J Am Soc Nephrol 10: 62–68, 1999 Pentosan Polysulfate Decreases Proliferation and Net Extracellular Matrix Production in Mouse Mesangial Cells

SHARON J. ELLIOT,*§ LILIANE J. STRIKER,*§ WILLIAM G. STETLER-STEVENSON,† TERRY A. JACOT,* and GARY E. STRIKER*§ *Renal Cell Biology Section, Metabolic Disease Branch, National Institute of Diabetes and Digestive and Kidney Diseases, and †Extracellular Matrix Pathology Section, Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland; and §Department of Medicine, University of Miami, Miami, Florida.

Abstract. Glomerulosclerosis is characterized by extracellular types I and type IV were significantly decreased in the media matrix accumulation and is often associated with mesangial (P Ͻ 0.0001) and cell layer (P Ͻ 0.005) after treatment with cell proliferation. Heparin-like molecules have been shown to PPS but not after treatment with heparin. By zymography, decrease glomerulosclerosis in vivo, although their cellular site MMP-2 was significantly increased after treatment with PPS and mechanism of action is still unclear. In this study, a line of (P Ͻ 0.001) and heparin (P Ͻ 0.05). PPS and heparin also glomerular mesangial cells derived from normal mice was used decreased MMP-9 (P Ͻ 0.001) after treatment. Reverse zy- to determine whether pentosan polysulfate (PPS) inhibited mography showed the presence of tissue inhibitors of metal- proliferation and altered extracellular matrix turnover. Cells loproteinases (TIMP)-1 and -2 in control mesangial cells. treated with PPS showed a decrease in cell number beginning Treatment with PPS and heparin increased TIMP-1. In addi- 24 h after treatment, which was maintained for 5 d. For matrix tion, TIMP-3 was found in the medium of treated but not accumulation and degradation studies, cells were treated for control cells. In conclusion, PPS alters extracellular matrix 5 d and collagen types I and IV protein were measured by turnover through the induction of MMP-2 and alterations in the enzyme-linked immunosorbent assay as well as matrix metal- TIMP profile and may be useful in decreasing progressive loproteinases (MMP) measured by zymography. Collagen glomerulosclerosis.

Human and experimental glomerulosclerosis is characterized ulosclerosis because of its anticoagulant properties, and low by a gradual increase in glomerular mesangial extracellular molecular weight heparin preparations, which have reduced matrix (ECM) (1–4). The turnover of mesangial matrix is anticoagulant activity, must be parenterally administered. Re- normally slow, and several investigators have postulated that cently, an oral pentosan polysulfate (PPS) preparation, essen- an imbalance between accumulation and degradation of ECM tially devoid of anticoagulant activity, has become available could contribute to the development of glomerulosclerosis. (Elmiron®). This product has been shown to decrease glomer- Heparin and heparin-like molecules have been shown to ulosclerosis in a mouse model of glomerulosclerosis (12). To decrease the proliferation of several types of cells in vitro, determine the cellular site of action, we chose to study glo- including mesangial cells (5), smooth muscle cells (6,7), and merular mesangial cells because they are a major contributor to glomerular epithelial cells (8). In addition, heparin and/or low ECM turnover (16–18). molecular weight heparin fragments have been shown to slow or prevent the progression of glomerulosclerosis in vivo (5,9– 12). The mechanism of action on net ECM production in vivo Materials and Methods is unclear. However, heparin decreases collagen synthesis by Cell Culture and Experimental Conditions smooth muscle cells in vitro (13). Finally, it has recently been Glomerular mesangial cells were isolated from normal 4-wk-old shown that heparin influences the induction of metalloprotein- mice (C57B1/6J X SJL/J) and maintained as described previously ase (MMP) production