Intrathymic Islet Cell TransplantationReduces p-Cell Autoimmunity andPrevents Diabetes in NOD/Lt MiceIVAN C. GERLING, DAVID V. SERREZE, SHERRI W. CHRISTIANSON, AND EDWARD H. LEITER

Intrathymic transplantation of syngeneic islets intoadolescent NOD/Lt mice was performed to establishwhether the thymus would serve as animmunoprivileged site for p-cell engraftment, andwhether this treatment would prevent the developmentof diabetes by eliciting tolerance to islet antigens.Intrathymic injection of cells from 200 NOD islets into4-wk-old female NOD/Lt mice produced a significantreduction in the severity of insulitis at 24 wk of age.Furthermore, diabetes development was stronglysuppressed (11% incidence) compared with controls(100% incidence). Both thymus histology and thymicinsulin content revealed a rapid loss of the implantedp-cells with < 1 % remaining 1 wk posttransplantation.Despite the rapid loss of thymus-implanted islet cells,evidence for tolerance induction to islet cell antigenswas obtained by adoptive transfer of splenicleukocytes from these mice into NOD-scidlscidrecipients. After adoptive transfer of splenicleukocytes from 24-wk-old untreated prediabeticdonors, 4 of 5 MOD-scidlscid recipients developeddiabetes within 4 wk, and none of the recipientsbecame diabetic after transfer of splenocytes fromintrathymic islet-implanted donors. Intrathymic islettransplantation did not lead to reduction of sialitis infemales with reduced severity of insulitis, indicatingthat the protective effect was tissue specific. This alsowas reflected in adoptive transfer experiments,because equal severity of sialitis was observed inNOD-sc/d/sc/d recipients of spleen cells from eitherislet transplanted or control NOD/Lt mice. Inconclusion, the data suggest that intrathymic injectionof islet cells prevents diabetes by stimulatingimmunological tolerance to p-cells. Diabetes41:1672-76,1992

From The Jackson Laboratory, Bar Harbor, Maine.Address correspondence and reprint requests to Edward H. Leiter, The

Jackson Laboratory, Bar Harbor, ME 04609.Received for publication 1 September 1992 and accepted 17 September

1992.STZ, streptozocin; APC, antigen-presenting cell; PBS, phosphate-buffered

saline; RIA, radioimmunoassay.

Recent studies have demonstrated that intrathy-mic transplantation of allogeneic islets can re-duce hyperglycemia in STZ-induced diabeticrats (1). Moreover, intrathymic engraftment ofsyngeneic islets into prediabetic, diabetes-prone BB ratsprotects against diabetes development (2,3). T-cellsmaturing within the thymus normally are rendered toler-ant to self-antigens through interactions with both he-matopoietically derived APCs and thymic epithelial cells(4). The generation of p-cell autoreactive effectors in BBrats results in part from defects expressed at the level ofintrathymic APCs (5). Hence, intrathymic transplantationof prediabetic BB rats with syngeneic islets may overridethis defect by rendering T-cells that are maturing in thethymus tolerant to p-cell autoantigens, thus blocking thedevelopment of diabetes (3). Indeed, intrathymic trans-plantation of allogeneic islets allows overtly diabetic BBrats to retain donor-matched islets that are subsequentlyengrafted intrahepatically (1). However, these studies inBB rats did not conclusively demonstrate that the mech-anism of protection was by specific induction of T-celltolerance to pancreatic p-cells. Indeed, the finding that 8of 23 diabetes-prone BB rats still became diabetic,despite the retention of syngeneic islets engrafted intothe thymus (2), suggested the possibility that the thymussimply may be an immunoprivileged site in BB rats, fromwhich normoglycemia is maintained by insulin secretedfrom the intrathymic islet grafts.

Although the BB rat is severely T-lymphopenic,whereas the NOD mouse has an abundance of periph-eral T-cells, APCs from NOD mice also fail to presentp-cell autoantigens in a completely tolerogenic fashion(6,7). Hence, APC dysfunctions likely underlie the gen-eration of p-cell autoreactive effectors in both models.Therefore, this study was conducted to evaluate if thethymus also is an immunoprivileged site for engraftmentof syngeneic islets in NOD/Lt mice, and if intrathymic

1672 DIABETES, VOL. 41, DECEMBER-1992

I.C. GERLING AND ASSOCIATES

implantation of NOD islets renders the developing T-cellpopulation tolerant to p-cell autoantigens and, therefore,prevents development of insulitis and diabetes.

