Toll-Like Receptor 3 Is Critical to the Pancreatic Islet Milieu That Is Required for Coxsackievirus B4-Induced Type 1 Diabetes in Female Nonobese Diabetic Mice.

OBJECTIVE: Genetic and environmental influences play a role as triggers of type 1 diabetes mellitus (T1DM). Female nonobese diabetic (NOD) mice are useful for studying T1DM as they spontaneously develop T1DM, which can be accelerated by some viruses. Toll-like receptor 3 (TLR3) is believed to play a critical role in viral-induced T1DM and β-cell destruction, because female Tlr3 knockout (Tlr3-/-) NOD mice are protected from Coxsackievirus B4 (CVB4)-induced acceleration of T1DM. However, the exact role(s) TLR3 plays in the pathogenesis of CVB4-induced T1DM remain unknown.
METHODS: This longitudinal study used immunostaining, laser capture microdissection, and reverse transcription real-time polymerase chain reaction of islets from female uninfected and CVB4-infected Tlr3+/+ and Tlr3-/- NOD mice.
RESULTS: Islets isolated from female Tlr3+/+ NOD mice 4 to 8 weeks of age had higher amounts of insulitis, Cxcl10, Il1b, Tnfa, and Tgfb1 expression compared with Tlr3-/- NOD mice. After CVB4 infection, Tlr3+/+ NOD mice had higher amounts of insulitis and T-cell infiltration at 3 days after infection compared with Tlr3-/- CVB4-infected NOD mice.
CONCLUSIONS: Toll-like receptor 3 is necessary for establishment of a pancreatic islet inflammatory microenvironment by increasing insulitis and cytokine expression that facilitates CVB4-induced T1DM in female NOD mice.

Type 1 diabetes mellitus (T1DM) is an autoimmune disorder marked by destruction of insulin-producing β cells by one's own immune system and/or by an environmental insult. Both genetic and environmental factors play a role in the onset of T1DM, with considerable evidence that suggests viruses are one of the key players.[1-22] Specifically, coxsackie B viruses (CVBs), which are positive single-stranded RNA viruses, have been shown to trigger T1DM in susceptible individuals and in mouse models[4,9,17,19,23-26]; however, the exact mechanisms by which CVBs trigger T1DM onset have yet to be elucidated.

In mammals, one of the mechanisms through which viruses are recognized is via toll-like receptors (TLRs), a specific class of innate immune receptors responsible for recognizing particular pathogen-associated and/or damage-associated molecular patterns (ie, PAMPs and DAMPs, respectively), such as viral double-stranded RNA (dsRNA), and eliciting an innate immune response.[27] In particular, toll-like receptor 3 (TLR3), which recognizes dsRNA, is known to be present in both immune and nonimmune cells, particularly in β cells of both the human and murine pancreas.[27,28] Toll-like receptor 3 is responsible for recognizing both viral dsRNA and endogenous dsRNA released from damaged or dying host cells[29-31] and signaling activation results in the production of type 1 interferons, inflammatory cytokines, and chemokines, and can lead to β-cell apoptosis.[29,31-34] Previous reports show that when left in a sterile environment, female Tlr3−/− nonobese diabetic (NOD) mice will develop T1DM at a similar rate as female Tlr3+/+ NOD mice.[23,32] However, when a preexisting mass of autoreactive T cells is present within pancreatic islets, that is, after a “critical threshold” of insulitis has been reached (ie, usually by 8 weeks of age),[4] and then female Tlr3+/+ NOD mice are infected with CVB4, they rapidly develop T1DM.[4,17,23,26] In contrast, female Tlr3−/− NOD mice are protected from this viral acceleration of T1DM,[23] indicating that TLR3 is key for viral acceleration of T1DM in female NOD mice. Thus, although TLR3 is key to CVB4 acceleration of T1DM in female NOD mice, the mechanism(s) by which TLR3 mediates viral acceleration of T1DM in female NOD mice remains elusive. Herein, we used female Tlr3+/+ and Tlr3−/− NOD mice to offer the first direct line of evidence to explain how TLR3 is involved in the establishment of the “critical threshold” of insulitis that is permissive for CVB4-induced acceleration of T1DM in female NOD mice and how TLR3 affects a specific inflammatory islet environment 3 days after CVB4 infection.

