Epithelial to Mesenchymal Transition and Peritoneal Membrane Failure in Peritoneal Dialysis Patients: Pathologic Significance and Potential Therapeutic Interventions
Luiz S. Aroeira*,
Abelardo Aguilera*,
José A. Sánchez-Tomero,
M. Auxiliadora Bajo,
Gloria del Peso,
José A. Jiménez-Heffernan,
Rafael Selgas and
Manuel López-Cabrera*
* Unidad de Biología Molecular and Servicio de Nefrología, Hospital Universitario de la Princesa, and Servicio de Nefrología, Hospital Universitario La Paz, Madrid, and Departamento de Patología, Hospital Universitario de Guadalajara, Guadalajara, Spain
Correspondence: Dr. Manuel López-Cabrera, Unidad de Biología Molecular, Hospital Universitario de la Princesa, C/ Diego de León 62. 28006-Madrid, Spain. Phone: +34-91-5202334; Fax: +34-91-5202374; E-mail: mlopez.hlpr{at}salud.madrid.org
Peritoneal dialysis (PD) is a form of renal replacement andis based on the use of the peritoneum as a semipermeable membraneacross which ultrafiltration and diffusion take place. Nevertheless,continuous exposure to bioincompatible PD solutions and episodesof peritonitis or hemoperitoneum cause acute and chronic inflammationand injury to the peritoneal membrane, which progressively undergoesfibrosis and angiogenesis and, ultimately, ultrafiltration failure.The pathophysiologic mechanisms that are involved in peritonealfunctional impairment have remained elusive. Resident fibroblastsand infiltrating inflammatory cells have been considered themain entities that are responsible for structural and functionalalterations of the peritoneum. Recent findings, however, demonstratedthat new fibroblastic cells may arise from local conversionof mesothelial cells (MC) by epithelial-to-mesenchymal transition(EMT) during the inflammatory and repair responses that areinduced by PD and pointed to MC as protagonists of peritonealmembrane deterioration. Submesothelial myofibroblasts, whichparticipate in inflammatory responses, extracellular matrixaccumulation, and angiogenesis, can originate from activatedresident fibroblasts and from MC through EMT. This heterogeneousorigin of myofibroblasts reveals new pathogenic mechanisms andoffers novel therapeutic possibilities. This article providesa comprehensive review of recent advances on understanding themechanisms that are implicated in peritoneal structural alterations,which have allowed the identification of the EMT of MC as apotential therapeutic target of membrane failure.
In the past decades, peritoneal dialysis (PD) has become anestablished alternative to hemodialysis for the treatment ofESRD, and the number of patients who are in PD programs hasincreased progressively worldwide, especially in some Asiancountries. One of the most important challenges in PD is thelong-term preservation of the peritoneal membrane integrity.1,2Damage to the peritoneum is a serious event in PD, because itmay jeopardize the organ on which the whole treatment is based.1–3The morphology of the peritoneum is simple; A single layer ofmesothelial cells (MC) covers a submesothelial region that iscomposed of connective tissue with few fibroblasts, mast cells,macrophages, and vessels.4 The nonphysiologic nature of dialysisfluids and the uremic status are considered the main etiologicfactors that lead to the functional decline of the peritonealmembrane.1 They induce a sustained situation of peritoneal chronicinflammation that can be exacerbated periodically by recurrentor acute episodes of peritonitis or hemoperitoneum. Closelylinked to the inflammatory response is the reparative process.Its activation is responsible for many of the structural abnormalitiesof the peritoneal membrane, including loss of MC monolayer,submesothelial fibrosis, angiogenesis, and hyalinizing vasculopathy.5–8Such alterations are considered the major cause of ultrafiltrationfailure and loss of the dialytic capacity of the peritoneum.
There are two different pathologic forms of PD-related fibrosis.The most common is simple peritoneal sclerosis, which appearsin almost all patients. The degree of fibrosis is usually mildand shows a relation with time on dialysis.4 On the other endof the spectrum is encapsulating peritoneal sclerosis, whichis an uncommon form of sclerosis that evolves rapidly with intensefibrosis, inflammation, and fibrin deposits. It is a life-threateningcondition that in many cases evolves to visceral encapsulationand progresses even if the patient is removed from PD.1–3It is well established, however, that fibrosis is not the uniquestructural alteration of the peritoneal membrane induced byPD. Besides this alteration, the peritoneum shows an increaseof capillary number (angiogenesis) and vasculopathy.5 Thereis increasing evidence that fibrosis, angiogenesis, and probablyaugmented vessel permeability are key determinants of ultrafiltrationdysfunction.1,9,10 The relationship between peritoneal fibrosisand angiogenesis has not been clearly defined. In animal modelsof PD, it has been shown that fibrosis and angiogenesis maybe two separate responses to peritoneal injury.8 However, inPD patients, it is likely that fibrosis and angiogenesis areintimately and closely related in the response of the peritoneumto continuous injury.8
Recent works have begun to identify the mechanisms that areinvolved in the pathogenesis of peritoneal membrane failureduring long-term PD. Resident stromal fibroblasts and inflammatorycells have been classically considered the main cells responsiblefor structural and functional peritoneal alterations, whereasMC have been considered mere victims of peritoneal injury. However,more recently, it has been shown that MC also play an activerole in peritoneal membrane alteration. It has been demonstratedthat, soon after PD is initiated, peritoneal MC show a progressiveloss of epithelial phenotype and acquire myofibroblast-likecharacteristics by an epithelial-mesenchymal transition (EMT).11MC that have undergone an EMT acquire higher migratory and invasivecapacities, which allow these cells to invade the submesothelialstroma, where they contribute to peritoneal fibrosis and angiogenesisand ultimately lead to peritoneal membrane failure.11–13The myofibroblastic conversion of MC has been confirmed in anin vivo animal model based in the injection of an adenovirusvector that transferred active TGF-1 in rat peritoneum.14 Inthese studies, EMT appears as the central point in the earlypathogenesis of peritoneal damage associated with PD. Giventhat there is no definitive treatment for the progressive lossof the dialytic capacity of the peritoneum, the identificationof the EMT of MC as a key process in the onset and progressionof peritoneal fibrosis and angiogenesis opens new insights fortherapeutic intervention.