in vitro, with an inhibitory effect in (19). Except for growth curves, the experiments were performed in some cell types (14) or a stimulatory effect in others (15). six-well plates (Nunc Nalge International, Naperville, IL). The cells Heparin is not commonly used in the treatment of glomer- were plated so that the cell number in each well was approximately the same at the end of the experiment and that the cells were semi- confluent (approximately 150,000 cells/well). Briefly, the cells were plated in medium with 20% fetal bovine serum (FBS; Life Technol- Received November 17, 1997. Accepted July 24, 1998. Correspondence to Dr. Liliane J. Striker, Renal Cell Biology Lab, University of ogies, Gaithersburg, MD) and allowed to attach. The serum concen- Miami School of Medicine, P.O. Box 016960 (R126), Miami, FL 33101. Phone: tration was decreased to 0.1% overnight before each experiment, and 305-243-2811; Fax: 305-243-2810; E-mail: [email protected] the following day the cells were exposed to fresh medium containing ␮ ® 1046-6673/1001-0062$03.00/0 0.1% FBS alone or with the addition of 100 g/ml PPS (Elmiron , Journal of the American Society of Nephrology IVAX Corp., Miami, FL) or heparin (sodium salt, Sigma, St. Louis, Copyright © 1999 by the American Society of Nephrology MO). PPS or heparin was added each day for 5 d. At the end of the J Am Soc Nephrol 10: 62–68, 1999 PPS Decreases Proliferation and ECM Production 63

5 d, media was collected for the assessment of metalloproteinases or tion was determined on quadruplicate wells processed individually collagen production, and the cell layer was used to determine cell from two separate experiments. Cell number was determined on number, prepare RNA, or measure collagen accumulation. duplicate wells from treated and control cells. For growth curves, cells were plated at equal density in medium containing 20% serum in 24-well plates. Twenty-four hours later, Zymography for Metalloproteinases 0.1% FBS-containing medium with or without increasing concentra- The amount of MMP-9 and MMP-2 gelatinases in the medium was ␮ tions of PPS (5 to 100 g/ml) were added to assess dose–response. assessed by using 10% zymogram gels (Novex, San Diego, CA). ␮ Subsequent growth curves were performed with PPS (100 g/ml) or Briefly, medium was diluted to normalize for cell counts (approxi- ␮ heparin (100 g/ml) or 0.1% FBS-containing medium. Cell number mately 50,000 cells/ml) before 5ϫ Laemmli buffer (without ␤-mer- was assessed on days 1, 3, and 5. captoethanol) was added to each sample. After electrophoresis, the gels were washed for1hin2.5% sodium dodecyl sulfate and incubated overnight in 50 mM Tris buffer as described previously Media and Cell Layer Collection (18). The gels were stained with Coomassie blue and air-dried. Gels Metalloproteinase Production and RNA Preparation. On the were also incubated in 50 mM Tris buffer, with the addition of 25 mM day of collection, medium from those wells designated for zymogra- ethylenediaminetetraacetic acid to check for nonmetalloproteinase- phy was centrifuged to remove cell debris and frozen at Ϫ70°C for dependent bands of proteolytic activity. Densitometry was performed later use. The cell number from these same wells was determined by using NIH image 1.6 to quantify the relative gelatinase activities. direct cell counting, and the remaining cells were used to prepare total ® RNA by using Tri Reagent (Molecular Research Center, Cincinnati, Reverse Zymography OH) according to the manufacturer’s directions. The protocol was TIMP were assessed by reverse zymography in the supernatant and modified by the addition of muscle glycogen (5 PRIME-3 PRIME, the cell layer, as described (21). Briefly, gels containing gelatinase A Inc., Boulder, CO) as a nonspecific carrier to increase the yield of total were prepared. Medium was diluted to normalize for cell number as RNA. The final RNA pellet