|
|
||||||||
Bench |
Department of Molecular Cell Biology and Immunology,1 VU University Medical Center, Amsterdam, The Netherlands; Department of Molecular Biology,2 University Hospital La Princesa, Madrid, Spain; Department of Nephrology,3 VU University Medical Center, Amsterdam, The Netherlands a Now at Department of Nephrology, University Medical Center Groningen, Groningen, The Netherlands.
Correspondence to: M.N. Schilte, Department of Molecular Cell Biology and Immunology, H269, VU University Medical Center, Van der Boechorststraat 7, 1081 BT Amsterdam, The Netherlands. m.schilte{at}vumc.nl
| ABSTRACT |
|---|
|
|
|---|
Background: Peritoneal dialysis (PD) is associated with
functional and structural alterations of the peritoneal membrane, particularly
new vessel formation and fibrosis. In addition to anticoagulant effects,
heparin displays anti-inflammatory and angiostatic properties. Therefore, the
effects of administration of heparins on function and morphology of the
peritoneal membrane were studied in a rat PD model.
Methods: Rats received 10 mL conventional PD fluid
(PDF) daily, with or without the addition of unfractionated heparin (UFH) or
low molecular weight heparin (LMWH) in the PDF (1 mg/10 mL intraperitoneally)
via a mini access port. Untreated rats served as controls. After 5 weeks, a
90-minute functional peritoneal transport test was performed and tissues and
peritoneal leukocytes were taken.
Results: PD treatment induced loss of ultrafiltration
(p < 0.01), a twofold increase in glucose absorption (p
< 0.03), increased urea transport (p < 0.02), and loss of
sodium sieving (p < 0.03), which were also found in the PDF +
heparin groups. Increased peritoneal cell influx and hyaluronan production
(p < 0.02) as well as an exchange of mast cells and eosinophils
for neutrophils after PD treatment were observed in PD rats; addition of
heparin did not affect those changes. Mesothelial regeneration, submesothelial
blood vessel and matrix formation, and accumulation of tissue macrophages were
seen in PD animals. Spindle-shaped vimentin-positive and cytokeratin-negative
cells indicated either partial injury and denudation of mesothelial cells or
epithelial-to-mesenchymal transition. Neither UFH nor LMWH affected any of
these morphological changes.
Conclusion: Within 5 weeks, PD treatment induces a
chronic inflammatory condition in the peritoneum, evidenced by high transport,
leukocyte recruitment, tissue remodeling, and induction of spindle-shaped
cells in the mesothelium. Addition of LMWH or UFH to the PDF did not prevent
these adverse PDF-induced peritoneal changes.
KEY WORDS: Angiogenesis; epithelial-to-mesenchymal transition; fibrosis; heparin; mesothelium; morphology; peritoneal transport.
Peritoneal dialysis (PD) is an effective treatment for patients with end-stage renal failure. During this treatment, the peritoneal membrane functions as a dialyzing organ and is constantly exposed to PD fluid (PDF). It is well known that long-term PD induces loss of mesothelial cells, submesothelial thickening, and neoangiogenesis (1). The nonphysiological environment in peritoneum caused by chronic PDF instillation sustains low-grade of sterile inflammation (1). This low-grade inflammatory state may, in time, cause functional deterioration of the peritoneal membrane, with increasing peritoneal permeability to small solutes and reduced ultrafiltration capacity (2,3). At the molecular level there is increasing evidence that cytokines, chemokines, and growth factors play key roles in regulating and sustaining this proinflammatory condition (4). After long-term PD treatment, peritoneal mesothelial cells show a progressive loss of the epithelial phenotype and acquire fibroblast-like characteristics, suggesting epithelial-to-mesenchymal transition (EMT) (5).
