h; (C) trityl chloride (1.5 equiv.), pyridine,
50 °C, 18 h; (D) aliphatic acid (1.05 equiv.), DCC (1.05 equiv.), DMAP (0.3 equiv.), CHCl3, r.t., overnight; (E) BF3.Et2O (4 equiv.), CHCl3, 0 °C, 3 h; (F)
POCl3 (1.1 equiv.), pyridine (2 equiv.), THF, 0 °C, 2-3 h; (G) choline tetraphenyl borate (2 equiv.), TPS (2.5 equiv.), pyridine, 70 °C, 1 h, and then r.t.,
3 h.
the corresponding lyso-phosphocholine in high yield. Similarly,
direct conjugation of cholesteryl hemisuccinate and lyso-
phosphatidylcholine led to 6a-d (Scheme 6). It should be noted
that the lyso-lipid should be completely dissolved in the dry
ethanol-free chloroform to obtain a high yield.
Since this new category of lipids do not fall within any of
the categories identified in the most recent lipid classification
recommendations,27 we designate the chimeric phospholipids
abbreviations according to the rules for the common names
of glycerolphospholipids. For example, 1-cholesterylcarbon-
oyl-2-palmityl-glycero-3-phosphatidylcholine is abbreviated
as ChcPePC, wherein (1) groups at the sn-1/sn-2 positions
are represented by capitalized abbreviations (for example,
“Ch” for cholesterol, “P” for palmitoyl) in the sequence of
their substitution positions; (2) the subscript letter or
lowercase letter is used to indicate the linkage type such as
“c” for carbonate, “e” for ether, “a” for carbamate, and the
blank (or no subscript letter) for ester according to the
existing convention; (3) the head groups are named according
to the convention such as “PC” for phosphocholine.
Thermotropic Phase Behavior of Bilayers Composed of
SML and Synthetic Diacylphosphatidylcholines. The addition
of free cholesterol into a bilayer composed of saturated
phospholipids will alter the thermotropic phase behavior of the
bilayer resulting in the elimination of the phase transition at
about 33 mol % cholesterol.4 We employed differential scanning
calorimetry (DSC) to quantify the effect an SML exerts on the
phase transition of synthetic lipids when the aliphatic chain
length in the SML is the same as the aliphatic chain length in
the diacyl phospholipids. Accordingly, SMLs containing C-16
(27) Fahy, E.; et al. J. Lipid Res. 2005, 46, 839-861.
9
15704 J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008

Sterol-Modified Phospholipids
A R T I C L E S
Scheme 4. Synthesis of 4a-da
a Reagents and conditions: (A) NaH (1.7 equiv.), bromoalkane (0.9 equiv.), toluene, 120 °C, overnight; (B) HCl in MeOH (2 M), reflux, 1.5 h; (C) trityl
chloride (1.5 equiv.), pyridine, 50 °C, 18 h; (D) cholesteryl chloroformate (1.2 equiv.), DMAP (1.2 equiv.), CHCl3, r.t., overnight; (E) BF3.Et2O (4 equiv.),
CHCl3, 0 °C, 3 h; (F) POCl3 (1.1 equiv.), pyridine (2 equiv.), THF, 0 °C, 2-3 h; (G) choline tetraphenyl borate (2 equiv.), TPS (2.5 equiv.), pyridine, 70
°C, 1 h, and then r.t., 3 h.
Scheme 5. Synthesis of 5a-da
1B). (See the Supporting Information, Figures 1-7, for the DSC
thermograms of other SMLs).
The major transition temperature (Tm) and enthalpy (∆H)
were plotted against the percentage of the cholesterol in the
lipid mixture (panels C and D in Figure 1, and Table 1 in the
Supporting Information). Not surprisingly, there was no detect-
able phase transition in the temperature range of 10-80 °C for
a Reagents and conditions: (A) cholesteryl chloroformate (2.3 equiv.),
all the pure SMLs tested regardless of the chemical structure
DMAP (4 equiv.), CHCl3, r.t., 16 h.
of the SMLs. Thus, a homogeneous liquid ordered bilayer of
high cholesterol concentration (67%) is achievable if cholesterol
chain with various linkages (1b-5b) and SMLs of one linkage
type group with different acyl chains (5a-5d) were chosen for
the DSC study. All of the SMLs used in the following studies
were readily dispersed in water from the dry state and formed
lipid vesicles as evidenced by dynamic light scattering, freeze
fracture electron microscopy, and negative stain transmission
electron microscopy (see the Supporting Information, Figure 9).
When a SML is mixed with a diacyl phospholipid, there is
an additional acyl chain in SML that must be accounted for
when calculating the total moles of aliphatic chains so that the
computation of cholesterol mole percentage is consistent with
that used in the conventional free cholesterol-diacyl lipid
mixtures. The equivalent free cholesterol mole percentage in a
mixture of SML with a diacyl lipid is calculated according to
the following formula
and the aliphatic chain are positioned close to each other. In
mammalian cells, a similar domain could be formed in the
plasma membrane if the membrane protein has a strong
affinity for both the aliphatic lipid chain and free cholesterol.
The addition of SML to the corresponding diacyl lipids
broadened the transition peak in a similar manner as does
free cholesterol,4,5,28 and decreased both the transition
temperature and transition enthalpy (panels C and D in Figure
1, and Table 1 in the Supporting Information). This effect
was dependent on the mole fraction of the SML and
eventually led to the elimination of the phase transition when
the equivalent free cholesterol reached a certain mole fraction.
The condensing effects of SMLs of different linkages are
similar, although there are noticeable differences on Tm and
∆H especially between the isomers ChcPePC and PeChcPC.
ChcPePC was able to rapidly lower the Tm and ∆H and eliminate
the phase transition completely at 30% equivalent cholesterol,
whereas other SMLs with the same chain length need at least
35% cholesterol. It seems that the 2-ether linkage of ChcPePC
is responsible for the difference because ChcPaPC and PeChcPC
are similar in effect on Tm and ∆H but differ from the effect of
ChcPePC on Tm and ∆H. The addition of OChcPC to DSPC
resulted in very broad transition peaks (see the Supporting
Information, Figure 7) at less than 20% cholesterol. This is
significantly different from the effect of SChcPC on DSPC
(Figure 1A), and may be the result of mismatch in chain length
or chain packing due to the kinked unsaturated oleoyl chain
and the saturated stearoyl chain. A similar result has been
observed in the effects of different side-chain analogs of
cholesterol on the thermotropic phase behavior of stearoylo-
leoylphosphatidylcholine.29
Chol% ) nSML/ 1.5 × nSML + ndiacyl × 100
(
where nSML refers to the moles of SML and ndiacyl refers to
the moles of diacyl lipids.
A pure SML has one cholesterol and one acyl chain, thus
the equivalent free cholesterol percentage is 1/1.5 × 100 ) 67.
Liposomes composed of the SML alone have a greater mole
fraction of cholesterol than the maximum that can be achieved
by mixing free Chol with diacyl chain lipids. In a 1:1 mol
mixture of a SML and a diacyl lipid, the equivalent free Chol
percentage is 1/(1.5 + 1) × 100 ) 40. The same calculation
method was applied to all other SML formulations.
