Edited by: Aida Martinez-Sanchez, Imperial College London, United Kingdom
Reviewed by: Ken Coppieters, Novo Nordisk, Denmark; Jason Collier, Pennington Biomedical Research Center, United States
*Correspondence: Lars Krogvold,
†These authors have contributed equally to this work and share first authorship
This article was submitted to Diabetes: Molecular Mechanisms, a section of the journal Frontiers in Endocrinology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
The Diabetes Virus Detection (DiViD) study has suggested the presence of low-grade enteroviral infection in pancreatic tissue collected from six of six live adult patients newly diagnosed with type 1 diabetes. The present study aimed to compare the gene and protein expression of selected virally induced pathogen recognition receptors and interferon stimulated genes in islets from these newly diagnosed type 1 diabetes (DiViD) subjects vs age-matched non-diabetic (ND) controls.
RNA was extracted from laser-captured islets and Affymetrix Human Gene 2.0 ST arrays used to obtain gene expression profiles. Lists of differentially expressed genes were subjected to a data-mining pipeline searching for enrichment of canonical pathways, KEGG pathways, Gene Ontologies, transcription factor binding sites and other upstream regulators. In addition, the presence and localisation of specific viral response proteins (PKR, MxA and MDA5) were examined by combined immunofluorescent labelling in sections of pancreatic tissue.
The data analysis and data mining process revealed a significant enrichment of gene ontologies covering viral reproduction and infectious cycles; peptide translation, elongation and initiation, as well as oxidoreductase activity. Enrichment was identified in the KEGG pathways for oxidative phosphorylation; ribosomal and metabolic activity; antigen processing and presentation and in canonical pathways for mitochondrial dysfunction, oxidative phosphorylation and EIF2 signaling. Protein Kinase R (PKR) expression did not differ between newly diagnosed type 1 diabetes and ND islets at the level of total RNA, but a small subset of β-cells displayed markedly increased PKR protein levels. These PKR+ β-cells correspond to those previously shown to contain the viral protein, VP1. RNA encoding MDA5 was increased significantly in newly diagnosed type 1 diabetes islets, and immunostaining of MDA5 protein was seen in α- and certain β-cells in both newly diagnosed type 1 diabetes and ND islets, but the expression was increased in β-cells in type 1 diabetes. In addition, an uncharacterised subset of synaptophysin positive, but islet hormone negative, cells expressed intense MDA5 staining and these were more prevalent in DiViD cases. MxA RNA was upregulated in newly diagnosed type 1 diabetes vs ND islets and MxA protein was detected exclusively in newly diagnosed type 1 diabetes β-cells.
The gene expression signatures reveal that pathways associated with cellular stress and increased immunological activity are enhanced in islets from newly diagnosed type 1 diabetes patients compared to controls. The increases in viral response proteins seen in β-cells in newly diagnosed type 1 diabetes provide clear evidence for the activation of IFN signalling pathways. As such, these data strengthen the hypothesis that an enteroviral infection of islet β-cells contributes to the pathogenesis of type 1 diabetes.
