Brief Communication

Gala1,3Gal expression on porcine pancreaticislets, testis, spleen, and thymus

Introduction

Gala1,3Gal (Gal) epitopes are the major target fornatural antibodies causing hyperacute rejection ofdiscordant xenografts in human or non-humanprimates [1]. Oriol et al. [2] and McKenzie et al. [3]described the distribution of Gal in different pigtissues, both groups concentrating on tissues andorgans of relevance to the potential transplantationof pig organs or tissues into humans. Subsequentstudies have been made by others [4,5]. Gal hasconsistently been shown to be expressed on thevascular endothelium in all tissues throughout thebody, particularly on arterioles, capillaries andvenules. At other sites within organs Gal isdistributed heterogeneously [2–5].We have assessed Gal expression in four porcine

organs – the pancreas (since its expression in thisorgan is of interest for the clinical xenotransplanta-tion of islets), testis (in view of the growing interestin Sertoli cells as ameans of providing local immuneprotection to other donor-specific cells [6–8]), spleen(in view of the potential of this organ to induce astate of tolerance after transplantation [F.J.M.F.Dor, et al. unpublished data]), and thymus (which

does not appear to have been investigated previ-ously and is being explored for its potential to inducexenogeneic immune tolerance [9]). Although pigsnegative for Gal expression may be available forxenotransplantation research in the near future, wehere record the results of this small study.

Materials and methods

Animals

Twenty-two pigs, 1 to 22 months old (age distri-bution in Fig. 1), were selected from the Massa-chusetts General Hospital (MGH) herd of majorhistocompatibility complex (MHC)-inbred minia-ture swine (n ¼ 19) and from the Imutran herd ofcross-breed Large White/Landrace pigs transgenicfor human decay-accelerating factor (n ¼ 3).

Tissues

Biopsies of pancreas, testis, spleen, and thymuswere obtained from newly euthanized pigs, andfixed in 10% formalin immediately. Fixed porcinetissues were dehydrated from 70% EtOH gradually

Dor FJMF, Cheng J, Alt A, Cooper DKC, Schuurman H-J. Gala1,3Galexpression on porcine pancreatic islets, testis, spleen, and thymus.Xenotransplantation 2004; 11: 101–106. � Blackwell Munksgaard, 2004

Abstract: Gala1,3Gal (Gal) is the first target in antibody-mediatedrejection of pig-to-non-human primate xenograft. Its expression mayvary between organs and constituents of organs. Gal expression wasstudied in pancreas, testis, spleen and thymus of 22 pigs, with agesranging from 1 to 22 months. The immunoperoxidase technique usingthe biotinylated lectin, Griffonia simplicifolia (IB4), was used. In thepancreas, neither endocrine (islet cells) nor exocrine cells expressed Gal.The Sertoli cells in the testis were negative. The spleen capsule andtrabeculae did not stain for Gal, although both splenic T and Blymphocytes expressed Gal (B > T). Thymocytes were weakly positive,whereas thymic epithelial cells were negative for Gal. No age-relateddifferences were seen in any tissues. Porcine islets of Langerhans, Sertolicells, and the splenic and thymic structural frameworks did not expressGal, and therefore, should be relatively resistant to anti-Gal antibody-mediated rejection. The availability of pigs deficient in Gal as a source ofislets may therefore not be beneficial in extending islet graft survival innon-human primate models.

F. J. M. F. Dor,1 J. Cheng,2 A. Alt,2

D. K. C. Cooper1 andH-J. Schuurman21Transplantation Biology Research Center,Massachusetts General Hospital/Harvard MedicalSchool, Boston, MA, USA, 2ImmergeBiotherapeutics, Inc., Charlestown, MA, USA

Key words: Gala1,3 Gal epitopes – pancreaticislets – pig – spleen – testis – thymus –xenotransplantation

Address reprint requests to H-J. Schuurman, Ph.D.,Immerge Biotherapeutics, Inc., 300 TechnologySquare, Cambridge MA 02139 USA(E-mail: Henk.Schuurman@immergebt.com)

Received 6 January 2003;Accepted 21 February 2003

Xenotransplantation 2004: 11: 101–106Printed in UK. All rights reserved

Copyright � Blackwell Munksgaard 2004

XENOTRANSPLANTATION

101

to 100% EtOH, cleared with Xylene, embedded inparaffin, and cut into 4-lm-thick sections. Thesections were then deparaffinized with Xylene andrehydrated with EtOH from 100% EtOH graduallyto water.