MATERIALS AND METHODS

Animal Husbandry/Glucose Measurements/CVB4 Infection

This work was conducted with approval from the Ohio University Institutional Animal Care and Use Committee in accordance with accepted standards of humane animal care. Three-week-old female Tlr3+/+ NOD mice were obtained from the Jackson Laboratory (Bar Harbor, Maine) and housed in a sterile/germ-free facility. Breeding pairs of Tlr3−/− NOD mice were kindly provided by Dr Li Wen (Yale University), and animals were generated as previously described.[32] Tlr3−/− NOD mice used in this study were from our breeding colony of these mice at Ohio University. Glucose measurements were conducted weekly using the FreeStyle Lite glucose monitoring system (Abbott Diabetes Care, Inc, Alameda, Calif). Blood glucose measurements were nonfasting and were consistently measured at the same time of day (ie, early morning approximately 8:00 a.m.). Diabetes was defined as blood glucose values greater than 240 mg/dL on consecutive days; however, no mice in this study were diabetic by the end point (ie, 8 weeks +3 days of age or 3 days after CVB4 infection). Eight-week-old female Tlr3+/+ and Tlr3−/− NOD mice received an intraperitoneal injection of 5 × 105 plaque forming units of CVB4 (kindly provided by Dr Roger Loria, Virginia Commonwealth University).[35]

Insulin Staining of Pancreatic Tissue

Harvested pancreata from 4-, 6-, and 8-week-old uninfected Tlr3+/+ and Tlr3−/− NOD and uninfected and CVB4-infected Tlr3+/+ and Tlr3−/− 8 weeks +3 days of age NOD mice underwent formalin-fixed paraffin embedded protocols and slides containing tissue sections were deparaffinized, rehydrated, and antigen retrieval using sodium citrate (10 mM, pH 6) was completed. Following the protocol provided from Abcam EXPOSE rabbit-specific horseradish peroxidase (HRP)/3,3′-diaminobenzidine detection kit (Abcam, Cambridge, Mass), tissues were incubated in a 1:1000 dilution of rabbit antimouse insulin antibody (Abcam). Goat antirabbit HRP-conjugated secondary antibody (from kit) was added directly onto the tissues followed by a 1:50 dilution of DAB plus chromogen (from kit) and a counterstain with hematoxylin. Slides were dehydrated after opposite order of the rehydration steps and mounted with permount and a coverslip (ThermoFisher, Carlsbad, Calif).

CD3 Immunostaining of Pancreatic Tissue

Slides of tissues were rehydrated following the same steps as insulin staining. Antigen retrieval using Proteinase K (Abcam) was followed by a 3% hydrogen peroxide block and protein block (either a 5% goat serum or 10% donkey serum, 1% BSA in PBS). Rabbit antimouse CD3 (1:200 dilution, Abcam) was added and incubated overnight at 4°C. Tissues were then incubated in goat antirabbit-specific HRP-conjugate (from EXPOSE kit) secondary antibody at a dilution of 1:500. For CD3 staining, a 1:50 dilution of DAB plus chromogen (from EXPOSE kit) was added followed by counterstain with hematoxylin, and slides were mounted as described above.

Insulitis Scoring and Percent CD3+ T-Cell Infiltration of Islets

Sections of insulin-immunostained pancreas were assessed for insulitis on Nikon Eclipse 80i microscope (Melville, NY) with Evolution MP color camera (Media Cybernetics, Rockville, Md) at 400× magnification and ranking of insulitis for each islet was based on a previously described scale.[23] Sections of insulin-immunostained pancreas were assessed for percentage of immune cell infiltration using ImageJ Software Version 1.53e (National Institutes of Health, Rockville, Md).[36] Percent of immune cell infiltration was calculated by subtracting the area of insulin staining in the islet section from the total area of the islet section and then dividing by the total area of the islet section and multiplying by 100. An average of 28 islets per mouse were analyzed. CD3 immunohistochemical staining was done to confirm that the immune cell infiltration areas of the islets that were measured were CD3+ T cells.

