Syndecan-1 Deficiency Aggravates Anti-Glomerular Basement

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector original article http://www.kidney-international.org & 2007 International Society of Nephrology

Syndecan-1 deficiency aggravates anti-glomerular basement membrane nephritis AL Rops1,MGo¨tte2, MH Baselmans1, MJ van den Hoven1, EJ Steenbergen3, JF Lensen4, TJ Wijnhoven4,5, F Cevikbas6, LP van den Heuvel5, TH van Kuppevelt4, JH Berden1 and J van der Vlag1

1Nephrology Research Laboratory, Division of Nephrology, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands; 2Department of Obstetrics and Gynecology, Mu¨nster University Hospital, Mu¨nster, Germany; 3Department of Pathology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands; 4Department of Matrix Biochemistry; Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands; 5Department of Pediatrics, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands and 6Department of Dermatology, Mu¨nster University Hospital, Mu¨nster, Germany

During the heterologous phase of experimental The syndecans belong to a family with four members of type anti-glomerular basement membrane (anti-GBM) nephritis, I transmembrane heparan sulfate (HS) proteoglycans, which leukocyte influx peaks within hours, whereas albuminuria consist of an extracellular N-terminal domain with several occurs within 1 day. In the subsequent autologous phase, HS side chains, a single hydrophobic transmembrane endogenous anti-GBM IgG develops and albuminuria domain, and a short C-terminal cytoplasmic domain.1 persists. Heparan sulfate (HS) proteoglycans like syndecan-1 Syndecans are expressed in a developmental and cell-type- play multiple roles during inflammation and we evaluate its specific pattern and act as co-receptors for various ligands role in experimental anti-GBM disease using syndecan-1 like growth factors and matrix molecules.1–3 Syndecan-1 knockout (sdc-1/) mice. During the heterologous phase, (sdc-1; CD138) is expressed on epithelial cells, endothelial glomerular leukocyte/macrophage influx was significantly cells, and leukocytes4 and appears to play a role in higher in the sdc-1/ mice and this was associated with tumorigenesis.5 During inflammation, the expression of higher glomerular endothelial expression of specific HS sdc-1 on both endothelial cells and leukocytes can be domains. In the autologous phase, glomerular influx of upregulated, which suggests that sdc-1 also plays a role in CD4 þ /CD8 þ T cells was higher in the sdc-1/ mice and the interaction of leukocytes with endothelium.6,7 Sdc-1- these mice had persistently higher albuminuria and serum deficient mice5,8 develop normally and are fertile, but show creatinine levels than wild-type mice. This resulted in a more an enhanced adhesion of leukocytes in the ocular vasculature sever glomerular injury and increased expression of and in mesentery venules after tumor necrosis factor (TNF)-a extracellular matrix proteins. The sdc-1/ mice developed treatment compared to wild type (WT) mice.9 The role of higher plasma levels and glomerular deposits of total mouse sdc-1 in renal physiology and pathology remains elusive, but Ig and IgG1 anti-rabbit IgG, whereas the levels of mouse it has been reported that the sdc-1 ectodomain is present in IgG2a anti-rabbit IgG were lower. Furthermore, decreased the circulation of patients with type I diabetes and early Th1 and higher Th2 renal cytokine/chemokine expression diabetic nephropathy.10 were found in the sdc-1/ mice. Our studies show that HS binds and modulates the activity of many inflam- syndecan-1 deficiency exacerbates anti-GBM nephritis matory molecules such as chemokines, selectins, integrins, shifting the Th1/Th2 balance towards a Th2 response. and adhesion molecules.11,12 The linear HS side chains Kidney International (2007) 72, 1204–1215; doi:10.1038/sj.ki.5002514; are composed of about 150 U of glucosamine-uronic acid published online 5 September 2007 disaccharides that can be sulfated at the N-position, 2-, KEYWORDS: chemokine receptor; chemokine; adhesion molecule; heparan 3-, and/or 6-O-position, and epimerized at the C-5 position sulfate; glomerulonephritis by a large family of HS-modifying enzymes.13 The number of possible combinations of modifications within one HS chain can give rise to an enormous structural diversity, which is important for the specific binding of a myriad of factors. The importance of specific N- and 6-O-sulfated HS domains on Correspondence: J van der Vlag, Division of Nephrology, Nephrology activated endothelium for the adhesion of polymorpho- Research Laboratory (279), Nijmegen Centre for Molecular Life Sciences, 12 Radboud University Nijmegen Medical Centre, Geert Grooteplein 26, nuclear granulocytes (PMNs) has been documented. 6525 GA Nijmegen, The Netherlands. E-mail: [email protected] Structural analysis of HS in kidney, lung, and spleen from Received 31 October 2006; revised 14 June 2007; accepted 3 July 2007; both WT and sdc-1-deficient mice revealed no differences in 14 published online 5 September 2007 HS disaccharide composition.

1204 Kidney International (2007) 72, 1204–1215 AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis original article

The experimental model of passive anti-glomerular base- 1 Collagen IV ment membrane (GBM) nephritis is induced by the EHS-laminin 0.8 HS intravenous injection of heterologous anti-GBM antibodies, Agrin Fibronectin which bind immediately to the GBM with subsequent 0.6 complement activation. In the early, acute heterologous, nm 450 0.4

phase, that is within hours after anti-GBM IgG administra- OD tion, glomerular PMN influx is peaking, whereas albuminuria is manifested within 1 day.15–17 At high doses of 0.2 heterologous anti-GBM serum, the albuminuria and glomerular lesions in the acute phase of anti-GBM nephritis are 0 100 200 300 400 500 strictly PMN-dependent and complement-independent,18 Rabbit anti-mouse GBM IgG (g/ml) whereas at lower doses of anti-GBM antibodies, also Figure 1 | Rabbit anti-mouse GBM antibodies react with purified complement-dependent mechanisms may be operative.19 In collagen IV, laminin, HS, agrin, and fibronectin. Results are particular, the cationic neutral proteinases elastase and expressed as means7s.e.m. optical density values at 450 nm from three experiments. cathepsin G, which are released by activated PMNs, contribute to glomerular damage and albuminuria in the acute heterologous phase of anti-GBM nephritis, as beige mice that are deficient in these proteinases failed to develop a fine granular staining along the GBM after 1 h, which albuminuria after induction of anti-GBM nephritis.20 The became linear after 2 h (Figure 2c and d; Table 1). Some integrin Mac-1, expressed by PMNs, probably interacts with segmental staining of Bowman’s capsule and tubular base- deposited C3 complement,21 which subsequently may lead to ment membranes was observed, which was also noted in mice the release of elastase and cathepsin G by the PMNs.20,21 that received control rabbit IgG and has been described Subsequently, these cationic proteinases bind to the GBM, previously.15 In both WT and sdc-1-deficient mice, focal and probably to anionic HS,22–24 which leads to increased GBM segmental deposits of fibrin could be observed after 4 h, permeability and albuminuria.21 The autologous phase of which increased until day 8 (Figure 2e and f; Table 1). The anti-GBM nephritis is initiated by the immune response of autologous phase started after 4 days as revealed by the the host to the injected heterologous anti-GBM antibodies, presence of mouse Ig anti-rabbit IgG along the capillary wall that is about 1 week after administration of anti-GBM anti- in both mice strains. However, in sdc-1-deficient mice about bodies,25 and is characterized by persistent severe albumi- two-fold more mouse Ig anti-rabbit IgG was present along nuria. The development of antibodies to the heterologous the capillary wall after 4 and 8 days (Figure 2g–i; Table 1). anti-GBM antibodies has been identified as a Th1 response.26 Administration of control rabbit IgG to WT and sdc-1- The deposition of mouse Ig anti-rabbit IgG along the GBM deficient mice did not lead to glomerular binding and may be important for the persistence of albuminuria during deposition of C3c, fibrin or mouse Ig at all evaluated time the autologous phase of anti-GBM nephritis.27 points (data not shown). These data indicate that sdc-1 It is well established that HS proteoglycans like sdc-1 play deficiency does not affect the binding/deposition of rabbit a crucial role in the leukocyte–endothelium interaction anti-mouse GBM IgG, C3c, and fibrin during the hetero- during inflammation in different vascular beds.28,29 However, logous and autologous phase of anti-GBM nephritis. there is little knowledge about the role of HS(PG)s in However, sdc-1 deficiency leads to an enhanced glomerular glomerulonephritis, therefore, we evaluated experimental deposition of mouse Ig anti-rabbit IgG during the autologous anti-GBM nephritis in both WT and sdc-1-deficient mice. phase.

