Mice that Lack Endothelial Nitric Oxide Synthase Are Protected against Functional and Structural Modifications Induced by Acute Peritonitis

Animals, Experimental Groups, and Acute Peritonitis Model

Experiments were conducted using male C57 BL/6J mice (Iffa Credo, Brussels, Belgium). The C57 BL/6J eNOS^+/+ (wild-type [WT]) and eNOS^−/− (knockout [KO]) mice were generated as described previously (17) and obtained from the Jackson Laboratory (Bar Harbor, ME). Male Wistar rats, aged 8 to 10 wk, were used for comparative studies (6,10,14⇓⇓). All animals had access to appropriate standard diet and tap water ad libitum. The experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the local Ethics Committee.

A group of 12 control mice was used to characterize the permeability parameters at the basal state, in comparison with those of control rats (n = 12) and mice with acute peritonitis (n = 6). Another group of mice with acute peritonitis (n = 6) was used to assess the differential leukocyte counts in the dialysate at day 1, day 3, and day 7. In a second series of experiments, 11 pairs of eNOS WT and KO mice matched for age were investigated in control conditions (n = 5) or after acute peritonitis (n = 6). Acute peritonitis was generated by insertion of a peritoneal catheter as described previously (6,14⇓). At day 0, mice were anesthetized with ketamine (100 mg/kg subcutaneously; Merial, Brussels, Belgium) and xylazine (10 mg/kg subcutaneously; Bayer, Brussels, Belgium). A silicone catheter (Terumo, Leuven, Belgium) was implanted into the peritoneal cavity without aseptic precautions and subcutaneously tunneled to the neck. A daily infusion of 2 ml of dialysate (3.86% Dianeal; Baxter, Nivelles, Belgium) was performed for 6 d. In addition, animals were intraperitoneally administered Staphylococcus epidermidis (10⁷/ml colony-forming units, diluted in 3.86% dialysate) at day 1 and day 3. On day 7, mice were submitted to a 2-h PD exchange to measure permeability parameters followed by tissue sampling.

Peritoneal Permeability Measurement and Tissue Sampling

A peritoneal equilibration test similar to that used in rat (6,14⇓) was used to investigate peritoneal permeability parameters in mice. After anesthesia with ketamine and xylazine, mice were placed on a thermopad at 37°C. The right common carotid artery was cannulated for the measurement of mean arterial BP using an isotonic BP transducer (Harvard Apparatus, Holliston, MA). The right jugular vein was catheterized for saline infusion (0.9% NaCl, 0.3 ml/h). After 30 min of stabilization, a silicon catheter (Venflon 22 GA, Baxter) was inserted into the peritoneal cavity and 2.0 ml of a standard dialysate (Dianeal; Baxter) containing either 3.86% or 7% glucose was instilled. Blood and dialysate samples (100 μl) were taken from the carotid artery and the PD catheter at time 0 and at 30 min, 60 min, and 120 min of dwell time. Before each sampling, 50 to 100 μl of the dialysate was flushed back and forth, and the abdomen was agitated gently to facilitate fluid mixing. Hematocrit was measured before PD exchange. At the end of the dwell, the dialysate was recovered from the peritoneal cavity through the catheter, whereas the remaining portion was subsequently removed from the open cavity with gauze tissues that were weighted. As described previously (6), net UF corresponds to the difference between the total volume of dialysate collected (sampling through catheter and dialysate weight) and the volume instilled in the cavity. The intra-assay and interassay variability parameters of the latter method in mice are 6% and 5%, respectively.

Dialysate white blood cells (WBCs) were counted using a hemocytometer (Marienfeld, Lauda, Germany), and differential leukocyte counts were obtained from peritoneal cells plated to a glass slide by cytospin (Thermoshandon, Pittsburgh, PA) and stained with Papanicolaou. Urea, glucose, sodium, and total protein were assayed using a Kodak Ektachem DT60 II and DTE II analyser (Eastman Kodak Company, Rochester, NY). PD parameters were also obtained in rats, using 15 ml of standard dialysate containing either 3.86% glucose (n = 6) or 7% glucose (n = 6) as described previously (6,14⇓).

At the end of the dwell, animals were killed and peritoneum samples were processed for fixation and protein extraction as described previously (6,10,14⇓⇓). Samples from the visceral and parietal peritoneum were fixed for 3 h at 4°C in 4% paraformaldehyde in 0.1 M phosphate buffer, rinsed, and embedded in paraffin. Samples from the visceral peritoneum were dissected, snap-frozen in liquid nitrogen, and stored at −80°C.

