Differential expression of receptors for advanced glycation end-products in peritoneal mesothelial cells exposed to glucose degradation products

Abstract

Autoclaving peritoneal dialysate fluid (PDF) degrades glucose into glucose degradation products (GDPs) that impair peritoneal mesothelial cell functions. While glycation processes leading to formation of advanced glycation end-products (AGE) were viewed commonly as being mediated by glucose present in the PDF, recent evidence indicates that certain GDPs are even more powerful inducers of AGE formation than glucose per se. In the present study, we examined the expression and modulation of AGE receptors on human peritoneal mesothelial cells (HPMC) cultured with GDPs, conventional PDF or PDF with low GDP content. HPMC cultured with GDPs differentially modulated AGE receptors (including RAGE, AGE–R1, AGE–R2 and AGE–R3) expression in a dose-dependent manner. At subtoxic concentrations, GDPs increased RAGE mRNA expression in HPMC. 2-furaldehyde (FurA), methylglyoxal (M-Glx) and 3,4-dideoxy-glucosone-3-Ene (3,4-DGE) increased the expression of AGE–R1 and RAGE, the receptors that are associated with toxic effects. These three GDPs up-regulated the AGE synthesis by cultured HPMC. In parallel, these GDPs also increased the expression of vascular endothelial growth factor (VEGF) in HPMC. PDF with lower GDP content exerted less cytotoxic effect than traditional heat-sterilized PDF. Both PDF preparations up-regulated the protein expression of RAGE and VEGF. However, the up-regulation of VEGF in HPMC following 24-h culture with conventional PDF was higher than values from HPMC cultured with PDF containing low GDP. We have demonstrated, for the first time, that in addition to RAGE, other AGE receptors including AGE–R1, AGE–R2 and AGE–R3 are expressed on HPMC. Different GDPs exert differential regulation on the expression of these receptors on HPMC. The interactions between GDPs and AGE receptors may bear biological relevance to the intraperitoneal homeostasis and membrane integrity.

INTRODUCTION

The peritoneal cavity is lined by a monolayer of mesothelial cells that provide a large surface for potential fluid movement between peritoneal capillaries and the peritoneal cavity. During chronic peritoneal dialysis, peritoneal mesothelial cells are exposed repeatedly to a non-physiological hypertonic environment. Long-term exposure to non-physiological dialysate may lead to peritoneal fibrosis and ultrafiltration failure in continuous ambulatory peritoneal dialysis (CAPD) [1,2].

d-glucose is a reactive compound in peritoneal dialysis fluid (PDF). Two pathways of glucose degradation play an important role in peritoneal mesothelial biology: (i) degradation into glucose degradation products (GDPs) during heat sterilization and storage; and (ii) formation of advanced glycation end-products (AGEs) following non-enzymatic reaction with free-amino groups of proteins [3]. Accumulation of AGEs in the peritoneal tissues of CAPD patients promotes peritoneal expression of various growth factors and subsequently deteriorates ultrafiltration capacity in CAPD [4].

Recently, several receptor systems mediating AGE–related biology have been elucidated, including a specific AGE–receptor complex composed of R1 (also termed p60) [5], R2 (also termed p90) [5] and R3 (also termed Galectin-3, Mac-2 or carbohydrate binding protein-35) [6], receptor for AGE (RAGE) [7] and scavenger receptors such as CD-36 [8] and SCR-II [9]. These AGE–receptor systems contribute to cell activation and act as an intracellular signal transducing or proinflammatory peptide leading to endothelial leakage and atherosclerosis. These systems may be regulated by diabetic factors such glucose, AGEs and reactive oxygen species [10].

In this study, we examined whether different AGE–receptor systems are present in cultured human peritoneal mesothelial cells (HPMC) and in the mesothelial lining of peritoneal tissue. We investigated whether different GDPs in the PDF would modulate the expression of these receptors and, hence, exert pathophysiological effect on the biology of the peritoneal mesothelial cells.

METHODS

Materials

Medium 199 and fetal bovine serum (FBS) were obtained from Life Technologies (Rockville, MD, USA). Consumables for electrophoresis were obtained from Bio-Rad Laboratories (Hercules, CA, USA). The GDP preparations of methylglyoxal (M-Glx), 5-hydoxymethyl-2-furaldehyde (5-HMF) and 3,4-dideoxyglucosone-3-Ene (3,4-DGE) were obtained from Gambro AB (Lund, Sweden). All other chemicals including other GDPs were obtained from Sigma (St Louis, MO, USA). The following antibodies were used to identify different AGE receptors. Anti-AGE–R1 is a rabbit polyclonal antiserum raised against a peptide mapping at the amino terminus of human OST-48 [10]. Anti-AGE–R2 is a rabbit polyclonal antiserum raised against human recombinant 80K-H [5]. Anti-AGE–R3 is a rat monoclonal antibody specific to murine Gal-3 purified from culture supernatant of hybridoma M3/38 (ATCC TIB166). Anti-RAGE is a goat polyclonal antiserum from Chemicon International (Temecula, CA, USA). These antibodies were specific for antigens of both human and mice AGE–receptors. Monoclonal mouse antihuman mesothelial cells (clone HBME-1), vimentin, polyclonal antihuman factor VIII and secondary antibodies for histochemical staining were obtained from Dako (Carpinteria, CA, USA). Monoclonal antihuman cytokeratin 18 was obtained from ICN Biochemicals (Aurora, OH, USA). Peritoneal dialysis fluid (PDF) (Gambrosol 40 and Gambrosol Trio 10) was from Gambro Lundia AB (Lund, Sweden). Each litre of electrolyte fluid is composed of 5·4 g sodium chloride, 4·6 g sodium lactate and minute quantities of calcium chloride, magnesium chloride and sodium hydroxide. Gambrosol 40 is contained in a bag with a single compartment and is sterilized by autoclave. Gambrosol Trio 10, contained in a bag with three compartments with lower GDP contents, is reconstituted immediately before use. After reconstitution, the final glucose concentration in both fluids is 83 m m (1·5%).

