Chronic Uremia Induces Permeability Changes, Increased Nitric Oxide Synthase Expression, and Structural Modifications in the Peritoneum

Home , Uremia

J Am Soc Nephrol 12: 2146–2157, 2001 Chronic Uremia Induces Permeability Changes, Increased Nitric Oxide Synthase Expression, and Structural Modifications in the Peritoneum

SOPHIE COMBET,*‡ MARIE-LAURE FERRIER,* MIEKE VAN LANDSCHOOT,§ MARIA STOENOIU,* PIERRE MOULIN,† TOSHIO MIYATA,¶ NORBERT LAMEIRE,§ and OLIVIER DEVUYST* Departments of *Nephrology and †Pathology, Universite´Catholique de Louvain Medical School, Brussels, Belgium; ‡Department of Cell Biology, CEA, Saclay, France; §Department of Nephrology, Rijksuniversiteit Gent, Gent, Belgium; and ¶Institute of Medical Science and Department of Internal Medicine, Tokai University School of Medicine, Kanagawa, Japan.

Abstract. Advanced glycation end products (AGE), growth failure. Focal areas of vascular proliferation and fibrosis were factors, and nitric oxide contribute to alterations of the perito- detected in uremic rats, in relation with a transient up-regula- neum during peritoneal dialysis (PD). These mediators are also tion of vascular endothelial growth factor and basic fibroblast involved in chronic uremia, a condition associated with in- growth factor, as well as vascular deposits of the AGE car- creased permeability of serosal membranes. It is unknown boxymethyllysine and pentosidine. Correction of anemia with whether chronic uremia per se modifies the peritoneum before erythropoietin did not prevent the permeability or structural PD initiation. A rat model of subtotal nephrectomy was used to changes in uremic rats. Thus, in this rat model, uremia induces measure peritoneal permeability after 3, 6, and 9 wk, in parallel permeability and structural changes in the peritoneum, in par- with peritoneal nitric oxide synthase (NOS) isoform expression allel with AGE deposits and up-regulation of specific NOS and activity and structural changes. Uremic rats were charac- isoforms and growth factors. These data suggest an indepen- terized by a higher peritoneal permeability for small solutes dent contribution of uremia in the peritoneal changes during and an increased NOS activity due to the up-regulation of PD and offer a paradigm to better understand the modifications endothelial and neuronal NOS. The permeability changes and of serosal membranes in uremia. increased NOS activities correlated with the degree of renal

The unique structure of the peritoneal membrane (PM) allowed glucose concentrations, with eventual accumulation of ad- the development of peritoneal dialysis (PD), a therapeutic vanced glycation end products (AGE), release of growth fac- modality that now accounts for 15% of the total number of tors, and increased production of nitric oxide (NO) (4–6). The patients worldwide who are undergoing dialysis (1). The long- latter is mediated by regulation of specific NO synthase (NOS) term use of PD is limited by progressive alterations in the PM. isoforms within the PM (6). These alterations include fibrosis and vascular proliferation, Thus far, the potential contribution of uremia per se to the the latter being responsible for an increased effective peritoneal changes of PM structure and function has not been considered. surface area (EPSA), an increased permeability for small sol- Among a constellation of symptoms known for decades, ure- utes and glucose, and, thereby, an impaired ability to sustain mia is associated with an increase in the permeability of serosal ultrafiltration (2). The relevance of the problem has been membranes with eventual pleural or pericardial effusion (7). confirmed by recent studies demonstrating that high PM per- Several molecular mechanisms associated with uremia might meability is associated with increased mortality and morbidity have potential effects on the PM, irrespective of exposure to in patients undergoing PD (3). Studies in both animal models glucose-based dialysate. NO is an obvious candidate, because and patients undergoing PD have shown that these structural it regulates vascular tone and reactivity and is involved in changes are mainly related to the exposure of the PM to high vascular proliferation (8). However, the issue of the L-arginine–NO pathway in uremia appears to be complex. Evidence for decreased excretion of NO metabolites (9) and reduced Received October 19, 2000. Accepted March 14, 2001. whole-body NO production (10) have suggested low basal NO Correspondence to Dr. Olivier Devuyst, Division of Nephrology, Universite´ production in uremia. In contrast, other studies have shown Catholique de Louvain Medical School, Avenue Hippocrate 10, B-1200 Brus- increased NO production in the systemic circulation of uremic sels, Belgium. Phone: ϩ32-2-764-1855; Fax: ϩ32-2-764-5455; E-mail: [email protected] rats (11) and patients (12). The availability of L-arginine, the 1046-6673/1210-2146 NOS substrate, in uremia has also been debated (9,12), al- Journal of the American Society of Nephrology though renal synthesis of endogenous arginine appears to be Copyright © 2001 by the American Society of Nephrology maintained in uremic rats (13) and patients with end-stage J Am Soc Nephrol 12: 2146–2157, 2001 Peritoneal Modifications in Chronic Uremia 2147 renal disease (12). Interestingly, an up-regulation of NOS erlands) implanted in the carotid artery and connected to a Linear- isoforms in extrarenal tissues has been documented in uremic corder polygraph (Ankersmit, Groot-Bijgaarden, Belgium). The BP rats (14). Uremic patients are also characterized by high levels values were obtained after 30-min equilibration, i.e., before starting of circulating AGE (15) and vascular endothelial growth factor the dwell. Blood and dialysate samples were obtained before (t0) and at 30 min (t ), 60 min (t ), and 120 min (t ) of dwell time. At