Background

Prolonged exposure of human peritoneal mesothelial cells (HPMC) to high glucose concentrations in peritoneal dialysate is the principal factor leading to matrix accumulation and thickening of the peritoneal membrane, accompanied by progressive deterioration of transport functions. These changes are mediated in part through protein kinase C (PKC) activation and the induction of transforming growth factor-beta 1 (TGF-1). Emodin (3-methyl-1,6,8 trihydroxyanthraquinone) has previously been demonstrated to reduce cell proliferation and fibronectin synthesis in cultured mesangial cells. How emodin modulates glucose-induced abnormalities in HPMC has not been elucidated and thus constitutes the theme of this study.

Methods

We investigated the effects of emodin on the expression of PKC, TGF-1, fibronectin, and collagen type I in HPMC, and its effects on HPMC proliferation under physiologic (5 mmol) or high (30 mmol) glucose concentrations.

Results

Exposure of HPMC cultured with 5 mmol or 30 mmol D-glucose to emodin (20 g/mL) resulted in an initial lag of proliferation by 2.3 to 2.7 days, but did not affect cell viability or morphology at confluence. D-glucose (30 mmol) induced TGF-1 secretion in a time-dependent manner (3.72 0.29 and 4.30 0.50 pg/g cellular protein at 24 hours and 48 hours respectively, compared to 2.13 0.23 and 2.65 0.32 pg/g cellular protein at 24 hours and 48 hours, respectively for 5 mmol glucose; P < 0.001 at both time points). Such induction was ameliorated by emodin (20 g/mL) (TGF-1 concentration 2.25 0.15 and 2.96 0.33 pg/g cellular protein at 24 hours and 48 hours, respectively, in the presence of emodin and 30 mmol D-glucose; P < 0.001 compared to 30 mmol D-glucose alone at both time points). Induction of TGF-1 synthesis by 30 mmol D-glucose was associated with induction of PKC, phosphorylation of cAMP-responsive element binding protein (CREB) and activating transcription factor-1 (ATF-1), and increased fibronectin and type I collagen translation. Emodin abrogated all these effects of concentrated glucose. Immunohistochemical staining showed that 30 mmol D-glucose induced cytoplasmic, perinuclear, and extracellular fibronectin and type I collagen expression by HPMC. Emodin reduced 30 mmol D-glucose-induced cytoplasmic and extracellular matrix synthesis to near basal levels.

Conclusion

Our findings demonstrate that emodin ameliorates the undesirable effects of concentrated glucose on HPMC via suppression of PKC activation and CREB phosphorylation, and suggest that emodin may have a therapeutic potential in the prevention or treatment of glucose-induced structural and functional abnormalities in the peritoneal membrane.

METHODS

Materials

All chemicals were purchased from Sigma Chemical Co. (St. Louis, MO, USA) unless otherwise stated and were of the highest purity. Tissue culture flasks were purchased from Falcon (Becton Dickinson, Hong Kong) and culture media and supplements from Invitrogen Hong Kong Ltd. (Hong Kong, China). Antisera to human FN (clone IST-4), Coll-I (clone COL-1 or C-18), TGF-1 (sc-146), PKC (clone MC5), and actin (sc-8432) were purchased from Sigma Chemical Co. and TWC BioSearch International (Santa Cruz, CA, USA). The PhosphoPlus CREB (Ser133) antibody kit (New England BioLabs, Beijing, China; TWC Biosearch International) was used to detect CREB and phosphoCREB. Tyrosine kinase assay kit was purchased from Sigma Chemical Co. Cytotoxicity detection kit [lactate dehydrogenase (LDH)] was purchased from Boehringer Mannheim (Roche Diagnostics GmbH, Mannheim, Germany). [³H]-thymidine (methyl-[³H-thymidine) was purchased from Amersham Biosciences China Limited (Hong Kong, China). Emodin was extracted from Rheum officinale and fraction-distilled to above 98% purity, and characterized by high-performance liquid chromatography (HPLC), nuclear magnetic resonance (NMR), and mass spectrometry³³. Endotoxin-free emodin was dissolved in 0.1 mol/L NaOH at a stock concentration of 4 mg/mL, and diluted in Medium 199 prior to use in experiments. Human umbilical vascular endothelial cells (HUVEC) and normal human bronchial epithelial cells (NHBE) were purchased from Clonetics (San Diego, CA, USA), and human pleural mesothelial cells (MeT-5A) and mink lung epithelial cells (MLEC; CCL-64) were purchased from American Type Culture Collection (ATCC) (Manassas, VA, USA).

Cytochemical analysis of peritoneal biopsies

Parietal peritoneal biopsies were obtained from 5 patients at the time of initial insertion of the PD catheter, and from 8 patients who had been on PD for more than 12 months after informed consent, at the time of removal, or reinsertion of the peritoneal dialysis catheter. Paraffin sections (10 m) were incubated with primary antibodies against human FN, Coll-I, or TGF-1 (1:50 for all antibodies) overnight at room temperature, and incubated with relevant secondary antibodies conjugated to horseradish peroxidase for 1 hour at 37°C. Signal detection and visualization was achieved using the peroxidase-anti-peroxidase (PAP) method (Dako; Gene Company, Ltd., Hong Kong, China), counterstained with hematoxylin, and specimens examined using a Leica DMLB microscope (Leica Microsystems, Ltd., Hong Kong, China). Specimens were photographed with Kodak Max 100 film using a MPS60 camera.

Culture of cells

HPMC were obtained by enzymatic digestion of omental specimens from nonuremic patients who underwent abdominal surgery, and cultured in Medium199 supplemented with 10% fetal calf serum (FCS) as previously described^5,6. All experiments were performed on cells of the second passage that had been growth arrested for 72 hours. After synchronization of the cell cycle, HPMC were cultured under physiologic D-glucose concentration (5 mmol), or elevated D-glucose concentration (30 mmol) for periods up to 1 week, or exposed to phorbol 12-myristate 13-acetate (PMA) (100 nmol) for up to 48 hours, in the presence or absence of emodin (20 g/mL). D-glucose (30 mmol) represented the intraperitoneal glucose concentration after 1 hour of equilibration following the instillation of conventional peritoneal dialysate, which contained glucose at 75 mmol or 120 mmol. Mannitol or L-glucose was used as hexose control in parallel experiments, at concentrations identical to those of D-glucose. In separate experiments, TGF-1 neutralizing antibody (1 g/mL) was added to HPMC 1 hour prior to stimulation by elevated glucose or PMA. Human peritoneal fibroblasts (HPF) were cultured according to the method of Beavis et al³⁴, and human mesangial cells (HMC), glomerular epithelial cells (GEC), and proximal tubular epithelial cells (PTEC) by differential sieving of nephrectomized renal specimens^35,36,37.

Assessment of cell proliferation

[³H]-thymidine incorporation

Cells were seeded into 96-well plates at a density of 10,000/cm² and cultured in Medium 199 containing 10% FCS until 80% confluent. Cells were washed twice with PBS to remove unattached cells and further incubated with Medium 199 containing either 5 mmol or 30 mmol D-glucose and [³H]-thymidine (1 Ci/mL) in the presence or absence of emodin (0 to 200 g/mL) for 24 hours. At the end of the time point, the supernatant was decanted, cells trypsinized with 0.05% trypsin/0.02% ethylenediaminetetraacetic acid (EDTA), incorporated radioactivity precipitated with 10% trichloroacetic acid (TCA) (100 L), washed twice with TCA, and dissolved with 0.1 mol NaOH (200 L). [³H]-thymidine incorporation was detected by beta liquid scintillation counting.

Direct cell counting

Cells were seeded into 24-well plates at a density of 10,000/cm² and cultured for 24 hours in Medium 199 containing 10% FCS. Cells were washed twice with PBS to remove unattached cells and further incubated with Medium 199 containing either 5 mmol or 30 mmol D-glucose in the presence or absence of emodin for periods up to 7 days. At selective time periods, HPMC were trypsinized as described above, and counted using a Neubauer chamber (Sigma Chemical Co.).

