Angiotensin-Converting Enzyme Inhibition but Not Angiotensin II Receptor Blockade Regulates Matrix Metalloproteinase Activity in Patients with Glomerulonephritis

Home , Fosinopril

J Am Soc Nephrol 14: 2861–2872, 2003 Angiotensin-Converting Enzyme Inhibition but not Angiotensin II Receptor Blockade Regulates Matrix Metalloproteinase Activity in Patients with Glomerulonephritis

NADE` GE LODS,* PAOLO FERRARI,* FELIX J. FREY,* ANDREAS KAPPELER,† CELINE BERTHIER,* BRUNO VOGT,* and HANS-PETER MARTI* *Division of Nephrology and Hypertension, Inselspital Bern, Bern, Switzerland; and †Institute of Pathology, University of Bern, Bern, Switzerland

Abstract. Equivalent long-term effects on the kidney are attrib- periods of 4 wk each without therapy. Untreated patients with uted to angiotensin-converting enzyme inhibitors (ACEI) and glomerulonephritis displayed distinctively higher serum levels angiotensin II type 1 receptor blockers (ARB). Nevertheless, it of MMP-2 but much lower MMP-1/-8/-9 concentrations com- is unknown to which degree effects of these compounds on pared with healthy control subjects. Immunohistology of individual inflammatory mediators, including matrix metallo- MMP-2 and MMP-9 in kidney biopsy specimen was accord- proteinases (MMP), are comparable. On the basis of structural ingly. However, these patients excreted higher amounts of and functional differences, it was hypothesized that ACEI and MMP-2 and MMP-9 in urine than healthy control subjects, ARB differentially regulate MMP activity. In a randomized, possibly reflecting ongoing glomerular inflammation. In pa- prospective crossover trial, the effect of an ACEI (fosinopril; tients with glomerulonephritis, ACEI significantly reduced 20 mg/d) and of an ARB (irbesartan; 150 mg/d) on MMP overall MMP serum activity to 25%, whereas ARB did not activity was evaluated. Ten hypertensive patients with glomer- show any effect. Activities of MMP-1/-2/-8/-9 were also sig- ulonephritis and normal or mildly reduced creatinine clearance nificantly inhibited by fosinopril but not by irbesartan. Levels were studied. MMP activity and tissue inhibitors of metallo- of TIMP-1/-2 remained unaffected. In conclusion, ACEI and proteinase (TIMP) levels were analyzed in serum and urine: ARB differentially regulate MMP activity, which may ulti- without therapy, with ACEI, with ARB, and with both agents mately have consequences in certain types of MMP-dependent combined. Treatment periods continued for 6 wk separated by glomerulonephritis.

Glomerular inflammatory diseases represent frequent and often long-term effects with regard to the kidney is not yet fully difficult-to-treat causes of end-stage kidney failure. Hyperten- resolved, and the identification of patient groups in which one sion and proteinuria are important independent determinants of these agents demonstrates potential advances over the other and risk factors for progression of these diseases (1). There- one remains of pivotal clinical interest. fore, these two conditions are prime targets for the therapy of Angiotensin II plays a key role in the pathogenesis of all types of glomerulonephritis. chronic injury to the kidneys resulting in loss of function. Pharmacologic inhibition of the renin-angiotensin system by Beyond hemodynamic effects, angiotensin II also has direct angiotensin-converting enzyme inhibitors (ACEI) and angio- effects on glomerular cells, including induction of cytokine tensin II type 1 receptor blockers (ARB) has been particularly production. Angiotensin II supports mesangial cell prolifera- well demonstrated to treat high BP and proteinuria successfully tion and induces the expression of TGF-␤, leading to extracel- (1). Moreover, the decline of kidney function occurring in lular matrix (ECM) accumulation (1,4,5). The degradation of diabetic as well as in nondiabetic glomerulopathies was greatly the ECM occurs mainly by the action of extracellularly active reduced by these agents in many investigations (2,3). Although neutral proteinases, the matrix metalloproteinases (MMP), in ARB have only recently been introduced to common clinical conjunction with other proteolytic enzymes (6–8). Impor- use, several studies reported equivalent renoprotection for both tantly, MMP seem to play an increasingly prominent role as classes of agents, ACEI and ARB (1). However, the issue of inflammatory mediators in the regulation of glomerular ECM proteins and of cell proliferation (6,7,9). Agents that are given over prolonged periods to patients with Received December 20, 2002. Accepted August 1, 2003. inflammatory diseases need to be evaluated for their effect on Correspondence to Dr. Hans-Peter Marti, Division of Nephrology/Hyperten- sion, University of Bern, Inselspital Bern, CH-3010, Bern, Switzerland. Phone: MMP activity. Taking into account the diverse effects of the 031-632-31-44; Fax: 031-632-94-44; E-mail: [email protected] renin-angiotensin system on glomerular cells and on ECM 1046-6673/1411-2861 regulating cytokines or growth factors, it is conceivable that Journal of the American Society of Nephrology ACEI and ARB influence MMP activity in patients with glo- Copyright © 2003 by the American Society of Nephrology merulonephritis. ACEI, such as captopril or ramiprilat, have DOI: 10.1097/01.ASN.0000092789.67966.5C already been demonstrated by us and others to inhibit directly 2862 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 2861–2872, 2003

