A study [33] investigating the miRNA expression profile using plasma samples of 33 T1D patients reported that several miRNAs such as miR-21-5p, miR-101-3p, miR-103a-3p, miR-148b-3p, miR-155-5p, miR-200a-3p, miR-210-3p and miR-1275 were significantly upregulated. Among these miRNAs, the expressions of miR-21 and miR-101 were induced by inflammatory cytokines (TNF and IL-1β) [34], showing their role in beta-cell destruction mediated by cytokines. Further, miR-200a-3p regulates TP53 in the apoptosis pathway along with other genes involved in the insulin pathway and innate immune system; miR-210-3p regulates PTPN1 (tyrosine-protein phosphatase non-receptor type 1) in the insulin pathway; and IRS2 (Insulin receptor substrate 2) is regulated by both miR-200a-3p and miR-103a-3p [33]. Interestingly, there was no difference detected in the miRNA expression between the control group and T1D patients diagnosed for more than 5 years. Barutta et al. [35] conducted a cross-sectional nested case–control study to examine the differential expression of miRNAs in the serum samples of type 1 diabetic patients. It was reported that miR-140-3p, miR-574-3p, miR-139-5p, miR-106a, miR-17, miR-486-3p, miR-16, miR-222 and miR-885-5p were significantly upregulated. CXCL10 is targeted by miR-16, resulting in the inhibition of pancreatic β-cells [36]. Vascular homeostasis is regulated by miR-222 by negatively regulating STAT5A (signal transducer and activator of transcription 5A) in T1D patients, showing its anti-angiogenic properties [37]. In a study [38] using isolated human islets exposed to high glucose concentrations, and the serum samples of 22 patients with T1D, miR-23a and miR-98 were reported to be highly upregulated and targeted important signaling pathways involved in β-cell regulation and islet homeostasis, indicating that they might be useful biomarkers for T1D.

miR-197 was seen to be highly overexpressed in the serum samples of children with new-onset T1D [39]. It was also suggested that residual beta-cell function can be efficiently predicted by miR-197. The study further reported a significant partial correlation between miR-197 and stimulated C-peptide levels, indicating the critical role of miR-197 in insulin production. Moreover, using human activated CD4+ T cells, Scherm et al. [40] reported that miR-142-3p played a crucial role in the inhibition of T-cell activation, causing impaired stability of Treg cells. Elevated levels of miR-142-3p were reported to be linked to epigenetic remodeling in Treg cells, leading to islet autoimmunity, which is an indication of the onset of T1D.

lncRNA SRA was reported [41] to be overexpressed in the plasma samples of 25 patients with T1D as well as in CD4+ regulatory T cells (Tregs) induced with high glucose in the CD4+ MOLT4 human T lymphoblast cell line. lncRNA SRA modulates the expression of miR-146b, which is downstream of interleukin-1 receptor-associated kinase 1 (IRAK1)/protein kinase B (AKT)/S6 Kinase 1 (S6K1) signaling. However, lncRNA SRA and miR-146b reciprocally regulate each other, which indicates that lncRNA participates in crosstalk with IRAK1/AKT/mTOR signaling by competing with miR-146b. A positive association between lactate levels and hemoglobin A1c (HbA1c) was observed in the plasma of T1D patients. lncRNA SRA played a major role in the pathogenesis of T1D by inhibiting miR-146b via promotion of the IRAK1/lactate dehydrogenase A (LDHA)/phosphorylated LDHA (pLDHA) signaling pathway.

