Abstract
Aims/hypothesis
Type 2 diabetes is characterised by islet amyloid and toxic oligomers of islet amyloid polypeptide (IAPP). We posed the questions, (1) does IAPP toxicity induce an islet response comparable to that in humans with type 2 diabetes, and if so, (2) what are the key transcriptional drivers of this response?
Methods
The islet transcriptome was evaluated in five groups of mice: beta cell specific transgenic for (1) human IAPP, (2) rodent IAPP, (3) human calpastatin, (4) human calpastatin and human IAPP, and (5) wild-type mice. RNA sequencing data was analysed by differential expression analysis and gene co-expression network analysis to establish the islet response to adaptation to an increased beta cell workload of soluble rodent IAPP, the islet response to increased expression of oligomeric human IAPP, and the extent to which the latter was rescued by suppression of calpain hyperactivation by calpastatin. Rank-rank hypergeometric overlap analysis was used to compare the transcriptome of islets from human or rodent IAPP transgenic mice vs humans with prediabetes or type 2 diabetes.
Results
The islet transcriptomes in humans with prediabetes and type 2 diabetes are remarkably similar. Beta cell overexpression of soluble rodent or oligomer-prone human IAPP induced changes in islet transcriptome present in prediabetes and type 2 diabetes, including decreased expression of genes that confer beta cell identity. Increased expression of human IAPP, but not rodent IAPP, induced islet inflammation present in prediabetes and type 2 diabetes in humans. Key mediators of the injury responses in islets transgenic for human IAPP or those from individuals with type 2 diabetes include STAT3, NF-κB, ESR1 and CTNNB1 by transcription factor analysis and COL3A1, NID1 and ZNF800 by gene regulatory network analysis.
Conclusions/interpretation
Beta cell injury mediated by IAPP is a plausible mechanism to contribute to islet inflammation and dedifferentiation in type 2 diabetes. Inhibition of IAPP toxicity is a potential therapeutic target in type 2 diabetes.
Graphical abstract
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
The islet in type 2 diabetes is characterised by islet amyloid derived from islet amyloid polypeptide (IAPP), a protein co-expressed with insulin by beta cells that when misfolded and in aggregate form may contribute to beta cell failure [1,2,3,4]. Human IAPP (hIAPP) toxicity is most potently mediated by small intracellular membrane permeant oligomers [5]. Species with amyloidogenic IAPP, such as humans, non-human primates and cats, share vulnerability to type 2 diabetes, while those with non-amyloidogenic IAPP, such as mice and rats, do not [6]. While numerous hypotheses have been put forward to explain the wide-ranging changes in islets in type 2 diabetes [7], there is a consensus that misfolded protein stress induced by toxic oligomers of amyloidogenic proteins initiate these changes in neurodegenerative diseases.
Given the known proximal role of misfolded protein stress in neurodegenerative diseases, and the connection of the risk factors for type 2 diabetes to misfolded protein stress, we hypothesised that hIAPP misfolded protein stress may be a proximal cause of the wide-ranging changes in islets in individuals with type 2 diabetes. Risk factors for type 2 diabetes include insulin resistance [8] and low birthweight [9]. Low birthweight may lead to low adult beta cell mass [10], which together with insulin resistance, predicts a high insulin and IAPP expression rate per beta cell [11]. Beta cell misfolded protein stress is induced when expression of hIAPP per cell exceeds the cellular threshold for clearing misfolded proteins [12]. This threshold declines with ageing [13], a risk factor for both type 2 diabetes and neurodegenerative diseases. hIAPP overexpression in isolated mouse islets in vitro can modify islet gene expression with relevance to type 2 diabetes [14].
In the present study we evaluated the islet transcriptome from a mouse model of beta cell hIAPP toxicity [12] before diabetes onset in order to avoid the confounding effects of hyperglycaemia. To control for the increased burden of IAPP expression, we evaluated the transcriptome from mice overexpressing rodent IAPP (rIAPP) [15]. We then compared the changes in the transcriptome of hIAPP or rIAPP islets to those in humans with prediabetes (defined as an HbA1c level of 42 to 47 mmol/mol [6% to 6.5%]) or type 2 diabetes to establish if the changes in the islet in type 2 diabetes are potentially attributable in part to hIAPP protein misfolding stress.