was dissolved in 10 ␮l of diethyl pyro- described for zymography. Gels were washed for1hin2.5% sodium carbonate water and stored at Ϫ70°C. dodecyl sulfate after electrophoresis and incubated overnight at 37°C. Collagen Analysis. The media and cell layers of separate but Coomasie blue staining and air-drying were described for zymogra- parallel wells were collected for determination of ␣1 (I) and ␣1 (IV) phy. To quantify the TIMP activity, densitometry was performed collagens at the same time as zymography and RNA analysis. Ascor- using NIH Image 1.6. bic acid (50 ␮g/ml) and ␤-aminopropionitrile (80 ␮g/ml) were added daily to the cells in culture. As previously described (18), the medium was collected and centrifuged to remove cell debris, and the super- Reverse Transcription and mRNA Determination natant was stored with protease inhibitors (ethylenediaminetetraacetic One microgram of total RNA was reversed-transcribed, as de- acid, phenylmethylsulfonyl fluoride, N-ethylmaleimide) at Ϫ70°C. scribed previously, using a First Strand cDNA synthesis kit (Boehr- The cell layer was collected in 1 ml of 5 M guanidine, 0.1 M Tris-HCl, inger Mannheim, Indianapolis, IN) (18). The reversed-transcribed ␮ pH 8.6, and protease inhibitors. Cell numbers were determined by material was diluted to a working concentration of 33 ng/ l. A typical direct cell counts in parallel wells. PCR reaction contained 66 ng of reversed-transcribed material in a ␮ Collagen Assay by Enzyme-Linked Immunosorbent Assay. final volume of 50 l. Standard PCR was performed for MMP-9, ␣ Medium was incubated for 24 h in a 96-well Nunc-Immuno Maxisorp MMP-2, and 1 (I) collagen using primers previously published ␣ plate (Nalge Nunc International, Naperville, IL) followed by washing (18,22,23). Competitive PCR was performed for 1 (IV) collagen and and blocking with 3% bovine serum albumin/phosphate-buffered sa- glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as described line for an additional 24 h. A total of 100 ␮l of an antibody against previously (23). The gels were photographed using an IS-1000 Digital collagen type IV (1:750) or collagen type I (1:1000) (Biodesign, Imaging System and saved as a Tiff file. Densitometry was performed Kennebunk, ME) was applied for 1.5 or 2 h, respectively. After five using the computer program NIH image 1.6. The standard curves for washes with 0.05% Tween/phosphate-buffered saline, biotinylated competitive PCR were generated as described (22,23). For standard goat anti-rabbit IgG (Sigma) was applied for 1.5 h. After five washes, PCR, the densitometry units were divided by the number of attomoles streptavidin alkaline phosphatase followed by p-nitrophenyl phos- of GAPDH to normalize for the differences between samples. phate was used for color development. The plate was read in a Titertek Mulitskan plate reader at 405 nm. Conditions were identical to those Statistical Analyses described previously(18), except that the standard curves were mod- Data from three independent experiments were analyzed. Differ- ified. The concentrations of the type I (Collaborative Biomedical ences between PPS treatment, heparin, and control cells were mea- Products, Bedford, MA) standards were 5 to 40 ng/well and the type sured using the t test or one-way ANOVA (Prism, GraphPad, San IV standards 4 to 32 ng/well. The curves were linear (r ϭ 0.98). Diego, CA). Collagen Production. Mesangial cells were plated as described above in 6-well plates and maintained with or without PPS for 5 d. On Results the fourth day of treatment, the medium was removed and proline-free Dose Response and Growth Curve Dulbecco’s modified Eagle’s medium with 0.1% FBS was added to Cells were treated for 4 d with increasing concentrations of each well and incubated for 4 h. After this washout period, the PPS. Proliferation decreased in a dose-dependent manner (Fig- medium was changed to proline-free Dulbecco’s modified