Agmatine Inhibits Cell Proliferation and Improves Renal Function in Anti–Thy-1 Glomerulonephritis

J Am Soc Nephrol 11: 2256–2264, 2000 Agmatine Inhibits Cell Proliferation and Improves Renal Function in Anti–Thy-1 Glomerulonephritis

SHUNJI ISHIZUKA, ROBYN CUNARD, SIRIA POUCELL-HATTON, LUCINDA WEAD, MARK LORTIE, SCOTT C. THOMSON, FRANCIS B. GABBAI, JOSEPH SATRIANO, AND ROLAND C. BLANTZ Division of Nephrology-Hypertension and Pathology, University of California, Department of Medicine, San Diego, and VA San Diego Healthcare System, La Jolla, California.

Abstract. Changes in the expression of alternate arginine met- production in Thy-1 nephritis rats by 23% and 41% at days 1 abolic pathways have been implicated in the pathogenesis of and 4, respectively. Agmatine treatment also reduced ODC experimental glomerulonephritis. Agmatine, decarboxylated activity and glomerular 3H-thymidine incorporation on days 1, arginine, has been shown in vitro to suppress both inducible 4, and 7. Histologic evaluation revealed a decline in mesangial nitric oxide synthase and the rate-limiting enzyme of poly- cell proliferation and extracellular matrix accumulation asso- amine biosynthesis, ornithine decarboxylase (ODC). This ciated with agmatine treatment administered before or 24 h study was undertaken to determine whether agmatine admin- after Thy-1 antibody, and this was confirmed by a reduction in istration could reduce tissue injury by decreasing nitric oxide, the number of cells expressing proliferating cell nuclear anti- and reduce cell proliferation, by diminishing ODC activity, in gen on days 4 and 7. These studies provide the first in vivo experimental mesangial proliferative glomerulonephritis evidence that agmatine administration can reduce cellular pro- (Thy-1 nephritis). Agmatine treatment (50 mg/kg per d intra- liferation in Thy-1 nephritis and attenuate the initial reduction peritoneally) in Thy-1 nephritis rats prevented a reduction in in renal function associated with this model. GFR at day 1. Agmatine treatment decreased nitric oxide

