Calcium depletion-mediated protease inhibition and apical-junctional- complex disassembly via an EGTA-conjugated carrier for oral insulin delivery

Article abstract of DOI:10.1016/j.jconrel.2012.11.011

Calcium (Ca²⁺) has a crucial role in maintaining the intestinal protease activity and in forming the apical junctional complex (AJC) that preserves epithelial barrier function. Ethylene glycol tetraacetic acid (EGTA) is a Ca²⁺-specif

Full text of DOI:10.1016/j.jconrel.2012.11.011

Journal of Controlled Release 169 (2013) 296–305
Contents lists available at SciVerse ScienceDirect
Journal of Controlled Release
journal homepage: www.elsevier.com/locate/jconrel
Calcium depletion-mediated protease inhibition and apical-junctional-complex
disassembly via an EGTA-conjugated carrier for oral insulin delivery
Er-Yuan Chuang a,1, Kun-Ju Lin b,c,1, Fang-Yi Su a,1, Hsin-Lung Chen a, Barnali Maiti a, Yi-Cheng Ho d,
Tzu-Chen Yen c, Nilendu Panda a, Hsing-Wen Sung a,
⁎
a
Department of Chemical Engineering/Institute of Biomedical Engineering, National Tsing Hua University, Hsinchu, Taiwan, ROC
Department of Medical Imaging and Radiological Sciences, Chang Gung University, Taoyuan, Taiwan, ROC
Department of Nuclear Medicine and Molecular Imaging Center, Chang Gung Memorial Hospital, Taoyuan, Taiwan, ROC
b
c
d
Department of Biotechnology, Vanung University, Chungli, Taoyuan, Taiwan, ROC
a r t i c l e i n f o
a b s t r a c t
Calcium (Ca2+) has a crucial role in maintaining the intestinal protease activity and in forming the apical
junctional complex (AJC) that preserves epithelial barrier function. Ethylene glycol tetraacetic acid (EGTA)
is a Ca2+-specific chelating agent. To maintain the concentration of this chelator in areas where enzyme in-
hibition and paracellular permeation enhancement are needed, this study synthesized a poly(γ-glutamic
acid)–EGTA conjugate (γPGA–EGTA) to form nanoparticles (NPs) with chitosan (CS) for oral insulin delivery.
The results of our molecular dynamic (MD) simulations indicate that Ca2+ ions could be specifically chelated
to the nitrogen atoms, ether oxygen atoms, and carboxylate oxygen atoms in [Ca(EGTA)]2− anions. By che-
lating Ca2+, γPGA–EGTA conferred a significant insulin protection effect against proteases in intestinal tracts
isolated from rats. Additionally, calcium depletion by γPGA–EGTA could stimulate the endocytosis of AJC
components in Caco-2 cell monolayers, which led to a reversible opening of AJCs and thus increased their
paracellular permeability. Single-photon emission computed tomography images performed in the
biodistribution study clearly show the 123I-insulin orally delivered by CS/γPGA–EGTA NPs in the heart,
aorta, renal cortex, renal pelvis and liver, which ultimately produced a significant and prolonged hypoglyce-
mic effect in diabetic rats. The above results confirm that this γPGA–EGTA conjugate is a promising candidate
for oral insulin delivery.
Article history:
Received 30 September 2012
Accepted 20 November 2012
Available online 27 November 2012
Keywords:
Oral protein delivery
Chelating agent
Calcium depletion
Enzyme inhibition
Apical junctional complex
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
indicated that this CS/γPGA–DTPA NP system effectively enhances in-
sulin absorption throughout the entire small intestine [2].
Ease of administration coupled with high patient compliance
makes the oral route the preferred mode of drug administration.
However, the drawbacks of orally administered hydrophilic macro-
molecules include enzymatic degradation in the gastrointestinal
(GI) tract and barrier to physical absorption in the epithelium [1].
To address these issues, we recently reported a pH-responsive nano-
particle (NP) system composed of chitosan (CS) and poly(γ-glutamic
acid) conjugated with diethylene triamine pentaacetic acid (γPGA–
DTPA) [2]. CS, a mucoadhesive polycation, enhances paracellular
permeability by transiently opening tight junctions (TJs) between
epithelial cells [3], whereas DTPA inhibits intestinal proteases and
disrupts epithelial TJs by nonspecific chelation of divalent metal
ions such as Ca2+, Mg2+ and Zn2+ [4]. Our experimental results
Studies showed that calcium (Ca2+) has crucial roles in chymo-
trypsin activity and in maintaining the thermodynamic stability of
trypsin [5,6]. Trypsin and chymotrypsin are the major proteolytic en-
zymes present in the intestinal lumen, where cleavage of insulin mol-
ecules occurs [7]. Extracellular Ca2+ is also required to form the apical
junctional complex (AJC) needed for epithelial barrier function.
Therefore, depletion of Ca2+ in the intestinal lumen is advantageous
for oral administrated insulin to prevent enzymatic degradation and
enhance penetration through intestinal epithelial cells.
EGTA, a Ca2+-specific chelator, has significantly greater Ca2+ affini-
ty compared to other chelating agents [8]. Due to its Ca2+-binding
specificity, EGTA is more favorable for oral insulin delivery than DTPA.
To take advantage of its Ca2+-binding specificity, an amine-group
containing EGTA derivative was first synthesized and then conjugated
on γPGA to form NPs with CS for oral insulin delivery. Structural
changes in the as-prepared NPs were examined in different pH environ-
ments by dynamic light scattering (DLS), transmission electron micros-
copy (TEM), and small angle X-ray scattering (SAXS). Effects of γPGA–
EGTA on enzymatic inhibition and on paracellular permeability were
⁎
Corresponding author at: Department of Chemical Engineering, National Tsing Hua
University, Hsinchu 30013, Taiwan, ROC. Tel.: +886 3 574 2504.
E-mail address: hwsung@che.nthu.edu.tw (H.-W. Sung).
The first three authors (E.Y. Chuang, K.J. Lin and F.Y. Su) contributed equally to this
1
work.
0168-3659/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jconrel.2012.11.011

E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
297
evaluated in vitro in isolated rat intestinal segments and in Caco-2 cell
monolayers, respectively. Furthermore, the mechanism of TJ opening
induced by test NPs was studied by visualizing the ultra-structural
changes in the epithelial AJC via TEM and confocal laser scanning mi-
croscopy (CLSM). The biodistribution of insulin orally delivered by the
developed NPs was investigated by single-photon emission computed
tomography (SPECT); their pharmacokinetic (PK) and pharmacody-
namic (PD) profiles were then analyzed in a diabetic rat model.