Morphological studies

The study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional ethics committee for studies in humans. Full informed consent was obtained from patients undergoing elective abdominal surgery for removing small pieces of omentum for experimental studies. Mouse peritoneal membrane was obtained from the parietal walls of 6-week-old male C57L6B mice. The protocol was approved by the Animal Ethic Committee of the University of Hong Kong. Four µm thick tissue sections were deparaffinized with xylene and then rehydrated through a descending gradient of ethanol. AGE–receptor expression on the peritoneal mesothelial monolayer was determined by immunohistochemical staining using specific antibodies for AGE–receptors. Briefly, the slides were incubated with 0·5% hydrogen peroxide for removal of endogenous peroxidase activity. Non-specific binding was blocked by incubation of the slides for 30 min with blocking buffer [5% normal goat serum and 3% bovine serum albumin in phosphate buffered saline (PBS)]. The sections were then incubated with anti-RAGE, anti-AGE–R1, anti-AGE–R2 or anti-AGE–R3 antibody (0·5 µg/ml) overnight. The bound rabbit anti-AGE receptor antibody was visualized as an orange-brown colour using the Dako Envision Plus System (Dako). To ensure the specificity of the staining, the following labelling controls were performed: (i) the primary antibodies were substituted with preimmune rabbit immunoglobulins; and (ii) staining was carried out without either the primary antibodies or the peroxidase-labelled polymer. Some sections were counterstained with haematoxylin before mounting.

Double immunohistochemical staining was performed on paraffin sections to confirm that the staining of AGE receptors was located in mesothelial cells using techniques described previously [11]. Briefly, sections were first stained for AGE receptor as described above. After the first labelling and colour development (brown colour for positive AGE receptor signal), sections were treated by 10 min of microwave heating in 0·01 m sodium citrate buffer (pH 6·0) at 2450 MHz and 800 W. The sections were blocked again with blocking buffer and were incubated with mouse antihuman mesothelial cell MoAb (clone HBME-1, 1 µg/ml) for 1 h. The sections were then incubated sequentially with alkaline phosphatase-conjugated goat antimouse IgG and mouse alkaline phosphatase anti-alkaline phosphatase complexes (APAAP; Dako) and then developed with 4-benzoylamino-2,5-diethoxybenzenediazonium chloride hemisalt (Fast Blue BB Salt) that gives a blue colour for positive reaction.

Culture of human peritoneal mesothelial cells

Human peritoneal mesothelial cells were isolated and characterized using procedures described previously [12]. Immunohistochemical staining was performed with monoclonal antibodies for human cytokeratin 18 and vimentin, as well as a polyclonal antibody for human factor VIII. All cells were positive for cytokeratin and vimentin (positive staining of cytokeratin 18 excludes the presence of fibroblasts [13]), but factor VIII antigen was not detected (excluding the presence of endothelial cells). The cells were maintained in medium 199 (Life Technologies) supplemented with penicillin (100 U/ml), streptomycin (100 µg/ml), l-glutatmine (2 m m), transferrin (5 µg/ml), insulin (5 µg/ml) and 10% v/v FCS. Human peritoneal mesothelial cells were incubated at 37°C in a humidified atmosphere with 5% CO₂. Once monolayers of HPMC reached confluence, the culture medium was then removed and medium 199 containing 0·1% v/v FBS were added to the cells for 48 h prior to further culture experiments. Under these conditions, the HPMC remained in a non-proliferative viable condition for up to 96 h [14]. Confluent cells were split at a ratio of 1 : 3 and all experiments performed with cells of first to second passage.

