Ultrastructural changes of the peritoneum in a rabbit model of peritoneal dialysis

Background/Aim. The number of patients with end-stage renal diseases treated with chronic dialysis is increasing over the last years. Long-term peritoneal dialysis is associated with progressive development of structural and functional alterations of peritoneal membrane. The aim of the study was to analyze ultrastructural alterations of mesothelial monolayer and submesothelial tissue in a modified nonuremic experimental model of peritoneal dialysis in rabbits. Methods. The study was performed on 5 healthy Chinchilla rabbits. Surgical procedures of implantation and removal of peritoneal catheter, prevention of catheter clothing, prevention of infection and dialysate instillation were performed according to previously described protocols. Peritoneal tissue samples were collected upon catheter placement and removal after a 5-week follow-up and processed for transmission electron microscopy (TEM) examination. Results. The rabbits tolerated anesthesia, surgical procedure and the applied regimen of dialysate instillations well. The animals recovered completely and no adverse effects were noted. In the animals treated with peritoneal dialysis instillations, TEM revealed alterations of the mesothelial monolayer and submesothelial tissue. The mesothelial cells in direct contact with dialysis fluid were prone to shrinking. They lost the typical cobblestone morphology and assumed a flattened shape. The mesothelial cells were often detached from the basement membrane. These cells showed euchromatic nuclei, higher number of microvilli in their apical part and very numerous vesicles. A higher quantity of collagen fibers was noticed in the peritoneal lamina propria in close relation to the basement membrane of mesothelium. The nuclei of the fibroblasts were also euchromatic. Numerous mitochondria, granules and vesicles were present in their cytoplasm. Conclusion. The used rabbit model of peritoneal dialysis is simple, practical to perform, reproducible, not expensive and not requiring advanced devices. It is suitable for obtaining peritoneal tissue samples for histological examination and can be used to analyze the effects of dialysis solutions on the rabbit peritoneal membrane.

TEM revealed alterations of the mesothelial monolayer and submesothelial tissue.The mesothelial cells in direct contact with dialysis fluid were prone to shrinking.They lost the typical cobblestone morphology and assumed a flattened shape.The mesothelial cells were often detached from the basement membrane.These cells showed euchromatic nuclei, higher number of microvilli in their apical part and very numerous vesicles.A higher quantity of collagen fibers was noticed in the peritoneal lamina propria in close relation to the basement membrane of mesothelium.The nuclei of the fibroblasts were also euchromatic.Numerous mitochondria, granules and vesicles were present in their cytoplasm.Conclusion.The used rabbit model of peritoneal dialysis is simple, practical to perform, reproducible, not expensive and not requiring advanced devices.It is suitable for obtaining peritoneal tissue samples for histological examination and can be used to analyze the effects of dialysis solutions on the rabbit peritoneal membrane.

Introduction
The number of patients with end-stage renal diseases treated with chronic dialysis methods has been increasing over the past years 1,2 .In spite to significant scientific and technological advances in dialysis treatment, it is still associated with high morbidity and mortality rates 3 .Peritoneal dialysis (PD) has been an established form of renal replacement therapy for patients with end-stage renal disease for three decades now and it is currently performed in about 15% of uremic patients around the world 4 .
The peritoneal membrane is a live biological membrane, structurally similar to other mesothelial membranes in the body 5 .It is affected by uremia per se, but also by chronic exposure to dialysis fluid during PD 6 .Water and solvents exchange during PD is performed through the peritoneal membrane.The failure of peritoneal membrane to provide adequate dialysis correlates with structural changes in the peritoneal membrane 7 .The mesothelial layer of the peritoneal membrane is stimulated and suffers injury in long-term PD patients.Furthermore, structural changes are also observed in the submesothelial layer in the peritoneum of PD patients 5,[8][9][10][11] .These ultrastructural changes of the mesothelium and submesothelial tissue alter the transport characteristics of the peritoneum as dialysis membrane and can ultimately cause ultrafiltration failure 12 .
A limitation of peritoneal alterations study resulting from PD are technical and ethical difficulties to obtain biopsies of the peritoneum.Therefore, various uremic and nonuremic, acute and chronic animal models of peritoneal dialysis enable investigation of structural and functional changes of peritoneum exposed to PD.
The aim of this study was to analyze ultrastructural alterations of mesothelial monolayer and submesothelial tissue in a modified non-uremic infusion experimental model of peritoneal dialysis in rabbits.

