1 Introduction

Exosomes are released from various types of cells under both normal and pathological conditions and the size ranges from 50 to 150 nm. Exosomes are secreted by many types of cells including tumor cells [1], mast cells [2], and adipose cells [3]. Exosomes are now recognized as major mediators of intercellular communication, although they had previously been thought to be inert cell debris [4–6]. Because of these physiological characteristics, some researchers recently suggested that exosomes are promising disease biomarkers [7]. The mechanisms by which exosomes may mediate intercellular signaling involves the activation of receptors on the plasma membrane of the recipient cells or influx of exosomes into the recipient cells [4, 8]. Alternatively, exosomes may transmit signals to recipient cells by directly transferring bioactive molecules via vesicles. Furthermore, expanding efforts in the field of exosomes research may explain the interaction among endocrine organs. Moreover some authors have characterized exosomes from mouse adipose tissue [3], and 3T3-L1 adipocyte [9], and rat primary adipocytes secreting exosomes [10]. Exosome like vesicles that are released by adipocytes of ob/ob mice not only activate monocytes but also promote differentiation and proliferation of bone marrow-derived macrophages (BMDMs) [3]. In addition, lipid synthesis and storage in small adipocytes are stimulated by microvesicles from large adipocytes by transferring RNA and glycosylphophatidylinositol (GPI)-anchored proteins [11]. Recently, adiponectin, a well-known adipokine, was shown to be associated with exosomes in vivo [12] and was detected in the exosomes from murine adipocytes and adipose tissue using proteomic approach [3, 9]. Indeed, exosomes that are released from adipocytes may be present in the blood circulation [12]. Taken together, adipocytes-derived exosomes are involved in the development of metabolic diseases by mediating cell communications. Nonetheless, the function and characteristics of exosomes that are released from adipocytes during diabetes have yet to be elucidated.

Because obesity is frequently related with the development of metabolic diseases such as type 2 diabetes and vascular complications, it has become a global problem for health in the worldwide. Adipose tissue performs an endocrine function by producing signaling and mediator proteins known as adipokines, via which the adipose tissue communicates with other tissues and organs to maintain systemic homeostasis [13]. By secreting adipokines which mediate cell signaling, adipose tissue communicates with the liver, skeletal muscle, heart, brain, and vasculature [14, 15]. Recent data indicate that these adipokines constitute a complex interconnected network mediating the cross-talk among the above-mentioned tissues and organ. In obesity, expansion of adipose tissue has been implicated in inflammation of adipose tissue and in dysregulation of adipokine secretion. This chronic stage of inflammation indicates a crucial pathogenic connections between obesity and metabolic syndrome such as type 2 diabetes. Regional distribution of body fat also believed to be a major reliable risk factor of cardiovascular diseases [16]. In addition, biological role of adipose tissue is to produce physiologically active substances as well as to stores excess energy [17]. It is thought that visceral adipose tissue is involved in the pathogenesis of metabolic disorders and type 2 diabetes related complications [18]. Such signals converge on target tissues, for example, on the liver and affect glucose production or on beta cells of pancreas, thereby modulating insulin production. The genes that are up-regulated during adipocytic hypertrophy may participate in the development of obesity and its complications [19, 20]. Dysregulated production of adipokines contributes to the pathogenesis of obesity-associated metabolic syndrome. In order to identify proteins that are related to the progression of common diseases, various proteomic approaches have been used to characterize the secretome of rodent and human adipocytes and adipose tissues [21].

In this study, to identify adipocyte exosome-associated proteins, we used obese diabetic Otsuka Long-Evans Tokushima fatty (OLETF) rats and their counterparts, Long-Evans Tokushima Otsuka (LETO) rats as an experimental model [22, 23]. The OLETF rat represent a status of non-insulin dependent diabetes mellitus (NIDDM) by Kawano et al. [24]. Because OLETF rats show insulin resistance, obesity, hypertension, hyperinsulinemia, and hyperglycemia, they seem to be the most suitable animal for our study. The control strain LETO rats, which were derived from Long-Evans rats, are mostly healthy (never develop diabetes) but obese. We isolated adipose tissue from visceral fat and collected exosomes released by the adipocytes.