Expression of AGE receptors on cultured HPMC exposed to different GDPs

Human peritoneal mesothelial cells were grown to confluence in six-well culture plates (1 × 10⁶ cells per well) (Nalge Nunc International, Naperville, IL, USA) and exposed to different GDPs including acetaldehyde (AcA), formaldehyde (FoA), 2-furaldehyde (FurA), glyoxal (Glx), M-Glx, 5-HMF, 3-deoxyglucosone (3-DG) and 3,4-DGE at increasing concentrations (0·1, 1, 10, 100 or 1000 µg/ml) for a defined time period (24 h) at 37°C. The optimal or non-toxic dose of GDPs was determined by cell viability measured by a 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay (R&D Systems, Minneapolis, MN, USA). Briefly, growth-arrested HPMC were seeded in 96-well plates (1 × 10⁵ cells per well) before being exposed to GDPs for 48 h. MTT reagent was then added and incubated at 37°C until purple precipitate was clearly visible under an inverted microscope. Detergent reagent (2% Triton X-100) was finally added and incubated in the dark at 37°C until the purple precipitate solubilized. The absorbance was measured using 550 nm as the primary wavelength and 650 nm as the reference wavelength. In order to study whether synthesis of AGE receptors was inducible with extended exposure to different GDPs, similar experiments were performed with cells cultured for exposure to non-toxic concentrations of GDPs, AGE–BSA (10 µg/ml), glucose (30 m m) or H₂O₂ (50 µm) for defined time periods (0, 4, 8 or 24 h) at 37°C. The gene expression on and protein synthesis by the HPMC was determined by reverse transcription-polymerase chain reaction (RT-PCR) and immunoblotting as described below.

AGE–receptor gene expression by cultured HPMC

Total RNA was extracted from HMPC using a Qiagen RNeasy kit (Qiagen GmbH, Hilden, Germany). All RNA samples prepared from 1 × 10⁶ cells were dissolved in 20 µl DEPC-H₂0 and were stored at −70°C until RT-PCR. The quality of RNA was checked by formaldehyde agarose; gel electrophoresis RNA was quantified by absorbance at 260 nm. Five µg total RNA was reverse transcribed to cDNA with Superscript II reverse transcriptase (Life Technologies, Paisley, UK) in a 20 µl reaction mixture containing 160 ng oligo(dT)₁₂₋₁₈, 500 µm of each dNTP and 40 U RNase inhibitor for 10 min at 37°C, 60 min at 42°C and 5 min at 99°C. cDNA was stored at −20°C until further use.

Specific primers for RAGE, AGE–R1, AGE–R2, AGE–R3, CD36, vascular endothelial growth factor (VEGF) and GAPDH (glyceraldehyde-3-phosphate-dehydrogenase) were designed from known GenBank accession numbers (RAGE BC020669, AGE–R1 XM_001844, AGE–R2 NM_002743, AGE–R3 BC001120, CD36 S67532, VEGF XM_004512 and GAPDH AF261085). The sequences of each primer were as follows: (i) RAGE, sense primer 5′-TGTGTGGCCACCCATTCCAG and antisense primer 5′-GCCCTCCAGTACTACTCTCG; (ii) AGE–R1, 5′-TGGAGACCATCAGTGCCTTTAT and antisense primer 5′-AGTTCATA-GTTGCCTGTCTGGG; (iii) AGE–R2, 5′-GGCCGTAA-GGAGAGAGAGTCCC and antisense primer 5′-AGGTGGCGTCTGTCTGTGTGTC; (iv) AGE–R3, 5′- AGCACCTGC-ACCTGGAGTCTA and antisense primer 5′- TCAAATGGGAAAAGGGAAAACCG-ACTGT; (v) CD36, 5′-CAGGAAGTGATGAACAGC and antisense primer 5′-TATGTGTCGATTATGGCAACT; VEGF, 5′-GGCAGAAT CATCACG-AAGTGGTG and antisense primer 5′-CTGTAG GAAGCTCATCTCTCC; (vi) GAPDH, sense primer 5′-TGAAGG-TCGGAGTCAACGGATTTGGT and antisense primers 5′-CATGTGGGCCA-TGAGGTCCACCAC. The polymerase chain reaction (PCR) was carried out in the following profiles: first cycle, 94°C for 3 min, 58°C for 1 min, 72°C for 1 min; second cycle to 30 cycles, 95°C for 45 s, 58°C for 40 s, 72°C for 45 s. The final cycle was 94°C for 1 min and 72°C for 10 min. The PCR reaction from AGE receptors or VEGF and control (GAPDH) amplicons were mixed and separated by 1·5% wt/vol agarose gels, stained with ethidium bromide and the gel image was captured and analysed using the Gel Doc 1000 Gel Documentation System and Quantity One software (Bio-Rad Laboratories Ltd, Hercules, CA, USA). The result of AGE–receptor or VEGF mRNA yield was expressed as a ratio of the specific amplicon to GAPDH amplicon. Precautions were taken to ensure the validity of the results, as described previously [11].

Western blot analysis

Homogenized HPMC incubated with different GDPs were spun at 1000 g for 10 min at 4°C to remove nuclei and cell debris. The cell extracts were pelleted at 150 000 g for 60 min and dissolved in protein extraction buffer containing protease inhibitor cocktails, and the protein concentrations were measured by a modified Lowry method using bovine serum albumin (BSA) as standard (DC protein assay kit, Bio-Rad). Five µg of total protein from the extract were electrophoresed through a 15% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) gel before transferring to a polyvinylidene difluoride (PVDF) membrane. After blocking for 1 h at room temperature in blocking buffer (1% gelatin in PBS with 0·05% Tween 20), the membrane was incubated for 16 h with anti-AGE receptor antibodies (1 : 1000) in PBS-Tween. The membrane was washed and incubated for 2 h at room temperature with a peroxidase-labelled goat anti-rabbit immunoglobulin, peroxidase-labelled mouse anti-goat immunoglobulin or peroxidase-labelled rabbit anti-rat immunoglobulin (Dako). After further washing, the membrane was detected with ECL chemiluminescence (Amersham Pharmacia Biotech, Arlington, IL, USA). The Western blotting images were scanned on a flatbed scanner and the density of the bands was quantified using ImageQuant software (Molecular Dynamic, Sunnyvale, CA, USA). Densitometry results were reported as average arbitrary integrated values (units).

