protection against autoimmune diabetes with oral insulin is

Protection Against Autoimmune Diabetes WithOral Insulin Is Associated With the Presence ofIL-4 Type 2 T-Cells in the Pancreas and Pancreaticlymph NodesCorinne Ploix, Isabelle Bergerot, Nicole Fabien, Stephanie Perche, Valerie Moulin, and Charles Thivolet

Oral administration of antigens has been proposed inthe prevention and treatment of autoimmune diseases.We reported that oral administration of 0.8 mg ofrecombinant human insulin to 6-week-old NOD miceevery other day for a month generated regulatory T-cellsthat were able to reduce the severity of insulitis and thepercentage of clinical diabetes in naive irradiatedrecipients when co-injected with diabetogenic T-cells. Inthe present study, immunohistochemical analysis of thepancreatic glands revealed that injection of T-cellsfrom insulin-fed mice upregulated the number of inter-leukin (IL)-4-secreting cells within the islets. Usingtwo strains of NOD mice congenic at the Theta, orThyl, locus, we observed a higher proportion of T-cellsfrom insulin-fed mice in both the spleen (7.73 ± 0.3 vs.5.57 ± 0.2%; P < 0.001) and the pancreatic lymph nodes(10.1 ± 0.8 vs. 7.2 ± 0.7%; P < 0.05) of cotransferredmice. By reverse transcription-polymerase chain reac-tion (RT-PCR) analysis, mice reconstituted with T-cellsfrom insulin-fed animals had detectable amounts of IL-4 mRNA, specifically in the pancreatic lymph nodes (8of 9 experimental mice vs. 1 of 9 control mice) and thepancreas (3 of 3 experimental mice vs. 0 of 3 controlmice). 7-Interferon mRNA was detectable in allcotransferred animals, but IL-10 mRNA and transform-ing growth factor p mRNA were undetectable. Theseresults suggested a shift from a T-helper 1 (Thl) to aTh2 pattern of cytokine expression and underlined therole of pancreatic lymph nodes in the protection.Repeated injections of 500 ug s.c. of anti-IL-4 mono-clonal antibody led to an accentuation of the severity ofislet infiltration and to the development of clinical dia-betes. We concluded that oral administration of insulincan induce the presence of regulatory T-cells in thepancreas and the corresponding draining lymph nodes,initiate the secretion of IL-4 in this microenvironmentsufficiently to suppress the activity of Thl autoreactiveT-cell clones, and ultimately provide protection againstautoimmune diabetes. Diabetes 47:39-44, 1998

From INSERM 449, Faculty de Medecine RTH Lae'nnec, Lyon, France.Address correspondence and reprint requests to Dr. C. Thivolet,

INSERM 449, Faculty de Medecine RTH Laennec, Rue Guillaume Paradin,69372 Lyon Cedex 08, France.

Received for publication 28 March 1997 and accepted in revised form 9September 1997.

FACS, fluorescence-activated cell sorter; FITC, fluorescein isothio-cyanate; IFN--y, -y-interferon; IL, interleukin; PBS, phosphate-bufferedsaline; PCR, polymerase-chain reaction; RT, reverse transcription; TGF-(3,transforming growth factor (3; Thl/2, T-helper 1/2.

IDDM is a T-cell mediated autoimmune disease local-ized to the endocrine pancreas that occurs sponta-neously in genetically predisposed individuals (1) andresults in mononuclear cell infiltration of the pancre-

atic islets and specific (3-cell destruction. As in other organ-specific autoimmune diseases, lack of tolerance to (3-cellantigens results in the generation of autoreactive T-cells thatare able to transfer adoptively the disease in both BB rats andNOD mice, two experimental models of spontaneous dia-betes with immunopathological features resembling thoseof the human disease (1,2). The immune system can beviewed as a dynamic balance in which T-cells are importantpositive or negative regulators of the immune responses.Rupture of the balance between autoreactivity and tolerancecan result in autoimmunity. Oral administration of self-anti-gens or peptides to induce oral tolerance has been used as atherapy for several allergic and autoimmune disorders tocompensate for lack of suppression (3). Although the exactmechanism of protection (i.e., clonal deletion, anergy, oractive suppression) may rely on the composition and dosageof the orally administered antigen (4), oral tolerance inexperimental models of autoimmunity has underlined theimportance of anti-inflammatory cytokines within the targetorgan. Several antigens have been identified in the patho-genesis of IDDM. Among them, insulin is specificallyexpressed in (3-cells and elicits both humoral and cellularimmune responses in NOD mice (5,6) and humans (7,8). Oraladministration of insulin has been shown to prevent diabetesin NOD mice (9). We have previously characterized CD4+

regulatory T-cells that reduce the ability of diabetogenic T-cellsto transfer diabetes during cotransfer experiments (10). Theaim of the present study was to characterize the mechanismsof active cellular suppression leading to diabetes preventionafter oral insulin treatment in the NOD mouse.