Determination of lactate dehydrogenase (LDH) release

Confluent HPMC were cultured under control and experimental conditions in 96-well tissue culture plates for 24 hours. Supernatants were collected, centrifuged for 10 minutes at 2000g, and assessed for LDH release using a commercially available cytotoxicity kit according to the manufacturer's instructions. Cytotoxicity of control and experimental samples were expressed as the percentage of LDH release compared to total intracellular LDH content, the latter determined by lysis of representative cell monolayers using 2% Triton X-100 (vol/vol) according to the manufacturer's instructions.

Measurement of cellular protein concentration

HPMC cultured in triplicate in 96-well plates under control or experimental conditions were lysed with 4 mol urea buffer, 20 mmol sodium acetate, pH 6.0 containing 1% (vol/vol) Triton X-100 (50 L). Protein concentration in cell lysate was determined using a modified Lowry assay according to the manufacturer's instructions (Bio-Rad Pacific, Ltd., Hong Kong, China).

Determination of TGF-1 concentration in culture supernatant

Samples (1 mL) of culture supernatant were lyophilized, reconstituted to 200 L in PBS containing 1% fatty acid free bovine serum albumin (BSA), and assayed using a commercial TGF-1 enzyme-linked immunosorbent assay (ELISA) kit according to the manufacturer's instructions (Onwon Trading, Ltd., Hong Kong). Cross-reactivity with other TGF- isoforms was less than 2%.

Determination of TGF-1 bioactivity

The mink lung epithelial cell (MLEC) bioassay for TGF-1 was performed as described by Ha and colleagues¹³ with slight modification. Briefly, MLEC (passage 50–55) were subcultured into 96-well culture plates at 10,000 cells/cm² and cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% FCS for 2 to 4 hours. MLEC were then exposed to serially diluted recombinant human TGF-1 (0 to 1000 pg/mL) or spent culture media from HPMC obtained under control and experimental conditions for 72 hours at 37°C. Thereafter, [³H]-thymidine (1 Ci/mL) was added to each well and the incubation continued for another 24 hours. At the end of the incubation period, cells were washed three times with ice-cold PBS, trypsinized with 0.05% trypsin/0.02% EDTA, incorporated radioactivity precipitated with 10% TCA (100 L), washed twice with TCA, and dissolved with 0.1 mol NaOH (200 L). [³H]-thymidine incorporation was detected by beta liquid scintillation counting. A standard curve was constructed from the varying degrees of inhibition on MLEC growth observed with corresponding concentrations of recombinant human TGF-1.

Determination of tyrosine kinase activity

Confluent HPMC cultured in 35 mm culture dishes under control and experimental conditions were washed twice with cold PBS and lysed at 4°C with 50 mmol N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES) buffer pH 7.4, containing 1% Triton X-100, 10% glycerol, 1 mmol dithiothreitol, 1 mmol activated sodium vanadate (100 g/mL) in the presence of proteinase inhibitors (final concentration 1 mmol benzamidine, 10 g/mL leupeptin, and 2 g/mL pepstatin A). Cell lysates were centrifuged at 10,000g for 15 minutes at 4°C to remove cell debris, and assayed for tyrosine kinase according to the manufacturer's instructions (Sigma Chemical Co.). Tyrosine kinase activity is often associated with receptor tyrosine kinases, such as epidermal growth factor receptor (EGFR). Thus, EGFR (4U) was used as the positive control and incubated with emodin (20 g/mL) on ice for 10 minutes prior to assay.

Determination of FN and Coll-I synthesis by HPMC using enzyme-linked immunosorbent assay

HPMC were cultured in duplicate in 35 mm culture dishes under control or experimental conditions for 48 hours. The supernatants were decanted and cells lysed with 4 mol urea buffer, 20 mmol sodium acetate, pH 6.0 containing 1% (vol/vol) Triton X-100 (100 L). Cell lysates were desalted, buffer-exchanged into PBS using Ultrafree Centrifugal Concentrators (Millipore China Ltd., Hong Kong, China), and assayed for FN and Coll-I using "in-house" ELISAs³⁸. Briefly, 96-well microtiter plates (Immunlon 2; Dynex Technologies, Inc., Chantilly, VA, USA) were coated in duplicate with FN or Coll-I (0 to 500 ng/mL) in serial dilutions, or HPMC supernatant (starting dilution 1:10) or cell lyaste (starting dilution 1:10) in 15 mmol Na₂CO₃ and 35 mmol NaHCO₃, pH 9.6 overnight at 4°C. Plates were washed three times with PBS/0.1% Tween-20 in between all steps. The plates were blocked with 3% bovine serum albumin for 1 hour at 37°C and incubated with primary antibodies (1:1500 for mouse anti-human FN and 1:1000 for goat anti-human Coll-I) for 1 hour at 37°C. The relevant horseradish peroxidase–conjugated secondary antibody was then added for 1 hour at 37°C. Bound immunoglobulins were detected by the addition of Sigma Fast™ (Sigma Chemical Co., St. Louis, MO, USA) o-phenylenediamine dihydrochloride tablets dissolved according to the manufacturer's instructions, and the absorbance read at a wavelength of 450 nm. Prior experiments demonstrated <5% cross-reactivity of primary antibodies with other matrix proteins. Intra-assay and interassay coefficients of variance were 4.08 1.5% and 5.7 2.7%, respectively, for FN, and 6.48 2.9% and 6.95 3.4%, respectively, for Coll-I.

Preparation of cytosolic and membrane extracts

Separation of cytosolic and membrane proteins was achieved as described by Ma and colleagues³⁹ with slight modifications. Briefly, confluent HPMC exposed to control and experimental conditions for time periods up to 48 hours were scraped into PBS using a rubber policeman and collected by centrifugation at 1500g for 10 minutes. Cell pellets were sonicated five times for 10 seconds each period in 200 L PBS containing protease inhibitors⁶, and fractionated by ultracentrifugation at 100,000g for 30 minutes at 4°C. Cytosolic proteins appeared in the supernatant, while pelleted membrane proteins were resuspended in 20 mmol sodium acetate, pH 6.0, containing 4 mol urea and 1% Triton X-100 (100 L). The protein concentration of each sample was determined by a modified Lowry assay (Bio-Rad Pacific, Ltd.) prior to Western blot analysis.

Preparation of nuclear extract

Confluent HPMC exposed to control and experimental conditions for 48 hours were scraped into PBS using a rubber policeman and collected by centrifugation at 1500g for 10 minutes as previously described. Nuclear extracts were prepared as described by Abmayr and Workman⁴⁰ with slight modifications. Briefly, the cell pellet was resuspended in 10 mmol HEPES (pH 7.9) containing 1.5 mmol MgCl₂, 10 mmol KCl, 0.5 mmol dithiothreitol, and 0.2 mmol phenylmethylsulfonyl fluoride, and homogenized with a tissue tearor (BioSpec Products, Inc., Bartlesville, OK, USA) at 12,000 rpm for 2 minutes. Release of nuclei was monitored under phase contrast microscopy. Isolated nuclei were separated from cytoplasmic contents by centrifugation at 3300g for 15 minutes and nuclear proteins extracted with 20 mmol HEPES (pH 7.9) containing 25% (vol/vol) glycerol, 1.5 mmol MgCl₂, 0.8 mmol KCl, 0.2 mmol EDTA, 0.5 mmol dithiothreitol, and 0.2 mmol phenylmethylsulfonyl fluoride. The samples were then dialyzed against 20 mmol HEPES buffer (pH 7.9) containing 20% (vol/vol) glycerol, 100 mmol KCl, 0.5 mmol dithiothreitol, and the aforementioned protease inhibitors. The protein concentration of each sample was determined by a modified Lowry assay (Bio-Rad Pacific, Ltd.) prior to Western blot analysis.

