Peritoneal Changes after Exposure to Sterile Solutions by Catheter
Michael F. Flessner*,
Kimberly Credit*,
Karla Henderson*,
Heather M. Vanpelt*,
Rebecca Potter*,
Zhi He,
Jeffrey Henegar and
Barry Robert
Departments of * Medicine and Pathology, University of Mississippi Medical Center, Jackson, Mississippi; and Pennington Biomedical Research Center, Baton Rouge, Louisiana
Correspondence: Dr. Michael F. Flessner, Division of Nephrology, Department of Medicine, University of Mississippi Medical Center, 2500 North State Street, Jackson, MS 39216-4505. Phone: 601-984-5670; Fax: 601-984-5765; E-mail: mflessner{at}umsmed.edu
Received for publication December 30, 2006.
Accepted for publication April 15, 2007.
Most current animal models that are used to study effects oflong-term peritoneal exposure to dialysis solutions use an indwellingcatheter for daily injections. It was hypothesized that thepresence of a foreign body in the peritoneal cavity (PC) mightalter the inflammatory response to the solutions and that theresponse would depend on exposure duration. For addressing these,long-term injections were carried out for 2 to 8 wk in 90 Sprague-Dawleyrats: 40 via a subcutaneous port connected to a silicone cathetertunneled to the PC, 40 via direct needle injection, and 10 noninjected,age-control rats. Daily volumes were 30 to 40 ml of filter-sterilized,bicarbonate-buffered solutions that contained 4% dextrose. After2, 4, 6, and 8 wk, anesthetized rats underwent transport experimentswith a chamber affixed to the abdominal wall to determine masstransfer coefficients of mannitol (MTCmannitol) and albumin(MTCBSA), osmotic filtration flux (Josm), and hydrostatic pressure–drivenflux. After the rats were killed, tissues were collected formeasurement of peritoneal thickness, vascular density, and immunohistochemicalstaining. ANOVA demonstrated significant (P < 0.01) differencesin thickness, vessel density, MTCmannitol, and MTCBSA amongthe groups at the various time intervals and in overall means.Differences among the groups were less pronounced for hydrostaticpressure–driven flux and Josm. Vessel density, MTCmannitol,MTCBSA, and Josm were dependent on injection duration (P <0.01). There were marked differences between the needle injectionand catheter injection groups at various intervals in the expressionof three cytokines. It is concluded that the histologic andfunctional response depends on the duration of injection withanimals that are exposed for as little as 2 wk demonstratingalterations. These findings confirm the hypothesis that thepresence of a PC catheter increases inflammatory response tosterile solutions as evidenced by the structural and functionalchanges in the peritoneal barrier.
Exposure of the peritoneum to sterile solutions that containglucose results in changes in cellular characteristics, histology,and barrier function. Long-term peritoneal dialysis (PD) withoutevidence of peritonitis alters the phenotypic appearance ofthe human peritoneum in 8 to 12 mo.1 Biopsy studies of long-term(>3 yr) PD patients have demonstrated marked thickening overtime of the subcompact zone, a region between the mesothelialcells and their basement membrane, with corresponding changesin the underlying microvasculature.2 Exposure of patients toglucose-based dialysis solutions resulted in significant functionalchanges in transperitoneal transport in a 2-yr study of patientswho were treated with automated PD.3 Studies in animal modelsparallel these changes and have demonstrated benefit of solutionsthat are thought to be more biocompatible.4–6
Studies in our laboratory using an animal model with a subcutaneousinjection port connected to a silicone catheter tunneled tothe cavity demonstrated angiogenesis and thickening of the peritoneumafter 8 wk of daily exposure to sterile solutions.7 A portionof the control group was made up of animals that had a catheterbut were exposed only to a small volume of a dilute heparinsolution (10 µ/ml isotonic Krebs bicarbonate) once a week.Despite having no exposure to clinical or glucose solutions,the animals with catheters did show minor structural alterationsin the peritoneum. A recent study by Musi et al.,8 who performeddaily intraperitoneal injections without a catheter, demonstratedfar less striking differences in the peritoneum after exposureto various dialysis solutions. This suggested that experimentalfindings in animal models might depend on the presence of acatheter.
Catheters are foreign bodies inserted through the abdominalwall into the cavity. Bacteria as well as human cells are knownto grow in a biofilm and coat these catheters.9 Although biofilmis often associated with an infected catheter, careful observationsin PD patients have demonstrated severe biofilm formation withoutdetectable infection.10 We therefore hypothesized that the cathetermay have a significant effect on the inflammatory process inthe peritoneum, as evidenced by the thickness of the peritoneumand the vascular density. We further hypothesized that inflammatorychange might produce different effects that depend on the durationof chronic exposure to PD solutions. Variable structural changeswould lead to changes in transperitoneal transport of solutesand water. We therefore carried out careful studies of dailyinjection of the same sterile glucose solution for 2, 4, 6,and 8 wk in rats with and without catheters to study the earlymanifestations of the inflammatory process. The results supportour hypotheses.