Two well-described pathways of L-arginine metabolism in extracellular matrix (3). Therefore, the production of ornithine inflammation include the conversion of arginine to nitric oxide and its subsequent metabolism to polyamines and proline are (NO) by nitric oxide synthase (NOS), and the breakdown of important elements of the repair phase. L-arginine to urea and L-ornithine by arginase. These path- A third arginine metabolic pathway, the conversion of argi- ways are temporally regulated in inflammatory models of nine to agmatine by arginine decarboxylase (ADC), has re- wound healing and glomerulonephritis (GN) (1,2). In the early cently been described in mammals (7). ADC activity is prev- injury phase of inflammation, there is induction of inducible alent in the kidney, liver, and brain. Similarly, agmatine is NOS (iNOS) expression and generation of NO (3–5). High- present in the plasma and has been observed in multiple tissues output NO generation from iNOS during periods of cellular and cell types (8–10). Agmatine exerts functional effects stress is known to exert cytostatic and cytotoxic effects. This is within the kidney, elevating single nephron filtration rate followed by the production of ornithine, which initiates the (SNGFR) via vasodilatation by constitutive NOS (cNOS) and repair phase (1). Ornithine is converted to polyamines via the ryanodine channel-dependent mechanisms (10,11). Agmatine rate-limiting polyamine biosynthetic enzyme, ornithine decar- also suppresses NO generation by iNOS in vitro (12,13; Satri- boxylase (ODC). Polyamines are required components of cell ano et al., submitted). Studies have recently shown that agma- cycle entry and progression and thus are essential for cellular proliferation (6). Ornithine is also metabolized by ornithine tine inhibits ODC activity and polyamine transport by an aminotransferase (OAT) to proline, an important constituent of ODC-antizyme–dependent mechanism (14). These data sug- gest that agmatine could exert important effects on both NO production and cell proliferation in inflammatory conditions. Received July 27, 1999. Accepted April 23, 2000. It is widely recognized that NO production is increased after Correspondence to Dr. Roland C. Blantz, Professor and Head, Division of induction of accelerated nephrotoxic-serum GN, active Hey- Nephrology-Hypertension, UCSD & VA Medical Center, 3350 La Jolla Vil- lage Drive, San Diego, CA 92161. Phone: 619-552-7528; Fax: 619-552-7549; mann GN, in situ immune complex GN, and anti–Thy-1 anti- E-mail: [email protected] body–induced GN (Thy-1 nephritis) (4,5,15). Thy-1 nephritis A portion of this work was presented at the annual meeting of the American is a model of acute mesangial proliferative GN in rats that is Society of Nephrology, 1998. characterized by early mesangial cell (MC) injury, followed by S.I. and R.C. share first-author status. MC proliferation (16,17). Augmentation of arginase activity 1046-6673/1112-2256 Journal of the American Society of Nephrology and ODC and OAT expression in the repair phase result in Copyright © 2000 by the American Society of Nephrology cellular proliferation and extracellular matrix production in this J Am Soc Nephrol 11: 2256–2264, 2000 Effects of Agmatine in Thy-1 Nephritis 2257 model (3,18–20). Therefore, we chose the Thy-1 nephritis showing moderate lucency (25 to 50% disruption of MC) with pres- model to investigate the in vivo effects of agmatine on injury ervation of the underlying glomerular tuft architecture; 3 ϭ MA and repair. showing severe lucency (50 to 75%) with degeneration and disruption ϭ Our results provide the first evidence that agmatine can limit of MC; 4 MA showing complete dissolution (75 to 100%) with MC proliferation and improve renal function in experimental disappearance of MC, usually in association with microaneurysm formation. Matrix score: 0 ϭ decrease in mesangial matrix (MM); 1 mesangial proliferative GN. ϭ no increase in MM; 2 ϭ slight increase in MM; 3 ϭ moderate increase in MM; 4 ϭ almost confluent appearance of MM. Each score Materials and Methods reflects changes in the extent rather than in the intensity of MM Disease Model and Experimental Protocol staining. Proliferation score: 0 ϭ no nests of proliferating cells (PC); Animal experiments were performed using male Sprague-Dawley 1 ϭ normal number of PC; 2 ϭ slight increase in PC; 3 ϭ moderate rats that weighed 200 to 225 g. Rats were allowed free access to increase in PC; 4 ϭ substantial increase in PC. To compare the normal rat chow and tap water. The mouse monoclonal antibody glomerular cell number quantitatively, we individually counted nuclei against rat Thy-1.1 (OX7) was a generous gift from Dr. T. Yamamoto and polymorphonuclear leukocytes (PMN) in 30 glomeruli from each (Niigata University School of Medicine, Japan). This antibody binds rat. to the Thy-1 antigen expressed on the MC membrane and leads to mesangiolysis followed by marked MC proliferation (16,17). The Immunohistochemical Staining for Proliferating Cell animals were divided into three groups. The Thy-1 nephritis group Nuclear Antigen ϭ (Thy-1, n 18) received a single 0.1 ml/kg body wt intravenous Tissue sections in metharcarn solution were incubated overnight at injection of anti–Thy-1 antibody on day 0. The agmatine-treated 4°C with 19A2, a murine IgM monoclonal antibody against human ϩ ϭ Thy-1 nephritis group (Thy-1 Ag, n 18) received a 50 mg/kg body proliferating cell nuclear antigen (PCNA; Coulter cytometry), fol- wt intraperitoneal injection of agmatine (Sigma Chemical Co., St. lowed by a peroxidase-conjugated rat anti-mouse IgM (Zymed Lab- Louis, MO) 2 h before administration of anti–Thy1-antibody and then oratory, San Francisco, CA). PCNA is an acidic nuclear protein that ϭ daily. The agmatine group (Ag, n 18) served as a positive control increases from the late G1 to S phases of the cell cycle, followed by and received the same daily intraperitoneal injections of agmatine as a decrease in G2/M. Black nuclear staining was detected using dia- ϩ the Thy-1 Ag group. Untreated rats were used as a normal control minobenzidine (Dakopatts, Glostrup, Denmark), and slides were ϭ group (Control, n 6). In a separate group of animals, agmatine was counterstained with methyl green. The negative control consisted of administered 24 h after disease induction to confirm that the alter- substitution of the primary antibody with a normal mouse IgM mono- ations in proliferation were not due to agmatine’s ability to limit initial clonal antibody (Dakopatts). To determine the number of proliferating ϭ tissue injury (N 5). At each time point, rats were anesthetized with cells in individual animals, a blinded observer evaluated 30 to 40 sodium brevital (65 mg/kg intraperitoneally) and killed immediately glomerular profiles containing more than 20 discrete capillary seg- after the kidneys were removed at days 0 (Control), 1 (24 h), 4, and ments, and positive staining was defined as black nuclear staining. 7 after anti–Thy-1 antibody injection. The kidneys were immediately fixed for microscopic studies. Kidney cortical slices were obtained for Immunofluorescence Staining for Macrophage/ measurement of ODC activity, and glomeruli isolated from the cortex were used for the cell proliferation, nitrite, and guanosine 3Ј,5Ј-cyclic Monocytes and Anti–Thy-1 Antibody monophosphate (cGMP) assays. To assess whether agmatine reduces anti–Thy-1 antibody binding to glomeruli, thereby causing an alteration in the extent of injury, we performed immunofluorescence studies. Anti–Thy-1 antibody has Awake Animal Studies of GFR and Urinary been shown to bind maximally to glomeruli in the first 24 h after Protein Excretion administration (16). Therefore, we assessed the amount of Thy-1 To measure 3H-inulin clearance, cannulated rats were prepared as antibody binding to mesangial cells 24 h after disease induction. The previously described (21,22). In brief, animals were anesthetized and kidney cortices were snap-frozen in OCT compound, sectioned at 3 catheters were placed in the left femoral artery, vein, and bladder. ␮m, fixed with acetone for 7 min, washed twice in phosphate-buffered After a 3-d recovery from the cannulation procedure, rats were di- saline (PBS), and then incubated with goat FITC-labeled anti-mouse vided into three experimental groups: Thy-1–treated rats, Thy-1ϩAg– IgG (Sigma Chemical Co.) to detect the binding of anti–Thy-1 anti- treated rats, and agmatine-only–treated rats (n ϭ 6 per group). Three body to the MC surface. The sections were also incubated with measurements of GFR in each animal were performed at days 0, 1, 4, anti-ED1, a mouse monoclonal FITC-labeled anti-rat monocyte/mac- and 7. For determination of urinary protein excretion, a 24-h urine rophage (Serotec, Oxford, England) antibody. The glomerular ED-1 collection was filtered, precipitated by 10% trichloroacetic acid positive cells were counted in 30 glomeruli from each rat. The (TCA), collected by centrifugation, and measured by the method of intensity of binding of anti–Thy-1 antibody was graded from 0 to 4ϩ. Lowry et al. (23). ODC Activity Assay Renal Morphology The activity of ODC was determined by measuring the conversion 14 14 Kidney tissue was fixed in 10% buffered formalin and embedded in of L-(1- C)-ornithine to CO2. After the kidneys were removed, paraffin. The paraffin-embedded tissues were sectioned and stained sections of the cortex were immediately suspended in ODC reaction with periodic acid Schiff. Three observers, blinded to the treatments, buffer (10 mM Tris [pH 7.4], 2.5 mM dithiothreitol, 0.3 mM pyri- semiquantitatively graded glomerular injury, extracellular matrix ac- doxal-5-phosphate, 0.1 mM ethylenediaminetetraacetate) and homog- cumulation, and cellular proliferation of each quadrant in 30 glomeruli enized for 10 s. The homogenate was centrifuged at 30,000 ϫ g for 40 per kidney on a scale from 0 to 4 using the following scales. Injury min, and the supernatants were assayed for ODC activity as described score: 0 ϭ absence of mesangiolysis; 1 ϭ mesangial area (MA) (14). Briefly, large bore tubes capped with rubber stoppers fitted with showing slight lucency (0 to 25% disruption of MC); 2 ϭ MA metabolic wells (Kontes, Vineland, NJ) containing trapping agent 2258 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 2256–2264, 2000

