Am J Physiol Renal Physiol 306: F116 –F122, 2014. First published November 6, 2013; doi:10.1152/ajprenal.00543.2013.
Epithelial-to-mesenchymal transition and slit function of mesothelial cells are regulated by the cross talk between mesothelial cells and endothelial cells Shigehisa Aoki,1 Toshiaki Takezawa,2 Ayumi Oshikata-Miyazaki,2 Satoshi Ikeda,1 Hiroyuki Kuroyama,3 Tomoyuki Chimuro,3 Yuji Oguchi,3 Mitsuru Noguchi,4 Yutaka Narisawa,5 and Shuji Toda1 1
Department of Pathology and Microbiology, Faculty of Medicine, Saga University, Saga, Japan; 2Division of Animal Sciences, National Institute of Agrobiological Sciences, Ibaraki, Japan; 3Kanto Chemical Company. Incorporated, Tokyo, Japan; 4Department of Urology, Faculty of Medicine, Saga University, Saga, Japan; and 5Department of Dermatology, Faculty of Medicine, Saga University, Saga, Japan Submitted 11 October 2013; accepted in final form 31 October 2013
Aoki S, Takezawa T, Oshikata-Miyazaki A, Ikeda S, Kuroyama H, Chimuro T, Oguchi Y, Noguchi M, Narisawa Y, Toda S. Epithelial-to-mesenchymal transition and slit function of mesothelial cells are regulated by the cross talk between mesothelial cells and endothelial cells. Am J Physiol Renal Physiol 306: F116 –F122, 2014. First published November 6, 2013; doi:10.1152/ajprenal.00543.2013.—Peritoneal dysfunction is a major factor leading to treatment failure of peritoneal dialysis (PD). However, the precise mechanism of the peritoneal diffusion changes related to PD remains to be elucidated. To this end, we have established a novel peritoneal diffusion model in vitro, which consists of a three-dimensional culture system using a collagen vitrigel membrane chamber and a fluid-stream generation system. This artificial peritoneal model revealed that high-glucose culture medium and fluid flow stress promoted the epithelial-mesenchymal transition (EMT) process of mesothelial cells and that endothelial cells inhibited this mesothelial EMT process. Mesothelial cells in the EMT state showed high expression of connective tissue growth factor and low expression of bone morphogenic protein-7, while non-EMT mesothelial cells showed the opposite expression pattern of these two proteins. In addition, these protein expressions were dependent on the presence of endothelial cells in the model. Our model revealed that the endothelial slit function was predominantly dependent on the covering surface area, while the mesothelial layer possessed a specific barrier function for small solutes independently of the surface area. Notably, a synergic barrier effect of mesothelial cells and endothelial cells was present with low-glucose pretreatment, but high-glucose pretreatment abolished this synergic effect. These findings suggest that the mesothelial slit function is not only regulated by the high-glucose-induced EMT process but is also affected by an endothelial paracrine effect. This peritoneal diffusion model could be a promising tool for the development of PD. collagen vitrigel membrane; epithelial-mesenchymal transition; peritoneal permeability
a major factor leading to treatment failure of peritoneal dialysis (PD) (5, 6). Dialysis works on the principle of both solute diffusion of dissolved substances and ultrafiltration of fluid across a semipermeable membrane, and PD uses the patient’s peritoneal membrane in the abdomen as the semipermeable membrane. Loss of diffusion capacity of the peritoneal membrane leads to a reduction in the clearance of solutes and causes insufficient efficiency of dialysis. This dysfunction is a common consequence of inadequate peritoneal permeability secondary to inappropriate regeneration of the
PERITONEAL DYSFUNCTION IS
Address for reprint requests and other correspondence: S. Aoki, Dept. of Pathology and Microbiology, Faculty of Medicine, Saga Univ., 5-1-1 Nabeshima, Saga 849-8501, Japan (e-mail: [email protected]). F116
peritoneal membrane (16). However, the precise relationship between peritoneal permeability and pathological morphology remains to be clarified. The peritoneal membrane is divided into three components that each has a specific function, namely, the mesothelial layer, interstitial collagen, and small vessels (Fig. 1, A and B). These components have specific pores and a resistance that contributes to solute diffusion (12, 17, 23). It is known, however, that the small pores of endothelial cells are the most important slits for solute diffusion (13, 22), and the slit function of both the mesothelial layer and the interstitial collagen remains unexplained in vitro. The epithelial-mesenchymal transition (EMT) of peritoneal mesothelial cells is recognized to play an important role in denudation of the peritoneal membrane (11, 28). However, the role of EMT of mesothelial cells in the slit function of the peritoneal membrane remains uncertain. Interstitial tissues generally consist of bundles of collagen, and their influence on the diffusion coefficient of the peritoneal membrane has been poorly understood. To clarify the changes in the slit function in the peritoneal denaturation state, it is necessary to establish an appropriate permeability system in which the three fundamental factors of mesothelial cells, collagen scaffold, and endothelial cells are combined and which replicates a mesothelial cell EMT state. It remained difficult to reproduce the EMT process in vitro until we established a fluid flow stress culture model that can replicate mesothelial hyperplasia and EMT under culture conditions (1, 2). It was also difficult to create collagen-based membrane chambers that allow permeability tests. Since most collagen reagents exist in sol and gel states, it is difficult to apply a chamber membrane created from commercial collagen products. Meanwhile, Takezawa et al. (25) developed a novel biomaterial consisting of high-density collagen fibrils equivalent to connective tissues in vivo, and named it “collagen vitrigel” because its preparation required a vitrification process. Recently, we succeeded in developing a collagen vitrigel membrane chamber useful for reconstructing culture models, such as “tissue sheets” composed of epithelial cells, mesenchymal cells, or endothelial cells alone, and “organoid plates” composed of more than two types of cells (24, 26, 27). This novel collagen material has enabled us to make a new culture chamber that replicates the microenvironment of organs specifically related to the extracellular matrix. In this study, we have established a novel peritoneal diffusion model to clarify the precise relationship between peritoneal permeability and pathological changes of cell constituents of the peritoneum. In the model, mesothelial cells, collagen
1931-857X/14 Copyright © 2014 the American Physiological Society
http://www.ajprenal.org
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
A
B
Peritoneum
F117
Solute diffusion PD solution
Mesothelial layer
Mesothelial cell
Collagen fiber
Collagen fiber
Endothelial cell
Blood vessel Small vessel
C
Peritoneal diffusion model
D
Continuous fluid flow stress generation system
Inner chamber wall
Outer dish
5% CO2, 37°C PD solution
Fig. 1. A: histology and schematic illustration of the major cellular components of the peritoneum. The surface is covered by a mesothelial layer. The submesothelial layer contains a collagen layer and small vessels. B: schematic representation of solute movement between a blood vessel and peritoneal dialysis (PD) solution. C: schematic representation of a vitrigel-peritoneal membrane (PM) model. 2MG,  2-micorglobulin. D: schematic representation of the fluid flow stress-generating system.
2D-rotation
PD solution
Collagen vitrigel Collagen vitrigel
Urea, creatinine, uric acid and β2MG were added to the culture medium
speed = 40 rpm, r = 2.0 cm
Rotatory shaker
vitrigel membrane, and endothelial cells replicate an artificial peritoneal membrane, which we refer to as a vitrigel-PM model. Fluid flow stress and high-glucose stimulation induce a mesothelial EMT state of the vitrigel-PM model in vitro. Permeability tests of this artificial membrane in the EMT state revealed an unknown peritoneal slit factor and close cell-cell interactions between mesothelial cells and endothelial cells. This artificial peritoneum diffusion model will likely open up a new field for peritoneal cellular kinetics and specific slit function. METHODS
Animals and cell line. All procedures involving animal materials were performed in accordance with the regulations laid down by Saga University as Ethical Guidelines. Collagen vitrigel membrane chamber. A collagen vitrigel membrane was prepared by the following three processes as previously reported (25): 1) a gelation process in which 2.0 ml of 0.25% type I collagen sol was allowed to form an opaque soft gel in a culture dish with a diameter of 35 mm; 2) a vitrification process in which the gel became a rigid material through sufficient drying; and 3) a rehydration process in which the vitrified material was converted into a thin transparent gel membrane with enhanced gel strength by moisture supplementation. Subsequently, a collagen xerogel membrane defined as a dried collagen vitrigel membrane without free water was prepared by simply revitrifying a collagen vitrigel membrane on a separable sheet. The collagen xerogel membrane was pasted onto the bottom edge of a plastic cylinder with an inner-outer diameter of 11–13 mm and a length of 15 mm, and two hangers were connected to the top edge of the cylinder, resulting in the fabrication of a collagen xerogel membrane chamber that easily allowed conversion into a collagen vitrigel membrane chamber by rehydration, as previously reported (24). For cell culture on the outer surface of the collagen vitrigel membrane chamber, another plastic cylinder was connected to the
ON OFF