the (VEGF) (16). Several growth factors, including the basic fi- 30 60 120 end of the dwell, animals were killed by exsanguination, and the broblast growth factor (FGF2) and VEGF, have been impli- dialysate was collected. Urea, creatinine, hematocrit, glucose, sodium, cated in structural changes induced by experimental uremia in potassium, total protein, and osmolality were assayed by standard the remnant kidney (17) or in the heart (18). methods (22). Plasma levels of prealbumin were measured in 6-wk In this study, we tested the hypothesis that chronic uremia control and uremic rats (four randomly selected samples in each induces functional and structural changes in the PM of rats that group) by nephelometry (Behring Nephelometer Analyzer II, Dade have undergone subtotal nephrectomy. This procedure pro- Behring, Marburg, Germany) and amino acids by ion-exchange chro- vides a classical model of chronic uremia that has been exten- matography (Biochrom 20 Analyzer, Pharmacia Biotech, Rosendaal, sively used for descriptive and interventional studies focused The Netherlands). White blood cells were counted in a Bu¨rker cham- thus far on the remnant kidney or other target organs such as ber, and dialysate cultures were routinely obtained (22). Identical the heart (18) or the bone (19). Our investigations of peritoneal samples from the visceral and parietal peritoneum were processed for permeability parameters were correlated with determinations fixation in 4% paraformaldehyde or protein extraction as previously detailed (22,23). All experiments were conducted in accord with local of NOS expression and enzymatic activities, as well as with prescriptions and the NIH Guide for the Care and Use of Laboratory expression of AGE and growth factors in the peritoneum. The Animals. role of anemia in the genesis of the modifications was further investigated in uremic rats treated with erythropoietin (EPO). Antibodies The NOS isoforms were detected with monoclonal antibodies Materials and Methods raised against human endothelial NOS (eNOS) and neuronal NOS Rat Model of Chronic Uremia (nNOS) and mouse inducible NOS (Transduction Laboratories, Lex- Adult (8 wk) male Sprague-Dawley rats (Iffa Credo, Brussels, ington, Kentucky) (23). Other antibodies included affinity-purified Belgium) weighing 300 Ϯ 3 g were randomly assigned to uremic and rabbit antibodies against FGF2 and a monoclonal antibody against sham-operated groups. Uremia was induced by a standard procedure VEGF (both from Santa Cruz Biotechnology, Santa Cruz, California); of tissue removal (5/6 nephrectomy followed 1 wk later by contralat- a goat anti–type III collagen (Southern Biotechnology Associates, eral nephrectomy), as described previously (20). The amount of renal Birmingham, Alabama); a monoclonal antibody against human tissue removed (approximately 85%) was quantified by weighting the ␣–smooth muscle actin (Dako, Glostrup, Denmark); affinity-purified tissue, and less than 10% variation was observed between series (20). rabbit antibodies against carboxymethyllysine (CML) and pentosidine Sham-operated (control) rats were subjected to flank incisions sepa- (24), and a monoclonal antibody against CML (25); and a monoclonal rated by 1 wk. All rats had access to a standard rat laboratory diet antibody against ␤-actin (Sigma, St. Louis, Missouri). Positive im- (Pavan, Oud-Turnhout, Belgium) containing 21% protein (1.3% L- munoblot controls included lysates from bovine aortic endothelial arginine) and tap water ad libitum until the evening before they were cells (eNOS), mouse macrophages (iNOS), rat pituitary gland killed. A first series of uremic and control rats were studied 3, 6, and (nNOS), and recombinant FGF2 and VEGF (Santa Cruz 9 wk after the last surgery/incision. Subsequently, a second series of Biotechnology). 8-wk-old male Sprague-Dawley rats underwent the same procedure of subtotal nephrectomy to investigate the effects of a subcutaneous Western Blot Analyses injection of EPO (100 IU/kg, Boehringer, Mannheim, Germany) or Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and normal saline twice a week from the time of nephrectomy and until 6 immunoblotting were performed as described previously (22). Effi- wk thereafter. The cumulative mortality, including the nephrectomy/ ciency of transfer to nitrocellulose was routinely tested by Ponceau sham procedure, was 6% (1/17) in controls versus 34% (18/53) in red (Sigma) staining and ␤-actin immunoreactivity. The membranes uremic rats. About 60% of the deaths (11/18) in the uremic group were blocked for 30 min at room temperature and then incubated with occurred within 10 d, i.e., during or immediately after the two-step the primary antibody at 4°C for 16 to 18 h. The membranes were then surgery. This mortality is expected for this degree of tissue removal in washed and incubated for1hatroom temperature with the appropri- large rats that are prone to pre- or postoperative bleeding (20,21). ate peroxidase-labeled secondary antibodies (Dako) before visualiza- Accordingly, the other rats (7/18) died during the 9-wk follow-up tion with enhanced chemiluminescence (Amersham, Little Chalfont, period, corresponding to the chronic uremic state. Four rats died UK). Densitometry analyses were performed with a studioStar Scan- within the first week, one during the fifth week, and two during the ner (Agfa-Gevaert, Mortsel, Belgium) by use of the NIH-Image eighth week of follow-up. Light microscopy examination of the peri- V1–57 software. The relative optical densities (in %, relative to the toneum of rats that died within the first week of uremia did not show age-matched control) were obtained in duplicate. signs of acute peritonitis or any specific change. Immunohistochemistry PD and Biologic Parameters: Tissue Sampling Immunostaining was performed on 6-␮m sections from peritoneum