RESEARCH DESIGN AND METHODSMice. NOD mice were bred in our facilities under standard conditions. Diabeteswas diagnosed by glycosuria (Urine Chemstrips; Bayer Diagnostics, Germany) andpersistent hyperglycemia (Blood Glucose Chemstrips; Ufescan, Roissy, France).The incidence of diabetes in our colony reached 80% in females and 20% in malesby age 30 weeks. Diabetic females served as donors of autoreactive T-cells. Con-genic NOD-N Thyl, 1 mice initiated from a cross between NOD/Lt and a diabetes-resistant strain NON/Lt were obtained from Leiter (Bar Harbor, ME). Spontaneousincidence of diabetes was identical to Thyl,2 NOD mice.

DIABETES, VOL. 47, JANUARY 1998 39

Dow

nloaded from http://diabetesjournals.org/diabetes/article-pdf/47/1/39/363286/47-1-39.pdf by guest on 04 January 2022

IL-4 AND ORAL TOLERANCE IN NOD MOUSE

Oral insulin treatment. Six-week-old female NOD mice were fed 400 pi con-taining 20 units (0.8 mg) of human insulin (NovoNordisk, Baegsvaerd, Denmark)in phosphate-buffered saline (PBS) or PBS alone by gastric intubation with an 18-gauge stainless steel feeding needle. Mice were fed three times/week during 30days. Additional control experiments were performed using 0.8 mg of chicken oval-bumin (Sigma, St-Quentin Fallavier, France) administrated orally threetimes/week during 30 days.Cell preparation and cotransfer experiments. Splenocytes were isolated inHanks' balanced salt solution and treated with Tris-ammonium chloride solutionto eliminate red blood cells. T-cells were enriched by panning on plastic dishescoated with a rabbit anti-mouse IgG (H + L) antibody diluted 1:500 (Biosys, Com-piegne, France) eluting 20-25% of the initial cell population. More than 90% of non-adherent cells were from the Thyl,2+ phenotype during flow cytometric analysis.Next, 5 X 106 T-cells from diabetic donors were mixed in equal proportion withT-cells from treated mice and injected intravenously to irradiated (750 rads) 8- to10-week-old naive males. In separate experiments, nondiabetic recipient mice weretreated intraperitoneally with 200 mg/kg cyclophosphamide (Sigma) to testwhether the protection was linked to an active cellular mechanism of suppression,as previously described (11).

T-cell studies. One month after repeated oral administrations of 0.8 mg insulinor PBS, splenic T-cells from Thyl,2+ NOD female mice were mixed with dia-betogenic Thy-l,l+T-cells and injected into irradiated Thy-l,l+ NOD male recip-ients. To evaluate the influence of insulin feeding on the destination of regulatoryT-cells, the percentage of Thy-1,2+ T-cells was determined 30 days after cell trans-fer in the spleen, thymus, mesenteric, and pancreatic lymph nodes of recipient miceby fluorescence-activated cell sorter (FACS) analysis as described previously(12) using fluorescein isothiocyanate (FITC)-conjugated anti-rat Thyl,l+ mono-clonal antibody (clone MRC OX-7; Cedarlane, Ontario, Canada), anti-Thyl,2+ ratmonoclonal antibody (clone 30H12), and FITC-conjugated anti-rat IgG kappaantibody (Mark 1; Biosys, Compiegne, France). Double immunostainings of pan-creatic sections were performed using FITC-conjugated anti-rat Thyl,l+ mono-clonal antibody and phycoerythrin (PE)-conjugated anti-mouse Thyl,2+ mono-clonal antibody (clone 30H12; Sigma). T-cell subset analysis was performed byFACS using anti-CD4 (clone GK1.5) or anti-CD8 (clone 53-67) rat mAbs andFITC-conjugated anti-rat IgG kappa antibody.

