In vitro and in vivo evaluation of positively charged liposaccharide derivatives as oral absorption enhancers for the delivery of anionic drugs

Article abstract of DOI:10.1002/jps.21999

Oral delivery of hydrophilic, ionisable drugs remains a major challenge in drug development and a number of active pharmaceuticals fail to reach the market of oral drugs because of a lack of absorption and/or stability issues. One possible approach to improving the bioavailability of such drug candidates is to increase their lipophilicity, which is a key parameter in the permeation across cell membranes. However, modifying the chemical structure by adding lipid residues often results in changes in activity. With ionised molecules, ion-pairing can be considered to associate charged lipid moieties with the parent drug without altering its structure and therefore activity. This study presents the results of in vitro and in vivo evaluation of a series of synthetic, positively charged liposaccharide derivatives combined with an anionic model drug, piperacillin. The antimicrobial activity, plasma stability, Caco-2 cell permeability and oral absorption of the conjugates were assessed. Increases in apparent permeability were observed in vitro for three of the tested formulations, while retaining the antibacterial activity of the drug. However, the in vivo intestinal absorption of piperacillin formulated with the liposaccharide derivatives was found unchanged, possibly due to molecular dissociation, early degradation or structural differences between the intestinal epithelium and cultured monolayers.

Full text of DOI:10.1002/jps.21999

In Vitro and In Vivo Evaluation of Positively Charged
Liposaccharide Derivatives as Oral Absorption
Enhancers for the Delivery of Anionic Drugs
JULIE A. BERGEON,1 ZYTA M. ZIORA,2 ADEL S. ABDELRAHIM,1 NIKLAS U. PERNEVI,1 ANNE R. MOSS,1 ISTVAN TOTH1,2
1School of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane, Qld 4072, Australia
2Centre for Integrated Preclinical Drug Development (CIPDD), The University of Queensland, Brisbane, Qld 4072, Australia
Received 22 May 2009; revised 23 September 2009; accepted 28 September 2009
Published online 12 November 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21999
ABSTRACT: Oral delivery of hydrophilic, ionisable drugs remains a major challenge in drug
development and a number of active pharmaceuticals fail to reach the market of oral drugs
because of a lack of absorption and/or stability issues. One possible approach to improving the
bioavailability of such drug candidates is to increase their lipophilicity, which is a key parameter
in the permeation across cell membranes. However, modifying the chemical structure by adding
lipid residues often results in changes in activity. With ionised molecules, ion-pairing can be
considered to associate charged lipid moieties with the parent drug without altering its structure
and therefore activity. This study presents the results of in vitro and in vivo evaluation of a series
of synthetic, positively charged liposaccharide derivatives combined with an anionic model drug,
piperacillin. The antimicrobial activity, plasma stability, Caco-2 cell permeability and oral
absorption of the conjugates were assessed. Increases in apparent permeability were observed
in vitro for three of the tested formulations, while retaining the antibacterial activity of the drug.
However, the in vivo intestinal absorption of piperacillin formulated with the liposaccharide
derivatives was found unchanged, possibly due to molecular dissociation, early degradation or
structural differences between the intestinal epithelium and cultured monolayers. ß 2009 Wiley-
Liss, Inc. and the American Pharmacists Association J Pharm Sci 99:2333–2342, 2010
Keywords: charged liposaccharide; ion-pairing; Caco-2 cells; calorimetry (ITC); absorption
enhancer; piperacillin; oral drug delivery; bioavailability
INTRODUCTION
ability, essentially by passive diffusion, have been
investigated.3,4 Considering the lipidic nature of
the bilayered cell membrane, the derivatisation of
drugs by lipidation has been extensively investigated
and improvements in enzymatic stability and mem-
brane permeability have been reported.5,6 Although
increased lipophilicity can facilitate absorption across
membranes, it is also frequently associated with
solubility difficulties in aqueous media.
Another approach to overcome absorption problems
revolves around the ionisability of polar molecules.
Hydrophilic drugs are commonly commercialised as
salts (e.g., sodium, acetate, chloride salts), more
soluble and often more stable7 than the parent
molecule, which means that the drug itself can
potentially exhibit charges depending on pH condi-
tions (both in solution and in situ). This offers ample
possibilities to associate lipophilic counter-ionic
moieties by ion-pairing in order to enhance the
overall lipophilicity of these hydrophilic drugs while
preserving their biologically active chemical struc-
ture.8 Our recent investigations have demonstrated
Recent advances in drug development have shown the
crucial need for developing new active drugs to treat
and/or prevent diseases which currently resist con-
ventional therapies.
However, the promising therapeutical profiles of a
large number of potential drug candidates (including
ionic/ionisable molecules, peptides and protein-like
pharmaceuticals with a high hydrophilic character) is
often dampened by issues such as low oral bioavail-
ability and poor in situ stability.1,2 To overcome these
difficulties, different strategies such as structural
modifications by adjunction of various protecting
groups to improve stability, or use of surfactants and
absorption enhancers to increase membrane perme-
Correspondence to: Julie A. Bergeon (Telephone: 617-3346-9893;
Fax: 617-3365-4273; E-mail: j.bergeon@uq.edu.au)
Journal of Pharmaceutical Sciences, Vol. 99, 2333–2342 (2010)
ß 2009 Wiley-Liss, Inc. and the American Pharmacists Association
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010
2333

2334
BERGEON ET AL.
the potential of charged glycolipids to form ion-pairs
with a model drug like piperacillin.9,10
liposaccharide derivatives were mixed with the
commercial salt and dissolved in pure glacial acetic
acid; the excess acid was removed under vacuo and
the mixture was then lyophilised; (ii) with piperacillin
acid, the liposaccharide derivatives and the drug were
dissolved in water/acetonitrile (1:1) and lyophilised.