PD for ESRD treatment has been used for almost 30 yr.3 The twofirst decades of the modern era of PD were dedicated to catheterdesign, technical aspects, and peritonitis prevention. The thirddecade was devoted to improving the biocompatibility of thesolutions with the expectancy of diminishing their adverse effectson peritoneal morphology and function.1 PD is a successful alternativeto hemodialysis, and several studies have confirmed equivalentadequacy, mortality, and fluid balance status with both modalities,at least for the first 4 to 5 yr.15 However, the growth of PDcontinues to be limited by the membrane incapacity to performadequate diffusive and/or convective transports in the longterm.1,2 Peritonitis and ultrafiltration failure, with a clinicalresult of extracellular volume overload and an increased cardiovascularrisk, are the major factors that contribute to technique dropouts.1,3,9
Our overview of the morpho-functional relationship in peritonealmembrane failure has changed as a result of some recent experimentaldata. Some studies of peritoneal biopsies have suggested thathyalinizing vasculopathy and angiogenesis are the most characteristicfeatures in PD-related peritoneal pathology, at least in patientswith severe membrane failure.5,6 In contrast, a more recentreport by Sherif et al.10 showed that in stable, uncomplicatedPD patients, vascular density does not increase, whereas intactvessels decrease with time of treatment and severe vasculopathypredominate mostly in long-term PD. Other studies have demonstratedthat before any other lesions are evident, human peritonealbiopsies show an EMT of the MC and the consequent environmentalchange, mainly represented by the accumulation of collagen Iand fibronectin.12 In fact, the submesothelial thickening isthe constant change found in peritoneal biopsies after a timeon PD.5–7,12 In animal models that seem to reproduce thesequence of phenomena that occur in PD patients, Margetts etal.16 demonstrated that the transduction of TGF- into rat peritoneumcauses peritoneal sclerosis and angiogenesis, and both processesare followed by the EMT of MC.14
It is worthy of remark that our understanding of peritonealfunctional outcome in response to PD requires a deep knowledgeof the starting point because of initial functional diversityof the human peritoneum. It can be hypothesized that the diversityin solute and water transports at PD initiation may representdifferent tissue responses to similar components of dialysates.Furthermore, different starting points followed by diverse reactionsto PD at midterm can generate a wide spectrum of peritonealmorphofunctional scenarios.17 After molecular and cellular biologystudies of the intimate processes that take place into the peritoneum,a refinement in therapeutic strategies is expected. Therefore,it can be proposed that the fourth decade of PD should be conductedto explore molecular aspects of the peritoneal response.
PERITONEAL INFLAMMATORY RESPONSE TO INJURY AND INFECTION
Exposure to bioincompatible fluids and episodes of bacterialand fungal infection or hemoperitoneum induce situations ofacute and chronic inflammation that cause damage to the peritonealtissue. The successful repair of injured tissue requires a tightlycontrolled response to limit the structural alteration. Theperitoneal immune response to injury or infection involves,among other cells, MC and resident macrophages that work ina coordinated manner to recruit other inflammatory cells, includingmononuclear phagocytes, lymphocytes, and neutrophils.18 MC andinfiltrating immune cells can produce a wide number of cytokines,growth factors, and chemokines to establish a complex networkthat feeds back, resulting in acute or chronic inflammation,which leads to membrane deterioration.18–21 Many of theseinflammatory mediators, such as TNF-, IL-1, IL-8, TGF-, andfibroblast growth factor-2 (FGF-2), may have a role in peritonealfibrosis by stimulating resident fibroblasts proliferation andextracellular matrix (ECM) component deposition22 and by inducingEMT of MC,11 which further increases the number of peritonealfibroblasts.12 In addition, IL-8, FGF-2, and especially vascularendothelial growth factor (VEGF) may induce an increase of peritonealcapillary number and probably vessel permeability,13,19 causingan increase of small-solute transport, which characterize ultrafiltrationfailure.
Even in the absence of infection or hemoperitoneum, it is thoughtthat PD patients have a low-grade chronic inflammation thatpromotes a progressive structural alteration of the peritonealmembrane. This could be due to the mechanical injury duringPD fluid exchanges and also to nonphysiologic composition ofthe fluids. Prolonged exposure of the peritoneum to glucose,the most common osmotic agent in PD fluids, or to Amadori adducts,formed by the condensation between glucose and reactive aminogroups of proteins, induces the synthesis of inflammatory, fibrogenic,and angiogenic factors by macrophages23 and MC.24,25 Glucosemay also promote peritoneal inflammation through a leptin-dependentmechanism.26 In this context, adipocytes that are exposed toglucose produce leptin, which in turn promotes TGF- productionby MC.27 This evidence suggest that peritoneal inflammationis a key triggering process for peritoneal membrane failure;therefore, the knowledge of the mechanisms that are involvedin its regulation may serve in the design of therapeutic interventions.
