A Critical Role for the Type I Interferon Receptor in Virus-Induced Autoimmune Diabetes in Rats

Type 1 diabetes (T1D) is a T-cell–mediated autoimmune disease that destroys insulin-producing pancreatic β-cells (1). It is heritable but non-Mendelian, and genetic susceptibility loci are insufficient for predicting diabetes onset; most people with risk alleles never become diabetic (2). Interaction of genes with environmental factors has been invoked as a determinant of disease (3,4). Viral infection, particularly with enterovirus, is believed to be a key environmental modulator of T1D, and its possible role in pathogenesis has been reviewed in detail (5,6). The mechanisms that underlie viral triggering of T1D remain unclear; β-cell infection, bystander activation, antigenic spreading, and molecular mimicry have been proposed. Alternatively, viruses could prevent T1D through immunoregulation or induction of protective immunity (7).

To gain better insight into the mechanism of virus-induced diabetes, we used a rat model of the disease. Rats are the only naturally occurring virus-induced T1D model that closely resembles that of human T1D in terms of histopathology, pathogenesis, lack of sex bias, and MHC class II association (8). Type 1–like autoimmune diabetes, both spontaneous and inducible, is relatively common among inbred rat strains, which, like humans, express a high-risk class II MHC haplotype; in rats, this is designated RT1B/Du. Among susceptible rat strains, the LEW.1WR1 strain has been particularly useful. About 2.5% of LEW.1WR1 rats develop T1D spontaneously, typically during their early reproductive years; both sexes are affected, and islets show insulitis (9). Various perturbations of the immune system, however, can efficiently trigger autoimmune diabetes in up to 100% of animals. Perturbants include regulatory T-cell (Treg) depletion, innate immune activation with thrice-weekly doses of polyinosinic:polycytidylic acid (poly I:C), and infection with Kilham rat virus (KRV) or rat cytomegalovirus. LEW.1WR1 rats also develop diabetes at an increased rate (18%) after Coxsackie B serotype 4 (CVB4) infection, but only if pretreated with a low dose of poly I:C daily for 3 days before viral challenge (10). Of note, weanling but not adult LEW.1WR1 rats are prone to diabetes after any viral challenge, making these animals a faithful model of the generally juvenile aspect of T1D.

After poly I:C challenge or during viral infection, a cascade of cytokines, including type I interferon (IFN) (i.e., IFN-α/β [Supplementary Fig. 1]), could contribute to the induction of diabetes in LEW.1WR1 rats. Furthermore, genome-wide association studies (GWASs) have established associations between the risk for human T1D and polymorphisms in genes that mediate type I IFN responses, including IFIH1 (11,12), Ebi2 (13), and Tyk2 (14–16). Thus, we aimed to define the role of type I IFN on the development of autoimmune diabetes in LEW.1WR1 rats by disrupting the IFNAR1 subunit of the type I IFN receptor complex, a key component of IFN signaling. The generation of knockout rats has previously been challenging, but advances in zinc finger nuclease gene targeting (17) and, more recently, clustered regularly interspaced short palindromic repeats (CRISPR)–associated protein 9 (Cas9) technology (18) have made this feasible. We generated Ifnar1^−/− LEW.1WR1 rats using CRISPR-Cas9 gene editing and challenged weanling wild type (WT) and Ifnar1^−/− rats with either poly I:C or KRV and assessed for the development of diabetes. We found that IFNAR1 deficiency protects against diabetes.

LEW.1WR1 rats (RT1B/Du) were from Biomere (Worcester, MA). They develop spontaneous diabetes at a rate of ∼2.5% (9), but treatment with poly I:C (9) or infection with viruses from several families (10) increases the frequency of diabetes to 30–100%. Animals were housed in viral antibody–free conditions, confirmed monthly to be serologically free of rat pathogens (19), and maintained in accordance with institutional and national guidelines (20).