To identify the exosomal proteome in the adipocytes of obese rats and rats with type 2 diabetes, we used the nano-liquid chromatography with quantitative time-of-flight tandem mass spectrometry (nanoLC-Q-TOF–MS/MS) along with a label-free relative quantification method. We identified 509 proteins in adipocyte-derived exosomes, and some of the identified proteins were either known to be present in or were detected in mouse adipocytes. In addition, we compared protein expression between OLETF and LETO rats; we analyzed 128 upregulated and 72 downregulated proteins, and changes in the expression of some of the differentially expressed proteins were confirmed by western blotting.

2 Materials and Methods

2.1 Animal Preparation

All experimental procedures were conducted in accordance with the Guidelines for Animal Experimentation of our institution. Four-week-old male OLETF and male LETO rats were obtained from Tokushima Research Institute of Otsuka Pharmaceutical Co. (Tokushima) and maintained at an animal facility. The animals were fed standard chow until 32 weeks of age and kept at controlled temperature with a 12-h lighting cycle.

2.2 Isolation of Adipocytes and Primary Culture of Rat Adipocytes

Adipocytes were isolated by collagenase digestion from epididymal fat pads of male LETO (n = 3) and OLETF (n = 3) rats under sterile conditions according to procedures described previously [25]. The adipocytes were resuspended in 20 mL of adipocyte buffer (0.14 M NaCl, 4.7 mM KCl, 2.5 mM CaCl2, 1.2 mM MgSO4, 1.2 mM KH2PO4, 20 mM HEPES/KOH at pH 7.4) supplemented with 0.2 % (w/v) bovine serum albumin (BSA), 1 % antibiotic/antimycotic solution, and 1 mM sodium pyruvate, and incubated in a shaking water bath. Subsequently, the cells were washed by flotation with 50 mL of adipocyte buffer 3 times, then cells were resuspended in 10 mL of adipocyte buffer, and incubated. To minimize variation in the cell confluence and viability, the cells were subjected to washing step and flotation in same buffer. Adipocytes obtained from each rat (three rats per group) were cultured into individual dish.

2.3 Preparation of Adipocyte-Derived Exosomes

The cell culture medium was collected every 24 h for 3 days. The pooled conditioned medium (10 mL × 3 times = 30 mL) from each dish was subjected to serial centrifugations (Fig. 1): 300×g for 3 min, 1500×g for 15 min, and 3000×g for 15 min; after that, the medium was filtered through a membrane (0.2-μm pore) and centrifuged at 200,000×g for 1 h. For washing, we resuspended the pellet in PBS and then pelleted the EVs at 200,000×g for 1 h again. Exosomes were enriched in this pellet. The pellet was resuspended in PBS for further analysis. Total exosomal protein was quantified using the BCA protein assay (Thermo Fisher Scientific). Equal amount (1 μg) of exosomal proteins were used for each MS run, and equal amount (5 μg) of exosomal proteins were loaded on each lane of SDS-PAGE for western blot analysis (Fig. 1).

Fig. 1
figure 1

A flow chart of the procedure Adipocytes were isolated from visceral adipose tissue of Long-Evans Tokushima Otsuka (LETO) rats (n = 3) and Otsuka Long-Evans Tokushima fatty (OLETF) rats (n = 3) and were cultured for 3 days. The culture medium was collected and centrifuged to enrich it in extracellular vesicles (EVs). The isolated EV fractions were analyzed using liquid chromatography with tandem mass spectrometry (LC-MS/MS). For protein profiling and protein quantification, the raw data files were processed in the MASCOT and IDEAL-Q software

Full size image

2.4 Electron Microscopy

The enriched exosomes were fixed in 4 % paraformaldehyde and deposited onto pure carbon-coated electron-microscope grids. The vesicle-coated grids were washed and incubated with 50 mM glycine in PBS. After staining with 3 % uranyl acetate, we dried the grids at room temperature and examined them at 6000× and 10,000× magnification under a transmission electron microscope (Hitachi H-7000).