Methods
Mouse models
Animal studies were approved by the University of California, Los Angeles (UCLA) Office of Animal Research Oversight. The transgenic mice homozygous for human IAPP (hIAPP) [12] were originally from Pfizer (available from Jackson Laboratory, Bar Harbor, ME, USA: IMSR cat. no. JAX:008232, RRID:IMSR_JAX:008232) and wild-type FVB (WT) mice (IMSR cat. No. CRL:207, RRID:IMSR_CRL:207) from Charles Rivers Laboratory (Wilmington, MA, USA). The generation of the transgenic mice expressing rodent Iapp (rIAPP), human calpastatin (hCAST), and both human IAPP and CAST (hIAPP:hCAST) on FVB background has been described elsewhere [15, 16]. Mice were bred and maintained at UCLA on 12 h day/night rhythm, Harlan Teklad Rodent Diet 8604 (Placentia, CA, USA) and water ad libitum; diabetes was monitored as described [16]. hIAPP transgenic mice develop diabetes (fasting blood glucose >6.9 mmol/l) after 9 weeks of age, while rIAPP mice remained non-diabetic until 18 weeks of age, the end of observation. See electronic supplementary material (ESM) Fig. 1. Only non-diabetic 9–10-weeks-old male mice were used (ESM Tables 1–4). Expression of IAPP (sum of endogenous and transgenic) is comparable in the rIAPP and hIAPP mice [15]. Mice were either subjected to metabolic studies with fasting blood glucose measurements and GTT, or islets and pancreases were collected for analysis of RNA by bulk islet RNA sequencing (RNA-seq) or qPCR, or analysis of protein levels by western blotting (whole cell lysate in RIPA buffer) or immunostaining (4 μm thick sections of frozen in OCT 4% paraformaldehyde fixed tissue) [16, 17]. See ESM Methods.
RNA-seq of mouse islets
RNA samples from three mice per group were used for RNA-seq (ESM Table 1). Total RNA was extracted from islet samples (176 ± 11 islets per mouse) using the RNeasy Mini Kit (Qiagen, Germantown, MD, USA). RNA integrity was confirmed using the Agilent Bioanalyzer 2100 (RNA integrity number [RIN] range: 6.8–8.9). RNA-seq libraries were prepared using the TruSeq with Ribo-Zero treatment (Illumina, San Diego, CA, USA) to deplete ribosomal RNA. cDNA libraries were generated using the NuGEN Ovation kit (NuGEN, Redwood City, CA, USA). Illumina’s NextSeq 500 platform was used to generate 75 bp, paired-end reads (64 ± 1.4 million reads per sample). Short reads were aligned to the mouse reference genome build GRCm38 (mm10) using the Spliced Transcripts Alignment to a Reference software (STAR aligner) [18]. Between 65% and 75% (mean 70%) of the reads mapped uniquely to the mouse genome. The HT-Seq package [19] was used to count the number of fragments aligned to known exonic regions. Gene expression was measured as total fragment counts per gene. Sample clustering of islet RNA-seq using multidimensional scaling largely reflected genotype (ESM Fig. 2).
RNA-seq of human islets
RNA-seq data from human pancreatic islets were downloaded from the Gene Expression Omnibus (GEO) (GSE50244) [20]. Data from 77 samples with available HbA1c values were analysed. Read counts were normalised via the trimmed mean method prior to differential expression analysis using the edgeR package. One type 2 diabetes sample was excluded as an outlier (GSM1216834); therefore 76 samples were included in this manuscript: 51 from normoglycaemic donors with HbA1c levels below 42 mmol/mol (<6%) (HbA1c 5.4 ± 0.1%; BMI 26 ± 0.3; age 56 ± 2; 18 female/33 male), 15 prediabetic donors with HbA1c levels 42 to 47 mmol/mol (6% to 6.5%) (HbA1c 6.1 ± 0.03%; BMI 26 ± 1; age 61 ± 2; 6 female/9 male), and ten donors with type 2 diabetes with HbA1c levels 48 mmol/mol (>6.5%) or higher (HbA1c 7.5 ± 0.3%; BMI 30 ± 1; age 61 ± 3; 6 female/4 male). No additional information about the donors or islet morphology was available.
RNA-seq data analysis
The RNA-seq data from mouse and human islets were subjected to differential expression analysis [21] to find differentially expressed genes (DEGs) between mouse groups and between human groups. Rank-rank hypergeometric overlap (RRHO) [22] analysis was then used to compare human and mouse gene expression signatures to evaluate between-species similarity and differences (ESM Methods). In order to understand the biological processes informed by the DEGs, we performed functional enrichment analysis (ESM Methods) to identify over-representation of Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways [23]. DEGs from mouse islets were further assessed for enrichment of genes associated with Mendelian or common form of diabetes and metabolic syndrome (ESM Methods).
To identify genes with coordinated expression in the form of co-expression modules that are related to diabetes development in mouse models, we utilised weighted gene co-expression network analysis (WGCNA; ESM Methods) [24, 25].
To identify cell types contributing to the gene expression changes, we subjected DEGs and co-expression modules (restricting analysis to genes with module membership >0.5 and false discovery rate [FDR] <0.05) to two types of analysis. First, we carried out cell type marker enrichment analysis (ESM Methods) using PanglaoDB cell marker compendium [26] and GeneOverlap R tool [27]. Second, we applied CibersortX [28] to deconvolute mouse bulk islet RNA-seq data into cell type proportions using single cell RNA-seq mouse islet data from GEO (GSM2230762) as reference (ESM Methods).
