Transcript Expression Profiles and MicroRNA Regulation Indicate an Upregulation of Processes Linked to Oxidative Stress, DNA Repair, Cell Death, and Inflammation in Type 1 Diabetes Mellitus Patients

Type 1 diabetes (T1D) arises from autoimmune-mediated destruction of insulin-producing β-cells leading to impaired insulin secretion and hyperglycemia. T1D is accompanied by DNA damage, oxidative stress, and inflammation, although there is still scarce information about the oxidative stress response and DNA repair in T1D pathogenesis. We used the microarray method to assess mRNA expression profiles in peripheral blood mononuclear cells (PBMCs) of 19 T1D patients compared to 11 controls and identify mRNA targets of microRNAs that were previously reported for T1D patients. We found 277 differentially expressed genes (220 upregulated and 57 downregulated) in T1D patients compared to controls. Analysis by gene sets (GSA and GSEA) showed an upregulation of processes linked to ROS generation, oxidative stress, inflammation, cell death, ER stress, and DNA repair in T1D patients. Besides, genes related to oxidative stress responses and DNA repair (PTGS2, ATF3, FOSB, DUSP1, and TNFAIP3) were found to be targets of four microRNAs (hsa-miR-101, hsa-miR148a, hsa-miR-27b, and hsa-miR-424). The expression levels of these mRNAs and microRNAs were confirmed by qRT-PCR. Therefore, the present study on differential expression profiles indicates relevant biological functions related to oxidative stress response, DNA repair, inflammation, and apoptosis in PBMCs of T1D patients relative to controls. We also report new insights regarding microRNA-mRNA interactions, which may play important roles in the T1D pathogenesis.


Introduction
Type 1 diabetes (T1D) or insulin-dependent diabetes mellitus (IDDM) is a polygenic disorder possibly triggered by environmental factors that results from a T cell-mediated autoimmune attack against the insulin-producing β-cells localized in the pancreatic islets of Langerhans [1]. The typical pathological lesion is a destructive immune cell infiltrate (insulitis), affecting insulin-producing β-cells at several levels and impairing insulin synthesis as a consequence [2]. During this process, islet-infiltrating mononuclear cells release proinflammatory cytokines and specific biochemical markers in the serum of patients, which have been exploited as potential markers for the pathogenesis of T1D. In fact, the levels of inflammatory markers have been significantly upregulated in T1D patients compared to healthy human subjects [3].
The hallmark of T1D is a decreased insulin secretion that is subsequently succeeded by chronic hyperglycemia, which has been implicated in long-term complications affecting several organs, including the kidneys, eyes, heart, nerves, and blood vessels [4]. Chronic hyperglycemia induces the production of reactive oxygen species (ROS), which in excess can overwhelm the antioxidant system and lead to oxidative stress [5,6]. There is evidence that oxidative stress can increase the release of proinflammatory cytokines, subsequently leading to inflammation and β-cell destruction, contributing to T1D progression [7]. Nevertheless, studies have reported that T1D patients present elevated oxidative stress markers [8][9][10] and decreased antioxidant capacity [7,9,11]. Besides, T1D patients have shown higher levels of DNA damage and oxidative DNA damage than controls [12,13], indicating that DNA repair mechanisms may be compromised in those patients. However, there is still scarce information about molecular signaling pathways and genes implicated in biological processes related to oxidative stress responses and DNA repair in T1D.
In the present study, we applied the microarray method to study the mRNA transcript expression profiles displayed by peripheral blood mononuclear cells (PBMCs) of T1D patients compared to healthy subjects. We also aimed to identify potential mRNAs targets that are mainly associated with responses to oxidative stress and DNA repair pathways and those associated with differentially expressed microRNAs (that were previously reported by Takahashi et al. [21]).

Study Subjects.
A total of 19 patients with type 1 diabetes (7 women and 12 men, with age ranging from 18 to 37), recruited while undergoing regular follow-up at the Outpatient Endocrinology of the Clinical Hospital of the Faculty of Medicine of Ribeirão Preto (HC/FMRP-USP), Brazil, and 11 healthy subjects (control group) (6 women and 5 men, with age ranging from 20 to 31) participated in the present study. The main characteristics of all participants are described in Tables 1 and 2. All patients were receiving treatment with human insulin, and those presenting recent episodes of ketoacidosis and late diabetic complications were excluded from the study. The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Local Ethics Committee of Clinical Hospital-Ribeirão Preto Medical School, University of São Paulo (protocol code no. 13314/2011 in November 23, 2011). Informed written consent was obtained from all participants.