Eagle’s medium containing 0.1% FBS, 10 ng/ml ascorbic acid, and 5 ␮Ci/well ure 1A). Cell number was equally suppressed after the addition 3 ␮ ␮ of L-[2,3- H]proline (Amersham, Arlington Heights, IL), and PPS in of PPS (100 g/ml) or heparin (100 g/ml) (Figure 1B). Cell those cells being treated. The medium and cell layer were collected number decreased 24 h after the addition of PPS or heparin. By 24 h later and processed according to Diegelmann and Peterkofsky day 3, cell number was 50% of control and this difference was (20), using purified bacterial collagenase (Sigma). Collagen produc- maintained up to5dinthepresence of either agent. Viability 64 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 62–68, 1999

Figure 2. ␣1 (I) collagen synthesis of mesangial cells. Cells treated with PPS for 5 d were collected as described. Medium (top: PPS, f, n ϭ 11; heparin, u, n ϭ 7; control, Ⅺ, n ϭ 10) and cell layers (bottom: PPS, n ϭ 7; heparin, n ϭ 4; control, n ϭ 7) from triplicate experiments were measured by enzyme-linked immunosorbent assay. There was a significant decrease in both the medium (P Ͻ 0.0001) and the cell layer (P Ͻ 0.005) in cells treated with PPS. Data are repre- Figure 1. (A) Dose–response to pentosan polysulfate (PPS). Glomer- sented as percentage of control. ular mesangial cells were plated in 20% serum-containing medium to allow for attachment. Twenty-four hours later, medium was changed to medium containing 0.1% fetal bovine serum alone (f) or with increasing concentrations (5 ␮g/ml [Œ], 10 ␮g/ml [ƒ], 50 ␮g/ml [Ⅺ], cells treated with PPS (Figure 3). Heparin did not have a and 100 ␮g/ml [F]) of PPS. Duplicate wells were counted on days 1, significant effect on collagen type I or IV in the media or cell 3, and 4. (B) Comparison of PPS and heparin. Twenty-four hours after layer. attachment, glomerular mesangial cells were exposed to 0.1% serum f ␮ Œ containing medium alone ( ) or PPS (100 g/ml; ) or heparin (100 Effects of PPS on Collagen Synthesis and Production ␮g/ml; E). Duplicate wells were counted on days 1, 3, and 5. The percentage of collagen-sensitive material in semicon- fluent cells (150,000/well) was equal in cells treated with PPS (28.6 Ϯ 4.3%; n ϭ 3) and control cells (26.1 Ϯ 2.42%; n ϭ 4). was not affected, as assessed by trypan blue staining. The Experiments performed on sparse cells revealed similar results antiproliferative effect of5dofPPStreatment was reversible (cells treated with PPS, 21.98 Ϯ 2.32, n ϭ 4; control cells, by reexposing the cells to medium containing 20% FBS (data 21.53 Ϯ 2.59, n ϭ 4). not shown). Zymography Collagen Determination There was an increase in secreted, soluble MMP-2 (72-kD The amount of collagen types I and IV was determined in gelatinase) (Figure 4) after treatment with PPS (E, P Ͻ 0.001) media and cell layers by enzyme-linked immunosorbent assay. and heparin (H, P Ͻ 0.05) compared with control cells (C). There was a significant decrease of ␣1 (I) collagen in the There was also an increase in the active form of this gelatinase medium (P Ͻ 0.0001) and monolayer (P Ͻ 0.005) of cells most notably with PPS (Figure 4A, lower band). In contrast, treated with PPS (Figure 2). ␣1 (IV) collagen was also signif- MMP-9, the 92-kD gelatinase (Figure 4C) was markedly de- icantly decreased in the medium and cell layer (P Ͻ 0.0001) in creased in the presence of PPS and heparin (P Ͻ 0.001). J Am Soc Nephrol 10: 62–68, 1999 PPS Decreases Proliferation and ECM Production 65

Figure 3. ␣1 (IV) collagen synthesis. PPS treatment (f) induced a significant reduction of ␣1 (IV) collagen synthesis in the medium and cell layer (P Ͻ 0.0001) as measured by enzyme-linked immunosor- bent assay, compared with control cells (Ⅺ). Data are expressed as percentage of control of triplicate experiments. Medium (top: PPS, n ϭ 12; heparin, n ϭ 7; control, n ϭ 11) and cell layer (bottom: PPS, n ϭ 8; heparin, n ϭ 4; control, n ϭ 7).