To identify shared regulatory factors between islets from hIAPP transgenic mice and islets from humans with type 2 diabetes, we performed transcription factor (TF) network analysis using the Enrichr tool [29] with the TF-Gene Co-occurrence extension. To identify additional non-TF regulators, we utilised the key driver analysis function from the Mergeomics R package [30] to identify regulatory genes for the DEG sets and for the co-expression modules using gene regulatory Bayesian network based on a χ2 like statistic (ESM Methods).
Statistical analysis
Statistical analysis was performed as described in the figure legends and ESM.
Results
Similarity in islet transcriptome in prediabetes or type 2 diabetes and IAPP overexpressing mice
There is striking concordance of the islet transcriptome between individuals with prediabetes and those with type 2 diabetes (Fig. 1a). Since ~80% of individuals with prediabetes do not develop diabetes [22], this finding implies that a high proportion of the changes in islets in type 2 diabetes are adaptive rather than causal of diabetes. There was also a close concordance of changes in islet gene expression in type 2 diabetes and mouse islets with increased hIAPP or rIAPP expression (Fig. 1b–d, ESM Fig. 2).
Since by design neither the hIAPP or rIAPP mice had diabetes when islets were sampled, we further compared the changes in islet transcriptome in the mouse islets with those in human islets with prediabetes. There was again close concordance in transcriptome in both rIAPP and hIAPP islets with those from individuals with prediabetes (Fig. 1e,f). These findings imply that a more detailed analysis of the transcriptome in response to increased expression of rIAPP and hIAPP might shed light on adaptive vs disease causal changes in type 2 diabetes.
Islet transcriptome in response to increased beta cell rIAPP or hIAPP expression
To investigate beta cell adaptation to an increased workload of non-amyloidogenic rIAPP expression, we compared the transcriptome of islets from rIAPP vs WT mice. Gene expression was increased in WGCNA modules M8 (Mitogen activated protein kinase [MAPK] signalling), M9 (TGF-β signalling), M13 (cell migration) and decreased in M7 (cell cycle) in rIAPP islets (Fig. 2, Table 1, ESM Fig. 3). Differential expression analysis identified 2731 DEGs (1306 up- and 1425 downregulated in rIAPP islets; FDR < 0.05) (Fig. 3a). Of the 2731 DEGs, 14 are implicated in monogenetic diabetes and 39 are located in genomic regions associated with type 2 diabetes by genome-wide association studies (GWAS), consistent with the overlap between rIAPP and type 2 diabetes by RRHO analysis (Fig. 1c, ESM Fig. 4). Prominently upregulated genes in rIAPP islets include those required for protein synthesis and those that adapt cells to an increased burden of protein folding and quality control, collectively referred to as the unfolded protein response (UPR) (Fig. 3c, ESM Fig. 5, 6).
Since there is a considerable overlap in the changes in transcriptome in hIAPP and rIAPP mice (Fig. 1d), and hIAPP but not rIAPP mice develop diabetes, we next compared these transcriptomes to discern changes related to oligomer toxicity vs adaptation to an increased burden of IAPP expression. Five co-expression modules were differentially expressed between hIAPP and rIAPP islets: M1 (inflammation), M4 (oxidative stress) and M6 (cell cycle; cell signalling) expression was increased, while M5 (RNA processing) and M8 (MAPK signalling; cell adhesion) were downregulated in hIAPP (Fig. 2b,c, Table 1). This pattern of changes is consistent with islet inflammation reported in type 2 diabetes [31]. We identified 2011 DEGs between hIAPP and rIAPP islets, with 1031 upregulated and 980 downregulated in hIAPP islets (Fig. 3b). Among these, 13 DEGs have been implicated in monogenetic diabetes and 194 are located in genomic regions associated with type 2 diabetes by GWAS (ESM Fig. 4). A number of interesting trends were found in genes involved in critical pathways associated with type 2 diabetes development (Fig. 3, ESM Fig. 4).
UPR
Consistent with the comparable increase in IAPP expression [15], islets in hIAPP mice share a comparable increase in the UPR to rIAPP mice (Fig. 3c). The evaluation of changes in expression of the same genes in humans with type 2 diabetes or prediabetes compared with non-diabetics show consistently upregulated BHLHA15 (MIST1), a potent endoplasmic reticulum (ER) stress-inducible transcriptional regulator upregulated by beta cell Ca2+ overload [32].
Inflammation
The pronounced signal for inflammation in hIAPP, but not rIAPP islets is also present in human prediabetes and type 2 diabetes (Fig. 3). Consistent with a proinflammatory state, several key cell surface antigens (Cd4, Cd68, Cd84), immune sensors (Cx3cr1, Clec7a), and macrophage genes (Axl, Slc11a1) were upregulated in hIAPP islets and in type 2 diabetes. The top hub gene in the M1 module that is most highly activated in hIAPP islets is Cx3cr1 [33]. This prominence of islet inflammation in hIAPP as compared with rIAPP is further shown by the lack of difference between rIAPP and WT in module M1 (immune response) (Fig. 2b,c).
