Abstract
← Background
← Methods
← Results
← Conclusions
Peritoneal dialysis (PD) has been used for about two decades as a renal replacement therapy for treating patients with end-stage renal disease. Preservation of the peritoneal membrane as a dialyzing membrane is critical for this treatment. The peritoneum of patients on long-term PD has characteristic alterations, with mesothelial cell (MC) loss, deposition of collagen, and neovascularization as main events (1-3), ultimately leading to a loss of ultrafiltration (4).
Mesothelial cells play an important role in regulating the inflammatory response in the peritoneal cavity by producing proinflammatory cytokines (5) and chemoattractants, such as interleukin (IL)-8 and monocyte chemoattractant protein-1 (MCP-1) (6,7). By secreting these chemokines at the apical surface, MCs contribute to the recruitment of leukocytes (8). Moreover, the process of leukocyte recruitment is also regulated by the expression of adhesion molecules. Mesothelial cells are shown to express intercellular adhesion molecule-1 (ICAM-1), CD44, and, upon stimulation, vascular cell adhesion molecule-1 (VCAM-1) (9,10). Migration of leukocytes across MCs cultured on transwell filters has been showed to be dependent on ICAM-1 (11,12).
Conventional PD fluids make use of the osmotic capacity of glucose. These fluids contain glucose degradation products (GDPs) that are formed during heat sterilization (13). Up to now, several of these products have been identified and quantified, such as 5-hydroxymethylfurfural (5-HMF) and other small molecular weight aldehydes (14), as well as glyoxal, methylglyoxal (MGO), and 3-deoxyglucosone (3-DG) (15,16). In vitro studies comparing heat- with filter-sterilized fluids, the latter containing no or very low levels of GDPs (15), indicated a cytotoxic effect of GDPs (17,18). Furthermore, heat-sterilized PD fluid caused a reduction in cell function, as assessed by bacterial defense capacity and cytokine release by peripheral blood cells (19) and peritoneal cells (17,18,20). In vitro exposure of MCs to a mixture of isolated GDPs caused decreased proliferation and a reduced IL-6 production after IL-1β stimulation (17). Furthermore, clinical and animal studies using multicompartment bags containing fewer GDPs show promising results in reducing the adverse effects of PD fluid (21,22).
In addition to direct cytotoxic effects of these GDPs, the GDP-induced formation of advanced glycation end-products (AGEs) may also contribute to toxicity. Both MGO and 3-DG are reported to be important intermediate products of glycation (23,24) and accelerators of the process of AGE formation (15,25,26). In addition to these harmful effects of constituents of PD fluid, carbonyl stress related to uremia in patients with end-stage renal failure contributes to cell injury and AGE formation (27). AGEs are detected in the peritoneal membrane and correlate with the duration of PD treatment (28,29). AGEs may be responsible for thickening and fibrosis of the peritoneal membrane, leading to long-term complications during PD, such as ultrafiltration failure.
In the present study, we evaluated the effects of two GDPs, MGO and 3-DG, on MC biology, with emphasis on some immune parameters. It appeared that a single dose MGO and 3-DG dose-dependently decreased MC proliferation, but did not impair protein synthesis. Furthermore, both MGO and 3-DG increased VCAM-1 expression, and MGO increased the production of IL-6 and IL-8. These findings suggest that these GDPs signal to MCs to mount a proinflammatory response.
Materials and Methods
Reagents
Methylglyoxal was prepared from 1,1-dimethoxyacetone (Sigma Chemical Co., St. Louis, Missouri, USA) as previously described (30). 3-Deoxyglucosone was prepared as previously described (31) and purified in a Kieselgel column (Merck, Darmstadt, Germany), with pentane/ethylacetate/methanol (2:1:1) as the solvent system (15). Tumor necrosis factor alpha (TNFα) was obtained from Endogen (Woburn, Massachusetts, USA). Antibodies against cytokeratin (clone MNF116) were obtained from Dako A/S (Glostrup, Denmark), against VCAM-1 and CD44 from Pharmingen (San Diego, California, USA), and against ICAM-1 from CLB (Amsterdam, The Netherlands). Cycloheximide was obtained from Sigma and used at a concentration of 50 μmol/L.