MAJOR MEDIATORS OF PERITONEAL MEMBRANE DETERIORATION
Chronic and acute inflammations are believed to be key determinantsin the onset and progression of peritoneal membrane structuraland functional alteration. Among the cytokines and growth factorsproduced during peritoneal inflammation, TGF-, a strong profibroticcytokine,28 is considered the master molecule in the genesisof peritoneal fibrosis, because its overexpression has beencorrelated to worse PD outcomes.19 The relevance of TGF- inperitoneal fibrosis has been demonstrated in an in vivo ratmodel, in which TGF- gene was transduced into the peritonealcavity with an adenovirus vector, reproducing the structuraland functional alterations that are observed in PD patients.16Along with TGF-, other proinflammatory and profibrotic cytokinesas well as angiotensin II (AngII) have been shown to be upregulatedduring peritonitis episodes and may contribute in the settingof peritoneal fibrosis.8,29 In regard to peritoneal angiogenesis,it has been shown that local production of the proangiogenicand vasoactive factor VEGF during PD plays a central role inincreased solute transport across the peritoneum and ultrafiltrationfailure.13,30,31
The most important factor of the PD solutions that are responsiblefor peritoneal deterioration seems to be glucose degradationproducts, which, through the formation of advanced glycationend products (AGE), stimulate the production of (ECM) componentsas well as the synthesis of cytokines and growth factors, includingTGF- and VEGF.8 Furthermore, AGE may also induce EMT of MC.32Several studies have demonstrated the appearance of AGE in theperitoneal effluents of PD patients, which correlated with thetime on PD treatment. Biopsy studies have confirmed the accumulationof AGE in the peritoneal tissues of PD patients. The intensityof AGE accumulation is associated with fibrosis and ultrafiltrationdysfunction.8
The uremic status also contributes to the accumulation of AGEand may affect the anatomy of the peritoneal membrane and itstransport characteristics. In this context, the peritoneum ofpartially nephrectomized rats showed increased permeabilityassociated with upregulation of nitric oxide synthase isoforms,VEGF, and FGF-2 and accumulation of AGE.33 Despite these findingsin animal models, the effect of uremia itself on the peritoneumin humans is controversial. Two human peritoneal biopsy studieshave shown a modest compact zone thickening and vasculopathyin predialysis renal patients.5,7 In contrast, in other studies,no significant fibrosis or vasculopathy was observed in uremicnon-PD patients.12
Fibroblasts represent a dynamic population of cells that showfunctional and phenotypic diversity. Among the various fibroblasticphenotypes, the myofibroblast is of the greatest importance.The term "myofibroblast" defines a cell with intermediate featuresbetween a fibroblast and a smooth muscle cell and is characterizedby the expression of -smooth muscle actin (-SMA). Myofibroblastshave been reported as important protagonists of almost all situationsof repair and fibrosis in various pathologies. Their capacityto synthesize ECM, growth factors, cytokines, and participationin the inflammatory response, as well as their contractile properties,convert them in the most important fibroblastic phenotype.34
In normal peritoneum or in the peritoneum of uremic non-PD patients,myofibroblasts are not present, but the resident fibroblastsshow an intense expression of CD34, an antigen characteristicof bone marrow stem cells (Figure 1). The expression of CD34gradually disappears in PD patients during the onset of peritonealfibrosis.12 The significance of the loss of CD34 expressionis unknown but seems to correlate with the appearance of themyofibroblastic phenotype.12 Tissue CD34+ fibroblasts are closelyrelated to circulating CD34+ fibrocytes and reflect a bone marroworigin. Fibrocytes were described as a small subpopulation ofcirculating leukocytes that express collagen I, CD45RO, CD13,CD11b, CD34, CD86, and MHC class II, which transform into myofibroblastswhen exposed to TGF-in vitro.35 However, peritoneal CD34+ fibroblastsdo not express these fibrocyte markers, suggesting that theyare simply residual embryonic mesenchymal cells that remainedin the peritoneal tissue after organogenesis (Figure 1). Incontrast to normal peritoneum, myofibroblasts can be easilydetected in the peritoneal membrane of many patients who undergoPD treatment,6 even at very early stages of PD and precedingthe appearance of fibrosis (Figure 1).12
Figure 1. Subpopulations of peritoneal fibroblasts in peritoneal membranes of normal and peritoneal dialysis (PD) patients. (A) The morphology of normal peritoneum is simple, with a single layer of mesothelial cells (MC) that covers a submesothelial region that is composed of connective tissue with few fibroblasts and some vessels. In normal peritoneum, myofibroblasts are not present, whereas resident fibroblasts show an intense expression of CD34, an antigen that is characteristic of bone marrow stem cells. Peritoneal CD34+ fibroblasts do not express fibrocyte markers, suggesting that they are simply residual embryonic mesenchymal cells that remained in the peritoneal tissue after organogenesis. (B) In PD patients, there are many structural abnormalities of the peritoneal membrane, including loss of MC monolayer, increased number of fibroblasts, submesothelial fibrosis, and augmented vessel number. In contrast to normal peritoneum, myofibroblasts can be easily detected in the peritoneal membrane of many patients who undergo PD. The myofibroblasts may originate from activated resident fibroblasts, from circulating cells (fibrocytes), and from MC through an epithelial-to-mesenchymal transition (EMT). The myofibroblastic conversion of MC can be observed by immunohistochemical analysis, which reveals the presence of fibroblast-like cells embedded in the compact zone expressing mesothelial markers such as cytokeratins, intercellular adhesion molecule-1, and calretinin. Peritoneal myofibroblasts express vascular endothelial growth factor (VEGF), extracellular matrix (ECM), and cyclooxygenase-2 (COX-2), indicating that these cells are implicated in peritoneal structural alterations that are induced by PD. The proportions of peritoneal myofibroblasts that originated from resident fibroblasts, circulating fibrocytes, and MC remain to be established.
There is evidence that myofibroblasts may originate from differentsources. First, they may differentiate from resident fibroblasts.34Second, they may derive from circulating cells (fibrocytes)that are recruited to injured tissues.35 Finally, they can originatefrom nearby epithelial cells through an EMT. The contributionof EMT as an important source of myofibroblasts was initiallydescribed in animal models of renal fibrosis.36 In the peritoneum,the proportions of myofibroblasts that originated from residentfibroblasts, circulating fibrocytes, and MC remain to be established.Similarly, the dynamics in which EMT takes place has to be defined.Most probably, it is not a uniform process, and several factorsmay result in variations of its intensity. Among others, theepisodes of peritonitis must have an influence on EMT.12
Turning an epithelial cell into a mesenchymal cell is a complexand step-wise process that requires alterations in cellulararchitecture and behavior and a profound molecular reprogrammingwith new biochemical instruction.36–38 As illustratedin Figure 2, the EMT starts with the dissociation of intercellularjunctions, as a result of downregulation of adhesion moleculessuch as E-cadherin, claudins, occludins, zona occludens-1, anddesmoplakin, and with the loss of microvilli and apical-basalpolarity. Then, the cells adopt a front-back polarity, as aresult of cytoskeleton reorganization, and acquire -SMA expressionand increased migratory capacity. In the latest stages of EMT,the cells acquire the capacity to degrade the basement membraneand to invade the fibrotic stroma by upregulating the expressionof matrix metalloproteinases (MMP). Other commonly used molecularmarkers for EMT include the downregulation of cytokeratins;upregulation of vimentin, N-cadherin, and transcription factorsnail; and increased production of ECM components.
Figure 2. Key events during EMT. The diagram shows four key steps that are essential for the completion of the entire EMT course and the most commonly used epithelial and mesenchymal markers.