Cell cycle
Activation of cell replication is a key component of injury repair programmes. Cell cycle related genes are increased in hIAPP islets but decreased in rIAPP islets (Fig. 3), likely reflecting the injury-mediated signalling in hIAPP islets vs adaptive UPR in rIAPP islets [34]. Prominent amongst upregulated genes in hIAPP islets are those that enhance cell replication directly (Cdk1, Cep55) or indirectly (Hmmr, Anln). In contrast, these genes are downregulated in islets from prediabetes and type 2 diabetes, which is perhaps consistent with epigenetic silencing of cell cycle genes in beta cells in adult humans compared with 9-week-old mice [35].
Beta cell dedifferentiation
Genes important for maintaining beta cell differentiation were decreased in islets of both rIAPP and hIAPP mice compared with islets from WT mice (Fig. 3), a pattern reproduced in islets from humans with prediabetes and type 2 diabetes. In mice, beta cell dedifferentiation was confirmed by mRNA and protein analysis (Fig. 4a–c). To establish if this partial dedifferentiation impacts glucose tolerance, we performed IPGTTs. As expected, mice transgenic for hIAPP were glucose intolerant compared with WT mice. Although to a lesser extent, mice transgenic for rIAPP were also glucose intolerant, consistent with the observed partial beta cell dedifferentiation and known action of IAPP to inhibit insulin secretion [36] (Fig. 4d).
Contribution of calpain hyperactivation to beta cell hIAPP toxicity
Calpain hyperactivation has been widely reported as a mediator of amyloidogenic protein induced cytotoxicity, presumably as a consequence of aberrant Ca2+ signalling that results from nonselective ion channel activity of toxic oligomers [37]. Concurrent beta cell-specific overexpression of human calpastatin (CAST) (hIAPP:hCAST), which inhibits calpain, delays or prevents diabetes in hIAPP transgenic mice [16].
Functional enrichment analysis revealed the sets of genes partially rescued by CAST overexpression in hIAPP mice were those that mediate UPR and inflammation (Fig. 5). Sustained calpain hyperactivation may activate proinflammatory signalling pathways mediated by NF-κB and endothelial nitric oxide synthase (eNOS) [38].
To further evaluate the role of calpain activation in type 2 diabetes and prediabetes, we correlated the differential expression patterns and found that the type 2 diabetes islet profile is best reflected by the hIAPP islet profile, outperforming the rIAPP and hIAPP:hCAST profiles (Fig. 5e). Similarly, RRHO analysis showed a marked decrease in shared transcriptome between hIAPP and type 2 diabetes islets after introduction of human calpastatin (Fig. 5f vs Fig. 1b,e). These results imply that calpain hyperactivation may play a prominent role in the shared transcriptome between hIAPP and type 2 diabetes.
Cell type analysis for IAPP misfolded protein stress
To explore the cell types that contribute to the transcriptomic signals in the bulk RNA-seq analysis, we conducted cell type marker enrichment analysis of the DEGs and modules as well as cell proportion deconvolution of bulk islet RNA-seq (Fig. 6). These analyses emphasised the impact of beta cell IAPP misfolded protein stress on multiple islet cell types including beta, alpha, endothelial, macrophage and stellate. Both the DEG and module enrichment analysis for cell type markers (Fig. 6a,b) point to a role of stellate cells with marker enrichment for module M13 (cell migration) and downregulated DEGs in hIAPP:hCAST vs hIAPP mice, possibly implicating the role of calpastatin in reducing stellate cell population. The deconvolution results showcase a reduction in beta cell and a rise in alpha cell populations in both rIAPP and hIAPP (Fig. 6c), which alludes to partial beta cell dedifferentiation (Fig. 5d). The increase in endothelial cell populations exemplifies an increase in vascularisation within the islet consistent with response to injury and dedifferentiation. Similarly, the stellate cell populations also follow this trend, and rescue by hIAPP:hCAST shows a reduction in both endothelial and stellate cell (Fig. 6d), which is consistent with reduced inflammation (Fig. 5d). In addition, there is a subtle increase in macrophages in hIAPP (Fig. 6d), matched by macrophage marker enrichment in module M1 (immune response) and in DEGs between hIAPP and rIAPP as well as between hIAPP and hIAPP:hCAST (Fig. 2b,c). These results may explain the differences in inflammation between groups. Complementing this, we found that module M7, associated with beta cells (Fig. 6b), was also enriched for type 2 diabetes GWAS candidate genes (ESM Fig. 4b).