chamber by taping the edges of the two cylinders together via the collagen xerogel membrane. Vitrigel-PM model and permeability test. For the reconstructed peritoneum conditions, 5 ⫻ 104 mesothelial cells of the cell line MeT-5A (CRL-9444; American Type Culture Collection, Manassas, VA) and 5 ⫻ 104 endothelial cells of the cell line MS-1 (CRL-2279; American Type Culture Collection) were seeded onto the inner and outer surfaces of the collagen vitrigel membrane chamber, respectively. We prepared three types of vitrigel-PM models: 1) mesothelial cells with endothelial cells (vitrigel-ME); 2) mesothelial cells alone (vitrigel-M); and 3) endothelial cells alone (vitrigel-E). To mimic normal and fibrotic peritoneum conditions, the cell-seeded chambers were cultured for 2 or 28 days in complete medium, comprising low-glucose (LG) DMEM containing 1,000 mg/l of glucose (Sigma, St. Louis, MO) and high-glucose (HG) DMEM containing 4,500 mg/l of glucose, both supplemented with 10% FBS, 100 g/ml of streptomycin, and 100 g/ml of penicillin under fluid flow stress (1, 2). To replicate the PD conditions, 0.5 ml of PD solution (Dianeal-N PD-2-1.5; Baxter, Tokyo, Japan) was poured into the inner side of the reconstructed chambers (Fig. 1C). The chambers were then settled in 12-well plates in 1.5 ml of complete medium described above containing 1.5 mg/ml of urea (Kanto Chemical, Tokyo, Japan), 0.15 mg/ml of creatinine (Kanto Chemical), 0.2 mg/ml of uric acid (Kanto Chemical), or 0.5 g/ml of 2-microglobulin (2MG; Sigma). Transport of low- and medium-molecular weight solutes was assessed by the dialysate-over-medium (D/M) ratio for urea, creatinine, uric acid, and 2MG. Transport of glucose was assessed by the end-to-initial dialysate concentration (D/D0) ratio. Before the permeability assay, the chamber membranes for observation were fixed in neutral-buffered 10% formalin and vertically embedded in paraffin to allow confirmation of the phenotypic changes. In the diffusion test, dialysate and culture medium samples were collected at 0, 0.5, 1.0, 2.0, and 4.0 h of dwell time. The concentrations of urea, creatinine, uric acid, and 2MG in the PD solution and culture medium were measured by a clinical chemistry analyzer
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
F118
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
(BioMajesty JCA-BM1650; JEOL, Tokyo, Japan) using CicaLiquid-N UN (Kanto Chemical) with a detection limit of 2–2,000 mg/l and a coefficient of variation of 5%; CicaLiquid-N CRE with a detection limit of 0.2–1,000 mg/l and a coefficient of variation of 5%; CicaLiquid UA with a detection limit of 0 –200 mg/l and a coefficient of variation of 5%; and Runpia Latex 2MG (Kyokuto Pharmaceutical Industrial, Tokyo, Japan) with a detection limit of 0.1–16.0 mg/l and a coefficient of variation of 10%, respectively. Fluid flow stress-generating system. The fluid flow stress-generating system was described as part of our previous papers (2, 3). Briefly, mesothelial cells and peritoneal tissue fragments were incubated under 5% CO2 and 20% O2 at 37°C in a CO2 incubator. To generate fluid flow stress, the culture dishes were placed on a rotatory shaker (MIR-S100C; Panasonic, Tokyo, Japan), which rotated at a speed of 40 rpm with a radius of 2.0 cm (Fig. 1D). Immunohistochemistry. Immunohistochemistry was performed as described previously (2). For identification of epithelial and fibrosis markers, the following antibodies were used: mouse monoclonal anti-cytokeratin AE1/AE3 (CK AE1/AE3; Dako, Glostrup, Denmark) and mouse monoclonal anti-vimentin, mouse monoclonal anti-vascular endothelial growth factor (VEGF), rabbit monoclonal anti-bone morphogenic protein (BMP), and anti-connective tissue growth factor (CTGF; Abcam, Cambridge, UK). Antigen retrieval was performed by microwave treatment in 0.01 mol/l citrate buffer (pH 6.0) at 95°C for 4 min. Statistical analysis. Data obtained from three to five independent experiments were analyzed by the Scheffé test. Values are presented as means ⫾ SE, together with the number of experiments carried out. Values of P ⬍ 0.05 were considered to indicate statistically significant differences. Averages of replicates in experiments were used to determine the statistical significance. All statistical analyses were performed using SPSS for Windows 20.0 (SPSS, Chicago, IL). RESULTS
Phenotypic analysis of the vitrigel-PM model. In this study, we prepared three types of vitrigel-PM models: vitrigel-ME, vitrigel-M, and vitrigel-E. These three vitrigel-PM models were cultured for 2 (D2) or 28 days (D28) in LG or HG media under fluid flow stress. As shown in Fig. 2A, the mesothelial cells of vitrigel-ME in both LG and HG media and vitrigel-M in LG medium showed a monolayer structure, while the mesothelial cells of vitrigel-M in HG medium showed mild hypertrophy and a multilayer structure at D2. The hyperplastic change in the mesothelial cells in vitrigel-M in LG and HG media was further intensified by D28. In contrast, the hyperplastic change was significantly inhibited in vitrigel-ME, even in HG medium, at D28 (Fig. 2, A and B). These mesothelial phenotypic changes are consistent with our previous reports (1, 2). It has been known that mesothelial cells located in the omentum under pathological conditions and the hernia sac show a hyperplastic phenotype and multilayer formation (23, 29). In contrast, the endothelial cells did not show any obvious phenotypic changes under any of the conditions examined. Replication of the mesothelial EMT process in the vitrigel-PM model. The mesothelial cells of vitrigel-ME and vitrigel-M expressed the epithelial marker CK AE1/AE3 independently of the glucose concentration at D2, while those of vitrigel-M in LG and HG media showed weak and absent CK AE1/AE3 expression, respectively, at D28 (Fig. 3). The vimentin expression of vitrigel-M in LG and HG media was higher than that of vitrigel-ME at D2. The expression intensity of vimentin increased with time, and vitrigel-M in HG medium showed high vimentin expression. The vimentin expression
level was inversely associated with the CK AE1/AE3 expression pattern. These phenotypic changes in mesothelial cells were consistent with the EMT process (28). The mesothelial cells of vitrigel-ME in HG medium and vitrigel-M in LG and HG media showed VEGF expression at D2. At D28, all mesothelial cells expressed VEGF, and the HG concentration synergistically promoted the VEGF expression of mesothelial cells. The mesothelial cells of vitrigel-ME showed high expression of BMP-7, while faint BMP-7 expression was seen in the mesothelial cells of vitrigel-M at D2 and D28. The mesothelial cells of vitrigel-ME in both LG and HG media showed weak or faint expression of CTGF, while the mesothelial cells of vitrigel-M in LG and HG media showed distinct expression of CTGF at D2. At D28, the mesothelial cells of vitrigel-ME in HG medium and vitrigel-M showed stronger positivity for CTGF compared with those of vitrigel-ME in LG medium. The vitrigel-PM models replicated the EMT process and reactivity to glucose of mesothelial cells, and the cellular dysfunction of endothelial cells. The models showed a clear inverse correlation between BMP-7 and CTGF expressions in mesothelial cells, which depended on the presence of endothelial cells. Vitrigel-PM-based peritoneal diffusion model. To evaluate the permeability of the vitrigel-PM, we performed permeability tests using a PD solution. The three types of vitrigel-PM models were precultured for 2 or 28 days (D2, D28) in LG or HG medium under fluid flow stress before the permeability tests. To replicate PD conditions, 0.5 ml of PD solution was poured into the inner side of the reconstructed chambers. Transport of low- and medium-molecular weight solutes was assessed by the D/M ratio for urea, creatinine, uric acid, and 2MG. Transport of glucose was assessed by the D/D0 ratio. The control vitrigel membrane without a cell component showed predominantly higher permeability for the tested items under all D2 conditions examined. The line plots for D/M and D/D0 are shown in Fig. 4. Urea has a molecular weight of 60.06, and its D/M ratio in vitrigel-E was significantly higher than that in vitrigel-M and vitrigel-ME in both LG and HG media for 4 h in the D2 group. Vitrigel-ME in LG medium showed a significantly higher urea D/M ratio than vitrigel-ME in HG medium and vitrigel-M in LG and HG media at 0.5 h. In the D28 group, the D/M ratio of urea in vitrigel-E was significantly higher than that in vitrigel-M and vitrigel-ME in LG and HG media for 4 h. Vitrigel-ME in LG medium showed a significantly higher urea D/M ratio than vitrigel-ME in HG medium and vitrigel-M in both LG and HG media for 4 h. Creatinine has a molecular weight of 113.12, and its D/M ratio in vitrigel-E was significantly higher than that in vitrigel-ME and vitrigel-M in LG and HG media until 2 h in the D2 group. Vitrigel-ME showed a significantly higher creatinine D/M ratio than vitrigel-M in LG medium at 0.5, 2.0, and 4.0 h. In the D28 group, the D/M ratio of creatinine in vitrigel-E was significantly higher than that in vitrigel-ME and vitrigel-M in LG and HG media for 4 h. Vitrigel-ME in LG medium showed a significantly higher creatinine permeability than vitrigel-ME in HG medium at 2.0 and 4.0 h. Vitrigel-M in LG medium also showed a higher creatinine permeability than vitrigel-M in HG medium at 2.0 and 4.0 h.
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
F119
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
A
Vitrigel-ME
Vitrigel-M
Vitrigel-E Collagen vitrigel
Collagen vitrigel
Collagen vitrigel
LG D2 HG
LG D28 HG 50 µm
B Thickness of mesothelial layer (µm)
** **
40
** **
30
20
**
10
0 Vitrigel-cell types
ME
ME
M
M
Culture medium
LG
HG
LG
HG
D28
Fig. 2. A: mesothelial cells cultured with endothelial cells of vitrigel-mesothelial cells with endothelial cells (ME) in both low-glucose (LG) and high-glucose (HG) media and mesothelial cells of vitrigel-M in LG medium show a monolayer structure, while the mesothelial cells of vitrigel-M in HG medium show mild hypertrophy and a multilayer structure at day 2 (D2). The mesothelial cells of vitrigel-M show an augmented hyperplastic change in LG and HG media at day 28 (D28). On the other hand, the mesothelial hyperplasia is significantly inhibited in vitrigel-ME, even in HG medium, at D28. The endothelial cells of vitrigel-ME and endothelial cells of vitrigel-E do not show any obvious phenotypic changes under all conditions examined. B: endothelial cells inhibit the hyperplastic change in mesothelial cells. Bar ⫽ 50 m. Values are means ⫾ SE. **P ⬍ 0.001.