Heparin is a highly sulfated polysaccharide belonging to the glycosaminoglycan family and is synthesized by mast cells. Although it is most known for its anticoagulant effect, heparin has a number of immunomodulatory and anti-inflammatory activities, including binding of cytokines, chemokines, and growth factors (6). Earlier reports demonstrated that heparin can inhibit leukocyte rolling along the vessel wall, a process known to involve selectins (7,8). The function of selectins in this process depends on recognition of specific glycoprotein ligands. However, a variety of carbohydrate structures, including heparin, may bind these lectin-like molecules and interfere with ligand binding (9). In addition to unfractionated heparin (UFH), low molecular weight heparin (LMWH) is one of the most commonly used types of heparin in clinical settings. Fractions of LMWH are prepared from standard UFH and are similar to UFH in many aspects. The main advantages of this class of antithrombotic agent compared with UFH are improved bioavailability, prolonged half-life (which may permit one single daily subcutaneous injection), and improved efficacy (10,11).
In the clinical PD setting, heparin is used to prevent catheter obstruction and the formation of fibrin during peritonitis. Results of studies on the effects of intraperitoneal heparin administration on peritoneal transport and inflammation have been contradictory (12–16). Therefore, we studied the effects of long-term addition of UFH, as well as LMWH, to the PDF on function, defense, and tissue remodeling of the peritoneal membrane to determine whether heparin can improve peritoneal transport and whether it can also prevent morphological changes in peritoneal tissues. In the present study, either UFH or LMWH was added to the PDF for 5 weeks in a chronic peritoneal exposure model in the rat (17,18).
| MATERIAL AND METHODS |
|---|
|
|
|---|
EXPERIMENTAL DESIGN
Fluids were instilled via a peritoneal catheter connected to an implanted
subcutaneous mini access port, as previously described
(19). Rats that were not
surgically treated and received no fluid instillation served as the control
group (n = 8). The rest of the animals (n = 34) received 2
mL saline with 1 IU/mL heparin to allow wound healing during the first week
after their operation. Thereafter, during a 5-week period
(20–23),
rats were instilled daily with 10 mL standard PDF (Dianeal PD4, 3.86% glucose,
pH 5.2; Baxter R&D, Utrecht, The Netherlands; PDF-control; n =
14), PDF with the addition of UFH (PDF-UFH; n = 10), or PDF with the
addition of LMWH (nadroparin calcium; PDF-LMWH; n = 10). The
concentrations used for UFH and LMWH (1 mg/rat/day: equivalent to 175 IU
UFH/rat/day and 100 IU LMWH/rat/day) are consistent with the concentrations
used in the clinical setting and proportional to the dose used in deep venous
thrombosis prophylaxis in humans
(24). To check whether the
animals were not over-heparinized due to daily heparin administration, blood
samples were taken after 4 weeks of treatment 3 hours after PDF + heparin
instillition to measure APTT clotting times and anti-factor Xa levels. At the
end of the experiment, a functional 90-minute peritoneal equilibrium test
(PET) was performed (see below) under a mixture of fentanyl–fluanisone
(0.05 mL/100 g) and midazolam (0.08 mL/100 g) anesthesia. After the animals
were sacrificed, tissues and peritoneal leukocytes were taken to analyze
functional, morphological, and cellular parameters.
ANALYSIS OF PERITONEAL CELLS AND EFFLUENTS
After 5 weeks of treatment, a PET was performed by injecting 30 mL standard
PDF into the peritoneal cavity via a direct intraperitoneal catheter (Venflon
Pro; BD Medical Systems, Franklin Lakes, NJ, USA). After 90 minutes, the PDF
was drained and ultrafiltration capacity was calculated. Cells were isolated
by centrifugation and cell number, and viability was determined by trypan blue
exclusion. Cytocentrifuge preparations were stained with
May–Grünwald–Giemsa and cells were differentiated. After the
PET, the animals were sacrificed and a cardiac puncture was performed.
Glucose, sodium, creatinine, and urea concentrations were analyzed in serum
samples and in the cell-free effluents at t = 90 minutes of the PET.