25 mL) was subsequently mixed with N,N-diisopropylethylamine
(iPr2NET, 2.2 g) and MsCl (1.9 g) and reacted for 24 h. The pure com-
pound in DMF (15 mL) was then mixed with K2CO3 (1.6 g), NaI
(1.8 g) and 4-nitrocatechol (0.8 g) and stirred for 24 h; the product
was dissolved in methanol and hydrogenated with 10% Pd/C. The
obtained EGTA derivative was purified by column chromatography
and then conjugated on γPGA (the γPGA–EGTA conjugate, Fig. 1a) as
described below.
The γPGA–EGTA conjugate was prepared by reacting the acid form
of γPGA (0.1 g) with an equimolar amine-group of the EGTA derivative
in the presence of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide
hydrochloride (EDC, 0.2 g) and N-hydroxysuccinimide (NHS, 0.1 g) in
dimethyl sulfoxide (DMSO) at room temperature for 48 h. The product
was purified by dialysis against deionized (DI) water for 3 days and
then lyophilized. The Boc-deprotection reaction was performed by
mixing the freeze-dried product with trifluoroacetic acid (TFA, 5 mL)
at room temperature for 48 h. Finally, the synthesized γPGA–EGTA con-
jugate was purified by dialysis against DI water for 3 days and lyophi-
lized before further use.
2. Materials and methods
2.1. Materials
The 85%-deacetylated CS (MW 60 kDa) was purchased from Koyo
Chemical Co. (Japan), whereas γPGA (MW 100 kDa) was acquired
from Vedan Co. (Taichung, Taiwan). Other chemicals and reagents used
were of analytical grade and obtained from Sigma-Aldrich (St. Louis,
MO, USA).
2.2. Synthesis and characterization of γPGA–EGTA
The synthesized γPGA–EGTA conjugate was characterized by 1H
NMR (Varian Unityionva 500 NMR Spectrometer, MO, USA) and
FT-IR (Perkin-Elmer Spectrum RX 1 FT-IR System, Buckinghamshire,
UK). The degree of substitution (DS) of EGTA on γPGA was estimated
by 1H NMR spectroscopy. The Ca2+ binding efficiency of γPGA–EGTA
was determined by complexometric titration [10].
An amine-group containing EGTA derivative was first prepared
by incorporating an amino-functionalized aromatic moiety on the
oxoethylenic bridge of EGTA [9]. Briefly, ethanolamine (0.9 g)
was reacted with KHCO3 (3.8 g) and tert-butyl bromoacetate
(BrCH2COOtBu, 7.4 g) in dimethylformamide (DMF, 25 mL) at room
temperature for 24 h. The crude product in ethyl acetate (CH3CO2Et,
The conformation of γPGA–EGTA in the presence of Ca2+ and
Mg2+ was examined by molecular dynamic (MD) simulations [11].
a
b
γPGA
3293 (-NH2)
1734 (C=O stretch of acid)
1570 (amide I)
1594 (amide II)
EGTA derivative
Ethanolamine
1733 (C=O sretch of acid)
3014 (-NH2)
1406 (C-H stretch of aromatic)
1516 (C-H stretch of aromatic)
γPGA-EGTA
1404
(C-H stretch of aromatic)
3419 (-NH2)
1634 (C=O stretch of amide)
1515 (C-H stretch of aromatic)
4000
3500
*
3000
2500
2000
1500
1000
500
Wavenumber (cm-1
)
O
1
H
N
2
c
1
3
*
3
n
γPGA
2
O
OH
g
OH
NH2
a
a
e
f
O
b
d
N
a
DMSO
O
c
b
O
O
O
OH b
c
f
e
HO
O
a
c
g
N
d
EGTA derivative
a
OH
O
H
2
3
N
*
OH
*
n
1
g
a
O
NH
a
f
e
d
O
b
N
a
O
c
b
O
O
N
OH
HO
O
γPGA-EGTA
c
a
c
b
1
O
a
d e f
7
3
2
OH
8
6
5
4
3
2
1
0
PPM
Fig. 1. Synthesis and characterization of the γPGA–EGTA conjugate: (a) the reaction scheme showing the formation of an amino-functionalized EGTA derivative and its subsequent
conjugation on the γPGA backbone; (b) the FT-IR and (c) 1H NMR spectra of the free-form γPGA, EGTA derivative and synthesized γPGA–EGTA conjugate. γPGA: poly(γ-glutamic
acid) and EGTA: ethylene glycol tetraacetic acid.

298
E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
a
(n = 6)
b
Ca2+
γ PGA-EGTA
Mg2+
Insulin
140
120
Insulin + EGTA
Insulin +γPGA-EGTA
100
Mg2+
Ca2+
80
Chelated Ca2+
60
40
20
0
0
1
2
3
Time (h)
c
(n = 6)
Removal of Samples
120
γ PGA
Ca2+
100
80
60
40
20
0
Mg2+
control
EGTA at pH 7.4
C
O
H
N
γ
PGA-EGTA at pH 7.4
0
4
8
12
16
20
24
Time (h)
Fig. 2. (a) Results of molecular dynamic simulations showing the conformation of γPGA–EGTA in the presence of Ca2+ and Mg2+; (b) effects of EGTA and γPGA–EGTA on insulin
protection against intestinal proteases and (c) transepithelial electrical resistance (TEER) in Caco-2 cell monolayers. Control: the group without any treatment.
a
c
b
pH 6.6
pH 7.4
CS
γ PGA-EGTA
Insulin Molecule
5.7 nm
100
10
1
-4
300 nm
300 nm
NPs at pH 6.6
NPs at pH 7.4
120
100
80
60
40
20
0
Removal of NPs
qm
0.1
γ
γ
γ
CS/ PGA-EGTA NPs at pH 6.6
CS/ PGA-EGTA NPs at pH 7.0
CS/ PGA-EGTA NPs at pH 7.4
0.1
1
0
4
8
12
16
20
24
q (nm-1)
Time (h)
Fig. 3. (a) TEM micrographs and (b) small angle X-ray scattering (SAXS) profiles showing the pH-responsive characteristics of insulin-loaded CS/γPGA–EGTA nanoparticles (NPs);
(c) effects of these NPs on the TEER of Caco-2 cell monolayers at distinct pH environments. CS: chitosan.

E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
299
In body fluids, the predominant divalent metal ions are Ca2+ and
Mg2+ [12], and Ca2+ and Mg2+ ion levels are higher in gastrointesti-
nal (GI) mucosae than in blood [13,14]. Atomistic pictures of mole-
cules were generated using UCSF Chimera.