An Ifnar1 target region in exon 4, encoding the IFN-binding domain, was disrupted in the genome of the LEW.1WR1 rat using the CRISPR-Cas9 method. The IFNAR1ex4_guide RNA (gRNA) 2 target site (AGGAGAGATGTAGACTA|GTATGG) includes an overlapping SpeI restriction site. The IFNAR1ex4_gRNA3 target site (TCAATTACACGATACGG|ATCTGG) includes an overlapping XhoII restriction site. Note that for both target sites, the cleavage site is indicated with a vertical line and the protospacer adjacent motif (PAM) sequence is underlined. To confirm high activity of the single guide RNA (sgRNA)/Cas9 nucleases before embryo injection, the guide sequences were cloned into plasmid pX330 (Plasmid #42230; Addgene, Cambridge, MA) (21), using the following primers:

• IFNAR1ex4_gRNA2_F 5′-gtggaaaggacgaaacaccgAGGAGAGATGTAGACTAGTA-3′

• IFNAR1ex4_gRNA3_F 5′-gtggaaaggacgaaacaccgTCAATTACACGATACGGATC-3′

• IFNAR1ex4_gRNA2_R 5′-ctatttctagctctaaaacTACTAGTCTACATCTCTCCT-3′

• IFNAR1ex4_gRNA3_R 5′-ctatttctagctctaaaacGATCCGTATCGTGTAATTGA-3′

The full target sites, including PAM sequence, were cloned into the nuclease reporter plasmid M427 (provided by M. Porteus, Stanford University) (22), using the following primers:

• IFNAR1ex4_gRNA2_M427F 5′-gaattcgacgacggcccagGAGAGATGTAGACTAGTATGG-3′

• IFNAR1ex4_gRNA3_M427F 5′-gaattcgacgacggcccTCAATTACACGATACGGATCTGG-3′

• IFNAR1ex4_gRNA2_M427R 5′-aaaattgtgcgctcctgCCATACTAGTCTACATCTCTCCT-3′

• IFNAR1ex4_gRNA3_M427R 5′-aaaattgtgcgctcctgCCAGATCCGTATCGTGTAATTGA-3′

The M427 reporter plasmid expresses green fluorescent protein after cotransfection with a nuclease that cleaves the target site. Nuclease activity was confirmed by examining green fluorescent protein–positive cells after cotransfection into 293T cells with the corresponding nuclease plasmid or with a negative control.

Capped and tailed Cas9 mRNA was prepared using the mMESSAGE mMACHINE T7 Ultra Kit (Ambion, Austin, TX), and gRNA was prepared using the HiScribe T7 High Yield RNA Synthesis Kit (New England Biolabs, Ipswich, MA) as previously described (23). Linear DNA templates for Cas9 mRNA synthesis and sgRNA synthesis through T7 RNA polymerase were prepared by PCR using the pX330 sgRNA plasmids and the following oligonucleotides:

• Cas9T7_F 5′-TAATACGACTCACTATAGGGAGAATGGACTATAAGGACCACGAC-3′

• Cas9T7_R 5′-GCGAGCTCTAGGAATTCTTAC-3′

• IFNAR1ex4_gRNA2_vT7F 5′-ttaatacgactcactataggAGGAGAGATGTAGACTAGTA-3′

• IFNAR1ex4_gRNA3_vT7F 5′-ttaatacgactcactataggTCAATTACACGATACGGATC-3′

• gRNA_RsEq 5′-AAAAAAgcaccgactcggtgccac-3′

Two independent in vitro–transcribed sgRNAs (50 ng) were coinjected with Cas9 mRNA (50 ng) in 0.5-day-old LEW.1WR1 single-cell embryos by intracytoplasmic microinjection to create site-specific DNA double-strand breaks, thereby stimulating targeted gene disruptions. After injection, LEW.1WR1 embryos were transferred into pseudopregnant Sprague Dawley female rats. Embryonic injections and transfers were performed at the University of Massachusetts (UMass) Medical School Transgenic Animal Modeling Core facility. Genomic DNA was isolated from tail samples. The genotypes of individual pups (i.e., presence of insertions/deletions [indels]) were determined by PCR, restriction enzyme digests, and sequencing (Macrogen, Rockville, MD).

