Despite the advances in biotechnology and allied fields, successful oral administration of protein/peptide drugs has remained to be an elusive goal for the researchers around the globe. Orally administered protein drugs encounter several difficulties such as rapid pre-systemic degradation and poor absorption in the gastrointestinal (GI) tract [1]. Therefore, a delivery system is needed to improve the oral bioavailability of such drugs. An ideal delivery system for oral administration of protein drugs should provide the intact drug to the systemic circulation and reversibly increase the permeability of the mucosal epithelium to enhance the absorption of protein drugs [2]. Additionally, this delivery system should not show any adverse systemic effects or irreversibly damage the intestinal epithelium.
In a recent study, we reported a pH-responsive nanoparticle (NP) system shelled with chitosan (CS) for the oral delivery of insulin via the paracellular pathway [3, 4]. CS, a cationic polysaccharide, can adhere to the epithelial surfaces to impart transient opening of the tight junctions between contiguous cells [5]. The results obtained in a diabetic rat model indicated that the prepared NPs could effectively adhere to the mucosal surface, increase the intestinal absorption of insulin and produce a slower, but prolonged hypoglycemic effect. Additionally, the in vivo toxicity study indicated that the test NPs were safe even at a dose 18 times higher than that used in the pharmacodynamics/ pharmacokinetics (PD/PK) study [4].
CS is a natural-origin polymer, which has been widely used as a weight-loss product and been developed as a safe excipient in drug formulations [6]. Additionally, CS, an excellent mucoadhesive agent, has been widely used as a paracellular permeability enhancer for improving the intestinal absorption of hydrophilic macromolecules [6, 7]. CS has been proven to be safe and biocompatible [8]. However, it is often argued that the paracellular permeation enhancement by CS may result in absorption of unwanted toxins present in the GI tract [9]. Therefore, the purpose of this study was to evaluate the effects of test NPs on the oral absorption of unwanted toxins present in intestinal lumen.
Lipopolysaccharide (LPS, also known as endotoxin) is a component of the outer membrane of gram-negative bacteria. If delivered to the systemic circulation, LPS can trigger a systemic inflammatory response that may progress to septic shock and sometimes death in vivo [10]. Although disintegrating enteric bacteria continuously release LPS in the intestinal lumen, the intact intestinal epithelium constitutes a barrier that prevents significant absorption of this amphiphilic molecule [11]. Hence, the bacterial LPS was selected as a model toxin in this study.
The biodistribution of the orally administered LPS was studied in a rat model using the single-photon emission computed tomography (SPECT)/computed tomography (CT). To understand how the intestinal epithelium prevents the in vivo absorption of LPS, we investigated the absorption of fluorescent LPS in rats using the confocal laser scanning microscopy (CLSM). Then, the effects of test NPs on the enteral absorption of LPS were investigated using SPECT/CT and CLSM. Finally, the adverse effects of orally administered LPS and effects of test NPs on them were investigated in an in vivo toxicity study.
2. Materials and Methods
2.1. Preparation and characterization of test NPs
CS (MW 80 kDa) with a degree of deacetylation of approximately 85% was acquired from Koyo Chemical Co. Ltd. (Japan), while γ-PGA (MW 60 kDa) was purchased from Vedan Co. Ltd. (Taichung, Taiwan). A sample of 100 mg of insulin (from bovine pancreas, 27.4 IU/mg, Sigma-Aldrich, St. Louis, MO, USA) was dissolved in 10 mL of 0.01N HCl and this solution was neutralized with 0.1N NaOH [12]. The insulin solution was then diluted with deionized (DI) water to make a stock solution of 1 mg/mL insulin. Test NPs were prepared by a simple ionic-gelation method previously reported by our group [4, 13].
In brief, the prepared insulin stock solution (1 mL) was premixed with an aqueous γ-PGA (2 mg/mL, 1 mL). Subsequently, TPP (1 mg) and MgSO4 (2 mg) were added into this mixture and thoroughly stirred for 30 min. The mixed solution was then added by flush mixing with a pipette tip into an aqueous CS solution (1.5 mg/mL, 8 mL, pH 6.0) under magnetic stirring at room temperature. Self-assembled NPs were collected and washed three times with DI water by centrifugation at 8000 rpm for 50 min. The centrifuged NPs were redispersed in DI water and stored at 4°C until used. The mean particle size and zeta potential value of NPs were measured using a Zetasizer (Nano ZS, Malvern Instruments Ltd., Worcestershire, UK) [14].
