- Combined transcriptome and proteome profiling of the pancreatic β -cell response to palmitate unveils key pathways of β -cell lipotoxicity. - To unravel critical mediators of lipotoxicity upstream of the palmitate-modified genes, we identified overrepresented transcription factor binding sites and performed network inference analysis. - 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. - Full list of author information is available at the end of the article. - The identification of critical drivers and pathways of the β - cell lipotoxic response points to novel therapeutic targets for type 2 diabetes.. - The prevalence of type 2 diabetes has dramatically risen over the past few decades, reaching currently 9% of the world population [1]. - This rise, expected to continue in the coming years, can be in large part explained by the global epidemic of obesity. - Obesity increases the risk of the developing type 2 diabetes 7-fold [2] and is associ- ated with increased circulating levels of free fatty acids (FFA), due to resistance to the anti-lipolytic effect of in- sulin and increased adipose tissue mass [3]. - In large-scale epidemiological studies in the United States [12, 13] and Europe [14], high circulating levels of palmitic acid were associated with higher risk of type 2 diabetes, insulin resistance and inflammation [12].. - RNA-sequencing (RNA-seq) enables the interrogation of the whole tran- scriptome and has become the method of choice for transcriptome profiling. - In the present study, we crossed this tran- scriptomic study (with the addition of one more experi- ment) with a proteome analysis of palmitate-treated rat INS-1E cells. - We used both targeted and unbiased bio- informatic analyses to identify critical pathways and reg- ulators of the β-cell response to lipotoxicity that are preserved among two different species, suggesting rele- vant functional impact. - RRHO heatmaps allow to visualize the pattern and signifi- cance of the overlap. - Perfectly correlated gene expres- sion changes in two datasets generate a strong positive signal along the diagonal in the RRHO heat- map (Supplementary Fig. - 1A), whereas perfect overlap limited to the most up- and downregulated genes translates to a significant signal at the bottom left and upper right corners of the map, respectively (Supplementary Fig. - 1b), rem- iniscent of the map obtained when two independent data- sets have identical expression changes (Supplementary Fig. - These data indicate that gene signatures between rat INS-1E cells and human islets exposed to palmitate are significantly correlated, rendering their joint analysis in the present study pertinent.. - The level map colors represent -log p -values for overlap between ranked genes, with an indication of the smallest FDR corrected p- value for coordinates with the most statistically significant overlap between genes up-regulated in both datasets (bottom left corner) and down- regulated in both (top right corner). - Palmitate downregulated DHCR24, which catalyzes the final step in cholesterol biosynthesis, induced expression of the LDL receptor and attenuated ABCG1 expression.. - In previous studies by us and others, chronic exposure to palmitate has been shown to activate signaling through the three canonical branches of the ER stress response under the control of PERK, IRE1 and ATF6, and to elicit ER stress-induced apoptosis [15, 26].. - CREB3L2 and other members of the CREB3/ATF fam- ily of transcription factors have been proposed as novel ER stress transducers, functioning in a cell- and stimulus-specific manner [27]. - Palmi- tate has been shown to inhibit protein trafficking between ER and the Golgi, inducing ER stress by subse- quent protein overload in the ER [31]. - Palmitate also up- regulated ERO1A, an oxidoreductase involved in oxidative protein folding in the ER. - Finally, palmi- tate downregulated EIF2B5, which codes for the catalytic eIF2Bε subunit of the eukaryotic initiation of translation factor eIF2B [33]. - our current findings also suggest transcriptional and transla- tional regulation of the protein.. - It upregulated CLIC1, an intracellular chloride ion channel, previously shown to act as a down- stream effector of insulin [34], and it enhanced expres- sion of members of the SLC7 family of amino-acid transporters. - Palmitate inhibited expression of the GLP1 receptor in our –omic studies. - Pathway analysis and enriched transcription factors Further to our manual annotation, we used bioinfor- matic approaches to obtain an unbiased overview of the biological pathways modified in lipotoxic conditions.. - IPA indicated that upregulated genes were involved principally in the oxi- dative stress response, fatty acid β-oxidation, mitochon- drial dysfunction, the ERK/MAPK pathway and amino- acid metabolism (Fig. - To obtain this prior network, we searched IPA and DAVID for putative upstream regula- tors in the set of 207 differentially expressed genes/pro- teins. - 5, we obtained a network of 416 regulations involv- ing 190 genes/proteins: 