The Identification of Potential Factors Associated with the Development of Type 2 Diabetes

Abstract

Type 2 diabetes (T2D) arises when pancreatic β-cells fail to compensate for systemic insulin resistance with appropriate insulin secretion. However, the link between insulin resistance and β-cell failure in T2D is not fully understood. To explore this association, we studied transgenic MKR mice that initially develop insulin resistance in skeletal muscle but by 8 weeks of age have T2D. In the present study, global islet protein and gene expression changes were characterized in diabetic MKR versus non-diabetic control mice at 10 weeks of age. Using a quantitative proteomics approach (isobaric tags for relative and absolute quantification (iTRAQ)), 159 proteins were differentially expressed in MKR compared with control islets. Marked up-regulation of protein biosynthesis and endoplasmic reticulum stress pathways and parallel down-regulation in insulin processing/secretion, energy utilization, and metabolism were observed. A fraction of the differentially expressed proteins identified (including GLUT2, DNAJC3, VAMP2, RAB3A, and PC1/3) were linked previously to insulin-secretory defects and T2D. However, many proteins for the first time were associated with islet dysfunction, including the unfolded protein response proteins (ERP72, ERP44, ERP29, PPIB, FKBP2, FKBP11, and DNAJB11), endoplasmic reticulum-associated degradation proteins (VCP and UFM1), and multiple proteins associated with mitochondrial energy metabolism (NDUFA9, UQCRH, COX2, COX4I1, COX5A, ATP6V1B2, ATP6V1H, ANT1, ANT2, ETFA, and ETFB). The mRNA expression level corresponding to these proteins was examined by microarray, and then a small subset was validated using quantitative real time PCR and Western blot analyses. Importantly ∼54% of differentially expressed proteins in MKR islets (including proteins involved in proinsulin processing, protein biosynthesis, and mitochondrial oxidation) showed changes in the proteome but not transcriptome, suggesting post-transcriptional regulation. These results underscore the importance of integrated mRNA and protein expression measurements and validate the use of the iTRAQ method combined with microarray to assess global protein and gene changes involved in the development of T2D.

The prevalence of type 2 diabetes (T2D)1 is reaching epidemic proportions and presents a severe health burden worldwide (1, 2). The pathogenesis of T2D is thought to be complicated, involving multiple genetic, metabolic, and environmental factors. In high risk T2D subjects, the earliest detectable abnormality is insulin resistance in the skeletal muscle (3–5), which is characterized by impaired insulin-mediated signaling, glucose usage, gene expression, and glycogen synthesis as well as the accumulation of intramyocellular triglycerides (4, 6–10). However, not all individuals with insulin resistance develop T2D, and it is now clear that T2D only ensues when insulin-producing pancreatic β-cells fail to compensate for the increased metabolic demands associated with insulin resistance (11–13). The exact mechanisms underlying β-cell failure associated with most human forms of T2D remain to be identified (14, 15).

Several relatively rare monogenic forms of diabetes characterized by defects in β-cells have been identified (5, 15–20). In these forms, diabetes is caused by disruptions or mutations in transcription factors that regulate gene expression in β-cells, leading to an early onset of the disease. However, in most cases of T2D, it is predicted that more subtle alterations in multiple genes and proteins that control glucose-stimulated insulin secretion (GSIS) and β-cell survival play a prominent role in determining susceptibility (21). Thus, in many forms of T2D it is predicted that multiple genes interact to diminish the ability of β-cells to respond to changes in metabolic demand. The potential importance of such complex gene-gene interactions in controlling β-cell function has been supported by several polygenic mutations in animal models (22–25).

A substantial effort toward understanding insulin resistance and β-cell dysfunction in T2D has been made using animal models and human subjects (5, 15–20, 26–30). To this end, DNA microarray analyses provide a powerful tool to search for clues regarding the molecular mechanisms associated with the pathogenesis of T2D. Gene array studies have been performed on islets to study aspects of T2D in animal models (31–33); however, it is ultimately changes at the protein level that affect cellular function. To date, limited information is available concerning large scale dynamic protein expression changes in pancreatic islets linked to this disease. Using two-dimensional gel electrophoresis combined with mass spectrometry, Sanchez et al. (34) reported nine differentially expressed proteins between ob/ob (diabetic) and lean mouse islets. Qiu et al. (35) found three differentially expressed proteins associated with the high fat diet-induced T2D using mouse pancreatic lysates.

To further understand the link between insulin resistance and β-cell dysfunction in T2D, we have studied a mouse model of insulin resistance that progressively develops diabetes (36). One unique feature of the MKR mouse is that it does not harbor a genetic defect in β-cells but rather has a dominant-negative insulin-like growth factor-I receptor mutation specifically targeted to the skeletal muscle (36). Our previous studies revealed that the mutation induces a progressive systemic insulin resistance that leads to compensatory increases in islet and β-cell mass, defective GSIS, β-cell dysfunction, and T2D by 8 weeks of age (36–38). Therefore, examining the protein expression pattern of diseased MKR compared with healthy islets may provide important clues to the molecular events associated with the dynamic transition of β-cell dysfunction to failure. In this study, we characterized T2D MKR islets by applying an integrated quantitative iTRAQ proteomics and DNA microarray approach combined with Western blot and quantitative real time PCR for validation. A total of 159 proteins were dysregulated in diabetic MKR islets compared with controls. Functional cluster analysis of these proteins revealed a marked up-regulation of protein biosynthesis and endoplasmic reticulum (ER) stress pathways and a concomitant down-regulation in insulin processing and secretion, as well as mitochondrial energy metabolism pathways in MKR islets. In addition to the affirmation of known diabetogenic proteins, this study revealed novel proteins involved in ER stress and mitochondrial oxidative metabolism that may be associated with β-cell dysfunction and T2D.

EXPERIMENTAL PROCEDURES

Animal Care

Mice were maintained in a standard 12-h light/dark cycle and had free access to water and food (diet number 8664; Harlan Tekland, Madison, WI). MKR mice were genotyped by PCR analysis of tail DNA (37). All studies were performed on male mice. Wild-type (WT) FVB mice (Charles River, Wilmington, MA) were used as controls. Animal care procedures were conducted according to protocols and the standards of the Canadian Council on Animal Care and approved by the Animal Care and Use Committee at the University of Toronto.

Assessment of Mouse Weight and Non-fasting Glucose and Insulin Levels

Mouse blood glucose and insulin levels were measured from tail vein blood using a glucometer (Bayer, Toronto, Ontario, Canada) and radioimmunoassays (Linco Research, St. Charles, MO), respectively, under non-fasting conditions (39).

Pancreatic Islet Morphology Study and Islet Isolation

Pancreatic islets were isolated from age-matched male MKR and WT mice by collagenase digestion as described previously (39). Briefly the pancreatic duct was perfused with 3 ml of type V collagenase (Sigma). The pancreas was then dissected and digested by incubating for 13–15 min at 37 °C in 5 ml of type V collagenase. Islets were hand-picked three times under a dissecting microscope to remove as much exocrine tissue contamination as possible. Islets were either cultured in RPMI 1640 medium (containing 10% fetal bovine serum and 11.1 mm glucose) or processed for protein and RNA preparation immediately.

Pancreatic islet morphology was determined in 3- and 10-week-old mice. Mice were sacrificed, and the pancreas was removed, fixed in 4% paraformaldehyde before being mounted in paraffin blocks, and sectioned for immunostaining with an insulin antibody as described previously (40). Images of freshly isolated islets were taken using a Zeiss LSM510 laser scanning microscope.

Islet Secretion

Glucose-stimulated insulin secretion studies were performed in 3- and 10-week-old mice as reported previously (38, 40). The isolated islets were hand-picked and cultured overnight prior to glucose stimulation. Insulin secretion was measured from groups of 20 islets using Krebs-Ringer bicarbonate buffer solution containing 2.8 or 20 mm glucose. Islet DNA was extracted using acid-ethanol, and DNA was calculated for normalization. Insulin concentrations were measured by radioimmunoassays.

iTRAQ Studies

An overview of the iTRAQ work flow is shown in Fig. 2A. For a complete description of the iTRAQ labeling reaction and the methods for analyses please refer to Refs. 41 and 42. To decrease eventual biases caused by biological variations, islets from 8–10 mice of each group were pooled yielding one sample. Islet protein samples were prepared in cell lysis buffer (Cell Signaling Technology Inc., Beverly, MA) and kept in liquid N₂ until use in the proteomics study. The protein concentration was measured using a BCA™ protein assay kit (Pierce). Three iTRAQ analyses using three independently isolated pancreatic islet samples each containing islets from 8–10 mice were performed.

Fig. 2.

