Epigenetics in kidney diseases

Abstract

Epigenetics examines heritable changes in DNA and its associated proteins except mutations in gene sequence. Epigenetic regulation plays fundamental roles in kidney cell biology through the action of DNA methylation, chromatin modification via epigenetic regulators and non-coding RNA species. Kidney diseases, including acute kidney injury, chronic kidney disease, diabetic kidney disease and renal fibrosis are multistep processes associated with numerous molecular alterations even in individual kidney cells. Epigenetic alterations, including anomalous DNA methylation, aberrant histone alterations and changes of microRNA expression all contribute to kidney pathogenesis. These changes alter the genome-wide epigenetic signatures and disrupt essential pathways that protect renal cells from uncontrolled growth, apoptosis and development of other renal associated syndromes. Molecular changes impact cellular function within kidney cells and its microenvironment to drive and maintain disease phenotype. In this chapter, we briefly summarize epigenetic mechanisms in four kidney diseases including acute kidney injury, chronic kidney disease, diabetic kidney disease and renal fibrosis. We primarily focus on current knowledge about the genome-wide profiling of DNA methylation and histone modification, and epigenetic regulation on specific gene(s) in the pathophysiology of these diseases and the translational potential of identifying new biomarkers and treatment for prevention and therapy. Incorporating epigenomic testing into clinical research is essential to elucidate novel epigenetic biomarkers and develop precision medicine using emerging therapies.

1. Introduction

Kidney diseases relate to the organs’ inability to clear the blood of toxins, which are multistep processes associated with the accumulation of numerous molecular alterations. These molecular changes impact renal cell function and its microenvironment. Kidney diseases have been associated with different genetic alterations (mutations, loss of heterozygosity, deletions, insertions, aneuploidy, etc.) [1]. These lead to altered gene expression that ultimately result in decreased renal function. However, the landscape of genetic alterations is insufficient to explain the pervasive gene expression changes and alterations of renal function in some kidney diseases, i.e., acute kidney injury, chronic kidney disease (CKD), diabetic kidney disease and renal fibrosis. Recent advances to identify alternations in renal cell epigenome that drive the onset and progression of renal diseases have extended our understanding of the pathophysiology of kidney disease progression in general.

Epigenetic alterations are heritable traits that impact the phenotype by interfering with gene expression independent of DNA sequence [2]. Epigenetic changes are pervasive in kidney diseases and likely responsible as a source of phenotypic variation [3]. Major epigenetic mechanisms include DNA methylation and hydroxymethylation, histone protein post-translational modification, chromatin remodeling, noncoding ribonucleic acid (ncRNA) regulation and RNA editing [4,5]. Collectively, epigenetic mechanisms determine the chromatin architecture, accessibility of genetic loci to transcriptional machinery and gene expression.

Epigenetic effects on cellular and physiological phenotypic traits may result from external and environmental factors, or be part of normal development. Epigenetic events usually act together with other molecular processes in normal states or disease to have stable and long-term gene expression and functional alterations [6,7]. For example, epigenetic alterations that impact DNA methylation and transcription factor binding can explain genome-wide transcription dysregulation independent of genetic variation [8–10]. Genome-wide characterization of epigenetic aberrations with functional impacts on gene expression in individual kidney disease remain key targets of interest, which is crucial for the development of new therapeutic agents to reverse specific alterations to the epigenetic landscape and to identify new biomarkers for different kidney diseases.

This chapter summarizes the epigenetic mechanisms in four kidney diseases. These include acute kidney injury, CKD, diabetic kidney disease and renal fibrosis. We focus on current knowledge about the genome-wide profiling of DNA methylation and histone modification, and epigenetic regulation of specific gene(s) in their pathophysiology and the translational potential of identifying new biomarkers and treatment for the prevention and therapy of these disorders as well as end-stage kidney disease.

2. Epigenetic mechanisms in kidney cell biology

There are three basic epigenetic mechanisms related to kidney cell biology and kidney development, including DNA methylation, histone modification and microRNAs.

2.1. DNA methylation

The methylation of DNA at CpG dinucleotides was among the first genome modifications described and is an attractive mechanism for the regulation of kidney cell biology and kidney development [11]. DNA methylation involves the covalent addition of a methyl radical (CH3) from S-adenyl methionine (SAM) to the 5-carbon on cytosine residues in CpG dinucleotides of DNA to form 5-methylcytosine (5mC) [12,13]. DNA methylation is catalyzed by a family of DNA methyltransferases (DNMTs) [14,15]. In vertebrates, there are five known DNMTs that differ in structure and function. The ubiquitously expressed DNMT1 functions to maintain the DNA methylation patterns established by the DNMT3 subfamily, comprising DNMT3a and DNMT3b, on unmethylated DNA [14–16] during DNA replication and DNA repair [14–18]. DNMT3a and DNMT3b were thought to establish the methylation patterns at the early development which is further maintained through somatic cell divisions by DNMT1, acting on the hemimethylated CpG sites generated by DNA replication [19–21]. Mice lacking Dnmt3a die at about 4 weeks of age, while the expression of DNMT3b is less in majority of differentiated tissues and knockout of Dnmt3b induces embryonic lethality [15,22], which suggest that DNMT3a is required for normal cellular differentiation, while DNMT3b is required during early development. The cofactor DNMT3L1 stimulates the activity of DNMT3A and DNMT3B, but by itself lacks enzymatic activity [23,24]. Recently, mouse studies show that DNMT3L is expressed during gametogenesis and is required for the establishment of maternal genomic imprints [24,25] and mice lacking DNMT3L die early during development [24]. The fifth member of the DNMT family, DNMT2, has very weak activity toward DNA [26,27].

DNA methylation or hypomethylation on specific CpG sites plays critical roles during different stages of normal kidney development, whereas aberrant methylation or hypermethylation on specific CpG sites can occur in disease kidneys [28]. Under normal conditions, dynamic changes in DNA methylation, lead to stable and unique patterns that regulate tissue-specific gene transcription in somatic cells. It has been showed that maintenance of DNA methylation plays a key role during nephron development [29]. Loss of DNMT1 in nephron progenitor cells led to a strong reduction of DNA methylation in all cells originating from the cap mesenchyme (CM). Deletion of DNMT1 in nephron progenitor cells (in contrast to deletion of DNMT3a or DNMT3b) mimics nutritional models of kidney growth restriction and results in a substantial reduction of nephron number as well as renal hypoplasia at birth [29]. In disease conditions, these methylation patterns have been known to change; either preceding the disease, or occurring as a consequence of it. For example, site-specific DNA methylation changes have been detected in patients with CKD in general and diabetic nephropathy [30,31]. Since DNA methylation patterns are known to be fairly stable and unique in differentiated cells, evaluating the differences in methylation status on regulatory DNA sequences (promoters, insulators and enhancers) in the genome of patients and animal models could extend our understanding of kidney development and the pathophysiology of kidney diseases.

2.2. Histone modifications

Histones are highly conserved proteins in chromatin of eukaryotic cells, which pack and order the DNA into structural units called nucleosomes. Histone modification involves distinct types of covalent post-translational modifications (PTMs) to core histones, including acetylation, methylation, phosphorylation, ubiquitination, sumoylation, citrullination, biotinylation, crotonylation and ADP ribosylation [32,33], which usually occur on critical amino acids of histone tails that extend out of the nucleosome [34]. Histone modifications result in either open or closed conformations of chromatin and drive different accessibility to transcription factors and regulatory proteins to specific genes [35]. Histone acetyltransferases/deacetylases and methyltransferases/demethylases can either write or erase histone acetylation and methylation, respectively [36,37]. Acetylation of histones strongly results in transcription activation, whereas deacetylation of histones mediated by histone deacetylases (HDACs) represses gene expression in almost all experimental systems [38,39]. Methylation of histones can either activate or repress gene transcription, depending on which specific residues of histones are modified [33,40–42]. In addition, modified histones can be selectively bound by different chromatin remodeling factors. For example, bromodomain-containing proteins (BRDs) selectively read and bind to acetylated histone lysine residues to promote gene expression [43], while chromodomain proteins selectively read and recognize methylated lysine residues to promote hetero-chromatin formation and to suppress gene transcription [44].

Many of these epigenetic regulators involving in histone modifications are dysregulated to affect cell homeostasis pathways in diseased cells and tissues [45]. Because histone marks have stable covalent structures, they can be inherited during cell division and DNA duplication to potentially serve as disease markers [46]. Genome-wide epigenetic analysis is essential to determine the source of alterations to the distribution of histone markers, which could provide useful information for a better understanding of epigenetic regulation of cellular processes in kidneys. In addition, histone modifications on specific gene(s) could affect its expression during kidney development and kidney disease progression, For example, the methylation of histones on lysine (K) or arginine (R) residues has been recognized as a key epigenetic mechanism regulating gene expression during kidney development and disease [47]. Therefore, to detect various histone modifications on specific genes during kidney disease progression should help us to develop histone modifying enzyme-targeted therapy for kidney diseases.

2.3. Non-coding RNA (ncRNA) and MicroRNAs

Non-coding RNAs (ncRNAs) are functional RNA molecules that are transcribed from DNA but not translated into proteins [48,49]. High-throughput sequencing technology confirmed that over 98% of the human genome is transcribed into ncRNAs, which are divided into two main groups: the small non-coding RNAs (<200 nucleotides) and the long non-coding RNAs (lncRNAs) (>200 nucleotides) [50]. In general, ncRNAs play a role in hetero-chromatin formation, histone modification and DNA methylation, leading to regulate gene expression at the transcriptional and post-transcriptional level [51]. Epigenetic related ncRNAs include miRNA, siRNA, piRNA and lncRNA.

miRNAs comprise a large family of highly conserved ncRNAs of ~22 nucleotides in length, acting to target 3′-UTR of many mRNAs to induce mRNA degradation and repress the translation [52]. The function of miRNA depends on many factors, including the subcellular location, relative expression and processing of the miRNA and the regulation of miRNA-target interactions. Biosynthesis of miRNAs begins in the nucleus, where a primary miRNA (pri-miRNA) transcript is first transcribed by RNA polymerase II. Then, the Drosha-Dgcr8 complex processes pri-miRNAs into precursor miRNAs (pre-miRNAs). Subsequently, pre-miRNAs are exported into the cytoplasm and processed to mature miRNA by Dicer [53–55]. A single miRNA can potentially target hundreds or thousands of mRNAs, regulating crucial functions in numerous biological processes, including development, differentiation, stress responses and apoptosis.

Long non-coding RNAs (lncRNAs) refer to a group of large ncRNAs (>200 nucleotides and up to ~100kb in length) that have many similarities with mRNAs. lncRNAs can also modulate gene expression at the transcriptional, post-transcriptional, and translational levels. In addition, lncRNAs may serve as the microRNA host or sponge that inhibits their binding to the actual mRNA targets [56]. Furthermore, lncRNAs act as modular molecules with individual domains that enable them to specifically associate with DNA, RNA, and/or proteins to affect gene expression [57]. For example, lncRNAs can bind to DNA and to chromatin remodeling complexes, which are associated with complex alterations in the distribution of nucleosomes [58]. The functional role of specific ncRNAs on kidney epigenetics is an active area of research.

3. Epigenetics in acute kidney injury (AKI)

Acute kidney injury (AKI) is a common renal disorder, which can occur in a number of ways, including from decreased blood flow caused by vasculitis (inflammation of blood vessels), organ failure, surgery or blockage of the urinary tract, or from acute allergic reaction to certain medications, or even from trauma sustained in a car accident. AKI is a critical complication in patients with increased mortality [59–61]. Kidney injury results cell death and inflammation in the initial phase, following by a recovery phase involving epithelial cell de-differentiation, proliferation and re-differentiation, which may lead to functional and structural recovery of kidney or lead to CKD and end-stage renal disease (ESRD) [62–66]. Although great progress has been made in understanding the cellular and molecular basis of AKI, mortality remains high [67,68]. Currently, there are no effective treatments for AKI other than supportive care such as dialysis. Therefore, it is urgent to understand the pathogenesis of AKI and develop effective therapeutic strategy for AKI treatment.

Abnormal epigenetic modifications, which may be triggered as cellular adaptive responses toward hypoxia, oxidative stress and mitochondrial injury, on DNA methylation, histone modification and non-coding RNAs have been reported to be involved in the pathogenesis of AKI [69–73]. However, how epigenetics is involved in the onset, progression, and treatment of AKI is still in its infancy. In this section, we discuss the links between epigenetic mechanisms and AKI, and the known epigenetic changes to kidney injury and repair in AKI.