(Solvable, Packard Instruments, Meriden, CT) were used to incubate The supernatants obtained from the aliquots were ether extracted and the supernatants at 37°C for1hinthepresence of 0.1 ␮Ci (1-14C)- lyophilized, and the remaining pellets were used for RIA determina- ornithine (Dupont-NEN, Boston, MA). TCA was added to terminate tion of cGMP with an RIA kit (Dupont-NEN). the reaction, and after an additional 1 h, equilibration ␤ scintillation counting was performed. Aliquots of the supernatants in the Thy-1 Statistical Analyses 6 group were also incubated for1hwith1mML-N-(1-iminoethyl)- All data are expressed as the mean Ϯ SEM. Histologic data were lysine, (L-NIL; Alexis Biochemicals, San Diego, CA), a specific averaged across three raters, across 30 glomeruli within each quad- iNOS inhibitor, to determine in vitro whether changes in NO could rant, and across quadrants to yield mean injury, proliferation, and alter ODC activity. matrix scores for each animal. The data were then evaluated with a two-way ANOVA and post hoc Tukey tests. GFR was measured Isolation and Incubation of Glomeruli repeatedly in each animal, and the data were analyzed by repeated After the kidneys were removed, they were decapsulated and measures ANOVA. The remainder of the data were analyzed by bisected and the cortex was carefully dissected free. Cortices were two-way ANOVA followed by Tukey test for multiple group gently pressed through a 180- then a 106-␮M stainless steel sieve. The comparisons. resulting material was suspended in Krebs buffer at 4°C and gassed

with 95% O2/5% CO2 . The buffer consisted of 125 mM NaCl, 5 mM Results KCl, 1 mM CaCl2, 0.75 mM NaH2PO4, 0.75 mM Na2HPO4, 0.5 mM GFR and Urinary Protein Excretion MgCl2, 10 mM glucose, and 15 mM NaHCO3. Then the suspension In Figure 1, the GFR in each group is compared with its ␮ was passed through a 75- M sieve. The glomeruli trapped on the control. On day 1, the GFR in the Thy-1 group declined to 83% sieve were washed and pelleted by centrifugation at 3000 RPM for 1 of control levels. However, in the Thy-1 rats treated with min. After graded sieving, glomeruli were suspended in Dulbecco’s Ͻ modified Eagle’s medium (DMEM) without phenol red for the cell agmatine, there was a significant increment in the GFR (P ϩ proliferation and nitrite assays. The remaining pellet was resuspended 0.05). This elevation in GFR persisted in the Thy-1 Ag group in Krebs buffer, aliquoted, and used for the cGMP assay. on day 4, but the values did not reach statistical significance. In the agmatine-only–treated group, there was no appreciable Glomerular Cell Proliferation Assay change in GFR. Proteinuria in the Thy-1 and Thy-1ϩAg The proliferation assay was performed using the method described groups was similar on day 1 (167 Ϯ 32 versus 158 Ϯ 28, by Ketteler et al. (3). After graded sieving, cell strainers were placed respectively); however, by days 5 through 7, there was a in a six-well plate and glomeruli resuspended in DMEM (3 ml/well) significant decline in protein excretion (day 7: 113 Ϯ 5 versus were pipetted onto the cell strainers. 3H-thymidine (1 ␮Ci/well) was 86 Ϯ 8, P Ͻ 0.05, Thy-1 versus Thy-1ϩAg). added to the cell strainers and incubated for 24 h. The cell strainers were then removed from the six-well plate, rinsed with cold PBS, Histologic Studies placed into the six-well plate again, and washed three times for 10 min The glomeruli on day 1, as assessed by light microscopy, with 2 ml of 5% TCA per well. Then 2 ml of 0.5 M NaOH/0.1% showed signs of injury with diminished cell numbers in both sodium dodecyl sulfate was added to each well, and the plates were agitated at 100 RPM for 30 min. Samples were neutralized with 0.6 mol/L HCl, and radioactivity in 500-␮l aliquots was determined.