Of nine rats born, five founder animals that contained monoallelic or biallelic mutations in Ifnar1 were identified by nuclease screening and used for further breeding (Fig. 1B) to establish two distinct homozygous lines, designated IFNAR1^Δ81 and IFNAR1^Δ81+4, that were based on the indels present. Total RNA was extracted from heart and brain samples of encephalomyocarditis virus (EMCV)–infected rats 2 days postinfection using TRIzol reagent (Sigma-Aldrich, St. Louis, MO). One microgram of RNA was reverse transcribed to cDNA using the QuantiTect Reverse Transcription Kit (QIAGEN, Valencia, CA) according to the manufacturer’s protocol. For RT-PCR analysis, primers targeting a 574-base pair (bp) region between exons 2 and 5 of the Ifnar1 gene (IFNAR1-F [5′-CCGTAGCCTCAGGTGAAGAC-3′] and IFNAR1-R [5′-GCTGTGTCTCTGAAGCGATG-3′]) were designed using Primer-BLAST (www.ncbi.nlm.nih.gov/tools/primer-blast). As a control, primers targeting a 543-bp region between exons 1 and 5 were designed for the Ifnar2 gene (IFNAR2-F [5′-AAGCCCAGAACAGGGGAAAC-3′] and IFNAR2-R [5′-CCAACCACTCGTCAGTCACA-3′]). The reference cDNA sequences used for rat Ifnar1 and Ifnar2 are NM_001105893.1 and XM_006248107.2, respectively. A 50-μL PCR master mix comprising HotStartTaq Master Mix (QIAGEN), nuclease-free water, forward and reverse primers (0.2 μmol/L each), and cDNA (2 μL per reaction) was prepared. For PCR cycling conditions, initial denaturation at 95°C for 15 min was followed by 40 cycles of 95°C for 40 s, 59°C for 45 s, and 72°C for 45 s, with a final extension at 72°C for 10 min. Amplified products were run on a 1.5% agarose gel and visualized with ethidium bromide under ultraviolet illumination after electrophoresis.

Spleens were isolated from rats of either sex and immediately washed with PBS (Corning, Manassas, VA) and then minced and passed through a 40-μm sterile nylon mesh with a 3-mL rubber syringe plunger. Cells were collected by centrifugation at 1,500 rpm for 3 min. The supernatant was discarded, and cells were resuspended in red blood cell lysis buffer (Sigma, St. Louis, MO) for 7 min at room temperature. After removal of erythrocytes, the splenocytes were washed once with PBS, centrifuged and resuspended in appropriate volumes of RPMI medium supplemented with 10% FBS, counted, and seeded in 96-well plates at a density of 1 × 10⁶ cells/well. Cultured splenocytes were stimulated with 1,000 units/mL of recombinant rat IFN-β or IFN-α (PBL Assay Science, Piscataway, NJ) for 18 h, then cell lysates were harvested for total RNA preparation using the RNeasy Plus Mini Kit (QIAGEN).

cDNA was synthesized from 100 ng of total RNA using the QuantiTect Reverse Transcription Kit according to the manufacturer’s protocol. Gene expression was quantified by quantitative RT-PCR (RT-qPCR) with the QuantiTect Primer Assay (Rn_Isg15_1_SG, Rn_Oas1a_1_SG) for interferon-stimulated gene (ISG) 15 and OAS1a. Expression levels were normalized to GusB (Rn_GusB_1_SG QuantiTect Primer Assay). The QuantiFast SYBR Green PCR Kit (QIAGEN) was used for real-time PCR amplification according to the manufacturer’s protocol on a Mastercycler ep realplex (Eppendorf, Hauppauge, NY).

To determine splenic KRV transcript levels, total RNA was extracted from spleens of KRV-infected WT and IFNAR1^Δ81 rats. cDNA was synthesized by using 1 μg of total RNA followed by RT-qPCR as described above. We used previously published primers (24) with the following sequences: forward primer 5′-GGAAACGCTTACTCCGATGA-3′ and reverse primer 5′-AACCGATGTCCTTCCCATTT-3′. The expression levels of viral transcripts were normalized to GusB.

All real-time PCR reactions were run in duplicate, including a no-template control reaction. Fold changes in gene expression of test and control samples were determined by using the 2^−ΔΔCt method.

Studies of induced diabetes were performed in WT LEW.1WR1 and Ifnar1^−/− LEW.1WR1 rats (backcrossed ≥ F5) using two different perturbants known to trigger autoimmunity in WT animals and one not previously tested. High-molecular-weight poly I:C (InvivoGen, San Diego, CA) was administered to rats of either sex at 21–25 days of age by using a dose of 1 μg/g body weight by intraperitoneal (i.p.) injection three times weekly for 3 weeks as previously described (9). Rats were monitored for a total of 39 days after the first injection with poly I:C, defined as day 0. Spleens, pancreata, pancreatic lymph nodes (PLNs), and sera were collected at day 4 from age-matched animals injected on days 0 and 2 with either saline or poly I:C in one experiment.

