Online Issues

ILAR Journal Vol 45(3)

Piscine Islet Xenotransplantatio

James R. Wright, Jr., Bill Pohajdak, Bao-You Xu, and Joseph R. Leventhal

James R. Wright, Jr., M.D., Ph.D., is a Professor in the Departments of Pathology, Surgery, and Biomedical Engineering, Dalhousie University Faculty of Medicine and IWK Health Centre, Halifax, Nova Scotia, Canada. Bill Pohajdak, Ph.D., is a Professor in the Department of Biology, Dalhousie University, Halifax. Bao-You Xu, M.D., Ph.D., is a Postdoctoral Fellow in the Department of Pathology, Dalhousie University Faculty of Medicine and IWK Health Centre, Halifax. Joseph R. Leventhal, M.D., Ph.D., is an Assistant Professor in the Division of Transplantation, Department of Surgery, Northwestern University Medical Center, Chicago, Illinois.

Abstract

Tilapia, a teleost fish species with large anatomically discrete islet organs (Brockmann bodies; BBs) that can be easily harvested without expensive and fickle islet isolation procedures, make an excellent donor species for experimental islet xenotransplantation research. When transplanted into streptozotocin-diabetic nude or severe combined immunodeficient mice, BBs provide long-term normoglycemia and mammalian-like glucose tolerance profiles. However, when transplanted into euthymic recipients, the mechanism of islet xenograft rejection appears very similar to that of islets from "large animal" donor species such as the very popular fetal/neonatal porcine islet cell clusters (ICCs). Tilapia islets are more versatile than ICCs and can be transplanted (1) into the renal subcapsular space, the cryptorchid or noncryptorchid testis, or intraportally as neovascularized cell transplants; (2) as directly vascularized organ transplants; or (3) intraperitoneally after microencapsulation. Unlike the popular porcine ICCs, BBs function immediately after transplantation; thus, their rejection can be assessed on the basis of loss of function as well as other parameters. We have also shown that transplantation of tilapia BBs into nude mice can be used to study the possible implications of cross-species physiological incompatibilities in xenotransplantation. Unfortunately, tilapia BBs might be unsuitable for clinical islet xenotransplantation because tilapia insulin differs from human insulin by 17 amino acids and, thus, would be immunogenic and less biologically active in humans. Therefore, we have produced transgenic tilapia that express a "humanized" tilapia insulin gene. Future improvements on these transgenic fish may allow tilapia to play an important role in clinical islet xenotransplantation.

Key Words: Brockmann bodies; diabetes; discordant; islet xenotransplantation; physiological incompatibilities; teleost; tilapia; transgenic fish

Introduction

In nondiabetic individuals, the beta cells in the pancreatic islets constantly monitor blood glucose levels and respond to moment-to-moment fluctuations by adjusting insulin secretion. In type 1 diabetes, these insulin-producing beta cells have been destroyed. For the past 80 yr, type 1 diabetes mellitus has been routinely treated by daily insulin injections. Although this treatment controls the acute aspects of the disease, chronic complications eventually occur, primarily because of imprecise control of blood glucose levels (DCCT 1993). Although insulin therapy combined with careful diet and regular exercise can result in normal average blood glucose levels, these patients are not truly normoglycemic in that their glucose levels fluctuate too much throughout the day and, as a result, chronic complications develop over time.

Theoretically, transplantation of new pancreatic islets should ameliorate or prevent the complications described above because it is a more physiological treatment than insulin injections. However, the great promise of islet transplantation has been very slow to materialize. Nevertheless, recent reports definitively show that clinical islet allotransplantation is now very feasible (Ricordi 2003; Ryan et al. 2002; Shapiro et al. 2000), although it is less clear that it will become a widespread treatment. The major issue now is the very limited availability of human islets. In humans, islets comprise approximately 1 to 2% of the pancreas, and it is labor intensive and very expensive to "isolate" them from the exocrine pancreas. Currently, it generally requires two donor pancreases to obtain sufficient yield for a transplant. The most critical problem blocking future islet allotransplantation is a severe shortage of cadaveric donors.

One attractive alternative is clinical islet xenotransplantation (i.e., transplantatation of animal islets). However, this alternative creates a number of problems including the following: (1) selection of an appropriate donor species; (2) robust graft rejection; (3) the potential risk of xenozoonotic disease transmission; (4) animal welfare issues; (5) complicated ethics related to assessment of the rights of transplant recipients, their close contacts, and society as a whole; (6) very thorny institutional, national, and international regulatory issues; (7) very high potential cost; and (8) whether healthcare dollars could be better spent, an issue even more pertinent in countries like Canada, where the government pays medical care costs (Cooper and Lanza 2000; Fishman et al. 1998; Platt 2002). It is beyond the scope of this review to cover all of these topics comprehensively. Our review focuses primarily on the first issue (i.e., selection of the donor species), but we also briefly address many of these other issues in the context of how they affect this selection process.

Experimental Islet Xenotransplantation: Large Animal Donors

Donor Selection and the Importance of Biological Relevance

The criteria used to select an appropriate donor species for islet xenotransplantation (IXTx¹) differ depending on whether the goal is experimental or clinical work. The obvious criteria for experimental work are generally the same as those delineating a good animal model (Leader and Padgett 1980). The most critical determinant is "biological relevance" of the model. Before the 1990s, almost all of the experimental IXTx literature comprised studies in which rat islets were transplanted into mice (Hering 1992). This combination undoubtedly facilitated good (i.e., publishable) results. Rat islets were relatively easy to isolate; however, it soon became apparent that this system did not provide very rigorous testing, because relatively simple graft immunomodulation methods or short-term immunosuppression resulted in marked prolongation or even indefinite graft survival. When similar methods were applied to islet grafts from large animal donors, they were not effective. Because entirely negative results are rarely presented or published, it was not widely recognized that there was such a huge discrepancy between the results obtained using concordant and discordant species combinations; however, by the mid-1990s, this discrepancy became painfully clear. Therefore, except for studies in which the object is to compare rejection of islet allografts, concordant islet xenografts, and discordant islet xenografts, the rat-to-mouse islet xenograft model is no longer biologically relevant.

Biological relevance is hierarchical in nature. Unquestionably, the most clinically relevant IXTx models are those that use spontaneously diabetic nonhuman primates as recipients. Although such studies may represent the absolute "gold standard" for preclinical testing, they require very rigorous ethical review and specialized animal facilities, are extremely expensive, invariably involve very small numbers of recipients and experimental groups, and are very limited in the variables that can be manipulated. Thus, primate studies have limited applicability to answer more fundamental questions on the immunobiology of islet xenograft rejection, which often can be better dissected using large animal donors and specialized recipient models such as knockout mice. In general, studies using nonhuman primates as recipients are beyond the capabilities of most laboratories working in the field of experimental islet transplantation and are generally limited to large centers with clinical islet transplant programs. They are not discussed further in this article.