Figure 4. Matrix metalloproteinase-2 (MMP-2; 72-kD gelatinase) and Reverse Zymography MMP-9 (92-kD gelatinase). (A) Representative zymogram of mesan- Analysis of TIMP by reverse zymography showed the pres- gial cells treated with PPS (E), heparin (H), or control cells (C). S, ence of TIMP-1 and -2 in the medium of mouse mesangial cells standard. (B) Analysis of triplicate experiments showing an increase (Figure 5). Treatment of cells with PPS significantly upregu- of MMP-2 (PPS, f, n ϭ 10, P Ͻ 0.001 versus control; heparin, u, lated TIMP-1 (P Ͻ 0 0.05). Heparin also increased TIMP-1 n ϭ 5, P Ͻ 0.05 versus control; control, Ⅺ, n ϭ 7). (C) There is a (P Ͻ 0 0.01). Treatment had no statistically significant effect decrease of MMP-9 in triplicate experiments (PPS, f, n ϭ 8; heparin, on TIMP-2 release. TIMP-3, however, was only detectable in u, n ϭ 5, P Ͻ 0.001 versus control; control, Ⅺ, n ϭ 6) in mesangial the medium of cells treated with PPS or heparin. cells. mRNA Determination ␣1 (I) collagen was not significantly affected by treatment Discussion with PPS. By competitive PCR, the ratio ␣1 (IV)-collagen: The administration of the heparinoid PPS to a line of mouse GAPDH mRNA did not differ between treated cells and con- glomerular mesangial cells significantly modified ECM syn- trol cells (Figure 6, P ϭ NS). Figure 7 shows a graph of PCR thesis and degradation. It also decreased proliferation. We results for MMP-9. Cells treated with PPS had a significantly found a significant decrease in type I and IV collagen proteins lower (P Ͻ 0.05) MMP-9 mRNA level than did control cells. in the medium and the cell layer after5doftreatment with Heparin also significantly lowered mRNA levels of MMP-9 PPS. Heparin has also been shown to decrease synthesis of (P ϭ 0.001, data not shown). There was no significant differ- type I and type IV collagen in chrondrocytes and vascular ence in MMP-2 mRNA levels between PPS and control mes- smooth muscle cells (13) in vitro (24). Surprisingly, the hep- angial cells. arin preparation we used did not significantly decrease colla- 66 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 62–68, 1999

Figure 6. Competitive PCR of ␣1 (IV) collagen mRNA levels. Cells treated with PPS (f, n ϭ 5) versus control cells (Ⅺ, n ϭ 5) from triplicate experiments were reverse-transcribed, as described in Ma- terials and Methods. Competitive PCR was performed for ␣1 (IV) collagen and glyceraldehyde-3-phophate dehydrogenase (GAPDH). Results were expressed as attomoles of ␣1 (IV) collagen divided by GAPDH. Data are represented as the ratio of ␣1 (IV) collagen and GAPDH (P ϭ not significant).

Figure 7. PCR data comparing PPS-treated (f) and untreated (Ⅺ) cells. MMP-9 mRNA levels were significantly lower (P Ͻ 0.05) in cells treated with drug for 5 d. The data are expressed as percentage of control for each experiment. Data were analyzed from triplicate experiments (PPS, n ϭ 7; control, n ϭ 8).

Although PPS significantly lowered type I and type IV collagen net protein levels in both the cell monolayer and conditioned media, there was no significant change in their mRNA levels. This finding suggests that either the secretion of Figure 5. Reverse zymography. (A) Representative reverse zymo- collagen or its extracellular stability was affected. Heparin has gram. (B) Densitometric analysis of duplicate experiments for tissue not been shown to decrease collagen type I and IV mRNA and inhibitors of metalloproteinase (TIMP)-1 and TIMP-2. Cells treated with the following: PPS (E), f, n ϭ 7; heparin (H), u, n ϭ 6; control protein levels in human mesangial cells (4), although it did cells (C), Ⅺ, n ϭ 6. S, standard. *P Ͻ 0.05; **P Ͻ 0.01. decrease their levels in smooth muscle cells (13). Thus, the effects of heparin may differ depending on the vascular bed from which the cells are isolated. Therefore, it may be neces- gen synthesis in our experiments, although it was as effective sary to compare human smooth muscle cells isolated from as PPS in inhibiting cell proliferation. Thus, as