Abstract

Oral therapy utilizing engineered microorganisms has shown promise in the treatment of many diseases. By microencapsulation, viable cells can overcome the harsh gastrointestinal (GI) environment and secrete needed therapeutics into the gut. These engineered cells should be encased without escaping into the GI tract for safety concerns, thus robust microcapsule membrane is requisite. This paper examined the GI performance of a novel microcapsule membrane using a dynamic simulated human GI model. Results showed that the genipin cross-linked alginate-chitosan (GCAC) microcapsules possessed strong resistance to structural disintegration in the simulated GI environment. Leakage of encapsulated high molecular weight dextran, a model material to be protected during the simulated GI transit, was negligible over 72 h of exposure, in contrast to considerable leakage of dextran from the non-cross-linked counterparts. These microcapsules did not alter the microflora and enzymatic activities in the simulated human colonic media. This study suggested the potential of the GCAC microcapsules for oral delivery of live microorganisms and other biotherapeutics.

1. Introduction

Advances in molecular biology research have introduced a wide range of genetically engineered (GE) microorganisms with a superior capacity to produce disease-modifying substrates, such as cytokines, enzymes, vaccines, hormones, antibodies, growth factors, and other therapeutic products [1, 2]. The use of these microorganisms opens up new hopes of treating a wide array of human diseases. Because the secreted biologics are generally fragile and easily degraded or denatured [3], encapsulation technology may offer significant advantages over the conventional biotechnological production methods. Being protected against external stresses, encapsulated bacteria remain viable and functional. They can be delivered proximally to the target site in vivo and continuously secrete therapeutic products to the host at a more effective concentration [4]. Recent research on the microencapsulation of GE cells has demonstrated great potential in the treatment of kidney failure, cancers, hypercholesteraemia, and many other diseases [513].

Oral ingestion is usually a preferred route of administration for therapy; however, microcapsules containing bacterial cells and other biotherapeutic molecules can be disrupted in the harsh gastrointestinal (GI) system by a number of means such as low pH, antimicrobial substances and mechanical stress [14]. Furthermore, the eruption of microcapsules and the subsequent release of engineered bacteria could induce many adverse effects on the body [2]. It was previously shown that oral administration of repeated doses of bacteria may stimulate a host immune response [15, 16]. Propagation of foreign bacteria in the GI tract may cause uncontrolled and persistent production of harmful substances, and may detrimentally disrupt and/or replace the natural microflora [2, 17]. There are also risks of immunomodulation and gene transfer associated with the use of novel microorganisms [18]. Therefore, it is essential that GE bacteria be encased in the microcapsules, perform the therapeutic functions during the GI transit, and be excreted along with the intact microcapsules in feces without being retained in the body, even though these GE cells are classified as nonpathogenic [7]. To fulfill these requirements, it is important to maintain the structural integrity of microcapsules and thus prevent cell leakage during the GI transit. Current literature reports on oral delivery systems mainly focus on the controlled release of encapsulated contents, for example, probiotics and drugs [4, 1923]. Scanty research is available on microcapsules intended to retain cells throughout the GI transit.

Alginate and chitosan are biomaterials widely studied for cell encapsulation because of their excellent biocompatibility, status as FDA approved food additives, and mild process conditions [2428]. It was reported that the ionically linked alginate-chitosan (AC) membrane improved gastric survival of probiotics, but some limitations such as inadequate stability, susceptibility to degradation and cell leakage persist [2933]. Genipin and its derivatives, extracted from gardenia fruits [34], have traditionally been used as a herbal medicine and a natural colorant in the food industry [35]. Previous research has demonstrated its low cytotoxicity and potential in protein and live cell delivery [3638]. We have previously developed a novel covalently cross-linked microcapsule system composed of a calcium alginate core with a genipin cross-linked chitosan membrane [36, 39]. Our recent data showed that this microcapsule membrane possessed strong membrane stability and resistance to enzymatic degradation [40]. The purpose of this study is to further evaluate the potential of this genipin cross-linked alginate-chitosan (GCAC) microcapsule system for GI applications by using a dynamic human GI model. For comparison, the frequently used AC microcapsules were also tested.

2. Materials and Methods

2.1. Chemicals

Sodium alginate (low viscosity), and fluorescein isothiocyanate (FITC) labeled dextran ( 2,000 KD) were supplied by Sigma-Aldrich, USA. Chitosan (low viscosity, , degree of deacetylation or DDA = 73.5%) and genipin were purchased from Wako BioProducts, USA. 4-nitrophenyl- -D galactopyranoside, 4-nitrophenyl- -D glucopyranoside, and 4-nitrophenyl- -D galactopyranoside were obtained from Acros Organics, USA. 4-nitrophenol, 4-nitrophenyl- -D glucopyranoside, and 4-nitrophenyl- -D glucuronide were purchased from Sigma-Aldrich, USA. All other reagents and solvents were of reagent grade and used as received without further purification.