MMP activity in vitro by chelating the zinc ion in the active Ten age- and gender-matched volunteers (aged 49 Ϯ 3 yr, mean Ϯ center of these proteases (10–12). Moreover, the ACEI cilaza- SEM; eight men and two women) without any health problems, pril was shown to decrease microalbuminuria and MMP-9 including the absence of abnormalities with respect to kidney function Ͼ plasma concentrations in eight patients with non–insulin-de- (creatinine clearance 80 ml/min) or to urinalysis, were investigated pendent diabetes (13). It is interesting that angiotensin II treat- as healthy control subjects. A second group of untreated patients with documented noninflammatory kidney disease (autosomal dominant ment of C57B1/6 female mice did not affect MMP-2 and polycystic kidney disease [ADPKD]) in the absence of concomitant MMP-9 activity (14). Conclusive data on the effect of ARB on illness (aged 43 Ϯ 4 yr, mean Ϯ SEM; seven men and three women) MMP and TIMP are still missing. were used as an additional control group. These patients had a similar MMP expression and activity can be analyzed in tissue degree of renal failure (creatinine, 147 Ϯ 29 ␮mol/L; creatinine sections and in serum or plasma. To our knowledge, there exist clearance, 72 Ϯ 10 ml/min, mean Ϯ SD; Ͼ30 ml/min in all cases) as no studies conclusively assessing the relationship between the study group. circulating and tissue MMP levels, especially in patients with glomerulonephritis. However, to show effects of pharmaco- Study Design logic agents on MMP activity per se, the use of serum seems The clinical study was performed in a prospective, randomized, to be appropriate and the most practicable solution, as in the triple-crossover manner, as reported previously (15). After an initial case of the present study. 6-wk run-in control period without antihypertensive agents, our 10 It is still unknown whether ACEI and ARB show identical patients with glomerulonephritis were observed during three treatment effects on these ECM regulating proteases. However, on the periods of 6 wk each. During these treatment periods, they received basis of structural and functional differences of ACEI and randomly once daily at 8:00 a.m. fosinopril 20 mg, irbesartan 150 mg, ARB, we hypothesized that these agents may well regulate or fosinopril 20 mg plus irbesartan 150 mg, in a prospective, random- MMP activity in a different manner. Thus, it was the prime aim ized, open, blinded end point (PROBE) study. Each of the three active of our investigation to analyze the effect of an ACEI, fosino- treatment phases was separated by a washout period lasting 4 wk. At the end of the control run-in phase and of each active treatment pril, and an ARB, irbesartan, alone or in combination, on MMP phase (weeks 6, 12, 22, and 32), MMP activities were analyzed activity in serum of patients with chronic glomerulonephritis. separately in serum and in morning spot urine samples. Similar The secondary aim of our study was to compare MMP activity analyses of healthy control subjects and of control patients with in serum and urine of our untreated patient group with age- and ADPKD, both groups without therapy, were performed on one gender-matched healthy control subjects and with a second occasion. control group consisting of subjects with noninflammatory renal failure. Blood and Urine Samples Samples containing 5 ml of blood were collected and centrifuged at 3000 rpm for 10 min. Thereafter, serum was stored in aliquots at Materials and Methods Ϫ Patients and Healthy Control Subjects 20°C before analyses. Urine spot samples (5 ml) were centrifuged in an identical manner to remove any suspended particles and stored in Ten consecutive patients (aged 49 Ϯ 4 yr, mean Ϯ SEM; eight men aliquots at Ϫ20°C. All creatinine measurements as well as analyses of and two women) who had biopsy-proven glomerulonephritis from our serum proteins were performed in the central chemical laboratory of renal clinic and met the entry criteria participated in the study (15). our institution. Urinary protein concentrations were measured using Relevant biochemical data are given in Table 1; creatinine clearance the BCA protein assay reagent kit (Pierce, Lausanne, Switzerland) was Ͼ30 ml/min in all patients. Renal histology was membranous (16). glomerulonephritis (n ϭ 4), focal segmental glomerulosclerosis (n ϭ 4), IgA nephropathy (n ϭ 1), and membranoproliferative glomerulonephritis (n ϭ 1). Pregnant women and patients who had diabetes or Antibodies and Reagents were on immunosuppressive therapy were excluded. Before enroll- For Western blot analysis, mouse mAb to human MMP-9 (Ab-3) ment, all antihypertensive medications were withdrawn for at least 2 were obtained from Oncogene (Juro, Lucerne, Switzerland), and the wk, except diuretics prescribed at a constant dose to control edema. secondary goat anti-mouse horseradish peroxidase–conjugated anti-

Table 1. Characterization of patients with glomerulonephritis: serum and urine values at baseline without therapy and during different treatment periods (mean Ϯ SD)a

Baseline ACEI ARB ACEI ϩ ARB

Total protein (g/L) 66 Ϯ 10 66 Ϯ 11 67 Ϯ 12 69 Ϯ 9 Albumin (g/L) 32 Ϯ 833Ϯ 833Ϯ 735Ϯ 4 Creatinine (␮mol/L) 132 Ϯ 63 145 Ϯ 82 133 Ϯ 70 140 Ϯ 75 Proteinuria (g/d) 7.9 Ϯ 7.2 5.3 Ϯ 5.2b 5.0 Ϯ 4.9b 3.3 Ϯ 3.7c Creatinine clearance (ml/min) 77 Ϯ 27 73 Ϯ 32 80 Ϯ 30 73 Ϯ 32

a ACEI, anglotensin-converting enzyme inhibitor; ARB, angiotensin II type I receptor blocker. b P Ͻ 0.01, c P Ͻ 0.005 versus baseline (15). J Am Soc Nephrol 14: 2861–2872, 2003 Angiotensin and Matrix Metalloproteinases 2863

