C57BL/6 male specific pathogen-free mice, 6-8 weeks old, were bred in the Animal Facility of the University of São Paulo at Ribeirão Preto Medical School. The animals were maintained on a 12-hour light/dark cycle and were given free access to food and autoclaved water. Each experiment was performed in triplicate. The following six groups of mice were studied: control-injected with phosphate-buffered saline (PBS); DNA-HSP65 - injected with the DNA-HSP65 vaccine; DNAv - injected with the empty vector; STZ - injected with STZ on 5 consecutive days; DNA-HSP65+STZ - injected with DNA-HSP65 prior to STZ; DNAv+STZ - injected with DNAv prior to STZ.

The DNA-HSP65 vaccine was derived from the pcDNA3 plasmid (Invitrogen, Carlsbad, CA), which had been previously digested with BamHI and NotI (Gibco BRL, Gaithersburg, MD) by inserting a 3.3-kb fragment corresponding to the M. leprae Hsp65 gene and the cytomegalovirus intron A 31. The empty pcDNA3 plasmid was used as a control. Luria-Bertani liquid medium (Gibco BRL) containing ampicillin (100 μg/ml) was used to culture DH5α Escherichia coli, transformed either with pcDNA3 plasmid or with the plasmid containing the Hsp65 gene (DNA-HSP65). Plasmids were purified using the Concert High Purity Maxiprep System (Gibco BRL). Plasmid concentration was determined by spectrophotometry at λ = 260 and at λ = 280 nm using the Gene Quant II apparatus (Pharmacia Biotech, Cambridge, UK). Mice were intramuscularly injected with three doses (100 μg each) of DNA-HSP65 or DNAv given at 2-week intervals. Control animals were injected with PBS.

In order to induce diabetes, male C57BL/6 mice were given daily intra-peritoneal injections of STZ diluted in citrate buffer, (40 mg/kg, Sigma-Aldrich, St. Louis, MO) for five consecutive days as described before 32. Serum glucose concentration in blood obtained from a tail vein, was measured using Prestige LX Smart System Test-strips (Home Diagnostic, Inc., Fort Lauderdale, FL) once a week, beginning 7 days after the last STZ dose. Blood glucose levels of 200 mg/dl (12 mmol/l) or above on two consecutive weeks was considered confirmation of diabetes development. The incidence of diabetes was expressed as the percentage of animals that presented diabetes.

Splenocytes were harvested, cultured in RPMI 1640 medium (5 × 10⁶/ml; Life Technologies, Grand Island, NY) and stimulated in vitro with 20 μg/ml of recombinant Hsp65 protein or concanavalin A. Cytokine levels in culture supernatants were evaluated 48 hours later by ELISA according to manufacturer instructions. Monoclonal antibodies derived from the clones G281-2626 and MP6-XT3 were used as capture and biotinylated antibodies, respectively, for TNF-α quantification, whereas mAb clones JES5-2A5 and SXC-1 were used to detect IL-10. The detection limit was 15 pg/ml. Antibodies and recombinant cytokines were purchased from Pharmingen (San Diego, CA).

T-cell subsets were determined with mAbs specific for cell surface markers. Phycoerythrin (PE)- labeled mAbs included anti-CD8 mAbs (clone 53-6.7, rat IgG2a) and anti-CD25 mAbs (PC61, rat IgG1). Fluorescein isothiocyanate (FITC)-labeled mAbs were also used: anti-CD4 (clone H129.19); anti-CD103 (M290, rat IgG2a); and anti-CTLA-4 (clone 1B8, hamster IgG). In addition, a Percp-labeled, anti-CD4 mAb (clone RM4-5) was used. Hamster IgG and rat IgG2a, together with rat IgG2b labeled with PE, FITC or Percp, were used as isotype controls. All mAbs were purchased from PharMingen and were used according to manufacturer instructions.

