MicroRNAs Associated with Chronic Kidney Disease in the General Population and High-Risk Subgroups—A Systematic Review

The potential utility of microRNAs (miRNAs) as diagnostic or prognostic biomarkers, as well as therapeutic targets, for chronic kidney disease (CKD) has been advocated. However, studies evaluating the expression profile of the same miRNA signatures in CKD report contradictory findings. This review aimed to characterize miRNAs associated with CKD and/or measures of kidney function and kidney damage in the general population, and also in high-risk subgroups, including people with hypertension (HTN), diabetes mellitus (DM) and human immunodeficiency virus (HIV) infection. Medline via PubMed, Scopus, Web of Science, and EBSCOhost databases were searched to identify relevant studies published in English or French languages on or before 30 September 2022. A total of 75 studies fulfilled the eligibility criteria: CKD (n = 18), diabetic kidney disease (DKD) (n = 51) and HTN-associated CKD (n = 6), with no study reporting on miRNA profiles in people with HIV-associated nephropathy. In individuals with CKD, miR-126 and miR-223 were consistently downregulated, whilst in DKD, miR-21 and miR-29b were consistently upregulated and miR-30e and let-7a were consistently downregulated in at least three studies. These findings suggest that these miRNAs may be involved in the pathogenesis of CKD and therefore invites further research to explore their clinical utility for CKD prevention and control.


Introduction
The incidence of chronic kidney disease (CKD) is on the rise globally, and it is expected that CKD will be the fifth leading cause of death by 2045 [1]. This is partly attributable to the high burden of diabetes mellitus (DM) and hypertension (HTN), which are the leading causes of CKD, as well as other causes, including human immunodeficiency virus (HIV) infection and advanced age [2]. Chronic kidney disease is described as a silent condition due to a lack of obvious clinical symptoms, particularly in its early stages. As a result, most affected individuals are unaware of their disease status, and are often only detected at an advanced stage of the disease [3]. Furthermore, CKD is an independent risk factor for cardiovascular disease (CVD), and individuals with CKD are more likely to die of CVDs than progress to end-stage kidney disease (ESKD) [4]. Early diagnosis of CKD and effective screening of high-risk individuals is critical to mitigate disease progression and substantially reduce related poor health outcomes [5].

Search Strategy
A comprehensive and systematic search of Medline via PubMed, Scopus, Web of Science, and EBSCOhost databases was conducted to identify eligible studies, published in English or French languages on or before 30 September 2022, without a starting date. The search strategy made use of keywords and phrases such as "microRNAs, miRNA, miR-NAs, chronic kidney disease, CKD, chronic kidney injury, chronic renal disease, chronic renal injury, renal failure, end-stage renal disease, diabetic kidney disease, diabetic nephropathy, hypertensive nephrosclerosis, chronic kidney failure, chronic renal failure, end-stage renal failure, HIV-associated nephropathy, HIVAN, HIV-associated renal disease, HIV-associated kidney disease, serum creatinine, serum cystatin C, estimated glomerular filtration rate, urinary albumin excretion rate (UAER), urinary albumin-to-creatinine ratio (UACR), urinary albumin" in combination with Boolean operators (AND/OR/NOT) (refer to Additional files: Tables S1-S4). Furthermore, we manually scanned reference lists of the included studies for additional studies.

Data Collection
Two authors (DDM and DMM) independently conducted the database searches and screened studies by title, abstract and full text to identify those meeting the inclusion criteria, as shown in Figure 1. Disagreements encountered were resolved through discussions or consultation with a third author (CG). Studies were included if they: (i) were original articles reporting on miRNAs associated with prevalent CKD and/or measures of kidney function (serum creatinine, serum cystatin C, eGFR) or kidney damage (urinary albumin excretion rate, albumin-to-creatinine ratio, urinary protein) in the general adult population and/or high-risk subgroups (HTN, DM, HIV-infection), ii) written in English and French languages, (iii) clearly described the type of sample in which miRNA analysis was done, methods used for miRNA detection and quantification, as well as the normalization control used, (iv) with clearly defined cases and controls. Studies were excluded if they were: (i) conducted in animal or cell models, (ii) qualitative in nature (reviews, case reports, newspaper articles, editorials, commentaries, book chapters), or (iii) pre-prints or unpublished research.