The DSC thermogram (Figure 1A) of SChcPC and its
mixtures with 1,2-distearoyl-sn-glycero-phosphatidylcholine
(DSPC) was the typical result observed for most of the SMLs;
SChcPC at about 35 mol % cholesterol eliminated the phase
transition of DSPC (Figure 1A). Replacing the C-18 acyl chain
with a C-14 acyl chain did not reduce the ability of MChcPC
to eliminate the phase transition of DSPC (Figure 1B). Thus a
cholesterol-like effect was observed rather than a phase separa-
tion of the MChcPC and the longer chain length DSPC (Figure
(28) Lai, M. Z.; Vail, W. J.; Szoka, F. C. Biochemistry 1985, 24, 1654–
1661.
(29) Vilcheze, C.; McMullen, T. P.; McElhaney, R. N.; Bittman, R.
Biochim. Biophys. Acta 1996, 1279, 235–42.
9
J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008 15705

A R T I C L E S
Huang and Szoka
Scheme 6. Synthesis of 6a-da
a Reagents and conditions: (A) cholesteryl hemisuccinate (1.2 equiv.), DMAP (1 equiv.), DCC (1.2 equiv.), CHCl3, r.t., 24 h.
Figure 1. Differential scanning calorimetry study of SML liposomes. (A) Thermograms of SChcPC with various amounts of equivalent cholesterol. (B)
Thermograms of MChcPC/DSPC. (C) Effect of cholesterol on enthalpy. (D) Effect of cholesterol on transition temperature. (See Table 1 in Supporting
Information for the tabulated data.
Leakage of SML Liposomes. The inclusion of cholesterol in
a lipid composition used to make a lipid vesicle will generally
make the liposome less leaky in vitro6,7 and in vivo.30,31
Measuring the leakage of contents from liposomes under an
osmotic gradient is an effective way to evaluate the elastic
deformation and critical failure of lipid membranes.32,33 When
liposomes are subjected to high outwardly directed transbilayer
osmotic pressure, the membrane will swell and burst at the
critical point to rapidly release a portion of the contents. Vesicle
will then reseal into a mechanically stable structure once
sufficient contents have been expelled to bring the system into
osmotic equilibrium. The leakage of calcein from selected
liposomes under an osmotic gradient was monitored at 37 °C
(panels A and B in Figure 2). The leakage profiles of SML
liposomes with different chain length but the same linkage have
been measured along with the traditional diacyl lipid/cholesterol
liposomes (Figure 2A). SML liposomes showed the stability
trend of C18 > C16 > C14 for the specific linkage tested.
Liposomes containing C-14 chain showed different leakage
profiles depending on the linkage type (Figure 2B). Liposome
of ChcMaPC exhibited similar leakage profiles to that of 1,2-
dimyristoyl-sn-glycero-phosphatidylcholine (DMPC)/Cholesterol
(3:2) liposomes and better stability than CheMPC and MChcPC.
The cholesterol-free DMPC liposome was the leakiest composi-
tion. The influence of chain length on the liposome stability
may change when the linkage type varies. For example,
ChcMaPC liposome is less leaky than SChcPC liposome. This
may be due to the stabilizing effect of hydrogen bonding from
the carbamate linkage. Generally, SML liposomes of appropriate
linkage maintain their contents at least as well as the corre-
sponding cholesterol/diacyl lipid mixtures under the osmotic
gradients tested.
(30) Gregoriadis, G.; Davis, C. Biochem. Biophys. Res. Commun. 1979,
89, 1287–1293.
(31) Mayhew, E.; Rustum, Y. M.; Szoka, F.; Papahadjopoulos, D. Cancer
Treat. Rep. 1979, 63, 1923–8.
(32) Mui, B. L. S.; Cullis, P. R.; Evans, E. A.; Madden, T. D. Biophys. J.
1993, 64, 443–453.
The physiological environment is a significant barrier for in vivo
liposome drug delivery because of the propensity of serum protein
and biological membranes to extract free cholesterol from the
(33) Shoemaker, S. D.; Vanderlick, T. K. Ind. Eng. Chem. Res. 2002, 41,
324–329.
9
15706 J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008

Sterol-Modified Phospholipids
A R T I C L E S
Figure 3. Relative rates of cholesterol exchange at 37 °C. All formulations
have 40% free cholesterol or the equivalent cholesterol in the SML.
stark contrast to the control liposome: DMPC/Chol (3:2) released
70% contents in the first week and DSPC/Chol liposome had a
60% content release after the 4 week incubation. The contents
release rate of SML liposomes may be controlled by the structure
of the SML. For example, compared with SChcPC, MChcPC
liposomes released 50% of the content in about 18 days. The results
of SML liposomes leakage in serum are consistent with the osmotic
leakage profiles. For liposome of DMPC/Chol, despite its high
resistance to the osmotic pressure, it is the least stable formulation
in the presence of 30% serum that may be caused by the transfer
of cholesterol from the lipid bilayer to serum proteins.
Cholesterol Exchange. The desorption of free cholesterol from
the donor lipid-water interface is rate-limiting for the overall
transfer process and the rate of this step is influenced by
interactions of free cholesterol molecules with neighboring
phospholipid molecules.21,22 Thus, we predicted that cholesterol
covalently bound to the phospholipid would have a reduced
exchange rate between membranes. To investigate the choles-
terol exchange rate between donor and acceptor liposomes, we
formulated SML liposomes and control liposomes with 10%
negatively charged phosphatidylglycerol and 40% cholesterol
(or the equivalent from SML) and 50% diacyl lipid. 10-fold
mole excess of neutral 1-palmitoyl-2-oleoyl-sn-glycero-3-phos-
phocholine (POPC) liposome was used as the acceptor. All SML
liposomes exhibited very little exchange of cholesterol, par-
ticularly compared with control liposomes of Chol/1,2-dipalmi-
toyl-sn-glycero-phosphatidylcholine (DPPC)/1,2-dipalmitoyl-sn-
glycero-phosphatidylglycerol (DPPG) or Chol/DMPC/1,2-
dimyristoyl-sn-glycero-phosphatidylglycerol (DMPG) (Figure
3). These data confirm that covalently linked cholesterol in SML
compounds does not transfer between bilayers at a significant
rate, whereas the exchange of free cholesterol from a conven-
tional liposome has a half-time of approximate 2 h. This result
provides a plausible explanation for why SML liposomes are
more stable than liposomes containing free cholesterol in
biological fluids.
Figure 2. Leakage profile of calcein loaded SML liposomes. (A) Effect
of chain length on osmotic stress-induced leakage. (B) Effect of linkage
and formulation on osmotic stress-induced leakage. The fraction of calcein
remaining in the liposome was calculated and plotted versus the osmotic
gradient. The error of data is within 0.5%. (C) Contents leakage in 30%
fetal bovine serum at 37 °C was monitored by measuring the fluorescence
intensity change. The fraction of calcein remaining in the liposome was
calculated and plotted versus the time of incubation. The control formula-
tions have 40 mol % cholesterol.
liposome bilayer, resulting in contents leakage. The retention of
contents by SML liposomes was tested in 30% fetal bovine serum
at 37 °C and compared to liposomes with conventional composi-
tions of DSPC/Chol (3:2) or DMPC/Chol (3:2). A striking
observation is that the ChcMaPC liposomes retained their contents,
whereas the DSPC/Chol (3:2) liposome gradually released its
contents over a 4 week period (Figure 2C). Liposomes composed
of SChcPC or CheMPC showed leakage profiles similar to that of
ChcMaPC with less than a 10% content loss in 4 weeks. The
prolonged retention of contents of the SML liposomes stands in
Cytotoxicity of SML Liposomes in Culture Cells. The unique
structures of the SML raised the prospect that they may have
toxic effects on cells due to their potential for interfering with
lipid degradation pathways.34 As a first toxicity screen, we
examined the effects of SML liposomes on the metabolic activity
of C26 colon carcinoma cultured cells using the MTT assay.
