Neither B Lymphocytes Nor Antibodies Directed Against Self Antigens of the Islets of Langerhans Are Required for Development of Virus-Induced Autoimmune Diabetes

Generation of animals and screening

Transgenic RIP-GP and RIP-NP animals deficient of B lymphocytes were generated by breeding μMT/μMT (B cell^−/−) mice (21) with RIP-LCMV transgenic mouse lines (8, 9). Both the μMT mice (purchased from The Jackson Laboratory, Bar Harbor, ME) and the RIP-NP and RIP-GP mice were bred to C57BL/6J H-2^b mice for at least 8 generations before they were used. Offspring of RIP-LCMV × μMT mice were backcrossed for four generations to RIP-LCMV transgenic mice before being used in diabetes studies.

PCR (21, 22) or Southern blotting (9) determined the genotype of transgenic animals. Two independent standard PCRs were performed to test for the presence of the wild-type μ-chain gene and the disrupted copy of this locus (μMT) (21). The PCR primer pair used to detect the wild-type μ-chain gene was IgM-sense (5′-CTCTGTAAGGAGTCACCACC) and IgM-antisense (5′-CTTCCTCCTCAGCATTCACC). The disrupted μ-chain gene (μMT) (21) was detected by the PCR primer combination neomycin-sense (5′-CCGGCCACAGTCGATGAATCC) and IgM-antisense. The presence of the RIP-GP and RIP-NP transgenes was determined either by Southern blotting as described (9) or by standard PCR (22). For detection of the RIP-GP transgene by PCR, the primer pair GP-sense (5′-TGGACAGGCTCAGATGGCAAGA) and GP-antisense (5′-CTCAAAGCAGCCTTGTTGTAGTC) was used. To test for the presence of the RIP-NP transgene, the primers NP-sense (5′-CAGTTATAGGTGCTCTTCGGC) and NP-antisense (5′-AGATCTGGGAGCCTTGCTTTG) were used.

Induction of IDDM in transgenic animals and blood glucose measurements

IDDM was induced in age-matched (5- to 6-wk-old) transgenic RIP-LCMV × μMT mouse lines (H-2^b) by a single i.p. injection of 10⁵ PFU LCMV strain Armstrong (clone 53b) (8). Virus used for IDDM induction was plaque purified three times on Vero cells, and viral stocks were prepared by a single passage on BHK-21 cells as described (8).

After inoculation with virus, mice were checked weekly for diabetes by determining their blood glucose concentration. Animals were considered diabetic when glucose values were ≥300 mg/dl. Blood samples were obtained from the retro-orbital plexus of mice, and their glucose concentrations assessed with the ACCUCHECK III system (Boehringer Mannheim, Indianapolis, IN) (9).

Detection of autoantibodies specific to glutamate decarboxylase (GAD), ICA512/IA2, and insulin

Serum autoantibodies in RIP-LCMV mice were measured using radiobinding assays similar to those for detection of autoantibodies in human type 1 diabetes (23, 24). Mouse recombinant GAD65 (whole molecule), human recombinant GAD67 (whole molecule), human recombinant ICA512/IA2 (cytoplasmic domain), and human recombinant insulin (IAA) were used as Ags in the radiobinding assays. The human Ags used displayed a high homology to the corresponding mouse protein. The Ag insulin was purchased radiolabeled with ¹²⁵I as mono-iodoTyr-A14-insulin. GAD65, GAD67, and ICA512/IA2 were radiolabeled using ³⁵S-Met as described (25). The complex of Ag (containing 30,000 cpm per sample) and Ab (4 μl of mouse serum) was allowed to form for 24 h (GAD65, GAD67, ICA512/IA2) or 72 h (IAA). Ag-Ab complexes were precipitated with 10% TCA. Samples were run in duplicate. Mouse Ig-Ag complex was precipitated with Protein G-Sepharose (Zymed, San Francisco, CA). Pellets were washed 5 times, and the radioactivity was counted. Sera containing known amounts of autoantigens were run concurrently as positive controls as were negative controls from pooled wild-type (nontransgenic) mouse sera. All assays had signal-to-noise ratio (positive control cpm/negative control cpm) of >20. Indices for insulin autoantibodies were calculated as described (24).