2.5 Nanoparticle Tracking Analysis

Suspensions containing exosomes from the culture medium were analyzed using a Nano-Sight LM10 instrument (NanoSight). For the detection of particle size and concentration, a laser beam (wave length 405 nm) was passed through the dilute suspension of the EVs and the data were obtained in video recording. A video of 60-s duration was shot at the rate of 30 frames/s, and exosome movements were analyzed using the NTA software (version 2.2, NanoSight). NTA data acquisition settings were optimized at camera level 6–7, and detection threshold 5–6 for EV detection and kept constant during analysis of all samples, and acquired video was analyzed to obtain an estimate of the size distribution of isolated EVs.

2.6 Protein Digestion

The EV pellets from rat adipocytes were digested using the gel-assisted digestion method, as described previously [26, 27]. Briefly, the samples were mixed with acrylamide solution and solidified into gel, then cut into pieces, and washed with gel washing buffer (25 mM Triethylammonium bicarbonate (TEABC) in 50 % acetonitrile) 3 times, and dried. After drying, the proteins were treated with dithiothreitol (DTT) at 10 mM for reduction of disulfide bond, then treated with iodoacetamide (IAA) at 55 mM in the dark. The gel slices were washed and then dried. Trypsin (Promega) in TEABC was added (protein:trypsin 50:1) to gel and allowed to be soaked at 4 °C for 1 h, and then incubated at 37 °C overnight. The protein digestion was boosted by addition of enzyme again and followed by incubation for another 3 h. The peptides were extracted twice with buffer A (0.1 % formic acid/50 mM TEABC/50 % acetonitrile) and buffer B (0.1 % formic acid/50 mM TEABC) twice, in turn. These solutions were concentrated.

2.7 Mass Spectrometry

The digested samples were analyzed under optimized conditions as described previously [27]. Briefly, LC runs of the peptide mixtures of EV fractions were performed using a ultra-performance liquid chromatography (UPLC) (nanoAcquity system, Waters Corporation) equipped with a C18 trap column (5 μm, 20 mm × 180 μm) and a C18 analytical reversed-phase column (1.7 μm, 25 cm × 75 μm) (Waters Corporation). The separation of peptide samples were processed with a gradient 3–40 % of mobile phase B (0.1 % FA in acetonitrile) at 300 nL/min. The lock mass [Glu1-fibrinopeptide B solution was supplied through the reference spray of the NanoLockSpray source of the instrument and used for accurate mass detection. The peptide samples were analyzed using Q-TOF Premier mass spectrometer (Waters Corporation). The data acquisition was performed via data-dependent analysis (DDA) mode to automatically switch between the full MS scan (m/z 150–1600, 0.6 s) and the three MS/MS scans (m/z 100–1990, 0.2 s per scan) on the three most intense peaks.

2.8 Data Processing and Quantification

To create peak list file from MS raw data, MASCOT Distiller (Matrix Science; version 2.1) was used for DDA data. Subsequently, the data search of the MS/MS peak list files was performed in MASCOT (Matrix Science; version 2.2.1). MASCOT was set up to search against IPI_RAT_3.72 database, assuming trypsin as the digestion enzyme with 0.1 Da of parent ion tolerance and 0.05 Da of fragment ion mass tolerance. The modifications with carbamidomethylation of cysteine and oxidation of methionine as variable modification and the two missed cleavage of trypsin were allowed for analysis. Assignment of protein identification was determined when the protein identified using >2 peptides with >95 % probability or a single peptide with >99 % probability. During MASCOT analysis, each peak list data were performed and used for a search against a randomized decoy database which is automatically generated by MASCOT, that resulted in a <2 % false discovery rate according to decoy searches. For label-free quantification, the IDEAL-Q software (version 1.0.1.1) [28] was used. The raw data files from mass spectrometry were converted into files of mzXML using massWolf softwarer (Institute for Systems Biology). The search results were exported in the XML data format. In the data processing of IDEAL-Q, the identified proteins with a score ≥34 were assigned for label-free quantification. To determine peptide abundance, the detected peptide peak cluster is processed by validation criteria (signal-to-noise, charge state, and isotopic distribution) as described previously [26, 28, 29]. If a peptide passes these validation, the peak cluster is used to construct peptide abundance. To quantify the relative peptide abundance, extracted ion chromatography (XIC) areas of an assigned peptide normalized by XIC area of internal standard were calculated. The relative protein ratio for each protein was resulted using average abundance among the corresponding peptides.