Bacterial Cultures and Cell Counts after Injection Period Table 1 shows the number of positive cultures and the numberof rats analyzed in each group (n = 10 rats per injection groupper duration of injection). Surprising, the age-control (AC)group, which was not administered an injection except just beforethe transport experiment, had three positive cultures. Althoughspontaneous bacterial peritonitis cannot be ruled out, the cellcounts in each of these rats was low (mean 71 cells/mm3) andthe positive cultures were likely the result of accidental bowelperforation at the intraperitoneal injection, just before thetransport experiments. A relatively low cell count was alsoobserved for nearly half of the rats that received a needleinjection (NI) and whose culture was positive. All culture-positiverats were deleted from the analysis. No adhesions were notedin any of the rats, including those with positive cultures.
Table 1. Number of infected rats from each experimental group (positive cultures) and the number of rats analyzeda
Figure 1 displays the total white blood cell (WBC) count foreach group and the fraction of neutrophils in each WBC. Exceptfor the 8-wk groups, the catheter injection (CI) group had higherWBC count than the NI group, but the differences were statisticallynonsignificant. Fractions of neutrophils were greater in theCI group and were statistically different from the NI groupin the two-way ANOVA (P < 0.0001). The overall means forthe CI and NI groups were significantly higher that that ofthe AC group (P < 0.0001).
Figure 1. Left axis: Total white blood cell count per mm3 of the peritoneal fluid just before transport experiment versus the duration in weeks of daily injection. Mean values are the time-averaged means of each injection group (, needle injection [NI]; , catheter injection [CI]) and that of the age-control (AC) group (). Right axis: The fraction of neutrophils (PMN fraction) versus the NI group () and the CI group () and the AC group (). *Significance in the two-way ANOVA between groups (P < 0.0001); +significant difference among the overall means of groups for the PMN fractions (one-way ANOVA, P < 0.0001).
Structural Changes in the Peritoneum Figure 2 displays the means ± SEM of the thickness ofthe submesothelial compact zone for NI rats, CI rats, and ACrats. The mean thickness for the NI rats was not significantlydifferent from that for the AC rats. However, the thicknessfor the CI rats was significantly different (P < 0.0001)from that for the NI rats and approximately two times that forthe AC rats (P < 0.0001).
Figure 2. Peritoneal thickness (µm) versus duration of injection and overall means. , NI; , CI; , AC. The differences between CI and NI were significant (*P < 0.0001, two-way ANOVA). The overall means were significantly different as well in a one-way ANOVA (P < 0.0001).
Figure 3 shows the vascular densities of the subcompact zonefor each group in terms of number of vessels per linear millimeterof peritoneum. When the peritoneal vessel number (CD31 staining)per square millimeter of peritoneum was plotted versus durationof injection, the plot was the same as in Figure 3, with significantdifferences between the CI and NI groups (data not shown, P< 0.0001, two-way ANOVA). Overall means for CI rats are significantlygreater than those for either NI or AC rats (P < 0.0001).At each time interval, the CI rats have a greater vessel numberthan the NI rats. The differences between the injection groupsare statistically significant (P < 0.0001). Within the NIgroup or the CI group, duration of injection affected the vesselcount per linear millimeter of peritoneum (P < 0.03) butnot the vessel density per square millimeter. Both NI and CIrats had at all times significantly higher number of vesselsthan AC rats (P < 0.002).
Figure 3. Peritoneal vessel number (CD31 staining) per linear millimeter of peritoneum versus duration of injection. , NI; , CI; , AC. The differences between CI and NI were significant (*P < 0.0002, two-way ANOVA). The variation within injection groups versus duration of injection was significant (#P < 0.03, two-way ANOVA). The overall means were significantly different as well in a one-way ANOVA (+P < 0.0001).
Immunohistochemical Studies of the Peritoneum
Immunohistochemical stains of the various groups demonstratedifferent patterns, which depended on the presence of a catheter.Table 2 is the semiquantitative measurement of the stains ineach group, and Figure 4 shows typical images of the three cytokines.Both the CI and NI groups demonstrated significantly greaterstaining than untreated AC rats (P < 0.005) for each of thecytokines. In a two-way ANOVA, the CI rats had increased stainingover the NI rats (P < 0.02), but within each injection group,there was no significant trend in staining for basic fibroblastgrowth factor (bFGF), vascular endothelial growth factor (VEGF),and TGF-1 with duration of injection.
Figure 4. Comparison of vascular endothelial growth factor (VEGF; A), basic fibroblast growth factor (bFGF; B), and TGF- (C) staining (brown) in CI and NI groups at 2, 4, 6, and 8 wk. Bars = 100 µm. AC rats did not receive injections and were held for 8 wk with subsequent transport experiments and tissue sampling. All tissue samples are from rats with negative fluid and catheter cultures and were sampled at a location remote from the location of the catheter. Magnification, x400.