3 regulations inferred from the RNA-seq and proteomics dataset, all present in the prior network. - 97 inferred from the RNA-seq dataset, of which 44 were present in the prior network. - and 316 inferred from the proteomics dataset, of which 129 were present in the prior network. - HNF1A mutations cause maturity-onset diabetes of the young (MODY) type 3 and rare variants in this gene increase type 2 dia- betes risk [42]. - The length of the bars is proportional to the significance of the association between the set of genes and the pathway, expressed by the negative logarithm of the p-value. - This study, as well as microarray studies of human islets [18, 46] and clonal β-cells [47, 48] by other groups, have helped to expand our under- standing of the underlying mechanisms of lipotoxicity. - In the present study, we used state-of-the art, highly sensitive transcriptome and proteome profiling tech- nologies. - These allowed deep coverage of the transcrip- tome and detection of least 2–3 times more proteins compared to previous proteomic studies of palmitate- treated β-cells . - The length of the bars is proportional to the significance of the overrepresentation of potential binding sites for the indicated transcription factors in the modified genes, expressed by the negative logarithm of the p-value. - The red line indicates the fold enrichment of the palmitate-modified genes compared to a random set of genes from the human genome. - The palmitate-induced lipid metabolism gene/protein expression changes identified in the present study may hence contribute to inhibit insulin. - Transcription factor enrichment analysis by DAVID identified PPARα as a critical player in the transcriptional response. - PPARα is predicted in silico to bind to the large majority of the palmitate-modulated genes related to lipid metabolism.. - Our network inference analysis indicated FOXO1 as a major regulator of the β-cell response to palmitate.. - Regulatory networks were inferred in the RNA-seq and proteomics data separately using a random forest algorithm to score predictors and then the 2 networks were combined. - Both in the DAVID and network inference analyses, BACH1 was predicted to be a key transcription factor mediating the response to palmitate. - This points to a role of BACH1 in regulating insulin se- cretion, through the modulation of the oxidative stress response.. - Metabolomic studies of β-cells, basally and in the face of metabolic stress, will shed further light on the biological impact of amino-acid alterations.. - First, transcriptome was performed in human is- lets, while proteome was performed in clonal rat INS-1E β-cells, because of the amount of material needed for protein profiling. - It has been shown, however, that there is a high correlation between β-cell and islet-expressed genes (r = 0.94) and that 87% of the variance in β-cell gene expression can be explained using islet expression as a proxy [79]. - Taking into account that RNA-seq is more sensitive in detecting quantitative changes in gene expression, we applied less strict criteria in the definition for differential expression from the pro- teomics database.. - This combined transcriptomic and proteomic study of β-cells provides a comprehensive overview of the gene expression changes induced by palmitate and highlights significant pathways implicated in the response to lipo- toxicity. - Our bioinformatic analyses unveil tran- scription factors that may act as drivers of the response to FFA and could serve as targets for future investigations.. - The islets were treated with 0.5 mM palmitate (Sigma, Schnelldorf, Germany) or control (ethanol) in the same medium con- taining 1% charcoal-absorbed BSA but no FBS for 48 h.. - The average percentage of β-cells in the human islet prepara- tions, examined by insulin immunofluorescence [15], was 50%.. - RNA-seq. - iTRAQ is based on the covalent labeling of the N-terminus and side chain amines of peptides from pro- tein digestions with tags of varying mass. - The fragmentation of the attached tag gener- ates a low molecular mass reporter ion that was used to relatively quantify the peptides and proteins from which they originated [83]. - A detailed description of the method is provided in the additional material (Supple- mentary Methods).. - Protein expression was first divided by the expression mean of the control samples (i.e. - A 95% confidence interval for the ratio between the control samples was estimated (2 to the power of ± 1.96*standard deviation of the log2 transformed ratio, as- suming normality) and used as cut-off values for up and down-regulation (1.24 and 0.8 respectively). - Proteins were considered outliers and excluded from analyses if the ratio of the control samples exceeded the estimated 95% confidence interval. - A protein was considered dif- ferentially expressed if its expression was higher/lower than these cut-offs in any of the time points of exposure in at least one experiment. - The p-values of the overlap- ping genes were assessed by a two-tailed