Most SMLs showed no detectable cytotoxic effects at concen-
tration as high as 1 mM although the C16 and C18 SMLs
9
J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008 15707

A R T I C L E S
Huang and Szoka
Figure 4. Therapeutic data and biodistribution of doxorubicin-loaded SML liposome on BALB/c mice bearing C-26 tumor. A single dose of 15 mg/kg (10
mg/kg for free doxorubicin) was administered i.v. on day 8 after the tumor inoculation (A) Survival curve. SML-Dox: doxorubicin encapsulated in liposome
composed of PChcPC/PEGDSPE/R-tocopherol (94.8/5/0.2). Log-rank analysis of the paired survival data: p ) 0.0025 for SML-Dox vs PBS and Doxil vs
PBS, p ) 0.0018 for SML-Dox vs Free Dox and Doxil vs Free Dox. No significant difference between SML-Dox and Doxil (p ) 0.94). (B) Tumor growth
curve. The tumor growth delays among the treatment groups (SML-Dox and Doxil) are significantly greater than the control group (PBS) based on
Student-Newman-Keuls method, p < 0.001. (C) Liposomal doxorubicin biodistribution 48 h after i.v. injection at 15 mg/kg. (D) Blood concentration of
doxorubicin after the i.v. injection of liposomal doxorubicin.
containing the carbamate linkage (ChcPaPC and ChcSaPC)
inhibited metabolic activity at 100 µM concentration (see the
Supporting Information, Figure 8). All other SMLs tested were
as well-tolerated as liposome dispersions prepared from free
cholesterol and a diacylphosphatidylcholine.35
animals survived to 22 days (Figure 4). Animals treated with
the maximum tolerated dose of nonencapsulated doxorubicin
(10 mg/kg body weight) had a median survival time of 24 days
and all animals died by day 26. Animals treated with Doxil or
doxorubicin encapsulated in the SML liposome at 15 mg/kg
had a median survival time greater than 90 days and 4 of 5
animals survived greater than day 90. The effect on animal
survival comparing the liposome formulations to either the
vehicle control or to the nonencapsulated drug was significant
(log-rank test, P < 0.01) for both SML liposome and Doxil.
There was no significant difference on animal survival between
the two liposome formulations. Animals treated by the liposomal
doxorubicin (Doxil or SML-Dox) have significantly greater
tumor growth delay39 than the control groups (PBS or free
doxorubicin) (P < 0.001, Student-Newman-Keuls pairwise
comparison). Actually, four out of five mice treated by SML-
Dox or Doxil were cured with the complete elimination of the
tumor. We further compared the biodistribution and blood
circulation of the SML-Dox and Doxil. Although SML-Dox has
a shorter circulation time than Doxil, the amount of doxorubicin
accumulated in C-26 tumor 48 h after the i.v. injection of SML-
Dox is about 70% that of Doxil (Figure 4). The SML
formulation might not seem like a big advantage compared with
the current Doxil formulation, but there is one less lipid
component in the SML formulation, which could improve the
manufacturing and quality control stage in the preparation of
the encapsulated drug.
Comparison of the Effect on Tumor Progression and
Animal Survival of Doxorubicin Encapsulated in SML
Liposomes. Optimal therapeutic outcome of an anticancer
liposome formulation requires the precise control of the
cholesterol concentration.36,37 The SML lipids enable this precise
control. We encapsulated doxorubicin in a liposome composed
of PChcPC/PEG-DSPE/R-tocopherol (R-T): 94.8/5.0/0.2 and
compared the effect of this formulation on tumor progression
and animal survival in the BALB/C mice tumored with C-26
colon carcinoma against the effect of nonencapsulated doxo-
rubicin or Doxil. The Doxil formulation is a commercial
cholesterol-containing liposome formulation of doxorubicin that
is approved by the Food and Drug Administration and was
originally tested in the C-26 model.38
In this model, animals treated with the nondrug containing
vehicle, PBS had a median survival time of 20 days and no
(34) Allen, T. M.; Murray, L.; MacKeigan, S.; Shah, M. J. Pharmacol.
Exp. Ther. 1984, 229, 267–75.
(35) Parnham, M. J.; Wetzig, H. Chem. Phys. Lipids 1993, 64, 263–74.
(36) Mayer, L. D.; Tai, L. C. L.; Ko, D. S. C.; Masin, D.; Ginsberg, R. S.;
Cullis, P. R.; Bally, M. B. Cancer Res. 1989, 49, 5922–5930.
(37) Drummond, D. C.; Noble, C. O.; Hayes, M. E.; Park, J. W.; Kirpotin,
D. B. J. Pharm. Sci. 2008, DOI 10.1002/jps.21358.
(38) Huang, S. K.; Mayhew, E.; Gilani, S.; Lasic, D. D.; Martin, F. J.;
Papahadjopoulos, D. Cancer Res. 1992, 52, 6774–81.
(39) Schluep, T.; Hwang, J.; Cheng, J. J.; Heidel, J. D.; Bartlett, D. W.;
Hollister, B.; Davis, M. E. Clin. Cancer Res. 2006, 12, 1606–1614.
9
15708 J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008

Sterol-Modified Phospholipids
A R T I C L E S
VA) Horizon HPFC system with prepacked silica gel columns (60
Å, 40-63 µm). Unless noted otherwise, the ratios describing the
composition of solvent mixtures represent relative volumes. 1H
NMR spectra were acquired on a Varian 400 MHz instrument.
Chemical shifts are expressed as parts per million using tetram-
ethylsilane as internal standard. J values are in Hertz. MALDI-
TOF mass spectra were obtained at the Mass Spectrometry Facility,
University of California San Francisco. The general procedures used
in the synthesis were described below.
After an extensive review of the literature we have been
unable to find examples of phospholipids where a sterol is
attached to either the glycerol or sphingosine.40-44 Indeed, the
recently introduced lipid classification system27 does not have
a category for the SML and to the best of our knowledge such
lipids have not as yet been identified or isolated from biological
systems. This is perplexing; it would seem incorporation into
biomembranes of a rigid amphipathic lipid with a slow transfer
would provide a consistent permeability barrier for cells. The
disadvantage of SML would be the inability to rapidly respond
to changes in the environment by physical mechanisms, i.e.,
the rapid transfer of the sterol from the membrane.12,45 Unlike
cholesterol, the SML do not transfer rapidly, are unlikely to
undergo a facile transbilayer flip,23 and exhibit a similar
membrane condensing effect on membranes composed of
diacylphospholipids regardless of the nature of the aliphatic
group paired with the sterol. Thus to respond to environment
changes, a cell having two classes of lipids with distinctively
different properties (phospholipids and sterols) would be able
to employ a combinatorial approach to fine-tune membrane
properties.12,45 This could be thermodynamically more efficient
than having to degrade and resynthesize lipids to meet the
demands on the cell to maintain internal homeostasis in the face
of a changing environment.
