Abstract
It is unknown whether there is a gene signature in pancreas which is associated with type 1 diabetes (T1D). We performed partial pancreatectomies on 30-day preinsulitic, diabetes-prone BioBreeding (BBdp) rats to prospectively identify factors involved in early prediabetes. Microarrays of the biopsies revealed downregulation of endoplasmic reticulum (ER) stress, metabolism and apoptosis. Based on these results, additional investigations compared gene expression in control (BBc) and BBdp rats age ~8, 30 and 60 days using RT-qPCR. Neonates had increased ER stress gene expression in pancreas. This was associated with decreased insulin, cleaved caspase-3 and Ins1 whereas Gcg and Pcsk2 were increased. The increase in ER stress was not sustained at 30 days and decreased by 60 days. In parallel, the liver gene profile showed a similar signature in neonates but with an early decrease of the unfolded protein response (UPR) at 30 days. This suggested that changes in the liver precede those in the pancreas. Tnf and Il1b expression was increased in BBdp pancreas in association with increased caspase-1, cleaved caspase-3 and decreased proinsulin area. Glucagon area was increased in both 30-day and 60-day BBdp rats. Increased colocalization of BIP and proinsulin was observed at 60 days in the pancreas, suggesting insulin-related ER dysfunction. We propose that dysregulated metabolism leads to ER stress in neonatal rats long before insulitis, creating a microenvironment in both pancreas and liver that promotes autoimmunity.
Introduction
Type 1 diabetes (T1D) occurs in a small subset of genetically susceptible individuals whose immune systems destroy most of the pancreatic β-cells. Recent studies suggest that T1D begins in humans at a very early age, but the reasons why the β-cells are targeted and when this begins are unclear (Atkinson et al. 2015). Individuals who develop T1D have a smaller pancreas than nondiabetic individuals and studies indicate involvement of both endocrine and exocrine pancreas (Campbell-Thompson et al. 2015, Husseini et al. 2015). An important question is whether there are defects in the whole pancreas of subjects who progress to T1D that distinguishes them from susceptible individuals who remain disease free (Scott et al. 2017). There have been attempts to investigate early changes in diabetes-prone rodents often with an emphasis on immune-related genes, but there is no consensus on which markers predict onset of T1D. These studies frequently use non-physiological concentrations of pro-inflammatory cytokines to mimic the autoimmune attack on isolated islets. However, there are only a few studies on the natural course of pancreas gene expression in prediabetes (Crevecoeur et al. 2017, Henschel et al. 2018).
One way to identify differences in the target pancreas before T1D development is to surgically remove tissue from young, preinsulitic animals, analyze it and monitor disease development. One can then designate with certainty whether the tissue is from a prediabetic or nondiabetic individual. It is important to perform this analysis before insulitis is present to avoid confounding effects of infiltrating immune cells. In the BioBreeding diabetes-prone (BBdp) rats from our colony, classic insulitis is not apparent before 50 days of age and pancreas biopsies taken at 30 days are histologically normal, have minimal immune cell involvement and represent a very early point in diabetogenesis.
In the present study, microarray and bioinformatics analyses identified differences in genes involved in the endoplasmic reticulum (ER) stress response, metabolism and apoptosis in prediabetic pancreas of BBdp rats. Candidate genes and associated pathways were further investigated in pancreas and liver from non-diabetes-prone (BBc) and BBdp rats at age ~8, 30 and 60 days. Differences in gene expression in both pancreas and liver of prediabetic individuals suggested that diabetogenesis begins very early in the pancreas and may involve crosstalk with other organs.
Materials and methods
Animals
We used the diabetes-prone BioBreeding (BBdp) rat inbred model of spontaneous diabetes to study the prediabetes, preinsulitic period because classic insulitis is absent in 30-day animals. Insulitis is not apparent until ~50 days and approximately 2/3 of the animals develop diabetes before 150 days. Rats were maintained as described previously (Pound et al. 2015) and monitored twice weekly for diabetes development; those with a fasting blood glucose ≥11.1 mmol/L were deemed diabetic. Control BioBreeding (BBc) rats, which do not develop spontaneous diabetes despite sharing a similar genetic background, were used in a follow-up study, which compared nondiabetes-prone BBc vs diabetes-prone BBdp rats at ages ~8, 30 and 60 days. All procedures pertaining to this study were approved by the University of Ottawa and Ottawa Hospital Research Institute Animal Care Committees.
Analysis of pancreas biopsies
We performed 30% partial pancreatectomies (PPx) in BBdp rats aged 30 days (Fig. 1A) essentially as described by Logothetopoulos (Logothetopoulos et al. 1984). When larger biopsies (60–90%) are removed, a regenerative program is induced that results in the formation of numerous focal areas, also known as tubular complexes (Wang et al. 2005). These were infrequent and did not differ comparing nondiabetic vs prediabetic rats. Following laparotomy, the pancreas tail was removed and either fixed for immunohistochemistry (n = 12, cohort 1) or used for gene expression analyses (n = 14, cohort 2). Postoperatively, animals were treated with analgesics: temgesic (0.3 mg/kg) every 12 h and melloxicam (0.2 mg/kg) every 24 h for 3 days. This treatment did not affect diabetes incidence. Animals were followed until 150 days or development of diabetes and then the biopsies were classified as originating from prediabetic or nondiabetic rats.
Microarray analysis of prospective pancreas biopsies from prediabetic and nondiabetic rats. (A) Survival analysis: Kaplan-Meier survival curves were analyzed by log-rank test using Prism 4.0 (GraphPad, La Jolla, USA). No difference in survival was observed between sham-operated animals (n = 23, black line) and rats that received a 30% PPx (n = 26, red line) (49.3 vs 54.6% incidence). (B) Increased endothelium-associated-CD68+ cells were observed in prediabetic rats (red circles) compared to nondiabetic rats (open circles) (n = 6/group). Gene expression in prediabetic and nondiabetic pancreas (n = 6/group) was analyzed using GeneChip™ Rat Gene 2.0 ST Arrays and compared using Expression Console and Transcriptome analysis software. Genes of interest were those that had a 1.5-fold-change or greater (P ≤ 0.1). (C) PCA analysis of probe signals. An animal with late-onset diabetes had a different gene expression profile compared to other prediabetic rats and was excluded from subsequent analysis. (D) Hierarchical clustering analysis of genes of interest demonstrated the pancreatic gene expression signature in prediabetic rats differs from nondiabetic rats prior to the development of insulitis. A full colour version of this figure is available at https://doi.org/10.1530/JOE-18-0328.
Citation: Journal of Endocrinology 239, 2; 10.1530/JOE-18-0328
Microarray analysis
RNA was isolated using a Nucleospin RNA kit (Machery-Nagel, Bethlehem, USA). Gene expression analyses were performed using GeneChip Rat Gene 2.0 ST arrays (ThermoFisher Scientific) of pancreas RNA from prediabetic and nondiabetic BBdp rats (n = 6/group). Candidate genes and pathways were identified using the GeneChip Expression Console and Transcriptome Analysis Console (ThermoFisher Scientific). Because we expected only subtle differences, we identified genes of interest as those ≥1.5-fold-change in expression and P ≤ 0.1. Ontology of gene processes was analyzed using the Gene Ontology enRIchment anaLysis and visuaLizAtion (GOrilla) tool (Eden et al. 2009). Protein interactions were analyzed using multiple protein search in STRING version 10.5 (Szklarczyk et al. 2017).
Immunohistochemistry staining of pancreas biopsy samples
Pancreatic tissues obtained by biopsy were fixed in Universal Molecular Fixative (n = 6/group, Sakura Finetek Inc, Torrance, USA) embedded in paraffin and cut into 5 μm sections, which were labeled using the avidin-biotin complex method (Patrick et al. 2013). Antibodies to the following proteins were used: CD68 and CD8α (Bio-Rad Antibodies). Northern Eclipse (Empix Imaging, Mississauga, Canada) was used to quantify CD68+ and CD8α+ cells (Husseini et al. 2015).