Solute Transport versus Treatment Group Figure 5 displays the mass transfer coefficients for mannitol(MTCmannitol). Within either the CI or NI group, there weresignificant trends with respect to duration of injection (P< 0.02). However, MTCmannitol from the NI group was significantlyless than that of the CI group (P < 0.0001) at each timeinterval. The overall means of each group were significantlydifferent as well (P < 0.0001).
Figure 5. Mass transfer coefficients of mannitol (MTCmannitol; cm/min x 103) versus duration of injection. , NI; , CI; , AC. The differences between CI and NI were significant (*P < 0.0001, two-way ANOVA). The variation within injection groups versus duration of injection was significant (#P < 0.02, two-way ANOVA). The overall means were significantly different as well in a one-way ANOVA (+P < 0.0001).
Figure 6 shows the mass transfer coefficients for BSA (MTCBSA).The CI groups had higher mean values at each interval than theNI groups (P < 0.003). Within each injection group, therewas a significant decrease in magnitude (P < 0.0001) withduration of injection, particularly between weeks 2 and 4. Theoverall means demonstrated a significant difference (P <0.002) with magnitudes of injected rats being greater than thoseof AC rats.
Figure 6. MTCBSA (µl/min per cm2) versus duration of injection. , NI; , CI; , AC. The differences between CI and NI were significant (*P < 0.003, two-way ANOVA). The variation within injection groups versus duration of injection was significant (#P < 0.0001, two-way ANOVA). The overall means were significantly different as well in a one-way ANOVA (+P < 0.002).
Osmotic and Hydrostatic Pressure Effects on Transperitoneal Fluid Flow Figure 7 displays the fluid flux as a result of osmotic pressureexerted in the test chamber for each injection group and eachtime interval. Within the NI and CI groups, there was a decreasingand significant trend in osmotic filtration flux (Josm) versusduration of injection (P < 0.003). Within time injectionintervals, there were no significant differences; neither werethere significant differences among the overall means for AC,CI, and NI rats. Solution osmolalities varied between 477 and493 mOsm/kg among all of the groups.
Figure 7. Osmotic flux (µl/min per cm2) versus duration of injection. , NI; , CI; , AC. The variation within injection groups versus duration of injection was significant (#P < 0.003, two-way ANOVA). The overall means were not significantly different.
Fluid flow across the peritoneum as a result of hydrostaticpressure is displayed in Figure 8. Although there was a decreasein mean flux from weeks 2 to 4 in each injection group, thetrends in each group were not statistically significant. Therewere no differences between injection groups at each injectioninterval. The overall means of the CI and NI groups were significantlydifferent from AC rats (P < 0.0001).
Figure 8. Hydrostatic pressure–driven flux (µl/min per cm2) versus duration of injection. , NI; , CI; , AC. Neither the variation between injection groups nor that within injection groups versus duration of injection was significant. The overall means were significantly different in a one-way ANOVA (+P < 0.0001).
Effect of Catheter on Peritoneal Structure and Its Transport Properties
Significant differences in peritoneal structure, histology,and function were noted between the two injection techniques.The peritoneal thickness, vascular density, and transport ofboth small solutes and albumin revealed marked differences betweenNI and CI rats. There were higher mean values for total WBCin the CI rats with a significantly higher fraction of neutrophilsthan in the NI group. Cytokine staining was more intense inCI rats during the first 4 wk of solution exposure. Functionally,the MTCmannitol and the MTCBSA were higher in the CI group thanin the NI group. The solution was designed to have a high concentrationof glucose (4%), high osmolality and low glucose degradationproducts (presumed but not measured), and neutral pH. The solutionresulted in inflammation in both animal models, but no statementscan be made concerning direct effects of the makeup of the solutionbecause there was no comparison of different solutions. Becausethe same solution was used in each group, the different resultspoint to the catheter as the cause of the variation in measurements.
The catheter seems to set up inflammation before the infusionsof the solution and may interact with the solution to enhanceperitoneal inflammation. Isolated leucocytosis with an elevatedfraction of neutrophils has been found after catheter implantationin humans.11 We observed this in earlier experiments7 in animalswith the same catheter but no infusions other than a weeklyinfusion of dilute heparin. Although heparin infusions are consideredto be anti-inflammatory,12,13 the use of a heparinized catheterdecreased bacterial colonization but not the degree of inflammationor peritonitis rate over 4 wk of twice-daily exchanges.14 Injectionof isotonic, buffered solutions has been observed to promoteinflammation in the abdominal wall peritoneum15 but to preservecapillary morphology in the omentum16; the latter study wasperformed with heparin in the solution, which could have influencedthe inflammatory response.
