Tissue Engineering and Banking Laboratory, National Centre for Cell Science, Ganeshkhind, Pune 411 007, India

*Corresponding author (Fax, 91-20-5672259; Email, aahardikar@hotmail.com).

Techniques for immunoisolation and immobilization of viable cells within semipermeable microcapsules have been developed using highly sophisticated droplet generator systems. We propose here an indigenously designed, simple and efficient droplet generator system for encapsulation of the pancreatic islets employing chitosan-alginate matrix. The droplet generator system comprises of a needle assembly, a 3-way valve with extended rubber tubing and a filtration unit connected to a pressure pump. Microbeads of the size of around 400 m m diameter or even lesser (minimun attainable size 20·2 m m) could be easily generated using the droplet generator system proposed here. Islet microcapsules cultured in Roswell Park Memorial Institute (RPMI) 1640 with 10% fetal calf serum showed around 98% viability, comparable to that of the non-encapsulated islets. Transplantation of microencapsulated islets to streptozotocin (STZ)-induced diabetic mice, resulted in disappearance of hyperglycemia and restoration of normoglycaemia during a 30-day follow-up suggesting graft functionality. No graft failures were observed in any of the transplanted mice (n = 15) and none of them showed membrane associated fibrous overgrowth, which can be attributed to the fibroblast-growth inhibitory properties of chitosan. The proposed design appears to be superior in its simplicity and cost effectiveness and comparable in performance with the microcapsule generator designs proposed so far. The proposed system can be further exploited for encapsulation and immunoisolation of various cell types in alginate based matrices.

1. Introduction

Type 1 diabetes mellitus essentially involves a total/ partial destruction of islet b cells leading to an insulin deficiency and related diabetic complications. The present line of therapy for such subjects involves daily administration of insulin with routine monitoring of plasma glucose levels. Pancreatic/islet transplantation was looked upon as a seemingly efficient and less traumatizing alternative. However, recipients had to be kept on immunosupressive drugs for a considerable length of time. Therefore, immunoisolation devices progressively gained importance over the traditional pancreatic/islet transplant procedures. Immunoisolation of transplanted cells by artificial barriers that permit crossover of low molecular weight substances, such as nutrients, electrolytes, oxygen and bio-secretory products, but not of immunocytes and other transplant rejection effector mechanisms, provides great promise for developing new technologies to overcome these problems in a reasonable time frame.

Microcapsules from alginate and other polymers were reported (Boag and Sefton 1987) and some droplet generator designs were proposed (Klein et al 1983; Sefton et al 1987). However most of these rely on a highly sophisticated instrument setup. We propose here a simple, cost effective and efficient droplet generator system
for islet encapsulation, using chitosan-alginate based matrices.

2. Materials and methods

2.1 Design of microcapsule generator system

Basic design of the droplet generator is shown in figure 1. The assembly is composed of three major parts viz., a three-way valve (B Braun, Germany) with an extension tubing, Swinny filter 0·2 m m (Tarson, India) and a 14 cm long, 16 G flat end stainless steel needle (Hispan, India). A sterile syringe (Dispovan, India) containing the cell suspension was attached to the basal end of the needle that passed through the extended rubber tubing of a 3 way valve (figure 1). The piston was connected to a single drive syringe pump (Razel Sci. ins., Stanford, USA) that offered a constant flow rate of 1 ml/min. The distal end of the needle passed through the rubber tubing separated at a radial distance of 0·6 mm from its inner wall. The needle was fixed at its distal end with two support stubs (see figure 1), each with dimensions of 5 mm ´  0·6 mm ´ 0·6 mm. Air was made to pass through the space between the exterior of needle and interior of the tubing. The air flowing through this chamber was ensured to be sterile after passing through a sterile 0·2 m m swinny filter. Due to the application of coaxial airflow, the cell suspension was forced to come out in the form of a spray of very fine droplets that gel to form miniature beads immediately on contact with a counter ion solution.

2.2 Islet isolation

Islets were isolated from 8 weeks old male BALB/c mice, following the protocol of Gotoh et al (1985). The animals were sacrificed by decapitation and the pancreata obtained were dissected, minced and subjected to collagenase digestion using collagenase P (Boehringer Mannheim, Germany). Islets were cultured in RPMI 1640 with 10% FCS for 48 h and then hand-picked to obtain a pure islet population for encapsulation.