Gene and protein expression of candidates and pathways in BBdp vs control BBc rats
Candidate genes and pathways were further analyzed in pancreas and liver from normoglycemic ~8, 30 and 60 day BBc and BBdp rats (n = 11–15/group) using RT-qPCR. Whole pancreas and liver tissues were obtained at necropsy and RNA was prepared as described earlier. Relative gene expression was quantified with an Applied Biosystems 7500 Real-Time PCR System (ThermoFisher Scientific) using TaqMan primers and Actb as an endogenous control (ThermoFisher Scientific). Gene expression was normalized to the mean value of age-matched control BBc rats and presented as fold-change (2−ΔΔCt).
Immunohistochemistry staining of BBc and BBdp whole pancreas
Pancreas tissues from ~8-, 30- and 60-day BBc and BBdp rats (n = 10–15/group) were fixed in Bouin’s fixative, embedded in paraffin and cut into 5 µM sections as described earlier. Antibodies to the following proteins were used: insulin (Agilent Technologies Canada Inc., Mississauga, Canada), proinsulin (Bio-Techne, Minneapolis, USA), glucagon (MilliporeSigma), BIP (Abcam Inc.). ImageJ (National Institute of Health) was used to quantify the insulin+ and proinsulin+ area. Insulitis scoring was performed as described previously (Husseini et al. 2015).
Confocal laser scanning microscopy was performed with a Zeiss LSM 800/AxioObserver Z1 system. BIP and proinsulin were labeled using Cy3- or Alexa 488-conjugated secondary antibodies, respectively. Image analysis used surface rendering and 3D reconstruction with seven full Z-stacks acquired at a z-distance of 0.3 μm and surface-surface volume colocalization analysis was performed with IMARIS 9.0.2 (Bitplane Zurich, Switzerland).
Immunoblotting
Immunoblotting was performed using whole pancreas lysate from 8-day BBc (n = 7) and BBdp (n = 11) or 30-day BBc (n = 10) and BBdp (n = 10) rats. Protein extracts were prepared in RIPA buffer supplemented with Halt Protease Inhibitor Cocktail (ThermoFisher Scientific). Total pancreas protein (30–50 µg) was resolved on 4–12% Bis-Tris gels (ThermoFisher Scientific) in MES buffer (ThermoFisher Scientific) and transferred to PVDF membranes (MilliporeSigma). Antibodies to the following proteins were diluted in 5% bovine serum albumin in Tris-buffered saline with 0.1% Tween-20 and incubated overnight at 4°C: BIP (Abcam), insulin (Cell Signaling Technology), LC3B (MilliporeSigma), Sequestosome-1 (SQSTM1) (Bio-Techne, Minneapolis, USA), caspase-1 (SantaCruz Biotechnology), caspase-3 (Cell Signaling Technology) and cleaved caspase-3 (Cell Signaling Technology). Membranes were incubated in the corresponding HRP-conjugated secondary antibodies, visualized using ECL, imaged with a ChemiDoc MP imaging system and quantified with Image Lab (Bio-Rad). Bands of interest were normalized to β-actin-HRP (Cell Signaling Technology) for semi-quantitative analysis of each blot. These values were normalized to the mean of control (BBc) rats and are presented as fold-change.
Statistical analyses
Kaplan–Meier survival curves were analyzed by log-rank test using Prism 4.0 (GraphPad). Gene expression, immunohistochemistry and immunoblotting data were compared using unpaired t-test with Welch’s correction for unequal variance when applicable (GraphPad). Insulitis was analyzed by Kruskal–Wallis one-way ANOVA with Dunn’s multiple comparison post hoc test. Statistical outliers were identified using the Grubbs’ test (GraphPad) and removed from subsequent analyses. Data are expressed as mean ± s.d.
Results
We have previously demonstrated that the gastrointestinal tract plays an important role in the development of type 1 diabetes (Lefebvre et al. 2006, Mojibian et al. 2009, Patrick et al. 2013, Crookshank et al. 2015, Scott et al. 2017). Therefore, we performed prospective PPx in 30 day BBdp rats, approximately 1 week after weaning, which is a major event in the development of gut microbiota and the immune system and represents a time when insulitis is absent. PPx did not impact survival compared with sham-operated animals (Fig. 1A) although diabetes incidence in both groups (PPx, 55% (n = 26); sham, 49% (n = 23)) was lower than the historic norm for our colony which is ~65%. The number of islet- and duct-associated CD68+ and CD8α+ cells was similar in prediabetic (n = 6) and nondiabetic rats (n = 6) (data not shown). However, vascular endothelium-associated CD68+ cells were increased in prediabetic rats (Fig. 1B). Therefore, biopsies were obtained before the onset of classic insulitis and CD68+ macrophage cells are likely the initiating immune cells (Hanenberg et al. 1989).
Microarray study reveals unique pancreas gene signature in confirmed prediabetic rats
Prediabetic and nondiabetic BBdp rats (n = 6/group) mostly formed separate clusters (Fig. 1C); one outlier was removed. Seventy-three genes were differentially expressed in prediabetic rats, 44 downregulated and 29 upregulated (Supplementary Table 1, see section on supplementary data given at the end of this article). The largest fold-change (down 8.4 fold) was observed in Fkbp5, a pro-apoptotic and pro-autophagic co-chaperone of HSP90 (Gassen et al. 2015). Genes involved in metabolic processes were also affected such as Pdk4 (down 4.1-fold), a negative regulator of the pyruvate dehydrogenase complex (Zhang et al. 2014). The transcription factor, Zbtb16, has been reported to be down in PBMC from T1D patients (Jia et al. 2017) and was decreased 3.8-fold in prediabetic rats. The cytochrome P450 gene, Cyp2c22, was downregulated 3.5-fold as was the ER stress gene Atf3 (2.5 fold). Gnpat, which encodes the rate-limiting enzyme in phospholipid biosynthesis, was downregulated 2.2 fold, similar to NOD mice (Lindfors et al. 2009). Numerous genes associated with disease resistance were identified that were involved in islet function: Rgs2, Cyth3, Rph3al, Epha7, protection from oxidative stress: JundD, Dusp16, apoptosis and autophagy: Zbtb16, Ip6k2 Klf6, Atf3. Upregulated genes displayed only modest increases. Hierarchical clustering analysis also showed distinct patterns of gene expression in prediabetic BBdp rats compared with BBdp animals that remained disease-free (Fig. 1D). Multiple GO terms were enriched with most related to apoptosis/ER stress (33.3%) and metabolism (29.2%) (Fig. 2A and B). Several genes, including Foxo3, Pdk4, Pik3r1, Txnip, Ern1 and Aft3 were implicated in multiple GO processes (Fig. 2A).
Bioinformatic analyses of prediabetic pancreas microarrays. (A) Genes that demonstrated a 1.5-fold-change or greater and P ≤ 0.1 were analyzed using the GOrilla tool with the P-value threshold set to 10−3. Multiple gene processes were enriched in the list of differentially expressed genes. (B) Enriched GO terms were related mainly to apoptosis/ER stress (33.3%) and metabolism (29.2%). Enrichment in development- and lipid phosphorylation-associated terms was also observed (12.5 and 8.3%). (C) Candidates were analyzed by STRING using medium confidence for the minimum required interaction score. The resulting network had 23 edges (P < 0.0001), demonstrating interactions between the candidates at the protein level. Twelve proteins related to apoptosis and ER stress were altered in prediabetic rats. A full colour version of this figure is available at https://doi.org/10.1530/JOE-18-0328.