At the time they were killed, rats were anesthestized with subcu- samples embedded in paraffin (22,23). After blocking in 0.3% H2O2 taneous Nembutal and placed on a heated pad to perform a 2-h and incubation with 10% normal serum, sections were incubated exchange PD, as described previously (22). The rats were weighed, successively for 45 min each with the primary antibody, biotinylated and their BP was measured by means of a transducer (Pressure Set IgG (Vector Laboratories, Burlingame, California), and avidin-biotin Uniflow type 43-600F, Bentley Laboratories Europe, Uden, The Neth- peroxidase (Vector). Immunolabeling was visualized by use of ami- 2148 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 2146–2157, 2001 noethylcarbazole (Vector). Sections were viewed under a Leica DMR Data Analyses coupled to a Leica MPS60 photomicrographic system (Leica, Heer- Data are presented as mean Ϯ SEM. Comparisons between results brugg, Switzerland). The specificity of the immunolabeling was con- from different groups were performed by use of the t test or one-way firmed by incubation without primary antibody and with nonimmune ANOVA, as appropriate. Statistical significance was defined as P Ͻ IgG (Dako) (22,23). 0.05.

Quantification of eNOS Immunoreactivity Results Immunostaining for eNOS was quantified by image analysis (6). Clinical, Biologic, and Functional Parameters Random fields (a mean of eight per slide) were selected in at least four The evolution of clinical and biologic parameters in the first different slides in each group and digitized through a Leitz Ortholux series of rats is shown in Table 1. Chronic renal failure sec- microscope (Wetzlar, Germany) coupled with a JVC KY-F58 color ondary to subtotal nephrectomy was documented by signifi- digital camera (JVC Americas, Wayne, NJ). Images were analyzed by cantly higher urea and creatinine levels in uremic versus con- a KS-400 system (Zeiss, Munich, Germany) to isolate labeled vessels trol rats throughout the study, whereas BP remained similar in from the background and quantify the density of stained capillaries all groups. Uremic rats had a significant degree of anemia and (N/field). The slides were coded and analyzed on a single-blind basis a lower weight. However, body weight gains were similar in by the same operator. control and uremic rats (ϩ7% versus ϩ8% between weeks 3 and 6 and ϩ32% versus ϩ40% between weeks 3 and 9, respectively), as were plasma prealbumin levels at 6 wk (1.9 Ϯ NOS Activity Assay 0.03 versus 2 Ϯ 0.07 mg/dl, respectively). Plasma arginine NOS enzymatic activities were measured using the L-citrulline levels were similar in 6-wk control and uremic rats (129 Ϯ 6 assay as described elsewhere (23). Briefly, 25 ␮l of tissue extract (400 Ϯ ␮ ␮ ␮ versus 146 36 mol/L, respectively), whereas citrulline was g of total protein) containing 20 mM CHAPS were added to 200 l Ϯ Ϯ ␮ significantly higher in the latter (61 5 versus 249 45 of Tris buffer (50 mM, pH 7.4) containing 10 mM DTT, 10 g/ml ␮ ϭ calmodulin, 1 mM NADPH, 4 ␮M flavin adenine dinucleotide, 4 ␮M mol/L, P 0.01). Renal function recovered partially, with a flavin adenine mononucleotide, 2 ␮M L-arginine, and 10Ϫ3 mCi/ml higher body weight and hematocrit level between weeks 6 and L-[3H]-arginine (NEN Life Science, Zaventem, Belgium). Assays 9. Absence of peritonitis was demonstrated by a clear dialysate were performed for 30 min at 37°C and stopped with 2 ml of ice-cold at the end of the dwell, similar dialysate white blood cell stop buffer (20 mM CH3COONa [pH 5.5], containing 2 mM ethyl- counts, and negative dialysate cultures. enediaminetetraacetic acid, 0.2 mM ethyleneglycotetraacetic acid, and Functional parameters evaluated at 6 wk showed that re- 1 mM L-citrulline). L-[3H]-citrulline was separated by cation-ex- moval of urea and creatinine from the plasma (Figure 1A), as change chromatography (Bio-Rad) and quantified by liquid scintilla- well as absorption of glucose from the dialysate (Figure 1B), tion (SL3000, Intertechnique, Plaisir, France). The NOS activity were significantly higher in uremic than in control rats. Sodium (pmol citrulline/mg protein/min) was determined by subtracting sieving (i.e., the percentage of decrease in the dialysate-over- counts obtained with or without 1 mM NG-monomethyl-L-arginine. plasma ratio for sodium during the first 30 min of the dwell) Assays were performed with (1 mM CaCl2) or without (0 mM CaCl2, 2 mM ethyleneglycotetraacetic acid, and 2 mM ethylenediaminetet- fell significantly in uremic rats (Figure 1C), despite a signifi- ϩ ϩ Ϫ raacetic acid) Ca2 to measure total versus Ca2 -independent NOS cant glucose gradient (Dglucose Pglucose at 30 min: 2541 activities and to calculate Ca2ϩ-dependent NOS activity. Determina- mg/dl). Total protein loss in the dialysate was increased in tions were performed in duplicate on samples randomly selected in uremic versus control rats (35 Ϯ 7 versus 18 Ϯ 1 mg, respec- each group. tively, P ϭ 0.09). Similar permeability alterations (increased

Table 1. Clinical and biologic parameters of sham-operated (control) and uremic rats. Two series of 8-wk-old male Sprague- Dawley rats were used for evaluating the effects of uremia (A) and, subsequently, the role of anemia (B)a

Plasma Plasma Dialysate Series Groups Time n Body Weight BP Urea Creatinine Hematocrit WBC (wk) (g) (mm Hg) (%) 3 (mg/dl) (mg/dl) (cells/mm )

A Control 3 6 375 Ϯ 18 123 Ϯ 15 22 Ϯ 1 0.23 Ϯ 0.02 49 Ϯ 1 362 Ϯ 69 Control 6 4 404 Ϯ 8 130 Ϯ 524Ϯ 1 0.26 Ϯ 0.05 48 Ϯ 2 498 Ϯ 139 Control 9 6 496 Ϯ 11d 125 Ϯ 727Ϯ 1d 0.26 Ϯ 0.03 48 Ϯ 1 735 Ϯ 59d Uremic 3 7 297 Ϯ 17b 142 Ϯ 4 144 Ϯ 26b 1.7 Ϯ 0.2b 33 Ϯ 3b 416 Ϯ 62 Uremic 6 6 321 Ϯ 23b 129 Ϯ 7 183 Ϯ 56b 1.9 Ϯ 0.3b 36 Ϯ 2b 342 Ϯ 92 Uremic 9 7 418 Ϯ 13b,c 123 Ϯ 783Ϯ 9b 1.0 Ϯ 0.1b,c 44 Ϯ 1b,c 726 Ϯ 82c B Uremic EPOϪ 6 7 370 Ϯ 13 109 Ϯ 8 123 Ϯ 23b 1.5 Ϯ 0.3b 36 Ϯ 2b 546 Ϯ 101 Uremic EPOϩ 6 7 373 Ϯ 17 113 Ϯ 9 126 Ϯ 26b 1.4 Ϯ 0.4b 48 Ϯ 1c 599 Ϯ 85

a EPO, erythropoietin; WBC, white blood cells; EPO (ϩ) or control (Ϫ) injections were given to the second series of 6wk-uremic rats. b P Ͻ 0.05 versus control at corresponding time; c P Ͻ 0.001 versus 6-wk uremic; d P Ͻ 0.005 versus 3-wk control. J Am Soc Nephrol 12: 2146–2157, 2001 Peritoneal Modifications in Chronic Uremia 2149

permeability for urea and glucose, decreased sodium sieving) were observed in uremic rats at 3 and 9 wk. Analysis of permeability parameters in all groups showed a significant correlation between the increased permeability for small solutes and the degree of renal failure (Figure 2A). Significant correlations were also obtained with the other permeability parameters (including the dialysate-over-plasma ratio at 120 min) and when plasma urea levels were used instead of plasma creatinine (data not shown).

NOS Enzymatic Activities In comparison with controls, total NOS activity in the visceral peritoneum of uremic rats was significantly increased at 3 and 6 wk and returned to control levels at 9 wk (Table 2). Determination of Ca2ϩ-independent NOS activity yielded similar and low values in all groups. As a consequence, the differences in total NOS activity observed between uremic and control rats at 3 and 6 wk were solely due to Ca2ϩ-dependent NOS activity (Table 2). The level of total NOS activity within the peritoneum was significantly correlated with plasma creatinine levels (Figure 2B). Significant correlations were also observed with Ca2ϩ-dependent NOS and when plasma urea levels were used instead of creatinine (data not shown).