Quantitative iTRAQ proteomics approach. A, flow chart of iTRAQ proteomics approach. B and C, PDI was up-regulated 2.43-fold in MKR islets. Quantitative information is encoded in the low mass-to-charge ratio portion of the MS/MS spectrum. The MKR islet sample was labeled with iTRAQ-117, and the WT islet sample was labeled with iTRAQ-114. Relative peak areas of the two marker ions were used to quantify the PDI levels (B). For each MS/MS spectrum, y- and b-type fragment ions (containing the C and N termini of the peptide, respectively) enable the identification of the peptide sequence (C).

iTRAQ Sample Labeling—

The iTRAQ reagent labeling was performed according to the manufacturer's instructions (Applied Biosystems, Foster City, CA). Before performing islet sample labeling, a defined six-protein mixture (Applied Biosystems) was labeled and used to confirm the accuracy of ratiometric quantitation of the iTRAQ reagents. Islet protein lysates from MKR and WT mice were purified by acetone precipitation. 150 μg of islet protein from each group was dissolved in 20 μl of dissolution buffer and 1 μl of denaturant reagent. The samples were reduced by addition of 2 μl of reducing reagent and incubation at 60 °C for 1 h. Reduced cysteine residues were then blocked by addition of 1 μl of cysteine blocking reagent and incubated at room temperature for a further 10 min. Tryptic digestion was initiated by the addition of 10 μl of trypsin solution (Applied Biosystems; prepared as 0.5 μg/μl in water solution with enzyme and substrate ratio of ∼1:30) and incubated at 37 °C for 12–16 h. To label the peptides with iTRAQ reagents, one vial of label reagent was thawed and reconstituted in 70 μl of ethanol. The reagent solution was added to the digest (reagent 117 for MKR sample and reagent 114 for WT control) and incubated for 1 h at room temperature. The labeled samples were then mixed together before LC/MS/MS analysis.

Sample Fractionation—

To remove excess, unbound iTRAQ reagent and to simplify the peptide mixture, the labeled peptide mixture was purified and fractionated using an off-line strong cation exchange column (PolySulfoethyl A column, 2.1 × 200 mm, 5 μm, 300 Å) on an Agilent 1100 HPLC system. The mixed sample was diluted in loading buffer (25% (v/v) acetonitrile, 10 mm potassium phosphate, pH 3.0) and injected into the strong cation exchange column. After being washed isocratically for 10 min at 200 μl/min to remove excess reagent, the peptides were eluted with the gradient from 0% buffer A (25% ACN, 10 mm potassium phosphate, pH 3.0) to 25% buffer B (25% ACN, 10 mm potassium phosphate, 350 mm KCl, pH 3.0) in 30 min and then from 25% B to 100% B in 20 min. The fractions were collected at 1-min intervals. The fractions were completely dried in a SpeedVac (Thermo Electron Corp., Waltham, MA). The samples were combined into 20–30 fractions, desalted using peptide cleanup C₁₈ spin tubes (Agilent Technologies, Palo Alto, CA), and vacuum-dried before being sent for LC/MS/MS analysis.

LC/MS/MS Analysis—

Dried sample fractions were resuspended in 5 μl of 0.1% (v/v) formic acid and loaded onto a New Objective C₁₈ PicoFrit column (75 μm × 10 cm, 5-μm particle size, 300 Å) using an UltiMate micropump (LC Packings, Amsterdam, Netherlands) with a flow rate of 200 nl/min. The eluent gradient consisted of 0–60% buffer B for 60 min and then 60–80% buffer B for 10 min and then was maintained at 80% buffer B for 20 min (buffer A, 0.1% formic acid in 5% acetonitrile; buffer B, 0.1% formic acid in acetonitrile). Data were acquired on a QStar XL mass spectrometer (Applied Biosystems) using data-dependent acquisition in which every 9 s the instrument cycled through acquisition of a full-scan TOF mass spectrum (1 s), and three MS/MS spectra (2, 3, and 3 s) were recorded sequentially on the most abundant three ions present in the initial MS survey scan; the three most abundant multiply charged peptides (2+ to 4+) with threshold count above 10 in the MS scan with m/z between 400 and 2000 Da were selected for MS/MS.

Data Analysis—

The complete set of data files (*.wiff) from the iTRAQ experiments were searched against the mouse Swiss-Prot protein database (version uniprot_sprot_20070123) using ProteinPilot™ 2.0 software (Applied Biosystems, Software Revision 50861) with the exclusion of common contaminants as set by the software (42). The ProteinPilot software uses the Paragon™ algorithm to perform protein identification and Pro Group™ algorithm to perform a statistical analysis on the peptides found to determine the minimal set of confident protein identifications. Search parameters within ProteinPilot were set with trypsin cleavage specificity, methyl methanethiosulfate-modified cysteine as fixed modifications, biological modification “ID focus” settings, and a protein minimum confidence score of 95% (Detected Protein Threshold >95% (Unused ProtScore >1.3)). The features such as common modifications, substitutions, and cleavage events are built-in functions in the software (42). ProteinPilot software pooled data from all LC/MS runs for each iTRAQ experiment. For protein identification, this software calculates a percentage of confidence that reflects the probability that the hit is a false positive so that at the 95% confidence level there is a false positive identification rate of around 5% (42, 43). Although the software automatically accepts all peptides with a confidence of identification >1%, only proteins that had at least one peptide with >95% were initially recorded. These low confidence peptides therefore do not identify a protein by themselves but may support the presence of a protein identified using other peptides (42, 43).

The ratio of the areas under the signature peaks of 114 and 117 Da that are the masses of the tags that correspond to the iTRAQ reagents is used for relative quantification of each peptide. The ratio for each protein was calculated as a weighted average value combining averages of the peptide quantifications for this protein from all fractions; correction for experimental bias and shared peptides are excluded for protein quantification. The accuracy of each protein ratio is given by a calculated “error factor (EF)” in the software and a p value to assess whether the protein is significantly differentially expressed. The actual value for the average protein ratio is expected to be found between (reported average ratio) × (error factor) and (reported average ratio)/(error factor) 95% of the time. The error factor is calculated as follows: error factor = 10^{95% confidence error} where this 95% confidence error = S_MW × (Student's t factor for n − 1 degrees of freedom). S_MW is the weighted standard deviation of the weighted average of log ratios where n is the number of peptides contributing to protein relative quantification. The p value is determined by calculating Student's t factor where t = (weighted average of log ratios − log bias) divided by the weighted standard deviation, allowing determination of the p value with n − 1 degrees of freedom again where n is the number of peptides contributing to protein relative quantification. The calculation details can be obtained in the literature (42, 43) and ProteinPilot 2.0 software on-line help instructions (Applied Biosystems).

About 590 unique proteins were identified at 95% protein minimum confidence in our iTRAQ study. For each protein identification and quantification, the ProteinPilot software listed all peptides assigned to this protein, including redundant peptides, shared peptides, and peptides with no quantification because of low signals. Thereafter all proteins were verified manually to confirm the unique peptide number assigned to this protein. The quantitatively defined proteins were restricted to the proteins containing at least two unique peptides across two of three independent iTRAQ experiments (supplemental Table S1). For the selection of differentially expressed proteins we considered the following situation: 1) proteins must contain at least two unique high scoring peptides (peptide confidence >90%); 2) the proteins must have a p value <0.05 and EF <2 across two of three independent iTRAQ experiments; 3) in the case of the proteins with EF <2 but no p value due to containing low spectra number, we manually inspected the peptide ratio and the correspondent spectra using raw data (*.wiff) and Analyst QS 1.1 software (Applied Biosystems) for correctness; and 4) the final protein average ratio must meet a 1.3- or 0.75-fold cutoff. All identifications of differentially expressed proteins were manually inspected for correctness. A detailed list of proteins detected in this study together with their molecular sequence coverage and the number of unique peptides is provided in supplemental Tables S1 and S2. Data were further analyzed for protein subcellular location and functional cluster using the GoMiner program (44) based on the Gene Ontology Consortium.

Immunoblot Analysis

Western blot analysis was conducted as described previously (45). The primary antibodies used in this study included antibodies against BIP (GRP78, BD Transduction Laboratories; a gift from Dr. Allen Volchuk), KDEL (GRP94) and PDI (Stressgen, British Columbia, Canada; gifts from Dr. Allen Volchuk), VAMP2 (a gift from Dr. Herbert Y. Gaisano), PC1 and CPE (gifts from Dr. Savita Dhanvantari), PC2 (a gift from Dr. Nabil Seidah), GLUT2 (Chemicon International), and pyruvate carboxylase (PCX) (a gift from Dr. Brian Robinson). Proteins from freshly isolated islets were subjected to SDS-PAGE, transferred to polyvinylidene difluoride membranes (Fisher), and probed using the indicated antibodies. Protein loading was normalized to β-actin expression (Sigma). The chemiluminescent signals (PerkinElmer Life Sciences) were captured on film and quantified using the NIH software ImageJ.

RNA Extraction and Gene Expression Profile

Sample processing and microarray experiments were performed as described previously (45). Islets from 8–10 mice of each group were pooled yielding one sample to decrease eventual biases caused by biological variations. Islet RNA was extracted using TRIzol reagent (Invitrogen). The quality of total RNA was assessed using gel electrophoresis and an Agilent Bioanalyzer 2100 (Agilent Technologies) before labeling for hybridization to Affymetrix Mouse Genome 430 2.0 arrays (Affymetrix, Santa Clara, CA). Affymetrix GeneChip Operating Software (GCOS) 1.2 software (Affymetrix) was used to scan and quantitatively analyze the scanned image. Spotfire (Spotfire, Cambridge, MA) and Microsoft Excel software were also used for array data analysis. GCOS was used for absolute and comparison analysis to calculate signal values and to provide “detection” calls as “present,” “marginal,” or “absent” for each probe set. For estimation of regulated genes, pairwise comparisons of test versus control were performed, resulting in a quantitative signal log ratio and a qualitative change call. Signal log ratio is the logarithmic (base = 2) ratio of intensities from test and control samples. Detection calls are determined from statistical calculations of the difference in hybridization signals between perfect match oligonucleotides and their corresponding control mismatch sequence. Only genes/expressed sequence tags with a detection call of present in both MKR and WT triplicates in at least two of three independent experiments were subsequently used for comparison analysis.