2.9 Western Blot

Protein extracts from EV fractions and cell were loaded to a polyacrylamide gel (8 %) and transferred to nitrocellulose membranes. Then the transferred membrane were incubated on a rotating shaker for 1 h in a blocking solution (5 % skim milk). After blocking, we incubated each blot overnight at 4 °C with a primary antibody. For validation of the differential expression of proteins, we used specific antibodies against CD63 (Abcam), caveolin (Abcam), lipoprotein lipase (Abcam), AQP7 (aquaporin 7; Abcam), AK2 (Santa Cruz Biotechnology), catalase (Abcam), and liver carboxylesterase (Abcam). Each blot was washed 3 times, followed by incubation with a secondary antibody (conjugated to horseradish peroxidase): a goat anti-mouse IgG antibody (Cell signaling) or a goat anti-rabbit IgG antibody (Santa Cruz Biotechnology). The blots were visualized with enhanced ECL detection reagents and ECL hyperfilm. For analysis of densitometry, the each blots was measured and normalized with CD63 for EV protein or GAPDH for cell lysates.

2.10 Statistical Analysis

The SPSS 17.0 software (IBM) was used for statistical analysis. The quantified ratios were calculated of 3 biological replicates per group and the p values were determined Student’s t test between LETO and OLETF rats. For image analysis of western blot, Student’s t test was used method if significance corresponded to p < 0.05 (mean ± SD, n = 3 to validate differences between LETO and OLETF rats).

3 Results

3.1 Isolation of Exosome from Cultured Adipocytes

To characterize proteins of adipocyte-derived EVs, we used the proteomic approach shown in the flow chart of Fig. 1. Two animal models of the disease: OLETF (n = 3) and matched control (LETO, n = 3) rats served as a source of adipose cells in this study. The isolated adipocytes from single rat were cultured into one dish, and the culture medium of each dish was pooled every 24 h during 3 days to isolate the adipocyte-derived exosomes. The EV fractions were prepared by modified differential ultracentrifugation method from previously described method [30]; to further confirm the presence of exosomes after the centrifugation, we immunoblotted each fraction (obtained using the differential centrifugation method shown in Fig. 2a, top) for CD63, an exosome marker protein [31]. The immunoblot results showed that the exosomes were enriched in the pellet after centrifugation at 200,000×g. To evaluate the size of the isolated EVs, we measured the vesicle size using NanoSight. The isolated adipocyte-derived EVs had a known exosome size that ranged between 50 and 150 nm (Fig. 2b). The isolated vesicles were then analyzed by transmission electron microscopy (Fig. 2c); the images showed that our exosome preparations contained vesicles that were surrounded by a lipid bilayer and had a diameter of ~100 nm. Taken together, our results indicated that the majority of the EV fractions prepared by differential ultracentrifugation were enriched in exosomes.

Fig. 2
figure 2

Characterization of the isolated extracellular vesicles (EVs). a The pellets after centrifugation at 3000×g and 200,000×g and the supernatant after centrifugation at 200,000×g (the differential centrifugation method) as well as cell lysates (CL) were separated in an 8 % polyacrylamide gel and transferred onto a membrane. The transferred blots were immunoblotted with an anti-CD63 or anti-GAPDH antibody. b The distribution of vesicle sizes in the 200,000×g pellet was measured using NanoSight NTA software. c Representative images of transmission electron microscopy of EVs from adipocytes of Otsuka Long-Evans Tokushima fatty (OLETF) rats. The scale bar is 100 nm