RESEARCH DESIGN AND METHODSHusbandry conditions for the NOD/Lt colony and thedevelopment of the NOD congenic stock homozygous forthe severe combined immunodeficiency (scid) mutation,now designated NOD/LtSz-sc/tf, have been describedpreviously (8). NOD-scid/scid mice do not develop func-tional T- or B-cells, and, therefore, are insulitis- anddiabetes-free. NOD-scid/scid females at the 10th back-cross generation (following outcross from the C.B-17-scid/scid strain) were used as splenocyte recipients inthe adoptive transfer studies described below.Preparation and transplantation of islet cells. Isletswere isolated from 6-7-wk-old NOD/Lt females as de-scribed previously (9) and treated for 5 min with 0.05%trypsin and 0.5mM EDTA in Hanks' balanced salt solutionto yield islet cell suspensions. Trypan blue dye exclusionindicated a viability of -90% in the cell suspensions. Thecells were washed and resuspended in 0.1 ml PBS/1000islets, and 0.01 ml of this suspension was injected intoeach thymic lobe (total of 200 islets/thymus) of 28-29-day-old female NOD/Lt mice that were anesthesized withtribromoethanol (250 mg/kg, i.p.). Controls were injectedwith 0.01 ml PBS in each thymic lobe.Analysis of diabetes development. Female NOD/Ltmice were tested weekly for glycosuria by using Tes-Tape (Lilly, Indianapolis, IN). Mice were classified asdiabetic when glycosuria of 3+ was detected. Aglycosu-ric mice were considered prediabetic if, after an over-night fast, their plasma glucose concentration was >27.8mM, 60 min after injection of 3 g glucose/kg body weight.Plasma glucose levels were determined with a GlucoseAnalyzer II (Beckman, Palo Alto, CA) using blood fromthe retro-orbital sinus.

At the indicated times, mice were killed; —50% of eachpancreas and each thymic lobe were fixed in Bouin'ssolution, and granulated p-cells were stained with alde-hyde fuchsin. All islets sampled from three nonoverlap-ping pancreatic levels were assigned an insulitis scoreas follows: 4, end stage; 3, >25% of the islet infiltrated; 2,

INTRATHYMIC ISLET CELL TRANSPLANTATION

TABLE 1Antidiabetogenic effect of intrathymic islet transplantation in 24-wk-old female NOD/Lt mice

Pancreatic insulinTreatment Mouse Diabetic Insulitis index (pM/jig protein)

None

Intrathymic islets

NoNoNoYesYesYes

NoNoNoNoNoNoNoNoYes

0.89 ± 0.021.00 ±0.000.94 ± 0.03

ND*ND*ND*

0.94 ± 0.04f

0.12 ±0.020.31 H0.15 d0.10 d0.88 d0.90 d0.23 d0.30 d1.00 d0.44 d

t0.03t0.03t0.04t0.03t 0.01fc0.04t0.05tO.OOfcO.13tt

25512232644

42093

210±66t

2700174027817140891548

16372667621

2303 ± 713t*

Values are means ± SE.*Not determined (insulitis index of chronically diabetic females usually is 1.0).tMean of the means for individual sample determinations.jlntrathymic islet-treatment group means are significantly different from nontreated control group at P < 0.05 by Student's t test.

eating the presence of significant residual p-cell mass.Because all vehicle-injected controls depicted in Fig. 1had developed diabetes and were, therefore, necropsiedbefore 24 wk, aglycosuric age-matched NOD/Lt femaleswith impaired glucose tolerance by the criteria describedabove were identified in our colony and used as controls.The mean insulitis index for this group of age-matchedunmanipulated prediabetic NOD Lt females (mean,0.94 ± 0.04) was significantly higher than in intrathymi-cally transplanted mice, correlating with significantlylower pancreatic insulin contents (Table 1).

Intrathymic injection of trypsin/EDTA dispersed synge-neic submandibular gland cells into 5 28-day-oldNOD/Lt females, unlike dispersed islet cells, failed to

==* 800-,

| 700-

H 600-

- 500-z3 400-ẑ 300-

co? 200-O2 100-% o"

770±254

42.4110.2| T | 6.62±2.12 4.44±2.18

0 2 7 14

DAYS POST ISLET TRANSPLANTATION

FIG. 2. Insulin content In thymlc extracts from NOD/Lt recipients ofIntrathymic Islet cell transplants. The thymus was removed within5-10 mln (day 0), or at 2, 7, and 14 days postinjection of islet cells.Means ± SE (for 2 -4 mice) are Indicated.