Quantitative Real Time PCR

Quantitative real time PCR (qPCR) was performed as described previously (45). Primers were designed using Primer Express version 2.0 software (Applied Biosystems), and the sequences are listed in supplemental Table S3.

Comparison of mRNA and Protein Expression

To compare the protein changes with their corresponding mRNA levels, we mapped protein accession to its Swiss-Prot ID and then identified the corresponding Affymetrix Mouse Genome 430 2.0 probe set using NetAffx (Affymetrix) and its annotation tables (July 2006). Note that this automated mapping does not guarantee that every protein is mapped to a probe set ID. We excluded the proteins without gene probe in the Mouse Genome 430 2.0 arrays or those that gave an absent signal. A custom genome/proteome array was created for the identified proteins including accession number and annotation. For the mRNA and protein expression level correlation study, we compared the average protein ratios detected by the iTRAQ study with their corresponding gene ratios measured by microarray analyses in MKR islets. Gene/protein cluster analysis was performed using GoMiner programs (44) and Cluster 3.0 (46) and visualized using Treeview (47).

Statistical Analysis

Results are expressed as mean ± S.D. or S.E. Statistical significance was determined by two-tailed Student's t test. Correlations were determined using Pearson's correlation coefficient calculation with a two-tailed test of significance. A value of p <0.05 was considered significant.

RESULTS

MKR Mouse Phenotypic Analysis—

Ten-week-old MKR mice that have developed overt T2D and 3-week-old mice that are insulin-resistant were used in this study. Table I summarizes their biological characteristics. In young MKR animals (3 weeks of age), blood glucose concentrations were only 11% higher than control (p = 0.08), whereas insulin levels were 4-fold higher in MKR mice (p < 0.001), suggesting severe insulin resistance in these mice. Consistent with previous studies, 10-week-old MKR mice exhibited hyperglycemia, hyperinsulinemia, and a significantly leaner body weight compared with the WT controls (36, 38). At 3 weeks of age pancreatic histology revealed that MKR and WT mice had islets of similar size (Fig. 1, A and B). However, at 10 weeks of age, MKR islets were significantly larger, demonstrating β-cell hypertrophy and hyperplasia (Fig. 1, A and B).

Table I.

Ages, weights, and non-fasting blood glucose and insulin

Age	Weight	Blood glucose	Blood insulin
weeks	g	mM	ng/ml
WT
3 (n = 10)	26.0 ± 0.3	10.6 ± 0.7	0.8 ± 0.2
10 (n = 20)	31.7 ± 1.0	9.9 ± 0.5	1.9 ± 0.2
MKR
3 (n = 9)	19.7 ± 0.4a	11.8 ± 0.3	3.5 ± 1.4a
10 (n = 15)	23.3 ± 1.2b	18.5 ± 2.5b	36.0 ± 3.1b

^ap < 0.01 compared with age-matched WT mice.

^bp < 0.001 compared with age-matched WT mice.

Fig. 1.

Morphological characterization of MKR and WT pancreatic islets at 3 and 10 weeks of age. A, images of freshly isolated islets were taken with a confocal microscope. B, immunostaining for insulin (magnification, ×100) in pancreatic sections. C and D, glucose-stimulated insulin secretion from isolated islets was determined in response to 2.8 and 20 mm glucose (Error bars indicate the standard error of the mean calculated on the insulin secretion from three independent experiments with >5 mice per genotype). ***, p < 0.001.

Glucose homeostasis was assessed in vivo by a glucose tolerance test (38) and ex vivo by measuring insulin secretion from isolated islets. Ten-week-old MKR mice displayed glucose intolerance with loss of first phase insulin secretion, whereas 3-week-old MKR mice retained some first phase insulin response to glucose (38). Ex vivo MKR versus control islets showed a similar -fold GSIS at 3 weeks of age but secreted 51.1% less under high glucose conditions at 10 weeks of age (p < 0.001) (Fig. 1, C and D). Collectively these results demonstrated the presence of β-cell dysfunction in 10-week-old MKR mice with a significantly attenuated capacity to secrete insulin in the presence of high glucose.

Proteomics Profiling of MKR/WT Islets—

To investigate the molecular consequences of insulin resistance and T2D in MKR islets, we conducted a comprehensive proteomics analysis by determining the protein expression profile in freshly isolated islets of 10-week-old mice using a quantitative proteomics iTRAQ method (48). To compensate for the extreme sample complexity in islet protein levels, a batch of 60 fractions was separated per iTRAQ experiment using strong cation exchange chromatography. These fractions were then combined in 20–30 samples and analyzed by LC/MS/MS. A schematic flow of the iTRAQ proteomics approach is illustrated in Fig. 2A. Fig. 2, B and C, show the representative MS spectra for protein quantification and identification. A total of 590 unique islet proteins were identified with 95% confidence by the ProteinPilot search algorithm (Applied Biosystems) (42) against the mouse Swiss-Prot protein database (uniprot_sprot_20070123). To minimize incorrect protein quantification, we used stringent criteria (see “Experimental Procedures”) for selecting differentially expressed proteins. This filtering measure resulted in a final set of 159 differentially expressed proteins in MKR versus WT islets with an approximate -fold change of 1.3–4.3 in either direction. Of those, 92 proteins were increased and 67 were decreased in MKR islets (Table II).

Table II.

Differentially expressed proteins in MKR islets

159 proteins were differentially expressed in diabetic MKR islets versus controls. Categorical analysis was based on biological process functions in Gene Ontology by GoMiner (44). Some proteins are listed more than once because of multiple Gene Ontology annotations.