The possible influence of viral infections in the development of type 1 diabetes was first postulated by Harris as long ago as 1899, who described a person in whom diabetes developed soon after a bout of mumps (
In the DiViD study, pancreatic specimens of optimal quality have been collected, with written consent, from six live patients (age 24-35 years) with newly diagnosed type 1 diabetes (
The objectives of the present study were to explore whether additional antiviral responses are also detectable in the islets of Langerhans of individuals with type 1 diabetes. This was achieved by extracting RNA directly from islets of Langerhans using laser capture microdissection and then employing global gene expression profiling as a tool to gain insight into the biological and pathological processes associated with the onset of type 1 diabetes. Secondly, led by the RNA data, the expression of three antiviral tissue response proteins (Interferon-inducible Myxovirus resistance Protein A (MxA), Protein kinase R (PKR) and Melanoma differentiation-associated protein 5 (MDA5) was examined in islets from DiViD subjects and compared with that in islets from similarly aged non-diabetic nPOD-controls (
Pancreatic samples from the six type 1 diabetic patients included in the DiViD study were used in the present investigation. The DiViD study was approved by The Norwegian Governments Regional Ethics Committee. Written informed consent was obtained from all cases after oral and written information from the diabetologist and the surgeon separately. Three women and three men, three to nine weeks after diagnosis, (aged 24-35 years) participated. Detailed clinical characteristics have been described (
Six otherwise healthy, non-diabetic organ donors from within the network of Pancreatic Organ Donors (nPOD) biobank were included as controls for immunofluorescent examinations, and the same 6 plus an additional 12 cases were included as controls for the islet laser-capture experiments (
Eight µm thick cryo-sections from the DiViD-cases and the 18 nPOD controls were fixed and dried. 10-20 randomly chosen islets were captured from each section by LCM. The same islets were captured from six consecutive sections. RNA of optimal quality for further analyses was obtained from DiViD-cases 2-6. We have previously described isolation of RNA and obtaining transcriptome data on the Affymetrix Human Gene 2.0 ST arrays (
4μm formalin fixed paraffin embedded sections were obtained from 6/6 DiViD pancreatic biopsies and six similarly aged nPOD controls. The presence and localisation of viral response proteins was then examined by combined immunofluorescent labelling. In brief, sections were dewaxed by submersion in Histoclear™ (Fisher Scientific, UK); antigens were unmasked by heating sections for 20 minutes under pressure in 10mM Citrate (pH 6) and non-specific binding was ameliorated by incubating sections in 5% normal goat serum (Vector, UK). Sections were then incubated with primary antibodies specifically directed against the viral response proteins of interest, in combination with antibodies targeting the various islet hormones (described in
After filtering, statistical analysis identified 500 differentially expressed genes in micro dissected islets when comparing the five DiViD-cases employed vs 18 controls (Student’s t-test, p<0.001, fold change >1.1 and false discover rate < 0.1%). Among these 500 genes, several Gene Ontologies (GO) were enriched significantly (
Gene Ontology analyses of gene expression in DiViD cases compared to controls.
Ontologies that showed significant enrichment in cases compared to non-diabetic controls | ||
---|---|---|
Biological processes: | ||
Viral reproduction genes: | 37 genes | adjP=5.85e-09 |
Viral reproductive process | 33 genes | adjP=2.50e-09 |
Viral infectious cycle | 24 genes | adjP=1.80e-11 |
Viral genome expression | 21 genes | adjP=9.27e-12 |
Viral transcription | 21 genes | adjP=9.27e-12 |
Metabolic processes: | ||
Translation genes | 43 genes | adjP=7.70e-10 |
Translational initiation | 23 genes | adjP=1.94e-13 |
Translational termination | 15 genes | adjP=1.69e-09 |
Translational elongation | 18 genes | adjP=1.80e-11 |
Nuclear-transcribed mRNA catabolic process | ||
Nonsense-mediated decay | 16 genes | adjP=5.85e-09 |
Molecular function: | ||
RNA binding | 40 genes | adjP=7.18e-09 |
Ribosome binding | 5 genes | adjP=1.70e-03 |
Translation factor activity, nucleic acid binding | 8 genes | adjP=1.20e-03 |
Structural molecule activity | 33 genes | adjP=1.15e-08 |
Structural constituent of ribosome | 26 genes | adjP=8.35e-18 |
NADH dehydrogenase activity | 5 genes | adjP=1.08e-02 |
ATPase activity | 14 genes | adjP=1.61e-02 |
Cellular component: | ||
Macromolecular complex | 107 genes | adjP=2.66e-11 |
Intracellular | 230 genes | adjP=1.11e-11 |
Intracellular organelle part | 154 genes | adjP=2.99e-11 |
Cytoplasm | 194 genes | adjP=1.95e-13 |
Cytoplasm part | 157 genes | adjP=7.07e-12 |
Ribonucleoprotein complex | 38 genes | adjP=1.37e-13 |
Ribosomes | 29 genes | adjP=4.36e-18 |
Ribosomal subunit | 22 genes | adjP=4.42e-15 |
Cytosolic ribosome | 16 genes | adjP=9.06e-12 |
Pathway and Transcription factor analysis of gene expression in DiViD cases compared to controls.