Antibodies

The NOS isoforms were detected with monoclonal antibodies against eNOS, neuronal NOS (nNOS), and iNOS (Transduction Laboratories, Lexington, KY) (10). Other antibodies included a rabbit antibody against AQP1 (Chemicom International, Temecula, CA), a goat antibody against CD31 (Santa Cruz Biotechnology, Santa Cruz, CA), a rabbit antibody against vascular endothelial growth factor (VEGF; Santa Cruz), a goat antibody against vascular cell adhesion molecule-1 (VCAM-1; Santa Cruz), and a monoclonal antibody against β-actin (Sigma, St. Louis, MO).

Immunoblot and ELISA Analyses

SDS-PAGE and immunoblotting were performed as described (6,10⇓). Protein extracts from the visceral peritoneum were separated on 7.5% or 12% acrylamide slabs and transferred to nitrocellulose. Efficiency of transfer to nitrocellulose was tested by Ponceau red (Sigma) staining and immunoreactivity for β-actin. The membranes were blocked for 30 min, incubated overnight with the primary antibody at 4°C, washed, incubated for 1 h with peroxidase-labeled secondary antibodies (Dako, Glostrup, Denmark), and visualized with enhanced chemiluminescence (Amersham, Little Chalfont, UK). The immunoblots were performed at least in duplicate, and specificity of the signal was determined by comparison with positive controls. Densitometry analysis was performed with a StudioStar Scanner (Agfa-Gevaert, Mortsel, Belgium) using the NIH-Image V1-57 software. The concentrations of VEGF in mouse plasma and dialysate were measured by quantitative sandwich ELISA specific for mouse VEGF (R&D Systems, Minneapolis, MN).

Tissue Staining and Immunohistochemistry

Trichrome blue staining and immunostaining were performed as described previously (6,14⇓). After blocking in 0.3% H₂O₂ and incubation with 10% normal serum, sections were incubated successively for 45 min each with the primary antibody, biotinylated IgG, and avidin-biotin peroxidase (Vector Laboratories, Burlingame, CA). The MOM immunodetection kit (Vector) was used for monoclonal antibodies. Antigen retrieval was performed by incubating sections for 30 min in 0.01 M citrate buffer (pH 5.8), in a water bath heated at 97°C, before rinsing in water. Immunolabeling was visualized using aminoethylcarbazole (Vector). Sections were viewed under a Leica DMR coupled to a Leica DC 300 camera (Leica, Heerbrugg, Switzerland). The specificity of immunolabeling was confirmed by incubation without primary antibody and with nonimmune IgG (Dako).

Quantification of Vascular Immunoreactivity

The use of CD31 to quantify microvascular density has been previously reported in detail (18,19⇓). Peritoneal sections stained for CD31 were viewed through a Zeiss microscope coupled to a DAGE-MTI CCD 72 camera (Michigan City, IL) and analyzed through a KS400 system (Kontron, Munich, Germany). The stained areas of the visceral peritoneum were digitized, to outline the vessel walls. The density of vascular structures (in N vessels/mm²), the relative endothelial area (i.e., the ratio between the cumulative endothelial area and the peritoneum area, in %), and the mean vascular radius (in μm) were measured as described previously (20). All slides were coded and analyzed on a simple-blind basis by the same operator. The analysis included a mean of 10 fields per sample and four mice in each group. The intensity of the mononuclear cells infiltrate was graded as 0, absent; 1, moderate; 2, mild; and 3, severe in WT versus KO eNOS mice (n = 5 in each group).

Statistical Analyses

Data are presented as mean ± SEM. Comparisons between results from different groups were performed using two-tailed t test or Mann-Whitney U test, as appropriate.

Clinical and Biological Parameters

Mice were similar in terms of body weight, hematocrit, plasma urea, and glycemia at baseline (Table 1), and none of them died prematurely during the protocol. In comparison with controls, mice with acute peritonitis were characterized by positive dialysate cultures and cloudy dialysates with increased WBC counts (Table 1). None of the cultures obtained from controls, eNOS WT, and eNOS KO mice was positive. The differential leukocyte counts in the dialysate of mice with peritonitis showed that the majority (60 to 80%) of inflammatory cells were polymorphonuclear leukocytes at day 1. That proportion decreased to 34 to 45% at day 3 and further to 8 to 20% at day 7, in parallel with a progressive increase in mononuclear leukocytes (macrophages and lymphocytes). A comparison between eNOS WT and KO mice did not show any significant difference in the differential leukocyte count at day 7 (data not shown). As expected (17), eNOS KO mice were characterized by a mean arterial BP higher than eNOS WT mice in control conditions and, to a lower extent, during acute peritonitis (Table 1).

