Epigenetic modifications and the development of kidney graft fibrosis

Abstract

Purpose of review

To outline recent discoveries in epigenetic regulatory mechanisms that have potential implications in the development of renal fibrosis following kidney transplantation.

Recent findings

The characterization of renal fibrosis following kidney transplantation has shown TGFβ/Smad signaling to play a major role in the progression to chronic allograft dysfunction. The onset of unregulated proinflammatory pathways are only exacerbated by the decline in regulatory mechanisms lost with progressive patient age and comorbidities such as hypertension and diabetes. However, significant developments in the recognition of epigenetic regulatory markers upstream of aberrant TGFβ-signaling has significant clinical potential to provide therapeutic targets for the treatment of renal fibrosis. In addition, discoveries in extracellular vesicles and the characterization of their cargo has laid new framework for the potential to evaluate patient outcomes independent of invasive biopsies.

Summary

This review summarizes the main findings in epigenetic machinery specific to the development of renal fibrosis and highlights therapeutic options that have significant potential to translate into clinical practice.

Introduction

End-stage-renal disease (ESRD) continues to be a major world health problem and has been projected to increase by 29–68% in the next 10 years [1]. While dialysis and supportive care can prolong ESRD, kidney transplantation remains the best alternative to improve patient wellbeing and quality of life. However, complications of surgery or of the donor organ has reinstated many patients back to ESRD management or awaiting a secondary transplantation. These outcomes partially originate from the lack of criteria for evaluating the donor organ and the current limitation to predict long-term graft survival. Presently, clinical outcomes in kidney transplant patients are evaluated not only on clinical parameters (elevated systolic blood pressure, depressed glomerular filtration rate, albumin-to-creatinine ratio in urine) but also by the progressive development of interstitial fibrosis and tubular atrophy (IFTA) in successive post-transplant biopsies [2]. Developments in functional MRI have improved non-invasive diagnostics but have yet to replace the need for biopsies [3]. Instead, in an attempt to non-invasively monitor the kidney graft, transplant researchers now search for alternative methods of evaluating the graft and have partially turned toward the assessment of molecular biomarkers present in urine, plasma, and serum. Although molecular studies still, to some degree, rely upon graft biopsies, the improved evaluation of circulating markers can have a profound impact on reducing the number of biopsies done [4]. These studies, which are further strengthened by the advancement of new methodological advances such as single-cell RNAseq and the increased awareness of epigenetic fluctuations specific to both the donor organ and the recipient, aim to reframe how clinicians predict clinical outcomes [5]. In this review, we outline the major recent developments in renal fibrosis and characterize the extensive epigenetic machinery specific to fibrosis development in kidney transplant grafts, highlighting the current understanding of molecular biomarkers and the potential impact of such on clinical practice.

Pathogenesis of chronic allograft dysfunction to renal fibrosis

Late graft loss continues to be a major problem after kidney transplantation (KT), mainly as consequence of death with a functioning graft and intrinsic allograft failure (or chronic allograft dysfunction (CAD)). CAD remains a major cause of allograft attrition over time, resulting in reinstitution of end-stage renal disease care. CAD is considered by many to be a variant of chronic kidney disease (CKD), with both immune and non-immune mechanisms, contributing to the development of IFTA and progressive loss of graft function. Early development of graft fibrosis is predictive of late graft function. (Figure 1)

Figure 1.

Kidney transplantation remains one of the best options to restore patient wellbeing and quality of life following the induction of kidney disease. However, non-immunological stressors such as progressive age, prolonged cold ischemia time, donation after cardiac/brain death, or procurement of a kidney which suffered from a previous AKI can prompt epigenetic modifications that instigate dysfunctional recovery mechanisms following transplantation. These mechanisms then propagate towards the migration/activation of leukocytes, fibroblast activation, and deposition of ECM, ultimately culminating in the development of IFTA and decreased GFR through glomerulosclerosis. After the initiation of IFTA and decrease in GFR, the kidney graft will eventually progress to CAD and the patient will need to be returned to ESRD-management or re-enlisted on the transplant waiting list.

