The circulating exosomal microRNAs related to albuminuria in patients with diabetic nephropathy

Diabetic nephropathy (DN) is one of the most important chronic complications found in patients with diabetes mellitus (DM) [1], and is a leading cause of end stage renal disease (ESRD) worldwide [2, 3]. Because of its increasing prevalence and the excessive risk of cardiovascular mortality, DN has become a burden to global public health [4]. Therefore, tremendous efforts have been made to understand the pathogenesis of DN and to identify therapeutic targets in order to develop new treatments for DN. Current treatments for DN include medications such as renin–angiotensin–aldosterone system inhibitors, sodium-glucose cotransporter 2 inhibitors, and glucagon-like peptide-1 receptor agonists that can reduce cardiovascular mortality and limit the progression of DN [5‐7]. However, even the current best clinical practice for DN does not fully prevent DM patients suffering from complications and a progression to ESRD. Thus, further studies are warranted to unveil additional mechanisms to retard the progression of DN.

MicroRNAs (miRNAs) are endogenous, single-stranded, non-coding RNAs that regulate gene expression via post-transcriptional mechanisms. Currently, more than 2000 mature human miRNAs have been identified, and it is estimated that at least 60% of all human protein-coding genes are regulated by miRNAs [8]. Early studies examining the expression patterns of miRNAs in DN focused on miRNAs which were highly and specifically expressed in the kidney [9]. However, obtaining kidney tissues through a renal biopsy is highly invasive and hence these tissues sample are not readily available. Another important consideration is that DN is not only a kidney-specific disease, but also a consequence of the systemic complications of DM. Accordingly, some researchers have tried to investigate whether the levels of circulating miRNAs could be used for the early diagnosis and treatment of DN [10, 11]. However, the cell-free miRNAs present in circulation are highly heterogeneous. Cell-free miRNAs in the blood stream include the Ago2 protein bound form, the extracellular vesicle (EV)-enclosed form, and the vesicle-free form [12]. EV-incorporated miRNAs and whole cell-free miRNAs clearly differ from one another. miRNAs in EVs may have high chance to provide signatures reminiscent of their cell origin and are protected from the activity of extracellular RNases [13]. Furthermore, these exosome-enclosed miRNAs most likely function in intercellular communication and they could play a role in the pathology of various diseases [14‐16]. To the best of our knowledge, there are few studies examining the circulating levels of exosomal miRNAs in patients with DN [17]. Moreover, all previous studies exploring the levels of miRNAs in DN have used real-time quantitative PCR (qPCR) [17]. However, because of the inherent limitations in this technique unidentified miRNAs are often not detected [18].

Here, we investigated circulating exosomal miRNA profiles in DM patients with DN, DM patients without nephropathy, and in heathy volunteers (HVs). The levels of miRNAs were compared by RNA sequencing in order to identify differentially expressed miRNAs that could highlight the potential pathways involved in the progression of DN in DM patients, as well as to possibly provide new targets for the treatment of DN.

In this study, we conducted RNA sequencing to examine the profile of circulating exosomal miRNAs in patients with DN. As a result, we identified eight DN-specific miRNAs whose levels were up-regulated in circulating exosomes. In addition, the levels of these miRNAs were found to be significantly correlated with the degree of albuminuria. To our knowledge, this is the first study to examine miRNA profiles specifically in circulating exosomes in patients with DN.

Several previous studies have tried to characterize the circulating miRNAs in patients with DN [21‐27]. However, the miRNAs found in this study did not overlap with the miRNAs found in these previous studies. There are several possible explanations for this. First, all of the previous studies conducted used RNA isolation and quantification from whole serum or plasma, but we specifically measured RNA expression levels in circulating exosomes. The progression of diabetes is a consequence of pathological alterations in several tissues such as the liver, skeletal muscle, and adipose tissue, which can cause glucose intolerance, β-cell dysfunction, and the development of complications [28]. Therefore, inter-organ communications are important for the development of diabetes. Many previous studies have reported diverse circulating factors including hormones, cytokines, and growth factors that can modulate inter-organ communications in diabetes [29]. Moreover, recent studies have also revealed that circulating exosomal miRNAs can regulate gene expression in distant organs [30, 31]. Therefore, circulating exosomal miRNAs, rather than non-specific circulating miRNAs, may be more reliable candidates to explore the underlying disease mechanisms in patients with DN. Second, in all the previous studies RNA quantification was carried out using qPCR and microarrays, whereas we used RNA sequencing in this study. While RNA sequencing is not a replacement for other RNA quantification methods it is complementary to them. However, RNA sequencing has the advantages of highly reproducibility, low background, large dynamic range, and the ability to perform untargeted analysis [18, 32]. Therefore, for this study we believed that it was better to use RNA sequencing to discover unknown circulating exosomal miRNAs associated with DN [18]. Third, the different characteristics of patients may affect the different miRNA profiles. Some of the previous studies conducted their analysis using type 1 diabetic patients [23, 26], whereas other studies enrolled only diabetic patients, but not HVs [21, 27, 33]. Because of the excessive cardiovascular risk, the survival of DN patients is dramatically reduced as renal function declines [34]. However, the renal damage in diabetes forms a part of the systemic vascular complications that arise in this disease. A previous study has reported that the development of microalbuminuria, an early marker of DN, is a valid prognostic factor that can be used to predict future cardiovascular events [35]. Therefore, it is important to evaluate the unique characteristics of diabetic patients with vascular complications compared to those patients without vascular complications and healthy population. Accordingly, we selected seven miRNAs that were significantly up-regulated in DM patients with DN when compared to either HVs or to DM patients without nephropathy. In addition, one miRNA (miR-4449) was found to be commonly up-regulated in DN patients when compared to both HVs and DM patients without nephropathy. Moreover, the expression levels all of these miRNAs were found to be significantly correlated with the degree of albuminuria.