In a second set of experiments, weanling WT and Ifnar1^−/− rats 21–25 days of age were infected with a single i.p. dose of KRV-UMass strain (1 × 10⁷ plaque-forming units [PFU]) on day 0 and monitored for a total of 39 days for diabetes. KRV was prepared as previously described (25). Spleens, pancreata, PLNs, and sera were collected from age-matched uninfected (i.e., injected with culture media) or KRV-infected animals at day 5 postinoculation in one experiment.

In a third set of experiments, weanling WT LEW.1WR1 rats were infected with a single dose of 1 × 10⁷ PFU of EMCV (ATCC strain VR-129 propagated in BHK-21 cells) by i.p. injection on day 0 and monitored for diabetes for a total of 40 days. EMCV has not been previously tested in the LEW.1WR1 rat model. In one experiment, adult WT and Ifnar1^−/− LEW.1WR1 rats were infected with 1 × 10⁷ PFU of EMCV i.p. and monitored for 14 days. Rats were euthanized if they exhibited gross signs of illness (e.g., ruffling, hunching). Some adult WT and Ifnar1^−/− rats were euthanized 48 h after infection with 1 × 10⁷ PFUs of EMCV, and serum and select organs were harvested and stored at −80°C until used for quantifying viral titers.

Blood glucose concentrations were measured at least three times weekly with a glucometer (Breeze2; Bayer, Carlsbad, CA). Rats were diagnosed as diabetic when the blood glucose concentration exceeded 250 mg/dL on 2 consecutive days.

EMCV was measured in rat serum and organs based on previously published methods (26).

A ProcartaPlex kit (Affymetrix, Santa Clara, CA) was used according to the manufacturer’s instructions to measure cytokines and chemokines (CCL2, interleukin 1β [IL-1β], CCL5, CXCL10) in rat samples. Insulin was measured in serum samples by using an ultrasensitive insulin ELISA kit (ALPCO, Salem, NH).

After the diagnosis of diabetes or at the conclusion of an experiment, rats were euthanized, and pancreata were removed and fixed in 10% buffered formalin. Paraffin-embedded sections of pancreas were sectioned and prepared for light microscopy in the UMass Medical School Morphology Core laboratory (www.umassmed.edu/morphology/protocols). Sections stained with hematoxylin-eosin (H-E) were scored for insulitis as previously described (9) by an experienced reader (J.P.M.) who was not aware of the animal’s glycemic status. Intensity of insulitis was scored as follows: 0, no inflammatory mononuclear cell (MNC) infiltration; 1+, small numbers of infiltrating MNCs with preservation of islet architecture; 2+, moderate infiltrating MNCs with preservation of architecture; 3+, many MNCs, with most islets affected and distortion of islet architecture; 4+, florid infiltration and distorted islet architecture or end-stage islets with or without residual inflammation. Histology images in Fig. 3 were adjusted for clarity by setting the white point for each image using Adobe Photoshop CS6.

Statistical procedures were carried out with either GraphPad Prism version 6 (GraphPad Software, La Jolla, CA) or SPSS version 19 (IBM Corporation, Armonk, NY) software. Survival and disease-free survival were analyzed using Kaplan-Meier methodology; equality of survival distributions was tested by the log-rank statistic (27). Parametric data are given as arithmetic means ± 1 SD or ± SE as indicated in the figure legends and Table 1. Fisher exact test was used to analyze 2 × 2 tables, and the χ² test was used for larger tables. For comparisons of three or more means, we used one-way and two-way ANOVAs and either Bonferroni correction or the least significant differences procedure for posteriori contrasts (27). P < 0.05 was considered statistically significant.

We induced mutations in rat Ifnar1 using a CRISPR-Cas9 strategy. Two sgRNAs were designed to target exon 4 of rat Ifnar1, which, together with exon 5, encodes an IFN-binding domain (Fig. 1A). The sgRNAs were coinjected with Cas9 mRNA into single-cell LEW.1WR1 rat embryos. We assayed Ifnar1 somatic mutations in F0 pups and identified either deletions spanning the region between the two sgRNA/Cas9 target sites (Fig. 1B, lower bands) or small indels at the individual target sites (data not shown). Two lines with germline mutations due to error-prone nonhomologous end joining repair were established and designated IFNAR1^Δ81 and IFNAR1^Δ81+4, and confirmed by sequencing rat genomic DNA from F0 founder, F1 heterozygous, and F2 homozygous animals. Sequences of IFNAR1^Δ81 and IFNAR1^Δ81+4 F2 homozygous animals are shown in Fig. 1C. PCR on cDNA using primers spanning exons 2–5 of Ifnar1 yielded the appropriate-sized product in WT rats but truncated products in IFNAR1^Δ81 and IFNAR1^Δ81+4 rats (Fig. 1D, left panel). Sequence analysis of Ifnar1 PCR products confirmed that the amplicons contained the predicted mutations (data not shown). In contrast, PCR analysis of Ifnar2 cDNA, which was not targeted, revealed the expected products for WT and both IFNAR1 rat lines for a region spanning exons 1–5 (Fig. 1D, right panel).