Uric acid has a molecular weight of 168.11, and its D/M ratio in vitrigel-E was significantly higher than that in vitrigel-ME and vitrigel-M in both LG and HG media until 2 h in the D2 group. Vitrigel-ME showed a significantly higher uric acid D/M ratio than vitrigel-M in both LG and HG media at 0.5 h. At 2.0 and 4.0 h, the uric acid permeability in vitrigel-ME was higher than that in vitrigel-M in HG medium. In the D28 group, the permeability of uric acid in vitrigel-E was significantly higher than that in vitrigel-ME in HG medium and vitrigel-M in both LG and HG media for 4 h. Vitrigel-ME in LG medium showed a significantly higher uric acid D/M ratio than vitrigel-ME in HG medium and vitrigel-M in both LG and HG media for 4 h. The D/M ratio of uric acid in vitrigel-E in LG medium was significantly higher than that in vitrigel-E in HG medium at 2.0 h. 2MG has a molecular weight of 11,800, and its D/M ratio in vitrigel-E was significantly higher than that in the other
models for 4 h in the D2 group. There was no difference between vitrigel-ME and vitrigel-E in both LG and HG media. In the D28 group, the permeability of 2MG in vitrigel-ME in HG medium was significantly lower than that in vitrigel-ME in LG medium and vitrigel-M in both LG and HG media at 1.0 and 2.0 h. The D/M ratio of 2MG in vitrigel-E in LG medium was significantly higher than that in vitrigel-E in HG medium at 1.0 h. Glucose has a molecular weight of 180.16, and its D/D0 ratio in vitrigel-E was significantly lower than that in vitrigel-M in both LG and HG media for 4 h in the D2 group. The D/M ratio of glucose in vitrigel-E in LG medium was significantly higher than that in vitrigel-E in HG medium at 0.5 h. There was no difference between vitrigel-ME in LG and HG media for 4 h. Vitrigel-ME in LG medium possessed a significantly higher permeability to glucose than vitrigel-M in LG medium at 4.0 h. In the D28 group, the D/D0 ratio of glucose
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
F120
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
Vitrigel-ME
CK AE1/AE3
LG
HG
Vitrigel-M LG
HG D2 D28
Vimentin
D2 D28
VEGF
D2 D28
BMP-7
D2 D28
CTGF
D2
50 µm
D28
Fig. 3. Immunostaining of vitrigel-PM models. All mesothelial cells show cytokeratin (CK) AE1/AE3 expression at D2, while those of vitrigel-M in LG and HG media show weak and absent CK AE1/AE3 expression, respectively, at D28. The intensity of vimentin expression increases with time, and vitrigel-M in HG medium shows high expression of vimentin. The mesothelial cells of vitrigel-ME in HG medium and vitrigel-M in LG and HG media show VEGF expression at D2. At D28, all mesothelial cells express VEGF, and a HG concentration synergistically promotes VEGF expression in mesothelial cells. Bone morphogenic protein (BMP-7) expression levels are high in the mesothelial cells of vitrigel-ME and low in those of vitrigel-M. The CTGF expression pattern shows an inverse correlation with that of BMP-7 expression. Bar ⫽ 50 m.
in vitrigel-ME in HG medium and vitrigel-M in LG and HG media was significantly higher than that in the other models. At 4.0 h, vitrigel-ME in LG medium showed a significantly higher glucose D/D0 ratio than vitrigel-E in LG and HG media. The glucose permeability in vitrigel-ME in LG medium was significantly higher than that in vitrigel-ME in HG medium and vitrigel-M in both LG and HG media for 4 h. These findings revealed that the barrier function of the mesothelial layer was present, while that of the endothelial layer was minimal in this model. Notably, a synergic barrier effect of mesothelial cells and endothelial cells was present in LG medium at D28, but this synergic effect was abolished in HG medium at D28. These findings suggest that glucose concentration is a critical factor for solute diffusion in this vitrigel-PM model. DISCUSSION
In the present study, we have established a novel peritoneal diffusion model in vitro and demonstrated for the first time that
mesothelial cells and endothelial cells each have a specific slit function. In addition, the mesothelial slit function is not only regulated by the EMT process induced by high glucose but is also affected by an endothelial paracrine effect. A limitation of this study may lie in the discrepancy between solute transport in our artificial model and clinical findings for long-term PD patients. Over time, PD patients develop increased transport of low-molecular-weight solutes across the peritoneal membrane (18). The enhanced diffusion is recognized by an increase in the number of peritoneal vessels, which frequently leads to an increased vascular surface area in longterm PD. This pathological change in the peritoneum causes an elevation of the D/P ratio (14, 19). On the other hand, the surface area of our vitrigel-PM model is fixed, and it can only be used to evaluate changes in solute diffusion with respect to the identical surface area. In other words, our vitrigel-PM model cannot increase its surface area in a manner that mimics the increase in vascular surface area in vivo. This study has demonstrated that high glucose stimulation slightly affects the endothelial slit function. However, our findings are consistent with the common recognition that diffusion coefficients are largely dependent on the vascular surface area. Endothelial cells have been widely recognized as having specific pores and are supposed to play a key role in solute transport (12, 17, 23). In contrast, the mesothelial layer was not regarded as making a major contribution to the barrier function of the peritoneum (9, 10). Flessner (8) reported that the parietal peritoneum, including the mesothelial layers, which usually serves as a barrier, did not function against small solutes. Several reports have described mesothelial cells undergoing EMT and suggested that this may be responsible for high peritoneal transport rates (4, 7). Despite speculation that a phenotypic change, especially EMT, of mesothelial cells may affect peritoneal permeability, the precise mechanism has remained unclear. Our simplified in vitro system revealed that endothelial cells inhibited the EMT process and the permeability loss of mesothelial cells in LG medium and that this effect was abolished by HG medium in a short time period. With the abundant neovascular vessels present in peritoneal fibrosis, some research into the cell-cell interactions between endothelial cells and mesothelial cells has been carried out (21), but their cross talk has not been demonstrated. Both BMP-7 and CTGF play major roles in peritoneal fibrosis and have opposite effects and expression patterns (15, 20, 29). In this study, endothelial cells were shown to promote BMP-7 expression and inhibit CTGF expression in mesothelial cells. In addition, these expression patterns were strongly correlated with the morphological EMT change in mesothelial cells. These findings suggest that the high glucose concentrations in dialysates would not only lead to mesothelial dysfunction brought on by EMT but also reduce the protective effect afforded by endothelial cells. This breakdown of the homeostasis of the microenvironment between endothelial cells and mesothelial cells and the mesothelial EMT would synergistically promote peritoneal fibrosis, leading to ultrafiltration failure in long-term PD. In conclusion, we have established a peritoneal diffusion model. Permeability tests of the vitrigel-PM model demonstrated for the first time that mesothelial cells and endothelial cells each play a role in the slit function related to peritoneal permeability. In addition, the mesothelial slit function is not
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
D2
D28 †,∗
1.2
††,∗
††,∗
D/MUA
††,∗ ¶¶,∗
0.8
1.0
§§§,∗∗
†,∗
§§,∗∗
0.8
§§,∗∗ §§§,∗∗
††,∗
†,∗∗
††,∗∗
††,∗∗
§§,∗
§,∗
§§§,∗∗
§§,∗ §§§,∗∗
††,∗
0.6
§§§,∗∗
0.4
0.2
0.2
Vitrigel-ME in HG Vitrigel-M in LG
Vitrigel-E in LG
§§,∗
0.4
Vitrigel-ME in LG
Vitrigel-M in HG
§§,∗
§§§,∗∗
0.6
†,∗∗
1.2
§§§,∗∗
§§,∗∗
1.0
F121
§§§,∗∗
Vitrigel-E in HG Vitrigel control
0
0 0
0.5
1.0
2.0
3.0
0
4.0
0.5
1.0
2.0
3.0
4.0
hours †,∗∗
1.2
D/MCre
††,∗
††,∗
0.8
†,∗∗
§§,∗∗
§§,∗
§§§,∗∗
1.2
§§§,∗∗
§§,∗∗
1.0
1.0
§§§,∗∗
†,∗∗
¶,∗
††,∗∗
§§,∗
¶,∗
§§§,∗∗
§§,∗∗ ††,∗∗
0.8
§§§,∗∗
††,∗∗
0.6
0.6
§§,∗
†: Vitrigel-ME in LG vs Vitrigel-ME in HG ††: Vitrigel-ME in LG vs Vitrigel-M in LG ¶: Vitrigel-M in LG vs Vitrigel-M in HG
§§§,∗∗
¶¶: Vitrigel-ME in HG vs Vitrigel-M in HG
§§,∗ §§§,∗∗
§: Vitrigel-E in LG vs Vitrigel-E in HG
§§§,∗∗
0.4
0.4
§§: Vitrigel-ME in LG vs Vitrigel-E in LG
0.2
0.2
§§§: Vitrigel-ME in HG vs Vitrigel-E in HG
0
0 0
0.5
1.0
2.0
3.0
0
4.0
0.5
1.0
2.0
3.0
4.0 †,∗
1.2
1.2
D/MUrea
††,∗
§§,∗∗
§§,∗∗ §§§,∗∗
0.8
††,∗
0.8
§§§,∗∗
††,∗∗ §§,∗
0.6
0.6
0.4
0.4
0.2
0.2
0
§§§,∗∗
§§,∗ §§§,∗∗
§§,∗ §§§,∗∗
§§§,∗∗
0 0
D/Mβ2MG
§§,∗
††,∗ †,∗
1.0
§§§,∗∗
Fig. 4. Permeability test for creatinine in vitrigel-PM models. The tested vitrigel-PM models were precultured for D2 and D28 in both LG and HG media. Three types of vitrigel-PM models were used: vitrigel-ME; vitrigel-M; and vitrigel-E. Permeability was assessed by the dialysate-over-medium (D/M) ratio. Values are means ⫾ SE. *P ⬍ 0.05, **P ⬍ 0.001.