Besides these classic markers, there are some other markersas well as a number of biochemical changes that define the EMT(summarized in Table 1). EMT can be easily engaged by combinationsof a wide spectrum of extracellular stimuli, including cytokines,growth factors, AGE, MMP, and ECM components such as collagenI. It is worthwhile to point out that most of these EMT regulatorshave been identified by using in vitro cell culture system inwhich epithelial cells were treated with purified factors, eitheralone or in combination, at high concentrations. These fractionatedstudies, although necessary to define the potential role ofeach individual factor in regulating EMT, did not mimic thereal situation in vivo, in which the EMT regulators are presentat low concentrations. Therefore, an EMT in vivo may resultfrom an integration of diverse signals triggered by multiplefactors, being difficult to assign priorities or hierarchy.
Table 1. Molecular and functional patterns of EMTa
Receptor-mediated signaling in response to these factors triggersthe activation of a complex network of intracellular effectormolecules, such as Ras/Rho GTPases, Rho-activated kinase, tyrosine-kinaseSrc, integrin-linked kinase (ILK), Wnt-1, Smad 2 and 3, themitogen-activated protein kinases p38 and extracellular–signalregulated kinases, and phosphatidylinositol-3-kinase (Table 1).These effectors orchestrate the dissociation of intercellularadhesion complexes, the changes in cytoskeletal organization,and the acquisition of migratory and invasive capacities thatoccur during EMT.36–38
A central target of some of these signaling pathways is glycogen-synthasekinase-3 (GSK-3), which has been shown to phosphorylate -cateninand the transcriptional repressor Snail, leading to their ubiquitinizationand degradation via the proteasome. The phosphorylation of GSK-3by extracellular–signal regulated kinases, ILK, Wnt-1,or phosphatidylinositol-3-kinase leads to its functional inhibition.As a result, -catenin is stabilized and localizes to the nucleus,where it feeds into the Wnt signaling pathway by interactingwith lymphoid enhancer factor-1/T cell factor. In addition,the inhibition of GSK-3 drives the stabilization and nucleartranslocation of Snail, a potent transcriptional repressor ofE-cadherin and other intercellular adhesion molecules, and inducerof cell growth arrest and survival.39 The regulation of signalingpathways also results in the activation and nuclear translocationof other transcription factors, such as Smads 2 and 3, NF-B,and ligand-bound FGF receptor 1, which in conjunction with -catenin,lymphoid enhancer factor-1/T cell factor, and Snail, repressthe expression of epithelial markers and engage the EMT transcriptome.36–38It is important to point out that EMT is a reversible process,at least in the early stages. Therefore, molecules that negativelyregulate EMT and that promote mesenchymal-to-epithelial transitionmust exist (Table 1). In this context, two endogenous factors,namely hepatocyte growth factor (HGF) and bone morphogeneticprotein-7 (BMP-7), have been demonstrated to block and reverseEMT by inducing the expression of the transcriptional co-repressors,in the case of HGF, or by activating Smad-5, in the case ofBMP-7, which interfere with TGF-–activated Smad-2/3. Furthermore,Smad-7 is another molecule that negatively regulates EMT byblocking Smad-2 phosphorylation and activation.36–38
The presence of EMT in the peritoneum of PD patients was firstdemonstrated in a landmark paper that published in 2003.11 Itwas described that confluent MC cultures from PD effluents mayshow epithelioid and nonepithelioid (fibroblast-like) morphologies.The frequency of nonepithelioid MC was associated with timeon PD and with episodes of peritonitis or hemoperitoneum. MCfrom effluents showed high expression of intercellular adhesionmolecule-1 independent of their morphology. In contrast, thismesothelial marker was negligible on cultured peritoneal fibroblasts,supporting that effluent nonepithelioid cells had a mesothelialorigin and were not fibroblast contaminations. The analysisof the epithelial markers cytokeratins and E-cadherin was alsoimportant to determine more precisely the nature of effluent-derivedcells. High expression of cytokeratins and E-cadherin was observedonly in omentum-derived MC, whereas effluent-derived cells showeda progressive reduction in the expression of these molecules,although even fibroblast-like MC maintained a small populationof positive cells. Fibroblasts were completely negative forthese two markers. The phenotype changes in effluent MC wereindicative of an EMT. Additional evidence that the PD-inducedchanges of the MC were due to an EMT came from the analysisof the expression of the transcription factor Snail, a potentinducer of EMT. Omentum-derived MC did not express Snail mRNA,whereas a progressive expression of this mRNA was observed ineffluent MC preparations along the transdifferentiation process.11Besides these molecules, there are some other markers that mayhelp to distinguish effluent MC from contaminating fibroblastsas well as to establish the various stages of transdifferentiationof MC (summarized in Table 2).
Table 2. Molecular markers of mesothelial cells and fibroblastsa
The mesenchymal conversion of MC could also be observed in vivoin the peritoneum as a response to PD. Immunohistochemical analysisof peritoneal biopsies from PD patients revealed the presenceof fibroblast-like cells embedded in the compact zone expressingmesothelial markers. In addition, these peritoneal biopsiesshowed expression of -SMA in the fibrotic stroma, especiallyin the upper submesothelial level, and in many cases, thesemyofibroblasts showed coexpression of cytokeratins (Figure 1).12These results indicated that new myofibroblastic cells may arisefrom local conversion of MC by EMT during the repair responsesthat take place in PD. Recently, the myofibroblastic conversionof MC was confirmed in vivo by injection of an adenovirus vectorthat transferred active TGF-1 in rat peritoneum.14
PATHOLOGIC SIGNIFICANCE OF EMT IN PERITONEAL FIBROSIS AND ANGIOGENESIS
It has been shown that during the progression of EMT, MC acquirehigher ex vivo and in vitro ability to synthesize componentsof the matrix, such as fibronectin and collagen I.13 In addition,MC that undergo an EMT express high levels of cyclooxygenase-2(unpublished data, Aroeira et al.), which has been implicatedin tissue remodeling and fibrosis processes.40 Furthermore,it can be observed that the grade of peritoneal fibrosis andsmall-solute transport rate of PD patients correlate with theexpression of mesothelial markers within the fibrotic stroma(unpublished data, del Peso et al.). These results support thenotion that transdifferentiated MC play an essential role inthe initiation of fibrosis and subsequent peritoneal functionaldecline.