We tested several commercially available antibodies against rat IFNAR1 but were unable to validate their target specificity. Therefore, we confirmed the IFNAR1 deficiency phenotype of our mutant rats using two approaches. First, we isolated splenocytes from WT IFNAR1^Δ81+4 rats and IFNAR1^Δ81 rats and challenged them in vitro with either recombinant rat IFN-β or IFN-α. At 18 h postchallenge with IFN-β or IFN-α, robust induction of Isg15 was present with WT but not with IFNAR1^Δ81+4 and IFNAR1^Δ81 splenocytes (Fig. 2A). A second ISG, Oas1a, was similarly induced by IFN-β in WT but not IFNAR1-deficient rat splenocytes. The inability to respond to recombinant type I IFNs indicates that the cognate type I IFN receptor was lacking in the Ifnar1-targeted rats.

Second, we examined rat survival after challenge with EMCV, which has been shown to induce type I IFN by engaging the IFIH1-encoded melanoma differentiation–associated protein 5 (MDA5) (28,29). Ifnar1^−/− mice are highly susceptible to EMCV infection compared with control mice (28,30). In addition, viral titers in Ifnar1^−/− mice are much higher than in controls after challenge with viruses such as vesicular stomatitis virus and Semliki Forest virus (30,31). Thus, we expected that rats lacking IFNAR1 would have a heightened susceptibility to EMCV. After EMCV challenge, all WT adult rats survived >14 days postinfection without exhibiting signs of disease. In contrast, 100% of IFNAR1^Δ81+4 rats and 80% of IFNAR1^Δ81 rats died by 4–5 days postinfection (P = 0.0013 and 0.0070, respectively) (Fig. 2B). Serum and heart from IFNAR1-targeted rats also showed high viral titers compared with WT rats 48 h after inoculation with EMCV (Fig. 2C and D), indicating that the type I IFN response is impaired in these rats. The consistent findings between the two IFNAR1-targeted lines suggest that off-target effects from the CRISPR-Cas9 editing are unlikely. From this point on, we considered the two lines equivalent and henceforth term these Ifnar1^−/− rats.

In mice, only certain strains of EMCV are diabetogenic (32), whereas EMCV pathogenesis in rats has been only partly characterized (33). Therefore, we challenged weanling WT LEW.1WR1 rats (n = 7) with 1 × 10⁷ PFU EMCV and monitored for diabetes for 40 days. None of the rats became diabetic over the course of the experiment (data not shown).

Previous studies established that spontaneous diabetes in LEW.1WR1 rats occurs with a cumulative frequency of ∼2.5%, but administration of poly I:C to weanling rats leads to diabetes in 100% of rats (9). We challenged WT and Ifnar1^−/− rats with poly I:C and monitored them for diabetes. Poly I:C administration resulted in diabetes in 13 of 15 (87%) WT rats by 23 days after the first dose (Fig. 3A). In contrast, only 2 of 11 (18%) Ifnar1^−/− rats became diabetic and not until day 28 at the earliest. This difference was highly significant (P < 0.0001). Of note, the difference between WT and Ifnar1^−/− rats was statistically significant regardless of whether the rats were of IFNAR1^Δ81+4 or IFNAR1^Δ81 lineage (Supplementary Fig. 2), again indicating that protection from diabetes is not simply because of an off-target effect.