Spleen cells (1 × 10⁷/ml) were initially incubated for 30 min at 4°C with Fc Block (1 μg/10⁶ cells; PharMingen). The cells were then incubated with the proper mAb (0.75 μg/10⁶ cells) for 30 minutes at 4°C in total darkness. Lymphocytes were analyzed through flow cytometry using the CellQuest computer program (FACSort; Becton Dickinson, San Jose, CA). Ten thousand events per sample were collected, and double-color fluorescence-activated cell sorter analysis was performed as described elsewhere 33.

Five weeks after the final STZ injection, the pancreas were removed and sections (4-5 μm) were stained with hematoxylin-eosin (Merck, Whitehouse Station, NJ). The degree of insulitis was scored using the following scale: 0 = intact islet; 1 = peri-insulitis; 2 = moderate insulitis (< 50% of the islet infiltrated); 3 = severe insulitis (≥ 50% of the islet infiltrated); and 4 = destructive insulitis. At least 20 islets per pancreas were analyzed by two independent examiners. Staining for CD4, CD8, IL-10 and TNF-α was performed on cryostat acetone-fixed pancreatic sections by incubation for 1 hour with either rat anti-mouse primary mAb (Pharmingen) diluted 1:200 in 3% bovine serum albumin (BSA)-PBS, or with rat anti-mouse CD25 biotinylated mAb (Pharmingen) diluted 1:200 in 3% BSA-PBS. This was followed by 1 hour of incubation with biotin-conjugated rabbit anti-rat antibody (Pharmingen). The color was revealed using the 3,3'-diaminobenzidine revelation system (Vector kit; Vector Laboratory, Burlingame, CA). For the control of nonspecific staining, pancreas sections were treated without anti-mouse primary mAb. In addition, the pancreas immunohistochemistry was further quantified by using the image processing program (ImageJ, U.S. National Institutes of Health, Bethesda, Maryland, USA) 34.

The results are presented as mean ± SD. Statistical significance of diabetes incidence and degree of insulitis was determined by using Fisher's exact test. For ELISA and fluorescence-activated cell sorter results, the statistical significance was determined by One-Way analysis of variance followed by Tukey's test. Values of P < 0.05 were considered significant.

Glucose concentrations were analyzed weekly after the final STZ injection. As shown in Fig. 1A-C, administration of DNA-HSP65 or DNAv did not induce diabetes or significantly altered diabetes incidence induced by STZ in mice. Notably, administration of DNA-HSP65 clearly delayed diabetes development (DNA-HSP65+STZ group) during the first week (Fig. 1A and 1B). Diabetes incidence at this period was 25% in DNA-HSP65+STZ group, compared with 75% in the STZ group (Fig. 1A and 1B). However, four weeks later, diabetes incidence was similar in the three groups that received STZ (Fig. 1A and 1C). The histological pancreas analysis documented in Fig. 1D, clearly demonstrates that previous DNA-HSP65 vaccination determined a protective effect. As expected, mice from the STZ group presented a clear pattern of insulitis, characterized by peri-insulitis (in 25% of the islets), moderate insulitis (in 19%), invasive insulitis (in 3.3%) and destructive insulitis (in 2.7%). A very similar islet damage profile was observed in the DNAv+STZ group. However, DNA-HSP65+STZ group presented considerably less islet injury, characterized by a significantly higher percentage of intact islets, i.e. 81% in comparison to 50% in the STZ group and 56% in DNAv+STZ group. Also, no destructive insulitis was observed in this previously vaccinated group.

To understand the prophylactic effect of DNA-HSP65 in the STZ model, we initially analyzed the production of TNF-α and IL-10 in spleen cell cultures. TNF-α levels were higher in STZ, DNAv+STZ and DNA-HSP65+STZ derived splenic cultures stimulated with recombinant Hsp65 or concanavalin A than in non-stimulated cultures, however, there was no significant difference among these three groups (Fig. 2A). Nevertheless, in the cultures from mice injected with DNA (vector and vaccine) before STZ diabetes triggering, the stimulation with rhsp65 determined a very significant increase in IL-10 production. Interestingly, production of IL-10 was significantly higher in vaccinated mice in comparison to the control vector injected group (Fig. 2B).