Significant downregulation of miR-155, miR-92a and miR-126 was seen in the serum samples of T1D patients [35]. miR-92a regulates NF-κB and other inflammatory signaling pathways, aiding in the progression of cardiovascular disease associated with diabetes [42]. Further, miR-126 plays a role in maintaining endothelial homeostasis, which is a common characteristic in diabetes [43]. In addition, miR-126 controls endothelial inflammation in patients with micro/macrovascular disorders associated with diabetes, which provides a link between diabetes-associated complications and lower levels of miR-126 [44].

miR-29a was upregulated in the serum samples of type 2 diabetic patients [45]. It was concluded that the overexpression of miR-29a represses glucose uptake via insulin stimulation, eventually causing resistance to insulin. In another study investigating the dysregulation of miR-34a, using the plasma samples of patients with T2D [46], it was observed that miR-34a was significantly upregulated, which was associated with senescence in pancreatic β-cells with a decrease in SIRT1 activity [47]. Interestingly, circulating miR-1 and miR-133a showed a positive association with myocardial steatosis in a study including T2D patients [48]. The elevated expression of miR-1 and hsa-miR-133a in the serum samples of T2D patients was an indicator of myocardial steatosis. Amr et al. reported that the plasma level of miR-210 was critically elevated in T2D patients with coronary artery disease. Indeed, miR-210 is linked with the hypoxia pathway and upregulated in response to hypoxia-inducible factor HIF1α [49].

LncRNA-p3134 was highly upregulated in a study [50] conducted using whole blood samples of T2D patients. Using bioinformatic analysis, it was reported that lncRNA-p3134 targets mRNAs, viz., TNS1, ASZ1, DIAPH1, IFNA14, ZNF436, MTMR3, CDK1, PPARD, TCF7, HUWE1 and CCND2 and regulates their biological functions. Sathishkumar et al. [51] studied the differential expression of lncRNAs in the peripheral-blood mononuclear cells (PBMCs) from patients with T2D wherein the lncRNAs ANRIL, ENST00000550337, GAS5, HOTAIR, lincRNA-p21 and PLUTO were significantly overexpressed in diabetic patients. The lncRNAs ANRIL, GAS5 and HOTAIR show a positive correlation with senescence-related biomarkers (p16, p21, p53 and β-galactosidase) and pro-inflammatory biomarkers (IL1- β, MCP1, IL6 and TNF-α) [51].

In a study comprising of 327 T2D patients with varying levels of urinary albumin-to-creatinine ratios (UACR) [52], it was observed that miR-130b was highly suppressed in the serum samples of the patient group with the highest UACR. It was also observed that serum miR-130b was negatively correlated with HbA_1c and HIF1-α levels. Additionally, the circulating serum miR-130b was closely related to blood glucose levels and insulin resistance inT2DM. The study further reported that miR-130b can be a biomarker for the early identification of diabetic nephropathy in T2D. Significant downregulation of let-7b was seen in the peripheral whole-blood samples of T2D patients [53]. miRNA let-7b targeted T2D-susceptibility genes such as IGF2BP2, TSPAN8, CDKN2A and KCNQ1. Additionally, downregulation of miR-214-3p and miR-27-3p was reported in the whole-blood samples of patients with T2D and also in a population with T2D risk factors [54], suggesting that they might be potential biomarkers for T2D.

A study [55] was conducted in 200 participants divided into two groups (diabetic and healthy volunteers). The differential expression of lncRNA VIM-AS1 and CTBP1-AS2 was studied in the PBMCs of these participants, and significant downregulation of both these lncRNAs was reported. Further, lncRNA CTBP1-AS2 was positively correlated with low-density-lipoprotein cholesterol (LDL-C) levels and negatively correlated with high-density-lipoprotein cholesterol (HDL-C) levels. No significant correlation between metabolic features and lncRNA VIM-AS1 was observed in the study. The target genes of the lncRNAs CTBP1-AS2 and VIM-AS1 were significantly enriched in pathways relevant to glucose homeostasis such as the mTOR, Hippo, TGF-β, insulin and Wnt signaling pathways.