When comparing models of equal biological relevance, the remaining issues such as ease of use and cost are practical. There are several obvious obstacles to using large animal donors, a category generally encompassing pigs, cows, dogs, nonhuman primates, and humans (i.e., human islets not needed for clinical transplantation). Unlike rodent islets, isolating islets from large animal donors is labor intensive, costly, and requires considerable expertise to obtain high-quality islets.

Pigs are generally considered the most promising source animals for future clinical IXTx (Lacy 1995; Larsen and Rolin 2004; Smith and Mandel 1998). However, they make particularly difficult donors (Larsen and Rolin 2004; Ricordi et al. 1990), and only a small number of laboratories around the world are recognized for consistently isolating high-quality adult porcine islets. This situation has led to the use of islet tissue derived from fetal or neonatal pigs (Deol and Tuch 1999; Korbutt et al. 1996; Korsgren et al. 1991; Mandel 1999; Weir et al. 1997), which are now probably the most widely used large animal models to study islet xenograft rejection. The major advantage of these systems is the ease of harvesting the "islet" tissue. Basically, entire fetal or neonatal pancreases are removed at necropsy and chopped/digested into small fragments (islet cell clusters [ICCs¹]). The ICCs contain primitive ductal tissue and islet precursors that will continue to proliferate and develop into islets after transplantation. The major disadvantage is that the islet tissue does not initially contain many beta cells and does not function for 1 mo or more after transplantation. The ability of these ICCs to continue dividing and differentiating after transplantation is possibly an advantage in the context of clinical IXTx because this process may decrease the total number of litters of donor piglets required per transplant.

Nevertheless, in the context of studying the mechanism of xenograft rejection, the use of ICCs poses a major disadvantage. Because unencapsulated islets from pigs and other large animal donors functionally reject after transplantation into euthymic mice in roughly 1 wk, it is not generally possible to transplant porcine ICCs into diabetic recipients and follow them for functional rejection. Instead, studies using ICCs are often performed by killing recipients at regular intervals and then retrieving the grafts for some combination of histological, immunological, chemical, or molecular analyses. Because islet xenograft rejection between discordant species has proven very difficult to solve, many laboratories and biotechnology companies have focused their efforts on developing microencapsulation devices and other bioartificial pancreas technologies. The unifying concept behind these technologies is that the islet tissue is placed behind a semipermeable barrier that allows small molecules such as glucose, insulin, oxygen, and nutrients to pass freely but restricts the entry of larger molecules such as immunoglobulins, some complement components, and cells of the immune system (Lanza and Chick 1994; Kuhtreiber et al. 1999). Both innate (nonspecific) and adaptive (specific) immunity are believed to play critical roles in the response to encapsulated islet xenografts. The inability to assess graft function during "rejection" of encapsulated islets also hampers studies attempting to dissect this process.

Practical Issues

In addition to the caveats discussed above, several other important practical issues relate to the use of large animal donors. Because they are "more sentient" than rodents, obtaining approval by institutional animal use and ethical review committees can be difficult. The use of large animal donors also results in a "feast or famine" scenario. For example, the pancreas from an adult pig or pancreases from a litter of fetal/neonatal pigs may yield up to several hundred thousand islets. To utilize all of these islets optimally and to avoid waste, large numbers of experiments much be batched and performed simultaneously; alternatively, extra islets can be cultured, cyropreserved, or shipped to investigators for use at other institutions. However, culture and cryopreservation are additional variables that may have immunomodulatory effects. Under ideal circumstances, it would be better to be able to isolate the correct number of islets for each experiment at the time they are needed; however, this ideal is impractical with large animal donors. Finally, the cost of purchasing and housing large animal donors, especially when combined with very high islet isolation costs, can make this work exceedingly expensive. In summary, experimental IXTx using large animal donors is biologically relevant but problematic, especially for smaller laboratories with lower budgets.

Experimental Islet Xenotransplantation: Tilapia as Donors

Advantages of the System

Since the early 1990s, our laboratory has used tilapia (Figure 1), a large tropical teleost (i.e., bony) fish as islet donors for experimental IXTx (Wright 1992; Wright and Pohajdak 2001). We and others have shown that tilapia islets transplanted under the kidney capsule of streptozotocin (STZ¹)-diabetic athymic nude mice or severe combined immunodeficient mice function immediately and provide long-term normoglycemia routinely (Leventhal et al. 2004; Morsiani et al. 1995; Wright et al. 1992). The advantages of this system for studying the mechanism of islet xenograft rejection are numerous. Most importantly, xenotransplantation using tilapia donors passes the biological relevance test, because rejection in euthymic rodent recipients is temporally and appears mechanistically very similar to rejection of pig or human islets (Dickson et al. 2003). When transplanted into inbred strains of euthymic mice, mean graft survival times are between 6.7 and 8.0 days (standard deviation 1.4-2.2) (Wright et al. 1994b). Rejected grafts showed extensive infiltration by macrophages, eosinophils, and T cells (both CD4⁺ and CD8⁺) (Wright et al. 1997). Because of the ease of harvesting tilapia islets (see below), this model proved to be an excellent and quick way to screen various methods that have been reported to prolong islet allograft, concordant (rat-to-mouse) xenograft, or discordant (pig-to-mouse) islet xenograft survival (Wright and Pohajdak 2001). We have screened treatments including administration of various immunosuppressive drugs; depletion of various cytokines, T cell subsets, or eosinophils using mAb treatments; macrophage depletion; treatments with eicosanoid inhibitors or free radical scavengers; and islet culture, irradiation, or cryopreservation to deplete passenger leukocytes before transplantation.

Figure 1 Tilapia (Oreochromis niloticus) are a popular aquaculture species worldwide.

As reported by others transplanting islets from large animal donors, all immunomodulation methods directed at decreasing graft immunogenicity before transplantation have been ineffective (O'Hali et al. 1997; Wright and Kearns 1995b; Yang et al. 1995), and temporary immunosuppression methods have provided very modest prolongation at best (Wright et al. 1994a). However, continuous immunosuppression with high-dose, potent immunosuppressive drugs has provided more prolonged graft survival, albeit with toxic systemic effects (e.g., lethal bone marrow suppression or post-transplant lymphoproliferative disorder) within 1 to 2 mo (Wright and Kearns 1995a; Yang et al. 2002b). To date, the only highly effective, tested treatment for prolonging survival of nonencapsulated tilapia islet grafts is CD4 depletion of recipients (Dickson et al. 2003), a method that is also highly effective in the context of pig-to-mouse IXTx. Additional studies using adoptive transfer of T cell subpopulations into immunodeficient mice confirms the critical role for CD4⁺ T cells in tilapia islet rejection (Leventhal et al. 2003). In each of these instances, our results suggest that the tilapia-to-mouse model provides similar results to those obtained with the popular and biologically relevant pig-to-mouse model (Simeonovic 1999). However, it is the practical aspects of our model that make it excel for experimental IXTx.