Enrichment in the following KEGG pathways | |||
---|---|---|---|
Oxidative phosphorylation (27/132 genes) | p-value 1.08 E-27 | ||
Ribosome (20/92 genes) | p-value 2.08 E-21 | ||
Metabolic Pathways (53/1130 genes) | p-value 6.71 E-21 | ||
Enrichment of these canonical pathways: | |||
Mitochondrial dysfunction (30/171 genes) | p-value 5.74 E-22 | ||
Oxidative phosphorylation (25/109 genes) | p-value 1.67 E-21 | ||
EIF2 Signaling (26/185 genes) | p-value 1.04 E-16 | ||
Antigen presentation (11/37) | p-value 1.62 E-11 | ||
Enrichment of promoter and |
|||
Transcription factors involved in cellular growth/apoptosis: | |||
Elk-1, 83 genes | p-value 2.40 E-8 | ||
Tel-2, 44 genes | p-value 1.93 E-4 | ||
Mitochondrial function (GABP), 93 genes | p-value 1.93 E-4 | ||
Interferon regulatory factor 7 (IRF7), 44 genes | p-value 9.85 E-4 |
RNA encoding MxA was present at significantly higher levels in the islets of individuals with type 1 diabetes vs ND islets (p < 0.00005). To verify that this also resulted in an increase in MxA protein, immunostaining profiles were examined in 124 type 1 diabetes islets (70 with residual beta cells) vs 26 islets from 6 non-diabetic subjects. MxA was expressed exclusively in type 1 diabetes islets and was detected in 91.4% (64 of 70) of those islets with residual insulin immunopositivity. MxA immunopositivity was restricted only to β-cells and it was not present in other islet endocrine cells (
Micrographs illustrating the immunofluorescent staining of representative islets from an individual without diabetes (ND; column 1) and two islets from a DiViD patient with diabetes, either with or without residual beta cells (columns 2 and 3 respectively). Green, MxA; Red, Glucagon; Cyan, Insulin; Blue, DAPI. Scale bars, 25um.
No differences in RNA expression of
Micrographs showing immunofluorescence images of representative islets from an individual without diabetes (ND; column 1) and from two different DiViD patients with diabetes (columns 2 & 3). PKR positive beta cells are indicated (yellow lines). Green, PKR; Red, Glucagon; Cyan, Insulin; Blue, DAPI. Scale bars, 25um.
RNA encoding MDA5 was significantly increased in type 1 diabetes islets compared to controls (p < 0,005). However, marked differences in expression were noted between subjects, with the highest levels found in one individual (DiViD case 5) while MDA5 mRNA was elevated less strongly in the remainder.
Among 124 islets examined histologically, MDA5 protein was highly expressed in some α-cells of all DiViD subjects and controls. The protein was rarely detected in the β-cells of controls (it was detected weakly in only 1 of 25 islets imaged), however among the 33 ICIs imaged in the DiViD subjects, 25 (75.8%) exhibited a marked upregulation of MDA5 within β-cells (
Micrographs showing immunofluorescent images of a representative islet from an individual without diabetes (ND; column 1) and from two DiViD patients with diabetes (columns’ 2-3). MDA5+ alpha cells are indicated with arrows and MDA5+ beta cells with dotted lines (columns 2 & 3). Green, MDA5; Red, glucagon; Cyan, Insulin; Blue, DAPI. Scale bars, 25um.
Examples of MDA5+ cells which are synaptophysin + but hormone negative. Green, MDA5; Red, synaptophysin; Cyan, hormone cocktail; insulin, glucagon and somatostatin. Scale bars, 25um.