Citation: Journal of Endocrinology 239, 2; 10.1530/JOE-18-0328
To search for known and predicted protein–protein interactions, the 73 genes were analyzed using STRING (Fig. 2C) with rat and human databases (Szklarczyk et al. 2017), at medium confidence level (0.35) for the minimum required interaction score. Using the human database, we identified a network of 21 proteins with 23 edges; 15 proteins formed a common network. This was higher than the expected value of 7 (P < 0.0001), indicating a biological connection among these proteins. Interestingly, 12 proteins in this common network are associated with apoptosis and ER stress and 11 of the 12 proteins were downregulated in prediabetic rats. Using the rat database revealed 24 edges compared with the expected value of 8 (P < 0.0001). These data are consistent with the pathways identified by GOrilla. Overall, this analysis further suggested that ER stress, apoptosis and metabolism genes were differentially expressed in prediabetic animals.
Different age-related gene signature in diabetes-prone BBdp compared to control BBc rats
ER stress, apoptosis and metabolism were identified as pathways of interest in both the GOrilla and STRING analyses of microarray data from prospective whole pancreas biopsies from nondiabetic and prediabetic BBdp rats. Based on bioinformatic analyses and a review of the literature, we selected genes involved in these pathways for further investigation. To further clarify whether these processes are linked to diabetes, we examined strain differences between BBdp and control BBc rats, which do not develop spontaneous diabetes despite having originated from the same Wistar colony (Lefebvre et al. 2006). We obtained tissues from 8-day neonates, a time associated with pancreas remodeling (Trudeau et al. 2000), recently weaned 30-day rats and from 60-day rats to examine changes in gene expression related to the onset of insulitis. All animals were normoglycemic, however, 60-day BBdp rats had increased insulitis compared with age-matched controls (Supplementary Fig. 1A).
Pancreatic transcriptional signature in 8-, 30- and 60-day BBc and BBdp rats
There were age-dependent, strain differences in ER stress, apoptosis, metabolism and immune genes. In diabetes-prone BBdp neonates (Fig. 3A), ER stress genes were increased including the molecular chaperone, Hspa5 (BIP), as well as Eif2ak3 (PERK) and Atf6, which control two of the three UPR branches (Brozzi & Eizirik 2016, Clark & Urano 2016); Ern1 (IRE1α), controller of the third branch was not different. Downstream effector genes, Edem1, Xbp1s and Ddit3 (CHOP) were also increased. Unresolved ER stress can result in the activation of apoptosis. Apoptosis-related genes such as Casp3, Plin2, a modulator of lipid homeostasis (Chen et al. 2017), Foxo3 (Morris et al. 2015) and Txnip, which regulates glucose uptake (Waldhart et al. 2017) and activates the inflammasome (Oslowski et al. 2012), were also increased in neonatal BBdp rats. In contrast, expression of anti-apoptotic Bcl2 was increased in 8-day BBdp pancreas, while expression of pro-apoptotic Bax was not changed. Protective elements may also be engaged in neonatal pancreas as reflected by the decreased Bax:Bcl2 ratio (Fig. 3E).
ER stress and apoptosis-related genes in the pancreas of prediabetic rats. Expression of target genes was compared in pancreas samples from BBc (open circle) and BBdp rats (red filled circles) at (A) 8 (B) 30 and (C) 60 days of age (n = 11–15 animals per group). (D) Increased (orange) or decreased (blue) pancreatic gene expression of ER stress and apoptosis associated genes, relative to age-matched BBc rats. (E) Bax:Bcl2 ratio was calculated using 2−ΔCt values; individual animals were normalized to the average of age-matched BBc rats. BBc and BBdp Bax:Bcl2 ratios were compared at each time point but are shown on the same x-axis for simplicity. Data were analyzed by unpaired t-test with Welch’s correction for unequal variance when applicable (GraphPad) and are expressed as mean ± s.d. A full colour version of this figure is available at https://doi.org/10.1530/JOE-18-0328.
Citation: Journal of Endocrinology 239, 2; 10.1530/JOE-18-0328
Although Hspa5 was decreased in 30-day BBdp rats, there were no additional strain-related differences in ER stress genes (Fig. 3B). Expression of the pro-apoptotic genes Txnip, Hrk and Casp9 was decreased in diabetes-prone rats but Foxo3 and Casp3 did not differ. However, expression of Bcl2 was also lower, resulting in a pro-apoptotic Bax:Bcl2 ratio (Fig. 3E). Hspa5 remained down at 60 days (Fig. 3C). Interestingly, Ern1, Eif2ak3 and Atf6 were also decreased at 60 days, as was expression of Edem1 and Xpb1s while Ddit3 was unchanged. This was consistent with our finding of reduced ER stress capacity in prediabetic rats (Fig. 2) (Bellmann et al. 1997, Herbert & Laybutt 2016). Downregulation of Plin2 is associated with decreased ER stress and prevention of β-cell apoptosis (Chen et al. 2017), and it was increased in 60-day BBdp rats. Gene signatures are summarized in panel D of Figs 3, 4, 5 and 6.
Genes related to metabolism and immunity in the pancreas of prediabetic rats. Expression of target genes was compared in pancreas samples from BBc (open circle) and BBdp rats (red filled circles) at (A) 8 (B) 30 and (C) 60 days of age (n = 11–15 animals per group). (D) Increased (orange) or decreased (blue) pancreatic gene expression of metabolism and immunity-associated genes, relative to age-matched BBc rats. (E) Ins1:Gcg ratio was calculated using 2−ΔCt values; individual animals were normalized to the average of age-matched BBc rats. BBc and BBdp Ins1:Gcg ratios were compared at each time point but are shown on the same x-axis for simplicity. Data were analyzed by unpaired t-test with Welch’s correction for unequal variance when applicable (GraphPad) and are expressed as mean ± s.d. A full colour version of this figure is available at https://doi.org/10.1530/JOE-18-0328.
Citation: Journal of Endocrinology 239, 2; 10.1530/JOE-18-0328
ER stress and apoptosis-related genes in the liver of prediabetic rats. Expression of target genes was compared in liver samples from BBc (open circle) and BBdp rats (red filled circles) at (A) 8 (B) 30 and (C) 60 days of age (n = 11–15 animals per group). (D) Increased (orange) or decreased (blue) hepatic gene expression of ER stress and apoptosis associated genes, relative to age-matched BBc rats. (E) Bax:Bcl2 ratio was calculated using 2−ΔCt values; values were normalized to the average of age-matched BBc rats. BBc and BBdp Bax:Bcl2 ratios were compared at each time point but are shown on the same x-axis for simplicity. Data were analyzed by unpaired t-test with Welch’s correction for unequal variance when applicable (GraphPad) and are expressed as mean ± s.d. A full colour version of this figure is available at https://doi.org/10.1530/JOE-18-0328.
Citation: Journal of Endocrinology 239, 2; 10.1530/JOE-18-0328
Genes related to metabolism and immunity in the liver of prediabetic rats. Expression of target genes compared in liver samples from BBc (open circle) and BBdp rats (red filled circles) at (A) 8 (B) 30 and (C) 60 days of age (n = 11–15 animals per group). (D) Increased (orange) or decreased (blue) hepatic gene expression of metabolism and immunity associated genes, relative to age-matched BBc rats. (E) Serpinb1:Plin2 ratio was calculated using 2−ΔCt values; individual animals were normalized to the average of age-matched BBc rats. BBc and BBdp Serpinb1:Plin2 ratios were compared at each time point but are shown on the same x-axis for simplicity. Data were analyzed by unpaired t-test with Welch’s correction for unequal variance when applicable (GraphPad) and are expressed as mean ± s.d. A full colour version of this figure is available at https://doi.org/10.1530/JOE-18-0328.