ESI-MS and MS/MS analyses were carried out on a
PE Sciex API3000 triple quadrupole mass spectro-
meter, using a mixture of solvent A (1% formic acid in
water) and B (1% formic acid in 9:1 acetonitrile/water)
In an earlier study, we reported that the in vitro
permeability of piperacillin across Caco-2 monolayers
could be notably improved by associating this anionic
b-lactam antibiotic to a positively charged liposac-
charide containing a D-glucose moiety coupled to a 12-
carbon long lipoamino acid (LAA) (a-amino acid with
lipophilic alkyl side-chain).10 Positive results were
also observed when varying the lipid chain length.
LAAs are versatile molecules with the unique ability
to be linked through either their amino or carboxyl
extremities. They can be readily obtained from alkyl
bromide precursors in a few, short and easy synthetic
steps. LAAs combined with carbohydrates offer
certain advantages versus lipids alone; for example,
it allows for a modulation of the degree of lipophilicity
brought by the lipid chain via the hydroxyl residues
present on the carbohydrate moiety, hence preserving
the water solubility characteristics of the modified
drug while still increasing its lipophilicity.
at 0.1 mL/min. For LC–MS/MS experiments,
Phenomenex luna C18 column (5 mm, 50 mm Â
2.0 mm) equipped with C18 guard column
a
a
(4 mm  3.0 mm) was attached to the mass spectro-
meter and a 30–100% gradient of B in A over 4 min
was used for elution at a flow rate of 0.5 mL/min with
a 1:10 splitter upstream from the ionisation source
(Shimadzu LC-10AT system).10 Data were acquired
with Analyst 1.4 software (Applied Biosystems/MDS
Sciex, Toronto, Canada).
The amine function of the LAAs is subject to acido-
basic equilibrium and therefore is not constantly
protonated, which can be problematic as this residual
charge is the key element of the interaction between
the drugs of interest and counter ionic liposacchar-
ides. Quaternisation of the terminal amine was
therefore considered to create a constantly, positively
charged residue. Also, as the sugar moiety introduced
a significant degree of hydrophilicity, additional
lipophilic components were considered in the design
of new charged liposaccharides for a better, adjus-
table balance of lipophilicity versus hydrophilicity.
Isothermal Titration Microcalorimetry (ITC)
ITC measurements were performed on a MicroCal
VP-ITC microcalorimeter (Northampton, MA), with
Origin 5.0 and VP viewer 2000 software. Experiments
were performed at 378C in purified water (MiiliQ
Gradient A10, Millipore, North Ryde, NSW, Austra-
lia) with degassed, sonicated samples. The reference
cell was loaded with water while the sample cell
(1.4395 mL) was charged with piperacillin acid (0.05–
0.5 mM) and rotated at 300 rpm. The liposaccharide
derivatives (4 mM) were injected into the sample cell
using a microsyringe at a rate of 3–10 mL every 6 min
(total injection volume 300 mL).
MATERIALS AND METHODS
Chemistry and Analytical Characterisation
Antimicrobial Assays (MIC)
All solvents and reagents were obtained at the
highest available purity from Sigma–Aldrich (Castle
Hill, NSW, Australia) and used directly without
further purification. The liposaccharide derivatives
presented in this publication were prepared according
to the methodology described by Abdelrahim et al.11
Commercially available piperacillin sodium was
converted into piperacillin acid using an IR-120-
[Hþ] cationic-exchange resin (Rohm Haas, Philadel-
phia, PA). The salt was dissolved in water and
vigorously stirred with the resin for 30 min, until pH
stabilised at 3 and piperacillin acid precipitated. The
precipitate was then dissolved in acetonitrile and the
resin was filtered off, washed with acetonitrile, and
the filtrate was then lyophilised on an Alpha 2-4/LSC
(Martin Chris, Osterode am Harz, Germany) at
À808C, <150 psi. Two methodologies for lyophilisa-
tion were used; (i) with piperacillin sodium, the
Minimum inhibitory concentration (MIC) experi-
ments were carried using a broth dilution method.
Two strains of bacteria were used, Escherichia coli
(MC4100) and Pseudomonas aeruginosa (A01). Bac-
teria were cultured in Luria-Bertani (LB) broth
medium (1%, w/v NaCl, 1%, w/v tryptone, 0.5%, w/v
yeast extract, adjusted to pH 7 and autoclaved at
1218C for 20 min before use; reagents purchased from
Sigma–Aldrich). Bacterial growth was monitored by
optical density (OD) measurements, performed on an
Ultrospec 2000 (Pharmacia Biotech, Uppsala, Swe-
den). Bacteria were grown to an OD ꢀ 1 then diluted
with LB broth and subcultured again to reach an OD
of 0.3–0.8. They were then diluted to a count of
1 Â 106 cfu/100 mL (OD 0.1) and plated in 96-well
round bottom plates (TPP, Zurich, Switzerland).