Histology and immunohistochemistry

Tissue for histologic evaluation was fixed in 10% Bouin’s solution, and paraffin-embedded tissue sections were stained with hematoxylin and eosin. Immunochemical studies were conducted on 8-μm cryomicrotome sections fixed in ice-cold ethanol (95%) as described (22). Briefly, tissue sections were incubated in 2% FBS and PBS for 1 h and treated with an avidin-biotin blocking kit (Vector Laboratories, Burlingame, CA) to avoid nonspecific staining. T cells, B cells, and macrophages were detected with cell type-specific rat mAbs (anti-mouse-CD4 (L3T4), anti-mouse-CD8 (Ly-2 and Ly-3.2), anti-mouse-CD19 (ID3), and anti-mouse-B220 (RA3–6B2) (all obtained from PharMingen, La Jolla, CA), and anti-mouse-F4/80 (C1:A3-1; Serotec, Raleigh, NC). Staining for insulin was performed with a guinea pig anti-mouse insulin Ab (Dako, Carpinteria, CA). For Ab detection, sections were incubated with a biotinylated rat anti-mouse IgG1 Ab (Boehringer Mannheim). After incubation with an anti-rat or anti-guinea pig biotinylated secondary Ab (Vector Laboratories) and a streptavidin-HRP complex (Vector Laboratories), sections were stained with diaminobenzidine (DAB; Zymed). Counterstaining was performed with hematoxylin (Sigma, St. Louis, MO). Slides were embedded in AquaMount (Fisher, Pittsburgh, PA) before being photographed.

CTL assays

Primary CTL activity of splenocytic lymphocytes obtained from RIP-LCMV × μMT/μMT (H-2^b) or RIP-LCMV (H-2^b) mice was determined 7 days after infection with LCMV. CTL activity was measured using an in vitro ⁵¹Cr release assay (22). In brief, syngeneic (MC57; H-2b) and allogeneic (BALB/Cl7; H-2^d) target cells were infected either with LCMV strain Armstrong (multiplicity of infection = 1) or with a recombinant vaccinia virus expressing LCMV-NP (8). After 48 h, infected target cells were labeled with ⁵¹Cr (Amersham, Arlington Heights, IL) for 1 h, washed with medium, and transferred to 96-well dishes (10⁴ cells per well). In other assays, uninfected target cells were similarly labeled with ⁵¹Cr and coated with 10 μg/ml MHC class I D^b-restricted LCMV peptides GP aa 33–41 and NP aa 396–404. The ratios of splenic effector cells to ⁵¹Cr-labeled target cells were 50:1, 25:1, 12.5:1, and 6:1. Target cells, which were incubated with media only or with 1% Nonidet P-40, defined the spontaneous release background and total ⁵¹Cr release, respectively. After a 5 1/2 hour incubation of target with effector cells, culture supernatants cleared of cellular debris were recovered, and ⁵¹Cr counts were measured (22). Specific lysis of target cells was calculated by the following formula: (measured value − spontaneous release) × 100/(total release − spontaneous release). Each sample was run in triplicate with a variance <15%.

ELISA

Serum obtained from RIP-LCMV × μMT/μMT and RIP-LCMV mice was serially diluted and tested for the presence of mouse Ig in 96-well ELISA plates coated with an anti-mouse IgG Ab. Plates were washed three times, and a secondary anti-mouse Ig Ab conjugated with HRP was added for 30 min. After washing the plates three times, an ortho-phenylenediamine (OPD; Sigma) color reaction was performed to visualize bound mouse Ig, which was quantified using an ELISA reader at 492 nm and DeltaSoft 3 software (BioMetallics, Princeton, NJ).