Western blot analyses—Cytosolic and membrane fractions for PKC activation, total cell lysate for matrix protein expression, and nuclear extract for CREB phosphorylation

Whole cell lysates were obtained by solubilizing HPMC in 4 mol of urea buffer (500 L). Whole cell lysates, cytosolic, membrane, or nuclear extracts (10 g total protein content) was denatured in sample buffer at 95°C for 5 minutes and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Samples were electrophoresed on 7.5% and 10% acrylamide gels, respectively, to investigate FN and Coll-I synthesis, or 12% gels to investigate translation of PKC and CREB. Proteins were transferred onto nitrocellulose membrane using a mini-gel transfer system (Bio-Rad Pacific, Ltd.) at 100 V for 1 hour at 4°C. Equal loading of proteins was confirmed by staining of membranes with 1% (weight/vol) Ponceau S solution. Membranes were immunoblotted as previously described³⁵ using respective primary (1:750 for antibodies to PKC or FN; 1:1000 for antibodies to Coll-I, total CREB, or phosphoCREB) and secondary antibodies. The bands were visualized by enhanced chemiluminescence (ECL), semiquantitated by densitometry using ChemiGenius analysis software (Syngene, Ltd., Cambridge, UK) and unless otherwise stated expressed as arbitrary densitometric unit (DU). Expression of PKC, FN, and Coll-I was normalized to actin synthesis, and phosphoCREB was normalized to total CREB.

Immunohistochemical staining of HPMC

Control and stimulated HPMC were fixed with cold acetone/methanol (1:1) for 5 minutes prior to washing with PBS⁴¹. Cells were blocked with 3% BSA in PBS for 1 hour at room temperature followed by washing with PBS for 30 minutes. Cells were then incubated with primary antibody (1:50 for antibody to PKC, TGF-1, or Coll-I; 1:100 for antibody to FN) for 1 hour at 37°C in a humidified chamber. Cells were subsequently washed with PBS and the relevant fluoroisothiocyanate (FITC)-conjugated secondary antibody applied for 1 hour at 37°C in a darkened humidified chamber. After washing with PBS the cells were mounted with fluorescence mountant (Dako; Gene Company), and epifluorescence was viewed using an Axiovert Zeiss 135 inverted microscope.

Statistical analysis

Results are expressed as mean SD. Unless otherwise stated, all experiments were performed four times using HPMC from four separate donors. Data from experimental groups and controls were compared by the Student t test. P < 0.05 was considered statistically significant.

RESULTS

Expression of FN, Coll-I, and TGF-1 in peritoneal biopsies

Cytochemical staining of peritoneal biopsy specimens demonstrated FN and Coll-I synthesis in mesothelial cells Figure 1. Increased expression of both FN and Coll-I was noted in chronic PD patients Figure 1, compared to those at the time of initiation of PD Figure 1, and was accompanied by increased TGF-1 expression Figure 1.

Figure 1.

Representative illustration comparing FN (A andB), Coll-I (C andD), and TGF-1 (E andF) expression in peritoneal biopsy specimens from patients entering the PD program (renal failure controls) (A, C, and E) or patients on long-term PD (B, D, and F). Paraffin sections were subjected to cytochemical staining using the PAP method. Immunoreactive products were visualized using diaminobenzidine as chromagen and counterstained with hematoxylin. Constitutive expression of FN, Coll-I, and TGF-1 was observed in peritoneal specimens from controls (A, C, and E, respectively). Peritoneal specimens from patients on long-term PD demonstrated enhanced FN, Coll-I, and TGF-1 expression (B, D, and F, respectively), both within the mesothelium (–) and in the submesothelium (*). Abbreviations are: FN, fibronectin; Coll-I, collagen type I; TGF-1, transforming growth factor-beta 1; PD, peritoneal dialysis; PAP, peroxidase-anti-peroxidase. Original magnification, 400.

Full figure and legend (185K)

The effects of high glucose concentration and emodin on HPMC proliferation, expression of PKC, synthesis of TGF-1, FN, and Coll-I, and the mechanisms leading to these changes, were investigated using an in vitro model whereby HPMC were preconditioned in either physiologic (5 mmol) or elevated (30 mmol) D-glucose concentrations for up to 1 week in the presence or absence of emodin. Optimal dose of emodin used in all experiments was 20 g/mL (IC₅₀ = 71.4 43.6 g/mL) Figure 2 a to c. Emodin at this dose did not affect HPMC thymidine uptake or cell morphology Figure 2 a and c. In addition, it was not cytotoxic to HPMC cultured under physiologic Figure 2b or elevated glucose concentrations, as well as a wide variety of other cell types, including HPF, HMC, GEC, PTEC, HUVEC, NHBE, and MeT-5A Figure 2d.

Figure 2.

The effects of emodin on HPMC proliferation (A) and membrane integrity (B) as determined by [³H]-thymidine incorporation and LDH release respectively after incubation for 24 hours. (C) Morphology of HPMC with or without emodin (0, 20, 50, and 100 g/mL) was assessed under phase contrast microscopy, respectively; original magnification, 200). Emodin at 20 g/mL did not affect HPMC proliferation, viability, or morphology. (D) Furthermore, emodin at this dose was not cytotoxic to fibroblasts, mesangial cells, epithelial, or endothelial cells, as demonstrated by insignificant changes in LDH release in the presence () or absence () of emodin. Abbreviations are: HPMC, human peritoneal mesothelial cells; LDH, lactate dehydrogenase. Data represented are mean SD of four individual experiments. *P < 0.001 emodin vs. control (no emodin); #P < 0.0001 emodin vs. control.

Full figure and legend (79K)

Effect of emodin on HPMC proliferation

HPMC proliferation was significantly reduced after incubation with 30 mmol D-glucose for 48 hours, compared to cells incubated with 5 mmol D-glucose (direct cell count, 43,200 4100 vs. 31,200 3600 cells, 5 mmol vs. 30 mmol D-glucose, P < 0.001). After 7 days, cells exposed to 30 mmol D-glucose exhibited 14.28% reduction in proliferation compared to control cells. Administration of emodin to cells cultured under 5 mmol and 30 mmol D-glucose reduced cell proliferation for both groups transiently Figure 3 a and b, resulting in an initial lag by 2.3 0.3 days and 2.7 0.7 days for 5 mmol and 30 mmol D-glucose, respectively. Thereafter, cell proliferation proceeded at the same rate as cells not exposed to emodin, and they reached confluency 2.4 1.3 days later than control cells. Cell morphology of confluent cells remained unaltered in the presence or absence of emodin Figure 3c. Unless otherwise stated, all data were normalized to g cellular protein.

Figure 3.

Proliferation of HPMC as assessed by direct cell counting, under 5 mmol (physiologic) (A) or 30 mmol D-glucose (B), in the presence () or absence () of emodin. Proliferation was significantly reduced in HPMC cultured under 30 mmol D-glucose after 48 hours, compared to cells cultured in 5 mmol D-glucose. Addition of emodin delayed cell proliferation by 48 to 72 hours. On reaching confluency, no change was observed in cell morphology (C) in the presence (lower panels) or absence (upper panels) of emodin under 5 mmol (left panels) or 30 mmol (right panels) D-glucose. HPMC is human peritoneal mesothelial cell. *P < 0.001, glucose alone vs. glucose and emodin at the same time point; #P < 0.001 5 mmol vs. 30 mmol D-glucose at the same time point, original magnification, 200. Data represented are mean SD of six individual experiments.

Full figure and legend (95K)

Effect of emodin on tyrosine kinase and PKC expression in HPMC under high glucose concentration

Previous studies have demonstrated in human bronchial epithelial cells that emodin is a tyrosine kinase inhibitor²⁸. Our results also showed that emodin inhibited the activity of pure tyrosine kinase (EGFR) Figure 4. Elevated glucose concentration induced tyrosine kinase activity in HPMC, and emodin abrogated this induction by high glucose concentration Figure 4.

Figure 4.

Inhibition of tyrosine kinase activity by emodin. Cell lysates obtained under control or experimental conditions after 48 hours of stimulation were analyzed for tyrosine kinase activity as described. Purified tyrosine kinase (EGFR, 4U) was used as the positive control. Results are expressed as mean SD of three separate experiments.