Tissue Preparation for LCM of Islets

All equipment, including microtome, blades, water bath, brushes, and polyethylene naphthalate membrane (PEN-membrane) slides (Applied Biosystems, Foster City, Calif), was sprayed with RNase Away (ThermoFisher) before tissue sectioning. Tissue sections were sliced at a 12-μm thickness (separated by 36 μm), deparaffinized, rehydrated, and stained with hematoxylin and eosin. Slides were air dried, followed by brief incubation on a hot plate and immediately taken to the Leica LMD6000 microscope (Buffalo Grove, Ill) for islet isolation as previously described.[37]

RNA Extraction From Islets Isolated by LCM

To obtain sufficient RNA to run reverse transcription real-time polymerase chain reaction (RT-qPCR), islets were pooled from 3 to 4 different mice per group; thus, no statistical analysis can be performed. Because of islet pooling, the following modifications to the traditional delta-delta Ct analysis were made. The Livak and Schmittgen's delta-delta Ct method for one housekeeping gene (Rn18s) was used.[38] Delta Ct (∆Ct) was calculated by subtracting the Ct of the housekeeping gene from the Ct of the gene of interest. Because the samples were pooled, no calibrator/reference sample was available to further calculate delta-delta Ct. 1/∆Ct provided a way to compare relative gene expression between groups. Thus, RT-qPCR results are represented as 1/∆Ct to compare relative gene expression between groups. Islet numbers ranged from 25 to 65/mouse. RNA extraction was performed using the RNease FFPE Kit (Qiagen, Hilden, Germany). RNA integrity and quantity were assessed using the Agilent 2100 Bioanalyzer (Santa Clara, Calif) at the Ohio University Genomics Facility.

cDNA Synthesis

cDNA was synthesized using Applied Biosystems High Capacity cDNA Reverse Transcription kit with RNAse inhibitor (ThermoFisher) according to the manufacturer's instructions.

qPCR of Islet RNA

All cDNA was preamplified for target genes using TaqMan PreAmp Master Mix (ThermoFisher) with the following gene expression assays: Cxcl10, Mm00445235_m1; Il1b, Mm00434228_m1; Ifnb1, Mm00439552_s1; Tnfa, Mm00443258_m1; Tgfb1, Mm01178820_m1; and all FAM labeled. cDNA preamplification consisted of 14 cycles of 95°C for 15 seconds, followed by 60°C for 4 minutes. Expression of the abovementioned genes was detected using TaqMan Gene Expression Master Mix (ThermoFisher) and gene expression assays previously mentioned. In addition, 18S ribosomal RNA (Rn18s, Mm03928990_g1; VIC labeled) was used as the housekeeping gene in duplex with each gene of interest. Each sample was run in duplicate on BioRad CFX384 Touch Real-Time PCR Detection System (Hercules, Calif). The PCR cycles were 40 cycles of 95°C for 15 seconds, followed by 60°C for 1 minute.

Statistical Analyses

Statistical analyses were performed using Statistica V13.3 software (Tibco, Palo Alto, Calif). Independent Student t tests or 1-way analyses of variance were used for comparisons. P ≤ 0.05 was considered significant or as indicated.

RESULTS

Quantification of Insulitis and CD3+ T-Cell Infiltration in Pancreatic Islets of Young Female Tlr3+/+ and Tlr3−/− NOD Mice With Impact of CVB4 Infection in 8-Week-Old Mice 3 Days After Infection

Multiple previous studies have shown that 8 weeks is the age when Tlr3+/+ NOD mice have achieved a “critical threshold of insulitis” (in approximately 30%–50% of mice) needed for the CVB4 virus to trigger an acute acceleration of T1DM. We observed that Tlr3+/+ NOD mice exhibited significantly higher insulitis scores at 8 weeks and significantly greater CD3+ T-cell infiltration of islets at ages 4 weeks and 8 weeks, compared with Tlr3−/− NOD mice (Figs. 1A, B). At the critical age of 8 weeks, both Tlr3+/+ and Tlr3−/− female NOD mice were infected with CVB4 and insulitis scores and CD3+ T-cell infiltration of islets were assessed at 3 days after infection. Insulitis scores and CD3+ T-cell infiltration in CVB4-infected Tlr3+/+ (Figs. 1A, B, red bars, P ≤ 0.001) and Tlr3−/− (Figs. 1A, B, dashed black bar, P ≤ 0.001) female NOD mice roughly doubled in both groups at 3 days after CVB4 infection compared with the 8-week-old mice. Insulitis scores and CD3+ T-cell infiltration of islets were again significantly (Figs. 1A, B, solid black bars, P ≤ 0.01 and P ≤ 0.001, respectively) higher in the Tlr3+/+ NOD mice compared with Tlr3−/− NOD mice at 3 days after CVB4 infection.