Protection of 3-Hydroxy Group of 1-Substituted Glycerol.
A mixture of 1-substituted glycerol and triphenyl chloride (1.5
equiv.) in anhydrous pyridine was stirred at 50 °C for 18 h under
anhydrous condition. After cooling to room temperature (r.t., ca.
23 °C), the mixture was poured into ice-cold water, and extracted
with 3 portions of hexane. Undissolved triphenylmethanol was
removed by filtration. The filtrate was washed 3 times with water
and dried over anhydrous sodium sulfate. The solvent was
evaporated and the residue was dissolved in minimum amount of
hexane. Additional triphenyl methanol was precipitated from the
solution by standing overnight at 4 °C. The solid was filtered off,
and the filtrate was evaporated to dryness. The residue was dried
over high vacuum and used directly for next step reaction. The
reaction was monitored by TLC and yield was generally 80-90%.
Removal of Trityl Group from 1,2-Substituted 3-Trityl Glyc-
erol. 1,2-substituted-3-trityl glycerol in chloroform was treated with
boron trifluoride diethyl etherate (4 equiv.) at 0 °C for 3 h. The
solution was washed with water/chloroform/methanol (2:2:1).
The organic layer was dried over sodium sulfate and evaporated.
The residue was dried and used directly for next step in the reaction.
The reaction was monitored by TLC and yield was above 90%.
Phosphorylation of 1,2-Subsitituted Glycerol. A solution of 1,2-
substituted glycerol and anhydrous pyridine (2 equiv.) in anhydrous
tetrahydrofuran (THF) was added dropwise to the freshly distilled
phosphorus oxychloride (1.1 equiv.) in THF with stirring at 0 °C.
Stirring was continued for 2-3 h at 0 °C. Then 10% sodium
bicarbonate (ca. 5 equiv.) was added, and the mixture was stirred
for 15 min at 0 °C. The solution was then poured on ice water,
acidified with HCl (pH ca. 2), and extracted with diethyl ether.
The product in aqueous layer was precipitated by adding acetone
into water. The precipitate was combined with product from the
ether extract, azeotropically dried with toluene, and used directly
for next step reaction. Yield was generally above 90%.
1,2-Substituted Glycero-phosphocholine. 1,2-Substituted-glycero
phosphate, choline tetraphenyl borate (2 equiv.), and 2,4,6-
triisoproylbenzene sulfonyl chloride (TPS) (2.5 equiv.) were
dissolved in anhydrous pyridine with brief warming, and then stirred
for 1 h at 70 °C and 3 h at room temperature. After the addition of
water, the solvents are removed by rotary evaporation. The residue
was extracted with diethyl ether twice. The extract was combined
and evaporated. The crude product was purified by HPFC. Yield
of this step is generally 80-90%.
Synthetic Route 1 (1a-c). (1) 1-O-Trityl-sn-glycero-3-phospho-
choline (7). Zinc chloride powder (anhydrous, 25 g, 175 mmol)
was added to the suspension of glycerophosphocholine (50 g, 185
mmol) in anhydrous DMF (500 mL). The mixture was stirred at
r.t. for 30 min, and trityl chloride (53 g, 190 mmol) was added at
4 °C. The reaction was kept at 4 °C for 10 h. The crude product
was then precipitated by the addition of 1 L diethyl ether. The oily
product was dissolved in 1 L of chloroform/isobutanol (2:1), washed
with 300 mL of 4% aqueous ammonia, and dried over anhydrous
sodium sulfate. After the evaporation of the volatiles, the crude
product was azeotropically dried with toluene. The residue was
triturated with acetonitrile for 5 h at r.t. The white precipitate was
then collected and dried over high vacuum. Yield: 45.2 g (49%
wrt glycerophosphocholine). TLC: Rf ) 0.08 (eluent A).1H NMR
(MeOH-d4): δ 3.12 (m, 2H); 3.15 (s, 9H); 3.56 (m, 2H); 3.90 (m,
2H); 4.01 (m, 1H); 4.20 (m, 2H); 7.27 (m, 9H); 7.42 (m, 6H).
MALDI-MS: calcd for C27H35NO6P+ [M + H]+ 500.22, found
500.31.
Conclusions
We have synthesized six series of SMLs created from
biocompatible components and demonstrated that the attachment
of cholesterol to the glycerol backbone does not interfere with
cholesterol’s membrane condensing properties. We also dem-
onstrated that the cholesterol exchange rate was greatly reduced
when cholesterol was attached to the phospholipid glycerol
backbone. Finally, we documented that SMLs are biocompatible
and liposome prepared from SML can be used as a drug carrier.
The SMLs library could be expanded to have thousands of new
lipids by varying the combination of chain length, position of
substitution, type of linkage, and headgroup. SMLs provide a
rich supplement to the existing lipid world. They contribute new
properties to expand the possible uses lipids have in biology
and commerce.
Materials and Methods
Chemistry. Representative procedures are described here. De-
tailed procedures of synthesis and characterization of the lipids of
different chain lengths are provided in the Supporting Information.
Glycerophosphocholine was from BACHEM (Torrance, CA). Lyso-
phospholipids were purchased from Avanti Polar Lipids (Alabaster,
AL). Other reagents were from Aldrich (Milwaukee, WI). Solvents
were used either directly or purified and dried before use according
to the standard protocol. TLC analyses were performed on 0.25
mm silica gel F254 plates using a variety of developing systems:
(A) CHCl3/MeOH/NH4OH (65/25/4), (B) CHCl3/MeOH/NH4OH
(65/35/5), (C) CHCl3/MeOH/H2O (65/25/4), (D) hexane/EtOAc (2/
1), (E) hexane/EtOAc (10/1), (F) hexane/EtOAc (5/1), (G) toluene/
ether (9/1), (H) toluene/ether (1/1). High-performance flash chro-
matography (HPFC) was carried out on a Biotage (Charlottesville,
(40) Urata, K.; Takaishi, N. Eur. J. Lipid Sci. Technol. 2001, 103, 29–39.
(41) Salunke, D. B.; Hazra, B. G.; Pore, V. S. Curr. Med. Chem. 2006,
13, 813–47.
(42) Guo, X.; Szoka, F. C. Acc. Chem. Res. 2003, 36, 335–341.
(43) Bhattacharya, S.; Bajaj, A. Curr. Opin. Chem. Biol. 2005, 9, 647–55.
(44) Huang, Z.; Szoka F. C. Liposome Technology, 3rd ed.; Gregoriadis,
G., Ed.; Informa Healthcare: New York, 2007; Vol. 1, pp 165-196.
(45) Lange, Y.; Ye, J.; Steck, T. L. J. Biol. Chem. 2005, 280, 36126–
36131.
9
J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008 15709

A R T I C L E S
Huang and Szoka
1-O-Trityl-2-stearylcarbamoyl-sn-glycero-3-phosphocholine (8a).