Full figure and legend (12K)

Since induction of PKC plays a pivotal role in mediating the stimulatory effect of glucose on the synthesis of TGF-1 and matrix proteins in various cells, we examined PKC expression and the effects of emodin on PKC activation in our in vitro model. Western blot analysis showed that 30 mmol D-glucose significantly increased PKC expression in the cytosolic and membrane fractions in a time-dependent manner, commencing 12 hours after stimulation (data not shown). Maximum increase in PKC expression was observed 48 hours after stimulation, in which cytosolic and membrane fractions increased 1.4-fold and 3.7-fold, respectively, compared to 5 mmol D-glucose (Figure 5a, Table 1). While the addition of emodin to 5 mmol D-glucose had no effect on PKC expression in either the cytosolic or membrane fraction (Figure 5a, lanes 1 and 2; 0.515 0.142 DU vs. 0.415 0.095 DU, P = 0.504; and 0.370 0.015 DU vs. 0.331 0.015 DU, P = 0.487, respectively), emodin lowered PKC expression in HPMC exposed to 30 mmol D-glucose, resulting in levels similar to baseline (Figure 5a, 0.503 0.035 DU and 0.669 0.048 DU for cytosolic and membrane fraction, respectively, P < 0.05 for both). To confirm the inhibitory effect of emodin on PKC, PKC activation was achieved by exposing HPMC to PMA for 48 hours (Figure 5a, lane 6; 0.737 0.154 DU and 1.718 0.214 DU for cytosolic and membrane fraction, respectively, P < 0.05, compared to HPMC not exposed to PMA); the concomitant presence of emodin and PMA returned PKC activity to basal levels (Figure 5a, lane 5; 0.588 0.156 DU and 0.437 0.057 DU, P < 0.05, compared to lane 6). Immunohistochemical analysis confirmed the increase in PKC expression and its translocation to the cell membrane under elevated glucose concentration Figure 5b, and the abrogation of these changes by emodin Figure 5b.

Figure 5.

Western blot analysis of PKC expression in HPMC. (A) Cytosolic and membrane fractions were purified from 48 hour–stimulated HPMC as described in Methods, 10 g total protein were subjected to SDS-PAGE and probed with PKC antibody. Activation of PKC- was associated with translocation to the cell membrane (lanes 3 and 6). Emodin did not alter PKC expression under 5 mmol D-glucose, but ameliorated the increase in PKC expression in the membrane fraction induced by 30 mmol D-glucose or PMA. One representative blot of four is shown. (B) Immunohistochemical analysis of HPMC for PKC under (a) 5 mmol D-glucose, (b) 5 mmol D-glucose and emodin, (c) 30 mmol D-glucose, or (d) 30 mmol D-glucose and emodin. Cells were cultured under control or experimental conditions for 48 hours prior to fixation with cold acetone/methanol and were subsequently probed with specific antibody to PKC (arrow). D-glucose (30 mmol) induced membrane translocation of PKC (depicted by arrow). Emodin suppressed 30 mmol D-glucose–induced PKC expression to basal levels, but did not affect constitutive PKC expression. Original magnification, 600. Abbreviations are: PKC, protein kinase C ; HPMC, human peritoneal mesothelial cells; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; PMA, phorbol 12-myristate 13-acetate.

Full figure and legend (131K)

Table 1 - Densitometric intensity of cytosolic and cell membrane–associated PKC- under control and experimental conditions.

Full table

Effect of glucose, PMA, and emodin on TGF-1 synthesis by HPMC

Immunohistochemical analysis showed enhanced TGF-1 synthesis in HPMC cultured under 30 mmol D-glucose with predominant perinuclear staining, compared to cells cultured under 5 mmol D-glucose Figure 6a. While emodin has no effect on constitutive TGF-1 expression Figure 6a, concomitant addition of emodin ameliorated the induction of TGF-1 by 30 mmol D-glucose, so that both the intensity and localization of TGF-1 staining were similar to control Figure 6a.

Figure 6.

(A) Immunohistochemical analysis of HPMC for TGF-1 expression under (a) 5 mmol D-glucose, (b) 5 m D-glucose and emodin, (c) 30 mmol D-glucose, or (d) 30 mmol D-glucose and emodin. Cells were cultured under control or experimental conditions for 48 hours prior to fixation with cold acetone/methanol and were subsequently probed with specific antibody to TGF-1. Emodin suppressed 30 mmol D-glucose–induced TGF-1 expression to basal levels, but did not affect constitutive TGF-1 expression. Original magnification, 600. (B) Effect of glucose and emodin on TGF-1 secretion by HPMC. HPMC were cultured in 5 mmol D-glucose in the presence () or absence () of emodin, or in 30 mmol D-glucose with () or without () emodin for time periods up to 48 hours. TGF-1 content was measured in the supernatant (N = 4). Emodin had no effect on TGF-1 secretion by HPMC cultured under 5 mmol (physiologic) D-glucose, but abrogated the increase in TGF-1 secretion induced by 30 mmol D-glucose to levels similar to cells exposed to physiologic glucose concentration. *P < 0.001, 30 mmol D-glucose vs. 5 mmol D-glucose at T = 24 and 48 hours. (C) Effects of PMA-stimulated PKC activation, emodin, and neutralizing TGF-1 antibody on TGF-1 secretion by HPMC. HPMC were stimulated with PMA (100 nmol) in the presence or absence of emodin or TGF-1 neutralizing antibody for 48 hours. TGF-1 secretion was assessed in the supernatant (N = 4). Emodin suppressed PMA-stimulated TGF-1 secretion to basal levels, and its suppressive effect was more pronounced compared to TGF-1 neutralizing antibody. Abbreviations are: HPMC, human peritoneal mesothelial cells; TGF-1, transforming growth factor-1; PMA, phorbol 12-myristate 13-acetate; PKC, protein kinase C.

Full figure and legend (93K)

The secretion of TGF-1 by HPMC was determined using a specific TGF-1 ELISA, which demonstrated that the majority of TGF-1 (>95%) in culture supernatants under control or experimental conditions was detected only after acid-activation, implying that it was synthesized mainly in its latent form. TGF-1 synthesized by cells cultured in 5 mmol or 30 mmol D-glucose showed a time-dependent increase Figure 6b. Exposure of HPMC to 30 mmol D-glucose resulted in significantly higher supernatant concentrations of "latent" TGF-1 at 24 hours and 48 hours compared to cells cultured under physiologic glucose concentration (3.72 0.29 pg/g cellular protein vs. 2.13 0.23 pg/g cellular protein at 24 hours, P < 0.001; and 4.30 0.50 pg/g cellular protein vs. 2.65 0.32 pg/g cellular protein at 48 hours, P < 0.001). Exposure of HPMC to 30 mmol mannitol or L-glucose did not alter TGF-1 synthesis significantly compared to 5 mmol D-glucose (data not shown). Emodin did not affect the constitutive secretion of TGF-1 under 5 mmol D-glucose, but it abrogated the induction by 30 mmol D-glucose (2.25 0.15 pg/g cellular protein at 24 hours, P < 0.001; 2.96 0.33 pg/g cellular protein at 48 hours, P < 0.001) Figure 6b. PMA markedly increased TGF-1 secretion compared to 5 mmol D-glucose (6.92 2.98 pg/g cellular protein vs. 2.21 0.21 pg/g cellular protein, P < 0.0001) Figure 6c, but such increase was completely abrogated by emodin (2.58 0.21 pg/g cellular protein, P < 0.0001). Prior incubation of HPMC with neutralizing antibody to TGF-1 (1 g/mL) resulted in an 81.4% reduction of PMA-induced TGF-1 secretion (3.41 0.31 pg/g cellular protein, P < 0.0001). No further change was observed with higher concentrations of neutralizing antibody to TGF-1 (data not shown). Concomitant exposure of HPMC to PMA, neutralizing antibody to TGF-1 and emodin resulted in supernatant TGF-1 levels similar to those when HPMC was exposed to 5 mmol D-glucose alone.

Three to five percent of the total TGF-1 secreted by HPMC under control or experimental conditions was in the "active" form and was detectable by ELISA without prior acid activation. Their bioactivity was confirmed using the mink lung epithelial cell growth inhibition assay Table 2. This bioactive fraction increased 3.60- and 4.12-fold when the cells were cultured in 30 mmol D-glucose for 24 hours and 48 hours, respectively Table 2. Mannitol or L-glucose (30 mmol) showed no effect (data not shown). In comparison, stimulation of HPMC with PMA for 24 hours and 48 hours induced bioactive TGF-1 by 6.69- and 7.50-fold, respectively, compared to cells exposed to 5 mmol D-glucose. Emodin ameliorated the increase in bioactive TGF-1 induced by elevated glucose or PMA Table 2. To ascertain whether residual PMA emodin remaining in the conditioned media after incubation with HPMC could induce MLEC TGF-1 synthesis, and thus alter MLEC growth, TGF-1 levels in the conditioned media were determined using ELISA before and after administration to MLEC. Our results showed that the residual PMA emodin had no significant effect on MLEC TGF-1 synthesis or cell growth (data not shown).