2.2. Evaluation of the micelle forming characteristics of LPS
LPS molecules are composed of a hydrophilic oligosaccharide chain of varying length and a hydrophobic, membrane-anchoring moiety, termed lipid A []. The amphiphilic nature of LPS results in the formation of micelles in an aqueous environment. To confirm these characteristics, different concentrations of LPS were suspended in PBS (pH 7.4). The formation of micelles was confirmed by measuring the particle size and zeta potential of the resulting suspensions using Zetasizer (Nano ZS, Malvern Instruments Ltd., Worcestershire, UK).
2.3. Animal studies
Animal studies were performed in compliance with the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Resources, National Research Council, and published by the National Academy Press, revised in 1996.
2.3.1. Biodistribution study
The biodistribution of LPS and the insulin loaded in NPs was studied in rats using the SPECT/CT (n = 3). In this study, LPS (Escherichia coli, Serotype 0111: B4, Sigma-Aldrich, St. Louis, MO, USA) was radiolabeled with 99mTc-pertechnetate (99mTc, emitting 140KeV photons) using a stannous chloride (SnCl2) method [15]; the labeling efficiency of 99mTc to LPS was determined by the instant thin layer chromatography (ITLC). The insulin was radiolabeled with 123iodine (123I, emitting 159KeV photons) using an iodogen-tube (Pierce Iodination Tubes, Thermo Fisher Scientific, Rockford, IL, USA) method, as per the manufacturer's instructions. The 123I-insulin was separated from the free-form 123I using a centrifugal dialysis device (MWCO: 3 kDa, Amicon Ultra 4, Millipore, Billerica, MA, USA); its labeling efficiency was determined by a reversed-phase HPLC system equipped with a gamma counter [16]. The obtained 123I-insulin was then used to prepare test NPs as described above.
To study the biodistribution, rats were fed with 99mTc-LPS (?) alone or with 123I-insulin-loaded NPs (?). For the group receiving both 99mTc-LPS and 123I-insulin-loaded NPs, test NPs were administered 1 h after the ingestion of 99mTc-LPS, to mimic the natural conditions where LPS is evenly distributed in the GI tract. The detailed protocol used in the image acquisition was previously described by our group [17, 18]. In brief, animal images were acquired using a dual modality system, NanoSPECT/CT (Bioscan Inc., USA), which is capable of detecting two kinds of isotopes simultaneously at high resolution (~0.6 mm). Animals were kept under the controlled temperature (37°C) and anesthesia (1.5% isoflurane in 100% oxygen) during imaging. Dual isotope dynamic SPECT images were acquired at 30-min intervals up to 24 h after the oral administration of test samples. Additional CT images were collected for anatomical references and used to investigate the details of radiotracer distribution in rats. The co-registered dynamic scintigraphy and CT images were displayed and analyzed using the PMOD v2.9 image analysis software (PMOD Technologies Ltd., Zurich, Switzerland).
Areas of the whole body and GI tract were selected for the volume of interest (VOI) analysis to obtain their biodistribution profiles. To calculate the percentage of initial dose (% ID) within each area, the corresponding VOIs were manually drawn on the co-registered dynamic SPECT and reference CT images. The peripheral tissue/plasma compartment was defined as the whole body excluding the GI tract and urinary bladder. The biodistribution data were expressed as %ID using the following formula:
(total counts in the target organ) ï‚´ 100 %
ingested counts
% ID/organ =
2.3.2. Ultrastructural examination of the in vivo tight junction opening activity of CS-NPs
The in vivo tight junction opening activity of test NPs was studied using transmission electron microscopy (TEM). For this, CS was labeled with quantum dots (QD) as per a previously reported method [19]. Briefly, carboxyl QDs (40 µL, 0.6nM, Qdot® ITKâ„¢, Invitrogen, USA) were activated in the presence of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC, 60 µL, 50mM) and N-hydroxysuccinimide (NHS, 30 µL, 25mM) for 15 min under gentle stirring. The resulting NHS-activated QDs were covalently linked to the primary amines on CS at pH 6.0. The reaction was carried out under gentle mixing for 4 h. The final QD-CS conjugate was purified by the centrifugal spin filtration and resuspended in DI water. QD-CS-NPs were then prepared using the method described above. The QD-CS-NPs (1 mg/mL, 1 mL) were then administered to the overnight-fasted male ICR mice (33-40 g, National Laboratory Animal Center, Taipei, Taiwan). Animals were sacrificed 3 h later and the intestinal segments were dissected and washed three times with isotonic saline. The dissected intestinal segments were then fixed in 4% paraformaldehyde. The fixed intestinal segments were cut into small pieces, washed with s-Collidine buffer and treated with 2% lanthanum nitrate for 2 h at room temperature. After being rinsed in s-Collidine and then PBS, the tissue samples were processed for TEM. Test samples were postfixed in 1% osmium tetroxide, dehydrated in a graded series of ethanol, and embedded in the Spurr resin. Ultrathin sections were then cut with a diamond knife and loaded onto the TEM grids. The sections were examined by a Philips CM10 electron microscope (Philips Electron Optics B.V.) at accelerating voltage of 120 kV and micrographs were taken.
