IRI is inevitable for kidney transplantation from deceased donors to recipients (20) and plays a pivotal role in occurrence of DGF, which negatively affects long term function of the transplant (21). DNA methylation has been reported to be implicated in renal IRI (14). In 2006, Pratt and colleagues demonstrated the demethylation at the C3 complement gene promoter in rat kidney with 24 hours of cold ischemia and a 2 hours of reperfusion in an isolated ex-vivo circuit, which led to local C3 synthesis and tissue injury (22). Their follow-up study showed that this demethylation of C3 promoter influenced pathological change in chronic allograft nephropathy (23). In contrast, Mehta et al. demonstrated the hypermethylation of CALCA (Calcitonin Related Polypeptide Alpha) gene in urine of KT patients with biopsy-proven acute tubular necrosis, suggesting the biomarker potential of the hypermethylation (24). In 2018, Heylen et al. (25) took advantage of microarray to analyze paired pre- and post-ischemic biopsies of brain-dead transplants in a longitude cohort to find out an overall increase in methylation after ischemia. Further cross-sectional cohort genome-wide DNA methylation analysis showed a cold ischemia time-dependent hypermethylation in majority of methylated CpG sites. Functionally, IRI-induced hypermethylation predominantly involved in suppression of anti-apoptotic and anti-fibrotic genes and correlated with chronic allograft injury. For example, the highest increase in methylation is in the DDR1 promoter, which is known to be involved in apoptosis and kidney fibrosis. These findings indicate that DNA methylation represents a promising therapeutic target for preventing cold ischemia-associate acute injury and chronic allograft injury.

Acute rejection is recognized as one of the most important cause of graft loss and it is composed of aforementioned ACR and AMR (9). During ACR, mononuclear cells like CD4⁺ or CD8⁺ T cells accumulate in interstitium, along with cytokines or chemokines such as interferon γ (IFN γ), tumor necrosis factor β (TNF β), TNF α, etc. (10). CD4⁺Foxp3⁺ regulatory T (Treg) cells are essential for the maintenance of immune tolerance through suppressing excessive immune response (26) and their characteristic transcriptional factor Foxp3 is regulated epigenetically to modulate immune homeostasis (27, 28). Bestard et al. reported that demethylation of Foxp3 is associated with intra-graft higher expression of Treg cells and a favorable long-term allograft outcome in subclinical rejection patients (histologic evidence of rejection exists in biopsy but lack of clinical kidney dysfunction) (29). In contrast, Boer et al. reported that hypermethylation of the immune inhibitory receptor programmed death 1 (PD1) in CD27^- memory CD8⁺ T cell correlates positively with acute rejection at 3 months (30). Therefore, different genes may be subjected to the regulation by methylation in different T cells affecting the immune response and outcome of kidney transplantation. In 2020, Zhu et al. analyzed DNA methylation in peripheral blood mononuclear cells (PBMC) of patients with or without AR-induced allograft dysfunction and demonstrated that hypermethylated genes are enriched in T cells and mechanistic target of rapamycin signaling (mTOR) pathways in AR-related graft dysfunction. Further work using mice AR model of renal transplantation showed that the DNMT inhibitor decitabine could ameliorate kidney allograft inflammatory injury via demethylating and enhancing the negative regulators of mTOR signaling (31). These results indicate that inhibition of DNA methylation may provide therapeutic effects against acute renal allograft rejection.

