Grafting of Encapsulated BDNF-Producing Fibroblasts into the Injured Spinal Cord Without Immune Suppression in Adult Rats ABSTRACT

Transcription

1 JOURNAL OF NEUROTRAUMA Volume 18, Number 3, 2001 Mary Ann Liebert, Inc. Grafting of Encapsulated BDNF-Producing Fibroblasts into the Injured Spinal Cord Without Immune Suppression in Adult Rats CHRIS A. TOBIAS, 1 NIKHIL O. DHOOT, 2 MARGARET A. WHEATLEY, 2,3 ALAN TESSLER, 1,4 MARION MURRAY, 1 and ITZHAK FISCHER 1 ABSTRACT Grafting of genetically modified cells that express therapeutic products is a promising strategy in spinal cord repair. We have previously grafted BDNF-producing fibroblasts (FB/BDNF) into injured spinal cord of adult rats, but survival of these cells requires a strict protocol of immune suppression with cyclosporin A (CsA). To develop a transplantation strategy without the detrimental effects of CsA, we studied the properties of FB/BDNF that were encapsulated in alginate-poly-l-ornithine, which possesses a semipermeable membrane that allows production and diffusion of a therapeutic product while protecting the cells from the host immune system. Our results show that encapsulated FB/BDNF, placed in culture, can survive, secrete bioactive BDNF and continue to grow for at least one month. Furthermore, encapsulated cells that have been stored in liquid nitrogen retain the ability to grow and express the transgene. Encapsulated FB/BDNF survive for at least one month after grafting into an adult rat cervical spinal cord injury site in the absence of immune suppression. Transgene expression decreased within two weeks after grafting but resumed when the cells were harvested and re-cultured, suggesting that soluble factors originating from the host immune response may contribute to the downregulation. In the presence of capsules that contained FB/BDNF, but not cell-free control capsules, there were many axons and dendrites at the grafting site. We conclude that alginate encapsulation of genetically modified cells may be an effective strategy for delivery of therapeutic products to the injured spinal cord and may provide a permissive environment for host axon growth in the absence of immune suppression. Key words: alginate; cyclosporin A; encapsulation; immune protection; neurotrophin; recombinant retrovirus; transgene expression INTRODUCTION SPINAL CORD INJURY (SCI) results in the disruption of ascending and descending axons that produces a devastating loss of motor and sensory function. Central nervous system (CNS) regeneration is thought to fail because of the lack of tropic support, the presence of inhibitory molecules in the glial scar and CNS myelin, and the loss of neurons as a result of retrograde cell death (reviewed by Murray, 2001). Several strategies have been 1 Department of Neurobiology and Anatomy, MCP Hahnemann University, 2 Department of Chemical Engineering and 3 School of Biomedical Engineering, Science and Health Systems, Drexel University, and 4 Department of Veterans Affairs Hospital, Philadelphia, Pennsylvania. 287

2 TOBIAS ET AL. FIG. 1. Schematic diagram of the properties of alginate capsules. The semipermeable alginate poly-l-ornithine membrane that encapsulates the FB/BDNF allows nutrients to diffuse into the capsule, metabolic waste products and recombinant BDNF to exit, and provides immunoisolation from host immune mediating cells. used to provide trophic, tropic, and antiapoptotic molecules that can alter the environment of the injured CNS to one that is more favorable for axon growth. These strategies have included fetal CNS transplants (Bregman et al., 1993; Himes et al., 1994), neural stem cell transplants (Liu et al., 1999b), transplants of non-neuronal cells (Tuszyinski et al., 1994; Ramon-Cueto et al., 2000; Liu et al., 1999a), and delivery of blocking antibodies to myelin inhibitory molecules (Schnell and Schwab, 1990). Some of the most promising results of these interventions have come from the use of neuronal and non-neuronal cells that have been genetically engineered (ex vivo) to secrete a neurotrophin at the injury site (Grill et al., 1997; Liu et al., 1999a). Since the genetic composition of the transplanted cells usually differs from that of the host (nonautologous), a strict immunosuppressive protocol utilizing cyclosporin A is often necessary to prevent rejection of the graft. The disadvantages of immune suppression include making the recipient vulnerable to infections, potential tumorigenesis of the transplanted cells and of the host, and possible adverse effects on regeneration and recovery. Autologous fibroblasts support regeneration of corticospinal axons after transplantation into the injured spinal cord of inbred Fischer 344 rats without immune suppression (Grill et al., 1997). For many experimental and clinical applications, however, autologous transplants are overly time consuming, expensive, and impractical. The ability to transplant nonautologous engineered cells in the absence of immune suppression would be a valuable tool for use in ex vivo gene therapy for SCI, and is best studied in outbred rat strains such as Sprague-Dawley. One approach is to encapsulate nonautologous cells in a semipermeable membrane that would allow free passage of nutrients, oxygen, and therapeutic products, but prevent the host immune response from damaging the nonautologous cells inside (Fig. 1). Alginate is a water-soluble polysaccharide extracted from brown seaweed. It is composed of alternating blocks of 1 4 linked a-l-guluronic and b-d-mannuronic acid residues. In the presence of multivalent cations, the carboxylate groups on the a-l-guluronic acid residues crosslink via the cations to form a hydrogel (Morris et al., 1973). A majority of the carboxylate groups on the surface of the gel are not involved in the crosslinking with the multivalent cations and are available for interactions with other cations, such as the polycations poly- L-lysine and poly-l-ornithine. A polycation coating can act as a size exclusion membrane that regulates the substances that pass into and out of the gel. The size exclusion characteristics of the coating can be altered by varying the coating material, its molecular weight, concentration in the coating solution, and the coating time (King et al., 1987). Alginate has been previously used for encapsulation of unmodified and genetically modified cells that deliver products for the treatment of diseases such as diabetes (Sun et al., 1996), Parkinson s disease (Winn et al., 1991), liver failure (Hirai, et al., 1993), porcine dwarfism (Cheng et al., 1998), and hemophilia B (Hortelano et al., 1996). Previous experiments from this laboratory have shown that primary fibroblasts genetically modified to produce BDNF (FB/BDNF) survive in the injured spinal cord of adult Sprague-Dawley rats immune suppressed with cyclosporin A, rescue axotomized neurons, promote regeneration, and contribute to recovery of locomotor function (Liu et al., 1999a; Y. Liu et al., unpublished observations). These grafts were rejected in the absence of immune suppression. In the present study, we examined the effects of alginate encapsulation on the properties of FB/BDNF that contribute to their usefulness as intraspinal transplants. Specifically, we studied whether encapsulated FB/BDNF would survive and continue to express the transgene in the absence of immune suppression and whether they would provide an environment that is permissive for axon growth upon grafting into the injured cervical spinal cord of the adult Sprague- Dawley rat. 288