Direct modulatory effect of GDPs on the expression of AGE receptors on cultured HPMC

While glycation processes were viewed commonly as being mediated by glucose present in the PDF, recent evidence indicates that certain GDPs are even more powerful inducers of AGE formation than glucose per se [15,16]. In order to determine whether the GDPs exert any direct modulatory effect on the expression of AGE receptor in addition to that induced indirectly following AGE formation from GDPs, parallel experiments were conducted with the addition of aminoguanidine at a dose that maximally diminishes the formation of AGE but remains non-toxic to the cultured cells. The difference in values represented the direct effect of GDPs on AGE–receptor expression independently of AGE formation.

Measurement of AGE and VEGF in supernatants of cultured HPMC

The supernatant AGE concentration was detected by a non-competitive enzyme-linked immunosorbent assay [17]. All samples were carried out in the same batch to avoid interbatch variation. Briefly, each well of a 96-well microtitre plate was coated with 100 µl of samples diluted at 1 : 2 with 0·05 m carbonate buffer (pH 9·6) overnight at 4°C and washed three times with 0·01 m PBS containing 0·05% Tween 20 (PBS-Tween). Each well was then blocked for non-specific reaction with 200 µl of 2% skimmed milk, 0·5% BSA and 0·5% gelatin in 0·05 m carbonate buffer (pH 9·6) for 1 h and washed three times with PBS-Tween. Each well was incubated for 1 h with 50 µl of an anti-AGE monoclonal antibody (clone 6D12; Wako, Osaka, Japan) diluted 1 : 1000 with PBS-Tween containing 0·3% bovine serum albumin (BSA). The wells were then washed three times with PBS-Tween and incubated with 100 µl of alkaline phosphatase-labelled goat antimouse IgG antibody (Dako) for 1 h. After washing three times with PBS-Tween, the wells were reacted with 100 µl of p-nitrophenylphosphate as a colorimetric substrate for 1 h. The absorbance at 405 nm was read using an enzyme-linked immunosorbent assay (ELISA) plate reader. The supernatant concentration of VEGF was measured by an ELISA (R&D Systems) with a detection limit of 50 pg/ml and a coefficient of variation of 8·5%.

Statistical analysis

All data (from five experiments) were expressed as means ±standard deviation (s.d.). Inter-group differences for continuous variables were assessed by the unpaired t-test except in GDP concentrations exceeding the in vivo relevance. The gene expression or protein synthesis of AGE receptors in cultured HPMC following exposure to different concentrations of GDPs were analysed with multivariate anova for repeated measures. Statistical analysis was performed using statistical software (Statview, SAS Intelligence, Cary, NC, USA). All P-values quoted are two-tailed and the significance is defined as P < 0·05.

RESULTS

Expression of AGE receptors in cultured HPMC and peritoneal membrane

Figure 1a depicts results of PCR for RAGE, AGE–R1, AGE–R2, AGE–R3 and CD36. Constitutive expression of mRNA for RAGE, AGE–R1, AGE–R2 and AGE–R3 was detected in cultured HPMC from three separate donors. PCR product for CD36 (another receptor for AGE that presented on macrophage and human mesangial cells) was not detected in these cultured HPMC. However, a band of PCR product 470 kDa in size was detected in U937 cells (monocytic cell line) and cultured human mesangial cells that served as RT-PCR positive controls for CD36.

Fig. 1.

(a) Expression of RAGE mRNA (PCR product: 310 bp), AGE–R1 mRNA (PCR product: 535 bp), AGE–R2 mRNA (PCR product: 442 bp) and AGE–R3 (PCR product: 519 bp) in RNA from three separate preparations of human peritoneal mesothelial cells (HPMC). CD36 mRNA (PCR product: 470 bp) is detected in human mesangial cells (HMC) and a monocytic cell line (U937) but not in HPMC. (b) Higher magnification examination of the human peritoneum staining for AGE receptors which was located in peritoneal mesothelial cells (arrow) (orange-brown colour and magnification ×400). (c) Higher magnification examination of the mouse peritoneum staining for AGE receptors which was located in peritoneal mesothelial cells (arrow) (orange-brown colour and magnification ×400). (d) Double staining of human peritoneum for AGE receptors (brown) and mesothelial cells (blue) (magnification ×1000). To view this figure in colour, see the online article on http://www.blackwell-synergy.com.

To localize the in vivo site of AGE synthesis within the peritoneal membrane, paraffin sections of omental tissue were stained immunohistochemically using specific antibodies for AGE receptors. RAGE, AGE–R1, AGE–R2 and AGE–R3 were detected in the mesothelial monolayer of human and mouse peritoneum (Fig. 1b,c). The localization of AGE receptor protein in peritoneal mesothelial cells was confirmed by double immunohistochemical staining (Fig. 1d).