Expression of NOS Isoforms Immunoblot analyses were performed to identify which NOS isoform(s) was (were) involved in the increased Ca2ϩ- dependent NOS activity measured in uremic peritoneum. Rep- resentative blots shown on Figure 3A demonstrated that the expression of both nNOS (155 kD) and eNOS (140 kD) in the peritoneum was increased in uremic versus control rats. Den- sitometry analysis (Figure 3B) confirmed a significant up- Figure 1. Functional parameters of peritoneal dialysis in 6 wk control regulation of nNOS in uremic rats at 3 wk (132% increase; P and uremic rats. The dialysate-to-plasma (D/P) ratio of urea and Ͻ 0.006) and 6 wk (70% increase; P Ͻ 0.05), with a return to creatinine (A), the absorption of glucose from the dialysate (D/D0) control levels at 9 wk. In contrast, eNOS expression was (B), and the sodium sieving (expressed in % decrease of D/P sodium significantly higher at 6 wk (60% increase; P ϭ 0.01) and 9 wk at the 30th min of the dwell) (C) were determined in control (open (52% increase; P ϭ 0.04). No signal was detected when the symbols) and uremic (filled symbols) rats during a 2-h exchange with blots were probed with an anti-iNOS (data not shown), as 15 ml of 7% glucose. All parameters were obtained for six rats in each group. *P Ͻ 0.05 between control and uremic rats. expected, given that none of the rats suffered from peritonitis and Ca2ϩ-independent NOS activity was very low (23).

Table 2. Nitric oxide synthase (NOS) enzymatic activities in the peritoneum of sham-operated (control) and uremic rats

NOS Activity (pmol citrulline/mg protein per min) Series Groups N Total Ca2ϩ-Independent Ca2ϩ-Dependent

A 3-wk control 4 0.11 Ϯ 0.02 0.03 Ϯ 0.01 0.08 Ϯ 0.01 3-wk uremic 4 0.21 Ϯ 0.02a 0.03 Ϯ 0.01 0.18 Ϯ 0.01a 6-wk control 3 0.17 Ϯ 0.02b 0.03 Ϯ 0.01 0.15 Ϯ 0.03b 6-wk uremic 3 0.24 Ϯ 0.03a 0.02 Ϯ 0.01 0.22 Ϯ 0.03a 9-wk control 4 0.15 Ϯ 0.02 0.01 Ϯ 0.05 0.14 Ϯ 0.02b 9-wk uremic 4 0.16 Ϯ 0.03c 0.02 Ϯ 0.01 0.15 Ϯ 0.03c B 6-wk uremic EPOϪ 4 0.20 Ϯ 0.01 0.02 Ϯ 0.01 0.18 Ϯ 0.01 6-wk uremic EPOϩ 4 0.20 Ϯ 0.01 0.03 Ϯ 0.01 0.17 Ϯ 0.01

a P Ͻ 0.02 versus control at corresponding time; b P Ͻ 0.005 versus 3-wk control; c P Ͻ 0.02 versus 6-wk uremic. 2150 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 2146–2157, 2001

tions (Figure 4, F through H). The specificity of the signal was confirmed by the absence of staining when incubation was performed with control mouse IgG (Figure 4I). Examination of trichrome- and reticulin-stained sections from 6-wk rats showed focal areas of increased collagen matrix in submesothelial and perivascular tissue of uremic versus control rats (Figure 4, compare J with K and L, and compare N with O and P). The focal areas were located both in the isolated visceral peritoneum (Figure 4, J and K) and above the omentum (Figure 4, N and O). Staining on serial sections demonstrated strong immunoreactivity for type III collagen in these areas (Figure 4, M and Q), whereas staining for ␣–smooth muscle actin was consistently negative. These structural changes were detected in 3- and 9-wk uremic rats, but they were more marked at 6 wk. Computer-assisted morphometry in random sections of the visceral peritoneum confirmed a significant increase in the density of capillaries stained for eNOS (6.6 Ϯ 1.2 versus 2.7 Ϯ 0.5 stained capillaries/field, P ϭ 0.03) in 6-wk uremic versus control rats. To substantiate the link among uremia, carbonyl stress, and AGE deposition in the peritoneum (6,15), we performed immunostaining for CML and pentosidine in representative sections of the peritoneum (Figure 5). A weak staining for CML and pentosidine was detected in the peritoneum of 3-wk (Fig- ure 5A) and 6-wk (Figure 5D) control rats; the staining intensity was significantly increased in uremic rats at 3 (Figure 5, B and C) and 6 wk (Figure 5, E through G), mainly located in the arterial walls. Although the labeling for CML was consistently more intense than pentosidine, staining on serial sections allowed us to demonstrate that both AGE accumulated on the same structures (Figure 5, compare B with C and F with G). The AGE immunostaining was specific, as was demonstrated by complete inhibition of staining in the presence of an excess of CML- or pentosidine-conjugated bovine serum albumin Figure 2. Peritoneal permeability and total nitric oxide synthase (data not shown). (NOS) activity as a function of renal failure. Relations between plasma creatinine and the area under the curve for the D/P ratio of Expression of Growth Factors urea between t0 and t120 (A) and the total NOS activity (L-citrulline assay) in the peritoneum (B). Regressions were calculated for all rats Immunoblot analysis confirmed that expression of type III for which both parameters were available. Open symbols are control collagen (150 kD) was increased in the peritoneum of uremic rats; filled symbols are uremic rats. Squares: 3-wk rats; circles: 6-wk rats, mostly at 6 wk (Figure 6A). Compared with controls, rats and triangles: 9-wk rats. expression of both dimeric VEGF and FGF2 (45 and 17 kD, respectively, identified by comigration with full-length recombinant human factors) markedly increased in 3-wk uremic rats Structural Changes of the Peritoneum (171% and 549% increase, respectively) (Figure 6, B and C). Immunostaining for NOS isoforms was performed in 6-wk The VEGF expression returned to control levels in 6- and 9-wk rats (Figure 4, A through I). As previously described (6,22), a uremic rats. By contrast, expression of FGF2 was sustained in faint signal for eNOS was located in the endothelium that lined 6-wk uremic rats (Figure 6C), essentially located within vas- peritoneal blood vessels of control rats (Figure 4A). The signal cular and capillary walls within the peritoneum (Figure 4R). intensity, as well as the density of stained capillaries, were increased in uremic rats (Figure 4B), and focal areas of capil- Effects of EPO Treatment in Uremic Rats lary proliferation were observed in the latter (Figure 4C). A The potential effect of anemia on the changes observed in very faint staining for nNOS could be detected in the nerve uremic peritoneum was evaluated by treating a second series of sections found in the peritoneum of control rats (Figure 4, D uremic rats for 6 wk with EPO or vehicle (Table 1). These and E). The signal was more intense in uremic rats in large 6-wk uremic rats showed slight (but NS) differences for body nerve sections and also within the media of peritoneal arteries weight, BP, and plasma urea, and creatinine levels when com- (Figure 4F). The absence of cross-reactivity between antibod- pared with 6-wk uremic rats investigated in the first series, but ies against nNOS and eNOS was demonstrated on serial sec- their hematocrit was exactly similar (36 Ϯ 2%). Apart from the J Am Soc Nephrol 12: 2146–2157, 2001 Peritoneal Modifications in Chronic Uremia 2151