3 GRAFTING OF ENCAPSULATED MODIFIED FIBROBLASTS MATERIALS AND METHODS Retroviral Vector Adult rat Sprague-Dawley fibroblasts have been genetically modified by a recombinant retrovirus to secrete BDNF (FB/BDNF). The retroviral construct contains the human BDNF transgene along with the reporter gene LacZ, which codes for the bacterial enzyme b-galactosidase (Fig. 2). The LTRs and IRES sequence within the retroviral construct allow for expression of the transgene (BDNF) and the reporter gene LacZ through transcription of a polycistronic mrna. The construction of the recombinant retrovirus and characterization of the modified fibroblast cells have been reported previously (Liu et al., 1999a). Preparation of Alginate Capsules FB/BDNF were cultured in 10-cm tissue culture dishes (Becton Dickinson Labware, Franklin Lakes, NJ), harvested with trypsin (Life Technologies, Grand Island, NY), washed three times with sterile 0.9% NaCl (saline), and resuspended at a concentration of cells/ml in a solution of low-viscosity sodium alginate (Keltone- LV), a gift from The Nutrasweet Kelco Company (Chicago, IL), to obtain a final alginate concentration of 1.5% (w/v). The FB/BDNF-alginate mixture was then sprayed from a 5-mL syringe (, 3.5 ml/min) into a solution of 1.3% (w/v) CaCl 2 solution through a needle in a custom-made jethead, which allowed a coaxial stream of sterile air to pass around the needle, using a Sage syringe pump (model 355; Orion Research Inc., Beverly, MA). The alginate stream from the jethead was sheared by the coaxial stream of air (, 2 slpm). Spherical microcapsules were formed due to the cross-linking of the alginate droplets by Ca 21 ions. Alginate fills the microcapsules, and the cells are embedded within the alginate matrix. The microcapsules were allowed to harden in the cross-linking solution for 1 h, after which they were washed twice with saline. The microcapsules were then coated with a 0.5 mg/ml solution of poly-l-ornithine (molecular weight 15,000 30,000; Sigma Chemicals, St. FIG. 2. Structure of the recombinant BDNF retrovirus. The retroviral LTR promoters drive expression of the human BDNF cdna and the GEO gene (fusion gene of LacZ and neomycinphosphotransferase) through the EMCV internal ribosomal entry site (IRES), which allows transcription of a polycistronic mrna. Louis, MO) by gentle shaking for 6 min. The volume of the coating solution was six times the volume of alginate used. The microcapsules were again washed with saline to remove any unreacted poly-l-ornithine followed by a 15-min treatment with 0.12% (w/v) sodium alginate. The microcapsules were washed with saline to remove unreacted alginate and transferred to 10-cm tissue culture dishes. The size of the microcapsules was approximately mm in diameter. This same encapsulation procedure was repeated in the absence of cells to provide capsules filled with alginate alone that served as controls (cell-free capsules). Capsules were maintained in culture and transplanted 48 h after preparation. Freezing and Thawing of FB/BDNF Alginate Capsules Capsules containing FB/BDNF growing in culture were pelleted by a 5-min centrifugation at 1,000 rpm. The capsules were resuspended in freezing media (DMEM 1 30% FCS 1 10% DMSO) and placed in cryotubes. The tubes were placed at 270 C for 24 h, then transferred to liquid nitrogen. Capsules were thawed by transfer from liquid nitrogen to a 37 C waterbath. Capsules were then placed in a 10-cm dish with 20 ml of prewarmed growth media (DMEM 1 10% FCS, standard growth media) and incubated at 37 C. After 4 h, the capsule medium was replaced with fresh prewarmed culture media. Viability of the cells was assessed by X-gal staining and growth in culture after thawing. Capsule membrane integrity was assessed for continuity after thawing by its appearance in culture. X-gal Staining of Capsules In Vitro After 2, 7, and 28 days in culture, capsules containing FB/BDNF were processed for X-gal histochemistry, as previously described (Liu et al., 1997). Briefly, capsules were fixed with 0.5% glutaraldehyde for 10 min in a 15- ml conical tube (Becton Dickinson Labware), washed three times with PBS, then incubated with X-gal reagent (Molecular Probes, Eugene, OR), 1 mg/ml final concentration, and X-gal mixer (35 mm K 3 Fe(CN) 6, 35 mm K 4 Fe(CN) 6? 3H 2 O, 2 mm MgCl 2 ) in PBS overnight at 37 C. After 24 h, the capsules were transferred to a 10- cm plate for examination and photographed. In a separate experiment, FB/BDNF cells were released from capsules by changing media in the culture dish to versene (0.2 g EDTA? 4Na/L in PBS; Gibco BRL, Rockville, MD) for 20 min on the fifth day in culture. Cells released from the capsules were then placed in standard growth media and processed for X-gal histochemistry 4 days after release. 289