Full size image

3.2 Protein Profiling of Adipocyte Exosomes

Each EV fraction from adipose cells of LETO rats (n = 3) and OLETF rats (n = 3) was used for the proteomic analysis. Lysed samples from each group were used for gel-assisted digestion and were assessed by LC–MS/MS analysis in triplicate. 276, 257, and 252 proteins from LETO group, and 279, 280, and 312 proteins from OLETF group were identified (p < 0.05, protein score ≥34, false discovery rate <1 %). To identify whole of EV proteins from adipocyte, we combined all of data from MS runs (3 biological × 3 analytical replicates of each group = 18 runs in Fig. 1 and Supplementary Table 1), resulting in the identification of 509 non-redundant proteins. Protein profiles were compared to the entries in the ExoCarta database; we found that approximately 78 % of the identified proteins were listed in the database (Fig. 3a). Using the DAVID gene ontology-based functional annotation analysis, the identified proteins were categorized by their subcellular location (Fig. 3b). According to the classification by subcellular locations, plasma membrane and extracellular-space proteins constituted a half of the exosomal content. In addition, the protein identification revealed 44 % of cytosolic proteins and 5 % of nuclear proteins. Seven functional clusters were built by the DAVID annotation tool from our 509 identified proteins: the first cluster corresponded to developmental processes, the second to localization, the third to establishment of localization, the fourth to organization of cellular components, the fifth to metabolic processes, the sixth to biogenesis of cellular components, and seventh cluster corresponded to cellular processes (Fig. 3c). The molecular functions of the identified proteins indicated that the exosomes contained diverse types of proteins such as transporters, antioxidants, electron carriers, enzymes, and structural molecules (Fig. 3d). We also compared our results to another study where the authors analyzed human-adipocytic-secretory-adipokine profiles using LC–MS/MS [21]. Eighty-one proteins, including adiponectin, fatty acid synthase, and catalase among the identified exosome-associated proteins were identified as adipokines in that profiling study [21] (Supplementary Table 2). Additionally, several adipose-tissue-specific proteins were successfully identified here in the EV fraction: carboxylesterase 3, caveolin 1, fatty acid-binding protein 4 (FABP4), and lipase (hormone-sensitive). FABP4, an adipokine which plays a crucial role in the pathogenesis of chronic metabolic diseases, was identified in the secretome of 3T3-L1 adipocytes in our previous study [32]; a recent study showed that this protein is secreted from adipocytes through secretory vesicles [33].

Fig. 3
figure 3

Characterization and classification of the identified proteomic profiles of adipocyte-derived exosomes. a Comparison of the identified proteins with entries in the ExoCarta database. Using gene ontology analysis (DAVID; web-based bioinformatic analysis tool), we calculated the percentage of proteins or the number of proteins, by b subcellular location, c molecular function, or d biological process

Full size image

3.3 Quantification of Identified Proteins

For quantification of disease-related proteins from the exosomes, the 509 identified proteins were quantified using the IDEAL-Q software (version 1.0.6.2) for label-free relative quantification as we have described previously [26]. To analyze the differentially expressed adipocytic-exosome-associated proteins, we quantified 3 replicates per group in IDEAL-Q. The cutoff levels that we determined for the differentially expressed proteins were filtered if the p value <0.05 and to 2 coefficients of variation (CV) of the relative protein ratio. In this experiment, 2 CV equaled 24.8 %, resulting a 1.2-fold change. According to these criteria, 165 proteins were expressed significantly differentially between OLETF and LETO rats. The resulting upregulated (128 entries) and downregulated proteins (72 entries) in the exosomes are listed in Table 1.

Table 1 Statistically significant proteins that were found to be expressed differentially between adipocytic exosomes of Otsuka Long-Evans Tokushima fatty (OLETF) rats and Long-Evans Tokushima Otsuka (LETO) rats
Full size table