Citation: Journal of Endocrinology 239, 2; 10.1530/JOE-18-0328
We also investigated the expression of genes related to metabolic and immune functions. Ins1 was decreased in 8-day BBdp rats, while expression of Pcsk2, a proprotein convertase that participates in both insulin and glucagon production, was upregulated (Jonsson et al. 2012) (Fig. 4A); Pcsk1 was unchanged. Both Ins1 and Ins2 were decreased in diabetes-prone animals at 30 days, while Pcsk2 remained elevated. Expression of Gcg, which encodes glucagon, was increased, resulting in a decreased Ins1:Gcg ratio in BBdp rats (Fig. 4E). Because glucoregulation could be impaired, we investigated expression of glucose transporters. Slc2a1 was increased, while Slc2a2 was decreased in 8-day BBdp rats. Expression of Slc2a4 was decreased in both 30- (Fig. 4B) and 60-day (Fig. 4C) BBdp rats. In neonates, Pik3r1 was increased, as was Gnpat. Tfeb, required for the formation of lysosomes (Nabar & Kehrl 2017) and promotion of autophagy was down at 8 and 60 days. There were no strain-related differences in the inflammasome gene, Nlrp3, or the pro-inflammatory cytokines Il1b or Tnf at 8 days. Increased Il1b and Tnf was first detected in 30-day BBdp rats, long before insulitis (Fig. 4B). Mtor was decreased at all ages (Fig. 4A, B and C). By 60 days, Nlrp3 and Tnf were increased as was Il1b in a subset of animals. By 60 days, expression of Ins1, Ins2, Gcg and Pcsk2 returned to control levels (Fig. 4C), whereas Gnpat and Tfeb were suppressed. However, the Ins1:Gcg ratio remained lower in BBdp rats at both 30 and 60 days (Fig. 4E) suggesting only partial success in normalizing metabolism.
Hepatic transcriptional signature in 8-, 30- and 60-day BBc and BBdp rats
Gene expression in prediabetic liver in T1D has not been reported. Because several of our candidate genes were implicated in insulin resistance and T2D, we investigated genes in liver of 8-, 30- and 60-day BBc and BBdp rats. As in pancreas (Fig. 3), ER stress genes were increased in neonatal BBdp liver (Fig. 5A). We also observed increased Txnip, Hrk, Bax, Casp3 and Foxo3, suggesting that the increase in UPR genes was associated with increased expression of pro-apoptotic genes. Hspa5, Ern1, Eif2ak3 and Atf6 were downregulated in 30-day BBdp rat liver but not in pancreas (Figs 3B and 5B). The downstream effector, Edem1, was similarly downregulated, but Plin2 and Ddit3 were not different. Txnip, Hrk, Foxo3 and Casp9 were down in 30-day BBdp rats (Fig. 5B). However, Bcl2 was also decreased in 30-day BBdp, resulting in a pro-apoptotic Bax:Bcl2 ratio (Fig. 5E). While Hspa5 and Plin2 were increased in 60-day BBdp rats (Fig. 5C), there was no difference in Eif2ak3, Atf6 or Ddit3, and both Ern1 and Edem1 were decreased. Txnip, Hrk, Casp3 and Casp9 were also down at 60 days.
We also investigated metabolic and immune-related genes in the liver. Mtor was suppressed in 30- and 60-day BBdp rats (Fig. 6B and C). Slc2a2 was unchanged and Slc2a4 was down at all ages in BBdp liver (Fig. 6) whereas Slc2a1 was increased in neonates and decreased at 60 days (Fig. 6C). Pik3r1 was increased at all ages (Fig. 6). Gnpat was up in 8-day BBdp, but decreased by 30 and 60 days (Fig. 6). Fkbp5 was elevated at 8 and 60 days (Fig. 6A and C). G0s2, which modulates apoptosis and proliferation and negatively regulates triglyceride catabolism, was increased in 8-day rats only (Heckmann et al. 2013) (Fig. 6A). Serpinb1, a hepatic factor that stimulates β-cell proliferation (El Ouaamari et al. 2016), was decreased in 8- and 30-day BBdp rats (Fig. 6A and B). Plin2 was up at 8 and 60 days. The ratio of Serpinb1 and Plin2, genes coding for hepatic secreted proteins that influence β-cells (El Ouaamari et al. 2016, Chen et al. 2017), was low in neonates and 30-day BBdp rats suggesting a β-cell inhibitory microenvironment (Fig. 6E). Tfeb was decreased in both 30- and 60-day BBdp rats (Fig. 6B and C). Nlrp3 was only increased in 30-day BBdp animals (Fig. 6B). Although hepatic expression of Il1b was decreased in neonatal BBdp rats, expression of Tnf was elevated (Fig. 6A). However, we observed increased Il1b at 30- and 60-days, while Tnf was normalized (Fig. 6B and C).
Metabolic and apoptotic proteins in pancreas
We observed dysregulated metabolic gene expression in very young BBdp rats (Fig. 4A and B) and therefore, we investigated levels of insulin, proinsulin and glucagon. Insulin+ area at 30 or 60 days was not different between BBc and BBdp rats (Fig. 7A, B and C). However, proinsulin area was decreased in 30-day BBdp rats, resulting in a smaller ratio of proinsulin:insulin (Fig. 7A and B). Concomitantly, glucagon+ area was larger in BBdp islets (Fig. 7B and C). Although colocalization of BIP and proinsulin by confocal microscopy was not different at 30 days, it was increased by 60 days (Fig. 8A). This suggested that lower levels of proinsulin can be managed by BIP chaperone in younger animals, however, restoration of proinsulin levels at 60 days (Fig. 7C) could exceed the capacity of BIP and the ER machinery resulting in unresolved ER stress (Szabat et al. 2016).
Metabolic hormones in the pancreas of prediabetic rats. Proinsulin, insulin and glucagon were quantified using IHC in 30 and 60 day BBc and BBdp rats (n = 10–15). (A) Representative images from BBc and BBdp rats. (B) 30 day BBdp rats had decreased proinsulin area, but no difference in insulin area. Glucagon area was increased in BBdp rats. (C) No strain-related differences were observed in insulin or proinsulin area at 60 days. However, glucagon area was increased. Data were analyzed using unpaired t-test with Welch’s correction for unequal variance when applicable (GraphPad) and are expressed as mean ± s.d. A full colour version of this figure is available at https://doi.org/10.1530/JOE-18-0328.
Citation: Journal of Endocrinology 239, 2; 10.1530/JOE-18-0328
Protein expression of identified targets in the pancreas of 30 and 60 day old BBdp rats. (A) Confocal analysis of BBc and BBdp islets showed increased colocalization of BiP (Cy3, pseudo-colored red) and proinsulin (Alexa 488, pseudo-colored green) in 60 day BBdp rats (n = 6, 14 islets) compared with BBc animals (n = 6, 11 islets); no difference in colocalization was observed at 30 days (BBc, n = 3, 6 islets; BBdp, n = 4, 11 islets). Hoechst (pseudo-colored blue) was used to stain nuclei. The insets show magnified views of BiP and proinsulin. Representative western blots from day 8 (B) and day 30 BBc and BBdp rat pancreas (D) Protein bands were detected by chemiluminescence. Band intensity was normalized to β-actin and individual animals were normalized to the average value of age-matched BBc controls. (C) Neonatal BBdp rats had decreased caspase-1, cleaved caspase-3 and insulin protein compared with BBc rats. (E) 30 day old BBdp rats had increased caspase-1 and cleaved caspase-3 but decreased LC3-I and LC3-II protein compared with BBc rats. Data were analyzed using unpaired t-test with Welch’s correction for unequal variance when applicable (GraphPad) and are expressed as mean ± s.d. A full colour version of this figure is available at https://doi.org/10.1530/JOE-18-0328.