Histological studies. Pancreatic and salivary glands were excised andprocessed for conventional histological studies after fixation in Bouin's alco-holic solution or directly frozen in liquid nitrogen. Next, 5-um sections werestained with hematoxylin-eosin as described previously (12). The severity ofinsulitis was scored on at least 25 islets for each specimen by two observers blindto the experimental protocol, according to the following scale: 0, when islet cellshad no visible sign of inflammation; 1, when islets had lymphocytes at the periph-ery or peri-insulitis; 2, when islets were mildly infiltrated (<40%); or 3, whenislets were severely infiltrated. Sialitis was scored on the following scale: 0, no lym-phoid infiltration; 1, mild infiltration; or 2, severe infiltration.Immunohistochemical detection of cytokines. Pancreatic glands wereremoved 30 days after cotransfer and stored at -80°C. Serial frozen sections (5um) were fixed in cold acetone for 10 min andp-formaldehyde for 15 min. Slideswere rehydrated in PBS with 0.1% saponin and then incubated overnight at 4°Cwith rat mAbs against mouse IL4 (clone 11B11; ATCC, Rockville, MD), mouse 7-interferon (IFN-7) (clone R46A2; Pharmingen, San Diego, CA), mouse IL-10(clone JES5-2A5; Pharmingen), or a rabbit anti-transforming growth factor (31 poly-clonal antibody (a gift from J. Saez, Lyon, France) diluted 1:50. Sections werewashed in PBS/0.1% saponin and incubated with a biotinylated rabbit anti-rat Igmouse-absorbed antibody (Vector, Burlingame, CA) diluted 1:200 or a biotinylatedswine anti-rabbit Ig antibody (Dako, Trappes, France) diluted 1:50 for 30 min atroom temperature. At this stage, endogenous peroxidase activity was quenched.Sections were then washed in PBS and incubated in 0.1% hydrogen peroxide/50%methanol in PBS for 15 min at room temperature and washed in PBS/0.1% saponin.The antibody-biotin complexes were detected with a streptavidin-horseradishperoxidase complex diluted 1:50 for 30 min (Amersham, U.K.). Enzyme activity wasrevealed with diaminobenzidine (Sigma), and sections were counterstained withhematoxylin. Cytokine studies were performed in two different series of experi-ments with 11 mice in each group. About 23 ± 2 islets for each mouse were exam-ined. Two different observers did the analysis.

RNA extraction and PCR. Active RNAs from pancreases and lymph nodes wereobtained with slight modifications of the method described by Chirgwin et al. (13).After ductal dilatation, pancreatic glands were digested in 2 mg/ml of collagen-ase P (Boehringer Mannheim, Meylan, France) DNAase (Sigma) at 37°C for 5 min.After filtration and sedimentation at unit gravity, islet-enriched cell preparationswere processed for RNA extraction. For all samples, the absorption ratio was1.7-2.0 (260-280 nm). Total RNA was dissolved in sterile water and stored at -80°Cuntil analysis. The first strand cDNA was synthesized in reaction mix containing1 ug of total RNA, 15 pmol of anti-sense primer (Eurogentec, Angers, France), 0.2mmol/1 each of dNTP, 0.9 mmol/1 MnCl2, and 5 units of thermostable reversetranscriptase (Tth DNA polymerase; Promega, Charbonnieres, France) in 1 X RT

buffer. The mixture was overlaid with mineral oil, and, after a 3-min incubationat 60°C followed by 20 min at 72°C, the samples were heated at 99°C for 5 min andthen quickly chilled on ice. The 20-ul sample of the RT reaction was used for cDNAamplification. PCR was carried out in 1 x chelating buffer supplemented with 0.2mmol/1 dNTP, 15 pmol of 5' and 3' specific primers, 1.5 mmol/1 MgCl2, and 2 unitsof Taq polymerase (Life Technologies, Cergy Pontoise, France) in a final volumeof 50 pi. The mixture was amplified after 37 cycles with denaturation at 94°C for1 min, hybridization for 1.5 min at 56°C for IL4 or 60°C for IFN-7, and elongationat 72°C for 1 min in the thermal cycler (Omnigene; Hybaid, Teddington, U.K.). ThePCR products were analyzed by agarose gel electrophoresis (2%) in the presenceof ethidium bromide (0.5 pg/ml). The control of the RT-PCR reaction was per-formed using a multispecific plasmidic RNA (a gift from D. Shire, Sanofi, Labege,France) corresponding to an amplicon length of 322 bp.Anti-IL-4 treatment. The day after cell transfer, recipients were injected with500-pg i.p. t.i.d. every other day anti-IL-4 mAbs diluted in PBS. The control groupswere the mice receiving PBS or an irrelevant rat IgG mAb (TIB122; ATCC). Dia-betes incidence and severity of insulitis were studied after 1 month of treatment.Murine antibodies against IL-4 were purified ascites from clone 11B11 (ATCC).Spleens from experimental animals were subjected to T-cell subset analysis byFACS using anti-Thyl,2+ (clone 30H12), anti-CD4+ (clone GK1,5), and anti-CD8+(clone 53-67) rat monoclonal antibodies and FITC-conjugated anti-rat IgG K anti-body (MARK-1; Biosys).