Prominent regulatory factors shared between hIAPP islets and type 2 diabetes islets
Having established that hIAPP-induced beta cell toxicity in mouse islets results in an islet transcriptome mimicking that of human islets in type 2 diabetes, we evaluated regulatory cascades in hIAPP and type 2 diabetes-associated transcriptomic alterations by TF network and non-TF gene network analysis. TF network analysis uncovered four upstream hub transcriptional factors NF-κB, ESR1, STAT3 and CTNNB1 active in both type 2 diabetes and hIAPP islets (Fig. 7). NF-κB activation has been implicated as a core transcriptional mediator of neuronal cell inflammatory responses to amyloidogenic misfolded stress in neurons and can be protective or contribute to toxicity of injured beta cells [39] [40] and is activated by aberrant Ca2+ signalling and calpain hyperactivation [41]. NF-κB1 was assigned as an upstream regulator to co-expression module M7 that is enriched for beta cell markers (Fig. 6b) and type 2 diabetes GWAS candidate genes (ESM Fig. 4b). NF-κB is a known target of calpain, consistent with calpain hyperactivation in beta cells in hIAPP mouse islets and type 2 diabetes. STAT3 is activated by aberrant Ca2+ signalling and has been reported as a key regulator of inflammation in neurodegenerative diseases [42]. STAT3 and NF-κB cooperate as transcriptional regulators to induce angiogenesis, cellular proliferation and pro-survival metabolic remodelling, the latter through activation of hypoxia-inducible factor 1 α (HIF1α) [43]. CTNNB1 encodes β-catenin that has been implicated in tissue repair and regeneration responses in gut and beta cells, inducing cell proliferation, cell migration repair of cytoskeleton and regulation of intracellular Ca2+ dynamics [34]. ESR1 signalling is protective of beta cells in response to injury in both human and mouse islets [44].
Complementary to the TF network analysis, which is not tissue specific, we utilised an islet gene regulatory network constructed using population-based genetic and transcriptomic datasets, which captures network regulators that are not necessarily TFs. Here, we uncover the interconnectivity between modules/processes for their role in islet pathogenesis and their associated regulatory genes (Fig. 8; ESM Fig. 7). With stellate and endothelial cell markers enriched for M13 (Fig. 6b), hub genes such as COL3A1 (extracellular matrix), NID1 (wound healing) and CXCL12 (leucocyte trafficking, angiogenesis and vascular repair) are plausible regulatory genes underlying stellate and endothelial cell contribution to islet pathogenesis. Moreover, module M7, which is enriched for beta cell markers (Fig. 6b) and type 2 diabetes GWAS candidates (Fig. 8; ESM Fig. 4b), highlights hub genes such as ZNF800, which is closely associated with PAX4 [45], important in the development and differentiation of beta cells.
Discussion
The synthesis, folding and processing of insulin is close to the limit of the biosynthetic capacity of beta cells [46]. hIAPP is highly oligomer prone and readily assembles into membrane permeant toxic oligomers if the rate of expression exceeds the capacity of the cell to fold and traffic newly expressed protein. The propensity of IAPP to form toxic oligomers defines the relative vulnerability of a species to develop type 2 diabetes. Taken together, these observations suggest protein misfolding may contribute to beta cell failure leading to type 2 diabetes under conditions of insulin resistance (Fig. 9).
A striking finding from the studies is the high degree of coordinate transcriptome between islets of humans with prediabetes and type 2 diabetes. Since most individuals with prediabetes do not progress to diabetes, these findings imply much of the islet transcriptome in type 2 diabetes may reflect protective pro-survival changes. Furthermore, we found a close overlap in transcriptome between islets from mice with beta cell overexpression of rIAPP and islets from humans with either prediabetes or type 2 diabetes. Since rIAPP overexpressing mice do not develop diabetes, these findings imply that much of the islet transcriptome in prediabetes and type 2 diabetes reflects adaptive changes to increased beta cell protein synthesis. The importance of successful adaptation of beta cells to increased expression of secretory protein is further illustrated by the large number of genes linked to vulnerability to type 2 diabetes, by GWAS or Mendelian association, that were differentially expressed in rIAPP compared with WT islets.
To better understand the potential role of hIAPP toxicity in beta cell failure and loss in type 2 diabetes, we evaluated the transcriptome of islets in mice overexpressing hIAPP with that in islets of humans with type 2 diabetes. In both hIAPP islets and islets from individuals with type 2 diabetes, there was a strong inflammatory signal, consistent with an ongoing injury response. Network analysis reveals that shared key pro-survival gene networks (Fig. 7) are activated in hIAPP islets and islets from individuals with type 2 diabetes, consistent with the slow progression of beta cell loss in type 2 diabetes. Notably, a decrease in expression of genes that confer beta cells their identity (dedifferentiation) is apparent in both rIAPP mouse and prediabetes human islets, implying that beta cell dedifferentiation maybe a pro-survival adaptation.
By comparing the successful adaptation of beta cells to rIAPP vs hIAPP overexpression, the most prominent difference was the increased expression of genes ascribed to inflammation that were markedly increased in hIAPP islets and type 2 diabetes but only modestly increased in rIAPP islets and in prediabetes. Type 2 diabetes is a heterogeneous disease and there are likely multiple pathways to beta cell toxicity, including toxic actions of lipids and hyperglycaemia [47]. In a recent partial pancreatectomy study in rats, the resulting marked decrease in beta cell mass and modest hyperglycaemia induced many of the same signals observed in response to beta cell hIAPP toxicity in the present study, including inflammation and partial beta cell differentiation [48]. In an in vitro acute injury study of human islets exposed to glucolipotoxicity, there was also some overlap with islets from humans with type 2 diabetes by RRHO analysis [49], but the overlap was weaker compared with IAPP overexpression in the current study. The difference could be due to in vivo vs in vitro conditions or intrinsic differences between IAPP and glucolipotoxicity.