Viability of HPMC cultured with GDPs

The cells were viable at a concentration of 10 µg/ml for all tested GDPs measured by proliferation in HPMC following incubation of different GDPs. However, the MTT conversion fell significantly (less than 20% of that of the medium control) when the GDP concentration was raised to 100 µg/ml with maximum toxicity in FurA, FoA and 3,4-DGE.

Differential effect of GDPs on expression of AGE receptors in HPMC

After establishing the maximal non-toxic dosage of GDPs, GDP was added to the culture medium in which cells were incubated for different time periods (4 h for RNA and 24 h for protein synthesis). At a concentration of 1 µg/ml (for FoA, FurA and M-Glx) or 10 µg/ml (for AcA, 5-HMF and 3,4-DGE), these GDPs up-regulated the gene expression and protein synthesis of RAGE in cultured HMPC by 17·5%, 36·9%, 26·7%, 30·5%, 62·4% and 63·6%, respectively (for protein) (Table 1). Similarly, at these concentrations FurA, 3,4-DGE and 3-DG up-regulated the gene expression and protein synthesis of AGE–R1 in cultured HMPC by 31·3%, 26·6% and 22·5%, respectively (for protein) (Table 1). 2-Furaldehyde and Glx increased the gene expression and protein synthesis of AGE–R2 by 95% and 116%, respectively (for protein) (data not shown). At a concentration of 1 µg/ml, only FurA and Glx up-regulated the gene expression and protein synthesis of AGE–R3 in HMPC by 38·5% and 34%, respectively (for protein) (data not shown). Glucose (30 m m) or H₂O₂ (50 µm) did not affect the gene and protein expression of all AGE receptors. AGE–BSA (10 µg/ml) increased the gene expression and protein synthesis of RAGE by 34% and 59·6% (Table 1). Gene and protein expression of AGE–R1 was increased by 27·7% and 26·7% under the same concentration of AGE–BSA. However, gene and protein expression of AGE–R2 and AGE–R3 were not altered by AGE–BSA (data not shown).

Table 1.

Gene and protein expression of RAGE and AGE–R1 in HPMC incubated with different GDPs, PDF, glucose and AGE

	RAGE		AGE–R1
	mRNA‡	Protein§	mRNA‡	Protein§
Medium	0·838 ± 0·011	47·34 ± 4·27	0·907 ± 0·017	59·73 ± 2·32
AcA	1·030 ± 0·084*	55·61 ± 3·80†	0·899 ± 0·020	60·65 ± 3·71
FoA	0·993 ± 0·041*	61·79 ± 4·87*	0·774 ± 0·034	54·57 ± 3·63
FurA	0·971 ± 0·045*	73·30 ± 4·85*	1·019 ± 0·069†	78·46 ± 7·23*
Glx	0·894 ± 0·031	67·70 ± 4·10*	0·914 ± 0·033	64·32 ± 4·00
M-Glx	0·929 ± 0·048*	77·50 ± 3·95**	0·918 ± 0·043	67·39 ± 4·70†
5-HMF	0·928 ± 0·026†	64·83 ± 5·07*	0·920 ± 0·034	66·72 ± 5·44
3,4-DGE	0·998 ± 0·053*	59·99 ± 4·43*	0·977 ± 0·021†	75·61 ± 4·58*
3-DG	0·913 ± 0·038	57·29 ± 2·65*	0·947 ± 0·012†	73·18 ± 6·13*
Medium + 10%FBS	0·860 ± 0·029	49·83 ± 6·96	0·903 ± 0·042	63·44 ± 3·18
Gambrosol Trio + 10%FBS	0·909 ± 0·046	58·57 ± 3·93*	0·906 ± 0·039	61·77 ± 3·48
Gambrosol 40 + 10%FBS	0·944 ± 0·043†	64·62 ± 3·76*	0·910 ± 0·048	62·08 ± 3·03
Glucose	0·845 ± 0·025	48·25 ± 5·69	0·912 ± 0·039	60·14 ± 3·28
AGE–BSA	1·124 ± 0·079*	72·56 ± 5·12*	1·158 ± 0·035*	75·65 ± 5·38*
H2O2	0·814 ± 0·025	45·26 ± 5·89	0·899 ± 0·047	56·25 ± 6·89

^‡mRNA expression of RAGE and AGE–R1 in HPMC incubated with different GDPs (1 µg/ml for FoA, FurA, Glx or M-Glx and 10 µg/ml for AcA, 5-HMF, 3,4-DGE or 3-DG), PDF, 30 mm glucose, 10 µg/ml AGE–BSA or 50 µm H₂O₂ for 4 h. Results were expressed as target mRNA/GAPDH mRNA ratio.

^§Protein expression of RAGE and AGE–R1 in HPMC incubated with different GDPs (1 µg/ml for FoA, FurA, Glx or M-Glx and 10 µg/ml for AcA, 5-HMF, 3,4-DGE or 3-DG), PDF, 30 mm glucose, 10 µg/ml AGE–BSA or 50 µm H₂O₂ for 24 h. Results were expressed as arbitrary unit. All results represent the mean ± s.d. of five separate experiments.

^†P < 0·05;

^*P < 0·01;

^**P < 0·0001 versus plain medium control.