Nitrite Assay Nitrite was measured using the Griess reaction (24). After the isolated glomeruli were incubated for 24 h in DMEM in a 24-well plate, the supernatants were harvested and centrifuged at 14,000 ϫ g for 5 min to remove the glomeruli. Total nitrite was then measured at 550 nm in an autoplate reader. Isolated glomeruli in the Thy-1 group were also incubated for 24 h in DMEM with 100 ␮M L-NIL or 100 ␮M agmatine to assess in vitro their inhibitory effects on NO pro- duction. Assays were done in triplicate.

cGMP Assay We measured cGMP, a second messenger of NO, to estimate indirectly glomerular NO activity. The aliquots (50 ␮l/tube) of glo- meruli resuspended in Krebs buffer were incubated at 37°C for 10 min and then were added to 50 ␮l of Krebs buffer with 1 mM 3-isobutyl- 1-methyl-xanthine (Sigma Chemical Co.) to inhibit degradation of cGMP. The aliquots were incubated for an additional 4 min, then 100 Figure 1. GFR in Thy-1–treated and –untreated rats and effects of ␮l of 10% TCA was added to terminate the reaction. The aliquots agmatine administration. Each point is the mean Ϯ SEM of six rats, were centrifuged at 3000 RPM at 4°C for 20 min. The pellet was and the control value is the mean GFR of each group on day 0. On day extracted in 2 N NaOH and used for protein determination. The 1, there was a significant increase in GFR in the Thy-1 rats treated aliquots in the Thy-1 group were also preincubated for 10 min with 10 with agmatine as compared with the Thy-1 group. The agmatine- mM sodium nitroprusside (SNP; Sigma Chemical Co.) to assess only–treated group had no significant change in GFR during the study whether a direct NO donor could elicit an increase in cGMP levels. period. *, P Ͻ0.05 compared with Thy-1 group. J Am Soc Nephrol 11: 2256–2264, 2000 Effects of Agmatine in Thy-1 Nephritis 2259

Figure 2. Light microscopic assessment of renal morphology in Thy-1 rats. (A) A normal glomerulus. Glomeruli from rats in the Thy-1 group developed severe mesangiolysis (B, day 1), characterized by reduced nuclear number (arrowhead). On day 4 (C), there was focal and segmental hypercellularity in the mesangium and increased mesangial matrix (arrowhead). By day 7, there was a marked increase in mesangial cell (MC) number and extracellular matrix (D, nests of proliferating cells [arrowhead]). Mesangiolysis in the Thy-1ϩAg group (E, day 1, microaneurysm formation [arrowhead]) was similar as compared with the Thy-1 group. However, there was a significant decrease in MC proliferation and extracellular matrix accumulation in the Thy-1ϩAg group on days 4 and 7 (F, G). Periodic acid-Schiff (PAS), ϫ500. 2260 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 2256–2264, 2000 the Thy-1 and Thy-1ϩAg groups as compared with the con- were not secondary to changes in the initial degree of injury. trols (Figure 2, B and E, versus Figure 2A, respectively), and The matrix and proliferation scores again were reduced in the the histologic appearance of the two Thy-1–treated groups was agmatine-treated group, and ODC activity was reduced by similar. Our evaluation showed no significant difference in the almost 50% by agmatine treatment (Table 2). These findings injury score in the two groups at day 1, but the sample size may suggest that agmatine’s ability to reduce cellular proliferation have limited the power to detect a small difference (Figure 3). and matrix accumulation in this model is not due to a reduced However, other studies support that there was an equivalent degree of injury. degree of injury in both treatment groups. First, immunofluo- rescence studies revealed no qualitative difference in distribu- tion or intensity of the staining of anti–Thy-1 antibody or in the Glomerular Cell Proliferation and ODC Activity number of ED-1–positive cells between the Thy-1 and Thy- Proliferation in isolated glomeruli was measured by 3H- 1ϩAg groups (Table 1). Furthermore, there was no difference thymidine incorporation. Glomerular cell proliferation in the in the total number of nuclei and PMN on day 1 (Table 1), and Thy-1 group was significantly increased on day 4 (3403 Ϯ 183 there was no alteration in the degree of proteinuria (167 Ϯ 32 versus 2424 Ϯ 54, Thy-1 versus Control, P Ͻ 0.01), and the versus 158 Ϯ 28, Thy-1 versus Thy-1ϩAg), suggesting that Thy-1ϩAg group exhibited significantly reduced cellular pro- agmatine did not influence the initial degree of injury. By day liferation on days 1, 4, and 7 as compared with the Thy-1 group 4, the glomeruli in the Thy-1 group were characterized by focal (2130 Ϯ 52, 2630 Ϯ 175, 2197 Ϯ 57, P Ͻ 0.05 versus Thy-1, and segmental hypercellularity in the mesangium with occa- Figure 5A). The group that was treated only with agmatine sional disintegration of the mesangial architecture (Figure 2C). served as a time control on days 1 through 7. We also measured This was followed at day 7 by a marked increase in extracel- ODC activity, the rate-limiting polyamine biosynthetic en- lular matrix and MC proliferation (Figure 2D). In contrast, zyme, as another index of the effect of agmatine on MC there was a decrease in the proliferation and matrix scores at proliferation. ODC activity in the Thy-1 group achieved sig- day 4, which became statistically significant at day 7 in the nificant peak levels on day 1 (25559 Ϯ 2243 versus 1168 Ϯ Thy-1ϩAg group (Figures 2, F and G, and 3). Reductions both 252 cpm/mg protein, day 1 versus Control, P Ͻ 0.01) and was in the number of cells that stained for PCNA (Figure 4) and in maintained above control levels on days 4 and 7 (22000 Ϯ 3H-thymidine incorporation (Figure 5A) in the Thy-1ϩAg 2211, 16719 Ϯ 1923, P Ͻ 0.01). ODC activity in the Thy- group as compared with the Thy-1 group provide additional 1ϩAg group was also markedly elevated above the control evidence that agmatine treatment diminished cellular prolifer- during the experimental period (11430 Ϯ 527, 11536 Ϯ 726, ation at days 4 and 7. 2151 Ϯ 213, P Ͻ 0.01 versus Control) but was significantly lower than that of the Thy-1 group (P Ͻ 0.05 versus Thy-1, Administration of Agmatine after Induction of Figure 5B) on all days evaluated. Addition of L-NIL to the Thy-1 Nephritis supernatants from the Thy-1 group did not cause a significant Agmatine was also administered 24 h after induction of difference in ODC activity during the experimental period Thy-1 nephritis to provide evidence that the alterations in (data not shown). This suggests that reductions in NO were not proliferation and matrix accumulation induced by agmatine responsible for alterations in ODC activity.