Statistical analysis. Diabetes incidences among groups were compared usinga two-tailed Wilcoxon's rank-sum test and Fisher's exact test. Degrees of insuli-tis were compared using Student's t test for unpaired samples. The level of signi-ficance was set a tP < 0.05.

RESULTS

Splenic T-cells from mice orally treated with insulin transferactive protection against diabetes. During cotransfer exper-iments, mice receiving a mixture of splenic T-cells frominsulin-fed mice and diabetogenic T-cells were protectedagainst diabetes at day 30 in comparison with controls (0 of6 vs. 5 of 6 diabetic mice; P < 0.01 using Fisher's exact test).In this experiment, severity of insulitis was significantlyreduced, whereas severity of sialitis was unaffected (Fig. 1).Additional experiments with longer periods of observationwere performed, indicating a significant delay of clinical dia-betes with only 2 of 6 diabetic mice (33%) in the insulin groupversus 9 of 12 (75%) in the control group developing diabetes60 days after cell transfer. Nonislet proteins were also givenorally as additional controls in a separate experiment. Micereceiving splenic T-cells from ovalbumin-fed mice were notprotected from diabetes: 3 of 5 (60%) and 5 of 5 (100%) micewere found diabetic at days 30 and 38, respectively, versus 1of 5 mice receiving T-cells from insulin-fed mice during thesame observation period. To evaluate the T-cell autoreactiv-ity of protected animals, we performed second transferexperiments using 7 X 106 splenic T-cells. As shown in Fig. 2,T-cells from three nondiabetic mice harboring T-cells frominsulin-fed mice were unable to transfer diabetes in com-parison with control mice (1 of 6 vs. 5 of 6; P = 0.04 usingFisher's exact test), which had a diabetes incidence curve sim-ilar to that of mice reconstituted with 2 X 106 diabetogenic T-cells. Injection of cyclophosphamide in all protected miceinduced diabetes 10 days later.Localization and distribution of regulatory T-cells dur-ing cotransfer experiments. To determine whether pro-tective T-cells generated by oral insulin feeding could migratepreferentially toward the pancreas, we performed cotransferexperiments using two strains of NOD mice congenic at theThyl locus, and studied the contribution of the two differentT-cell phenotypes 30 days after T-cell inoculation. As shownin Table 1, the percentage of Thyl,2+ T-cells originating fromorally treated mice was significantly higher in pancreaticlymph nodes and spleens of mice receiving splenic T-cells

40 DIABETES, VOL. 47, JANUARY 1998

Dow


C. PLOIX AND ASSOCIATES

-1,5

1 "egV)

0,5

FIG. 1. Organ specificity of the protection observed following oraladministration of insulin. T-cells from insulin-fed mice (2?) were ableto reduce the severity of insulitis ( • ) but not sialitis ( • ) in compar-ison with control mice (.4). **P < 0.01.

from insulin-fed mice than in those from the control group.This result corresponded to an increase in the number ofThyl,2+ T-cells in both pancreatic lymph nodes (1.42 ± 0.2 vs.1.01 ± 0.2 X 104; P < 0.05) and spleens (3.13 ± 0.29 vs. 2.16 ±0.25 X 106; P < 0.05). No difference was noted in the numberof Thyl,l+ T-cells (4 ± 0.25 vs. 3.56 ± 0.27 X 106), revealing anidentical distribution of committed T-cells in syngeneic irra-diated animals. Comparison of the percentages of thymic T-cells from both phenotypes revealed no difference amongexperimental mice. Histological analysis of the pancreaticislets of the recipients indicated a reduction in severe insuli-tis in animals reconstituted with T-cells from insulin-fedmice, but the proportions of Thyl,2+ and Thyl,l+ T-cells infil-trating the islets were identical during double-immunostain-ing experiments.In situ studies of cytokines. The cytokines present in theislets of the recipients were analyzed by immunocytochem-istry. Studies of mice reconstituted with T-cells from insulin-fed mice revealed that insulitis contained high levels of IL-4(Fig. 3), whereas staining was negative in the control group.In contrast, insulitis from control mice was positive for IFN-7, but was negative or weakly stained in the insulin group.Stainings for IL-10 or TGF-3 were negative in both situations.RT-PCR analysis revealed that IL-4 mRNA transcripts weresignificantly increased in pancreases from the insulin groupand absent in pancreases from the control group (Fig. 4). Onthe other hand, IFN-7 transcripts were detected in allcotransferred mice. Detectable levels of IL-4 mRNA werealso found in the pancreatic lymph nodes of mice from theinsulin group. As shown in Fig. 5, 8 of 9 mice in the insulingroup were positive for IL-4 transcripts, in contrast to 1 of 9mice in the control group. Interestingly, the one mouse thatwas negative for IL-4 mRNA expression in the insulin grouphad severe islet cell infiltration and was diabetic. TGF-(3 andIL-10 mRNAs were not detected in pancreatic glands andpancreatic lymph nodes of any mice. In the mesentericlymph nodes, no IL-4 mRNA was seen, but IFN-7 mRNAswere detected in all mice.Abrogation of oral tolerance by anti-IL-4 treatment. T-cells from insulin-fed animals were able to delay significantlythe occurrence of diabetes during cotransfer experiments, asshown in Fig. 6. By day 45, only 1 of 10 mice in the insulingroup developed diabetes in contrast to 10 of 12 in the con-trol group. Anti-IL-4 treatment accelerated diabetes in thecontrol group (four of five mice) and abrogated the protec-