Differential effect of GDPs on release of AGE and VEGF in HPMC

To confirm that the GDPs exert pathophysiological effects on HPMC, the synthesis of AGE and the release of VEGF were determined in cells incubated with an optimal concentration of GDPs. The supernatant concentrations of AGE are shown in Fig. 2. Advanced glycation end-products up-regulated the synthesis of AGE by cultured HPMC in a autocrine fashion. Different GDPs including AcA, FoA, FurA, M-Glx, 3,4-DGE and 3-DG induced AGE formation by the HPMC with quantities equivalent to 12%, 23%, 29%, 40%, 46% and 17%, respectively, of that induced by AGE at a concentration of 10 µg/ml. Methylglyoxal and 3,4-DGE up-regulated the gene expression of VEGF in cultured HPMC (Fig. 3a). Incubating HPMC with FoA, FurA, M-Glx and 3,4-DGE induced two-, three- and fourfold increases in VEGF synthesis, respectively, when compared with cells incubated in culture medium alone (Fig. 3b).

Fig. 2.

Supernatant concentration of AGE in cultured HPMC incubated with different GDPs (1 µg/ml for FoA, FurA, Glx or M-Glx and 10 µg/ml for AcA, 5-HMF, 3,4-DGE or 3-DG), AGE (10 µg/ml), BSA (10 µg/ml) or PDF for 24 h. All GDPs (except Glx) increased AGE formation (P < 0·01). HPMC incubated with Gambrosol 40 and 10% FBS synthesize more AGE than HPMC incubated with BSA alone (P = 0·0019) or with culture medium and 10% FBS (P = 0·008). The results represent the mean ± s.d. of five separate experiments. **P < 0·0001; *P < 0·01 versus plain medium control.

Fig. 3.

(a) Gene expression of VEGF in HPMC incubated with different GDPs (1 µg/ml for FoA, FurA, Glx or M-Glx and 10 µg/ml for AcA, 5-HMF, 3,4-DGE or 3-DG) or PDF for 4 h. Up-regulation of VEGF mRNA in cultured HMC incubated with M-Glx and 3,4-DGE (P < 0·05 and < 0·01, respectively). VEGF mRNA was increased in HPMC incubated with either conventional PDF or PDF with low GDP content. The results represent the mean ± s.d. of five separate experiments. (b) Protein synthesis of VEGF in HPMC incubated with different GDPs (1 µg/ml for FoA, FurA, Glx or M-Glx and 10 µg/ml for AcA, 5-HMF, 3,4-DGE or 3-DG) or PDF for 24 h. Up-regulation of VEGF synthesis in cultured HMC incubated with FoA, FurA, M-Glux and 3,4-DGE (P < 0·01). VEGF synthesis was increased in HPMC incubated with either conventional PDF or PDF with low GDP content and the VEGF synthesis was significantly lower in the latter (P = 0·003). The results represent the mean ± s.d. of five separate experiments. **P < 0·0001; *P < 0·01; ♯P < 0·05 versus plain medium control..

Pathophysiological difference between traditional PDF and PDF with low GDP contents

We then examined whether PDF with lower GDP content exerted less cytotoxic effect on HPMC by conducting cell culture experiments in traditional heat-sterilized PDF (Gambrosol 40) or non-heat-sterilized PDF with lower GDP content (Gambrosol Trio). Compared with HPMC cultured in M199 medium, cells cultured in either PDF had a lower proliferative index or decreased cell viability due to the non-physiological nature of PDF (Fig. 4). As reported previously [18], adding FBS to the medium or PDF with low GDP content improved cell viability in a dose-dependent manner. The beneficial effect of FBS supplement was not observed in traditional PDF except when the cells were cultured in PDF diluted with equal volume of medium with FBS at a final concentration of 10% v/v. At similar concentrations of FBS supplement, cell viability was significantly higher in HPMC cultured with low-GDP PDF (Gambrosol Trio) than with traditional PDF (Gambrosol 40).

Fig. 4.

Cell proliferation or viability measured by MTT conversion of HPMC cultured with different preparations of PDF or medium for 24 h 10%. †PD fluid diluted with equal volume of medium with FBS at a final concentration of 10% v/v. Cell viability improved when HPMC were cultured in PDF with low GDP content when compared with cells cultured in conventional PDF when the PDF were supplemented with 0, 5 or 10% FBS (P = 0·001). The results represent the mean ± s.d. of five separate experiments. *P < 0·01; **P < 0·0001 versus plain medium control.

HPMC incubated in either PDF supplemented with 10% FBS exhibited an enhanced synthesis of RAGE when compared with cells incubated with medium alone or with medium supplemented with equal amounts of FBS (P < 0·001) (Table 1). The up-regulatory effect on RAGE expression was apparently higher in HMPC cultured in traditional PDF than in PDF with lower GDP content, but the difference failed to reach statistical difference (P = 0·068). A significant modulatory effect by either PDF was not observed on the expression of AGE–R1 (Table 1), AGE–R2 or AGE–R3 (data not shown). HPMC incubated in either PDF supplemented with 10% FBS increased the synthesis of AGE when compared with cells incubated with medium alone or with medium supplemented with equal amounts of FBS (P < 0·01) (Fig. 2). There was apparently less AGE formation in HPMC cultured with PDF with low GDP content compared with conventional PDF, although the difference just failed to reach statistical significance. HPMC incubated in either PDF supplemented with 10% FBS exhibited an enhanced synthesis of VEGF when compared with cells incubated with medium alone or with medium supplemented with equal amounts of FBS (P < 0·001) (Fig. 3). There was an increased release of VEGF from HPMC cultured in traditional PDF compared with cells cultured in PDF with lower GDP content (P = 0·003).