FIGURE 1

Uninfected and CVB4-infected Tlr3+/+ NOD mice have more severe insulitis as well as elevated T-cell infiltration of pancreatic islets compared with Tlr3−/− NOD mice. Pancreata from uninfected and CVB4-infected female Tlr3+/+ and Tlr3−/− NOD mice were isolated and evaluated by immunohistochemistry. Insulitis score and amount of immune cell infiltration of islets were assessed for individual islets from each mouse (an average of 28 islets/mouse were analyzed). A, Average insulitis scores and (B) percent T-cell infiltration of islets for all mice within each group were calculated and represented. C and D, Examples of images of pancreas/islets stained for insulin (C) and CD3 (D) by immunohistochemistry and counterstained with hematoxylin and eosin that were used for assessment of insulitis and T-cell infiltration of islets. Arrow in (D) is pointing at the small islet in the picture. All pictures are at 40× magnification. Scale bars are 50 μm. All data are expressed as means ± standard error of the mean (n = 5–11 mice/group). P values were calculated using a 1-way analysis of variance or Student t test where appropriate. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001. Red bars indicate differences between Tlr3+/+ groups at different time points as indicated, dashed black bars indicate differences between Tlr3−/− groups at different time points as indicated, solid black bars indicate differences between Tlr3+/+ and Tlr3−/− groups at the same time points as indicated. v−, no CVB4 infection; v+, CVB4 infection.

Representative images of insulin and CD3+ immunohistochemical staining of pancreas preparations obtained from the Tlr3+/+ and Tlr3−/− NOD mice at 8 weeks of age (ie, before CVB4 infection) are depicted (Figs. 1C, D, respectively; top rows). At 8 weeks of age, insulin immunostaining in islets from Tlr3+/+ NOD mice was lower than that in Tlr3−/− NOD mice before CVB4 infection (Fig. 1C, top row). CD3+ T-cell infiltration of islets were significantly greater in the Tlr3+/+ NOD mice compared with Tlr3−/− NOD mice at 8 weeks of age (Fig. 1B); the immunostaining pattern of CD3+ T-cell infiltration (insulitis) in the islets obtained from Tlr3+/+ and Tlr3−/− NOD mice at 8 weeks of age roughly matches the percentages of CD3+ T-cell infiltration (23% in Tlr3+/+ vs 15% Tlr3−/− NOD mice) at 8 weeks of age (Fig. 1D, top row).

At 3 days after CVB4 infection, islet immunohistochemical staining revealed that islets are still intact in both cohorts and insulin staining was still present in both Tlr3+/+ NOD mice and Tlr3−/− NOD mice (Fig. 1C; bottom row). CD3+ T-cell infiltration was again higher in the Tlr3+/+ NOD mice and greater exocrine pancreatic inflammation is apparent in the Tlr3+/+ NOD mice compared with Tlr3−/− NOD mice (Fig. 1D; bottom row) after CVB4 infection at 3 days.

Cytokine and Chemokine Expression in Islets of Uninfected and CVB4-Infected Tlr3+/+ and Tlr3−/− NOD Mice

We next evaluated the expression of the key chemokine, Cxcl10, and cytokines (Il1b, Ifnb1, Tnfa, and Tgfb1) between the ages of 4 to 8 weeks during CD3+ T-cell infiltration (insulitis) of islets isolated from uninfected and CVB4-infected female Tlr3+/+ and Tlr3−/− NOD mice, as well as their expression 3 days after CVB4 infection (Figs. 2–6). Each of these factors has been shown previously to be involved in T1DM development.[39-42] The current studies were performed to determine the influence of TLR3 on the expression of each of these genes during insulitis between 4 and 8 weeks of age in female Tlr3+/+ and Tlr3−/− NOD mice and determine how they each collectively or independently contribute to establishing the “critical threshold” of insulitis at 8 weeks of age and what happens to these genes at 3 days after CVB4 infection. Moreover, this was done to determine whether these genes are involved in the acceleration of CVB4-induced diabetes observed in female Tlr3+/+ NOD mice or conversely the protection of female Tlr3−/− NOD mice from CVB4-induced onset of diabetes observed 14 days after virus infection.[23]