Sample Collection, Isolation of Peripheral Blood
Mononuclear Cells (PBMCs), and RNA Extraction. Peripheral blood samples (20 mL) were collected from all participants, followed by isolation of PBMCs using Histopaque-1077 (Sigma-Aldrich, Inc., USA). Total RNA was extracted using Trizol reagent (Invitrogen, Life Technologies, Carlsbad, CA, USA) according to manufacturer's instructions. The quality and quantification of RNA samples were measured using the NanoDrop ND-1000 Spectrophotometer (Uniscience, São Paulo, Brazil). The integrity of RNA samples was evaluated using Agilent RNA Nano 6000 chips onto the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). RNA samples that were protein-and phenol-free, and RNA integrity number ðRINÞ ≥ 7:0 were considered for the microarray and qRT-PCR analysis.
2.3. mRNA/miRNA Microarrays and qRT-PCR. The microarray technique and the expression data normalization and statistical analysis were performed as previously described [21,22]. The microarray data from all samples used in this study are publicly available in the ArrayExpress database (http://www.ebi.ac.uk/arrayexpress) under the accession numbers E-MEXP-3348 (T1D group) and E-MEXP-3963 (control group). The qRT-PCR method was used to validate the expression results obtained for mRNA and microRNA microarrays. For microRNA expression level analysis, 10 ng of total RNA was reverse transcribed using Taqman micro-RNA Reverse Transcription kit (Applied Biosystems, Foster City, CA). The qRT-PCR was performed using Taqman Universal PCR Master Mix with AmpErase uracil Nglycosylase (UNG) (Applied Biosystems) and Taqman Micro-RNA Assays (Applied Biosystems) for the following microRNAs: hsa-miR-101 (MIMAT0000099) (002253), hsa-miR-148a (MIMAT0000243) (000470), hsa-miR-27b (MIMAT0000419) (000409), hsa-miR-424 (MIMAT0001341)   using BRB-ArrayTools (developed by Richard Simon and BRB-ArrayTools Development Team) to identify gene groups with significant expression (GSA) and significantly enriched gene sets (GSEA), according to Gene Ontology (GO) terms. The LS/KS permutation test and Efron-Tibshirani's GSA maxmean test with random 1,000 permutations were carried out with a threshold of 0.005 to filter statistically significant (GSA) and significantly enriched (GSEA) gene sets, respectively.

Results
After normalization and adjustment of data, the analysis performed by rank products revealed 277 differentially expressed genes (DEGs) in PBMCs of patients with T1D compared to controls (220 upregulated and 57 downregulated) ( Table S1). Some of these genes have already been widely described as associated with T1D, including those related to inflammatory processes (Table 3). In addition, the GSA and GSEA analysis indicated 49 significantly expressed gene sets (p < 0:005) and 55 statistically enriched gene sets (p < 0:005) in PBMCs of T1D patients compared to the control group. Biological processes related to apoptosis (apoptotic signaling pathway, release of cytochrome c from mitochondria), oxidative stress (regulation of oxidoreductase activity), inflammatory processes (interleukin-2 production), and cell death (positive regulation of neuron death) were found significantly upregulated in T1D patients. Pathways associated with endoplasmic reticulum stress and unfolded protein response (response to endoplasmic reticulum stress, cellular response to topologically incorrect protein/unfolded protein), reactive nitrogen species (nitric oxide metabolic and biosynthetic process), and DNA repair (double-strand break repair via homologous recombination and recombinational repair) were found statistically enriched for T1D patients (Figures 1(a) and 1(b)). The complete list of significant and statistically enriched gene sets can be found in Table S2.
Besides, to identify genes that were specifically related to the biological processes "response to oxidative stress" (GO: 0006979) and "DNA repair" (GO: 0006281), the list of DEGs was submitted to the gene prioritization tool Endeavor, which allowed the integration of the results in several gene databases. The 16 best ranked genes in each process and the respective fold-change values are highlighted (Table 4).
We also verified whether any of those 277 DEGs were included as possible targets for the 44 differentially expressed microRNAs (AUC ≥ 0:90), which were previously reported for the same group of patients [21] using the VENNY tool. We found 144 (52%) genes that may be potentially regulated by the differentially expressed microRNAs. Of these, 75 are possibly regulated by the upregulated micro-RNAs, seven by the downregulated microRNAs, and 62 by both up-and downregulated microRNAs (Table 5). Interestingly, among the possible targets, nine and thirteen genes belong to "response to oxidative stress" and "DNA repair" biological processes, respectively, such as UCP3, PTGS2, ATF3, FOSB, DUSP1, and TNFAIP3 genes.
To confirm the expression of some mRNAs, we selected six genes (UCP3, PTGS2, ATF3, FOSB, DUSP1, and 1 Controls whose samples were used for the study of microRNA expression profiles. All 11 samples were used in the analysis of mRNA expression data. Cell cycle regulation, response to cytokine stimulation and oxidative stress, biosynthetic process of prostaglandin 1.57 0:00E + 00 0:00E + 00