There is an abundant literature on the effects of bacterialbiofilm on catheters and its link to peritonitis.9,17 The catheterwas cultured with the fluid in each rat at the end of the injectionperiod, and only culture-negative rats were used in the reportedstudies. Although the 48-h culture should have produced bacterialcolonies if planktonic bacterium was present, most biofilmsrequire electron microscopy or specialized confocal microscopyfor their study.18 Clinical studies have reported that biofilmformation does not require the presence of infection,10 butour study did not include examination of the catheters withelectron microscopy.19 Without more detailed study, we can onlyassume that the catheters resulted in a greater degree of structuralchanges in the peritoneum and parallel alterations in transportrates of solutes than the corresponding NI rats. It is possiblethat different catheter materials may vary the inflammatoryresponse, but test of bacterial adherence to three biomaterialsurfaces demonstrated differences between organisms rather thanmaterials.20 These findings suggest that careful examinationof the animal models that are used to test biocompatibilityof solutions is necessary for correct interpretation of theeffects that can be exclusively attributed to a particular solutionon the peritoneum and its barrier function.
The catheter that we used was made of silicone and designedfor intravenous insertion. We used sterile technique to implantthe catheter and waited until the skin over the titanium portwas completely healed before initiating injection. Use of thecatheter before this often resulted in wound dehiscence.7 Thatthe catheter was in the peritoneal cavity for 3 to 4 wk witha weekly flush of heparin and saline likely set up a reactionin the peritoneum to the commencement of the daily injections.We did not determine the condition of the peritoneum or itstransport function before the injections. Therefore, we cannotwith absolute certainty quantify its interactive effect withthe solution on our results. However, it is clear from the 2-and 4-wk results that the processes of peritoneal thickeningand angiogenesis in the subcompact zone and inflammation wereenhanced in the CI rats when compared with the NI rats. We notethat many of our patients undergo catheter placement with aweekly heparin flush for 3 to 4 wk before their training todo PD.
We speculate that the catheter is a platform for the activationof mesothelial cells that are chronically shed from their basementmembrane and their transition from an epithelial phenotype toa fibroblastic type with amplification of cytokine productionwithin the cavity.1 Inflammatory substances that are secretedby these cells would be absorbed through the peritoneum duringthe 24 h between injections. In this model, that there are noexchanges of dialysate fluid to "wash out" these mediators mayexacerbate the inflammatory response in this animal model.
Time Course of Inflammatory Reaction from Sterile Solutions
A major goal was the study of the early time course of the inflammatoryreaction of the peritoneum to the sterile solution. The solutionwas carefully prepared and filtered to minimize glucose degradationproducts, which arise in the heat sterilization of glucose solutions.21The solution was refrigerated and cultured before and afteruse; none of these cultures grew bacteria. Before injection,the solution was handled in a laminar airflow hood and injecteddaily into anesthetized rats to investigate their effect onthe structure and function of the peritoneal barrier. As illustratedin Figure 1, there was no clear pattern of WBC versus time,and the total WBC of the CI and NI groups were not significantlydifferent from that of the AC group.
The transport data and immunohistochemistry did display significantchanges with time. Presumably because of the presence of thecatheter for 3 to 4 wk before long-term injections, the CI grouphad marked differences from the NI group in the progressionof bFGF, VEGF, and TGF-1, with the NI rats lagging 4 wk behindthe CI group. In the 2-wk groups, the CI rats displayed thegreatest thickness, which subsequently decreased with time andon the average was significantly greater than that of AC rats.The mean thickness in the NI rats did not vary significantlywith duration of injection; neither did these rats have a significantdifference from the AC rats (see Figure 2). This again demonstratesthe effect of the catheter. In the NI group, the vascular densityincreased between 4 and 6 wk, whereas the CI group showed adecrease after the initial 2-wk period, demonstrating possibledifferences in cytokine expression versus time. MTCmannitol,MTCBSA, and Josm also tended to decrease in magnitude afterthe first 2 to 4 wk and had significant trends over the 8 wk.This pattern was also seen in the fluid fluxes and MTCBSA ofthe NI rats. Both injection groups demonstrated significantdifferences in vascular density, solute transport, and cytokinestaining from noninjected AC rats.
Characterization of Animal Models
Comparison of our results with studies from other laboratoriesis made difficult because of the different techniques of injection,including volume and frequency, and because of the method ofmeasuring transport via a chamber instead of the whole cavity.As far as we know, no one group has performed the side-by-sidecomparison of CI versus NI contained in this article. It isinteresting to note the varying results from two other laboratories.One group performed twice-daily NI of commercial solutions withand without glucose degradation products (GDP); they observedno significant differences in solute transport, ultrafiltration,or tissue thickness but showed higher vascularity with the morebiocompatible solution.8 A second group implanted polyurethanecatheters connected to a subcutaneous port and subsequentlyadministered injection to the animals twice daily for 12 to20 wk; they observed improved ultrafiltration and less angiogenesiswith the more biocompatible solutions.6 From these two studies,the reader arrives at completely different conclusions concerningthe effects of low-GDP, more "biocompatible" solutions. Ourstudy with a presumed low-GDP solution points to the presenceof a catheter, or "foreign body," in the cavity as an additionalfactor in the immune response of the peritoneum and suggeststhat a broader examination of interactions between the catheterand immune elements of the peritoneal cavity are required forfull characterization of the usefulness of relatively expensive,"biocompatible" solutions.