Direct modulatory effect of GDPs on the expression of AGE receptors on cultured HPMC

The effects of GDPs on the expression of AGE receptors were studied indirectly by conducting parallel experiments with the addition of aminoguanidine at a dose (50 µm) that maximally diminished the formation of AGE. Our preliminary experiments showed that 100 µm aminoguanidine was toxic to cultured HPMC with markedly decreased cell viability. Based on the studies on cell viability, expression of RAGE and AGE–R1 and VEGF synthesis, FurA, M-Glx and 3,4-DGE were selected as GDPs for further studies. With the addition of (50 µm) aminoguanidine, AGE synthesis by cultured HPMC induced by FurA, M-Glx and 3,4-DGE was reduced by 30·8%, 28·4% and 29·1%, respectively (Fig. 5a). Simultaneously, RAGE expression was reduced only by 11·0%, 16·8% and 13·8%, respectively (Fig. 5b). Similarly, VEGF synthesis was decreased by 19·1%, 23·2% and 22·1%, respectively (Fig. 5c). The difference in the values represented the direct stimulatory of individual GDP independent of AGE formed by glycation of GDPs.

Fig. 5.

(a) Supernatant concentration of AGE in cultured HPMC incubated with three different GDPs (1 µg/ml for FurA or M-Glx and 10 µg/ml for 3,4-DGE) in the presence or absence of aminoguanidine (500 µm). Aminoguanidine (AG) significantly reduced the GDP-induced AGE formation in HPMC. The results represent the mean ± s.d. of five separate experiments. (b) RAGE expression in cultured HPMC incubated with three different GDPs (1 µg/ml for FurA or M-Glx and 10 µg/ml for 3,4-DGE) in the presence or absence of aminoguanidine (500 µm). Aminoguanidine (AG) significantly reduced the GDP-induced RAGE expression in HPMC but the percentages of reduction were lower than that of AGE in the supernatant. (c) Supernatant concentration of VEGF in cultured HPMC incubated with three different GDPs (1 µg/ml for FurA or M-Glx and 10 µg/ml for 3,4-DGE) in the presence or absence of aminoguanidine (500 µm). Aminoguanidine (AG) significantly reduced the GDP-induced VEGF synthesis in HPMC but the percentages of reduction were lower than that of AGE in the supernatant.

DISCUSSION

Mesothelial cells (MCs) play an important role in regulating the inflammatory response in the peritoneal cavity by producing proinflammatory cytokines and chemoattractants, such as interleukin (IL)-1β, IL-6, IL-8 and monocyte chemoattractant protein-1 (MCP-1). By secreting these chemokines at the apical surface, MCs contribute to the recruitment of leucocytes following the expression of adhesion molecules [19].

Conventional PDF makes use of the osmotic gradient generated by glucose. These fluids contain GDPs that are formed primarily during heat sterilization [20]. Several of these products have been identified and quantified [21] and, recently, 3,4-DGE has been found to be the main candidate responsible for the clinical biocompatibility caused by conventional PDF [22]. Evidence has suggested that the mesothelial toxicity of PDF is related primarily to GDPs and not to glucose per se [23]. High glucose levels do not stimulate the release of VEGF from cultured HPMC but instead AGE increase VEGF mRNA expression and antigen secretion in a dose- and time-dependent manner [24]. In vitro exposure of MCs to a mixture of isolated GDPs caused decreased proliferation and a reduced IL-6 production after IL-1β stimulation [25]. Welton and coworkers [26] observed enhanced cytotoxic damage in MCs exposed to M-Glx or 3-DG with an increased proinflammatory response, evidenced by increased vascular cell adhesion molecule-1 expression and elevated production of IL-6 and IL-8. Paradoxically, M-Glx and 3-DG can induce apoptotic and antiproliferative effects in MCs through elevation of intracellular oxidative stress [26]. Exposure to GDP may decrease protein synthesis by HPMC and hence impairs the healing process [27]. In addition, M-Glx was found recently to stimulate the production of VEGF that enhances vascular permeability and angiogenesis [28]. In addition to direct cytotoxic effects of these GDPs, the GDP-induced formation of AGEs may also contribute to toxicity, and certain GDPs (in particular M-Glx and 3-DG) are even more powerful inducers of AGE formation than glucose per se [26,29].

As VEGF and carboxymethyl lysine (formed from GDPs) co-localized immunohistochemically in the mesothelial layer and vascular walls of the peritoneal membrane of chronic peritoneal dialysis patients, it was suggested that GDPs might be involved directly in peritoneal neoangiogenesis [28]. AGE formed by GDPs may also be responsible for thickening and fibrosis of the peritoneal membrane, leading to long-term complications during peritoneal dialysis, such as ultrafiltration failure. What remains uncertain is the mechanism(s) by which GDPs exert their toxicity and proinflammatory effects on MCs. Animal studies showing co-localization of AGEs and RAGE in sites of diabetic microvascular injury suggests that this ligand–receptor interaction may represent an important mechanism in the genesis of diabetic complication [30]. Therefore, we hypothesized that GDPs, directly or indirectly, mediate their pathophysiological effect through up-regulation of different AGE receptors in MCs. AGE receptors of interest include RAGE, AGE–receptor complexes composed of R1, R2 and R3 and scavenger receptors such as CD36. The characterization of these receptors has been summarized in recent publications [31,32].