IF/TA is a main signature of late allograft dysfunction and the mechanism leading to IF/TA includes inflammation, renal fibroblast activation, and deposition of extracellular matrix (32). Several lines of evidence indicate that DNA methylation plays a role in development and progression of IF/TA in renal allograft. In 2017, Bontha and colleagues firstly used integrative multi-omics approach (methylation arrays, gene expression arrays and miRNA arrays) to evaluate the overall distribution of epigenetic modifications across genes and showed their relationship with renal graft IF/TA, function and long-term outcome. By comparing the biopsies of IF/TA and non-fibrosis/atrophy (NFA) at 24 months post-transplantation, they observed that enrichment of hypomethylated CpGs sites correlated with inflammation in IF/TA biopsies, whereas hypermethylation corresponded to the inhibition of kidney repair. This study also showed that hypomethylation could also modulate immune genes expression by regulating miRNAs (33). As mentioned above, Heylen et al. suggested that ischemic insult caused general hypermethylation and predicted CAD 1 year after transplant (25). Their further study unveiled the mechanism underlying aging donor related kidney allograft dysfunction. Donor aging associated differentially methylated regions were frequently located in the genes involved in the Wnt/β-catenin signaling pathway, which affect glomerulosclerosis and interstitial fibrosis as well as graft function at one year after transplantation (34). In addition, Rodriguez et al. performed genome-wide analysis of DNA methylation in PBMCs from KT patients with chronic rejection or operational tolerance (OT), a condition of stable graft acceptance without the need for immunosuppression therapy. They found that OT is associated with demethylation in genes involved in immune function like T, B cell activation and Th17 differentiation, while chronic rejection is related to demethylation in intracellular signaling and ubiquitination pathways (35). The question is arising how epigenetic mechanisms can affect operational tolerance. Previous transcriptomic analysis showed that OT was associated with B cell profile (enrichment of B cells especially naïve and transitional B cells which can produce IL10 and inhibit CD4+ T cells) (36, 37). Consistently, the DNA methylation profile of OT patient showed demethylation of CD20 encoding genes to result in survival of transitional B cells and expansion of these populations (35). Operational tolerance is also related to poor Th17 response and reduced TCR signaling (38, 39). Although partially, the authors observed some molecular pathways associated with Th17 functions. For example, they found demethylation of PD1 and BTLA and the genes associated with negative regulation of ERK and NF-κB, which led to damage of T-cell activation and Th17 response (35).

A most recent study compared epigenome-wide methylation modification between healthy patients and renal failure replacement therapy patient (dialysis and kidney transplantation) at baseline (right before renal failure replacement therapy) and over 12 months treatment, they found that uremic milieu drives genome-wide methylation changes but partially reversed with kidney failure replacement therapy. However, 413 CpG sites remained differentially methylated at follow-up in renal replacement therapy, which merit further investigation (40). These observations suggest a role of DNA methylation in chronic inflammation, IF/TA and chronic allograft dysfunction in kidney transplantation. Future investigation needs to pinpoint the specific genes and related mechanism underlying their pathogenic functions.

In 2008, Marumo et al. firstly demonstrated that renal ischemia in mice induced a transient decrease in histone acetylation in proximal tubules. During reperfusion, HDAC5 was downregulated in parallel with the recovery of histone acetylation. Downregulation of HDAC5 was associated with the acetylation and expression of BMP7, which contributed to kidney repair and regeneration (45). In 2015, Levine et al. reported that the pan-HDAC inhibitor trichostatin (TSA) and class I specific inhibitor MS-275 ameliorated renal IRI in acute phase and diminished fibrosis formation in the long term. They also conducted renal syngeneic transplantation with prolonged cold ischemia in mice to figure out the anti-fibrotic effect of TSA (46). To elucidate the specific role of individual HDAC isoforms, their most recent study showed that global or renal tubule conditional (PAX8^cre) inducible ablation of HDAC2 (but not HDAC1) protected against IRI-induced renal injury and formation of fibrosis. Likewise, in cold storage/transplantation model, mice received HDAC2^-/- isografts showed superior survival compared with recipients of WT isografts and had the trend of long-term functional protection and less fibrosis. These results support a pathogenic role of HDAC2 in renal IRI during kidney transplantation (47). Our latest study found that HDAC3 specific inhibitor RGFP966 diminished cold storage/transplantation injury and improved renal allograft function in mice kidney transplantation. This inhibitor and HDAC3 knockdown also protected against cold storage/rewarming injury in rat kidney proximal epithelial cells, indicating HDAC3 also takes part in IRI of kidney transplantation (48). In the sirtuin family, SIRT1 (49, 50), SIRT3 (51) and SIRT5 (52) were reported to reduce kidney IRI mainly through mitochondrial mechanisms (mitochondria dynamic and biogenesis). In view of the crucial role of mitochondria in warm and cold renal IRI in kidney transplantation (53, 54), agonists or activators of SIRT1, SIRT3, and SIRT5 may mitigate injury and improve graft function in kidney transplantation. There are considerable studies implicating HDACs in kidney fibrosis (55–57), but very limited is known about the involvement of HDACs in IF/TA in renal transplantation. In aforementioned reports by Levine and colleagues, TSA not only ameliorated acute ischemia-reperfusion injury, but also reduced fibrosis substantially in kidney syngeneic transplantation with prolonged cold storage (46). Specifically, HDAC2 knockout grafts showed the trend of less fibrosis than WT grafts (47). Zou et al. established F344-Lewis rat CAD model to find that Sirt1 (58) and Sirt3 (59) were both decreased at 3 months after transplantation, and their expression negatively correlated with renal failure, inflammatory chemokines expression and interstitial fibrosis severity, suggesting that sirt1 and sirt3 may serve an important protective role in the early stage of CAD. However, the underlying mechanisms and the roles of other HDACs in chronic renal rejection and associated fibrogenesis remains unclear. Obviously, further research in this area will provide new insights into allograft function loss and may lead to new therapeutic target.