prevent diabetes, with 3 of 5 recipients (60%) diabetic by18 wk.Survival of p-cells in the thymus. Marked declines inthymic insulin content within 48 h postimplantation indi-cated a rapid loss of the injected islet cells from thethymus (Fig. 2). Within 1 wk posttransplantation, theinsulin content of the thymus was < 1 % of the initial (day0) content. Histological examination confirmed the rapidelimination of the islet syngrafts. At day 0, most islet cellswere found to be dispersed within the thymic subcapsule(Fig. 3/4). By 48 h, the numbers of granulated p-cellsremaining in this location were markedly reduced (Fig.38), such that by day 7, only an occasional granulatedp-cell was detected (not shown). Elimination was notlimited to p-cells, because the non-p-islet endocrinecells, well represented at day 0, also were lost concom-itant with the p-cell population (not shown). At 24 wk (20wk postimplantation), no surviving islet cells were detect-able in thymuses of mice described in Fig. 1 and Table 1.The insulin content in the thymus extracts of most of theislet cell transplant recipients was below the detectionlimit of 0.015 nmol/ml. Only 1 mouse had a thymus insulincontent >0.030 nmol/ml at 24 wk of age.Adoptive transfer of spleen cells to NOD-scidlscidmice. Four of five NOD-scid/scid recipients of pooledsplenocytes from 3 prediabetic age-matched controls(control mice 1-3 from Table 1) developed diabeteswithin 4 wk posttransfer. The one nondiabetic NOD-scid/scid recipient of control splenocytes exhibited 1 +

glucosuria by Tes Tape at wk 5 and an insulitis index of0.965, indicating severe islet cell damage and incipientdiabetes. In contrast, adoptive transfer of pooled sple-nocytes from 3 of the islet cell-transplant recipients (isletrecipient mice 1-3 in Table 1) failed to transfer diabetes

1674 DIABETES, VOL. 41, DECEMBER 1992

I.C. GERLING AND ASSOCIATES

FIG. 3. Insulin staining (dark) cells In thymus of Intrathymic Islet celltransplant recipients. Thymus was removed either (A) 5-10 min afterIslet cell Injection (x475) or (0) 48 h after Islet cell Injection (x950).

to any of 5 NOD-sc/d/sc/d recipients by 5 wk posttrans-fer. Reduced numbers of p-cell autoreactive T-cells inthe spleens of the islet-transplanted group was con-firmed by the reduced insulitis index (0.307 ± 0.017) inNOD-sc/d/sc/d recipients of splenocytes from these micecompared with the group receiving splenocytes from non-islet-implanted mice (insulitis index, 0.991 ± 0.005), P<0.001.Inflammation in submandibular glands. Histologicalexamination of the submandibular salivary glands ofNOD/Lt females aged to 24 wk revealed widespread,chronic sialitis in all mice without any noticeable differ-ence between recipients of islet transplant and controls(data not shown). Unmanipulated NOD-sc/d/sc/d miceare free of sialitis. All NOD-sc/d/sc/d recipients of sple-nocytes also developed focal sialitis within the 5-wkpostadoptive transfer. No difference in the severity of

sialitis was detected between recipients of spleen cellsfrom islet-transplanted mice versus recipients of spleencells from non-islet-implanted mice (data not shown).

DISCUSSIONAs in BB rats (2,3), intrathymic transplantation of synge-neic islet cells protects NOD mice from development ofdiabetes. However, in contrast to BB rats, we found thatsyngeneic islet cells were rapidly cleared from the thy-mus of NOD/Lt mice. This finding clearly eliminatesectopic secretion of insulin as the mechanism behind theprevention of diabetes in NOD/Lt recipients of intrathymicislet cells. Although p-cell rest by prolonged insulintherapy started at an early age can prevent diabetes inNOD mice (11), the rapid clearance of islet cells from thethymus in our experiments precludes this mechanism.The difference in p-cell survival in BB rats and NOD/Ltmice could be attributable to our use of dispersed isletcells instead of whole islets, or it could reflect differencesin the intrathymic environment between NOD mice andBB rats. The rapid loss of dissociated islet cells from thethymus of autoimmune NOD mice is not unique to thisstrain; syngeneic islet cells are eliminated from the thy-mus of nonautoimmune CBA/J mice, in a time coursesimilar to that for NOD/Lt (data not shown). Although theincidence of diabetes was markedly reduced in NOD/Ltrecipients of intrathymic islet cells compared with vehi-cle-injected controls at 24 wk of age, insulitis was notcompletely circumvented. The variable levels of bothinsulitis and pancreatic insulin content in the islet cell-transplant recipients indicate that this protocol inhibits,but does not completely block, the development ofp-cell-autoreactive T-cells within the thymus. This con-clusion is supported further by the finding that althoughthey retained granulated p-cells and remained diabetes-free, NOD-sc/d/sc/d recipients injected with spleen cellsfrom NOD mice that had received intrathymic islet celltransplants were not totally free of insulitis. NOD-sc/d/scid mice do not develop diabetes or insulitis spontane-ously, but diabetes can be transferred adoptively by i.v.injections of splenic T-cells from NOD/Lt mice (12). Theseverity of sialitis was not affected in recipients of in-trathymic islet cell transplantation, indicating that theeffect of the intrathymic islet cell transplants was p-cellspecific. Similarly, the severity of sialitis was equivalent inNOD-sc/d/sc/d females injected with splenocytes fromboth islet cell-transplant recipients and NOD/Lt femalecontrols.