Table 1. The number of glomerular cells, infiltrating cells, and quantitative assessment of the binding of anti–Thy-1 antibodya

Light Microscopy Immunofluorescence Group N No. of Nuclei No. of PMN Per No. of ED-1 Cells Per Goat Anti-Mouse Per Glomerulus Glomerulus Glomerulus IgG Staining

Thy-1 day 1 3 56.5** Ϯ 2.3 1.75 Ϯ 0.50 12.5 Ϯ 1.4 4ϩ day 4 3 61.2 Ϯ 2.3 0.55 Ϯ 0.14 day 7 3 77.1 Ϯ 8.8 0.73 Ϯ 0.15 Thy-1ϩAg day 1 3 56.4** Ϯ 1.6 1.48 Ϯ 0.28 10.8 Ϯ 1.2 4ϩ day 4 3 55.3 Ϯ 2.4 0.3 Ϯ 0.09 day 7 3 65.7 Ϯ 3.9 0.5 Ϯ 0.11 Controls 3 64.0 Ϯ 1.3 0.43 Ϯ 0.02

a PMN, polymorphonuclear leukocytes. The number of nuclei and PMNs were counted individually in 30 glomeruli in PAS stained slides. There was no significant difference in histologic variables on day 1 between the Thy-1 and Thy-1ϩAg groups. However, both groups on day 1 had a reduced nuclear number as compared with the control. The values are expressed as mean Ϯ SEM of three rats. **P Ͻ 0.05 compared with control group. J Am Soc Nephrol 11: 2256–2264, 2000 Effects of Agmatine in Thy-1 Nephritis 2261

day 1 of the Thy-1 GN model, it is likely that despite elevated NO, the decline in cGMP levels is more likely caused by a reduction in GTP availability or altered redox environment due to mesangial injury. By day 7 in both the Thy-1 and Thy-1ϩAg groups, there was a significant increase in cGMP levels as compared with the control group (P Ͻ 0.01), suggesting res- toration of guanylate cyclase activity or GTP levels during the MC proliferative phase.

Discussion Agmatine, decarboxylated arginine, has been shown to exert multiple effects on arginine metabolic pathways (10,14,25–27; Satriano et al., submitted). The majority of this work defining the actions of agmatine has been performed in vitro. Herein we describe that agmatine can exert beneficial renal functional and glomerular cellular effects in vivo, in the Thy-1 nephritis model. Figure 3. Semiquantitative analysis of glomerular tissue injury, cel- Thy-1 nephritis has an early injury phase characterized func- lular proliferation, and extracellular matrix scores. Each quadrant in tionally and histologically by a reduction in GFR and by 30 glomeruli per rat was graded on a scale from 0 to 4 on PAS-stained slides. There was no difference in the injury score between the Thy-1 mesangiolysis. This is followed by the repair phase in which and Thy-1ϩAg groups on day 1. However, the Thy-1ϩAg group there is extensive cellular proliferation and matrix accumula- exhibited significant declines in both proliferation and matrix scores tion. In this study, administration of agmatine significantly on day 7. The histogram represents the average score per quadrant of attenuated the decline in GFR without appreciably altering the 30 glomeruli assessed in each rat. Each point is mean Ϯ SEM of three initial degree of injury. During the repair phase, which is rats. *, P Ͻ 0.05 compared with the Thy-1 group. associated with markedly elevated ODC activity, agmatine substantially reduced ODC activity in the Thy-1–treated rats. Similarly, there was a decrease in cellular proliferation and Nitrite Production in Isolated Glomeruli matrix accumulation in the rats treated with agmatine, and Nitrite production in glomeruli isolated from rats in the these alterations did not seem to be attributable to reduced Thy-1 group achieved significant peak levels on day 1 (172.9 binding of anti–Thy-1 antibody to MC or to reduced glomer- Ϯ 22.2 versus 6.5 Ϯ 0.9 nmol/mg protein, Thy-1 versus ular infiltration by monocytes/macrophages. Control, P Ͻ 0.01), remained high on day 4 (73.1 Ϯ 9.7, P Ͻ The first phase of Thy-1 nephritis lasts up to 1 d after 0.01), and declined by day 7 (17 Ϯ 1.4, not significant). Nitrite administration of the Thy-1 antibody. Glomerular injury is production in the Thy-1ϩAg group was also elevated above control, but these levels were only 77% and 59% of those measured in the Thy-1 group on days 1 and 4, respectively (Figure 6). Addition of L-NIL to glomeruli isolated from the Thy-1–treated rats greatly reduced NO products to 17% on day 1 and 11% on day 4 (data not shown), suggesting that the increase in NO production was due to iNOS activity. When compared with the L-NIL effects, in vitro addition of agmatine demonstrated only weak inhibitory effects on NO production (79% at day 1 and 76% at day 4 of Thy-1 levels).