Non-histone acetylation is complicated with diverse human diseases and key cellular processes. Mostly extensively studied proteins that can be acetylated includes p53, tubulin, p65, heat shock protein-90, etc., with more and more non-histone proteins being recognized nowadays (44). Foxp3 acetylation is an important modulatory mechanism for immune response in solid organ transplantation (60). Acetylated Foxp3 is more stable than its non-acetylated form, because it prevents ubiquitination at target lysine residues and, in turn, proteasomal degradation ( Figure 3 ). Acetylation also improves the ability of Foxp3 binding to chromatin and augments its function as a transcriptional regulator in Treg cells (61). In 2016, Levine et al. (62) reported that inhibition of SIRT1 led to the stabilization of the Treg phenotype by increasing Foxp3 acetylation resulting in better renal allograft survival and function in mice. Specifically, in a mouse model of renal allograft transplantation, CD4-Cre directed ablation of sirtuin-1 as well as a sirtuin-1 specific inhibitor (EX-527) improved the survival and kidney function in recipient mice. It is worth noting that EX-527 at higher doses did not improve renal allograft function but trend to pose inferior effects, suggesting that SIRT1 may act protectively on non-T cells (49, 50, 63). These studies indicate that the time window and dose of Sirt1 inhibitors used in targeting immune response merit careful consideration. Similarly, it is plausible to speculate that inhibition of Zn-dependent HDACs has the potential to fortify immune tolerance by increasing Foxp3 acetylation in Tregs. In this regard, Levine and colleagues published a series of impressive studies using the model of cardiac allograft transplantation. They initially reported that the pan-HDAC inhibitor TSA in conjunction with Rapamycin induced permanent, Treg-dependent cardiac allograft survival and donor-specific graft tolerance (64). Then, they found that recipient mice receiving HDAC6-knockout Treg showed better cardiac allograft survival than those receiving WT Tregs (65). Moreover, Treg-specific deletion of HDAC2 (66), HDAC10 (67) and HDAC11 (68) promoted Treg function and prolonged cardiac allograft survival significantly. Despite these remarkable observations in cardiac allografts, there is no report about HDAC inhibitors in augmenting immune tolerance in kidney transplantation. It is still a long way for clinical use of HDAC inhibitors for the difficulty to have specific, non-toxic inhibitors and the complexity of biological effects of HDACs.