We found that the miRNA let-7c-5p was significantly up-regulated in patients with DN when compared to HVs. Previous studies have reported that the miRNA let-7 family members (let-7b and let-7c) play functional roles in renal fibrosis in DM through the regulation of transforming growth factor-β (TGF-β) signaling [36, 37]. In addition, one study also reported that mesenchymal stem cells, engineered to overexpress and deliver the miRNA let-7c via exosomes, attenuated renal fibrosis via the TGF-β1 pathway [38]. Therefore, it is likely that the up-regulated levels of let-7c-5p in our study are a physiological response to counteract the renal fibrosis seen in DN, and so let-7c may be a potential therapeutic target for the diagnosis and treatment of DN. Meanwhile, Pezzolesi et al. reported contradictory results that decreased circulating let-7c-5p was associated with increased risk of ESRD in type 1 diabetes patients that had been treated for more than 20 years [23]. Considering that TGF-β1 signaling also suppressed let-7c expression [37], subjects in the study of Pezzolesi et al. might already have prevalent renal fibrosis with highly activated TGF-β signaling, which can lead to down-regulated let-7c-5p. Further research is needed to describe the interaction between let-7c and TGF-β1 in progression of DN. The miRNA miR-150 was also evaluated in a previous study by Ranganathan et al. who reported that the deletion of miR-150 is associated with the suppression of inflammation and apoptosis in endothelial cells in the kidney seen as a result of myocardial infarction [39]. Interestingly, the downregulation of platelet miR-150 has also been reported in patients with atrial fibrillation (AF) and chronic heart failure [40]. AF increases the risk of stroke or mortality. Thus, the dysregulation of miR-150 may be one of the regulators that mediates cardiovascular damage. miR-642a-3p has also previously been reported to be a key regulator of adipogenesis [41]. Using deep sequencing, miR-642a-3p was shown to be significantly up-regulated during the adipogenic differentiation of human adipose tissue-derived stem cells. Adipocytes are not simply spaces for storing fat, but they also exert systemic metabolic effects by secreting adipokines via exosomes and extracellular vesicles [42], and are a main source of exosomal miRNAs [31]. Therefore, miR-642a-3p, which was up-regulated in patients with DN in our study, may a candidate for the inter-organ communication between adipose tissue and the kidney in diabetic patients. Another recent study reported that the levels of let-7i-3p in urinary extracellular vesicles was significantly increased in patients with DN, and is associated with degree of albuminuria [43]. Considering that let-7i-3p was also up-regulated in circulating exosomes and correlated with albuminuria in our study, increased let-7i-3p levels in serum and urine may indicate early progression in DN. The other miRNAs we found have not previously been reported to be associated with metabolic disease, and therefore further studies are warranted in order to understand the relationship between these miRNAs and DN.

We also analyzed the canonical biological pathways associated with the eight miRNAs we identified using miRSystem and DIANA-miRPath. These pathways have previously been reported to be involved in the pathogenesis of diabetic complications. The mRNA levels of mitogen-activated protein kinase (MAPK) signaling pathway components have been shown to be significantly increased in the kidney of diabetic patients compared to healthy population [44], and reduced levels of p38 MAPK signaling have been found to be associated with protection against renal injury and this signaling pathway plays a distinct pathogenic role in the progression of DN [45]. In addition, the integrin-vascular endothelial growth factor axis is a key factor in pathological angiogenesis including tumor metastasis, tissue remodeling, and diabetic complications [46]. Activator protein-1, along with nuclear factor-κB, has been reported to be associated with the increased release of inflammatory cytokines and the activation of monocytic cells in DN [47, 48]. Moreover, the olfactory receptor is expressed in pancreatic β-cells and renal tubules, and modulates systemic glucose metabolism, renin secretion, and GFR [49, 50]. Further studies are needed to evaluate the association between these miRNAs and the biological pathways we have identified, and it is hoped that this may reveal the hidden mechanisms underlying the progression of DN.

There are several limitations in our study that should be discussed. First, a relatively small number of participants were enrolled in this study; further studies with a larger cohort are needed to confirm our findings. Second, some demographic factors and clinical parameters were significantly different between groups, and these factors might affect circulating exosomal miRNA profiles. Third, although we conducted exosomal RNA sequencing using validated commercial kits, the procedures used to isolate exosomes are not standardized and a systematic bias in library preparation may lead to the false-positive identification of miRNAs [18, 51]. Further research with complementary methods such as qPCR is therefore needed to validate the novel miRNAs we identified. Fourth, our raw RNA sequencing data were normalized using DESeq 2, but another statistical approach may more accurately assess RNA expression level. Fifth, although we showed correlation between clinical parameters and the novel miRNAs we identified in DN, we only could suggest potential molecular pathways via bio-informatic analysis. Future experimental studies directly testing the relationships between miRNAs and clinical parameters are warranted. Finally, the circulating exosomes we isolated are likely to be derived from a various cell populations and we are therefore unable to investigate the true origin of the exosomes we isolated. It is likely that the exosomes shed from different cells have different miRNA profiles.

We identified a unique profile of circulating exosomal miRNAs in patients with DN. Our findings might form the basis for uncovering the inter-organ communication and pathogenesis in the progression of various diabetic complications including DN. Therefore, further studies examining the function of these miRNAs are warranted to understand the physiological mechanisms occurring in DN and to improve the outcomes in these patients.