2.3.3. GI absorption of FITC-LPS and Cy3-insulin-loaded-NPs
FITC-labeled-LPS (FITC-LPS, Sigma-Aldrich, St. Louis, MO, USA) and Cy3-labeled-insulin were used to visualize their in vivo absorption characteristics using CLSM (TCS SL, Leica, Germany). The Cy3-insulin was synthesized as per the method described in the literature [20]. Briefly, Cy3-NHS ester (GE Healthcare, Pittsburgh, PA, USA) dissolved in DMSO (1 mg/mL, 1 mL) was slowly added to an aqueous solution of bovine insulin (1% w/v in 0.01N HCl, 4 mL) and stirred overnight at 4°C. To remove the unconjugated Cy3, the synthesized Cy3-insulin was dialyzed in the dark against 5 L of 0.01N HCl and replaced on a daily basis until no fluorescence was detected in the dialysis medium. The resultant Cy3-insulin was lyophilized in a freeze dryer. Fluorescent NPs were then prepared for the subsequent in vivo CLSM studies as per the procedure described above.
FITC-LPS (2 mg/mL, 0.5 mL) alone or in combination with a mucolytic agent (N-acetylcystein) were administered to the overnight-fasted male Wistar rats (n = 3). To study the effects of CS-NPs on the LPS absorption, Cy3-insulin-loaded NPs (2 mg/mL, 0.5 mL) were administered 1 h after the administration of FITC-LPS. Rats were sacrificed 3 h later and intestinal segments were then dissected and washed three times with isotonic saline. The isolated intestinal segments were fixed in Methanol-Carnoy's fixative and processed for paraffin-embedding. The paraffin-embedded sections were dewaxed, hydrated and stained with Alexa-633-labeled wheat-germ-agglutinin and SYTOX blue (Invitrogen, ?, ?, USA) to visualize the mucus and nuclei, respectively. Finally, the stained sections were examined under CLSM.
2.3.4. In vivo toxicity study
It is well known that LPS, a potent inflammatory agent, has many physiological and biochemical effects in vivo [21]. Therefore, it was decided to evaluate the toxicity of the orally administered LPS in ICR mice. Animals were randomly divided into four groups (n = 7 for each studied group). The experimental groups received an oral dose of LPS (5 mg/kg) with or without blank NPs (10 mg/kg); the group without any treatment was used as a control. Additionally, a group receiving intraperitoneal (IP) LPS (5 mg/kg) served as a reference for the extent of toxicity produced by the systemic LPS. All animals were fed with normal chows and water ad libitum. Animals were observed carefully for the onset of any signs of toxicity and monitored for changes in body weight. At the end of the treatment period, animals were anaesthetized (tribromoethanol, IP, 240 mg/kg) and blood samples were collected via cardiac puncture for the determination of alanine aminotransferase and aspartate aminotransferase using a FUJI DRI-CHEM 3500s serum-chemistry analyzer.
2.4. Statistical analysis
Comparison between groups was analyzed by the one-tailed Student's t-test (SPSS, Chicago, Ill). All data are presented as a mean value with its standard deviation indicated (mean ± SD). Differences were considered to be statistically significant when the p values were less than 0.05.