Characterization of Peritoneal Permeability in Mouse: Effects of Acute Peritonitis and eNOS Deletion

Control mice were submitted to a 2-h exchange with 3.86% versus 7% glucose dialysate (Figure 1, Table 2). Exposure to both dialysates induced a progressive increase in the dialysate-to-plasma ratio for urea (Figure 1A), a progressive reabsorption of glucose from the dialysate (Figure 1B), and a fall in the dialysate-to-plasma ratio of sodium during the first 30 min of the dwell (sodium sieving; Figure 1C). At drainage, the dialysate was clear in all animals. Exposure to 7% dialysate induced a significant increase in the sodium sieving (Figure 1C, Table 2), whereas it had no effect on other parameters (Figure 1, A and B, Table 2). It must be noted that the PD parameters obtained in mice were remarkably similar to those obtained in rats, as shown by similar values for cumulative urea and glucose transport, sodium sieving, and net UF (normalized to body weight; Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Functional parameters of peritoneal dialysis (PD) in mice at baseline. The dialysate-to-plasma (D/P) ratio of urea (A) and sodium (C) and the progressive removal of glucose from the dialysate (D/D₀ glucose; B) were determined during a 2-h exchange with 2 ml of 3.86% (▪) versus 7% glucose (▴) dialysate. Exposure to 7% glucose induces a significant increase in the sodium sieving, whereas the permeability to urea and glucose remains similar. *P < 0.05 versus 3.86% glucose dialysate.

The effects of acute peritonitis were investigated in control and eNOS mice (Figure 2, Table 3). Acute peritonitis induced a major increase in peritoneal permeability for urea, a faster glucose absorption from the dialysate, and a loss of the sodium sieving in control and eNOS WT mice (Figure 2). In both groups, the changes were reflected by a similar increase in the cumulative transport of urea and glucose, a fall in UF, and an increased protein loss in the dialysate (Table 3). Thus, the magnitude of the changes induced by acute peritonitis was similar in control and eNOS WT mice. However, in comparison with the latter, eNOS KO mice with acute peritonitis were characterized by a significant reduction in the hyperpermeability to urea and glucose, a significant increase in UF and sodium sieving, and a decreased protein loss in the dialysate (Figure 2, Table 3). At baseline, the deletion of eNOS was not reflected by changes in peritoneal permeability parameters (Table 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Functional parameters of PD in mice: Effect of acute peritonitis and endothelial nitric oxide synthase (eNOS) deletion. The D/P ratio of urea (A) and sodium (C), the progressive removal of glucose from the dialysate (D/D₀ glucose; B), and net ultrafiltration (UF; volume out–volume in; D) were determined in eNOS wild-type (WT) mice (□), eNOS WT mice with peritonitis (WT-p; ▪), and eNOS knockout (KO) mice with peritonitis (KO-p; ▴) during a 2-h exchange with 2 ml of 3.86% glucose dialysate. Acute peritonitis induces a major increase in the permeability for urea and glucose, with a loss of the sodium sieving and a loss of UF. All of these modifications are significantly reversed in eNOS KO mice. #P < 0.05 between WT and WT-p mice; *P < 0.05 between WT-p and KO-p mice; **P < 0.05 between KO-p and WT mice.

Structural Changes in the Mouse Peritoneum: Morphology and Immunostaining

Morphologic examination of the visceral and parietal peritoneum showed that acute peritonitis in mice was reflected by a massive, submesothelial infiltrate of mononuclear cells, together with a discrete edema (Figure 3). These structural modifications were similar in eNOS WT and KO mice, although the mononuclear cell infiltrate was less marked in the latter (median intensity of the mononuclear cells infiltrate, 3.0 versus 2.0; n = 5; P = 0.033, Mann-Whitney U test).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Effects of acute peritonitis on the structure of the peritoneum in eNOS WT and KO mice. Trichrome blue staining in the visceral (A through D) and parietal (E through G) peritoneum of eNOS WT (A, B, E, and F) and eNOS KO (C, D, and G) mice in control conditions (A, C, and E) or acute peritonitis (B, D, F, and G). The deletion of eNOS is not reflected by structural changes of the peritoneal membrane at baseline (A versus C). Acute peritonitis is reflected by a massive infiltrate of mononuclear cells in the visceral and parietal peritoneum. The changes induced by acute peritonitis are similar in eNOS WT and KO mice, although the infiltrate is less important in the latter. Magnifications: ×150 in A through D, ×300 in E through G.