Lyophilised preparations of piperacillin and lipo-
saccharides were dissolved in 5% dimethylsulfoxide
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010
DOI 10.1002/jps

CHARGED LIPOSACCHARIDE DERIVATIVES AS ORAL ABSORPTION ENHANCERS
2335
(DMSO)/water and serially diluted (1:3) with LB
broth to reach a final piperacillin concentration of 0–
25 mM for E. coli and 0–52 mM for P. aeruginosa. The
bacteria (100 mL) were then mixed with the drug
solutions (100 mL) and incubated for 16–20 h at 378C,
after which bacterial growth was assessed by OD
measurements at 600 nm on a Spectramax 250 plate
reader (Molecular Devices, Sunnyvale, CA). The
MICs for each formulation was then determined by
plotting the OD values versus piperacillin concentra-
tions.
Caco-2 cells from the American type culture
collection (Rockville, MD) were maintained in Dul-
becco’s modified Eagle’s medium (DMEM) supple-
mented with 10% foetal bovine serum (FBS), and 1%
nonessential amino acids at 95% humidity and 378C
in an atmosphere of 5% CO2. The medium was
changed every second day. After reaching 80%
confluence, the cells were subcultured using 0.2%
ethylenediaminetetraacetic acid (EDTA) and 0.25%
trypsin. One hundred microliters of a solution
containing ꢁ106 cells/mL (passage number 31–40)
were seeded onto polycarbonate Transwell1 inserts
supplied by Costar (Cambridge, MA; mean pore
size ¼ 0.45 mm, 6.5 mm diameter, 0.33 cm2 surface
area) and cultivated in DMEM supplemented with
10% FBS, 1% nonessential amino acids, 100 U/mL
penicillin and 100 mg/mL streptomycin. The cells
were allowed to grow for 21–28 days and the medium
was changed every other day.
The tested compounds were dissolved in HBSS–
HEPES (HBSS buffered with 25 mM 4-(2-hydro-
xyethyl)piperazine-1-ethanesulfonic acid (HEPES)
at pH 7.4) to a final concentration of 200 mM. Prior
to the transport studies, the Caco-2 monolayers were
washed three times with prewarmed HBSS–HEPES.
The integrity of the monolayers was assessed prior to,
and after each experiment by measuring the transe-
pithelial electrical resistance (TEER) values using a
Millicell-ERS system (Millipore Corp., Bedford, MA).
TEER values (V cm2) were calculated by the following
equation:
Haemolytic Assays
Haemolytic assays were undertaken with approval
from the University of Queensland Ethics Committee
(approval # 2006000950). Full venous blood samples
were taken from healthy adult volunteers. Human
red blood cells (hRBCs) were isolated from the whole
blood by centrifugation at 2576 G for 15 min (Sigma
2–5 centrifuge, Sigma Laborzentrifugen GmbH,
Osterode am Harz, Germany). Plasma was removed
and the cells were washed several times with
phosphate-buffered saline (PBS) before being resus-
pended in PBS (to original blood volume) and
aliquoted (100 mL) in 96-well flat bottom plates (TPP).
Solutions of the tested compounds were prepared in
PBS (0.2–20 mM) and dispensed in the wells contain-
ing the hRBCs preparations. The plates were then
incubated at 378C for 60 min on a Heidolph Shaking
Incubator & Titramax 1000 (Schwabach, Germany)
at 400 rpm, after which they were centrifuged at
2129g for 15 min; 75 mL of the supernatant was then
removed and placed into new plates before absor-
bance was measured at 540 nm using a Spectramax
250 microplate reader. PBS was used as a negative
control; water and Sodium dodecylsulphate (SDS)
were used as positive controls. The percentage of
haemolysis caused by each of the tested formulation
was calculated by the following equation:
TEER ¼ ðRtotal À RblankÞ Â A
where Rtotal is the resistance of the tested monolayer,
R
blank the resistance of the filter (background) and A
the filter surface area. Monolayers with TEER values
<250 V cm2 were discarded.
The drug solutions (100 mL) were added to the
donor side of the monolayers and the Transwell1
plates were placed in a shaking incubator set to
400 rpm and 378C for the duration of the experiment.
At predetermined time points (30, 90, 120 and
150 min), samples (400 mL) were taken from the
receiver chamber and replaced with the same amount
of HBSS–HEPES. All compounds were tested in two
to three independent assays, using four wells each
time. Concentrations in piperacillin were determined
by LC–MS/MS. The transport of paracellular marker
14C-mannitol (200 mM) was also measured separately
(n > 3) as a control of membrane integrity under assay
conditions for each passage number of cells used.
After dilution of samples (400 mL) with 4 mL of Wallac
OptiPhase HiSafe 3 liquid scintillation cocktail
(Montreal, Canada), radioactivity was measured
using a Beckman Coulter LS650 Multipurpose liquid
scintillation spectrometer (Beckman Instruments,
Fullerton, CA) and permeability values determined
A540 À minA540
%haemolysis ¼
 100
maxA540 À minA540
where minA540 represents the absorbance at 540 nm
of PBS alone and maxA540 represents the absorbance
at 540 nm of SDS.
Caco-2 Permeability Assay
Cell culture reagents were purchased from Gibco-
BRL (Grand Island, NY) except of Hank’s balanced
saline solution (HBSS) and 14C-labelled mannitol
which were supplied by Sigma–Aldrich and Amer-
sham Biosciences (Piscataway, NJ), respectively.
Tissue culture flasks (75 cm2) were obtained from
BD Bioscience (Franklin Lakes, NJ).