To a solution of 7 (1 g, 2 mmol) in dimethylsulfoxide (anhydrous,
10 mL) was added octadecylisocyanate (0.6 g, 2 mmol). The
reaction mixture was stirred at 100 °C under N2 for 24 h. After the
evaporation of the solvent, the residue was extracted with methanol.
The white solid was filtered off and the filtrate was evaporated to
dryness. The Crude product was purified by HPFC (CHCl3/MeOH/
H2O, 35/13/2). Yield: 450 mg (28.3% wrt 7). TLC: Rf ) 0.25
(eluent A).1H NMR (CDCl3): δ 0.89 (t, J ) 6.4, 3H); 1.25-1.31
(m, 30H); 1.47 (m, 2H); 3.01 (m, 2H); 3.20 (s, 9H); 3.27 (m, 2H);
3.72 (m, 2H); 4.03-4.12 (m, 3H); 4.24 (m, 2H); 5.07 (br, 1H);
7.20 (m, 9H); 7.41 (m, 6H). MALDI-MS calcd for C46H72N2O7P+
[M + H]+ 795.51, found 795.52.
1-Hydroxy-2-stearylcarbamoyl-sn-glycero-3-phosphocholine (9a).
Compound 8a (420 mg, 0.52 mmol) was treated with trifluoroacetic
acid (TFA, 1 mL) in chloroform (5 mL) at r.t. for 4 h. The volatiles
were evaporated and the residue was purified by HPFC (CHCl3/
MeOH/H2O, 10/5/1). Yield: 320 mg (99% wrt 8a). TLC: Rf ) 0.05
(eluent B).1H NMR (CDCl3): δ 0.88 (t, J ) 6.4, 3H); 1.21-1.31
(br, 30H); 1.47 (m, 2H); 3.03 (m, 1H); 3.11 (m, 1H); 3.29 (s, 9H);
3.66 (m, 2H); 3.76 (m, 2H); 3.99 (m, 2H); 4.30 (m, 2H); 4.78 (m,
1H); 6.51 (br, 1H). MALDI-MS calcd for C27H58N2O7P+ [M +
H]+ 553.40, found 553.38.
1-Cholesterylcarbonoyl-2-stearylcarbamoyl-sn-glycero-3-phos-
phocholine (1a, ChcSaPC). To a solution of solution of 9a (0.3 g,
0.54 mmol) and diisopropylethylamine (DIPEA, 0.5 mL, 2.8 mmol)
in dry ethanol-free chloroform (10 mL), was added dropwise the
solution of cholesteryl chloroformate (1.21 g, 2.7 mmol) in ethanol-
free chloroform (5 mL) at r.t. After 16 h reaction at r.t., volatiles
were evaporated and the residue was purified by HPFC (CHCl3/
MeOH/H2O, 40/18/3). Yield: 438 mg (84% wrt 9a). TLC: Rf )
0.28 (eluent A).1H NMR (CDCl3): δ 0.69 (s, 3H); 0.85-1.65 (m,
68H); 1.78-2.01 (m, 5H); 2.38 (m, 2H); 3.03 (m, 1H); 3.17 (m,
1H); 3.38 (s, 9H); 3.88 (m, 2H); 4.0 (m, 2H); 4.25 (m, 1H); 4.36
(m, 4H); 5.08 (m, 1H); 5.39 (1H, d, J ) 4.4); 6.02 (br, 1H).
MALDI-MS calcd for C55H102N2O9P+ [M + H]+ 965.73, found
965.68.
(s, 3H); 0.85-1.65 (m, 68H); 1.78-2.01 (m, 5H); 2.40 (m, 2H);
3.52 (m, 2H); 3.60 (m, 2H); 3.68 (m, 1H); 4.22 (m, 2H); 4.43 (m,
+
1H); 5.40 (1H, d, J ) 4.4); MALDI-MS calcd for C49H89O5 [M
+ H]+ 757.67, found 757.68.
(4) 1-Cholesterylcarbonoyl-2-stearyl-rac-glycero-3-phosphate (13a).
This compound was synthesized according to the general procedure
of phosphorylation. TLC: Rf ) 0.05 (eluent A). MALDI-MS calcd
for C49H89NaO8P+ [M + Na]+ 859.62, found 859.60.
1-Cholesterylcarbonoyl-2-stearyl-rac-glycero-3-phosphocho-
line (2a, ChcSePC). This compound was synthesized according to
the general procedure of phosphocholine synthesis. TLC: Rf ) 0.3
(eluent A). 1H NMR (CDCl3): δ 0.69 (s, 3H); 0.85-1.65 (m, 68H);
1.78-2.04 (m, 5H); 2.38 (m, 2H); 3.41 (s, 9H); 3.52 (m, 2H); 3.68
(m, 1H); 3.88 (m, 4H); 4.19 (m, 1H); 4.35 (m, 4H); 5.39 (1H, d, J
) 4.4). MALDI-MS calcd for C54H101NO8P+ [M + H]+ 922.73,
found 922.74.
Synthetic Route 3 (3a-d) (1) 3-(2,3-Isopropylidene-1-glyceryl)
cholesterol (14). A mixture of cholesteryl tosylate (50 g, 90 mmol)
and solketal (250 mL, 2 mol) in toluene (50 mL) was stirred at
80-90 °C for 4 h under nitrogen. After cooling to r.t., toluene (300
mL) was added to the mixture. The mixture was washed with brine
(300 mL). After separation, additional 200 mL toluene was added
to the organic layer. The organic layer was then washed with brine
(300 mL), dried, and evaporated to dryness. The crude product was
used directly for next step reaction. TLC: Rf ) 0.55 (eluent G).
(2) 1-Glyceryl Cholesterol (15). The crude product of 14 was
dissolved in the mixed solvents of THF (130 mL)-TFA (40
mL)-HCl (conc., 20 mL). The mixture was kept at r.t. for 4 h.
The volatiles were evaporated under reduced pressure. The residue
was dissolved in CHCl3/MeOH (400 mL/100 mL), and washed with
water (100 mL). The organic layer was then dried over sodium
sulfate, filtered, and evaporated. The crude product was purified
by recrystallization from ethanol at - 20 °C. TLC: Rf ) 0.12 (eluent
H).1H NMR (CDCl3): δ 0.69 (s, 3H); 0.85-1.65 (m, 33H);
1.78-2.31 (m, 7H); 3.10 (m, 1H); 3.45-3.70 (m, 4H); 3.79 (m,
+
1H). 5.34 (1H, d, J ) 4.4); MALDI-MS calcd for C30H53O3 [M
+ H]+ 461.40, found 461.44.