The higher IL-10 production in mice that received DNA before STZ suggested the possible contribution of regulatory T cells in this protection against insulitis. In this context, we initially characterized the populations and phenotypes of T cells in the spleen. Clear differences were observed in the amounts of CD4⁺ and CD8⁺ T cells (Fig. 3). Animals that were immunized with DNA-HSP65 presented higher numbers of CD4⁺ cells but not of CD8⁺ cells. This alteration was not seen in animals injected with the empty vector. Injection of STZ alone clearly increased the number of both cell types and interestingly prior administration of DNA-HSP65 or DNAv was associated with significantly lower numbers of these cells (Figs. 3A and 3B). The analysis of phenotypic markers indicative of regulatory ability, including CD4⁺CD25^highCD103⁺ and CD4⁺CD25^highCTLA-4⁺, clearly showed that these cells were induced or activated in mice injected with the DNA-HSP65 but not with the empty vector (Figs. 3C and 3D). Moreover, DNA-HSP65 injection followed by STZ clearly increased the number of cells showing the regulatory phenotypic markers in comparison to STZ group. No such up-regulation was observed in the group of mice that received the DNAv before induction of diabetes with STZ. Interestingly, in the spleens of STZ-injected mice, high numbers of CD4⁺CD25^high cells were observed (data not shown), although the numbers of CD4⁺CD25^highCD103⁺ and CD4⁺CD25^highCTLA-4⁺ cells remained at similar levels as the control group (Figs. 3C and 3D).

We next considered the possibility that regulatory cells were migrating to the pancreas to block insulitis. In the STZ group there was pronounced pancreatic infiltration by CD8⁺ cells (Fig. 4D) but not by CD4⁺ cells (Fig. 4C). By evaluating the stained area for CD8⁺ and CD4⁺ cells we observed that CD8⁺ cell staining was 2 fold higher than CD4⁺ cell in STZ group (2,45% and 1,20% of stained area for CD8 and CD4, respectively). In DNA-HSP65+STZ and DNAv+STZ groups of mice, this profile was dramatically changed showing a marked infiltration by CD4⁺ cells (Figs. 4E and 4G, respectively) associated with a quite discrete infiltration by CD8⁺ cells (Figs. 4F and 4H, respectively). These groups presented CD4⁺ cell staining three fold higher than CD8⁺ cell staining (2,40% and 0,79% for CD4 and CD8, respectively, in DNA-HSP65+STZ group and, 2,21% and 0,71% for CD4 and CD8, respectively, in DNAv+STZ group).

Since a higher number of CD4⁺ cells and a lower number of CD8⁺ cells were present in the islets of DNA-HSP65-injected and DNAv-injected mice in relation to STZ group, we supposed that CD4⁺ cells with a regulatory phenotype could be involved in down modulation of diabetes development. To confirm the hypothesis that an anti-inflammatory and regulatory cytokine was involved in protection against insulitis in this model, we investigated the local production of IL-10 and TNF-α. In the pancreatic islets of STZ-injected mice there was abundant local TNF-α production and no IL-10 production (Figs. 5C and 5D, respectively). Analysis of stained area showed that 1,25% of total area was staining for TNF-α. The DNA-HSP65+STZ and DNAv+STZ groups also showed local TNF-α production. Labeled areas were 0,31% and 0,56% for vaccinated and vector injected groups, respectively (figs. 5E and 5G). However, DNA-HSP65+STZ and DNAv+STZ groups of mice presented a marked IL-10 production in the islets. The IL-10 stained area in DNA-HSP65+STZ and DNAv+STZ was 1,52% and 1,36%, respectively. This represented an increase of 5 and 2,5-folds, respectively, when compared with TNF-α labeled area (Figs. 5F and 5H).