3.1. Pathogenesis of AKI

Cellular and molecular mechanisms of AKI have been extensively investigated in a variety of experimental models and the pathogenesis of AKI involves in the injury and death of renal tubular cells, particularly of cells in the proximal tubule [74,75]. Following injury, a repair response is activated, which involves in epithelial cell de-differentiation, proliferation and re-differentiation [64]. The plasma membrane is a primary site of damage in AKI. With necrotic cell injury, compromise of cell membrane integrity results in extrusion of intracellular contents, and with apoptosis there is translocation of phospholipids from the inner to the outer membrane leaflet, followed by membrane blebbing [76]. Major components of the nucleus including DNA and chromatin, transcription factors, and epigenetic regulators of gene expression are among the key targets in AKI [77]. Cell cycle regulatory proteins are also important in the pathogenesis of AKI. Recent studies have highlighted the importance of epigenetic modifications, including microRNAs, as determinants for the outcome of AKI and the AKI-to-CKD transition [78,79]. Alterations in the cytoskeleton and the endoplasmic reticulum stress due to accumulation of unfolded proteins have also been described in the pathogenesis of AKI [80].

In order to move sodium and other solutes against strong electrochemical gradients, epithelial cells of the nephron-in particular, proximal tubular cells-are rich in mitochondria [81]. The spectrum of experimental rodent models of AKI demonstrates widespread mitochondrial injury and the onset of mitochondrial pathology typically precedes detectable loss of renal function [82,83]. Although further studies are needed, therapies that inhibit fission, favor fusion, or enhance mitophagy of mitochondrial may be beneficial in one or more phases of AKI. Peroxisomes are small single-membrane organelles that are rich in enzymes such as catalase. They are highly dynamic and metabolically active and are required for regenerating tubular epithelial cells. The role of peroxisomes in AKI is further substantiated by findings that overexpression of Sirt1, an NAD dependent protein deacetylase that regulates intracellular metabolism and attenuates reactive oxygen species (ROS)-induced cell death, mediates these protective effects through maintaining peroxisomal number and function [84].

3.2. DNA methylation in AKI

DNA methylation is the typical epigenetic modification that affects DNA [85,86] and is not only occurs predominantly in CpG dinucleotides but also occur at low frequency in non-CpG sites [87,88]. Abnormal DNA methylation has been linked to human diseases, such as cancer. Abnormal DNA methylation has been also reported in AKI.

3.2.1. Genome-wide profiling of DNA methylation in AKI

The development of improved DNA methylation arrays, high-throughput sequencing technologies, and powerful approaches for data analysis prompt genome-wide profiling of DNA methylation in AKI. A recent study found that the global level of 5-hydroxymethylcytosine (5hmC) was reduced in mouse kidney insulted by ischemia reperfusion (IR), whereas the level of 5-methylcytosine (5mC) was not changed [89]. The profiling of DNA hydroxymethylome revealed that 5hmC was enriched in genic regions but depleted from intergenic regions in IRI mouse kidneys as examined by hydroxymethylated DNA immunoprecipitation (hMeDIP-seq). The gene body enrichment of 5hmC is positively associated with the gene expression level in mouse kidneys. In particular, genes associated with IRI showed significantly higher 5hmC enrichment in their gene body regions than those unchanged genes in IRI mouse kidneys [90]. Another study using reduced representation bisulfite sequencing (RRBS) to identify the DNA methylome of renal tissues during IRI found that both acute and chronic IRI induced genome-wide DNA demethylation and the regulation of promoter methylation was an important mechanism underlying persistent alteration of gene expression [91]. The differences between the two studies might be due to the different detection methods used and different degrees of injury induced. The latter study using RRBS also identified 200 and 191 genes that had differentially methylated promoter regions in acute and chronic IRI, respectively. Of these genes, 79 were common to both conditions, and further investigation indicated that promoter methylation alterations contributed to differences in their expression. A third study using RRBS to analyze genome-wide DNA methylation during cisplatin-induced AKI identified 215 differentially methylated regions in cisplatin-treated kidneys compared to the control animals. Treatment with demethylating compound 5-aza-2′-deoxycytidine increased cisplatin-induced apoptosis in a rat kidney proximal tubular cell line. Conditional knockout of DNMT1 in renal proximal tubules exhibited more severe AKI during cisplatin treatment than wild-type mice. In addition, cisplatin treatment induced hypomethylation of interferon regulatory factor 8 (Irf8), leading to increase its expression, which contributed to renal tubular cell apoptosis [92].

3.2.2. DNA methylation in select genes related to AKI

In addition to genome-wide profiling of DNA methylation, a number of experimental and clinical studies indicate the existence of a link between DNA methylation at specific gene loci with AKI. It has been reported that demethylation of a cytosine residue occurs in the INF-γ response element within the compliment C3 promoter in response to cold ischemia in the rat kidneys and additional demethylation occurs during further warm reperfusion [93]. In addition, demethylation of the C3 promoter in rats with transplanted kidneys lasted for at least 6 months. A study to examine the DNA methylation of gene promoters in urines of kidney transplant patients found that DNA methylation levels in the calcitonin (CALCA) promoter was higher in urines of kidney transplant recipients compared to those in healthy controls. This study also found that patients with acute tubular necrosis (proven by biopsy) had higher DNA methylation levels than those patients with acute rejection and slow graft function [94]. With developed TaqMan-based assay, it has been reported that plasma levels of unmethylated DNA corresponding to the promoter of Slc22a12, a urate transporter specifically expressed in proximal tubular cells, were undetectable at baseline and were significantly elevated after acute kidney cortex necrosis [95]. In addition, it was observed that the promoter CpG methylation of KLK1 (encode renal kallikrein) in blood and urinary DNA from patients with established AKI was significantly increased compared to healthy/non-hospital controls [96].

3.3. Histone modifications in AKI

Histone modification is the second identified epigenetic mechanism that affects histones. Histones are highly conserved proteins found in eukaryotic cell nuclei that package and arrange DNA, which wraps around the eight histones, into structural units called nucleosomes [97]. There is increasing evidence that different types of AKI are associated with changes in histone modifications.

3.3.1. Histone acetylation and deacetylation in AKI

There is increasing evidence that different types of AKI are associated with changes in histone acetylation. A study showed that renal ischemia–reperfusion injury (IRI) induces the downregulation histone deacetylase 5 (HDAC5), which increases histone acetylation and bone morphogenetic protein 7 (BMP7) induction in the recovery phase, and highlights HDAC5 as a modulator of the regenerative response after ischemia [98]. Another study reported a progressive gene-activating H3 acetylation after 1 day of reperfusion, which could persist for 3 weeks [99]. In a mouse model of bilateral IRI, an initial increase of histone acetyltransferases, JADE1S, JADE1L, and HBO1 was found, which returned to baseline during renal recovery [100]. The temporal expression of JADE1S correlated with the acetylation of histone H4 on lysines 5 and 12, but not with acetylation of histone H3 on lysine 14, demonstrating that the JADE1S-HBO1 complex specifically marks H4 during epithelial cell proliferation [100]. Studies also demonstrated a role of histone modifications at the promoter HMG-CoA reductase (HMGCR) contributed to its upregulation. Upregulation of HMGCR may resulted in cytoprotective effects in ischemic AKI [101]. Patients with AKI had increased HMG-CoA reductase activity associated with increased H3K4m3 at HMGCR compared with patients without AKI, which supported this hypothesis. Furthermore, it has been found that activating transcription factor 3 (ATF3)-deficient mice had higher renal I/R-induced mortality, kidney dysfunction, inflammation and apoptosis compared with wild-type mice. Overexpression of ATF3 to the kidney rescued the renal I/R-induced injuries through recruitment of HDAC1 to the promoter regions of IL6 and IL12b, resulting in the suppression of IL6 and IL12b expression and inhibiting the inflammatory response [102]. In addition, upregulation of plasminogen activator inhibitor type 1 (PAI 1), which contributes to the susceptibility of male mice to ischemic AKI, was associated with a decrease in HDAC11 expression and increased acetylation of histone H3, suggesting that HDAC11 is a novel target for I/R injury [103].

Histone acetylation is also involved in other forms of AKI other than renal IRI. For example, decreased H3 acetylation has been reported in both in vivo and in vitro experimental models of septic AKI, and administration of dexmedetomidine reduced sepsis-induced AKI by decreasing TNF-α and MCP-1 and increasing BMP-7, which is associated with the decrease of HDAC2 and HDAC5, as well as the increase of histone H3 acetylation [104]. p300/CBP-associated factor (PCAF), a histone acetyltransferase, was overexpressed in the kidneys of lipopolysaccharide (LPS)-injected mice and was associated with increased acetylation of histones, such as H3K18Ac, leading to increase the expression of inflammatory genes in the kidneys of LPS-injected mice [105]. The expression of peroxisome proliferator activated receptor-γ co-activator 1α (PGC1α) in folic-acid-induced AKI mouse model was mediated through the inflammatory cytokine TWEAK, in which TWEAK promoted H3 deacetylation at the Ppargc1a promoter, leading to repression of PGC1α expression [106]. In addition, the downregulation of Klotho, another renoprotective factor, was also due to TWEAK mediated H3 deacetylation at the Kl gene promoter [107]. In cisplatin-induced AKI model, induction of the class III HDAC, SIRT1, occurred at 6 h after cisplatin administration and was also associated with deacetylation of histone H3 [108]. Overexpression of SIRT1 could protect against cisplatin-induced AKI by retaining peroxisome function [84]. Taken together, these studies support an important role of histone acetylation in various forms of AKI.

3.3.2. Histone methylation and demethylation in AKI

Histone methylation regulates transcription though creating docking sites for chromatin modifiers and has diverse outcomes-active, poised, or repressive status of chromatin and transcriptional marks. Compared with histone acetylation, the role of histone methylation in AKI and kidney repair has been much less studied. Methylation of histone lysine or arginine residues is modulated by methyltransferases and demethylases. Histone methyltransferases include two major types, lysine methyltransferase (KMT) and arginine methyltransferase (RMT).

Upregulation of enhancer of zeste homolog 2 (EZH2), a histone methyltransferase that specifically mediates trimethylation of lysine 27 on histone H3 (H3K27m3), has been documented in fibrotic kidneys from mice with unilateral ureteral obstruction (UUO) and in patients with CKD, suggesting that this methyltransferase has profibrotic functions [109,110]. Inhibition of EZH2 with 3-deazaneplanocin A (3-DZNeP) protects against renal tubular cell injury and death in murine models of AKI induced by I/R and FA. In addition, downregulation of EZH2 with either 3-DZNeP or its specific siRNA potentiates cell apoptosis in cultured renal proximal tubular cells. EZH2-mediated cell death is associated with epigenetic silencing of several genes including E-cadherin, tissue inhibitor of metalloproteases-3 (TIMP3) and Raf-1 kinase inhibitor protein (RKIP). As AKI can progress to CKD and promote CKD to develop to ESRD, this study further examined the role of EZH2 in renal fibrosis, a key pathological process for AKI progression to CKD. The development of renal fibrosis is accompanied by increased expression of EZH2 and vimentin, a mesenchymal marker, and decreased expression of E-cadherin, phosphatase and tensin homolog (PTEN) and Smad7, whereas pharmacological inhibition of EZH2 is able to stabilize the expression of E-cadherin, PTEN and Smad7 and prevent upregulation of vimentin in the kidney after UUO injury [109]. These data suggest that EZH2 is a critical mediator of acute and chronic kidney injury.

3.3.3. Other histone modifications in AKI

Knowledge of the role of histone phosphorylation in AKI is also very limited. One study demonstrated that histone H3 phosphorylation induced chromatin condensation and promoted the damage and death of renal proximal tubular cells in response to oxidative stress [111]. Prevention of H3 phosphorylation by inhibiting MAPK kinase (with the selective MAPK inhibitor PD98059) or by inhibiting poly (ADP-ribose) polymerase (PARP) (with 3-aminobenzamide) promoted renal tubular cell survival [111]. Phosphorylation of the histone H2A variant H2AX at serine 139, indicative of double-strand DNA damage, has been linked to AKI in various models, including cisplatin-induced AKI and renal IRI [112–114].

Histone crotonylation involves the addition of a crotonyl group to lysine residues on core histones by the actions of histone crotonyltransferases. Similar to histone acetylation, histone crotonylation neutralizes the positive charge of lysine residues and stimulates transcription; however, the mechanisms governing crotonylation versus acetylation remain largely unclear [115]. One study demonstrated an increase in histone crotonylation in kidney tissues in models of AKI induced by folic acid and cisplatin [116]. Further analysis indicated that histone crotonylation increased the expression of PGC1α and SIRT3, promoting protective pathways.

3.4. Non-coding RNAs in AKI

Non-coding RNAs act as crucial epigenetic modulators of gene expression other than translated into proteins. Non-coding RNAs, particularly microRNAs and lncRNAs, have emerged as important epigenetic regulators of AKI and kidney repair.