(3) 1-Cholesteryl-3-trityl Glycerol (16). The protection of 3-hy-
droxy group of 15 with trityl group was carried out according to
the general procedure. Product was purified by HPFC (9%-25%
ethyl acetate in hexane). TLC: Rf ) 0.38 (eluent F).1H NMR
(CDCl3): δ 0.69 (s, 3H); 0.85-1.65 (m, 33H); 1.86 (m, 3H); 2.0
(m, 2H); 2.18 (m, 1H); 2.32 (m, 1H); 2.42 (br, 1H); 3.11-3.22
(m, 3H); 3.57 (m, 2H); 3.93 (m, 1H); 5.35 (1H, d, J ) 4.4); 7.28
Synthetic Route 2 (2a-d). (1) 1,3-Benzylidene-2-stearyl-glycerol
(10a). A solution of 1,3-benzylidene glycerol (7.2 g, 40 mmol) in
toluene (100 mL) was added to NaH (60% in mineral oil, 1.92 g,
48 mmol, washed with hexane) suspension in toluene (30 mL) at
r.t. with stirring. Then 1-iodo-octadecane (20 g, 50 mmol) in toluene
(40 mL) was added dropwise into the reaction mixture. After the
addition, the mixture was refluxed under nitrogen overnight and
cooled to r.t. Excessive NaH was destroyed by careful addition of
water into the mixture. The reaction mixture was then washed with
water (100 mL × 2). The organic layer was collected and dried
over sodium sulfate. Solvent was evaporated and the residue was
used directly for the next reaction step.
+
(m, 9H); 7.43 (m, 6H). MALDI-MS calcd for C49H67O3 [M +
H]+ 703.51, found 703.53.
(4) 1-Cholesteryl-2-stearoyl-3-trityl Glycerol (17a). To a solution
of 16 (2.11 g, 3 mmol), stearic acid (0.94 g, 3.15 mmol), and
4-dimethylaminopyridine (DMAP, 0.13 g) in dry ethanol-free
chloroform (20 mL) was added DCC (0.65 g, 3.15 mmol) at 0 °C.
The reaction mixture was stirred at 0 °C for 30 min, and then r.t.
overnight. White precipitate was filtered off, and the filtrate was
evaporated to dryness. The crude product was purified by HPFC
(1%-10% ethyl acetate in hexane). TLC: Rf ) 0.5 (eluent E).1H
NMR (CDCl3): δ 0.69 (s, 3H); 0.85-1.65 (m, 66H); 1.81 (m, 3H);
2.0 (m, 2H); 2.19 (m, 1H); 2.26 (m, 1H); 2.35 (t, J ) 7.2, 2H);
3.12 (m, 1H); 3.25 (m, 2H); 3.67 (m, 2H); 5.17 (m, 1H); 5.32 (d,
J ) 4.4, 1H); 7.27 (m, 9H); 7.44 (m, 6H). MALDI-MS calcd for
(2) 2-Stearyl-glycerol (11a). The crude product of 10a was
hydrolyzed by refluxing in the mixed solution of HCl (conc., 30
mL) and methanol (270 mL) for 5 h. The reaction mixture was
cooled to r.t. and evaporated under reduced pressure. The residue
was dissolved in diethyl ether (300 mL) and washed consecutively
with sodium hydroxide solution (0.5 M, 100 mL) and water (150
mL × 2). The ether layer was then dried and evaporated. The crude
product was purified by HPFC (30-80% ethyl acetate in hexane).
Yield: 11.3 g (82% wrt 1,3-benzylidene glycerol). TLC: Rf ) 0.17
(eluent D).1H NMR (CDCl3): δ 0.86 (t, J ) 6.4, 3H); 1.29 (br,
30H); 1.58 (m, 2H); 3.44-3.78 (m, 7H). MALDI-MS calcd for
+
C67H101O4 [M + H]+ 969.77, found 969.73.
(5) 1-Cholesteryl-2-stearoyl Glycerol (18a). The removal of trityl
group was carried out according to the general procedure. The crude
product was used directly for next step reaction. TLC: Rf ) 0.08
(eluent E).
+
C21H45O3 [M + H]+ 345.34, found 345.33.
(3) 1-Cholesterylcarbonoyl-2-stearyl-glycerol (12a). To a solution
of 11a (0.7 g, 2 mmol), DIPEA (0.5 mL, 2.8 mmol), and DMAP
(0.12 g, 1 mmol) in dry ethanol-free chloroform (10 mL) was added
dropwise cholesteryl chloroformate (0.94 g, 2.1 mmol) chloroform
solution (10 mL) at 0 °C. The reaction mixture was stirred at 0 °C
for 0.5 h, and then at r.t. overnight. The volatiles were evaporated,
and the crude product was purified by HPFC (5-15% ethyl acetate
in hexane). TLC: Rf ) 0.41 (eluent E).1H NMR (CDCl3): δ 0.69
(6) 1-Cholesteryl-2-stearoyl-rac-glycero-3-phosphate (19a). This
compound was synthesized according to the general procedure of
phosphorylation. TLC: Rf ) 0.05 (eluent A).
(7) 1-Cholesteryl-2-stearoyl-rac-glycero-3-phosphocholine (3a,
CheSPC). This compound was synthesized according to the general
procedure of phosphocholine synthesis. TLC: Rf ) 0.31 (eluent
9
15710 J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008

Sterol-Modified Phospholipids
A R T I C L E S
1
A). H NMR (CDCl3): δ 0.68 (s, 3H); 0.85-1.65 (m, 66H); 1.84
2H); 4.24 (m, 1H); 4.41 (m, 3H); 4.55 (m, 1H); 5.21 (m, 1H); 5.40
(d, J ) 4.4, 1H). MALDI-MS calcd for C57H103NO10P+ [M + H]+
992.73, found 992.65.
(m, 3H); 2.0 (m, 2H); 2.12 (m, 1H); 2.30 (m, 3H); 3.15 (m, 1H);
3.39 (s, 9H); 3.63 (m, 2H); 3.81 (m, 2H); 4.25-4.45 (m, 5H); 5.33
(d, J ) 4.4, 1H). MALDI-MS calcd for C53H99NO7P+ [M + H]+
892.72, found 892.73.
Synthetic Route 4 (4a-d) (1) 1-Stearyl-3-trityl Glycerol (22a).
The 3-trityl group was introduced according to the general
Differential Scanning Calorimetry. Differential scanning calo-
rimetric (DSC) measurements were carried out by using a high-
temperature MC-DSC 4100 calorimeter (Calorimetry Sciences
Corp., Lindon, UT) with three reusable Hastelloy sample ampoules
and a reference ampule. Data were collected typically over a range
of 5-85 °C at 0.5 °C/min with Milli-Q water as the reference.
The CpCalc 2.1 software package from Calorimetry Sciences Corp
was used to convert the raw data into molar heat capacity (MHC).
Data were then imported into Origin 6.0 (Microcal, Northampton,
MA) for further processing and calculation. Liposomes used for
DSC measurement were prepared by hydrating the lipid film (10
µmol) in Milli-Q water (200 µL) at 65 °C under argon for 15 min
with intermittent vortex. Samples were then cooled to room
temperature, degassed, and loaded into the sample ampule using
gastight Hamilton syringe (100 µL per sample). Samples were
scanned through a heating-cooling-heating cycle and the second
heating scan data were used for analysis.
Encapsulation of Calcein in Liposome. Calcein (2.49 g, 4 mmol)
was dissolved in Tris-HCl buffer (10 mM, pH 7.5, 6 mL) after the
addition of 50% sodium hydroxide (695 µL, 13.2 mmol). This stock
solution was then loaded on a Sephadex LH-20 column (2.5 cm ×
40 cm) and eluted with Tris-HCl buffer (10 mM, pH 7.5). The
concentration of pooled fraction of calcein was determined by
measuring the absorbance (494 nm) of diluted sample at pH 9. The
purified calcein (56 mM) was then encapsulated into the liposomes
for the leakage assay. Generally, the dry lipid film(10 µmol) of
given formulation was hydrated in 1 mL calcein containing buffer
at 60 °C under argon for 15 min with intermittent vortex. The
sample was then extruded through a polycarbonate membrane 11
times at 60 °C followed by eluting from a Sephadex G-50 column
with the corresponding isosmotic eluent. The pooled liposome
fractions were then analyzed to determine the calcein concentration,
and diluted to the linear fluorescence range for leakage study.