bodies were obtained from SantaCruz Biotechnologies (Heidelberg, activation of pro-MMP forms by substances such as p-aminophe- Germany). nylmercuric acetate. PMSF for inhibition of serine proteinases and the zinc-chelating agent 1,10-phenantroline were obtained from Sigma-Aldrich (Grogg Overall MMP and MMP-2 Serum Activities Chemie, Bern, Switzerland). Purified human recombinant proMMP-2 Analyses of overall MMP activity and of MMP-2 activity in serum and proMMP-9 were acquired from Oncogene (Grogg Chemie, Bern, were performed by a continuously recording fluorescence assay, as Switzerland). For the assessment of MMP activities by continuously published previously (17,18). Serum samples were analyzed in the recording fluorescence assay, the quenched broad-spectrum fluoro- presence or absence of the serine protease inhibitor PMSF (2 mM) or genic peptide for overall MMP activity [(7-methoxycoumarin-4-yl)A- the zinc chelator 1,10-phenantroline (5 mM). The quenched broad- cetyl-Pro-Leu-Gly-Leu-(3-[2,4-dinitrophenyl]-L2,3-diaminopropio- spectrum fluorogenic peptide assessing overall MMP activity or the nyl)-Ala-Arg-NH2 (Bachem, Bubendorf, Switzerland); examples are quenched fluorogenic peptide specific for MMP-2 was added at the ϭ Ϫ1 Ϫ1 ϭ MMP-2 (kcat/Km 629,000 s M ) and MMP-7 (kcat/Km final concentration of 5 ␮M (17,18). Emission of fluorescence oc- Ϫ1 Ϫ1 169,000 s M )] or the quenched fluorogenic peptide specific for curred by MMP-induced cleavage of the Gly-Leu bond in the former MMP-2 activity [(7-methoxycoumarin-4-yl)Acetyl-Pro-Leu-Ala-L- or of the Ala-Nva bond in the latter peptide. Fluorescence was Nva-(3-[2,4-dinitrophenyl]-L2,3-diaminopropionyl)-Ala-Arg-NH2 quantified by subtracting the increase in emission of the blank sam- ϭ ϫ 5 Ϫ1 Ϫ1 (Calbiochem, Lucerne, Switzerland); kcat/Km 3.97 10 M s ] ples consisting of reagent buffer only from the samples containing was used. patient serum. The kinetic of fluorescence activity demonstrating For in vitro studies, fosinopril and fosinoprilat (Bristol-Myers linearity of cleavage rates over time was measured (excitation wave- Squibb, Baar, Switzerland) as well as irbesartan (Sanofi-Synthélabo, length, 328 nm; emission wavelength, 393 nm) every 2 min for1hat Meyrin, Switzerland) were prepared and used according to instruc- 37°C by Spectra max Gemini XS (Paul Bucher Company, Basel, tions of the manufacturers. The MMP inhibitor RO111-3456 (Mr Switzerland). Results were expressed as Vmax (units/s) using linear 425.89 g/mol), a peptide-based hydroxamic acid derivative (Roche regression for MMP activity. Thereafter, results were adjusted to the Bioscience, Palo Alto, CA), was prepared and used as described respective total protein concentration. Thus, MMP activity was ex- previously (10). pressed as Vmax/g total protein. A nonspecific interference of differences in serum protein concentrations per se with the test results has been excluded by separate experiments (data not shown). Immunohistologic Analyses of MMP-2 and MMP-9 ϭ ϭ Tissue samples from patients (n 10) and control subjects (n 5), TIMP-1, TIMP-2, and TGF-␤ Serum Levels namely normal renal cortex from nephrectomy specimen of patients Serum concentrations of TIMP-1, TIMP-2, and TGF-␤ were mea- with well-circumscribed kidney tumors, were routinely fixed in neu- sured by respective Quantikine assay kits (R&D Systems), according tral buffered 4% formaldehyde solution and embedded in paraffin. For to the instructions of the manufacturer. The optical density resulting routine analysis, hematoxylin and eosin stainings were performed from substrate cleavage was read at 450 nm, with a ␭ correction at 540 according to standard practice. nm to remove any background activity, by a spectrophotometer (Spec- For immunohistochemistry, 2- to 3-␮m paraffin-embedded sec- tra Rainbow Thermo; Tecan, Hombrechtikon, Switzerland). Serum tions were dewaxed, rehydrated, and pretreated by boiling in 100 mM concentrations were determined by interpolation from a standard Tris with 5% urea (pH 9.5) in a microwave oven. Sections were then curve. After adjustments to the respective total protein content, results (and after all subsequent steps) washed in Tris-buffered saline (TBS) were expressed in ng/mg protein (TIMP-1/-2) or in pg/mg protein and incubated with the primary antibody, either a mouse anti–MMP-2 (TGF-␤). antibody, clone A-Gel VC2 (working concentration, 2 ␮g IgG/ml; NeoMarkers, Fremont, CA), or a mouse anti–MMP-9 antibody, clone 56-2A4 (1 ␮g IgG/ml; Oncogene, San Diego, CA), diluted in TBS Inhibition of MMP-2 and MMP-9 In Vitro with 0.5% casein and 5% normal goat serum, for 60 min at room Inhibition of MMP-2 and MMP-9 in vitro was assessed by gelatin temperature. Sections were then incubated for 45 min with a 1:300 substrate zymography, as described previously (17,19). MMP-1 and dilution (same buffer as primary antibody) of a biotinylated goat MMP-8 were not included, because gelatin does not represent their anti-mouse Ig antiserum (DakoCytomation, Glostrup, Denmark). preferred substrate. SDS-PAGE of aliquots of 10 ␮l of serum adjusted Thereafter, a streptavidin-biotin-complex/alkaline phosphatase (1:200 with water to a protein content of 30 ␮g from untreated patients and in TBS; DakoCytomation) was applied for 60 min. Finally, sections healthy control subjects was performed on a 10-well, 10% polyacryl- were developed in new fuchsin-naphthol AS-BI (Sigma) for 30 min, amide minigel containing 0.1% gelatin (wt/vol). For inhibition stud- counterstained with hematoxylin, and mounted. Histology slides were ies, the ACEI fosinopril (200 ␮M and 600 ␮M) or fosinoprilat (100 analyzed and selected by a single, blinded pathologist (A.K.). nM and 1 ␮M) or the ARB irbesartan (600 ␮M) were added to the proteolysis buffer. Coomassie blue counterstaining revealed clear bands of gelatinolytic activity, which were quantified by densitomet- MMP-1, MMP-8, and MMP-9 Serum Activities ric analysis (10,17). In a separate series of experiments, aliquots of Endogenous MMP-1, MMP-8, and MMP-9 activities were mea- purified human MMP-2 (10 ng) and MMP-9 (5 ng) were subjected to sured by the use of respective Fluorokine activity assay kits (R&D zymography. Gels were treated with proteolysis buffer containing ␮ ␮ ␮ Systems, Wiesbaden-Nordenstadt, Germany), according to the in- fosinopril (200 M and 600 M), fosinoprilat (500 nM and 5 M), or ␮ structions of the manufacturer. After subtraction of background ac- the ARB irbesartan (600 M) for 18 h; as positive control, 500 nM tivities, concentrations of active MMP were determined by interpola- RO 111-3456 (MMP inhibitor) was included (10,17). tion from a standard curve. After adjustments to the total protein content, results were expressed in ng/g protein (MMP-1, MMP-8) or Western Blot Analysis of MMP-9 in Serum in ng/mg protein (MMP-9). Endogenous MMP activity refers to the Results of MMP-9 analyses were confirmed and extended by naturally occurring MMP serum activity, without exogenous in vitro Western blot analysis of serum. Aliquots of 10 ␮l of serum, adjusted 2864 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 2861–2872, 2003 with water to a protein content of 10 ␮g, were diluted in 10 ␮lof2ϫ the comparison of untreated patients with glomerulonephritis sample buffer (0.06 M Tris-HCl [pH 6.8], 2% SDS, 10% glycerol, versus age- and gender-matched healthy control subjects and 0.025% bromophenol blue) according to Laemmli (20). The samples control patients with noninflammatory renal failure (ADPKD) were heated to 95°C and centrifuged at 13,000 rpm for 1 min, and regarding the differential effects of the ACEI fosinopril followed by 10% SDS-PAGE. Subsequently, the gel was transferred and the ARB irbesartan on MMP/TIMP in patients during the to a polyvinylidene fluoride microporous membrane (Millipore, different treatment phases. Finally, the direct inhibitory effect Volketswil, Switzerland) for2hat250mA.Thereafter, the membrane was blocked for 30 min at 37°Corfor2hatroom temperature by of the ACEI fosinopril was confirmed in vitro. gently shaking in 5% nonfat dry milk (wt/vol) in TBS-T (TBS with 0.1% Tween 20; vol/vol). Untreated Patients versus Healthy Control Subjects After two washes in TBS-T, the blot was incubated overnight at and versus Subjects with Noninflammatory 4°C by gently shaking in a solution containing monoclonal mouse Renal Failure anti–MMP-9 antibodies diluted 1:1000 in TBS-T. Subsequently, the Overall MMP, MMP-1, and MMP-8 Activities in Serum. blot was washed extensively in TBS-T during1h(6ϫ 10 min) and Results of overall MMP, MMP-1, and MMP-8 activities in was incubated with goat anti-mouse horseradish peroxidase–conju- serum are summarized in Table 2. MMP activities in groups of gated secondary antibody. Finally, the membrane was incubated with healthy control subjects and untreated patients were of a sim- enhanced chemiluminescent reagents for 1 min (Amersham Pharma- ilar magnitude. Probably as a result of a higher variability of cia Biotech, Dübendorf, Switzerland), followed by autoradiography the activities observed in patients, the higher MMP activity in (Hyperfilm ECL; Amersham Pharmacia Biotech, Dübendorf, Switzerland). untreated patients of 25% versus control subjects did not reach statistical significance (P Ͼ 0.05). As summarized in Table 2, MMP-2 and MMP-9 Urine Activities there were differences between patients and healthy control Endogenous, naturally occurring MMP-2 and MMP-9 activities in subjects with respect to serum MMP-1 and MMP-8 activities. aliquots of 100 ␮l of urine were measured by use of Biotrak activity Both MMP were notably decreased in patients’ serum: MMP-1 assay kits (Amersham Pharmacia Biotech), according to the instruc- by 52% and MMP-8 by 32%, as compared with control sub- tions of the manufacturer. The cleavage of the chromogenic peptide jects. MMP activities of subjects with ADKPD were not dif- substrate resulted in a color change, read at 405 nm by a spectropho- ferent as compared with healthy control subjects. tometer (Spectra Rainbow Thermo). Concentrations of active MMP MMP-2/TIMP-2 and MMP-9/TIMP-1 in Serum. There were determined after subtraction of background activities by inter- were distinct differences between untreated patients and polation from a standard curve and were expressed in ␮g/␮mol healthy control subjects regarding MMP-2 and its main inhib- creatinine. itor TIMP-2, as depicted in Figure 1 and Table 2. Patients with glomerulonephritis displayed clearly increased serum MMP-2 Statistical Analyses activity by one third (32.5 units versus 21.3 units/s per g Differences between means were assessed by t test or ANOVA for protein; P ϭ 0.001). Levels of TIMP-2 demonstrated a clear analysis of continuous variables and by nonparametric Mann-Whitney tendency toward an increase but just did not differ to a signif- U test for variables that were not normally distributed. All statistical icant degree (P ϭ 0.05). tests were two sided. Statistical analysis was performed with the GraphPad Prism Software version 3.02 for Windows (Graphpad Soft- There were even more distinct differences between untreated ware, San Diego, CA). Results are in mean Ϯ SD for continuous patients and healthy control subjects regarding MMP-9 and its variables. For all experiments, P values were calculated; values of main inhibitor TIMP-1, as depicted in Figure 1 and Table 2. P Ͻ 0.05 were regarded as significant. In vitro experiments were Patients with glomerulonephritis displayed an approximately performed three times in duplicate. sixfold lower serum MMP-9 activity (21.1 ng versus 125 ng/mg serum protein; P ϭ 0.0003). In addition, TIMP-1 in Results patients was increased by almost 50% (4.2 ng versus 2.96 The results obtained from the investigations of MMP and ng/mg protein; P ϭ 0.007). It is interesting that in healthy TIMP were analyzed and reported separately with respect to control subjects as well as in patients, TIMP-1 levels were