3.4 Validation of Quantified Proteins Between LETO and OLETF

To characterize exosomal proteins from adipocytes of LETO and OLETF rats, we used western blotting to confirm the relative amounts of differentially expressed proteins of interest in exosomes and cultured-adipose-tissue lysates. Depending on the available commercial antibodies, and according to bioinformatic tools such as Gene Ontology Biological Process and Molecular Function annotation and some literature, we selected 6 proteins for further validation considering their involvement in obesity, type 2 diabetes, or lipid metabolism. These were (1) caveolin 1, which plays a role in vesicular transport and cholesterol homeostasis; (2) lipoprotein lipase, which participates in the metabolism and transport of lipids; (3) aquaporin 7 (AQP7), which modulates glycerol permeability of adipocytes by controlling triglyceride accumulation and fat-cell size; (4) adenylate kinase 2 (AK2), which may perform an important function in energy homeostasis in mature adipocytes; (5) catalase, which is an antioxidant enzyme; and (6) liver carboxylesterase, which controls cellular cholesterol esterification levels and detoxification. In the western blot analysis shown in Fig. 4, protein expression levels of caveolin 1, lipoprotein lipase, and aquaporin 7 were significantly higher in exosomes and cells of OLETF rats than in those of LETO rats. In contrast, AK2, catalase, and liver carboxylesterase were expressed more weakly in OLETF rats than in LETO rats. The upregulation or downregulation of proteins were confirmed to be similar between the exosomal fraction and the tissue lysate, even though the protein fold change in the densitometric analysis showed slight differences between the exosomes and the lysate of the tissue of origin. In addition, our western blot analysis showed similarity of protein expression changes between the exosomal fraction and the tissue lysate according to protein quantification analyses in IDEAL-Q; this result suggested that the findings of our nanoLC–MS/MS analysis provided reliable disease-related proteomic profiling.

Fig. 4
figure 4

Validation of the proteins that were expressed differentially between Otsuka Long-Evans Tokushima fatty (OLETF) rats and Long-Evans Tokushima Otsuka (LETO) rats. The expression levels of some upregulated or downregulated proteins were confirmed in the fraction of adipocyte-derived extracellular vesicles (EVs) (a) and in the corresponding adipose-cell lysate. b Adipocytes that were isolated form visceral fat of LETO and OLETF rats were cultured for 3 days and the culture supernatants were collected every 24 h. The EV fractions were collected by ultracentrifugation, and the cells were lysed

Full size image

4 Discussion

The importance of exosome secretion and the release of proteins from the cell has now been established in many cell types and under various physiological conditions, for example, during bone growth, platelet activation, immune responses, and especially in the tumor microenvironment. It is likely that the released exosomes carry a variety of cell-derived information and perform active functions. Adipose tissue produces various biologically active molecules collectively known as adipokines, e.g., adiponectin and FABP4 (adipocyte lipid-binding protein) [21]. Because adipose tissue may be involved in endocrine disorders, especially diabetes (which presents with aberrations in many organs), one possible cause of endocrine disorders may involve secreted microvesicles including exosomes derived from a primary disease organ. It has been demonstrated that the para-/endocrine regulation of adipocyte-derived extracellular vesicles, including exosomes and microvesicles that contain active mRNAs and proteins, entails a transfer of these vesicles to other cells and stimulation of intracellular communication during such processes as macrophage activation [3], angiogenesis [34], and lipid synthesis [9, 35].). Moreover, recent studies showed that in human hepatocarcinoma HepG2 cells, dysfunction of the TGF-β signaling pathway can be induced by exosomes from obesity-related adipocytes but not by exosomes from lean-tissue adipocytes [36]. Those authors hypothesized that the hepatocyte dysfunction can be induced by exosomes from visceral adipose tissue of obese patients; this was a novel paradigm for obesity-related liver diseases. By analyzing protein profiles of exosomes derived from adipocytes in present study, we confirmed the results of other studies that tested the possibility that adipocyte-derived exosomes participate in communication with other cell types.

One of the main aims of our study was to identify the proteome of extracellular vesicles from adipocytes of LETO and OLETF rats because this information may help to determine disease-related functions of exosomes in this animal model of diabetes. To date, proteomic analyses involving diabetic OLETF rats have been focused on the tissue proteome or oxidative stress-related protein modifications such as carbonylation, in comparison with LETO rats [37, 38]. Various studies involving 2-dimensional PAGE with LC–MS/MS as a targeted proteomic approach were designed to identify high-molecular-mass proteins with modifications in various tissues [38, 39]. As a result, several meaningful carbonyl proteins were identified, which show differential expression in comparison with LETO rats. On the other hand, we identified a great variety of proteins using our nanoLC–MS/MS coupled with relative quantitative proteomics. For the first time, proteomic analysis of adipocyte-derived exosomes was performed in both a diabetes animal model (OLETF rats) and controls (LETO rats) using LC–MS/MS. A total of 81 proteins, including adiponectin from adipocyte-derived exosomes, were identified previously by means of high-throughput proteomic analysis of adipokines from human primary adipocytes [21]. In fact, some in vivo studies have shown that there is an association of adiponectin to blood-circulating exosomes in genetically obese mice (ob/ob) and the corresponding lean mice (wild-type) [12]. Those researchers found that adiponectin concentrations in the circulating exosomes between the lean mice and obese mice shows no differences. In contrast, the ratio of adiponectin concentration to protein amount in the exosome is significantly lower in ob/ob mice than in wild-type mice [12]. Similarly, in the present study, there is no significant difference between LETO and OLETF rats in adiponectin levels in the exosomal proteome, according to quantitative proteomic analysis.