RESULTS Experimental anti-GBM nephritis leads to a higher Induction of anti-GBM nephritis in WT and sdc-1-deficient glomerular expression of sdc-1 and -4 mice and identification of the heterologous and autologous Immunohistochemical analysis of sdc-1 and sdc-4 expression phase in the normal kidney revealed a low expression level in The generated rabbit anti-mouse serum stained at higher glomeruli whereas the expression was significantly upregu- dilutions exclusively the GBM of both sdc-1-deficient and lated after 2 and 18 h during anti-GBM nephritis, which WT mice (data not shown). Furthermore, the rabbit anti- decreased after 4 and 8 days to normal levels (Figure 3a and b). mouse GBM antibodies reacted in enzyme-linked immuno- Furthermore, the sdc-1 and -4 mRNA expression in renal sorbent assay with purified agrin, HS, laminin, fibronectin, cortex significantly increased after 2 h during anti-GBM and collagen type IV (Figure 1). Administration of rabbit nephritis, which persisted after 8 days (Table 2). Both at the anti-GBM antibodies to WT and sdc-1-deficient mice protein and the mRNA level, the glomerular expression of resulted in a binding of rabbit IgG along the GBM already sdc-4 increased in sdc-1-deficient mice after 2 and 18 h, after 1 h (Figure 2a and b; Table 1), which remained which decreased to normal levels after 4 and 8 days (Figure unaffected until day 8. Rabbit IgG binding was accompanied 3b; Table 2). It appeared that the mRNA expression of sdc-4 with C3c deposition in both mice strains, which started with in sdc-1-deficient mice was lower compared with sdc-4

Kidney International (2007) 72, 1204–1215 1205 original article AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis

WT Sdc-1–/– concluded that both sdc-1 and -4 play a role during anti- ab GBM nephritis and that the other syndecan members do not compensate for the sdc-1 deficiency. Rabbit anti-GBM IgG (2 h) Sdc-1 deficiency leads to a higher glomerular leukocyte influx and a higher expression of inflammatory HS domains on glomerular endothelium in anti-GBM nephritis cd Next, we evaluated the glomerular influx of PMNs, being one of the key determinants of glomerular damage and albuminuria in the early phase of anti-GBM nephritis, C3c (2 h) and the glomerular influx of macrophages and CD4 þ and CD8 þ T cells. Glomerular PMN and macrophage influx could be observed 1 h after administration of rabbit anti- GBM IgG (Figure 4a and b), which was maximal after 2 h ef and quickly decreased after 4–18h. However, after 2 h, the glomerular PMN and macrophage influx was significantly Fibrin (day 8) higher in sdc-1-deficient mice compared with WT mice. Furthermore, the macrophage influx in the interstitium was significantly increased after 4 and 8 days and significantly higher in sdc-1-deficient mice compared with WT mice gh (data not shown). The glomerular influx of CD4 þ and CD8 þ T cells was observed after 18 h and increased after Mouse IgG (day 8) 4 and 8 days, whereas this influx was higher in sdc-1-deficient mice (Figure 4c and d). Recently, we have identified an increased expression of inflammatory N- and 6-O-sulfated HS domains on cultured i ∗ glomerular endothelial cells after activation with TNF-a, 12 000 which were directly involved in the adhesion of granulo- WT 30 Sdc-1–/– cytes. The anti-HS antibodies AO4B08, EW3D10, and 8000 EW4G2 identified these glomerular endothelial HS domains, ∗ respectively. Now, we evaluated the expression of these

4000 inflammatory HS domains on glomerular endothelium by double staining with these specific anti-HS antibodies and an Glomerular mouse Ig deposits (a.u.) anti-CD31 antibody. It appeared that the glomerular endo- 0 Day 4 Day 8 thelial expression of the N- and 6-O-sulfated HS domains Figure 2 | Characteristics of the heterologous and autologous was significantly increased after 2 h in both WT and sdc-1- phase of anti-GBM nephritis in WT and sdc-1-deficient mice. deficient mice compared with untreated mice or mice that Immunofluorescence staining 2 h after anti-GBM administration received control rabbit IgG (Figure 5a–d). Interestingly, the revealed binding of (a, b) rabbit anti-GBM IgG and (c, d) mouse glomerular endothelial expression of the inflammatory HS complement 3c (C3c) along the capillary wall, which was similar between (a, c) WT and (b, d) sdc-1-deficient mice. (e, f) At day 8 after domains was significantly higher in sdc-1-deficient mice induction of anti-GBM nephritis, the glomerular deposition of fibrin compared with WT mice (Figure 5b–d). Expression analysis was similar between (e) WT and (f) sdc-1-deficient mice, whereas the of the endothelial activation marker intercellular adhesion deposition of mouse anti-rabbit Ig along the capillary wall was higher molecule-1 and the HS ligand L-selectin on leukocytes in (h) sdc-1-deficient mice compared with (g) WT mice (original magnification Â 400). (i) Glomerular deposition of mouse Ig revealed a higher expression in sdc-1-deficient mice com- anti-rabbit IgG was higher in sdc-1-deficient mice compared with pared with WT mice after 2 h (Table 2). Taken together, WT mice at 4 and 8 days after administration of rabbit anti-GBM IgG. sdc-1-deficient mice show a higher number of glomerular 7 The titers (means s.e.m.) are shown from eight WT and eight PMNs, macrophages, and CD4 þ and CD8 þ T cells, and sdc-1-deficient mice. *Po0.05 and **Po0.001 versus anti-GBM injected WT mice. WT, wild type; Sdc-1/, syndecan-1-deficient. a higher glomerular endothelial expression of inflammatory N- and 6-O-sulfated HS domains. expression in scd-1-deficient mice during anti-GBM nephri- Sdc-1 deficiency leads to higher albuminuria, a worse renal tis, whereas sdc-4 protein expression was similar in WT and function, and more extensive glomerular injury in anti-GBM sdc-1-deficient mice. Sdc-2 and -3 were not expressed in nephritis glomeruli of normal kidneys of WT and sdc-1-deficient mice A significant albuminuria was already observed after 18 h in (not shown), whereas at the mRNA level, the expression was the heterologous phase in both WT and sdc-1-deficient mice not changed during anti-GBM nephritis (Table 2). It can be (Figure 6a), which was significantly higher in sdc-1-deficient

1206 Kidney International (2007) 72, 1204–1215 AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis original article

Table 1 | Presence of rabbit IgG, C3c, fibrin, and mouse Ig anti-rabbit IgG during anti-GBM nephritis in wild type and syndecan-1-deficient mice

Phase during anti-GBM nephritis Time Rabbit IgGa C3ca Fibrina Mouse Ig anti-rabbit IgG WT Sdc-1/ Heterologous 1 h ++++ ++ - 2 h ++++ +++ 7 4 h ++++ +++ + 18 h ++++ + ++

Autologous 4 days ++++ + +++ + ++ 8 days ++++ + ++++ ++ ++++ C3c, complement 3c; Sdc-1/, syndecan-1-deficient; WT, wild type. aDeposition in WT and syndecan-1-deficient mice is identical.

Syndecan-1 10 mice had signiﬁcantly higher plasma creatinine levels compared with WT mice. Mice that received control rabbit IgG 8 did not develop albuminuria or renal function impairment (Figure 6a and b). 6 Renal histology after 18 h was similar in WT and sdc-1- deﬁcient mice, that is a minority of glomeruli showed 4 segmental or global endothelial swelling with occlusion of the

2 capillary lumens due to periodic acid-Schiff (PAS)-positive

Staining intensity (a.u.) fibrinous material. At day 4, approximately 50% of the 0 glomeruli of the sdc-1-deficient mice showed globally 2 h 18 h 4 days 8 days occluded and distended capillaries filled with PAS-positive fibrinous material, which was sporadically observed in WT Syndecan-4 mice. In addition, proteinacious tubular casts were more 10 frequently observed in sdc-1-deficient mice compared with 8 WT mice (Figure 7). At day 8, the differences between WT and sdc-1-deficient mice were even more pronounced, that is 6 in sdc-1-deficient mice, the majority of glomeruli showed globally distended capillaries with PAS-positive material, 4 occasionally cellular crescents, whereas tubular casts were frequent. In contrast, WT mice showed only thickening of the 2