For poly I:C experiments, animals were euthanized at the time of diabetes or at the end of the study, and serial sections of pancreas were stained for H-E, insulin, and glucagon. Histopathological analysis revealed that insulitis was more severe in diabetic WT animals after poly I:C treatment compared with nondiabetic animals. Islet pathology for all WT animals was associated with a mean insulitis score of 3.11 ± 0.45 (Table 1). End-stage insulitis was present in diabetic WT animals, which was associated with distorted islet architecture, shrunken size, and presence of few residual infiltrating lymphocytes (Fig. 3B). Moderate insulitis was present (2+) in one WT animal that was normoglycemic at the end of the study (Fig. 3B); other WT normoglycemic animals had no evidence of insulitis. In contrast, all normoglycemic Ifnar1^−/− rats were completely free of insulitis, with normal islet size and structure (Fig. 3B). However, insulitis was observed in the two Ifnar1^−/− rats that were diabetic. The mean insulitis score was 0.64 ± 0.43 among all poly I:C–treated Ifnar1^−/− animals (Table 1). The overall concordance between the diabetes phenotype and insulitis scores agrees with our previous findings (34).

Immunohistochemical staining on samples from all WT animals revealed abundant glucagon-positive cells, with a marked decrease in insulin-positive cells in diabetic WT rats and moderate loss of insulin-positive cells in normoglycemic poly I:C–treated WT animals (Fig. 3B). Both insulin and glucagon were abundant in islets of all normoglycemic poly l:C–treated Ifnar1^−/− animals, with some loss of insulin-positive cells only in the small number of Ifnar1^−/− diabetic animals (Fig. 3B). Insulin was measured in all available terminal serum samples; Table 1 shows the mean serum insulin values for WT versus Ifnar1^−/− rats regardless of diabetes status. Overall, WT rats had significantly lower levels of terminal serum insulin compared with Ifnar1^−/− rats, as anticipated.

We assessed cytokines and chemokines in organs harvested from WT and Ifnar1^−/− animals 4 days after the administration of poly I:C (or saline control). CXCL10 and CCL5, whose expression is mediated by IFN, were significantly decreased in spleens from poly I:C–challenged Ifnar1^−/− rats compared with poly I:C–challenged WT rats (Fig. 3C). In contrast, differences were not observed in IL-1β or CCL2 in spleens of WT versus Ifnar1^−/− rats after poly I:C challenge (Fig. 3C). Cytokines and chemokines were also measured in total pancreata, PLNs, and sera. Ifnar1^−/− rats challenged with poly I:C had decreased levels of CXCL10, CCL5, and CCL2 in pancreata as well as decreased CXCL10 in sera compared with poly I:C–challenged WT rats (Supplementary Fig. 3A).

KRV infection induces autoimmune diabetes in LEW.1WR1 rats (10). We conducted KRV infection studies with weanling Ifnar1^−/− rats with two goals: 1) to establish whether rats deficient in IFNAR1 could survive infection with this parvovirus and 2) to see whether the frequency of diabetes would be reduced compared with WT rats. Diabetes was observed in 3 of 16 (19%) Ifnar1^−/− rats infected with KRV and monitored over a 40-day period (Fig. 4A). No other morbidities occurred in these animals over the course of the study. Thus, although Ifnar1^−/− rats succumbed to EMCV infection, they tolerated and survived KRV infection. Nine of 17 (53%) WT rats became diabetic after infection with KRV (Fig. 4A), a rate consistent with that previously reported in KRV-infected LEW.1WR1 rats (10). The difference in the frequency of diabetes in WT versus Ifnar1^−/− rats was statistically significant (P = 0.0461), indicating that the absence of IFNAR1 is partially protective against KRV-induced diabetes. Pancreatic sections from all the normoglycemic KRV-infected Ifnar1^−/− rats at the end of the study showed the complete absence of insulitis, whereas all diabetic rats, whether WT or IFNAR1 deficient, had classical end-stage insulitis (data not shown).

KRV has been reported to be present in spleens of LEW.1WR1 rats after infection by the i.p. route (35). To establish whether IFNAR1 deficiency results in altered cytokine responses after KRV challenge, we measured select cytokines and chemokines in spleens collected from weanling rats that were uninfected or 5 days after KRV infection. CXCL10 and CCL5 were each significantly decreased in spleens from KRV-infected Ifnar1^−/− rats compared with KRV-infected WT rats (Fig. 4B), illustrating an overall diminished type I IFN–driven response in splenic cells after KRV infection in Ifnar1^−/− rats. Of note, levels of IL-1β and CCL2 were significantly higher in KRV-infected Ifnar1^−/− rat spleens compared with KRV-infected WT rat spleens (Fig. 4B). We measured KRV transcript levels in these WT and Ifnar1^−/− rat spleens to determine whether viral replication was altered in the absence of functional IFNAR. Although viral transcript levels were slightly (approximately twofold) higher in Ifnar1^−/− rat spleens compared with WT, the difference was not statistically significant (Fig. 4C). Total pancreatic, PLN, and serum cytokines and chemokines were also measured from these animals, but only limited differences were observed between KRV-infected WT and Ifnar1^−/− rats (Supplementary Fig. 3B).