TNFAIP3
Tumor necrosis factor, alpha-induced protein 3 Antiapoptosis, negative regulation of NF-κB transcription factor activity, TNF production, and inflammatory response   Journal of Diabetes Research   Table  5: Potential microRNA-mRNA interactions selected from both lists of differentially expressed targets in PBMCs of T1D patients compared to controls.

10
Journal of Diabetes Research TNFAIP3) to perform analysis by qRT-PCR. Five genes, except UCP3, were significantly upregulated in T1D patients compared to the controls, similarly as we have found in the microarray experiments. UCP3 expression was not significantly different between patients and controls (Figures 2(a)-2(f) and 3). Regarding the expression of microRNAs, we confirmed the microRNA expression (obtained in the microarray analysis) by performing the qRT-PCR for the following four microRNAs: hsa-miR-101, hsa-miR148a, hsa-miR-27b, and hsa-miR-424. All of them were significantly upregulated in T1D patients compared to the controls, compatible with the results obtained by the microarray method [21] (Figures 4(a)-4(d) and 5).

Discussion
The exact cause of T1D has not yet been elucidated, thus requiring a search for potential disease biomarkers and their possible functions related to the molecular mechanisms associated with the disease, providing new insights towards their application in clinical practice.
Among the 277 differentially expressed genes, it is important to note that we found some genes (HLA-DQB1, CD69, and TNFAIP3) that have been reported to be implicated in the pathogenesis of T1D [24]. While HLA-DQB1 was downregulated, CD69 and TNFAIP3 were found upregulated. In addition, a series of genes highlighted in this study are involved in molecular and cellular events associated with T1D (IL1B, TNF, IFNG, GZMB, and GZMH) [25,26] and inflammatory processes (IL1B, IFNG, IL6, PTX3, CCL20, CXCL2, and DUSP2) [3,27], endorsing other studies in the literature. Interestingly, our results indicated differential expression of genes related to T1D even years after diagnosis, such as CCL3L3, CCL4, CXCL1, CXCL3, and IL8 genes.
In addition, the analysis by gene sets (GSA and GSEA) revealed an upregulation of biological processes related to ROS generation, oxidative stress, inflammation, cell death, ER stress, and DNA repair, among others. The excessive generation of reactive species (derived from oxygen or nitrogen) leads to a redox imbalance in the organism. Consequently, oxidative stress disturbs endoplasmic reticulum (ER) homeostasis and activates the unfolded protein response (UPR). Oxidative stress can also increase DNA damage, which may activate DNA repair and, depending on the scenario, may also lead to apoptosis [28,29]. Other studies have described upregulation of DNA repair, inflammation, ER-stress response, and apoptosis pathways in T1D patients [30,31]. There is also evidence that oxidative stress and inflammation occur even several months before the onset of T1D in children, favoring the autoimmunity and induced cellular damage in β-cells [32].
In this study, the high expression levels of DUSP1, PTGS2, TNFAIP3, ATF3, and FOSB genes in T1D compared with the control group were identified by the microarray analysis and confirmed by the qRT-PCR method. DUSP1 (dual-specificity protein phosphatase-1), also known as MKP1, encodes a phosphatase protein that dephosphorylates and inactivates MAPKs (mitogen-activated protein kinases), playing an essential role in cell proliferation, cellular growth, inflammation, cell cycle arrest, innate immune function, and cellular response to oxidative damage [33]. In fact, it has been shown an upregulation of DUSP1 under stress conditions that lead to apoptosis [34]. The gene PTGS2 (prostaglandin synthase cyclooxygenase 2) encodes cyclooxygenase 2 (COX2) that generates prostaglandin E2 (PGE2). Stress conditions such as hyperglycemia and ROS induce the expression of PTGS2 increasing PGE2 levels, which further stimulate ROS production [35,36]. Elevated expression of PTGS2 is commonly observed in many chronic inflammatory diseases, and high COX2 levels are also associated with microvascular complications of diabetes, including endothelial dysfunction and renal injuries [37].
TNFAIP3 (tumor necrosis factor (TNF) α-induced protein-3) encodes a zinc finger protein (A20), which inhibits NFKB signaling, decreases NO production, and has demonstrated a critical role in protecting β-cells from apoptosis [38][39][40]. Interestingly, we found an upregulation of NFKB Inhibitor Alpha (NFKBIA) and NFKB Inhibitor Zeta (NFKBIZ) genes, both with inhibitory functions upon NFKB, which may be related to the high level of inflammation in T1D patients [41]. TNFAIP3 has also been indicated as a candidate gene for T1D [42]. Some authors suggest that the SNP rs2327832 of TNFAIP3 may help to predict glycemic control and disease progression in T1D children [39].
ATF3 (activating transcription factor 3) gene is an adaptive-response gene activated upon various stress stimuli, including ROS, DNA damage, genotoxic agents, and cytokines [43]. There is evidence that ATF3 silencing increases apoptosis in β-cells, thus suggesting a protective role regarding antiapoptotic effects [44]. ATF3 also represses inflammatory responses (by inhibiting NFKB), several proinflammatory cytokines, and NO synthesis [45].
Regarding FOSB (FosB Proto-Oncogene, AP-1 Transcription Factor Subunit), this gene encodes leucine zipper proteins that form dimers with other proteins (e.g., Jun and ATF/CREB) to compose the AP-1 complex which plays a crucial role in DNA repair mechanism [46]. Altogether,