Data Extraction, Assessment, and Synthesis
The following data were independently extracted by two reviewers (DDM and DMM) from the eligible studies: publication details (first author, year of publication, country); study details [design, sample size, demographics (age, sex)]; disease outcome [CKD (of unspecified cause), diabetic kidney disease (DKD)/HTN-associated CKD/HIV-associated nephropathy (HIVAN) or ESKD]; population [general or high-risk subgroups (HTN, DM and HIV)]; participant clinical characteristics [body mass index (BMI), C-reactive protein (CRP), smoking status, alcohol consumption, lipid profile (low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides and total cholesterol)]; clinical outcomes (diagnostic criteria, classification/staging, medication status); CKD diagnostic criteria (eGFR or proteinuria/albuminuria and eGFR equation used); miRNA analysis [sample type, molecular techniques, inclusion of screening and validation cohorts, expression pattern (upregulated or downregulated) and normalization control used]. Any inconsistencies or disagreements were resolved by discussions or consultation with a third author (CG). Furthermore, we assessed the quality of studies using the Newcastle-Ottawa Quality Assessment Scale for observational studies (NOS) tool [24]. The assessment was done based on a critical appraisal of three domains, namely: (i) participant group selection, (ii) how comparable the groups are, and (iii) determination of the exposure of interest. The quality of evidence was then assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework [25]. Given that very few studies investigated the association of the same miRNA with CKD risk or markers of kidney function or damage in the same sample type, making use of different normalization controls and miRNA quantification techniques as well as variabilities in disease outcome measures, and attempting to pool studies was meaningless. We, therefore, instead opted for a narrative synthesis of evidence.

Search Results
We obtained a total of 2860 related citations (663 from Scopus, 906 from EBSCOhost, 693 from Web of Science and 598 from PubMed) from database searches. Of these, 2732 citations were excluded for various reasons ( Figure 1). The remaining 128 articles were further assessed for eligibility and 53 studies were subsequently excluded from the review because they were irrelevant to our review (n = 44); they did not report on the quantitative reverse transcription polymerase chain reaction (RT-qPCR) normalization controls used (n = 7), did not include a control group (n = 1) and did not meet our reporting language restrictions (n = 1). Ultimately, 75 studies fulfilled the eligibility criteria and were retained for the systematic review ( Figure 1). The eligible studies were classified according to disease outcome, as follows: CKD (n = 18), DKD (n = 51) and HTN-associated CKD (6). Database searches did not return studies reporting on miRNA profiles in humans with HIVAN.