In addition to acetylation, there is evidence supporting the role of methylation and phosphorylation of histone in renal IRI or kidney transplantation. Histone methylation is the addition of methyl group to lysine or arginine of core histones, which is mediated by histone methyltransferase (HMT) (69). Different from histone acetylation, histone methylation may be associated with gene transcriptional permission or repression. The effect depends on the amino acid residue methylated and the extent of modification (mono-, di-, and tri-) (70). For example, H3K4 methylation is associated with activation of transcription, while methylation of H3K27 and H3K9 results in transcriptional silencing (71). Naito et al. reported that in unilateral IRI mice, H3K4m3, H3K9ac and H2A.Z variant of HMG CoA reductase gene were increased at their promoters and exons. These permissive modifications allow for tubular cholesterol accumulation and renal protection (72). In contrast, another study showed that unilateral IRI in rat induced the recruitment of H3K4m2 to the promoter of TGF-β, which is consistent with TGF-β transcription and kidney injury (73). Matrix chromatin immunoprecipitation showed that renal IRI induced permissive histone marks (H3K9Ac, H3K18Ac, H3K27Ac, H4K5/8/12/16Ac, H3K4m3, H3K4m2 and H3S10ph) at the proinflammatory gene TNF-α, whereas repressive histone mark (H3K27m3) at the gene was decreased (74). In terms of histone methyltransferase, enhancer of zeste homolog 2 (EZH2) (75, 76) and G9a (77) catalyze methylation of H3K27 and H3K9/H3K27, respectively. Their inhibitors were reported to ameliorate renal IRI. What’s more, EZH2 inhibitor suppressed immune response of alloreactive T cell and inflammatory cytokine activation, resulting in attenuation of acute allograft rejection in rat kidney transplantation (78). Zhang et al. used chromatin immunoprecipitation with DNA microarray (CHIP-chip) to determine H3K4 methylation of monocytes in acute rejection or non-AR rats. They detected significant difference in H3K4m3 levels in 141 probes, but their correlation with AR after kidney transplantation needs further clarification (79).

Histone phosphorylation involves phosphorylation of serine, threonine and tyrosine residues on histones mediated by protein kinases and phosphatases (80). The best-known function of histone phosphorylation is in DNA damage response, where H2AX variant is phosphorylated on serine 139 to mediate DNA double strand break repair (81). Besides, histone phosphorylation can also function in transcriptional regulation and chromatin compaction, the latter is associated with mitosis/meiosis and apoptosis (80). Our study in 2014 firstly demonstrated the induction of phosphorylation of H2AX and its kinase ATM in renal IRI, indicating involvement of DNA damage response in this disease (82). Of note, histone modification by these mechanisms is a largely untapped area to explore in kidney transplantation.

In 2010, our laboratory generated a proximal tubule (PT) specific Dicer ablation mouse model and demonstrated the resistance of these mice to renal IRI, suggesting a pathogenic role of miRNAs (119). In the same study, our microarray analysis further identified the specific microRNAs that were up- or down- regulated in renal IRI (119). Lee et al. extracted peripheral blood from kidney transplant recipient right before transplant surgery and immediately after reperfusion, in the purpose of comparing the microRNA profile of non-AKI and AKI in kidney transplantation. They observed 3 upregulated and 6 downregulated miRNAs post-reperfusion. But their function in IRI of kidney transplantation need further clarification (87). In 2019, Khalid et al. used microarray analysis of “first-pass urine” sample (taken immediately post-transplant) from DGF and non-DGF patients and identified 6 upregulated microRNAs (miR-9, -10a, -21, -29a, -221, and -429) in DGF (88). Another independent cohort further verified the increase trend of the 6 candidate microRNAs in urine samples at varying time in the first week post-transplantation, with miR-21 being the most evident (88). This study suggested the biomarker potential of the urinary microRNAs for early prediction or diagnosis of DGF following kidney transplantation. Besides, miRNAs in peripheral blood have been suggested as promising non-invasive biomarkers. For example, upregulation of miR-142-5p, miR-142-3p and miR-223 and downregulation of miR-10b were detected in peripheral blood mononuclear cells (PBMC) at 3-4 weeks earlier than renal dysfunction shown by serum creatinine increase in human kidney transplantation patients. The elevations of miR-142-5p and miR-142-3p were also associated with later cystatin C increases. These results indicate that miRNAs in PBMC may be useful in predicting renal allograft dysfunction than conventional biomarkers, such as serum creatinine and cystatin C (89). Recently, exosomes, a type of extracellular vesicle carrying abundant contents including miRNAs, have attracted extensive interest (120). Wang et al. applied high-throughput sequencing to explore the miRNA expression profile in exosomes from peripheral blood of allograft recipients with or without DGF. They observed the up-regulation of hsa-miR-33a-5p_R-1, hsa-miR-98-5p, and hsa-miR-151a-5p in DGF recipients and demonstrated a positive correlation of hsa-miR-151a-5p with first-week renal function post-transplantation (90).