3. Results and discussions
The efficacy of oral delivery of peptides/proteins is often limited because of their inherent instability in the GI tract and their low permeability across the epithelial membranes. The high molecular weight of this class of drugs coupled with their hydrophilic nature restricts their transcellular permeation [1]. Thus, enhancement of the paracellular permeation is a more feasible alternative for oral delivery of peptides/proteins [22]. CS is a well-known mucoadhesive agent with the capability of transiently and reversibly opening tight junctions between epithelial cells [7].
Formulating CS into NPs has advantages over the traditional tablet or powder formulations, as the NPs can readily infiltrate into the mucus layer and deliver the protein drugs to the actual site of absorption (i.e., the tight junctions between epithelial cells). CS is generally regarded as a safe and biocompatible material for drug delivery. However, its effects on the absorption of unwanted toxin present in the GI tract have never been investigated. It is often argued that the disruption of tight junctions by CS may promote the intestinal absorption of such toxins, resulting in systemic immunogenic reactions [9]. To our knowledge, this is a first such study evaluating the effects of CS-NPs on the intestinal absorption of LPS (a model for unwanted toxins present in the GI tract).
3.1. Characterization of the insulin-loaded NPs
The prepared NPs had a mean particle size of 253.2  4.8 nm with a positive zeta potential of 28.2 ± 1.3 mV and their insulin LE and LC were 72.4  3.9 % and 17.9  2.1 %, respectively (n = 6 batches). The NPs were stable in the pH range of 2.0-7.0. The NPs prepared using QD-CS or FITC-CS exhibited similar physicochemical properties as the NPs prepared with unmodified CS.
3.2. Oral absorption and biodistribution of LPS
The absorption and biodistribution of the orally administered LPS was investigated by the SPECT/CT. In the study, 99mTc was used to label LPS and its labeling efficiency was determined by the ITLC. It was found that more than 99% of the 99mTc was bound to the LPS (Fig. 1a). The SPECT/CT images of orally administered 99mTc-LPS are shown in Fig. 1b. As shown, the radioactivity (99mTc-labeled LPS) propagated from the stomach, small intestine and then to the large intestine with time. Overall, the 99mTc-labeled LPS appeared to be restricted within the GI tract, with no quantifiable presence in the plasma and peripheral tissues. These results confirmed the previously reported findings that the orally administered LPS was not able to be absorbed into the systemic circulation in rabbits [23].
It is known that the epithelium lining on the GI tract provides a regulated, selectively permeable barrier between the external environment (the intestinal lumen) and the systemic circulation [24]; it transports nutrients, ions, and fluid tans cellular ly, but prevents the entry of toxins, antigens, and microorganisms. The paracellular route is restricted by the tight junctions at the apical poles of enterocytes that limit the passage of macromolecules. In general, the permeability of the intestinal epithelium depends on the regulation of intercellular tight junctions [22]. Therefore, it is important to study the effects of tight junction opening activity of CS on the transport of antigens (e.g., LPS) present in the GI tract.
3.3. Effects of CS-NPs on the absorption and biodistribution of LPS and insulin
Effects of CS-NPs on the biodistribution of 99mTc-LPS and the loaded drug (123I-insulin) were studied simultaneously using the dual isotope dynamic SPECT/CT in a rat model following oral administration. To mimic the physiological conditions, 99mTc-LPS was ingested 1 h prior to the administration of 123I-insulin-loaded NPs. As shown in Fig. 2a, with time progressing, 99mTc-LPS propagated from the stomach, small intestine and then to the large intestine; the radioactivity associated with LPS was mainly limited to the intestinal lumen throughout the entire course of the study. On the other hand, 123I-insulin started to appear in the urinary bladder at 3 h post administration. These results suggest that the insulin loaded in NPs can traverse the intestinal epithelium and enter the systemic circulation, whereas LPS was unable to do so, hence retaining within the intestinal lumen.
The biodistribution data in Fig. 2a was further processed for the quantification of LPS and insulin entering into the plasma/peripheral tissues. For this, a 3D VOI was constructed; the plasma/peripheral tissue compartment was defined as the whole body (blue contour) excluding the GI tract and urinary bladder (green contour). Dynamic 99mTc-LPS SPECT/CT images overlay with VOI contours in the coronal, sagittal, and transverse views are presented in Fig. 2b. The contour marked with blue line was identified as the whole body, the contour marked with green line was identified as the GI tract and urinary bladder. As previous mentioned, the PP compartment was defined as volume between blue and green contours. The radioactivity counts of 99mTc-LPS and 123I-insulin within the PP compartment were calculated and normalized to each initial ingested dose (%ID/compartment) for comparison. As shown in Fig. 2b (iv), the distribution of 123I-insulin in the PP compartment increased slowly and reached to its maximal value (~8%ID) by 6 h post administration. On the other hand, no measurable 99mTc-LPS could be observed in the PP compartment, indicating that the paracellular permeation effect of test NPs was specific for the loaded insulin. To determine the specific reasons for inability of LPS to pass through the opened paracellular route, we evaluated the physicochemical properties of the LPS in aqueous environment.