Besides fibrosis, an increased number of capillaries is relatedto peritoneal membrane failure.1,8 It was proposed that localproduction of VEGF during PD plays a central role in the processesthat lead to peritoneal angiogenesis and functional decline.30,31However, the main source of VEGF in the PD patients as wellas the mechanisms that are implicated in VEGF upregulation duringPD remained elusive. A number of studies have shown that MCfrom omentum have the capacity to produce VEGF in vitro in responseto a variety of stimuli.13,41 A recent study demonstrated thattransdifferentiated MC are an important source of VEGF in PDpatients and that the underlying mechanism of VEGF upregulationin MC is the mesenchymal conversion of these cells.13 It wasshown that MC from effluents with fibroblast-like phenotypeproduce much more VEGF ex vivo than MC with epithelial phenotypeand that patients who drain fibroblast-like cells have higherblood VEGF levels than patients with MC with epithelial phenotypein their effluents. Moreover, a correlation between ex vivoand in vivo VEGF levels and the rate of peritoneal transportin patients who were on PD could be demonstrated.13 These resultssuggest a direct and active role of MC not only in fibrosisbut also in peritoneal angiogenesis. In addition, these studiesindicated that the EMT of MC reflects peritoneal functionaldecline better than previous tests, such as measurement of effluentcancer antigen 125 (CA125).13 Although CA125 has been classicallyused as an index of MC mass and peritoneal health in PD patients,the value of effluent CA125 has been questioned recently becauseit correlates with peritoneal transport rate only in short-termbut not in long-term PD.42
The insights of the molecular mechanisms that regulate chronicinflammation that is induced by peritoneal injury will allowthe identification of potential therapeutic targets. However,from the clinical nephrologist's point of view, perhaps themost important aspect is the identification of the EMT of MCas a key event in peritoneal membrane failure, because thisprocess can be manipulated with a wide range of agents and pharmaceuticalproducts.37 The therapeutic strategies may be designed eitherto prevent or to reverse the EMT itself or to treat its effectssuch as cellular invasion, ECM accumulation, or angiogenesisfactor synthesis.
It can be proposed that at least six steps that are relatedto the EMT process of the MC can be clinically managed, aloneor in combination, to prevent peritoneal membrane failure (Figure 3).Because EMT is reversible, treatments may be tailored eitherto prevent (step 1) or to reverse (step 2) this process. Thesetherapeutic approaches have been proved to be effective in animalmodels of renal fibrosis. The endogenous factors HGF and BMP-7have been demonstrated to block EMT both in vitro and in vivoand to prevent renal interstitial fibrosis.37 In addition, thesetwo proteins reversed the phenotypic conversion of tubular epithelialcells that was induced by TGF- by restoring E-cadherin expressionand downregulating the expression of mesenchymal markers suchas -SMA, vimentin, and fibronectin.37 The reversion from mesenchymalto epithelial phenotype by BMP-7 has also been demonstratedfor transdifferentiated MC.43 Mechanistically, HGF interfereswith TGF-–mediated EMT by inducing the expression of thetranscriptional co-repressors such as SnoN and TG interactingfactor that interact with activated Smad-2/4 complex and blockthe expression of Smad-dependent genes, including ILK.37 Theunderlying mechanism of BMP-7 blockade of EMT is by activationof Smad-5 protein that counteracts with TGF-–activatedSmad-2/3.36
Figure 3. Therapeutic strategies for peritoneal membrane failure based on EMT of MC. EMT of MC in vivo results from integrated signals that are induced by multiple stimuli. These include elevated glucose and glucose degradation products (GDP) and concentration of PD fluids, which through the formation of advanced glycation-end products (AGE) stimulate the transdifferentiation of MC. The formation of AGE may also be due to the uremic status of the PD patients. The low pH of the dialysates and the mechanical injury during PD fluid exchanges may cause tissue irritation and contribute to chronic inflammation of the peritoneum, which promote EMT of MC. Episodes of bacterial or fungal infections or hemoperitoneum cause acute inflammation and upregulation of cytokines and growth factors such as TGF-, IL-1, fibroblast growth factor-2 (FGF-2), TNF-, and angiotensin II (AngII), among others, which are strong inducers of EMT. The therapeutic strategies may be designed either to prevent or to reverse the EMT itself or to treat its effects such as cellular invasion, fibrosis, or angiogenesis. The diagram illustrates six steps related to the EMT process of the MC that can be clinically managed, alone or in combination, to prevent peritoneal membrane failure. See text for details.
Besides targeting EMT by supplementation of endogenous factors,this process can be disrupted by using small molecule inhibitors.For instance, the inhibition of Rho-activated kinase, a downstreameffector kinase of RhoA, resulted in suppression of -SMA expressionand renal interstitial fibrosis in a mouse model of ureteralobstruction.37 Pharmacologic inhibition of AngII, a potent EMTpromoter, also attenuated EMT and reduced renal fibrosis.37More recently, it was shown that paricalcitol, a vitamin D analogue,blocked the mesenchymal conversion of tubular epithelial cellsand ameliorated renal fibrosis in animal models of obstructivenephropathy.44
Whereas these endogenous factors and pharmacologic compoundshave been widely demonstrated to block transdifferentiationof tubular epithelial cells and to ameliorate renal fibrosis,their effects on EMT of MC and in the prevention of peritonealmembrane failure are just starting to be tested in vitro andin animal models of PD. In this context, it has been demonstratedthat AngII induces the expression of TGF- and fibronectin inMC and that inhibitors of angiotensin-converting enzyme andAngII type 1 receptor attenuate the production of VEGF in thesecells.29,45 In addition, it was previously shown that intraperitonealadministration of an inhibitor of angiotensin-converting enzymeattenuated structural and functional alteration of the peritoneumin a rat PD model.46
It should be taken into consideration that EMT of MC is a physiologicprocess that is necessary for wound healing during the aggressionof the peritoneal membrane induced by PD; therefore, it is plausiblethat chronic blockade of EMT would result in inefficient tissuerepair. Therefore, alternative therapeutic approaches may beaddressed to treat the consequences of the EMT of MC, such ascellular invasion (step 3), fibrosis (steps 4 and 5), or angiogenesis(step 6; Figure 3).
Induction of EMT of MC is accompanied by upregulated expressionof MMP such as MMP-2 and MMP-9, which would degrade the basalmembrane and the connective tissue, allowing the submesothelialinvasion by the transdifferentiated cells.14 It can be expectedthat MMP inhibitors may prevent the accumulation MC-derivedmyofibroblasts in the submesothelial compartment, which in turnwould diminish the structural alteration of the peritoneal membrane(step 3).