As for lncRNAs, Pang et al. examined a mouse syngeneic kidney transplantation model and demonstrated the upregulation of a specific LncRNA called maternally expressed gene 3 (MEG3) in the first few days of post-engraftment (105). Notably, silencing MEG3 resulted in the attenuation of kidney damage and dysfunction, suggesting an injurious function of MEG3 in this model. Mechanistically, MEG3 was shown to compete with miR-181b to induce TNF-a and related acute renal allograft injury (105). Nagarajah et al. collected blood samples from deceased donor kidney transplant with DGF or immediate graft function (IGF). In these samples, lower lncRNA MGAT3-AS1 was associated with DGF, indicating its predictive role in short-term outcome of transplantation (106). Zou et al. compared prolonged cold ischemia (16h) rat allogeneic kidney transplantation with short ischemia (2h) rat syngeneic or allogeneic transplantation. They found that lncRNA PRINS (Psoriasis susceptibility-related RNA gene induced by stress) was elevated in allograft after 16h of cold ischemia and correlated with acute tubule damage (107). However, in the absence of prolonged ischemia syngeneic transplantation group, it’s hard to discriminate whether the lncRNA PRINS play a role in AKI or acute rejection.

Pang et al. found that injection of immature dendritic cells-derived exosomes to renal allograft mice promoted Treg cell differentiation and immune tolerance (91). The protective effect was shown to be mediated largely by miR-682 in the exosomes, supporting the beneficial effect of miR-682 in AR (91). In allogeneic renal transplantation mice, miR-15b mediated the protective effect of Bortezomib (a highly selective proteasome inhibitor) against AR through inhibiting T follicular helper (Tfh) cell proliferation and differentiation, suggesting a critical role of miR-15b in suppressing Tfh cell activity and related acute immune response (92). For biomarker application, Liang et al. established F344-Lewis rat AR model to demonstrate that plasma miR-155 was increased during AR and correlated with the severity of AR (93). Gielis et al. further showed that miR-155-5p and miR-615-3p in urine sediments, together with chemokine CXCL-9 in urine supernatant, may discriminate kidney transplant rejection from stable graft conditions, suggesting the combined use of urinary molecular biomarkers for the detection of rejection (94). Similarly, another study using logistic regression analysis showed that urinary miR-155-5p could identify patients with potential high risk of rejection at early stage of post-transplantation (95). Alfaro et al. showed that miR-150-5p was decreased in acute rejection kidney transplantation patients, the decrease could be related to activation and proliferation of the immune system’s cells (96). A pilot study showed that in the first 3 months post transplantation with tacrolimus therapy, urinary exosome derived miR-155-5p and miR-223-3p correlated with tacrolimus dose, miR-223-3p with serum creatinine, and miR-223-3p and miR-1228-3p with blood leukocytes, indicating that differentially expressed urinary exosome miRNAs might monitor tacrolimus therapy and renal graft function (97). In contrast to the leading role of ACR in kidney transplantation, AMR characterized by microvascular inflammation (MVI) is less studied. However, a recent study unraveled 6 differently expressed miRNAs between AMR biopsies and non-AMR. Integrative omics profiling of miRNAs and mRNAs and single-cell RNA sequencing revealed new pathways involved in MVI and AMR in different cell types, including endothelial cells, epithelial cells and immune cells (98). Moreover, as a histological hallmark of AMR, transplant glomerulopathy is associated with down-regulation of miRNAs like miR-1224-5p, -4508, -320, -378a. The downregulation of these miRNAs results in upregulation of their target genes and activation of T helper cells, dendritic cell maturation and Th1/Th2 pathways (99).