Li et al. [56] explored novel non-proliferative diabetic retinopathy (NPDR)-specific circulating miRNAs in T2D mellitus patients. It was observed that miR-190a was highly expressed in the serum samples of NPDR patients. Further, it was noted [57] that miR-2116-5p and miR-3197 were critically upregulated in the serum samples of diabetic patients with retinopathy, which made it an important factor in differentiating between patients with and without DR. Notch homolog 2 (NOTCH2) was reported to be a target gene of miR-2116-5p in human embryonic kidney cells (HEK293T). Enhanced miR-20a-5p and -223-3p expression was seen in the serum samples of T2D patients with retinopathy, suggesting their critical role in the early detection of diabetic retinopathy [58].

A study [59] using the serum samples of 80 diabetic patients with either non-diabetic retinopathy (NDR), NPDR or proliferative diabetic retinopathy (PDR) was conducted to investigate the differential expression of lncRNAs. lncRNA HOTAIR and lncRNA MALAT1 (metastasis-associated lung adenocarcinoma transcript 1) were reported to be significantly upregulated in all groups. MALAT1 is a highly conserved lncRNA playing an important role in diabetes progression by modulating genes at the transcription level. MALAT1 has also been reported to be a regulator of the hyperglycemia-induced inflammatory process [60]. HOX transcript antisense intergenic RNA (HOTAIR), located on chr.12q13.13, is expressed abnormally in several disorders involving T1D. It is involved in distinguishing between diabetic and non-diabetic retinopathy as the expression levels of HOTAIR are increased in diabetic retinopathy [59]. The lncRNA HOTAIR also plays a role in regulating glucose levels by inducing insulin transcription-related genes, suggesting that glucose metabolism is closely associated with HOTAIR [61]. Li et al. [62] reported upregulation of lncRNA-MIAT in plasma samples of patients with NPDR and patients without diabetic retinopathy, as well as in human adult retinal pigment epithelial cells (ARPE-19) in a high-glucose environment. It was suggested that overexpressed lncRNA-MIAT was associated with diabetic retinopathy and reduced viability of ARPE-19 cells by upregulating TGFβ1. Upregulation of lncRNA KCNQ1OT1 in the serum samples of patients with PDR and NPDR, as well as in human retinal endothelial cells (hRECs), was observed. lncRNA KCNQ1OT1 promoted the expression of EGFR by inhibiting miR-1470. The lncRNA KCNQ1OT1 also plays a role in the proliferation and epithelial-to-mesenchymal transition (EMT) of lens epithelium, thus aiding in the progression of diabetes retinopathy [63].

Prado et al. [64] examined differential miR-320a expression in the plasma samples of diabetes patients with and without diabetic retinopathy. Downregulation of miR-320a was observed in patients with diabetic retinopathy. Further, it was reported that miR-320 targeted about 16 genes: YOD1, YWHAE, PTEN, TSC1, CDK6, MAPK8IP3, IGF2BP3, ARPP19 and FOXM1, which play roles in the regulation of cellular response to stress, the cell cycle and the negative regulation of the cellular protein metabolic process. PDX3 (a homolog of PDX1) and GXYLT1 have been implicated in regulating pancreatic development [65,66,67,68]. Li et al. [56] determined the expression of miRNAs from the peripheral serum of 21 T2D patients and reported that miR-485-5p, miR-4448, miR-9-5p and miR-338-3p were suppressed in patients with diabetic retinopathy. It was suggested that these miRNAs can be useful in early diabetic retinopathy risk prediction. Molecular pathway studies and analysis revealed that the aforementioned miRNAs target four genes, viz., SIRT1, FOXO1, FOXO3 and NFKB1. Target gene prediction revealed that SIRT1 (an important regulator of the revascularization of ischemic tissues) is the potential target of miR-338 and miR-9-5p, whereas FOXO1 and FOXO3 (regulating vascular morphogenesis) were targets for miR-485-5p, miR-338-3p and miR-9-5p. It has been reported [69] that miR-423 is negatively associated with the progression of diabetic retinopathy since it is highly downregulated in the serum samples of patients with proliferative diabetic retinopathy.