2.3 Bead formation

All encapsulation procedures were carried out under strictly aseptic conditions. Islets were hand-picked and suspended in sodium alginate solution (1·2% w/v Alginic acid; Sigma Chemical Co., USA, in 0·85% saline) in a ratio of 500 islets/ml alginate solution. These were then transferred into a 5 ml syringe that was connected to the syringe pump. The pressure pump was connected to the droplet generator via a sterile swinny filter (see figure 1). After starting the pressure pump, the syringe pump was started. The resulting fine spray of alginate solution was collected in a 1 L teflon beaker containing 100 ml of the bath mixture. The bath mixture composed of 0·2% (w/v) chitosan (ICN Biomedicals, Ohio, USA) in 0·1 N acetic acid and CaCl₂ (0·15 M in distilled water). Gelling was observed as soon as the cell suspension came in contact with this counter ion bath. The beaker was kept on a magnetic stirrer with a continuous stirring speed of 200 rpm. After the alginate solution in the syringe was over the pumps were stopped. Turbidity in the crosslinking bath indicated microcapsule formation. Stirring was continued for another 30 min and beads were collected by centrifuging the bath contents at 200 rpm for 3 min. Beads were then washed in phosphate buffered saline (pH 7·2) and suspended in Roswell Park Memorial Institute (RPMI) 1640 with 10% FCS and observed under phase contrast microscope.

2.4 Islet viability

The viability of encapsulated islets, cultured in RPMI 1640 with 10% FCS for 48 h, was checked by trypan blue dye exclusion test (Warburton and James 1995), using 0·4% (w/v) trypan blue (ICN Pharmaceuticals, Inc., USA). Blue stained islets were scored as non-viable as compared to the unstained viable islets. The percentage viability was calculated after counting 100 different islets in 10 fields.

2.5 Image analysis

In order to quantify the microcapsule size, we prepared microcapsules using the aforementioned protocols and took these capsules for calculation of their diameter. Computations were done on a Kontron image analysis system (Kontron Elektronik GmbH, München, Germany) connected to a Ziess microscope. Microcapsules were observed, images captured with a Variocam, Germany PCO CCD imaging camera and quantifications of capsule diameter were made on the processed binary images using the Kontron Image analysis software version 2·04.

2.6 Islet transplantation

Functionality of the encapsulated islets was tested by transplantation of islet-capsules to diabetic mice. Male BALB/c mice were rendered diabetic with single injection of 200 mg/kg body weight of streptozotocin (STZ) (Sigma Chemicals Co., Dorseth, UK). Animals were fasted overnight and anaesthetized by intraperitoneal (i.p.) administration of thiopental sodium at a dose of 40 mg/kg body weight. Around 2 mm long incisions were made on the abdomen and microcapsules, with or without islets, were introduced. Animals in the experimental group received 400 islet equivalents (1 I.E. = an islet of 150 mm diameter) while diabetic control animals received an equivalent number of empty capsules. Incisions were sutured using absorbable 6–0 catgut sutures (Stericat Gutstrings, Delhi) and autoclipper (Becton Dickinson, Bedford, USA). All animals (control and experimental) received an i.p. injection of gentamycin (3 mg/kg body weight), Ampicillin and cloxacillin (20 mg/kg body weight) and Diclofenac sodium (0·5 mg/kg body weight), for 3 days (starting from the day of operation) in addition to the topical ointments (Soframycin).

2.7 Glucose estimations

Fasting plasma glucose of all the transplanted animals were recorded for a period of 10 days post-transplantation using a glucometre (Reflolux/S, Boehrringer Mannheim, Germany) with compatible glucose detection strips (Haemo-Glukotest 20–800 R, Boehrringer Mannheim).

2.8 Statistical analysis

Computations were performed using Jandel Scientific; Sigma-stat statistical package (SPSS Inc., Chicago, USA, version 4·04 for Windows 95). Results were expressed as mean ± SEM for normally distributed data or median and inter-quartile range (25%–75%) when the data was not normally distributed. Differences between groups were tested using Mann-Whitney test or ANOVA as appropriate.

3. Results

Using the proposed assembly, islet beads of the size of around 400 m m diameter or smaller were attained. The mean diameter of the microcapsules generated was found to be 398 ±  13 m m. In the present study, the assembly was used to encapsulate mouse pancreatic islets (figure 2b) which are around 300 m m in diameter. The size of the islet was a limiting factor to the size of the microcapsule as the microcapsule diameter cannot be smaller than the size of the islet. In order to estimate the minimum possible attainable size of the microcapsules, we estimated the diameter of the empty microcapsules (figure 2a). These empty capsules had median diameter of 78 m m (43·6–147·4), the minimum attainable diameter being 20·2 m m. The bead generator assembly used in the present study thus generated fairly uniform and spherical beads. Under the conditions mentioned above, the microcapsules encapsulated pancreatic islets in almost one-islet/bead proportion (figure 2b) without any significant run-to-run variation.