Swiss-Prot ID	Protein name	Gene symbol	Fold ± S.D. (MKR/WT)
			Protein ratioa	mRNA ratiob
Protein metabolism
Q9Z2W0	Aspartyl aminopeptidase	Dnpep	2.13 ± 0.41	1.42 ± 0.20
Q9R1P1	Proteasome subunit β type 3	Psmb3	1.65 ± 0.09	1.00 ± 0.12
Q9D8N0	Elongation factor 1-γ	Eef1g	1.49 ± 0.47	1.20 ± 0.05
Q9D8E6	60 S ribosomal protein L4	Rpl4	1.74 ± 1.09	1.07 ± 0.07
Q9D1Q6	Thioredoxin domain-containing protein 4 precursor	Txndc4	1.71 ± 0.24	1.39 ± 0.14
Q9D1M7	FK506-binding protein 11 precursor	Fkbp11	2.73 ± 0.69	2.30 ± 0.16
Q9CZX8	40 S ribosomal protein S19	Rps19	1.33 ± 0.06	1.00 ± 0.07
Q9CYN2	Signal peptidase complex subunit 2	Spcs2	1.37 ± 0.05	1.39 ± 0.21
Q99KV1	DnaJ homolog subfamily B member 11 precursor	Dnajb11	2.47 ± 0.79	1.80 ± 0.32
Q91YW3	DnaJ homolog subfamily C member 3	Dnajc3	3.04 ± 0.82	1.72 ± 0.36
Q91YQ5	Dolichyl-diphosphooligosaccharide–protein glycosyltransferase subunit 1 precursor	Rpn1	1.70 ± 1.06	1.32 ± 0.09
Q8R2E9	ERO1-like protein β precursor	Ero1lb	1.52 ± 0.28	1.38 ± 0.05
Q8BHN3	Neutral α-glucosidase AB precursor	Ganab	1.36 ± 0.10	1.12 ± 0.09
Q6P5E4	UDP-glucose:glycoprotein glucosyltransferase 1 precursor	Ugcgl1	1.39 ± 0.22	1.32 ± 0.00
Q6NSR8	Probable aminopeptidase NPEPL1	Npepl1	1.44 ± 0.23	1.07 ± 0.19
Q66JS6	Eukaryotic translation initiation factor 3 subunit 1	Eif3s1	2.12 ± 0.40	1.39 ± 0.21
Q01853	Transitional endoplasmic reticulum ATPase	Vcp	1.50 ± 0.17	1.15 ± 0.08
P99024	Tubulin β-5 chain	Tubb5	1.57 ± 0.43	1.07 ± 0.07
P63242	Eukaryotic translation initiation factor 5A-1	Eif5a	1.51 ± 0.20	1.19 ± 0.05
P62908	40 S ribosomal protein S3	Rps3	1.59 ± 0.95	1.10 ± 0.12
P62301	40 S ribosomal protein S13	Rps13	1.30 ± 0.14	0.96 ± 0.04
P62270	40 S ribosomal protein S18	Rps18	1.56 ± 0.09	1.10 ± 0.12
P61961	Ubiquitin-fold modifier 1 precursor	Ufm1	1.94 ± 0.51	1.53 ± 0.21
P59325	Eukaryotic translation initiation factor 5	Eif5	1.45 ± 0.22	1.10 ± 0.09
P57759	Endoplasmic reticulum protein ERp29 precursor	Erp29	2.40 ± 0.41	1.45 ± 0.16
P49722	Proteasome subunit α type 2	Psma2	1.46 ± 0.04	1.12 ± 0.09
P48024	Eukaryotic translation initiation factor1	Eif1	1.66 ± 0.22	0.96 ± 0.04
P47955	60 S acidic ribosomal protein P1	Rplp1	1.34 ± 0.16	1.07 ± 0.07
P53026	60 S ribosomal protein L10a	Rpl10a	1.34 ± 0.29	1.00 ± 0.07
P25444	40 S ribosomal protein S2	Rps2	1.38 ± 0.10	0.96 ± 0.10
P24369	Peptidyl-prolyl cis-trans isomerase B precursor	Ppib	1.61 ± 0.27	1.52 ± 0.11
P45878	FK506-binding protein 2 precursor	Fkbp2	1.69 ± 0.51	1.33 ± 0.15
P20029	78-kDa glucose-regulated protein precursor	Bip	2.25 ± 0.72	1.53 ± 0.27
P54775	26 S proteasome regulatory subunit 6B	Psmc4	1.40 ± 0.27	1.02 ± 0.04
P62245	40 S ribosomal protein S15a	Rps15a	1.61 ± 0.29	1.05 ± 0.04
P14211	Calreticulin precursor	Calr	1.63 ± 0.37	1.26 ± 0.10
P12970	60 S ribosomal protein L7a	Rpl7a	1.51 ± 0.13	1.05 ± 0.09
P10126	Elongation factor 1-α 1	Eef1a1	1.77 ± 0.43	1.10 ± 0.09
P09103	Protein-disulfide isomerase precursor	PDI	2.43 ± 0.51	1.42 ± 0.17
P08113	Endoplasmin precursor	Grp94	2.86 ± 0.89	1.56 ± 0.26
P08003	Protein-disulfide isomerase A4 precursor	Erp72	2.59 ± 0.69	1.91 ± 0.15
O70251	Elongation factor 1-β	Eef1b2	1.65 ± 0.71	1.23 ± 0.09
O08795	Glucosidase 2 subunit β precursor	Prkcsh	1.80 ± 0.25	1.45 ± 0.06
Q9Z1Q9	Valyl-tRNA synthetase	Vars	0.74 ± 0.03	—
Q9QXV0	Proprotein convertase subtilisin/kexin type 1 inhibitor	Pcsk1n	0.31 ± 0.15	0.74 ± 0.18
Q7TQI3	Ubiquitin thioesterase protein OTUB1	Otub1	0.68 ± 0.05	0.89 ± 0.10
Q00493	Carboxypeptidase E precursor	Cpe	0.59 ± 0.08	1.10 ± 0.09
P63239	Neuroendocrine convertase 1 precursor	Pc1	0.65 ± 0.20	1.20 ± 0.05
P62806	Histone H4	Hist1h4a	0.52 ± 0.19	—
P21661	Neuroendocrine convertase 2 precursor	Pc2	0.54 ± 0.14	0.92 ± 0.09
P17742	Peptidyl-prolyl cis-trans isomerase A	Ppia	0.71 ± 0.08	0.79 ± 0.03
P12961	Neuroendocrine protein 7B2 precursor (Secretogranin V)	Sgne1	0.56 ± 0.22	1.15 ± 0.00
Amino acid and derivative metabolism
Q9WVL0	Maleylacetoacetate isomerase	Gstz1	1.49 ± 0.17	1.03 ± 0.14
Q922W5	Pyrroline-5-carboxylate reductase 1	Pycr1	2.85 ± 0.62	2.98 ± 0.44
Q922Q4	Pyrroline-5-carboxylate reductase 2	Pycr2	1.52 ± 0.28	1.29 ± 0.05
P09103	Protein-disulfide isomerase precursor	PDI	2.43 ± 0.51	1.42 ± 0.17
P05202	Aspartate aminotransferase, mitochondrial precursor	Got2	0.69 ± 0.08	0.87 ± 0.21
Q9Z1Q9	Valyl-tRNA synthetase	Vars	0.74 ± 0.03	—
P26443	Glutamate dehydrogenase 1, mitochondrial precursor	Glud1	0.58 ± 0.17	0.87 ± 0.06
Carbohydrate metabolism
Q8K0C9	GDP-mannose 4,6-dehydratase	Gmds	2.32 ± 0.30	2.15 ± 0.15
Q8BHN3	Neutral α-glucosidase AB precursor	Ganab	1.36 ± 0.10	1.12 ± 0.09
P97807	Fumarate hydratase, mitochondrial precursor	Fh1	1.46 ± 0.39	0.94 ± 0.18
P70699	Lysosomal α-glucosidase precursor	Gaa	1.97 ± 0.33	1.05 ± 0.11
P16858	Glyceraldehyde-3-phosphate dehydrogenase	Gapdh	1.72 ± 0.68	0.98 ± 0.08
P17182	α-Enolase	Eno1	2.35 ± 0.96	1.27 ± 0.22
P14152	Malate dehydrogenase, cytoplasmic	Mdh1	1.31 ± 0.16	1.38 ± 0.05
P00687	α-Amylase 1 precursor	Amy1	1.35 ± 0.29	0.81 ± 0.23
O88844	Isocitrate dehydrogenase (NADP) cytoplasmic	Idh1	1.41 ± 0.23	1.26 ± 0.10
Q9CZU6	Citrate synthase, mitochondrial precursor	Cs	0.64 ± 0.06	0.98 ± 0.04
Q9ERT9	Protein phosphatase inhibitor 1	Ppp1r1a	0.59 ± 0.24	0.66 ± 0.09
Q64521	Glycerol-3-phosphate dehydrogenase, mitochondrial precursor	Gpd2	0.70 ± 0.05	0.64 ± 0.15
Q60597	2-Oxoglutarate dehydrogenase E1 component, mitochondrial precursor	Ogdh	0.73 ± 0.08	0.89 ± 0.04
P01326	Insulin-2 precursor	Ins2	0.74 ± 0.29	0.93 ± 0.06
P01325	Insulin-1 precursor	Ins1	0.50 ± 0.13	0.96 ± 0.04
Q05920	Pyruvate carboxylase, mitochondrial precursor	Pcx	0.43 ± 0.15	0.59 ± 0.08
P54071	Isocitrate dehydrogenase (NADP), mitochondrial precursor	Idh2	0.64 ± 0.10	0.76 ± 0.05
P14246	Solute carrier family 2, facilitated glucose transporter member 2	Slc2a2	0.42 ± 0.13	0.76 ± 0.00
Lipid metabolism
Q91V92	ATP-citrate synthase	Acly	1.64 ± 0.75	0.78 ± 0.13
Q8VDJ3	Vigilin	Hdlbp	1.39 ± 0.29	1.73 ± 0.38
Q99JY0	Trifunctional enzyme subunit β, mitochondrial precursor	Hadhb	0.62 ± 0.04	0.87 ± 0.06
Q61425	Hydroxyacyl-coenzyme A dehydrogenase, mitochondrial precursor	Hadhsc	0.63 ± 0.09	0.81 ± 0.06
Q05920	Pyruvate carboxylase, mitochondrial precursor	Pcx	0.43 ± 0.15	0.59 ± 0.08
Nucleobase, nucleoside, nucleotide, and nucleic acid metabolism
Q99KV1	DnaJ homolog subfamily B member 11 precursor	Dnajb11	2.47 ± 0.79	1.80 ± 0.32
Q8K0C9	GDP-mannose 4,6-dehydratase	Gmds	2.32 ± 0.30	2.15 ± 0.15
Q01768	Nucleoside-diphosphate kinase B	Nme2	1.33 ± 0.38	1.29 ± 0.11
P63028	Translationally controlled tumor protein	Tpt1	1.36 ± 0.12	—
P62821	Ras-related protein Rab1A	Rab1	1.48 ± 0.11	1.36 ± 0.15
P50096	Inosine-5`-monophosphate dehydrogenase 1	Impdh1	2.75 ± 0.32	1.59 ± 0.06
Q9Z1N5	Spliceosome RNA helicase Bat1	Bat1a	0.72 ± 0.09	0.99 ± 0.15
Q9D7G0	Ribose-phosphate pyrophosphokinase I	Prps1	0.61 ± 0.21	0.98 ± 0.04
Q60668	Heterogeneous nuclear ribonucleoprotein D0	Hnrpd	0.74 ± 0.10	0.87 ± 0.06
Q9Z1Q9	Valyl-tRNA synthetase	Vars	0.74 ± 0.03	—
P67778	Prohibitin	Phb	0.54 ± 0.08	0.85 ± 0.09
P63158	High mobility group protein B1	Hmgb1	0.71 ± 0.25	0.96 ± 0.07
P62806	Histone H4	Hist1h4a	0.52 ± 0.19	—