††,∗
†,∗
§§,∗∗
1.0
0.5
1.0
2.0
3.0
4.0
0
1.2
1.2
1.0
1.0
0.8
0.5
1.0
2.0
3.0
4.0
3.0
4.0
0.8 §§,∗∗ §§§,∗∗
0.6
0.6
§§,∗∗
0.4
§§,∗∗
†,∗
§§§,∗∗
0.4
§§,∗∗ §§§,∗∗ §§§,∗∗
0.2
0.2
0
†,∗
¶¶,∗
††,∗
§,∗
§§§,∗ §§,∗
†,∗
0 0
0.5
1.0
2.0
3.0
4.0
0
0.5
1.0
2.0
††,∗
†,∗∗
§§§,∗∗
††,∗
†,∗
1.2
1.2
††,∗
††,∗
§,∗
1.0
§§§,∗∗
§§,∗∗
D/D0Glu
§§§,∗∗
0.8
§§,∗∗
††,∗
§§§,∗∗
§§,∗∗ §§§,∗∗
0.6
0.4 0.2 0.5
2.0
3.0
4.0
§§,∗∗ §§§,∗∗
0 1.0
††,∗
0.6
0.2
0
†,∗∗
§§§,∗
0.8
0.4
0
§§§,∗
1.0
0
0.5
1.0
2.0
3.0
4.0
only regulated by the EMT process induced by high glucose but is also affected by an endothelial paracrine effect. This alternative peritoneal model could become a promising tool for further PD development. More specifically, focusing on the dysfunctional peritoneal homeostasis in PD patients can help one to realize that the distinct cross talk between endothelial dysfunction and cellular endothelial-mesothelial cross talk in peritoneal fibrosis may represent a novel therapeutic target for peritoneal fibrosis.
GRANTS
ACKNOWLEDGMENTS
Author contributions: S.A., T.T., Y.O., M.N., Y.N., and S.T. provided conception and design of research; S.A., A.O.-M., H.K., T.C., and Y.O. performed experiments; S.A., A.O.-M., S.I., H.K., T.C., Y.O., and M.N. analyzed data; S.A., T.T., A.O.-M., M.N., and S.T. interpreted results of experiments; S.A. and S.I. prepared figures; S.A. drafted manuscript; S.A.,
We thank H. Ideguchi, M. Nishida, F. Mutoh, S. Nakahara, and I. Nanbu for excellent technical assistance. We are grateful to K. Tokaichi for refining the English of the manuscript.
This work was supported in part by an Agri-Health Translational Research Project (No. 6110) from the Ministry of Agriculture, Forestry and Fisheries of Japan and a Grant-in-Aid from the Japanese Ministry of Education, Culture, Sports, Science, and Technology for Scientific Research (No. 25461701 to S. Aoki). DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the authors. AUTHOR CONTRIBUTIONS
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
F122
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
T.T., Y.N., and S.T. edited and revised manuscript; S.A., T.T., A.O.-M., S.I., H.K., T.C., Y.O., M.N., Y.N., and S.T. approved final version of manuscript. REFERENCES 1. Aoki S, Ikeda S, Takezawa T, Kishi T, Makino J, Uchihashi K, Matsunobu A, Noguchi M, Sugihara H, Toda S. Prolonged effect of fluid flow stress on the proliferative activity of mesothelial cells after abrupt discontinuation of fluid streaming. Biochem Biophys Res Commun 416: 391–396, 2011. 2. Aoki S, Makino J, Nagashima A, Takezawa T, Nomoto N, Uchihashi K, Matsunobu A, Sanai T, Sugihara H, Toda S. Fluid flow stress affects peritoneal cell kinetics: possible pathogenesis of peritoneal fibrosis. Perit Dial Int 31: 466 –476, 2011. 3. Aoki S, Udo K, Morimoto H, Ikeda S, Takezawa T, Uchihashi K, Nishijima-Matsunobu A, Noguchi M, Sugihara H, Toda S. Adipose tissue behavior is distinctly regulated by neighboring cells and fluid flow stress: a possible role of adipose tissue in peritoneal fibrosis. J Artif Organs 16: 322–331, 2013. 4. Aroeira LS, Aguilera A, Selgas R, Ramirez-Huesca M, Perez-Lozano ML, Cirugeda A, Bajo MA, del Peso G, Sanchez-Tomero JA, JimenezHeffernan 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. 5. Churchill DN, Thorpe KE, Nolph KD, Keshaviah PR, Oreopoulos DG, Page D. Increased peritoneal membrane transport is associated with decreased patient and technique survival for continuous peritoneal dialysis patients. The Canada-USA (CANUSA) Peritoneal Dialysis Study Group. J Am Soc Nephrol 9: 1285–1292, 1998. 6. Davies SJ, Bryan J, Phillips L, Russell GI. Longitudinal changes in peritoneal kinetics: the effects of peritoneal dialysis and peritonitis. Nephrol Dial Transplant 11: 498 –506, 1996. 7. Del Peso G, Jimenez-Heffernan JA, Bajo MA, Aroeira LS, Aguilera A, Fernandez-Perpen A, Cirugeda A, Castro MJ, de Gracia R, SanchezVillanueva R, Sanchez-Tomero JA, Lopez-Cabrera M, Selgas R. Epithelial-to-mesenchymal transition of mesothelial cells is an early event during peritoneal dialysis and is associated with high peritoneal transport. Kidney Int Suppl: S26 –S33, 2008. 8. Flessner MF. Osmotic barrier of the parietal peritoneum. Am J Physiol Renal Fluid Electrolyte Physiol 267: F861–F870, 1994. 9. Flessner MF, Dedrick RL, Schultz JS. Exchange of macromolecules between peritoneal cavity and plasma. Am J Physiol Heart Circ Physiol 248: H15–H25, 1985. 10. Flessner MF, Fenstermacher JD, Blasberg RG, Dedrick RL. Peritoneal absorption of macromolecules studied by quantitative autoradiography. Am J Physiol Heart Circ Physiol 248: H26 –H32, 1985. 11. Friedman SL, Sheppard D, Duffield JS, Violette S. Therapy for fibrotic diseases: nearing the starting line. Sci Transl Med 5: 1–17, 2013. 12. Gotloib L, Digenis GE, Rabinovich S, Medline A, Oreopoulos DG. Ultrastructure of normal rabbit mesentery. Nephron 34: 248 –255, 1983. 13. Grotte G. Passage of dextran molecules across the blood-lymph barrier. Acta Chir Scand Suppl 211: 1–84, 1956.
14. Honda K, Nitta K, Horita S, Yumura W, Nihei H. Morphological changes in the peritoneal vasculature of patients on CAPD with ultrafiltration failure. Nephron 72: 171–176, 1996. 15. Kalluri R, Neilson EG. Epithelial-mesenchymal transition and its implications for fibrosis. J Clin Invest 112: 1776 –1784, 2003. 16. Kim YL. Update on mechanisms of ultrafiltration failure. Perit Dial Int 29, Suppl 2: S123–S127, 2009. 17. Leak LV, Rahil K. Permeability of the diaphragmatic mesothelium: the ultrastructural basis for “stomata”. Am J Anat 151: 557–593, 1978. 18. Margetts PJ, Churchill DN. Acquired ultrafiltration dysfunction in peritoneal dialysis patients. J Am Soc Nephrol 13: 2787–2794, 2002. 19. 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. 20. Mizutani M, Ito Y, Mizuno M, Nishimura H, Suzuki Y, Hattori R, Matsukawa Y, Imai M, Oliver N, Goldschmeding R, Aten J, Krediet RT, Yuzawa Y, Matsuo S. Connective tissue growth factor (CTGF/ CCN2) is increased in peritoneal dialysis patients with high peritoneal solute transport rate. Am J Physiol Renal Physiol 298: F721–F733, 2010. 21. Perez-Lozano ML, Sandoval P, Rynne-Vidal A, Aguilera A, JimenezHeffernan JA, Albar-Vizcaino P, Majano PL, Sanchez-Tomero JA, Selgas R, Lopez-Cabrera M. Functional relevance of the switch of VEGF receptors/co-receptors during peritoneal dialysis-induced mesothelial to mesenchymal transition. PLoS One 8: e60776, 2013. 22. Renkin EM. Relation of capillary morphology to transport of fluid and large molecules: a review. Acta Physiol Scand Suppl 463: 81–91, 1979. 23. Rippe B, Haraldsson B. Capillary permeability in rat hindquarters as determined by estimations of capillary reflection coefficients. Acta Physiol Scand 127: 289 –303, 1986. 24. Takezawa T, Aoki S, Oshikata A, Okamoto C, Yamaguchi H, Narisawa Y, Toda S. A novel material of high density collagen fibrils: a collagen xerogel membrane and its application to transplantation in vivo and a culture chamber in vitro. In: 24th European Conference on Biomaterials. Bologna, Italy: Medimond, 2012, p. 181–185. 25. Takezawa T, Ozaki K, Nitani A, Takabayashi C, Shimo-Oka T. Collagen vitrigel: a novel scaffold that can facilitate a three-dimensional culture for reconstructing organoids. Cell Transplant 13: 463–473, 2004. 26. Uchino T, Takezawa T, Ikarashi Y. Reconstruction of three-dimensional human skin model composed of dendritic cells, keratinocytes and fibroblasts utilizing a handy scaffold of collagen vitrigel membrane. Toxicol In Vitro 23: 333–337, 2009. 27. Yamaguchi H, Kojima H, Takezawa T. Vitrigel-eye irritancy test method using HCE-T cells. Toxicol Sci 135: 347–355, 2013. 28. Yanez-Mo M, Lara-Pezzi E, Selgas R, Ramirez-Huesca M, DominguezJimenez 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 epithelialto-mesenchymal transition of mesothelial cells. New Engl J Med 348: 403– 413, 2003. 29. Zarrinkalam KH, Stanley JM, Gray J, Oliver N, Faull RJ. Connective tissue growth factor and its regulation in the peritoneal cavity of peritoneal dialysis patients. Kidney Int 64: 331–338, 2003.