Characteristics of Included Studies
Tables 1-3 detail the main characteristics of studies that were included in the systematic review according to disease outcome. Table 1 summarizes the 18 studies that quantified miRNA expression patterns in CKD compared to controls, whilst Table 2 is a summary of the 51 studies that quantified miRNA expression patterns in DKD relative to controls. A summary of the six studies that quantified miRNA expression patterns in individuals with HTN-associated CKD compared to controls is shown in Table 3. All 75 studies were published between 2013 and 2022, and from diverse geographical locations, including China (n = 26), the United States of America (n = 6), Spain (n = 4), Egypt (n = 6), South Africa (n = 1), Germany (n = 3), Italy (n = 2), Austria (n = 2), Japan (n = 3), Iran (n = 3), Belgium (n = 1), Sweden (n = 1), Turkey (n = 1), Poland (n = 1), Bahrain (n = 2), Brazil (n = 2), United Kingdom (n = 1), India (n = 2), Romania (n = 2), France (n = 1), Canada (n = 1), Ireland (n = 1), Republic of Korea (n = 1), Netherlands (n = 1) and Malaysia (n = 1). The design of most studies was either case-control or cross-sectional, with study participant numbers ranging between 28 to 1385 in CKD, 11 to 1018 in DKD and 30 to 150 in HTNassociated CKD. The included studies used varying diagnostic methods to classify kidney disease, with 20% of the studies using only eGFR to classify CKD, 42% defined CKD by the level of albuminuria alone, whilst 34% of the studies used both albuminuria and eGFR to classify CKD. The remaining 4% of studies were not clear on the methods used for CKD classification. For estimation of GFR, most studies used the Modification of Diet in Renal Disease Study (MDRD) equation [26] (44% of the studies) and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [27] (38% of the studies). Of those included, 26% of the studies further validated their findings in a separate cohort [CKD (n = 3), DKD (n = 14) and HTN with CKD (n = 1)].      Of the included studies, 29% performed initial miRNA expression discovery using next-generation sequencing (NGS) techniques [CKD (n = 2); DKD (n = 2), HTN with CKD (n = 4)] and microarrays [CKD (n = 1); DKD (n = 15)], followed by a validation step using a PCR-based technique. PCR-based techniques were used for the quantification of miRNAs in instances where the miRNA was already identified by NGS and microarrays in previous studies. A wide range of normalization techniques were employed by the studies included in this review, with 31% using an exogenous spike-in control containing non-mammalian synthetic miRNAs such as Caenorhabditis elegans-miR-39 (cel-miR-39) (n = 22) and 52% of the studies used endogenous controls such as small non-human ubiquitous miRNA (U6) (n = 28) or miR-16 (n = 8). Of the included studies, 15% of the studies used more than one Only 17% of the studies quantified miRNA expression in two or more sample types, and this included two studies in CKD, 10 in DKD and one in HTN-associated CKD.
The expression patterns of 288 miRNAs were investigated across the 75 studies included in this review. Of the 288 miRNAs, 67 miRNAs were evaluated in populations with prevalent CKD, with 53 miRNAs found to be dysregulated (25 downregulated and 28 upregulated). Of the 193 miRNAs evaluated in populations with DKD, 155 miRNAs were found to be dysregulated (67 downregulated and 88 upregulated), whilst 13 (10 downregulated and 3 upregulated) of the 28 miRNAs evaluated in populations with HTN-associated CKD were dysregulated. The dysregulation discussed below refers only to miRNAs evaluated in three or more studies, of which HTN-associated CKD had none.

Dysregulated miRNAs in CKD
Of the 53 differentially expressed miRNAs in individuals with CKD, four miRNAs (miR-126, miR-223, miR-155 and miR-21) were reported in at least three studies (Table 4). Of these, serum miR-126 [21,28,32] and serum miR-223 [21,28,32,43] were consistently downregulated in individuals with CKD across studies. Inconsistent expression patterns were observed for miR-155, two studies found the miRNA to be downregulated in serum and urine, [29,30] and two studies found it upregulated in serum [30,31] in CKD, with one study reporting no difference [41] in the expression of miR-155 in both urine and serum, in individuals with and without CKD. Of the studies reporting on the expression pattern of miR-21 in CKD, one study showed a downregulation in this miRNA in serum sample [30], whilst two studies evaluated miR-21 in urine and urine exosomes and found it upregulated [30,34], and one study found no difference in urine and plasma samples between the case and control groups [40].