hypergeometric test and false discovery rate corrected by the Benjamini and Yekutieli method. - The RRHO R package was modi- fied to better take into account the multiplicity of min- imal p-values, null p-values, the up- and down-regulated genes going in opposite direction and the asymmetry be- tween the number of genes up- or down-regulated in the two datasets.. - Gene regulatory networks were obtained by combining inferred networks from the expression profiles of the proteomics and RNA-seq datasets and a prior network, obtained from literature knowledge.. - To obtain a prior network using literature data, we searched for putative upstream regulators of the set of 207 differentially expressed genes/proteins using IPA (QIAGEN, Redwood City) and DAVID. - A prior regulatory network was obtained by associating the enriched tran- scription factors to the respective targets, and including regulations obtained in the TRANSFAC [85] and RegNetwork [86] databases, involving the novel set of 258 genes/proteins. - In the end, a prior network of 3082 regulations between 258 genes/proteins was obtained (1877 regulations from DAVID, 232 regulations from IPA, 938 regulations from TRANSFAC, 551 regulations from RegNetwork).. - A regulatory network was inferred in the RNA-seq and proteomic datasets separately. - In the RNA-seq data, fold change values were used (the minimum RPKM was set to 0.1). - In the proteo- mics dataset, the inference was directed, making use of the fact that different time points were used. - In the RNA-seq dataset, the inference was undirected, and the regulation score between two genes was the maximum of the two scores obtained when each of the genes was considered as target.. - 0.05 and the regulation was present in the prior network. - For these experiments, human is- lets were cultured in the same medium as described above (see section human islets and rodent β-cells). - Exposure to palmitate (0.5 mM) in the presence of 1% charcoal- absorbed BSA or precomplexed to 0.67% FFA-free BSA results in similar unbound FFA concentrations [81].. - At least 400 cells were counted per experimental condi- tion by two investigators, one of them unaware of the conditions, with an agreement between them of >. - For gene expression data, the same tests were applied after loga- rithmic transformation of the data. - To exemplify and facilitate interpretation of the RRHO plots, RRHO maps were generated for 3 different hypothetical conditions: (A) Identical gene expression changes in two unrelated transcriptome datasets X and Y that generate perfect overlap. - Characteristics of the organ donors and human islet preparations used for RNA-seq. - available in the ProteomeXchange Consortium via JPOST partner repository with dataset identifier PXD020851.. - Human islet collection and handling were approved by the Ethical Committee of the University of Pisa, Pisa, Italy. - Written informed consent was obtained from relatives of the deceased organ donors, none of whom were under the age of 16 years.. - Nutrient-induced metabolic stress, adaptation, detoxification, and toxicity in the pancreatic β -cell.. - Adaptive changes of human islets to an obesogenic environment in the mouse. - Prospective association of fatty acids in the de novo lipogenesis pathway with risk of type 2 diabetes: the. - Differences in the prospective association between individual plasma phospholipid saturated fatty acids and incident type 2 diabetes: the EPIC-InterAct case-cohort study. - RNA sequencing identifies dysregulation of the human pancreatic islet transcriptome by the saturated fatty acid palmitate. - An intracellular role for ABCG1-mediated cholesterol transport in the regulated secretory pathway of mouse pancreatic beta cells. - Hyperactivity of the Ero1alpha oxidase elicits endoplasmic reticulum stress but no broad antioxidant response. - The linkage and association of the gene encoding upstream stimulatory factor 1 with type 2 diabetes and metabolic syndrome in the Chinese population. - Functional investigations of HNF1A identify rare variants as risk factors for type 2 diabetes in the general population. - The BTB and CNC homology 1 (BACH1) target genes are involved in the oxidative stress response and in control of the cell cycle. - Role of the saturated nonesterified fatty acid palmitate in beta cell dysfunction. - Oleate protects beta-cells from the toxic effect of palmitate by activating pro- survival pathways of the ER stress response. - Human beta-cell proliferation and intracellular signaling: driving in the dark without a road map. - Major species differences between humans and rodents in the susceptibility to pancreatic beta-cell injury. - Species-related differences in the proteome of rat and human pancreatic Beta cells. - Cell-type, allelic, and genetic signatures in the human pancreatic beta cell transcriptome. - Towards better understanding of the contributions of overwork and glucotoxicity to the beta-cell inadequacy of type 2 diabetes
Xem thử không khả dụng, vui lòng xem tại trang nguồn hoặc xem
Tóm tắt