MC that have undergone an EMT produce higher amounts of ECMcomponents, including fibronectin and collagen I,13 and displayless fibrinolytic capacity as a result of unbalanced ratio betweentissue plasminogen activator and plasminogen activator inhibitor-1.47Therapeutic interventions of peritoneal fibrosis may be designedeither to prevent ECM synthesis (step 4) or to increase fibrinolysis(step 5). A number of antifibrotic drugs, including pentoxifylline,dipyridamole, troglitazone, diltiazem, and emodin, have directinhibitory effects on ECM protein synthesis or on TGF- expressionand activity in MC.48–52 The statin simvastatin is anothercandidate molecule for therapeutic treatment of peritoneal fibrosisbecause it exhibits a dual effect: (1) Inhibition of ECM synthesisand (2) fibrinolytic activity by inducing tissue plasminogenactivator synthesis and inhibiting plasminogen activator inhibitor-1expression.53
Finally, it has been demonstrated that during the progressionof EMT, MC produce higher amounts of VEGF, which correlatedwith increased peritoneal transport.13 Therefore, therapeuticintervention may also be directed to prevent peritoneal angiogenesisand vessel permeability, via interrupting VEGF expression orits effects on endothelial cells (step 6). For instance, theinhibitors of angiogenesis TNP-470 and endostatin have beenshown to suppress the progression of peritoneal deteriorationin mouse experimental models.54,55
Recent studies using ex vivo cultures of effluent-derived MC,in conjunction with immunohistochemical analysis of peritonealbiopsies, have allowed the identification of the EMT of MC asa key process in peritoneal membrane failure. Although it canbe argued that effluent MC are not representative of the mesotheliumbecause they might be cells that are committed to detach fromthe tissue, there is increasing evidence that MC that are releasedinto PD effluent are representative of the MC population thatremains attached to the peritoneum. In fact, it could be demonstratedthat effluent-derived MC reflected more precisely the functionalstatus of the peritoneal tissue than previous tests.
It can be expected that proteomics and functional genomics analysisof the EMT process of MC will provide fine biomarkers for accuratefollow-up of the progressive peritoneal membrane deteriorationand for the identification of master molecules that govern themesenchymal conversion of MC. These molecular profiles of theEMT process might also become excellent tools to test the biocompatibilityof new PD fluids, both in in vitro studies and in animal modelsof PD. In addition, the increasing knowledge of the events thatculminate in EMT and its consequences will provide future hopefor long-term preservation of peritoneal membrane through thedevelopment of preventive and therapeutic approaches. However,the design of new therapeutic strategies still requires extensiveanalysis, in vitro and in animal models, of the various drugsand factors with potential effects on EMT and its deleteriousconsequences before starting with clinical trials.
This work was supported by grant FIS 03/0599 to R.S. and bygrant SAF 2004-07855 to M.L.-C. This work was also partiallysupported by Fresenius Medical Care and by a research awardfrom Fundación Instituto de Credito Oficial to M.L.-C.
Footnotes
Published online ahead of print. Publication date availableat www.jasn.org.
Krediet RT, Lindholm B, Rippe B: Pathophysiology of peritoneal membrane failure.
Perit Dial Int 20[Suppl 4]
: S22
–S42, 2000
Krediet RT: The peritoneal membrane in chronic peritoneal dialysis.
Kidney Int 55
: 341
–356, 1999[CrossRef][Medline]
Gokal R, Khanna R, Krediet RT, Norph KD:
The Textbook of Peritoneal Dialysis, Dordrecht, Kluwer Academic Publishers, 2000
Di Paolo N, Sacchi G: Atlas of peritoneal histology.
Perit Dial Int 20[Suppl 3]
: S5
–S96, 2000
Williams JD, Craig KJ, Topley N, Von Ruhland C, Fallon M, Newman GR, Mackenzie RK, Williams GT: Morphologic changes in the peritoneal membrane of patients with renal disease.
J Am Soc Nephrol 13
: 470
–479, 2002[Abstract/Free Full Text]
Mateijsen MA, van der Wal AC, Hendriks PM, Zweers MM, Mulder J, Struijk DG, Krediet RT: Vascular and interstitial changes in the peritoneum of CAPD patients with peritoneal sclerosis.
Perit Dial Int 19
: 517
–525, 1999[Abstract/Free Full Text]
Plum J, Hermann S, Fussholler A, Schoenicke G, Donner A, Rohrborn A, Grabensee B: Peritoneal sclerosis in peritoneal dialysis patients related to dialysis settings and peritoneal transport properties.
Kidney Int Suppl 78
: S42
–S47, 2001[Medline]
Margetts PJ, Bonniaud P: Basic mechanisms and clinical implications of peritoneal fibrosis.
Perit Dial Int 23
: 530
–541, 2003[Abstract/Free Full Text]
Krediet RT, Zweers MM, van der Wal AC, Struijk DG: Neoangiogenesis in the peritoneal membrane.
Perit Dial Int 20[Suppl 2]
: S19
–S25, 2000
Sherif AM, Nakayama M, Maruyama Y, Yoshida H, Yamamoto H, Yokoyama K, Kawakami M: Quantitative assessment of the peritoneal vessel density and vasculopathy in CAPD patients.
Nephrol Dial Transplant 21
: 1675
–1681, 2006[Abstract/Free Full Text]
Yanez-Mo M, Lara-Pezzi E, Selgas R, Ramirez-Huesca M, Dominguez-Jimenez C, Jimenez-Heffernan JA, Aguilera A, Sanchez-Tomero JA, Bajo MA, Alvarez V, Castro MA, del Peso G, Cirujeda A, Gamallo C, Sanchez-Madrid F, Lopez-Cabrera M: Peritoneal dialysis and epithelial-to-mesenchymal transition of mesothelial cells.
N Engl J Med 348
: 403
–413, 2003[Abstract/Free Full Text]
Jimenez-Heffernan JA, Aguilera A, Aroeira LS, Lara-Pezzi E, Bajo MA, del Peso G, Ramirez M, Gamallo C, Sanchez-Tomero JA, Alvarez V, Lopez-Cabrera M, Selgas R: Immunohistochemical characterization of fibroblast subpopulations in normal peritoneal tissue and in peritoneal dialysis-induced fibrosis.