Figure 3. Uremia and expression of NOS isoforms in the peritoneum: immunoblot analysis. (A) Representative immunoblots for the neuronal NOS (nNOS, 155 kD) and the endothelial NOS (eNOS, 140 kD) in the visceral peritoneum of control and uremic rats at 3, 6, and 9 wk. Together with a positive control, peritoneum samples (30 ␮g of protein/lane) were run on sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE, 7.5% gel), transferred, and probed with monoclonal antibodies against NOS. The faint bands below 150 kD observed on nNOS blots correspond to weak cross-reactivity with other NOS isoforms (23). (B) Densitometry analysis showed a 132% increase in nNOS expression in uremic rats at 3 wk and a 70% increase at 6 wk, whereas the increase observed at 9 wk was NS. Increased eNOS expression was also observed in uremic rats, being significant at 6 and 9 wk (60% and 52% increase, respectively). Determinations were performed in duplicate and normalized against time-matched control rats. subcutaneous injection of EPO/vehicle twice a week, all the containing solutions. The changes parallel an increased NOS 6-wk uremic rats were identical in terms of age, race, gender, activity due to the up-regulation of specific NOS isoforms. The and experimental procedure. The differences might thus reflect peritoneum of uremic rats is also characterized by structural a certain degree of variability in the response to surgery, modifications that include angiogenesis, fibrosis, and a major anesthesia, or preparation for the dwell between two batches of increase in AGE deposits. The increased permeability and rats. Treatment with EPO managed to fully correct anemia in NOS activities correlate with the degree of uremia. The the 6-wk uremic rats, whereas BP, weight, and other biologic changes are independent of renal failure–induced anemia as (Table 1) parameters remained unchanged. Also, treatment demonstrated in EPO-treated uremic rats. with EPO had no effect on permeability parameters: in com- The functional characteristics of the PM were evaluated parison with control rats, EPO-treated uremic rats were still with the clinical peritoneal equilibration test (20,22). The characterized by an increased permeability for urea (Figure PM of uremic rats is characterized by an increased perme- 7A) and increased removal of glucose from the dialysate (Fig- ability for small solutes such as urea and creatinine, a faster ure 7B). Similarly, correction of anemia with EPO was not reabsorption of glucose from the dialysate, and a loss of free reflected by significant changes in NOS activities (Table 2) or water permeability (attested by a lower sodium sieving in in the expression levels of nNOS, eNOS, VEGF, and FGF2 the presence of an effective osmotic gradient) (Figure 1). (Figure 7C). These permeability features, which are explained by an increased EPSA, are similar to those observed in acute Discussion peritonitis (22) and long-term PD (6). Acute peritonitis, a In this article, we demonstrate that chronic renal failure potential complication of any animal model, was ruled out induced by subtotal nephrectomy in rats induces significant in our uremic rats by negative cultures, absence of infiltrate changes in peritoneal permeability before exposure to glucose- on pathologic examination, and NOS studies (vide infra). 2152 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 2146–2157, 2001

Figure 4. Expression of NOS isoforms and structural modifications in the peritoneum of uremic rats. (A through C) Staining for eNOS in the visceral peritoneum of 6-wk control (A) and uremic (B and C) rats. Staining for eNOS is located in the endothelium lining blood vessels and capillaries, whereas the mesothelium (m) is negative (A). The signal intensity, as well as the density of capillaries stained for eNOS are increased in uremic rats (B), with distinct areas of capillary proliferation (C). The mesothelium (m) remains unstained. (D and E) Serial sections J Am Soc Nephrol 12: 2146–2157, 2001 Peritoneal Modifications in Chronic Uremia 2153

Figure 5. Immunostaining for carboxymethyllysine and pentosidine in the peritoneum of control and uremic rats. (A through C) Peritoneal arteries of 3-wk control (A) and uremic (B and C) rats stained with polyclonal antibodies against carboxymethyllysine (CML) (A and B) or pentosidine (C). The weak staining for CML detected in the arterial walls of controls is strongly enhanced in uremic rats. Staining on serial sections demonstrates the colocalization of CML and pentosidine. (D through G) Peritoneal arteries of 6-wk control (D) and uremic (E through G) rats stained with polyclonal antibodies against pentosidine (D through F) and a monoclonal antibody against CML (G). The staining for pentosidine is significantly increased in 6-wk uremic rats. Staining on serial sections demonstrates that pentosidine (F) and CML (G) accumulated in the arterial walls. Note that some nerve sections are also stained. Magnifications: ϫ250 for A through C: ϫ275 for D and E; ϫ300 for F and G.

The relation between the increased peritoneal permeability neum, mediated by specific NOS isoforms, is a critical factor and the degree of renal failure thus includes the peritoneum for increased EPSA (22,26). This study further illustrates the among serosal membranes modified by uremia. potential link between peritoneal permeability and NOS activ- Studies performed in rat models and patients undergoing PD ity; both parameters are directly correlated with the degree of have shown that an enhanced NO release within the perito- uremia in the rats (Figure 2). The demonstration of increased

of the visceral peritoneum of 6-wk control rats stained for nNOS (D) or with hemalun-eosine (E). nNOS is detected in nerve sections (*) within the peritoneum. (F through H) Serial sections of the visceral peritoneum of 6-wk uremic rats stained for nNOS (F), trichrome blue (G), and eNOS (H). The signal for nNOS (located in nerve sections, *) is more intense than in control rats, in particular within the media of large peritoneal arteries (F and G). The absence of cross-reactivity between nNOS (nerve staining, *) and eNOS (endothelium staining, arrowheads) is demonstrated by comparing the serial sections. (I) The specificity of the immunostaining is shown by the lack of signal when sections of 6-wk uremic rats are incubated with control mouse IgG. (J and K) Trichrome blue staining of the visceral peritoneum of 6-wk control (J) and uremic (K) rats. Focal areas of fibrosis, with thickening of the visceral peritoneum, are identified in uremic rats. (L and M) Serial sections with (K), stained with reticulin (L) or type III collagen (M). The colocalization of these markers confirm the focal area of fibrosis within the visceral peritoneum of 6-wk uremic rats. (N and O) Trichrome blue staining of the visceral peritoneum of 6-wk control (N) and uremic (O) rats, showing the perivascular fibrosis in uremic rats (vascular lumen, *). (P and Q) Serial sections with (O), stained with reticulin (P) or type III collagen (Q). (R) Fibroblast growth factor (FGF2) location in the vascular and capillary walls located in the visceral peritoneum of a 6-wk uremic rat. No signal is observed in the mesothelium (m). Magnifications: ϫ335 for A and B; ϫ225 for C; ϫ200 for D and E; ϫ315 for F through H; ϫ250 for I and N through Q; ϫ125 for J through M; ϫ270 for R 2154 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 2146–2157, 2001

Figure 6. Expression of type III collagen and growth factors in the peritoneum: immunoblot analysis. (A) Time course of the expression of type III collagen (150 kD) in samples from the visceral peritoneum. Samples (30 ␮g per lane) from 6-wk control and 3-, 6-, and 9-wk uremic rats were run on SDS-PAGE (7.5% gel), transferred, and probed with antibodies against type III collagen. (B) Representative immunoblots for dimeric vascular endothelial growth factor (VEGF; 45 kD) and FGF2 (17 kD) in the peritoneum of control and uremic rats at 3, 6, and 9 wk. Thirty ␮g of samples were loaded on 14% gels. Dimeric VEGF was investigated in nonreducing conditions. (C) Densitometry analysis showed a transient, 171% increase in VEGF expression in uremic rats at 3 wk. A major and sustained increase in FGF2 expression was observed in 3- and 6-wk uremic rats (549% and 459% increase, respectively). Determinations were performed in duplicate and normalized against time-matched control rats.