3.4. Micelle forming characteristics of LPS
It has been reported that the amphiphilic character of LPS results in the formation of micelles in an aqueous environment above their critical micellar concentration (CMC) [25]. The reported CMC values for LPS vary from 10 nM to 1.6 µM depending on the source of LPS [26]. Based on this concentrations and the fact that the LPS aggregates are usually highly stable, it was suggested that the aggregate forms of LPS should predominate in the concentration range relevant for biological responses [25]. To evaluate the micelle forming characteristics of LPS, we measured the size and zeta potential of the aqueous suspensions of LPS at different concentrations.
Figure 3 shows the effects of the concentration on the average size and particle size distribution of the LPS micelles. With an increase in the concentration, the average size of LPS micelles increased significantly. This could be attributed to the aggregation of multiple micelles at the higher concentrations, which was also evident from the particle size distribution profiles of these micelles. As shown, the size distribution for LPS at lower concentration (0.2 mg/mL) was narrow indicating that the LPS predominantly exist as dispersed micelles at lower concentration. However, at higher concentrations of LPS (2 mg/mL and 10 mg/mL), the particle size distribution became broader indicating the formation of bigger aggregates of LPS. The zeta potential of these micelles was anionic at all concentrations. The anionic charge on the LPS micelles could be a result of two phosphate groups on the LPS structure (Fig. 3).
Based on these findings, we hypothesized that the anionic micelles and aggregates of LPS would be repelled by the negatively charged mucus layer on intestinal epithelium, thus preventing them from entering the systemic circulation. On the other hand, the test NPs with mucoadhesive CS on their surfaces may adhere and infiltrate into the mucus. The infiltrated NPs then disintegrate near the cell surfaces due to their pH-responsive characteristics and release the loaded insulin. The released insulin could then enter the systemic circulation due to the TJ opening activity of CS (Figure 4).
3.5. Ultrastructural examination of TJ opening activity of CS in vivo
The ability of CS to increase the permeability of model drug compounds across Caco-2 cell monolayers has been investigated in numerous studies. However, very few studies have focused on the ultrastructural examination of its TJ opening activity in vivo. Figure 5a shows the TEM micrographs of the intestinal sections from mice treated with QD-CS NPs. As shown, the QD-CS-NPs were adhered to the microvilli and goblet cells indicating that they were able to infiltrate the mucus layer. However, a very small fraction of QD-CS could be observed within paracellular spaces. It has been reported that the size of TJs opened by CS is less than 20 nm [22]; this could be the reason for poor permeation of QD-CS in the paracellular spaces. Therefore, to visualize the TJ opening activity of CS, we examined the movement of lanthanum (La3+) through the opened TJs. La3+ is an electron dense element with a hydrated radius (0.4 nm), which has been widely used a surface coating material for several tissues [27]. As shown in Figure 5b, La3+ could penetrate the paracellular spaces, indicating that the TJ were indeed opened by the QD-CS NPs. A careful examination by TEM revealed no detectable alterations in the morphology of the tight junctions or lateral intercellular spaces, indicating that the TJ opening activity of CS was transient and reversible. These results supported our hypothesis that the test NPs could infiltrate the mucus layer on the intestinal epithelium and could release the loaded drugs near cell surface to promote its absorption.
3.6. In vivo CLSM study
Mucus is a viscoelastic gel layer that protects tissues that would otherwise be exposed to the external environment [28]. The thickness of the intestinal mucus layer has been determined in several studies. Atuma and co-workers reported that the mucus layer was thickest in the colon (~830 µm) and thinnest in the jejunum (~123 µm). They also reported the presence of an inner, firmly adherent mucus layer in the stomach (~80 µm), and colon (~116 µm). In the small intestine, this layer was very thin (~20µm) or absent [29]. The protective functions of this adherent mucus gel layer have been widely investigated [30]. In the stomach and duodenum, it provides a stable unstirred layer supporting surface neutralization of luminal acid by mucosal bicarbonate secretion. In this way, it maintains a pH gradient from acid in the lumen to near neutral pH at the mucosal surface [31]. In addition, the mucus layer has been shown to act as a physical barrier to the enteric bacteria in colon and hinders their access to the underlying epithelium [Johansson, PNAS 2008]. However, the protective effects of the mucus layer against the LPS are still unknown.