4 TOBIAS ET AL. Bioassay of Recombinant BDNF Activity To determine if the recombinant BDNF produced by the encapsulated FB/BDNF was functional, we analyzed supernatant with an E8 chick dorsal root ganglion (DRG) explant assay. This assay has been previously described (Horie et al., 1991) and used to study FB/BDNF growing in culture (Liu et al., 1999a). Briefly, FB/BDNF growing at 50% confluency, FB/BDNF growing in capsules, and cell-free capsules from which the standard growth media had been removed, were washed three times with sterile saline, and then returned to fresh media containing 0.1% fetal calf serum. Twenty-four hours later, the conditioned media were removed and studied. At that time, 400 ml of conditioned media from either FB/BDNF capsules, cell-free capsules, FB/BDNF cells adherent in culture, or control media (DMEM 1 0.1% FCS) was added to separate wells of a 12- well plate containing the DRGs. Conditioned media were also assayed at dilutions of 1:10 and 1:100. DRG neurite outgrowth was observed after h and compared to that elicited by media containing 15, 45, or 450 ng recombinant human BDNF (Upstate Biotechnology, Lake Placid, NY). Surgical Procedures Twenty-eight female Sprague-Dawley rats weighing g (Taconic, Germantown, NY) received a partial hemisection at the C3/C4 segment of the spinal cord, as described previously (Liu et al., 1999a). Briefly, the rats were anesthetized and the C4 segment of the spinal cord was exposed by a laminectomy. The dura was cut in the midline and the dorsolateral portion of the right side of the C4 spinal cord was removed by aspiration. Either alginatepoly-l-ornithine-alginate capsules containing FB/BDNF (n 5 22) or cell-free alginate-poly-l-ornithine-alginate capsules (n 5 6) were then injected into the lesion cavity through a 16-gauge syringe needle. Between capsules were required to fill the lesion cavity. The dura was closed with interrupted 10-O sutures (Ethicon), and the muscle and skin were closed in layers with 4-O sutures (Ethicon). At the time of injury, all rats received one bolus intravenous injection of Methylprednisolone (30 mg/kg BW, i.v.; Pharmacia & Upjohn Company, Kalamazoo, MI) through the tail vein. After surgery, rats were kept on heating pads and observed until fully awake, then returned to their home cages. All procedures were carried out in accordance with a protocol approved by MCP Hahnemann University Institutional Animal Care and Utilization Committee and followed the NIH guidelines for the care and use of laboratory animals. X-gal Staining of Harvested Capsules Two (n 5 3) or four weeks (n 5 3) after implantation of capsules into a partially hemisected cervical spinal cord, rats were deeply anesthetized with Nembutal and transcardially perfused with 200 ml of saline. Capsules were harvested, washed in PBS, and then either processed for X-gal staining immediately or after growing in culture in standard growth media for 1 week. Versene (Gibco BRL, Rockville, MD) was added to some cultures to rupture the capsules and release FB/BDNF. FB/BDNF cells were processed for X-gal histochemistry 4 days after release from the capsule. Immunocytochemical Staining of Transplant Tissue Rats that received encapsulated FB/BDNF (n 5 8, each time-point) or cell-free capsules (n 5 3, each timepoint) were deeply anesthetized with Nembutal at 2 and 4 weeks postoperatively and transcardially perfused with 200 ml of saline, then 500 ml of 4% paraformaldehyde in 0.1 M phosphate buffer (PB) solution, ph 7.4. Spinal cords were dissected and the transplant region was washed with PB for 2 h, then placed in 0.1 M PB containing 30% sucrose for h at 4 C. The spinal cord tissue was then frozen with O.C.T. compound (Tissue Tek) and serially sectioned at 20 mm on a freezing microtome. Sectioned tissue was processed for Nissl staining and fluorescence immunocytochemistry as described previously (Liu et al., 1999a; Chow et al., 2000). Primary antibodies were polyclonal CGRP (1:5,000; Peninsula Laboratories, Inc., Belmont, CA), monoclonal RT-97 (1:100; Boehringer Mannheim, GmbH, Germany), polyclonal MAP-1B (1:40,000) (Black et al., 1994), polyclonal MAP-2 (1:10,000) (Fischer et al., 1991), and monoclonal GFAP (1:100) (Boehringer Mannheim). Fluorescent secondary antibodies (diluted 1:200) were either FITC or rhodamine conjugated goat-anti-mouse (Jackson ImmunoResearch, West Grove, PA) for monoclonal antibodies or FITC or rhodamine conjugated goat-anti-rabbit (Jackson ImmunoResearch) for polyclonal antibodies. Slides were then coverslipped with Vectashield (Vector, San Diego, CA) containing the nuclear counterstain DAPI. Image Analysis Stained sections were examined under a Leica DMRBE microscope, and images were captured using a Photometric Sensys KAF-1400 CCD camera (Photometric Inc.). Images were processed on a Macintosh Power PC 8500 with IP Lab (Scanalytics) and NIH image analysis software packages. To quantitate RT-97 immunostaining, we first outlined a square field containing an area of mm 2 within the central portions of the grafted region with the highest amount of staining in animals with survivals of 2 and 4 weeks. A micrometer at 290

5 GRAFTING OF ENCAPSULATED MODIFIED FIBROBLASTS the same magnification was used to convert the number of pixels to microns, and RT-97 immunostaining was traced within the outlined field providing a measure of length to the amount of staining present in a section. RESULTS In Vitro Properties of Encapsulated FB/BDNF FB/BDNF were encapsulated at a density of cells/ml of alginate. This concentration provided a number of cells/capsule (approximately 300 cells/300 mm diameter capsule) that allowed room for the cells to proliferate. Most capsules formed complete spheres with continuous walled membranes; the few that did not were removed from the culture dish. The FB/BDNF predominantly appeared as individual cells immediately after encapsulation (Fig. 3A), but later began to form spheroid groups of cells which have previously been described for fibroblasts growing in alginate capsules (Chang et al., 1994). These spheroid aggregates could be clearly seen by day 5 postencapsulation (Fig. 3B). The FB/BDNF continued to divide within the capsules as indicated by increases in number and size of the spheres over a two week period (Fig. 3C). To examine the growth properties of encapsulated FB/BDNF, on day 5 after encapsulation one set of capsules was placed in a solution of EDTA to disrupt the alginate gel and release the spheroid aggregates. After release, the spheres attached to the tissue culture plate (Fig. 3D) and within 2 days began to grow (Fig. 3E). To examine transgene expression, the released spheres were stained for b-galactosidase 4 days after release from the capsules. Numerous b-gal positive cells were present throughout the attached sphere (Fig. 3F) and in adjacent fields where the cells had migrated away from the spheres (Fig. 3G). b-galactosidase staining was very similar in cells that had been released from encapsulation and in FB/BDNF that had not been encapsulated (data not shown). FIG. 3. Encapsulated FB/BDNF growth and reporter gene expression after release in vitro. (A E) Phase contrast. (F,G) Brightfield. Encapsulated FB/BDNF on the day of encapsulating (A), after 5 days (B), and after 12 days (C) in vitro. Note the increased number and size of FB/BDNF spheres at 5 and 12 days. FB/BDNF released after 5 days of encapsulation remained as spheres immediately after release (D). FB/BDNF sphere attached to surface of plate two days after release (day 7 after encapsulation) (E), b-galactosidase (reporter gene) staining of attached sphere four days after release (9 days after encapsulation) (F), and b- galactosidase staining of cells that had migrated away from the spheres (G), shown in F on the same day. Bar mm. 291