Virchows Arch 444
: 247
–256, 2004[CrossRef][Medline]
Aroeira LS, Aguilera A, Selgas R, Ramirez-Huesca M, Perez-Lozano ML, Cirugeda A, Bajo MA, del Peso G, Sanchez-Tomero JA, Jimenez-Heffernan JA, Lopez-Cabrera M: Mesenchymal conversion of mesothelial cells as a mechanism responsible for high solute transport rate in peritoneal dialysis: Role of vascular endothelial growth factor.
Am J Kidney Dis 46
: 938
–948, 2005[CrossRef][Medline]
Margetts PJ, Bonniaud P, Liu L, Hoff CM, Holmes CJ, West-Mays JA, Kelly MM: Transient overexpression of TGF-beta1 induces epithelial mesenchymal transition in the rodent peritoneum.
J Am Soc Nephrol 16
: 425
–436, 2005[Abstract/Free Full Text]
Fenton SS, Schaubel DE, Desmeules M, Morrison HI, Mao Y, Copleston P, Jeffery JR, Kjellstrand CM: Hemodialysis versus peritoneal dialysis: A comparison of adjusted mortality rates.
Am J Kidney Dis 30
: 334
–342, 1997[Medline]
Margetts PJ, Kolb M, Galt T, Hoff CM, Shockley TR, Gauldie J: Gene transfer of transforming growth factor-beta1 to the rat peritoneum: Effects on membrane function.
J Am Soc Nephrol 12
: 2029
–2039, 2001[Abstract/Free Full Text]
Gillerot G, Goffin E, Michel C, Evenepoel P, Biesen WV, Tintillier M, Stenvinkel P, Heimburger O, Lindholm B, Nordfors L, Robert A, Devuyst O: Genetic and clinical factors influence the baseline permeability of the peritoneal membrane.
Kidney Int 67
: 2477
–2487, 2005[CrossRef][Medline]
Topley N, Liberek T, Davenport A, Li FK, Fear H, Williams JD: Activation of inflammation and leukocyte recruitment into the peritoneal cavity.
Kidney Int Suppl 56
: S17
–S21, 1996[Medline]
Lai KN, Lai KB, Lam CW, Chan TM, Li FK, Leung JC: Changes of cytokine profiles during peritonitis in patients on continuous ambulatory peritoneal dialysis.
Am J Kidney Dis 35
: 644
–652, 2000
Sakamoto N, Sugimura K, Kawashima H, Tsuchida K, Takemoto Y, Naganuma T, Tatsumi S, Nakatani T: Influence of glucose and inflammatory cytokines on TGF-beta1 and CTGF mRNA expressions in human peritoneal mesothelial cells.
Int J Mol Med 15
: 907
–911, 2005[Medline]
Masunaga Y, Muto S, Asakura S, Akimoto T, Homma S, Kusano E, Asano Y: Ascites from patients with encapsulating peritoneal sclerosis augments NIH/3T3 fibroblast proliferation.
Ther Apher Dial 7
: 486
–493, 2003[CrossRef][Medline]
Wen Y, Gu J, Li SL, Reddy MA, Natarajan R, Nadler JL: Elevated glucose and diabetes promote interleukin-12 cytokine gene expression in mouse macrophages.
Endocrinology 147
: 2518
–2525, 2006[Abstract/Free Full Text]
Nevado J, Peiro C, Vallejo S, El-Assar M, Lafuente N, Matesanz N, Azcutia V, Cercas E, Sanchez-Ferrer CF, Rodriguez-Manas L: Amadori adducts activate nuclear factor-kappaB-related proinflammatory genes in cultured human peritoneal mesothelial cells.
Br J Pharmacol 146
: 268
–279, 2005[CrossRef][Medline]
Wang T, Chen YG, Ye RG, Li SL, Li HQ, Chen YX: Effect of glucose on TGF-beta 1 expression in peritoneal mesothelial cells.
Adv Perit Dial 11
: 7
–10, 1995[Medline]
Axelsson J, Carrero JJ, Avesani CM, Heimburger O, Lindholm B, Stenvinkel P: Adipokine signaling in the peritoneal dialysis patient.
Contrib Nephrol 150
: 166
–173, 2006[Medline]
Leung JC, Chan LY, Tang SC, Chu KM, Lai KN: Leptin induces TGF-beta synthesis through functional leptin receptor expressed by human peritoneal mesothelial cell.
Kidney Int 69
: 2078
–2086, 2006[CrossRef][Medline]
Border WA, Noble NA: Transforming growth factor beta in tissue fibrosis.
N Engl J Med 331
: 1286
–1292, 1994[Free Full Text]
Kiribayashi K, Masaki T, Naito T, Ogawa T, Ito T, Yorioka N, Kohno N: Angiotensin II induces fibronectin expression in human peritoneal mesothelial cells via ERK1/2 and p38 MAPK.
Kidney Int 67
: 1126
–1135, 2005[CrossRef][Medline]
Pecoits-Filho R, Araujo MR, Lindholm B, Stenvinkel P, Abensur H, Romao JE Jr, Marcondes M, De Oliveira AH, Noronha IL: Plasma and dialysate IL-6 and VEGF concentrations are associated with high peritoneal solute transport rate.
Nephrol Dial Transplant 17
: 1480
–1486, 2002[Abstract/Free Full Text]
Zweers MM, de Waart DR, Smit W, Struijk DG, Krediet RT: Growth factors VEGF and TGF-beta1 in peritoneal dialysis.
J Lab Clin Med 134
: 124
–132, 1999[CrossRef][Medline]
De Vriese AS, Tilton RG, Mortier S, Lameire NH: Myofibroblast transdifferentiation of mesothelial cells is mediated by RAGE and contributes to peritoneal fibrosis in uraemia.
Nephrol Dial Transplant 21
: 2549
–2555, 2006[Abstract/Free Full Text]
Combet S, Ferrier ML, Van Landschoot M, Stoenoiu M, Moulin P, Miyata T, Lameire N, Devuyst O: Chronic uremia induces permeability changes, increased nitric oxide synthase expression, and structural modifications in the peritoneum.
J Am Soc Nephrol 12
: 2146
–2157, 2001[Abstract/Free Full Text]
Powell DW, Mifflin RC, Valentich JD, Crowe SE, Saada JI, West AB: Myofibroblasts. I. Paracrine cells important in health and disease.
Am J Physiol 277
: C1
–C9, 1999[Medline]
Bucala R, Spiegel LA, Chesney J, Hogan M, Cerami A: Circulating fibrocytes define a new leukocyte subpopulation that mediates tissue repair.