Staining intensity (a.u.) capillary wall, mostly open capillary lumens and no crescents, 0 whereas tubular casts were sporadically observed (Figure 7). WT Sdc-1–/– WT Sdc-1–/– WT Sdc-1–/– WT Sdc-1–/– Next, we evaluated the expression of extracellular matrix 2 h 18 h 4 days 8 days components. As our anti-GBM serum reacts with collagen IV, Figure 3 | Glomerular expression of sdc-1 and -4 in WT and laminin, agrin, HS, and fibronectin (Figure 1), we used sdc-1-deficient mice during anti-GBM nephritis. Glomerular digoxigenin-labeled rabbit anti-GBM serum, as a probe to (a) sdc-1 and (b) sdc-4 protein expression increased B3-fold at 2 and 18 h after the induction of anti-GBM nephritis in WT mice. At days stain renal sections of untreated mice (WT and sdc-1- 4 and 8, glomerular expression of sdc-1 and -4 became deficient mice) and mice with anti-GBM nephritis at 2 h and normal (indicated by the dotted line). (b) Glomerular expression of 8 days. These data revealed that there is no difference in sdc-4 was also increased in sdc-1-deficient mice at 2 and 18 h after expression between untreated WT and sdc-1-deficient mice, the induction of anti-GBM nephritis and became normal at days 4 but an increased expression of extracellular matrix compo- and 8. ^Po0.05 versus untreated mice or mice injected with control rabbit IgG. WT, wild type mice; Sdc-1/, syndecan-1-deficient mice. nents in sdc-1-deficient mice compared with WT mice at day 8 due to expansion of GBM matrix (Figure S1). At the mRNA level, the expression of the glomerular extracellular matrix mice. Albuminuria increased further in the autologous phase proteins collagen XVIII, fibronectin, collagen IV a3 through after 4 and 8 days, whereas the significant difference between a5, and laminin a5, b2, g1 was two- to three-fold higher in WT and sdc-1-deficient mice persisted (Figure 6a). The the sdc-1-deficient mice compared with WT mice at day 8 renal function, that is the plasma creatinine concentration, (Table 2). The expression of agrin and the non-GBM collagen was unaffected during the heterologous phase (Figure 6b), IV chains a1anda2, however, was not different. Further- whereas plasma creatinine levels increased during the more, we found that the expression of the matrix remodeling autologous phase (Figure 6b). After 8 days, sdc-1-deficient enzymes matrix metalloproteinase-7 and matrix metallopro-

Kidney International (2007) 72, 1204–1215 1207 original article AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis

Table 2 | Quantitative mRNA expression of syndecans, extracellular matrix components, metalloproteases, growth factors, cytokines, and chemokines in the renal cortex of wild type and syndecan-1-deficient mice during anti-GBM nephritisa 2 h 8 days WT Sdc-1/ WT Sdc-1/ Sdc-1 6.8172.57&& ND 2.1070.36&& ND Sdc-2 0.8370.39 0.8470.13 0.9470.20 1.0270.16 Sdc-3 0.7270.24 1.0370.21 1.0670.33 1.1670.27 Sdc-4 6.7071.95** 3.5270.88 2.7470.36** 1.7770.51 ICAM-1 22.2072.49 32.1573.41& 11.2973.54 10.1472.44 L-selectin 4.4271.27 14.0872.21& 1.4170.21 3.3070.92& Fibronectin 1.0470.87 2.0571.18 6.5671.34 25.6172.46& Agrin 1.5170.67 2.1571.22 3.2371.08 4.8471.27 Collagen XVIII 2.3171.01 1.6771.21 3.8771.36 8.4772.18* Laminin a5 1.7170.54 1.4670.98 3.1471.52 6.2371.11* Laminin b2 1.2470.63 1.2670.41 2.1171.54 7.2472.50* Laminin g1 1.0670.13 1.6270.23 1.7470.72 3.4271.09* Collagen IV a1 1.6670.3 1.4970.29 5.9171.43 5.2871.24 Collagen IV a2 0.9870.06 1.5670.16 1.5971.18 1.3571.43 Collagen IV a3 0.6670.13 1.5770.79 2.4871.11 4.9671.55* Collagen IV a4 0.6670.21 1.1070.55 2.0370.21 4.1671.27* Collagen IV a5 0.9570.12 1.9570.91 1.0370.32 3.3371.33* MMP-7 1.5770.69 1.2370.22 28.2875.27 90.1578.42& MMP-9 2.4170.98 3.6970.91 5.7472.18 13.8674.32* TGF-b1 1.9770.22 2.267096 4.4371.38 3.7470.92 FGF-1 1.2370.38 1.8370.49 1.6770.09 1.4770.32 CTGF 1.0770.32 0.9070.23 7.8271.88 10.1872.11 IL-1b 12.6070.25 46.9070.32& 0.7370.21 0.9570.19 TNF-a 1.1070.09&& 0.4370.11 1.0270.21^ 0.4470.17 IFN-g 1.4470.14 1.5670.31 3.0270.09&& 0.9870.11 MCP-1 67.01715.42 192.67742.45^ 3.6271.32 13.6974.23^ MIP-1a 140.45731.37^^ 41.18718.39 0.9970.19 0.7870.22 IL-2 1.0370.27 0.8370.13 1.6670.13&& 0.8570.02 IL-4 0.4370.10 0.5870.21 0.4970.07 1.1970.08& IL-6 24477568 37117654* 1.1370.23 1.1070.17 IL-10 0.6970.12 0.6670.08 0.7070.16 4.7570.18& IL-12a 0.7170.13^^ 0.2770.15 0.7670.04&& 0.3770.05 IL-12b 0.4670.12^^ 0.1170.08 1.0070.06&& 0.4070.09 Tnfrsf1a 1.0770.23 0.9670.16 4.0471.13^^ 1.8470.12 Tnfrsf1b 3.2270.84 2.1970.97 17.2475.43** 8.5673.27 CTGF, connective tissue growth factor; GBM, glomerular basement membrane; FGF, fibroblast growth factor; ICAM-1, intercellular adhesion molecule-1; IFN-g, interferon-g; IL, interleukin; MCP-1, monocyte chemotactic protein-1; MIP-1a, macrophage inflammatory protein 1a; MMP, matrix metalloproteinase; ND, not detectable; Sdc-1/, syndecan-1-deficient mice; TGF, transforming growth factor; TNF-a, tumor necrosis factor-a, Tnfrsf, TNF receptor; WT, wild type mice. The mean expression7s.d. of four mice per group are depicted. There were no differences in expression between untreated WT and sdc-1-deficient mice. The significant highest expression is given in bold. *Po0.05 versus anti-GBM injected WT mice; ^Po0.01 versus anti-GBM injected WT mice; &Po0.001 versus anti-GBM injected WT mice; **Po0.05 versus anti-GBM injected sdc-1-deficient mice; ^^Po0.01 versus anti-GBM injected sdc-1-deficient mice; &&Po0.001 versus anti-GBM injected sdc-1-deficient mice. aThe relative expression compared with untreated WT or sdc-1/ mice is shown. teinase-9 was significantly higher in sdc-1-deficient mice at rabbit IgG response when mice received control rabbit IgG. day 8 (Table 2). Quantitative mRNA expression analysis of Irrespective of the presence of sdc-1, the response against the growth factors fibroblast growth factor, transforming rabbit anti-GBM IgG was significantly higher than the growth factor-b and connective tissue growth factor, how- response against control rabbit IgG (data not shown). It ever, revealed no significant differences between WT and sdc- has been described that the autologous phase in experimental 1-deficient mice after 2 h or 8 days. anti-GBM is characterized by a Th1 response.26 We determined plasma levels and glomerular deposition of two Ig Sdc-1 deficiency leads to a shift from a Th1 to a Th2 response isotypes, IgG2a that is related to a Th1 response and IgG1 in anti-GBM nephritis that is related to a Th2 response. It appeared that in sdc-1- Sdc-1-deficient mice showed significantly more mouse deficient mice plasma levels (Figure 8a and b) and glomerular Ig anti-rabbit IgG along the capillary wall compared with deposits (Figure 8c and d) of mouse IgG1 anti-rabbit IgG WT mice (Figure 1i). Evaluation of plasma levels of total mouse were significantly higher, whereas plasma levels and glome- Ig anti-rabbit IgG revealed in sdc-1-deficient mice two-fold rular deposits of mouse IgG2a anti-rabbit IgG were signi- higher levels than in WT mice at days 4 (Figure 8a) and 8 ficantly lower compared with WT mice after 4 (Figure 8a and (Figure 8b) after administration of rabbit anti-GBM IgG. c) and 8 days (Figure 8b and d). We further analyzed Th1/ However, there was no difference in total mouse Ig anti- Th2 (promoting) cytokines/chemokines. At the mRNA level,