Animals and Experimental Groups
Sprague-Dawley female rats (200 to 300 g; Charles River Laboratories,Wilmington, MA) were divided into three groups. One group ofrats (NI) received direct intraperitoneal injections daily fordurations of 2, 4, 6, or 8 wk. The other group of rats (CI)received daily peritoneal injections through a subcutaneousport connected to a peritoneal catheter for durations of 2,4, 6, or 8 wk. A third group was used as age controls (AC) andfollowed for 8 wk. The relative age of the rats in each groupwas similar. No animal lost weight during the period of injection.Average weight gains in both groups were 21 g at 2 wk, 30 gat 4 wk, 43 g at 6 wk, and 45 g at 8 wk.
The experimental protocols for this study were approved by theInstitutional Animal Care and Use Committee of the Universityof Mississippi Medical Center and were carried out in accordancewith both the Guide for the Care and Use of Laboratory Animalsfrom the National Institutes of Health and the guidelines ofthe Animal Welfare Act.
Catheter Implantation
Aseptic techniques were used to implant a catheter for long-termPD in each rat that was to receive the daily injections throughthe catheter. The peritoneal catheter was placed in the lowerright quadrant of the abdomen and connected to a subcutaneouschamber (Rat-o-Port CP6-0S; Access Technologies, Skokie, IL),as described in our previous publication.7 The chamber is madefrom implant grade 4 titanium, and the silicone tubing is thesame as is used in human applications (Dow Corning Medical-GradeSilicone Rubber, Corning, NY). The material is of similar makeupas the Silastic silicone rubber, which is used in silicone dialysiscatheters and intravenous catheters for human use. The dailyinjections of dialysis solution were not started until the skinwound over the port was completely healed (usually 3 to 4 wkafter implantation, which is not unlike the placement of a catheterin humans). Heparin (10 U in 1 ml of PBS) was injected at weeklyintervals during this healing period to maintain catheter patency.
Daily Injection of Dialysis Solution
To minimize animal activity that might increase the potentialfor contamination during this process, each rat was anesthetizedwith isoflurane (1 to 3%, 3 to 5 min daily exposure). Each ratwas given 30 to 40 ml of sterile dialysis solution daily forup to 8 wk. The volume of PD solution used in these studieswas selected by scaling the typical volume used in humans (approximately2 L) by the factor (body weight).7,22 The rationale and justificationfor this scaling criteria are detailed in by Dedrick and colleagues.22,23From previous experiments, this volume will touch the entireperitoneum during the 24 h between injections, but the contactarea at any one time will be a maximum of approximately 40%of the anatomic peritoneum.7,24,25 The peritoneal pressuresobserved with this volume are well within the range seen inhumans.26
The solution that was administered for PD was prepared weeklyfrom Krebs-Ringer solution with 4% glucose (average osmolarity480 to 500 mOsm). The composition of the Krebs-Ringer solutionwas as follows (in mM): 120 NaCl, 10 KCl, 2 CaCl2, 25 NaHCO3,0.28 KH2PO4, and 1.2 MgSO4. The dialysis solution was sterilefiltered (0.2-µm filter) in a laminar-flow hood and storedat 4°C until use. The solution was presumed to be low inGDP and to be more biocompatible than commercial solutions,which are typically heat-sterilized with a low pH.8 Bottlesof dialysis solution were handled throughout with aseptic technique,and syringes were filled within a laminar-flow hood just beforeinjection. Each solution was cultured for bacteria at the firstuse of the bottle and at the end of each bottle. All solutioncultures were negative.
Needle Injection.
The lower abdominal wall was prepared with alcohol and the solutionwas administered to rats in the NI group by direct intraperitonealinjection with a syringe and 23-G needle. Skin puncture waslimited to the lower quadrants of the abdomen, with the leftand right lower quadrants used alternately for the daily injections.Care was taken to avoid puncture of the intestine or subcutaneousinjection of the dialysis solution. Injection sites were inspecteddaily. One rat developed a subcutaneous hematoma, which resolvedduring a 7-d period and did not affect the peritoneal cavity.
Catheter Injection.
Dialysis solution was administered through the subcutaneousport to rats in the CI group. The skin over the port was sterile-preparedwith alcohol, and Huber needles were used to inject the solution.Tissue around the injection port was examined carefully forany sign of swelling or scabbing as a result of infection orbleeding.
Each rat was carefully monitored at daily injections and weighedweekly. Weight loss (>10%), abnormal activity (lack of movementin the cage), or development of a hematoma or apparent infectionat the injection site or port was noted in a written log.
Bacterial Cultures before Transport Studies
After the requisite duration (2, 4, 6, or 8 wk) of daily injections,each rat was anesthetized by intramuscular injection of sodiumpentobarbital (60 mg/kg). After sterile preparation of the abdomen,30 ml of isotonic sterile Krebs solution was injected intraperitoneally.After 10 min, using sterile technique, an incision was madein the midline of the abdominal wall. The residual peritonealfluid and catheter (if appropriate) were collected for cellcount and bacterial culture. Cells were stained with Wright'sstain and counted with a hemacytometer. Fluid was cultured onTSA II 5% SB plates (Becton Dickinson Microbiology Systems,Cockeysville, MD) for 48 h at 37°C. Positive bacterial cultureswere identified by inspection of the plates.
Transport Studies
Polyethylene catheters (PE-50) were placed in the femoral veinto maintain anesthesia with additional pentobarbital and inthe femoral artery for collection of blood samples during thetransport studies. Temperature was maintained at 37°C, andBP was monitored to ensure that the mean pressure remained >80mmHg (average mean arterial pressures varied between 90 and110 mmHg).
For carrying out transport studies of solutes and water acrossthe peritoneum, a plastic chamber was affixed to the serosaof the abdominal wall to examine the transport characteristicsof the peritoneum using the procedures described previously.27,28The chamber was affixed to the upper left quadrant of the abdominalwall remote from the location of the catheter tip or needlepunctures. The chamber removes the peritoneal contact area asa variable and therefore provides a measure of the intrinsicmass transfer characteristics of selected portions of the peritoneum;recent data in rodents have shown similitude in the measurementswith the chamber method in mice and rats.29 Because the peritonealcontact area may be altered by inflammation, there may not bea direct correlation between chamber measurements and transportstudies performed with the traditional whole-cavity technique.
Small Solute Transport and Osmotic Filtration.
The osmotic water flux from the blood flowing through the tissueinto the chamber and the transport of a small solute, [14C]mannitol,from the chamber into the tissue were measured for a periodof 120 min. The MTC was calculated from the experimental databy our previously published method.27,28
Protein Transport.
Transport of protein from the blood into the chamber was determinedduring the same 120-min period by intravenous injection of FITC-labeledBSA (FITC-BSA) as a bolus (40 mg/ml). Changes in the plasmaconcentration and chamber concentration were measured duringthe 120-min period. The albumin flux and MTC were calculatedfrom these measurements as described previously.7
Hydrostatic Pressure–Driven Convection.
The chamber was washed with Krebs-Ringer solution and then filledto a depth of 6 cm with Krebs-Ringer solution that contained5% BSA, 0.05% Evans blue dye, and 125I-IgG (1 µCi/ml)to determine hydrostatic pressure–driven water transport.The volume of the chamber solution and the concentration of125I-IgG were measured every hour for a period of 3 h. At theend of experiment, the chamber was removed and the tissue underthe chamber was identified from the staining with Evans bluedye. Tissue from this region was recovered, and its radioactivecontent was measured to estimate transport of 125I-IgG intothe tissue. The hydrostatic pressure–driven flux of waterwas calculated from these measurements as described previously.7All calculations of transport rates were performed as previouslypublished.7,29
Materials
The tracer molecule that was used for small solute experimentswas 14C-mannitol purchased from Moravek Biochemicals (Brea,CA), and the tracer was stated to be at least 97% pure by themanufacturer and was used as received. 14C-mannitol was detectedby liquid scintillation (Packard Tricarb 2500TR, Ramsey, MN).
FITC-BSA was purchased from Sigma Chemicals (St. Louis, MO)and was used as delivered. Fluorescence was detected with aTurner TD700 spectrophotometer. Checks of label purity7 havedemonstrated that there is no detectable separation of the fluorescencelabel from the protein.
125I-IgG (anti-rabbit, immunoabsorbed for human and rat antigens)was purchased from Amersham (Piscataway, NJ) and purified dailyaccording to our previously published method.30 Detection wascarried out by counting (Packard Cobra II Auto-Gamma Counter).
Histology and Image Analysis
As in our previous publication,7 trichrome staining was usedfor determinations of the submesothelial compact zone thickness,and CD31 immunohistochemistry was used to stain endothelialcells for determination of apparent vascular density.
All quantitative measurements were averaged from observationsof three independent observers. Samples of abdominal wall tissuewere collected in a standardized manner for the evaluationsto minimize variation as a result of the sampling process. Sampleswere taken from the right upper quadrant of the abdominal wallopposite the site for the transport studies and remote fromthe location of the CI or NI so as to exclude effects from mechanicalirritation. Three to five horizontal sections from the midportionof the abdominal wall were prepared. Sections of peritonealtissue were excised and immediately fixed in neutral pH-buffered4% formalin solution and stained as described previously.31
Expression of growth factors was evaluated using standard immunohistochemistryprotocols. Antibodies to VEGF, bFGF, and TGF-1 were obtainedfrom Santa Cruz Biotechnology (Santa Cruz, CA). After deparaffinizationand rehydration, tissue sections were processed for quenching(treatment with 3% hydrogen peroxide for 30 min), antigen retrieval(steam heating sections for 30 min in Ag Retrieval Citra Plusbuffer; BioGenex, San Ramon, CA), and biotin blocking (DakoCytomation,Carpinteria, CA). Incubation with blocking solution (VectaStainsystem [Vector Laboratories, Burlingame, CA] or Protein BlockSerum-Free, Avidin-Biotin Block system [Dako, Carpinteria, CA])was performed to block nonspecific binding. Sections were incubatedwith primary antibody at 4°C overnight and processed fortreatment with the secondary antibody after washing with buffersolution. An avidin-biotinylated horseradish peroxidase reagentand 3,3'-diaminobenzidine tetrahydrochloride were used for detection.The sections were counterstained with hematoxylin. The degreeof tissue staining with antibodies to these growth factors wasevaluated as follows: No stain, 0; minimal stain, +1; moderatestain, +2; large stain, +3; and intense stain, +4. At leastfive sections of each group were graded and averaged for thesingle score.