Intracellular Glomerular cGMP Concentration cGMP, the second messenger of NO, was measured to assess indirectly the alterations in NO production. Paradoxically, de- spite the high levels of NO measured in the Thy-1–treated groups at days 1 and 4, there was no physiologically significant increase in cGMP as compared with the controls (agmatine- only–treated group, Figure 7). To examine the response of guanylate cyclase to an NO donor, SNP was added to glomer- Figure 4. Quantification of the number of cells stained with prolifer- uli isolated from the Thy-1 group. In parallel with the in vivo ating cell nuclear antigen (PCNA) in 30 glomeruli per animal. There data, there was no significant increase in cGMP levels after was a significantly fewer number of cells stained with PCNA in the administration of SNP on days 1 and 4, but this responsiveness Thy-1ϩAg group at days 4 and 7 as compared with the Thy-1 group. was restored at day 7 (data not shown). Because other enzyme *, P Ͻ 0.05 compared with Thy-1 group. Each point is mean Ϯ SEM systems such as iNOS and ODC were intact and upregulated on of four rats. 2262 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 2256–2264, 2000

Figure 6. Evaluation of nitric oxide (NO) generation. Nitrite produc- tion in isolated glomeruli was measured after incubation in Dulbec- co’s modified Eagle’s medium for 24 h. Nitrite levels in the Thy- 1ϩAg grouped appeared lower than those in the Thy-1 group at day 1 but reached significance only by day 4. The agmatine only–treated group was used as a time control. Each point is mean Ϯ SEM of six rats. #, P Ͻ 0.01 compared with control. *, P Ͻ 0.05 compared with Thy-1 group.

suggests that the arginine/NO pathway is involved in the patho- genesis of tissue injury. In fact, inhibition of NO synthesis with NG-monomethyl-D-arginine (L-NMMA), a nonspecific NOS inhibitor, has been reported to reduce MC lysis significantly Figure 5. Evaluation of glomerular cell proliferation and ornithine (28). decarboxylase (ODC) activity. (A) Proliferation was assessed by Agmatine has been shown in vitro to inhibit iNOS, resulting 3 H-thymidine incorporation. Isolated glomeruli were incubated with in decreased NO production (12,13; Satriano et al., submitted). 3 ␮ H-thymidine (1 Ci/well) for 24 h. When compared with the Thy-1 However, in this study, there was only a 23% and 41% decline ϩ group, the Thy-1 Ag group had a significant decline in glomerular in NO production in the agmatine-treated Thy-1 rats as com- 3H-thymidine incorporation on all days evaluated. (B) ODC activity pared with untreated Thy-1 rats on days 1 and 4. The lack of a evaluated in renal cortices. ODC levels in the Thy-1ϩAg group were significantly lower than those measured in the Thy-1 group on all days significant alteration in the degree of injury may have been due evaluated. Each point is mean Ϯ SEM of six rats. #, P Ͻ 0.01 to this limited inhibition of NO production. However, it is compared with control. *, P Ͻ 0.05 compared with Thy-1 group. possible that the levels of NO measured in this study may have underestimated in vivo levels, because they were performed ex Table 2. Evaluation on day 7 of effect of administration of vivo on isolated glomeruli, incubated for 24 h in DMEM agmatine 24 hours after induction of Thy-1 without exogenous agmatine. nephritisa In the present study, agmatine treatment significantly re- versed the decline in GFR in the Thy-1–treated rats (Figure 1) ϩ Thy-1 Group Thy-1 Agmatine and the reduction in NO production by agmatine treatment on Nuclear number 82.5 Ϯ 7.2 73.6 Ϯ 4.5 day 1 was probably not sufficient to explain the preservation of Proliferation score 2.4 Ϯ 0.08 1.9 Ϯ 0.13* GFR observed (Figure 6). Similarly, the agmatine-mediated Matrix score 2.3 Ϯ 0.11 1.9 Ϯ 0.14* improvement in GFR cannot be attributed to a difference in ODC levels 20164 Ϯ 2565 11187 Ϯ 2340* initial injury (Figures 2 and 3) or to a reduction in the recruit- ment of macrophage/monocytes into glomeruli (Table 1) on a ODC, ornithine decarboxylase. Values expressed are means Ϯ day 1. It is interesting that the control rats treated with 50 Ͻ SEM of five rats, * P 0.05 compared with Thy-1 group. mg/kg of agmatine intraperitoneally did not exhibit an increase in GFR, which seems at odds with our earlier findings that agmatine increases SNGFR (10). The most likely reason for partially due to the effects of high-output NO generation by this finding is that the doses of agmatine achieved intrarenally iNOS, and the expression of this inducible enzyme has been in this study probably were much lower than the 1 ␮M urinary localized to infiltrating macrophages and polymorphonuclear space concentrations achieved in the previous microperfusion leukocytes rather than intrinsic glomerular cells (3,5,19). This experiments. J Am Soc Nephrol 11: 2256–2264, 2000 Effects of Agmatine in Thy-1 Nephritis 2263