Osmotic Stress-Induced Leakage. Liposomes used for this study
were prepared according to the above-mentioned method with high
concentration content (56 mM calcein, 10 mM Tris, 711 mM NaCl),
extruded through 100 nm membrane, and eluted with the isosmotic
buffer (50 mM HEPES, 775 mM NaCl). Liposomes containing 40%
(mole) free cholesterol such as DMPC/Chol (3:2), DSPC/Chol (3:
2) and DPPC/Chol (3:2) were used as the positive control in the
leakage study of SML liposomes. Solutions of various osmotic
concentrations were prepared by mixing the calcein free isosmotic
buffer (1600 mOsm) and a 50 mOsm dilution buffer (50 mM
HEPES). Liposomes were then exposed to solutions of various
osmotic concentrations by mixing 10 µL of liposome with 990 µL
testing buffer at 37 °C. Fluorescence signal at 517 nm (ext.: 494
nm) was read after 5 min equilibration by using a Quantech
fluorometer (Barnstead/Thermolyne, Dubuque, IA). Liposomes were
then lysed by adding 100 µL of 10% Triton X-100 to release calcein
completely. The fluorescence of the total calcein was measured and
used as 100% signal (F100%). The fraction of calcein remaining in
the liposome before lysis was defined as 1- (Fsignal - Fblank)/(F100%
- Fblank), where Fsignal is the fluorescence intensity of the sample
and Fblank is the fluorescence intensity of liposome in the isosmotic
buffer.
1
procedure. TLC: Rf ) 0.11 (eluent E). H NMR (CDCl3): δ 0.87
(t, J ) 7.2, 3H); 1.27 (br, 30H); 1.55 (m, 2H); 2.40 (br, 1H); 3.18
(m, 2H); 3.32-3.53 (m, 4H). 3.94 (m, 1H); 7.23 (m, 9H); 7.43 (m,
6H). MALDI-MS calcd for C40H59O3 [M + H]+ 587.45, found
+
587.44.
(2) 1-Stearyl-2-cholesterylcarbonoyl-3-trityl Glycerol (23a). To
a solution of 22a (2.4 g, 4 mmol) and DMAP (0.6 g) in dry ethanol-
free chloroform (10 mL) was added dropwise the solution of
cholesteryl chloroformate (2.2 g, 4.8 mmol) in chloroform (5 mL)
at r.t. The reaction mixture was stirred at r.t. overnight. Then a
mixture solvent of CHCl3/MeOH/H2O (40 mL/20 mL/30 mL) was
added to the reaction mixture. The organic layer was dried over
sodium sulfate, filtered, and evaporated. The crude product was
purified by HPFC (0-10% ethyl acetate in hexane). TLC: Rf )
0.43 (eluent E).1H NMR (CDCl3): δ 0.69 (s, 3H); 0.85-1.65 (m,
68H); 1.79-2.02 (m, 5H); 2.42 (m, 2H); 3.25 (m, 2H); 3.38 (m,
2H); 3.60 (m, 2H); 4.48 (m, 1H); 5.04 (m, 1H); 5.39 (d, J ) 4.4,
+
1H); 7.27 (m, 9H); 7.43 (m, 6H). MALDI-MS calcd for C68H103O5
[M + H]+ 999.78, found 999.75.
(3) 1-Stearyl-2-cholesterylcarbonoyl Glycerol (24a). This com-
pound was synthesized according to the general procedure of
removal of trityl group. TLC: Rf ) 0.63 (eluent D).
(4) 1-Stearyl-2-cholesterylcarbonoyl-rac-glycero-3-phosphate (25a).
This compound was synthesized according to the general procedure
of phosphorylation. TLC: Rf ) 0.57 (eluent C).
(5) 1-Stearyl-2-cholesterylcarbonoyl-rac-glycero-3-phosphocho-
line (4a, SeChcPC). This compound was synthesized according to
the general procedure of phosphocholine synthesis. TLC: Rf ) 0.53
(eluent A). 1H NMR (CDCl3): δ 0.68 (s, 3H); 0.85-1.65 (m, 66H);
1.84-2.05 (m, 5H); 2.36 (m, 2H); 3.39 (s, 9H); 3.44 (m, 2H); 3.61
(m, 2H); 3.83 (m, 2H); 4.01 (m, 2H); 4.35 (m, 2H); 4.44 (m, 1H);
4.98 (m, 1H); 5.39 (d, J ) 4.4, 1H). MALDI-MS calcd for
C54H101NO8P+ [M + H]+ 922.73, found 922.73.
Synthetic Route 5 (5a-d). (1) 1-Stearoyl-2-cholesterylcarbonoyl-
sn-glycero-3-phosphocholine (5a, SChcPC). To a solution of
1-stearoyl-2-hydroxy-sn-glycero-phosphocholine (1 g, 1.91 mmol)
and DMAP (1 g) in ethanol-free dry chloroform (50 mL) was added
dropwise the chloroform solution (10 mL) of cholesteryl chloro-
formate (2 g, 4.45 mmol) at r.t. After 16 h of reaction at r.t., the
solvent was evaporated and the residue was purified by HPFC
(CHCl3 to CHCl3-MeOH-H2O 65/25/4). Yield, 1.57 g, 88%. TLC:
Rf ) 0.54 (eluent C). 1H NMR (CDCl3/MeOH-d4/pyridine-d5, 10:
2:1): δ 0.68 (s, 3H); 0.85-1.65 (m, 66H); 1.84-2.05 (m, 5H); 2.33
(t, J ) 7.6 Hz, 2H); 2.39 (m, 2H); 3.28 (s, 9H); 3.67 (m, 2H); 4.10
(m, 2H); 4.24 (m, 1H); 4.32 (m, 2H); 4.47 (m, 2H); 5.11 (m, 1H);
5.41 (d, J ) 4.4, 1H). MALDI-MS calcd for C54H99NO9P+ [M +
H]+ 937.61, found 937.67.
Synthetic Route 6 (6a-d). (1) 1-Stearoyl-2-cholesterylhemisuc-
cinoyl-sn-glycero-3-phosphocholine (6a, SChemsPC). To a solution
of 1-stearoyl-2-hydroxy-sn-glycero-phosphocholine (2 g, 3.82
mmol) and cholesterylhemisuccinate (2.68 g, 4.58 mmol) in ethanol-
free dry chloroform (60 mL) at room temperature were added
DMAP (0.47 g, 3.82 mmol) and DCC (0.95 g, 4.58 mmol). The
reaction mixture was stirred at r.t. for 24 h. The mixture was filtered
and the filtrate was diluted with 100 mL of a mixture of chloroform
and methanol (2/1, v/v), washed with 30 mL of 1 M hydrochloride.
The organic layer was dried over anhydrous sodium sulfate,
concentrated by rotary evaporation. The residue was applied to
HPFC for purification (CHCl3 to CHCl3-MeOH-H2O 65/25/4).