At t = 0 and t = 45 minutes, sodium concentrations in the
effluents were analyzed to measure sodium sieving. The amount of hyaluronic
acid in the supernatant of the peritoneal effluent was determined using an
ELISA-based assay according to Fosang et al.
(25).
MORPHOLOGIC ANALYSIS
Mesentery and Omentum: A part of omental and mesenteric tissue was
dissected and spread on a glass slide for fluorescence microscopic
examination. Both tissues were stained with antibodies to CD31 (PECAM;
Serotec, Oxford, United Kingdom) to visualize vasculature and to ED2 (Serotec)
to visualize macrophages. As a negative control, conjugate controls
(Invitrogen, Carlsbad, CA, USA) without the first antibody were used. Images
were analyzed by computerized digital image analysis (AnalySIS; Soft Imaging
System; Olympus, Hamburg, Germany). The areas positive for CD31 and ED2 were
calculated as percentages of the total area of the tissue.
Parietal Peritoneum: The parietal peritoneum was taken at the contralateral side to the tip of the implanted catheter using a standardized method. Cryostat sections (7 µm) were cut and stained for CD31 and ED2. Images were analyzed by computerized digital image analysis and the areas positive for CD31 and ED2 were calculated as percentages of the total area of the submesothelial matrix layer. Frozen sections were also used to quantify fibrosis formation. The thickness of the submesothelial extracellular matrix was determined after Van Gieson staining (Merck, Darmstadt, Germany) as the mean of 10 independent measurements for each animal.
Liver Imprints: Imprints of the mesothelial monolayer of the liver were taken with 3% gelatin-coated slides according to a previously described method (26) and stained with May–Grünwald–Giemsa. Mesothelial cell density per 0.1 mm2 area was counted using a scored eyepiece; the mean of 15 areas was calculated for each slide and expressed as cells per square millimeter. Liver imprints were also stained for vimentin (Serotec) and cytokeratin (DakoCytomation, Glostrup, Denmark) to determine spindle-shaped cells that underwent EMT.
STATISTICAL ANALYSIS
All data are presented as median and interquartiles. The
Kolmogorov–Smirnov test for normality showed that the groups were not
normally distributed; therefore, differences between groups were analyzed
using the nonparametric Mann–Whitney U-test. We made three comparisons,
namely, control versus PDF-control, PDF-control versus PDF-UFH, and
PDF-control versus PDF-LMWH. According to a Bonferroni correction, probability
values of p < 0.03 were considered significant. Correlation
analysis was performed using the nonparametric Spearman's rho test. Survival
analysis was performed using the Kaplan–Meier method.
| RESULTS |
|---|
|
|
|---|
TRANSPORT PARAMETERS AND COMPOSITION OF PERITONEAL CELLS
Table 1 shows the transport
parameters measured in the PET effluents at t = 90 minutes for all
four groups. The PET in the control animals demonstrated a net ultrafiltration
capacity of 8.6 mL per 90 minutes. Exposure to PDF reduced peritoneal
performance significantly to a net ultrafiltration volume of 6.5 mL. Addition
of UFH or LMWH further reduced the ultrafiltration capacity. In addition to
the reduced net ultrafiltration capacity, PDF induced a significantly enhanced
glucose absorption (p < 0.03) and an increased urea transport
(p < 0.02). Sodium measurements at t = 0, t =
45, and t = 90 minutes of the PET showed a reduction in the sodium
dip upon PDF treatment (control vs PDF-control at t = 90: p
< 0.03). Addition of UFH or LMWH did not restore PDF-induced changes in
peritoneal transport (Table
1).