The generated microcapsules showed sufficient mechanical strength so as to withstand usual centrifugation procedures and also the hostile intra-peritoneal conditions. Chitosan-alginate capsules were transparent and provided clear microscopic examination of the encapsulated cells (figure 2b).

Islet cells encapsulated in this matrix, using this method were found to be viable (98%). Diabetic mice receiving islet capsules showed normoglycemia (fasting plasma glucose < 120 mg/dl) following 24 h of transplantation (figure 3). Control diabetic animals receiving empty chitosan-alginate capsules (figure 2a) remained diabetic (fasting plasma glucose > 200 mg/dl) throughout the period of study (figure 3). Animals which received empty capsules could not survive the uncontrolled hyperglycaemic status for more than 15 days and died later on. However animals receiving the encapsulated islet grafts survived and remained normoglycemic (fasting plasma glucose 91·2 ±  5·26 mg/dl) even 30 days after the day of transplant.

4. Discussion

Several studies till now have been carried out reporting complex droplet generator systems for cell encapsulation (Lacik et al 1998; Siebers et al 1995). Chick et al (1977) reported the first successful attempt to immunoisolate transplanted islets using a hollow fibre device. Since then, many immunoisolation devices like hollow fibres (Soldani et al 1992), chamber surrounding shunts (Galletti et al 1981) and microcapsules (Lim and Sun 1980) have been proposed amongst which microencapsulation seems to be a promising method. However, these encapsulation procedures require highly sophisticated instrument setup for carrying out the encapsulation procedures. We have attempted to simplify the entire procedure using a coaxial airflow system for the generation of microcapsules. The entire assembly is designed from commonly available accessories that can be fitted into at any laboratory and used for encapsulation of cells.

In our present study we have used pancreatic islets as a model system for trying out the encapsulation protocol. Encapsulated islets showed very high viability (98%) which was similar (P > 0·05) to that of the control/non-encapsulated islets (data not shown) suggesting that neither the bead forming protocol nor the matrix practically introduced any measurable degree of abnormality in the system. There have been several reports till now regarding the use of alginate based matrices in islet encapsulation (Soon-Shiong et al 1993). However, to the best of our understanding, only one group (Rha et al 1984) till now has reported the use of chitosan-alginate matrices for islet encapsulation. Chitosan, a cationic polysaccharide obtained by alkaline deacetylation of chitin, a principle component of exoskeleton in organisms like crustaceans (Chandy and Sharma 1990), is well documented for its anti-bacterial and anti-fungal properties (Allan and Hadwiger 1979; Stossel and Leuba 1984). Moreover chitosan has been shown to inhibit adhesion and growth of fibroblasts in culture (Malette
et al 1986). In case of type 1 diabetes mellitus, islet transplantation has been looked upon as a promising alternative to repeated insulin injection therapy. However no report till now has documented permanent reversal of diabetes by a single intraperitoneal injection of encapsulated islets. The major obstacle in this success of encapsulated islet transplants has been the issue of fibrous overgrowth associated with the capsule membrane. A lot of inconsistent results have been published on rodent model (Soon-Shiong 1994), some demonstrating long-term success, while others reporting severe fibrous outgrowth and graft malfunction within a short period following implantation. Chitosan has been very well documented to inhibit fibrotic growth in vitro (Schmidt et al 1993). Our transplantation studies also did not show any fibrotic growth on the capsules (figure 2) or any graft failure during the tenure of study. Islet microcapsules also showed proper graft functionality upon transplantation (figure 3) suggesting proper glucose sensing and insulin release by the encapsulated islets. The microcapsule membrane thus allowed the passage of low molecular weight substances (mainly glucose and insulin) through the capsular membrane but not of immunocytes and other transplant rejection effector mechanisms.

Recent human gene therapy relies on genetic modification of the patient’s own cells. An alternative to autologous approach is to use universal cell lines engineered to secrete therapeutic products. Protection with immuno-isolation devices such as these would allow the delivery of the product from the encapsulated cell line for different patients. The size of the microcapsules obtained by using the proposed assembly was around 400 mm diameter for pancreatic islets, the smallest attainable capsule size being 20 mm. This may be of great value in encapsulation of such single cells and their transplantation into host sites like the hepatic blood vessels where the size of the transplant is a critical factor.

Acknowledgements

Authors wish to thank Dr G C Mishra for supporting this work and would like to make a special mention of
Dr Ashutosh A Hardikar, Cardiothorasic Surgery Unit, KEM Hospital, Bombay, for his help. The entire work is a part of the Ph.D thesis of AAH.

References

MS received 28 December 1998; accepted 25 May 1999