FIGURE 2

Cxcl10 gene expression in islets of uninfected and CVB4-infected female Tlr3+/+ and Tlr3−/− NOD mice. Multiple (25–65) individual islets from each female Tlr3+/+ and Tlr3−/− NOD mouse (n = 3–4 mice/group) were isolated using laser capture microdissection, and then all islets from all mice/group (the total number of islets represented per group = 75–260) were pooled to isolate enough RNA for subsequent Taqman RT-qPCR analysis of Cxcl10 gene expression. All data are represented as 1/∆Ct. Red bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ groups at different time points as indicated, dashed black bars indicate ≥1.5-fold change in gene expression between Tlr3−/− groups at different time points as indicated, and solid black bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ and Tlr3−/− groups at the same time points as indicated. Numbers above bars indicate the actual fold change in gene expression between indicated groups having a ≥1.5-fold change. v−, no CVB4 infection; v+, CVB4 infection.

FIGURE 6

Tgfb1 gene expression in islets of uninfected and CVB4-infected female Tlr3+/+ and Tlr3−/− NOD mice. Multiple (25–65) individual islets from each female Tlr3+/+ and Tlr3−/− NOD mouse (n = 3–4 mice/group) were isolated using laser capture microdissection, and then all islets from all mice/group (the total number of islets represented per group = 75–260) were pooled to isolate enough RNA for subsequent Taqman RT-qPCR analysis of Tgfb1 gene expression. All data are represented as 1/∆Ct. Red bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ groups at different time points as indicated, dashed black bars indicate ≥1.5-fold change in gene expression between Tlr3−/− groups at different time points as indicated, and solid black bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ and Tlr3−/− groups at the same time points as indicated. Numbers above bars indicate the actual fold change in gene expression between indicated groups having a ≥1.5-fold change. v−, no CVB4 infection; v+, CVB4 infection.

Cxcl10 expression in isolated islets obtained from uninfected female Tlr3+/+ NOD and Tlr3−/− NOD mice was higher in the Tlr3+/+ NOD islets compared with Tlr3−/− NOD islets between the age of 4 and 8 weeks with 6 weeks of age having the largest fold change (Fig. 2). Cxcl10 expression in islets of Tlr3+/+ NOD mice increased dramatically 3 days after CVB4 infection (2.2-fold), whereas the Cxcl10 expression in response to the CVB4 infection in islets of Tlr3−/− NOD mice was less (1.7-fold) at 3 days after CVB4 infection compared with that in the Tlr3 NOD mice (Fig. 2). Moreover, the Cxcl10 expression was lower in the islets of the Tlr3 NOD mice (1.6-fold) than that in the Tlr3+/+ NOD mice at 3 days after CVB4 infection (Fig. 2).

Il1b expression in islets of nonvirus-infected Tlr3+/+ NOD mice was highest at 4 weeks of age and greater than in Tlr3−/− NOD mice at each age; however, expression did not appreciably increase between 4 and 8 weeks of age in either cohort of mice (ie, Tlr3 and Tlr3 NOD mice; Fig. 3). Il1b expression increased dramatically in isolated islets from both Tlr3+/+ and Tlr3−/− NOD mice at 3 days after CVB4 infection (2.0- and 2.5-fold, respectively) and there was no appreciable difference in Il1b expression between Tlr3 and Tlr3 NOD mice at this time point (Fig. 3).

FIGURE 3

Il1b gene expression in islets of uninfected and CVB4-infected female Tlr3+/+ and Tlr3−/− NOD mice. Multiple (25–65) individual islets from each female Tlr3+/+ and Tlr3−/− NOD mouse (n = 3–4 mice/group) were isolated using laser capture microdissection, and then all islets from all mice/group (the total number of islets represented per group = 75–260) were pooled to isolate enough RNA for subsequent Taqman RT-qPCR analysis of Il1b gene expression. All data are represented as 1/∆Ct. Red bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ groups at different time points as indicated, dashed black bars indicate ≥1.5-fold change in gene expression between Tlr3−/− groups at different time points as indicated, and solid black bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ and Tlr3−/− groups at the same time points as indicated. Numbers above bars indicate the actual fold change in gene expression between indicated groups having a ≥1.5-fold change. v−, no CVB4 infection; v+, CVB4 infection.