Table 2. Serum levels of MMP/TIMP in healthy control subjects and in control subjects with noninflammatory renal failure (ADPKD) versus untreated patients with GN (mean Ϯ SD)a

MMP/TIMP Control subjects ADPKD Patients with GN

Overall MMP (Vmax/g protein) 7.0 Ϯ 0.2 6.8 Ϯ 0.5 8.8 Ϯ 1.1 MMP-1 (ng/g protein) 4.5 Ϯ 0.3d 5.7 Ϯ 0.9d 2.1 Ϯ 0.2 MMP-8 (ng/g protein) 28.9 Ϯ 2.9b 32.7 Ϯ 1.7b 19.7 Ϯ 1.8 TIMP-1 (ng/mg protein) 2.9 Ϯ 0.2c 2.9 Ϯ 0.3c 4.2 Ϯ 0.4 TIMP-2 (ng/mg protein) 2.0 Ϯ 0.1 1.75 Ϯ 0.08 2.7 Ϯ 0.4

a MMP, Matrix Metalloproteinase; TIMP, tissue inhibitor of metalloproteinase; ADPKD, autosomal dominant polycystic kidney disease; GN, glomerulonephritis. b P Ͻ 0.05, c P Ͻ 0.01, d P Ͻ 0.001 versus patients with GN. J Am Soc Nephrol 14: 2861–2872, 2003 Angiotensin and Matrix Metalloproteinases 2865

Figure 2. (A) Zymography of MMP-9 in serum. Results of three representative patients (P1 to P3) with their age- and gender-matched control subjects (C1 to C3) are depicted; M, 100 ng of purified human MMP-9, used as positive control. Densitometry of gelatinolytic activity in 10 healthy control subjects (Ⅺ) versus 10 untreated patients with glomerulonephritis (u) is depicted below. (B) Western blot analysis of MMP-9 in serum. Again, results of three representative patients (P1 to P3) with their matched control subjects (C1 to C3) are given; M, 100 ng of purified human MMP-9. Densitometry of bands Figure 1. Matrix metalloproteinase-2 (MMP-2) and MMP-9 serum in 10 healthy control subjects (Ⅺ) versus 10 untreated patients with levels of healthy control subjects (Ⅺ), control subjects with nonin- glomerulonephritis (u) is given below. *P Ͻ 0.05 compared with flammatory renal failure as a result of autosomal dominant polycystic control group. kidney disease (Ⅵ), and untreated patients with glomerulonephritis (u). *P Ͻ 0.05 compared with both control groups. healthy control subjects (i.e. mean 41 versus 63 arbitrary units; P ϭ 0.001). Similarly, densitometry of Western blot analysis higher than the respective TIMP-2 concentrations (P Ͻ 0.03 demonstrated a relevant decrease of MMP-9 protein in patients for all comparisons). in the order of 35% (388 versus 584 arbitrary units; P ϭ The decrease in the serum level of MMP-9 in untreated 0.002), as demonstrated in Figure 2B. MMP-2/-9 activities and patients compared with healthy control subjects was confirmed TIMP-1/-2 levels of patients with ADKPD were very similar to by gelatin zymography and by Western blot analysis. Figure healthy control subjects, as also shown in Figure 1 and Table 2A shows results of gelatin zymography of three representative 2. patients, each with the matched healthy control subject. Den- Concentration of Active MMP-2 and MMP-9 in Urine. sitometric analysis of all respective gelatinolytic bands re- Untreated patients excreted in their urine distinctively higher vealed a decrease by one third of total MMP-9 activity in the levels of both MMP-2 and of MMP-9, as compared with 10 patients with glomerulonephritis compared with the 10 healthy control subjects (Figure 3). The respective values for 2866 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 2861–2872, 2003

Figure 3. MMP-2 and MMP-9 levels in urine of healthy control subjects (Ⅺ) and untreated patients with glomerulonephritis (u). *P Ͻ 0.05 compared with the control group.

MMP-2 were 69.8 versus 21 ␮g/␮mol creatinine (P ϭ 0.007) and for MMP-9 amounted to 227 versus 23.9 ␮g/␮mol creatinine (P ϭ 0.008). Importantly, there was no clear correlation between the amount of proteinuria and the level of activity derived from either MMP-2 or MMP-9 (data not shown). Urinary MMP levels in subjects with ADPKD may be difficult to interpret, because they have a tubular disease. In this Figure 4. Immunohistology of MMP-2 (A to C) and MMP-9 (D to F) respect, tubular MMP-2 was reduced in an animal model of in two patients with glomerulonephritis and in control tissue from a polycystic kidney disease (21). Therefore, we did not investi- nephrectomy specimen. (A) MMP-2 in control kidney. (B) MMP-2 in gate urinary MMP-2 levels in subjects with ADPKD. However, a patient with membranous glomerulonephritis. (C) MMP-2 in a the urinary MMP-9 level of 28 Ϯ 4 ␮g/␮mol creatinine was subject with membranoproliferative glomerulonephritis. (D) MMP-9 almost identical to the level of healthy control subjects reported in control tissue. (E) MMP-9 in the same patient with membranous above. glomerulonephritis. (F) MMP-9 in the same subject with membrano- Immunohistology of MMP-2 and MMP-2 in Kidney Bi- proliferative glomerulonephritis. Magnification, ϫ400. opsies. Immunohistologic analyses of MMP-2 and MMP-9 were performed in the kidney biopsies of all of our patients with glomerulonephritis and in normal-appearing renal cortex from tumor nephrectomy specimen. There was distinctly stron- ger staining for MMP-2 but a clearly weaker detection of MMP-9 in essentially all glomeruli of each patient investigated, as compared with the control tissue. Representative histologic examples are depicted in Figure 4.