Cell Culture
Mesothelial cells were isolated from biopsies of human omentum, obtained during abdominal surgery, as described earlier (6,32), and cultured in M199 medium supplemented with Hanks balanced salt solution, 10% vol/vol heat-inactivated fetal calf serum (FCS), 50 U/mL penicillin, 50 μg/mL streptomycin, and 2 g/L NaHCO3. Medium was refreshed after 2 days of culture. When confluence was reached, cells were passed and cultured on fibronectin-coated culture flasks. All experiments were performed using cells in the second passage. Purity of MCs was checked by phase-contrast microscopy and by immunocytochemistry with antibodies against cytokeratins.
Analysis of MGO and 3-DG in Culture
Cells were incubated with 50 μmol/L MGO and 100 μmol/L 3-DG in complete M199 medium or in M199 with a low serum (0.1%) concentration. As a control, medium containing equal concentrations of MGO and 3-DG was incubated in the absence of cells. At time points t = 0, 1, 4, 6, and 24 hours, supernatant was removed and directly mixed with an equal volume of 1.2 mol/L perchloric acid. The amount of MGO and 3-DG was determined by reverse-phase HPLC, with gradient elution after derivatization to their respective dimethoxy-quinoxaline derivatives, as previously described (15).
Immunocytochemistry for MGO adducts in MCs
We recently developed a specific mouse monoclonal antibody against MGO adducts using MGO-modified keyhole limpet hemocyanin (KLH). (The generation and characterization of this monoclonal antibody will be described in full detail in a separate report.) This antibody is specific for MGO-modified proteins. It clearly distinguishes MGO-modified proteins from native protein, as demonstrated by Western blot (see Figure 1). In addition, we found a strong and time-dependent cross-reactivity with dihydroxyacetone- and glyceraldehyde-modified proteins, intermediates known to be precursors of MGO.
Detection of methylglyoxal (MGO)-modified human serum albumin (HSA) by Western blot. HSA was incubated for various times with 10 mmol/L MGO. Immunostaining with a recently developed mouse monoclonal against MGO adducts showed generation of a specific product.
For immunocytochemistry, MCs were cultured on LabTek chamber glass slides (Nalgene Nunc International, Naperville, Illinois, USA) coated with fibronectin. After reaching confluence, MCs were incubated in complete medium, with or without 500 μmol/L MGO, for 6 hours at 37°C in a humidified atmosphere. Cells were fixed with 80% acetone and incubated for 45 minutes at room temperature with anti-MGO adduct monoclonal antibody diluted in phosphate-buffered saline (PBS) with 0.1% bovine serum albumin (BSA) (PBS/BSA). After three washes with PBS/BSA, cells were incubated with alkaline phosphatase-conjugated goat anti-mouse immunoglobulins (Southern Biotechnology Associates, Birmingham, Alabama, USA) for 45 minutes at room temperature. Alkaline phosphatase activity was visualized by incubation with AS-BI substrate in 0.1 mol/L Tris pH 8.7 buffer for 15 minutes, resulting in red staining. Counterstaining was performed with hematoxylin. Staining was completely prevented by competition with MGO-albumin adducts.
Apoptosis
Cells were incubated with various concentrations of MGO or 3-DG in 24-well cluster plates for different time intervals. Cells were harvested by trypsinization then incubated for 15 minutes in a propidium iodide solution (5 μg/mL propidium iodide, 0.3% saponin, 5 mmol/L EDTA, and 50 μg/mL RNase) on ice. Based on propidium iodide content, cells can be divided into G1 and G2 cells by FACS analysis. Cells undergoing apoptosis showed impaired propidium iodide uptake due to fragmentation of DNA. Apoptotic cells were quantified as percentage of total cells.
MTT Assay
The modified colorimetric MTT assay is based on the selective ability of living cells to reduce MTT salt [3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium bromide; Sigma] to formozan by active mitochondrial dehydrogenases, thus representing the number of viable cells.