Specifically, renal fibrosis following transplantation is a multifactorial injury which originates from not only the damage incurred through the transplantation process but also via emerging mechanisms associated with endothelial-to-mesenchymal transition (EMT), cell cycle arrest in the G2/M phase, and metabolic disorders [6,7,8]. Upstream of these aberrant mechanisms are a milieu of dysregulated signaling pathways. Two of which, the TGFβ/Smad [9,10] and the Wnt/β-Catenin signaling axis [11,12], encompass most downstream effectors that are specific to kidney transplantation. It has been described that the degree of renal fibrosis and the associated risk to progress to CAD correlates to the length of cold ischemia time [13,14,15]. This would suggest that the severity of the ischemia-reperfusion injury (IRI) associated with the transplantation procedure instigates many of the dysregulated signaling pathways lending to dysfunctional recovery mechanisms and lays the foundation for the fibrotic scar.

IRI is associated with damage to the tubular epithelial cells characterized by the loss of cell-adhesion molecules and a denuded basement membrane, further contributing to damage of the proximal tubular cells and cast formation [16]. In ideal conditions, the activation of hypoxia-inducible factor-1-α (HIF1α) in concert with the ischemic damage to the proximal tubular cells would result in an upregulation of miRNA-23 from epithelial cells that function as a promoter of neutrophil and monocyte infiltration and induction of tubulointerstitial inflammation [17,18]. Resolution of macrophages from a pro-inflammatory phenotype (M1) to pro-repair (M2) signals the initiation of the re-epithelialization process [8]. However, secondary to prolonged cold ischemia or transplantation from marginal or so-called non-ideal organ donors (i.e., donation after cardiac/brain death, donor of advanced age) can trigger dysfunctional repair mechanisms. Ischemic damage, specifically, can upregulate the expression of long-noncoding RNA (lncRNA) Erbb4-IR which directly binds and inhibits Smad-7, a crucial regulator of TGFβ1’s inflammatory effects, leading to increased transcription of collagen I and α-smooth muscle actin (αSMA) [9]. With a currently ageing population, the number of donor organs from older recipients has increased during last years and is expected to continue, representing an important percentage of the available donor organ pool for kidney transplantation. In aged donors (from donors older than 65 yo.), aberrant TGFβ1 signaling is only exacerbated by the unregulated Wnt9a activity that is accompanied with progressive age [19].

Historically, therapeutic developments have targeted the TGFβ1/Smad signaling axis as a potential to halt the progression to fibrosis. More recently, it was found that TGFβ1 signaling could be redirected from profibrotic to anti-inflammatory through the delivery of T-cell factor, which enhanced the interaction between β-catenin and FoxO, thereby bolstering the differentiation of TGFβ1-induced regulatory T-cells (iTregs) [20*]. However, aberrant TGFβ1 signaling could be downstream to dysfunctional epigenetic regulatory mechanisms. In a study by Zhou et al., TGFβ1-induced EMT and the epithelial cell cycle arrest in the G2/M phase that are characteristic of progressive renal fibrosis were suppressed through the silencing of enhancer of zeste homolog-2 (EZH2), a methyltransferase that induces the trimethylation of histone H3 lysine 27 (H3K27m3). In fact, suppression of EZH2 led to the downregulation of key fibrotic genes, Snail-1 and Twist-1, suggesting that EZH2 might be upstream of multiple signaling pathways and could be a therapeutic target for treatment of renal fibrosis [21*]. (Figure 2)

Figure 2.

Prolonged cold ischemia time is directly correlated with the progression to IFTA and CAD in a transplanted kidney. Hypermethylation of the graft can induce dysregulated TGFβ-signaling which upregulates downstream effectors Smad 1–4, thereby contributing to the initiation of EndMT, cell cycle arrest in G2/M phase, and a host of metabolic disorders. This in concert with upregulated Wnt/β-Catenin signaling secondary to progressive age or co-morbidities, such as hypertension and diabetes, can further magnify TGFβ aberrant mechanisms through an increase in macrophage infiltration and shift polarization toward a pro-inflammatory M1 phenotype. However, recent therapeutic developments in the use of T-cell factor or the downregulation of EZH2 can shift TGFβ-signaling towards the anti-inflammatory Smad-7 and the increase differentiation of iTregs.