Islet Ifnb1 expression was relatively low in both Tlr3+/+ and Tlr3−/− NOD mice at 4 weeks of age, increased in both cohorts at 6 weeks of age, and was higher in the Tlr3−/− NOD mice at 8 weeks of age compared with infected Tlr3+/+ (Fig. 4). After CVB4 infection, Ifnb1 expression increased in islets of both cohorts of mice (Fig. 4, red and dashed black bars); however, response to the virus was almost twice as high (1.8-fold) in the Tlr3−/− NOD mice at 3 days after CVB4 infection compared with Tlr3+/+ NOD mice (Fig. 4, solid black bar).

FIGURE 4

Ifnb1 gene expression in islets of uninfected and CVB4-infected female Tlr3+/+ and Tlr3−/− NOD mice. Multiple (25–65) individual islets from each female Tlr3+/+ and Tlr3−/− NOD mouse (n = 3–4 mice/group) were isolated using laser capture microdissection, and then all islets from all mice/group (the total number of islets represented per group = 75–260) were pooled to isolate enough RNA for subsequent Taqman RT-qPCR analysis of Ifnb1 gene expression. All data are represented as 1/∆Ct. Red bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ groups at different time points as indicated, dashed black bars indicate ≥1.5-fold change in gene expression between Tlr3−/− groups at different time points as indicated, and solid black bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ and Tlr3−/− groups at the same time points as indicated. Numbers above bars indicate the actual fold change in gene expression between indicated groups having a ≥1.5-fold change. v−, no CVB4 infection; v+, CVB4 infection.

Islet expression of both Tnfa and Tgfb1 were higher in islets of Tlr3+/+ NOD mice compared with that in islets of Tlr3−/− NOD mice at 4 and 6 weeks of age; however, at 8 weeks of age, there were no appreciable differences in islet expression of either cytokine (Figs. 5, 6, respectively). At 3 days after CVB4 infection, the expression of Tnfa increased dramatically (Fig. 5, red and dashed black bars) and was essentially equivalent in islets from both the Tlr3+/+ and Tlr3−/− NOD mice (Fig. 5). Similar to Ifnb1 expression (Fig. 4), Tgfb1 expression in response to CVB4 infection was greater (1.6-fold) in Tlr3−/− NOD mice at 3 days after infection compared with Tlr3+/+ NOD mice (Fig. 6, solid black bar).

FIGURE 5

Tnfa gene expression in islets of uninfected and CVB4-infected female Tlr3+/+ and Tlr3−/− NOD mice. Multiple (25–65) individual islets from each female Tlr3+/+ and Tlr3−/− NOD mouse (n = 3–4 mice/group) were isolated using laser capture microdissection, and then all islets from all mice/group (the total number of islets represented per group = 75–260) were pooled to isolate enough RNA for subsequent Taqman RT-qPCR analysis of Tnfa gene expression. All data are represented as 1/∆Ct. Red bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ groups at different time points as indicated, dashed black bars indicate ≥1.5-fold change in gene expression between Tlr3−/− groups at different time points as indicated, and solid black bars indicate ≥1.5-fold change in gene expression between Tlr3+/+ and Tlr3−/− groups at the same time points as indicated. Numbers above bars indicate the actual fold change in gene expression between indicated groups having a ≥1.5-fold change. v−, no CVB4 infection; v+, CVB4 infection.

In short, islet expression of the cytokines, Il1b, Ifnb1, Tnfa, and Tgfb1, and the chemokine Cxcl10 tended to be higher in islets of uninfected Tlr3+/+ NOD mice compared with islets of uninfected Tlr3−/− NOD mice from 4 to 6 weeks of age (Figs. 2–6). By 8 weeks of age, islet expression of nearly all the genes was comparable in Tlr3+/+ and Tlr3−/− NOD mice (Figs. 2–6), with the exception of Ifnb1 which was more highly expressed (3.3-fold) in islets of uninfected Tlr3−/− NOD mice relative to islets of age-matched, uninfected Tlr3+/+ NOD mice (Fig. 4). While CVB4 infection caused increases in islet expression of all genes evaluated at 3 days after CVB4 infection compared with their expression in islets of 8-week-old, uninfected mice (Figs. 2-6), Cxcl10 was the only gene that was more highly expressed in islets of CVB4-infected Tlr3+/+ NOD mice (Fig. 2), a large percentage of which develop diabetes by 14 days after infection,[23] compared with islets of CVB4-infected Tlr3−/− NOD mice, which do not develop diabetes by 14 days after infection.[23] However, Ifnb1 and Tgfb1 expression was increased in islets of CVB4-infected, Tlr3−/− NOD mice compared with islets of CVB4-infected Tlr3+/+ NOD mice (Figs. 4, 6, respectively).

