Epigenetic changes mediating transition to chronic kidney disease: Hypoxic memory

Acute kidney injury (AKI) was classically recognized as a disease of transient nature—an event which resolves spontaneously and never leaves any scar in the future. However, this long-held view has been challenged by a number of recent epidemiological studies. According to a recent meta-analysis, there is an 8.8-fold increase in risk for chronic kidney disease (CKD) and a 3.1-fold increase in risk for end-stage kidney disease (ESKD) in patients surviving an AKI episode.1 Underlying mechanisms accounting for the transition of AKI to CKD have been eagerly sought. Candidates so far include nephron loss, inflammation, endothelial injury with vascular rarefaction and cell cycle arrest in epithelial cells. Among them, hypoxia is increasingly recognized as a common pathway mediating transition from AKI to CKD.2 As such, tubular epithelial cells express hypoxia-inducible factor (HIF)-1 immediately after occlusion of blood flow and at some time after reflow – 3 and 7 days after injury. Rarefaction of peritubular capillaries, as well as microvascular insufficiency, is regarded as an important event underlying tubulointerstitial hypoxia. A recent study using fluorescence microangiography demonstrated that reductions in the total perfused areas after an AKI episode were associated with reductions in capillary number and individual capillary calibre. A comparison with CD31 immunostaining demonstrated that some capillaries expressed CD31, yet failed to exhibit effective flow, suggesting that some capillaries were experiencing insufficient perfusion. During the AKI-to-CKD transition, hypoxia is recorded as epigenetic changes in the cell and has a long-term effect, which is called "hypoxic memory".3 These changes include DNA methylation, histone modification, changes in chromosome conformation, alterations in levels of long non-coding (lnc) RNAs and microRNAs. Pathogenic involvement of epigenetic changes in renal ischaemia is to be envisaged by a series of preceding studies looking at histone and chromatin modifiers that influence the expression of inflammatory and fibrotic genes.2 Three weeks after ischaemia-reperfusion injury in mouse kidneys, the expression of inflammatory and profibrotic genes, such as monocyte chemoattractant protein-1 (MCP-1), transforming growth factor (TGF)-β1 and collagen III, increased significantly. This increase was associated with histone modification that drives gene activation, such as H3K4me3 and H3K9/14ac, as well as an increase in binding of the chromatin-remodelling enzyme, Brahma-related gene 1, to the promoter of these genes. In addition, the ischaemia-reperfusion injury upregulated endothelin-1 (ET-1) and facilitated gene-activating histone modifications near the transcriptional start site, which likely contributes to a vicious cycle of hypoxia by narrowing capillary calibre. Furthermore, a positive feedback mechanism of the hypoxic gene induction, through epigenetic changes and HIF, is reported in the transcription of the glucose transporter 3 (GLUT3) gene. Using chromosome conformation capture (3C) assay, it was demonstrated that HIF-1 and lysine-specific demethylase 3A (KDM3A) are recruited to the GLUT3 gene loci, cooperatively demethylate H3K9me2 and construct long-range looping interactions between the promoter and the distal enhancer in order to maximize the hypoxic induction of GLUT3.4 Non-coding RNAs also appear to play a role in kidney fibrosis. microRNAs, short RNA molecules of 22 nucleotide length, target complementary sequences on the 3′ untranslated region of a cluster of target mRNAs for post-transcriptional repression, which constructs a network of gene regulation. On the other hand, lncRNAs are non-coding transcripts of over 200 nucleotides in length, with their putative roles being identified in conditions such as AKI, glomerular diseases, acute allograft rejection and renal fibrosis. Using RNA-seq, non-coding RNAs commonly up- or downregulated were identified at multiple time points in the unilateral ureteral obstruction (UUO) fibrotic kidneys. In vitro, overexpression studies confirmed that lncRNA 3110045C21Rik can influence the expression of fibrosis-related proteins, such as E-cadherin, α-smooth muscle actin (α-SMA) and TGF-β1.5 A more comprehensive understanding of the spectrum of hypoxia-inducible, non-coding RNAs was obtained in vitro, using non-biased, genomewide approaches. A combinational analysis of RNA-seq and chromatin immunoprecipitation (ChIP)-seq identified 44 lncRNAs in multiple tubular cell lines. Among them, DARS-AS1 (DARS [aspartyl-tRNA synthetase] antisense 1) was most strikingly hypoxia inducible. Several HIF binding motifs were found in the 5′ flanking sequences relative to the DARS AS1 transcriptional start site. Functionally, DARS-AS1 suppressed apoptotic cell death in hypoxia.6 Histone methyl transferases are deeply involved in the "hypoxic memory," which also facilitates transition and progression to CKD. In keeping with this view, a pharmacological targeting of histone methyltransferase that induces histone H3 lysine 27 trimethylation revealed potential therapeutic application of such inhibitors in fibrotic mouse kidneys. Enhancer of zeste homolog 2, EZH2, is a histone-lysine N-methyltransferase which usually causes transcriptional repression. In mouse UUO kidneys, EZH2 was upregulated in the nuclei of αSMA-positive interstitial cells. In immunohistochemistry of human kidney biopsy sections, interstitial expression of EZH2 was also observed in patients with focal segmental glomerulosclerosis (FSGS) and weakly in those with IgA nephropathy, but not in those with minimal change nephrotic syndrome (MCNS), suggesting that the expression was associated with fibrosis. Pharmacological inhibition of EZH2 with 3-deazaneplanocin A (3-DZNeP) abrogated deposition of extracellular matrix proteins and αSMA in the UUO kidney.7 In summary, there is an emerging concept of hypoxia-driven, epigenetic changes that accounts for the transition from AKI to CKD (Figure 1). Efforts are under way to prevent the AKI-to-CKD transition targeting hypoxia. Along similar lines, there is growing interest in preventing fibrosis by directly targeting epigenetic changes. A more detailed, comprehensive understanding of epigenetic changes in the pathogenesis of ischaemic kidney injury would facilitate discovery of such therapies in the future. This study was supported by a Grant-in-Aid for Scientific Research (C) 17K09688 (T.T.) by Japan Society for the Promotion of Science. None declared.