20 30 40 50 60Days after cell transfer

FIG. 2. 7 x 106 splenic T-cells from irradiated NOD mice previouslycotransferred with diabetogenic T-cells and T-cells from insulin-fedmice (O) or PBS-fed mice ( • ) were injected into naive irradiatedrecipients. Incidence of diabetes was compared with incidence curvesobtained after the injection of 7 x 10° ( ) or 2 x 10° ( ) T-cellsfrom diabetic female mice.

tive effect of insulin feeding (1 of 10 in the sham-treated vs.9 of 11 in the anti-IL-4-treated insulin recipient groups; P <0.001). Histopathological analysis of the pancreatic glands wasconcordant with the clinical situation, with a significantincrease in severe insulitis (49.2 ± 7.6 vs. 11.1 ± 2.9%; P < 0.001)and decrease in the percentage of islets with peri-insulitis(15.87 ± 3.81 vs. 43.72 ± 3.73%; P < 0.001) during anti-IL4 treat-ment. The decrease in the percentage of normal islets did notreach statistical significance (8.58 ± 8.6 vs. 11.23 ± 4.3%). T-cell subset analysis of the spleen of experimental animalsrevealed no bias in the distribution of CD4+ and CD8+ T-cellsubsets.

DISCUSSIONIn the present study, we showed that oral administration ofinsulin delayed autoimmune diabetes in the NOD mouse dueto regulatory T-cells that can adoptively transfer the protection.

As in other autoimmune disorders, several lines of evi-dence have linked the pathogenesis of diabetes in NOD miceto the activation of CD4+ Thl T-cells that release predomi-nantly pro-inflammatory cytokines, such as WN-y, and to theinhibition of CD4+ Th2 cells, which release IL-4 and IL-10 (14).Studies of islet infiltrates of both NOD mice (15) and BB rats

TABLE 1Analysis of the repopulation of irradiated Thyl,l+ NOD-N recip-ients by Thyl,2+ T-cells 30 days after the inoculation of diabeto-genic Thyl,l+ T-cells and identical numbers of Thyl,2+ T-cellsfrom insulin- or PBS-fed NOD mice

Recipient Thy 1,1 mice

Feeding protocol ofdonor Thyl,2 mice

Insulin (%) PBS (°A

SpleenThymusPancreatic lymph nodes

7.73 ± 0.3366.24 ± 4.3210.11 ±0.78

5.57 ± 0.21 <0.00168.83 ± 1.58 NS7.22 ± 0.71 <0.05

Data are means ± SD of six individual mice from two independentexperiments.


Dow


IL-4 AND ORALTOLERANCE IN NOD MOUSE

FIG. 3. Immunohistochemical studies of insulitis from cotransferredmice. Islet infiltrates from animals reconstituted with diabetogenic T-cells and T-cells from PBS-fed mice {a,b~) were positive for IFN-7 (o)but negative for IL-4 (6), whereas in the insulin group (c,tf) insulitis(-») was weakly stained for IFN-7 (c) but strongly positive for IL-4 (d).Original magnifications 250 x.