|
The total cell numbers and differentiation of peritoneal cells recovered from the PET dialysate are shown in Table 2. Significantly more cells were found in the peritoneal effluents of the three PDF-treated groups compared to the control animals (p < 0.02). No difference in cell numbers was seen between PDF-control and the heparin groups. Cell differentiation revealed an exchange of mast cells and eosinophils for neutrophils after PD treatment. Treatment with UFH as well as LMWH seemed to nearly abolish the PD-induced inflammation by reducing neutrophil recruitment. Furthermore, the addition of UFH or LMWH resulted in slightly reduced hyaluronan synthesis compared to the PDF-control group (Table 2).
|
MORPHOLOGICAL CHANGES
Mesentery and Omentum: All groups exposed to PDF showed a
significant increase (p < 0.01) in accumulation of mature
macrophages (ED2) and new blood vessel formation (CD31) compared to the
control group (Figure 1).
Neither UFH nor LMWH was able to prevent the increased density of macrophages
and blood vessels seen in the mesentery and omentum. When all four groups were
combined, the increased accumulation of macrophages was positively correlated
to increased blood vessel density: r = 0. 64 and r = 0.89
for mesentery and omentum respectively; both p < 0.01.
|
|
Liver Imprints: Mesothelial cell density was measured as a marker for mesothelial cell regeneration. The number of mesothelial cells on the liver showed a significant twofold increase (up to 2045 ± 1167 cells/mm2) in the PDF-control group compared to the untreated animals (1168 ± 273 cells/mm2, p < 0.01) [Figures 3(a) and 3(b)]. Neither of the two heparins was able to influence mesothelial cell regeneration. Spindle-shaped cells were found between mesothelial cells in PDF-treated animals [Figure 3(b)]. An average of 7% – 10% of the cells found on the liver imprints of PDF-treated animals showed this phenotype, whereas these cells were not found in control animals (p < 0.01). No differences in the numbers of spindle-shaped cell were found between the three PDF-exposed groups. Vimentin-positive and cytokeratin-negative staining [Figures 3(c) and 3(d)] identified those cells as fibroblast-like cells, indicating either that the imprint picked up submesothelial fibroblasts due to partial injury and denudation of mesothelial cells, or possibly mesothelial cells undergoing EMT [Figure 3(d)].
|
| DISCUSSION |
|---|
|
|
|---|
We decided to add heparins to the PDF in order to obtain maximal local availability and to minimize systemic effects of the heparin intervention. Nonsignificant changes in APTT clotting times and anti-factor Xa levels indicated that the animals were not over-heparinized during the experiment, whereas increased plasma factor Xa levels after LMWH administration showed that adequate working doses were administered. The concentrations used for UFH and LMWH were clinically relevant doses and were similar or even higher compared to the literature (15,24,27–29). Higher concentrations of heparin might have led to bleeding complications or heparin-induced thrombocytopenia (30). The normal clotting times in animals that received daily heparin might be due in part to the fact that intraperitoneal heparin barely reached the circulation or that it already left the circulation within 3 hours after administration. In line with our results are the unaffected APTT and anti-factor Xa measurements of PD patients that received an intraperitoneal injection of LMWH (14,31,32), whereas increased plasma factor Xa levels were seen after subcutaneous injection (32). These data indicate that intraperitoneal administration of heparin has only minor and partial systemic effects (31,32).
Although our intention to avoid systemic effects of heparin treatment was successful, none of the PDF-induced peritoneal changes could be (partly) prevented by the local heparin treatment. Since we were especially interested in the long-term effects of heparin treatment and wanted to avoid direct effects of heparin, we treated the animals daily during 5 weeks and performed the PETs without heparin 1 day after the last heparin administration. An identical methodological approach has been successfully applied in the past with intraperitoneal aminoguanidine treatment (21). This is an important difference with some other studies and might explain why others reported beneficial effects of glycosaminoglycans, such as heparin or hyaluronan, present in PDF during the PET procedure (15,33), which is explained by a reduction of the peritoneal fluid reabsorption rate due to the formation of a filter-cake of the glycosaminoglycan chains at the tissue–cavity interface.