6 TOBIAS ET AL. FIG. 4. FB/BDNF reporter gene expression of encapsulated cells before and after storage in liquid nitrogen. b-galactosidase staining of encapsulated FB/BDNF after 48 h (A), after 7 days (B), and after 4 weeks (C) in culture. Note the increase number and size of the b-galactosidase stained spheres through 1 month in vitro. b-galactosidase staining of encapsulated FB/BDNF 48 h after thawing from liquid nitrogen (D) and after an additional 3 weeks in culture (E). Note that the appearance of the thawed capsules is indistinguishable from freshly prepared capsules with respect to capsule integrity and level of b-galactosidase staining. Bar mm. To determine the effect of encapsulation on transgene expression, capsules containing FB/BDNF were stained for b-gal expression 2 days (Fig. 4A), 14 days (Fig. 4B), and 28 days (Fig. 4C) after encapsulation. As the time in culture increased, the number and/or size of the b- gal stained spheres also increased, indicating that the encapsulated FB/BDNF-like unencapsulated FB/BDNF continued to divide and grow. During the 4 weeks in culture, some free FB/BDNF were observed attached to the bottom of the tissue culture dish, probably from the small number of broken capsules. To test the viability of encapsulated FB/BDNF after cryopreservation, capsules were frozen in media containing DMSO and stored in liquid nitrogen. After thawing, the growth characteristics of encapsulated FB/BDNF appeared unchanged. Sphere size increased within capsules during 3 weeks in culture, and the cells expressed the b-galactosidase reporter gene after 48 h (Fig. 4D) and for at least 3 weeks (Fig. 4E). In Vitro Bioactivity of BDNF from Encapsulated FB/BDNF To determine if BDNF released from the semipermeable alginate capsule retained bioactivity, an E8 chick DRG assay was performed on unconditioned media, and conditioned media from cell-free capsules, encapsulated FB/BDNF, unencapsulated FB/BDNF, and media containing commercially available recombinant human BDNF. Unconditioned media (Fig. 5A) and conditioned media from cell-free capsules (Fig. 5C) failed to induce neurite outgrowth in this assay. In contrast, conditioned media from encapsulated FB/BDNF (Fig. 5D), and unencapsulated FB/BDNF (Fig. 5E) induced neurite outgrowth, as did media containing recombinant human BDNF, the positive control (Fig. 5B). Neurite outgrowth is indicated by processes extending from the DRG (Fig. 5B,D,E) and not the nonspecific Schwann cell proliferation that is present in Figure 5A,C. These results indicate 292

7 GRAFTING OF ENCAPSULATED MODIFIED FIBROBLASTS FIG. 5. Analysis of recombinant BDNF bioactivity. Analysis of biologically active BDNF in an E8 chick DRG explant assay was carried out with unconditioned media, used as the negative control (A), recombinant human BDNF, used as the positive control (B), cultures containing cell-free capsules (C), encapsulated FB/BDNF (D), and unencapsulated FB/BDNF (E). Encapsulated FB/BDNF secreted biologically active BDNF, as did unencapsulated FB/BDNF, and elicited neurite outgrowth (arrows), while the unconditioned media (A) and the cell-free capsule media (C) did not. All figures at 3100 magnification. that BDNF secreted by the encapsulated FB/BDNF exits the semipermeable membrane and remains bioactive. This assay s results were similar to those that we obtained previously for unencapsulated FB/BDNF in which the production of secreted recombinant BDNF was approximately 24 ng/10 6 cells/24 h (Liu et al., 1999a). In Vitro Analysis of FB/BDNF Capsules Harvested from a Transplant To determine whether the semipermeable alginatepoly-l-ornithine membrane protected encapsulated cells from the cytotoxic effects of the host immune system, we examined the survival of the encapsulated cells and expression of the reporter gene (b-galactosidase) after transplantation into the injured spinal cord of a host that had not been immune suppressed. While rats in this study received one bolus dose of methylprednisolone to minimize the inflammatory response due to the injury, none of the animals were subjected to chronic immune suppression. Unencapsulated fibroblasts are rejected in the absence of immune suppression (Liu et al., 1998). Capsules containing FB/BDNF were transplanted into the lesion site of a partially hemisected cervical spinal cord of adult Sprague Dawley rats and harvested 2 weeks or 4 weeks after grafting. At 2 weeks, the spheres within the capsules had increased in size and number compared to those initially grafted. b-galactosidase reporter gene expression was evident (Fig. 6A), but decreased from that observed just prior to transplantation (Fig. 4A). At 4 weeks, the harvested capsules contained large spheres within intact capsules (Fig. 6D). The encapsulated FB/BDNF continued to stain weakly for b-galactosidase (Fig. 6C). Capsules harvested at 2 weeks (Fig. 6B) and at 4 weeks (Fig. 6E) after transplantation and returned to culture for an additional week contained cells that showed a strong increase in b-galactosidase staining compared to that 293

8 TOBIAS ET AL. FIG. 6. Analysis of reporter gene expression of encapsulated FB/BDNF after grafting. Encapsulated FB/BDNF were grafted into an injured spinal cord in the absence of immune suppression. b-galactosidase staining of capsule harvested at 2 weeks (A) and following 1 week in culture (B). Phase contrast after harvest at 4 weeks (D), b-galactosidase staining of capsule harvested at 4 weeks (C), and following 1 week in culture (E). Reporter gene expression in encapsulated FB/BDNF was decreased after grafting of capsules in vivo (A,C), but downregulation was reversible after 1 week in culture (B,E). (F) b-galactosidase staining of FB/BDNF that were harvested 4 weeks after grafting and released from capsules in culture. Note the similarities in appearance and reporter gene expression between these cells that were released and encapsulated FB/BDNF grown only in vitro and released (Fig. 3G). Bar mm (A F). seen when the cells were first dissected from the spinal cord (Fig. 6A,C). These results suggest that down-regulation of a transgene in vivo is due to soluble factors since the encapsulated FB/BDNF had no direct contact with the host spinal cord. The results also show that, in this experimental paradigm, the down-regulation of transgene expression is reversible even when cells have remained in the host spinal cord for 1 month. To test the growth and reporter gene expression of the encapsulated FB/BDNF after 1 month in vivo, capsules were harvested, placed in culture, and the FB/BDNF were released from capsules. The growth and b-galactosidase staining of these cells (Fig. 6F) were similar to those of FB/BDNF that had not been encapsulated. These results suggest that encapsulated cells that have survived in damaged host spinal cord for up to 4 weeks are protected from permanently harmful interactions with the host. 294