For LncRNAs, Dai and colleagues identified 5 transcriptional factors that were associated with 12 miRNAs and 32 lncRNAs in biopsies of 3 patients with AR (108). Their subsequent microarray analysis of lncRNA in AR biopsies revealed 5 differentially expressed lncRNAs (109). However, these studies are limited by small sample size and absence of assessing the predictive value of these lncRNAs. Qiu et al. (110) detected a higher level of lncRNA-ATB (lncRNA activated by transforming growth factor β) in ACR biopsies. LncRNA-ATB inversely correlated with miR-200c and acted as a sponge for miR-200c. In biological fluid, Lorenzen et al. conducted genome-wide analysis of RNA from urine of patients with ACR. They figured out a novel lncRNA named RP11-354P17.15-001 that significantly increased in ACR patients compared with healthy controls and normalized to control level after successful anti-rejection therapy. Moreover, RP11-354P17.15-001 was associated with a more severe decline in glomerular filtration rate at 1 year post-transplantation and it allowed detection of subclinical rejection patients that would have been missed by routine serum creatinine measurement (111). Genome-wide lncRNA analysis in circulating peripheral blood (PB) uncovered 23 deregulated lncRNAs which could discriminate AR from those without AR. Among them, AF264622 and AB209021 showed the best diagnostic capabilities (112). Nafar et al. showed that lncRNA FAS-AS1 was higher in rejection patients when compared with non-rejection ones in males but not in females, suggesting its putative role in the pathogenesis of renal transplant rejection (113). In addition, circulating plasma LNC-EPHA6 (a vascular injury related lncRNA) showed higher expression in ACR than that in stable patients (114). For extracellular vesicular lncRNAs, Wang et al. showed that bone marrow mesenchymal stem cell (BMSC)-derived small extracellular vesicles (sEVs) mitigated inflammation and renal dysfunction in a Sprague-Dawley (SD) to Wistar rat renal transplantation model, and the effect was mainly mediated by LncRNA Loc108349490 (115). Similarly, in a mouse renal allograft model, BMSC-derived exosomes (BMSC-Ex) implantation promoted Treg cell quantity and improved immune tolerance, which was mediated by lncRNA DANCR (differentiation antagonizing non-protein coding RNA) (116). CircRNAs exist abundantly in eukaryotic transcriptome and function mainly by sponging and sequestering miRNAs (86). A recent study analyzed the global circRNA expression pattern in urine of patient with ACR and their control transplants. It was found that hsa_circ_0001334 upregulation not only correlated with acute kidney rejection, but predicted loss of kidney function 1 year after transplantation (117).

To identify the molecular basis underlying operational tolerance (OT), Cabral et al. used low density array to reveal 3 higher level (miR-885-5p, -331-3p, -27a-5p) and 4 lower level of miRNAs (miR-1233-3p, -572, -638, -1260a) in serum of OT patients compared with that of chronic rejection patients (100). Gene set enrichment analysis indicated that these miRNAs target cell death regulation/survival signaling and transplantation tolerance (100). Xiong et al. reported that miR378 was decreased in renal allograft with IF/TA. Functionally, miR378 reduced IRI injury, inflammatory cell infiltration, and subsequent fibrosis in renal allografts (101). Moreover, miR-21, miR-210 and miR-4639 in plasma exosomes were significantly higher in CAD, and these 3 circulating exosomal miRNAs in combination may predict post-transplant renal graft dysfunction (102). It is worth noting that miR-21 upregulation in urine not only predicts DGF in the early phase post-transplantation as mentioned above (88), its overexpression in urine (103) and plasma exosomes (104) was associated with moderate to severe IF/TA and worse allograft function in the long term.

Very little is known about lncRNA in IF/TA following kidney transplantation. Nonetheless, 6 lncRNAs candidates (AC126763.1, RP11-280K24.1, LINC01137, WASIR2, RP1-276N6.2 and AD000684.2) were suggested to predict the development and progression of CAD using the GEO datasets of 407 biopsies from 3 different studies (118).