Epigenetics in diabetic nephropathy, immunity and metabolism

Abstract

When it comes to the epigenome, there is a fine line between clarity and confusion-walk that line and you will discover another fascinating level of transcription control. With the genetic code representing the cornerstone of rules for information that is encoded to proteins somewhere above the genome level there is a set of rules by which chemical information is also read. These epigenetic modifications show a different side of the genetic code that is diverse and regulated, hence modifying genetic transcription transiently, ranging from short- to long-term alterations. While this complexity brings exquisite control it also poses a formidable challenge to efforts to decode mechanisms underlying complex disease. Recent technological and computational advances have improved unbiased acquisition of epigenomic patterns to improve our understanding of the complex chromatin landscape. Key to resolving distinct chromatin signatures of diabetic complications is the identification of the true physiological targets of regulatory proteins, such as reader proteins that recognise, writer proteins that deposit and eraser proteins that remove specific chemical moieties. But how might a diverse group of proteins regulate the diabetic landscape from an epigenomic perspective? Drawing from an ever-expanding compendium of experimental and clinical studies, this review details the current state-of-play and provides a perspective of chromatin-dependent mechanisms implicated in diabetic complications, with a special focus on diabetic nephropathy. We hypothesise a codified signature of the diabetic epigenome and provide examples of prime candidates for chemical modification. As for the pharmacological control of epigenetic marks, we explore future strategies to expedite and refine the search for clinically relevant discoveries. We also consider the challenges associated with therapeutic strategies targeting epigenetic pathways.

Keywords: Chromatin; Diabetes; Diabetic complications; Diabetic nephropathy; EWAS; Epigenetics; Histone; Innate immune memory; Vascular.

Fig. 1

Codified signature of the diabetic epigenome. Readers, writers and erasers in diabetes. Modification of the diabetic epigenome includes post-translational modifications to the tails of histones, carried out by histone-modifying enzymes (known a ‘writers’), such as SET7 [24, 26, 30, 102, 103, 106, 107], SETDB1 [108], SUV39H1 [109, 110], EZH2 [40, 111], KAT2A [42, 112] and GLYATL1 [113]. Experimental studies that provide mechanistic insights for specific determinants are grouped to include the enzyme and corresponding modified histone, whereas informative profiling studies using clinical cohorts are separated with examples such as SET7 [30], SUV39H1/H2 [107], H3K9 acetylation [31] and H3K9me2 [82]. The epigenetic code is dynamic and eraser enzymes are implicated in diabetes such as KDM6B [114], PHF2 [115], KDM1A [26, 108], HDAC3 [99, 116], HDAC4 [117] and HDAC7 [118]. Protein readers such as CTCF recognise post-translational histone modifications including methylation of cytosine residues in CpG dinucleotides [94, 119]. Genome readers regulate transcriptional responses and include KLF4 [38], SIRT1 [47], as well as non-CpG methylation by DNMT3B [120]. The DNA template is subject to modification and recent experimental studies have shown an association with 5mC [, , , –130] and 5hmC [131]. Clinical profiling studies for DNA modification have also shown an association with 5mC [, , , , , –149] and 5hmC [150]. Post-transcriptional gene regulation by RNA modifications include the writers, erasers and readers of N ⁶-methyladenosine (m6A). FTO [–153] is an m6A eraser implicated in metabolic homeostasis [154] and is associated with type 2 diabetes [155] CARM1, coactivator associated arginine methyltransferase 1; DNMT3B, DNA methyltransferase 3B; FTO, fat mass and obesity-associated protein; GLYATL1, glycine-N-acyltransferase like 1; HNRNP, heterogeneous nuclear ribonucleoprotein; KAT2A, K (lysine) acetyltransferase 2A; KDM, lysine (K)-specific demethylase; NSD, nuclear receptor binding SET domain protein; MBD, methyl-CpG binding domain protein; PHF2, PHD finger protein 2; PRMT, protein arginine methyltransferase; SETDB, SET domain bifurcated; SETMAR, SET domain and mariner transposase fusion protein; SIRT1, sirtuin 1; SUV39H, suppressor of variegation 3-9 homologue; YTH, YTH domain protein. Blank fields in the mechanistic and profiling studies refer to either enzymes or modified determinants that were not reported in the studies listed

Fig. 2

Epigenetic and metabolic memory nexus. Persistent epigenetic changes in the context of transient, medium- and long-term metabolic memory. In certain cell types, hyperglycaemic memory exists. For example, in response to hyperglycaemia, SET7 methyltransferase enzyme writes mono-methylation of histone H3 lysine 4 (H3K4me1) in vascular endothelial cells, and this methylation is retained over the long term. The effectiveness of transient hyperglycaemic stimuli to tightly control the expression of target genes implicated in vascular dysfunction and inflammation relies on ROS-mediated pathways. Indeed, a hyperglycaemic sensor (SET7) and the transfer of a chemical group (mono-methylation on H3K4) to a gene are two common principles implicated in metabolic memory. These regulatory events serve to remodel chromatin to precisely decorate target genes such as the p65 subunit of NFkB, with H3K4me1 corresponding with persistent transcriptional activation [24, 26, 30, 102]. The long-term contribution of hyperglycaemic signalling cues derived from the monocytes of type 1 diabetic individuals from the DCCT and EDIC trials show HbA_1c levels and H3K9ac are also tightly linked [31]. The control of genes related to the NFkB inflammatory pathway by modifications exhibits common principles of epigenetic control that serves as a paradigm for metabolic memory. The mechanism conferring H3K9ac and metabolic memory remains poorly characterised. Clearly, this explains only part of the epigenetic complexity, in the metabolic memory example and using monocytes derived from the DCCT and EDIC trials, genomic modification, specifically, 5mC is also an important determinant in the control of gene expression [54]. The mechanism underlying 5mC modification in metabolic memory remains poorly characterised. Recent advances in our understanding of the mechanisms of genomic modification show 5hmC, a DNA base derived from 5mC by oxidation by TET enzymes, is implicated in transposon activity associated with exposure to adverse in utero programming and gestational diabetes [150]. These findings emphasise that cytosine residue modification plays an important role in the regulation of genes implicated in diabetes