FACS staining

Splenocytes were phenotyped by FACS analysis using mAbs conjugated with FITC and recognizing murine CD45R (RA3-6B2), CD4 (L3T4), and CD8 (Ly-2)(PharMingen, San Diego, CA) as described (22). Intracellular detection of IFN-γ was performed as described (22, 26). Splenocytes were obtained from mice infected with 10⁵ PFU LCMV 7 days previously. Cells were incubated for 6 h in the presence of LCMV peptides (1 ng/ml) and brefeldin A (1 μg/ml; Sigma). Peptides used were the MHC class I (D^b)-restricted GP aa 33–41 or NP aa 396–404 or the MHC class II (I-A^b)-restricted GP aa 61–80 or NP aa 309–328. After surface staining with anti-CD4 (L3T4, CyChrome-conjugated) and anti-CD8 (Ly-2, APC conjugated) Abs, intracellular IFN-γ was detected using a murine-specific Ab (XMG1.2, PE-conjugated) (22, 26).

Determination of viral titers

Spleen and kidney tissues were weighed, homogenized in medium to make a 10% solution, and cleared of tissue debris by low speed (1500 × g) centrifugation. Tissue samples and sera were obtained from four to five individual animals per group at 5, 14, and 65 days after infection with 10⁵ PFU LCMV i.p. Viral titers were determined using log dilutions of materials plaqued on Vero cells (22).

B lymphocytes and their products are present in pancreatic infiltrates of diabetic RIP-LCMV mice

We demonstrated the presence of B lymphocytes and their products in the mononuclear islet infiltrates in virally infected RIP-NP and RIP-GP mice (Fig. 1⇓). In contrast, nontransgenic littermates similarly infected with the virus had no such infiltration in the islets of Langerhans. As pictured in Fig. 1⇓, the infiltration of B lymphocytes was primarily confined to the islets of Langerhans. Furthermore, the B lymphocytes infiltrating into pancreatic islets produced Igs, evident as deposition of mouse IgG1 Ab (see Fig. 1⇓). B cell infiltration into the pancreas occurred before the complete destruction of the β cells of the islets shown by a diminished, but not complete absence of staining for insulin (Fig. 1⇓). Compared with numbers of CD8⁺ and CD4⁺ T cells (9, 22), fewer B lymphocytes infiltrated into individual pancreatic islets. Neither B cell infiltration nor IgG production was detectable in islets from RIP-NP × μMT/μMT (Fig. 1⇓) or RIP-GP × μMT/μMT mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

B lymphocytes infiltrate the islets of Langerhans. Detection of B lymphocytes (first row) and mouse IgG1 Ig (second row) in the pancreas of diabetic RIP-GP (left) and RIP-NP (right) mice. The third row shows diminished staining for insulin in the infiltrated islets. The bottom row demonstrates the absence of staining for B lymphocytes and Ig in the infiltrated pancreatic islets of RIP-NP mice deficient for B cells. The observed infiltration of B lymphocytes was restricted to the islets of Langerhans. The pancreata of RIP-GP and RIP-NP mice were analyzed 14 and 45 days after infection with LCMV. Infiltration of B cells into the pancreatic islets was absent in nontransgenic littermates infected with LCMV (data not shown), and incubation of pancreatic tissue sections with an anti-rat (secondary) Ab did not cause any staining (data not shown). B cells were detected by immunohistochemical staining with a rat Ab to B220. The presence of Igs in the islets was visualized using an Ab to mouse IgG1. All sections shown were counterstained with hematoxylin. See Materials and Methods for details.