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
Copyright of American Journal of Physiology: Renal Physiology is the property of American Physiological Society and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Epithelial-to-mesenchymal transition and slit function of mesothelial cells are regulated by the cross talk between mesothelial cells and endothelial cells Shigehisa Aoki,1 Toshiaki Takezawa,2 Ayumi Oshikata-Miyazaki,2 Satoshi Ikeda,1 Hiroyuki Kuroyama,3 Tomoyuki Chimuro,3 Yuji Oguchi,3 Mitsuru Noguchi,4 Yutaka Narisawa,5 and Shuji Toda1 1
Department of Pathology and Microbiology, Faculty of Medicine, Saga University, Saga, Japan; 2Division of Animal Sciences, National Institute of Agrobiological Sciences, Ibaraki, Japan; 3Kanto Chemical Company. Incorporated, Tokyo, Japan; 4Department of Urology, Faculty of Medicine, Saga University, Saga, Japan; and 5Department of Dermatology, Faculty of Medicine, Saga University, Saga, Japan Submitted 11 October 2013; accepted in final form 31 October 2013
Aoki S, Takezawa T, Oshikata-Miyazaki A, Ikeda S, Kuroyama H, Chimuro T, Oguchi Y, Noguchi M, Narisawa Y, Toda S. Epithelial-to-mesenchymal transition and slit function of mesothelial cells are regulated by the cross talk between mesothelial cells and endothelial cells. Am J Physiol Renal Physiol 306: F116 –F122, 2014. First published November 6, 2013; doi:10.1152/ajprenal.00543.2013.—Peritoneal dysfunction is a major factor leading to treatment failure of peritoneal dialysis (PD). However, the precise mechanism of the peritoneal diffusion changes related to PD remains to be elucidated. To this end, we have established a novel peritoneal diffusion model in vitro, which consists of a three-dimensional culture system using a collagen vitrigel membrane chamber and a fluid-stream generation system. This artificial peritoneal model revealed that high-glucose culture medium and fluid flow stress promoted the epithelial-mesenchymal transition (EMT) process of mesothelial cells and that endothelial cells inhibited this mesothelial EMT process. Mesothelial cells in the EMT state showed high expression of connective tissue growth factor and low expression of bone morphogenic protein-7, while non-EMT mesothelial cells showed the opposite expression pattern of these two proteins. In addition, these protein expressions were dependent on the presence of endothelial cells in the model. Our model revealed that the endothelial slit function was predominantly dependent on the covering surface area, while the mesothelial layer possessed a specific barrier function for small solutes independently of the surface area. Notably, a synergic barrier effect of mesothelial cells and endothelial cells was present with low-glucose pretreatment, but high-glucose pretreatment abolished this synergic effect. These findings suggest that the mesothelial slit function is not only regulated by the high-glucose-induced EMT process but is also affected by an endothelial paracrine effect. This peritoneal diffusion model could be a promising tool for the development of PD. collagen vitrigel membrane; epithelial-mesenchymal transition; peritoneal permeability
a major factor leading to treatment failure of peritoneal dialysis (PD) (5, 6). Dialysis works on the principle of both solute diffusion of dissolved substances and ultrafiltration of fluid across a semipermeable membrane, and PD uses the patient’s peritoneal membrane in the abdomen as the semipermeable membrane. Loss of diffusion capacity of the peritoneal membrane leads to a reduction in the clearance of solutes and causes insufficient efficiency of dialysis. This dysfunction is a common consequence of inadequate peritoneal permeability secondary to inappropriate regeneration of the
PERITONEAL DYSFUNCTION IS
Address for reprint requests and other correspondence: S. Aoki, Dept. of Pathology and Microbiology, Faculty of Medicine, Saga Univ., 5-1-1 Nabeshima, Saga 849-8501, Japan (e-mail: [email protected]). F116
peritoneal membrane (16). However, the precise relationship between peritoneal permeability and pathological morphology remains to be clarified. The peritoneal membrane is divided into three components that each has a specific function, namely, the mesothelial layer, interstitial collagen, and small vessels (Fig. 1, A and B). These components have specific pores and a resistance that contributes to solute diffusion (12, 17, 23). It is known, however, that the small pores of endothelial cells are the most important slits for solute diffusion (13, 22), and the slit function of both the mesothelial layer and the interstitial collagen remains unexplained in vitro. The epithelial-mesenchymal transition (EMT) of peritoneal mesothelial cells is recognized to play an important role in denudation of the peritoneal membrane (11, 28). However, the role of EMT of mesothelial cells in the slit function of the peritoneal membrane remains uncertain. Interstitial tissues generally consist of bundles of collagen, and their influence on the diffusion coefficient of the peritoneal membrane has been poorly understood. To clarify the changes in the slit function in the peritoneal denaturation state, it is necessary to establish an appropriate permeability system in which the three fundamental factors of mesothelial cells, collagen scaffold, and endothelial cells are combined and which replicates a mesothelial cell EMT state. It remained difficult to reproduce the EMT process in vitro until we established a fluid flow stress culture model that can replicate mesothelial hyperplasia and EMT under culture conditions (1, 2). It was also difficult to create collagen-based membrane chambers that allow permeability tests. Since most collagen reagents exist in sol and gel states, it is difficult to apply a chamber membrane created from commercial collagen products. Meanwhile, Takezawa et al. (25) developed a novel biomaterial consisting of high-density collagen fibrils equivalent to connective tissues in vivo, and named it “collagen vitrigel” because its preparation required a vitrification process. Recently, we succeeded in developing a collagen vitrigel membrane chamber useful for reconstructing culture models, such as “tissue sheets” composed of epithelial cells, mesenchymal cells, or endothelial cells alone, and “organoid plates” composed of more than two types of cells (24, 26, 27). This novel collagen material has enabled us to make a new culture chamber that replicates the microenvironment of organs specifically related to the extracellular matrix. In this study, we have established a novel peritoneal diffusion model to clarify the precise relationship between peritoneal permeability and pathological changes of cell constituents of the peritoneum. In the model, mesothelial cells, collagen
1931-857X/14 Copyright © 2014 the American Physiological Society
http://www.ajprenal.org
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
A
B
Peritoneum
F117
Solute diffusion PD solution
Mesothelial layer
Mesothelial cell
Collagen fiber
Collagen fiber
Endothelial cell
Blood vessel Small vessel
C
Peritoneal diffusion model
D
Continuous fluid flow stress generation system
Inner chamber wall
Outer dish
5% CO2, 37°C PD solution
Fig. 1. A: histology and schematic illustration of the major cellular components of the peritoneum. The surface is covered by a mesothelial layer. The submesothelial layer contains a collagen layer and small vessels. B: schematic representation of solute movement between a blood vessel and peritoneal dialysis (PD) solution. C: schematic representation of a vitrigel-peritoneal membrane (PM) model. 2MG,  2-micorglobulin. D: schematic representation of the fluid flow stress-generating system.
2D-rotation
PD solution
Collagen vitrigel Collagen vitrigel
Urea, creatinine, uric acid and β2MG were added to the culture medium
speed = 40 rpm, r = 2.0 cm
Rotatory shaker
vitrigel membrane, and endothelial cells replicate an artificial peritoneal membrane, which we refer to as a vitrigel-PM model. Fluid flow stress and high-glucose stimulation induce a mesothelial EMT state of the vitrigel-PM model in vitro. Permeability tests of this artificial membrane in the EMT state revealed an unknown peritoneal slit factor and close cell-cell interactions between mesothelial cells and endothelial cells. This artificial peritoneum diffusion model will likely open up a new field for peritoneal cellular kinetics and specific slit function. METHODS
Animals and cell line. All procedures involving animal materials were performed in accordance with the regulations laid down by Saga University as Ethical Guidelines. Collagen vitrigel membrane chamber. A collagen vitrigel membrane was prepared by the following three processes as previously reported (25): 1) a gelation process in which 2.0 ml of 0.25% type I collagen sol was allowed to form an opaque soft gel in a culture dish with a diameter of 35 mm; 2) a vitrification process in which the gel became a rigid material through sufficient drying; and 3) a rehydration process in which the vitrified material was converted into a thin transparent gel membrane with enhanced gel strength by moisture supplementation. Subsequently, a collagen xerogel membrane defined as a dried collagen vitrigel membrane without free water was prepared by simply revitrifying a collagen vitrigel membrane on a separable sheet. The collagen xerogel membrane was pasted onto the bottom edge of a plastic cylinder with an inner-outer diameter of 11–13 mm and a length of 15 mm, and two hangers were connected to the top edge of the cylinder, resulting in the fabrication of a collagen xerogel membrane chamber that easily allowed conversion into a collagen vitrigel membrane chamber by rehydration, as previously reported (24). For cell culture on the outer surface of the collagen vitrigel membrane chamber, another plastic cylinder was connected to the
ON OFF