P61979	Heterogeneous nuclear ribonucleoprotein K	Hnrpk	0.74 ± 0.15	0.98 ± 0.10
P60041	Somatostatin precursor	Sst	0.29 ± 0.04	0.59 ± 0.05
Electron/proton transport, ATP synthesis
Q9DCW4	Electron transfer flavoprotein subunit β	Etfb	0.66 ± 0.05	1.15 ± 0.08
Q9DC69	NADH dehydrogenase	Ndufa9	0.54 ± 0.05	0.96 ± 0.04
Q99LC5	α-ETF	Etfa	0.57 ± 0.06	0.82 ± 0.16
Q8BW75	Monoamine oxidase type B	Maob	0.33 ± 0.07	0.71 ± 0.00
Q8BVE3	Vacuolar ATP synthase subunit H	Atp6v1h	0.48 ± 0.06	0.80 ± 0.14
Q8BH59	Calcium-binding mitochondrial carrier protein Aralar1	Slc25a12	0.60 ± 0.09	0.90 ± 0.13
P99028	Ubiquinol-cytochrome c reductase complex 11-kDa protein, mitochondrial precursor	Uqcrh	0.66 ± 0.12	1.13 ± 0.17
P62814	Vacuolar ATP synthase subunit B, brain isoform	Atp6v1b2	0.54 ± 0.05	0.87 ± 0.00
P56135	ATP synthase f chain, mitochondrial	Atp5j2	0.75 ± 0.09	1.05 ± 0.09
P51881	ADP/ATP translocase 2	Slc25a5	0.75 ± 0.14	1.15 ± 0.08
P48962	ADP/ATP translocase 1	Slc25a4	0.63 ± 0.12	1.00 ± 0.07
P19783	Cytochrome c oxidase subunit 4 isoform 1, mitochondrial precursor	Cox4i1	0.70 ± 0.27	0.96 ± 0.07
P12787	Cytochrome c oxidase subunit 5A, mitochondrial precursor	Cox5a	0.69 ± 0.22	0.91 ± 0.04
P09671	Superoxide dismutase (manganese), mitochondrial precursor	Sod2	0.75 ± 0.13	0.91 ± 0.07
P00405	Cytochrome c oxidase subunit 2	mt-Co2	0.66 ± 0.15	—
Transport
Q9DBH5	Vesicular integral membrane protein VIP36 precursor	Lman2	1.44 ± 0.12	1.07 ± 0.00
Q9R0Q3	Transmembrane emp24 domain-containing protein 2 precursor	Tmed2	2.12 ± 0.42	1.45 ± 0.60
Q9D662	Protein transport protein Sec23B	Sec23b	1.63 ± 0.78	1.24 ± 0.17
Q9DC16	Endoplasmic reticulum-Golgi intermediate compartment protein 1	1200007D18Rik	1.70 ± 0.72	1.94 ± 0.43
Q9D1D4	Transmembrane emp24 domain-containing protein 10 precursor	Tmed10	1.64 ± 0.26	1.33 ± 0.16
Q9D0F3	ERGIC-53 protein precursor	Lman1	1.47 ± 0.35	1.40 ± 0.30
Q9CXE7	Transmembrane emp24 domain-containing protein 5 precursor	Tmed5	2.55 ± 0.54	1.43 ± 0.27
Q99PL5	Ribosome-binding protein 1	Rrbp1	1.83 ± 0.15	2.44 ± 0.51
Q8VDJ3	Vigilin	Hdlbp	1.39 ± 0.29	1.73 ± 0.38
Q8R2E9	ERO1-like protein β precursor	Ero1lb	1.52 ± 0.28	1.38 ± 0.05
Q8R1V4	Transmembrane emp24 domain-containing protein 4 precursor	Tmed4	1.86 ± 0.62	1.18 ± 0.05
Q01853	Transitional endoplasmic reticulum ATPase	Vcp	1.50 ± 0.17	1.15 ± 0.08
P99024	Tubulin β-5 chain	Tubb5	1.57 ± 0.43	1.07 ± 0.07
P70670	Nascent polypeptide-associated complex subunit α, muscle-specific form	Naca	1.69 ± 0.24	0.96 ± 0.04
P62897	Cytochrome c, somatic	Cycs	1.35 ± 0.03	1.45 ± 0.11
P62821	Ras-related protein Rab1A	Rab1	1.48 ± 0.11	1.36 ± 0.15
P50396	Rab GDP dissociation inhibitor α	Gdi1	1.48 ± 0.10	0.93 ± 0.06
P10639	Thioredoxin	Txn1	1.53 ± 0.28	1.39 ± 0.14
O55029	Coatomer subunit β`	Copb2	1.40 ± 0.47	1.29 ± 0.13
P35700	Peroxiredoxin-1	Prdx1	1.34 ± 0.28	1.29 ± 0.11
P01326	Insulin-2 precursor	Ins2	0.74 ± 0.29	0.93 ± 0.06
P01325	Insulin-1 precursor	Ins1	0.50 ± 0.13	0.96 ± 0.04
Q99LC5	Electron transfer flavoprotein subunit α, mitochondrial precursor	Etfa	0.57 ± 0.06	0.82 ± 0.16
Q9R0Q1	Synaptotagmin-like protein 4	Sytl4	0.59 ± 0.18	0.61 ± 0.10
Q9DCW4	Electron transfer flavoprotein subunit β	Etfb	0.66 ± 0.05	1.15 ± 0.08
Q8VEM8	Phosphate carrier protein, mitochondrial precursor (PTP)	Slc25a3	0.70 ± 0.06	1.03 ± 0.11
Q8BVE3	Vacuolar ATP synthase subunit H	Atp6v1h	0.48 ± 0.06	0.80 ± 0.14
Q8BH59	Calcium-binding mitochondrial carrier protein Aralar1	Slc25a12	0.60 ± 0.09	0.90 ± 0.13
Q64331	Myosin-6	Myo6	0.52 ± 0.05	0.65 ± 0.10
P99028	Ubiquinol-cytochrome c reductase complex 11-kDa protein, mitochondrial precursor	Uqcrh	0.66 ± 0.12	1.13 ± 0.17
P84086	Complexin-2	Cplx2	0.67 ± 0.03	0.73 ± 0.10
P63158	High mobility group protein B1	Hmgb1	0.71 ± 0.25	0.96 ± 0.07
P63044	Vesicle-associated membrane protein 2	Vamp2	0.59 ± 0.09	1.07 ± 0.07
P56135	ATP synthase f chain, mitochondrial	Atp5j2	0.75 ± 0.09	1.05 ± 0.09
P62814	Vacuolar ATP synthase subunit B, brain isoform	Atp6v1b2	0.54 ± 0.05	0.87 ± 0.00
P63011	Ras-related protein Rab3A	Rab3a	0.69 ± 0.04	1.10 ± 0.09
P51881	ADP/ATP translocase 2	Slc25a5	0.75 ± 0.14	1.15 ± 0.08
P48962	ADP/ATP translocase 1	Slc25a4	0.63 ± 0.12	1.00 ± 0.07
P14246	Solute carrier family 2, facilitated glucose transporter member 2	Slc2a2	0.42 ± 0.13	0.76 ± 0.00
P12961	Neuroendocrine protein 7B2 precursor	Sgne1	0.56 ± 0.22	1.15 ± 0.00
O55143	Sarcoplasmic/endoplasmic reticulum Ca-ATPase 2	Atp2a2	0.63 ± 0.26	1.00 ± 0.26
Cell development/cycle/proliferation/death
Q64727	Vinculin	Vcl	1.53 ± 0.19	1.15 ± 0.00
Q01853	Transitional endoplasmic reticulum ATPase	Vcp	1.50 ± 0.17	1.15 ± 0.08
P63028	Translationally controlled tumor protein	Tpt1	1.36 ± 0.12	—
P63242	Eukaryotic translation initiation factor 5A-1	Eif5a	1.51 ± 0.20	1.19 ± 0.05
P62897	Cytochrome c, somatic	Cycs	1.35 ± 0.03	1.45 ± 0.11
P50096	Inosine-5`-monophosphate dehydrogenase 1	Impdh1	2.75 ± 0.32	1.59 ± 0.06
P27773	Protein-disulfide isomerase A3 precursor	Pdia3	1.62 ± 0.35	1.40 ± 0.24
P14211	Calreticulin precursor	Calr	1.63 ± 0.37	1.26 ± 0.10
Q64331	Myosin-6	Myo6	0.52 ± 0.05	0.65 ± 0.10
P60041	Somatostatin precursor	Sst	0.29 ± 0.04	0.59 ± 0.05
Other functions
Q99PL5	Ribosome-binding protein 1	Rrbp1	1.83 ± 0.15	2.44 ± 0.51
Q9D819	Inorganic pyrophosphatase (pyrophosphate phosphohydrolase)	Pyp	1.73 ± 0.23	1.55 ± 0.12
Q3UHX2	28-kDa heat- and acid-stable phosphoprotein	Pdap1	1.31 ± 0.15	1.01 ± 0.14
P81117	Nucleobindin-2 precursor	Nucb2	2.25 ± 0.41	2.47 ± 0.29
Q9EPS2	Peptide YY precursor	Pyy	0.48 ± 0.22	0.38 ± 0.04
Q91WD9	Secretagogin	Scgn	0.73 ± 0.09	0.78 ± 0.08
Q7TMM9	Tubulin β-2A chain	Tubb2a	0.72 ± 0.07	—
P55095	Glucagon precursor	Gcg	0.23 ± 0.10	0.91 ± 0.04
Q03517	Secretogranin-2 precursor	Scg2	0.70 ± 0.21	1.18 ± 0.05
P48036	Annexin A5	Anxa5	0.70 ± 0.15	1.00 ± 0.07
P15626	Glutathione S-transferase Mu2	Gstm2	0.73 ± 0.14	0.60 ± 0.06
Unknown function
Q9CXI5	ARMET protein precursor	Armet	1.81 ± 0.54	1.46 ± 0.20
Q9CYA0	Cysteine-rich with EGF-like domain protein 2 precursor	5730592L21Rik	3.54 ± 1.32	2.71 ± 0.29
Q9QXT0	MIR-interacting saposin-like protein precursor	Tmem4	1.80 ± 0.40	1.45 ± 0.16
Q922Q8	Leucine-rich repeat-containing protein 59	Lrrc59	1.67 ± 0.41	1.45 ± 0.11
Q91VW5	Golgin subfamily A member 4	Golga4	1.86 ± 0.04	1.15 ± 0.08
Q61543	Golgi apparatus protein 1 precursor	Glg1	1.30 ± 0.04	0.94 ± 0.13
Q68FD5	Clathrin heavy chain	Cltc	1.48 ± 0.71	0.87 ± 0.11
P63325	40 S ribosomal protein S10	Rps10	1.39 ± 0.16	1.03 ± 0.08
P62852	40 S ribosomal protein S25	Rps25	1.53 ± 0.19	1.05 ± 0.04
P26516	26 S proteasome non-ATPase regulatory subunit 7	Psmd7	1.47 ± 0.25	1.15 ± 0.14
Q62465	Synaptic vesicle membrane protein VAT-1 homolog	Vat1	0.72 ± 0.03	0.98 ± 0.10
Q8VCM8	Nicalin precursor	Ncln	0.58 ± 0.06	1.06 ± 0.23
Q8CAQ8	Mitochondrial inner membrane protein	Immt	0.61 ± 0.16	0.98 ± 0.10
Q04447	Creatine kinase B-type	Ckb	0.55 ± 0.18	1.18 ± 0.09
P47867	Secretogranin-3 precursor	Scg3	0.37 ± 0.14	1.00 ± 0.07
P26339	Chromogranin A precursor	Chga	0.31 ± 0.16	0.96 ± 0.10
P16014	Secretogranin-1 precursor	Chgb	0.56 ± 0.19	1.07 ± 0.00
P21460	Cystatin-C precursor	Cst3	0.58 ± 0.06	1.02 ± 0.08
O08553	Dihydropyrimidinase-related protein 2	Dpysl2	0.73 ± 0.17	0.68 ± 0.09