Table 2 - Effect of D-glucose, PMA, and emodin on TGF-1 bioactivity.

Full table

Effect of glucose, PMA, and emodin on FN and Coll-I synthesis by HPMC

Using "in-house" FN and Coll-I ELISA, we observed up-regulation of FN synthesis and secretion after stimulation of HPMC with 30 mmol D-glucose for 48 hours, compared to cells cultured under 5 mmol D-glucose (36.3 3.6 vs. 157.8 9.1 ng/g cellular protein for 5 mmol and 30 mmol D-glucose, respectively, for secreted FN; 19.4 5.2 vs. 79.4 12.2 ng/g cellular protein for 5 mmol and 30 mmol D-glucose, respectively, for cell-associated FN, P < 0.05 for both) Table 3. FN synthesis was also significantly increased upon PKC activation (217.0 21.8 and 91.5 15.3 ng/g cellular protein for secreted and cell-associated FN, respectively, P < 0.05 compared to control). Emodin had no effect on FN synthesis under physiologic glucose concentration, but ameliorated both glucose-induced and PMA-stimulated FN synthesis. The inhibitory effect of emodin was similar to TGF-1 neutralizing antibody, and was more marked with elevated glucose concentration than PMA Table 3.

Table 3 - Determination of secreted and cell-associated FN and Coll I in HPMC under control and experimental conditions.

Full table

Elevated glucose concentration and PMA also modulated the translation of Coll-I. Coll-I was not constitutively secreted from HPMC cultured under 5 mmol D-glucose, nor was Coll-I secretion altered by high glucose concentration or PMA stimulation (data not shown). D-glucose (30 mmol) induced the expression of only the cell-associated Coll-I, compared to 5 mmol D-glucose (56.3 12.5 vs. 16.8 8.1 ng/g cellular protein, P < 0.05), and such stimulation was abrogated by emodin. Exposure of HPMC to PMA markedly increased cell-associated Coll-I (70.4 10.2, P < 0.05 compared to control), which was again ameliorated by emodin and anti-TGF-1 antibody Table 3. Western blot analysis corroborated the results on FN and Coll-I by ELISA Figure 7 a and b.

Figure 7.

Western blot analysis of FN and Coll-I synthesis by HPMC showing the effects of glucose, PMA, and emodin. (A) Culture supernatant or total cell lysate was extracted from HPMC after 48 hours of stimulation, subjected to SDS-PAGE, and probed with antibodies to FN or Coll-I. Emodin ameliorated the increase in FN (secreted and cell-associated) and Coll-I (cell-associated only) synthesis induced by 30 mmol D-glucose and PMA to control levels. One representative blot of four separate experiments is shown. (B) The intensity of the detected bands of secreted FN (), cell-associated FN (), and cell-associated Coll-I () were analyzed by densitometric scan, normalized to actin, and expressed as arbitrary units. (C) Immunohistochemical analysis of HPMC for FN and Coll-I under control or experimental conditions. HPMC were cultured in 5 mmol D-glucose (a, e, i, m), 5 mmol D-glucose and emodin (b, f, j, n), 30 mmol D-glucose (c, g, k, o), or 30 mmol D-glucose and emodin (d, h, l, p) for 72 hours (a through h) or one week (i through p) prior to fixation with cold acetone/methanol. Cells were subsequently stained for FN (a through d, i through l) or Coll-I (e through h, m through p). Cytoplasmic and extracellular deposition of FN, the latter in fibrillar form, was noted after exposure to 5 mmol D-glucose for up to one week (a vs. i), but was accelerated when cells were exposed to 30 mmol D-glucose (c and k). Fibrillar FN deposition was evident 72 hours after exposure to elevated glucose concentration (c). D-glucose (30 mmol) also induced Coll-I synthesis in a time-dependent manner similar to FN (e, g, m, o). Addition of emodin to 30 mmol D-glucose for 72 hours reduced cytoplasmic and fibrillar FN, as well as Coll-I expression, although perinuclear staining of FN and Coll-I was still evident (c and d, g and h, respectively). An extensive matrix was deposited by cells cultured in 30 mmol D-glucose for one week, containing predominantly FN (k) and to a lesser extent Coll-I (o). Both were suppressed to near basal levels when emodin was present concomitantly (l and p). Original magnification, 400. Abbreviations are: FN, fibronectin; Coll-I, collagen type I; HPMC, human peritoneal mesothelial cells; PMA, phorbol 12-myristate 13-acetate; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

Full figure and legend (146K)

Localization of FN and Coll-I in HPMC cultured under control or experimental conditions for periods up to 1 week was assessed using immunohistochemical staining Figure 7c. Significantly increased intracellular FN and Coll-I was observed after 48 hours of 30 mmol D-glucose stimulation (data not shown), while induction of FN and Coll-I deposition into the extracellular milieu was observed at 72 hours, and the latter progressed to form an extensive matrix by 1 week Figure 7c. Emodin reduced the increased matrix synthesis induced by high glucose concentration to near basal levels, and the ameliorating effect was evident even in HPMC exposed to 30 mmol D-glucose for 1 week. Despite the reduction in matrix synthesis by emodin, cell attachment was unaffected, as determined by cell counting (62,000 967 cells/cm², 62,630 1010 cells/cm², 61,567 1009 cells/cm², and 61,789 1034 cells/cm² for HPMC cultured under 5 mmol D-glucose, 5 mmol D-glucose with emodin, 30 mmol D-glucose, and 30 mmol D-glucose with emodin, respectively; P = 0.70, 0.69, and 0.78, respectively, when compared to 5 mmol D-glucose alone after 1 week of culture under control or experimental conditions). Control experiments using 30 mmol mannitol or L-glucose did not show any significant change in FN or Coll-I expression compared to 5 mmol D-glucose (data not shown).

Effect of glucose and emodin on CREB phosphorylation

CRE of the FN gene has been demonstrated to mediate increased FN synthesis in mesangial cells induced by high glucose concentration or TGF-1^31,32. We therefore investigated CREB phosphorylation in nuclear extracts isolated from HPMC cultured under physiologic and elevated glucose concentrations using Western blot analysis. Anti-phosphoCREB (pCREB) antibody used in this study was raised against the phosphoserine in a portion of the CREB protein that shared 100% homology with the phosphoserine site of the CREB-related protein activated transcription factor-1 (ATF-1)⁴². Consequently, ATF-1 could also be detected by this antibody. Our results demonstrated that while total CREB content remained unaltered under control and experimental conditions, 30 mmol D-glucose increased CREB phosphorylation at serine 133 by 1.84-fold, compared to 5 mmol D-glucose (0.46 0.25 DU vs. 0.25 0.09 DU). This was accompanied by a 1.36-fold increase in ATF-1 at serine 63 (Figure 8a, lanes 1 and 3). PMA also activated CREB and ATF-1 (Figure 8a, lane 6; 0.57 0.01 and 0.22 0.01 DU, respectively). In comparison, emodin significantly reduced the induction of CREB and ATF-1 phosphorylation by both high glucose and PMA (Figure 8a, lanes 4 and 5, respectively). In control experiments, neither 30 mmol mannitol nor L-glucose altered phosphorylation of CREB or ATF-1, as compared to 5 mmol D-glucose (data not shown).

Figure 8.