Citation: Journal of Endocrinology 239, 2; 10.1530/JOE-18-0328
Western blots of pancreas demonstrated decreased caspase-1 and cleaved caspase 3 protein in BBdp neonates, although there was no difference for caspase-3 (Fig. 8B and C). Neonatal BBdp rats also had lower levels of insulin protein. There was no difference in BIP, or the autophagy markers, LC3-I and LC3-II, in neonates (Fig. 8B and C) nor was the ratio of LC3-I:LC3-II different. By 30 days, LC3-I and LC3-II were down, but the ratio was not different nor were there differences in sequesterome-1 (Fig. 8D and E). Caspase-1 and cleaved caspase-3 were increased whereas total caspase-3 was not different. BIP tended to be less expressed but this only reached statistical significance if the two highest points were removed (P = 0.004) (Fig. 8D and E).
Discussion
The early events in the pancreas which predispose β-cells to autoimmune attack are currently unknown. Most studies examine early changes in pancreatic ER stress using heat shock, chemical inducers or analyze isolated islets which have undergone stress during preparation. Our prospective analysis of pancreatic biopsies in BBdp rats enabled us to identify differences in ER stress, apoptosis and metabolic pathways in prediabetic animals before insulitis and in the absence of stress inducers. These animals displayed different disease outcomes despite being at similar genetic risk and in the same environment as their disease-free littermates. Follow-up studies comparing control BBc vs BBdp rats revealed ER stress/UPR genes were increased in the pancreas and liver of neonates. However, this increase was not sustained in older animals, which also showed signs of dysregulated metabolism and increased pro-inflammatory cytokines. Immune genes were not increased in neonatal pancreas but Tnf was increased in liver at this age.
PERK, IRE1α and ATF6 cooperatively detect errors in protein folding by halting protein translation, promoting degradation of misfolded protein and increasing expression of molecular chaperones that facilitate proper protein folding (Brozzi & Eizirik 2016). Several studies have demonstrated that persistent, unrelieved ER stress leads to β-cell apoptosis, which in turn promotes inflammation and T1D (Eizirik et al. 2013). Islets from individuals with overt T1D had increased C/EBP homologous protein (CHOP, Ddit3) (Marhfour et al. 2012) while CHOP knockdown reduced cytokine-induced apoptosis in primary β-cells and the Ins1 cell line (Allagnat et al. 2012). Inhibiting the activity of IRE1α partially reversed diabetes in NOD mice (Morita et al. 2017). Conversely, there is evidence that the UPR is impaired during diabetes development. Engin et al. demonstrated deficient ATF6 and sXBP1 in two murine models of T1D and in pancreas sections from patients with TID. Additionally, augmenting ATF6 activity using a chemical chaperone reduced diabetes incidence in both murine models (Engin et al. 2013). Loss-of-function mutations in PERK (Eif2ak3) result in diabetes development in both mice and humans (Delepine et al. 2000, Harding et al. 2001). Notably, Eif2ak3 -/- mice had increased IRE1α activity in addition to reduced insulin secretion and hyperglycemia.
Other studies support the concept that ER/UPR response is impaired in BBdp animals (Bellmann et al. 1997, Wachlin et al. 2002, Pino et al. 2009, Ravelli et al. 2013). Heat stressed isolated islets from 45- to 52-day BBDP rats were unable to upregulate HSP70 (Bellmann et al. 1997) and similar results were observed in high glucose challenged islets at a very early age with decreased insulin secretion being a common feature following heat stress (Wachlin et al. 2002). In both these cases, age-related differences were observed under high stress conditions. The present study indicates that diabetes-prone rats upregulate ER stress/UPR protective mechanisms early in life (8 days) but cannot sustain this response. It is rather an impairment of UPR that distinguishes animals that are prone to T1D from those that remain diabetes free (Figs 1D and 2). Thus, our study not only agrees with the previous reports that UPR is impaired in postnatal BBdp rats, but also provides a detailed genetic analysis and reveals increased UPR in day 8 neonates.
In BBdp neonates, we saw increased ER stress genes in pancreas and liver associated with decreased Ins1, insulin protein, Slc2a2 and increased Gcg, Txnip, Pcsk2. The increased ER stress response was not maintained at 30 days despite continued downregulation of Ins1 and Ins2 while Gcg and Pcsk2 expression remained elevated whereas proinsulin+ area was decreased and the ratio of proinsulin:insulin was also down (Fig. 7B). In a previous report, we demonstrated insulin defects in this model. Compared with BBc rats, neonatal and young BBdp rats had lower insulin content per islet (Malaisse et al. 2000). While β-cell proliferation was increased, β-cell growth factors, Pdx1 and Ngfb, were decreased in islets from 30-day BBdp rats (Kauri et al. 2007). Of note, nerve growth factors were shown recently to be essential for glucose-induced insulin secretion (Houtz et al. 2015, Pingitore et al. 2016). Preinsulitic BBDRlyp/lyp rats display impaired insulin secretion in response to glucose challenge (Medina et al. 2017).
Szabat et al. demonstrated that reducing insulin expression lowers ER stress and induces β-cell proliferation (Szabat et al. 2016). Although insulin gene expression was reduced, insulin protein levels were maintained through depletion of proinsulin stores. This agrees with our observations of decreased Ins1 and Ins2 gene expression, decreased proinsulin+ area and increased β-cell proliferation in 30-day BBdp rats (Kauri et al. 2007) in association with reduced ER stress gene expression. A recent pseudotime analysis of gene expression in single β-cells from normoglycemic individuals identified distinct subpopulations, including a population with high insulin gene expression/low UPR gene expression (Inshi/UPRlo) and unexpectedly, a population of Inslo/UPRhi β-cells, which showed higher proliferation and higher antioxidant activity than the other subpopulations (Xin et al. 2018). Thus, they theorized that β-cells in the normal pancreas fluctuate between an active, insulin-producing state and a recovery state. It is possible that development of T1D in the BBdp rat is associated with disruptions among these states, and it would be intriguing to perform similar analyses in this model.
High insulin production in response to fluctuating glucose makes pancreatic islets particularly susceptible to ER stress due to protein misfolding. Unresolved or sustained ER stress can invoke programmed cell death (Pirot et al. 2007, Pino et al. 2009, Brozzi & Eizirik 2016, Herbert & Laybutt 2016, Szabat et al. 2016). Increased ER stress response-related genes in neonatal pancreas were associated with a corresponding increase in pro-apoptotic genes but decreased levels of caspase-1 and cleaved caspase-3 compared with control animals. This is significant as the neonatal period is a time of pancreas remodeling (Trudeau et al. 2000, Wang et al. 2005, Kauri et al. 2007) and suggests that this process may be impaired in BBdp animals. In contrast, pro-apoptotic proteins were increased in 30-day rats in association with increased pro-inflammatory cytokine expression.
BBdp rats had lower levels of CD68+ cells in total pancreas at day 30 (Supplementary Fig. 1B), consistent with diminished capacity to clear apoptotic debris (O’Brien et al. 2002), which could polarize CD68+ cells to M1, pro-inflammatory macrophages. The increase in pro-inflammatory cytokines, Tnf and Il1b, persisted in 60 day BBdp rats, in association with downregulated UPR, apoptosis and autophagy genes and an increase in colocalization of proinsulin and BIP proteins (Fig. 8A). Thus, insulitis could develop as a response to impaired clearance of apoptotic debris (O’Brien et al. 2002), resulting in a pro-inflammatory cell death pathway, known as pyroptosis (Miao et al. 2011). We hypothesize that failure of apoptosis, which is meant to clear damaged cells results in danger-associated molecular patterns which induce inflammation and contribute to autoimmunity. Additionally, induction of ER stress has been demonstrated to upregulate activity of post-translational modifying enzymes including tissue-transglutaminase-2 and peptidyl arginine deiminase (Marre et al. 2018). This was associated with the generation of post-translationally modified epitopes which were recognized by T cell clones from patients with T1D. It would be interesting to investigate whether increased UPR in neonates activates these enzymes, generating neoepitopes that promote early development of autoimmunity.