We demonstrated changes in the expression of many kinds of proteins (128 upregulated and 72 downregulated) in the exosomes of diabetic rats. These differentially expressed proteins in the adipocytic exosomes isolated from OLETF and LETO rats include caveolin 1 (upregulated in OLETF rats), lipoprotein lipase (upregulated in OLETF rats), aquaporin 7 (upregulated in OLETF rats), adenylate kinase 2 (downregulated in OLETF rats), catalase (downregulated in OLETF rats), and liver carboxylesterase (downregulated in OLETF rats). Some authors suggested that caveolin 1 is involved in the pathogenesis of obesity and adipose-tissue-related insulin signaling in humans [40, 41]. Upregulation of caveolin 1 mRNA expression was reported in obese patients with type 2 diabetes compared to lean controls [40]. Lipoprotein lipase performs a major function in the metabolism and transport of lipids and is the enzyme responsible for hydrolysis of core triglycerides [42]. Furthermore, upregulation of mRNA expression of lipoprotein lipase was shown in visceral adipose tissue of OLETF rats compared to LETO rats [43]. Moreover, 1 study demonstrated that a reduced lipoprotein lipase amount and activity correlate inversely with the extent of visceral fat accumulation.

Aquaporin 7 is a glycerol channel mainly expressed in adipocytes [44, 45]. Recently, an increase (2.5-fold) of mRNA expression of aquaporin 7 was demonstrated in adipose tissue of OLETF rats compared to LETO rats [44]. AK2, a mitochondrial enzyme, regulates adenine nucleotide interconversion. Upregulation of the AK2 protein is strongly associated with adipocytic differentiation [46]. In our study, the exosomes and adipose cells show a lower level of the AK2 protein expression in OLETF rats compared to LETO rats. This finding can be explained as follows: downregulation of AK2 causes perturbation of energy metabolism in mitochondria.

Catalase is an antioxidant enzyme expressed in various tissues; it protects cells from harmful effects of hydrogen peroxide by converting it to oxygen and water [47]. In adipose tissue of obese mice, expression of catalase significantly decreases [48] just as in our present proteomic and western blot analysis. It was also reported that the activity of catalase decreases during oxidative stress judging by lipid peroxide content of the renal cortex in OLETF rats [47]. Although liver carboxylesterase (carboxylesterase 1) has not been detected in adipose tissue of OLETF rats, carboxylesterase 1/esterase-x-deficient mice become obese and hyperlipidemic and develop hepatic steatosis even on a standard diet [49, 50]. Some researchers showed that phosphorylation of AKT, which is a downstream kinase of insulin signaling, is significantly reduced in adipose tissue of carboxylesterase 1/esterase-x-deficient mice [50]. We can theorize that the weaker expression of liver carboxylesterase in adipocytes of OLETF rats is indicative of insulin resistance.

We can conclude that our proteomic analysis and bioinformatic search can provide a means for gathering comprehensive data on exosomal proteins in diabetes. Ours seems to be the first report on proteomic characterization of adipocytic exosomes from obese rats with or without diabetes (OLETF and LETO). Our results support the utility of protein profiling of adipose-tissue-derived exosomes and suggest that diseased-tissue exosomes reflect the status and functions of the tissue of origin. Moreover, independent confirmation of our findings should advance the understanding of obesity- and diabetes-related diseases.