DISCUSSION

Female NOD mice spontaneously develop an autoimmune, T-cell–mediated insulitis, with gradual β-cell destruction/loss and onset of T1DM beginning at age approximately 15 weeks in a nonsterile environment, and by 25 weeks of age nearly 80% of the cohort will be diabetic.[43] Recently, the important role of NOD mice as an animal model in T1DM research has been reviewed.[44] Viruses and CVB4, in particular, have been investigated as principle triggers initiating, as well as accelerating, the onset of T1DM. Multiple investigations have demonstrated that CVB4 infection in Tlr3 (ie, wild-type) female NOD mice can induce the rapid onset of diabetes within 2 weeks after infection; however, there is a “critical threshold” of spontaneous preexisting insulitis (at 8–10 weeks of age),[4] required for this CVB4 acceleration of T1DM to occur.[4,17,23,26] However, Tlr3 female NOD mice are protected from CVB4 acceleration of T1DM.[23] The results of these studies indicate that Tlr3 female NOD mice have increased insulitis and an increase in Cxcl10, Il1b, Tnfa, and Tgfb1 expression between 4 and 6 weeks of age compared with Tlr3 female NOD mice. Taken together, our data offer insight into how TLR3 contributes to the establishment of the “critical threshold of insulitis” necessary for viral acceleration of T1DM.

Toll-like receptor 3 is one of a family of pattern recognition receptors also involved in the innate immune response, which recognizes viral or endogenous dsRNA released by virus-damaged β cells. Toll-like receptor 3 activation and signaling results in the production of key cytokines and chemokines that are known to play a role in the pathogenesis of T1DM in the female NOD mouse; however, although present in islets by 2 weeks of age, and during the insulitis, which occurs between 4 and 8 weeks of age, we previously demonstrated that CVB4 induction/acceleration of diabetes onset in female wild-type, Tlr3+/+, NOD mice occurred after CVB4 infection at 8 weeks of age, yet similarly CVB4-infected Tlr3−/− NOD mice are protected from CVB4-accelerated induction of diabetes at this age.[23] Furthermore, Tlr3+/+ NOD female mice demonstrate higher concentrations of CD4+ and CD8+ T cells at 14 days after CVB4 infection compared with Tlr3−/− NOD mice.[23] This study demonstrates that between 4 and 8 weeks of age, islets isolated from female Tlr3+/+ NOD mice exhibit significantly greater islet inflammation (insulitis [Fig. 1A] and T-cell infiltration [Fig. 1B]) than Tlr3−/− NOD mice. Islet expression of the chemokine Cxcl10 (Fig. 2) as well as the cytokines Il1b (Fig. 3), Tnfa (Fig. 5), and Tgfb1 (Fig. 6) were also higher between 4 and 6 weeks of age in uninfected Tlr3+/+ NOD mice compared with islets of uninfected Tlr3−/− NOD mice, yet by age 8 weeks, the earliest age at which the “critical threshold” for CVB4 acceleration of T1DM in the Tlr3+/+ NOD mice occurs, there was no appreciable differences in expression of these chemokines and cytokines between the 2 cohorts (Figs. 2–6). CXCL10 is a major contributor to insulitis, as it is secreted from damaged β cells signaling dendritic cells, macrophages, and other T cells to sites of destruction.[39,45,46] This increase in expression leading up to 8 weeks of age may aid in making the pancreatic microenvironment ideal for CVB4-accelerated T1DM to occur. In addition, interleukin 1β and tumor necrosis factor α also contribute to this hostile environment by increasing inflammation and have been implicated in β-cell destruction and diabetes development.[40,42,47-52] However, the increase in transforming growth factor-β may have been trying to counteract the proinflammatory cytokines by preventing self-reactive T cells from triggering T1DM onset[41,51,53] as a decrease in transforming growth factor-β signaling has led to an acceleration of T1DM in NOD mice.[41]