Yield, 3.41 g, 90%. TLC: Rf ) 0.48 (eluent C). 1H NMR (CDCl3):
δ 0.68 (s, 3H); 0.85-1.65 (m, 66H); 1.84-2.05 (m, 5H); 2.29-2.31
(m, 4H); 2.55-2.62 (m, 4H); 3.31 (s, 9H); 3.78 (m, 2H); 4.04 (m,
Leakage in 30% Fetal Bovine Serum. Calcein was encapsulated
into the liposome by the method described above. Liposomes were
extruded through 200 nm membrane and the free calcein was
removed by passing the liposomes through the Sephadex G-50
column using HEPES buffer (10 mM HEPES, 140 mM NaCl, pH
7.4) as the isosmotic eluent. Conventional liposome formulations
containing 40% cholesterol were used as the control in the long
term leakage assay. An aliquot of liposome sample (20-50 µL)
was diluted by 30% fetal bovine serum to a total volume of 2 mL.
Samples were then sealed in the glass tube and incubated at 37 °C.
Fluorescence intensities of samples were monitored at different time
9
J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008 15711

A R T I C L E S
Huang and Szoka
points and the fraction of calcein remaining in the liposome was
determined by the method described in the osmotic stress induced
leakage.
C26 Murine Adenocarcinoma Model. All animal experiments
were performed in compliance with the NIH guidelines for animal
research under a protocol approved by the Committee on Animal
Research at the University of California, San Francisco. For all
chemotherapy experiments, on day 0, BALB/c mice were given
subcutaneous injections of C-26 tumor cells (4 × 105 cells per
mouse) in the right flank and were then randomized with 5 mice
per group and numbered. On day 8, post-tumoring mice received
0.2 mL via a single tail vein injection of either phosphate buffered
saline, doxorubicin at 10 mg/kg, Doxil at 15 mg/kg, or doxorubicin
at 15 mg/kg encapsulated in 94.8:5.0:0.2 mol ratio of PChcPC-
PEG-DSPE-RT. Mice were weighed and tumor sizes were moni-
tored daily during the experimental period. The tumor volume was
estimated by measuring three orthogonal diameters (a, b, and c)
with calipers; the volume was calculated as (a × b × c) × 0.5
cm3. Tumors that were just palpable were defined as 1 mm × 1
mm × 1 mm. In each experiment, the mice were monitored for up
to 90 days postinoculation or until one of the following conditions
for euthanasia was met: (1) their body weight dropped below 15%
of their initial mass; (2) their tumor was greater than 2.0 cm across
in any dimension; (3) they became lethargic or sick and unable to
feed; or (4) they were found dead. On day 90, all surviving mice
were euthanized. All animals that survived 60 days also survived
until day 90.
Cholesterol Exchange Experiment. Unilamellar liposomes were
prepared by the extrusion method.46 The donor liposomes consisted
of 40% cholesterol (or equivalent from SML), 10% negatively
charged corresponding phosphatidylglycerol (PG), and 50% 1,2-
diacyl-sn-glycero-3-phosphocholine (PC) (or the equivalent from
SML). Specifically, the five donor liposomes were formulated at
the following molar ratios: (1) PChcPC/DPPC/DPPG (5/4/1), (2)
MChcPC/DMPC/DMPG (5/4/1), (3) ChcMaPC/DMPC/DMPG (5/
4/1), (4) Chol/DPPC/DPPG(4/5/1), and (5) Chol/DMPC/DMPG(4/
5/1). A 10-fold molar excess of neutral POPC liposome was used
as the acceptor liposome. After extrusion, the diameter of the donor
liposomes ranged from 97 to 129 nm with a narrow size distribution
(see the Supporting Information, Table 2), and 115 nm for the
acceptor liposome. The unilamellarity of the liposomes were
confirmed by 31P NMR47 (see the Supporting Information, Figure 10).
For the exchange experiments, 1 mL of donor liposomes (10 mM)
and 1 mL of acceptor liposomes (100 mM, 10-fold) were warmed
at 37 °C first, and then mixed and incubated at 37 °C. An aliquot
(250 µL) of the mixture was sampled at a given time point and
applied to a small (ca. 2 cm in length) anion exchange column
(Q-Sepharose XL). The column was pretreated with 0.1 mL of 10
mM POPC before the loading of the exchange sample to reduce
the nonspecific binding of the neutral liposome. The column was
eluted with 1 mL pH 7.4 10 mM NaCl, 10 mM HEPES buffer.
The eluate was lyophilized and analyzed by the cholesterol assay48
to quantify the amount of cholesterol exchanged. Briefly, the
lyophilized lipid powder was dissolved in 100 µL of distilled water
and 50 µL was transferred to screw-capped glass tube (13 mm ×
100 mm). After the addition of 5 mL of cholesterol assay reagent
(339 mg ferric perchlorate hexahydrate in 300 mL of ethyl acetate
mixed with 200 mL concentrated sulfuric acid at 4 °C), the mixture
was heated at 100 °C for 90 s, immersed immediately into
ice-water. The absorbance at 610 nm was measured and the amount
of cholesterol was calculated according to the standard curve.
Cytotoxicity Determination. The cytotoxicity of the SML lipids
was evaluated on C26 cells. After 3 days incubation, cell viability
was measured with the standard MTT (3-(4, 5-dimethylthiazolyl-
2)-2, 5-diphenyltetrazolium bromide) assay method49 as described.50
Statistical Analysis. Statistical analysis was performed using
SigmaPlot version 11 (San Jose, CA). To evaluate the therapeutic
efficacy, one-way ANOVA and Student-Newman-Keuls pairwise
comparisons were performed using tumor growth delay data39
derived from the tumor growth curve. Survival data was analyzed
by the log-rank test.
Acknowledgment. We thank the NIH EB003008 and GM61851
for supporting this research. We acknowledge the excellent technical
assistance of Dr. Mahmoud Reza Jaafari and Nichole Macaraeg.
Thanks are also extended to the UCSF Mass Spectrometry Facility
(A.L. Burlingame, Director) supported by NIH NCRR RR0161.
Supporting Information Available: Synthesis of SMLs, DSC
thermograms, and the tabulated data of Tm and ∆H, cytotoxicity
of SMLs, electron microscope of SML liposomes, size and 31
P
NMR of liposomes, biodistribution study of liposomal doxo-
rubicin, and complete ref 27. This material is available free of
charge via the Internet at http://pubs.acs.org.
(46) Szoka, F.; Olson, F.; Heath, T.; Vail, W.; Mayhew, E.; Papahadjopou-
los, D. Biochim. Biophys. Acta 1980, 601, 559–571.
(47) Frohlich, M.; Brecht, V.; Peschka-Suss, R. Chem. Phys. Lipids 2001,
109, 103–12.
(48) Wybenga, D. R.; Pileggi, V. J.; Dirstine, P. H. J. D. Clin. Chem. 1970,
16, 980–984.
(49) Mosmann, T. J. Immunol. Methods 1983, 65, 55–63.
JA8065557
(50) Huang, Z.; Li, W.; MacKay, J. A.; Szoka, F. C., Jr. Mol. Ther. 2005,
11, 409–17.
9
15712 J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008