−20 °C. Non-specific binding was blocked in 5% normal goat serum
(Jackson ImmunoResearch Laboratories, West Grove, PA, USA) in PBS
for 60 min at 37 °C. The cells were then incubated overnight in primary
antibodies [mouse anti-claudin-4 (Invitrogen) and donkey anti-E-
cadherin (Abcam, Cambridge, MA, USA)] at 4 °C. The results were visual-
ized by applying appropriate Alexa-Fluor-conjugated secondary anti-
bodies (Invitrogen); the cells were also counterstained with SYTOX blue
(Invitrogen) for examination by inverted CLSM.
Extracellular calcium levels were measured by staining the treated
cells with 40 mM Alizarin Red S (ARS), pH 4.0, for 10 min, and then
washing with 70% ethanol and then with PBS to remove non-specific
staining [19]. Samples were then visualized by an inverted phase-
contrast light microscope (Nikon Eclipse-Ti, Nikon Instruments Inc.,
Melville, NY, USA).
2.3. Inhibition of intestinal proteolytic degradation
In vitro studies of intestinal protease inhibition by γPGA–EGTA
were performed using proximal sections of small intestine freshly iso-
lated from rats (Wistar, 180–200 g) [2]. The intestinal tissues were
incubated with a Krebs–Ringer buffer (KRB) solution containing insu-
lin (1 mg/mL, 1 mL) and γPGA–EGTA (5 mg/mL) at 37 °C; free-form
EGTA (2 mg/mL) at 37 °C, which had a similar Ca2+ binding capacity,
was used as a control. To quantify the amount of intact insulin
remaining, test solutions (50 μL) withdrawn at different time inter-
vals were immediately analyzed by high-performance liquid chroma-
tography (HPLC).
2.8. Animal study
Animal studies were performed according to the “Guide for the
Care and Use of Laboratory Animals” prepared by the Institute of
Laboratory Animal Resources, National Research Council and pub-
lished by the National Academy Press, revised in 1996.
2.4. Permeability of epithelial AJCs
The permeability of the epithelial AJCs in Caco-2 cell monolayers
after treatment with test samples was measured in terms of
transepithelial electrical resistance (TEER). The cell monolayers were
incubated individually with EGTA (4 mg/mL, 0.5 mL) and γPGA–EGTA
(10 mg/mL, 0.5 mL), which had equivalent Ca2+ binding capacities.
The change of TEER for the tightness of cell monolayers was measured
with a Millicell®-Electrical Resistance System (Millipore Corp., Bedford,
MA, USA).
The biodistribution of insulin orally delivered by CS/γPGA-EGTA
NPs was studied in rats (n=3) by using SPECT. In the study, 123iodine
was used to radiolabel insulin (123I-insulin) using iodogen precoated
tubes (Thermo Fisher Scientific, Rockford, IL, USA). Following the
last SPECT scan, additional whole-body CT images of the heart,
lungs, great vessels, liver and kidneys were acquired after intravenous
administration of the Conray™ contrast agent (Mallinckrodt, MO,
USA). To maximize the amount of contrast fluid stasis in the vascula-
ture, the contrast agent was slowly injected via the tail vein (2 mL
over 40 s). After receiving half of the Conray™ dose, each animal
was sacrificed by co-injection of potassium chloride under deep anes-
thesia. The detailed protocols used in the image scanning and its
quantification were previously described by our group [20,21]. To
calculate the percentage of initial dose (% ID) within each area, the
corresponding volumes of interest (VOIs) were manually drawn on
the co-registered reference CT and dynamic scintigraphic images.
Biodistribution data were expressed as % ID using the following formula:
2.5. Preparation and characterization of CS/γPGA–EGTA NPs
A simple ionic-gelation method was used to prepare the insulin-
loaded CS/γPGA–EGTA NPs [2]. The particle size and zeta potential of
the prepared NPs were measured by DLS (3000 HS, Malvern Instru-
ments Ltd., Worcestershire, UK), and morphology study was performed
by TEM (JEOL 2010F, Tokyo, Japan). The loading efficiency (LE) and
loading content (LC) of insulin in NPs were calculated using a method
reported previously [15]. The internal structures of CS/γPGA–EGTA
NPs were examined by SAXS [16] after resuspending the test NPs
in phosphate buffered saline (PBS) of different pH values. Addition-
ally, TEER measurements of Caco-2 cell monolayers incubated with
CS/γPGA–EGTA NPs (0.02% w/v, 0.5 mL) were performed at pH values
of 6.6, 7.0, or 7.4 (to simulate the pH environments in different portions
of the small intestine) in the donor compartment and at pH 7.4 in the
receiver compartment.
decaycorrectedradioactivitywithintheVOIateachtimeframe ꢀ 100%
% ID ¼
:
wholebodyradioactivityatthe1stimageframe
Diabetes was induced in rats (Wistar, 250 g) via intraperitoneal
injection of streptozotocin (STZ) as described previously [22]. The
rats with induced diabetes were then fasted overnight and for the
remainder of the experiment but with free access to water. The fol-
lowing formulations were administered individually to the diabetic
rats: an enteric-coated capsule containing free-form insulin powder
(30 IU/kg) and trehalose (oral); an enteric-coated capsule filled
with CS/γPGA-EGTA NPs containing 30 IU/kg insulin (oral); and a
subcutaneous (SC) injection of free-form insulin solution (5 IU/kg,
n=6 for each studied group). Blood sampling from the tail veins of
rats was performed before administering the drug and at varying
time intervals after dosing. The blood glucose levels were then mea-
sured with a glucose meter (LifeScan, Milpitas, CA, USA). Plasma
2.6. Ultra-structural changes in AJCs
After treating test NPs with EGTA conjugate (CS/γPGA–EGTA NPs,
0.2 mg/mL) or without EGTA conjugate (CS/γPGA NPs), ultra-structural
changes in epithelial AJCs in Caco-2 cell monolayers were examined by
TEM; paracellular permeability was then depicted by lanthanum nitrate
staining. Lanthanum is an electron-dense rare-earth element widely
used as a tracer for studying junctional barrier permeability [17]. The
cell monolayers were fixed in 3.7% paraformaldehyde, washed with
s-Collidine buffer, and incubated with 2% lanthanum for 2 h at room
temperature. After rinsing in s-Collidine and PBS, samples were
processed for TEM examination as detailed previously [18].
Table 1
Mean particle sizes (nm), polydisperse index, and zeta potential values (mV) of
CS/γPGA–EGTA nanoparticles (NPs) at distinct pH environments (n=5 batches). N/A:
data were not available due to the observed disintegration of NPs.
2.7. Translocation of AJC proteins
pH value
Mean particle size
Polydisperse index
Zeta potential
The Caco-2 cells were grown to confluence in 35-mm dishes with
glass cover-slip bottoms and then treated with test NPs for 2 h; translo-
cation of their AJC proteins was investigated by immunofluorescence
staining. The treated cells were washed 3 times with pre-warmed
PBS and then fixed and permeabilized in 100% acetone for 10 min at
6.0
6.6
7.0
7.4
328.6 2.3
354.3 3.1
1549.3 33.9
N/A
0.23 0.02
0.25 0.05
0.39 0.08
0.48 0.13
38.7 0.2
12.1 0.4
7.1 0.6
1.3 0.9

300
E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
Control
CS/γ PGA NPs
CS/γ PGA-EGTA NPs
AJC
La3+
200 nm
La3+
La3+
AJC
AJC
La3+
Fig. 4. TEM micrographs showing the effects of CS/γPGA NPs or CS/γPGA–EGTA NPs on the AJC opening and its subsequent paracellular permeability in the cell monolayers at pH 6.6
and 7.4. AJC: apical junctional complex; blue arrow: indicating an intact AJC, and red arrow: indicating the lanthanum (La3+) permeability through the opened intercellular spaces.
insulin levels and the relative bioavailability (BAR) of insulin after oral
administration were calculated using the following formula [16].
insulin absorption by the small intestine, thus enhancing the bioavail-
ability of insulin.
3.1. Synthesizing and characterizing γPGA–EGTA
AUCðoralÞ ꢀ DoseðSCÞ
BAR
¼
ꢀ 100%
AUCðSCÞ ꢀ DoseðoralÞ
Fig. 1b shows the FT-IR spectra of free-form γPGA, the synthesized
EGTA derivative, and γPGA–EGTA. The γPGA spectra exhibited absor-
bance peaks at 1734 cm−1 for carboxyl groups, 1594 cm−1 for amide
II, and 1570 cm−1 for amide I, while those of the EGTA derivative
where AUC is the total area under the plasma insulin concentration
versus time curve.
showed absorbance peaks characteristic of –COOH (1733 cm−1
)
2.9. Statistical analysis
and aromatic C–H stretching (1516 and 1406 cm−1). For γPGA–
EGTA, the presence of the peaks for amide bond (1634 cm−1) and
aromatic C–H stretching (1515 and 1404 cm−1) confirms the suc-
cessful conjugation of EGTA derivative on γPGA.
The two groups were compared by using statistical software
(SPSS, Chicago, Ill) to perform one-tailed Student t-test. All data
were expressed as means with standard deviations indicated
(mean SD). A difference of Pb0.05 was considered statistically
significant.
In further analysis of the synthesized γPGA–EGTA by 1H NMR
(Fig. 1c), the spectrum of the unmodified γPGA displayed chemical
shifts characteristic of α-CH (δ=4.01 ppm), β-CH2 (δ=1.96 and
1.83 ppm) and γ-CH2 (δ=2.22 ppm) while that of the EGTA deriva-
tive showed shifts indicating aromatic phenyl (7.87, 7.74 and
7.16 ppm) and methylene and primary amine protons (4.21, 3.16,
3.63 and 3.38 ppm). In the γPGA–EGTA conjugate, the characteristic
peaks for aromatic protons (7.03 and 6.93 ppm) and methylene pro-
tons (4.31, 3.68 and 3.85 ppm) were for EGTA derivative, and those
for α-CH (δ=4.02 ppm), β-CH2 (δ=1.95 ppm and 1.83 ppm), and
γ-CH2 (δ=2.25 ppm) were from γPGA. These 1H NMR results con-
firm the successful conjugation of the EGTA derivative on the γPGA
backbone. According to the 1H NMR spectrum of γPGA–EGTA, the
DS of EGTA derivative on γPGA was an estimated 50%, which indicat-
ed the presence of 0.5 mol of EGTA for each mole of γPGA monomer.
The complexometric titration results showed that the Ca2+-binding
capacity of γPGA–EGTA at neutral pH (simulating the intestinal pH en-
vironment) was 1666.7 8.3 μmol/g conjugate (n=6), which was
equivalent to an EGTA-to-Ca2+ molar ratio of 1:1. The chelation of
Ca2+ by γPGA–EGTA results from proton release from the EGTA struc-
ture at neutral pH [26,28].
3. Results and discussion
Calcium depletion by chelating agents such as EGTA is known to in-
crease paracellular permeability [23] and has a crucial role in inhibiting
intestinal enzyme activities [24]. The EGTA, a hexadentate ligand, is best
known for its selectivity for Ca2+ over Mg2+ (ca. l06-fold) [25]. In intes-
tinal pH environments (pH 6.4–7.4), acidic groups of EGTA are
deprotonated, which accelerates the association kinetics of Ca2+ [26].
To concentrate the chelating agent in areas where paracellular
permeation enhancement and enzyme inhibition are needed, immo-
bilization of this auxiliary agent on the drug-carrier matrix is cru-
cial [27]. Our approach is to synthesize a γPGA–EGTA conjugate
(Fig. 1a) to form NPs with CS for oral insulin delivery. Such a delivery
system can concentrate the activities of permeation enhancement
and enzyme inhibition in areas where drug release occurs in the in-
testinal mucus layer. This study shows that this CS/γPGA–EGTA NP
system inhibits intestinal proteolytic degradation and promotes

E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
301
Control
CS/γPGA NPs
CS/γPGA-EGTA NPs
40 µm
10 µm
10 µm
Fig. 5. Photomicrographs showing the extracellular calcium levels (as indicated by the red ARS-labeled Ca2+) and the immunofluorescence staining of the cell monolayers before
and after receiving CS/γPGA NPs or CS/γPGA–EGTA NPs at pH 7.4. ARS: Alizarin Red S.
The EGTA reportedly has a high binding affinity for Ca2+ (stability
constant log K=11.0) and a relatively low affinity for Mg2+ (log K=
5.2) [29]. The confirmation of the γPGA–EGTA chelated with Ca2+ ions
alone, obtained by MD simulations, is shown in Fig. 2a. The Ca2+ in
[Ca(EGTA)]2− anions binds to the nitrogen atoms (Ca–N), ether oxygen
atoms (Ca–O), and carboxylate oxygen atoms (Ca–O) at average dis-
tances of 2.66 Å, 2.49 Å, and 2.52 Å, respectively. The Ca2+/Mg2+ dis-
crimination of EGTA presumably stems from its binding cavity, which
is the right size for Ca2+ but does not further envelop Mg2+ tightly be-
cause the carboxylate anions at each end of the chain begin butting
into each other [30]. The DTPA has similar stability constants for Ca2+
(log K=10.7) and Mg2+ (log K=9.3) [29] and therefore can chelate
both Ca2+ and Mg2+ ions [4].
freshly isolated from rats and placed in KRB, a simulated intestinal
fluid [34].
Fig. 2b shows that the control group of free-form insulin was rap-
idly degraded by the intestinal proteolytic enzymes (approximately
90% insulin degradation within 3 h). Adding EGTA, or γPGA–EGTA
conferred a significant protective effect of insulin against proteases
(Pb0.05). EGTA and γPGA–EGTA conferred much larger insulin pro-
tection effects compared to their DTPA counterparts reported in our
previous study (8% vs. 35% insulin degradation, respectively Pb0.05)
[2]. These experimental results confirm that EGTA and γPGA–EGTA
are superior to their DTPA counterparts in terms of enzyme inhibition
as they have specific capability to deprive Ca2+ from the structure of
enzymes, which ultimately helps prevent insulin degradation.
3.2. Inhibition of intestinal enzyme degradation by γPGA–EGTA
3.3. Enhancement of the paracellular permeability
Insulin is very sensitive to trypsin and chymotrypsin, which are
Ca2+-dependent enzymes present in the intestinal fluid and mucus
layer [31–33]. One strategy for inhibiting intestinal enzyme activities
is removing essential metal ions such as Ca2+ from the enzyme struc-
ture [24]; the affinity for Ca2+ correlates with the strength of the
enzyme inhibition. The inhibitory activity of γPGA–EGTA against in-
testinal proteases was investigated in proximal intestinal tracts
Calcium is critical for maintaining cell–cell junctions in various cell
types [35]. The use of complexing agents for chelating extracellular
Ca2+ can disrupt intercellular junctions and can increase paracellular
permeability. The well-established in vitro Caco-2 cell model was
used to evaluate absorption enhancement across the intestinal epi-
thelium by assessing their TEER [36]. Fig. 2c shows that Caco-2 cell
monolayers treated with free-form EGTA or γPGA–EGTA produced a

302
E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
a
b
0.5 h
E
2 h
4 h
6 h
24 h
CT Images
L
0.5 h
2 h
4 h
6 h
24 h
E
E
8
L
St
St
St
K
St
Rc
Rc
4
0
Int
Rv
A
Rv
A
B
B
Int
Lv
B
E: esophagus; L: lung; St: stomach; Int: intestine; K: kidney ; B: bladder
9
8
7
6
5
4
3
2
1
0
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
c
H
H
(n = 3)
Lv
Stomach
Intestine
Liver
Heart
Kidney
St
St
B
L: lung; St: stomach; Rc: renal cortex; Rv: renal pelvis;
A: aorta; H: heart; Int: intestine; Lv: liver
0
4
8
12
16
20
24
Time (h)
Fig. 6. (a) Soft-tissue contrast CT images (in gray) superimposed with 3D volume-rendering SPECT images (in red) showing the distribution of 123I-insulin in various body com-
partments obtained at different time points; (b) biodistribution of 123I-insulin illustrated by contrast-enhanced tomographic images. The concentrations of 123I-insulin in different
organs are shown in rainbow pseudo-color scale and (c) their percentages of ingested dose per volume (% ID/cm3) vs. time profiles of 123I-insulin in the stomach (left y-axis) and
other organs (right y-axis) are also demonstrated.
significantly larger reduction in TEER compared to the control cells
(Pb0.05). The TEER reduction induced by EGTA and γPGA–EGTA
was larger than that induced by their DTPA counterparts [2]. Again,
the larger reduction may have resulted from their binding specificity
towards Ca2+. After removing the test samples and replenishing the
medium, TEER increased gradually and recovered almost completely
within 24 h, indicating that the transient calcium depletion may
allow a reversible modulation of paracellular permeability.
γPGA–EGTA. Since most of the insulin molecules were not contained
in the NPs, their spatial correlation peak was no longer observable in
the SAXS profile.
Fig. 3c shows the TEER variations observed in the Caco-2 cell
monolayers after receiving CS/γPGA–EGTA NPs at pH 6.6, 7.0 or 7.4.
Regardless of pH exposure, all groups showed remarkably decreased
TEER, indicating that the EGTA-conjugated NPs developed in the
study may facilitate paracellular permeation in the pH environments
from the proximal duodenum to the distal ileum (pH 6.4–7.4) along
the intestinal tract. Conversely, the CS is well known to be inadequate
for opening intercellular TJs in neutral pH environments [37], which
limits its potential use as a permeation enhancer only in the duode-
num section.
3.4. Characteristics of CS/γPGA–EGTA NPs
The CS/γPGA–EGTA NPs prepared in DI water (pH 6.0) had a par-
ticle size of 328.6 2.3 nm and a zeta potential of 38.7 0.2 mV;
their insulin LE and LC were 78.7 0.4% and 17.5 1.5%, respectively
(n=5 batches). The pH-sensitivity of test NPs was evaluated at dis-
tinct pH environments, representing the small intestine conditions.
As pH was increased, the size of the test NPs significantly increased
until they eventually disintegrated at pH 7.4 (Fig. 3a); however,
their zeta potential values significantly decreased (Pb0.05, Table 1)
due to CS deprotonation.
Fig. 3b shows the SAXS profiles in a log–log plot of CS/γPGA–EGTA
NPs dispersed in different pH environments. Close study of the
SAXS profiles of test NPs at pH 6.6 showed peak centering at about
1.1 nm−1. Earlier, the authors showed that this peak is attributable
to the characteristic correlation between the insulin molecules con-
tained within the individual NP [18], where the relatively high con-
centration of insulin molecules in test NPs produces an interaction
peak. The characteristic correlation distance between the insulin mol-
ecules calculated from the peak position via d=2π/qm approximated
5.7 nm. No scattering peak occurred when the pH of the medium was
increased to 7.4. At this pH value, the disintegrating test NPs released
their originally embedded insulin molecules to the bulk media be-
cause CS deprotonation weakened their electrostatic attraction to
3.5. Ultra-structural study of AJC opening and paracellular permeability
Ultra-structural study of the AJC opening and its subsequent
paracellular permeability in the cell monolayers was performed by
TEM after treatment with CS/γPGA NPs or CS/γPGA–EGTA NPs at pH
6.6 and 7.4. Lanthanum (La3+) was used as the electron microscopic
stain because its interaction with negatively charged cell-surface glyco-
proteins makes it effective for delineating intercellular spaces [17].
Fig. 4 shows that before treatment with test NPs (controls), the
intact AJCs (as pointed by the blue arrow) appeared as highly
electron-dense structures underneath the apical cell membrane and
between adjacent cells. When treated with the test NPs without
EGTA conjugate (CS/γPGA NPs) at pH 6.6, lanthanum staining of the
paracellular spaces between adjacent cells was clearly observed (red
arrow), suggesting the opening of intercellular junctions mediated
by CS [18]. At pH 7.4, the AJCs remained intact, and limitation of
the lanthanum staining to the brush-border surface suggested that
CS/γPGA NPs do not effectively enhance permeation at neutral pH
values. Whereas CS (pKa 6.5) facilitates paracellular transport only

E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
303
(n = 6)
Free-form Insulin Powder (Oral, 30 IU/kg)
a
140
120
100
80
γ
CS/ PGA-EGTA NPs (Oral, 30 IU/kg)
Free-form Insulin Solution (SC, 5 IU/kg)
60
40
20
0
2
4
6
8
10
12
Time (h)
(n = 6)
b
Free-form Insulin Powder (Oral, 30 IU/kg)
160
140
120
100
80
γ
CS/ PGA-EGTA NPs (Oral, 30 IU/kg)
Free-form Insulin Solution (SC, 5 IU/kg)
60
40
20
0
0
2
4
6
8
10
12
Time (h)
Fig. 7. (a) Plasma insulin level vs. time profiles and (b) blood glucose change vs. time profiles of diabetic rats treated with different formulations of insulin. Oral: oral administration
and SC: subcutaneous injection.
after protonation [3], treatment with the EGTA-conjugated (CS/
γPGA–EGTA) NPs caused the opening of AJCs and significant penetra-
tion of lanthanum into the lateral intercellular spaces at both test pH
environments, possibly due to the deprivation of extracellular Ca2+
by EGTA. Calcium depletion by EGTA reportedly increases paracellular
permeability [29].
(CS/γPGA NPs) did not significantly affect either extracellular calcium
levels or translocation of E-cadherin and CLDN4.
In contrast, treatment with EGTA-conjugated NPs (CS/γPGA–EGTA
NPs) substantially diminished extracellular calcium levels; additionally,
internalization of both E-cadherin (red color) and CLDN4 (green color)
to a cytosolic compartment surrounding the nucleus suggested that
calcium depletion by EGTA stimulates endocytosis of AJC components.
Internalization of AJs and TJs via a clathrin-dependent pathway appears
to be a common mechanism to rapidly modulate the cell–cell sealing,
which has been documented in the literature [40].
3.6. AJC disassembly induced by calcium depletion
The AJC, which is comprised of adherens junctions (AJs) and TJs, is
an important regulator of cell structure and function [38]. AJs bring
two plasma membranes together physically and enhance TJ assembly;
thus, alteration of AJs modulates the TJ structure and the epithelial
paracellular barrier function. E-cadherin, a Ca2+-dependent adhesion
molecule, is the major trans-membrane protein of AJs, and claudin-4
(CLDN4) is a TJ protein [3,39]. Neither AJs nor TJs have a static structure;
they can be rapidly disassembled and reorganized in response to vari-
ous extracellular stimuli.
Fig. 5 shows the extracellular calcium levels (as indicated by the
red ARS-labeled Ca2+) and the immunofluorescence staining of the
cell monolayers before and after receiving CS/γPGA NPs or CS/
γPGA–EGTA NPs at pH 7.4. Extracellular Ca2+ is essential for AJC
formation and for maintaining cell–cell junctions [40]. Compared to
the control group, treatment with test NPs without EGTA conjugate
3.7. Biodistribution of insulin
Fig. 6a presents the radioactivity distribution of 123I-insulin orally de-
livered by CS/γPGA–EGTA NPs over the 24-h period after ingestion in a rat
model. The 3D volume-rendering SPECT images of 123I-insulin (in red)
were superimposed onto the soft-tissue contrast-volume-rendering CT
images (in gray). Air in the hollow organs, including the bilateral lungs,
stomach and intestine, was clearly visible in the volume-rendering CT
images, whereas 123I-insulin was identified in the GI track, kidneys and
bladder via the SPECT images.
Additional details of the biodistribution of 123I-insulin are visible
in the coronal tomographic images (Fig. 6b), in which the radioactiv-
ity concentration of 123I-insulin is presented in rainbow pseudo-color
scale and superimposed on the CT images to distinguish anatomic

304
E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
Table 2
containing free-form insulin powder produced negligible insulin ab-
sorption and hypoglycemic effects. Conversely, administration of cap-
sules containing insulin-loaded CS/γPGA–EGTA NPs produced a slow
but prolonged reduction in blood glucose levels, indicating that the
pharmacological activity of the absorbed insulin was still intact; their
AUC(0–12 h) was 244.5 33.7 μIU h/mL, and the corresponding BAR
was 21.3 1.5% (Table 2).
Pharmacokinetic parameters of insulin in diabetic rats after subcutaneous (SC) admin-
istration of free-form insulin solution, oral administration of capsules containing
free-form insulin powder or insulin-loaded CS/γPGA–EGTA NPs. Cmax: maximum plas-
ma concentration; Tmax: time at which Cmax is attained; AUC(0–12 h): area under the
plasma concentration–time curve; BAR: relative bioavailability (n=6).
SC free-form
Oral free-form
Oral
insulin solution insulin powder CS/γPGA–EGTA NPs
Dose (IU/kg)
Cmax (μIU/mL)
Tmax (h)
5.0
30.0
1.8 0.7
5.0
30.0
44.7 13.0
5.0
119.63 3.8
1.0
3.9. Functionality of γPGA–EGTA
AUC(0–12 h) (μIU h/mL) 191.31 30.7
BAR (%) 100
6.5 2.7
0.5 0.3
244.5 33.7
21.3 1.5
Fig. 8 schematically depicts the functionality of γPGA–EGTA in fa-
cilitating oral insulin delivery according to the above results. By spe-
cific deprivation of extracellular Ca2+ from the intestinal lumen
(Figs. 2a and 5), γPGA–EGTA can significantly reduce the enzymatic
degradation of insulin (Fig. 2b). Additionally, via causing the endocy-
tosis AJCs (Fig. 5), γPGA–EGTA displays the ability to modulate the
transepithelial resistance (Fig. 2c), enhance the paracellular perme-
ation (Fig. 4), and promote the insulin absorption (Fig. 6a–c), thus
increasing its oral bioavailability (Fig. 7a and b).
localizations. The vascular structures, including the heart, aorta, renal
cortex, renal pelvis and liver were visible in the contrast-enhanced CT
images, while 123I-insulin was clearly recognizable in the above-
mentioned organs in the SPECT images. Fig. 6c shows the radioactiv-
ity concentrations of 123I-insulin (% ID/cm3) observed in different
organs. After ingestion, the radioactivity concentration of 123I-insulin
present in the stomach diminished gradually, while those in the
intestine, liver, heart and kidneys increased with time, reaching
their maximum values at 3–4 h following ingestion, revealing the ab-
sorption of insulin into the systemic circulation.
4. Conclusions
A calcium-specific depletion agent, γPGA–EGTA, was successfully
synthesized to formulate a carrier for oral insulin delivery. Our results
indicate that γPGA–EGTA plays a significant role in the prevention of
enzymatic degradation of insulin and in the enhancement of its
paracellular permeability through the modulation of epithelial AJC
disassembly, ultimately producing a significant and prolonged hypo-
glycemic effect. These results confirm the efficacy of the proposed
EGTA-conjugated carrier in promoting the bioavailability of orally ad-
ministered insulin.
3.8. In vivo PK and PD profiles
The SC injection of free-form insulin solution resulted in a maximum
plasma concentration at 1 h post administration. Blood glucose levels
then sharply decreased within 1–3 h before returning to their basal
levels (Fig. 7a and b). Subsequent oral administration of capsules
Before γPGA-EGTA
After γPGA-EGTA Treatment
Treatment
Ca2+-depletion enzyme inhibition
Intestinal enzymes with
Mucus layer
cofactor Ca2+
Ca2+-depletion-mediated AJC
endocytosis
TJ
AJ
CLDN4
E-cadherin
AJC redistribution
Intact AJC
AJC
Disassembly
Nucleus
Cytoplasm
Ca2+
Activated Ca2+-
dependent enzymes
Unactivated Ca2+-
dependent enzymes
γPGA-EGTA
E-cadherin CLDN4
Fig. 8. Schematic illustrations displaying the specific deprivation of extracellular Ca2+ from the intestinal lumen by γPGA–EGTA and its subsequent effects on the inhibition of in-
testinal proteases and the enhancement of paracellular permeation.

E.-Y. Chuang et al. / Journal of Controlled Release 169 (2013) 296–305
305
[20] K. Sonaje, K.J. Lin, S.P. Wey, C.K. Lin, T.H. Yeh, H.N. Nguyen, C.W. Hsu, T.C. Yen, J.H.
Juang, H.W. Sung, Biodistribution, pharmacodynamics and pharmacokinetics of
insulin analogues in a rat model: oral delivery using pH-Responsive nanoparticles
vs. subcutaneous injection, Biomaterials 31 (2010) 6849–6858.
[21] H.N. Nguyen, S.P. Wey, J.H. Juang, K. Sonaje, Y.C. Ho, E.Y. Chuang, C.W. Hsu, T.C.
Yen, K.J. Lin, H.W. Sung, The glucose-lowering potential of exendin-4 orally deliv-
ered via a pH-sensitive nanoparticle vehicle and effects on subsequent insulin se-
cretion in vivo, Biomaterials 32 (2011) 2673–2682.
Acknowledgments
This work was supported by a grant from the National Science
Council (NSC 100-2120-M-007-003), Taiwan, Republic of China. The
molecular-imaging study was partially supported by a grant from
Chang Gung Memorial Hospital (CMRPG391512).
[22] S. Sajeesh, K. Bouchemal, V. Marsaud, C. Vauthier, C.P. Sharma, Cyclodextrin
complexed insulin encapsulated hydrogel microparticles: an oral delivery system
for insulin, J. Control. Release 147 (2010) 377–384.
[23] P.H. Johnson, D. Frank, H.R. Costantino, Discovery of tight junction modulators:
significance for drug development and delivery, Drug Discov. Today 13 (2008)
261–267.
[24] A. Bernkop-Schnurch, The use of inhibitory agents to overcome the enzymatic
barrier to perorally administered therapeutic peptides and proteins, J. Control.
Release 52 (1998) 1–16.
[25] C.K. Schauer, O.P. Anderson, Calcium-selective ligands. 2. Structural and spec-
troscopic studies on calcium and cadmium complexes of EGTA4-[H4EGTA=
3,12-bis(carboxymethyl)-6,9-dioxa-3,12-diazatetradecanedioic acid], J. Am.
Chem. Soc. 109 (1987) 3646–3656.
[26] M. Naraghi, T-jump study of calcium binding kinetics of calcium chelators, Cell
Calcium 22 (1997) 255–268.
[27] A. Bernkop-Schnurch, Chitosan and its derivatives: potential excipients for per-
oral peptide delivery systems, Int. J. Pharm. 194 (2000) 1–13.
[28] E. Permyakov, K. Murakami, L. Berliner, On experimental artifacts in the use of
metal ion chelators for the determination of the cation binding constants of
alpha-lactalbumin. A reply, J. Biol. Chem. 262 (1987) 3196–3198.
[29] T. Matsukawa, S. Ikeda, H. Imai, M. Yamada, Alleviation of the two-cell block of
ICR mouse embryos by polyaminocarboxylate metal chelators, Reproduction
124 (2002) 65–71.
[30] R.Y. Tsien, New calcium indicators and buffers with high selectivity against mag-
nesium and protons: design, synthesis, and properties of prototype structures,
Biochemistry 19 (1980) 2396–2404.
[31] T. Yamagata, M. Morishita, N.J. Kavimandan, K. Nakamura, Y. Fukuoka, K.
Takayama, N.A. Peppas, Characterization of insulin protection properties of com-
plexation hydrogels in gastric and intestinal enzyme fluids, J. Control. Release 112
(2006) 343–349.
[32] F. Madsen, N.A. Peppas, Complexation graft copolymer networks: swelling prop-
erties, calcium binding and proteolytic enzyme inhibition, Biomaterials 20 (1999)
1701–1708.
[33] M. Morishita, T. Goto, K. Nakamura, A.M. Lowman, K. Takayama, N.A. Peppas,
Novel oral insulin delivery systems based on complexation polymer hydrogels:
single and multiple administration studies in type 1 and 2 diabetic rats, J. Control.
Release 110 (2006) 587–594.
References
[1] C. Damge, P. Maincent, N. Ubrich, Oral delivery of insulin associated to polymeric
nanoparticles in diabetic rats, J. Control. Release 117 (2007) 163–170.
[2] F.Y. Su, K.J. Lin, K. Sonaje, S.P. Wey, T.C. Yen, Y.C. Ho, N. Panda, E.Y. Chuang, B. Maiti,
H.W. Sung, Protease inhibition and absorption enhancement by functional
nanoparticles for effective oral insulin delivery, Biomaterials 33 (2012) 2801–2811.
[3] T.H. Yeh, L.W. Hsu, M.T. Tseng, P.L. Lee, K. Sonjae, Y.C. Ho, H.W. Sung, Mechanism
and consequence of chitosan-mediated reversible epithelial tight junction open-
ing, Biomaterials 32 (2011) 6164–6173.
[4] L. Shen, H.Y. Zhao, J. Du, F. Wang, Anti-tumor activities of four chelating agents
against human neuroblastoma cells, In Vivo 19 (2005) 233–236.
[5] M. Morishita, T. Goto, N.A. Peppas, J.I. Joseph, M.C. Torjman, C. Munsick, K.
Nakamura, T. Yamagata, K. Takayama, A.M. Lowman, Mucosal insulin delivery
systems based on complexation polymer hydrogels: effect of particle size on in-
sulin enteral absorption, J. Control. Release 97 (2004) 115–124.
[6] F.C. Wu, M. Laskowski, The effect of calcium on chymotrypsins [alpha] and B,
Biochim. Biophys. Acta 19 (1956) 110–115.
[7] R.J. Schilling, A.K. Mitra, Degradation of insulin by trypsin and alpha-chymotrypsin,
Pharm. Res. 8 (1991) 721–727.
[8] C.C. Kim, S. Falkow, Delineation of upstream signaling events in the salmonella path-
ogenicity island 2 transcriptional activation pathway, J. Bacteriol. 186 (2004)
4694–4704.
[9] L. Tei, Z. Baranyai, M. Botta, L. Piscopo, S. Aime, G.B. Giovenzana, Synthesis and so-
lution thermodynamic study of rigidified and functionalised EGTA derivatives,
Org. Biomol. Chem. 6 (2008) 2361–2368.
[10] A. Bernkop-Schnurch, C. Paikl, C. Valenta, Novel bioadhesive chitosan–EDTA con-
jugate protects leucine enkephalin from degradation by aminopeptidase N,
Pharm. Res. 14 (1997) 917–922.
[11] Z.X. Liao, S.F. Peng, Y.C. Ho, F.L. Mi, B. Maiti, H.W. Sung, Mechanistic study of
transfection of chitosan/DNA complexes coated by anionic poly(γ-glutamic
acid), Biomaterials 33 (2012) 3306–3315.
[12] C. Leypold, M. Reiher, G. Brehm, M. Schmitt, S. Schneider, P. Matousek, M. Towrie,
Tetracycline and derivatives—assignment of IR and Raman spectra via DFT calcu-
lations, Phys. Chem. Chem. Phys. 5 (2003) 1149–1157.
[13] X. Wang, Aspirin-like drugs cause gastrointestinal injuries by metallic cation che-
lation, Med. Hypotheses 50 (1998) 227–238.
[14] M. Cassidy, C. Tidball, Calcium and magnesium contents of gastrointestinal tissues
in six species, Am. J. Physiol. 217 (1969) 674–679.
[34] J. Lee, Intraluminal distension pressure on intestinal lymph flow, serosal transu-
dation and fluid transport in the rat, J. Physiol. 355 (1984) 399–409.
[35] R.C. Brown, T.P. Davis, Calcium modulation of adherens and tight junction func-
tion, Stroke 33 (2002) 1706–1711.
[15] Y.H. Lin, K. Sonaje, K.M. Lin, J.H. Juang, F.L. Mi, H.W. Yang, H.W. Sung,
Multi-ion-crosslinked nanoparticles with pH-responsive characteristics for
oral delivery of protein drugs, J. Control. Release 132 (2008) 141–149.
[16] K. Sonaje, Y.J. Chen, H.L. Chen, S.P. Wey, J.H. Juang, H.N. Nguyen, C.W. Hsu, K.J.
Lin, H.W. Sung, Enteric-coated capsules filled with freeze-dried chitosan/poly
([gamma]-glutamic acid) nanoparticles for oral insulin delivery, Biomaterials
31 (2010) 3384–3394.
[17] A.N. Flynn, O.A. Itani, T.O. Moninger, M.J. Welsh, Acute regulation of tight junction
ion selectivity in human airway epithelia, Proc. Natl. Acad. Sci. U. S. A. 106 (2009)
3591.
[18] K. Sonaje, K.J. Lin, M.T. Tseng, S.P. Wey, F.Y. Su, E.Y. Chuang, C.W. Hsu, C.T. Chen,
H.W. Sung, Effects of chitosan-nanoparticle-mediated tight junction opening on
the oral absorption of endotoxins, Biomaterials 32 (2011) 8712–8721.
[19] Y. Zhang, J.R. Venugopal, A. El-Turki, S. Ramakrishna, B. Su, C.T. Lim, Electrospun
biomimetic nanocomposite nanofibers of hydroxyapatite/chitosan for bone tissue
engineering, Biomaterials 29 (2008) 4314–4322.
[36] E. Roger, F. Lagarce, E. Garcion, J.P. Benoit, Lipid nanocarriers improve paclitaxel
transport throughout human intestinal epithelial cells by using vesicle-mediated
transcytosis, J. Control. Release 140 (2009) 174–181.
[37] X. Zhao, L. Yin, J. Ding, C. Tang, S. Gu, C. Yin, Y. Mao, Thiolated trimethyl chitosan
nanocomplexes as gene carriers with high in vitro and in vivo transfection effi-
ciency, J. Control. Release 144 (2010) 46–54.
[38] R. Vogelmann, W.J. Nelson, Fractionation of the epithelial apical junctional complex:
reassessment of protein distributions in different substructures, Mol. Biol. Cell 16
(2005) 701–716.
[39] P. Pulimeno, C. Bauer, J. Stutz, S. Citi, PLEKHA7 is an adherens junction protein
with a tissue distribution and subcellular localization distinct from ZO-1 and
E-cadherin, PLoS One 5 (2010) e12207.
[40] A.I. Ivanov, A. Nusrat, C.A. Parkos, Endocytosis of epithelial apical junctional proteins
by a clathrin-mediated pathway into a unique storage compartment, Mol. Biol. Cell
15 (2004) 176–188.