Brockmann Bodies

In many types of teleost fish (including tilapia), the islet tissue is anatomically distinct from the exocrine pancreas and forms multiple discrete islet organs called Brockmann bodies (BBs¹). This anatomical separation of islet tissue in teleosts has been exploited for diabetes research, and even diabetes treatment, for more than a century (Wright 2002a,b). Harvesting tilapia BBs is simple and does not require expensive, fickle islet isolation procedures. The BBs are scattered throughout the adipose tissue surrounding the common bile duct and the pancreatic ducts that reside between the liver, stomach, spleen, and gall bladder (i.e., the "BB region") (Figures 2 and 3) (Wright 1994). The larger tilapia islets can be harvested easily by removing the whole BB region and microdissecting it while it is visualized through a dissecting microscope (Wright 1994). However, this process is slow, and smaller BBs are missed. Alternatively, tilapia BBs can be enzymatically mass-harvested (Yang and Wright 1995) by removing BB regions from multiple donor fish simultaneously, placing them in a tube containing a type 2 collagenase (i.e., generally sold for harvesting adipocytes) solution, and then placing the tube in a shaker water bath at 37°C. After roughly 10 min, the digestion is stopped by adding cold Hanks' balanced salt solution; the adipocytes float and the islets sink forming a pellet. Because of the large size of some of the BBs (up to 5 mm diameter in large donors), we shred or chop them until the fragments are roughly the diameter of mammalian islets; after overnight culture, these fragments "round up" and look like cultured mammalian islets (Yang et al. 1997a). Although this fragmentation process does decrease islet yield, it does not affect islet function adversely because of the high degree of architectural redundancy of the tilapia BBs (Yang et al. 1999). Like mammalian islets, fragmented tilapia BBs can be maintained in tissue culture under a variety of conditions (Wright and Kearns 1995b), or they can be cryopreserved (O'Hali et al. 1997) before transplantation.

Figure 2 Dissection of a female tilapia with the right ovary and the omentum reflected downward reveals the roughly triangular "BB region" (outlined by arrows) surrounded by the liver (L) anteriorly, stomach (ST) superiorly, and spleen (S) and gall bladder (G) inferiorly. BB, Brockmann bodies. Reprinted with permission from Yang H, Wright JR Jr. 1995. A method for mass harvesting islets (Brockmann bodies) from teleost fish. Cell Transplant 4:621-628.

Figure 3 Whole mount produced by processing an entire "BB region" for histology. Sections were cut at three different levels through the block to provide a three-dimensional view. Sections were stained with hematoxylin and eosin. The regions are composed of adipose tissue (A), bile and pancreatic ducts (D), blood vessels (V), and Brockmann bodies (BB). Twelve BBs can be identified in the center frame. Original magnification ×1, photographic enlargement ×7. Reprinted with permission from Yang H, Wright JR Jr. 1995. A method for mass harvesting islets (Brockmann bodies) from teleost fish. Cell Transplant 4:621-628.

Because fish are less sentient and probably have less capacity to feel pain than mammals, our model is generally viewed very favorably by institutional animal use and ethics review committees, because the use of tilapia as donors supports the widely accepted Russell and Burch "3 Rs" tenet of refining, reducing, and replacing animals used in research (Flecknell 2002). Because there is a direct linear relationship between donor body weight and islet cell numbers (Dickson et al. 1998), this relationship can be exploited to determine how much islet tissue is present in each donor fish. Thus, sufficient islet tissue, without waste, can be harvested on demand for experimentation on a daily basis. A "transplantable unit" of fragmented tilapia BBs (i.e., sufficient to assure long-term normoglycemia after transplantation under the kidney capsule of a 25- to 30-g STZ-diabetic athymic nude mouse) is the equivalent of all of the chopped islet tissue harvested from 800 g of donor tilapia (~1.5 million islet cells after fragmentation). This unit could be derived from one 800-g donor, two 400-g donors, or, for example, one tenth of the total islet tissue harvested from multiple donors with an aggregate body weight of 8000 g. Although we generally use donor fish weighing between 500 and 1000 g, this linear relationship holds for donors weighing between 100 g and 1.8 kg (n.b., islet yields appear to plateau when harvesting donors > 2 kg) (Dickson et al. 1998).

Tilapia islets, like hard-to-isolate adult porcine islets, function immediately after transplantation. Therefore, unlike the popular porcine ICCs, graft function can be monitored directly, and rejection can be assessed on the basis of function. Also like adult porcine islets, unencapsulated tilapia islets can be transplanted into multiple sites, including the renal subcapsular space, the liver via the portal vein, and the testis. After transplantation into these sites, the grafts gradually become revascularized when recipient capillaries invade the graft ("neovascularization"). Very recently, we have developed a microvascular surgery technique that allows tilapia islets to be transplanted as cluster grafts. This process allows a direct comparison of islet xenograft rejection processes in directly vascularized versus neovascularized tilapia islets (Yu et al. 2003). The work is particularly interesting in the context of our recent findings, that tilapia cells do not express α(1,3)gal, an antigen present on mammalian cells (except humans and Old World apes) that precipitates hyperacute rejection of vascularized whole organ xenografts in humans (Leventhal et al. 2004).

Fragmented tilapia BBs are very suitable for encapsulation, and these technologies markedly prolong islet xenograft survival in diabetic rodents (Yang and Wright 2002; Yang et al. 1997b). Current encapsulation technologies are only moderately effective when islets are transplanted between discordant species; thus, they are still evolving to improve their biocompatibility, minimize graft hypoxia, and decrease their tendency to shed antigen. Furthermore, much basic research pertaining to the possible entry of small molecules such as cytokines, chemokines, and smaller elements of the complement system remain to be performed. One of our laboratory's current major research foci is studying the mechanism of eventual graft failure after transplantation of encapsulated tilapia islets. Recently, we have found that coencapsulation of tilapia islets with rodent Sertoli cells, which are known to produce a number of immunomodulatory products, promotes graft survival (Yang and Wright 1999; Yang et al. 2002a). We have also used encapsulated tilapia islets to study the possible role of autoimmunity in the loss of islet xenografts in spontaneously diabetic NOD mice (Xu et al. 2003).