Collectively, these data underscore the importance of type I IFN-mediated signaling for the development of autoimmune diabetes. The role of IFN signaling has been implicated in the initiation of islet autoimmunity and development of T1D (36). A strong IFN gene signature was identified in the peripheral blood of at-risk children before initiation of islet autoimmunity (37,38). The current findings in Ifnar1^−/− LEW.1WR1 rats firmly define a role for type I IFN and downstream signaling pathways in poly I:C– and KRV-induced diabetes in the rat model and agree with a previous study in which anti-IFN-α antibody administration to T1D-prone BB rats resulted in a trend toward a later onset of poly I:C–induced diabetes (39). They also support human GWAS findings for human T1D risk associations with single nucleotide polymorphisms (SNPs) in genes such as IFIH1 that participate in IFN-regulated pathways. Specifically, a nonsynonymous SNP in IFIH1 resulting in an amino acid change of alanine to threonine at 946 (A946T) of MDA5 is associated with an increased risk of diabetes (11). MDA5 recognizes cytoplasmic long double-stranded RNA (dsRNA) intermediates generated during the replication cycle of CVB or intracellularly delivered synthetic dsRNA analog poly I:C, leading to potent IFN-α/β induction (40). Funabiki et al. (41) characterized a constitutively active form of MDA5 caused by the amino acid mutation G821S that results in type I IFN hyperexpression and causes lupus-like nephritis; they also similarly associated hyperexpression of type I IFN with A946T . The resulting increase in IFN is associated with severe autoimmune disease.

In contrast, four rare-variant nonsynonymous SNPs in IFIH1 were found to be protective against T1D in GWAS (12); at least two of these variants, E627X and I923V, are predicted to decrease MDA5 function with loss of type I IFN responses after viral challenge (42). Recent data have also shown that NOD mice heterozygous for MDA5 were protected from T1D when infected with CVB4; this protective effect was attributed to a unique type I IFN signature that led to expansion of Tregs at the site of autoimmunity (43). Similarly, Ebi2, a regulator of the interferon regulatory factor 7 (IRF7)–driven inflammatory network (IDIN), including human IFIH1 (13), was associated with an increased risk of T1D; the specific Ebi2 polymorphism is associated with increased expression of IDIN genes (13). Finally, the human Tyk2 gene, which was mapped to the possible T1D susceptibility locus (15), encodes the IFNAR1-associated molecule tyrosine kinase 2, and its deficiency results in a reduced antiviral response (14). These data support the critical contribution of IFN-regulated pathways in the development of T1D following environmental insult.

Rat models are particularly useful for understanding the pathogenesis of T1D, specifically in defining the roles of genetic and environmental factors, including viral infection. The mechanism by which KRV induces diabetes in rats has been partially dissected and involves both innate and adaptive immune responses (8). KRV infection of LEW.1WR1 rat primary islets and splenic cells reportedly activates the Toll-like receptor (TLR) 9 signaling pathway, leading to the activation of two major transcription factors, IRF7 and nuclear factor-κB (NF-κB) through the adaptor protein MyD88 (35,44). Treg depletion synergizes with KRV to induce diabetes (45). In our studies, the interferon-stimulated response element (ISRE)–regulated chemokines CCL5 and CXCL10 are decreased in spleens of KRV-infected Ifnar1^−/− rats, whereas NF-κB–driven IL-1β and CCL2 are significantly increased (Fig. 4B), confirming roles for both IRF7 and NF-κB in KRV infection. Of importance, IL-1β is an inflammatory cytokine that has long been implicated in the development of T1D (46,47). The robust IL-1β induction suggests that IL-1β may not be a major contributor in the early stages of diabetes in this model. Of note, loss of IFNAR1 does not affect the overall survival of KRV-infected rats in the manner that it affects survival of EMCV-infected rats. Type I IFN production is elicited by several pathways in the infected host, depending on the specific virus, and each virus may have a unique means of antagonizing or evading the IFN response. For example, a murine parvovirus has been shown to efficiently evade host type I IFN (48), so rodent parvoviruses may possess unique mechanisms to counteract IFN-induced antiviral effectors. Although the impact of type I IFN on KRV replication is not completely defined, ISGs are induced during KRV infection and could contribute to autoimmunity.