1208 Kidney International (2007) 72, 1204–1215 AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis original article

phase. Furthermore, the expression of TNF receptor 1a and 100 100 WT TNF receptor 1b was lower in the sdc-1-deﬁcient mice 80 80 Sdc-1–/– (Table 2). Using cytokine/chemokine arrays, we could 60 60 conﬁrm the mRNA analysis data for interferon-g, IL-4, IL- 40 40 10, and monocyte chemotactic protein-1 at the protein level

20 20 in the autologous phase (Table 3), whereas also an increased No. of PMNs/50 No. glomeruli 0 0 IL-13 expression was present in sdc-1 deficient mice. Finally, we found at the protein level, in the sdc-1-deficient mice, an 100 ∗∗ 100 increased expression of the tyrosine kinase Axl and the 80 80 ∗∗ chemokine epithelial cell-derived neutrophil-activating pep- 60 ∗∗ 60 tide (ENA)-78/CXCL5, which both are involved in inflam- 31,32 40 40 matory renal diseases. In summary, sdc-1 deficiency leads to an increased 20 20 albuminuria and impaired renal function in anti-GBM

No. of macrophages/50 No. glomeruli 0 0 nephritis, which may be explained by a higher glomerular 50 50 PMN and macrophage inﬂux and by higher plasma levels and ∗∗ 40 40 ∗∗ glomerular deposition of mouse Ig anti-rabbit anti-GBM 30 30 IgG. Furthermore, sdc-1 deﬁciency leads to an increased antibody response against rabbit anti-GBM IgG and to a clear 20 20 shift from a Th1 to a Th2 response. 10 10

No. of CD4+ T cells/50 of CD4+ No. glomeruli 0 0 DISCUSSION

50 50 Previously, it has been shown that the HS proteoglycan sdc-1 ∗∗ is important for leukocyte trafﬁcking on activated endothe- 40 40 lium.9,33 Therefore, we evaluated the role of sdc-1 in 30 30 ∗ experimental anti-GBM nephritis. The rabbit anti-GBM 20 20 serum we generated was effective in binding to the GBM 10 10 and led to glomerular C3c and ﬁbrin deposition in the

No. of CD8+ T cells/50 of CD8+ No. glomeruli 0 0 heterologous and autologous phase of anti-GBM nephritis, 12418 48 which is in accordance with previous studies15,18,20 and which Time (h) Time (days) was not affected by sdc-1 deficiency. Importantly, we found Figure 4 | Glomerular leukocyte influx in WT and sdc-1-deficient an increased glomerular sdc-1 expression 2 h after rabbit mice during anti-GBM nephritis. (a) Glomerular PMN and (b) macrophage influx, analyzed by immunohistochemical staining, anti-mouse GBM IgG administration in WT mice. The peaked 2 h after rabbit anti-GBM IgG administration in both WT and increased glomerular sdc-1 expression persisted until 18 h, sdc-1-deficient mice and at this time point was significantly higher in whereas sdc-1 expression became normal after 4 days. We sdc-1-deficient mice. After 18 h, (a, b) 4 and 8 days, the glomerular PMN and macrophage influx was similar to untreated mice (indicated found the same for sdc-4 expression, in this case, in both WT by the dotted line) or control mice injected with rabbit IgG. (c) and sdc-1-deficient mice. However, during anti-GBM Glomerular CD4 þ T-cell and (d) CD8 þ T-cell influx was significantly nephritis, the mRNA expression of sdc-1 and sdc-4 expres- higher in sdc-1-deficient mice after 4 and 8 days. The results are sion in WT mice was still increased after 8 days. This may be expressed as the means7s.e.m. from eight WT and eight explained by a higher turnover of the syndecans during the sdc-1-deficient mice. ^Po0.05 versus untreated mice or mice injected with control rabbit IgG, *Po0.01 and **Po0.001 versus heterologous phase. Furthermore, in contrast to protein anti-GBM-injected WT mice. WT, wild type; Sdc-1/, expression during anti-GBM nephritis, mRNA expression of syndecan-1-deficient. sdc-4 was B2-fold higher in WT mice compared with sdc-1- deficient mice in both the heterologous and autologous phase. These data suggest that under inflammatory condi- an increased expression of the Th2 promoting and B cell tions, sdc-1 regulates the expression of the gene encoding stimulating proinflammatory cytokine interleukin (IL)-1b sdc-4 without a clear effect on sdc-4 protein expression. The was observed in the sdc-1-deficient mice in the heterologous underlying mechanisms remain unclear, but similar observa- phase. Furthermore, Th2 (promoting) cytokines IL-4, IL-6, tions have been made for sdc-3 and sdc-4.34 Nevertheless, it IL-10, IL-13, and monocyte chemotactic protein-1 were can be concluded that both sdc-1 and sdc-4 play a role during higher in the sdc-1-deficient mice compared with WT mice anti-GBM nephritis and that other syndecans do not in the heterologous and/or autologous phase. Moreover, compensate for the loss of sdc-1. Th1 (promoting/maintaining) cytokines TNF-a, interferon-g, In both WT and sdc-1-deficient mice, glomerular PMN/ IL-2, IL-12a, IL-12b, and macrophage inflammatory protein macrophage influx was maximal 2 h after rabbit anti-mouse (MIP)-1a were lower in the sdc-1-deficient mice compared GBM IgG administration, which is consistent with previous with the WT mice in the heterologous and/or autologous studies.15,18,20 However, the glomerular PMN/macrophage

Kidney International (2007) 72, 1204–1215 1209 original article AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis

AO4B08 (red) EW3D10 (red) EW4G2 (red) CD31 (green) CD31 (green) CD31 (green)

0 h

2 h

AO4B08 (GlcNS6S–IdoA2S–GlcNS6S) EW3D10 (GlcNS6S) ∗ 10 ∗ 10

8 8

6 6

4 4

2 2 Staining intensity (a.u.) Staining intensity (a.u.) 0 0 WT Sdc-1–/– WT Sdc-1–/–

EW4G2 (GlcNS6S–GlcA–GlcNS6S) 10 ∗ 8

2 Staining intensity (a.u.) 0 WT Sdc-1–/– Figure 5 | Glomerular endothelial expression of inflammatory, N- and 6-O-sulfated, HS domains during the acute heterologous phase of anti-GBM nephritis. (a) Immunofluorescence double staining of N- and 6-O-sulfated HS domains recognized by the anti-HS antibodies AO4B08 (red), EW3D10 (red), EW4G2 (red), and the endothelium marker CD31 (green) showed an increased glomerular endothelial expression of inflammatory HS domains in sdc-1-deficient mice 2 h after anti-GBM IgG administration. Original magnification Â 400. Semiquantitative analysis of the glomerular endothelial expression of inflammatory HS domains recognized by (b) AO4B08, (c) EW3D10, and (d) EW4G2 colocalizing with CD31 showed a higher expression of these HS domains in sdc-1-deficient mice compared with WT mice 2 h after anti-GBM IgG administration. The expression of CD31 was rather constant at the different time points both for the WT and the sdc-1-deficient mice. The epitopes of the single chain anti-HS antibodies are given within parentheses. Staining intensities are expressed as means7s.e.m. from eight WT and eight sdc-1-deficient mice in arbitrary units. Dotted lines indicate the base line score of untreated mice or mice injected with rabbit control IgG. *Po0.001 versus anti-GBM injected WT mice. GlcNS6S, N- and 6-O-sulfated glucosamine; GlcA, glucuronic acid; IdoA2S, 2-O-sulfated iduronic acid; WT, wild type; sdc-1/, syndelan/ deficient. influx in sdc-1-deficient mice was significantly higher com- adhesion of sdc-1-deficient leukocytes to activated HUVEC pared with WT mice. Furthermore, we found an enhanced in vitro.9,33 Therefore, the higher glomerular PMN influx in glomerular influx of CD4 þ and CD8 þ T cells in sdc-1- sdc-1-deficient mice may be caused by an enhanced adhesion deficient mice, in particular, during the autologous phase. of PMNs to the activated glomerular endothelium. This may Our findings are in accordance with previous data, as an be explained by the impossibility of shedding, that is increased adhesion of leukocytes has been shown in sdc-1- proteolytic cleavage, of the ectodomain of sdc-1.36–39 This deficient mice in a retina perfusion model and in acute lung soluble ectodomain that contains the HS side chains could inflammation.9,35 It appeared that the enhanced leukocyte compete for proinflammatory chemokines destined for the adhesion mainly could be attributed to the sdc-1 deficiency endothelium or could modify leukocyte adhesion by comp- of leukocytes, which was underscored by an enhanced etitive binding to adhesion molecules, such as selectins and