9 GRAFTING OF ENCAPSULATED MODIFIED FIBROBLASTS Histology of Transplant Tissue FIG. 7. Photomicrographs of encapsulated FB/BDNF transplanted into injured spinal cord. Nissl-stained parasagittal sections, 4-week survival. (A) Capsules contain FB/BDNF growing as spheres (arrow) and sheets (arrow heads); the ribbon-like substance ( * ) is alginate torn during tissue processing. (B) Capsules are intact and contain FB/BDNF growing in sheets on their inner walls (arrow heads). (C) Higher power view of an adjacent section showing FB/BDNF growing as spheres within the capsules (arrows). Bar mm (A), 100 mm (B,C). Capsule integrity and FB/BDNF survival within the capsules were examined at 2 weeks and 4 weeks after transplantation into injured spinal cord. Capsules were abundant (Fig. 7A) and surrounded by a matrix of nonneuronal cells that filled the lesion site completely and were particularly numerous immediately around the outer walls of the capsules in the graft area. The capsule walls in cross section were continuous, which suggested that they were intact (Fig. 7A,C). b-galactosidase histochemistry performed on spinal cord tissue showed no staining outside capsules containing FB/BDNF (data not shown). Encapsulated FB/BDNF grew as spheroid structures within the capsules (Fig. 7C) and as sheets attached to the inner walls (Fig. 7B). The encapsulation procedure thus protected these cells from rejection. These results resemble previous descriptions of genetically modified fibroblasts growing in capsules in the peritoneum of mice (Chang et al., 1993). FB/BDNF appear to fill the capsules less completely than they fill capsules harvested and placed in culture (Fig. 6C,D), but this difference is probably due to the loss of alginate (in which centrally located FB/BDNF are embedded) during tissue processing (Fig. 7A). Immunohistochemical Analysis of the Host Interaction With the Transplanted Capsules To determine if the transplanted alginate capsules elicited an astrocyte reaction at the lesion site, we stained the spinal cords with an antibody against GFAP. We observed no consistent difference in GFAP staining between spinal cords containing cell-free capsules (Fig. 8A) and those containing FB/BDNF capsules (Fig. 8B) at 4 weeks after grafting. Astrocytes were aggregated at the lesion borders and sometimes around the outer wall of capsules at the lesion margin, but not associated with capsules in the central grafted area (Fig. 8A,B). Overall, the GFAP staining of the host cord resembled that seen after an injury to the adult rat spinal cord (Houle and Reier, 1988; Predy et al., 1988). To assay the permissiveness for axonal growth into the environment that contained encapsulated FB/BDNF, we stained for the presence of host axons using the RT-97 and MAP-1B antibodies, and for the presence of host dendrites using an antibody against MAP-2. At 2 weeks, a small amount of RT-97 staining was present in spinal cord tissue immediately adjacent to capsules at the margins of the injury (Fig. 9A,C). More abundant staining was present in this location at 4 weeks, and staining extended to capsules located deeper within the damaged spinal cord (Fig. 9B,D). Measurements of the lengths of axons immunoreactive for RT-97 within the central grafted region showed an 8.2-fold increase between 2 and 4 weeks, from 244 to 1,999 mm. RT-97 staining was present in the marginal tissue but not in the center of the region that contained cell-free capsules (data not shown). MAP-1B staining was not found in the transplant region until 4 weeks, when it was restricted to the area around the capsules containing FB/BDNF at the graft host in- 295

10 TOBIAS ET AL. DISCUSSION FIG. 8. Darkfield photomicrographs showing host astrocyte response to alginate capsules. Parasagittal sections immunostained for GFAP and counterstained with DAPI, 4-week survival. Intraspinal grafts consisted of cell-free control capsules (A) or FB/BDNF capsules (B). After both types of grafts, dense GFAP staining appears at the margins of the injury site (arrowheads) and around the occasional capsule found at the lesion periphery ( * ), whereas little staining appears in the central grafted area. Note that the staining at the margin of tissue containing FB/BDNF capsules (B) is not greater than that for cellfree capsules (A), indicated by arrow heads. Arrows indicate lateral margins of spinal cord tissue, H indicates host tissue, and G indicates the grafted region. Bar mm. terface (Fig. 9E). MAP-1B staining did not appear adjacent to cell-free capsules (data not shown). MAP-2 staining surrounded FB/BDNF containing capsules at the margin of the grafted region at 4 weeks but not at 2 weeks (Fig. 9F). We did not observe MAP-2 staining at either 2 or 4 weeks in the spinal cord tissue that received cellfree capsules (data not shown). These results show that host axons and dendrites can grow into the transplant region by 4 weeks posttransplantation and therefore suggest that FB/BDNF encapsulated in alginate provide a terrain permissive for neurite growth. The present investigation has shown that fibroblasts genetically modified to express BDNF (FB/BDNF) can be encapsulated in an alginate matrix where they continue to grow and express a functional transgene. Primary cells (Lim and Sun, 1980; Sun et al., 1986; Fu and Sun, 1989) and the PC12 cell line (Winn et al., 1991) have been previously reported to survive and proliferate in culture after encapsulation in alginate. Genetically modified fibroblasts encapsulated in alginate have also been observed to secrete growth factors (Chang et al., 1993; Cheng et al., 1998; Maysinger et al., 1993, 1996) and other peptides (Hughes et al., 1994; Liu et al., 1993) that were biologically active in vitro and in vivo. We also found that these encapsulated FB/BDNF continued to grow, express the transgene and maintain capsule integrity after they had been frozen, stored in liquid nitrogen and thawed. Similar results have been reported for porcine pancreatic islet cells (Zhou et al., 1997), rat hepatocytes (Guyomard et al., 1996), and human kidney 293 cells (Read et al., 1999), but these were primary cell lines, or cell lines that were not genetically modified to secrete a transgene. We believe that our results provide the first example in which a genetically modified primary cell line that had been encapsulated in alginate maintained capsule integrity and transgene expression after freezing and storage. The ability to freeze a highly characterized genetically modified primary cell line encapsulated in an immunologically resistant semipermeable matrix should greatly increase the clinical and experimental applications of this technology. After grafting into the injured adult rat spinal cord without immune suppression nonencapsulated modified fibroblasts survive poorly and the survivors show decreased transgene expression (Liu et al., 1998; B.T. Himes et al., unpublished observations). The present set of in vivo experiments show that the encapsulated FB/BDNF survive transplantation into the adult injured host spinal cord for at least 1 month (longest time-point examined) in the absence of immune suppression. These results are consistent with earlier reports that genetically modified cells encapsulated in alginate survived implantation in the peritoneum for long periods of time (Cheng et al., 1998; Al-Hendy et al., 1995), and that alginate encapsulated cells survived grafting into the lateral ventricle of adult mice (Ross et al., 1999), the cerebral cortex of BD-IX rats (Read et al., 1999), and the devascularized cerebral cortex of adult rats (Maysinger et al., 1994). Chromaffin cells encapsulated in an acrylic-based tubular semipermeable membrane have also produced prolonged reduction of pain after grafting into the lumbar subarachnoid space of sheep (Joseph et al., 1994), and 296