A limitation of the current study is the use of whole pancreatic islets for RNA-seq, which masks the cell types contributing to the transcriptional signals. Although we used cellular deconvolution and cell marker enrichment methods to mitigate this issue, single cell RNA-seq will be of added value in future studies to confirm the predicted cell type contributions from our deconvolution analysis. In addition, follow-up on the various genes found to be contributing to hIAPP beta cell toxicity through knockdown or overexpression studies will be of use to further confirm their causal role in disease. Another limitation of the present study is the potential bias caused by use of an inbred mouse FVB strain.
In conclusion, the present studies suggest that much of the islet transcriptome in type 2 diabetes is adaptive to the increased beta cell burden of protein synthesis and folding. Beta cell hIAPP toxicity induces a prominent islet inflammatory response, consistent with that observed in type 2 diabetes, implying protein misfolding stress may serve to initiate or contribute to beta cell injury in type 2 diabetes. There are also shared pro-survival gene networks in hIAPP and type 2 diabetes islets.
Taken together these studies suggest caution should be taken in interpreting transcriptome changes in islets in type 2 diabetes as therapeutic targets since many, if not most, of these changes are likely pro-survival adaptations. Strategies to suppress IAPP expression warrant further investigation due to the mounting evidence to suggest its role in type 2 diabetes pathogenesis.
Abbreviations
- DEG:
-
Differentially expressed gene
- ER:
-
Endoplasmic reticulum
- FDR:
-
False discovery rate
- GEO:
-
Gene expression omnibus
- GWAS:
-
Genome-wide association studies
- hCAST:
-
Human calpastatin
- hIAPP:
-
Human islet amyloid polypeptide
- IAPP:
-
Islet amyloid polypeptide
- MAPK:
-
Mitogen activated protein kinase
- rIAPP:
-
Rodent islet amyloid polypeptide
- RNA-seq:
-
RNA sequencing
- RRHO:
-
Rank-rank hypergeometric overlap
- TF:
-
Transcription factor
- UCLA:
-
University of California, Los Angeles
- UPR:
-
Unfolded protein response
- WGCNA:
-
Weighted gene co-expression network analysis
- WT:
-
Wild-type
References
Jurgens CA, Toukatly MN, Fligner CL et al (2011) β-Cell loss and β-cell apoptosis in human type 2 diabetes are related to islet amyloid deposition. Am J Pathol 178(6):2632–2640. https://doi.org/10.1016/j.ajpath.2011.02.036
Hull RL, Westermark GT, Westermark P, Kahn SE (2004) Islet amyloid: a critical entity in the pathogenesis of type 2 diabetes. J Clin Endocrinol Metab 89(8):3629–3643. https://doi.org/10.1210/jc.2004-0405
Kahn SE, Andrikopoulos S, Verchere CB (1999) Islet amyloid: a long-recognized but underappreciated pathological feature of type 2 diabetes. Diabetes 48(2):241–253. https://doi.org/10.2337/diabetes.48.2.241
Cooper GJ, Willis AC, Clark A, Turner RC, Sim RB, Reid KB (1987) Purification and characterization of a peptide from amyloid-rich pancreases of type 2 diabetic patients. Proc Natl Acad Sci 84(23):8628–8632. https://doi.org/10.1073/pnas.84.23.8628
Janson J, Ashley RH, Harrison D, McIntyre S, Butler PC (1999) The mechanism of islet amyloid polypeptide toxicity is membrane disruption by intermediate-sized toxic amyloid particles. Diabetes 48(3):491–498. https://doi.org/10.2337/diabetes.48.3.491
O'Brien TD, Butler PC, Westermark P, Johnson KH (1993) Islet amyloid polypeptide: a review of its biology and potential roles in the pathogenesis of diabetes mellitus. Vet Pathol 30(4):317–332. https://doi.org/10.1177/030098589303000401
Halban PA, Polonsky KS, Bowden DW et al (2014) beta-cell failure in type 2 diabetes: postulated mechanisms and prospects for prevention and treatment. J Clin Endocrinol Metab 99(6):1983–1992. https://doi.org/10.1210/jc.2014-1425
Haffner SM (1998) Epidemiology of type 2 diabetes: risk factors. Diabetes Care 21(Suppl 3):C3–C6. https://doi.org/10.2337/diacare.21.3.c3
Barker DJ, Hales CN, Fall CH, Osmond C, Phipps K, Clark PM (1993) Type 2 (non-insulin-dependent) diabetes mellitus, hypertension and hyperlipidaemia (syndrome X): relation to reduced fetal growth. Diabetologia 36(1):62–67. https://doi.org/10.1007/BF00399095