A Model to Study Cross-species Physiological Incompatibility in Xenotransplantation

To date, most xenotransplantation research has focused either on preventing rejection or on evaluating the risk of transmitting infectious diseases. A major question in the field of xenotransplantation is whether physiological differences between donors and recipients would prevent adequate graft function. Because evolution has resulted in variation and adaption to the environment over a long period of time, there are marked physiological differences between species. This issue is critical to the eventual success of xenotransplantation but has been largely ignored because it is almost impossible to study due to the immunological barriers that prevent long-term analysis of graft function (Hammer 1991; Hammer et al. 1998). Therefore, it is totally unknown whether organs or cells from phylogenetically distant species will function adequately in humans. This issue is particularly relevant to xenografts expected to maintain homeostasis when all or part of the function is endocrine. Peptide hormones are impotent unless they bind to an appropriate receptor. If feedback components and inhibitors do not function appropriately, hormonal chaos would be expected. Furthermore, such foreign peptides could function as antigens and precipitate antibody and immune complex formation.

We have recently shown for the first time that islet xenografts appear to adapt to their new environment (Morrison et al. 2003b). Using our highly discordant fish-to-diabetic nude mouse islet transplant model, we have shown that the cellular source of an irrelevant peptide (somatostatin [SST¹])-28, a product of the preproSST-II gene [either not present or not expressed in mammals]) is either rapidly lost or degranulates, and that the cellular composition of the remainder of the endocrine graft becomes increasingly more recipient-like. In stark contrast, staining for tilapia SST-14 (n.b., sequence is identical to mammalian SST [Nguyen et al. 1995]), a product of the preproSST-I gene, persists after transplantation. The relatively rapid loss or degranulation of SST-28⁺ cells suggests that sources of irrelevant peptides may be lost after xenotransplantation and, thus, may not serve as persistent antigens (Morrison et al. 2003b).

Mammals have two glucagon-like peptides: GLP-1 and GLP-2. Mammalian GLP-1 is produced by intestinal L cells and is secreted in response to nutrient ingestion (Doyle and Egan 2001). Fish, however, have only GLP-1, and it is produced in the pancreatic islets (Plistskaya and Mommsen 1996). Interest in GLP-1 (and its analogs) as a possible adjuvant treatment in type 2 diabetes has become intense since the early 1990s because exogenous administration has been found to reduce blood glucose levels in both normal subjects and diabetic patients by stimulating glucose-induced insulin secretion (and synthesis) and by inhibiting glucagon secretion. In mammals, GLP-1 acts directly on beta cells via a specific receptor (Doyle and Egan 2001). In fish, GLP-1 has minimal insulinotropic effect on beta cells (i.e., in stark contrast to humans) and acts primarily on the liver. In humans, GLP-1 has no direct effect on liver metabolism. Although the metabolic effects of GLP-1 are different in mammals and fish, administration of GLP-1 from either source elicits the species-appropriate effect in either (Plistskaya and Mommsen 1996).

We have recently found that tilapia islet cells that produce GLP-1 are abundant several months after transplantation into STZ-diabetic nude mice. This finding suggests that feedback mechanisms to maintain GLP-1 expression are intact. Because fish GLP-1 has a high degree of biological activity in mammals (Plistskaya and Mommsen 1996) and because GLP-1 is known to stimulate beta cell differentiation and neogenesis in mammals (Doyle and Egan 2001), we are currently looking for evidence of islet neogenesis (nesidioblastosis) in the native pancreata or altered GLP-1 expression in the intestinal L cells of STZ-diabetic nude mouse recipients with long-standing BB grafts. We also plan to measure serum levels in recipients. Because tilapia GLP-1 has only 68% sequence homology with murine GLP-1 and is three amino acids longer (Nguyen et al. 1995), we expect to be able to distinguish tilapia GLP-1 and mouse GLP-1 and measure them quantitatively using mass spectrometry. This work will be fascinating because GLP-1 circulates at levels in fish that are at least 10-fold higher than in mammals (n.b., perhaps due to slower clearance, increased secretion, or both) (Plistskaya and Mommsen 1996). These studies are examples of how this model can be used to study the physiological interactions across species after xenotransplantation.

Possible Suitability of Tilapia Islets for Clinical Islet Xenotransplantation

Tilapia islets are easy to harvest, and they provide long-term normoglycemia when transplanted into STZ-diabetic nude mice (Wright et al. 1992). These features have made them ideal for experimental IXTx. After working with the model for a few years, we began to wonder whether a case could be made to use tilapia BBs for clinical IXTx. Results of our analysis, as described below, are resoundingly affirmative.

Before we seriously considered clinical IXTx using tilapia, we needed to have a better understanding of their islet function and glucose homeostasis (reviewed in Wright et al. 2000). To that end, we initally compared mean fasting/nonfasting blood glucose levels in tilapia and humans and found them to be similar: 75.4/91.9 mg/dL versus 63/90 mg/dL, respectively. In another study, we transplanted equal numbers of rat, mouse, or tilapia islets into STZ-diabetic nude mice and observed them for 30 days. Mean nonfasting blood glucose levels in recipients of tilapia, rat, and mouse islets were 77.8 ± 4.4, 77.0 ± 1.3, and 115 ± 4.5 mg/dL, respectively. Next, we performed glucose tolerance tests (GTTs¹). All three groups had mean glucose disappearance rates (K values) of 4.3 to 5.7, indicating that tilapia islets are as glucose responsive as rodent islets (and, therefore, several-fold more glucose responsive than human or porcine islets) (Yang et al. 1997a). The observation that tilapia islets are very highly glucose responsive was puzzling, because teleost fish in general are notoriously glucose intolerant. In another study, we compared the results of GTTs in donor tilapia with those in STZ-diabetic nude mice cured by long-standing tilapia islet grafts. Amazingly, it took 3 days for the fish and only 30 min for the recipient mice to clear a glucose load (K values ~ 0.05 vs. 6.0). The most logical explanation was that the tilapia peripheral tissues were resistant whereas the mouse peripheral tissues were highly sensitive to the glucostatic effects of the secreted insulin.

In an effort to explain extreme peripheral resistance to insulin, we focused on glucose transporters, because there were already solid published data showing that tilapia skeletal muscle cells express functional insulin receptors. In mammals, glucose uptake in peripheral tissues is facilitated by the glucose transporters GLUT-1 and GLUT-4, which are coexpressed in muscle, heart, and adipose tissue. In mammals, GLUT-1 facilitates basal transport and is not insulin responsive, whereas GLUT-4, which is insulin responsive, is responsible for the rapid uptake of larger quantities of glucose. Because fish do not normally consume much glucose and because the small quantities of simple carbohydrates they do consume would not tend to be as large boluses, we hypothesized that fish may never have evolved a homolog of the insulin-sensitive GLUT-4 glucose transporter that allows mammals to handle hyperglycemia (Wright et al. 1998a, 2000). Initially, we were unable to identify any evidence of a GLUT-4-like transporter, but were easily able to demonstrate the presence of a GLUT-1 homolog in tilapia (Wright et al. 1998a) and have since cloned and sequenced it (Hrytsenko O, Mansour M, Wright JR Jr, Pohajdak B; Biology Dept., Dalhousie University, Halifax, NS, Canada). However, more recent developments suggest that a glucose transporter with considerable sequence homology to GLUT-4 may exist in fish muscle, but that its expression is insulin responsive only in red muscle (i.e., slow contraction fibers with primarily oxidative metabolism) but not in white muscle (i.e., fast contraction fibers with primarily glycolytic metabolism) (Capilla et al. 2002). If this finding in brown trout can be generalized to tilapia, it would nicely explain our GTT data described above, because the vast majority of tilapia body mass (as in many fish species) is composed of white muscle (n.b., tilapia have small amounts of red muscle subdermally along the lateral line) (Dal Pai-Silva et al. 2003).

"Ideal" Islet Tissue

Because there is no known method to prevent rejection of unencapsulated islet xenografts, clinical IXTx, for the foreseeable future, could only be accomplished using encapsulated islet grafts. Therefore, we have assessed the suitability of tilapia islets for encapsulation. According to Dionne and colleagues (1994), the ideal islet tissue for encapsulation would be one that "has a high insulin output, is correctly regulated by glucose and other secretogogues, has low metabolic demand, and is capable of functioning for extended periods without replacement. In addition, the cells must be procurable in high yield at reasonable cost with protocol meeting [US Food and Drug Administration (FDA¹)] FDA standards" (p. 123). Tilapia islets appear to meet many of these expectations; in fact, they are extremely well suited for encapsulation relative to mammalian islets because they tolerate exceedingly low oxygen tensions without loss of function or viability (Wright et al. 1998b). This characteristic is extremely important because even when encapsulation is functioning optimally, encapsulated islets function in a chronically hypoxic environment.

Within the mammalian pancreas, islets are a very well-vascularized tissue; however, once they have been isolated and encapsulated, the insulin-producing beta cells depend on diffusion of oxygen through the walls of the capillaries, through interstitial and peritoneal fluid to reach the capsule, through its wall, through the dead space between the islet and the capsule wall, and then through hypoxic non-beta cells that comprise the outer mantle of each mammalian islet. Therefore, gradual beta-cell attrition and loss of function due to the cumulative effects of hypoxia are unavoidable; in most instances, hypoxia is probably the ultimate cause of graft failure. Using side-by-side culture under hypoxic conditions, we have shown that tilapia islets are many fold more resistant to hypoxia than mammalian islets (Wright et al. 1998b). This resistance should both prolong graft function and decrease the islet mass required for transplantation.

Another important question is whether other foreign peptides secreted by tilapia islets are biologically active or immunogenic in humans. As discussed in the previous section, cells producing SST-28, a known irrelevant peptide, disappear after transplantation into a mammalian environment. However, GLP-1⁺ cells persist. Because we know that fish GLP-1s are biologically active in humans and it appears likely that this peptide may circulate, it is not unreasonable to speculate that it would exert its known effects on the recipient's native pancreas (i.e., promoting insulin secretion and islet neogenesis). In this context, it is critical to know whether tilapia GLP-1 can be inactivated in human recipients; if not, it could have an unregulated effect. Fortunately, unregulated production does not seem to be a problem because tilapia GLP-1, based on its known structure (Nguyen et al. 1995), appears to be a good substrate for dipetidyl peptidase-IV, which inactivates GLP-1 and is believed to play a critical role in regulating its biological activity. Secretion of GLP-1 could be very fortuitous, because GLP-1 analogs have shown considerable promise as a possible future treatment for type 2 diabetes, which is at least 10-fold more prevalent than type 1 diabetes. Although islet transplantation research has primarily focused on type 1 diabetes, there is solid clinical and experimental evidence supporting its use in some patients with type 2 diabetes (Ricordi et al. 1995; Thomas et al. 1993). Therefore, if GLP-1 circulates systemically and can be appropriately inactivated in humans, it is not implausible that tilapia islet xenografts, secreting both insulin and a biologically active insulinotropic GLP-1, could play a potential role in the future treatment of both type 1 and type 2 diabetes.

Genetically Modified Tilapia as Donors

Unfortunately, there is a major obstacle to the use of tilapia islets for clinical IXTx: tilapia islets secrete tilapia insulin, which differs from human insulin by 17 amino acids (Nguyen et al. 1995). This difference would create two problems: (1) Because fish insulins are only 30 to 50% biologically active in humans, two- or three-fold more islets would be needed per transplant; and (2) many recipients would likely develop antibodies to tilapia insulin, which could lead to severe insulin resistance.

To help circumvent these problems, we have produced transgenic tilapia that express a "humanized" tilapia insulin gene. Initially, we cloned and sequenced the tilapia insulin gene (Mansour et al. 1998). We then modified the coding region by site-directed mutagenesis, changing only the codons representing the 17 amino acids that differed in tilapia and humans. This process resulted in a tilapia insulin gene that coded for humanized insulin. We then microinjected 10⁶ copies of the modified gene into fertilized eggs at the single cell stage and screened for transgene incorporation. We have produced small numbers of founders with incorporation of our humanized insulin gene and have achieved germ-line transmission of the transgene. Using a radioimmunoassay that does not cross-react with tilapia insulin, we have shown high levels of circulating humanized insulin in serum of F1 offspring. Nonfasting blood glucose levels in transgenic tilapia are not significantly different from those in control tilapia, which suggests that insulin secretion is regulated. Using immunoperoxidase staining, beta cells in the BBs of transgenic tilapia, but not control tilapia, stain for human insulin (Pohajdak et al., 2004).

These fish are not yet the ideal islet donors, because their islets simultaneously produce humanized and tilapia insulins. We are now attempting to develop embryonic stem cell technology that should allow us to knock out the tilapia insulin; several lines of evidence suggest that tilapia that produce only human insulin should be healthy (Wright and Pohajdak 2002). Toward this goal, we have extensively characterized embryonic development in tilapia (Morrison et al. 2001, 2003a) and have been able to produce chimeric tilapia by microinjecting dispersed blastula cells into developing blastulas (MacNeil 2002).