antizyme (14). The resulting depletion of intracellular poly- amine levels by agmatine administration suppresses prolifera- tion (14). The data from the present studies demonstrate that ODC activity is markedly increased as early as day 1 after anti– Thy-1 antibody administration, and these high levels of activity persisted during the later cell proliferation phases (Figure 5B). Agmatine treatment in Thy-1 rats significantly attenuated the increase in ODC activity and resulted in reduced glomerular cell proliferation (Figures 3 and 5B). The histologic observa- tion of a decrease in cellular proliferation was supported by a decrease in both the number of PCNA-positive staining cells and in glomerular 3H-thymidine incorporation at days 4 and 7 (Figures 4 and 5A). On the basis of these data, it is likely that agmatine reduces MC proliferation via suppression of ODC. Agmatine therapy clearly reduced cell proliferation by day 7 Figure 7. Glomerular intracellular guanosine 3Ј,5Ј-cyclic monophos- in Thy-1 nephritis. We have argued that this reduction in phate (cGMP) concentrations. In isolated glomeruli on day 1, there cellular proliferation is a direct effect of agmatine therapy. was a small increase in cGMP generation in the Thy-1ϩAg group as However, the amelioration of cell proliferation also could have compared with the Thy-1 group. Because the NO levels were highest derived from lesser cell injury on day 1 as a consequence of at this time, it is unlikely that this increase was physiologically agmatine administration. Therefore, we made considerable ef- significant. By day 7, cGMP levels were elevated in both the Thy-1 forts to assess this issue and did not discern any difference in and Thy-1ϩAg groups as compared with the control and agmatine initial glomerular injury between untreated and agmatine- groups. Each point is mean Ϯ SEM of six rats. #, P Ͻ 0.01 compared with control. treated Thy-1 nephritis rats, using histologic scoring, evalua- tion of the number of nuclei, infiltrating PMN and macro- phages, and extent of binding of anti–Thy-1 antibody. ODC levels were reduced by agmatine therapy on day 1, but we We previously demonstrated that microperfusion of agma- interpret these findings as early indicators of cell proliferation tine into the renal interstitium or into the urinary spaces of rather than as indices of cellular injury. Nevertheless, we surface nephrons produces an increase in nephron filtration acknowledge that small differences in initial injury may be rate, which is largely mediated by an increase in plasma flow difficult to detect despite our objective efforts. Because of this caused by vasodilatation (10). These effects of agmatine were possibility, we designed other studies that obviate the possi- prevented by coadministration of a nonselective NOS blocker, indicating that the vasodilatation partially depends on consti- bility of an agmatine effect on initial MC injury. In a second tutive NOS (11). Other groups have shown that agmatine may group of rats, we began agmatine therapy 24 h after adminis- function as an endothelial NOS (eNOS) agonist (29). In the tration of the Thy-1 antibody, thereby guaranteeing that the Thy-1 nephritis model, we recently demonstrated that pretreat- initial injury was equivalent in both untreated and agmatine- ment with L-NIL, a specific iNOS inhibitor, did not prevent a treated rats. When these results are evaluated (Table 2), we reduction in GFR at day 1, whereas the nonselective inhibitor observed reductions in parameters of cell proliferation includ- L-NAME further decreased GFR, suggesting that activation of ing histologic evaluation and ODC activity in the rats that cNOS may preserve GFR (Ishizuka et al., unpublished obser- received agmatine beginning 24 h after the initial immune vations). Likewise, Goto et al. (19) reported an increase in insult. We conclude from these separate studies that agmatine eNOS activity in the early phase of this model. Recently, we exerts a significant antiproliferative effect that is independent demonstrated that production of high levels of NO in the of any influence on initial glomerular injury. lipopolysaccharide (LPS) model can downregulate the eNOS Herein we provide the first in vivo results of the effects of enzymes (isoform switching) (30). Pretreatment with agmatine agmatine on cellular injury and repair. Agmatine influences can prevent the reduction in BP and GFR in the LPS model MC proliferation and the accumulation of extracellular matrix (Satriano et al., submitted). These data suggest that agmatine in the Thy-1 nephritis rats, when given before and after disease preserves eNOS effects in pathologies associated with high NO induction. Although this model was not performed to assess production, such as GN. effects of agmatine on different NOS isoforms, further studies In addition to its effects on NOS isoforms, agmatine has should be undertaken to examine how the agmatine pathway been shown to alter polyamine synthetic pathways, important affects the balance of NOS isoform activities in GN and in cellular proliferation. Increased ODC expression has been regulates NO and polyamine synthesis to mediate the antipro- described in animal models of diabetes and LPS-induced in- liferative effects. flammation (31–33). Agmatine administration results in the The present study demonstrated that agmatine administra- inhibition of ODC activity and the concurrent inhibition of tion can improve GFR and, thus, renal function in this in vivo polyamine transport by promoting translation of the ODC model of inflammatory renal disease. Agmatine can also mark- 2264 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 2256–2264, 2000 edly reduce proliferative changes by inhibiting ODC activity, and exacerbates rat nephrotoxic nephritis. Kidney Int 49: 1090– an effect previously demonstrated in vitro. 