Patients Who Were Treated with the ACEI Fosinopril and/or the ARB Irbesartan Overall MMP Activity in Serum. Overall MMP activities in untreated patients and in patients who were treated with irbesartan were of the same magnitude (Figure 5). Treatment Figure 5. Overall MMP activity in serum of untreated patients (P) and with the ACEI fosinopril reduced MMP activity by a mean patients who were treated with irbesartan (P-ARB) or fosinopril value of approximately 60% (V ϭ 3.2 units/s per g total (P-ACEI) or in combination (P-ARB/ACEI). Results are reported for max Ⅺ protein) in comparison with the untreated patient group (P ϭ experiments with reagent buffer alone ( ), reagent buffer supple- u Ⅵ 0.0003), as measured in the assay using reagent buffer alone. mented with 2 mM PMSF ( ), or with 5 mM 1,10-phenantroline ( ). *P Ͻ 0.05 compared with P and P-ARB. Such a decline was observed in every individual patient. In comparison with ACEI alone, there was no further significant reduction of MMP activity by the combined use of ACEI with ϭ Ͼ the ARB irbesartan (Vmax 2.5 units/s per g total protein; P curred by MMP, the activity of each sample was measured in 0.05), as depicted in Figure 5. presence of the serine proteinase inhibitor PMSF and the For confirming that cleavage of fluorogenic peptide oc- zinc-chelating agent 1,10-phenantroline. As expected, the ad- J Am Soc Nephrol 14: 2861–2872, 2003 Angiotensin and Matrix Metalloproteinases 2867 dition of 2 mM serine proteinase inhibitor demonstrated no significant effect within the respective groups (P Ͼ 0.05), whereas 5 mM zinc chelator lead to a complete inhibition of ϭ cleavage of fluorogenic substrate (Vmax 0). Importantly, analyses of enzymatic activity in the presence of PMSF to further enhance assay specificity for MMP by the elimination of any intervening serine proteases (22) showed that ACEI reduced MMP activity to approximately 75% as compared with ϭ the untreated phase (Vmax 9.2 units/s per g protein versus 2 units/s per g protein; P Ͻ 0.0001). Overall MMP activity could not be compared with MMP-2 activity because of different fluorescence substrates containing different cleavage sites. MMP-1, MMP-2, MMP-8, and MMP-9 Serum Activities. Figure 6 depicts serum activities of interstitial collagenases (MMP-1, MMP-8) and of gelatinase B (MMP-9). Notably, MMP-9 activity was 1000-fold higher than MMP-8 and 10,000-fold higher than MMP-1 activity. Compared with the phase without any therapy, the ARB (irbesartan) showed no significant effect on MMP activity. The ACEI (fosinopril) reduced MMP-1 activity by 21% and MMP-8/-9 activity by 28% (P Ͻ 0.04), and the addition of irbesartan did not change these results (Figure 6). Figure 7 shows the results of MMP-2 activity by the continuously recording fluorescence assay (17). Again, irbesartan had no effect on MMP-2 activity, whereas ϭ fosinopril reduced MMP-2 activity by 40% (Vmax 19.7) ϭ ϭ compared with the untreated phase (Vmax 32.5; P 0.04). Serum Levels of TIMP-1 and TIMP-2. Serum concentrations of TIMP-1 and TIMP-2 remained unaffected by all drug regimens, as demonstrated in Figure 8. TIMP-1 levels remained stable at approximately 4 ng/mg total protein, and TIMP-2 levels stayed at approximately 2.85 ng/mg total protein. TIMP-1 levels were significantly higher than TIMP-2 during all four study phases (P Ͻ 0.03 for all comparisons). Serum Levels of TGF-␤. To extend the findings of ACEI and ARB on inflammatory mediators, we investigated the serum levels of TGF-␤, a key mediator of chronic glomerular inflammation. As reported in Table 3, there was a clear trend toward a decrease of circulating TGF-␤ concentrations in treated patients. The magnitude of TGF-␤ inhibition achieved by ACEI and ARB and their combination was very similar, although only therapy with ACEI just reached statistical significance.

Inhibition of MMP Activity In Vitro by ACEI Fosinoprilat Figure 6. MMP-1 (A), MMP-8 (B), and MMP-9 (C) activities in Exposure of Purified MMP-2 and MMP-9 to ACEI and serum quantified by Fluorokine activity kits in patients without treat- ARB. The direct effect of ACEI and ARB on MMP-2 and ment (P) and in the same subjects during irbesartan (P-ARB), fosi- Ͻ MMP-9 activities in vitro is shown in Figure 9. Fosinoprilat nopril (P-ACEI), and combined (P-ARB/ACEI) treatment. *P 0.05 compared with P. significantly inhibited both MMP-2 and MMP-9 activities by approximately 50% at a dose of 500 nM, and an even higher MMP inhibition occurred at 5 ␮M. Fosinopril showed a similar effect but only at very much higher concentrations. In accor- Exposure of Serum from Control Subjects and Patients dance with the results described above, irbesartan failed to to ACEI and ARB. The direct inhibitory effect of ACEI on exert any effect on either of the gelatinases. The known MMP MMP activity was confirmed in the case of MMP-9 by expo- inhibitor RO111-3456 (10) was used as positive control and sure of serum to fosinoprilat in vitro. Quantitative densitometry demonstrated inhibition of both gelatinases in the order of of gelatinolytic bands reflecting MMP-9 activity is summa- 80%. rized in Figure 10. MMP-9 inhibition in the order of 50% was 2868 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 2861–2872, 2003

randomized, clinical trial analyzing patients with biopsy- proven glomerular inflammatory disease. Our findings indicate that in serum of patients with glomerulonephritis, the ACEI fosinopril significantly reduced broad-spectrum endogenous MMP activity by 75% and that the individual MMP-1, MMP-2, MMP-8, and MMP-9 were also significantly inhibited by fosinopril but not by irbesartan. The degree of inhibition of these individual MMP did not reach the extent of the decrease in overall MMP activity. As it can be expected, this may indicate that our global test of MMP activity included additional MMP not represented in our selection of individual proteases. More- over, levels of TIMP-1 and TIMP-2 were not affected by either ACEI or ARB. Direct inhibition of MMP activity by ACEI but not by ARB was shown by straight exposure of purified MMP and of serum to fosinoprilat, the active metabolite of fosinopril Figure 7. Specific MMP-2 activity determined by fluorometry in and to irbesartan. In our in vitro assay, fosinoprilat showed a patients without treatment (P) and in the same subjects during irbe- 50% inhibitory effect on MMP activity in concentrations sartan (P-ARB), fosinopril (P-ACEI), and combined (P-ARB/ACEI) within the clinically observable range of blood levels, but treatment. *P Ͻ 0.05 compared with P and P-ARB. irbesartan failed to have this effect even in concentrations clearly above the therapeutic range (23,24). Progression of chronic forms of glomerulonephritis to end- stage kidney disease is closely related to the accumulation of ECM, leading to fibrosis and sclerosis (6). Therefore, MMP are increasingly recognized to play an important role in the pathogenesis of glomerulonephritis. Thus, the effects of ACEI and ARB on MMP are of pivotal clinical interest, especially in the case of glomerulonephritis, because MMP regulate physiologic and pathologic turnover of the ECM and of cell surface proteins (6,7). These proteolytic enzymes remodel the ECM in the extracellular space mainly in conjunction with serine proteinases and with a recently characterized gene family, the metalloproteinase disintegrins (8). MMP activity is regulated and contained by natural inhibitors, mainly TIMP-1 to -4 (7). Furthermore, the complex regulation of MMP activity and expression involves many cytokines or growth factors and also Figure 8. TIMP-1 (Ⅵ) and TIMP-2 (u) serum levels in patients the renin-angiotensin system. The renin-angiotensin system is without treatment (P) and in the same subjects during irbesartan involved in the regulation of MMP expression and activity in (P-ARB), fosinopril (P-ACEI), and combined (P-ARB/ACEI) treat- several ways. First, angiotensin II induces the synthesis of ment. *P Ͻ 0.05 compared with respective TIMP-1 levels. plasminogen activator inhibitor-1 (PAI-1) (1,25). PAI-1 itself inhibits the plasmin-mediated activation of MMP (1,26). Sec- achieved by the use of 1 ␮M fosinoprilat. Therefore, purified ond, ACEI have been demonstrated by us and others to inhibit MMP-9 and MMP-9 in serum were inhibited by fosinoprilat to MMP activity directly (10–12). ACEI are chelators of the a similar extent. As expected, irbesartan did not influence crucial zinc ion located in the active site of zinc metallopro- MMP activity. teases such as ACE and MMP (12,27,28). Given the potentially important therapeutic role of ACEI and Discussion ARB in glomerulonephritis, their effects on MMP were the The present study describes for the first time the differential prime focus of the present investigation. In serum of patients effects of ACEI and ARB on MMP activity in a prospective, with glomerulonephritis, the ACEI fosinopril significantly re-