Mesothelial cells were grown to confluence in 24-well cluster plates. Cells were exposed to increasing concentrations of MGO or 3-DG (50, 100, 250, 500, and 1000 μmol/L) for 20 hours. Subsequently, 0.5 mg/mL MTT was added to the cells and incubated at 37°C for 4 hours. The formozan product formed by the metabolic conversion of the MTT salt was dissolved in 150 μL dimethylsulfoxide (DMSO) with the addition of 25 μL glycine buffer (0.1 mol/L glycine, pH 10.5) (33). Absorbance of the converted dye was recorded at 540 nm.
[3H]-Thymidine Incorporation Assay
Cells were plated on a 24-well cluster plate at a subconfluent density, then incubated with MGO or 3-DG for 48 hours in complete medium. During the last 7 hours, cells were pulsed with 1 μCi/mL [3H]-thymidine (methyl-[3H]-thymidine; Amersham
Lab, Pharmacia Biotech Benelux, Roosendaal, The Netherlands). At the end of the incubation, supernatants were removed and cells were placed on ice and washed with PBS, followed by a methanol wash. Cells were precipitated with 5% TCA. The precipitate was resolved in 0.3 mol/L NaOH. Released radioactivity levels were measured in a beta-scintillation counter.
Protein Synthesis
De novo protein synthesis was measured by [35S]-methionine incorporation assay. Confluent cells were incubated for 24 hours with MGO or 3-DG; during this incubation, [35S]-methionine (1 μCi/mL; Amersham) was present. Cells were placed on ice and washed with PBS. Cells were precipitated with 10% TCA for 30 minutes, followed by 2 washes with 5% TCA. After the precipitate had been resolved in 0.3 mol/L NaOH, radioactivity was measured in a beta-scintillation counter.
Detection of Adhesion Molecules by Cell-Bound Elisa
Confluent cells were incubated for 24 hours with various concentrations of MGO or 3-DG. Cells were fixed in 0.025% glutaraldehyde in PBS for 15 minutes. After three washes with PBS, cells were incubated with primary monoclonal antibodies diluted in incubation buffer (PBS with 10% newborn calf serum and 0.05% Tween 20) for 60 minutes on a rocker platform at room temperature. Cells were washed with PBS-T (PBS supplemented with 0.05% Tween 20) five times. After incubation with peroxidase-labeled rat anti-mouse immunoglobulin (Dako) for 60 minutes on a rocker platform at room temperature, cells were washed again five times with PBS-T and subsequently incubated with the substrate 3,5,3',5'-tetramethyl-benzidine. Color development was stopped by the addition of 2 N H2SO4 and was measured with a microtiter plate reader at 450 nm.
Cytokine Production
Confluent monolayers were stimulated with GDPs for 24 hours. Concentrations of IL-6 and IL-8 were measured in the culture supernatant, in duplicate, by capture ELISA, using matched-antibody pairs (Biosource BV, Etten-Leur, The Netherlands), according to the manufacturer's guidelines. MCP-1 was determined by ELISA, as described previously (7). Values (ng/mL) were normalized to the medium control (100%) because of donor-related variation in basal cytokine production.
Statistical Analysis
Values were tested for statistical significance using the nonparametric Mann–Whitney test. Probability values < 0.05 were regarded as significant.
Results
Kinetics of Scavenging GDPs in Culture
To assess scavenging of GDPs in culture conditions over time, MC culture medium containing either 0.1% or 10% FCS was incubated for various time periods at 37°C with 50 μmol/L MGO and 100 μmol/L 3-DG. Medium samples were then deproteinized and levels of free MGO and 3-DG concentration were determined. Figure 2 demonstrates the results of a representative experiment (n = 3). We observed a very fast decrease in MGO concentration within the first hour that further declined over time, with barely detectable levels after 24 hours of incubation [Figure 2(a)]. In the absence of cells, the reduction in MGO levels was dependent on the FCS content, which is consistent with the ability of GDPs to cross-link with protein moieties. Importantly, the levels of MGO were more reduced in the presence of MCs irrespective of FCS concentration. In contrast, the concentration of 3-DG remained constant over time and was slightly reduced after 24 hours [Figure 2(b)]. Similarly to MGO, the reduction of 3-DG was increased in the presence of MCs, especially after 24 hours. This indicates adsorption of these GDPs to the cellular surface of MCs and/or uptake by MCs. To further substantiate this, we performed immunocytochemistry with a specific antibody against MGO adducts. After a 6-hour incubation of MCs with MGO, variable staining for MGO adducts was seen. Some MCs were highly positive for MGO adducts, but most MCs were weakly stained. In contrast, no staining was detected in control cells (Figure 3). This clearly indicates that MGO is cross-linked to mesothelial cellular proteins.
Time course of complexing 50 μmol/L methylglyoxal (MGO) (A) and 100 μmol/L 3-deoxyglucosone (3-DG) (B) in culture medium containing either 0.1% (squares) or 10% (diamonds) fetal calf serum in the absence (closed symbols) or presence (open symbols) of mesothelial cells. Data are presented as percentages of the initial value of glucose degradation product.
Immunocytochemistry of mesothelial cells with monoclonal antibody against methylglyoxal (MGO) adducts. Mesothelial cells were incubated for 6 hours with complete medium in the absence (upper) or presence (lower) of 500 μmol/L MGO, and stained for MGO adducts. (Original magnification, ∞200)
MGO and 3-DG Induce Apoptosis of MCs
Methylglyoxal is known to be involved in apoptosis in various cell types (34,35). However, the susceptibility is not known for MCs. We studied the effect of MGO and 3-DG on apoptosis in MC in two independent experiments. Basal percentage of apoptosis in MC culture was 1.73% at 24 hours and 3.03% at 72 hours of culture. The results are shown in Table 1 and clearly demonstrate that exposure of MCs to a single dose of 500 μmol/L MGO or 3-DG induces significant apoptosis.
Induction of Apoptosis by Glucose Degradation Products
ND = not determined.
MGO and 3-DG Inhibit Proliferation but do not Inhibit Protein Synthesis
In three independent experiments using MCs from different donors, cells were exposed to increasing concentrations of a single dose of MGO or 3-DG. Cells were analyzed for their proliferation after 24 hours using [3H]-thymidine incorporation and MTT assay, and for their synthesis of protein after 48 hours using [35S]-methionine incorporation. All three experiments showed the same results; one representative example is shown in Figure 4. Mesothelial cells demonstrated a dose-dependent decrease in cell proliferation compared to untreated cells [Figures 4(a) and 4(b)]. [3H]-Thymidine incorporation was significantly reduced to 50% after exposure to 500 μmol/L MGO or 1000 μmol/L 3-DG. In parallel experiments using MCs from the same isolation, no effects of MGO or 3-DG on protein synthesis were observed [Figure 4(c)]. This finding indicates that, despite a slight apoptotic response after exposure to MGO or 3-DG, most cells are still viable and capable of producing proteins, but are blocked in their proliferative capacity.
Effect of methylglyoxal (hatched bars) and 3-deoxyglucosone (solid bars) on mesothelial cell proliferation determined by [3H]-thymidine incorporation (A) and MTT assay (B), and protein synthesis analyzed by [35S]-methionine incorporation (C). Each graph shows a representative experiment. Data are presented as percentages (mean ±SD) compared to cells incubated with culture medium alone (= 100%).
MGO and 3-DG Induce VCAM-1 Expression
To address whether GDPs are able to affect the functionality of MCs, we studied levels of adhesion molecule expression using cell-bound ELISA. Cells were exposed to a single dose of MGO or 3-DG and analyzed 24 hours later for ICAM-1, VCAM-1, and CD44 expression. The MCs expressed ICAM-1 and CD44 constitutively, while VCAM-1 expression was very low on resting MCs (Figure 5). Exposure of MCs to MGO or 3-DG did not change ICAM-1 and CD44 expression, but VCAM-1 expression was increased two- to threefold. Even at the lowest concentration of MGO and 3-DG tested, induction of VCAM-1 expression was observed.
— Effect of methylglyoxal and 3-deoxyglucosone (3-DG) on expression of the adhesion molecules VCAM-1 (black bars), ICAM-1 (white bars), and CD44 (hatched bars) on mesothelial cells. Mesothelial cells were cultured for 24 hours in culture medium with or without glucose degradation products or tumor necrosis factor alpha (TNFα). Data are presented as mean OD (±SD), as measured by cell-bound ELISA [VCAM-1: n = 9, except 50 μmol/L 3-DG (n = 6) and 500 μmol/L 3-DG (n = 3); ICAM-1 and CD44: n = 6].
As a positive control, MCs were stimulated by TNFα, which had no effect on ICAM-1 and CD44 expression, but increased VCAM-1 expression on MCs fivefold. Stimulation by TNFα increased VCAM-1 expression after 2 hours of incubation and increased further over time (Figure 6). At all time points, cycloheximide inhibited the induction of VCAM-1 expression, while cycloheximide alone did not alter the basal expression of VCAM-1. The same held true for VCAM-1 induction by MGO and 3-DG in the presence or absence of cycloheximide. VCAM-1 expression after 24 hours of stimulation with MGO or 3-DG was reduced by 57% and 62%, respectively, by cycloheximide.
Upregulation of VCAM-1 is dependent on de novo protein synthesis. Mesothelial cells were incubated for the indicated times with 200 U/mL tumor necrosis factor alpha (TNFα) alone (gray bars), with cycloheximide (black bars) to block protein synthesis, or in combination (hatched bars). VCAM-1 expression was determined by cell-bound ELISA.
These data show that both MGO and 3-DG induce an increase in VCAM-1 expression, which is dependent on de novo protein synthesis.
3-DG Induces Production of Proinflammatory Cytokines, but MGO does not
Cytokine production was measured in cell-free supernatants of MCs that had been cultured for 24 hours in the presence of MGO or 3-DG. After 24 hours of culture, the control level of IL-6 was 3.7 ng/mL (range 2.5 – 7.5 ng/mL) and the control level of IL-8 was 1.55 ng/mL (range 0.8 – 3.2 ng/mL). The production of both cytokines was not changed by a single addition of 100 μmol/L MGO. In contrast, a single exposure for 24 hours to 500 μmol/L 3-DG resulted in a two- to threefold increase in production of IL-6 and IL-8 (Figure 7). MCP-1, a chemoattractant for monocytic cells, was produced in low levels in control cells (13.9 ng/mL, range 12.5 – 16 ng/mL). The level of MCP-1 production did not change with the addition of MGO or 3-DG.
Effect of glucose degradation products (GDPs) on cytokine production by mesothelial cells (MCs). Production of MCP-1 (white bars), interleukin (IL)-6 (gray bars), and IL-8 (dark gray bars) was analyzed by capture ELISA in supernatant of MCs cultured with GDPs for 24 hours. Data are presented as percentages (mean ±SD) compared to cells incubated with culture medium alone (= 100%). Both IL-6 and IL-8 were significantly increased after incubation with 500 μmol/L 3-deoxyglucosone (3-DG). MGO = methylglyoxal.
Discussion
In the present study, we assessed some of the effects of short-term exposure of MGO and 3-DG on MC biology. The cellular responses described in this article were found only with concentrations of GDPs above physiological levels, as found in standard PD fluids, especially for MGO. Nevertheless, we feel that our study contributes to the understanding of the proinflammatory condition within the peritoneum of continuous ambulatory PD patients. It should be realized that these observations were made after brief exposure to a single GDP, while in the PD patient, MCs are chronically exposed to lower concentrations of many GDPs, some unidentified. Furthermore, carbonyl stress from the uremic circulation may also be involved in cellular changes in the peritoneal cavity (27). GDPs can induce cellular effects, either directly or via the formation of AGEs. These AGEs stimulate a variety of cell responses via specific cell-surface receptors on several cell types, including MCs (36).
First, we determined the levels of MGO and 3-DG in culture supernatant, revealing a more pronounced reduction in the presence of MCs, suggesting that these GDPs could interact with MCs by binding and/or uptake. This could be substantiated by immunocytochemistry using an MGO adduct-specific monoclonal antibody. We demonstrated that, although staining intensity is high in some MCs, most MCs stained weakly. In contrast, no staining was detected in control cells. This is a strong indication that MGO adducts are cross-linked to mesothelial cellular proteins. At present, we do not know the reason for this variable staining. It might, however, be related to the phase of the cell cycle, induction of apoptosis, or the presence of certain receptor molecules.
Second, we demonstrated that VCAM-1 expression was increased two- to threefold under the influence of MGO and 3-DG, while ICAM-1 and CD44 expression was not altered. The upregulation of VCAM-1 was dependent on de novo protein synthesis, as upregulation was prevented by cycloheximide. Even at the lowest concentration of MGO tested, which is the approximate concentration in conventional glucose-containing PD fluids, VCAM-1 expression was slightly increased. It has been shown by others that VCAM-1 is upregulated after AGE-RAGE ligation (37) via a NF-κB–dependent pathway (38).
In the peritoneal cavity, MCs are an important source of chemoattractants, such as IL-8 and MCP-1 (6). Recently, it was shown that chemokines produced by peritoneal fibroblasts might also contribute to the intraperitoneal recruitment of leukocytes (39). In our study, we showed that production of IL-8 and IL-6 was doubled by 3-DG, while the production of MCP-1 was not changed. Together with increased expression of VCAM-1, the increased production of cytokines could favor recruitment of various leukocyte subsets in the peritoneal cavity. In contrast to the increased constitutive production of IL-6 after a single dose of 3-DG, Witowski et al. showed that long-term exposure of MCs to a GDP mixture decreased both constitutive and IL-1β–induced production of IL-6 (40). In THP-1 cells, MGO did not affect cytokine production, although MGO-modified human serum albumin could induce IL-1β production (41). These differences could be the result of the different susceptibilities of THP-1 and MCs in their response to GDPs/AGEs. Alternatively, MGO first has to modify proteins to exert an effect on cytokine production.
Finally, we confirmed the apoptotic and antiproliferative effects of MGO and 3-DG (17,34,35). We observed a significant induction of apoptosis, caused probably by the elevation of intracellular oxidative stress, in the presence of high concentrations of MGO or 3-DG. The reason that 3-DG led to apoptosis to a degree similar to MGO might be related to the fact that MGO is so highly reactive that, before entering the cell, most MGO reacts to serum proteins, in contrast to 3-DG which is not scavenged so quickly by serum proteins. This is clearly illustrated in Figure 2. Another possibility could be that MGO, in contrast to 3-DG, is converted to D-lactate in the cell by glyoxalase I. In addition, we observed a dose-dependent decrease in the proliferation of MCs, which is consistent with a previous study (17). We observed a significant reduction in MC proliferation at 500 μmol/L MGO or 3-DG. These GDPs may thus impair regeneration of MCs after damage during PD. In contrast, the decreased proliferation was not accompanied by changes in the level of protein synthesis, at least within the time frame tested. This is also confirmed by the de novo synthesis of VCAM-1, IL-6, and IL-8 after exposure to MGO or 3-DG.
In conclusion, our data contribute more understanding of the acute effects of MGO and 3-DG on MC biology. Our results indicate that both MGO and 3-DG cross-link to mesothelial cellular proteins. Furthermore, immunocytochemistry clearly shows the presence of MGO adducts in MCs. Upon short-term exposure to high levels of MGO and 3-DG, MCs demonstrated a proinflammatory response. Our data clearly show that short-term exposure to GDPs affects MC biology.
Footnotes
Acknowledgment
This work was supported by a grant from Baxter NL and the Dutch Kidney Foundation (NSN C97.17005b).