Figure 5 provides an overview of IFN signaling pathways in virus-induced diabetes in the context of this study as well as others (49,50). Viral nucleic acid structures are recognized by pattern recognition receptors, including MDA5 and TLR3, after viral infection of β-cells, leading to the activation of the key transcription factors IRF3, IRF7, and NF-κB (51), which is followed by induction of the transcription and synthesis of IFN-β, a subset of ISGs (40), and inflammatory cytokines; the magnitude of these responses may depend on the genetic susceptibility of the host. IFN-β binds to the type I IFN receptor and exerts its antiviral effects through the trimolecular ISGF3 (IFN-stimulated gene factor 3) complex STAT1-STAT2-IRF9, which binds to the ISRE and induces the expression of ISGs for the recruitment of immune cells (e.g., lymphocytes, monocytes, dendritic cells), leading to insulitis and diabetes. Type I IFNs also affect β-cell survival during infections with viruses linked to human T1D (49). We show that the absence of functional IFNAR tempers the onset of diabetes.

We found the incidence of diabetes in our WT LEW.1WR1 rats to be comparable to previously published reports: 38% for KRV (10) and close to 100% for poly I:C (9). Of note, IFNAR1 deficiency results in an incidence of diabetes of ∼18%, regardless of whether the trigger is poly I:C or KRV. This suggests a common mechanism of diabetes and insulitis in IFNAR1 deficiency. The mechanism behind diabetes during IFNAR1 deficiency remains undefined but may relate to IFN production that precedes IFNAR1 signaling events or that may depend on TLR-mediated activation of NF-κB [TLR9 for KRV (44); TLR3 for poly I:C (52)] with subsequent T-cell recruitment. Virus-induced diabetes that occurs independently of the IFNAR1 pathway can be further explored in the rat model and, perhaps, eventually defined through generation of double-knockout rat lines. We have already dissected important genetic factors in virus-induced diabetes in rats. The genome-encoded T-cell receptor element Vβ13 controls genetic susceptibility to diabetes (53); depletion of Vβ13⁺ T cells prevents poly I:C–induced diabetes in LEW.1WR1 rats (54). Additional genetic studies define diubiquitin as a susceptibility gene for virus-induced diabetes in rats (55). With continued advancements in genomics and in CRISPR-Cas9 genome editing, we anticipate exciting developments in autoimmune diabetes in the near future through rat models of diabetes.

In summary, these data advance our understanding of how innate immunity influences the development of T1D. These studies help us to better understand why certain individuals with specific polymorphisms in IFIH1 are either predisposed to or protected from T1D. Viral infection and innate immune activation may initiate early events in β-cells and/or immune cells that ultimately lead to autoimmune attack and T1D in genetically susceptible individuals. In the long term, findings from these studies could be transitioned to diabetes models involving human islets and human immune cells (56) in which type I IFN pathways are disrupted. The current data reinforce the need for novel approaches to diabetes prevention and treatment, such as viral vaccine development (57,58) or even cytokine-modulating therapies.

Acknowledgments. The authors thank Joe Gosselin and Stephen Jones of the UMass Medical School Transgenic Animal Modeling Core facility. They also thank Ping Liu and Melanie Trombly, both of UMass, for technical assistance and with assistance with preparation of the manuscript, respectively. Finally, the authors thank the UMass Medical School Morphology Core laboratory. S.R.O. and M.H.B. are members of the UMass Medical School Mutagenesis Core.

Funding. This work was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK-105837 to R.B.), the American Diabetes Association (1-16-ICTS-086 to J.P.M.), and National Institute of Allergy and Infectious Diseases (R01-AI-092105 and R56-AI-116920 to J.P.W.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. N.Q. designed and performed the experiments, analyzed data, and wrote and critically reviewed the manuscript. S.L. and G.R. designed and performed experiments, analyzed data, and critically reviewed the manuscript. C.Y. and S.R.O. performed experiments, analyzed data, and critically reviewed the manuscript. M.H.B., R.B., and J.P.M. designed the experiments and critically reviewed the manuscript. J.P.W. designed experiments and wrote and critically reviewed the manuscript. J.P.W. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 76th Scientific Sessions of the American Diabetes Association, New Orleans, LA, 10–14 June 2016.