Statistical Analyses
Data are presented as the means ± SEM. Data from ratswith positive bacterial cultures were excluded from the analysis.Statistical analyses of effects of different groups and theduration of injection were performed using two-way ANOVA. Overallmeans for each injection group and the AC group were comparedwith a one-way ANOVA. The probability of a type I error of P< 0.05 was considered significant. All tests were performedusing NCSS-97 (Kayesville, UT).
Yanez-Mo M, Lara-Pezzi E, Selgas R, Ramierez-Huesca M, Dominguez-Jimenez C, Jimenez-Heffernan JA, Aguilera A, Sanchez-Tomero JA, Bajo MA, Alvarez V, Castro MA, del Peso G, Cirujeda A, Gamallo C, Sanchez-Madrid F, Lopez-Cabrera M: Peritoneal dialysis and epithelial-to-mesenchymal transition of mesothelial cells.
N Engl J Med 348
: 403
–413, 2003[Abstract/Free Full Text]
Williams JD, Craig KJ, Topley N, Ruhland CV, Fallon M, Newman GR, Mackenzie RK, Williams GT: Morphologic changes in the peritoneal membrane of patients with renal disease.
J Am Soc Nephrol 13
: 470
–479, 2002[Abstract/Free Full Text]
Davies SJ, Brown EA, Frandsen NE, Rodrigues AS, Rodriguez-Carmona A, Vychytil A, MacNamara E, Ekstrand A, Tranaeus A, Filho JCD: Longitudinal membrane function in functionally anuric patients treated with APD: Data from EAPOS on the effects of glucose and icodextrin prescription.
Kidney Int 67
: 1609
–1615, 2005[CrossRef][Medline]
Hekking LH, Aalders MC, Van Gelderop E, Zweers MM, Struijk DG, Havenith CE, Beelen RH: Effect of peritoneal dialysis fluid measured in vivo in a rat-model of continuous peritoneal dialysis.
Adv Perit Dial 14
: 14
–18, 1998[Medline]
Jones S, Holmes C, Krediet RT, Mackenzie RK, Faict D, Tranaeus A, Williams JD, Coles GA, Topley N: Bicarbonate/lactate-based peritoneal dialysis solution increases cancer antigen 125 and decreases hyaluronic acid levels.
Kidney Int 59
: 1529
–1538, 2001[CrossRef][Medline]
Mortier S, Faict D, Lameire NH, deVriese AS: Benefits of switching from a conventional to a low-GDP bicarbonate/lactate-buffered dialysis solution in a rat model.
Kidney Int 67
: 1559
–1565, 2005[CrossRef][Medline]
Flessner MF, Choi J, Vanpelt H, He Z, Credit K, Henegar J, Hughson M: Correlating structure with solute and water transport in a chronic model of peritoneal inflammation.
Am J Physiol Renal Physiol 290
: F232
–F240, 2006[Abstract/Free Full Text]
Musi B, Braide M, Carlsson O, Wieslander A, Albrektsson A, Ketteler M, Westenfeld R, Floege J, Rippe B: Biocompatibility of peritoneal dialysis fluids: Long-term exposure of nonuremic rats.
Perit Dial Int 24
: 37
–47, 2004[Abstract/Free Full Text]
Dasgupta M: Biofilms and infection in dialysis patients.
Semin Dial 15
: 338
–346, 2003[CrossRef]
Swartz R, Messana J, Holmes C, Williams J: Biofilm formation on peritoneal catheters does not require the presence of infection.
ASAIO J 37
: 626
–634, 1991
Fijen J, Struik DG, Krediet RT, Boeschoten E, deVries J, Arisz L: Dialysate leucocytosis in CAPD patients without clinical infection.
Neth J Med 33
: 270
–280, 1988[Medline]
de Vriese A, Mortier S, Cornelissen M, Paniagua R, Vanacker NJ, Leyssens A, Faict D, De Ridder L, Lameire N: The effects of heparin administration in an animal model of chronic peritoneal dialysate exposure.
Perit Dial Int 22
: 566
–572, 2002[Abstract]
Mortier S, Lameire NH, deVriese AS: The effects of peritoneal dialysis solutions on peritoneal host defense.
Perit Dial Int 24
: 123
–138, 2004[Abstract]
Kim Y, Cho S, Kim J, Cho D-K, Kim Y-J, Larm O, Mathiesen T, Lindholm B, Bergstrom J: Effect in a rat model of heparinized peritoneal dialysis catheters on bacterial colonization and the healing of the exit site.