Sasaki H, Saisho Y, Inaishi J et al (2020) Associations of birthweight and history of childhood obesity with beta cell mass in Japanese adults. Diabetologia. https://doi.org/10.1007/s00125-020-05127-2
Costes S, Langen R, Gurlo T, Matveyenko AV, Butler PC (2013) beta-cell failure in type 2 diabetes: a case of asking too much of too few? Diabetes 62(2):327–335. https://doi.org/10.2337/db12-1326
Janson J, Soeller WC, Roche PC et al (1996) Spontaneous diabetes mellitus in transgenic mice expressing human islet amyloid polypeptide. Proc Natl Acad Sci U S A 93(14):7283–7288. https://doi.org/10.1073/pnas.93.14.7283
Vilchez D, Saez I, Dillin A (2014) The role of protein clearance mechanisms in organismal ageing and age-related diseases. Nat Commun 5:5659. https://doi.org/10.1038/ncomms6659
Hogan MF, Ziemann M, Harikrishnan KN et al (2019) RNA-seq-based identification of Star upregulation by islet amyloid formation. Protein Eng Des Sel 32(2):67–76. https://doi.org/10.1093/protein/gzz022
Huang CJ, Haataja L, Gurlo T et al (2007) Induction of endoplasmic reticulum stress-induced beta-cell apoptosis and accumulation of polyubiquitinated proteins by human islet amyloid polypeptide. Am J Physiol Endocrinol Metab 293(6):E1656–E1662. https://doi.org/10.1152/ajpendo.00318.2007
Gurlo T, Costes S, Hoang JD, Rivera JF, Butler AE, Butler PC (2016) beta cell-specific increased expression of calpastatin prevents diabetes induced by islet amyloid polypeptide toxicity. JCI Insight 1(18):e89590. https://doi.org/10.1172/jci.insight.89590
Rivera JF, Costes S, Gurlo T, Glabe CG, Butler PC (2014) Autophagy defends pancreatic beta cells from human islet amyloid polypeptide-induced toxicity. J Clin Invest 124(8):3489–3500. https://doi.org/10.1172/JCI71981
Dobin A, Davis CA, Schlesinger F et al (2013) STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29(1):15–21. https://doi.org/10.1093/bioinformatics/bts635
Anders S, Pyl PT, Huber W (2015) HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics 31(2):166–169. https://doi.org/10.1093/bioinformatics/btu638
Fadista J, Vikman P, Laakso EO et al (2014) Global genomic and transcriptomic analysis of human pancreatic islets reveals novel genes influencing glucose metabolism. Proc Natl Acad Sci U S A 111(38):13924–13929. https://doi.org/10.1073/pnas.1402665111
Law CW, Chen Y, Shi W, Smyth GK (2014) voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol 15(2):R29. https://doi.org/10.1186/gb-2014-15-2-r29
Cahill KM, Huo Z, Tseng GC, Logan RW, Seney ML (2018) Improved identification of concordant and discordant gene expression signatures using an updated rank-rank hypergeometric overlap approach. Sci Rep 8(1):9588. https://doi.org/10.1038/s41598-018-27903-2
Ogata H, Goto S, Fujibuchi W, Kanehisa M (1998) Computation with the KEGG pathway database. Biosystems 47(1–2):119–128. https://doi.org/10.1016/s0303-2647(98)00017-3
Langfelder P, Horvath S (2007) Eigengene networks for studying the relationships between co-expression modules. BMC Syst Biol 1:54. https://doi.org/10.1186/1752-0509-1-54
Langfelder P, Horvath S (2008) WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 9:559. https://doi.org/10.1186/1471-2105-9-559
Franzen O, Gan LM, Bjorkegren JLM (2019) PanglaoDB: a web server for exploration of mouse and human single-cell RNA sequencing data. Database (Oxford) 2019. https://doi.org/10.1093/database/baz046
Shen L (2014) GeneOverlap: an R package to test and visualize gene overlaps. R Package
Newman AM, Steen CB, Liu CL et al (2019) Determining cell type abundance and expression from bulk tissues with digital cytometry. Nat Biotechnol 37(7):773–782. https://doi.org/10.1038/s41587-019-0114-2
Chen EY, Tan CM, Kou Y et al (2013) Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics 14:128. https://doi.org/10.1186/1471-2105-14-128
Shu L, Zhao Y, Kurt Z et al (2016) Mergeomics: multidimensional data integration to identify pathogenic perturbations to biological systems. BMC Genomics 17(1):874. https://doi.org/10.1186/s12864-016-3198-9
Eguchi K, Nagai R (2017) Islet inflammation in type 2 diabetes and physiology. J Clin Invest 127(1):14–23. https://doi.org/10.1172/JCI88877
Hess DA, Strelau KM, Karki A et al (2016) MIST1 links secretion and stress as both target and regulator of the unfolded protein response. Mol Cell Biol 36(23):2931–2944. https://doi.org/10.1128/MCB.00366-16