Discussion
Recent studies have demonstrated that miRNAs are key mediators in the pathophysiology of CKD, suggesting that circulating miRNAs have potential utility as alternative markers for early detection and progression of CKD, as well as monitoring treatment responses. Circulating and urinary miRNAs are ideal minimally-or non-invasive biomarkers because they are stable in body fluids and exosomes and can be detected using validated techniques for quantification. However, miRNA profile studies in humans have shown contradictory results, with few miRNAs being consistently dysregulated across studies. We performed a systematic review of studies that evaluated miRNA expressions in CKD in the general population, and high-risk subgroups (individuals with HTN and DM). We found that two miRNAs (miR-126 and miR-223) were consistently downregulated in the general population with CKD, whilst miR-21 and miR-29b were consistently upregulated and let-7a-3p and miR-30e were consistently downregulated in individuals with DKD, in whole blood, plasma, serum, urine, or urinary exosomes. Although showing inconsistent data, miR-155, miR-192, miR-15a-5p, miR-29a, miR-29c were also commonly quantified in the studies included in this review. Of note, only a few studies quantified miRNA expression in individuals with HTN-associated CKD, and reported inconsistent findings and none in HIVAN.
MiR-126 is endothelial cell-specific and promotes vascular integrity and angiogenesis via regulation of vascular endothelial growth factor (VEGF) signalling and, as a result, inhibits vascular inflammation [100]. MiR-126, which was downregulated in CKD in the general population [21,28,32] and individuals with DM [48,76] and HTN [99] when quantified in serum and whole blood samples, was inversely associated with prevalent CKD [32,33] and positively associated with eGFR [21,33]. A prospective study showed that lower levels of miR-126 were associated with an increased risk of developing CKD and rapid decline in kidney function over a period of five years [33]. Zhou et al. also demonstrated that miR-126 has an atheroprotective role, as it increases vascular smooth muscle cells (VSMCs) turnover, thereby regulating the contractile phenotype of VSMC [101]. Taken together, the downregulation of miR-126 may result in vascular dysfunction, which is very common in early-stage CKD. MiR-126 may therefore be a potential biomarker for the early identification of CKD and a potential target for the prevention or treatment of CKD-related vascular complications. Contrary to these findings, a few studies reported upregulated expression of miR-126 individuals with DKD [53,60,64,76] when quantified in urine, plasma as well as serum. It is plausible that this may be a compensatory mechanism resulting in increased release of miR-126 when microvascular endothelial cells are exposed to stressful conditions [102]. Indeed, Beltrami and colleagues used in vitro analyses and showed that miR-126 is released from glomerular endothelial cells in response to DKDrelated cytokines [53].
Similarly to miR-126, miR-223 also plays a role in the regulation of VSMC proliferation [103]. This miRNA was consistently downregulated in the general population with CKD in serum samples [21,28,32,43]. Moreover, lower levels of this miRNA were associated with lower levels of eGFR [32], and prevalent CKD [21]. Although commonly considered to be involved in inflammatory pathways, evidence also suggests a protective role of miR-223 in VSMCs, through the inhibition of calcification by the regulation of interleukin-6 (IL-6)/signal transducer and activator of transcription 3 (STAT3) [63]. These findings imply that increased levels of miR-223 may improve kidney function, and thus may serve as a therapeutic strategy to improve CKD outcomes in the general population.
The involvement of miR-21 in kidney fibrosis is well established, although the mechanisms involved have not been completely clarified. miR-21 acts as a pro-fibrotic factor, and its upregulation induces kidney fibrosis through TGF-β signalling pathway regulation [104]. Consistent upregulation of miR-21 in individuals with DKD [51,61,64,67,71,76,77,89] as well as its inverse association with eGFR [71] and positive association with albuminuria [71] have been reported. Moreover, increased levels of miR-21 were associated with rapid progression to ESKD over a 10-year follow-up period [77]. Correspondingly, in vitro and in vivo knockdown of miR-21 ameliorated DKD by reducing albuminuria, kidney inflammation, podocyte loss and interstitial fibrosis, suggesting its value as a potential therapeutic target against DKD progression [105]. Although the observed findings imply that increased expression of miR-21 may be associated with the development and progression of DKD, the data were inconclusive in the case of CKD. In the general population with CKD, contrasting results were reported, with miR-21 upregulated in urine [30] and urine exosomes [34] but downregulated in serum samples [30]. Moreover, Fujii and colleagues found that increased levels of miR-21 were positively associated with eGFR and inversely associated with the risk of CKD in the general population [22]. Donderski et al. (2021) explained that the lower levels of miR-21 detected in serum samples of individuals with CKD may be as a result of suppression caused by increased fibrosis and TGF-β activity [30]. This is in line with the findings by Sun et al. (2018), using a murine kidney fibrosis model, they observed that miR-21 is the main regulator of fibroblasts activation through an auto-regulatory loop between miR-21 and programmed cell death protein 4 and activated protein 1, therefore suggesting that miR-21 may act as pro-or anti-fibrotic depending on the cell type [106]. It has been suggested that identification of the cellular source of miRNAs would be helpful instead of the biofluid sample to link the miRNAs to the specific disease process [107].