Mol Med 1
: 71
–81, 1994[Medline]
Kalluri R, Neilson EG: Epithelial-mesenchymal transition and its implications for fibrosis.
J Clin Invest 112
: 1776
–1784, 2003[CrossRef][Medline]
Liu Y: Epithelial to mesenchymal transition in renal fibrogenesis: Pathologic significance, molecular mechanism, and therapeutic intervention.
J Am Soc Nephrol 15
: 1
–12, 2004[Abstract/Free Full Text]
Barrallo-Gimeno A, Nieto MA: The Snail genes as inducers of cell movement and survival: Implications in development and cancer.
Development 132
: 3151
–3161, 2005[Abstract/Free Full Text]
Dubois RN, Abramson SB, Crofford L, Gupta RA, Simon LS, Van De Putte LB, Lipsky PE: Cyclooxygenase in biology and disease.
FASEB J 12
: 1063
–1073, 1998[Abstract/Free Full Text]
Mandl-Weber S, Cohen CD, Haslinger B, Kretzler M, Sitter T: Vascular endothelial growth factor production and regulation in human peritoneal mesothelial cells.
Kidney Int 61
: 570
–578, 2002[CrossRef][Medline]
Rodrigues A, Martins M, Santos MJ, Fonseca I, Oliveira JC, Cabrita A, Melo e Castro J, Krediet RT: Evaluation of effluent markers cancer antigen 125, vascular endothelial growth factor, and interleukin-6: Relationship with peritoneal transport.
Adv Perit Dial 20
: 8
–12, 2004[Medline]
Vargha R, Endemann M, Kratochwill K, Riesenhuber A, Wick N, Krachler AM, Malaga-Dieguez L, Aufricht C: Ex vivo reversal of in vivo transdifferentiation in mesothelial cells grown from peritoneal dialysate effluents.
Nephrol Dial Transplant 21
: 2943
–2947, 2006[Abstract/Free Full Text]
Tan X, Li Y, Liu Y: Paricalcitol attenuates renal interstitial fibrosis in obstructive nephropathy.
J Am Soc Nephrol 17
: 3382
–3393, 2006[Abstract/Free Full Text]
Sauter M, Cohen CD, Wornle M, Mussack T, Ladurner R, Sitter T: ACE inhibitor and at1-receptor blocker attenuate the production of VEGF in mesothelial cells.
Perit Dial Int 27
: 167
–172, 2007[Abstract/Free Full Text]
Duman S, Wieczorowska-Tobis K, Styszynski A, Kwiatkowska B, Breborowicz A, Oreopoulos DG: Intraperitoneal enalapril ameliorates morphologic changes induced by hypertonic peritoneal dialysis solutions in rat peritoneum.
Adv Perit Dial 20
: 31
–36, 2004[Medline]
Rougier JP, Guia S, Hagege J, Nguyen G, Ronco PM: PAI-1 secretion and matrix deposition in human peritoneal mesothelial cell cultures: Transcriptional regulation by TGF-beta 1.
Kidney Int 54
: 87
–98, 1998[CrossRef][Medline]
Chan TM, Leung JK, Tsang RC, Liu ZH, Li LS, Yung S: Emodin ameliorates glucose-induced matrix synthesis in human peritoneal mesothelial cells.
Kidney Int 64
: 519
–533, 2003[CrossRef][Medline]
Fang CC, Yen CJ, Chen YM, Chu TS, Lin MT, Yang JY, Tsai TJ: Diltiazem suppresses collagen synthesis and IL-1beta-induced TGF-beta1 production on human peritoneal mesothelial cells.
Nephrol Dial Transplant 21
: 1340
–1347, 2006[Abstract/Free Full Text]
Fang CC, Yen CJ, Chen YM, Shyu RS, Tsai TJ, Lee PH, Hsieh BS: Pentoxifylline inhibits human peritoneal mesothelial cell growth and collagen synthesis: Effects on TGF-beta.
Kidney Int 57
: 2626
–2633, 2000[CrossRef][Medline]
Hung KY, Chen CT, Huang JW, Lee PH, Tsai TJ, Hsieh BS: Dipyridamole inhibits TGF-beta-induced collagen gene expression in human peritoneal mesothelial cells.
Kidney Int 60
: 1249
–1257, 2001[CrossRef][Medline]
Peng Y, Liu H, Liu F, Liu Y, Li J, Chen X: Troglitazone inhibits synthesis of transforming growth factor-beta1 and reduces matrix production in human peritoneal mesothelial cells.
Nephrology (Carlton) 11
: 516
–523, 2006[CrossRef][Medline]
Haslinger B, Kleemann R, Toet KH, Kooistra T: Simvastatin suppresses tissue factor expression and increases fibrinolytic activity in tumor necrosis factor-alpha-activated human peritoneal mesothelial cells.
Kidney Int 63
: 2065
–2074, 2003[CrossRef][Medline]
Tanabe K, Maeshima Y, Ichinose K, Kitayama H, Takazawa Y, Hirokoshi K, Kinomura M, Sugiyama H, Makino H: Endostatin peptide, an inhibitor of angiogenesis, prevents the progression of peritoneal sclerosis in a mouse experimental model.
Kidney Int 71
: 227
–238, 2007[CrossRef][Medline]
Yoshio Y, Miyazaki M, Abe K, Nishino T, Furusu A, Mizuta Y, Harada T, Ozono Y, Koji T, Kohno S: TNP-470, an angiogenesis inhibitor, suppresses the progression of peritoneal fibrosis in mouse experimental model.
Kidney Int 66
: 1677
–1685, 2004[CrossRef][Medline]
This article has been cited by other articles:
R. Strippoli, I. Benedicto, M. L. Perez Lozano, A. Cerezo, M. Lopez-Cabrera, and M. A. d
1 in rat peritoneum.14 In these studies, EMT appears as the central point in the early pathogenesis of peritoneal damage associated with PD. Given that there is no definitive treatment for the progressive loss of the dialytic capacity of the peritoneum, the identification of the EMT of MC as a key process in the onset and progression of peritoneal fibrosis and angiogenesis opens new insights for therapeutic intervention. 
, IL-1, IL-8, TGF-
B, and ligand-bound FGF receptor 1, which in conjunction with 
Top
Abstract
Introduction