^aProtein ratio measured by iTRAQ method.

^bmRNA ratio measured by microarray study. —, no probe or absent signal in microarray studies (refer to “Experimental Procedures”).

Clustering analysis by GoMiner (44) based on Gene Ontology nomenclature revealed that the highest proportion of changed proteins was located in mitochondria (18.9%) (Fig. 3A). Proteins located in the extracellular region (18.4%) and endoplasmic reticulum (16.9%) were the two next largest groups differentially regulated in diabetic MKR islets (Fig. 3A). Categorical analysis based on molecular function revealed that the majority of changed proteins in diabetic islets were associated with protein binding and catalyst activity (Fig. 3B). Clustering analysis based on biological process revealed that proteins involved in primary metabolism and transport constituted the largest functional groups, comprising about 54.1 and 25.8%, respectively, of all of the differentially expressed proteins in the diabetic islets (Table II). Further detailed functional analysis of changed Gene Ontology terms demonstrated that the largest fraction of up-regulated proteins was associated with protein biosynthesis and folding. In contrast, the down-regulated proteins were mainly associated with insulin processing/secretion and energy metabolism, particularly mitochondrial oxidative metabolism (Table II). Fig. 3, C–F, list the representative differentially expressed proteins related to protein biosynthesis, protein folding, insulin secretion, and mitochondrial oxidative function in diabetic 10-week-old MKR islets.

Fig. 3.

Functional categorization and relative protein ratios of differentially expressed proteins in MKR islets. Differentially expressed proteins in diabetic MKR islets were sorted into subcellular location (A) and functional categories (B). Relative changes in the levels of proteins in MKR islets related to protein synthesis (C), ER stress (D), secretion (E), and mitochondrial defects (F).

Confirmation of iTRAQ by Western Blotting Study—

To provide confirmation of differentially expressed proteins, we performed Western blot analysis of selected proteins detected by the iTRAQ study. Western blot analysis was performed for nine proteins that were chosen to represent different metabolic pathways (protein folding/ER stress, glucose metabolism, and insulin processing and secretion pathway) (Tables II and III) and different -fold and directional changes (three up-regulated and six down-regulated). Fig. 4 shows the representative Western blot images with the quantification, and Table III lists the comparative summary of protein ratios detected by Western blot and iTRAQ. Collectively we observed a positive correlation for the direction of changes (Table III). For example, six tested proteins, GLUT2, VAMP2, PC1, PC2, CPE, and PCX, that were predicted to be down-regulated in MKR islets by the iTRAQ study were also significantly decreased in Western blot analyses. Additionally up-regulated proteins PDI, BIP, and GRP94 were also consistently increased in Western blot analyses. The corroboration by Western blotting provides evidence that the amine-specific isobaric tagging labeling method for the large scale protein quantification was reliable.

Table III.

Comparison of protein and mRNA ratio measured by iTRAQ/Western blot and microarray/qPCR studies

Gene symbol	-Fold (MKR/WT) ± S.D.
	Protein ratio		mRNA ratio
	iTRAQ	Western blot	Microarray	qPCR
Pcx	0.43 ± 0.15	0.72 ± 0.08a	0.59 ± 0.08	0.70 ± 0.09a
Vamp2	0.59 ± 0.05	0.53 ± 0.06b	1.07 ± 0.07	1.12 ± 0.12
PDI/Erp59	2.43 ± 0.51	1.80 ± 0.19c	1.42 ± 0.17	1.34 ± 0.26a
Slc2a2/Glut2	0.42 ± 0.13	0.56 ± 0.03b	0.76 ± 0.00	0.61 ± 0.15c
Grp94	2.86 ± 0.89	3.10 ± 0.54c	1.56 ± 0.26	1.76 ± 0.27c
Bip/Grp78	2.25 ± 0.72	3.14 ± 0.72c	1.53 ± 0.27	1.55 ± 0.16b
Pc1	0.65 ± 0.20	0.53 ± 0.06b	1.20 ± 0.05	0.94 ± 0.16
Pc2	0.54 ± 0.14	0.72 ± 0.10a	0.92 ± 0.09	0.98 ± 0.31
Cpe	0.59 ± 0.08	0.68 ± 0.08c	1.10 ± 0.09	0.86 ± 0.16
Slc25a4/Ant1	0.63 ± 0.12	d	1.00 ± 0.07	1.01 ± 0.18
Sytl4	0.59 ± 0.18	d	0.61 ± 0.10	0.63 ± 0.02a
Maob	0.33 ± 0.07	d	0.71 ± 0.00	0.62 ± 0.06a

^ap < 0.05 compared with age-matched WT mice.

^bp < 0.001 compared with age-matched WT mice.

^cp < 0.01 compared with age-matched WT mice.

^dnot measured.

Fig. 4.

Representative Western blotting images and quantification for differentially expressed proteins in MKR versus WT control islets. There was good correlation between iTRAQ and Western blot data, and this information is presented in Table III (n = 3–5 independent experiments with >5 mice per genotype). **, p < 0.01; ***, p < 0.001.

Correlation between mRNA and the Protein Level in the Detected Islet Proteins—

We sought to compare the changes in expression at the protein level with changes at the mRNA level in MKR versus WT islets. A comparative genome-wide analysis of transcripts from freshly isolated islets of 10-week-old mice was performed using Affymetrix Mouse Genome 430 2.0 arrays. Based on the calculation by GCOS and the filter set (see “Experimental Procedures”), ∼854 genes/expressed sequence tags showed differential expression in MKR versus WT islets (p < 0.05). The accuracy of microarray results was confirmed by performing qPCR analysis for a set of 100 selected genes. Table III lists the summary of the qPCR results in comparison with those detected by the microarray study. Additional details of the analytical studies including the classification of altered gene families and the validation of microarray results via qPCR will be reported elsewhere.