1210 Kidney International (2007) 72, 1204–1215 AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis original article

WT Sdc-1–/– ∗ 250 WT 250 ∗ Sdc-1–/– 200 200 150 150 Day 4 100 100 50 50

Albumin/creatinine ratio ratio Albumin/creatinine (mg/mg) 0 0 12418 48 Time (h) Time (days)

∗∗ 1.0 1.0 Day 8 0.8 0.8 0.6 0.6 0.4 0.4 Figure 7 | Histology of glomerular injury in WT and 0.2 0.2 sdc-1-deficient mice during anti-GBM nephritis. Glomeruli in

Plasma creatinine ratio Plasma creatinine ratio (mg/dl) 0.0 0.0 sdc-1-deficient mice showed more globally occluded and distended 12418 48 capillaries filled with PAS-positive fibrinous material (arrows) after Time (h) Time (days) 4 and 8 days. In addition, proteinacious tubular casts (asterisks) were more frequently observed in sdc-1-deficient mice compared Figure 6 | Sdc-1-deficient mice develop higher albuminuria and a with WT mice. At day 8, WT mice showed thickening of capillary walls worse renal function. (a) The albumin/creatinine ratio (mg/mg) (arrowhead) but mostly open capillary lumens. Original magnification (ACR) was significantly higher in sdc-1-deficient mice at 18 h, 4, and 8 Â 250. WT, wild type; Sdc-1/, syndecan-1-deficient. days after anti-GBM IgG administration. The ACR of WT and sdc-1-deficient mice at 1 h (0.4270.19 in WT mice, 0.3070.16 in Sdc-1/ mice), 2 h (0.5470.44 in WT mice, 0.9370.56 in Sdc-1/ mice), and 4 h (1.2670.52 in WT mice, 2.0170.32 in Sdc-1/ mice) was similar to untreated mice or mice injected with rabbit control IgG Mac-1, possibly by modulating intracellular cascades.40,41 (indicated by the dotted line). (b) Plasma creatinine concentration (mg/dl) was significantly higher in sdc-1-deficient mice than in WT However, we can conclude that sdc-1 apparently is not the mice at 8 days after anti-GBM IgG administration. Plasma creatinine HS proteoglycan core protein that carries HS chains with the concentrations of WT and sdc-1-deficient mice after 1, 2, 4, 18 h, and inflammatory, N- and 6-O-sulfated, HS domains, as these 4 days were similar to untreated mice or mice injected with control domains are also expressed on the glomerular endothelium of rabbit IgG (indicated by the dotted line). The means7s.e.m. are sdc-1-deficient mice. shown (from eight WT and eight sdc-1-deficient mice). ^Po0.001 versus untreated mice or control mice injected with rabbit IgG, During the acute heterologous phase that is after 18 h, *Po0.05 and **Po0.001 versus anti-GBM injected WT mice. albuminuria was higher in sdc-1-deficient mice compared / WT, wild type; Sdc-1 , syndecan-1-deficient. with WT mice. This may be explained by the higher glomerular PMN/macrophage influx, which could result in more release of proteinases by these leukocytes and thus integrins. An additional explanation for the higher PMN more destruction of the capillary wall.15 This increased influx influx in sdc-1-deficient mice may be the significantly higher led to a higher albuminuria at 4 h, which became significant glomerular endothelial expression of specific inflammatory, after 18 h. This difference in albuminuria persisted in the N- and 6-O-sulfated, HS domains in these mice compared autologous phase, whereas renal function of sdc-1-deficient with WT mice at the moment of maximal PMN influx, that is mice was worse after 8 days. In line with this, glomerular 2 h after anti-GBM administration. Recently, we found that injury in sdc-1-deficient mice was more severe compared these inflammatory HS domains are crucial for the firm with WT mice after 4 and 8 days, which was accompanied by adhesion of leukocytes to activated glomerular endothelial an increased matrix expansion as revealed by the higher cells in vitro.30 At this moment, we have no mechanistic expression of extracellular matrix components and extra- explanation for the higher expression of these inflammatory cellular matrix remodeling enzymes. The expression of the HS domains on activated glomerular endothelium in sdc-1- growth factors, transforming growth factor-b and connective deficient mice. We also found a higher expression of tissue growth factor, that promote glomerular scarring was intercellular adhesion molecule-1 in sdc-1-deficient mice increased after 8 days, but was not different between sdc-1- during the heterologous phase, which may indicate that the deficient and WT mice. We also identified an increased renal endothelium in sdc-1-deficient mice is more sensitive expression of the tyrosine kinase Axl and the chemokine for inflammatory stimuli in the heterologous phase of anti- neutrophil-activating peptide-78/CXCL5, which previously GBM nephritis. Apparently, sdc-1 on either the glomerular have been identified in inflammatory renal diseases.31,32 As endothelium and/or on leukocytes has anti-inflammatory discussed for the heterologous phase, this may be explained properties by controlling the expression of endothelial HS by more glomerular damage mediated by PMNs and macro- ligands for receptors on leukocytes, that is L-selectin and phages, initially. However, it has been reported that the

Kidney International (2007) 72, 1204–1215 1211 original article AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis

Sdc-1–/– WT 4 days 8 days ∗ ∗ 1000 1 000 000 ∗

∗ 100 ∗ 10 000 ∗

10 100 Ig isotype levels (U/ml) Ig isotype levels Ig isotype levels (U/ml) Ig isotype levels Plasma mouse anti-rabbit Plasma mouse anti-rabbit 1 1 Ig IgG1 IgG2a Ig IgG1 IgG2a

∗ 6000 30 000 ∗

4000 20 000

∗ 2000 10 000 ∗ Glomerular mouse Ig isotype deposits (a.u.) Glomerular mouse Ig isotype deposits (a.u.) 0 0 IgG1 IgG2a IgG1 IgG2a Figure 8 | Sdc-1-deficient mice develop higher levels of mouse Ig/IgG1 anti-rabbit IgG and lower levels of mouse IgG2a anti-rabbit IgG than WT mice during anti-GBM nephritis. The plasma levels of total mouse Ig and IgG1 anti-rabbit IgG in the sdc-1-deficient mice were higher, whereas plasma levels of mouse IgG2a anti-rabbit IgG were lower compared with WT mice (a) 4 and (b) 8 days after administration of rabbit anti-GBM IgG. Glomerular mouse IgG1 deposits were higher in sdc-1-deficient mice, whereas glomerular mouse IgG2a deposits were lower compared with WT mice (c) 4 and (d) 8 days after administration of rabbit anti-GBM IgG. The results (means7s.e.m.) are expressed as (a, b) titers of plasma levels in U/ml or as (c, d) titers of appropriate antibodies detecting glomerular deposits in arbitrary units from eight WT and eight sdc-1-deficient mice. *Po0.001 versus anti-GBM injected WT mice. WT, wild type; Sdc-1/, syndecan-1-deficient.