The miR-29 family has been well studied with regard to the TGF-β signalling pathway [104]. Two studies included in our review reported that miR-29b was upregulated in the urine samples of individuals with DKD [50,53], although one study reported no difference in the expression level of miR-29b when quantified in urine supernatant [75]. The lack of difference could be explained by the relatively lower abundance of miR-29b in urine supernatant sample reported in this study [75]. In individuals with HTN, increased expression of miR-29b was positively associated with albuminuria and inversely associated with kidney function [96]. However, these findings are contrary to previous studies that have reported on the protective role of this miRNA in DKD. Chen et al. (2014) showed that knockdown of miR-29b in diabetic mice was associated with increased albuminuria and TGFβ mediated fibrosis whereas overexpression of miR-29b attenuated kidney fibrosis in DKD through the regulation of TGFβ1/Smad3 pathway [108]. The upregulated expression of miR-29b in DKD observed in the included studies in our review may be due to the compensatory release of miR-29b. Beltrami and colleagues used in vitro analyses and observed increased release of miR-29b from glomerular endothelial cells in response to DKD-related cytokines [53]. Regarding the expression profile of miR-29a and miR-29c, contradictory results were observed when quantified in various samples of individuals with DKD. miR-29a was downregulated in plasma samples of individuals with severe DKD [51] and inversely associated with rapid progression to ESKD over a 10-year follow-up period [77], suggesting that this miRNA may have a protective effect against the progression of DKD. However, when quantified in urine supernatant, upregulated expression of miR-29a was observed in individuals with DKD [75]. Studies have previously highlighted that urine supernatant miRNAs inversely reflect intracellular miRNAs [109], which could be the possible reason for the discrepancy, however, tissue expression of miR-29a was not analyzed in this study [75]. On the other hand, Guo et al. (2017) [62] analyzed the expression of miR-29c in three different samples of individuals with DKD relative to those without, and found downregulated expression in urinary sediments and kidney tissues but upregulated expression in plasma. Cui and Cui (2020) found that relative to blood, urinary miRNAs highly reflected kidney tissue miRNAs and suggested that urine should be a better sample for kidney miRNAs analysis [110].
MicroRNAs miR-30e and let-7a were consistently downregulated in individuals with DKD relative to those without DKD, suggesting that increased expression of these miRNAs may have protective effects in the kidney and therefore may serve as possible diagnostic and prognostic markers of DKD. Accordingly, previous evidence suggests that the let-7 family of miRNAs may be a negative regulator of kidney fibrosis in DKD [111]. Muralidharan and colleagues [38] validated the expression of let-7a in an Alb/TGFβ mouse model of CKD and found that the miRNA was significantly downregulated further suggesting its possible role in the development or progression of CKD. MicroRNAs in the miR-30 family are highly enriched in kidney podocytes cells where they are involved in regulatory roles and are essential for structural and functional homeostasis [112]. In vitro and in vivo experimental studies showed that the expression of miR-30e was significantly decreased in those with DKD whereas overexpression of miR-30e was protective against the development of kidney fibrosis in DKD suggestive the potential role of this miRNA as a therapeutic target [113].
The miR-155 was commonly analyzed in the studies included in this review. Experimental studies have shown that suppressing miR-155 expression in DKD mice protects against kidney damage, attenuates hyperglycaemia-induced kidney damage and downregulates IL-17 expression by enhancing the suppression of cytokine signalling 1 (SOCS1) [114]. In line with these, upregulated expression of miR-155 was observed in the general population with CKD [30,31] and individuals with DKD 53. However, downregulated expression of miR-155 was commonly reported in individuals with DKD [19,47,52], as well as in the general population with CKD [29,30]. Furthermore, Donderski and colleagues reported on the positive association of miR-155 and eGFR [30]. These findings suggest that miR-155 may also play a role in the development of DKD. Wang et al. (2018) demonstrated that miR-155 is involved in the regulation of the autophagic process in DKD through the regulation of a signalling loop p53/miR-155-5p/Sirt1 and may therefore serve as a therapeutic strategy for DKD [115]. miR-192 has been shown to have a protective effect against kidney fibrosis. Downregulated expression of miR-192 was observed in individuals with DKD [49,70,71], and reduced levels of miR-192 were positively associated with kidney function [71] and inversely associated with kidney damage [70,71], suggesting that miR-192 levels may decrease with the increasing level of albuminuria and a decline in kidney function. Consistently, in vivo studies have shown that loss of miR-192 was associated with the development and progression of DKD through exacerbation of kidney fibrosis by enhancing TGF-β1 signalling pathway [116]. These findings suggest that reduced expression of miR-192 in the early stage may be associated with the development of DKD and therefore may serve as an early indicator of DKD. However, contrary to these findings, increased expression of miR-192 in individuals with DKD relative to controls has been observed in a few studies [64,76]. Jia and colleagues reported that the expression of miR-192 was increased during the early stages of DKD and decreased in the advanced stages of DKD [66]. They further showed that miR-192 was positively correlated with albuminuria and TGF-β1 levels [66], suggesting a profibrotic role of this miRNA. Jenkins et al. (2012) highlighted that miR-192 is pleiotropic, involved in multiple important roles in the kidney, and its role as an antifibrotic or profibrotic factor may be cell dependent [117].
This review provides an overview of miRNA dysregulation in CKD, including diabetic and hypertensive-related CKD in humans. It also highlights miRNAs that are associated with CKD and its clinical indicators. A few miRNAs showed consistent expression patterns in CKD relative to controls, whilst most of the frequently studied miRNAs showed contradictory findings. These discrepancies may be partly explained by the technical and methodological inter-study variabilities, such as the use of different biological samples, sample handling and processing procedures, miRNA extraction protocols, detection and quantification techniques, and normalizing controls [118]. Although the majority of included studies quantified miRNA expression in blood, recent evidence points to the superiority of urine miRNAs to serum/plasma miRNAs for CKD diagnosis with the non-invasive nature in which urine samples are collected, adding to its preference [119]. Another challenge is the lack of a standardized normalization control for miRNA expression studies. Although a wide range of endogenous and exogenous miRNAs are employed as normalizers, emerging evidence suggests that the use of synthetic spike in controls such as cel-miR-39 is preferable [120]. Therefore, to be able to identify reliable miRNA biomarkers, research findings need to be reproducible and comparable between studies. This can be achieved when normalization controls have been validated, and there is a standardization of robust protocols for sample processing and extraction [121].