2.2. Preparation of Microcapsules

The AC and GCAC microcapsules were prepared as per previously reported protocol [39]. Unless otherwise specified, the cross-linking reaction was performed by suspending the AC microcapsules in a genipin solution (2.5 mg/mL) at room temperature (RT). Sterile microcapsules were prepared similarly except that the entire encapsulation procedure was carried out in a biological containment hood and all solutions used were either 0.22  m filtered or autoclaved to ensure sterility. Microcapsules containing high molecular weight FITC-labeled dextran were prepared by mixing FITC-dextran with an alginate solution, making the final concentrations of alginate and FITC-dextran at 15 mg/mL and 2 mg/mL, respectively. The subsequent processes, including the formation of alginate beads, coating and cross-linking were performed using the aforementioned procedures.

2.3. Simulation of the Human Gastrointestinal (GI) Environment

The human GI conditions used in this study were simulated in vitro by means of five sequential bioreactors (Figure 1). Each compartment simulates a different part of the human GI tract: the stomach, the small intestine, the ascending colon, the transverse colon, and the descending colon [41, 44]. Human fecal slurries containing normal human GI bacterial cells were inoculated into the simulated colon (the last three vessels). Food content of human western diet suspension, composed of (per liter) 1 g arabinogalactan, 2 g pectin, 1 g xylan, 3 g potato starch, 0.4 g glucose, 3 g yeast extract, 1 g peptone, 4 g mucin, and 0.5 g cystein, was fed to the first vessel three times a day. After feeding, acidification of the stomach (pH 2) occurred, followed by neutralization (pH 6.8) in the second vessel and addition of simulated pancreatic juice (0.9 g pancreatin, 6 g bile salts, and 12 g NaHC per liter) to the simulated small intestine. Afterwards, the suspension was transferred to the simulated ascending colon, the transverse colon, and the descending colon, and finally excreted as effluent. The first two reactors were of the fill-and-draw principle with programmed periods of residence and stirring, and the last three were continuously agitated (approximately 250 rpm). The whole system was maintained under anaerobic conditions by flushing the headspace of each vessel with N2 for 15 minutes every day and the temperature of each vessel was kept constant at by a thermostat. The pH conditions, fluid volume, retention time at each stage, as well as the entire transit were simulated under computer control. This in vitro human GI model was validated against in vivo data by earlier studies [41].

2.4. Resistance of Microcapsules to the Simulated Human GI Transit

To study the microcapsule resistance to the simulated human GI transit, microcapsules (0.80 g) were exposed to the simulated human GI fluids for the estimated maximum period of time for the human GI transit (Table 1) [42]. Microcapsule samples were withdrawn at varied stages for morphological examination under an inverted microscope (LOMO, PC), and microphotographs taken as records using a digital camera (Canon Power shot G2, Japan). Percent defective microcapsules were estimated in three randomly picked observation fields.

To assess the recovery of microcapsules after the simulated human GI transit, microcapsules of known weight were placed in a sealed teabag-like container and exposed to the simulated GI media according to the timetable described in Table 1. At the final stage, the retrieved microcapsules were washed, dried using filter paper for approximately 10 min and weighed. Percent recovery was defined as: % recovery = ( W)/ *100, where and W are the weights of the microcapsules before and after being exposed to the simulated GI transit, respectively.

2.5. Retention of Encapsulated Macromolecules against Leaching to the Simulated Human GI Fluids

To evaluate the capacity of the GCAC microcapsules to retain enclosed large-sized materials, high molecular weight FITC-dextran was encapsulated as a model macromolecule. These FITC-dextran-containing microcapsules (  g) were exposed to 2 mL of the simulated human gastric fluid taken from the GI model and incubated in an Environ shaker at , 100 rpm for 1 hour, followed by a 72 hours incubation in 2 mL of the simulated human intestinal fluid from the GI model ( , 100 rpm). Samples of the incubation media were withdrawn periodically and leaking of the encapsulated FITC-dextran was assessed spectrofluorometrically using a Microplate Fluorescence Reader (FLx800, Bio-Tek Instruments, Inc.) at absorption and emission wavelengths of 485 nm and 528 nm, respectively. Volume of the incubation medium was kept constant by adding fresh simulated fluid after sampling. Data are presented as . from triplicate experiments.

2.6. Effect of Microcapsules on the Simulated Gut Microflora

To investigate the influence of oral administered microcapsules on gut microflora, sterile microcapsules (1.0 g) were mixed with the suspension of the simulated transverse colon (10 mL). After 0, 6, 12, and 24 hours of anaerobic incubation at , samples of the incubation medium were aseptically withdrawn and serially diluted with physiological saline. Bacterial enumeration for specific fecal marker microorganisms, Escherichia coli, Staphylococcus sp., Lactobacillus sp. as well as total aerobes and total anaerobes, was performed using an agar-plate-count assay. The plating media and incubation conditions used in the experiments are listed in Table 2. The simulated colonic suspension without microcapsules was used as control.