Table 3 | Semiquantitative protein expression of cytokines sdc-1-deficient mice showed a higher response against rabbit and chemokines in the kidneys of wild-type and anti-GBM IgG compared with control rabbit IgG. This can be syndecan-1-deficient mice at 8 days after rabbit anti-mouse a explained by interference of the anti-GBM antibodies with GBM IgG administration the immune response against rabbit anti-GBM IgG. The anti- WT Sdc-1/ GBM antibodies contained reactivities against HS and agrin, 12,42,43 IFN-g 11.2 6.60 which are both present on immunological cells and IL-4 14.4 24.3 thus may be potential targets of the anti-GBM antibodies. IL-10 o5 9.94 Another explanation may be that the anti-GBM rabbit IgG IL-13 8.15 13.2 serves as a planted antigen, whereas the control rabbit IgG MCP-1 8.86 17.6 ENA-78/CXCL5 9.81 22.6 remains in the circulation. Furthermore, the higher response Axl o5 10.8 against rabbit anti-GBM IgG in sdc-1-deficient mice may be ENA-78, epithelial cell-derived neutrophil activating peptide-78; IFN-g, interferon- g; the result of a combination of the intrinsic sdc-1 deficiency of IL, interleukin; MCP-1, monocyte chemotactic protein-1; Sdc-1/, syndecan-1- certain immunological cells such as B cells44,45 and immature deficient mice; WT, wild-type mice. 46 aThe average expression of three pooled kidneys is shown and is expressed as the dendritic cells and the potential effect of the anti-GBM integrated density. Signals lower than five (three times the signal of the negative antibodies on these immunological cells. controls) were not evaluated. Only cytokines and chemokines that gave a difference of more than 1.5-fold are shown. The response against rabbit anti-GBM IgG is quantita- tively and qualitatively different between WT and sdc-1- deficient mice. First, the response against rabbit anti-GBM development of mouse Ig anti-rabbit IgG and subsequent IgG is higher in sdc-1-deficient mice than in WT mice. deposition during the autologous phase may be a driving Furthermore, sdc-1-deficient mice had higher plasma levels force in persistence of albuminuria.27 Therefore, the observed of mouse IgG1 anti-rabbit IgG and lower plasma levels of two-fold higher glomerular deposition of mouse Ig anti- mouse IgG2a anti-rabbit IgG. This indicates a shift from a rabbit IgG in sdc-1-deficient mice may serve as an Th1 response, identified as the major T-cell response in anti- explanation for the higher albuminuria in sdc-1-deficient GBM,26 to a Th2 response in sdc-1-deficient mice. Indeed, mice during the autologous phase. Notably, both WT and we found higher renal levels of Th2 (promoting) cytokines

1212 Kidney International (2007) 72, 1204–1215 AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis original article

IL-1b, IL-4, IL-6, IL-10, IL-13, and monocyte chemotactic purified by protein-A affinity chromatography (Amersham Bio- protein-1 in sdc-1-deficient mice compared with WT mice in sciences, Roosendaal, The Netherlands) and concentrated to 30 mg/ml the heterologous and/or autologous phase. Moreover, Th1 by ultrafiltration with a XM-50 Diaflow membrane (Amicon (promoting/maintaining) cytokines TNF-a, interferon-g, Corporation, Danvers, MA, USA). IL-2, IL-12a, IL-12b, and MIP-1a were lower in the sdc-1-deficient mice compared with the WT mice in the Characterization of rabbit anti-mouse GBM antibodies in heterologous and/or autologous phase. In particular, enzyme-linked immunosorbent assay the severe effects of sdc-1-deficiency on the expression of Purified laminin (EHS; Campro Scientific, Berlin, Germany), agrin,52 HS (Seikagaku, Tokyo, Japan), fibronectin (Bio-Connect monocyte chemotactic protein-1 and MIP-a during the BV, Huissen, The Netherlands), or collagen type IV (Sanbio BV, heterologous phase point to an early skewing from Th1 to 1 47 Uden, The Netherlands) was coated overnight at 4 C in PBS (10 mg/ Th2. This skewing to Th2 has been described previously in well). Wells were washed with PBS/0.05% Tween 20 (PBS/Tween) 48 sdc-1-deficient mice with allergic lung inflammation. and blocked with 2% bovine serum albumin (BSA) (Sigma-Aldrich) However, at this moment, it remains difficult to explain in PBS for 2 h and washed with PBS/Tween. Serial dilutions of rabbit how a shift to a Th2 response leads to aggravation of anti-GBM anti-mouse GBM IgG in 2% BSA/PBS were added and incubated for nephritis, as Th2-induced antibodies in general are less effective 1 h, wells were washed with PBS/Tween and incubated with goat in complement activation than Th1-induced antibodies. anti-rabbit Ig (H þ L) antibody conjugated with horseradish peroxi- In conclusion, sdc-1 regulates the leukocyte adhesion to dase (Jackson ImmunoReseach Laboratories Inc., West Grove, PA, activated glomerular endothelium during the heterologous USA) in 2% BSA/PBS for 1 h. Finally, wells were washed with PBS/ phase of anti-GBM nephritis, either directly by shedding or Tween and incubated with tetramethylbenzidine substrate solution (SFRI Laboratories, Berganton, France). After exactly 15 min, the indirectly via regulation of glomerular endothelial expression reaction was stopped with 2 M H2SO4 and absorption was measured of inflammatory HS domains that promote the firm adhesion at 450 nm. of leukocytes. Furthermore, sdc-1 influences the immune response against rabbit anti-GBM IgG and sdc-1-deficiency Induction of anti-GBM glomerulonephritis in WT and leads to a shift from Th1 to Th2 during anti-GBM nephritis. sdc-1-deficient mice and determination of albuminuria and Overall, sdc-1 seems to have anti-inflammatory properties in creatinine anti-GBM nephritis. WT and sdc-1-deficient mice, 14- to 18-weeks-old, were injected in the tail vein with 8 mg rabbit anti-mouse GBM IgG or 8 mg control MATERIALS AND METHODS rabbit IgG. Mice were killed after 1, 2, 4 and 18 h, 4 and 8 days. Each Animals group comprised of eight mice. Urine was collected directly through C57BL/6 mice (Jackson Laboratories, Bar Habor, ME, USA), sdc-1- bladder puncture or after 18 h in metabolic cages. Collected kidneys deficient mice in a C57BL/6 background,5 and New Zealand white were fixed in 10% buffered formalin and paraffin embedded or snap rabbits were bred, housed, and handled according to the guidelines frozen in liquid nitrogen. Albumin was measured by radial of the local animal ethics committee. Genotyping of mice with gene- immunodiffusion (Mancini). Plasma and urinary creatinine con- specific primers was performed as described previously.49 centrations were determined enzymatically (Roche Diagnostics).

Isolation of mouse GBM RNA isolation and real-time reversed Glomeruli were isolated from C57/BL6 mice perfused with 0.01 M transcription-polymerase chain reaction m phosphate-buffered saline (PBS) pH 7.2 with 1.25% Fe3O4. Total RNA was isolated from 10 m frozen sections of renal cortex Glomeruli were sieved through a 75-mm sieve (Endecotts, London, using the RNeasy Mini protocol including a DNase I step (Qiagen, UK), pelleted by a magnet, washed in PBS, resuspended in excess Venlo, The Netherlands). Reversed transcription of 1 mg RNA was distilled water, and shaken for 2 h at 41C. Subsequently, the performed with Oligo dT primers and the Superscript First-Strand suspension was centrifuged for 15 min at 5910 g at 41C and the Synthesis System (Invitrogen Life Technologies, Breda, The Nether- pellet was sonicated three times for 15 s on ice with 45 s cooling lands). One-tenth of cDNA was used as template in real-time intervals. Finally, the suspension was treated with 3% Triton X-100, polymerase chain reaction using SYBR Green Supermix (Bio-Rad, 2 mg/ml deoxyribonucleaseI (Roche Diagnostics GmbH, Mann- Veenendaal, The Netherlands) with gene-specific primers (10 mM; heim, Germany) and 4% sodium deoxycholate for 18 h, washed Isogen, Ijsselstein, The Netherlands; Table S1) on the MyiQ poly- extensively in distilled water and lyophilized.50,51 merase chain reaction system (Bio-Rad). Expression of the house- keeping gene glyceraldehyde-3-phosphate dehydrogenase was used Immunization of rabbits with mouse GBM and purification to quantify gene expression using the DCt method. of rabbit IgG Rabbits were subcutaneously injected with 1 mg GBM emulsified in Immunofluorescence staining complete Freund’s adjuvant (Sigma-Aldrich Chemie BV, Zwijn- Frozen sections (2 mm) were fixed in ice-cold acetone for 10 min and drecht, The Netherlands), followed by subcutaneous booster incubated with primary antibodies diluted in PBS containing 1% injections of 1 mg GBM in incomplete Freund’s adjuvant after 3, BSA and 0.05% sodium azide (IF-buffer) for 45 min. Directly labeled 6, and 9 weeks. Control rabbit IgG was collected from rabbits that antibodies included goat anti-mouse C3c- and fibrinogen-fluore- were immunized with PBS in complete Freund’s adjuvant/ scein isothiocyanate (FITC) (Nordic, Tilburg, The Netherlands), incomplete Freund’s adjuvant. Pooled sera were centrifuged and goat anti-rabbit IgG Alexa 488 and goat anti-mouse IgG (H þ L), passed through a 0.4-mm filter. Subsequently, IgG fractions were IgG1 and IgG2a Alexa 488 (Invitrogen Life Technologies). Unlabeled