The actions of insulin are opposed by glucagon, and together these hormones maintain glucose homeostasis by regulating gluconeogenesis in the liver. Destruction of insulin-producing β-cells also causes hyperglucagonemia, resulting in uncontrolled hepatic glucose production. Glucagon receptor ablation has been shown to prevent diabetes, reflecting the importance of glucagon in diabetes pathogenesis (Lee et al. 2011). Surprisingly, expression of Gcg was increased in both neonatal and 30-day BBdp rats and the Ins1:Gcg ratio was decreased at all ages in diabetes-prone animals (Fig. 4). BBdp rats also had increased glucagon+ area at 30 and 60 days (Fig. 7B and C) in keeping with dysregulated glucagon and insulin responses in preinsulitic pancreas (Kanazawa et al. 1988). Thus, altered Gcg expression contributes to the abnormal transcriptional signature in the prediabetic BBdp rat.
A new concept concerning the decline of β-cell mass and function in diabetes is the contribution of inter-organ crosstalk (Shirakawa et al. 2017). The liver is of interest in this respect as it shares embryonic origin with the islets, produces factors that influence regeneration such as serpin B1 and PLIN2 and has major effects on glucose metabolism (El Ouaamari et al. 2016, Chen et al. 2017, Shirakawa et al. 2017). The low ratio of Serpinb1:Plin2 genes in 8- and 30-day BBdp rats would likely favor β-cell death. A previous study demonstrated alterations in lipid metabolism in liver of prediabetic BBDR lyp / lyp rats with decreased insulin levels contributing to altered gene expression (Regnell et al. 2017). We demonstrate here that hepatic genes are also altered in the prediabetic period. We observed a striking reduction of ER stress and metabolism genes in BBdp liver that coincided with increased expression of the inflammasome gene, Nlrp3 and increased pro-inflammatory cytokines, Il1b and Tnf. Of particular note, the UPR genes were downregulated in liver before pancreas (Figs 3 and 5). Alteration in genes involved in the PI3K-AKT signaling pathway, such as Pik3r1 and Mtor and changes in genes associated with glucose uptake suggest that insulin signaling and hepatic functions involving gluconeogenesis, lipid metabolism and fatty acid catabolism could be altered in BBdp rats. The gene signature in liver and pancreas suggests there is metabolic stress and a pro-inflammatory microenvironment in both organs by 30 days. Further studies are needed to determine whether the β-cell destructive process is initiated or promoted by the liver.
In summary, the gene signature in liver and pancreas demonstrates that dysregulated metabolic ER stress precedes a pro-inflammatory microenvironment and the development of insulitis. This temporal imbalance in pancreas and liver proteostasis long before insulitis may reflect very early metabolic dysregulation linked to dysregulated insulin and glucagon that the system attempts to correct by increasing the UPR. This is not sustained by 30 and 60 days at which time a pro-inflammatory milieu is present facilitating the autoimmune attack. Our analysis suggests that there are multiple points of failure in insulin and glucose homeostasis, ER stress response and immune tolerance that contribute to diabetes development. The involvement of the liver in these processes begs further investigation.
Supplementary data
This is linked to the online version of the paper at https://doi.org/10.1530/JOE-18-0328.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
C P was supported by a Queen Elizabeth II Graduate Scholarship in Science and Technology. Research in the laboratory of F W S was supported by the Canadian Institutes of Health Research (grants MOP-130485 and MOP-79531) and Cure Diabetes (Ottawa).
Author contribution statement
J A C designed the studies, performed gene expression analyses, analyzed the data and wrote the article. D S designed the studies, performed analyses, analyzed the data and wrote the article. G S W designed the studies, performed the partial pancreatectomies, monitored the animals, analyzed the data and contributed to writing the manuscript. C P designed the studies, assisted with partial pancreatectomies, monitored the animals, performed analyses, analyzed the data and contributed to writing the manuscript. B M performed analyses and analyzed the data. M-F P performed analyses, analyzed the data and helped write the manuscript. F W S designed and directed the study, analyzed the data, and wrote the article. All authors read and approved the article. F W S 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.
Acknowledgements
The authors thank Garnet Rodger, Pierre Bradley, Catherine Lépine-Bisson, Christina Gilmour and Tami Janveau for excellent care and maintenance of animals and assistance with experiments. They thank Samantha Rojas for technical assistance. They also thank Dr Holly Meredith Orlando, Director of Animal Care and Veterinary Services for support in maintaining our colonies and excellent care. They thank Drs Majambu Mbikay and Nabil Seidah from the Institut de recherches cliniques de Montréal for the gift of antibodies to PCSK1, PCSK2. Prior presentation: Part of this study was presented in abstract form at the 12th Immunology of Diabetes Society Meeting, Victoria, British Columbia, Canada, 15–19 June 2012.
References
Allagnat F, Fukaya M, Nogueira TC, Delaroche D, Welsh N, Marselli L, Marchetti P, Haefliger JA, Eizirik DL & Cardozo AK 2012 c C/EBP homologous protein contributes to cytokine-induced pro-inflammatory responses and apoptosis in β-cells. Cell Death and Differentiation 19 1836–1846. (https://doi.org/10.1038/cdd.2012.67)
Atkinson MA, von Herrath M, Powers AC & Clare-Salzler M 2015 Current concepts on the pathogenesis of type 1 diabetes – considerations for attempts to prevent and reverse the disease. Diabetes Care 38 979–988. (https://doi.org/10.2337/dc15-0144)
Bellmann K, Hui L, Radons J, Burkart V & Kolb H 1997 Low stress response enhances vulnerability of islet cells in diabetes-prone BB rats. Diabetes 46 232–236. (https://doi.org/10.2337/diab.46.2.232)
Brozzi F & Eizirik DL 2016 ER stress and the decline and fall of pancreatic beta cells in type 1 diabetes. Upsala Journal of Medical Sciences 121 133–139. (https://doi.org/10.3109/03009734.2015.1135217)
Campbell-Thompson M, Rodriguez-Calvo T & Battaglia M 2015 Abnormalities of the exocrine pancreas in type 1 diabetes. Current Diabetes Reports 15 79. (https://doi.org/10.1007/s11892-015-0653-y)
Chen E, Tsai TH, Li L, Saha P, Chan L & Chang BH 2017 PLIN2 is a key regulator of the unfolded protein response and endoplasmic reticulum stress resolution in pancreatic beta cells. Scientific Reports 7 40855. (https://doi.org/10.1038/srep40855)
Clark AL & Urano F 2016 Endoplasmic reticulum stress in beta cells and autoimmune diabetes. Current Opinion in Immunology 43 60–66. (https://doi.org/10.1016/j.coi.2016.09.006)
Crevecoeur I, Gudmundsdottir V, Vig S, Marques Camara Sodre F, D’Hertog W, Fierro AC, Van Lommel L, Gysemans C, Marchal K, Waelkens E, et al.2017 Early differences in islets from prediabetic NOD mice: combined microarray and proteomic analysis. Diabetologia 60 475–489. (https://doi.org/10.1007/s00125-016-4191-1)
Crookshank JA, Patrick C, Wang GS, Noel JA & Scott FW 2015 Gut immune deficits in LEW.1AR1-iddm rats partially overcome by feeding a diabetes-protective diet. Immunology 145 417–428. (https://doi.org/10.1111/imm.12457)
Delepine M, Nicolino M, Barrett T, Golamaully M, Lathrop GM & Julier C 2000 EIF2AK3, encoding translation initiation factor 2-alpha kinase 3, is mutated in patients with Wolcott-Rallison syndrome. Nature Genetics 25 406–409. (https://doi.org/10.1038/78085)
Eden E, Navon R, Steinfeld I, Lipson D & Yakhini Z 2009 GOrilla: a tool for discovery and visualization of enriched GO terms in ranked gene lists. BMC Bioinformatics 10 48. (https://doi.org/10.1186/1471-2105-10-48)
Eizirik DL, Miani M & Cardozo AK 2013 Signalling danger: endoplasmic reticulum stress and the unfolded protein response in pancreatic islet inflammation. Diabetologia 56 234–241. (https://doi.org/10.1007/s00125-012-2762-3)
El Ouaamari A, Dirice E, Gedeon N, Hu J, Zhou JY, Shirakawa J, Hou L, Goodman J, Karampelias C, Qiang G, et al.2016 SerpinB1 promotes pancreatic beta cell proliferation. Cell Metabolism 23 194–205. (https://doi.org/10.1016/j.cmet.2015.12.001)
Engin F, Yermalovich A, Nguyen T, Hummasti S, Fu W, Eizirik DL, Mathis D & Hotamisligil GS 2013 Restoration of the unfolded protein response in pancreatic beta cells protects mice against type 1 diabetes. Science Translational Medicine 5 211ra156. (https://doi.org/10.1126/scitranslmed.3006534)
Gassen NC, Hartmann J, Schmidt MV & Rein T 2015 FKBP5/FKBP51 enhances autophagy to synergize with antidepressant action. Autophagy 11 578–580. (https://doi.org/10.1080/15548627.2015.1017224)
Hanenberg H, Kolb-Bachofen V, Kantwerk-Funke G & Kolb H 1989 Macrophage infiltration precedes and is a prerequisite for lymphocytic insulitis in pancreatic islets of pre-diabetic BB rats. Diabetologia 32 126–134. (https://doi.org/10.1007/BF00505185)
Harding HP, Zeng H, Zhang Y, Jungries R, Chung P, Plesken H, Sabatini DD & Ron D 2001 Diabetes mellitus and exocrine pancreatic dysfunction in perk−/− mice reveals a role for translational control in secretory cell survival. Molecular Cell 7 1153–1163. (https://doi.org/10.1016/S1097-2765(01)00264-7)
Heckmann BL, Zhang X, Xie X & Liu J 2013 The G0/G1 switch gene 2 (G0S2): regulating metabolism and beyond. Biochimica et Biophysica Acta 1831 276–281. (https://doi.org/10.1016/j.bbalip.2012.09.016)
Henschel AM, Cabrera SM, Kaldunski ML, Jia S, Geoffrey R, Roethle MF, Lam V, Chen YG, Wang X, Salzman NH, et al.2018 Modulation of the diet and gastrointestinal microbiota normalizes systemic inflammation and beta-cell chemokine expression associated with autoimmune diabetes susceptibility. PLoS ONE 13 e0190351. (https://doi.org/10.1371/journal.pone.0190351)
Herbert TP & Laybutt DR 2016 A reevaluation of the role of the unfolded protein response in islet dysfunction: maladaptation or a failure to adapt? Diabetes 65 1472–1480. (https://doi.org/10.2337/db15-1633)
Houtz J, Borden P, Ceasrine A, Minichiello L & Kuruvilla R 2015 Neurotrophin signaling is required for glucose-induced insulin secretion. Developmental Cell 39 329–345. (https://doi.org/10.1016/j.devcel.2016.10.003)
Husseini M, Wang GS, Patrick C, Crookshank JA, MacFarlane AJ, Noel JA, Strom A & Scott FW 2015 Heme oxygenase-1 induction prevents autoimmune diabetes in association with pancreatic recruitment of M2-like macrophages, mesenchymal cells, and fibrocytes. Endocrinology 156 3937–3949. (https://doi.org/10.1210/en.2015-1304)
Jia X, Yu H, Zhang H, Si Y, Tian D, Zhao X, Luan J & Jia H 2017 Integrated analysis of different microarray studies to identify candidate genes in type 1 diabetes. Journal of Diabetes 9 149–157. (https://doi.org/10.1111/1753-0407.12391)
Jonsson A, Isomaa B, Tuomi T, Eriksson JG, Groop L & Lyssenko V 2012 Effect of a common variant of the PCSK2 gene on reduced insulin secretion. Diabetologia 55 3245–3251. (https://doi.org/10.1007/s00125-012-2728-5)
Kanazawa M, Ikeda J, Sato J, Natoya Y, Ito H, Komeda K, Kawazu S & Kanazawa Y 1988 Alteration of insulin and glucagon secretion from the perfused BB rat pancreas before and after the onset of diabetes. Diabetes Research and Clinical Practice 5 201–204. (https://doi.org/10.1016/S0168-8227(88)80089-5)
Kauri LM, Wang GS, Patrick C, Bareggi M, Hill DJ & Scott FW 2007 Increased islet neogenesis without increased islet mass precedes autoimmune attack in diabetes-prone rats. Laboratory Investigation 87 1240–1251. (https://doi.org/10.1038/labinvest.3700687)
Lee Y, Wang MY, Du XQ, Charron MJ & Unger RH 2011 Glucagon receptor knockout prevents insulin-deficient type 1 diabetes in mice. Diabetes 60 391–397. (https://doi.org/10.2337/db10-0426)
Lefebvre DE, Powell KL, Strom A & Scott FW 2006 Dietary proteins as environmental modifiers of type 1 diabetes mellitus. Annual Review of Nutrition 26 175–202. (https://doi.org/10.1146/annurev.nutr.26.061505.111206)
Lindfors E, Gopalacharyulu PV, Halperin E & Oresic M 2009 Detection of molecular paths associated with insulitis and type 1 diabetes in non-obese diabetic mouse. PLoS ONE 4 e7323. (https://doi.org/10.1371/journal.pone.0007323)
Logothetopoulos J, Valiquette N, Madura E & Cvet D 1984 The onset and progression of pancreatic insulitis in the overt, spontaneously diabetic, young adult BB rat studied by pancreatic biopsy. Diabetes 33 33–36. (https://doi.org/10.2337/diab.33.1.33)
Malaisse WJ, Olivares E, Laghmich A, Ladriere L, Sener A & Scott FW 2000 Feeding a protective hydrolysed casein diet to young diabetic-prone BB rats affects oxidation of L[U-14C]glutamine in islets and Peyer’s patches, reduces abnormally high mitotic activity in mesenteric lymph nodes, enhances islet insulin and tends to normalize NO production. International Journal of Experimental Diabetes Research 1 121–130. (https://doi.org/10.1155/EDR.2000.121)
Marhfour I, Lopez XM, Lefkaditis D, Salmon I, Allagnat F, Richardson SJ, Morgan NG & Eizirik DL 2012 Expression of endoplasmic reticulum stress markers in the islets of patients with type 1 diabetes. Diabetologia 55 2417–2420. (https://doi.org/10.1007/s00125-012-2604-3)
Marre ML, McGinty JW, Chow IT, DeNicola ME, Beck NW, Kent SC, Powers AC, Bottino R, Harlan DM, Greenbaum CJ, et al.2018 Modifying enzymes are elicited by ER stress, generating epitopes that are selectively recognized by CD4(+) T Cells in patients with Type 1 diabetes. Diabetes 67 1356–1368. (https://doi.org/10.2337/db17-1166)
Medina A, Parween S, Ullsten S, Vishnu N, Siu YT, Quach M, Bennet H, Balhuizen A, Akesson L, Wierup N, et al.2017 Early deficits in insulin secretion, beta cell mass and islet blood perfusion precede onset of autoimmune type 1 diabetes in BioBreeding rats. Diabetologia 61 896–905. (https://doi.org/10.1007/s00125-017-4512-z)
Miao EA, Rajan JV & Aderem A 2011 Caspase-1-induced pyroptotic cell death. Immunological Reviews 243 206–214. (https://doi.org/10.1111/j.1600-065X.2011.01044.x)
Mojibian M, Chakir H, Lefebvre DE, Crookshank JA, Sonier B, Keely E & Scott FW 2009 A diabetes-specific HLA-DR restricted pro-inflammatory T cell response to wheat polypeptides in tissue transglutaminase antibody negative patients with type 1 diabetes. Diabetes 58 1789–1796. (https://doi.org/10.2337/db08-1579)
Morita S, Villalta SA, Feldman HC, Register AC, Rosenthal W, Hoffmann-Petersen IT, Mehdizadeh M, Ghosh R, Wang L, Colon-Negron K, et al.2017 Targeting ABL-IRE1alpha signaling spares ER-stressed pancreatic beta cells to reverse autoimmune diabetes. Cell Metabolism 25 1207. (https://doi.org/10.1016/j.cmet.2017.04.026)
Morris BJ, Willcox DC, Donlon TA & Willcox BJ 2015 FOXO3: a major gene for human longevity – a mini-review. Gerontology 61 515–525. (https://doi.org/10.1159/000375235)
Nabar NR & Kehrl JH 2017 The transcription factor EB links cellular stress to the immune response. Yale Journal of Biology and Medicine 90 301–315.