MicroRNAs have been shown to be involved in a wide range of biological processes [117–120] and accumulating evidence suggests that microRNAs are critical regulators of renal development and pathophysiology [121] in AKI. The expression levels of microRNAs markedly alter in response to ischemic AKI in mice [79,122,123], which play a either protective or pathogenic role in these mice. In particular, miR-17–5p [124], miR-26a [125], miR-126 [126], miR-127 [127], miR-205 [128] and miR-489 [129] seems to protect against ischemic AKI while miR-24 [130], miR-150 [131] and miR-494 [132] contribute to kidney injury. miR-21 (an ubiquitously expressed miRNA in mammals) which is expressed in many organs such as heart, kidney, and lung [133] is the most frequently analyzed miRNA in the context of AKI and has been shown to influence apoptosis, inflammation, and fibrosis in kidneys. Studies have also revealed that direct administration of specific miRNAs or exosomal miRNAs could improve outcomes in experimental AKI [134–136]. For example, delivery of exosomes from human cord blood endothelial colony-forming (ECFC) cells reduces ischemic kidney injury via transferring miR-486–5p which targets PTEN, suggesting that miR-486–5p enriched exosomes may represent a therapeutic tool in acute kidney injury [135].

Evidence also supports a role of lncRNAs in AKI. The hypoxic and inflammatory lncRNAs were dysregulated in human proximal tubular epithelial cells (PTECs) by using unbiased whole transcriptome profiling [137]. Among these, three lncRNAs (MIR210HG, linc-ATP13A4–8, and linc-KIAA1737–2) were most markedly induced by hypoxia and cytokines and were detectable in human kidney biopsy samples from kidney transplant patients. It has also been identified that metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) is the most strongly upregulated lncRNA in proximal tubular cells of kidneys from mice subjected to inspiratory hypoxia [138]. MALAT1 was also upregulated in kidneys subjected to renal I/R injury and murine hypoxic kidneys as well as in cultured and ex vivo sorted hypoxic endothelial cells and tubular epithelial cells. In addition, MALAT1 was upregulated in kidney biopsies and plasma of patients with AKI. However, genetic ablation of MALAT1 suggests that MALAT1 is dispensable in the development of renal IRI and subsequent kidney repair [139].

The knowledge of the role of circulating non-coding RNAs (circRNAs) in kidney disease is even more at an early stage than that of other types of ncRNAs such as lncRNAs and miRNAs. circRNAs are induced in mouse kidneys after I/R-induced AKI [140] and have also been shown to be expressed in the blood of patients with cardiac disease and to be associated with disease progression [141,142]. Thus, circRNAs may represent potential biomarkers of AKI. Future investigations are required to deeply characterize the regulation of circRNAs during AKI development, to identify circRNAs specifically expressed in AKI kidneys, to evaluate the association between circulating circRNAs and AKI and to elaborate their potential in AKI diagnosis and therapy.

3.5. Epigenetic biomarkers in AKI

In the past 50 years, the diagnosis of AKI has been mainly based on serum creatinine [143,144]. Changes in serum creatinine or urinary output are neither sensitive nor specific for AKI, yet they are the cornerstone of our current diagnostic approach. Changes in serum creatinine lack sensitivity for AKI because in a healthy person, nearly 50% of glomerular filtration rate (GFR) must be lost before a change in serum creatinine is detectable. A large number of potential new AKI biomarkers have been extensively evaluated in past years. The most relevant potential markers include neutrophil gelatinase-associated lipocalin (NGAL), the cysteine protease inhibitor Cystatin C, kidney injury molecule 1 (KIM-1) and the product of insulin like growth factor binding protein 7 (IGFBP-7) and tissue inhibitor of metalloproteinase 2 (TIMP-2), which perform better than the standard markers, such as creatinine and blood urea nitrogen. However, most of these markers have not been established in clinical practice [143,144].

Non-coding RNAs, particularly microRNAs, lncRNAs and circulating non-coding RNAs, have emerged as potential biomarkers of AKI and kidney repair. In particular, a large number of miRNAs have been analyzed as the potential of biomarkers in AKI in animal and human studies over the last decade.

3.5.1. MiR-21

MiR-21 attenuates I/R injury of the kidney by inhibiting apoptosis [145]. Suppressing miR-21 expression leads to a reduced activity of tumor necrosis factor a and monocyte chemoattractant protein-1 and thereby renal inflammation [146]. A number of studies have shown that miR-21 is differentially expressed in plasma and/or urine in AKI [122,147]. However, as miR-21 is expressed ubiquitously, it is likely to be a rather unspecific marker of AKI.

3.5.2. MiR-30a and c

The miR-30 family consists of five highly conserved members termed miR-30a to e. This family has been shown to be involved in numerous cancers, such as lung cancer, breast cancer, and colorectal cancer [148]. In a rat model of AKI, as well as in patients after cardiac surgery, miR-30c-5p was an early diagnostic marker of AKI in urine [149]. It has, furthermore, been shown that podocyte injury is facilitated by downregulation of the miR-30 family and that this injury can be prevented by the administration of glucocorticoids [150].

3.5.3. MiR-107

MiR-107 is member of the miR-15/107 family which is a crucial regulator in AKI, including stress response and angiogenesis [151]. In patients suffering from septic induced AKI, miR-107 was elevated in circulating endothelial cells [152].

3.5.4. MiR-125b

MiR-125b has been shown to be involved in ischemia and reperfusion events of several organs, especially the heart [153]. Additionally, it has been shown that miR-125b is a protective factor in cardiac ischemia and reperfusion injury [154]. In AKI, miR-125b has been shown to be a potential marker in the diagnosis of ischemic kidney injury [155].

All those studies demonstrate the potential of miRNAs as biomarkers for AKI; however, their specificity for AKI requires further investigation. Some miRNAs are associated not only with AKI but also with other diseases, including other kidney diseases. The involvement of different miRNAs in AKI indicates that a panel of miRNAs would potentially be more useful than a single microRNA as a biomarker. Circulating lncRNAs might also represent biomarkers for the diagnosis and prognosis of various diseases, including AKI. Several circulating lncRNAs-including MALAT1, NEAT1 and transcript predicting survival in AKI (TapSAKI; also known as MGAT3-AS1) are upregulated in plasma and kidney biopsy samples of patients with AKI [156–158]. Importantly, levels of NEAT1 and TapSAKI are also associated with severity of AKI [157,158].

4. Epigenetics in chronic kidney disease (CKD)

CKD is the most common type of kidney disease with a progressive loss of renal function, characterized by a decline in GFR, which is typically associated with irreversible pathological changes within the kidney. Patients with CKD may eventually develop ESRD requiring renal replacement therapies (chronic dialysis and/or renal transplantation). Quantification of estimated GFR (eGFR) has been used to define renal clinical phenotypes. For most clinical and genetic epidemiological studies, CKD has been defined as eGFR <60mLmin⁻¹1.73m⁻², irrespective of the presence or absence of kidney damage, and this includes stage 3 CKD (eGFR = 30–59mLmin⁻¹ 1.73m⁻²), stage 4 CKD (eGFR = 15–29mLmin⁻¹1.73m⁻²), and stage 5 CKD (eGFR <15mLmin⁻¹1.73m⁻² or persons on chronic dialysis) (Fig. 1) [159]. Stage 5 CKD is also known as ESRD. The pathological features of CKD include glomerular hypertrophy, mesangial expansion, tubulointerstitial fibrosis, and excess accumulation of extracellular matrix proteins, basement membrane thickening, and podocyte loss. CKD patients have an increased risk of cardiovascular events, hospitalization, and death. Common etiologies for CKD include diabetes, hypertension, glomerulonephritis, polycystic kidney disease, and chronic pyelonephritis. The molecular etiology of CKD involves both inherited risk factors, which have typically been considered to be manifest as genetic variants in the DNA sequence and epigenetic factors, which will be reviewed below.

Fig. 1.

The stages of CKD based on the levels of eGFR in patients. Chronic kidney disease is defined as either kidney damage or eGFR ≤60mLmin1⁻¹·1.73m⁻² for ≥3 months. Stage 1 CKD is represented by the level of eGFR over 90mLmin⁻¹·1.73m⁻² or greater. Stage 2 CKD is defined by mild kidney damage and a level of eGFR between 60 and 89mLmin⁻¹·1.73m⁻². Stage 3 CKD is defined by mild to severe loss of kidney function and a level of eGFR between 30 and 59mLmin⁻¹·1.73m⁻². Stage 4 CKD is defined by severe loss of kidney function and a level of eGFR between 15 and 30mLmin⁻¹·1.73m⁻². Stage 5 CKD is represented by kidneyfailure and also known as end-stage renal disease, in which the level of eGFR lower than 15mLmin⁻¹·1.73m⁻²

4.1. Pathogenesis of CKD

CKD is an umbrella term for renal disease that results in the progressive loss of kidney function. Low-grade and sustainable inflammation has been recognized as an important pathogenesis of CKD. DNA methylation can modulate the inflammatory process in the injured kidney. In dialysis patients, global DNA hypermethylation (total DNA 5-mc) in blood samples was associated with elevated inflammatory markers, such as ferritin and procalcitonin, in which procalcitonin has been used as a marker of inflammation due to bacterial infections. Histone modification can also contribute to the inflammatory process associated with kidney damage. In adriamycin-induced renal fibrosis animal model, two kinds of HDAC inhibitors, trichostatin A (TSA) and valproic acid, reduced the expression of the chemokines mcp1 and mip1β then diminished renal macrophage infiltration. Similar effects of both HDAC inhibitors were found in the UUO model by reduction of CSF-1 in renal tubule interstitial. CSF-1 is a chemokine known to be involved in macrophage infiltration. Although a broad range of insults can initiate kidney injury, interstitial fibrosis is a final common pathologic mechanism of most causes of progressive CKD. Thus, targeting profibrogenic pathways is an important strategy to slow the progression of CKD. DNA methylation changes have been observed in the kidneys of a mouse model of renal fibrosis, such as hypermethylation of rasGAP-activating-like protein 1 (RASAL1) and KLOTHO. Many studies have found that histone acetylation participates in experimental renal fibrosis. Treatment with various HDAC inhibitors, such as trichostatin A or the selective class I HDAC inhibitor MS-275, reduced renal fibrosis by diminishing profibrotic markers (α-smooth muscle actin (SMA)) and accumulation of excess extracellular matrix (ECM) proteins, including fibronectin and type I collagen in UUO mouse model. Sustained injury to the kidney results in maladaptive responses, including the deposition of ECM, particularly collagen, in the glomerulus and tubulointerstitium of the kidney. Emerging evidence demonstrate the relationship between transforming growth factor-beta1 (TGF-β) signaling and miRNAs expression during renal diseases. TGF-β regulates expression of several microRNAs, such as miR-21, miR-132, miR-433, and miR-29. MiR-21, miR-132, and miR-433 which are positively induced by TGF-β signaling play a pathological role in kidney diseases [160–163]. In contrast, members in miR-29 families which are inhibited by TGF-β signaling protect kidneys from renal fibrosis by suppressing the deposition of ECM and preventing epithelial-to-mesenchymal transition [164].

4.2. DNA methylation in CKD

DNA methylation as a gene-silencing mechanism plays an important role during organ and disease development. Under normal conditions, dynamic changes in DNA methylation, lead to stable and unique patterns that regulate tissue-specific gene transcription in somatic cells. In disease conditions, these methylation patterns have been known to change; either preceding the disease, or occurring as a consequence of it. During embryonic development, DNA methylation is a dynamic yet tightly controlled process, which contributes to the regulation of cell fate transitions. DNMT3a and DNMT3b were thought to establish the methylation patterns at the early development which is further maintained through somatic cell divisions by DNMT1, acting on the hemimethylated CpG sites generated by DNA replication. In recent years, research into the genetics of CKD has shifted substantially away from single-gene studies and toward genome-wide linkage studies and the identification of epigenetic risk factors [165]. Several groups have identified associations between CKD and DNA methylation.

4.2.1. Genome-wide profiling of DNA methylation in CKD

A recent large-scale genome-wide evaluation of DNA methylation examined the whole blood-derived DNA of 255 patients with CKD via the Infinium Human Methylation 450K BeadChip (Illumina, San Diego, CA, USA) and found that DNA hypomethylation and hypermethylation were present at different loci in patients with CKD [166]. The identified differentially methylated genes, including CUX1, ELMO1, FKBP5, INHBA-AS1 and PTPRN2 in CKD patients [166], have previously been implicated in pathways involving transcription regulation, signaling, and apoptosis. It is reported that DNA hypermethylation may be linked to inflammation associated with bacterial infections in CKD patients with dialysis [167]. Another study examined cytosine methylation profiles in human kidney tissues collected for transplantation or surgical nephrectomy identified differentially methylated loci in renal tubular components of patients with and without CKD. The significant methylation differences were detected for 1061 genes, in which several of the genes are known to contribute to renal fibrosis, including TGFBR3, SMAD3 and SMAD6, some of the genes are involving the renal transcription factors, including HNF, TCFAP and SIX2, and some of the genes are markers of cell adhesion pathways, including collagens and laminins [30]. In addition, a study showed that the urine 5-Methyl-2′-deoxycytidine (5MedC) level in combination with albuminuria or the α1MG level significantly predicted the renal survival in CKD patients, suggesting that it can serve as a novel biomarker for predicting the renal outcome in CKD [168]. Moreover, an epigenome-wide association studies (EWAS) of kidney function and CKD in 4859 participants identified differential DNA methylation significantly and reproducibly associated with kidney function and CKD as well as with the clinical endpoint renal fibrosis [169]. The genes, PTPN6/PHB2, ANKRD11, TNRC18, PQLC2 and PRPF8 showed a significant association between DNA methylation and the degree of fibrosis in the corresponding kidney biopsies [169]. Importantly, for both whole blood and renal tissue, the direction of the association between DNA methylation and renal function was consistent, which suggested that blood DNA methylation profiles could be used to reflect DNA methylation in the kidney. These associations have provided insights into disease pathophysiology, severity and prognosis.