Western blot analysis showing the effects of glucose, PMA, and emodin on CREB and ATF-1 phosphorylation by HPMC after 48 hours of stimulation. (A) Nuclear extracts were subjected to SDS-PAGE and probed with antisera to phosphoCREB and phosphoATF-1. While no significant difference was observed for total CREB expression under different experimental conditions, emodin reduced 30 mmol D-glucose- or PMA-induced CREB and ATF-1 phosphorylation, but had no effect on the constitutive activation of the transcription factors. One representative blot of three separate experiments is shown. (B) The intensity of the detected bands of phosphoCREB () and phosphoATF-1 () were analyzed by densitometric scan, normalized to CREB, and expressed as arbitrary units. Abbreviations are: PMA, phorbol 12-myristate 13-acetate; CREB, cAMP-responsive element binding protein; ATF-1, activating transcription factor-1.

Full figure and legend (50K)

DISCUSSION

Mesothelial cells line the peritoneal cavity and constitute a protective bioactive barrier that plays a key role in maintaining the structural and functional integrity of the peritoneal membrane¹. In addition, the normal synthesis of peritoneal membrane constituents by peritoneal mesothelial cells is essential for peritoneal preservation. Surgical, chemical, or infective injury to the mesothelium initiates pathologic processes that culminate in disruptions of both structural and functional integrity of the peritoneal membrane^2,10,11,12. Conventional peritoneal dialysis solutions contain glucose at high concentrations as the osmotic agent. Continuous exposure of the peritoneum to these unphysiologic solutions is associated with abnormalities in cellular proliferation, synthesis of cytokines and growth factors, and increased synthesis of matrix components by the mesothelial cells and deposition of matrix within the submesothelium^3,9,13,14,15. Increased matrix deposition is precedent to peritoneal fibrosis and a major contributing factor to failure of peritoneal transport functions. In line with published work, we have demonstrated accumulation and deposition of FN and Coll-I within peritoneal biopsies obtained from chronic PD patients accompanied by enhanced TGF-1 expression, compared to biopsies obtained from new PD patients. While previous studies have implicated prolonged exposure of the peritoneum to glucose at high concentrations as a major contributing factor for these profibrotic processes, there is as yet no effective means to ameliorate or prevent the detrimental effects of concentrated glucose on peritoneal mesothelial cells.

HPMC cultured in vitro provide a relevant in vitro model to study the effect of PD solutions on cell proliferation and function, since previous in vitro studies have demonstrated that they possess the same immunohistochemical markers as peritoneal mesothelial stem cells⁴³. In addition to the suppressive effect of high glucose concentration on HPMC proliferation¹³, in this study we demonstrated for the first time that high ambient glucose concentration induced PKC autocrine phosphorylation of CREB and ATF-1 in HPMC, and induced TGF-1 mediated fibrillar and cytoplasmic FN and Coll-I synthesis. Activation of PKC, exemplified by its translocation from the cytosol to the cell membrane in HPMC by concentrated glucose, also up-regulated FN and TGF-1 gene transcription³⁰, but not that of Coll-I (unpublished data). Our results also showed that the synthesis of laminin and collagen types III and IV was similarly induced by concentrated glucose (unpublished data). These effects were a result of elevated D-glucose concentration, per se, rather than hyperosmolality, since there was no significant effect with concentrated mannitol or L-glucose compared to physiologic control. Similar effects of concentrated glucose have been reported in human mesangial cells, which have also been related to PKC activation^44,45,46. It has been suggested that glucose metabolism, either through the pentose phosphate shunt or through glycerolipid production, leads to an increase in diacylglycerol and, subsequently, activation of PKC. In this context, Ha et al¹³ have demonstrated that high glucose–induced diacylglycerol and PKC up-regulation mediated TGF-1 and FN synthesis by peritoneal mesothelial cells.

Activation of PKC has been shown to modulate transcription factors, for example, c-fos and c-jun, through the activated protein-1 (AP-1) binding site. Promotor regions of TGF-1, FN, and Coll-I have also been demonstrated to contain the AP-1 binding consensus sequence, which may therefore represent one mechanism by which high ambient glucose concentration up-regulates TGF-1 and matrix synthesis^47,48. On the other hand, it has been reported that PKC activation and CREB phosphorylation are involved in elevated glucose–mediated induction of FN transcription in human mesangial cells, mediated through a consensus CRE located at -170 bp of the FN gene^31,32. Our results in the present study showed that high glucose–induced FN synthesis was also associated with phosphorylation of CREB in HPMC. Phosphorylation of CREB at serine 133 appears to be essential for its activation, and can occur through numerous signaling pathways, for example, PKA stimulation, adenyl cyclase pathway, p42/44 MAP kinase, p38 kinase, or induction of intracellular Ca^2+49,50,51. Unlike mesangial cells^32,52, and in concordance with a previous study on FN synthesis¹³, we have found that concentrated glucose was able to induce FN and Coll-I synthesis in HPMC in the absence of exogenous TGF-1 or fetal calf serum (FCS). These results suggest that concentrated D-glucose per se plays a pivotal role in the induction of matrix synthesis in HPMC.

Other investigators have reported that emodin may possess antiproliferative and/or anti-inflammatory activities in mesangial cells, bronchial epithelial cells, lung squamous cell carcinoma cell line, and T lymphocytes, at concentrations that vary from 1.35 g/mL (5 mol) to 100 g/mL (370 mol)^25,26,28. The observed variability in the cellular effects of emodin may be attributed to species differences in its phytologic origin, or heterogeneity of response in different cell types. To address these issues, we have used highly purified emodin from a single batch and have examined its effects on cell proliferation and viability in different cell types. A prerequisite to further experiments in vivo is that the emodin species being tested must not adversely affect cell viability. Our data showed that emodin at 20 g/mL (74 mol) was nontoxic to HPMC exposed to either physiologic or elevated D-glucose concentration, as evidenced by assessment of cell proliferation and LDH release. We have, therefore, used this dose in subsequent experiments. Furthermore, our emodin at this concentration did not induce membrane damage in a wide variety of cell types, including peritoneal fibroblasts and other cells that may be involved in matrix synthesis or tissue fibrosis. It is of interest to note that emodin at 20 g/mL did not affect cell proliferation in 80% confluent HPMC, yet proliferation was transiently inhibited in newly seeded cells cultured under control and experimental conditions. This differential effect may suggest that subconfluent cells could be more sensitive to the effects of emodin. The antimitogenic effect of emodin in relation to cell density is currently being studied. A major finding in this study was that emodin consistently suppressed concentrated glucose-induced TGF-1 bioactivity, FN, and Coll-I synthesis in HPMC. These ameliorating effects of emodin on TGF-1 and matrix synthesis were genuine and not related to differences in cell number, since our results showed similar cell numbers under control and experimental conditions at confluence when the experiments were performed. Nevertheless, the observed reduction in mesothelial cell proliferation in the presence of 30 mmol D-glucose or emodin suggested alterations in the dynamic processes of cell turnover, which could be of particular relevance in the regenerative response to injury. In animal models of diabetic nephropathy, reduction of mesangial proliferation by antimitogenic agents was accompanied by amelioration of glucose-induced matrix synthesis⁵³. It remains to be investigated whether the ameliorating effect of emodin on D-glucose–induced matrix synthesis by HPMC might be causally related to its precedent antimitogenic effect that lasted for approximately 48 hours. It is noteworthy that, despite the initial inhibition of HPMC proliferation, emodin had no effect on the constitutive synthesis of TGF-1 and matrix by HPMC under 5 mmol D-glucose.

Our data demonstrated that the abrogation of elevated D-glucose–induced TGF-1 and matrix synthesis in HPMC by emodin was mediated through the inhibition of PKC activation. The results suggest a more pronounced inhibitory effect with emodin compared to neutralizing antibody to TGF-1, since the addition of emodin resulted in the complete suppression of PMA-induced TGF-1 and matrix synthesis. Since both soluble and cell-associated FN were decreased by emodin, it is likely that emodin suppressed FN synthesis, rather than inhibiting its accumulation and deposition within the extracellular milieu. Although we have demonstrated that PKC activation plays an important role in D-glucose–induced matrix synthesis in HPMC, it remains possible that other PKC isoforms may be involved in the pathogenesis of peritoneal fibrosis, as has been reported in diabetic nephropathy⁵⁴. In this regard, we are investigating the potential effects of emodin on other isoforms of PKC in the presence of different concentrations of D-glucose.