11 GRAFTING OF ENCAPSULATED MODIFIED FIBROBLASTS FIG. 9. Darkfield photomicrographs showing host neuron response to intraspinal transplant of encapsulated FB/BDNF. Parasagittal sections counterstained with DAPI. (A,C) Two weeks after grafting axons immunoreactive for RT-97 are rare in the central portion of the graft. (B,D) Four weeks postgrafting central staining for RT-97 has increased. C and D show the same fields as A and B with the addition of dashed lines to indicate the graft host interface and the symbols H and G to define the host and grafted area. At 4 weeks, axons immunoreactive for MAP-1B (E) and dendrites immunoreactive for MAP-2 (F) are closely associated with the outer capsule wall at the graft host interface. The symbol c identifies capsules with immunostaining on their outer walls. Bar mm (A D), 100 mm (E,F). CNTF-modified myoblasts encapsulated in alginate survived up to 1 month in the lumbar subarachnoid space of rats (Deglon et al., 1996). To our knowledge, the current study provides the first report that alginate-encapsulated cells survived after implantation into the injured spinal cord. We found that encapsulated FB/BDNF down-regulated expression of the b-galactosidase reporter gene by 2 weeks after implantation. This down-regulation was reversible because the cells stained strongly for the reporter gene after the capsules were harvested and placed in tissue culture for an additional week. One explanation for the apparent increase in reporter gene expression after 1 week in culture is that the FB/BDNF proliferated inside the capsules, producing new cells that expressed the transgene. This interpretation is probably incorrect because the FB/BDNF largely filled the capsules while in vivo (Fig. 6D) and the density of cells did not appear to increase while in vitro (Fig. 6E). These results suggest that downregulation in vivo is due to soluble factors, since encapsulated FB/BDNF had no cell-to-cell contact with the host tissue. Because host animals were not immune suppressed, these factors are likely to be inflammatory cytokines secreted by activated macrophages, microglia, or astrocytes, which are present in the injury area as early as a few days after grafting (reviewed by Schwab and 297

12 TOBIAS ET AL. Bartholdi, 1996). Cytokines would have access to encapsulated FB/BDNF because the semipermeable membrane allows free diffusion of substances smaller than 70 kd (King et al., 1987). Cytokines including interferon-g (Ghazizadeh et al., 1997), transforming growth factor-b, interleukin-1b, and tumor necrosis factor-a have been shown to decrease transcriptional expression of retroviral long terminal repeat (LTR) promoters in transgene viral constructs in genetically modified fibroblasts in vitro (Shinstine et al., 1997). In these studies, down-regulation was reversible upon removal of the cytokine and replacement with standard growth media. We administered methylprednisolone to reduce the inflammation that occurs in and around the injury site, but we did not immune suppress the animals. When unencapsulated FB/BDNF are grafted into the adult rat injured spinal cord in the presence of cyclosporin A, transgene expression is down-regulated considerably after 1 month (Liu et al., 1999a). These results suggest either that cytokines causing the retroviral transgene down-regulation are expressed despite immune suppression or that other factors are responsible for this regulation. However, expression of a neurotrophin transgene for 2 weeks to 1 month after grafting may be appropriate for treatment of spinal cord injury, since it may be adequate for regeneration (Liu et al., 1999a) while avoiding long-term exposure to neurotrophins that can have adverse side effects such as pain (Thompson et al., 1999) and thermal hyperalgesia (Shu and Mendell, 1999). Our interpretation of these results is that cytokines contribute to retroviral transgene down-regulation in vivo, but that other factors also play a role. An important point is that encapsulation allows the grafted cells to be identified and their transgene expression determined separately from that of host cells, which would be indistinguishable from harvested cells in the absence of encapsulation. Encapsulation also allows us to distinguish down-regulation of the transgene from loss of transgene due to cell death, which could be misinterpreted as down-regulation. Both the encapsulated FB/BDNF and cell-free capsules elicited a moderate host GFAP reaction similar to that reported previously in response to injuries to the adult rat spinal cord. Astrocytes identified by GFAP immunostaining were localized to the outer edges of the injury containing either cell-free capsules or FB/BDNF capsules. Occasionally, GFAP immunostaining surrounded capsules at the lesion border, but not associated with capsules in the central grafted region. Similar GFAP staining has also been associated with alginate-encapsulated PC12 cells implanted into the striatum of adult rats (Winn et al., 1991). Alginate-encapsulated FB/BDNF appeared to provide a permissive terrain for axonal and dendritic growth as indicated by immunostaining for RT-97 and MAP-2 found 4 weeks after injury in the spinal cord tissue. Quantitative analysis of the RT-97 staining showed an 8.2-fold increase in the total length of centrally located immunostaining between 2 weeks and 4 weeks after grafting. The source of the axons within the grafts is unknown. The axons could be regenerating axons or sprouting intact axons derived from ascending or descending spinal pathways or from dorsal roots. Calcitonin gene related peptide (CGRP) immunostaining was found within some grafts containing encapsulated FB/BDNF capsules, but the staining was mostly observed in the periphery of the injury site in contrast to centrally located RT-97 staining (data not shown). Because CGRP is a marker for small and medium sized dorsal roots (Gibson et al., 1984; McCarthy et al., 1990), most of the axons that penetrate into the central portion of the graft were likely to be of central rather than dorsal root origin. Previous reports have shown axon penetration into freeze dried alginate grafts in transected cat peripheral nerves (Suzuki et al., 1999b) and transected rat spinal cord (Suzuki et al., 1999a). However, we found axons and dendrites surrounding the outer wall of the capsule, but at no time observed immunoreactivity for RT-97, MAP-2, or MAP-1B inside the capsules. The most likely explanation for the differences between these results and those of Suzuki et al. is that the latter investigators used freeze-dried alginate alone as the graft, whereas we used alginate capsules that were coated with poly-l -ornithine. Poly-L -ornithine provided the capsules with structural strength and immune protection but is likely also to have prevented axons from penetrating the capsule membrane. An interesting property of these alginate capsules is that they can be coated with various substances. For instance, in the same way that we used poly-l -ornithine to provide structural strength and immune protection properties to the capsule, an extracellular matrix protein that is known to be permissive for axonal growth (e.g., laminin and/or L1) could be applied to make the environment of the injured spinal cord more supportive of growth. This coating capacity allows the rapid examination of the growthsupporting properties of extracellular matrix proteins in a spinal cord injury without the engineering of a cell line. These substances can be tested in combination with a neurotrophin (FB/BDNF) that has been shown to provide trophic support for long distance regeneration (Liu et al., 1999a) and cell survival (Y. Liu et al., unpublished observations; Diener and Bregman, 1994) after spinal cord injury. Alginate oligosaccharides have also been shown to stimulate growth of human endothelial cells in vitro (Kawada et al., 1999) and have been hypothesized to serve as a matrix for aggregation of platelets and erythrocytes by containing calcium ions (Suzuki et al., 1998). Alginate 298