we have shown various vascular beds of the same individual. In addition, it before, there is not a tight correlation between cell proliferation may be useful to compare different individuals, since we have and ECM synthesis (23). It has been shown that antiprolifera- shown the mesangial sclerosis is largely dependent on genetic tive activities vary widely between commercial sources of background in mice. heparin (25), and the heparin preparation we used has been Because the data suggested that PPS may alter ECM turn- reported to have an intermediate capability to inhibit cell over, we investigated the effect of PPS and heparin on MMP growth (26). and TIMP activity as detected by zymography and reverse J Am Soc Nephrol 10: 62–68, 1999 PPS Decreases Proliferation and ECM Production 67 zymography. We and others have shown that human (27), ated with decreased MMP-2 levels in rat mesangial cells mouse (18,28), and rat mesangial cells (29) produce MMP-2 (35,36) as well as in rabbit aortic smooth muscle cells (37). and MMP-9 in vitro. Both of these MMP degrade nonhelical Turck et al. (35) used antisense techniques to block MMP-2 in type IV collagen and denatured interstitial collagen (30,31). By cloned lines of rat mesangial cells, which produced only zymography, PPS significantly decreased MMP-9, and in- MMP-2, and found that they became quiescent after MMP-2 creased both the pro and active forms of MMP-2. The current inhibition. Other investigators using synthetic inhibitors of hypothesis regarding the activation of MMP-2 (30) is that MMP-2 were also able to inhibit proliferation in rabbit aortic proMMP-2 is activated on the cell surface by the membrane smooth muscle cells and rat mesangial cells (36,37). type matrix metalloproteinase (MT-MMP-1) in conjunction MMP play an important role in vivo in facilitating turnover with TIMP-2 (30,31). Because the mesangial cells treated with of ECM components. Because PPS seems to affect ECM by PPS had a decrease in net release of types I and IV collagens selective activation of MMP-2, this may be one mechanism by without change in the corresponding mRNA levels, the effect which it could act therapeutically in slowing progressive glo- was likely attributable to an increase in active MMP-2 in the merulosclerosis. extracellular space. Recently, Butler et al. (32) have shown that the complex of MT-MMP-1/TIMP-2 is required for activation of proMMP-2. An excess of TIMP-2, however, could inhibit References this activation. By reverse zymography, we did not detect an 1. Striker LJ, Doi T, Elliot S, Striker GE: The contribution of increase in TIMP-2. Our conclusion of increased local activa- glomerular mesangial cells to progressive glomerulosclerosis. Semin Nephrol 9: 318–328, 1989 tion of MMP-2 is consistent with these data and is further 2. Adler S, Striker LJ, Striker GE, Perkinson DT, Hibbert J, Couser supported by the observation that there were no differences in WG: Studies of progressive glomerular sclerosis in the rat. Am J the amount of collagenase-sensitive material between treated Pathol 123: 553–562, 1986 and control cells. These data suggest that the effect of PPS on 3. Morel-Maroger Striker L, Killen PD, Chi E, Striker GE: The ECM turnover occurs, at least in part, at the extracellular level. composition of glomerulosclerosis. 1. Studies in focal sclerosis, The synthesis of MMP-9 and MMP-2 have been shown to be crescentic glomerunephritis, and membranoproliferative glomer- independently regulated (30–33). This was most likely the case ulonephritis. Lab Invest 51: 181–192, 1984 in the current experiments. The decrease in MMP-9 protein and 4. Caenazzo C, Garbisa S, Onisto M, Zampieri M, Baggio B, mRNA levels after PPS and heparin treatment suggests that Gambaro G: Effect of glucose and heparin on mesangial alpha1 this effect was exerted at the transcriptional level. Our finding (IV) COLL and MMP-2/TIMP-2 mRNA expression. Nephrol of a significant increase of TIMP-1 after treatment with PPS or Dial Transplant 12: 443–448, 1997 5. Diamond