RIP-LCMV mice generate Abs against the islet-Ags GAD and insulin

Sera of RIP-NP and RIP-GP mice were tested for the presence of Abs to GAD (23, 27, 28), ICA512/IA2 (29, 30), and insulin (12). Abs to GAD67 were found in 6 of 8 RIP-NP and 4 of 6 RIP-GP mice after challenge with virus. Furthermore, in over 50% of these mice (4 of 6 for RIP-NP and 3 of 4 for RIP-GP transgenics), detection of Abs to GAD67 preceded an elevation of blood glucose levels (>300 mg/dl) in sera of these animals (Fig. 2⇓). Abs to heterologous GAD65 were elevated in some mice but was a less reliable and sensitive indicator. Fig. 2⇓B shows the serum levels of Abs specific to GAD65 and GAD67 for two representative individual RIP-NP and RIP-GP mice over the course of LCMV infection until development of IDDM.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Islet-specific Abs to GAD and to insulin are detected in RIP-LCMV transgenic mice before onset of IDDM. A, Detection and incidence of Abs to GAD67, GAD65, insulin, and ICA512/IA-2 in the sera of RIP-NP and RIP-GP mice is displayed. Mice were either uninfected or inoculated with LCMV. Abs specific for insulin, GAD67, and GAD65 were detected. Abs against ICA512/IA-2 were not found. In the absence of LCMV infection, none of the RIP-LCMV mice spontaneously generated Abs to GAD or to insulin. B, Detection of Abs specific for GAD65 (○) and GAD67 (▪) was measured in two representative RIP-NP mice (top) and two representative RIP-GP mice (bottom) over a time course (0–24 wk, RIP-NP; 0–3 wk, RIP-GP). The blood glucose levels measured during the timed detection of each autoantibody are given in the bottom row. Values (cpm) represent the amounts of radiolabeled Ag after immunoprecipitation with the tested sera. The dotted lines show the cut-off value (mean + 2 SD of negative control sera) for specific Ab detection. C, The insulin Ab index (24 ) and the blood glucose levels (bottom) appear for two representative RIP-NP mice over time (0–9 wk after infection with LCMV). Presence of Abs to GAD and to insulin that preceded elevated blood glucose levels is clearly seen.

For analysis of Abs to insulin and ICA512/IA2, nineteen RIP-NP and four RIP-GP mice were tested before and after infection with LCMV. None of the tested mouse sera were positive for Abs specific to ICA512/IA2. However, 9 of 19 RIP-NP mice (47%) and 1 of 4 RIP-GP animals (25%) had Abs to insulin. In 7 of the 9 RIP-NP mice with anti-insulin Abs, the Abs preceded a rise in blood glucose. Fig. 2⇑C displays the profiles of insulin Ab and blood glucose levels for two representative RIP-NP mice after viral infection, one showing a correlation with autoantibodies and onset of IDDM and the other no correlation (Fig. 2⇑, B and C).

IDDM occurs in RIP-LCMV mice in the absence of B lymphocytes and their products

The presence of B lymphocytes and their products (e.g., Abs) in the pancreatic islets of RIP-LCMV mice after infection with LCMV (Fig. 1⇑) and the generation of Abs to islets Ags (Fig. 2⇑) raised the possibility that B cells, in concert with T cells, might participate in the pathogenesis of type 1 diabetes in RIP-LCMV mice. To directly test this possibility, we bred B cell-deficient (μMT/μMT) mice (21) with RIP-GP and RIP-NP transgenic mice. B cell-deficient mice are homozygous for the μMT mutation that results from the targeted disruption of a membrane domain exon of the Ig μ-chain gene (21). Because these homozygous μMT/μMT (B cell^−/−) animals lack mature B lymphocytes, they cannot produce Ig. Double transgenic μMT/μMT mice expressing the LCMV transgenes had a disruption of the membrane exon of the μ-chain gene and no B lymphocytes but normal distributions of T cell subsets according to flow cytometric analysis. No Ig was present in sera of RIP-LCMV × B cell^−/− mice tested by ELISA. However, RIP-LCMV mice heterozygous for the μMT mutation (B cell^+/−) were indistinguishable from B cell^+/+ and wild-type mice in their relative and absolute quantities of splenic B and T lymphocytes and displayed normal and equivalent levels of serum Ig.