4.2.2. DNA methylation in select genes related to CKD

In addition to the whole genome methylation analysis in CKD, the methylation on several specific genes has been directly associated with CKD progression. Hypermethylation of RASAL1, an inhibitor of the Ras oncoprotein, has been associated with the perpetuation of fibroblast activation and fibrogenesis in the fibrotic kidneys [170]. The expression of RASAL1 was decreased in fibrotic mouse kidney fibroblasts compared to nonfibrotic fibroblasts in vitro, whereas the expression of DNMT1, but not of DNMT3a and DNMT3b, was increased in the folic acid-challenged kidneys, which suggests a role of DNMT1 in the methylation of RASAL1 and the progression of kidney fibrosis. In experimental kidney fibrosis mouse models, including UUO mice, CD1 mice that developed diabetic nephropathy through administration of streptozotocin, COL4A3-deficient Alport mice and 5/6 nephrectomy, researchers also found that renal fibrosis was associated with RASAL1 hypermethylation. In whole-kidney biopsies and corresponding fibroblasts, severe fibrosis was also associated with RASAL1 hypermethylation and transcriptional silencing of RASAL1 [171]. An eight-year follow-up study from 81 residents in Cadmium (Cd)-polluted and non-polluted Chinese areas indicated that hypermethylation of RASAL1 and KLOTHO was strongly positively associated with Blood-Cd and urinary-Cd. Moreover, increased methylation in RASAL1 is correlated to renal dysfunction [172]. These findings suggest that hypermethylation of RASAL1 plays a pertinent role in acute kidney injury (AKI) only when it progresses to irreversible fibrosis associated with CKD.

KLOTHO is an anti-aging gene involved in renal phosphate handling and resistance to oxidative stress [173]. Emerging evidence from clinical and basic studies reveals that CKD is a state of endocrine and renal Klotho deficiency, which may serve as an early biomarker and a pathogenic contributor to chronic progression and complications in CKD including vascular calcification, cardiac hypertrophy, and secondary hyperparathyroidism. The methylation of Klothos was increased in CDK patient kidneys and peripheral blood mononuclear cells [174]. The methylation of Klotho promoter was also increased in animal model of CKD as examined by methylation-specific polymerase chain reaction (PCR) [175]. Treatment with uremic toxins resulted in the up-regulation of DNA methyltransferases (DNMT1 and DNMT3a) and hyper-methylation of Klotho in uninephrectomized B-6 mice and HK2 cells [175]. In addition, this study also found that HK2 cells treated with TGF-β in presence of a DNMT inhibitor 5-aza-2-deoxicytidine increased the levels of Klotho and decreased the phosphorylated Smad3, a TGF-β- signaling activation marker that was reported to be inhibited by Klotho. These results suggested that enhanced Klotho promoter methylation in CKD mice model might be due to increased DNMT1 and DNMT3a.

In addition, due to that erythropoietin (EPO) is secreted from renal interstitial fibroblasts, anemia arises as a major complication of CKD. It has been found that hypoxia-inducible factor 2α (HIF2α) tightly regulates EPO production at the gene transcription level to maintain oxygen homeostasis [176]. During CKD progression, the promoters of EPO and HIF2α are highly methylated which further induce kidney damage.

Furthermore, genes related to ECM are also regulated by the methylation of their promoters. Interstitial fibrosis is the common endpoint of end-stage CKD leading to kidney failure. Collagen IV (key component of ECM), TGF-β and smad, have been recognized as important components related to ECM signaling. It is reported that the changes of the methylation of SMAD3 and SMAD6 were correlated with gene transcription levels in CKD samples [30]. Also, the imbalance of the methylation status in MMP9 and TIMPS in injured kidneys, leading to collagen deposition. Importantly, the treatment with a DNA methylation inhibitor, 5-aza-2-deoxycytidine (5-Aza), restored matrix-degrading enzyme MMP-9/TIMP imbalance and ameliorated renal fibrosis in animal models [177].

4.3. Histone modifications in CKD

Histone modifications usually occur on critical amino acids of histone tails which extend out of the nucleosome [34]. Histone modifications correlate with open or closed conformations of chromatin and drive different accessibility to transcription factors and regulatory proteins to specific genes [35]. The most important histone modifications are acetylation and methylation of target lysine residues of the histone tails. Acetylation of histones strongly results in transcription activation, whereas deacetylation of histones mediated by HDACs represses gene expression in almost all experimental systems [38,39]. Methylation of histones can either activate or repress gene transcription, depending on which specific residues, including H3K4, H3K9, H3K27, H3K36, etc., are modified. Many more histone modifications exist, including phosphorylation, ubiquitylation and glycosylation, which spread to other amino acid residues, such as arginine, serine and threonine [35].

Multiple epigenetic regulatory proteins write, erase and read histone marks to alter chromosomal structure by directly modifying and regulating DNA accessibility [178]. Histone acetyltransferases/deacetylases and methyltransferases/demethylases can either write or erase histone acetylation and methylation, respectively [36,37]. The BRDs selectively read and bind to acetylated histone lysine residues to regulate gene expression, while the chromatin remodeling factors, such as chromodomain protein, selectively read and recognize methylated lysine residues to promote hetero-chromatin formation and to suppress gene transcription [44]. In addition to the histone modifications, epigenetic regulators also modify nonhistone substrates and regulates the activity of these proteins, such as histone methyltransferase, SMYD2, methylate histone H3K4 and H3K36 [179], histone H4K20 [180], but it also could methylate nonhistone proteins including HSP90, p53, Rb, STAT3 and p65 [181,182]. Many of these epigenetic regulators are dysregulated to affect cell homeostasis pathways in diseased cells and tissues. Because histone marks have stable covalent structures, they can be inherited during cell division and DNA duplication and serve as disease markers. Genome-wide epigenetic analysis is essential to determine the source of alterations to the distribution of histone markers, which is critical to developing epigenetics therapies. Next, we will discuss how these enzymes and epigenetic modifications could participate in CKD.

4.3.1. Histone acetylation and deacetylation in CKD

Histone acetylation reduces the net positive charge of histones and weakens interactions with DNA. Many studies have found that histone acetylation participates in experimental renal fibrosis of CKD. In diabetic renal fibrosis progression, ATP-citrate lyase (ACL) was upregulated and translocated to the nuclear, which resulted in the induction of histone hyperacetylation and the upregulation of TGF-β1, TGF-β3, CTGF, and ECM proteins fibronectin and collagen type IV [36]. ACL can also promote H3K9/14 and H3K27 hyperacetylation, leading to up-regulation of several rate-limiting lipogenic enzymes and fibrogenic factors, and resulting in promoting obesity-related kidney injury [183]. In addition, HDACs were found to participate in the process leading to renal fibrosis. It has been reported that daily oral administration of FR276457, a pan-histone deacetylase (HDAC) inhibitor, prevented interstitial fibrosis in a rat model of unilateral ureteral obstruction [184]. Treatment with CG200745, a recently developed pan-HDAC inhibitor attenuated fibrotic and inflammation responses and suppressed UUO-induced production of TGF-β and phosphorylation of Smad-2/3 [185]. Administration of valproic acid (VPA), a class I HDAC inhibitor, also abrogated the decrease in glomerular acetylation in an adriamycin-induced experimental model of glomerulosclerosis, resulting in attenuating the progression of renal fibrosis and glomerulosclerosis [186]. In addition, treatment with FK228, a selective inhibitor of class I HDACs, significantly suppressed the production of ECM in the murine model of UUO [187]. Mechanically, FK228 inhibited renal fibroblast activation and proliferation via Smad and non-Smad pathways via the increase of histone H3 acetylation. The expression of class IIa HDAC isoforms (HDAC4, 5, 7, and 9) increased in a murine model of UUO-induced renal fibrosis, with HDAC4 being most abundant. In a rat model of CKD, treatment with TSA, a class II HDAC inhibitor, increased TFEB acetylation and nuclear localization and attenuated ER stress associated tubule epithelial cell death, which diminished tubulointerstitial collagenous matrix deposition [188]. Interestingly, Tubastatin A induced PPARγ acetylation and bind to Klotho promoter, which upregulation of Klotho transcription. Furthermore, HDAC1/2 inhibitor FK228 and selective HDAC3 inhibitor RGFP966 significantly enhances PPARγ acetylation, restores Klotho by up-regulating its transcription, and consequentially attenuates the renal injury in CKD mice [189]. CKD-506, is a novel HDAC 6-specific inhibitor, which selectively inhibited HDAC6 enzyme activity and induced acetylation of α tubulin [190]. Treatment with CKD-506 significantly improved survival rate and significantly decreased the kidney inflammation, resulting in suppression of lupus nephritis in mice. In addition, blocking class IIa HDACs by MC1568 attenuates renal fibrosis through inactivation of TGF-β/Smad3 and NF-κB signaling, further up-regulation of MMP-2/9, and inducing klotho expression levels in an UUO injured kidneys [191]. SB939 (pracinostat), a new orally active hydroxamate-based HDAC inhibitor, is currently in phase II clinical trials. It potently inhibits class I, II and IV HDACs with excellent pharmacokinetic properties. Treatment with SB939 markedly inhibited the accumulation of α-SMA and tissue injury in the obstructed kidney through inhibition of Smad-independent TGF-β signaling [192].

Sirtuins (Sirts), mammalian Sir2 orthologs, are a highly conserved family of nicotinamide adenine dinucleotide–dependent protein deacetylases. Sirt6 is a histone 3 lysine 9 and 56 (H3K9 and H3K56) deacetylase, which represses the transcriptional activities of several transcription factors involved in DNA repair and genomic stability, glucose and lipid metabolism, cellular senescence and aging, as well as inflammation. Blockage of Sirt6 deacetylase activity aggravates renal fibrosis induced by unilateral ureteral obstruction and renal ischemia-reperfusion (IR) [192]. Mechanically, Sirt6 interacted with β-catenin and then was recruited to the promoters of β-catenin target genes for deacetylating H3K56 to negatively regulate β-catenin pathway [193].

The family of bromodomain and extraterminal (BET) proteins is composed of four members: BRD2, BRD3, BRD4 (ubiquitously expressed), and BRDT. Several evidences suggest that inhibitors of BETs can present antifibrotic properties. The in vitro experiments carried out in tubular epithelial cells stimulated with TGF-β1 showed that the inhibition of BRD4 functions by silencing their gene or treatment with JQ1 leads to a decrease in the expression of fibrotic genes, such as α-smooth muscle actin and fibronectin. Recent studies confirmed the beneficial effect of with I-BET151, a selective and potent BET inhibitor, in fibrosis and inflammation consequent to the unilateral ureteral obstruction induced renal fibrosis murine model, by inhibiting the activation of Smad-3, STAT3 and NF-κB pathways, as well as the expression of c-Myc and p53 transcription factors in the kidney [194].

4.3.2. Histone methylation and demethylation in CKD

It is recognized that the methylation of histones on lysine (K) or arginine (R) residues is also a key epigenetic mechanism regulating gene expression during CKD progression [47]. It is reported that histone H3 dimethyl K79 (H3m2K79) was decreased in patients with diabetic nephropathy [195]. Previous study found that histone H3 K79 methyltransferase, Dot1l deficiency exacerbates STZ- and UUO-induced kidney fibrosis through recruitment of HDAC2 to the Edn1 promoter to regulate Edn1 transcription by abolishing H3K79m2 and increasing H3 acetylation at the Edn1 promoter [196]. In the experimental UUO model in mice, H3K9me3 methylation was increased mainly in proximal tubules and myofibroblasts of obstructed kidneys. In the same study, TGF-β stimulation increased the methylation of H3K9me3 which is associated to the induction of α-SMA expression in primary rat renal fibroblasts and proximal tubule cells (NRK-52e). Enhancer of zeste homolog 2 (EZH2) is a methyltransferase that induces histone H3 lysine 27 trimethylation (H3K27me3) and functions as an oncogenic factor in many cancer types. Pharmacologic inhibition of EZH2 with 3-deazaneplanocin A (3-DZNeP) effectively inhibited histone H3K27 methylation and abrogated the deposition of extracellular matrix proteins and the expression of α-smooth muscle actin in the UUO mice kidney [109].

4.3.3. Other histone modifications in CKD

Histone phosphorylation occurs predominantly on serine, threonine, and tyrosine or residues through a phosphoester bond formation as well as on histidine, lysine, and arginine through phosphoramidate bonds [197]. Histone H2AX phosphorylation (γH2AX) at serine 139 was induced by ataxia telangiectasia-mutated (ATM) protein kinase, and is a key marker of DNA damage [198]. Previous study in a model of diabetic nephropathy induced by STZ in Nox1-deficient mice found that the levels of γH2AX were decreased in glomerular and tubular extracts in these knockout mice compared to its elevation in diabetic mice, suggesting a possible role of oxidative stress in CKD [199]. It has also been found that the phosphorylation of histone H3 on serine residue 10 (H3Ser10) were elevated in the kidneys of streptozotocin-diabetic endothelial nitric oxide synthase knockout (STZ-eNOS^−/−) mice and in the glomeruli of humans with diabetic kidney disease [200]. In addition, H3Ser10 phosphorylation could mediate endothelial activation in diabetes, which suggests that histone phosphorylation promotes endothelial activation in diabetes [200]. Histone phosphorylation should be set alongside other more canonical histone modifications when considering chromatin modifications may be amenable to therapeutic intervention.

4.4. Non-coding RNAs in CKD

Non-coding RNAs has been extensively studied in CKD. However, in this section we only focus on the roles of microRNA in fibrosis in CKD. Aberrant expression of miRNAs can drive the initiation and progression of fibrosis in response to renal damage. As miRNAs are stable and can be measured reproducibly from various sources, the determination of miRNA profiles could be very useful to define CKD progression. The last decade has witnessed a dramatic increase in the identification of miRNAs with potential roles in kidney fibrotic transformation. We will review the role of miRNAs in fibrosis in CKD in two categories: miRNAs are fibrosis promotors in CKD and miRNAs are fibrosis repressors in CKD (Table 1).

Table 1.

The potential roles of miRNAs in chronic kidney disease.

Function	miRNA	Species	Expression	Model	Targets or downstream molecular	References
Fibrosis promotors	miR-21	Human, mice	Up	CKD patients, UUO model	PPARA, Mpv17, TGF-β1/SMAD3, Smad7, ERK/MAPK	[146,160,161,201]
	miR-214	Human, mice	Up	CKD patients, UUO model	E-cadherin, PTEN, AKT1	[202]
	miR-433	Mice	Up	UUO model	Azin1, TGF-β/Smad3	[163]
	miR-132	Human, mice	Up	CKD patients, folic acid-induced renal fibrosis, UUO model	Smad2/Smad3, TGF-β, Foxo3/p300	[162,203]
	miR-34a	Mice	Up	UUO model	Klotho, Bcl-2	[204]
Fibrosis repressors	miR-29a-c	Human, mice	Down	CKD patients, UUO model	Collagen I, Collagen IV, HDAC4	[164,205,206]
	miR-30	Human, mice	Down	UUO model, TIF mice and human fibrotic kidney tissues	UCP2, Snail, Slug, Zeb2, Sox9	[207–209]
	miR-126	Human, mice	Down	CKD patients, UUO model	EGFL7, VEGFA, IκBα	[210–212]
	Let-7	Mice	Down	UUO model	Fibronectin, Thrombospondin, N-cadherin, Col1a2, Col4a1, JAG1	[213,214]
	miR-34c	Mice	Up	UUO model	Notch/Jag1	[215]
	miR-101a	Mice	Down	UUO model	KDM3A, TGFβ/Smad3, NF-κB	[216,217]

4.4.1. miRNAs are fibrosis promotors in CKD

4.4.1.1. MiR-21

MiR-21 is one of the best characterized fibro-miRs in multiple organs and was one of the first to be described in the kidney. miR-21 is also one of the most highly expressed miRNAs in the healthy kidney and is expressed in many other uninjured organs [218]. In the normal kidneys, miR-21 is mainly expressed in the cortex. Under renal injury the expression of miR-21 is increased in tubular epithelial cells, which suggests that miR-21 could target genes in tubular epithelial cells and mostly during injury. miR-21 was upregulated in several distinct animal models of kidney disease and in both human AKI and CKD tissue samples [160,201]. Studies confirmed that TGFβ1 induced miR-21 expression via regulation of SMAD3 but not SMAD2. Inhibition of miR-21 by a short hairpin RNA blocked renal fibrosis in a UUO model. In TGFβ1 transgenic mice, which have increased circulating levels of TGFβ1, deficiency of miR-21 ameliorated renal lesions via repression of proapoptotic signals. In miR-21-knockout mice, various genes, including Col1a1 and Col3a1, were silenced following renal injury, resulting in reduced fibrosis compared with wild-type controls. However, the global analysis [218] of gene expression profiles in miR-21 WT and knockout mice in response to injury suggested that miR-21 affects fibrotic disease via regulation of metabolic pathways as most of the repressed genes after miR-21 silencing were those involved in regulating fatty acid and lipid oxidation metabolic pathways. The role of miR-21 in regulation of metabolic pathways is an interesting finding. Among several genes with the roles in lipid metabolism pathways, PPARα was identified as a direct target of miR-21 [218]. Anti-miR-21 did not protect against UUO-induced renal fibrosis in Pparα-knockout mice [161]. An anti-miR-21 molecule, lademirsen, is undergoing clinical trials in Alport syndrome, which may provide the ground for development of the new treatment strategy based on miR-21 inhibition in CKD.

4.4.1.2. MiR-214

MiR-214 was regulated by Twist in many fibrotic diseases. The expression of miR-214–3p was significantly increased in kidneys from UUO mice and patients with CKD. In the mouse UUO model, decreased expression of miR-214 protected against the development of fibrosis, and treatment of wild-type mice with mmu-miR-214–3p sponge AAV before UUO resulted in similar anti-fibrotic effects [202]. Moreover, miR-214–3p negatively regulated the expression of epithelial cadherin (E-cadherin) by binding the E-cadherin 3′UTR under hypoxic condition [202]. miR-214 can also promote fibrosis in a SMAD-TGFβ1-independent manner through silencing the mitochondrial genes MT-ND6 and MT-ND4L, leading to disrupt mitochondrial oxidative phosphorylation or targeting PTEN, an inhibitor of the FAK/AKT pathway [219].

4.4.1.3. MiR-433

MiR-433 was reported to be upregulated in renal and cardiac fibrosis, while knockdown of miR-433 attenuated the induction and progression of renal fibrosis in UUO mouse model. In addition, renal miR-433 was also elevated in anti-glomerular basement membrane glomerulonephritis (day 14) in wild-type mice, and in adriamycin nephropathy mouse model [163], which acted as a downstream mediator of TGF-β/Smad3-driven renal fibrosis. miR-433 forms a positive feedback loop to amplify TGF-β signaling by suppressing the expression of Azin1, which is a target of miR-433 and has a protective role during renal fibrosis.

4.4.1.4. MiR-132

MiR-132 is regulated by the cAMP response element binding protein (CREB), the expression of which is regulated by angiotensin II. miR-132–3p and some other miRNAs are highly expressed at different stages in a mouse model of folic acid-induced kidney injury and fibrosis and in kidneys of UUO mice model. miR-132–3p was expressed throughout the cytosol of cells in the tubules and glomeruli in the tissues from CKD patients by In situ hybridization assay [203]. Silencing of microRNA-132 decreased collagen deposition (35%) and tubular apoptosis. Antagomir-132-treated mice reduces renal fibrosis by selectively inhibiting myofibroblast proliferation via the regulation of the levels of phospho-RB1, Cyclin, RASA1, p21 and the activation of TGF-β signaling (Smad2/Smad3), STAT3/ERK pathways in CKD kidney [162].

4.4.1.5. MiR-34a

Mature miR-34s in mammals are divided into three types, miR-34a, miR-34b, and miR-34c. miR-34a is the most widely distributed and highly expressed in brain. Recent study found that the miR-34a expression began to increase on the first day after UUO surgery, and had a time-dependent elevation after that. The miR-34a overexpression promoted epithelial-to-mesenchymal transition (EMT) in cultured human renal tubular epithelial HK-2 cells, which was accompanied by sharp downregulation of Klotho though direct binding with the 3′ UTR of Klotho [204]. miR-34a deficiency attenuated the progression of renal fibrosis following UUO surgery. miR-34a was also increased in the microvesicles secreted by tubulointerstitial fibroblasts, which promoted renal tubular epithelial cell apoptosis and participate in renal interstitial fibrosis by inhibiting Bcl-2 [220].

4.4.2. miRNAs are fibrosis repressors in CKD

4.4.2.1. MiR-29 family members

The miR-29 family members, miR-29a, miR-29b and miR-29c, can regulates ECM production in several organs, including the kidney. miR-29a, miR-29b and miR-29c were downregulated by TGF-β1 in proximal tubular cells [164]. Moreover, the levels of miR-29 [205] are significantly downregulated in urinary exosomes isolated from CKD patients and Fabry disease patients [204]. Interestingly, it has been found that renal fibrosis was partially depressed in the UUO mice with intramuscular injection of exosome-mediated miR-29 [221]. Upregulation of miR-29b by rAAV6-mediated miR-29b delivery attenuated renal fibrosis in UUO model by suppressing Snail1 expression [222]. Many genes that have been identified as targets of the miR-29 family encode ECM proteins, such as collagens, fibrillins, laminins, β1 integrin and elastin. In proximal tubular cells and podocytes, TGF-β1 decreased the expression of miR-29a/b/c and that this was associated with increased expression of collagens I, III, and IV. The miR-29 family reduces collagen I, III, and IV expression by targeting the 3′UTR of these mRNAs to repress translation. In podocyte, knockdown of miR-29a promoted HDAC4 activity that led to podocyte apoptosis, proteinuria, and subsequent renal dysfunction [206].

4.4.2.2. MiR-30

The miR-30 family consists of five evolutionarily conserved members, miR-30a through −30e. The miR-30 family had been documented to be downregulated in podocytes of patients with FSGS. Recent study shown that miR-30e was downregulated in TGF-β1-induced proximal tubule cells, tubulointerstitial fibrosis mice and human fibrotic kidney tissues [207]. The miR-30 family also inhibits epithelial cell ECM production and phenotype changes by downregulating mitochondrial uncoupling protein 2 (Ucp2) [208]. In UUO mice model, UCP2 was upregulated while miR-30e was decreased in kidney tubular cells. UCP2 is required for miR-30e-mediated tubular-cell extracellular matrix production stimulated by TGF-β1 [208]. TGF-β1 treatment could upregulate UCP2 and downregulate miR-30e expression in cultured NRK-52E cells. miR-30e-mimic transfection resisted to TGF-β1-induced damage in NRK-52E cells, whereas miR-30e inhibitor promoted epithelial cell phenotype changes under TGF-β1 treatment. UCP2 mRNA is a direct target of miR-30e. Interestingly, transcription factors that induce EMT, such as snail, slug, and Zeb2, were also direct targets of miR-30e. Enhance of miR-30e transcriptional activity decreased the expression of Snail, Slug and Zeb2, thereby attenuating the EMT of proximal tubule cells during tubulointerstitial fibrosis [207]. In fibroblast cells, miR-30 expression was significantly inhibited by TGF-β1, adenovirus-mediated ectopic expression of miR-30 in kidney fibroblast greatly reduced UUO-induced renal fibrosis by targeting Sox9 [209].

4.4.2.3. MiR-126

miR-126 is one of the most abundantly expressed miRNAs in endothelial cells, and is a well-studied miRNA in vascular biology and diseases. Expression level of miR-126 was significantly lower in the CKD kidneys compared to kidneys with normal function. Moreover, individuals in the highest tertile of each miR-126 had a significantly lower risk for CKD compared with the lowest tertile of the miR-126 [223]. Decreased expression of circulating miR-126 was associated with the development of diabetic nephropathy in type 2 diabetic patients and suggested it may be a promising biomarker for the progression of diabetic nephropathy. It has been reported that miR-126 directly targets 3′-UTR of IκBα to decrease its expression, which may play an important role in hepatic fibrosis [210].

4.4.2.4. Let-7 family

miRNAs Let-7 and its family members are highly conserved across species and play a significant role in the regulation of cell reprogramming. Members of the Let-7 family include Let-7a, Let-c, Let-7f, Let-7b, Let-7d, Let-7e, Let-7g, Let-7i and miR-98. Recently, let-7 family members have been suggested to act as negative regulators of profibrotic processes in renal fibrosis and CKD disease. It has been reported that TGF-β1 induces profibrotic changes and reduces let-7b expression in renal epithelium cells [224]. The expression of let-7b was also decreased in STZ-diabetic apoE knockout mice, an experimental renal fibrosis model. Knockdown of let-7b in rat proximal tubular epithelial cells (pTECs) elevated TGF-β1 receptor 1 (TGFBR1) expression via binding to the two let-7b sites in the 3-UTR of the human TGFBR1 gene. In addition, let-7 family members also target Col1a2 and Col4a1 in mouse mesangial cells (MCs). Thus, let-7 family members negatively modulate ECM accumulation by regulating multiple players in the pathway starting from receptor down to final gene product [213]. The potential prognostic value of let-7c dysregulation in renal fibrosis is supported by that several let-7c target genes were upregulated in human CKD biopsies [214].

4.4.2.5. MiR-34c

miR-34c is now reported to have several functions against different biological processes. In renal fibrosis, it has been reported that miR-34c was up-regulated by ureteral obstruction and TGF-β by microarray assay. However, over-expression of miR-34c inhibited Snail1 up-regulation induced by TGF-β, and administration of miR-34c into mice with UUO suppresses fibrosis of kidneys presumably through suppression of Notch/Jag1 pathway [215]. These results suggest that TGF-β and UUO up-regulates miR-34c which competes with EMT.

4.4.2.6. MiR-101a

Recent study found that miR-101a was expressed at a low level in UUO-induced chronic renal fibrosis kidneys and aristolochic acid (AA) treated HK2 cells, while overexpression of miR-101a could prevent chronic renal fibrosis. Furthermore, the dual-luciferase reporter assay showed that KDM3A is a target gene of miR-101a, which is highly expressed in mice with chronic renal fibrosis. miR-101a overexpression inhibits renal fibrosis via suppression of KDM3A, in which KDM3A stimulates chronic renal fibrosis via activation of the YAP-TGF-β-Smad signaling pathway [216]. MiR-101a mimics exhibited protective effects against AngII-induced hypertensive renal fibrosis via inhibiting TGFβ/Smad3 and NF-κB signaling pathways [217].

4.5. Epigenetic biomarkers in CKD

First, DNA methylation patterns in specific genes could be relevant as potential biomarkers in CKD. The connective tissue growth factor (CTGF or CCN2), as a profibrotic factor, is a potential biomarker of renal damage and a potential therapeutic target in experimental models of renal injury [225]. Recently study found that the levels of urine 5MedC, a marker of DNA methylation and an epigenetic marker, was increased in the later CKD stages and was related to a reduced eGFR in patients with CKD. The urine 5MedC level in combination with albuminuria or the α1MG level significantly predicted the renal survival in CKD patients, suggesting that it can serve as a novel biomarker for predicting the renal outcome in CKD [168]. In hemodialysis patients, global DNA methylation was higher than that in hemodiafiltration patients and DNA methylation was also elevated in most of the inflamed patients [226]. A study to evaluate the association with CKD risk factors and gene methylation sites in peripheral blood cells, and to determine whether these epigenetic sites were still predictors of eGFR after adjustment identified three significant CpG sites, including cg26842024 in KLF2 gene, cg07426848 in the S100A3 gene, and cg17589341 in the SLC14A1 gene [227]. These three epigenetic markers were able to significantly predict eGFR after adjustment for other risk factors. Second, it has been reported that global H3K9 trimethylation (H3K9me3) was increased in mouse kidneys after 10 days UUO [228]. Expression of EZH2, a methyltransferase that induces histone H3 lysine 27 trimethylation (H3K27me3), and H3K27me3 are increased in fibrotic kidneys from mice with UUO and humans with CKD [110]. These found suggested that the histone modification may act as biomarkers in CKD. Third, circulating miRNAs are regarded as attractive biomarker candidates and may reflect kidney disease as well. The expression of miRNA-16, miRNA-21, miRNA-155, miRNA-210, and miRNA-638 is correlated with renal function in patients with CKD [229]. Recent studies emphasize the possibility of employing circulating miR-21 as an early biomarker of renal fibrosis [230]. Up-regulation of miR-21 urinary exosomes of CKD patients and in different podocyte injury models, indicating that urinary exosomal miR-21 could be used as a potential new non-invasive, prognostic biomarker for CKD [231].

5. Epigenetic in diabetic kidney disease (DKD)

Diabetic kidney disease (DKD) is characterized with elevated urine albumin excretion associated with progressive decline in GFR, increased systolic blood pressure, and high risk of kidney failure in people with diabetes [232]. DKD occurs in nearly 40% of patients with diabetes, including diabetes type 1 (T1D) and diabetes type 2(T2D) [233]. T1D is defined as an autoimmune disease, which results from the irreversible elimination of pancreatic beta-cells mediated by autoreactive T cells, leading to absolute insulin deficiency and subsequent hyperglycemia; while T2D is due to a progressive loss of beta-cell insulin secretion frequently on the background of insulin resistance [234]. DKD develops in T1D 15–20 years later, while patients with T2D may present with albuminuria at the time the diabetes is detected. In addition to changes of albuminuria and kidney function (eGFR), the main histopathological features of DKD include glomerular hypertrophy, mesangial matrix expansion, interstitial fibrosis caused by accumulation of extracellular matrix in renal tubules, basement membrane thickening, glomerulosclerosis and podocyte effacement [233]. The mechanism of occurrence and development of DKD related to a multifactorial interaction of metabolic and hemodynamic factors such as high blood glucose, advanced glycation end products (AGEs) deposition, abnormal lipid metabolism, oxidative stress, and the renin-angiotensin system (RAS) [235]. However, current treatments to manage DKD refer to modulate glycemic and blood pressure have limited efficacy in preventing the progression of DKD [233]. The emerging role of epigenetic modification in DKD has attracted a broad attention. An in-depth understanding of the epigenetic mechanisms in DKD should facilitate to identify novel therapeutic target(s) for DKD treatment.

5.1. Pathogenesis of DKD

Conventionally, it is accepted that renal hemodynamics changes, oxidative stress and inflammatory response are majorly responsible for the pathogenesis of DKD [236]. Hyperglycemia is the initiating factor of DKD, which lead to intra-renal hemodynamic changes in the early stage, with the result of glomerular hyperfiltration. Glomerular hyperfiltration induces the constriction of efferent arteriole, thus causing changes of autoregulation and glomerular hypertension [237]. At the same time, hyperglycemia promotes ROS production [238], activation of protein kinase C (PKC) [239] and advanced glycation end-products (AGEs)-mediated proinflammatory response [240]. Under the condition of DKD, oxidative stress occurs in both glomerular mesangial cells and proximal tubular cells. In both cell types, hyperglycemia increased the level of ROS, which induces protein modifications, lipid peroxidation and mtDNA damage, ultimately results in mitochondrial dysfunction [241]. Although hyperglycemia and oxidative stress are the driving forces for renal damage associated with diabetes, numerous pieces of evidence point out the key role of inflammation in the pathogenesis of DKD [236]. Inflammatory mediators (chemokines and cytokines) induce extracellular matrix deposition and activation of intracellular signaling pathways including TGFβ/Smad [242], nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) [243], Janus kinase-signal transducer and activator of transcription (JAK-STAT) [244] and Notch pathway [245], have found to be integrally associated with DKD. Importantly, epigenetic mechanisms have been involved in oxidative stress and inflammatory response. Increasing evidence also indicated that genes associated with signaling pathways in DKD can be regulated by epigenetic mechanisms in response to environment stimuli, including nutritional changes. Detection of epigenetic events during the early stages of DKD could be valuable for timely diagnosis and prompt treatment to prevent progression to end-stage renal disease.

5.2. DNA methylation in DKD

Growing experimental and clinical studies support the links between CpG DNA methylation and DKD. Understanding genome-wide profiling of DNA methylation in DKD and DNA methylation in select genes related to DKD should facilitate disease prognosis, patient classification and treatment.

5.2.1. Genome-wide profiling of DNA methylation in DKD

Several studies have reported an association between DNA methylation profile measured in body fluid (e.g., peripheral blood and saliva) samples and DKD by analysis of genome-wide profiling of DNA methylation in DKD. A study has performed to assess the persistence of methylation patterns over years in 63 patients previously studied in the Diabetes Control and Complications Trial (DCCT) and Epidemiology of Diabetes Interventions and Complications Study (EDIC), of which 32 cases without significant microvascular complications in EDIC year 10 were selected from conventional therapy group and 31 cases of significant retinopathy or albuminuria by EDIC year 10 were selected from intensive therapy group. DNA methylation profiling was analyzed using Infinium Human Methylation 450K BeadChip arrays on whole blood collected from these patients. Hundreds of differentially methylated loci (DML) were identified in both monocytes and whole blood between the two groups. Twelve of the DML exhibited similar differences that persist for 16–17 years from the DCCT into the EDIC Study, in which four of the differentially methylated loci were hypomethylated (GATAD1, BSN, PRKCE, and TXNIP), and eight of the DMLs demonstrated hypermethylation (LHX6, MAP7, CDH3, SMYD1, CLCN7, ZNF167 and CUEDC1). The hypomethylation observed on the gene of thioredoxin-interacting protein (TXNIP) is interesting, given this gene known to be associated with hyperglycemia and diabetic complications [246]. More importantly, the fact that DNA methylation at certain loci persist for about 17 years supports the role of epigenetic phenomena in the development of metabolic memory. Another study profiled the genome-wide DNA methylation in leukocytes from non-diabetic offspring form mothers or fathers with type 1 diabetes. Among 87 differently methylated CpG sites, DNMT1 (the key enzyme for the maintenance of DNA methylation) was hypomethylated in the offspring compared to that in mothers with T1D [247]. By analysis of DNA extracted from saliva samples collected from African American and Hispanic diabetic participants, of whom 24 had no DKD, via Human Methylation 27K BeadChip arrays, 187 differentially methylated genes were identified between the two groups on at least 2 CpG sites. Among these 187 genes, 39 were involved in DKD, or identified to be up-regulated in dialysis patients [248].

Mitochondrial function plays an important role in the pathogenesis of DKD [249]. A study to profile the DNA methylation using blood-derived DNA to examine genes encoding mitochondrial proteins in patients with T1D with DKD (case group) or without renal disease (control group), identified 54 differentially methylated CpG sites associated with DKD. Pathway analysis revealed that several of the top-ranked genes, including ATP5O, NDUFA10 and COX7C, are involved in oxidative phosphorylation [250].

DNA methylation has also been associated with the clinical manifestation of DKD. A study demonstrated a correlation between DNA methylation patterns and the time to onset of DKD in a group of patients with T1D. The genome-wide DNA methylation analysis using whole blood from those patients identified 19 differentially methylated CpG sites associated with a higher risk of DKD. One of these CpG sites is located 18bp upstream of the transcription start site of UNC13B, that has higher levels of methylation in DKD patients and is associated with kidney injury in patients with T1D [31]. A study enrolled in 123 patients with T2D and with or without albuminuria found that global DNA methylation in peripheral blood monocytes was significantly higher in those patients with albuminuria. In addition, the severity of albuminuria correlated with the degree of global DNA methylation. These findings implicate DNA hypermethylation as a dependent risk factor for albuminuria in patients with diabetes [251].

5.2.2. DNA methylation in select genes related to DKD

The methylation states of several important genes were correlated with the clinical index of patient with DKD. For example, the methylation on the promoter of connective tissue growth factor (CTGF), a well-known pro-fibrotic factor, was significantly decreased in the peripheral blood of patients with DKD compared with that in blood from controls [252]. In addition, the degree of methylation in the CTGF gene is positively associated with the concentrations of CTGF in serum, the urinary albumin-to-creatinine ratio, blood urine nitrogen and serum creatinine, but negatively correlated with eGFR [252]. In another study, the methylation of the tissue inhibitor of metalloproteinase 2 (TIMP-2) and Aldo-keto reductase family 1 member B (AKR1B1) genes was decreased in the DKD patients with macroalbuminuria [253].

Abnormal DNA methylation has also been observed in DKD animal models and cells. It has been found that in the db/db mice, a T2D model, transcriptional repression of the transcription factor KLF4 was associated with increased DNA methylation at the promoter region of the nephrin, along with podocyte apoptosis and proteinuria [254]. In addition, DNA hypomethylation at the promoter of Agt is associated increased expression of angiotensinogen in early stages of DKD in db/db mice [255]. In streptozotocin-induced DKD mice, the induction of hypermethylation of the Cldn1 gene could lead to downregulation of the tight junction protein Claudin-1 and improve the albuminuria. Furthermore, the cultured proximal tubule cells exposed to high glucose exhibited DNA hypomethylation of MIOX, which was associated with the expression of genes involved in oxidative stress, hypoxia and fibrosis [256]. Demethylation of the transforming growth factor beta 1 (TGF-β1) promoter was observed in primary mesangial cells from diabetic (db/db) mouse kidneys as well as in mesangial cells exposed to hydrogen peroxide, which indicates that aberrant DNA methylation of TGF-β1 due to ROS overproduction plays a key to mesangial fibrosis during DKD progression [257].

5.3. Histone modifications in DKD

Recent studies have demonstrated that abnormal expression or activation of enzymes responsible for histone modifications is associated with development of DKD (Table 2).

Table 2.

The histone modification enzymes related to the development and progression of DKD.

Enzyme	Species	Expression	Target genes/proteins	Effects	References
HDAC2	Diabetic db/db mice, STZ-induced rats, diabetic patients	Up	Fibronectin, Collagen I, α-SMA	ECM accumulation, EMT	[258]
HDAC4	Diabetic db/db mice, STZ-induced rats, diabetic patients	Up	STAT1, Beclin 1	Proinflammatory mediator generation, podocyte apoptosis, inhibition of autophagy	[259]
Sirt1	Diabetic mice	Down	NF-κB, STAT3, p53, FOXO4, and PGC1-α	Aggravated albuminuria and worsened kidney disease progression	[260]
Sirt6	Diabetic mice	Down	Notch1, Notch4	Podocyte injury and proteinuria	[261]
p300/CBP	Diabetic mice	Up	PAI-1, p21	Augment glomerular dysfunction	[262]
EZH2	Diabetic mice	Down	Jag1	Glomerular injury	[263]
SETD7	STZ-induced rats	Up	p21	Mesangial cell (MC) hypertrophy	[264]
SUV39H1	Patients with diabetic nephropathy	Down	Fibronectin, p21 (WAF1)	Cell hypertrophy	[265]
UTX	Diabetic db/db mice	Up	p53, KLF10	Elevation of inflammation and DNA damage	[266–268]

5.3.1. Histone acetylation and deacetylation in DKD

Many studies have found that the aberrant expression of several HDACs and histone acetyl transferases (HATs) are involved in the pathogenesis in DKD. HDAC2 was the first deacetylase that was upregulated in the kidney of different diabetic animal models [258]. The expression of HDAC2/4/5 was also increased in the kidney biopsies from diabetic patients and in kidneys from diabetic animal models, including streptozotocin-induced diabetic rats and diabetic db/db mice, and the mRNA levels of HDAC2/4/5 in patients with DKD were negatively correlated with eGFR [259]. Sirtuins, including SIRT1–7, belonging to Class III HDAC, have been implicated in aging, transcription, apoptosis, and inflammation. The levels of Sirt1/3/4/6 were reduced in kidneys from patients with DKD. Moreover, the expression of Sirt1 in proximal tubular and glomerular was negatively correlation with the proteinuria in diabetic OVE26 mice [260]. Deletion of Sirt6 aggravated the albuminuria, mesangial matrix expansion, glomerular basement membrane thickening, and podocyte injury in diabetic mice through the increase the acetylation of H3K9 and the activation of Notch signaling pathways [261]. TGF-β1 is a major mediator of high glucose (diabetic condition) induced gene expression in renal cells. Upregulation of p300/CBP, a histone acetyltransferase, in diabetic mouse kidneys has been associated with the increase of TGF-β1, PAI-1 and p21 gene expression via promoter activation, and the augmentation of glomerular dysfunction in diabetic nephropathy [262]. Over the past decade, studies regarding the renoprotective effects of broad-spectrum HDAC inhibitors have been tested in diabetic rats and mice. Inhibition of HDACs reduced albuminuria and mesangial expansion, ameliorated podocyte injury, attenuated glomerulosclerosis and decreased production of proinflammatory mediators in diabetic animals [269–271].

5.3.2. Histone methylation and demethylation in DKD

Histone lysine methyltransferases (HMTs) catalyze lysine methylation of histones, whereas methyl groups from histones can be removed by histone lysine demethylases (KDMs) [272]. Recent studies have demonstrated that abnormal expression or activation of histone methyltransferases is associated with the development and prognostic state of DKD. We will briefly summarize the roles of EZH2, SETD7, SUV39H1 and UTX in DKD.

5.3.2.1. EZH2

Enhancer of zeste 2 polycomb repressive complex 2 subunit (EZH2), the functional enzymatic component of polycomb repressor complex 2 (PRC2), is responsible for the trimethylation of histone H3 at lysine 27 (H3K27me3). Pharmacologic or genetic depletion of EZH2 decreased H3K27me3 levels at the promoter region of Jag1, which encodes the Notch ligand Jagged1, leading to the increase of Jag1 expression and thereafter the induction of podocyte injury and the augmentation of oxidative stress and proteinuria in diabetic mice [263].

5.3.2.2. SET7/9

Histone methyltransferase SET domain containing lysine methyltransferase 7/9 (SET7/9) specifically mono-methylates lysine 4 of histone H3. A direct correlation was identified between SET7/9 recruitment and increased H3K4me1 at the p21 promoter under diabetic conditions, in which this process is mediated by TGF-β1 [264].

5.3.2.3. SUV39H1

The histone methyltransferase SUV39H1 is a histone methyltransferase that specifically trimethylates lysine 9 of histone H3 and functions as a transcriptional repressor. SUV39H1 has been shown to protect vascular smooth muscle cells (VSMCs) from metabolic memory and proinflammatory phenotypes under the stimulation of high-glucose. Recent studies found that SUV39H1 were reduced in renal tubules of patients with diabetic nephropathy, indicating that its expression correlates with the development of DKD [265]. Genetic overexpression of SUV39H1 or pharmacological activation of SUV39H1 was protective against the progression of DKD [273]. These findings imply that dysregulation of SUV39H1 may contribute to DKD progression.

5.3.2.4. UTX

Histone demethylase UTX (ubiquitously transcribed tetratricopeptide repeat on chromosome X, also known as KDM6A) specifically remove di- and tri-methyl groups from lysine 27 residue of histone H3 (H3K27) [266]. Because di- and tri-methylation of H3K27 are associated with gene silencing, upregulated UTX is usually associated with gene activation [266]. UTX was upregulated in the renal mesangial and tubular cells of diabetic mice and DKD patients, and administration of GSK-J4, an H3K27 demethylase inhibitor, ameliorated the diabetes-induced renal dysfunction, abnormal morphology, inflammation, apoptosis and DNA damage in db/db mice [267]. In addition, KDM6A reinforced diabetic podocyte dysfunction by creating a positive feedback loop through up-regulation of its downstream target KLF10 [268].

5.4. Non-coding RNAs in DKD

Epigenetic related ncRNA includes microRNA (miRNA), long noncoding RNA (lncRNA), etc. Some of the ncRNAs play important roles in the regulation of the expression of genes associated with DKD.

5.4.1. MicroRNA

A recent systematic review and meta-analysis in 12 of previous DKD human studies [274], showed that 15 miRNAs were upregulated, including miRNA-21, miRNA-181b, and miRNA-194, and 7 miRNAs (miRNA-26a, miRNA-126, miRNA-424, miRNA-574–3p, miR-223, miR-155, and miR-192) were downregulated in DKD group compared with control groups. Among which, the expression of 6 miRNAs, including miR-133b, miR-342, miR-30, miR-192, miR-194 and miR-215, was significantly correlated with urinary albumin excretion rates, and the expression of twelve miRNAs 12 was correlated GFR levels, which indicated that miRNAs may participate in the pathogenesis of DKD.

5.4.1.1. MiR-21

MiR-21 is a multipotent miRNA that has been demonstrated to promote cell proliferation, inflammation, angiogenesis, and immune destruction [275]. Its expression was upregulated in both human DKD and mice models, and the expression of glomerular miR-21 positively correlated with albumin to creatinine ratio, a measure of renal dysfunction. In addition, miR-21 negatively regulates the expression of phosphatase and tensin homolog (PTEN) and matrix metalloproteinases. Overexpression of miR-21 enhances TGF-β1-induced EMT by downregulating Smad7 and upregulating Smad3 expression [276,277]. Recently, miR-21 was demonstrated to promote inflammatory responses and cell apoptosis by targeting TIMP3 in HG-treated podocytes [278].

5.4.1.2. MiR-192

MiR-192 is the most abundant miRNA in the kidney, in which its high expression protects the kidney function. A case-control study enrolled in 464 patients with type 2 diabetes mellitus (T2DM) and 127 healthy controls found that miR-192 expression was reduced during DN progression, and the expression of miR-192 was negatively correlated with TGF-β, fibronectin and urine albumin creatinine ratio (UACR) [279]. MiR-192 causes the degradation of TGF-β and fibronectin through targeting early growth response factor 1 (Egr1) to affect the progression of DKD [280].

5.4.1.3. MiR-93

MiR-93 is a metabolically regulated miRNA and is differentially downregulated in animal model of DKD [281]. MiR-93 is a direct regulator of the expression of vascular endothelial growth factor A (Vegfa) [281] and Msk2, a ribosomal S6 kinase of serine/threonine [282]. Overexpression of miR-93 prevented the chromatin remodeling in podocyte cells under hyperglycemia condition through suppressing Msk2 expression (a kinase that phosphorylates histone, H3S10) [282]. MiR-93 significantly reduced in renal tissue of DKD patients and in TGF-β1-stimulated HK2 cells, and miR-93 overexpression prevented TGF-β1-stimulated EMT and renal fibrogenesis by targeting Orai1 expression [283].

5.4.1.4. MiR-503

Podocyte injury plays a key role in the pathogenesis of DKD. MiR-503 was involved in podocyte injury induced by high glucose. The expression of miR-503 was higher in DKD mice, and overexpression of miR-503 could aggravate podocyte injury in DKD mice by targeting E2F transcription factor 3 (E2f3), a transcription factor that is involved in the regulation of cell apoptosis, differentiation, and development [284]. Inhibition of miR-503 prevented DKD progression in mouse model of DKD treated with Losartan [285].

5.4.2. lncRNA

Several lncRNAs are involved in the regulation of ECM accumulation under the condition of DKD. The lncRNA plasmacytoma variant translocation 1 (PVT1) had been linked to diabetic nephropathy by the finding that the variants in this gene are associated with the development of ESRD in both type 1 and 2 diabetes mellitus [286]. PVT1 hosts several miRNAs, including miR-1207–5p, which subsequently regulates ECM formation in parallel to the lncRNA itself. Hyperglycemia induced upregulation of PVT-1 caused mesangial cell expansion and the accumulation of ECM proteins, including fibronectin, TGFB1, type IV collagen (COL4A1), and type 1 plasminogen activator inhibitor (PAI1). Knockdown of PVT1 significantly decreased the expression of both ECM proteins and their transcriptional regulators PAI-1 and TGF-β1 [287,288]. The expression of another lncRNA, Lnc-megacluster (lnc-MGC), was increased in glomeruli of DKD mice and in cultured mouse mesangial cells exposed to high glucose. Lnc-MGC hosts nearly 40 miRNAs in the miR-379 cluster, and promotes mesangial cell ECM accumulation in mouse models of early DKD [289]. Nuclear Enriched Abundant Transcript-1 (NEAT1), a novel nuclear long non-coding RNA, also plays a role in ECM accumulation. NEAT1 is increased in a streptozotocin-mediated diabetic rat model and murine mesangial cells treated with high-glucose, and NEAT1 interacts with AKT/mTOR pathway [290]. Inhibition of NEAT1 expression resulted in a reduction of TGFB1, fibronectin, and COL4A1 production in an animal model of diabetic nephropathy [290].

Several reports have described a crosstalk between lncRNAs and miRNAs in the pathogenesis of DKD. For example, the lncRNA MALAT1 controls renal tubular epithelial pyroptosis in experimental models of DKD by inhibiting the expression of miR-23c, leading to increase the expression of its target Elavl1 which induce pro-inflammatory programmed cell death in tubular epithelial cells during tubular injury [57]. The pro-apoptotic role of lncRNA GAS5 was due by its competitive binding toward miR-27a [291]. In hyperglycemic condition, GAS5 inhibits the regulatory function of miR-27a toward the BCL2 Interacting Protein 3 (BNIP3), thus causing an upregulation of BNIP3 expression and renal tubular epithelial cell apoptosis [291]. The lncRNA 1700020I14Rik reduces cell proliferation and fibrosis in DKD through interaction with the miR-34a-5p–Sirt1–HIF-1α signaling pathway [292].

5.5. Epigenetic biomarkers in DKD

Recent studies show that that baseline GFR, albuminuria and blood biochemical parameters can fairly accurately predict kidney function decline in patient with DKD. Additional biomarkers in the blood or urine have also been identified to predict renal function decline, however none has been shown to outperform the baseline clinical parameters. Recently, studies have focused on identification of epigenetic biomarkers in DKD.

First, modified nucleosides in serum may act as the potential biomarkers to describe the development of DKD. A recent study detected the expression pattern of free modified nucleosides containing m6A/C, I/C, 5-mdC/C, 5-mC/C, pseU/U by HILIC-ESI-MS method in 189 serum samples from four groups of volunteers and patients. In this study, 43 healthy volunteers and 156 patients were divided into four groups according to the degree of proteinuria, in which can be considered as the different stages of the development of DKD. The results shown that m6A/C, I/C, 5-mC/C, and pseU/U can distinguish the different stages of the development of diabetic nephropathy, which may help the clinical diagnosis and treatment of DKD [293].

Second, circulating lncRNA may serve as new biomarkers for diagnosis of DKD. A study analyzed the circulating lncRNA and mRNA expression profiles in normal control, diabetes mellitus, and diabetic nephropathy patients, which showed that the levels of lncRNA-ARAP1-AS2 and ARAP1-AS2 were gradually increased during the progression of diabetes and DKD. These findings suggested that ARAP1-AS2 and ARAP1-AS2 may be new biomarkers for the development of DKD [294].

In addition, miRNA has gradually shown its potential to be a diagnostic marker for DKD. Several reports demonstrate comprehensive profiles of miRNAs in patient urine, urinary sediment, and serum that could be correlated with specific stages of diabetic nephropathy, fibrosis, and renal function decline (GFR). In particularly, exosomes in urine are an extremely valuable source for miRNA profiling in renal disorders because they are originated from most of renal cells [295]. Recent study suggests that plasma miRNA-21 can serve as an early marker for diagnosis and identifying DKD in T1D [296].

miRNA as biomarker has the following three advantages: (1) miRNA has expression stability and tissue specificity in specific diseases; (2) miRNA is highly expressed in urine and stable in tissues [297]; (3) Studies have found that under certain circumstances, miRNA expression will change in the early stages of DKD, which is related to renal fibrosis [298]. However, as far as clinical application is concerned, there are still problems that need to be solved urgently. For example, with the progress of renal failure, the reduction of miRNA expression can be attributed to the decline in synthesis ability, not the down-regulation of its own expression. Although several challenges remain, it is clearly that miRNAs are very attractive as simple and accurate biomarkers for DKD.

6. Epigenetics in renal fibrosis

Fibrosis is a common pathologic pathway of progressive kidney disease involving complex signaling networks. The identification of epigenetic alternation in response to environmental modifiers driving the onset and progression of renal fibrosis has extended our understanding of the pathophysiology of kidney fibrosis progression. In this section, we will summarize studies concerning the implications of epigenetic modification in regulation renal fibrosis progression and highlight some possible epigenetic therapeutic strategies for the treatment of renal fibrosis.

6.1. Epigenetic regulation of fibrosis in AKI

Following acute renal injury, a repair response should be activated. However, severe or sustained renal injury often results in maladaptive and incomplete repair, leading to cell proliferation, hypertrophy, exaggerated extracellular matrix production and ultimately progression to CKD or end-stage renal disease [299–301]. Transition of AKI to CKD is mediated by multiple mechanisms, including oxidative stress, mitochondrial injury, aberrant cell cycle arrest and hypoxia [64–66]. Of noted, during an AKI episode, modification occurs not only in the cellular metabolism, but also in the chromatin structure and in the binding of different transcription factors in tubular epithelial cells subjected to a hypoxic milieu [302,303], resulting in the increase of the expression of pro-inflammatory cytokines, such as tumor necrosis factor (TNF-α) and monocyte chemoattractant protein (MCP-1) [304,305]. Studies in animal models of AKI showed that the increment in the multiprotein chromatin remodeling complex such as SWItch/Sucrose Non-Fermentable (SWI/SNF) factor, which regulates chromatin structure by activating or inactivating gene expression, and progressive H3 acetylation was associated with the increase of proinflammatory/pro-fibrotic genes, including MCP-1, TNF-α, and TGF-β1, etc. [99]. In addition, lysine 4 trimethylation on histone 3 (H3K4m3) and upregulation of histone 2 variant (H2A.Z) also promote the proinflammatory/profibrotic genes expression, including MCP-1, TGF-β1 and collagen III, in unilateral I/R model [304]. Studies further showed that treatment with histone deacetylase inhibitor m4PTB accelerates recovery and reduces post injury fibrosis after IR-induced AKI, which is due to activation of transcriptional programs involved in promoting cell cycle progression and reducing G2/M [306,307]. All these studies suggested that epigenetic mechanisms are involved in renal inflammation and fibrosis in AKI, and epigenetic therapy may be a potential strategy for the treatment of renal fibrosis in acute injured kidneys.

6.2. Epigenetic regulation of fibrosis in CKD

Chronic progressive kidney disease are mainly caused by diabetes, hypertension and primary glomerulopathies [308]. Regardless of the primary underlying disease, CKD is histomorphologically characterized bytubulointerstitial fibrosis as a common downstream feature [309].We have deeply discussed the roles of miRNAs in regulating renal fibrosis in CKD above. Studies also found that other epigenetic mechanisms are involved in the progression of renal fibrosis in CKD. It has been reported that the rate of loss of renal function and changes in DNA methylation at CpG islands of genes, including NPHP4, IQSEC1 and TCF3, etc., are involved in EMT and renal fibrosis [310]. The DNA methylation profiling of tubule epithelial cells from CKD patients was significantly changed compare to normal individuals, which occurred predominantly at enhancers of specific genes, such as those key fibrotic genes of the TGFβ pathway [30]. Administration of bone morphogenic protein 7 (BMP7) has been linked to the Tet3-mediated reversal of pathologic hypermethylation at the Rasal1 promoter and mitigated the experimental kidney fibrosis [171]. There is significantly higher prevalence of hyperhomocysteinemia (HHcy) in the blood of kidney failure patients [311]. In hyperhomocysteinemia induced renal damage animal model, DNA hypermethylation were followed by the changes of gene expression that was involved in ECM regulation, including matrix metalloproteinase-9 (MMP-9) and downregulation of MMP inhibitors, such as metalloproteinase 1 (TIMP1) and TIMP2. The changes of methylation status of MMP9 and TIMPs might increase collagen deposition and contributed the renal fibrosis progression [177]. In addition, abnormal histone modifications, such as H3K4 methylation, on the genes in the TGFβ pathway has also been proved as an important mechanism of fibrosis in CKD, which could either promote or repress the transcription of pro-fibrotic genes in CKD [304,312]. These studies suggested that in addition to miRNAs, DNA methylation and histone modifications are also contribute to renal fibrosis in CKD.

6.3. Epigenetic regulation of fibrosis in DKD

Diabetic kidney disease (DKD) is the greatest risk factor for CKD and ESKD. In untreated diabetic patients and animal models, hyperglycemia leads to DKD with persistent microalbuminuria, glomerular basement thickening, podocyte injury, mesangial matrix expansion, and eventually glomerulosclerosis and tubulointerstitial fibrosis [31,313]. TGF-β1 mediated EMT is an outstanding mechanism in the progression of fibrosis in DKD. The profibrotic cytokines such as plasminogen activator inhibitor 1 (PAI-1) and connective tissue growth factor (CTGF) play an important role in the progression of excess deposition of ECM in DKD. It has been found that TGF-β1 induced elevation of H3K4 methylation and decrease H3K9 methylation at PAI-1 and CTGF gene promoters, and the accumulation of HMT SET7/9 to fibrotic and ECM gene promoters, resulting in the increase of the expression of these profibrotic proteins. In DKD, HDACs also play a role in the expression of pro-fibrotic genes [258,314,315]. Treatment with Apelin-13, a deacetylation agent, inhibits inflammation and the expression of TGFβ1 and NF-κB. Similarly, treatment with a broad class I and II HDAC inhibitor, TSA, decreases ECM accumulation and EMT in DKD. In addition, the study of human renal biopsies from patients with diabetic nephropathy showed that TGFβ-mediated upregulation of miR-192 in proximal tubule cells was linked to fibrosis [316], which via inhibition of E-box repressors [232]. Furthermore, aberrant DNA methylation and consequent deregulated gene expression could also be involved with the renal fibrosis in DKD. For example, methylation in specific CpG sites on genes, including WDR59, CFDP1, COX17, CD80, and PLA1, was associated with eGFR decline and fibrosis in diabetic American Indians [317]. In sum, growing evidence suggest that DNA methylation, histone modification and miRNAs are all involved in renal fibrosis progression in DKD.

6.4. Epigenetic treatments in renal fibrosis

We have briefly discussed the associations of DNA methylation, histone modifications and miRNA in renal fibrosis in kidney diseases. Pharmacological interference on aberrant epigenetic modifications has been shown to attenuate fibrogenesis and to be beneficial for renal function in human and animal models of renal fibrosis. Several studies have elucidated the efficiency of these drugs in the treatment of renal fibrosis as summarized in Table 3 [170,175,184,318–327]. First, methylation was shown to play an important role in the progression of renal fibrosis as described [170,171,318,328], supporting the role of pharmacological demethylation agents in preventing or slowing the development of renal fibrosis. 5-azacytidine and 5-aza-dCTP cytidine analogs are incorporated into genomic DNA and lead to irreversible covalent complex with DNMT1 formation, which results in genome-wide demethylation after cytosine excision [329]. These demethylating compounds are currently being tested in some clinical trials, most of them in cancer therapies. In experimental renal fibrosis, a potential role of 5-azacytidine in ameliorating renal fibrosis have been tested in folic acid-induced nephropathy and uremic toxins associated nephropathy mice, which has resulted a positive outcomes [170,175].

Table 3.

The epigenetic pharmacological interference in experimental kidney fibrosis.

Drugs	Target enzyme	Experimental model	Effects	References
5-Azacytidine	DNMT1/Tet3	Folic acid-induced Nephropathy; Unilateral ureteral obstruction; Diabetic db/db mice	Rasal1 promoter demethylation and normalizes the phenotype of fibrotic fibroblasts; Lower urinary albumin/creatinine ratio and serum creatinine	[170] [175] [318] [319]
Hydralazine	Tet3	Unilateral ureteral obstruction; IRI model	Rasal1 promoter demethylation and prevents AKI-CKD	[170] [318] [320]
5-Aza-2′-deoxycytidine	DNMT1	Indoxyl sulfate and pcresyl sulfate injected mice; 5/6-nephrectomized model	Klotho expression	[321] [321]
Trichostatin A	HDAC Class I HDAC Class II	Unilateral ureteral obstruction; DOX-induced nephropathy	Inactivation of renal interstitial fibroblasts and inhibition of renal tubular cell death	[322] [323] [324]
Valproic acid	HDAC Class II	Unilateral ureteral obstruction; DOX-induced renal fibrosis; STZ-induced diabetic kidneys	Induction of autophagy; Inhibition of the development of proteinuria; Repression of the myofibroblast Transformation and fibrogenesis	[323] [324] [325] [326]
MS-275	HDAC Class I	Unilateral ureteral obstruction	Inhibition of ECM protein synthesis by TGF-β and EGFR signaling	[327]
FR276457	Pan-HDAC	Unilateral ureteral obstruction	Inhibition MCP-1 production	[184]
I-BET151	BET	Unilateral ureteral obstruction	Inhibition of ECM protein synthesis	[194]

Second, in the efforts to erase aberrant histone acetylation, pharmacological agents such as TSA and MS-275, the inhibitor of Class I and II HDACs, have been tested in the experimental kidney fibrotic animals, and have shown an anti-inflammatory and anti-ECM protein synthesis effects in different models of kidney fibrosis [322–324]. Moreover, treatment with other inhibitors of Class I and II HDAC such as phenylbutyrate and valproic acid also have been shown to significant effective activity in the streptozotocin induced diabetic nephropathy and doxorubicin-induced nephropathy subjected to TGF-β1 [258,324,330].

However, we should point out that pharmacological agents show several side effects. In particular, 5-azacytidine and decitabine incorporated affecting methylation of not only aberrant methylated genes thus they have considerable cytotoxicity when incorporated into DNA [171,331]. In addition, even though different selective HDAC inhibitors have been developed, their clinical beneficial are not ensure yet. Consequently, these drugs are of limited utility in the setting of renal fibrosis prevention and future studies with epigenetic modulators in renal fibrosis would be necessary.

7. Conclusion and discussion

Kidney disease is increasingly becoming a great epidemiologic concern worldwide. Despite the considerable advances in research, the pathophysiologic pathways involved in the progression of these kidney diseases remain to be elucidated. In recent years, multiple studies in kidney diseases have shown that mis-regulation of epigenetic mechanisms contribute to kidney disease pathophysiology. It is becoming increasingly clear that DNA methylation and histone modifications impact cellular function within the kidney cells and its microenvironment and contribute to kidney disease progression. It is known that anomalous DNA methylation, aberrant histone alterations, and changes of microRNA expression all contribute to its pathogenesis. The manner and extent to which these factors modulate inflammation, fibrosis, apoptosis and the transition of mesenchymal to epithelial cells in kidney diseases are still being elucidated.

To better understand the pathophysiology of kidney diseases, it is important to identify the changes in DNA methylation and histone modifications in disease conditions compared to the health individuals, which is essential to finding hidden sources of variation in kidney diseases and therapeutic selection. The newly developed high-throughput measurement technologies enable unprecedented, quantitative measurements of the epigenetic state in normal and disease kidneys. For DNA methylation, these techniques can be applied to kidney samples from human and model organisms, in which the functional impact of methylation alterations can be assessed bioinformatically in targeted experiments on model organisms and across sample populations. In contrast, chromatin assays require higher-quality and quantity samples that are typical not feasible for preserved kidney samples or biopsies. As a result, chromatin measurements are typically limited to model organisms and cultured cell lines, which are essential to kidney epigenetics studies. Due to that DNA methylation and histone modifications are reversible process, identifying these changes may potentially serve as therapeutic targets for kidney disease treatment. Future studies using the already developed and newly emerging technologies should elucidate how alterations in the renal cell epigenome cooperate with genetic aberrations for kidney disease initiation and progression. In addition, we believe that incorporating epigenomic testing into the clinical research is essential to future studies with epigenetic biomarkers and precision medicine using emerging epigenetic therapies.