Downregulation of lncRNA AK077216 was reported in a study [70] conducted using the plasma samples of 60 diabetic retinopathy patients. Downregulated lncRNA AK077216 was observed only in diabetic retinopathy patients as compared to diabetic patients with no complications and controls. miR-383 was negatively associated with lncRNA AK077216 in diabetic retinopathy patients. Subsequently, ARPE-19 cells were used to analyze the expression of lncRNA AK077216 under high-glucose conditions. There was no significant effect of high glucose on lncRNA AK077216 in ARPE-19 cells; however, lncRNA AK077216 was reported to modulate ARPE-19 cell apoptosis by targeting miR-383.

Wu et al. [71] reported significant upregulation of miR-27a-3p in human HK-2 cells treated with high glucose. Further, it was observed that miR-27-3p inhibits the expression of prohibitin and transmembrane BAX inhibitor motif containing 6 (TMBIM6), which play a role in the progression of diabetic nephropathy. Thus, the upregulation of miR-27a-3p is positively associated with diabetic nephropathy via miR-27a-3p-prohibitin/TMBIM6 signaling axis.

LncRNA NR_038323 was overexpressed in high-glucose-induced HK-2 cells, which alleviated HG-induced renal fibrosis by targeting miR-324-3p [72]. It was suggested that miR-324-3p can mediate the high-glucose-induced upregulation of fibronectin, collagen I and collagen IV by targeting dual-specificity protein phosphatase-1 (DUSP1). Thus, lncRNA NR_038323 attenuated renal fibrosis by modulating the levels of fibronectin, collagen I and collagen IV via the miR-324-3p/DUSP1 axis. The study further verified these molecular changes in the human kidney samples of patients with diabetic nephropathy.

In a study [73] including 31 diabetic patients (18 with no diabetic complications and 13 with diabetic nephropathy), miR-31 was observed to be significantly downregulated in patients with diabetic nephropathy as compared to patients with no diabetic complications. The presence of diabetic nephropathy involved increased leukocyte adhesion and rolling flux and a lower rolling velocity. Leukocyte rolling velocity was positively associated with miR-31 levels, whereas leukocyte adhesion, as well as TNFα, ICAM-1 and IL-6 levels, were negatively associated with miR-31.

Yang et al. [74] reported the downregulation of long intergenic noncoding RNA for kinase activation (LINK-A) from the renal biopsies of patients with diabetic nephropathy. lncRNA LINK-A was found to be correlated with hypoxia-inducible factor 1α (HIF1α). The activation of HIF1α has been demonstrated to improve diabetic nephropathy [75]. Thus, HIF1α signaling activated by lncRNA LINK-A may improve the treatment of diabetic nephropathy.

Li et al. [76] investigated the dysregulation of miR-199a-3p and reported an overexpression of miR-199a-3p in the serum samples of patients with diabetic neuropathy. It was further noted that upregulated miR-199a-3p caused the inhibition of serine protease inhibitor (SerpinE2). miR-199a-3p induces diabetes neuropathy by promoting coagulation by targeting SerpinE2, which is known to prevent cartilage catabolism.

LncRNA MALAT1 was significantly upregulated in the peripheral-blood mononuclear cells of patients with diabetic neuropathy [77]. Further, a significant positive association between MALAT1 and C-X-C motif chemokine receptor 4 (CXCR4) was observed. It was earlier reported that CXCR4 chemokine receptor signaling mediates pain in diabetic neuropathy [78]. Therefore, the study suggests a possible involvement of the MALAT1/CXCR4 axis in the pathogenesis of diabetic neuropathy.

MiR-146a was reported [79] to be downregulated in the PBMCs of T2D patients with diabetic neuropathy. A negative correlation was observed between miR-146a and TNF-α and IL-6 levels. The downregulation of miR-146a was associated with diabetic neuropathy via activation of the NF-κB gene and an increase in the expression of TNF-α and IL-6.