Epigenetics: new insights into chronic allograft dysfunction

A major shift in understanding fibrosis in renal grafts has been the recognition that epigenetic markers, regulatory processes which control gene expression without changing DNA sequence, integrate the various intrinsic and extrinsic regulatory mechanisms which trigger fibrogenesis. The epigenome represents the merging point between genetics and environment and, in the case of transplantation, the environment shifts from donor to recipient thereby providing a clear alteration in epigenomic processes, including DNA methylation (DNAm) patterns, histone modifications, and the action of non-coding RNAs (ncRNA), with microRNAs (miRNA) being the more studied in this last group.

DNA methylation in the kidney graft

DNA methylation (DNAm) occurs by the addition of a methyl group to the 5-carbon position of the cytosine ring in genomic DNA, thereby creating a 5-methylcytosine (5mC), by methyltransferase enzymes (DNMTs). Typically, these enzymes target areas rich in cytosine-guanine repeats, termed ‘CpG islands’ [22]. CpG islands are associated with the promoter regions of genes and are present in two states: Heterochromatin, highly methylated DNA with a multitude of histone repressive modifications, or Euchromatin, unmethylated and transcriptionally permissive regions [23]. Still more intergenic methylation can be present at enhancer or insulator elements at adjacent sites termed ‘CpG island shores’. CpG island shores present more variable DNAm patterns as compared to CpG islands and are highly correlated with variations in gene expression [24–27].

DNAm represents a relatively stable and reversible process. Its removal can be initiated by ten-11 translocation (TET) enzymes, which convert 5mC to 5-hydroxymethylcytosine (5hmC) [28,29]. These epigenetic processes thus shift with the environment from donor to recipient and in turn to the exacerbation of oxidative and inflammatory stress. Fluctuations in DNAm are represented in a variety of pathologies which could negatively affect graft outcomes. Aging has specifically been found to be associated with a loss of kidney integrity and expressed transcriptional changes secondary to aberrant structural remodeling, decreased blood flow, and a progressive decrease in glomerular filtration [30, 31, 32*]. Heylen et al. investigated the role of age-related changes in DNA methylation patterns in relation to post-transplant renal fibrosis. In the study, the authors found that TET demethylation enzymes were downregulated thereby contributing to age-associated hypermethylation. In addition, 11.5% of the assessed CpG islands and shores were found to be significantly hypermethylated. In the same study, Heylen et al. reported that loss of TET-activity, independent of cold ischemia time, contributed to the accumulation of oxidative stress and the accompanied loss of DNA hydroxymethylation. Accumulation of these changes contributed to the downregulation of key inhibitors (specifically Dkk1 and Dkk2) of the Wnt/β-Catenin signaling pathway [30], thereby stimulating the progression of interstitial fibrosis, glomerulosclerosis, and CAD (33–37).

Unraveling the cause and effect of DNAm on graft survival independent of other confounding variables presented a major limitation in recognizing biomarkers. To this end, Bontha et al. investigated the use of a multidimensional systems approach that focused not only on global kidney graft DNAm, but also the downstream gene expression and variations in miRNA in pre- and post-transplant biopsies with progressive IFTA and declining graft function [38*]. Many of the mechanisms elucidated showed DNAm of CpG island shores to be the root cause for upregulating immune-mediated injury while suppressing metabolic mechanisms necessary for graft function. Further evaluations of data from our studies have shown significantly differentially hypomethylated DNAm patterns in kidney graft biopsies from transplant patients with histological evidence of IFTA after 24 months post-transplantation. When compared to gene expression patterns, the same biopsies showed lymphocyte activation, leukocyte migration, adaptive immune response, leukocyte activation involved in immune response, regulation of immune effector processes, positive regulation of hydrolase activity, positive regulation of cytokine production, lymphocyte migration, and regulation of leukocyte migration after evaluating gene enrichment functions (Figure 3). Otherwise, hypermethylated CpG genes related to decreased gene expression of metabolism-related genes (i.e. carbohydrate metabolic processes, cellular modified amino acid metabolic process, among other). These results strongly support a role of DNA methylation and regulation of downstream effectors during kidney graft fibrogenesis. (Figure 3)

Figure 3.