In present study, we hypothesized that the mucus layer plays an important role in preventing the intestinal absorption of LPS. To prove this hypothesis, the in vivo absorption of LPS was studied at a microscopic level using the CLSM. For this, the FITC-LPS with or without a mucolytic agent (N-acetylcystein) were administered to the overnight fasted rats. It was found that the mucus layer was intact under normal circumstances (Figure 6, upper panel). The administration of N-acetylcystein caused the mucus layer to detach from the intestine due to the mucolytic effects of this agent (Figure 6, lower panel). The FITC-LPS was found to be restricted outside the mucus layer, probably due to the charge repulsion between the negatively charged mucus and anionic LPS micelles (Figure 6, upper panel). The mucolytic agent decreases the viscosity of mucus layer which makes it easy to detach from the epithelial surfaces [32]. Therefore, the FITC-LPS administered with the mucolytic agent was able to infiltrate the mucus layer and had access to the epithelium surfaces, which lead to its absorption in to the systemic circulation, as indicated by its presence on the basolateral side of villi (Figure 6, lower panel). It has also been reported that the intestinal mucus layer is rich in LPS binding proteins [33], which might contribute to the inability of LPS to penetrate the mucus layer.
Figure 7 shows the CLSM images acquired from the rats administered with FITC-LPS followed by administration of Cy3-insulin loaded NPs. As expected, the Cy3-insulin was able to traverse the mucus layer and underlying epithelium, as indicated by its presence on the basolateral side of villi (magenta). In contrast, the FITC-LPS was still restricted away from the epithelial cell surfaces (green) by the mucus layer (red). These results supported our hypothesis that the paracellular permeation enhancement effect of CS-NPs was specific for the loaded insulin.
3.7. In vivo toxicity study
LPS is a potent proinflammatory agent, which play an important role in the pathogenesis of endoxic shock [11]. It can initiate the production of noxious mediators like acute-phase proteins, cytokines and prostaglandins, which are responsible for the endotoxic shock. However, under the experimental conditions the LPS can induce pathological processes only when it is administered parenterally [34]. Therefore, it was decided to study whether the test NPs can promote the toxic effects of orally administered LPS. To study this, the changes in body weight and the blood biochemistry parameters (alanine aminotransferase and aspartate aminotransferase) were evaluated.
As expected, the daily IP injections of LPS (5 mg/mL) lead to a significant decrease in body weight and increase in the SGPT levels. In contrast, the oral administration of LPS (5 mg/kg) did not produce any adverse effects as compared to the untreated controls (Figures 8a and 8b). It is a well-known fact that the LPS is poorly absorbed from the intestine. It was reported that the large doses of orally administered LPS failed to produce any toxic symptoms in rats [35]. In another study, the effects of orally administered LPS on the intestinal morphology and cell proliferation were examined in rats. It was found that, there was no evidence of epithelial changes, leukocyte infiltration or any other abnormality in the gut of LPS-treated as compared to the controls [11]. Similarly, the group receiving oral LPS (5 mg/kg) followed by test NPs (10 mg/kg) did not show any significant changes in the body weight, SGOT or SGPT levels as compared to the control. These results confirmed that the paracellular permeation enhancement by CS-NPs does not promote the intestinal absorption of LPS.
5. Conclusions
The purpose of this study was to determine whether the CS-NPs can enhance the oral absorption of LPS (a model toxin) present in the intestinal lumen. The biodistribution study indicated that orally administered LPS was unable to enter the systemic circulation, whereas, the insulin loaded in NPs could be absorbed. Additionally, the toxicity study confirmed that CS-NPs did not improve the oral absorption of LPS. On the basis of these results, it can be concluded that the CS-NPs can be used as a safe carrier for the oral protein delivery.
Acknowledgment
This work was supported by a grant from the National Science Council (NSC 96-2120-M-007-004), Taiwan, Republic of China.