Fifty percent of Tlr3+/+ NOD mice are known to develop diabetes by 14 days after CVB4 infection, whereas Tlr3−/− NOD mice seem to be “protected” from CVB4-accelerated diabetes.[23] After CVB4 infection, Tlr3+/+ NOD mice histologically again exhibited greater insulitis and T-cell infiltration at 3 days after infection than Tlr3 NOD mice (Figs. 1A, B) with similar amounts of insulin staining (Fig. 1C). These data support our previous findings at 14 days after CVB4 infection and are likely to, at least in part, contribute to the 50% incidence of diabetes in the CVB4-infected Tlr3+/+ NOD mice and the protection of the Tlr3 NOD mice that was observed previously.[23] Thus, this study agrees with our earlier studies demonstrating that CVB4-induced diabetes in the wild-type (ie, Tlr3) female NOD mice is TLR3 mediated[23] and show for the first time that TLR3 signaling enhances T-cell infiltration of young (4–8 weeks) islets along with expression of Cxcl10 and cytokines Il1b, Tnfa, and Tgfb1, as Tlr3+/+ NOD mouse islets exhibit more T cells/greater insulitis leading to the establishment of the “critical threshold” of insulitis required for CVB4-induced diabetes at 8 weeks of age. In addition, Ifnb1 and Tgfb1 levels increased dramatically in the Tlr3−/− NOD mice in response to CVB4 compared Tlr3+/+ NOD mice, suggesting that they might actually be protective from viral-induced diabetes in this mouse model. Previous reports have shown that after viral infection, IFN-β1 expression is induced via IRF3 activation, through the TLR3 pathway.[54] Moreover, recent studies have shown that IFN-β1−/− mice were more susceptible to West Nile virus and had increased mortality, compared with IFN-β1+/+ mice.[55,56] Exactly how/why Ifnb1 and Tgfb1 expression increases in the absence of Tlr3 in response to CVB4 is yet to be determined; however, one possibility may be due to compensation by other viral sensing pathways. Thus, although TLR3 is a key primary and major mediator of CVB4-induced T1DM in NOD mice, in its absence, there can be redundancy because some Tlr3−/− NOD mice still develop T1DM. This redundancy is likely through dsRNA recognition by other pattern recognition receptors that sense cytoplasmic dsRNA such as retinoic acid-inducible gene 1 (RIG1) and/or melanoma differentiation associated factor 5.[57]

As noted earlier, there was a drop in insulitis and T-cell infiltration of islets from 4 to 6 weeks of age in Tlr3 female NOD mice and a subsequent rise again at 8 weeks of age similar to the levels seen at 4 weeks of age (Figs. 1A, B, respectively). Insulitis is considered a dynamic lesion at all stages of the disease with a continuous influx and efflux of immunocytes and progresses over time in response to immunologic and environmental influences.[58] Thus, the vacillation in insulitis and T-cell infiltration from 4 to 8 weeks of age is likely a result of the fluctuation of different T-cell subtypes (and other CD3+ immunocytes) in and out of the islets of the younger female NOD mice (ie, up to 8 weeks of age) before the establishment of the critical threshold of insulitis necessary for CVB4 to trigger T1DM in these mice (that typically occurs between 8–10 weeks of age), and the insulitis then continues to increase over time. This fluctuation in insulitis and T-cell infiltration of islets was not observed in the female Tlr3 NOD mice, supporting the previously reported finding that Tlr3 female NOD mice are protected from CVB4 acceleration of T1DM[23] and the results of the studies described herein indicating that Tlr3 female NOD mice have decreased insulitis and decreased Cxcl10, Il1b, Tnfa, and Tgfb1 expression between 4 and 6 weeks of age, compared with Tlr3 female NOD mice.

This study was not without limitations. Harvested islets had to be pooled for gene expression assays; thus, results are relative comparisons of gene expression rather than statistical comparisons. The NOD mouse model is ideal for studying T1DM, however, because of the heterogeneity of these animals, not all develop diabetes at the same time after CVB4 infection.

However, these studies are the first of their kind to (1) offer insight into the basic mechanisms by which TLR3 contributes to the “critical threshold” of insulitis in female NOD mice, contributing to a better understanding of the natural progression of T1DM in these mice and providing a more detailed description of what constitutes the “critical threshold” of insulitis and (2) describe the role of TLR3 in mediating CVB4-induced acceleration of T1DM in female NOD mice.