Stirm et al. [80] studied the microRNA expression profiles of maternal and fetal WBCs from 30 women with gestational diabetes. It was observed that miR-340 was upregulated and PAIP1, a miR-340 target gene, was downregulated in women with gestational diabetes at the protein and mRNA levels. These results provide evidence for the miRNA-dependent programming of maternal WBCs in gestational diabetes.

Zhang et al. [81] studied the expression level of lncRNA MALAT1 in placental tissues from 78 gestational diabetes patients and 30 normal pregnant women. It was reported that lncRNA MALAT1 was upregulated in gestational diabetes patients. Further, they investigated the role of lncRNA in gestational diabetes using the placental trophoblastic-derived cell line HTR8, which was cultured in a high-glucose environment. It was observed that lncRNA MALAT1 downregulation in HTR8 cells in high-glucose conditions inhibited the cell proliferation, migration and invasion potency of HTR8. It was also reported that the downregulation of lncRNA MALAT1 led to the suppression of inflammatory factors and led to inhibition of the TGF/NF-κB pathway, suggesting that lncRNA MALAT1 affects the biological behavior of HTR8 through regulation of the TGF/NF-κB pathway.

Sun et al. [82] studied the expression of miR-29b in placental tissues from 204 gestational diabetes patients and reported that miR-29b was downregulated in these patients. It was also observed that the upregulation of miR-29b led to inhibition of the cell proliferation, viability, migration and invasion potency of trophoblastt cells. It was also reported that miR-29b overexpression led to the downregulation of HIF3A, which is the functional target of miR-29b. MiR-29b suppressed the expression of HIF3A by binding to its 3′ UTR region. Thus, miR-29b downregulation may be related to gestational diabetes development by increasing the expression of HIF3A.

A microarray analysis [83] of lncRNA expression in exosomes found in the umbilical cord blood of 23 patients with gestational diabetes was conducted wherein several lncRNAs, including ENST00000596839.1, lnc-EIF4ENIF1-1:1, lnc-TBC1D30-4:1 and lnc-ZNF800-1:1, were downregulated. miR-362-5p was reported to be a target of lnc-ZNF800-1:1. However, the functional relevance of this regulation is yet to be elucidated.

Investigation of the network topology of dysregulated ncRNAs in diabetes is imperative for biomarker discovery. In particular, the delineation of major target hubs based on the degree of the nodes, the directionality of the network (if known) and the presence of network motifs, including feed-forward loops, bifan motifs, etc., can shed tremendous light on novel biomarkers that can be exploited for theranostic and/or preventative intervention in diabetes. We classify dysregulated ncRNA biological networks in diabetes as follows:

(a)

Upregulated human microRNAs: Upregulated human microRNAs implicated in diabetes are summarized in Table 1. miRNA–miRNA and miRNA–target networks of upregulated human miRNAs in diabetes were constructed using an in silico method, as shown in Figure 1A and 1B, respectively.

(b)

Downregulated human microRNAs: Downregulated human microRNAs implicated in diabetes are summarized in Table 2. miRNA–miRNA and miRNA–target networks of downregulated human miRNAs in diabetes were constructed using an in silico method, as shown in Figure 2A and 2B, respectively.

(c)

Upregulated human lncRNAs: Upregulated human lncRNAs implicated in diabetes are summarized in Table 3. lncRNA–lncRNA and lncRNA–target networks of upregulated human lncRNAs in diabetes were constructed using an in silico method, as shown in Figure 3A and 3B, respectively.

(d)

Downregulated human lncRNAs: Downregulated human lncRNAs implicated in diabetes are summarized in Table 4. lncRNA–lncRNA and lncRNA–target networks of downregulated human lncRNAs in diabetes were constructed using an in silico method, as shown in Figure 4A and 4B, respectively.