Methods
The study was performed on five healthy Chinchillas rabbits (3 males and 2 females), weighing 2,932 ± 504 g at the beginning of the experiment.The rabbits were housed at room temperature (22 ± 2 o C) and 12 hours light cycles, in individual cages, which were cleaned daily.The animals were given stan-dard pellet rabbit food (Veterinary Institute, Subotica, Serbia) and water ad libitum.All the rabbits were allowed to adapt to new living conditions for at least five days prior to catheter insertion.During the study period of five weeks (one week for recovery following catheter placement and four weeks of dialysis) a diary of animal behavior was kept, including daily measurements of body mass, body temperature, food and water intake and defecation, antibiotics administration, other therapy and interventions (wound toilette, catheter suturing etc.).
Animals were anesthetized with tiopental BP 1G (Rotexmedica, Trittau, Germany; 0.5 mL/kg administered thorugh ear vein) for catheter placement at the beginning of experiment and for catheter removal, at the end of experiment.The peritoneal catheter in this study was made from the Tro-soluset infusion system (Troge Medical GMBH, Hamburg, Germany).The catheter was inserted in the abdominal cavity through a small incision on the front abdominal wall, below the left costal arc and parallel to the median abdominal line.It was then led through a subcutaneous tunnel to the exit site on the neck, according to a previously described procedure [13][14][15] .
To prevent infection cefuroxime (Nilacef ® , Hemofarm AD, Serbia) was administered intramuscularly daily three days before catheter placement and three days following catheter removal.The same antibiotic was administered daily through the peritoneal catheter during the four weeks instillation period.Antibiotic doses were calculated according to body mass.
Following a 7-day recovery period after catheter placement, the animals were instilled with dialysis solution (Dianeal PD4 Glucose, with 3.86% glucose; Baxter Vertriebs GmbH, Vienna, Austria), previously warmed at 37°C, once a day for 28 days [13][14][15] .To prevent dyspnea, the initial dose of dialysis solution of 60 mL was gradually increased by 10 mL/day until a total dose of 40 mL/kg was reached.
Peritoneal tissue samples for histological analysis were taken upon peritoneal catheter insertion and catheter removal.This tissue is extremely fragile and susceptible to mechanical irritation and environmental factors.Therefore, oval tissue samples, 18 3 mm, were taken with extreme caution, immediately after opening the abdominal cavity, to avoid any damage.
For transmission electron microscopic (TEM) analysis tissue samples were fixed for 24 h in 4% glutaraldehyde diluted in Sorensen's phosphate buffer 0.1M (pH 7.4), then rinsed in Sorensen and cacodylate buffer.The samples were then postfixed in 1% osmium tetroxide in 0.1M cacodylate buffer and left over night in 4% uranyl acetate.After dehydratation in ethanol and propylene-oxide, the samples were embedded in Epon.Fine sections were contrasted with uranyl acetate and lead-citrate and analyzed with a transmission electron microscope Morgagni 268D.
All experimental procedures were performed according to the European Council Directive (86/609/EEC) and with the permission from The Committee for Animal Care, University of Belgrade.
The results were statistically analyzed with Microsoft Office Excel 2006 and shown as mean values and standard deviations.

Results
All the animals tolerated aclimatization, surgical procedure and dilaysate instillations well.No peritonitis, nor high temperature or other signs of infections were noted.
During a 5-week study period the animals showed steady mass increase (Figure 1), while body temperature remained in physiological range (Figure 2).The animals had neither infection of the surgical wound nor peritonitis episode, and no catheter clothing, thanks to preventive use of antibiotic and heparin.
TEM analysis of the control samples of peritoneal tissue, taken before dialysate instillations, revealed flat or cubic irregular mesothelial cells.Adjacent mesothelial cells were connected with intercellular connections, mostly desmosomes (Figure 3).The apical plasmalema formed numerous cytoplasmic extensions -microvilli.Pinocytotic vesicles were observed in the apical and other parts of plasmalemma.The submesothelial connective tissue showed fibroblasts, collagen and elastic fibers.The fibroblasts were large and heavily branched, so only their parts of different shapes and sizes could be seen on individual sections.TEM analysis of peritoneal tissue samples taken after dialysate instillations showed alterations in both mesothelial and submesothelial layers.
Mesothelial cells exposed to dialysis fluid were flat, with widening of the perinuclear space (Figures 4 and 5).Their nuclei were euchromatic, with prominent nucleoli.Numerous transport vesicles were observed throughout cytoplasm, as well as increased number of microvilli on the apical plasmalemma of the mesothelial cells.Patches of peritoneum devoid of mesothelium were seen even at lower magnification.
The submesothelial layer was dominated by the bundles of cross-striated collagen fibers with a different course, surrounding fibroblasts (Figures 4 and 6).Fibroblasts had mostly euchromatic nuclei and numerous secretory granules, vesicles and mytochondria in their cytoplasm (Figure 6).