The fact that chronic treatment with either of the heparins did not prevent peritoneal membrane remodeling (new vessel formation, fibrosis, macrophage influx, and mesothelial EMT) might explain why the peritoneal transport values in the heparin-treated groups were not different from the PDF-control group, since molecular transport is highly dependent on vascular surface area and the thickness and composition of the submesothelial matrix (34). Positive effects of heparin on peritoneal transport are reported in studies by Pawlaczyk et al. (13) and Sjoland et al. (14). However, the first study showed only marginal and nonsignificant differences with heparin after 30 days of treatment and, to the best of our knowledge, the second study has not been confirmed by others.
The absence of any effect of chronic heparin treatment on peritoneal inflammatory markers such as cell influx might be explained by the fact that intraperitoneal cell recruitment occurs at the level of postcapillary venules (22) and via the omentum (23). We show that microvascular density in the peritoneal tissues was unchanged by heparin treatment. Moreover, the absence of heparin in the circulation during the PET allowed normal selectin–ligand interaction, which is the first step in leukocyte–endothelial interaction. There have been reports of anti-inflammatory properties of heparinoids in other peritoneal inflammatory models. However, in all those studies, the heparinoids were given intravenously and were present in systemic circulation, thereby inhibiting L- and/or P-selectin-dependent peritoneal cell influx (8,35–38). In line with our results, heparin showed no effects on PDF-induced neutrophil numbers (15), nor did it show any effect on inflammatory cells, fibrosis, or angiogenesis in an animal model of abdominal wound healing (29).
Concerning the role of heparins in angiogenesis, opposite results have been described (29,39). In general, tumor angiogenesis seems to be highly dependent on fibrin formation and tissue factor pathway, processes that are highly modulated by heparin (40). However, not all forms of new vessel formation are fibrin- and/or tissue factor-dependent. An impressive number of factors and mediators might be involved, depending on the model system used (41). The fact that chronic heparin treatment did not prevent neoangiogenesis suggests that intraperitoneal fibrin formation is not involved in our experimental PD model.
Another intriguing finding was the appearance of vimentin-positive cytokeratin-negative spindle-shaped cells within the mesothelial monolayer, which suggests EMT of the mesothelial cells. An alternative explanation would be that the spindle-shaped cells might be submesothelial fibroblasts picked up by the imprint and indicative of partial injury and denudation of mesothelial cells. Addition of either of the heparins did not modify this process and neither heparin affected mesothelial regeneration, as evidenced by increased mesothelial cell density on the liver imprints.
In conclusion, we have confirmed that PD induces significant damage to peritoneal tissues within 5 weeks. Moreover, we have shown an induction of spindle-shaped cells in the mesothelium, indicating either mesothelial damage or EMT. We also conclude that, based on our findings, addition of UFH or LMWH to PDF does not prevent these changes. Therefore, our data do not support long-term peritoneal administration of heparin during chronic PD treatment.
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
Received 7 January 2008; accepted 1 May 2008.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
P. Fabbrini, M. N. Schilte, M. Zareie, P. M. ter Wee, E. D. Keuning, R. H. J. Beelen, and J. van den Born Celecoxib treatment reduces peritoneal fibrosis and angiogenesis and prevents ultrafiltration failure in experimental peritoneal dialysis Nephrol. Dial. Transplant., December 1, 2009; 24(12): 3669 - 3676. [Abstract] [Full Text] [PDF] |
||||
|
|
M. N. Schilte, J. W.A.M Celie, P. M. t. Wee, R. H.J. Beelen, and J. van den Born FACTORS CONTRIBUTING TO PERITONEAL TISSUE REMODELING IN PERITONEAL DIALYSIS Perit. Dial. Int., November 1, 2009; 29(6): 605 - 617. [Abstract] [Full Text] [PDF] |
||||
|
|
P. Margetts HEPARIN AND THE PERITONEAL MEMBRANE Perit. Dial. Int., January 1, 2009; 29(1): 16 - 19. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |