An estimated 2 million people worldwide are affected by end-stage renal disease (ESRD), and the number of patients newly diagnosed with ESRD continues to increase at a rate of 5% to 7% per year. Approximately 11% of ESRD patients receive peritoneal dialysis (PD) as renal replacement therapy, with an annual growth rate of 8% [1]. PD offers therapeutic advantages for appropriate patients, including early survival benefits, the convenience of home therapy, and lower healthcare costs in general [2,3].

PD uses the patient’s own peritoneal membrane instead of a disposable dialysis membrane in hemodialysis, across which waste products are exchanged from the circulation and flushed out through a peritoneal catheter. It is of paramount importance to maintain a healthy peritoneum. However, cumulative evidence based on basic and clinical studies has highlighted that continuous exposure to bio-incompatible compositions of peritoneal dialysate, the presence of catheter per se, recurrent episodes of peritonitis, and other factors lead to functional and structural damage of the peritoneal structure over time [4]. Peritoneal biopsy studies have demonstrated morphologic changes in long-term PD patients, which included loss of peritoneal mesothelial cells (MCs), accumulation of extracellular matrices (ECM), increased angiogenesis, and fibrosis [5,6]. Peritoneal fibrosis with extensive submesothelial thickening leads to a functional loss of the peritoneum as a semi-permeable dialysis membrane, ultimately resulting in discontinuation of PD.

Therefore, one of the most important challenges in the PD community is preservation of peritoneal membrane integrity. To improve the outcome of PD therapy and favorably expand its clinical application, there is a compelling need to understand the early mechanisms that trigger peritoneal membrane damage during PD. Previous studies have suggested the phenotype transition of peritoneal MCs before evident morphologic change becomes recognizable under light microscopy [7-9]. During the process of phenotype transition, there is disruption of the tight junction and acquisition of migratory and invasive phenotypes in MCs, followed by production of ECM [10-12]. The phenotype transition of peritoneal MCs, such as epithelial-to-mesenchymal transition (EMT), is regarded as one of the earliest phenomena of peritoneal dysfunction in PD [7]. The molecular mechanism of EMT and its regulation are of interest as potential therapeutic targets to ameliorate peritoneal fibrosis [12,13].

In this review, the evidence of EMT in peritoneal fibrosis and its mechanisms will be discussed with the proposal of therapeutic modalities targeting peritoneal EMT.

Peritoneal MC cells are epithelial-like cells resting on a thin basement membrane that lines the entire abdominal cavity [14]. The MC monolayer (“mesothelium”) is recognized as the first-line barrier that provides a protective, non-adhesive surface on the abdominal cavity and organs. Peritoneal MCs predominantly adopt a polygonal cobblestone-like morphology and are supported by submesothelial connective tissue containing blood vessels, lymphatics, and resident fibroblast-like cells [15]. The boundaries between MCs consist of delicate junctional complexes, including tight junctions, adherens junctions, gap junctions, and desmosomes, which are crucial for maintenance of a tight semi-permeable diffusion barrier [14]. In addition, MCs are not just a simple barrier, but play a critical role in regulating peritoneal homeostasis by secreting diverse mediators governing immune surveillance, tissue repair, angiogenesis, inflammatory responses, and control of fluid and solute transport [15-17]. Therefore, functional and structural alterations in MCs may trigger signals that lead to irreversible peritoneal damage.