11
Journal of Diabetes Research the upregulation of the aforementioned genes in T1D group may indicate that T1D patients exhibit an adaptive response activating cellular stress response and DNA repair genes upon several conditions, such as ROS, DNA damage, oxidative stress, and inflammation.
Moreover, microRNAs are important posttranscriptional regulators of several biological processes in T1D patients.
Notably, it has been demonstrated that microRNAs show a specific signature for each form of diabetes (T1D, type 2 diabetes, or gestational diabetes) [47] and may be implicated in the pathogenesis and complications of diabetes [48]. Therefore, we crossed our differentially expressed mRNAs with the set of 44 differentially expressed microRNAs, which were previously reported for the same group of T1D patients [21]. Our results indicated that 52% of differentially expressed genes are predicted targets of the 44 microRNAs. Those target genes are related to inflammation, stress response, and DNA repair, showing the importance of microRNA regulation in these processes. In this study, we also confirmed the expression of four microRNAs (hsa-miR-101, hsa-miR148a, hsa-miR-27b, and hsa-miR-424) by qRT-PCR, all of them being upregulated in PBMCs of T1D patients.
miR-101 has been found to regulate DNA-dependent protein kinase (DNA-PKcs) and ATM (ataxia telangiectasia mutated) gene, with roles in suppressing DNA repair processes and sensitize cells to DNA damage induction [49]. Besides, there is evidence confirming that miR-101 targets DUSP1, promoting the activation of MAPKs and stimulating the production of proinflammatory cytokines [50]. On the other hand, miR-101 targets PTGS2, decreasing the release of proinflammatory cytokines [51]. miR-101 is also related to cytokine-mediated defective insulin production [52], βcell dysfunction [52,53], and increased apoptosis [52]. Interestingly, upregulation of miR-101 has been found in individuals with normal glucose levels, but testing positive for autoantibodies linked to T1D, suggesting that miR-101 precede the impairment of glucose homeostasis and may represent a potential biomarker for the onset of T1D [53]. In addition, miR-101 downregulation ameliorates insulin release and protects β-cells from cytokine-induced apoptosis [52].
Overexpression of miR-27b was found to inhibit NRF2 activation, a master regulator of antioxidant response [54], and suppress NFKB activation [55]. Furthermore, mir-27b upregulation has been associated with the onset of retinopathy [56], while its downregulation has been implicated in glucose tolerance impairment [57] and nephropathy [58]. The upregulation of mir-148a has been consistently found in T1D patients [59,60]. miR-148a has been described as a regulator of autoimmunity, β-cell tolerance [18], and insulin activity [61]. Recently, Tamara et al. [62] reported that upregulation of miR-424 in T1D patients was associated with increased levels of inflammatory cytokines and increased risk of cardiovascular diseases.
Altogether, our results indicate specific mRNA and microRNA expression profiles in PBMCs from T1D patients relative to healthy nondiabetic individuals. In addition, we have shown that a number of biological processes related to oxidative stress response, DNA repair, inflammation, and apoptosis are upregulated in T1D patients. Besides, we provide new data regarding potential microRNA-mRNA interactions in T1D, in particular involving genes associated with responses to oxidative stress and DNA repair, which might play relevant roles in the microRNA-target network in T1D patients.

Conflicts of Interest
The authors declare that they have no conflicts of interest.