Hematoxylin and eosin staining

The rehydrated sections were stained with Hema-toxylin-2 (Richard-Allan Scientific, Kalamazoo,MI, USA) for 15 min, rinsed in water, incubatedin eosin solution for 5 min, then dehydrated andmounted with Cytoseal 60 Mounting medium(Richard-Allen Scientific).

IB4 lectin staining

To detect Gal epitopes, the immunoperoxidasetechnique using the biotinylated lectin, Griffoniasimplicifolia (IB4), was used [2,10]. The rehydratedsections were treated in BorgDeclocker solution(Biocare Medical, Walnut Creek, CA, USA) for3 min at 120 �C in a pressure cooker (BiocareMedical) and incubated with biotinylated IB4lectin (Vector Laboratories, Burlingame, CA,USA) for 30 min. After washing with phosphate-buffered saline, the sections were incubated withavidin/biotin peroxidase complex (Vectastain EliteABC Kit, Vector Laboratories) for 30 min, visual-ized by using 3,3¢-diaminobenzidine substrate kit(Vector) and counterstained with Hematoxylin.

Keratin staining

The staining procedure is similar to that describedabove. Cytokeratin monoclonal antibody was used(Clone AE1/AE3, Dako Corporation, Carpiteria,CA, USA). Sections were treated with Pepsin

(Dako) for 5 min after being heated for 3 min at120 �C in the pressure cooker (Biocare Medical).

Results

There were no apparent differences in Gal expres-sion between the MGH and Imutran pigs. Galexpression on the various structures in each organis indicated in Table 1 and illustrated in Figs 2–5.

Pancreas (Fig. 2A, B)

All endocrine (alpha and beta) islet cells andexocrine cells were negative for Gal staining,whereas centro-acinar cells were weakly positive,and interstitial fibroblasts, endothelial cells, inter-lobular ducts, and vascular endothelial cells werepositive. Within the neonatal and adult population,no age-related differences in staining could be seen.

Testis (Fig. 3A, B)

In the undeveloped testis, in which there was noevidence of spermatogenesis, the Sertoli cells, as well

0 2 4 6 8 10

1–3

3–6

6–12

>12A

ge

(m

on

th)

Number of pigs

Fig. 1. Age distribution of the 22 pigs in this study.

Table 1. Differential expression of Gal in various pig organs

Tissue Location Intensity of Gal staining

Pancreas Islet alpha cells )Islet beta cells )Exocrine cells )Centro-acinar cells +Fibroblasts ++Interlobar ducts +Vascular endothelium ++

Testis Sertoli cells )Interstitial cells +Vascular endothelium ++

No spermatogenesis Cells inside seminiferous tubules )

Spermatogenesis Spermatids +Spermatogonia +

Spleen Capsule )Trabeculae )Stroma +Smooth muscle fibers ++White pulp

– Germinal centers +– T cells +– B cells ++– Vascular endothelium ++

Red pulp– Macrophages ++– Dendritic cells ++

Thymus Epithelial cells )Hassall's corpuscles )Thymocytes +Macrophages ++Dendritic cells ++Vascular endothelium ++

), negative; +, weakly positive; ++, strongly positive.

Dor et al.

102

as all other cells inside the seminiferous tubules,werenegative for Gal staining. Vascular endothelial cellsand interstitial cells, however, stained positive. Intestes from sexually mature animals, spermatogen-esis was seen; spermatids and spermatogonia insidethe seminiferous tubules, stained weakly positive,but Sertoli cells were negative for Gal.

Spleen (Fig. 4A, B)

The splenic capsule and trabeculae did not expressGal. The white pulp and vascular endothelial cellsstained strongly positive. Both B- and T-cell areasstained positive, but T lymphocytes in the peri-arteriolar lymphocyte sheaths exhibited weaker

staining than B lymphocytes in follicles. Germinalcenters were positive. In the red pulp, macrophagesand dendritic cells stained strongly positive.Stroma also gave a positive reaction. A pro-nounced staining was seen in smooth muscle fibersof the spleen.There was no change in the intensity and pattern

of staining between neonatal or adult spleens.

Thymus (Fig. 5A, B)

All vascular endothelial cells stained positive.Epithelial cells identified by staining for keratinwere negative for Gal, including those in andaround the Hassall’s corpuscles. Macrophages and

A B

Fig. 3. (A) IB4-lectin staining of testis in a 22-month-old pig, in which spermatogenesis is seen in the seminiferous tubules (ST) (·20).Note negative staining for Gal in Sertoli cells (SC), whereas both interstitial cells (Leydig cells; LC) and the vascular endothelium (VEin Fig. 2B, not indicated in 2 (A) do express Gal epitopes. Spermatids and spermatogonia (Sp) also stain positive. (B) IB4-lectinstaining of testis in a 1-month-old pig (·20) in which no spermatogenesis is seen. Note the difference in staining for the elements insidethe seminiferous tubules.

A B

Fig. 2. (A) H&E staining of porcine pancreas (magnification ·20), in which pancreatic islets are identified (I). (B) IB4-lectin stainingof the same sample of porcine pancreas (·20). Note the absence of staining for the Gal epitope in both endocrine and exocrinepancreatic cells. Staining for Gal can be seen in centro-acinar cells (CAC), fibroblasts (F), vascular endothelium (VE) and interlobarducts (not shown).

Gala1, 3Gal expression on porcine tissues

103

dendritic cells stained positive. Although thymo-cytes were positive, staining was much weaker thanon macrophages. There was no variation within theage range of the pigs we examined.

Discussion

Porcine pancreatic islets are of interest for theirpotential application in the treatment of diabetes inhumans. In the present study, pancreatic exocrineand endocrine tissues were found not to expressGal,and hence isolated islets would be predicted to berelatively protected from anti-Gal antibody-medi-ated injury when transplanted from pig to human.Oriol et al. [2] found that there was some expression

of Gal on the pancreatic ducts, which we confirmed.Several groups have demonstrated a difference inGal expression on porcine islet cells between adultand fetal cells [11]. Islet cell clusters (ICC) expressGal but, as they mature, expression is lost [12]; it ispossible that islets are not fully developed in the fetalporcine pancreas, and that ICC contain a largenumber of non-beta cells. McKenzie et al. [11] andothers [13,14], however, couldnot detectGal on fetalor adult pancreatic endocrine cells, but foundexpression limited to intra-islet ductal and endot-helial cells. Perhaps this variability can be attributedto differences in pig age and strain [15], although wecould not identify age- or strain-related differences.The absence ofGal expressionon adult porcine islets

Fig. 4. (A) H&E staining of a section of pig spleen (·20). The different elements in the spleen are identified – white pulp (WP) with acentral arteriole (CA) in a Malphigian corpuscle, a germinal center (G), red pulp (RP), smooth muscle fibers (SMF) and trabeculae(Tr). (B) IB4-lectin staining on the same section of spleen (·20). White pulp in general stains positive for Gal, with B cell areas (BC)displaying stronger staining than T cell areas (TC). In the red pulp (RP), dendritic cells and macrophages stain positive. A strikingpositive staining is seen in smooth muscle fibers (SMF). Trabeculae (Tr) are negative for Gal, as is the splenic capsule (not shown).

A B

Fig. 5. (A) Keratin staining of porcine thymus (·20). A Hassall’s corpuscle (HC) is identified. (B) IB4-lectin staining of the samesection of thymus (·20). The Hassall’s corpuscle (HC) and epithelium are negative for Gal. Thymocytes (T), dendritic cells (DC) andmacrophages (M) stain positive.

Dor et al.