chamber by taping the edges of the two cylinders together via the collagen xerogel membrane. Vitrigel-PM model and permeability test. For the reconstructed peritoneum conditions, 5 ⫻ 104 mesothelial cells of the cell line MeT-5A (CRL-9444; American Type Culture Collection, Manassas, VA) and 5 ⫻ 104 endothelial cells of the cell line MS-1 (CRL-2279; American Type Culture Collection) were seeded onto the inner and outer surfaces of the collagen vitrigel membrane chamber, respectively. We prepared three types of vitrigel-PM models: 1) mesothelial cells with endothelial cells (vitrigel-ME); 2) mesothelial cells alone (vitrigel-M); and 3) endothelial cells alone (vitrigel-E). To mimic normal and fibrotic peritoneum conditions, the cell-seeded chambers were cultured for 2 or 28 days in complete medium, comprising low-glucose (LG) DMEM containing 1,000 mg/l of glucose (Sigma, St. Louis, MO) and high-glucose (HG) DMEM containing 4,500 mg/l of glucose, both supplemented with 10% FBS, 100 g/ml of streptomycin, and 100 g/ml of penicillin under fluid flow stress (1, 2). To replicate the PD conditions, 0.5 ml of PD solution (Dianeal-N PD-2-1.5; Baxter, Tokyo, Japan) was poured into the inner side of the reconstructed chambers (Fig. 1C). The chambers were then settled in 12-well plates in 1.5 ml of complete medium described above containing 1.5 mg/ml of urea (Kanto Chemical, Tokyo, Japan), 0.15 mg/ml of creatinine (Kanto Chemical), 0.2 mg/ml of uric acid (Kanto Chemical), or 0.5 g/ml of 2-microglobulin (2MG; Sigma). Transport of low- and medium-molecular weight solutes was assessed by the dialysate-over-medium (D/M) ratio for urea, creatinine, uric acid, and 2MG. Transport of glucose was assessed by the end-to-initial dialysate concentration (D/D0) ratio. Before the permeability assay, the chamber membranes for observation were fixed in neutral-buffered 10% formalin and vertically embedded in paraffin to allow confirmation of the phenotypic changes. In the diffusion test, dialysate and culture medium samples were collected at 0, 0.5, 1.0, 2.0, and 4.0 h of dwell time. The concentrations of urea, creatinine, uric acid, and 2MG in the PD solution and culture medium were measured by a clinical chemistry analyzer
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
F118
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
(BioMajesty JCA-BM1650; JEOL, Tokyo, Japan) using CicaLiquid-N UN (Kanto Chemical) with a detection limit of 2–2,000 mg/l and a coefficient of variation of 5%; CicaLiquid-N CRE with a detection limit of 0.2–1,000 mg/l and a coefficient of variation of 5%; CicaLiquid UA with a detection limit of 0 –200 mg/l and a coefficient of variation of 5%; and Runpia Latex 2MG (Kyokuto Pharmaceutical Industrial, Tokyo, Japan) with a detection limit of 0.1–16.0 mg/l and a coefficient of variation of 10%, respectively. Fluid flow stress-generating system. The fluid flow stress-generating system was described as part of our previous papers (2, 3). Briefly, mesothelial cells and peritoneal tissue fragments were incubated under 5% CO2 and 20% O2 at 37°C in a CO2 incubator. To generate fluid flow stress, the culture dishes were placed on a rotatory shaker (MIR-S100C; Panasonic, Tokyo, Japan), which rotated at a speed of 40 rpm with a radius of 2.0 cm (Fig. 1D). Immunohistochemistry. Immunohistochemistry was performed as described previously (2). For identification of epithelial and fibrosis markers, the following antibodies were used: mouse monoclonal anti-cytokeratin AE1/AE3 (CK AE1/AE3; Dako, Glostrup, Denmark) and mouse monoclonal anti-vimentin, mouse monoclonal anti-vascular endothelial growth factor (VEGF), rabbit monoclonal anti-bone morphogenic protein (BMP), and anti-connective tissue growth factor (CTGF; Abcam, Cambridge, UK). Antigen retrieval was performed by microwave treatment in 0.01 mol/l citrate buffer (pH 6.0) at 95°C for 4 min. Statistical analysis. Data obtained from three to five independent experiments were analyzed by the Scheffé test. Values are presented as means ⫾ SE, together with the number of experiments carried out. Values of P ⬍ 0.05 were considered to indicate statistically significant differences. Averages of replicates in experiments were used to determine the statistical significance. All statistical analyses were performed using SPSS for Windows 20.0 (SPSS, Chicago, IL). RESULTS
Phenotypic analysis of the vitrigel-PM model. In this study, we prepared three types of vitrigel-PM models: vitrigel-ME, vitrigel-M, and vitrigel-E. These three vitrigel-PM models were cultured for 2 (D2) or 28 days (D28) in LG or HG media under fluid flow stress. As shown in Fig. 2A, the mesothelial cells of vitrigel-ME in both LG and HG media and vitrigel-M in LG medium showed a monolayer structure, while the mesothelial cells of vitrigel-M in HG medium showed mild hypertrophy and a multilayer structure at D2. The hyperplastic change in the mesothelial cells in vitrigel-M in LG and HG media was further intensified by D28. In contrast, the hyperplastic change was significantly inhibited in vitrigel-ME, even in HG medium, at D28 (Fig. 2, A and B). These mesothelial phenotypic changes are consistent with our previous reports (1, 2). It has been known that mesothelial cells located in the omentum under pathological conditions and the hernia sac show a hyperplastic phenotype and multilayer formation (23, 29). In contrast, the endothelial cells did not show any obvious phenotypic changes under any of the conditions examined. Replication of the mesothelial EMT process in the vitrigel-PM model. The mesothelial cells of vitrigel-ME and vitrigel-M expressed the epithelial marker CK AE1/AE3 independently of the glucose concentration at D2, while those of vitrigel-M in LG and HG media showed weak and absent CK AE1/AE3 expression, respectively, at D28 (Fig. 3). The vimentin expression of vitrigel-M in LG and HG media was higher than that of vitrigel-ME at D2. The expression intensity of vimentin increased with time, and vitrigel-M in HG medium showed high vimentin expression. The vimentin expression
level was inversely associated with the CK AE1/AE3 expression pattern. These phenotypic changes in mesothelial cells were consistent with the EMT process (28). The mesothelial cells of vitrigel-ME in HG medium and vitrigel-M in LG and HG media showed VEGF expression at D2. At D28, all mesothelial cells expressed VEGF, and the HG concentration synergistically promoted the VEGF expression of mesothelial cells. The mesothelial cells of vitrigel-ME showed high expression of BMP-7, while faint BMP-7 expression was seen in the mesothelial cells of vitrigel-M at D2 and D28. The mesothelial cells of vitrigel-ME in both LG and HG media showed weak or faint expression of CTGF, while the mesothelial cells of vitrigel-M in LG and HG media showed distinct expression of CTGF at D2. At D28, the mesothelial cells of vitrigel-ME in HG medium and vitrigel-M showed stronger positivity for CTGF compared with those of vitrigel-ME in LG medium. The vitrigel-PM models replicated the EMT process and reactivity to glucose of mesothelial cells, and the cellular dysfunction of endothelial cells. The models showed a clear inverse correlation between BMP-7 and CTGF expressions in mesothelial cells, which depended on the presence of endothelial cells. Vitrigel-PM-based peritoneal diffusion model. To evaluate the permeability of the vitrigel-PM, we performed permeability tests using a PD solution. The three types of vitrigel-PM models were precultured for 2 or 28 days (D2, D28) in LG or HG medium under fluid flow stress before the permeability tests. To replicate PD conditions, 0.5 ml of PD solution was poured into the inner side of the reconstructed chambers. Transport of low- and medium-molecular weight solutes was assessed by the D/M ratio for urea, creatinine, uric acid, and 2MG. Transport of glucose was assessed by the D/D0 ratio. The control vitrigel membrane without a cell component showed predominantly higher permeability for the tested items under all D2 conditions examined. The line plots for D/M and D/D0 are shown in Fig. 4. Urea has a molecular weight of 60.06, and its D/M ratio in vitrigel-E was significantly higher than that in vitrigel-M and vitrigel-ME in both LG and HG media for 4 h in the D2 group. Vitrigel-ME in LG medium showed a significantly higher urea D/M ratio than vitrigel-ME in HG medium and vitrigel-M in LG and HG media at 0.5 h. In the D28 group, the D/M ratio of urea in vitrigel-E was significantly higher than that in vitrigel-M and vitrigel-ME in LG and HG media for 4 h. Vitrigel-ME in LG medium showed a significantly higher urea D/M ratio than vitrigel-ME in HG medium and vitrigel-M in both LG and HG media for 4 h. Creatinine has a molecular weight of 113.12, and its D/M ratio in vitrigel-E was significantly higher than that in vitrigel-ME and vitrigel-M in LG and HG media until 2 h in the D2 group. Vitrigel-ME showed a significantly higher creatinine D/M ratio than vitrigel-M in LG medium at 0.5, 2.0, and 4.0 h. In the D28 group, the D/M ratio of creatinine in vitrigel-E was significantly higher than that in vitrigel-ME and vitrigel-M in LG and HG media for 4 h. Vitrigel-ME in LG medium showed a significantly higher creatinine permeability than vitrigel-ME in HG medium at 2.0 and 4.0 h. Vitrigel-M in LG medium also showed a higher creatinine permeability than vitrigel-M in HG medium at 2.0 and 4.0 h.
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
F119
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
A
Vitrigel-ME
Vitrigel-M
Vitrigel-E Collagen vitrigel
Collagen vitrigel
Collagen vitrigel
LG D2 HG
LG D28 HG 50 µm
B Thickness of mesothelial layer (µm)
** **
40
** **
30
20
**
10
0 Vitrigel-cell types
ME
ME
M
M
Culture medium
LG
HG
LG
HG
D28
Fig. 2. A: mesothelial cells cultured with endothelial cells of vitrigel-mesothelial cells with endothelial cells (ME) in both low-glucose (LG) and high-glucose (HG) media and mesothelial cells of vitrigel-M in LG medium show a monolayer structure, while the mesothelial cells of vitrigel-M in HG medium show mild hypertrophy and a multilayer structure at day 2 (D2). The mesothelial cells of vitrigel-M show an augmented hyperplastic change in LG and HG media at day 28 (D28). On the other hand, the mesothelial hyperplasia is significantly inhibited in vitrigel-ME, even in HG medium, at D28. The endothelial cells of vitrigel-ME and endothelial cells of vitrigel-E do not show any obvious phenotypic changes under all conditions examined. B: endothelial cells inhibit the hyperplastic change in mesothelial cells. Bar ⫽ 50 m. Values are means ⫾ SE. **P ⬍ 0.001.