Cost Analysis

Once it has been shown to be feasible and safe, cost will be a major factor in determining whether widespread implementation of clinical IXTx will occur. Therefore, we present below what we believe to be a conservative analysis of the relative costs on a "per transplant" basis and using either transgenic tilapia or adult porcine donors (n.b., there is currently not sufficient information available to attempt a cost estimate for the use of neonatal/fetal porcine islets. We believe that these costs would be substantially greater than with tilapia, but probably less than with adult pigs).

We estimate that clinical IXTx of encapsulated islets will require about 20,000 islet equivalents (IEQ¹)/kg of recipient body weight (Soon-Shiong et al. 1994). Therefore, 1.4 million IEQ would be required for a 70-kg recipient. Our analysis is based on several assumptions. If we assume an average yield of 140,000 IEQ/pig pancreas using 1-yr-old donors, then 10 pigs will be required per transplant. Based on transplantation into STZ-diabetic nude mice, one 800-g nontransgenic tilapia would yield the functional equivalent of about 2,000 human or porcine IEQ (Wright and Pohajdak 2001). Therefore, the islet tissue from 700 nontransgenic tilapia would be needed to transplant a 70-kg recipient. Fish insulins are 30 to 50% active in humans (Nguyen et al. 1995), and human insulin secreted by transgenic tilapia islets should be 100% active in humans. Based on this assumption, 350 (or fewer) transgenic donors should provide sufficient islets for a 70-kg recipient.

The FDA and other federal regulatory agencies require all "source animals" for any form of clinical xenotransplantation to be produced in indoor factories under specific pathogen-free (SPF¹) conditions; animals produced for human consumption obtained from slaughterhouses are specifically excluded (CBER 2003). Therefore, animal husbandry costs will comprise a significant component of total islet costs. Compared with pigs, tilapia are 2.5-fold more efficient at converting food into body mass, have shorter generation times (i.e., conception to sexual maturity, 6 vs. 12 mo), have larger "litter" sizes (i.e., 100s-1000s vs. <10), have shorter minimal intervals between "litters" (i.e., 2 weeks vs. 6 mo), and have minimal space requirements for housing (Wright and Pohajdak 2001). Assuming 10 donor pigs per transplant and a "per diem" rate of only $5/day/pig for food and indoor housing (estimate excludes the costs of SPF conditions), pig donor husbandry costs are estimated at $18,250 US dollars (USD¹) per clinical islet transplant. In contrast, 350 tilapia weighing 800 g each can be commercially raised from fertilized eggs in an indoor facility in Canada for about $560 USD (i.e., $2 USD/kg) (personal communication, Gary Chapman, Northern Tilapia Inc., Bond Head, Ontario, July 2003).

The cost of isolating porcine islets is staggering. Assuming considerable savings because of the ability to buy reagents in bulk, the total cost of reagents and technical support should still exceed $6,000 USD/pancreas (Wright and Pohajdak 2001). Therefore, total islet isolation costs per encapsulated islet xenotransplant should exceed $60,000 USD. Based on our experience, unskilled individuals can quickly be taught to harvest BB regions from tilapia at a rate of > 35 fish/hour; therefore, harvesting 350 fish should require about 10 hr of unskilled labor ($60 USD). The reagents needed to harvest BBs from these fish would cost about $580 USD. Therefore, total islet isolation costs (on a per transplant basis) would be about $640 USD for transgenic tilapia.

The American Association for the Accreditation of Laboratory Animal Care or other animal welfare agencies require humane housing. To meet minimum standards, one pig requires 4 m² of floor space. Therefore, to house the one million donor pigs needed to perform 100,000 clinical islet transplants per year, four million square meters (or 200 indoor factories of 20,000 m² floor space) would be required. In contrast, tilapia thrive in crowded conditions; we estimate that seven factories of the same size would house the 35,000,000 donor fish. Even with amortization over 5 to 10 yr of performing 100,000 clinical transplants per year, the cost of building almost 30-fold more factories would add substantially to the cost of producing pig islets.

Finally, the costs of the SPF conditions have not been estimated, and these costs will be significant. In pigs, vertical transmission of disease from mother to offspring can occur in utero, at delivery, or while nursing; Caesarian rederivation costs (i.e., the mother is killed and an aseptic hysterectomy performed in an approved operating room) will be huge. In contrast, tilapia eggs are fertilized, develop externally, and have a protective shell (Morrison et al. 2003b), which can be chemically treated. FDA-approved antibacterial, antifungal, and antiviral treatments for fish eggs already exist (Wright and Pohajdak 2001). On a per transplant basis, providing SPF conditions should be many fold cheaper for tilapia than pigs.

In summary, our analysis suggests that transgenic tilapia islet production costs should be at least 100-fold less expensive than porcine islets on a per clinical transplant basis. This advantage should greatly facilitate the widespread implementation of clinical IXTx.

Conclusions

The foregoing discussion summarizes the reasons that tilapia make an excellent donor species for biologically relevant experimental IXTx studies. Their islets are easy to harvest, can be transplanted into numerous sites, and, unlike the popular porcine fetal/neonatal ICCs, function immediately after transplantation. Evidence indicates that nude mouse recipients of tilapia islets can be used to gain insights into issues related to cross-species physiological incompatibilities in xenotransplantation. Importantly, transgenic tilapia that express a humanized insulin gene might prove to be suitable and economical donors for future clinical IXTx.

¹Abbreviations used in this article: BBs, Brockmann bodies; FDA, US Food and Drug Administration; GLP, glucagon-like peptide; GTT, glucose tolerance test; ICCs, islet cell clusters; IEQ, islet equivalents; IXTx, islet xenotransplantation; SPF, specific pathogen-free; SST, somatostatin; STZ, streptozotocin; USD, US dollars.

References

Capilla E, Diaz M, Gutierrez J, Planas JV. 2002. Physiological regulation of the expression of a GLUT-4 homolog in fish skeletal muscle. Am J Physiol Endocrinol Metab 283:E44-E49.

CBER [Center for Biologics Evaluation and Research]. 2003. Guidance for Industry: Source Animal, Product, Preclinical, and Clinical Issues Concerning the Use of Xenotransplantation Products in Humans. April 3 2003 (http://www.fda/gov/cber/guidelines.htm).

Cooper DKC, Lanza RP, eds. 2000. Xeno: The Promise of Transplanting Animal Organs into Humans. Oxford UK: Oxford University Press.

Dal Pai-Silva M, Carvalho RF, Pellizzon CH, Dal Pai V. 2003. Muscle growth in Nile tilapia (Oreochromis niloticus): Histochemical, ultrastructural and morphometric study. Tissue Cell 35:179-187.

DCCT [Diabetes Control and Complications Trial] Research Group. 1993. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 329:977-986.

Deol HS, Tuch BE. 1999. Immunomodulation of fetal pig islet-like cell clusters gamma irradiation. Xenotransplant 6:141-146.