We have demonstrated that RAGE, AGE–R1, AGE–R2 and AGE–R3 are present in HPMC constitutively by both cell culture and histological studies. In contrast with the findings of Nakamura et al. [33], who detected an AGE receptor on macrophage in omental tissue, we failed to demonstrate any gene expression for CD36 on cultured HPMC. Notably, GDPs with maximal cytotoxic effect (3·4-DGE and FurA) enhanced AGE–R1, AGE–R2, AGE–R3 (for FurA) and RAGE (for 3,4-DGE) gene expression in a dose-dependent manner. At a concentration of 10 µg/ml (the maximal non-toxic dosage), 3,4-DGE and FurA up-regulated the gene expression and protein synthesis of RAGE and AGE–R1 in cultured HMPC by 27–62% above the baseline value with culture medium alone. Following selective up-regulation of these AGE receptors by individual GDPs, a parallel increase in VEGF synthesis was observed in cultured HPMC. Not only we had confirmed that M-Glx and 3-DG induce more AGE formation than glucose per se, we found other GDPs (FoA, FurA and 3,4-DGE) are more powerful promoters of AGE formation than 3-DG and comparable to M-Glx with quantities equivalent to 23%, 29% and 46%, respectively, of that induced by AGE at a concentration of 10 µg/ml.

In the present study, we assessed some of the effects of short-term exposure of various GDPs on the biology of HPMC. Although the concentrations of GDPs used in our cell culture experiments were comparable to or slightly higher than that found in conventional PDF (Table 2), we feel that our in vitro findings still contribute to understanding of the proinflammatory condition within the peritoneum of CAPD patients. It should be noted that these observations were made after brief exposure to a single GDP, while in maintenance peritoneal dialysis HPMC are exposed to lower concentrations of multiple GDPs, yet in continuous manner. Furthermore, carbonyl stress from uraemic circulation may also contribute to cellular changes in the peritoneal cavity [34]. Hence, the use of a PDF of lower GDP content is attractive and less injurious to the peritoneal lining. Recent in vitro studies revealed that HPMC cultured in conventional PDF have decreased cell viability, reduced ability to generate IL-6, increased VEGF secretion and enhanced formation of procollagen III N-terminal peptide compared with HPMC cultured in PDF with a lower GDP content [22,35]. In the present study, we confirmed that HPMC cultured in PDF with low GDP content had better cell viability than cells cultured in conventional PDF. This was associated with a smaller release of VEGF. There was an apparent up-regulatory effect on RAGE expression in HMPC cultured in traditional PDF than in PDF with lower GDP content but the difference just failed to reach statistical difference. These in vitro studies by us and others strongly support the contributory role of GDPs in the toxicity of PDF to MCs. A low GDP content improves the biocompatibility of PDF with HPMC.

Table 2.

Concentration of GDPs in PDF and in culture medium

GDP	MW g/mol	Concentration in conventional PDF (µm)**	Concentration (µm) in cell culture experiments at 1a or 10bµg/ml	Concentration giving < 50% cell viability in cell culture experiments
AcA	44·05	120–420	227b	2270
FoA	30·03	6–15	33·3a	3330
FurA	96·09	0·05–2	10·4a	1040
Glx	58·04	3–14	17·2a	1720
M-Glx	72·08	2–23	13·9a	1390
5-HMF	126·1	6–30	78b	780
3-DG	162·1	118–154	61b	610
3,4-DGE	144·1	9–22	69b	690

^**Data adapted from [22] and [26]. Abbreviations: GDPs = glucose degradation products, PDF = peritoneal dialysis fluid, AcA = acetaldehyde, formaldehyde = FoA, FurA = 2-furaldehyde, Glx = glyoxal, methylglyoxal = M-Glx, 5-hydoxymethyl-2-furaldehyde = 5-HMF, 3,4-dideoxyglucosone-3-Ene = 3,4-DGE, 3-deoxyglucosone = 3-DG.

The action of different GDPs in the toxicity of PDF to MCs is mediated through the formation of AGE. We explored whether GDPs exerted direct cytotoxic effect independently of AGE formation by conducting parallel experiments with addition of aminoguanidine at a subtoxic dose. The findings that reduction of VEGF release and RAGE expression are consistently lower than expected values calculated from the suppression of AGE formation lead us to conclude that GDPs can exert a direct cytotoxic effect on MCs independent of AGE formation.

In summary, we have demonstrated for the first time gene and protein of AGE receptors in HPMC. Glucose degradation products up-regulate these receptors differentially directly or indirectly through AGE formation. Although the exact pathophysiological sequel to peritoneal membrane property in long-term CAPD remains to be elucidated, our findings may pave the way for future research into the mechanisms and prevention of ultrafiltration failure in CAPD.