(16) have shown a positive correlation between IFN-7mRNA levels and autoimmune (3-cell destruction, whereasincreased IL-4 mRNA levels were found in the islets of non-diabetic animals. Experimental conditions known to pro-tect NOD mice from diabetes, such as complete Freund'sadjuvant (CFA) administration (17) or insulin immuniza-tions (18), have been associated with reduction of IFN-7mRNA expression. In vivo administration of recombinant LL-4 in prediabetic NOD mice protected them from diabetes(19), and transgenic NOD-IL-4 mice are completely pro-tected from insulitis and diabetes (20). A possible explana-tion for the protective effects of IL-4 may be the inhibitionof IL-2 secretion by CD8+ TCI cells present in the islet infil-trate (21). Committed TCI cells present in the spleen of dia-betic NOD mice could be regulated in similar ways. However,the role of IL-10, another Th2 cytokine, seems to be more con-troversial. Although IL-10 administration has been reportedto potentiate the protective effects of IL-4 in inhibiting therecurrence of autoimmune diabetes in syngeneic islet-trans-planted NOD mice (22), pancreatic expression of IL-10 in con-junction with NOD major histocompatibility complex(MHC) homozygosity did not protect against insulitis or dia-betes, and development of insulitis in NOD mice could beinhibited by anti-IL-10 antibody treatment (23).

W

B

w

Insulin PBSFIG. 4. RT-PCR analysis of IL-4 (.4) and IFN-7 (£) mRNA in the pan-creatic glands of cotransferred mice, 30 days after the inoculation ofdiabetogenic T-cells and T-cells from insulin-fed or PBS-fed mice. Theamplicon length of the RT-PCR control (column C) was 322 bp. All pan-creatic glands tested from the insulin group were positive for IL-4 (216bp) and IFN-7 (227 bp). The W column corresponds to DNA molecu-lar weight markers.

Feeding antigens can cause an antigen-specific reduction ofT-cell responsiveness through the generation of regulatory T-cells that produce anti-inflammatory cytokines. In the presentstudy we showed that addition of regulatory T-cells frominsulin-fed mice to diabetogenic T-cells increased IL4 expres-sion in the islets and draining lymph nodes of recipient mice,thereby suggesting a local mechanism of regulation throughthe inhibition of Thl responses. In addition, the key role of IL-4 in this regulation is strengthened by the demonstration thatanti-IL-4 antibody treatment abrogated this protection in vivo.Additional cytokines associated with the Th2 phenotype havebeen involved during oral administration of insulin in NODmice (24) and other experimental models. Weiner et al. (25)showed that milligram amounts of myelin basic protein(MBP) administered orally can protect animals from experi-mental autoimmune encephalomyelitis (EAE), and have pro-posed a mechanism involving the recruitment from the gut intothe central nervous system of Th2 regulatory cells capable ofproducing cytokines, mainly TGF-fJ, that are known to antag-onize Thl-driven cell-mediated immune responses. We werenot able to detect the presence of TGF-P or LL-10 in the isletcell infiltrates by immunocytochemistry or RT-PCR. Usingenzyme-linked immunospot (ELJSPOT) assays, Von Herrathet al. (26) recently showed that oral administration of insulinin a model of virus-induced diabetes resulted in significantchanges in the islet infiltrates, with a higher proportion oflymphocytes producing LL-4, IL-10, and TGF-(3. The apparentdiscrepancy between the spontaneous situation and theaccelerated form of diabetes after cell transfer might reflectthe ability of IL-4-secreting T-cells to reach the islets in con-trast to other Th2-secreting T-cells and/or the importance ofIL-4 during oral administration of insulin.

The increased percentage of T-cells from insulin-fed animalsin the pancreatic lymph nodes following adoptive cell trans-


Dow


C. PLOIX AND ASSOCIATES

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 C W

B

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 C W

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 C W

Insulin PBS

FIG. 5. RT-PCR analysis of IL-4 G4), IFN-7 (J3), andp2-microglobulin (C) mRNAs in the pancreaticlymph nodes of cotransferred mice, 30 days afterthe inoculation of diabetogenic T-cells and T-cellsfrom insulin-fed or PBS-fed mice. IL-4 mRNA wasdetected in 8/9 mice in the insulin group and in 1/9mice in the PBS group. The W column correspondsto DNA molecular weight markers.

fer suggests that protective mechanisms might also occurbefore islet cell invasion. Dendritic cells have been shown totake up antigens in tissues and travel via the afferent lym-phatics to the draining lymph nodes, where they stimulateantigen-specific immune responses (27). Our group previ-ously demonstrated a significant accumulation and activationof committed T-cells in the pancreatic lymph nodes 2 daysafter adoptive cell transfer (12) that could be reduced byanti-L-selectin antibodies (28). We can postulate that regula-tory CD4+ T-cells inhibit autoreactive T-cells through local IL-4 secretion after exposure to pancreatic antigens presentedby specific dendritic cells. However, although strong evi-dence exists for the key role of CD4+ T-cells in this regulation,both CD4+ and CD8+ T-cells may function as regulatory T-cells.The dosage and route of administration of the antigensappear to be critical factors. Aerosol administration ofinsulin in NOD mice produced CD8+ T-cells that suppressed