1096, 1996 16. Yamamoto T, Wilson CB: Quantitative and qualitative studies of antibody-induced mesangial cell damage in the rat. Kidney Int Acknowledgments 32: 514–525, 1987 These studies were supported by grants from the National Institutes 17. Yamamoto T, Wilson CB: Complement dependence of antibody- of Health (DK 28602 and T32 HL 07261) and funds provided by the induced mesangial cell injury in the rat. J Immunol 138: 3758– Office of Research and Development, Medical Research Service of 3765, 1987 the Department of Veterans Affairs. 18. Peters H, Border WA, Noble NA: L-arginine supplementation increases mesangial cell injury and subsequent tissue fibrosis in References experimental glomerulonephritis. Kidney Int 55: 2264–2273, 1999 1. Albina JE, Mills CD, Henry WL Jr, Caldwell MD: Temporal 19. Goto S, Yamamoto T, Feng L, Yaoita E, Hirose S, Fujinaka H, expression of different pathways of L-arginine metabolism in Kawasaki K, Hattori R, Yui Y, Wilson CB, et al.: Expression and healing wounds. J Immunol 144: 3877–3880, 1990 localization of inducible nitric oxide synthase in anti-Thy-1 2. Cook HT, Jansen A, Lewis S, Largen P, O’Donnell M, Reaveley glomerulonephritis. Am J Pathol 147: 1133–1141, 1995 D, Cattell V: Arginine metabolism in experimental glomerulo- 20. Yamamoto T, Mundy CA, Wilson CB, Blantz RC: Effect of nephritis: Interaction between nitric oxide synthase and arginase. mesangial cell lysis and proliferation on glomerular hemodynam- Am J Physiol 267: F646–653, 1994 ics in the rat. Kidney Int 40: 705–713, 1991 3. Ketteler M, Ikegaya N, Brees DK, Border WA, Noble NA: 21. Blantz RC, Tuker BJ: Measurement of glomerular dynamics. In L-arginine metabolism in immune-mediated glomerulonephritis Methods in Pharmacology, Renal Pharmacology, edited by Mar- in the rat. Am J Kidney Dis 28: 878–887, 1996 4. Cattell V, Largen P, de Heer E, Cook T: Glomeruli synthesize tinez-Maldonado M, New York, Plenum, 1978, pp 141–163 nitrite in active Heymann nephritis: The source is infiltrating 22. Tucker BJ, Mendonca MM, Blantz RC: Contrasting effects of macrophages. Kidney Int 40: 847–851, 1991 acute insulin infusion on renal function in awake nondiabetic and 5. Cattell V, Lianos E, Largen P, Cook T: Glomerular NO synthase diabetic rats. J Am Soc Nephrol 3: 1686–1693, 1993 activity in mesangial cell immune injury. Exp Nephrol 1: 36–40, 23. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ: Protein mea- 1993 surement with the Folin phenol. J Biol Chem 193: 265–275, 1951 6. Auvinen M, Paasinen A, Anderson LC, Holtta E: Ornithine 24. Green LC, Wagner DA, Glogowski J, Skipper PL, Wishnok JS, decarboxylase activity is critical for cell transformation. Nature Tannenbaum SR: Analysis of nitrite, nitrite, and [15N]nitrate in 360: 355–358, 1992 biological fluids. Anal Biochem 126: 131–138, 1982 7. Li G, Regunathan S, Barrow CJ, Eshraghi J, Cooper R, Reis DJ: 25. Morrissey JJ, Klahr S: Agmatine activation of nitric oxide syn- Agmatine: An endogenous clonidine-displacing substance in the thase in endothelial cells. Proc Assoc Am Physicians 109: 51–57, brain. Science 263: 966–969, 1994 1997 8. Raasch W, Regunathan S, Li G, Reis DJ: Agmatine, the bacterial 26. Satriano J, Ishizuka S, Archer DC, Blantz RC, Kelly CJ: Regu- amine, is widely distributed in mammalian tissues. Life Sci 56: lation of intracellular polyamine biosynthesis and transport by ␣ ␥ 2319–2330, 1995 nitric oxide and cytokines TNF- and IFN- . Am J Physiol 276: 9. Morrissey J, McCracken R, Ishidoya S, Klahr S: Partial cloning C892–C899, 1999 and characterization of an arginine decarboxylase in the kidney. 27. Satriano J, Kelly CJ, Blantz RC: An emerging role for agmatine. Kidney Int 47: 1458–1461, 1995 Kidney Int 56: 1252–1253, 1999 10. Lortie MJ, Novotny WF, Peterson OW, Vallon V, Malvey K, 28. Narita I, Border WA, Ketteler M, Ruoshlati E, Noble NA: Mendonca M, Satriano J, Insel P, Thomson SC, Blantz RC: L-arginine may mediate the therapeutic effects of low protein Agmatine, a bioactive metabolite of arginine. Production, deg- diets. Proc Natl Acad Sci USA 92: 4552–4556, 1995 radation, and functional effects in the kidney of the rat. J Clin 29. Morrissey JJ, Klahr S: Effects of agmatine, an active metabolite Invest 97: 413–420, 1996 of arginine metabolism, on the kidney [Editorial]. Nephrol Dial 11. Schwartz D, Peterson OW, Mendonca M, Satriano J, Lortie M, Transplant 11: 1217–1219, 1996 Blantz RC: Agmatine affects glomerular filtration via a nitric 30. Schwartz D, Mendonca M, Schwartz I, Xia Y, Satriano J, Wilson oxide synthase-dependent mechanism. Am J Physiol 272: F597– CB, Blantz RC: Inhibition of constitutive nitric oxide synthase 601, 1997 (NOS) by nitric oxide generated by inducible NOS after lipo- 12. Auguet M, Viossat I, Marin JG, Chabrier PE: Selective inhibition polysaccharide administration provokes renal dysfunction in rats. of inducible nitric oxide synthase by agmatine. Jpn J Pharmacol J Clin Invest 100: 439–448, 1997 69: 285–287, 1995 31. Matsuzaki Y, Sugimoto H, Hamana K, Nagamine T, Matsuzaki 13. Galea E, Regunathan S, Eliopoulos V, Feinstein DL, Reis DJ: S, Mori M: Effects of eicosanoids on lipopolysaccharide-induced Inhibition of mammalian nitric oxide synthases by agmatine, an ornithine decarboxylase activity and polyamine metabolism in endogenous polyamine formed by decarboxylation of arginine. the mouse liver. J Hepatol 27: 193–200, 1997 Biochem J 316: 247–249, 1996 32. Pedersen SB, Bjorn SF, Richelsen B, Flyvbjerg A: Inhibition of 14. Satriano J, Matsufuji S, Murakami Y, Lortie MJ, Schwartz D, renal ornithine decarboxylase activity fails to reduce kidney size Kelly CJ, Hayashi S, Blantz RC: Agmatine suppresses prolifer- and urinary albumin excretion in diabetic rats with manifest ation by frameshift induction of antizyme and attenuation of kidney hypertrophy. Mol Cell Endocrinol 107: 123–128, 1995 cellular polyamine levels. J Biol Chem 273: 15313–15316, 1998 33. Lortie MJ, Ishizuka S, Schwartz D, Blantz RC: Bioactive prod- 15. Waddington S, Cook HT, Reaveley D, Jansen A, Cattell V: ucts of arginine in sepsis: Tissue and plasma composition fol- L-arginine depletion inhibits glomerular nitric oxide synthesis lowing LPS and iNOS blockade. Am J Physiol 2000, in press