NOS activities in the uremic peritoneum relies on the specific nNOS expression also increases in the peritoneum of uremic and sensitive L-citrulline assay, confirmed by expression stud- rats, in contrast with its stability in acute peritonitis (22). The ies for NOS isoforms (23). At variance with acute peritonitis, intricate regulation of nNOS is implicated in many processes, the increased NOS activity observed in uremic rats is only due including neuronal development and plasticity (27). One hy- to Ca2ϩ-dependent NOS (Table 2), associated with the up- pothesis is that nNOS up-regulation in the peritoneum some- regulation of both eNOS and nNOS (Figure 3). Just as in how reflects uremia-induced peripheral neuropathy. A similar long-term PD, increased eNOS expression probably reflects increase in both eNOS and nNOS has been reported in the vascular proliferation and increased endothelial area (6). The penis and pelvic ganglia of rats with subtotal nephrectomy J Am Soc Nephrol 12: 2146–2157, 2001 Peritoneal Modifications in Chronic Uremia 2155

(14). A specific role of the uremic state in NOS regulation is further suggested by the positive correlation between the level of NOS activity within the peritoneum and the degree of uremia (Figure 2). The state of the L-arginine–NO pathway in uremia remains debated, probably because (1) different methods are used to investigate the complex NO metabolism and (2) the release of NO by a given organ might be different from its systemic production. Studies have shown decreased excretion of NO oxidation products (9) and a reduction of basal whole-body NO production (10) in patients with chronic renal disease. In contrast, increased NO production has been documented in patients with end-stage renal disease (12) and also in the circulation of rats made uremic by subtotal nephrectomy (11). Whether NOS substrate deficiency occurs in uremia is another open question. Our data confirm previous reports of the main- tenance of (sub)normal plasma levels of L-arginine in uremia (9,12,28,29). This implies a compensatory mechanism, either increased synthesis of L-arginine in the remaining nephrons (12,13) or increased extrarenal release (29). The repeated observations of increased plasma levels of citrulline in uremia (9,12) might thus reflect an adaptation mechanism or, alterna- tively, originate from reduced renal clearance and arginine synthesis (28). At any rate, the plasma arginine levels observed ␮ here are well above the Km of NOS (32 to 35 mol/L), although this does not exclude a deficit in intracellular availability (28). Different structural modifications—vascular proliferation and fibrosis—are observed in focal areas of the uremic peritoneum (Figure 4). The structural changes were quantified by image analysis and confirmed by immunoblot analyses for eNOS (Figure 3) and type III collagen (Figure 6). The changes were more conspicuous when uremia peaked at 6 wk. The functional consequences of each structural change are different (30). Submesothelial and perivascular fibrosis increases the distance between the dialysate and the endothelium and, thus, decreases peritoneal permeability. In contrast, vascular proliferation increases EPSA and, thus, permeability for small solutes. The increased peritoneal permeability documented in uremic rats suggests that, quantitatively, vascular proliferation is functionally more important than fibrosis in this model. However, because fibrotic changes predominate in the peritoneum (O. Devuyst and N. Lameire, unpublished observations) and the myocardium (31) of uremic rats that survived over 6 mo, the nature of structural changes might change over time. Figure 7. Peritoneal permeability and expression of NOS and growth This study identifies several mediators potentially involved factors in uremic rats: influence of treatment with erythropoietin in the structural changes of the uremic peritoneum. The over- (EPO). (A and B) The D/P ratio of urea (A) and the removal of expression of angiogenic factors such as VEGF and FGF2 at glucose from the dialysate (D/D0) (B) were determined in 6-wk the third wk of uremia (Figure 6) suggests indeed that they are control (E) and uremic rats treated with EPO () or vehicle (}) instrumental in the vascular proliferation—and the fibrotic during a 2-h exchange with 15 ml of 7% glucose. All variables were changes—observed at 6 wk. Previous studies in human and rat Ͻ obtained for six rats in each group. *P 0.05 between control and have suggested a link between VEGF expression and vascular EPO-treated uremic rats. (C) Representative immunoblots for neuro- proliferation in the peritoneum (5,6). NO is necessary for the nal nNOS (155 kD), eNOS (140 kD), dimeric VEGF (45 kD), and FGF2 (17 kD) in the visceral peritoneum of 6-wk uremic rats treated biologic activity of VEGF (32), and enhanced expression of with vehicle (EPOϪ) or EPO (EPOϩ). Thirty ␮g of samples were run both mediators is thought to play a determining role in the on SDS-PAGE (7.5% gel for NOS; 14% gel for VEGF and FGF2). neovascularization observed in the peritoneum of patients un- The expression levels are similar in both groups. dergoing long-term PD (6). On the other hand, FGF2 is impli- 2156 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 2146–2157, 2001 cated in processes including angiogenesis, smooth muscle cell There is a progressive increase in total NOS activity—in rela- proliferation, wound healing, and tissue repair (33). It induces tion with the up-regulation of NOS isoforms—from3to6wk endothelial cell proliferation, migration, and angiogenesis, pos- of uremia, with a fall at 9 wk. Vascular proliferation and sibly under the control of VEGF (34). In addition, FGF2 is a functional alterations of the PM peak at 6 wk. It is also potent mitogen for various cell types, including fibroblasts important to point out that the changes are variable through (35), and its expression correlates with renal fibrogenesis (36). evolution. Although variability in renal function and expres- A similar increase in FGF2 expression has been detected in the sion of growth factors over time is characteristic in this rat peritoneum of uremic patients at time of PD catheter insertion model (21,42), our findings open perspectives because of po- (M. L. Ferrier and O. Devuyst, unpublished data). tential regression of peritoneal alterations in parallel with renal Immunostaining with specific antibodies (24,25) demon- functional improvement. strates a striking increase in CML and pentosidine deposits in the peritoneum of uremic rats (Figure 5). Both in vitro and in Acknowledgments vivo studies have shown that chronic uremia is associated with These studies were supported in part by the Belgian agencies FNRS high levels of circulating AGE, as a result of the accumulation and FRSM (cre´dit 9.4540.96 and convention 3.4566.97) and the ARC of reactive carbonyl compounds (15). Glucose-derived carbon- 00/05–260 (to O.D.) and Grants from Baxter (to S.C.) and the Socie´te´ yls promote the liberation of VEGF in cultured peritoneal cells de Ne´phrologie (to M.S.). We thank Prof. P. Courtoy, Dr. E. Goffin, (37), and AGE colocalizes with VEGF in the peritoneum Dr. W. van Biesen, and Prof. Ch. van Ypersele de Strihou for (6,37). Taken together with these observations, our data sug- suggestions and critiques. The expert technical assistance of Y. gest that AGE deposition in the peritoneum might contribute, Cnops, T. Dheuvaert, S. Ruttens, and L. Wenderickx is highly appre- via enhanced VEGF liberation, to the structural changes ob- ciated. S.C. and M.L.F. contributed equally to this study. served in this uremic rat model. These results were presented in part at the 33rd Annual Meeting of The putative role of anemia in VEGF/eNOS up-regulation and the American Society of Nephrology, October 11 to 16, 2000, To- ronto, Canada, and have been published in abstract form (JAmSoc vascular proliferation in uremic rats was specifically addressed. At Nephrol 11: 615A, 2000). 6 wk, the mean hematocrit level of uremic rats had fallen by 25% (Table 1). In vivo, both hypoxia and anemia are potent inducers of VEGF expression (38), and increased NO levels are associated References with anemia (39). Furthermore, chronic hypoxia induces perme- 1. Gokal R, Mallick NP: Peritoneal dialysis. Lancet 353: 823–828, ability changes in endothelial cells (40), as well as remodeling of 1999 the vasculature and surrounding tissues (including smooth muscle 2. Krediet RT: The peritoneal membrane in chronic peritoneal dialysis. Kidney Int 55: 341–356, 1999 proliferation and fibrosis) (41). Nevertheless, full correction of 3. Churchill DN, Thorpe KE, Nolph KD, Keshaviah PR, Oreopou- anemia with EPO failed to reverse the changes in PM permeabil- los DG, Page´ D: Increased peritoneal membrane transport is ity, NOS activity, and growth factors expression in uremic rats associated with decreased patient and technique survival for (Figure 7). Uremic anemia as such was thus ruled out in the continuous peritoneal dialysis patients. J Am Soc Nephrol 9: genesis of the observed modifications. 1285–1292, 1998 Thus, our data demonstrate that significant changes of per- 4. Nakayama M, Kawaguchi Y, Yamada K, Hasegawa T, Takazoe meability and structure of the PM occur in this uremic rat K, Katok N, Hayakawa H, Osaka N, Yamamoto H, Ogawa A, model. These modifications might be secondary to the adaptive Kubo H, Shigematsu T, Sakai O, Horiuchi S: Immunohistochem- response (e.g., renal compensatory growth) to renal ablation ical detection of advanced glycosylation end products in the (42). However, the correlations between the degree of uremia peritoneum and its possible pathophysiological role in CAPD. and changes in permeability or NOS activity, the similar effect Kidney Int 51: 182–186, 1997 5. Zweers MM, de Waart DR, Smit W, Struijk DG, Krediet RT: of uremia on NOS expression in target organs, and the involve- Growth factors VEGF and TGF-beta 1 in peritoneal dialysis. ment of a series of mediators—NO, VEGF, and AGE—known J Lab Clin Med 134: 124–132, 1999 to operate in uremia (9–18) support an independent contribu- 6. Combet S, Miyata T, Moulin P, Pouthier D, Goffin E, Devuyst tion of uremia in the PM changes that occur during PD. These O: Vascular proliferation and enhanced expression of endothelial findings, which confirm the earlier observation by Rubin et al. nitric oxide synthase in human peritoneum exposed to long-term (43) that peritoneal transport might be increased in uremic peritoneal dialysis. J Am Soc Nephrol 11: 717–728, 2000 patients when compared with nonuremic patients, are also 7. Bailey JL, Mitch WE: Pathophysiology of uremia. In: The Kid- important in terms of biocompatibility because most of the ney, 6th ed., edited by Brenner BM, Philadelphia, WB Saunders, studies evaluating dialysates are performed in normal rats. 2000, pp 2059–2078 Finally, the combination of functional and expression stud- 8. Cook JP, Mont-Reynaud R, Tsao PS, Maxwell AJ: Nitric oxide ies used here offers a paradigm to better understand the inter- and vascular disease. In: Nitric Oxide: Biology and Pathobiol- ogy, edited by Ignarro LJ, San Diego, Academic Press, 2000, pp play between various factors in the modification of serosal 759–783 membranes induced by uremia. Thus, the retention of carbonyl 9. Schmidt RJ, Baylis C: Total nitric oxide production is low in compounds leads to a significant accumulation of AGE in the patients with chronic renal disease. Kidney Int 58: 1261–1266, peritoneum, as early as 3 wk after induction of uremia. This is 2000 paralleled by a peak in VEGF and FGF2 expression within the 10. Wever R, Boer P, Hijmering M, Stroes E, Verhaar M, Kastelein PM, returning to normal within 6 (VEGF) or 9 wk (FGF2). J, Versluis K, Lagerwerf E, van Rijn H, Koomans H, Rabelink T: J Am Soc Nephrol 12: 2146–2157, 2001 Peritoneal Modifications in Chronic Uremia 2157