2.7. Effect of Microcapsules on the Microbial Enzyme Activities in the Simulated Colonic Media

To assess the effect of oral administered microcapsules on the GI microbial enzyme activities, the suspension from the simulated transverse colon (20 mL) was incubated anaerobically at in the presence of sterile microcapsules (2.0 g) for up to 24 hours. At different time points of 0, 12, and 24 hours, the enzymatic activities of -galactosidase, -glucosidase, -glucuronidase, -galactosidase, and -glucosidase in the incubation medium were analyzed spectrometrically using the method described earlier [43, 44]. The absorbance at 405 nm was recorded by a Quant multiplate reader (Bio-Tek Instruments). The simulated colonic fluid free of microcapsules was used as control. Results are expressed as percentage of enzymatic activities relative to the control at each time point. Data of the control samples at each time point were normalized to 100%.

Numerical values are shown as mean ± standard deviation. Statistical analyses used the two-tailed Student’s t-test with considered significantly different.

3. Results

3.1. Resistance of Microcapsule to the Simulated Human GI Transit

To assess the resistance to the GI environment, the GCAC microcapsules were exposed to the simulated GI media representing different phases of digestion for a length of time based on the estimated maximum retention in the human GI tract (Table 1). Figure 2 depicts the photomicrographs of the microcapsules after exposure to the simulated human GI fluids. The tested microcapsules remained morphologically stable during the simulated gastric incubation (2 hours, ) (Figures 2(a), 2(d)), but behaved differently in the subsequent simulated intestinal transit depending on the extent of cross-linking. The GCAC microcapsules cross-linked at swelled appreciably in the simulated small intestine (pH 7.2–7.4) (Figure 2(b)). They became fragile and adhesive in the simulated descending colon where shriveled or partially dissolved beads were observed. When leaving the simulated descending colon, 30–40% of the microcapsules lost their structural integrity (Figure 2(c)). In contrast, the majority of the GCAC microcapsules cross-linked at higher temperature (RT) maintained their physical stability during the entire simulated human GI transit (Figures 2(e) and 2(f)), with more than 80% recovery after 72 hours exposure (Figure 3). In comparison, the recovery of the GCAC microcapsules with less cross-link (cross-linked at ) was lower (61.4%) (Figure 3).

3.2. Retention of Encapsulated HMW Dextran during the Simulated GI Transit

To examine the capacity of the GCAC microcapsules to encase large-sized materials in the GI environment, encapsulated high molecular weight FITC-labeled dextran was exposed to the simulated human GI media. Being a large polymer of 2,000 KD, this fluorescent probe was indefinitely withheld inside the intact microcapsules and could not leak out unless the microcapsule membranes became defective or damaged [45]. During the first hour of exposure to the simulated gastric fluid, no FITC-dextran was detected in the incubation medium (data not shown). In the subsequent exposure to the simulated intestinal medium, the leakage of the encapsulated FITC-dextran from the non-cross-linked AC microcapsules increased gradually with the incubation time, attaining the fluorescence intensity of 70, 100 and 115 at 24, 48 and 72 hours, respectively. In contrast, negligible amounts of FITC-dextran escaped from the GCAC microcapsules, with very low fluorescence intensity of the medium detected (not exceeding 50) throughout the experiment (Figure 4), which significantly outran those of the AC microcapsules.

3.3. Oral Administration of Microcapsules on Gut Microflora and Enzymatic Activities

The effect of microcapsules on human GI microflora was assessed by investigating the population of fecal marker microorganisms and the activity of 5 representative GI microbial enzymes in the simulated colonic media containing sterile microcapsules. Since the simulated GI model is a dynamic system, static experiments were performed in this study to maximize the effects. We found no marked differences in the tested microbial population including total aerobes, total anaerobes, Escherichia coli, Staphylococcus sp., and Lactobacillus sp. in the stimulated colonic media containing tested microcapsules of either AC or GCAC formulation when compared to the control media in the absence of tested microcapsules (Table 3, all ).

Figure 5 shows the effect of the microcapsules on the activities of microbial enzymes in the simulated transverse colon suspension. As time elapsed, a slight decline in the tested enzymatic activities was found in the simulated colonic fluids in the presence of microcapsules. As an exception, a decrease of more than 20% in the activity of -glucuronidase was detected after 12 hours of contact with microcapsules, yet the loss remained at a similar level with extended incubation for up to 24 hours (Figure 5(c)). No significant differences in the alteration of enzymatic activities were found when the GCAC microcapsules were compared with the AC microcapsules ( ).