Because APCs from NOD mice appear to be defectivein tolerogenic presentation of p-cell autoantigens toT-cells (6,7), the processing of islet cell antigens directlyin the thymus may compensate for the APC functionaldefects. However, the results of our adoptive transferexperiments indicate that the single implantation of 200islets into the thymus of prediabetic NOD mice wasinsufficient to completely eliminate the development ofp-cell autoreactive T-cells. Further studies are needed toidentify the optimal conditions required to induce andmaintain immunological tolerance to p-cells by intrathy-mic islet transplantation and to identify the autoantigensinvolved.

DIABETES, VOL. 41, DECEMBER 1992 1675

INTRATHYMIC ISLET CELL TRANSPLANTATION

ACKNOWLEDGMENTSThis study was supported by National Institutes of HealthGrants DK-27722 and DK-36175.

The authors are grateful for the excellent technicalassistance of Anne Higgins, Bruce Regimbal, and DanielKrull. We are indebted to Dr. Leonard D Schultz forproviding the NOD-scid/scid mice.

REFERENCES1. Posselt AM, Naji A, Roark JH, Markman JF, Barker CF: Intrathymic

transplantation in the spontaneously diabetic BB rat. Ann Surg214:365-71, 1991

2. Koevary SB, Blomberg M: Prevention of diabetes in BB/Wor rats byintrathymic islet injection. J Clin Invest 89:512-16, 1992

3. Posselt AM, Barker CF, Friedman AL, Naji A: Prevention of autoim-mune diabetes in the BB rat by intrathymic islet transplantation atbirth. Science 256:1321-24, 1992

4. Sprent J, Webb SR: Function and specificity of T-cell subsets in themouse. Adv Immunol 41:39-133, 1987

5. Georgiou HM, Bellgrau D: Thymus transplantation and diseaseprevention in the diabetes-prone Bio-Breeding rat. J Immunol142:3400-405, 1989

6. Serreze DV, Leiter EH: Development of diabetogenic T-cells fromNOD/Lt marrow is blocked when an allo-H-2 haplotype is expressedon cells of hematopoietic origin, but not on thymic epithelium. JImmunol 147:1222-29, 1991

7. Leiter EH, Serreze DV: Antigen presenting cells and the immuno-genetics of autoimmune diabetes in NOD mice. Regional Immunol.In press

8. Prochazka M, Gaskins HR, Shultz LD, Leiter EH: The nonobesediabetic scid mouse: model for spontaneous thymomagenesis as-sociated with immunodeficiency. Proc NatlAcad Sci USA 89:3290-94, 1992

9. Serreze DV, Leiter EH, Worthen SM, Shultz LD: NOD marrow stemcells adoptively transfer diabetes to resistant (NOD x NON) F1mice. Diabetes 37:252-55, 1988

10. Schwizer RW, Leiter EH, Evans R: Macrophage-mediated cytotox-icity against cultured pancreatic islet cells. Transplantation 37:539-44, 1984

11. Atkinson MA, Maclaren NK, Luchetta R: Insulitis and diabetes inNOD mice reduced by prohylactic insulin therapy. Diabetes 39:933-37, 1990

12. Christianson SW, Shultz LD, Leiter EH: Adoptive transfer of diabetesinto immunodeficient NOD scid/scid mice: relative contributions ofCD4+ and CD8+ T lymphocytes from diabetic versus prediabeticNOD. NON Thy1a donors. Diabetes. In press.

1676 DIABETES, VOL. 41, DECEMBER 1992