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010

2336
BERGEON ET AL.
by the following equation:
of variance (ANOVA) of repeated measures to a
significance level of 0.05 (p < 0.05), followed by
Tukey’s post hoc test (multiple pairwise comparison
of means). Computations were realised using Graph-
Pad Prism v5.01 (GraphPad software, La Jolla, CA).
dC Vr
¼
dt AC0
Papp
where dC/dt is the steady-state rate of change in the
chemical in the receiver chamber (mM, or dpm mLÀ1),
Vr is the volume of the receiver chamber (mL), A is the
surface area of the cell monolayers (cm2) and C0 is the
initial concentration in the donor chamber (mM, or
dpm mLÀ1).
RESULTS AND DISCUSSION
Design and Preparation of a New Series of Charged
Liposaccharide Derivative Carriers
Oral Absorption Studies
A novel series of liposaccharide derivatives was
synthesised by modifying the amino extremity
of the liposaccharide. The resulting compounds
comprised a D-glucose (G) moiety and a 12-carbon
long LAA (C12) scaffold, onto which additional amino
residues, namely leucine, phenylalanine (two of the
most hydrophobic amino acids)12 and another C12
were attached. The N-terminus amino group was
then converted into a quaternary ammonium salt
using Amberlite 400 [HOÀ] as a base11 and the
obtained liposaccharide derivatives were lyophilised
together with the model drug piperacillin (Fig. 1).
Initially, lyophilisation was attempted directly
with the commercially available piperacillin sodium
salt under acidic conditions (glacial acetic acid);
however, preliminary observations (binding studies
and nuclear magnetic resonance analyses, results not
shown) suggested that limited interactions occurred
between the anionic drug and its cationic liposac-
charide counterparts. It is believed that this could be
due to strong ionic binding between the piperacillin
anion and the sodium cation, which is smaller and has
a greater charge density compared to the charged
liposaccharide derivatives (calculated Hu¨ckel charge
of 0.788 (GC12C12)); as a consequence, the exchange
between the two cations was likely to be thermo-
dynamically unfavourable. To facilitate this substitu-
Animal studies were undertaken with approval from
the University of Queensland Ethics Committee
(approval #SMMS/002/08/ARC). Eight-week-old male
Sprague–Dawley rats (300–350 g, four rats/com-
pound) were administered formulations of the tested
compounds in 5% DMSO/water by oral gavage, to a
level of 50 mg/kg of piperacillin for each rat. At
regular time points (0, 5, 10, 15, 30, 60, 90, 120 min),
blood samples (200 mL) were collected in heparinised
tubes by tail vein bleeding. The rats were euthanised
immediately after the experiment by inhalation of a
1:1 O2/CO2 mixture; death was further asserted by
cervical dislocation.
Blood samples were centrifuged for 20 min at 1717g
and the plasma was collected, treated with acetoni-
trile (1:1,v/v) to precipitate the proteins and centri-
fuged. The supernatant was then collected for
analyses by LC-MS/MS (or kept at À308C if analyses
could not be run immediately).
Statistical Analysis
Results were expressed as mean values Æ standard
error mean (SEM). Where applicable, statistical
analysis of values was performed by one-way analysis
Figure 1. Synthesised charged liposaccharide derivatives associated with model drug piper-
acillin.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010
DOI 10.1002/jps

CHARGED LIPOSACCHARIDE DERIVATIVES AS ORAL ABSORPTION ENHANCERS
2337
tion, piperacillin sodium was dissolved in water and
stirred with cation-exchange resin under
and therefore ion-pairing, if present, could not be
significantly assessed.
a
controlled pH to obtain piperacillin acid, which was
then immediately lyophilised with different liposac-
charide derivatives from a 1:1 acetonitrile/water
solution. The prepared conjugates were stored at
À208C and their integrity was monitored by analy-
All liposaccharide derivatives were therefore tested
in association with piperacillin in a 1:2 drug/counter-
ion ratio. A 1:1 ratio was also examined to see if even a
lower quantity of liposaccharide derivatives was
sufficient to generate an increase in oral absorption
(penetration enhancing effect).
1
tical examination by H-NMR.
Different ratios of drug/liposaccharide were con-
sidered in this study. Initial measurements of binding
by ITC revealed that two concurrent processes were
taking place, binding (ion-pairing) and aggregation. A
1:1 ratio did not seem to be sufficient to observe
optimum ionic interactions, possibly due to the
presence of multiple interaction sites on the model
drug. Indeed, when examining interactions between
piperacillin and GC12C12, a ratio close to 1:2 drug/
liposaccharide was found to give optimal ion inter-
actions (Fig. 2). The negative enthalpy observed
initially below a 1:2 ratio is indicative of favourable
ion-pairing, while above 1:2, the process becomes
endothermic as a result of aggregation occurring. As
the ratio of drug/liposaccharide was increased, the
aggregation of the lipidic cations was predominantly
observed over the binding process (results not shown)
Antimicrobial Activity
The antimicrobial activity of the drug conjugates was
assessed in vitro in two Gram-negative bacterial
strains, E. coli and P. aeruginosa. MICs were
determined by
a
broth dilution method13 and
compared to the MIC of piperacillin alone. Previously
synthesised 2-amino-N-(20-a,D-gluco)dodecanamide
(GC12)10 was also tested in the same 1:1 and 1:2
molar ratios.
After 16–20 h incubation with bacterial cultures at
378C, resistance to piperacillin alone was found to be
in agreement with reference values of 2–4 mg/mL for
E. Coli14 and 3–6 mg/mL for most strains of
P. aeruginosa.15,16 When the drug was tested in
association with the synthesised charged liposacchar-
ide derivatives, no significant change was observed in
the MIC values for P. aeruginosa, all averaging 7–
12 mg/mL (Fig. 3). For E. coli, slight, however not
significant, increases in the MICs were noticed when
piperacillin was formulated with GC12Phe (1:1 and
1:2) and GC12C12 (1:1).