10/12

Days after co-transfer

FIG. 6. Irradiated NOD mice were reconstituted with diabetogenic T-cells and T-cells from insulin-fed (O, A) or PBS-fed (A, • ) mice.Recipients were treated the day following cell transfer with either 500ug of anti-IL-4 niAb (A, A) or sham treated with an irrelevant mAb (O,• ) . The results presented were from two independent experiments.

disease transfer, suggesting differences between the mecha-nisms that mediate mucosal tolerance in the upper airwaysand gut (29).

In conclusion, our data suggest that the protective effectsof oral administration of insulin is mediated by the presenceof Th2 cells in the target organ and the corresponding lym-phoid system. As our understanding of the autoantigensinvolved in the pathogenesis of type 1 diabetes increases, thepossibility of using antigen-specific immunotherapies forintervention becomes more feasible. Present limitations onthe amount of antigen to be fed appear to be overcome by theuse of mucosal adjuvants. We recently obtained strong pro-tective effects against autoimmune diabetes using a single oraladministration of microgram amounts of a cholera toxoidinsulin conjugate (30). This may encourage the view that itmay now be possible to prevent or attenuate human T-cell-mediated autoimmune diseases by modulating the activ-ity of autoreactive T-cell clones.

ACKNOWLEDGMENTSWe thank H. Vidal for his advice on RT-PCR and A. Durand andA. Stefanuti for their excellent technical assistance.

REFERENCES1. Castano L, Eisenbarth GS: Type-1 diabetes: a chronic autoimmune disease of

human, mouse and rat. Annu Rev Immunol 8:647-679,19902. Makino S, Kunimoto K, Mureoka Y, Mizushima Y, Katagiri X, Tochino Y: Breed-

ing of a non-obese diabetic strain of mice. Exp Anivi 29:1-13,19803. Wells H: Studies on the chemistry of anaphylaxis. III. Experiments with iso-

lated proteins, especially those of hen's egg. J Infect Dis 9:147-151,19114. Friedman A, Weiner HL: Induction of anergy or active suppression following

oral tolerance is determined by antigen dosage. Pmc Natl Acad Sci USA91:6688-6692,1994

5. Michel C, Boitard C, Bach JF: Insulin auto-antibodies in nonobese diabetic(NOD) mice. Clin Exp Immunol 75:457-460,1989

6. Wegmann DR, Norbury-Glaser M, Daniel D: Insulin-specific T-cells are a pre-dominant component of islet infiltrates in the pre-diabetic NOD mice. Eiir JImmunol 24:1853-1857,1994

7. Palmer JP, Asplin CM, demons P, Lyen K, Tatpati 0, Raghu PK, Pacquette TL:Insulin antibodies in insulin dependent diabetics before insulin treatment. Sci-ence 222:1337-1339, 1983

8. Keller RJ: Cellular immunity to human insulin in individuals at high risk forthe development of type 1 diabetes mettitus. J Autoimmun 31:321-327,1990


Dow


IL-4 AND ORAL TOLERANCE IN NOD MOUSE

9. Zhang ZJ, Davidson L, Eisenbarth G, Weiner HL: Suppression of diabetes innon-obese diabetic mice by oral administration of porcine insulin. Pmc NatlAcad Sci USA 88:10252-10256,1991

10. Bergerot I, Fabien N, Maguer V, Thivolet C: Oral administration of humaninsulin to NOD mice generates CD4+ T-cells that suppress adoptive transferof diabetes. J Autoimmun 7:655-663,1994

11. Yasunami R, Bach JF: Anti-suppressor effect of cyclophosphamide on thedevelopment of spontaneous diabetes in NOD mice. Eur J Immunol18:481-484,1988

12. Fabien N, Bergerot I, Maguer-Satta V, Orgiazzi J, Thivolet C: Pancreatic lymphnodes are early targets of T-cells during adoptive cell transfer of diabetes inNOD mice. J Autoimmun 8:323-334,1995

13. Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ: Isolation of biologicallyactive ribonucleic acid from sources enriched in ribonuclease. Biochemistry18:5294-5299, 1979

14. Mosmann TR, Cherwinski H, Bond MW, Giedlin MA, Coffamn RL: Two typesof murine helper T-cell clone. I. Definition according to profiles of lymphokineactivities and secreted proteins. J Immunol 136:2348-2357,1986