Table 3. Serum levels of TGF-␤ in patients with glomerulonephritis at baseline without therapy and during different treatment periods (mean Ϯ SD)

Baseline ACEI ARB ACEI ϩ ARB

TGF-␤ (pg/mg total protein) 735 Ϯ 132 455 Ϯ 54* 512 Ϯ 60 537 Ϯ 91

* P Ͻ 0.05 for ACEI versus baseline. J Am Soc Nephrol 14: 2861–2872, 2003 Angiotensin and Matrix Metalloproteinases 2869

by irbesartan. A clear direct inhibition of MMP activity by ACEI but not by ARB was shown by exposure of purified MMP and of serum to these agents. Therefore, MMP inhibition most likely occurred in a direct and therefore angiotensin II–independent manner. Moreover, for these in vitro studies, we have used concentrations of fosinoprilat that were within the range of levels attained in human subjects (24). The constant levels of TIMP-1 and TIMP-2 that were not affected by either ACEI or ARB support the specific effect of ACEI on MMP. It is interesting that TIMP-1 levels were consistently higher than the respective TIMP-2 concentrations in treatment phases. The magnitude of MMP inhibition in our study was similar as reported in the case of trandolapril (29). This agent reduced plasma MMP-9 activity in a subgroup of patients with diabetes in the order of 37% (29). ARB showed only a mild trend toward a decrease in MMP activity. One can only speculate that this finding might have been the result of an angiotensin II–dependent limited inhibition of MMP expression (30). From Figure 9. Zymography of purified MMP-2 and MMP-9. (A) In pres- our studies, we were not able to determine to which extent ence of ACEI (fosinopril), ARB (irbesartan), or synthetic MMP MMP expression was also influenced by ACEI (31). inhibitor (RO 111-3456), as positive control, showing dose-dependent The effects of fosinopril and irbesartan on MMP were not inhibition of MMP-2 and MMP-9 by ACEI and also inhibition by RO related to different effects on BP, total serum protein, kidney 111-3456. (B) MMP-2 and MMP-9 inhibition by fosinoprilat in much function, and proteinuria, because both agents displayed iden- lower doses than the parent drug fosinopril. tical effects on these four parameters, as reported previously (15). In addition, the differential effects of ACEI and ARB on inflammatory mediators may well be somehow restricted to duced broad-spectrum endogenous MMP activity to a level of MMP, because TGF-␤ was downregulated to a similar extent only 25%, whereas the ARB irbesartan did not show any effect. by both agents, in accordance with previous publications The gelatinases (MMP-2, MMP-9) and the interstitial collag- (32–34). enases (MMP-1, MMP-8) represent particularly important It is interesting that in comparison with age- and gender- MMP (6,7). In accordance with the analyses of the overall matched healthy control subjects or with control subjects with MMP activity, the individual MMP-1, MMP-2, MMP-8, and comparable noninflammatory renal failure, untreated patients MMP-9 were also significantly inhibited by fosinopril but not displayed a distinctively lower serum MMP-9 activity but a higher MMP-2 activity. The significance of the lower MMP-9 levels in our patients was even enhanced by the concomitant increase in TIMP-1. The pathogenetic relevance of these findings remains to be defined in more detail, for instance by concurrent measurements of MMP activity in the respective renal tissue. However, our immunohistologic analyses were compatible with an increase of MMP-2 and a decrease of MMP-9 within the glomeruli of our patients. Therefore, circulating MMP levels may reflect tissue activities to a certain extent. Naturally, our studies that comprise only a limited number of patients do not allow any definite conclusions on MMP expression and activity in different types of human glomerulonephritis, as it was not the prime aim of our study. In addition to MMP-9, untreated patients with glomerulonephritis displayed distinctly decreased activities of the interstitial collagenases MMP-1 and MMP-8. One of the reasons for the lower MMP-9 activity may be an increased angiotensin II–mediated stimulation of PAI-1 synthesis (25), which leads to a decrease in the plasmin-mediated MMP-9 activation Figure 10. Densitometric analysis of zymograms of MMP-9 from (26,35). Furthermore, differential regulation of MMP has pre- serum of healthy controls (Ⅺ) and of untreated patients (u), in the viously been shown. In human mesangial cells, MMP-2 seems absence or presence of ACEI (fosinoprilat) or ARB (irbesartan). *P Ͻ to be constitutively secreted, whereas MMP-9 is affected by 0.05 compared with no ACEI/ARB. cytokines such as IL-1 (36). 2870 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 2861–2872, 2003