JR, Karnovsky MJ: Nonanticoagulant protective effect heparin may be an additional independent action of these drugs of heparin in chronic aminonucleoside nephrosis. Renal Physiol or may result from the decrease in MMP-9 protein. In contrast, 9: 366–374, 1986 it appears that regulation of MMP-2 protein occurs mainly at 6. Clowes AW, Karnovsky MJ: Suppression by heparin of smooth the level of proenzyme activation rather than the transcriptional muscle cell proliferation in injured arteries. Nature 265: 625– level. It is possible that PPS could directly influence 626, 1977 proMMP-2 binding to the cell surface if one compares its effect 7. Hoover RL, Rosenberg R, Haering W, Karnovsky MJ: Inhibition with those of heparin. Heparin has been shown to influence the of rat arterial smooth muscle cell proliferation by heparin. II. In C-terminal interactions between TIMP-2 and proMMP-2, vitro studies. Circ Res 47: 578–583, 1980 which may in turn influence cell surface binding (32,34). 8. Adler S: Inhibition of rat glomerular visceral epithelial cell Finally, PPS caused a shift of TIMP-3 into the medium from growth by heparin. Am J Physiol 255: F781–F786, 1988 the ECM. TIMP-3 is normally bound to the ECM, whereas 9. Floege J, Eng E, Young BA, Couser WG, Johnson RJ: Heparin suppresses mesangial cell proliferation and matrix expansion in TIMP-1 and TIMP-2 are soluble. The present data suggest that experimental mesangioproliferative glomerulonephritis. Kidney one of the effects of PPS may be to facilitate the degradation Int 43: 369–380, 1992 of collagens by MMP by removing TIMP-3 from the cell 10. Gambaro G, Cavazzana AO, Luzi P, Piccoli A, Borsatti A, surface. Crepaldi G, Marchi E, Venturini AP, Baggio B: Glycosamino- This differential effect of PPS and heparin on MMP-2 and glycans prevent morphological renal alterations and albuminuria MMP-9 production is consistent with previous data describing in diabetic rats. Kidney Int 42: 285–291, 1992 the effects of heparin on MMP as cell-type specific (14,15). 11. Purkerson ML, Tollefsen DM, Klahr S: N-desulfated/acetylated Kenagy et al. reported that heparin inhibited the production of heparin ameliorates the progression of renal disease in rats with MMP-9 without affecting MMP-2 in primate arterial smooth subtotal renal ablation. J Clin Invest 81: 69–74, 1988 muscle cells (14). 12. Striker GE, Lupia E, Elliot S, Zheng F, McQuinn C, Blagg C, We found that increased levels of MMP-2 in normal mouse Selim S, Vilar J, Striker LJ: Glomerulosclerosis, arteriosclerosis, and vascular graft stenosis: Treatment with oral heparinoids. mesangial cells were associated with decreased cell prolifera- Kidney Int 63: S120–S123, 1997 tion. We have previously shown in mesangial cells isolated 13. Tan EML, Dodge GR, Sorger T, Kovalszky I, Unger GA, Yang from mice transgenic for bovine growth hormone that in- L, Levine EM, Iozzo RV: Modulation of extracellular matrix creased MMP-2 levels were also associated with decreased cell gene expression by heparin and endothelial cell growth factor in proliferation (18). This effect may be species-specific, because human smooth muscle cells. Lab Invest 64: 474–482, 1991 others have found that decreased cell proliferation was associ- 14. Kenagy RD, Nikkari ST, Welgus HG, Clowes AW: Heparin 68 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 62–68, 1999

inhibits the induction of three matrix metalloproteinases (strome- Similar binding with different rates of [3H] thymidine incorpo- lysin, 92-kD gelatinase, and collagenase) in primate arterial ration. J Cardiovasc Pharmacol 25: 782–788, 1995 smooth muscle cells. J Clin Invest 93: 1987–1993, 1994 27. Martin J, Knowlden J, Davies M, Williams J: Identification and 15. Shibata Y, Abiko Y, Goto K, Moriya Y, Takiguchi H: Heparin independent regulation of human mesangial cell metalloprotein- stimulates the collagen synthesis in mineralized cultures of the ases. Kidney Int 46: 877–885, 1994 osteoblast-like cell line, MC3T3-E-1. Biochem Int 28: 335–344, 28. Carome MA, Striker