Progeny mice obtained by backcrossing RIP-LCMV × μMT mice to RIP-LCMV lines for four generations were used for diabetes studies. Fig. 3⇓A demonstrates that all experimental groups of RIP-GP × μMT mice, including those with no B cells (RIP-GP × B cell^+/+, ^+/−, ^−/−), developed diabetes after infection with LCMV; the incidence and kinetics of disease were similar. Thus, absence of B lymphocytes had no significant effect on the incidence of CD4-independent IDDM. We then analyzed the effect of an absence of B lymphocytes in the RIP-NP transgenic model. In contrast to RIP-GP mice, RIP-NP mice require the participation of both CD8⁺ and CD4⁺ T lymphocytes to develop autoimmune diabetes. In addition, RIP-NP mice develop disease more slowly because the majority of their anti-viral (self) high-affinity CTLs are deleted in the thymus due to thymic expression of the viral (NP) self-Ag (9). As in RIP-GP × μMT mice, all experimental groups of RIP-NP × μMT mice (RIP-NP × B cell^+/+, ^+/−, and ^−/−) developed diabetes (Fig. 3⇓B). No significant differences in disease incidence or kinetics appeared during comparison of RIP-NP/B cell-competent groups with RIP-NP/B cell-deficient mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Incidence of IDDM in RIP-GP and RIP-NP × μMT/μMT mice after infection with LCMV. Incidence (%) of autoimmune diabetes in RIP-GP × μMT/μMT (A) and RIP-NP × μMT/μMT (B) mice. Animals of all experimental groups (B cell^+/+, B cell^+/−, and B cell^−/−) were age matched when infected with 10⁵ PFU LCMV i.p. Blood glucose levels were measured weekly. Animals termed diabetic had blood glucose levels ≥300 mg/dl, and the numbers (n) of mice per experimental group is given. Statistical analysis revealed no significant differences between B lymphocyte-competent and B lymphocyte-deficient RIP-GP/RIP-NP mice at any time point.

Anti-LCMV CTL responses that terminate acute LCMV infection and cause IDDM are generated in RIP-LCMV × B cell^−/− mice

In the virus-induced RIP-LCMV model of diabetes, overt disease is dependent on the generation of anti-viral (self) CTL (8, 9, 20). Therefore, we determined the generation and effectiveness of the primary anti-LCMV CTL responses generated in RIP-LCMV × μMT mice 7 days after a primary LCMV infection. RIP-GP × B cell^−/− as well as RIP-NP × B cell^−/− mice produced vigorous anti-LCMV CTL responses of a quantity and quality that were indistinguishable from those of their B cell-competent littermates as judged by in vitro ⁵¹Cr lysis assay (Table I⇓). All groups also eliminated infectious virus from serum, spleen, and kidney by 14 days after LCMV challenge (data not shown). As expected, due to the expression of the NP transgene in the thymus of RIP-NP transgenic mice, CTL responses in RIP-NP × B cell mice were weaker than those found in their nontransgenic littermates who, in contrast, had high affinity LCMV CTL (Table II⇓). These observations were independent of the presence or absence of B lymphocytes.

We assessed the ability of CD8⁺ T lymphocytes recovered from the spleens of RIP-NP × B cell^−/− mice to secrete IFN-γ (Fig. 4⇓). Similar to our CTL results, the numbers of IFN-γ⁺/CD8⁺ T cells responding to the MHC class I D^b-restricted LCMV epitope NP aa 396–404 were significantly reduced in RIP-NP transgenic mice compared with nontransgenic littermates (2.8 × 10⁶ IFN-γ⁺/CD8⁺ T cells in RIP-NP⁺ × B cell-competent mice vs 5.6 × 10⁶ IFN-γ⁺/CD8⁺ T cells in nontransgenic B cell-competent mice). In contrast, the numbers of CD8⁺ T lymphocytes responding to LCMV peptide GP aa 33–41 by production of IFN-γ were indistinguishable in RIP-NP and nontransgenic mice competent in B cells. Compared with B cell-competent wild-type mice, B cell-deficient μMT/μMT mice have reduced numbers of splenocytes (26, 31). Accordingly, we detected a reduction in the total numbers of splenic CD4⁺ and CD8⁺ T lymphocytes producing IFN-γ in response to each of the tested LCMV peptide epitopes in B cell-deficient RIP-NP mice (Fig. 4⇓). Despite the differences in absolute T lymphocyte numbers, the relative proportions of Ag-specific T cells were similar in RIP-NP mice competent or deficient in B cells (Fig. 4⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Numbers of Ag-specific CD8⁺ and CD4⁺ T lymphocytes producing IFN-γ in RIP-NP × μMT compared with nontransgenic mice. The relative proportions (graph) and the total numbers (table) of CD8⁺ and CD4⁺ T cells specifically responding to MHC class I (GP aa 33–41 and NP aa 396–404) and MHC class II (GP aa 61–80 and NP aa 309–328) D^b-restricted LCMV peptides were determined by intracellular FACS staining for IFN-γ. Cells were recovered from the spleens of RIP-NP⁺ × B cell^−/− (open columns), RIP-NP⁺ × B cell^+/+ (gray columns), and nontransgenic, B cell-competent (filled columns) mice 7 days after infections with 10⁵ PFU LCMV. Three to four mice were used for each experimental group. The error bars indicate 1 SEM.