Using Swiss-Prot ID and Affymetrix net support, we cross-referenced the iTRAQ and microarray data sets using their respective gene product identifiers. 154 of the differentially expressed proteins were able to be mapped to probe IDs on the microarray. The five unmatched proteins either had no corresponding gene probe or gave an absent signal on the microarray (Table II). Fig. 5, A and B, show protein ratios detected by the iTRAQ study versus the corresponding mRNA expression ratios obtained by DNA microarrays. Overall the protein changes of MKR versus WT islets were moderately correlated with the corresponding mRNA (r = 0.72, p < 3.5 × 10⁻²⁶) (Fig. 5B). Fig. 5C graphically shows that about 45.2% of the differentiated proteins showed concordant changes in mRNA (i.e. changes in the same direction), 0.6% were discordant (i.e. having higher protein expression but lower mRNA expression), and notably 54.2% showed changes in the proteome but not in the transcriptome. The proteins involved in protein proteolysis, modification, and biosynthesis and mitochondrial oxidative metabolism were mainly included in the latter portion, suggesting that post-transcriptional or translational mechanisms are involved in the regulation of expression of these proteins.

Fig. 5.

Correlation of mRNA ratios and protein levels of differentially expressed proteins in MKR islets. A and B, scatter plots of average protein ratios determined by the iTRAQ method and mRNA ratios determined by the microarray study. C, a pictorial comparison of changed protein ratios detected by iTRAQ and microarray analysis together with functional cluster analysis. Hierarchical clustering was performed using the GoMiner program (44) based on the biological process category in the Gene Ontology Consortium. Colors represent average gene/protein expression changes (MKR/WT) relative to the median (46) with red and green representing an increase or decrease in fold expression, respectively. “iTRAQ” and “MA” represent protein -fold and gene -fold (MKR/WT), respectively.

DISCUSSION

When pancreatic β-cells fail to compensate for the increased metabolic demands associated with insulin resistance, hyperglycemia and T2D ensue (11–13). The exact mechanisms underlying β-cell failure in most forms of T2D remain to be characterized (14, 15). Our studies using the MKR mice demonstrate that these animals are hyperinsulinemic and hyperglycemic with impaired first phase insulin secretion upon a glucose challenge by 10 weeks of age. In addition, islet hyperplasia and hypertrophy developed in MKR pancreata, reflecting an effort by the islets to compensate for progressive insulin resistance and hyperglycemia. We propose that the molecular defects in MKR β-cells that impair their ability to respond to glucose are multifactorial and can be revealed by global proteomics and gene profiling strategies. To examine this hypothesis, we applied a recently developed quantitative proteomics strategy, iTRAQ (48), to profile diabetic MKR versus control islets. The advantage of this approach stems from the unique sample preparation method achieved by jointly performing stable isotope tag labeling, ion exchange, and reverse-phase chromatography (at the peptide level) with automated procedures for peptide selection and fragmentation via tandem mass spectrometry. This combination makes it possible to provide a quantitative comparison of hundreds of proteins from different metabolic states (48), such as diabetic versus non-diabetic. Based on the stringent criteria, in the present study, 159 islet proteins were identified to be differentially expressed in diabetic MKR islets with high confidence. Western blot validation of a selected group of proteins detected by the iTRAQ method showed good agreement, demonstrating that the iTRAQ-based quantitative proteomics approach is a feasible method to compare islet protein expression profiles under different physiological and pathophysiological conditions.

Using an integrated approach, we were able to compare protein abundance ratios with their corresponding mRNA levels determined by gene profiling. A total of 154 mouse islet protein-mRNA pairs were mapped in this study (Fig. 5 and Table II). Collectively a moderate correlation between protein ratios and mRNA expression was observed in MKR islets (r = 0.72, p < 3.5 × 10⁻²⁶). This is higher than two previous reports on yeast (48, 49) and lung cancer (50) that showed poor correlations of differentiated proteins and mRNA in disturbed/diseased states. It is interesting to note that about 54% of differentiated proteins in MKR islets showed changes in the proteome but not in the transcriptome, suggesting possible post-transcriptional regulation. For example, the proteins involved in the processing of proinsulin, including CPE, PC1/3, and PC2, were down-regulated in MKR islets in the iTRAQ study and Western blot analysis. However, the microarray study and qPCR validation revealed that their mRNA levels did not change. This observation is similar to a study in MIN6 β-cells exposed to FFA where only the protein levels of PC1/3 and PC2 changed (51). These results underscore the importance of integrated mRNA and protein expression measurements for understanding the complex mechanisms of transcriptional control in T2D.

In our study, functional cluster analysis of differentially expressed proteins in MKR islets using GoMiner (44) revealed that many biological processes were altered depicting a polygenic and multipathway disease with complex metabolic disturbances. The differentially expressed proteins in diabetic islets may reflect the severe insulin resistance (primary alteration) or the complication of hyperglycemia and other metabolic factors (secondary effects). Consistent with previous literature, some of the genes identified here have been linked to T2D; however, the majority were shown to be associated with islet dysfunction for the first time. We herein summarize some key proteins that were disturbed in diabetic MKR islets.

Protein Biosynthesis, Folding, and Degradation—

The largest group of differentially expressed proteins in the MKR islets were those involved in protein metabolism, containing 28.9% of the total changed proteins (Table II). Several components of the protein synthesis machinery, including eukaryotic initiation factors (EIFs) (52) and elongation factors (eEF1s) (53, 54) along with ribosomal proteins, were significantly up-regulated in MKR islets. EIFs play a key role in initiation of translation by binding to the 40 S subunit to prevent the reassociation of the 60 S to the 40 S subunit before formation of the 43 S preinitiation complex and by stabilizing the Met-tRNA, eIF2, and the GTP ternary complex (55, 56). GTP-dependent elongation factors eEF1s are also required for translation and mediate the binding of the cognate aminoacyl-tRNA to the A-site of the ribosome and its subsequent release (57). It has been reported that both initiation and elongation can be controlled by insulin (58). However, these proteins exhibited significantly discordant changes between the protein and mRNA levels in our study, implying a post-translational or post-transcriptional regulatory mechanism (59).

One prominent group of up-regulated proteins in MKR islets is that related to protein folding and ER stress. ER stress leads to accumulation of unfolded proteins in the ER (60, 61), which in turn evokes the unfolded protein response (UPR) (62). In diabetic models the β-cell has significantly increased ER activity and stress because of the increased demand of the peripheral tissues for insulin to prevent hyperglycemia. In this study we found that two major components of the ER stress response pathway, UPR and ER-associated degradation (ERAD), were highly activated in 10-week-old diabetic MKR islets. One group includes the peptide-binding molecular chaperones BIP/GRP78, GRP94 (61, 63), and calreticulin (64). BIP/GRP78 and GRP94 interact transiently with protein folding intermediates to prevent aggregation of a protein by keeping it in a folding-competent state (61, 63, 65). Interaction between the chaperones and proteins ensures that only proteins that are properly assembled and folded leave the ER compartment and thus alleviate the threat of cell death (61). The significant up-regulation of BIP and GRP94 at the protein level was reported in diabetic “Akita” mouse pancreatic islet β-cell lines (66). The second group that was up-regulated in diabetic MKR islets included members of the disulfide and peptidyl-prolyl isomerase families, which catalyze the rearrangement of disulfide bonds and isomerization of peptide bonds around Pro residues (67). The expression of the disulfide isomerase family, including PDI (ERP59), PDIA4 (ERP72), TXNDC4 (ERP44), and ERP29 (63, 68, 69), was increased 2–3-fold in MKR islets. These are involved in protein folding by functioning as oxidoreductases in the formation/isomerization of disulfide bonds and thereby increase the rate at which proteins attain their final folded conformation (63, 68). Several peptidyl-prolyl isomerase proteins, which catalyze the folding of proline-containing polypeptides (70), including PPIB, FKBP11, and FKBP2, were highly up-regulated in MKR islets. Under in vitro conditions, proline cis-trans isomerization may become rate-limiting in the folding of proteins (67). The third up-regulated protein group in MKR islets was the HSP40 family, including DNAJC3 and DNAJB11. The DnaJ family was proposed to bind to nascent polypeptides to prevent their premature folding and may catalyze protein disulfide formation, reduction, and isomerization due to an active dithiol/disulfide group (67, 71). DNAJC3-null mice exhibit pancreatic β-cell failure and diabetes (72), and the up-regulation of DNAJC3 at the protein level was recently reported in human T2D islets (73). However, a change in DNAJB11 in diabetic mouse islets was reported for the first time here. Interestingly all these UPR proteins highlighted above were up-regulated both in protein and mRNA levels in diabetic MKR islets.

Proteins involved in the ERAD pathway were increased in MKR islets, including the proteasome 26 S family, VCP/p97, and ubiquitin-fold modifier 1 (UFM1). VCP/p97 is a hexameric ATPase of the AAA (ATPases associated with various cellular activities) family that mediates numerous and diverse cellular functions, including ERAD via the ubiquitin-proteosome system (74). Loss of VCP causes polyubiquitinated cellular proteins to accumulate, indicating an impaired ability to present them to the proteosome (75). VCP has also been reported to be an apoptotic regulator and an essential target of Akt signaling and is now shown here to be associated with a model of T2D for the first time (76). UFM1 was recently identified in HEK293 cells and mouse tissues (77) as one of various ubiquitin-like modifiers to target proteins in cells through UBA5 (E1) and UFC1 (E2) (77). This protein was suggested to function with a unique set of alternate ubiquitin-conjugating enzyme complexes, although the role of this protein is not known (77). Except for UFM1, this group of proteins exhibited increased protein expression with a negligible change in transcript levels. The up-regulation of the Ufm1 gene in ER stress was reported in inflammation-induced heart disease (78). Collectively the up-regulation of proteins involved in protein biosynthesis, UPR, and ERAD may ultimately lead to defective insulin secretion.