15. Rabinovitch A, Suarez-Pinzon WL, Sorensen 0, Bleackley RC, Power RF: IFNgamma gene expression in pancreatic islet infiltrating mononuclear cells cor-relates with autoimmune diabetes in nonobese diabetic mice. J Immunol154:4874-4882, 1995

16. Rabinovitch A, Suarez-Pinzon WL, El-Sheikh A, Sorensen 0, Power RF:Cytokine gene expression in pancreatic islet-infiltrating leukocytes of BBrats: expression of Thl cytokines correlates with (3-cell destructive insulitisand IDDM. Diabetes 45:749-754,1996

17. Shehadeh NN, LaRosa F, Lafferty K: Altered cytokine activity in adjuvantinhibition of autoimmune diabetes. J Autoimmun 6:291-300,1993

18. Muir A, Peck A, Clare-Salzer M, Song YH, Cornelius J, Luchetta R, Krischer J,MacLaren NK: Insulin immunization of NOD mice induced a protective insuli-tis characterized by diminished intra-islet interferon gamma transcription. JClin Invest 95:628-634,1995

19. Rapoport MJ, Jaramillo A, Zipris D, Lazarus AH, Serreze DV, Leiter EH,Cyopick P, Danska JS, Delovitch TL: Interleukin 4 reverses T-cell proliferativeunresponsiveness and prevents the onset of diabetes in nonobese diabeticmice. JExp Med 178:87-99, 1993

20. Mueller R, Krahl T, Sarvetnick N: Pancreatic expression of interleukin-4 abro-gates insulitis and autoimmune diabetes in nonobese diabetic (NOD) mice. J

Exp Med 184:1093-1099,199621. Sad S, Mosmann TR: Interleukin (IL)-4 in the absence of antigen stimulation

induces an anergy like state in differentiated CD8+ TCI cells: loss of IL2 syn-thesis and autonomous proliferation but retention of cytotoxicity and synthesisof other cytokines. JExp Med 182:1505-1515, 1995

22. Rabinovitch A, Suarez-Pinzon WL, Sorensen O, Bleackley RC, Power RF,Rajotte V: Combined therapy with interleukin-4 and interleukin-10 inhibitsautoimmune diabetes recurrence in syngeneic islet-transplanted nonobese dia-betic mice: analysis of cytokine mRNA expression in the graft. Transplanta-tion 60:368-374,1995

23. Lee MS, Mueller R, Wicker LS, Peterson LB, Sarvetnick N: IL-10 is necessaryand sufficient for autoimmune diabetes in conjunction with NOD MHChomozygosity. JExp Med 183:2663-2668,1996

24. Hancock WW, Polanski M, Zhang J, Blogg N, Weiner HL: Suppression ofinsulitis in non obese diabetic (NOD) mice by oral insulin administration isassociated with selective expression of interleukin 4 and 10, transforminggrowth factor (3 and prostaglandin E. Am J Pathol 147:1193-1199,1995

25. Weiner HL, Friedman A, Miller A, Khoury SJ, Al-Sabbagh A, Santos L, SayeghM, Nussenblatt RB, Trentham DE, Hafler DA: Oral tolerance: immunologicmechanisms and treatment of animal and human organ-specific autoimmunediseases by oral administration of autoantigens: Ann Rev Immunol12:809-837, 1994

26. Von Herrath MG, Dyrberg T, Oldstone MBA: Oral insulin suppresses vims-induced antigen-specific destruction of (3 cells and prevents autoimmune dia-betes in transgenic mice. J Clin Invest 98:1324-1331,1996

27. MacKay CR, Marston W, Dudler W: Naive and memory T-cells show distinctpathways of lymphocyte recirculation. JExp Med 171:801-817,1990

28. Fabien N, Bergerot I, Orgiazzi J, Thivolet C: Lymphocyte function associatedantigen-1, integrin a4, and L-selectin mediate T-cell homing to the pancreas inthe model of adoptive transfer of diabetes in NOD mice. Diabetes45:1181-1186,1996

29. Harrison LC, Dempsey-Collier M, Kramer DR, Takahashi K: Aerosol insulininduces regulatory X.5 T-cells that prevent murine insulin-dependent diabetes.JExp Med 184:2167-2174,1996

30. Bergerot I, Ploix C, Petersen J, Moulin V, Rask C, Fabien N, Mayer A, Czerkin-sky C, Holmgren J, Thivolet C: A cholera toxoid insulin conjugate as oral vac-cine against spontaneous autoimmune diabetes. Proc Natl Acad Sci USA94:4610-4614,1997


Dow


protection against autoimmune diabetes with oral insulin is

Documents