In summary, in the absence of B lymphocytes, RIP-NP transgenic mice generated anti-LCMV-NP CTL that were efficient and effective in initiating β cell destruction in the islets of Langerhans, leading to diabetes. Because we observed reduced CTL responses in RIP-NP mice compared with nontransgenic littermates, it is likely that these CTL were of lower affinity as reported elsewhere (9).

CD8⁺ and CD4⁺ T cell and macrophage infiltration in the islets of Langerhans is equivalent in RIP-GP and RIP-NP mice whether they are B cell competent or deficient

Finally, we used immunohistochemistry to analyze the profile of mononuclear cells entering the pancreata of RIP-LCMV × B cell^+/+ and ^−/− mice (Figs. 5⇓ and 6⇓). After infection with LCMV, CD8⁺ as well as CD4⁺ T lymphocytes infiltrated the pancreas and specifically homed to the islets of RIP-GP × B cell^+/+ and ^−/− mice (Fig. 5⇓). The amounts of infiltrating T cells in the islets of RIP-GP animals either deficient or competent in B lymphocytes were equal and indistinguishable from one other. Corresponding findings were obtained when macrophages infiltrating the pancreas were identified by using F4/80 Ab (Fig. 5⇓). Thus, the proportions of individual mononuclear cell subsets (B lymphocytes excluded) remained similar independently of the presence or absence of B lymphocytes (see Fig. 5⇓). Results were comparable for CD8⁺ and CD4⁺ T lymphocytes and macrophages in the pancreatic islets of RIP-NP × B cell^−/− mice vs their B cell^+/+ littermates (Fig. 6⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Infiltration of mononuclear cells into the pancreatic islets of RIP-GP mice deficient or competent in B lymphocytes. Immunohistochemical localization of CD8⁺ T lymphocytes (left), CD4⁺ T lymphocytes (middle), and F4/80⁺ macrophages (Mφ; top right) in the pancreas of a RIP-GP × B cell^+/+ (top row) and a RIP-GP × B cell^−/− (bottom row) mouse 12 days after infection with 10⁵ PFU LCMV. A stain for B lymphocytes in the spleen (bottom right) and pancreas (data not shown) confirmed the predicted B lymphocyte phenotype of these animals. The amounts of CD4⁺ and CD8⁺ T lymphocytes and macrophages infiltrating the pancreatic islets of Langerhans were indistinguishable in these two experimental groups. Mononuclear cells were detected by cell type-specific anti-CD4, anti-CD8, F4/80, and B220 Abs and subsequent DAB color reaction. Sections were counterstained with hematoxylin. Similar results were obtained in the other 5–6 mice analyzed for each experimental group.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Localization of mononuclear cell types in RIP-NP mice deficient or competent in B lymphocytes. Infiltration of CD8⁺ T lymphocytes (top), CD4⁺ T lymphocytes (middle), and macrophages (Mφ; bottom) into the pancreas of a RIP-NP × B cell^+/+ mouse (left) and a RIP-NP × B cell^−/− mouse (right). Sections shown were obtained from diabetic animals 45–50 days after infection with 10⁵ PFU LCMV i.p. Neither differences in the distribution and amount of infiltrating mononuclear cells nor those in the CD8/CD4 T cell ratios were observed in B cell-competent vs B cell-deficient RIP-NP mice. This was also the case in the 4–5 additional mice analyzed for each experimental group. After DAB staining visualized CD4⁺ and CD8⁺ T cells and macrophages, tissue sections were counterstained with hematoxylin. A corresponding B lymphocyte stain using the B220 Ab confirmed the absence of B cells in the spleen and pancreas of RIP-NP × B cell^−/− mice (data not shown).