13 GRAFTING OF ENCAPSULATED MODIFIED FIBROBLASTS may therefore be useful for hemostasis and vascularization after trauma. These properties, together with its capacity of immunoisolation, neurotrophin delivery, and extracellular matrix coating, which may be used together to maximize regeneration and survival of the host tissue, make alginate encapsulation a very promising strategy in repair of spinal cord injury. DEGLON, N., HEYD, B., TAN, S.A., et al. (1996). Central nervous system delivery of recombinant ciliary neurotrophic factor by polymer encapsulated differentiated C2C12 myoblasts. Hum. Gene Ther. 7, DIENER, P.S., and BREGMAN, B.S. (1994). Neurotrophic factors prevent the death of CNS neurons after spinal cord lesions in newborn rats. Neuroreport 5, FISCHER, I., RICHTER-LANDSBERG, C., and SAFAEI, R. (1991). Regulation of microtubule associated protein 2 (MAP2) expression by nerve growth factor in PC12 cells. Exp. Cell Res. 194, ACKNOWLEDGMENTS This work was supported by the National Institutes of Health (grant 24707), the Christopher Reeve Research Foundation, the Veterans Administration, the Eastern Paralyzed Veterans Administration, the Spinal Cord Center of MCP Hahnemann University, and the Lesley Boyne Research Fellowship awarded to C.A.T. We thank Dr. L. Lillien for the generous gift of the LIG retrovirus vector. We thank Drs. S.Y. Chow, Y. Liu, and B.T. Himes, and Ms. T. Conners and L. Hodge for their invaluable advice and technical help. REFERENCES AL-HENDY, A., HORTELANO, G., TANNENBAUM, G.S., et al. (1995). Correction of growth defect in dwarf mice with nonautologous microencapsulated myoblasts an alternate approach to somatic gene therapy. Hum. Gene Ther. 6, BLACK, M.M., SLAUGHTER, T., and FISCHER, I. (1994). Microtubule-associated protein 1b (MAP1b) is concentrated in the distal region of growing axons. J. Neurosci. 14, BREGMAN, B.S., KUNKEL-BAGDEN, E., REIER, P.J., et al. (1993). Recovery of function after spinal cord injury: mechanisms underlying transplant-mediated recovery of function differ after spinal cord injury in newborn and adult rats. Exp. Neurol. 123, CHANG, P.L., SHEN, N., and WESTCOTT, A.J. (1993). Delivery of recombinant gene products with microencapsulated cells in vitro. Hum. Gene Ther. 4, CHANG, P.L., HORTELANO, G., TSE, M., et al. (1994). Growth of recombinant fibroblasts in alginate microcapsules. Biotechnol. Bioeng. 43, CHENG, W.T.K., CHEN, B.-C., CHIOU, S.-T., et al. (1998). Use of nonautologous microencapsulated fibroblasts in growth hormone gene therapy to improve growth of midget swine. Hum. Gene Ther. 9, CHOW, S.Y., MOUL, J., TOBIAS, C.A., et al. (2000). Characterization and intraspinal grafting of EGF/bFGF-dependent neurospheres derived from embryonic rat spinal cord. Brain Res. 874, FU, X.W., and SUN, A.M. (1989). Microencapsulated parathyroid cells as a bioartificial parathyroid. In vivo studies. Transplantation 47, GHAZIZADEH, S., CARROLL, J.M., and TAICHMAN, L.B. (1997). Repression of retrovirus-mediated transgene expression by interferons: implications for gene therapy. J. Virol. 71, GIBSON, S.J., POLAK, J.M., BLOOM, S.R., et al. (1984). Calcitonin gene related peptide immunoreactivity in the spinal cord of man and eight other species. J. Neurosci. 4, GRILL, R., MURAI, K., BLESCH, A., et al. (1997). Cellular delivery of neurotrophin-3 promotes corticospinal axonal growth and partial functional recovery after spinal cord injury. J. Neurosci. 17, GUYOMARD, C., RIALLAND, L., FREMOND, B., et al. (1996). Influence of alginate gel entrapment and cryopreservation on survival and xenobiotic metabolism capacity of rat hepatocytes. Toxicol. Appl. Pharmacol. 141, HIMES, B.T., GOLDBERGER, M.E., and TESSLER, A. (1994). Grafts of fetal central nervous system tissue rescue axotomized Clarke s nucleus neurons in adult and neonatal operates. J. Comp. Neurol. 339, HIRAI, S., KASAI, S., and MITO, M. (1993). Encapsulated hepatocyte transplantation for the treatment of D -galactosamine induced acute hepatic failure in rats. Eur. Surg. Res. 25, HORIE, H., BANDO, Y., CHI, H., et al. (1991). NGF enhances neurite regeneration from nerve-transected terminals of young adult aged mouse dorsal root ganglia in vitro. Neurosci. Lett. 121, HORTELANO, G., AL-HENDY, A., OFOSU, F.A., et al. (1996). Delivery of human factor IX in mice by encapsulated recombinant myoblasts: a novel approach towards allogeneic gene therapy of hemophilia B. Blood 87, HOULE, J.D., and REIER, P.J. (1988). Transplantation of fetal spinal cord tissue into the chronically injured adult rat spinal cord. J. Comp. Neurol. 269, HUGHES, M., VASSILAKOS, A., ANDREWS, D.W., et al. (1994). Delivery of a secretable adenosine deaminase through 299