Nitric oxide production is reduced in patients with chronic renal 26. Douma CE, de Waart DR, Struijk DG, Krediet RT: The nitric failure. Arterioscler Thromb Vasc Biol 19: 1168–1172, 1999 oxide donor nitroprusside intraperitoneally affects peritoneal per- 11. Aiello S, Noris M, Todeschini M, Zappella S, Foglini C, Benigni meability in CAPD. Kidney Int 51: 1885–1892, 1997 A, Corna D, Zoja C, Cavallotti D, Remuzzi G: Renal and 27. Dawson TM, Sasaki M, Gonzalez-Zulueta M, Dawson VL: Regu- systemic nitric oxide synthesis in rats with renal mass reduction. lation of neuronal nitric oxide synthase and identification of novel Kidney Int 50: 171–181, 1997 nitric oxide signaling pathways. Prog Brain Res 118: 3–11, 1998 12. Lau T, Owen W, Yu YM, Noviski N, Lyons J, Zurakowski D, 28. Schmidt RJ, Yokota S, Tracy TS, Sorkin MI, Baylis C: Nitric Tsay R, Ajami A, Young VR, Castillo L: Arginine, citrulline, and oxide production is low in end-stage renal disease patients on nitric oxide metabolism in end-stage renal disease patients. J Clin peritoneal dialysis. Am J Physiol 276: F794–F797, 1999 Invest 105: 1217–1225, 2000 29. Reyes AA, Karl IE, Klahr S: Role of arginine in health and in 13. Boudy N, Hassler C, Parvy P, Bankir L: Renal synthesis of renal disease. Am J Physiol 267: F331–F346, 1994 arginine in chronic renal failure: In vivo and in vitro studies in 30. Krediet RT: The physiology of peritoneal solute transport and rats with 5/6 nephrectomy. Kidney Int 44: 676–683, 1993 ultrafiltration. In: The Textbook of Peritoneal Dialysis, 2nd ed., 14. Abdel-Gawad M, Huynh H, Brock GB: The impact of chronic edited by Gokal R, Khanna R, Krediet RT, Nolph KD, Dordre- renal failure on nitric oxide synthase isoforms gene expression in cht, Kluwer Academic Publishers, 2000, pp 135–172 the penis and pelvic ganglia of rats. J Urol 162: 1473–1479, 1999 31. Amann K, Wiest G, Zimmer G, Gretz N, Ritz E, Mall G: 15. Miyata T, van Yperesele de Strihou C, Kurokawa K, Baynes JW: Reduced capillary density in the myocardium of uremic rats—a Alterations in nonenzymatic biochemistry in uremia: Origin and significance of “carbonyl stress” in long-term uremic complica- stereological study. Kidney Int 42: 1079–1085, 1992 tions. Kidney Int 55: 389–399, 1999 32. Papapetropoulos A, Garcia-Cardena G, Madri JA, Sessa WC: 16. Wasada T, Kawahara R, Katsumori K, Naruse M, Omori Y: Nitric oxide production contributes to the angiogenic properties Plasma concentration of immunoreactive vascular endothelial of vascular endothelial growth factor in human endothelial cells. growth factor and its relation to smoking. Method Clin Exp 47: J Clin Invest 100: 3131–3139, 1997 27–30, 1998 33. Basilico C, Moscatelli D: The FGF family of growth factors and 17. Muchaneta-Kubara EC, Sayed-Ahmed N, El Nahas AM: Subto- oncogenes. Adv Cancer Res 59: 115–165, 1992 tal nephrectomy: A mosaic of growth factors. Nephrol Dial 34. Bikfalvi A, Klezin S, Pintucci G, Rifkin DB: Biological roles of Transplant 10: 320–327, 1995 fibroblast growth factor-2. Endocrine Rev 18: 26–45, 1997 18. Amann K, Kronenberg G, Gehlen F, Wessels S, Orth S, Munter 35. Drummond IA, Mukhopadhyay D, Sukhatme VP: Expression of K, Ehmke H, Mall G, Ritz E: Cardiac remodeling in experimen- fetal kidney growth factors in a kidney tumor line: Role of FGF2 tal renal failure: an immunohistochemical study. Nephrol Dial in kidney development. Exp Nephrol 6: 522–533, 1998 Transplant 13: 1958–1966, 1998 36. Strutz F, Zeisberg M, Hemmerlein B, Sattler B, Hummel K, 19. Schmitt CP, Hessing S, Oh J, Weber L, Ochlich P, Mehls O: Becker V, Muller GA: Basic fibroblast growth factor expres- Intermittent administration of parathyroid hormone (1–37) im- sion is increased in human renal fibrogenesis and may mediate proves growth and bone mineral density in uremic rats. Kidney autocrine fibroblast proliferation. Kidney Int 57: 1521–1538, Int 57: 1484–1492, 2000 2000 20. Lameire N, Van Biesen W, Van Landschoot M, Wang T, Heim- 37. Inagi R, Miyata T, Yamamoto T, Suzuki D, Urakami K, Saito A, bu¨rger O, Bergstro¨m J, Lindholm B, Hekking LPH, Havenith van Ypersele de Strihou C, Kurokawa K: Glucose degradation CEG, Beelen RHJ: Experimental models in peritoneal dialysis: A product methylglyoxal enhances the production of vascular en- European experience. Kidney Int 54: 2194–2206, 1998 dothelial growth factor in peritoneal cells: Role in the functional 21. Kleinknecht C, Terzi F, Burtin M, Laouari D, Maniar S: Exper- and morphological alterations of peritoneal membranes in peri- imental models of nephron reduction: Some answers, many ques- toneal dialysis. FEBS Lett 463: 206–264, 1999 tions. Kidney Int 47: S51–S54, 1995 38. Minchencko A, Bauer T, Salceda S, Caro J: Hypoxic stimulation 22. Combet S, van Landschoot M, Moulin P, Piech A, Verbavatz JM, of vascular endothelial growth factor expression in vitro and in Goffin E, Balligand JL, Lameire N, Devuyst O: Regulation of vivo. Lab Invest 71: 374–379, 1994 aquaporin-1 and nitric oxide synthase isoforms in a rat model of 39. Anand IS, Chandrashekhar Y, Wander GS, Chawla LS: Endo- acute peritonitis. J Am Soc Nephrol 10: 2185–2196, 1999 thelium-derived relaxing factor is important in mediating the 23. Combet S, Balligand JL, Lameire N, Goffin E, Devuyst O: A high output state in chronic severe anemia. J Am Coll Cardiol 25: specific method for measurement of nitric oxide synthase enzymatic activity in peritoneal biopsies. Kidney Int 57: 332–338, 1402–1406, 1995 2000 40. Fischer S, Clauss M, Wiesnet M, Renz D, Schaper W, Karliczek 24. Horie K, Miyata T, Maeda K, Miyata S, Sugiyama S, Sakai H, GF: Hypoxia induces permeability in brain microvessel endothe- van Ypersele de Strihou C, Monnier VM, Witzum JL, Kurokawa lial cells via VEGF and NO. Am J Physiol 276: C812–C820, K: Immunohistochemical colocalization of glycoxidation prod- 1999 ucts and lipid peroxidation products in diabetic renal glomerular 41. Faller DV: Endothelial cell responses to hypoxic stress. Clin Exp lesions. J Clin Invest 100: 2995–3004, 1997 Pharmacol Physiol 26: 74–84, 1999 25. Miyata T, Oda O, Inagi R, Iida Y, Araki N, Yamada N, Horiuchi 42. Hostetter TH: Progression of renal disease and renal hypertro- S, Taniguchi N, Maeda K, Kinoshita T: ␤2-microglobulin mod- phy. Annu Rev Physiol 57: 263–278, 1995 ified with advanced glycation and products is a major component 43. Rubin J, Rust P, Brown P, Popovich RP, Nolph KD: A compar- of hemodialysis-associated amyloidosis. J Clin Invest 92: 1243– ison of peritoneal transport in patients with psoriasis and uremia. 1252, 1993 Nephron 29: 185–189, 1981