Kidney International (2007) 72, 1204–1215 1213 original article AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis

antibodies included anti-mouse CD31 (BD Biosciences, Alphen Aan Detection of renal cytokines/chemokines by antibody array De Rijn, The Netherlands), the vesicular stomatitis virus (VSV)- RayBio mouse cytokine antibody array C series 1000 (detects 96 tagged anti-HS antibodies AO4B08, EW3D10, and EW4G2 that cytokines, chemokines, and other proteins; http://www.raybiotech. recognize N- and 6-O-sulfated HS domains,53–55 anti-mouse CD4 com/map/mouse_C_Series_1000.pdf) was used according to the (L3T4), CD8a (Ly-2) (BD Biosciences), CD68 (MCA1957; Serotec manufacturer’s instructions (RayBiotech, Norcross, GA, USA). Ltd, Oxford, UK), and digoxigenin-labeled rabbit anti-mouse GBM. Renal extracts were prepared from three pooled half kidneys of Digoxigenin labeling was performed according to the manufacturer’s three separate mice by forcing the tissue through a 70 mm cell instructions (Roche). Sections were incubated with the appropriate strainer (BD Biosciences) in lysis buffer supplied by the manufac- IgG Alexa 488, anti-digoxigenin-FITC (Roche), or anti-VSV-Cy3 turer, which was supplemented with a cocktail of protease inhibitors (Sigma-Aldrich) antibodies in IF-buffer for 45 min. Finally, sections (Roche). Lysates were briefly sonicated on ice and centrifuged for were postfixed with 1% paraformaldehyde-PBS and embedded in 10 min at 15 000 g at 41C. The protein content of the supernatant VectaShield mounting medium H-1000 (Brunschwig Chemie, was determined by the Bicinchoninic acid assay (Sigma). Mem- Amsterdam, The Netherlands). The amount of rabbit IgG or mouse branes were incubated with 250 mg of pooled renal extract for 1 h at Ig, IgG1 and IgG2a deposited along the capillary wall was quantified room temperature. Finally, spots were developed and exposed for by titrating the appropriate detecting antibodies. The titer was different periods to Hyperfilm (GE Healthcare Europe, Diegem, defined as the reciprocal of the dilution, which was still positive and Belgium). The intensity of the spots on the films was quantified expressed as arbitrary units. The staining intensities of all antibodies using Image J software (http://rsb.info.nih.gov/ij). were scored in 50 glomeruli on a scale between 0 and 10 (0 ¼ no staining, 1 ¼ 10% staining intensity along the capillary wall or on Statistical analysis glomerular endothelial cells, with a maximum score of 10 for 100% Values are expressed as means7s.e.m. and significance was staining intensity). The scoring was performed by two independent evaluated by a one-way ANOVA and the post hoc analysis Tukey’s observers on blinded sections. multiple comparison test using GraphPad Prism, version 4.0 software (GraphPad Software Inc., San Diego, CA, USA). For Renal histology comparison between two groups, the student’s t-test was used. Renal sections (5 mm) were deparaffinized, rehydrated, and stained with PAS. Sections were counterstained with hematoxylin. At least 50 glomeruli per mouse were evaluated. ACKNOWLEDGMENTS We thank Dr Peter van Lent, Birgitte Oppers-Walgreen, and Cora Immunohistochemical staining Arndtz for their help with the paraffin embedding of the kidneys. Glomerular sdc-1 and -4 expression and glomerular PMN influx This work was supported by Grant 902-27-292 from The Dutch Organization for Scientific Research (NWO), Grant C05.5152 from the were determined on deparaffinized and rehydrated sections. For Dutch Kidney Foundation and Grant IMF GO¨ 120415 from Mu¨nster syndecan detection, the sections were treated with 0.5 mg/ml University Hospital. protease XXIV (Sigma-Aldrich) for 10 min at 371C, incubated with 0.02 IU/ml heparinase I and III (IBEX Technologies Inc., Montreal, Canada) for 1 h at 371C and blocked with 10% BSA, 1.7% NaCl in SUPPLEMENTARY MATERIAL PBS for 2 h at room temperature. Sections were incubated with rat Table S1. Primers used in real-time polymerase chain reaction. antibodies specific for sdc-1 (CD138; clone 281-2), sdc-4 (clone KY/ Figure S1. Anti-GBM staining is significantly higher in syndecan-1- 8.2; BD Biosciences), or granulocytes (NIMP-R14),56 overnight at deficient mice compared with wild type mice during anti-GBM nephritis due to expansion of GBM matrix. 41C or for 1 h, respectively. Subsequently, sections were incubated with a peroxidase-labeled anti-rat antibody (Dako, Glostrup, Denmark) in 5% normal mouse serum/PBS for 30 min followed REFERENCES by incubation with 1 mg/ml diaminobenzidine (Sigma-Aldrich) in 1. Carey DJ. Syndecans: multifunctional cell-surface co-receptors. Biochem J 1997; 327(Part 1): 1–16. 50 mM Tris-HCl pH 7.6 with 0.001% H2O2 for 10 min. Sections were 2. Bernfield M, Gotte M, Park PW et al. Functions of cell surface heparan counterstained with hematoxylin for 2 min. Two independent sulfate proteoglycans. Annu Rev Biochem 1999; 68: 729–777. observers counted PMNs on blinded sections in at least 50 3. Bernfield M, Kokenyesi R, Kato M et al. Biology of the syndecans: a family glomeruli per animal. of transmembrane heparan sulfate proteoglycans. Annu Rev Cell Biol 1992; 8: 365–393. 4. Kim CW, Goldberger OA, Gallo RL et al. Members of the syndecan Determination of plasma levels of mouse Ig, IgG1 and IgG2a family of heparan sulfate proteoglycans are expressed in distinct anti-rabbit IgG antibodies in enzyme-linked immunosorbent cell-, tissue-, and development-specific patterns. Mol Biol Cell 1994; 5: assay 797–805. Rabbit Ig was coated overnight at 41C (10 mg/well). Wells were 5. Alexander CM, Reichsman F, Hinkes MT et al. Syndecan-1 is required for Wnt-1-induced mammary tumorigenesis in mice. Nat Genet 2000; 25: washed with PBS/Tween and blocked with 2% BSA/PBS for 2 h. 329–332. Serial dilutions of mouse plasmas were added and incubated for 1 h. 6. Gallo R, Kim C, Kokenyesi R et al. Syndecans-1 and -4 are induced during Subsequently, wells were washed with PBS/Tween, incubated with wound repair of neonatal but not fetal skin. J Invest Dermatol 1996; 107: peroxidase-conjugated goat anti-mouse Ig (H þ L; Tebu-Bio, 676–683. 7. Elenius K, Vainio S, Laato M et al. Induced expression of syndecan in Heerhugowaard, The Netherlands), goat anti-mouse IgG1, or healing wounds. J Cell Biol 1991; 114: 585–595. IgG2a antibodies (Sanbio) in 2% BSA/PBS for 1 h. The remainder 8. Stepp MA, Gibson HE, Gala PH et al. Defects in keratinocyte activation of the enzyme-linked immunosorbent assay was performed as during wound healing in the syndecan-1-deficient mouse. J Cell Sci 2002; described above. The titer was defined as the reciprocal of the 115: 4517–4531. 9. Gotte M, Joussen AM, Klein C et al. Role of syndecan-1 in dilution of the plasma giving an optical density of 0.5 and expressed leukocyte–endothelial interactions in the ocular vasculature. Invest as U/ml. Ophthalmol Vis Sci 2002; 43: 1135–1141.