EMT is a biological process that allows an epithelial cell to acquire a mesenchymal cell phenotype through multiple biological processes resulting in enhanced migratory capacity, invasiveness, self-renewal properties, and increased production of ECM [18-20]. EMT is a physiologic process of organ development and wound healing through a finely programmed process. However, it can be pathologically continued in response to internal or external stimuli such as inflammation, eventually leading to organ fibrosis. EMT was once known as an irreversible process; however, several studies have revealed the reversible potential of peritoneal EMT [21]. During EMT, the earliest event involves loosening of cell-to-cell contacts between neighboring MCs [18], which is associated with downregulation of epithelial adherens junctions (like E-cadherin) and tight junction components (like zonula occludens-1, ZO-1). Tight junction proteins such as claudins and occludin are para-cellular components regulating cellular transport of peritoneal mesothelium. The expression and/or intracellular localization of tight junction proteins of MCs are altered by peritoneal dialysates in ESRD patients [22], which are mediated by induction of oxidative stress [23]. The alteration of adhesion and tight junction proteins can reflect an impairment of mesothelial integrity as a peritoneal barrier. Importantly, decrease of E-cadherin expression per se in MCs is known to upregulate the expression of mesenchymal junctional proteins (such as N-cadherin) and switch main cytoskeletal protein patterns of MCs from cytokeratin to vimentin with de novo synthesis of fibroblast-specific protein 1 and smooth muscle actin (α-SMA) [18-20,24]. It has been well demonstrated that EMT or mesothelial-to-mesenchymal transition of peritoneal MCs leads to accumulation of myofibroblasts that are distinct from peritoneal residential fibroblasts in phenotype. Loosening of intercellular tight adhesions also creates an environment that makes MCs and submesothelial tissue more vulnerable to the composition of bio-incompatible dialysates, reactive oxygen species (ROS), and inflammatory cytokines in PD patients (Fig. 1).

Despite recent controversy on the source of myofibroblasts in organ fibrosis, current evidence in animal PD models and patients supports the role of EMT as a key mechanism of the production of myofibroblast. In the seminal study by Yáñez-Mó et al [7], EMT was observed in MCs isolated from the effluent of PD patients even at the early time point of PD initiation, indicated by a progressive loss of epithelial morphology and a decrease in the expression of E-cadherin and cytokeratin. There are many direct/indirect findings supporting the presence of peritoneal EMT. Cultured MCs undergo transdifferentiation to mesenchymal cells upon exposure to various stimuli including high glucose, glucose degradation products (GDP), proinflammatory cytokines, and peritoneal dialysates. Furthermore, immunohistochemical analysis revealed the coexistence of epithelial and mesenchymal markers within MCs during peritoneal fibrosis [25-27]. Intraperitoneal transfection of transforming growth factor (TGF)-β1 induced EMT and peritoneal fibrosis in rats, which was associated with disruption of an intact mesothelial layer and appearance of cytokeratin+/α-SMA+ spindle-shaped cells in the submesothelial compact zone [28]. Consistent with this animal model, a clinical study with 35 PD patients demonstrated in situ evidence of EMT (submesothelial cytokeratin staining) in parietal peritoneal tissue of 17% of patients and loss of the mesothelial layer in 74% of patients, indicating that conversion of epithelial-like MCs to fibroblast-like phenotypes was frequent in the peritoneal membrane during PD therapy [29]. A number of subsequent studies demonstrated similar changes in the peritoneal membrane [30-33], suggesting local conversion of MCs to myofibroblasts after PD initiation.

Major mechanisms underlying peritoneal EMT include induction of oxidative stress, pro-inflammatory cytokines, endoplasmic reticulum (ER) stress and activation of the local renin-angiotensin system. Based on experimental data using TGF-β1, the most potent pro-fibrotic cytokine known to induce EMT, the balance among Smad2, Smad3, and Smad7 plays a role in the development of EMT of MCs [34]. Smad proteins, phosphorylated by TGF-β1 stimulation, control transcription of TGF-β1 responsive genes [34,35], many of which are identified as pro-fibrogenic genes and regulators of EMT in peritoneal fibrosis, including Snail [36,37], fibronectin [38,39], connective tissue growth factor [40], β-catenin [41], monocyte chemoattractant protein-1 [42], and matrix metalloproteinase-2 [43,44]. Therefore, TGF-β1-induced mesothelial transition leads to a characteristic myofibroblastic phenotype and complex modulation of gene expression, including cytoskeletal organization, cell adhesion, and ECM production [8,45]. There is also emerging evidence showing the role of non-Smad pathways in peritoneal EMT, which includes extracellular signal-regulated kinase (ERK)/nuclear factor kappa B [46], mitogen-activated protein kinase (MAPK) [21], Akt/mTOR [9], c-Jun N-terminal kinase [47,48], and Wnt/β-catenin [49]. TGF-β1 was also reported to induce ER stress [50], activate NLRP3 inflammasomes [51], and inhibit the expression of 5¢-adenosine monophosphate-activated protein kinase [27] in peritoneal MCs.