Dickson B, Yang H, Pohajdak B, Wright JR Jr. 1998. Quantification of tilapia islets: A direct relationship between islet cell number and body mass. Transplant Proc 30:621-622.

Dickson BC, Yang H, Savelkoul HFJ, Rowden G, van Rooijen N, Wright JR Jr. 2003. Islet transplantation in the discordant tilapia-to-mouse model: A novel application of alginate microencapsulation in the study of xenograft rejection. Transplantation 75:599-606.

Dionne K, Scharp D, Lysaght M, Hegre O, Lacy P. 1994. Macroencapsulation of islets for the treatment of diabetes. In: Lanza RP, Chick WL, eds. Pancreatic Islet Transplantation. Vol III. Immunoisolation of Pancreatic Islets. Austin: RG Landes Co. p 119-131.

Doyle ME, Egan JM. 2001. Glucagon-like peptide-1. Recent Prog Horm Res 56:377-399.

Fishman J, Sachs DH, Shaikh R, eds. 1998. Xenotransplantation: Scientific frontiers and public policy. Ann N Y Acad Sci Vol 862.

Flecknell P. 2002. Replacement, reduction and refinement. ALTEX 19:73-78.

Hammer C. 1991. Evolution: Its complexity and impact on xenotransplantation. In: Cooper DKC, Kemp E, Platt JL, White DJG, eds. Xenotransplantation: The Transplantation of Organs and Tissues Between Species. Berlin: Springer. p 716-735.

Hammer C, Linke R, Wagner F, Diefenbeck M. 1998. Organs from animals for man. Int Arch Allergy Immunol 116:5-21.

Hering BJ. Islet xenotransplantation. 1992. In: Ricordi C, ed. Pancreatic Islet Cell Transplantation: 1892-1992: One Century of Transplantation for Diabetes. Austin: RG Landes Co. p 313-335.

Korbutt GS, Elliott JF, Ao Z, Smith DK, Warnock GL, Rajotte RV. 1996. Large scale isolation, growth, and function of porcine neonatal islet cells. J Clin Invest 97:2119-2129.

Korsgren O, Jansson L, Eizirik, D, Andersson A. 1991. Functional and morphological differentiation of fetal porcine islet-like cell clusters after transplantation into nude mice. Diabetologia 34:379-386.

Kuhtreiber WM, Lanza RP, Chick WL, eds. 1999. Cell Encapsulation Technology & Therapeutics. Boston: Birkhauser.

Lacy PE. 1995. Treating diabetes with transplanted cells. Sci Am 273:54-58.

Lanza RP, Chick WL, eds. 1994. Pancreatic Islet Transplantation Series. Vol 3. Immunoisolation of Pancreatic Islets. Austin: RG Landes Co.

Larsen MO, Rolin B. 2004. Use of the Göttingen minipig as a model of diabetes, with special focus on type 1 diabetes research. ILAR J 45:303-313.

Leader RW, Padgett GA. 1980. The genesis and validation of animal models. Am J Pathol 101(3S): S11-S16.

Leventhal JR, Sun JD, Zhang J, Galili U, Chong A, Baker M, Kaufman DB, Wright JR Jr. 2004. Evidence that tilapia islets do not express α(1,3) Gal: Implications for islet xenotransplantation. Xenotransplant 11:276-283.

Leventhal J, Zhang J, Sun J, Barrett T, Wright J. 2003. Rejection of tilapia islets in mice is CD4⁺ cell dependent. (Abstract). Xenotransplant 10:498.

MacNeil SD. 2002. The development of tilapia (Oreochromis niloticus) chimeras and short-term culture of undifferentiated embryo cells. Master's Thesis, Dalhousie University, Halifax, Nova Scotia.

Mandel TE. 1999. Fetal islet xenotransplantation in rodents and primates. J Mol Med 77:155-160.

Mansour M, Wright JR Jr, Pohajdak B. 1998. Cloning, sequencing and characterization of the tilapia insulin gene. Comp Biochem Physiol 121B:291-297.

Morrison CM, Miyake T, Wright JR Jr. 2001. Development of embryo and early larva of Oreochromis niloticus (Pisces: Cichlidae). J Morphol 247:172-195.

Morrison CM, Pohajdak B, Wright JR Jr. 2003a. Structure and enzymatic removal of the chorion of embryos of the Nile tilapia Oreochromis niloticus. J Fish Biol 63:1439-1453.

Morrison CM, Yang H, Al-Jazaeri A, Tam J, Plisetskaya E, Wright JR Jr. 2003b. Xenogeneic milieu markedly remodels endocrine cell populations after transplantation of fish islets into streptozotocin-diabetic nude mice. Xenotransplant 10:60-65.

Morsiani E, Lebow LT, Rozga J, Demetriou AA. 1995. Teleost fish islets: A potential source of endocrine tissue for the treatment of diabetes. J Surg Res 58:583-591.

Nguyen T, Wright JR Jr, Nielsen PF, Conlon JM. 1995. Characterization of the pancreatic hormones from the Brockmann body of the tilapia: Implications to islet xenograft studies. Comp Biochem Physiol 111C:33-44.

O'Hali W, Yang H, Pohajdak B, LaPrairie A, Gross M, Wright JR Jr. 1997. Cryopreservation of fish islets: The effect on function and islet xenograft survival. Transplant Proc 29:1990-1991.

Platt JL, ed. 2002. Xenotransplantation: Basic Research and Clinical Application. Totowa NJ: Humana Press.

Plistskaya EM, Mommsen TP. 1996. Glucagon and glucagon-like peptides in fishes. Int Rev Cytol 168:187-257.

Pohajdak B, Mansour M, Hrytsenko O, Conlon JM, Dymond C, Wright JR Jr. 2004. Production of transgenic tilapia with Brockmann bodies secreting [desThrB30] human insulin. Transgenic Res (In Press).

Ricordi C. 2003. Islet transplantation: A brave new world. Diabetes 52:1595-1603.

Ricordi C, Angelico MC, Alejandro R, Bottino R, Linetsky E, Selvaggi G, Mintz DH, Tzakis A. 1995. Islet transplantation in type 2 diabetes. Transplant Proc 27:3166.

Ricordi C, Socci C, Davalli AM, Staudacher C, Baro P, Vertova A, Sassi I Gavazzi F, Pozza G, Di Carlo V. 1990. Isolation of the elusive pig islet. Surgery 107:688-694.

Ryan EA, Lakey JR, Paty BW, Imes S, Korbutt GS, Kneteman NM, Bigam D, Rajotte RV, Shapiro AMJ. 2002. Successful islet transplantation: Continued insulin reserve provides long-term glycemic control. Diabetes 51:2148-2157.