Insulin Secretion Defects—

In contrast to the up-regulation of proteins involved in protein biosynthesis, folding, and degradation, a significant down-regulation of proteins involved in insulin processing and secretion was observed in MKR islets (Table II). Insulin production is regulated primarily by glucose at the level of preproinsulin mRNA translation (79). Processing of proinsulin in the insulin secretory granule (ISG) yields the soluble, functional insulin hormone, which is regulated by the prohormone convertases 1/3 and 2 in concert with CPE. Stressing the biosynthesis and post-translational processing of prohormone convertases PC1/3 and PC2 leads to impaired proinsulin processing in T2D (80, 81). In our study, PC1/3, PC2, and CPE decreased by ∼35–46% in MKR islets without a change in their mRNA levels. This finding is similar to a study using the β-cell line MIN6 exposed to FFAs where there was post-transcriptional regulation of PC1/3 and PC2 (51). We also observed a 60–70% decrease in somatostatin (Sst) and glucagon (Gcg) expression in MKR islets in agreement with literature showing that mice with a disruption of PC1/3 or PC2 gene have a number of peptide processing defects including the processing of proinsulin, proglucagon, and prosomatostatin (82, 83). The abnormal processing of prohormones with incomplete conversion to hormones could result in disturbance of the proper functioning of the ISG that exacerbates protein misfolding and aggregation, which may in turn contribute to ER stress and impaired insulin secretion in MKR islets. On the other hand, down-regulation of several proteins of the granin family (Secretogranins I–III and V), one of the mediators of the regulated secretory pathway (84–86), was observed in MKR islets. The aggregation of these proteins induced by the millimolar concentration of calcium ions and an acidic pH may facilitate the condensation of regulated secretory proteins leading to the formation of dense core materials (84, 87).

MKR islets also displayed marked down-regulation in the expression of proteins involved in the regulation of vesicle trafficking and exocytosis. Pancreatic β-cells respond to increased circulating glucose levels by secretion of insulin from storage granules in a biphasic manner: first phase secretion is attributed to the fusion and release of insulin from granules clustered at the cell surface, whereas the second phase entails the mobilization and trafficking of intracellular storage pools of ISGs to the plasma membrane (88, 89). Fusion of ISGs is known to be regulated by soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) protein complexes at the plasma membrane (90, 91). Several SNARE complex-related proteins including VAMP2, RAB3, CPLX2, and synaptotagmin-like protein-4/granuphilin (SYTL4) were down-regulated in MKR islets. The importance of VAMP2 in the regulation of insulin secretion in β-cells was shown previously as the formation of a heterotrimeric SNARE core complex with syntaxin and SNAP-25 that was able to facilitate the fusion of the ISG with the plasma membrane and release of insulin (92). Furthermore cleavage of VAMP2 abolishes all insulin secretion (93, 94). Diminished expression of VAMP2 was reported in human T2D subjects (95) and other diabetic rodent models such as in Zucker fa/fa and Goto-Kakisaki rats (96, 97). Complexin (CPLX2), also known as synaphin, was originally linked to neurotransmitter release but was also reported to bind to syntaxin 1. Interestingly both overexpression and silencing of the complexin gene impaired stimulus-induced insulin secretion, and so its down-regulation in MKR islets could be linked to their impaired insulin secretion (98). Similarly overexpression of wild-type or mutant Rab3 reduced C-peptide secretion (99), and its expression was also decreased in MKR islets. Rab3 is thought to facilitate the dissociation of munc18 and syntaxin 1, thereby enabling interaction between the SNARE proteins (100). More recently, Rab3 via an interaction with calmodulin has been suggested to initiate the transportation of secretory granules to a readily releasable pool at the plasma membrane to undergo exocytosis (101). Thus, Rab3a-deficient mice are glucose-intolerant and exhibit loss of first phase insulin secretory response to glucose (102). One of the downstream effectors of Rab3 that is expressed and associated with insulin granules of the β-cells is Sytl4 (103), which was also significantly reduced in MKR islets as shown by iTRAQ and microarray studies. Sytl4 was also shown to interact with syntaxin 1A and munc18 in addition to the Rab GTPases (104, 105). Similarly drastically diminished expression of Sytl4 both in mRNA and protein levels was reported in INS-1 β-cells chronically exposed to supraphysiological glucose levels (106). Interestingly Sytl4-null mice show improved glucose tolerance as they secrete more insulin in response to a physiological glucose stimulus, and so its physiological effect is still under debate (107).

Because abrogated expression of all of these exocytotic proteins has been associated with a defect in insulin secretion and thus defines the MKR mouse phenotype, it is not entirely surprising that these proteins were reported to be down-regulated in the microarray and proteomics studies. However, whether this defect in insulin granule exocytosis is a primary one, a consequence of the persistent hyperglycemia, or downstream of another defect has yet to be explored.

Energy Utilization and Metabolism—

Islet β-cell glucose metabolism is essential for coupling glucose sensing to insulin release. The commonly recognized triggering pathway for GSIS is the K_ATP channel-dependent pathway (108, 109). The β-cell glucose transporter Glut2 was significantly down-regulated in MKR islets at both the protein and mRNA levels. Glut2-null mice are glucose-intolerant and lack first phase insulin secretion in response to glucose similar to diabetic MKR mice (110). Diminished Glut2 expression was also observed in T2D animal models and some human diabetic subjects as well as islets exposed to FFAs (111–115). Therefore, the impaired expression of Glut2 in MKR islets may reduce the amount of substrate available to mitochondria to generate ATP/ADP, a signal/stimulus for insulin release (116). Under glucose-stimulated conditions, β-cells shift from utilizing fatty acid to glucose as a fuel by inhibiting Cpt1 and causing conversion of glucose via Pcx, oxaloacetate, citrate, and acetyl-CoA to malonyl-CoA, which blocks the entry of long chain fatty acyl-CoA into the mitochondrion. The GSIS is tightly correlated with Pcx-catalyzed anaplerotic flux into the tricarboxylic acid cycle and stimulation of pyruvate cycling (117). The marked decrease in expression of Pcx both in protein and mRNA levels in MKR islets suggests impairment in the anaplerotic pathway and consequent defects in mitochondrial function.

Functional mitochondria are required for effective coupling of glucose metabolism to generate ATP and trigger insulin secretion in the β-cell (for a review, see Ref. 118). Glucose sensing requires oxidative metabolism, leading to the generation of ATP and other potential coupling factors (119). Down-regulation of multiple proteins involved in oxidative metabolism was observed in MKR islets, including several components of the mitochondrial respiratory chain (complex I subunit NDUFA9; complex III subunit UQCRH; complex IV subunits COX2, COX4I1, and COX5A; and complex V subunit ATP5J2) and electron transfer flavin proteins (ETFs). Reduced mRNA levels in multiple components in the mitochondrial respiratory chain have been reported in muscle and adipose tissue in human T2D patients (10, 120), and reduced expression of ATP synthase in INS-1 β-cells has been shown previously to inhibit GSIS (121). ETFs are necessary electron acceptors for many of the dehydrogenases in the mitochondria, and the deficiency of ETFs has been associated with multiple acyl-CoA dehydrogenation disorders (122). Point mutations or deletions in mtDNA have been associated with a large spectrum of diseases including mitochondrial diabetes (119). Furthermore we also observed the down-regulation of the adenine nucleotide translocators in MKR islets. Adenine nucleotide translocators transfer ATP to the cytosol in exchange for ADP. Collectively the down-regulation of these mitochondrial proteins in MKR islets is likely associated with decreased oxidative function and ATP production which could impair GSIS. These results are consistent with the observation that freshly isolated MKR islets contain significantly lower levels of ATP compared with WT controls.2

Conclusion—

This study represents the first integrated comprehensive global proteomics and genomics analysis of normal, healthy, and T2D pancreatic mouse islets. Our data provide insights into the pathological stages of pancreatic islet dysfunction induced by insulin resistance. Fig. 6 represents a model to explain the sequential and parallel changes that occur before β-cells fail and overt T2D presents. We propose that insulin resistance increases the demand on the β-cell to secrete insulin and may lead to 1) defective metabolic coupling and 2) cell stress. Defective metabolic coupling is caused by the dysregulation of several proteins involved in glucose uptake and oxidation ultimately decreasing the ATP/ADP. The up-regulation of UPR and ERAD proteins indicates inappropriate expression of insulin processing and secretory proteins leading to decreased insulin secretion. Concurrently an increase in ER stress would decrease synthesis of other proteins that also augment insulin secretion. Ultimately these changes cause a loss of glucose sensing and impaired insulin secretion. These initial changes cause β-cell failure, and the ensuing hyperglycemia likely exacerbates the initial defect, resulting in T2D.

Fig. 6.

A proposed model for the molecular and protein expression defects that lead to the dysfunctional islet metabolic phenotype in diabetic MKR islets. The proteins highlighted in the gray boxes were significantly changed in MKR islets.