Uric acid has a molecular weight of 168.11, and its D/M ratio in vitrigel-E was significantly higher than that in vitrigel-ME and vitrigel-M in both LG and HG media until 2 h in the D2 group. Vitrigel-ME showed a significantly higher uric acid D/M ratio than vitrigel-M in both LG and HG media at 0.5 h. At 2.0 and 4.0 h, the uric acid permeability in vitrigel-ME was higher than that in vitrigel-M in HG medium. In the D28 group, the permeability of uric acid in vitrigel-E was significantly higher than that in vitrigel-ME in HG medium and vitrigel-M in both LG and HG media for 4 h. Vitrigel-ME in LG medium showed a significantly higher uric acid D/M ratio than vitrigel-ME in HG medium and vitrigel-M in both LG and HG media for 4 h. The D/M ratio of uric acid in vitrigel-E in LG medium was significantly higher than that in vitrigel-E in HG medium at 2.0 h. 2MG has a molecular weight of 11,800, and its D/M ratio in vitrigel-E was significantly higher than that in the other
models for 4 h in the D2 group. There was no difference between vitrigel-ME and vitrigel-E in both LG and HG media. In the D28 group, the permeability of 2MG in vitrigel-ME in HG medium was significantly lower than that in vitrigel-ME in LG medium and vitrigel-M in both LG and HG media at 1.0 and 2.0 h. The D/M ratio of 2MG in vitrigel-E in LG medium was significantly higher than that in vitrigel-E in HG medium at 1.0 h. Glucose has a molecular weight of 180.16, and its D/D0 ratio in vitrigel-E was significantly lower than that in vitrigel-M in both LG and HG media for 4 h in the D2 group. The D/M ratio of glucose in vitrigel-E in LG medium was significantly higher than that in vitrigel-E in HG medium at 0.5 h. There was no difference between vitrigel-ME in LG and HG media for 4 h. Vitrigel-ME in LG medium possessed a significantly higher permeability to glucose than vitrigel-M in LG medium at 4.0 h. In the D28 group, the D/D0 ratio of glucose
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
F120
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
Vitrigel-ME
CK AE1/AE3
LG
HG
Vitrigel-M LG
HG D2 D28
Vimentin
D2 D28
VEGF
D2 D28
BMP-7
D2 D28
CTGF
D2
50 µm
D28
Fig. 3. Immunostaining of vitrigel-PM models. All mesothelial cells show cytokeratin (CK) AE1/AE3 expression at D2, while those of vitrigel-M in LG and HG media show weak and absent CK AE1/AE3 expression, respectively, at D28. The intensity of vimentin expression increases with time, and vitrigel-M in HG medium shows high expression of vimentin. The mesothelial cells of vitrigel-ME in HG medium and vitrigel-M in LG and HG media show VEGF expression at D2. At D28, all mesothelial cells express VEGF, and a HG concentration synergistically promotes VEGF expression in mesothelial cells. Bone morphogenic protein (BMP-7) expression levels are high in the mesothelial cells of vitrigel-ME and low in those of vitrigel-M. The CTGF expression pattern shows an inverse correlation with that of BMP-7 expression. Bar ⫽ 50 m.
in vitrigel-ME in HG medium and vitrigel-M in LG and HG media was significantly higher than that in the other models. At 4.0 h, vitrigel-ME in LG medium showed a significantly higher glucose D/D0 ratio than vitrigel-E in LG and HG media. The glucose permeability in vitrigel-ME in LG medium was significantly higher than that in vitrigel-ME in HG medium and vitrigel-M in both LG and HG media for 4 h. These findings revealed that the barrier function of the mesothelial layer was present, while that of the endothelial layer was minimal in this model. Notably, a synergic barrier effect of mesothelial cells and endothelial cells was present in LG medium at D28, but this synergic effect was abolished in HG medium at D28. These findings suggest that glucose concentration is a critical factor for solute diffusion in this vitrigel-PM model. DISCUSSION
In the present study, we have established a novel peritoneal diffusion model in vitro and demonstrated for the first time that
mesothelial cells and endothelial cells each have a specific slit function. In addition, the mesothelial slit function is not only regulated by the EMT process induced by high glucose but is also affected by an endothelial paracrine effect. A limitation of this study may lie in the discrepancy between solute transport in our artificial model and clinical findings for long-term PD patients. Over time, PD patients develop increased transport of low-molecular-weight solutes across the peritoneal membrane (18). The enhanced diffusion is recognized by an increase in the number of peritoneal vessels, which frequently leads to an increased vascular surface area in longterm PD. This pathological change in the peritoneum causes an elevation of the D/P ratio (14, 19). On the other hand, the surface area of our vitrigel-PM model is fixed, and it can only be used to evaluate changes in solute diffusion with respect to the identical surface area. In other words, our vitrigel-PM model cannot increase its surface area in a manner that mimics the increase in vascular surface area in vivo. This study has demonstrated that high glucose stimulation slightly affects the endothelial slit function. However, our findings are consistent with the common recognition that diffusion coefficients are largely dependent on the vascular surface area. Endothelial cells have been widely recognized as having specific pores and are supposed to play a key role in solute transport (12, 17, 23). In contrast, the mesothelial layer was not regarded as making a major contribution to the barrier function of the peritoneum (9, 10). Flessner (8) reported that the parietal peritoneum, including the mesothelial layers, which usually serves as a barrier, did not function against small solutes. Several reports have described mesothelial cells undergoing EMT and suggested that this may be responsible for high peritoneal transport rates (4, 7). Despite speculation that a phenotypic change, especially EMT, of mesothelial cells may affect peritoneal permeability, the precise mechanism has remained unclear. Our simplified in vitro system revealed that endothelial cells inhibited the EMT process and the permeability loss of mesothelial cells in LG medium and that this effect was abolished by HG medium in a short time period. With the abundant neovascular vessels present in peritoneal fibrosis, some research into the cell-cell interactions between endothelial cells and mesothelial cells has been carried out (21), but their cross talk has not been demonstrated. Both BMP-7 and CTGF play major roles in peritoneal fibrosis and have opposite effects and expression patterns (15, 20, 29). In this study, endothelial cells were shown to promote BMP-7 expression and inhibit CTGF expression in mesothelial cells. In addition, these expression patterns were strongly correlated with the morphological EMT change in mesothelial cells. These findings suggest that the high glucose concentrations in dialysates would not only lead to mesothelial dysfunction brought on by EMT but also reduce the protective effect afforded by endothelial cells. This breakdown of the homeostasis of the microenvironment between endothelial cells and mesothelial cells and the mesothelial EMT would synergistically promote peritoneal fibrosis, leading to ultrafiltration failure in long-term PD. In conclusion, we have established a peritoneal diffusion model. Permeability tests of the vitrigel-PM model demonstrated for the first time that mesothelial cells and endothelial cells each play a role in the slit function related to peritoneal permeability. In addition, the mesothelial slit function is not
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
D2
D28 †,∗
1.2
††,∗
††,∗
D/MUA
††,∗ ¶¶,∗
0.8
1.0
§§§,∗∗
†,∗
§§,∗∗
0.8
§§,∗∗ §§§,∗∗
††,∗
†,∗∗
††,∗∗
††,∗∗
§§,∗
§,∗
§§§,∗∗
§§,∗ §§§,∗∗
††,∗
0.6
§§§,∗∗
0.4
0.2
0.2
Vitrigel-ME in HG Vitrigel-M in LG
Vitrigel-E in LG
§§,∗
0.4
Vitrigel-ME in LG
Vitrigel-M in HG
§§,∗
§§§,∗∗
0.6
†,∗∗
1.2
§§§,∗∗
§§,∗∗
1.0
F121
§§§,∗∗
Vitrigel-E in HG Vitrigel control
0
0 0
0.5
1.0
2.0
3.0
0
4.0
0.5
1.0
2.0
3.0
4.0
hours †,∗∗
1.2
D/MCre
††,∗
††,∗
0.8
†,∗∗
§§,∗∗
§§,∗
§§§,∗∗
1.2
§§§,∗∗
§§,∗∗
1.0
1.0
§§§,∗∗
†,∗∗
¶,∗
††,∗∗
§§,∗
¶,∗
§§§,∗∗
§§,∗∗ ††,∗∗
0.8
§§§,∗∗
††,∗∗
0.6
0.6
§§,∗
†: Vitrigel-ME in LG vs Vitrigel-ME in HG ††: Vitrigel-ME in LG vs Vitrigel-M in LG ¶: Vitrigel-M in LG vs Vitrigel-M in HG
§§§,∗∗
¶¶: Vitrigel-ME in HG vs Vitrigel-M in HG
§§,∗ §§§,∗∗
§: Vitrigel-E in LG vs Vitrigel-E in HG
§§§,∗∗
0.4
0.4
§§: Vitrigel-ME in LG vs Vitrigel-E in LG
0.2
0.2
§§§: Vitrigel-ME in HG vs Vitrigel-E in HG
0
0 0
0.5
1.0
2.0
3.0
0
4.0
0.5
1.0
2.0
3.0
4.0 †,∗
1.2
1.2
D/MUrea
††,∗
§§,∗∗
§§,∗∗ §§§,∗∗
0.8
††,∗
0.8
§§§,∗∗
††,∗∗ §§,∗
0.6
0.6
0.4
0.4
0.2
0.2
0
§§§,∗∗
§§,∗ §§§,∗∗
§§,∗ §§§,∗∗
§§§,∗∗
0 0
D/Mβ2MG
§§,∗
††,∗ †,∗
1.0
§§§,∗∗
Fig. 4. Permeability test for creatinine in vitrigel-PM models. The tested vitrigel-PM models were precultured for D2 and D28 in both LG and HG media. Three types of vitrigel-PM models were used: vitrigel-ME; vitrigel-M; and vitrigel-E. Permeability was assessed by the dialysate-over-medium (D/M) ratio. Values are means ⫾ SE. *P ⬍ 0.05, **P ⬍ 0.001.