Perit Dial Int 21[Suppl 3]
: S357
–S358, 2001
Choi J, Credit K, Henderson K, Deverkadra R, Vanpelt H, Fless MF: Antibiotic prophylaxis in an animal model of chronic peritoneal exposure.
Perit Dial Int 26
: 249
–258, 2006[Abstract/Free Full Text]
Zweers MM, Splint LJ, Krediet RT, Struijk DG: Ultrastructure of basement membranes of peritoneal capillaries in a chronic peritoneal infusion model in the rat.
Nephrol Dial Transplant 16
: 651
–654, 2001[Abstract/Free Full Text]
Dasgupta M, Costerton J: Significance of biofilm-adherent bacterial microcolonies on Tenckhoff catheters of CAPD patients.
Blood Purif 7
: 144
–155, 1989[Medline]
Dasgupta M: Biofilms in peritoneal dialysis.
Perit Dial Int 21[Suppl 3]
: S213
–S217, 2001
Reid G, Khoury AE, Preston CA, Costerton JW: Influence of dextrose dialysis solutions on adhesion of Staphylococcus aureus and Pseudomonas aeruginosa to three catheter surfaces.
Am J Nephrol 14
: 37
–40, 1994[Medline]
Witowski J, Bender TO, Wisniewska-Elnur J, Ksiazek K, Passlick-Deetjen J, Breborowicz A, Jorres A: Mesothelial toxicity of peritoneal dialysis fluids is related primarily to glucose degradation products, not to glucose per se.
Perit Dial Int 23
: 381
–390, 2003[Abstract/Free Full Text]
Dedrick RL: Interspecies scaling of regional drug delivery.
J Pharm Sci 75
: 1047
–1052, 1986[CrossRef][Medline]
Dedrick RL, Flessner MF, Collins JM, Schultz JS: Is the peritoneum a membrane?
ASAIO J 5
: 1
–5, 1982
Flessner MF, Lofthouse J, Williams A: Increasing peritoneal contact area during dialysis improves mass transfer.
J Am Soc Nephrol 12
: 2139
–2145, 2001[Abstract/Free Full Text]
Flessner MF, Lofthouse J, Zakaria EL: Improving contact area between the peritoneum and intraperitoneal therapeutic solutions.
J Am Soc Nephrol 12
: 807
–813, 2001[Abstract/Free Full Text]
Flessner MF, Schwab A: Pressure threshold for fluid loss from the peritoneal cavity.
Am J Physiol 270
: F377
–F390, 1996[Medline]
Flessner MF: Small-solute transport across specific peritoneal tissue surfaces in the rat.
J Am Soc Nephrol 7
: 225
–233, 1996[Abstract]
Flessner MF, Deverkadra R, Smitherman J, Li X, Credit K: In vivo determination of diffusive transport parameters in a superfused tissue.
Am J Physiol Renal Physiol 291
: F1096
–F1103, 2006[Abstract/Free Full Text]
Flessner MF, Credit K, Li X, Tanksley J: Similitude of transperitoneal permeability in different rodent species.
Am J Physiol Renal Physiol 292
: F495
–F499, 2007[Abstract/Free Full Text]
Flessner MF, Choi J, He Z, Credit K: Physiological characterization of human ovarian cancer cells in a rat model of intraperitoneal antineoplastic therapy.
J Appl Physiol 97
: 1518
–1526, 2004[Abstract/Free Full Text]
Flessner MF, Choi J, Vanpelt H, He Z, Credit K, Henegar J, Hughson M: Correlating structure with solute and water transport in a chronic model of peritoneal inflammation.
Am J Physiol Renal Physiol 290
: F232
–F240, 2006[Abstract/Free Full Text]
Top
Abstract
Introduction
1 with duration of injection. 
4 wk behind the CI group. In the 2-wk groups, the CI rats displayed the greatest thickness, which subsequently decreased with time and on the average was significantly greater than that of AC rats. The mean thickness in the NI rats did not vary significantly with duration of injection; neither did these rats have a significant difference from the AC rats (see Figure 2). This again demonstrates the effect of the catheter. In the NI group, the vascular density increased between 4 and 6 wk, whereas the CI group showed a decrease after the initial 2-wk period, demonstrating possible differences in cytokine expression versus time. MTCmannitol, MTCBSA, and Josm also tended to decrease in magnitude after the first 2 to 4 wk and had significant trends over the 8 wk. This pattern was also seen in the fluid fluxes and MTCBSA of the NI rats. Both injection groups demonstrated significant differences in vascular density, solute transport, and cytokine staining from noninjected AC rats. 
counting (Packard Cobra II Auto-Gamma Counter). 
, needle injection [NI]; 
, catheter injection [CI]) and that of the age-control (AC) group (
). Right axis: The fraction of neutrophils (PMN fraction) versus the NI group (
) and the CI group (
) and the AC group (
). *Significance in the two-way ANOVA between groups (P < 0.0001); +significant difference among the overall means of groups for the PMN fractions (one-way ANOVA, P < 0.0001).