We have used both RNA expression analysis and protein immunodetection to monitor the induction of antiviral signaling in the islets of DiViD vs control subjects of equivalent ages. The data mining process revealed significant enrichment of several gene ontologies linked to viral infection in the DiViD tissue; findings which accord with gene and protein expression data revealing elevated levels of MxA and MDA5, as well as induction of PKR expression in a subpopulation of beta cells in the DiViD subjects. As such, the results reported here strengthen and expand our previous observations implying the presence of a sustained low-grade enteroviral infection in the pancreatic islet tissue of DiViD cases (
When considering the RNA expression data presented here, it is important to emphasize that, islets harvested by laser microdissection from the DiViD-cases were not pre-selected according to their insulin content or the extent of inflammation; rather, all available islets within any given section were captured to deliver an unbiased analysis. Inevitably, therefore, the islets comprised a heterogeneous population having variable proportions of alpha and beta cells, as well as differential immune cell infiltration (
It should also be emphasized that the data-mining technique is unbiased and observational, such that it is not bound by prior hypotheses. While this is a clear strength, it must also be acknowledged that it carries the risk that false positive differences might be highlighted. Conscious of this limitation, we chose a stringent level of significance for inclusion of gene expression changes (p<0.001). The corollary to these considerations is that we may have thereby omitted certain smaller scale differences in gene expression occurring between the DiViD subjects and controls and, in support of this, we noted that reducing the stringency (top>0.05) results in an increase in the number of differentially expressed genes from 500 to > 21 000. Thus, the application of stringent statistical limits gives confidence that any differences described here, represent genuine differences between the two groups.
The observed enhancement of genes involved in mitochondrial and ER stress pathways is consistent with previously published data. Islet cells from individuals with type 1 diabetes display a partial ER stress response, with evidence of the induction of some, but not all, components of the unfolded protein response (
One unexpected outcome of this study was that we also noted the presence of MDA5 positive, but somatostatin/insulin/glucagon immunonegative cells in type 1 diabetes cases. The identity of these cells has not been fully verified but, despite their lack of hormone staining, they were positive for the endocrine cell marker synaptophysin. Thus, it is possible that they could represent a population of dedifferentiating or nascent beta cells that either have lost, or not yet acquired, insulin immunopositivity. This observation in samples from the DiViD cases, using a different antibody against MDA5 (ab4544, Abcam), has recently been published (
In conclusion, an analysis of islet gene expression signatures in patients with newly diagnosed type 1 diabetes, showing the enrichment of gene ontologies for viral reproduction and the infectious cycle, supports the proposal that enteroviruses are involved in the pathogenesis of type 1 diabetes. The elevated levels of PKR, MxA and MDA5 in type 1 diabetes beta cells, together with our previous studies using these samples which revealed increased HLA-1 and STAT1 in the beta cells, where both nuclear and cytoplasmic was observed, provide firm support for the activation of IFN signaling and other antiviral pathways. As such, these new results strengthen the hypothesis that chronic enteroviral infection contributes to the pathogenesis of type 1 diabetes.
The datasets presented in this article are not readily available because of the sensitive nature of the data and possible high risks associated with patient confidentiality. Requests to access the datasets should be directed to the authors.
The studies involving human participants were reviewed and approved by The Norwegian Governments Regional Ethics Committee,
LK was responsible for clinical coordination and recruitment of patient, data collection, analysis and interpretation, and drafted the manuscript. IM, PL, MR, SR and NM performed, analyzed and interpreted the immunofluorescent analyses. IG, CM and NL were responsible for the laser capture, RNA extraction and subsequent gene expression profile data collection. All authors contributed in analysis and interpretation of the results, and in writing of the manuscript. KD-J was the principal investigator of the study, had the initial idea of the DiViD study, and participated in study design, funding, regulatory issues, international collaboration, data collection, analysis and interpretation, and in writing of the manuscript. All authors edited the manuscript. LK and KD-J are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors contributed to the article and approved the submitted version.
The project was funded by South-Eastern Norway Regional Health Authority (Grant to KDJ), The Novo Nordisk Foundation (Grant to KD-J) and through the PEVNET Study Group funded by the European Union’s Seventh Framework Programme [FP7/2007-2013] under grant agreement n°261441 PEVNET. The participants of the PEVNET consortium are described at
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
The authors thank specialist nurse Trine Roald, Oslo University Hospital, Norway, who has provided invaluable efforts in coordination of the study, nurses and doctors at the local hospitals, providing contact with the patients, and finally the patients who participated in this study.
The Supplementary Material for this article can be found online at:
MDA5, Melanoma differentiation-associated protein 5; MxA, Myxovirus resistance Protein A; ND, non-diabetic; PKR, Protein kinase R; STAT1, signal transducer and activator of transcription 1; VP1, enteroviral capsid protein 1.