4. Discussion

For successful exploitation of microcapsules as an oral delivery device, understanding of their performance under physiologically pertinent conditions that represent the human GI tract is essential. Although in vivo research using specific techniques such as histological sectioning [46], radiography [47] and gamma scintigraphy [48, 49] can progressively track the microcapsules in the GI tract, it remains difficult to follow the orally administered microcapsule at every stage of digestion on either animals or humans due to tedious processing, small size of the microcapsules, limited detection resolution and ethical constraints. In vitro simulation offers a number of advantages, for example, well-controlled experimental conditions and easy sampling, especially preferable for screening and examining a variety of samples. Buffered solutions, for example, with pH at 1-2 or 6.5-7.5, are frequently reported in literature as the simulated GI fluids [31, 50, 51]; however they only represent the pH in the stomach and in the intestine, and do not mimic the complex human GI microbial ecosystem. While other ex vivo and in vitro simulated models including USP apparatus were also reported [52], most of them were static systems, where fluids cannot be continuously transferred into the sequential GI compartments. Our present study used a dynamic computer-controlled human simulated GI model, which mimics the gradual transit of ingested materials through the simulated GI tract from feeding to discharge. It also maintains the microbial community actually presented in the human intestinal system, which allows for assessing the behavior of the microcapsules under more physiologically pertinent conditions.

As has been noted earlier, it is imperative that microcapsules maintain physical integrity during the GI transit to prevent the leaking of genetically engineered cells, which is strongly dependent on the microcapsule’s robustness and stability. So far few studies were reported to address this matter [44]. The present study investigated the behaviour of the GCAC microcapsules in the simulated human GI environment. When exposed to the simulated GI media, the microcapsules experienced the simulated transit through the GI tract including pH fluctuation, enzymatic degradation, microorganism actions, mechanical stresses, as well as other related chemical and physiological constraints. The exposure time was chosen according to the maximum period retention in the human GI tract that represents the most challenging case for the tested microcapsules. The better-preserved morphology and higher retrieval rate associated with the GCAC microcapsules cross-linked at RT (Figures 2 and 3) indicated improved membrane stability by higher degree of cross-linking. Moreover, negligible amounts of encapsulated fluorescent probe were released into the incubation GI media from the GCAC microcapsules (Figure 4), suggesting that the integrity of the microcapsule membrane was withheld. In contrast, the non-cross-linked AC microcapsules were found vulnerable to structural disruption in the harsh GI condition, likely due to the relatively weak polyelectrolyte complexation of membrane materials that are prone to gastric and proteolytic degradation. These findings corroborated the results reported previously [44, 5355]. The present study demonstrated that covalent cross-linking of microcapsule membranes by genipin substantially improved the resistance to membrane degradation in the simulated human GI environment.

Another important prerequisite of oral therapy utilizing microencapsulation is that the administration of microcapsules should not disturb the natural colonic flora, particularly when prolonged and repeated oral intake of a rather large quantity of microcapsules is required for therapy. Microcapsules are made of various materials and chemicals through complexation and cross-linking reactions, all of which may have an effect on the biocompatibility of the final microcapsules. In particular, the well-balanced gut microbiota are important in maintaining human health [56, 57] and should not be altered by the intake of microcapsules. When taken into account the static nature and a rather large dosage (1.0 g of microcapsules in 10 mL of intestinal fluid) of the experiments in this in vitro study, our results suggested that the materials used to construct the microcapsules did not evoke appreciable adverse effects on the human intestinal flora and that genipin cross-linked chitosan membranes did not compromise the biocompatibility of the microcapsules when compared to the non-cross-linked subjects. Decrease in the activities of the tested microbial enzymes in the simulated colonic media containing microcapsules was detected. It could possibly be attributed to the binding effect or diffusion of the enzymes to the microcapsules, though further research may continue to elucidate the consequence.

5. Conclusions

This study examined the performance of the genipin cross-linked alginate-chitosan microcapsules under physiologically pertinent GI conditions using a dynamic simulated human GI model. Results showed that these GCAC microcapsules possessed superior resistance against disintegration in the simulated GI environment and did not appreciably alter the normal gut flora. This novel microcapsule formulation is promising for oral delivery of genetically engineered bacteria; however, further studies on encapsulated cell viability in the gut lumen are needed before its full potential can be realized.

Acknowledgments

This paper was supported by the Canadian Institutes of Health Research (CIHR) MOP 64308. Postgraduate fellowships from the Natural Sciences and Engineering Research Council of Canada (to H. Chen and C. Martoni) and Fonds Québécois de la Recherche sur la Nature et les Technologies (to H. Chen and W. Ouyang) are greatly acknowledged. The authors also thank M. L. Jones for his help with the GI model set-up, T. Haque, M. Shafiei, and H. Wightman for assistance in the experiments.