Network of enriched terms when using significantly differentially hypomethylated DNAm patterns and over-expressed genes identified from same kidney graft biopsies from transplant patients with histological evidence of IFTA after 24 months post-transplantation (a) colored by cluster ID, where nodes that share the same cluster ID are typically close to each other, (b) network of enriched terms presented as pie charts, where pies are color-coded based on the identities of the gene lists. DNAm: DNA methylation; IFTA: Interstitial Fibrosis and Tubular Atrophy

Micro RNAs and graft fibrosis

As mentioned above, recently miRNAs have come into focus as powerful epigenetic regulators of gene expression. MiRNAs are small, non-coding RNAs that cause the repression of target genes through the posttranscriptional degradation of mRNA and/or translational inhibition of protein expression [39]. Numerous miRNAs have been identified to reinforce the pathogenesis of interstitial fibrosis. Studies demonstrate that mir-155–5p [40,41], mir-21 [41–45], miR-103a-3p [46], mir-142–3p [44], mir-153–3p [47], mir-196b-5p [48], and mir-184 [49] promote renal fibrogenesis, while the mir-29 family [42,50], the mir-30 family [42], mir-342–3p [51], mir-133a [41], mir-200b [41], and mir-148a [52] inhibit apoptosis and fibrosis through a myriad of interconnected mechanisms.

A complex epigenetic network exists whereby even miRNA expression is under epigenetic control. In our previous study, integrative approaches were used to further explore whether DNAm could regulate gene expression indirectly via the regulation of miRNA. Through gene expression analyses of paired kidney allograft biopsies with IFTA, genes were identified that were regulated by hypomethylated miRNAs (located in the TSS region). Most of these genes were related to metabolomic processes and were notably downregulated [37].

Several fibrosis-associated miRNAs have recently been identified in kidney disease. In a recent study, Schauerte et al., using a murine model of allogenic kidney transplantation with CAD, identified fibrosis-associated miR-21a-5p by whole miRNAome expression analysis to be among the most highly upregulated miRNAs. In renal fibroblasts cultured in vitro, miR-21a-5p was transcriptionally activated by interleukin 6–induced signal transducer and activator of transcription 3. Co-culture of LPS-activated macrophages with renal fibroblasts increased expression levels of miR-21a-5p and markers of fibrosis and inflammation. In addition, mature miR-21a-5p was secreted by macrophages in small vesicles, which were internalized by renal fibroblasts, thereby promoting profibrotic and proinflammatory effects. Notch2 receptor was identified as a potential target of miR-21a-5p and validated by luciferase gene reporter assays. Therapeutic silencing of miR-21a-5p in mice after allogenic kidney transplantation resulted in an amelioration of CAD, as indicated by a reduction in fibrosis development, inflammatory cell influx, tissue injury and BANFF lesion scoring. The results support an antagonistic role of miR-21a-5p having beneficial effects on kidney function. Further evaluations of potential miR-21a-5p silencing may therefore be a viable therapeutic option in the treatment of patients following kidney transplantation to avoid/decrease the development of CAD [53].

Long-noncoding RNAs and graft fibrosis

It has been recently demonstrated that long-noncoding RNAs and miRNAs are exceptionally interconnected, thus, many studies are currently attempting to define the intricate relationship between these circulating biomarkers. For instance, Lnc-1700020I14Rik was discovered to downregulate mir-34a [54], which was previously found to induce fibrosis through the TGF-β pathway [42]. Further, Chen et al. found that under fibrotic renal conditions, lncRNA LINC00667 expression was decreased and tightly regulated by mir-19–3p [55]. Similarly, lncRNA Erbb4-IR was revealed to promote fibrotic renal injury in diabetic mice by targeting miR-29b [50]. Erbb4-IR is further reported to be highly expressed in TGF-β/Smad3-mediated renal fibrosis [56]. Alternatively, increased expression of lnc-TSI [58], lnc-MEG3 [54], and lnc-1700020I14Rik [54] resulted in the downregulation of the TGF-β/Smad3 pathway, demonstrating a potential therapeutic strategy for decreasing renal injury. Although promising, the study of lncRNAs is still in its infancy. For instance, MALAT1 was once proposed to play a role in ischemia-reperfusion injury and AKI [59], however others were not able to confirm its significance in vivo [60]. Still, there is great potential for future clinical interventions involving noninvasive circulating biomarkers.