Shapiro AMJ, Lakey JR, Ryan EA, Korbutt GS, Toth E, Warnock GL, Kneteman NM, Rajotte RV. 2000. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med 343:230-238.

Simeonovic CJ. 1999. Xenogeneic islet transplantation. Xenotransplant 6:1-5.

Smith RM, Mandel TE. 1998. Transplantation treatment for diabetes. Immunol Today 19:444-447.

Soon-Shiong P, Heintz RE, Merideth N, Yao QX, Yao Z, Zheng T, Murphy M, Moloney MK, Schmehl M, Harris M. 1994. Insulin independence in a type 1 diabetic patient after encapsulated islet transplantation. Lancet 343:950-951.

Thomas F, Pittman K, McFadden T, Hiasch C, Peterson R, Thomas J. 1993. Reversal of type II diabetes by pancreas islet transplant in four separate animal models of type II diabetes. Transplant Proc 25:992-993.

Weir GC, Quickel RR, Yoon K-H, Tatarkiewicz K, Ulrich TR, Hollister-Lock J, Bonner-Weir S. 1997. Porcine neonatal pancreatic cell clusters (NPCCs): A potential source of tissue for islet transplantation. Ann Transplant 2:63-68.

Wright JR Jr. 1992. Experimental transplantation using principal islets of teleost fish (Brockmann bodies). In: Ricordi C, ed. Pancreatic Islet Cell Transplantation: 1892-1992: One Century of Transplantation for Diabetes. Austin: RG Landes Co. p 336-351.

Wright JR Jr. 1994. Procurement of fish islets (Brockmann bodies). In: Lanza RP, Chick WL, eds. Pancreatic Islet Transplantation Series. Vol l. Procurement of Pancreatic Islets. Austin: RG Landes Co. p l23-133.

Wright JR Jr. 2002a. Almost famous: E. Clark Noble, the common thread in the discovery of insulin and vinblastine. CMAJ 167:1391-1396.

Wright JR Jr. 2002b. From ugly fish to conquer death: JJR Macleod's fish insulin research, 1922-1924. Lancet 359:1238-1242.

Wright JR Jr, Bonen A, Conlon JM, Pohajdak B. 2000. Glucose homeostasis in the teleost fish tilapia: Insights from Brockmann body xenotransplantation studies. Am Zoologist 40:234-245.

Wright JR Jr, Kearns H. 1995a. Fish-to-mouse islet xenograft survival is markedly prolonged by l5-deoxyspergualin (DSG). Transplant Proc 27:294-295.

Wright JR Jr, Kearns H. 1995b. Long-term culture, low temperature culture, and hyperoxic culture do not prolong fish-to-mouse islet xenograft survival. Xenotransplantation 2:19-25.

Wright JR Jr, Kearns H, MacDonald AS. 1994a. Leflunomide and cyclosporin-A prolong fish-to-mouse islet xenograft survival in BALB/c mice. Transplant Proc 26:1310.

Wright JR Jr, Kearns H, Polvi S, MacLean H, Yang H. 1994b. Experimental xenotransplantation using principal islets of teleost fish (Brockmann bodies): Graft survival in selected strains of inbred mice. Transplant Proc 26:770.

Wright JR Jr, Kearns H, Yang H, Fraser RB, Colp P, Rowden G. 1997. Immunophenotyping fish-to-mouse islet xenograft rejection: A time course study. Ann Transplant 2:12-16.

Wright JR Jr, Pohajdak B. 2001. Cell therapy for diabetes using piscine islet tissue. Cell Transplant 10:125-143.

Wright JR Jr, Pohajdak B. 2002. Transgenic tilapia comprising a humanized insulin gene. US Patent No. 6,476,290 B1, issued 11/5/2002.

Wright JR Jr, Polvi S, MacLean H. 1992. Experimental transplantation using principal islets of teleost fish (Brockmann bodies): Long-term function of tilapia islet tissue in diabetic nude mice. Diabetes 41:1528-1532.

Wright JR Jr, O'Hali W, Yang H, Bonen A. 1998a. GLUT-4 deficiency and absolute peripheral resistance to insulin in the teleost fish tilapia. Gen Comp Endocrinol 111:20-27.

Wright JR Jr, Yang H, Dooley KC. 1998b. Tilapia--A source of hypoxia-resistant islets for encapsulation. Cell Transplant 7:299-307.

Xu B-Y, Yang H, Yu W, Wright JR Jr. 2003. Autoimmunity is not responsible for the rapid destruction of encapsulated islet xenografts by NOD mice. (Abstract). Xenotransplant 10:524.

Yang H, Al-Jazaeri A, Wright JR Jr. 2002a. The immunoprotective effect of Sertoli cells co-encapsulated with islet xenografts is not dependent upon Fas-ligand expression. Cell Transplant 11:799-801.

Yang H, Dickson B, O'Hali W, Kearns H, Wright JR Jr. 1997a. Functional comparison of mouse, rat, and fish islet grafts transplanted into diabetic nude mice. Gen Comp Endocrinol 106:384-388.

Yang H, Kearns H, Wright JR Jr. 1995. Immunomodulation with ionizing radiation does not prolong fish-to-mouse islet xenograft survival. Transplant Proc 27:3312-3313.

Yang H, McAlister VC, Al-Jazaeri A, Wright JR Jr. 2002b. Liposome-encapsulated tacrolimus produces better immunosuppressive effect in a discordant islet xenograft model. Transplantation 73:710-713.

Yang H, Morrison CM, Conlon JM, Laybolt K, Wright JR Jr. 1999. Immunocytochemical characterization of the pancreatic islet cells of the tilapia (Oreochromis niloticus). Gen Comp Endocrinol 114:47-56.

Yang H, O'Hali W, Kearns H, Wright JR Jr. 1997b. Long-term function of fish islet xenografts in mice by alginate encapsulation. Transplantation 64:28-32.

Yang H, Wright JR Jr. 1995. A method for mass harvesting islets (Brockmann bodies) from teleost fish. Cell Transplant 4:621-628.

Yang H, Wright JR Jr. 1999. Co-encapsulation of Sertoli enriched testicular cell fractions further prolongs fish-to-mouse islet xenograft survival. Transplantation 67:815-820.

Yang H, Wright JR Jr. 2002. Alginate encapsulation (Ca⁺²-induced gelation). In: Atala A, Lanza R, eds. Methods of Tissue Engineering. San Diego: Academic Press. p 787-801.

Yu W, Xu B-Y, Wright JR Jr. 2003. Directly vascularized pancreatic islet xenotransplantation: Is tilapia-to-nude mouse discordant? (Abstract). Xenotransplant 10:499.