Strength and Limitations
The main strength of this review is its comprehensive report of miRNAs dysregulated in CKD, their association with CKD as well as clinical markers of CKD in the general population as well as in high-risk individuals with HTN and/or DM for the very first time.
The review also provides a list of miRNAs that have been frequently studied in diverse geographical areas and showed consistent expression patterns across studies in CKD and DKD and therefore are worthy of future research.
The studies included in the review had their own shortcomings and, as such, impacted the review's overall quality of evidence. The most important limitation of the review was the inability to perform a pooled meta-analysis of our studies due to various factors, including insufficient raw data on fold changes or relative expression of miRNAs, technical and methodological variabilities between studies, such as the use of different biological samples, normalization control used, and different miRNA quantification techniques. Moreover, due to insufficient data, we could not report on the expression patterns of miRNAs across different stages of CKD. The language restriction on the inclusion criteria may have excluded other relevant studies, thus biasing our findings. Additionally, there were differences in the classification of CKD, wide sample size ranges, variability in participant demographic factors such as age, sex proportion, and race, as well as environmental and regional factors between studies.

Conclusions
MiRNAs detected in biofluids are promising as potential biomarkers of disease diagnosis and therapeutic targets for future clinical applications. However, understanding their role in CKD pathophysiology and how their expression pattern is regulated is still in its infancy. As such, further research is required to fully elucidate their roles before any extrapolation for clinical use. Prevention and early detection of CKD has been a topic of interest for many researchers and clinicians in the field. This review highlights several dysregulated miRNAs that were frequently studied and showed consistent findings across studies in CKD (miR-126 and miR-223) or DKD (miR-21, miR-29b, let-7a and miR-30e) with a potential for clinical application in CKD diagnosis/prognosis in the future. This consistent alteration of miRNAs with CKD/DKD and their stability and detectability in bodily fluids suggests that these miRNAs are promising potential non-invasive or minimally invasive diagnostic markers for early detection and therapeutic targets of CKD/DKD and warrant further scrutiny in future investigations. Besides these specific miRNAs, miR-155, miR-192, miR-15a-5p, miR-29a and miR-29c, despite their inconsistent expression patterns reported in different studies, were commonly dysregulated in CKD and/or DKD, and therefore may also play an important role in CKD. As such, their further exploration is warranted. Furthermore, it may be worthwhile for future studies to focus on identifying target genes and pathways of these frequently studied miRNAs, to get a complete understanding of their role in the development and progression of CKD, as well as to assess their potential value as diagnostic markers or therapeutic targets. Data Availability Statement: All data generated during this study is available upon reasonable request from the corresponding author.