Emodin inhibits different protein kinases, including casein kinases 1 and 2, PKA, and tyrosine kinase^55,56. Whether emodin can inhibit other kinases, such as MAPK and c-jun NH₂-terminal kinase (JNK), is currently being investigated. The inhibitory action of emodin is dependent on its ability to penetrate into the active site of the enzymes²⁹. Crystallography studies by Battistutta et al²⁹ demonstrated that emodin enters the nucleotide-binding site of casein kinase II in the proximity of the site normally occupied by the natural cosubstrate. This led to structural changes in the N-terminal lobe of the enzyme, resulting in the inability of casein kinase II to bind to its cosubstrate²⁹. It is speculated that similar molecular mechanisms may explain the inhibitory action of emodin on PKC in HPMC.

CONCLUSION

We have demonstrated that emodin can effectively ameliorate the detrimental effects of concentrated D-glucose on HPMC, namely the induction of TGF-1 bioactivity and matrix synthesis, through inhibition of PKC activation and CREB phosphorylation, without compromising cellular viability. Further studies will need to be be undertaken in an animal model of peritoneal fibrosis to confirm our in vitro findings. Amelioration of TGF-1 mediated tissue fibrosis using emodin may provide important clinical potential in the long-term preservation of structural and functional integrity of the peritoneal membrane in patients on PD.