Differences of MMP activities between patients and healthy accumulation (48). Therefore, it is highly likely that MMP control subjects were not a consequence of the difference in inhibition over a prolonged period may interfere with the kidney function per se, because our patients with glomerulo- course of the disease. nephritis experienced only mild renal failure and because sub- Subject to the actual stage of a given glomerular inflamma- jects with comparable noninflammatory kidney failure as a tory disease, too much or too little MMP activity may occur, result of ADPKD displayed similar results as the healthy and both of these features can be detrimental. Therefore, any control group. In this respect, the slightly higher levels MMP-1 recommendations on a preferential use of ACEI or ARB in and of MMP-9 were in line with a previous investigation of a glomerulonephritis would greatly depend on at least some somewhat older group of predominantly female patients with knowledge of MMP activity in the respective tissue. ADPKD (37). Moreover, our patients with glomerulonephritis In conclusion, ACEI and ARB have a differential effect on excreted significantly higher amounts of both MMP-2 and MMP activity that may have long-term consequences in certain MMP-9 in their urine than healthy control subjects, possibly types of glomerulonephritis depending on accurate MMP ac- reflecting ongoing glomerular inflammation. Also, urinary tivity for prevention of disease progression. MMP-9 was much higher compared with MMP-2, especially taking into account the lower MMP-9 serum levels, although Acknowledgments there was no correlation between proteinuria and MMP activity Portions of this study were published as an abstract (JAmSoc of any of the two proteases. Therefore, increased urinary ex- Nephrol 13: 14A, 2002) and were presented as oral communication at cretion of MMP may not just represent augmented MMP loss the 35th annual meeting of the American Society of Nephrology, accompanying increased proteinuria. Philadelphia, October 2002. In the early stage of diabetic nephropathy in humans, an We are thankful to J. Caulfield (Roche Bioscience, Palo Alto, CA) increased plasma MMP-9 activity was found to precede even for donating RO111-3456. the evolution of microalbuminuria in type 2 diabetes (13,38). This work was supported by the Swiss National Foundation for An increased plasma level of MMP-9 was also found in type 1 Scientific Research by the grant 31-55779.98 (to H.-P.M) and by an diabetes in the absence of significant vascular disease (39). unconditional grant from Sanofi-Synthélabo, Meyrin, Switzerland (to H.-P.M.). Beyond diabetes, there are reports about increased monocytic expression of MMP-9 in patients with IgA nephropathy and increased serum levels of MMP-3 in patients with lupus ne- References phritis (40,41). Nevertheless, the exact role of MMP in human 1. Taal MW, Brenner BM: Renoprotective benefits of RAS inhibi- types of glomerulonephritis remains to be defined, and there tion: From ACEI to angiotensin II antagonists. Kidney Int 57: 1803–1817, 2000 are no established blood levels for any MMP in such patients. 2. Lewis EJ, Hunsicker LG, Bain RP, Rohde RD: The effect of In this respect, our studies also helped to determine serum angiotensin-converting-enzyme inhibition on diabetic nephropa- MMP concentrations in populations of patients and healthy thy. N Engl J Med 329: 1456–1462, 1993 control subjects. 3. Ruggenenti P, Perna A, Gherardi G, Gaspari F, Benini R, Re- In animals, increased MT1-MMP and MMP-2 expression in muzzi G: Renal function and requirement for dialysis in chronic experimental mesangial proliferative glomerulonephritis has nephropathy patients on long-term ramipril: REIN follow-up been found by us and others in the past (19,42,43). Conse- trial: Gruppo Italiano di Studi Epidemiologici in Nefrologia quently, the inflammatory features of this disease, such as (GISEN): Ramipril Efficacy in Nephropathy. Lancet 352: 1252– increased mesangial cell proliferation and ECM deposition, 1256, 1998 were significantly attenuated by the use of an MMP inhibitor 4. Kagami S, Border WA, Miller DE, Noble NA: Angiotensin II (19). Furthermore, increased expression of MT1-MMP and stimulates extracellular matrix protein synthesis through induction of transforming growth factor-beta expression in rat glomer- MMP-2 were also found in rat anti–glomerular basement ular mesangial cells. J Clin Invest 93: 2431–2437, 1994 membrane disease (44). In addition, MMP-9 showed an en- 5. Ray PE, Bruggeman LA, Horikoshi S, Aguilera G, Klotman PE: hanced expression in Heymann nephritis, a model of membra- Angiotensin II stimulates human fetal mesangial cell prolifera- nous nephropathy (45). It is interesting that MMP-9 may also tion and fibronectin biosynthesis by binding to AT1 receptors. be protective, as shown for the development of fibrin-induced Kidney Int 45: 177–184, 1994 glomerular damage (46). In this study, susceptibility to the 6. Lenz O, Elliot SJ, Stetler-Stevenson WG: Matrix metalloprotein- accelerated model of anti–glomerular basement membrane ne- ases in renal development and disease. J Am Soc Nephrol 11: phritis was enhanced in MMP-9(Ϫ/Ϫ) compared with MMP- 574–581, 2000 9(ϩ/ϩ) mice. In nephrotic rats, it was suggested that metallo- 7. Woessner JF, Nagase H: Matrix Metalloproteinases and TIMPs, proteinases from different sources, such as intrinsic renal cells Oxford, Oxford University Press, 2000, pp 1–10, 109–135 and infiltrating phagocytes, may play a role in the development 8. Mahimkar RM, Baricos WH, Visaya O, Pollock AS, Lovett DH: Identification, cellular distribution and potential function of the of proteinuria by their direct action on the glomerular filtration metalloproteinase-disintegrin MDC9 in the kidney. JAmSoc barrier (47). Nephrol 11: 595–603, 2000 The ultimate consequences of the differential action of ACEI 9. Turck J, Pollock AS, Lee LK, Marti HP, Lovett DH: Matrix and ARB on MMP activity remain to be determined. MMP metalloproteinase 2 (gelatinase A) regulates glomerular mesan- have a dual role: maintenance of accurate, physiologic catab- gial cell proliferation and differentiation. J Biol Chem 271: olism of matrix proteins and prevention of pathogenic ECM 15074–15083, 1996 J Am Soc Nephrol 14: 2861–2872, 2003 Angiotensin and Matrix Metalloproteinases 2871