††,∗
†,∗
§§,∗∗
1.0
0.5
1.0
2.0
3.0
4.0
0
1.2
1.2
1.0
1.0
0.8
0.5
1.0
2.0
3.0
4.0
3.0
4.0
0.8 §§,∗∗ §§§,∗∗
0.6
0.6
§§,∗∗
0.4
§§,∗∗
†,∗
§§§,∗∗
0.4
§§,∗∗ §§§,∗∗ §§§,∗∗
0.2
0.2
0
†,∗
¶¶,∗
††,∗
§,∗
§§§,∗ §§,∗
†,∗
0 0
0.5
1.0
2.0
3.0
4.0
0
0.5
1.0
2.0
††,∗
†,∗∗
§§§,∗∗
††,∗
†,∗
1.2
1.2
††,∗
††,∗
§,∗
1.0
§§§,∗∗
§§,∗∗
D/D0Glu
§§§,∗∗
0.8
§§,∗∗
††,∗
§§§,∗∗
§§,∗∗ §§§,∗∗
0.6
0.4 0.2 0.5
2.0
3.0
4.0
§§,∗∗ §§§,∗∗
0 1.0
††,∗
0.6
0.2
0
†,∗∗
§§§,∗
0.8
0.4
0
§§§,∗
1.0
0
0.5
1.0
2.0
3.0
4.0
only regulated by the EMT process induced by high glucose but is also affected by an endothelial paracrine effect. This alternative peritoneal model could become a promising tool for further PD development. More specifically, focusing on the dysfunctional peritoneal homeostasis in PD patients can help one to realize that the distinct cross talk between endothelial dysfunction and cellular endothelial-mesothelial cross talk in peritoneal fibrosis may represent a novel therapeutic target for peritoneal fibrosis.
GRANTS
ACKNOWLEDGMENTS
Author contributions: S.A., T.T., Y.O., M.N., Y.N., and S.T. provided conception and design of research; S.A., A.O.-M., H.K., T.C., and Y.O. performed experiments; S.A., A.O.-M., S.I., H.K., T.C., Y.O., and M.N. analyzed data; S.A., T.T., A.O.-M., M.N., and S.T. interpreted results of experiments; S.A. and S.I. prepared figures; S.A. drafted manuscript; S.A.,
We thank H. Ideguchi, M. Nishida, F. Mutoh, S. Nakahara, and I. Nanbu for excellent technical assistance. We are grateful to K. Tokaichi for refining the English of the manuscript.
This work was supported in part by an Agri-Health Translational Research Project (No. 6110) from the Ministry of Agriculture, Forestry and Fisheries of Japan and a Grant-in-Aid from the Japanese Ministry of Education, Culture, Sports, Science, and Technology for Scientific Research (No. 25461701 to S. Aoki). DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the authors. AUTHOR CONTRIBUTIONS
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
F122
ESTABLISHMENT OF A PERITONEAL DIFFUSION MODEL IN VITRO
T.T., Y.N., and S.T. edited and revised manuscript; S.A., T.T., A.O.-M., S.I., H.K., T.C., Y.O., M.N., Y.N., and S.T. approved final version of manuscript. REFERENCES 1. Aoki S, Ikeda S, Takezawa T, Kishi T, Makino J, Uchihashi K, Matsunobu A, Noguchi M, Sugihara H, Toda S. Prolonged effect of fluid flow stress on the proliferative activity of mesothelial cells after abrupt discontinuation of fluid streaming. Biochem Biophys Res Commun 416: 391–396, 2011. 2. Aoki S, Makino J, Nagashima A, Takezawa T, Nomoto N, Uchihashi K, Matsunobu A, Sanai T, Sugihara H, Toda S. Fluid flow stress affects peritoneal cell kinetics: possible pathogenesis of peritoneal fibrosis. Perit Dial Int 31: 466 –476, 2011. 3. Aoki S, Udo K, Morimoto H, Ikeda S, Takezawa T, Uchihashi K, Nishijima-Matsunobu A, Noguchi M, Sugihara H, Toda S. Adipose tissue behavior is distinctly regulated by neighboring cells and fluid flow stress: a possible role of adipose tissue in peritoneal fibrosis. J Artif Organs 16: 322–331, 2013. 4. Aroeira LS, Aguilera A, Selgas R, Ramirez-Huesca M, Perez-Lozano ML, Cirugeda A, Bajo MA, del Peso G, Sanchez-Tomero JA, JimenezHeffernan 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. 5. Churchill DN, Thorpe KE, Nolph KD, Keshaviah PR, Oreopoulos DG, Page D. Increased peritoneal membrane transport is associated with decreased patient and technique survival for continuous peritoneal dialysis patients. The Canada-USA (CANUSA) Peritoneal Dialysis Study Group. J Am Soc Nephrol 9: 1285–1292, 1998. 6. Davies SJ, Bryan J, Phillips L, Russell GI. Longitudinal changes in peritoneal kinetics: the effects of peritoneal dialysis and peritonitis. Nephrol Dial Transplant 11: 498 –506, 1996. 7. Del Peso G, Jimenez-Heffernan JA, Bajo MA, Aroeira LS, Aguilera A, Fernandez-Perpen A, Cirugeda A, Castro MJ, de Gracia R, SanchezVillanueva R, Sanchez-Tomero JA, Lopez-Cabrera M, Selgas R. Epithelial-to-mesenchymal transition of mesothelial cells is an early event during peritoneal dialysis and is associated with high peritoneal transport. Kidney Int Suppl: S26 –S33, 2008. 8. Flessner MF. Osmotic barrier of the parietal peritoneum. Am J Physiol Renal Fluid Electrolyte Physiol 267: F861–F870, 1994. 9. Flessner MF, Dedrick RL, Schultz JS. Exchange of macromolecules between peritoneal cavity and plasma. Am J Physiol Heart Circ Physiol 248: H15–H25, 1985. 10. Flessner MF, Fenstermacher JD, Blasberg RG, Dedrick RL. Peritoneal absorption of macromolecules studied by quantitative autoradiography. Am J Physiol Heart Circ Physiol 248: H26 –H32, 1985. 11. Friedman SL, Sheppard D, Duffield JS, Violette S. Therapy for fibrotic diseases: nearing the starting line. Sci Transl Med 5: 1–17, 2013. 12. Gotloib L, Digenis GE, Rabinovich S, Medline A, Oreopoulos DG. Ultrastructure of normal rabbit mesentery. Nephron 34: 248 –255, 1983. 13. Grotte G. Passage of dextran molecules across the blood-lymph barrier. Acta Chir Scand Suppl 211: 1–84, 1956.
14. Honda K, Nitta K, Horita S, Yumura W, Nihei H. Morphological changes in the peritoneal vasculature of patients on CAPD with ultrafiltration failure. Nephron 72: 171–176, 1996. 15. Kalluri R, Neilson EG. Epithelial-mesenchymal transition and its implications for fibrosis. J Clin Invest 112: 1776 –1784, 2003. 16. Kim YL. Update on mechanisms of ultrafiltration failure. Perit Dial Int 29, Suppl 2: S123–S127, 2009. 17. Leak LV, Rahil K. Permeability of the diaphragmatic mesothelium: the ultrastructural basis for “stomata”. Am J Anat 151: 557–593, 1978. 18. Margetts PJ, Churchill DN. Acquired ultrafiltration dysfunction in peritoneal dialysis patients. J Am Soc Nephrol 13: 2787–2794, 2002. 19. 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. 20. Mizutani M, Ito Y, Mizuno M, Nishimura H, Suzuki Y, Hattori R, Matsukawa Y, Imai M, Oliver N, Goldschmeding R, Aten J, Krediet RT, Yuzawa Y, Matsuo S. Connective tissue growth factor (CTGF/ CCN2) is increased in peritoneal dialysis patients with high peritoneal solute transport rate. Am J Physiol Renal Physiol 298: F721–F733, 2010. 21. Perez-Lozano ML, Sandoval P, Rynne-Vidal A, Aguilera A, JimenezHeffernan JA, Albar-Vizcaino P, Majano PL, Sanchez-Tomero JA, Selgas R, Lopez-Cabrera M. Functional relevance of the switch of VEGF receptors/co-receptors during peritoneal dialysis-induced mesothelial to mesenchymal transition. PLoS One 8: e60776, 2013. 22. Renkin EM. Relation of capillary morphology to transport of fluid and large molecules: a review. Acta Physiol Scand Suppl 463: 81–91, 1979. 23. Rippe B, Haraldsson B. Capillary permeability in rat hindquarters as determined by estimations of capillary reflection coefficients. Acta Physiol Scand 127: 289 –303, 1986. 24. Takezawa T, Aoki S, Oshikata A, Okamoto C, Yamaguchi H, Narisawa Y, Toda S. A novel material of high density collagen fibrils: a collagen xerogel membrane and its application to transplantation in vivo and a culture chamber in vitro. In: 24th European Conference on Biomaterials. Bologna, Italy: Medimond, 2012, p. 181–185. 25. Takezawa T, Ozaki K, Nitani A, Takabayashi C, Shimo-Oka T. Collagen vitrigel: a novel scaffold that can facilitate a three-dimensional culture for reconstructing organoids. Cell Transplant 13: 463–473, 2004. 26. Uchino T, Takezawa T, Ikarashi Y. Reconstruction of three-dimensional human skin model composed of dendritic cells, keratinocytes and fibroblasts utilizing a handy scaffold of collagen vitrigel membrane. Toxicol In Vitro 23: 333–337, 2009. 27. Yamaguchi H, Kojima H, Takezawa T. Vitrigel-eye irritancy test method using HCE-T cells. Toxicol Sci 135: 347–355, 2013. 28. Yanez-Mo M, Lara-Pezzi E, Selgas R, Ramirez-Huesca M, DominguezJimenez 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 epithelialto-mesenchymal transition of mesothelial cells. New Engl J Med 348: 403– 413, 2003. 29. Zarrinkalam KH, Stanley JM, Gray J, Oliver N, Faull RJ. Connective tissue growth factor and its regulation in the peritoneal cavity of peritoneal dialysis patients. Kidney Int 64: 331–338, 2003.
AJP-Renal Physiol • doi:10.1152/ajprenal.00543.2013 • www.ajprenal.org
Copyright of American Journal of Physiology: Renal Physiology is the property of American Physiological Society and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