104

indicates that, after their transplantation into non-human primates, rejection is unlikely to be anti-Galantibody-mediated. Unfortunately, this suggeststhat islets from pigs deficient in Gal (homozygousfor a1,3-galactosyltransferase gene-knockout), thatare becoming available [16], may be equally vigor-ously rejected.There has been interest in xenogeneic Sertoli cell

transplantation as a constituent in the treatment ofParkinson’s disease [8], in combination with islettransplantation [7,17], and in reproductive biology.It has been postulated that Sertoli cells havenutritive, regulatory, trophic and immunosuppres-sive functions. There has been an interesting studyshowing that porcine Sertoli cells can survive long-term without immunosuppressive therapy in the ratbrain, which is an immunoprotected site [8]. It isbelieved that Sertoli cells secrete factors that maylead to local immune tolerance, such as Fas ligand,transforming growth factor beta, and clusterin,which are believed to have immunoprotective andmay be even tolerizing properties. McKenzie et al.[3] reported that all of the developing cells in thetestis (from Sertoli cells to spermatozoa) did notstain for IB4, although some of the basal cellsshowed weak staining. We confirm their findingthat Sertoli cells do not express the major xenoge-neic epitope, Gal, and that spermatozoa stain onlyvery weakly.We are currently investigating the potential of

spleen transplantation in the induction of allo-tolerance, with a view to extending this approach tothe pig-to-baboon model. Gal expression in thespleen has not been studied extensively to date.McKenzie et al. [3] reported thatGal expressionwasnot uniformly positive in the porcine spleen. In theirstudy, the capsule was weakly positive, and themoststriking expression was found on cellular compo-nents in the red pulp. Splenic trabeculae werenegative for Gal, and in the white pulp the peri-arteriolar lymphocyte sheats (T cells) were weaklyreactive. The endothelium with the central arteriolewas also positive, but not strongly. The present dataconfirm earlier studies on the distribution of Galexpression, (with the difference that in our studiesthe splenic capsule stained negative) and add furtherinformation on Gal expression on different func-tional and histological units of the spleen.To our knowledge, there has not been a detailed

histological report on the distribution of Gal inthe porcine thymus. Studies of pig-to-non-humanprimate vascularized thymus xenotransplantationare currently being undertaken in our center withthe aim being tolerance induction [9]. In earlierstudies, however, Gal expression on lymphocyteshad been investigated in the same miniature swine

herd used in the present study [18]; no difference inGal expression between haplotypes could be iden-tified. When CD4+ and CD8+ T cells wereexamined for Gal expression, larger (more mature)cells expressed higher levels of Gal. We documentthat the porcine thymic microenvironment ofepithelial cells lacks Gal expression, suggestingthat it will not be subject to hyperacute rejection,and therefore may continue to be able to play itsrole in the generation of the T-cell repertoire aftertransplantation into primates.In conclusion, Gal is a major target antigen in

pig-to-primate organ transplantation; in conjunc-tion with complement and components of thecoagulation cascade, natural anti-Gal antibodiescause hyperacute rejection due to the ubiquitouspresence of Gal on vascular endothelial cells. Withregard to non-vascularized cellular transplants andparenchymal components of vascular transplants,the expression of Gal is far more variable. Ourdata, and those of others, indicate that certainporcine cells and tissues might be much lesssusceptible to hyperacute rejection than vascular-ized organ transplants. The relatively short survi-val of pig islets currently being achieved in non-human primate models suggests that rejection isdirected against non-Gal antigens. This has majorimplications for the transplantation of islets takenfrom pigs deficient in Gal, where it might beanticipated that rejection will be equally severe andislet survival not significantly prolonged.

Acknowledgments

The authors thank Michel Awwad PhD and KenjiKuwaki MD for their comments on this manu-script. F.J.M.F. Dor MD is a recipient of grantsfrom the Prof. Michael van Vloten Fund, theNetherland–America Foundation, and the TerMeulen Fund of the Royal Netherlands Academyof Arts and Sciences. Work in our own laboratoryis supported in part by NIH Program Project 1PO1A145897 and by a Sponsored Research Agreementbetween the Massachusetts General Hospital andImmerge BioTherapeutics, Inc.