References

1. Dobbie, JW: Ultrastructural and pathology of the peritoneum in peritoneal dialysis, in The Textbook of Peritoneal Dialysis, 1994, edited by Gokal R, Nolph KD, The Netherlands, Kluwer Academic Publishers, pp 17–44
2. Levine, S, Saltzman, A: Repeated toxic injury of peritoneum: Accumulation of toxicity and adaptation to injury. J Appl Toxicol 2000 20: 431–434,  | Article | PubMed | ISI | ChemPort |
3. Chegini, N: The role of growth factors in peritoneal healing: Transforming growth factor beta (TGF-beta). Eur J Surg (1997 Suppl 577): 17–23,
4. Saed, GM, Zhang, W, Chegini, N, et al: Alteration of type I and III collagen expression in human peritoneal mesothelial cells in response to hypoxia and transforming growth factor-beta 1. Wound Repair Regen 1999 7: 504–510,  | Article | PubMed | ISI | ChemPort |
5. Stylianou, E, Jenner, LA, Davies, M, et al: Isolation, culture and characterization of human peritoneal mesothelial cells. Kidney Int 1990 37: 1563–1570,  | PubMed | ISI | ChemPort |
6. Yung, S, Thomas, GJ, Stylianou, E, et al: Source of peritoneal proteoglycans. Human peritoneal mesothelial cells synthesize and secrete mainly small dermatan sulfate proteoglycans. Am J Pathol 1995 146: 520–529,  | PubMed | ISI | ChemPort |
7. Cronauer, MV, Stadlmann, S, Klocker, H, et al: Basic fibroblast growth factor synthesis by human peritoneal mesothelial cells. Induction by interleukin-1. Am J Pathol 1999 156: 1977–1984,
8. Topley, N, Jörres, A, Luttmann, W, et al: Human peritoneal mesothelial cells synthesize interleukin-6: Induction by IL-1 and TNF. Kidney Int 1993 43: 226–233,  | PubMed | ISI | ChemPort |
9. Lai, KN, Lai, KB, Lam, CWK, et al: Changes of cytokine profiles during peritonitis in patients on continuous ambulatory peritoneal dialysis. Am J Kid Dis 2000 53: 644–652,
10. Verger, C, Brunschvicg, O, Charpentier, YL, et al: Structural and ultrastructural peritoneal membrane changes and permeability alterations during continuous ambulatory peritoneal dialysis. Proc Eur Dial Transplant Assoc 1981 18: 199–205,  | PubMed | ISI | ChemPort |
11. Williams, JD, Craig, KJ, Topley, N, et al: Morphologic changes in the peritoneal membrane of patients with renal disease. J Am Soc Nephrol 2002 13: 470–479,  | PubMed | ISI |
12. Honda, K, Nitta, K, Horita, S, et al: Morphological changes in the peritoneal vasculature of patients on CAPD with ultrafiltration failure. Nephron 1996 72: 171–176,  | PubMed | ISI | ChemPort |
13. Ha, H, Yu, MR, Lee, HB: High glucose-induced PKC activation mediates TGF-1 and fibronectin synthesis by peritoneal mesothelial cells. Kidney Int 2001 59: 463–470,  | Article | PubMed | ISI | ChemPort |
14. Hung, K-Y, Chen, C-T, Huang, J-W, et al: Dipyridamole inhibits TGF--induced collagen gene expression in human peritoneal mesothelial cells. Kidney Int 2001 60: 1249–1257,  | Article |
15. Medcalf, JF, Walls, J, Pawluczyk, IZA, Harris, KPG: Effects of glucose dialysate on extracellular matrix production by human peritoneal mesothelial cells (HPMC): The role of TGF-1. Nephrol Dial Transplant 2001 16: 1885–1892,  | PubMed | ISI | ChemPort |
16. Newton, AC: Protein kinase C: Structure, function and regulation. J Biol Chem 1995 270: 28495–28498,  | Article | PubMed | ISI | ChemPort |
17. Studer, RK, Craven, PA, DeRubertis, F: Role for protein kinase C in the mediation of increased fibronectin accumulation by mesangial cells grown in high-glucose medium. Diabetes 1993 42: 118–126,  | PubMed | ISI | ChemPort |
18. Border, WA, Ruoslahti, E: Transforming growth factor-beta in disease: The dark side of tissue repair. J Clin Invest 1992 90: 1–7,  | PubMed | ISI | ChemPort |
19. Harper, JR, Spiro, RC, Gaarde, WA, et al: Role of transforming growth factor beta and decorin in controlling fibrosis. Methods Enzymol 1994 245: 241–254,  | PubMed | ISI | ChemPort |
20. Fracasso, A, Baggio, B, Ossi, E, et al: Glycosaminoglycans prevent the functional and morphological peritoneal derangement in an experimental model of peritoneal fibrosis. Am J Kid Dis 1999 33: 105–110,  | PubMed | ISI | ChemPort |
21. Mezzano, SA, Ruiz-Ortega, M, Egido, J: Angiotensin II and renal fibrosis. Hypertens 2001 38: 635–638,  | ISI | ChemPort |
22. Hill, C, Flyvbjerg, A, Rasch, R, et al: Transforming growth factor-beta2 antibody attenuates fibrosis in the experimental diabetic rat kidney. J Endocrinol 2001 170: 647–651,  | Article | PubMed | ISI | ChemPort |
23. Yao, J, Li, LS, Zhou, H: Inhibition of fibronectin synthesis in cultured human mesangial cells by emodin. Chin J Nephrol Dial Transplant 1994 3: 349–352,
24. Yang, JW, Li, LS: Effects of rheum on renal hypertrophy and hyperfiltration on experimental diabetes in rats. Zhongguo Zhong Xi Yi Jie He Za Zhi 1993 13: 286–288,  | PubMed | ChemPort |
25. Kuo, YC, Tsai, WJ, Meng, HC, et al: Immune responses in human mesangial cells regulated by emodin from Polygonum hypoleucum ohwi. Life Sci 2001 68: 1271–1286,  | Article | PubMed | ISI | ChemPort |
26. Kuo, YC, Meng, HC, Tsai, WJ: Regulation of cell proliferation, inflammatory cytokine production and calcium mobilization in primary human T lymphocytes by emodin from Polygonum hypoleucum ohwi. Inflamm Res 2001 50: 73–82,  | Article | PubMed | ISI | ChemPort |
27. Kumar, A, Dhawan, S, Aggarwal, BB: Emodin (3-methyl-1,6,8-trihydroxyanthraquinone) inhibits TNF-induced NF-kappaB activation, IkapaB degradation, and expression of cell surface adhesion proteins in human vascular endothelial cells. Oncogene 1998 17: 913–918,  | Article | PubMed | ISI | ChemPort |
28. Chan, TC, Chang, CJ, Koonchanok, NM, Geahlen, RL: Selective inhibition of the growth of ras-transformed human bronchial epithelial cells by emodin, a protein-tyrosine kinase inhibitor. Biochem Biophys Res Commun 1993 30: 1152–1158,
29. Battistutta, R, Sarno, S, De Moliner, E, et al: The replacement of ATP by the competitive inhibitor emodin induces conformational modification in the catalytic site of protein kinase CK2. J Biol Chem 2000 275: 29618–29622,  | Article | PubMed | ISI | ChemPort |
30. Yung, S, Liu, ZH, Lai, KN, et al: Emodin ameliorates glucose-induced morphologic abnormalities and synthesis of transforming growth factor beta1 and fibronectin by human peritoneal mesothelial cells. Perit Dial Int 2001 21: S41–S47,  | PubMed | ISI |
31. Kreisberg, KI, Garoni, KN, Radnik, RA, Ayo, SH: High glucose and TGF-1 stimulates fibronectin gene expression through a cAMP response element. Kidney Int 1994 46: 1019–1024,  | PubMed | ISI | ChemPort |
32. Kreisberg, JI, Radnik, RA, Kreisberg, SH: Phosphorylation of cAMP responsive element binding protein after treatment of mesangial cells with high glucose plus TGF- or PMA. Kidney Int 1996 50: 805–810,  | PubMed | ISI | ChemPort |
33. Wang, XY, Sun, ZM: Extraction of emodin from Rheum officinale. Chin Contemp Appl Pharmacol 1997 14: 250–251,
34. Beavis, MJ, Williams, JD, Hoppe, J, Topley, N: Human peritoneal fibroblast proliferation in 3-dimensional culture: Modulation by cytokines, growth factors and peritoneal dialysis effluent. Kidney Int 1997 51: 205–215,  | PubMed | ISI | ChemPort |
35. Yung, S, Woods, A, Chan, TM, et al: Syndecan-4 up-regulation in proliferation in proliferative renal disease is related to microfilament organization. FASEB J 2001 15 (9): 1631–1633,  | PubMed | ChemPort |
36. Martin, J, Steadman, R, Knowlden, J, et al: Differential regulation of matrix metalloproteinases and their inhibitors in human glomerular epithelial cells in vitro. J Am Soc Nephrol 1998 9: 1629–1637,  | PubMed | ISI | ChemPort |
37. Phillips, AO, Steadman, R, Topley, N, Williams, JD: Elevated D-glucose concentrations modualte TGF-beta 1 synthesis by human cultured renal proximal tubular cells. The permissive role of platelet-derived growth factor. Am J Pathol 1995 147: 362–374,  | PubMed | ISI | ChemPort |
38. Yevdokimova, N, Wahab, NA, Mason, RM: Thrombospondin-1 is the key activator to TGF-1 in human mesangial cells exposed to high glucose. J Am Soc Nephrol 2001 12: 703–712,  | PubMed | ISI | ChemPort |
39. Ma, R, Kudlacek, PE, Sansom, SC: Protein kinase C participates in activation of store-operated Ca2+ channels in human glomerular mesangial cells. Am J Physiol Cell Physiol 2002 283: C1390–C1398,  | PubMed | ISI | ChemPort |
40. ABMAYR, SM, WORKMAN, JL: Preparation of nuclear and cytoplasmic extracts from mammalian cells, in Current Protocols in Molecular Biology, 1990, Vol 2, edited by AUSUBEL FM, BRENT R, KINGSTON RE, et al, New York, John Wiley and Sons, pp 12.1.1–12.1.9
41. Yung, S, Davies, M: Response of the human peritoneal mesothelial cell to injury: An in vitro model of peritoneal wound healing. Kidney Int 1998 54: 2160–2169,  | Article | PubMed | ISI | ChemPort |
42. Ginty, DD, Kornhauser, JM, Thompson, MA, et al: Regulation of CREB phosphorylation in the suprachiasmatic nucleus by light and a circadian clock. Science 1993 260: 238–241,  | PubMed | ISI | ChemPort |
43. Hjelle, JT, Golinska, BT, Waters, DC, et al: Isolation and propagation in vitro of peritoneal mesothelial cells. Perit Dial Int 1989 9: 341–347,  | PubMed | ISI | ChemPort |
44. Oh, JH, Ha, H, Yu, MR, Lee, HB: Sequential effects of high glucose on mesangial cell transforming growth factor-1 and fibronectin synthesis. Kidney Int 1998 54: 1872–1878,  | Article | PubMed | ISI | ChemPort |
45. Ayo, SH, Radnik, RA, Garoni, J, et al: High glucose causes an increase in extracellular matrix protein in cultured mesangial cells. Am J Pathol 1990 136: 1339–1348,  | PubMed | ISI | ChemPort |
46. Ziyadeh, FN, Sharma, K, Ericksen, M, Wolf, G: Stimulation of collagen gene expression and protein synthesis in murine mesangial cells by high glucose is mediated by activation of transforming growth factor-. J Clin Invest 1994 93: 536–542,  | PubMed | ISI | ChemPort |
47. Kim, SJ, Angel, P, Lafyatis, R, et al: Autoinduction of transforming growth factor beta 1 is mediated by AP-1 complex. Mol Cell Biol 1990 10: 1492–1497,  | PubMed | ISI | ChemPort |
48. Dean, DC: Expression of the fibronectin gene. Am J Respir Cell Mol Biol 1990 10: 1492–1497,
49. Lee, CQ, Yun, Y, Hoffer, JP, Habener, JF: Cyclic-AMP-responsive transcriptional activation involves interdependent phosphorylated subdomains. EMBO J 1990 9: 4455–4465,  | PubMed | ISI | ChemPort |
50. Tan, Y, Rouse, J, Zhang, A, et al: FGF and stress regulate CREB and ATF-1 via a pathway involving p38 MAP kinase and MAPKAP kinase-2. EMBO J 1996 15: 4629–4642,  | PubMed | ISI | ChemPort |
51. Sheng, M, Thompson, MA, Greenberg, ME: CREB: A Ca2+-regulated transcription factor phosphorylated by cadmodulin-dependent kinases. Science 1991 252: 1427–1430,  | PubMed | ISI | ChemPort |
52. Wahab, NA, Harper, K, Mason, RM: Expression of extracellular matrix molecules in human mesangial cells in response to prolonged hyperglycaemia. Biochem J 1996 316: 985–993,  | PubMed | ISI | ChemPort |
53. Gambaro, G, D'Angelo, A, Del Prete, D, et al: Mesangial cell proliferation in long-term streptozotocin-induced diabetes mellitus in the rat and the renoprotective activity of heparin. Am J Nephrol 1999 19: 530–534,  | Article | PubMed | ISI | ChemPort |
54. Koya, D, Jiousek, MR, Lin, Y-W, et al: Characterization of protein kinase C isoform activation on the gene expression of transforming growth factor-, extracellular matrix components, and prostanoids in the glomeruli of diabetic rats. J Clin Invest 1997 100: 115–126,  | PubMed | ISI | ChemPort |
55. Yim, H, Lee, YH, Lee, CH, Lee, SK: Emodin, an anthraquinone derivative isolated from the rhizomes of Rheum palmatum, selectively inhibits the activity of casein kinase II as a competitive inhibitor. Planta Med 1999 65: 9–13,  | Article | PubMed | ISI | ChemPort |
56. Fredenhagen, A, Mett, H, Meyer, T, et al: Protein tyrosine kinase and protein kinase C inhibition by fungal anthraquinones related to emodin. J Antibiot (Tokyo) 1995 48: 1355–1358,  | PubMed | ChemPort |

Acknowledgments

This work was supported by the Hong Kong Research Grant Council Earmarked Research Grant [7240/98M] and the University of Hong Kong School of Traditional Chinese Medicine Research Fund. The authors thank Dr. Kent-Man Chu, Dr. Kin-Wah Chu, Dr. Judy Ho, Dr. Sai Man Chu, Dr. Po Chor Tam, and their surgical and nursing team for the collection of omentum specimens. Part of this work was presented at the IX Congress of the International Society for Peritoneal Dialysis, Montreal, Canada, and the 2001 American Society of Nephrology/International Society of Nephrology Congress, and has been presented in abstract form (Perit Dial Int 2001; 21:S23 and J Am Soc Nephrol 2001; 12:722A).