Among these complicated networks of molecular pathway, induction of oxidative stress is an early key mechanism of peritoneal EMT [52-54]. A major source of ROS in MCs is NADPH oxidase (NOX), mitochondrial electron transport, xanthine oxidase, and nitric oxide synthase. TGFβ1 increased NOX activity with induction of both membranous NOX- and mitochondria-mediated ROS production with as little as 15 minutes of TGFβ1 stimulation [48]. MCs possess the unique ROS production machinery of NOX1 as the most abundant isoform and mitochondrial NOX4, which are differentially regulated in transcription, translation, and translocation of p47phox into the cell membranes and enhancement of p47phox-p22phox binding [51]. Peritoneal dialysate or high glucose is reported to induce an alteration in the expression of tight junction proteins such as claudin-1, claudin-2, and occludin with increased permeability through induction of oxidative stress [55].

Another important mechanism of peritoneal EMT involves mitigation of intrinsic defense mechanisms such as the peritoneal antioxidant system and anti-fibrotic peptide production in the peritoneal cavity. In the animal model of PD, increased oxidative stress through alleviation of mesothelial antioxidant production is expressed as a decrease in the ratio of reduced to oxidized glutathione (GSH/GSSG) and superoxide dismutase (SOD) activity and an increase of 8-hydroxy deoxyguanosine in the peritoneal dialysate. Particularly, SOD activity in peritoneal effluent was shown to be almost absent in rats on PD [27]. In addition, peritoneal MCs are known to constitutively synthesize hepatocyte growth factor (HGF) and bone morphogenic peptide-7 (BMP-7), which are well-known anti-fibrotic peptides. High glucose or TGFβ1 decreases the production of HGF and BMP-7, suggesting weakening of a natural anti-fibrotic mechanism of peritoneum by the process of PD [25].

The use of less bio-incompatible dialysates such as low-GDP or bicarbonate-based dialysate was reported to protect the peritoneum from EMT with better preservation of peritoneal morphology [31,38]. An alternative approach to maintaining the peritoneal MC barrier and preserving peritoneal integrity is therapy targeting EMT. Although the presence of peritoneal EMT and its mechanism have been extensively studied, there are not many studies showing the beneficial effect of therapy targeting peritoneal EMT per se. Table 1 shows the previous studies regarding EMT-targeting interventions using in vitro and in vivo experimental models [21,25-27,51,56-64].