Riopel M, Seo JB, Bandyopadhyay GK et al (2018) Chronic fractalkine administration improves glucose tolerance and pancreatic endocrine function. J Clin Invest 128(4):1458–1470. https://doi.org/10.1172/JCI94330
Keefe MD, Wang H, De la OJ, Khan A, Firpo MA, Murtaugh LC (2012) beta-catenin is selectively required for the expansion and regeneration of mature pancreatic acinar cells in mice. Dis Model Mech 5(4):503–514. https://doi.org/10.1242/dmm.007799
Tschen SI, Dhawan S, Gurlo T, Bhushan A (2009) Age-dependent decline in beta-cell proliferation restricts the capacity of beta-cell regeneration in mice. Diabetes 58(6):1312–1320. https://doi.org/10.2337/db08-1651
Young A (2005) Inhibition of insulin secretion. Adv Pharmacol 52:173–192. https://doi.org/10.1016/S1054-3589(05)52009-X
Ferreira A (2012) Calpain dysregulation in Alzheimer’s disease. ISRN Biochem 2012:728571. https://doi.org/10.5402/2012/728571
Ji J, Su L, Liu Z (2016) Critical role of calpain in inflammation. Biomed Rep 5(6):647–652. https://doi.org/10.3892/br.2016.785
Eldor R, Yeffet A, Baum K et al (2006) Conditional and specific NF-kappaB blockade protects pancreatic beta cells from diabetogenic agents. Proc Natl Acad Sci U S A 103(13):5072–5077. https://doi.org/10.1073/pnas.0508166103
Jones SV, Kounatidis I (2017) Nuclear factor-kappa B and Alzheimer disease, unifying genetic and environmental risk factors from cell to humans. Front Immunol 8:1805. https://doi.org/10.3389/fimmu.2017.01805
Lilienbaum A, Israel A (2003) From calcium to NF-kappa B signaling pathways in neurons. Mol Cell Biol 23(8):2680–2698. https://doi.org/10.1128/mcb.23.8.2680-2698.2003
Reichenbach N, Delekate A, Plescher M et al (2019) Inhibition of Stat3-mediated astrogliosis ameliorates pathology in an Alzheimer’s disease model. EMBO Mol Med 11(2). https://doi.org/10.15252/emmm.201809665
Grivennikov SI, Karin M (2010) Dangerous liaisons: STAT3 and NF-kappaB collaboration and crosstalk in cancer. Cytokine Growth Factor Rev 21(1):11–19. https://doi.org/10.1016/j.cytogfr.2009.11.005
Zhou Z, Ribas V, Rajbhandari P et al (2018) Estrogen receptor alpha protects pancreatic beta-cells from apoptosis by preserving mitochondrial function and suppressing endoplasmic reticulum stress. J Biol Chem 293(13):4735–4751. https://doi.org/10.1074/jbc.M117.805069
Kato N (2013) Insights into the genetic basis of type 2 diabetes. J Diabetes Investig 4(3):233–244. https://doi.org/10.1111/jdi.12067
Rege NK, Liu M, Yang Y et al (2020) Evolution of insulin at the edge of foldability and its medical implications. Proc Natl Acad Sci U S A 117(47):29618–29628. https://doi.org/10.1073/pnas.2010908117
Prentki M, Matschinsky FM, Madiraju SR (2013) Metabolic signaling in fuel-induced insulin secretion. Cell Metab 18(2):162–185. https://doi.org/10.1016/j.cmet.2013.05.018
Ebrahimi AG, Hollister-Lock J, Sullivan BA, Tsuchida R, Bonner-Weir S, Weir GC (2020) Beta cell identity changes with mild hyperglycemia: implications for function, growth, and vulnerability. Mol Metab 35:100959. https://doi.org/10.1016/j.molmet.2020.02.002
Marselli L, Piron A, Suleiman M et al (2020) Persistent or transient human beta cell dysfunction induced by metabolic stress: specific signatures and shared gene expression with type 2 diabetes. Cell Rep 33(9):108466. https://doi.org/10.1016/j.celrep.2020.108466
Authors’ relationships and activities
The authors declare that there are no relationships or activities that might bias, or be perceived to bias, their work.
Author information
Authors and Affiliations
Contributions
All authors substantially contributed to conception and design, acquisition of data, or critical analysis and interpretation of data, drafting the article or revising it critically for important intellectual content. All authors approved the final version to be published. GC, MB and TG are responsible for the integrity of the work as a whole.
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
ESM
(PDF 2.56 mb)
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Blencowe, M., Furterer, A., Wang, Q. et al. IAPP-induced beta cell stress recapitulates the islet transcriptome in type 2 diabetes. Diabetologia 65, 173–187 (2022). https://doi.org/10.1007/s00125-021-05569-2
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00125-021-05569-2