The pairing of the liposaccharide derivatives
to piperacillin therefore does not appear to have
negatively impacted the original antimicrobial activ-
ity against either bacterial strain. The absence of
significant variation in MICs (verified using a one-
Figure 3. MICs of piperacillin alone and formulated with
the synthesised liposaccharide derivatives for two Gram-
negative bacterial strains, E. coli and P. aeruginosa. Con-
centrations were determined by optical density measure-
ments at 600nm after 24 h incubation in bacterial culture
medium at 378C. (Values given as mean Æ SEM, n ꢀ 3.)
Figure 2. Isothermal titration microcalorimetry (ITC)
binding assay between piperacillin (acid form, 0.2mM)
and liposaccharide derivative GC12C12 at 4mM in purified
water.
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010

2338
BERGEON ET AL.
way Anova statistical comparison followed by Tukey’s
post hoc test) when using different ratios of liposac-
charides would indicate that the modified liposac-
charides do not have antimicrobial or antagonist
properties of their own.
second C12 moiety and the haemolytic properties of
the liposaccharide. At commonly used therapeutic
concentrations however (i.e., <1 mM when consider-
ing the standardised intravenous administration of
piperacillin sodium), most liposaccharide derivatives
tested showed <25% haemolysis, which is considered
to be the threshold for tolerable toxicity to hRBCs.19
However, piperacillin in association with GC12C12 in
1:1 and 1:2 ratios yielded synergistic effects resulting
in increased haemolysis, to levels comparable to SDS.
When examining the EC10 (drug concentration
causing 10% haemolysis of hRBCs; see Tab. 1—the
values were estimated from the graphical data
presented in Fig. 4), GC12 and GC12Leu were found
to be moderately haemolytic (0.5 mM < EC10 < 5 mM)
while the other two liposaccharide derivatives
GC12Phe and GC12C12 as well as the combined
Pip–GC12C12 were classified as strongly haemolytic
(0.5 mM < EC10), along with SDS.18
These findings are of great importance when
considering LAAs as penetration enhancer candi-
dates for general drug delivery. Although the con-
centrations used in this assay are far greater than
those used in vivo, the LAA-based molecules tested
here showed noticeable signs of toxicity against
hRBCs. However, as the compounds were designed
for oral delivery, the indication of haemolytic toxicity
was not a limiting factor for the in vivo experiments.
Haemolytic Activity
Although destined to oral drug delivery, the liposac-
charide derivatives and their ionic complexes with
piperacillin were nonetheless tested for signs of
haemolytic activity, which could cause disruptions
to the biological membranes and yield to toxicity.17,18
Haemolytic activity was assessed in triplicate against
hRBCs over 60 min at six different concentrations,
ranging from 0.1 mM to 10 mM; the tested solutions
were formulated in PBS and compared to piperacillin
alone. SDS, a surfactant known to be toxic to hRBCs
at the concentrations used in this assay, was also
included as a positive control. The percentage of
haemolysis was determined by absorbance readings
at 540 nm.
Initially, only the synthesised liposaccharide deri-
vatives were considered; one of the complexes, Pip–
GC12C12, was additionally evaluated to examine any
possible synergistic effect.
As shown in Figure 4, the synthetic liposaccharide
derivatives expectedly induced haemolysis of hRBCs
to some extent, yet the values remain significantly
less than those of SDS in the low concentration range.
Interestingly, the comparison between GC12 and
GC12C12 revealed a twofold increase on average in
haemolytic activity, suggesting a direct correlation
between the greater lipophilicity brought by the
In Vitro Permeability Assessment in Caco-2 Cells
Prior to conducting oral absorption studies, the
in vitro permeability of the synthesised compounds
was assessed in Caco-2 cells. This cell line originates
Figure 4. Haemolytic activity measured against hRBCs at 378C over 1h of piperacillin (alone
and with GC12C12) and the synthesised liposaccharide derivatives in PBS. The percentage of
haemolysis was calculated by absorbance readings at 540nm (results given as mean Æ SEM,
n ¼ 3).
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010
DOI 10.1002/jps

CHARGED LIPOSACCHARIDE DERIVATIVES AS ORAL ABSORPTION ENHANCERS
2339
Table 1. Estimated EC10 of drug, liposaccharide
derivatives and combined entities
Piperacillin alone expectedly showed very
low permeability across Caco-2 monolayers; the
observed values were higher than reported pre-
viously, 3.07(OH)-3.22(Na) Â 10À7 cm sÀ1 versus
Compound
EC10 (mM)
3.24(Na) Â 10À8 cm sÀ1 10
.
A similar 10-fold variation
Piperacillin
GC12
GC12Leu
GC12Phe
GC12C12
Pip–GC12C12 [1:1]
Pip–GC12C12 [1:2]
SDS
ꢂ10
3.75
1.04
0.37
0.32
0.11
0.13
0.17
was noticed for Pip–GC12 [1:1] and 14C-mannitol, this
variability could be explained by the difference in cell
passage numbers, since the cells used in the present
study were much younger than in the previous
examination. Reports in literature have shown that
the differentiation of the cells was affected by early
passage numbers, mainly a lower cohesion of the
monolayer and reduced tightness of the paracellular
junctions.21 This was also evidenced by measurement
of TEER,22 which were slightly lower than in previous
experiments.
As shown in Figure 5, piperacillin demonstrated a
marked increase in membrane permeability when
associated in a 1:1 ratio with GC12, which is in
agreement with previously reported data.10 Surpris-
ingly, the penetration of piperacillin was not sig-
nificantly improved when combined with GC12 in a
1:2 molar ratio. No satisfactory ITC data could be
obtained for GC12 so the ratio for optimal binding
could not be assessed accurately. It is also possible
that alternative phenomena occurred that could
decrease the permeability enhancing effects of
GC12 (e.g., efflux, aggregation, adhesion to mem-
branes).
When testing GC12C12 in a 1:1 ratio, a slight
increase was observed, lesser though than with
GC12 [1:2]. This could be correlated to the ITC data
which suggest that a 1:1 ratio was not sufficient in
the case of GC12C12 to obtain optimal binding
between the parent drug and the charged liposac-
charide derivatives. Indeed, when assessing the
drug in a 1:2 ratio with the liposaccharide derivative,
the permeability was greatly enhanced to reach a
mean value of 1.27 Â 10À5 cm sÀ1, nearly 40 times
higher than piperacillin alone. Drugs with apparent
permeability coefficients across Caco-2 monolayers
of 10À6 cm sÀ1 or greater usually present excellent
oral bioavailability;20 the results obtained in this
in vitro assay therefore suggest that Pip–GC12C12
[1:2] might be a good candidate for oral absorption
evaluation. The other liposaccharide derivatives
tested in formulation with piperacillin did not yield
significant changes in permeability, independent of
the ratio used.
from a human colorectal adenocarcinoma and is
commonly used as a model for predicting the
transport of drugs across the intestinal epithelium.
When cultured, Caco-2 cells spontaneously form
differentiated, polarised monolayers which express
most structural and functional characteristics of the
small intestine, including enzymes and efflux pro-
teins. Artursson et al.20 demonstrated that the
apparent permeability coefficients determined with
this model can be correlated to human oral drug
absorption values with good respect.
Two hundred micromolars solutions of piperacillin
associated with the different liposaccharide deriva-
tives (1:1 and 1:2 molar ratios) in HBSS were added to
the apical side of the Caco-2 monolayers and
permeability to the basolateral side of the monolayer
was studied over 150 min, using the paracellular
marker 14C-mannitol as a negative control (typical
permeability values of 14C-mannitol are around 1–
5 Â 10À7 cm sÀ1).20 Apparent permeability coefficients
(Fig. 5) were calculated from concentrations of
piperacillin determined by liquid chromatography
coupled to mass spectrometry (LC-MS/MS) against
an eight point standard curve while 14C-mannitol
was quantified by liquid scintillation counting
(b-emission).10
The major increase in permeability observed with
Pip–GC12C12 [1:2] can also, to a certain extent, be
linked to its toxicity, as evidenced earlier in the
haemolytic assay. The measurement of the TEER
values prior to, and after each experiment revealed a
drop in resistance across the monolayers utilised with
Pip–GC12C12, sign of a certain loss of integrity of the
membrane.
Figure 5. Apparent permeability coefficients determined
in 21- to 28-day-old monolayers of Caco-2 cells (passage 31–
40) over a 150 min period at 378C (results given ÆSEM,
n ¼ 2–3; four wells/exp).
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010

2340
BERGEON ET AL.
[1:2]. Male Sprague–Dawley rats (300–350 g) were
administered the formulations by oral gavage; blood
samples were taken from the tail vein over a 2 h
period and analysed by LC-MS/MS. Piperacillin given
both orally and intravenously (i.v.) was used as a
control. The i.v. curve (Fig. 7a) demonstrates a rapid
distribution of piperacillin after injection with mean
peak concentration of 70 mg mLÀ1 at around 5 min.
The antibiotic is then rapidly cleared from the
circulatory system with virtually no drug detected
after 1 h.
The oral uptake of piperacillin was calculated from
the AUC120 (mean area under the curve, see Tab. 2)
and was expectedly very low (0.5%). When examining
the absorption of the piperacillin conjugates, the
absorption curves were very similar to that of
piperacillin alone, administered orally. Surprisingly,
the improvements in permeability noticed in vitro
with the Caco-2 monolayer model were not observed
in vivo, either with GC12 (1:1) or GC12C12 (1:1 and
1:2). The determination of the AUCs showed 0.14% to
Figure 6. TEER variations in Caco-2 monolayers mea-
sured in DMEM at 160 min and expressed as a percentage of
the initial TEER values (results given ÆSEM, n ¼ 2–3; four
wells/exp).
As shown in Figure 6, the variations in TEER
values after assay (160 min) were in general not
significantly different from that observed in case of
piperacillin tested alone (and considered nontoxic to
cell membranes). The only notable variations were
observed with GC12C12-based formulations (in both
ratios) which did seem to affect the resistance of the
monolayer. Interestingly, Pip–GC12C12 [1:1],
although seemingly causing some disruption to the
monolayers, did not display an enhanced permeabil-
ity profile (Fig. 5). Increases in permeability also
observed with Pip–GC12 [1:2] could not here be
correlated to marked changes in the monolayer.
It is unknown whether the resistance
decreases observed with GC12C12 formulations were
permanent or if the monolayers would recover given
time and optimum culture conditions. The data
obtained could suggest that in the case of Pip–
GC12C12 [1:2], permeability increases could be simply
the result of penetration enhancing effects generated
by the concentration of lipid moieties present in the
formulation. A number of lipidic excipients and other
surfactants used in pharmaceutical formulations
have been shown to increase permeability by mem-
brane permeation enhancing effects.23,24 However,
most are reported to yield marked decreases in TEER
values, which in this study were only observed in the
case of GC12C12. Thorough investigations would be
required to establish the exact mechanism by which
some of the synthesised liposaccharide derivatives
seem to enhance the in vitro apparent permeability of
piperacillin.
Figure 7. (a) In vivo pharmacokinetics of piperacillin
sodium administered intravenously at 50 mg/kg in saline
(0.9% NaCl) over a 2 h period; piperacillin concentration in
plasma was determined by LC-MS/MS (results given
ÆSEM, n ¼ 3–5). (b) In vivo absorption and pharmacoki-
netics of oral formulations of piperacillin, alone or in asso-
ciation with liposaccharide derivatives GC12 and GC12C12
,
Oral Absorption in Rats
in 1:1 and 1:2 ratios in water þ5% DMSO, over a 2 h period.
A dose equivalent to 50 mg piperacillin was administered by
oral gavage to 300–350 g Sprague–Dawley rats and pipera-
cillin concentration in plasma determined by LC-MS/MS
(results given ÆSEM, n ¼ 3–5).
Based on the in vitro results and particularly the
permeability assays, four of the tested formulations
were evaluated in vivo, namely Pip–GC12 [1:1], Pip–
GC12 [1:2], Pip–GC12C12 [1:1] and Pip–GC12C12
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010
DOI 10.1002/jps

CHARGED LIPOSACCHARIDE DERIVATIVES AS ORAL ABSORPTION ENHANCERS
2341
Table 2. Area Under the Curve (AUC120) values for the
in vivo pharmacokinetics profile of piperacillin alone (Oral
and i.v.) and in association with liposaccharide
derivatives GC12 and GC12C12
ratios of piperacillin and amphiphilic conjugates to
use to observe strong interactions between the two
charged entities, and led to two ratios being
considered for further examinations, respectively
1:1 and 1:2 drug/liposaccharide.
AUC120
(mg h/mL)
AUC120
(rel %)
Antimicrobial assays revealed that the MIC of drug
was practically unchanged when piperacillin was
combined with the liposaccharide derivatives. When
examining the effect of the tested conjugates against
hRBCs, an increase of haemolytic activity was
observed, up to levels similar to those of SDS in the
case of Pip–GC12C12; the values observed denoted a
direct correlation between the lipophilicity of the
conjugates and their haemolytic activity. The in vitro
evaluation of the synthesised associates in Caco-2
monolayers showed that apparent membrane perme-
ability values were significantly improved with three
formulations, Pip–GC12 [1:1], Pip–GC12C12 [1:1] and
Pip–GC12C12 [1:2]. The latter displayed in vitro
permeability coefficients nearly forty times higher
than the parent piperacillin drug alone. This large
increase could be correlated to disrupted monolayers
as a decrease in TEER values was observed when
testing this compound, suggesting a mechanism
oriented towards permeation enhancement. How-
ever, none of the other synthetic liposaccharides
examined had any significant effect on the resistance
of the cell monolayers. Unfortunately, these promis-
ing results were not confirmed in vivo upon oral
administration by gavage. It is unclear at this stage
whether this could be due to molecular dissociation,
early degradation or structural differences between
the intestinal epithelium and cultured monolayers.
Current research is now being focused on investigat-
ing and characterising the nature of the interactions
between the drug and its counter-ionic liposaccharide
derivative, as well as their stability at various pH
values.
Compound
Piperacillin sodium (i.v.)
Piperacillin sodium (oral)
Pip þ GC12 (1:1)
1058.46
4.96
3.83
1.45
3.03
100
0.46
0.38
0.14
0.29
0.17
Pip þ GC12 (1:2)
Pip þ GC12C12 (1:1)
Pip þ GC12C12 (1:2)
1.79
Results are given as mean and relative values against intrave-
nous piperacillin levels.
0.38% absorption, which statistically is not signifi-
cantly different from the 0.5% absorption found for
commercial piperacillin sodium (Fig. 7b).
The literature abounds in reports correlating
in vitro predictions of permeability obtained with
the Caco-2 model of intestinal epithelium and in vivo
oral absorption. The variations observed in these
experiments are therefore more likely to be inherent
to the nature of the compounds tested. It can be
hypothesised that piperacillin might have been
released from its counter-ionic liposaccharide deri-
vatives postadministration (e.g., as a result of pH
variations, enzymatic degradation, etc.). The discre-
pancy between in vitro and in vivo results could also
be due to the structural differences existing between
Caco-2 monolayers and intestinal membranes, parti-
cularly the presence of a protective mucus layer,
inexistent in cell culture monolayers.23 Furthermore,
intestinal membranes are often found to be more
resistant to cytotoxic effects of permeation enhancers
than cultured monolayers, so the increases in
permeability observed in vitro as a possible result
of membrane disruption might not be observed
in vivo.23,25
ACKNOWLEDGMENTS
This work was supported by an ARC project grant
(DP0558334). The authors wish to thank Prof.
Michael F. Jennings for his kind assistance in provid-
ing facilities and equipment to conduct the microbiol-
ogy experiments, and Mr. Michael Moore for his help
with handling animals.
CONCLUSION
We report here the biological evaluation of a novel
series of ionic liposaccharide derivatives designed as
potential oral penetration enhancers. The synthetic
constructs were engineered to combine in a single
molecular entity a carbohydrate-based hydrophilic
moiety coupled to lipophilic residues, such as hydro-
phobic amino acids and 12-carbon long LAAs.