10. Daniel C, Duffield J, Brunner T, Steinmann-Niggli K, Lods N, 28. Salvetti A: Newer ACE inhibitors. A look at the future. Drugs Marti HP: Matrix metalloproteinase inhibitors cause cell cycle 40: 800–828, 1990 arrest and apoptosis in glomerular mesangial cells. J Pharmacol 29. Nakamura T, Ushiyama C, Suzuki S, Hara M, Shimada N, Exp Ther 297: 57–68, 2001 Ebihara I, Koide H: Urinary excretion of podocytes in patients 11. Sorbi D, Fadly M, Hicks R, Alexander S, Arbeit L: Captopril with diabetic nephropathy. Nephrol Dial Transplant 15: 1379– inhibits the 72 kDa and 92 kda matrix metalloproteinases. Kidney 1383, 2000 Int 44: 1266–1272, 1993 30. Chen H, Li D, Mehta JL: Modulation of matrix metalloprotein- 12. Reinhardt D, Sigusch HH, Hensse J, Tyagi SC, Korfer R, Figulla ase-1, its tissue inhibitor, and nuclear factor-kappa B by losartan HR: Cardiac remodelling in end stage heart failure: Upregulation in hypercholesterolemic rabbits. J Cardiovasc Pharmacol 39: of matrix metalloproteinase (MMP) irrespective of the underly- 332–339, 2002 ing disease, and evidence for a direct inhibitory effect of ACE 31. Li H, Simon H, Bocan TM, Peterson JT: MMP/TIMP expres- inhibitors on MMP. Heart 88: 525–530, 2002 sion in spontaneously hypertensive heart failure rats: The 13. Ebihara I, Nakamura T, Shimada N, Koide H: Plasma matrix effect of ACE- and MMP-inhibition. Cardiovasc Res 46: metalloproteinase-9 and diabetic microalbuminuria: Tip of the 298–306, 2000 iceberg? Am J Kidney Dis 32: 669–671, 1998 32. Yoshiji H, Kuriyama S, Yoshii J, Ikenaka Y, Noguchi R, Naka- 14. Tamarat R, Silvestre JS, Durie M, Levy BI: Angiotensin II tani T, Tsujinoue H, Fukui H: Angiotensin-II type 1 receptor angiogenic effect in vivo involves vascular endothelial growth interaction is a major regulator for liver fibrosis development in factor- and inflammation-related pathways. Lab Invest 82: 747– rats. Hepatology 34: 745–750, 2001 756, 2002 33. Yu CM, Tipoe GL, Wing-Hon Lai K, Lau CP: Effects of com- 15. Ferrari P, Marti HP, Pfister M, Frey FJ, Ferrari P: Additive bination of angiotensin-converting enzyme inhibitor and angio- antiproteinuric effect of combined ACE inhibition and angioten- tensin receptor antagonist on inflammatory cellular infiltration sin II blockade. J Hypertens 20: 125–130, 2002 and myocardial interstitial fibrosis after acute myocardial infarc- 16. Smith PK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH, tion. J Am Coll Cardiol 38: 1207–1215, 2001 Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ, Klenk 34. Peters H, Ruckert M, Gaedeke J, Liefeldt L, Ketteler M, Sharma DC: Measurement of protein using bicinchoninic acid. Anal AM, Neumayer HH: Angiotensin-converting enzyme inhibition Biochem 150: 76–85, 1985 but not beta-adrenergic blockade limits transforming growth 17. Steinmann-Niggli K, Lukes M, Marti HP: Rat mesangial cells factor-beta overexpression in acute normotensive anti-thy1 glo- and matrix metalloproteinase inhibitor: Inhibition of 72-kDa type merulonephritis. J Hypertens 21: 771–780, 2003 IV collagenase (MMP-2) and of cell proliferation. JAmSoc 35. Ramos-DeSimone N, Hahn-Dantona E, Sipley J, Nagase H, Nephrol 8: 395–405, 1997 French DL, Quigley JP: Activation of matrix metalloproteinase-9 18. Knight CG, Willenbrock F, Murphy G: A novel coumarin- (MMP-9) via a converging plasmin/stromelysin-1 cascade en- labelled peptide for sensitive continuous assays of the matrix hances tumor cell invasion. J Biol Chem 274: 13066–13076, metalloproteinases. FEBS Lett 296: 263–266, 1992 1999 19. Steinmann-Niggli K, Küng M, Ziswiler R, Marti HP: Inhibition 36. Martin J, Knowlden J, Davies M, Williams JD: Identification and of matrix metalloproteinases attenuates anti-Thy1.1 nephritis. independent regulation of human mesangial cell metalloprotein- J Am Soc Nephrol 9: 397–407, 1998 ases. Kidney Int 46: 877–885, 1994 20. Laemmli UK: Cleavage of structural proteins during the assem- bly of the head of bacteriophage T4. Nature 227: 680–685, 1970 37. Nakamura T, Ushiyama C, Suzuki S, Ebihara I, Shimada N, 21. Schaefer L, Han X, Gretz N, Hafner C, Meier K, Matzkies F, Kode H: Elevation of serum levels of metalloproteinase-1, tissue Schaefer RM: Tubular gelatinase A (MMP-2) and its tissue inhibitor of metalloproteinase-1, and plasma levels of metallo- inhibitors in polycystic kidney disease in the Han:SPRD rat. proteinase-9 in polycystic kidney disease. Am J Nephrol 20: Kidney Int 49: 75–81, 1996 32–36, 2000 22. Todor DR, Lewis I, Bruno G, Chyatte D: Identification of a 38. Yee J: Increased plasma metalloproteinase-9 concentrations pre- serum gelatinase associated with the occurrence of cerebral an- cede development of microalbuminuria in non-insulin-dependent eurysms as pro-matrix metalloproteinase-2. Stroke 29: 1580– diabetes mellitus. Am J Kidney Dis 32: 544–550, 1998 1583, 1998 39. Maxwell PR, Timms PM, Chandran S, Gordon D: Peripheral 23. Vachharajani NM, Shyu WC, Smith RA, Greene DS: The effects blood level alterations of TIMP-1, MMP-2 and MMP-9 of age and gender on the pharmacokinetics of irbesartan. Br J in patients with type 1 diabetes. Diabet Med 18: 777–780, Clin Pharmacol 46: 611–613, 1998 2001 24. O’Grady P, Yee KF, Lins R, Mangold B: Fosinopril/hydrochlo- 40. Koide H, Nakamura T, Ebihara I, Tomino Y: Increased mRNA rothiazide: Single dose and steady-state pharmacokinetics and expression of metalloproteinase-9 in peripheral blood monocytes pharmacodynamics. Br J Clin Pharmacol 48: 375–381, 1999 from patients with immunoglobulin A nephropathy. Am J Kidney 25. Feener EP, Northrup JM, Aiello LP, King GL: Angiotensin II Dis 28: 32–39, 1996 induces plasminogen activator inhibitor-1 and -2 expression in 41. Kotajima L, Aotsuka S, Fujimani M, Okawa-Takatsuji M, Ki- vascular endothelial and smooth muscle cells. J Clin Invest 95: noshita M, Sumiya M, Obata K: Increased levels of matrix 1353–1362, 1995 metalloproteinase-3 in sera from patients with active lupus ne- 26. Baricos WH, Cortez SL, El-Dahr SS, Schnaper HW: ECM deg- phritis. Clin Exp Rheumatol 16: 409–415, 1998 radation by cultured human mesangial cells is mediated by a 42. Lovett DH, Johnson RJ, Marti HP, Martin J, Davies M, Couser PA/plasmin/MMP-2 cascade. Kidney Int 47: 1039–1047, 1995 WG: Structural characterization of the mesangial cell type IV 27. Ondetti MA: Structural relationships of angiotensin converting- collagenase and enhanced expression in a model of immune enzyme inhibitors to pharmacologic activity. Circulation 77(6 Pt complex-mediated glomerulonephritis. Am J Pathol 141: 85–98, 2): I74–I78, 1988 1992 2872 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 2861–2872, 2003

43. Hayashi K, Osada S, Shofuda K, Horikoshi S, Shirato I, Tomino 46. Lelongt B, Bengatta S, Delauche M, Lund LR, Werb Z, Ronco Y: Enhanced expression of membrane type-1 matrix metallopro- PM: Matrix metalloproteinase 9 protects mice from anti-glomer- teinase in mesangial proliferative glomerulonephritis. JAmSoc ular basement membrane nephritis through its fibrinolytic activ- Nephrol 9: 2262–2271, 1998 ity. J Exp Med 193: 793–802, 2001 44. Hayashi K, Horikoshi S, Osada S, Shofuda K, Shirato I, Tomino 47. Davin JC, Davies M, Foidart JM, Foidart JB, Dechenne CA, Y: Macrophage-derived MT1-MMP and increased MMP-2 ac- Mathieu PR: Urinary excretion of neutral proteinases in ne- tivity are associated with glomerular damage in crescentic glo- phrotic rats with a glomerular disease. Kidney Int 31: 32–40, merulonephritis. J Pathol 191: 299–305, 2000 1987 45. McMillan JI, Riordan JW, Couser WG, Pollock AS, Lovett DH: 48. Andrews KL, Betsuyaku T, Rogers S, Shipley JM, Senior RM, Characterization of a glomerular epithelial cell metalloproteinase Miner JM: Gelatinase B (MMP-9) is not essential in the normal as matrix metalloproteinase-9 with enhanced expression in a kidney and does not influence progression of renal disease in a model of membranous nephropathy. J Clin Invest 97: 1094– mouse model of Alport syndrome. Am J Pathol 157: 303–311, 1101, 1996 2000