Immunostaining showed that acute peritonitis was associated with an increased signal for eNOS in capillary endothelium (Figure 4, A and B), whereas no specific staining was detected in eNOS KO mice (Figure 4C). Contrasting with the absence of signal in control mice, a strong staining for iNOS was detected in mononuclear cells infiltrating the peritoneum of mice with acute peritonitis (Figure 4, D and E). Of note, the intensity of iNOS staining was lower in eNOS KO versus eNOS WT mice (compare Figure 4, F and E). Acute peritonitis was reflected by a marked increase in the immunoreactivity for CD31, located in both endothelial cells and macrophages (Figure 4, G and H). The signal for CD31 was lower (intensity and density) in eNOS KO versus eNOS WT mice (compare Figure 4, I and H). The morphologic and immunocytochemical modifications induced by acute peritonitis were similar in the visceral and parietal peritoneum.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Effects of acute peritonitis and eNOS deletion on NOS isoforms and CD31 expression in the peritoneum. Immunostaining for eNOS (A through C), inducible NOS (iNOS; D through F), and CD31 (G through I) in the visceral peritoneum of eNOS WT mice (A, D, and G), eNOS WT mice with peritonitis (B, E, and H), and eNOS KO mice with peritonitis (C, F, and I). Immunoreactivity for eNOS (1:250) is located in the endothelium lining peritoneal blood vessels in eNOS WT mice (A). The staining intensity and the density of capillaries and blood vessels stained for eNOS are markedly increased in eNOS WT mice with peritonitis (B), whereas no signal is detected in eNOS KO mice (C). In contrast with the lack of signal in eNOS WT mice (D), an intense immunoreactivity for iNOS (1:100) is detected in mononuclear cells infiltrating the visceral peritoneum of eNOS WT mice with peritonitis (E, inset). The signal for iNOS, less intense than in WT mice, is also detected in eNOS KO mice with peritonitis (F, inset). Staining for CD31 (1:50) is detected in the endothelium lining peritoneal vessels and capillaries in eNOS WT mice (G). Acute peritonitis is associated with an increased density of CD31-positive blood vessels and capillaries that is significantly more marked in eNOS WT (H) than in eNOS KO (I) mice. Magnification, ×300.

Quantification of vascular density and relative endothelial area in the visceral peritoneum (Figure 5) confirmed that the deletion of eNOS in mice had no effect on both parameters at baseline. Peritonitis induced a significant increase in vascular density and relative endothelial area in both eNOS WT and KO mice, but the effect was significantly attenuated in eNOS KO mice. Peritonitis induced a significant increase in the mean radius of stained vessels in both eNOS WT-p (63 ± 8 versus 198 ± 19 μm; P = 0.03) and eNOS KO-p (64 ± 21 versus 115 ± 16 μm; P = 0.03) mice. The increase, which was predominantly observed in intermediate and large-sized vessels, was significantly attenuated in eNOS KO-p versus eNOS WT-p mice (115 ± 16 μm versus 198 ± 19 μm, respectively; P = 0.03).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Effects of acute peritonitis and eNOS deletion on the vascular density and relative endothelial area in the visceral peritoneum. Quantification of vascular density and relative endothelial area was performed on sections stained for CD31 (see text for details). The morphometry was performed on four samples in each group, with a mean of 10 fields analyzed per sample. #P < 0.05 versus eNOS WT; *P < 0.05 versus eNOS KO; §P < 0.05 versus eNOS WT-p.