The PPAR-γ agonist rosiglitazone was reported to preserve the MC monolayer with reduction of peritoneal fibrosis and angiogenesis in mice models of PD [56]. Anti-oxidants, including N-acetyl cysteine (NAC, ROS scavenger), apocynin (NOX inhibitor), and mitoQ (an inhibitor of mitochondrial electron transfer chain subunit I), were shown to alleviate TGF-β1-induced EMT of MCs [51]. Taurine-conjugated ursodeoxycholic acid, an ER stress blocker, also ameliorated TGF-β1-induced EMT in peritoneal MCs [50]. Interestingly, pre-treatment with tunicamycin or thapsigargin for 4 hours (ER stress pre-conditioning) also protected MCs from TGF-β1-induced EMT, demonstrating the role of ER stress as an adaptive response to protect MCs from EMT. Given the consideration that peritoneal MCs isolated from PD patients displayed an increase in the marker of ER stress GRP78/94 and was correlated with degree of EMT [50], modulation of ER stress in MCs could serve as a novel approach to ameliorate peritoneal damage in PD patients. Previous studies have also demonstrated how glucocorticoids and mineralocorticoids play a role in peritoneal EMT [21,26]. Dexamethasone inhibited TGF-β1-induced EMT and RU486, a glucocorticoid receptor antagonist, blocking the effect of dexamethasone on TGF-β1-induced EMT. The beneficial effect of dexamethasone on TGF-β1-induced EMT was mediated through amelioration of ERK and p38 MAPK phosphorylation. Dexamethasone also inhibited glycogen synthase kinase-3β phosphorylation and Snail upregulation induced by TGF-β1, which were ameliorated by inhibitors of MAPK. A recent study showed the beneficial effect of paricalcitol on the EMT of peritoneal MCs by ameliorating oxidative stress and inflammatory reactions [51]. Paricalcitol alleviated NOX activity by blocking the interaction of p47phox with p22 and inhibiting mitochondrial NOX4 mRNA expression and activity, resulting in an amelioration of TGFβ1-induced oxidative stress and activation of NLRP3 inflammasomes. Interestingly, blocking the expression and activity of NLRP3 inflammasomes in MCs also blocked the EMT of MCs [48]. In addition to the effort to alleviate peritoneal injury by extrinsic stress, enhancement of intrinsic anti-fibrotic defense in the peritoneal cavity has also been studied. HGF provided the dosage-dependent prevention of EMT and peritoneal fibrosis in animal PD models [25]. Both BMP-7 peptides and gene transfection with an adenoviral vector of BMP-7 also protected MCs from EMT. Furthermore, adenoviral BMP-7 transfection decreased peritoneal EMT and ameliorated peritoneal thickening in animal models of PD.

There are other pharmacologic agents such as cyclooxygenase-2 inhibitors [65,66], renin-angiotensin blockades [67], and heparin [68] showing beneficial effects on peritoneal fibrosis and ultrafiltration in animal models of PD.

Although several studies on the effect of intraperitoneal or oral pharmacologic intervention on peritoneal function have been conducted in PD patients [13], no prospective clinical studies have verified the benefit of clinical application of these agents.

Despite evidence of the presence of EMT in organ fibrosis including peritoneal fibrosis, there has been a major challenge in understanding EMT as a prerequisite for organ fibrosis. Since Iwano et al [69] first hypothesized EMT as the mechanism of renal fibrosis by the lineage-tracing strategy using transgenic reporter mice and bone marrow transplants, abundant studies providing experimental and clinical evidence for EMT as a key player of organ fibrosis have been published [70-76]. However, the role of EMT in renal and hepatic fibrosis has been recently challenged based on compelling findings in new lineage-tracing studies. Kriz et al [77] summarized these findings in their review and pointed out several “flaws” in previous studies of EMT, with a clear point of view that “unequivocal evidence supporting EMT as an in vivo process in renal fibrosis is lacking”. Similar questions have also been raised about the role of EMT in peritoneal fibrosis [78]. Therefore, it may be worth investigating other potential sources of peritoneal myofibroblasts [78]. Nonetheless, it is certain that peritoneal MCs undergo the process of phenotype transitions whether or not they take the appearance of myofibroblasts expressing both epithelial and mesenchymal cells.

Given the consideration of the reversible nature of EMT, it seems to be an attractive target for preventing or even reversing peritoneal damage. Although there are debates about whether EMT is indispensable for appearance of myofibroblast and peritoneal fibrosis, the phenotype transition of MCs clearly produces pro-inflammatory and pro-fibrotic microenvironments with induction of oxidative and ER stress in the abdominal cavity. Loosening of intercellular tight adhesions, which precedes loss of MCs and production of ECM, seems to be one of the targets to be reversed if it can be detected as early as possible. In addition, further studies regarding intervention targeting of EMT will be necessary through stimulation of active re-differentiation of MCs in the middle of EMT processes into healthy MCs (mesenchymal-to-epithelial transition) as well as inhibition of EMT.