O’Brien BA, Fieldus WE, Field CJ & Finegood DT 2002 Clearance of apoptotic beta-cells is reduced in neonatal autoimmune diabetes-prone rats. Cell Death and Differentiation 9 457–464. (https://doi.org/10.1038/sj/cdd/4400973)
Oslowski CM, Hara T, O’Sullivan-Murphy B, Kanekura K, Lu S, Hara M, Ishigaki S, Zhu LJ, Hayashi E, Hui ST, et al.2012 Thioredoxin-interacting protein mediates ER stress-induced beta cell death through initiation of the inflammasome. Cell Metabolism 16 265–273. (https://doi.org/10.1016/j.cmet.2012.07.005)
Patrick C, Wang GS, Lefebvre DE, Crookshank JA, Sonier B, Eberhard CE, Mojibian M, Kennedy CR, Brooks SP, Kalmokoff ML, et al.2013 Promotion of autoimmune diabetes by cereal diet in the presence or absence of microbes associated with gut immune activation, regulatory imbalance and altered cathelicidin antimicrobial peptide. Diabetes 62 2036–2047. (https://doi.org/10.2337/db12-1243)
Pingitore A, Caroleo MC, Cione E, Castanera Gonzalez R, Huang GC & Persaud SJ 2016 Fine tuning of insulin secretion by release of nerve growth factor from mouse and human islet beta-cells. Molecular and Cellular Endocrinology 436 23–32. (https://doi.org/10.1016/j.mce.2016.07.014)
Pino SC, O’Sullivan-Murphy B, Lidstone EA, Yang C, Lipson KL, Jurczyk A, diIorio P, Brehm MA, Mordes JP, Greiner DL, et al.2009 CHOP mediates endoplasmic reticulum stress-induced apoptosis in Gimap5-deficient T cells. PLoS ONE 4 e5468. (https://doi.org/10.1371/journal.pone.0005468)
Pirot P, Naamane N, Libert F, Magnusson NE, Orntoft TF, Cardozo AK & Eizirik DL 2007 Global profiling of genes modified by endoplasmic reticulum stress in pancreatic beta cells reveals the early degradation of insulin mRNAs. Diabetologia 50 1006–1014. (https://doi.org/10.1007/s00125-007-0609-0)
Pound LD, Patrick C, Eberhard CE, Mottawea W, Wang GS, Abujamel T, Vandenbeek R, Stintzi A & Scott FW 2015 Cathelicidin antimicrobial peptide: a novel regulator of islet function, islet regeneration, and selected gut bacteria. Diabetes 64 4135–4147. (https://doi.org/10.2337/db15-0788)
Ravelli RBG, Kalicharan RD, Avramut MC, Sjollema KA, Pronk JW, Dijk F, Koster AJ, Visser JTJ, Faas FGA & Giepmans BNG 2013 Destruction of tissue, cells and organelles in type 1 diabetic rats presented at macromolecular resolution. Scientific Reports 3 1804. (https://doi.org/10.1038/srep01804)
Regnell SE, Hessner MJ, Jia S, Akesson L, Stenlund H, Moritz T, La Torre D & Lernmark A 2017 Longitudinal analysis of hepatic transcriptome and serum metabolome demonstrates altered lipid metabolism following the onset of hyperglycemia in spontaneously diabetic biobreeding rats. PLoS ONE 12 e0171372. (https://doi.org/10.1371/journal.pone.0171372)
Scott FW, Pound LD, Patrick C, Eberhard CE & Crookshank JA 2017 Where genes meet environment-integrating the role of gut luminal contents, immunity and pancreas in type 1 diabetes. Translational Research 179 183–198. (https://doi.org/10.1016/j.trsl.2016.09.001)
Shirakawa J, De Jesus DF & Kulkarni RN 2017 Exploring inter-organ crosstalk to uncover mechanisms that regulate beta-cell function and mass. European Journal of Clinical Nutrition 71 896–903. (https://doi.org/10.1038/ejcn.2017.13)
Szabat M, Page MM, Panzhinskiy E, Skovso S, Mojibian M, Fernandez-Tajes J, Bruin JE, Bround MJ, Lee JT, Xu EE, et al.2016 Reduced insulin production relieves endoplasmic reticulum stress and induces beta cell proliferation. Cell Metabolism 23 179–193. (https://doi.org/10.1016/j.cmet.2015.10.016)
Szklarczyk D, Morris JH, Cook H, Kuhn M, Wyder S, Simonovic M, Santos A, Doncheva NT, Roth A, Bork P, et al.2017 The STRING database in 2017: quality-controlled protein-protein association networks, made broadly accessible. Nucleic Acids Research 45 D362–D368. (https://doi.org/10.1093/nar/gkw937)
Trudeau JD, Dutz JP, Arany E, Hill DJ, Fieldus WE & Finegood DT 2000 Neonatal beta-cell apoptosis: a trigger for autoimmune diabetes? Diabetes 49 1–7. (https://doi.org/10.2337/diabetes.49.1.1)
Wachlin G, Heine L, Kloting I, Dunger A, Hahn HJ & Schmidt S 2002 Stress response of pancreatic islets from diabetes prone BB rats of different age. Autoimmunity 35 389–395. (https://doi.org/10.1080/0891693021000014989)
Waldhart AN, Dykstra H, Peck AS, Boguslawski EA, Madaj ZB, Wen J, Veldkamp K, Hollowell M, Zheng B, Cantley LC, et al.2017 Phosphorylation of TXNIP by AKT mediates acute influx of glucose in response to insulin. Cell Reports 19 2005–2013. (https://doi.org/10.1016/j.celrep.2017.05.041)
Wang GS, Rosenberg L & Scott FW 2005 Tubular complexes as a source for islet neogenesis in the pancreas of diabetes-prone BB rats. Laboratory Investigation 85 675–688. (https://doi.org/10.1038/labinvest.3700259)
Xin Y, Gutierrez GD, Okamoto H, Kim J, Lee AH, Adler C, Ni M, Yancopoulos GD, Murphy AJ & Gromada J 2018. Pseudotime ordering of single human beta-cells reveals states of insulin production and unfolded protein response. Diabetes [epub]. (https://doi.org/10.2337/db18-0365)
Zhang S, Hulver MW, McMillan RP, Cline MA & Gilbert ER 2014 The pivotal role of pyruvate dehydrogenase kinases in metabolic flexibility. Nutrition and Metabolism 11 10. (https://doi.org/10.1186/1743-7075-11-10)