Extracellular vesicles in the progression of renal fibrosis

Extracellular vesicles (EVs) are involved in cell–to–cell communication and they can pass from the systemic circulation into endothelial cells and tubular epithelial kidney cells and into the urine. RNA can be detected in body fluids in a nuclease resistant form, mainly as part of EVs, making extracellular RNAs (exRNAs) excellent non-invasive markers for disease detection, risk prediction, and therapeutic intervention (61, 62, 63). The content of EVs include a multiplicity of proteins, miRNAs, lncRNA, mRNAs, and lipids [64,65]. MiRNAs are the most studied among the different classes of exRNAs and recent evidence suggest that c-miRNAs, not only function as disease biomarkers, but also as critical regulators of cellular crosstalk [66–69]. Previous studies have shown that EVs can transfer TGFβ1 mRNA and miRNA-34a between different resident cells of the kidney to propagate pro-fibrotic signaling through fibroblast activation, inflammation, and tubular atrophy [70–73]. In addition, urinary EVs are constitutively released and the selection of their cargo is likely to be representative of the intracellular state, thereby allowing EVs to potentially be used as biomarkers of renal injury [74]. Sonoda et al. explored the implication of acute kidney injury on the repertoire of cargo associated with urinary EVs. Using an acute ischemic model of injury, miR-16–5p, miR-24–3p, miR-200c-3p were increased at one day post-reperfusion. The target mRNA expression levels of these three miRNAs were collectively increased in the medulla suggesting that the EV cargo correlated with the intracellular state. However, three days following reperfusion, the assortment of miRNA expression shifted to favor those involved in TGFβ1 signaling mechanisms, such as the miR200-family, miR-148a-3p, and miR-9a-5p [71].

Herein, we also propose to evaluate tRNA fragments. These ncRNAs are mainly classified in groups: tRNA halves (30–40nt), tRF-3s or tRF-5s (18–22nt) or tRF-1s (variable length) from 3’ trailer sequence of tRNA [75] tRNA halves were described as associated with response to cellular stress [76–78]. Furthermore, tRNA halves are highly expressed in lymphoid tissues, acting as signaling molecules in immune responses [79].

Further studies characterizing the composition of EV cargo between pre- and post-transplant patients who have variable presentations of renal fibrosis can provide a deeper understanding of the cell-to-cell communication which is suspected to be driving fluctuations in epigenetic regulation and downstream progression of CAD.

Conclusions

Key discoveries regarding the shift in epigenetic regulatory mechanisms following kidney transplantation has significant implication to redefine the prediction of ultimate graft outcomes. In addition, recently reports about EVs and their potential to promote shifts in paracrine signaling secondary to cellular stress mechanisms has profound translational potential as new methods to capture and quantify EV cargo in circulation and urine can provide biomarkers specific to the transplanted kidney before the onset of renal fibrosis and the progression to CKD.

Key points:

• Renal fibrosis and the progression to chronic allograft dysfunction continues to return patients to managing end-stage-renal disease or awaiting transplantation.

• TGFβ/Smad signaling seems to be at the center of the onset of fibrosis and is further exacerbated by dysfunctional Wnt/β-catenin signaling that accompanies progressive age and comorbidities such as hypertension and diabetes.

• Significant progress has been made in recognizing the key DNAm, miRNA, and lncRNA fluctuations which accompany the onset of renal fibrosis.

• The characterization of EVs and their cargo has significant implications to expand our ability to assess graft outcomes independent of invasive procedures.