14 TOBIAS ET AL. microcapsules a novel approach to somatic gene therapy. Hum. Gene Ther. 5, JOSEPH, J.M., GODDARD, M.B., MILLS, J., et al. (1994). Transplantation of encapsulated bovine chromaffin cells in the sheep subarachnoid space: a preclinical trial for the treatment of cancer pain. Cell Transplant. 3, KAWADA, A., HIURA, N., TAJIMA, S., et al. (1999). Alginate oligosaccharides stimulate VEGF-mediated growth and migration of human endothelial cells. Arch. Dermatol. Res. 291, KING, G.A., DAUGULIS, A.J., FAULKNER, P., et al. (1987). Alginate-polylysine microcapsules of controlled membrane molecular weight cutoff for mammalian cell culture engineering. Biotechnol. Prog. 3, LIM, F., and SUN, A.M. (1980). Microencapsulated islets as bioartificial endocrine pancrease. Science 210, LIU, H.-W., OFOSU, F.A., and CHANG, P.L. (1993). Expression of human factor IX by microencapsulated recombinant fibroblasts. Hum. Gene Ther. 4, LIU, Y., HIMES, B.T., MOUL, J., et al. (1997). Application of recombinant adenovirus for in vivo gene delivery to spinal cord. Brain Res. 768, LIU, Y., HIMES, B.T., TRYON, B., et al. (1998). Intraspinal grafting of fibroblasts genetically modified by recombinant adenoviruses. Neuroreport 9, LIU, Y., KIM, D., HIMES, B.T., et al. (1999a). Transplants of fibroblasts genetically modified to express BDNF promote regeneration of adult rat rubrospinal axons and recovery of forelimb function. J. Neurosci. 19, LIU, Y., HIMES, B.T., SOLOWSKA, J., et al. (1999b). Intraspinal delivery of neurotrophin-3 using neural stem cells genetically modified by recombinant retrovirus. Exp. Neurol. 158, MAYSINGER, D., PICCARDO, P., FILIPOVIC-GRCIC, J., et al. (1993). Microencapsulation of genetically engineered fibroblasts secreting nerve growth factor. Neurochem. Int. 23, MAYSINGER, D., PICCARDO, P., LIBERINI, P., et al. (1994). Encapsulated genetically engineered fibroblasts: release of nerve growth factor and effects in vivo on recovery of cholinergic markers after devascularizing cortical lesions. Neurochem. Int. 24, terization of polysaccharide structure and interactions by circular dichroism. Order-disorder transition in the calcium alginate system. Chem. Commun MURRAY, M. (2001). Therapies to promote CNS repair, in: Axonal Regeneration in the Central Nervous System. N. Ingoglia, and M. Murray (eds), Marcel Dekker: New York, pps PREDY, R., MALHOTRA, S.K., and DAS, G.D. (1988). Enhanced expression of a protein antigen (JI-31 antigen, 30 kilodaltons) by reactive astrocytes in lacerated spinal cord. J. Neurosci. Res. 19, RAMON-CUETO, A., CORDERO, M.I., SANTOS-BENITO, F.F., et al. (2000). Functional recovery of paraplegic rats and motor axon regeneration in their spinal cords by olfactory ensheathing glia. 25, READ, T.-A., STENSVAAG, V., VINDENES, H., et al. (1999). Cells encapsulated in alginate: a potential system for delivery of recombinant proteins to malignant brain tumours. Int. J. Dev. Neurosci. 17, ROSS, C.J.D., RALPH, M., and CHANG, P.L. (1999). Delivery of recombinant gene products to the central nervous system with nonautologous cells in alginate microcapsules. Hum. Gene Ther. 10, SCHINSTINE, M., RAY, J., and GAGE, F.H. (1997). Potential effect of cytokines on transgene expression in primary fibroblasts implanted into the rat brain. Mol. Brain Res. 47, SCHNELL, L., and SCHWAB, M.E. (1990). Axonal regeneration in the rat spinal cord produced by an antibody against myelin-associated neurite growth inhibitors. Nature 343, SCHWAB, M.E., and BARTHOLDI, D. (1996). Degeneration and regeneration of axons in the lesioned spinal cord. Physiol. Rev. 76, SHU, X.-Q., and MENDELL, L.M. (1999). Neurotrophins and hyperalgesia. Proc. Natl. Acad. Sci. U.S.A. 96, SUN, A.M., CAI, Z., SHI, Z., et al. (1986). Microencapsulated hepatocytes as a bioartificial liver. Trans. Am. Soc. Artif. Organs 32, SUN, Y., MA, X., ZHOU, D., et al. (1996). Normalization of diabetes in spontaneously diabetic cynomologous monkeys by xenografts of microencapsulated porcine islets without immunosuppression. J. Clin. Invest. 98, MAYSINGER, D., KRIEGLSTEIN, K., FILIPOVIC-GRCIC, J., et al. (1996). Microencapsulated ciliary neurotrophic factor: physical properties and biological activities. Exp. Neurol. 138, SUZUKI, Y., NISHIMURA, Y., TANIHARA, M., et al. (1998). Evaluation of a novel alginate gel dressing: cytotoxicity to fibroblasts in vitro and foreign body reaction in pig skin in vivo. J. Biomed. Mater. Res. 39, MC CARTHY, P.W., and LAWSON, S.N. (1990). Cell type and conduction velocity of rat primary sensory neurons with calcitonin gene-related peptide-like immunoreactivity. Neuroscience 34, SUZUKI, K., SUZUKI, Y., OHNISHI, K., et al. (1999a). Regeneration of transected spinal cord in young adult rats using freeze-dried alginate gel. Neuroreport 10, MORRIS, E.R., REES, D.A., and THOM, D. (1973). Charac- SUZUKI, Y., TANIHARA, M., OHNISHI, K., et al. (1999b). Cat peripheral nerve regeneration across 50-mm gap repaired 300