Discussion
Animal models of PD have significantly contributed to better understanding of structural and functional changes of peritoneum exposed to PD.The modified rabbit PD model used in our study is practical, reproducible and did not require sophisticated technology.The animals tolerated the procedure well and no complications, such as peritonitis or catheter obstruction, were noted.This model is, therefore, adequate for further investigation of the long-term effects of dialysis solutions on the peritoneal membrane in rabbits [13][14][15] .
Long-term PD results in peritoneal injury with structural changes and functional decline 16 .The so-called conventional solutions have electrolyte content similar to serum and acidic reaction in order to prevent caramelization of glucose during heat sterilization.Glucose is widely used as osmotic agent in these solutions in concentrations 15 to 40 times higher than in physiological fluids.Therefore, PD patients are exposed to large quantities of glucose, even up to 100 kg per year 17 .High acidity and high glucose concentration, the presence of lactate as buffering agent and numerous glucose degradation products (GDP) formed during sterilization and storage, contibute to peritoneal injury.Typical structural alterations of peritoneum exposed to long-term PD include the loss of mesothelial cells, thickening of submesothelial tissue and various vascular changes 9,13,15,17,18 .
A number of in vitro and in vivo studies have shown adverse effects of GDP on the peritoneum.Accumulation of these products (especially toxic are 3,4dideoxyglucosone and methylglyoxal) in the peritoneum cause structural changes in the tissue, either by direct action or by inducing the formation of advanced glycosilated end products (AGE) 19,20 .The adverse effects of GDPs include: inhibition of cell proliferation and reparation of lesions; decreased IL6, fibronectin and collagen type 1 synthesis; promotion of apoptosis; increasing reactive oxygen species synthesis and carbonyle stress; increasing expression of receptors for AGEs, adhesion molecules, growth factors, vascular endothelial growth factor and IL8; cellular gluthatione depletion; reduced expression of intercellular tight-junctions; upregulation of HLA antigen expression on mesothelial cells and induction of epithelial-mesenchymal transition 6,21,22 .
The mesothelial cells are flat or cubic specialized epithelial cells lining the internal organs and peritoneal, pleural, pericardial and synovial cavities 5 .The shape of mesothelial cells in rabbits from our study, prior to exposure to dialysis solution, corresponds to these data (Figure 3).The luminal surface of mesothelial cells has numerous finger-like cytoplasmic extensions -microvilli, increasing the functional mesothelial surface for exchange between mesothelial cells and peritoneal cavity.These microvilli protect the delicate mesothelial surface from frictional injury by entrapping water and serous exudates, which act as lubricants for the cells.Microvilli remain present even during PD, thus increasing the surface for water and solutes exchange between peritoneal cavity and the cytoplasm of mesothelial cells (Figure 4).
The number of pinocytotic vesicles, which are normally present in the apical region of mesothelial cells, multiplies during PD due to intensive exchange processes (Figure 4).The apical suface of mesothelial cells also shows a single microcilia and cytosceleton from intermediar vimentin and cytokeratin filaments.A well developed Golgi apparatus, numerous perinuclear mytochondria, granulated endoplasmic reticulum and ribosomes, as well as euchromatic nuclei in mesothelial cells are indicative of their dynamic biosynthetic activity.This activity becomes even more significant during PD (Figure 5).A submesothelial layer of peritoneum consists of extracellular matrix made up of glycosaminoclycans (hyaluronic acid and chondroitin sulfate) and proteoglycans, collagen fiber bundles (mostly type I and III) and elastic fibers, with fibroblasts and free cells (macrophages, mastocytes, leukocytes and multipotent cells).Vascular structures and lymphatics are found in the subserous place.This tissue functions as a molecular filter regulating transition of various molecules and cells through peritoneum.The fibroblasts become activated during PD and enhance synthesis of collagen, which can be found in thick bundles in the submesothelial layer (Figure 4  and 6).
The mesothelial cells isolated from drained dialysate change their epithelial-like morphology into fibroblastlike cells.This mesothelial-fibroblast transformation is characterized by the loss of cadherin and cytokeratin markers, which are typical for the epithelial phenotype, and by appearance of alpha-smooth muscle actin.These transformed mesothelial cells are usually found in dialy-sate from PD patients treated with conventional dialysis solutions 6,23 .

Conclusion
The modified rabbit PD model used in our study is practical, reproducible and does not require sophisticated technology.The animals tolerated the procedure well and no complications, such as peritonitis or catheter obstruction, were noted.The studied experimental model is suitable for obtaining peritoneal tissue samples for histological examination.This model is, therefore, adequate for further investigation of long-term effects of dialysis solutions on the peritoneal membrane in rabbits.

Aknowlegments
The authors would like to express their deepest gratitude to the associates of the Institute of Pathophysiology, Faculty of Medicine, University of Belgrade, and especially to the laboratory technician Vladimir Miljkovi for his valuable help.
This study was conducted as a part of the research project No145070 funded by the grant from the Serbian Ministry of Education, Science and Technological Development.