LJ, Peten EP, Elliot SJ, Yang CW, Stetler- 1992 Stevenson WG, Reponen P, Tryggvason K, Striker GE: Assess- 16. Haralson MA, Jacobson HR, Hoover RL: Collagen polymor- ment of 72-kilodalton gelatinase and TIMP-1 gene expression in phism in cultured rat kidney mesangial cells. Lab Invest 57: normal and sclerotic murine glomeruli. J Am Soc Nephrol 5: 513–523, 1987 1391–1399, 1994 17. Striker G, Killen PD, Farin FM: Human glomerular cells in vitro: 29. Davies M, Thomas GJ, Martin J, Lovett DH: The purification Isolation and characterization. Transplant Proc 12: 88–99, 1980 and characterization of a glomerular-basement- membrane-de- 18. Jacot TA, Striker GE, Stetler-Stevenson MA, Striker LJ: Mes- grading neutral proteinase from rat mesangial cells. Biochem J angial cells from transgenic mice with progressive glomerulo- 251: 419–425, 1988 sclerosis exhibit stable, phenotypic changes including undetect- 30. Corcoran ML, Hewitt RE, Kleiner DE Jr, Stetler-Stevenson WG: able MMP-9 and increased type IV collagen. Lab Invest 75: MMP-2: Expression, activation and inhibition. Enzyme Prot 49: 791–799, 1996 7–19, 1996 19. Conti FG, Striker LJ, Elliot SJ, Andreani D, Striker GE: Synthe- 31. Nagase H: Activation mechanisms of matrix metalloproteinases. sis and release of insulinlike growth factor I by mesangial cells J Biol Chem 378: 151–160, 1997 in culture. Am J Physiol 255: F1214–1219, 1988 32. Butler GS, Butler MJ, Atkinson SJ, Will H, Tamura T, Schade 20. Diegelmann RF, Peterkofsky B: Collagen biosynthesis during van Westrum S, Crabbe T, Clements J, d’Ortho MP, Murphy G: connective tissue development in chick embryo. Dev Biol 28: The TIMP2 membrane type 1 metalloproteinase “Receptor” reg- 443–453, 1972 ulates the concentration and efficient activation of Progelatinase 21. Oliver GW, Leferson JD, Stetler-Stevenson WG, Kleiner DE: A. J Biol Chem 273: 871–880, 1998 Quantitative reverse zymography: Analysis of picogram amounts 33. Fabunmi RP, Baker AH, Murray EJ, Booth RFG, Newby AC: of metalloproteinase inhibitors using gelatinase A and B reverse Divergent regulation by growth factors and cytokines of 95 kDa zymograms. Anal Biochem 244: 161–166, 1997 and 72 kDa gelatinases and tissue inhibitors of metalloprotein- 22. Peten EP, Garcia-Perez A, Terada Y, Woodrow D, Martin BM, ases-1, -2 and -3 in rabbit aortic smooth muscle cells. Biochem J Striker GE, Striker LJ: Age-related changes in alpha 1- and alpha 315: 335–342, 1997 2-chain type IV collagen mRNAs in adult mouse glomeruli: 34. Crabbe T, O’Connell JP, Smith BJ, Docherty AJP: Reciprocated Competitive PCR. Am J Physiol 263: F951–957, 1992 matrix metalloproteinase activation: A process performed by 23. He CJ, Striker LJ, Tsokos M, Yang CW, Peten EP, Striker GE: interstitial collagenase and progelatinase A. Biochem 33: 14419– Relationships between mesangial cell proliferation and types I 14425, 1994 and IV collagen mRNA levels in vitro. Am J Physiol 269: 35. Turck J, Pollock AS, Lee LK, Marti HP, Lovett DH: Matrix C554–562, 1995 metalloproteinase 2 (gelatinase A) regulates glomerular mesan- 24. Hurley MM, Kessler M, Gronowicz G, Raisz LG: The interaction gial cell proliferation and differentiation. J Biol Chem 271: of heparin and basic fibroblast growth factor on collagen syn- 15074–15083, 1996 thesis in 21-day fetal rat calvariae. Endocrinology 130: 2675– 36. Steinmann-Niggli K, Lukes M, Marti HP: Rat Mesangial cells 2682, 1992 and matrix metalloproteinase inhibitor: Inhibition of 72-kD type 25. Lippmann MM, Mathews MB: Heparins: Varying effects on cell IV collagenase (MMP-2) and of cell proliferation. JAmSoc proliferation in vitro and lack of correlation with anticoagulant Nephrol 8: 395–405, 1997 activity. Fed Proc 36: 55–59, 1977 37. Southgate KM, Davies M, Booth RFG, Newby AC: Involvement of 26. Stohr S, Meyer T, Smolenski A, Kreuzer H, Buchwald AB: extracellular-matrix-degrading metalloproteinases in rabbit aortic Effects of heparin on aortic versus venous smooth muscle cells: smooth-muscle cell proliferation. Biochem J 288: 93–99, 1992