Expression of NOS Isoforms and AQP1 in the Mouse Peritoneum: Immunoblotting

Acute peritonitis induced a significant upregulation of the endothelial (140 kD) and inducible (130 kD) NOS isoforms, and a discrete upregulation of AQP1 (relative optical density versus control, 141 ± 17%; P = 0.20) in the mouse peritoneum (Figure 6A). As described previously (10), the anti-iNOS antibody cross-reacted with eNOS (see the upper band on the gel), but the specific band corresponding to iNOS (lower band of the gel) was identified by its molecular mass (130 kD) and co-migration with macrophage extract (positive control). The induction of iNOS was variable (6,10⇓), ranging from weak to very strong. In control conditions, eNOS and iNOS isoforms were not detected in the peritoneum of eNOS KO mice, whereas the expression of AQP1 was similar in eNOS WT and KO mice (data not shown). Acute peritonitis in eNOS WT and KO mice was reflected by a variable induction of iNOS in the peritoneum (Figure 6B). Of note, the expression of iNOS was lower in eNOS KO (relative optical density versus eNOS WT, 64 ± 22%; P = 0.55), whereas that of AQP1 was higher (relative optical density versus eNOS WT, 175 ± 54%; P = 0.42). A weak expression of nNOS was also detected within the peritoneum of control and eNOS mice and showed no significant variation in case of acute peritonitis (data not shown). Preliminary experiments (limited to three pairs of eNOS mice with peritonitis) also indicated that there was a 20% decrease in the expression of the adhesion molecule VCAM-1 in the peritoneum of eNOS KO mice versus eNOS WT mice with peritonitis (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6. Effects of acute peritonitis on the expression of eNOS and iNOS and aquaporin-1 (AQP1) in the peritoneum. (A) Representative immunoblots for eNOS (1:5,000), iNOS (1:5,000), and AQP1 (1:20,000) in peritoneum extracts prepared from control mice and mice with acute peritonitis (20 μg protein/lane). Extracts from bovine aortic endothelial cells (eNOS), mouse macrophages (iNOS), and rat kidney (AQP1) were used as positive controls (lane C). The blots were stripped and reprobed with an mAb against β-actin (1:10,000) for loading accuracy. A consistent upregulation of eNOS (140 kD) is detected in mice with acute peritonitis. The signal for iNOS (130 kD, lower band of the blot), never detected in control samples, is upregulated with variable intensity in mice with acute peritonitis. Note that the antibody against iNOS cross-reacts with eNOS (upper band at 140 kD), as previously reported (10). The core (28 kD) and glycosylated (35 to 50 kD) AQP1 isoforms are identified in the mouse peritoneum, with a slight upregulation in mice with peritonitis. Densitometry analysis confirms that peritonitis induces a major upregulation of eNOS (relative optical density, 312 ± 41%; P = 0.001), whereas the increase in AQP1 does not reach statistical significance (relative optical density, 141 ± 17%; P = 0.20). *P < 0.05 versus control mice. (B) Representative immunoblots for eNOS, iNOS, and AQP1 in the visceral peritoneum of eNOS WT and KO mice with peritonitis. The blots were performed as in A. As expected, eNOS KO mice do not express eNOS in the peritoneum. Although iNOS is detected in all samples from eNOS WT and KO mice with peritonitis, the upregulation is less intense in eNOS KO mice (relative optical density, 64 ± 22%; P = 0.55). The latter are also characterized by the upregulation (175 ± 54%; P = 0.42) of AQP1. However, these differences do not reach statistical significance.

Determination of VEGF Concentrations in the Plasma and Dialysate

A specific ELISA was used to determine the concentrations of VEGF in the plasma and dialysate of the eNOS mice in control conditions and after catheter-induced acute peritonitis (Table 4). The deletion of eNOS in mice was not reflected by significant changes in VEGF concentrations. Peritonitis induced a major increase in dialysate VEGF in both eNOS WT and eNOS KO mice; the increase in that parameter was slightly blunted in eNOS-KO mice, but the difference with eNOS WT mice was NS (36 ± 11 versus 61 ± 24 pg/ml, respectively; P = 0.42). Immunoblotting studies also failed to demonstrate a significant difference in the expression of VEGF in peritoneal samples from eNOS WT and KO mice (Figure 7).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7. Expression of vascular endothelial growth factor (VEGF) in the peritoneum of eNOS WT and KO mice with peritonitis. Representative immunoblot for monomeric VEGF (21 kD; 1, 1000) in the visceral peritoneum of eNOS WT and KO mice with peritonitis (20 μg protein/lane, reducing conditions). Recombinant VEGF was used as positive control (lane C). The blots were probed with Ponceau red (Sigma) for transfer accuracy. The band corresponding to VEGF is detected in all samples from eNOS WT and KO mice with peritonitis. The expression of VEGF in the peritoneum is similar in both groups (relative optical density, eNOS KO-p versus eNOS WT-p mice, 115 ± 16%; P = 0.63).