References

1. Good AH, Cooper DKC, Malcolm AJ et al. Identifica-tion of carbohydrate structures that bind human antipor-cine antibodies: implications for discordant xenograftingin humans. Transplant Proc 1992; 24: 559.

2. Oriol R, Ye Y, Koren E, Cooper DKC. Carbohydrateantigens of pig tissues reacting with human naturalantibodies as potential targets for hyperacute vascularrejection in pig-to-man organ xenotransplantation.Transplantation 1993; 56: 1433.

Gala1, 3Gal expression on porcine tissues

105

3. McKenzie IFC, Xing P-X, Vaughan HA et al. Distri-bution of the major xenoantigen [gal(a1-3)gal] for pig tohuman xenografts. Transpl Immunol 1994; 2: 81.

4. Alvarado CG, Cotterell AH, McCurry KR et al.Variation in the level of xenoantigen expression in porcineorgans. Transplantation 1995; 59: 1589.

5. Strokan V, Molne J, Svalander CT, Breimer ME.Heterogeneous expression of Gal alpha1-3Gal xenoanti-gen in pig kidney: a lectin and immunogold electronmicroscopic study. Transplantation 1998; 66: 1495.

6. KorbuttGS, Elliot JF, Rajotte RV. Cotransplantationof allogeneic islets with allogeneic testicular cell aggregatesallows long-term graft survival without systemic immu-nosuppression. Diabetes 1997; 46: 317.

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

8. Sanberg PR, Borlongan CV, Saporta S, CameronDF.Testis-derived Sertoli cells survive and provide localizedimmunoprotection for xenografts in rat brain. NatBiotechnol 1996; 14: 1692.

9. Barth RN, Yamamoto S, Vagefi PA et al. Xenogeneicvascularized thymic lobe transplantation: a novel strategyfor tolerance induction in a pig-to-baboon model(Abstract 166). XIX Int Cong Transpl Soc 2002.

10. Wood C, Kabat EA, Murphy LA, Goldstein IJ.Immunochemical studies of the combining sites of the twoisolectins A4 and B4, isolated from Bandeiraea simplicifoli.Arch Biochem Biophys 1979; 198: 1.

11. McKenzie IFC, Koulmanda M, Mandel TE, Xing PX,Sandrin MS. Pig-to-human xenotransplantation: the

expression of Gal alpha (1,3)Gal epitopes on pig islet cells.Xenotransplantation 1995; 2: 1.

12. Bennet W, Bjorkland A, Sundberg B et al. A com-parison of fetal and adult porcine islets with regard to Galalpha (1,3)Gal expression and the role of human immu-noglobulins and complement in islet cell cytotoxicity.Transplantation 2000; 69: 1711.

13. Mirenda V, Le Mauff B, Cassard A et al. Intact pigpancreatic islet function in the presence of human xeno-reactive natural antibody binding and complement acti-vation. Transplantation 1997; 63: 1452.

14. Rayat GR, Rajotte RV, Elliott JF, Korbutt GS.Expression of Gal alpha(1,3)Gal on neonatal porcine isletbeta-cells and susceptibility to human antibody/comple-ment lysis. Diabetes 1998; 47: 1406.

15. Heald KA, Carless N, Jay TR, Boucher N, Downing

R. Expression of the GALalpha(1-3)GAL epitope on pigislets. J Mol Med 1999; 77: 169.

16. Phelps CJ, Koike C, Vaught TD et al. Production ofa1,3-galactosyltransferase-deficient pigs. Science 2003;299: 411.

17. Valdes-Gonzalez RA, Elliot RB, Dorantes LM et al.Porcine islet xenografts can survive and function in type 1diabetic patients in the presence of both pre-existing andelicited anti-pig antibodies (Abstract 246). XIX Int CongTranspl Soc 2002.

18. Chae SJ, Kramer AD, Zhao Y et al. Lack of variation inalphaGal expression on lymphocytes in miniature swine ofdifferent genotypes. Xenotransplantation 1999; 6: 43.

Dor et al.

106