Piperacillin, a potent b-lactam antibiotic with known
low oral bioavailability, was used as a model drug and
associated in various molar ratios with the liposac-
charide derivatives by lyophilisation. ITC measure-
ments were performed to determine the optimal
REFERENCES
1. Trapani A, Garcia-Fuentes M, Alonso MJ. 2008. Novel drug
nanocarriers combining hydrophilic cyclodextrins and chito-
san. Nanotechnology 19:185101–185110.
2. Cano-Cebrian MJ, Zornoza T, Granero L, Polache A. 2005.
Intestinal absorption enhancement via the paracellular route
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010

2342
BERGEON ET AL.
by fatty acids, chitosans and others: A target for drug delivery.
Curr Drug Deliv 2:9–22.
3. Shaji J, Patole V. 2008. Protein and peptide drug delivery: Oral
approaches. Indian J Pharm Sci 70: 269–277.
4. Godin B, Touitou E. 2006. Enhancers for enteral delivery of
macromolecules with emphasis on surfactants. In: Touitou E,
Barry BW, editors. Enhancement in drug delivery. Boca Raton,
FL (USA): CRC Press. pp. 37–56.
5. Pouton CW, Porter CJH. 2008. Formulation of lipid-based
delivery systems for oral administration: Materials, methods
and strategies. Adv Drug Deliv Rev 60:625–637.
6. Wong A, Toth I. 2001. Lipid, sugar and liposaccharide based
delivery systems. Curr Med Chem 8:1123–1136.
7. Bastin RJ, Bowker MJ, Slater BJ. 2000. Salt selection and
optimisation procedures for pharmaceutical new chemical enti-
ties. Org Process Res Dev 4:427–435.
8. Neubert R. 1989. Ion-pair transport across membranes. Pharm
Res 6:743–747.
9. Falconer RA, Toth I. 2007. Design, synthesis and biological
evaluation of novel lipoamino acid-based glycolipids for oral
drug delivery. Bioorg Med Chem 15:7012–7020.
10. Violette A, Cortes DAF, Bergeon JA, Falconer RA, Toth I. 2008.
Optimized LC-MS/MS quantification method for the detection
of piperacillin and application to the development of charged
liposaccharides as oral penetration enhancers. Int J Pharm
351: 152–157.
11. Abdelrahim AS, Ziora ZM, Bergeon JA, Moss AR, Toth I. 2009.
Synthesis and calorimetric titration of new glucolipopeptides
as potential cationic penetration enhancers. Tetrahedron
65:9436– 9442.
12. Kyte J, Doolittle RF. 1982. A simple method for displaying the
hydropathic character of a protein. J Mol Biol 157:105–132.
13. Andrews JM. 2001. Determination of minimum inhibitory
concentrations. J Antimicrob Chemoth 48:5–16.
15. Hoogkampkorstanje JAA, Westerdaal NAC. 1979. Activity and
synergy of piperacillin and aminoglycosides against Pseudo-
monas-Aeruginiosa. Antonie Van Leeuwenhoek 45:617–618.
16. Korvick JA, Yu VL. 1991. Antimicrobial agent therapy for
Pseudomonas aeruginosa. Antimicrob Agents Chemother 35:
2167–2172.
17. Pignatello R, Noce C, Campisi A, Acquaviva R, Bucolo C,
Puglisi G, Toth I. 2007. Evaluation of cell tolerability of a series
of lipoamino acids using biological membranes and a biomem-
brane model. Curr Drug Deliv 4:109–121.
18. Ross BP, Braddy AC, McGeary RP, Blanchfield JT, Prokai L,
Toth I. 2004. Micellar aggregation and membrane partitioning
of bile salts, fatty acids, sodium dodecyl sulfate, and sugar-
conjugated fatty acids: Correlation with hemolytic potency and
implications for drug delivery. Mol Pharm 1:233–245.
19. Amin K, Dannenfelser RM. 2006. In vitro hemolysis: Guidance
for the pharmaceutical scientist. J Pharm Sci 95:1173–1176.
20. Artursson P, Palm K, Luthman K. 2001. Caco-2 monolayers in
experimental and theoretical predictions of drug transport.
Adv Drug Deliv Rev 46:27–43.
21. BriskeAnderson MJ, Finley JW, Newman SM. 1997. The influ-
ence of culture time and passage number on the morphological
and physiological development of Caco-2 cells. Proc Soc Exp
Biol Med 214:248–257.
22. Donna AV. 2008. Variability in Caco-2 and MDCK cell-based
intestinal permeability assays. J Pharm Sci 97:712–725.
23. Aungst BJ. 2000. Intestinal permeation enhancers. J Pharm
Sci 89:429–442.
24. He X. 2009. Integration of physical, chemical, mechanical, and
biopharmaceutical properties in solid oral dosage form devel-
opment. In: Qiu Y, Chen Y, Zhang GGZ, Liu L, Porter W,
editors. Developing solid oral dosage forms—Pharmaceutical
theory and practice. New-York, NY (USA): Elseviewer Inc.
pp. 409–442.
14. Nolting A, DallaCosta T, Rand KH, Derendorf H. 1996. Phar-
macokinetic pharmacodynamic modeling of the antibiotic effect
of piperacillin in vitro. Pharm Res 13:91–96.
25. Rege BD, Yu LX, Husain AS, Polli JE. 2001. Effect of common
excipients on Caco-2 transport of low-permeability drugs.
J Pharm Sci 90:1776–1786.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 5, MAY 2010
DOI 10.1002/jps