1214 Kidney International (2007) 72, 1204–1215 AL Rops et al.: Syndecan-1 deficiency in anti-GBM nephritis original article

10. Svennevig K, Kolset SO, Bangstad HJ. Increased syndecan-1 in serum is 34. Cornelison DD, Wilcox-Adelman SA, Goetinck PF et al. Essential and related to early nephropathy in type 1 diabetes mellitus patients. separable roles for Syndecan-3 and Syndecan-4 in skeletal muscle Diabetologia 2006; 49: 2214–2216. development and regeneration. Genes Dev 2004; 18: 2231–2236. 11. Gotte M. Syndecans in inflammation. FASEB J 2003; 17: 575–591. 35. Li Q, Park PW, Wilson CL et al. Matrilysin shedding of syndecan-1 12. Rops AL, van der Vlag J, Lensen JF et al. Heparan sulfate proteoglycans in regulates chemokine mobilization and transepithelial efflux of glomerular inflammation. Kidney Int 2004; 65: 768–785. neutrophils in acute lung injury. Cell 2002; 111: 635–646. 13. Esko JD, Selleck SB. Order out of chaos: assembly of ligand binding sites 36. Fitzgerald ML, Wang Z, Park PW et al. Shedding of syndecan-1 in heparan sulfate. Annu Rev Biochem 2002; 71: 435–471. and -4 ectodomains is regulated by multiple signaling pathways and 14. Ledin J, Staatz W, Li JP et al. Heparan sulfate structure in mice with mediated by a TIMP-3-sensitive metalloproteinase. J Cell Biol 2000; 148: genetically modified heparan sulfate production. J Biol Chem 2004; 279: 811–824. 42732–42741. 37. Kainulainen V, Wang H, Schick C et al. Syndecans, heparan sulfate 15. Assmann KJ, Tangelder MM, Lange WP et al. Anti-GBM nephritis in the proteoglycans, maintain the proteolytic balance of acute wound fluids. mouse: severe proteinuria in the heterologous phase. Virchows Arch A J Biol Chem 1998; 273: 11563–11569. Pathol Anat Histopathol 1985; 406: 285–299. 38. Elenius V, Gotte M, Reizes O et al. Inhibition by the soluble syndecan-1 16. Couser WG. Mechanisms of glomerular injury in immune-complex ectodomains delays wound repair in mice overexpressing syndecan-1. disease. Kidney Int 1985; 28: 569–583. J Biol Chem 2004; 279: 41928–41935. 17. Couser WG. Mechanisms of glomerular injury: an overview. Semin Nephrol 39. Brule S, Charnaux N, Sutton A et al. The shedding of syndecan-4 and 1991; 11: 254–258. syndecan-1 from HeLa cells and human primary macrophages is 18. Schrijver G, Assmann KJ, Bogman MJ et al. Antiglomerular basement accelerated by SDF-1/CXCL12 and mediated by the matrix membrane nephritis in the mouse. Study on the role of complement in metalloproteinase-9. Glycobiology 2006; 16: 488–501. the heterologous phase. Lab Invest 1988; 59: 484–491. 40. Chakravarti R, Sapountzi V, Adams JC. Functional role of syndecan-1 19. Sheerin NS, Springall T, Carroll MC et al. Protection against cytoplasmic V region in lamellipodial spreading, actin bundling, and cell anti-glomerular basement membrane (GBM)-mediated nephritis in migration. Mol Biol Cell 2005; 16: 3678–3691. C3- and C4-deficient mice. Clin Exp Immunol 1997; 110: 403–409. 41. Tkachenko E, Rhodes JM, Simons M. Syndecans: new kids on the signaling 20. Schrijver G, Schalkwijk J, Robben JC et al. Antiglomerular basement block. Circ Res 2005; 96: 488–500. membrane nephritis in beige mice. Deficiency of leukocytic neutral 42. Zhang J, Wang Y, Chu Y et al. Agrin is involved in lymphocytes proteinases prevents the induction of albuminuria in the heterologous activation that is mediated by alpha-dystroglycan. FASEB J 2006; 20: phase. J Exp Med 1989; 169: 1435–1448. 50–58. 21. Tang T, Rosenkranz A, Assmann KJ et al. A role for Mac-1 (CDIIb/CD18) in 43. Khan AA, Bose C, Yam LS et al. Physiological regulation of the immune complex-stimulated neutrophil function in vivo: Mac-1 immunological synapse by agrin. Science 2001; 292: 1681–1686. deficiency abrogates sustained Fcgamma receptor-dependent neutrophil 44. Tung JW, Mrazek MD, Yang Y et al. Phenotypically distinct B cell adhesion and complement-dependent proteinuria in acute development pathways map to the three B cell lineages in the mouse. glomerulonephritis. J Exp Med 1997; 186: 1853–1863. Proc Natl Acad Sci USA 2006; 103: 6293–6298. 22. Heeringa P, Van den Born J, Brouwer E et al. Elastase, but not proteinase 3 45. Sanderson RD, Lalor P, Bernfield M. B lymphocytes express and lose (PR3), induces proteinuria associated with loss of glomerular basement syndecan at specific stages of differentiation. Cell Regul 1989; 1: 27–35. membrane heparan sulphate after in vivo renal perfusion in rats. Clin Exp 46. Wegrowski Y, Milard AL, Kotlarz G et al. Cell surface proteoglycan Immunol 1996; 105: 321–329. expression during maturation of human monocytes-derived dendritic 23. Johnson RJ, Couser WG, Alpers CE et al. The human neutrophil serine cells and macrophages. Clin Exp Immunol 2006; 144: 485–493. proteinases, elastase and cathepsin G, can mediate glomerular injury in 47. Karpus WJ, Kennedy KJ. MIP-1alpha and MCP-1 differentially regulate vivo. J Exp Med 1988; 168: 1169–1174. acute and relapsing autoimmune encephalomyelitis as well as Th1/Th2 24. Klebanoff SJ, Kinsella MG, Wight TN. Degradation of endothelial cell lymphocyte differentiation. J Leukoc Biol 1997; 62: 681–687. matrix heparan sulfate proteoglycan by elastase and the 48. Xu J, Park PW, Kheradmand F et al. Endogenous attenuation of allergic myeloperoxidase-H2O2-chloride system. Am J Pathol 1993; 143: 907–917. lung inflammation by syndecan-1. J Immunol 2005; 174: 5758–5765. 25. Unanue ER, Dixon FJ. Experimental glomerulonephritis: immunological 49. Rops AL, van der Vlag J, Jacobs CW et al. Isolation and characterization events and pathogenetic mechanisms. Adv Immunol 1967; 6: 1–90. of conditionally immortalized mouse glomerular endothelial cell lines. 26. Kitching AR, Holdsworth SR, Tipping PG. IFN-gamma mediates crescent Kidney Int 2004; 66: 2193–2201. formation and cell-mediated immune injury in murine 50. Assmann KJ, Tangelder MM, Lange WP et al. Membranous glomerulonephritis. J Am Soc Nephrol 1999; 10: 752–759. glomerulonephritis in the mouse. Kidney Int 1983; 24: 303–312. 27. Nitta K, Horita S, Ogawa S et al. Resistance of CD28-deficient mice to 51. Baelde HJ, Bergijk EC, Bruijn JA. Isolation and characterization of autologous phase of anti-glomerular basement membrane mouse glomerular basement membrane. J Clin Lab Immunol 1990; 33: glomerulonephritis. Clin Exp Nephrol 2003; 7: 104–112. 17–20. 28. Parish CR. Heparan sulfate and inflammation. Nat Immunol 2005; 6: 861–862. 52. Vogtlander NP, Tamboer WP, Bakker MA et al. Reactive oxygen species 29. Wang L, Fuster M, Sriramarao P et al. Endothelial heparan sulfate deglycosilate glomerular alpha-dystroglycan. Kidney Int 2006; 69: deficiency impairs L-selectin- and chemokine-mediated neutrophil 1526–1534. trafficking during inflammatory responses. Nat Immunol 2005; 6: 53. Dennissen MA, Jenniskens GJ, Pieffers M et al. Large, tissue-regulated 902–910. domain diversity of heparan sulfates demonstrated by phage display 30. Rops AL, van den Hoven MJ, Baselmans MM et al. Heparan sulfate antibodies. J Biol Chem 2002; 277: 10982–10986. domains on activated cultured glomerular endothelial cells mediate the 54. van de Westerlo EM, Smetsers TF, Dennissen MA et al. Human single trafficking of leuckocytes. Kidney Int, in press. chain antibodies against heparin: selection, characterization, and effect 31. Fiebeler A, Park JK, Muller DN et al. Growth arrest specific protein 6/Axl on coagulation. Blood 2002; 99: 2427–2433. signaling in human inflammatory renal diseases. Am J Kidney Dis 2004; 43: 55. Lensen JF, Rops AL, Wijnhoven TJ et al. Localization and functional 286–295. characterization of glycosaminoglycan domains in the normal human 32. Schmouder RL, Streiter RM, Walz A et al. Epithelial-derived neutrophil- kidney as revealed by phage display-derived single chain antibodies. activating factor-78 production in human renal tubule epithelial cells J Am Soc Nephrol 2005; 16: 1279–1288. and in renal allograft rejection. Transplantation 1995; 59: 118–124. 56. McLaren DJ, Strath M, Smithers SR. Schistosoma mansoni: evidence that 33. Gotte M, Bernfield M, Joussen AM. Increased leukocyte–endothelial immunity in vaccinated and chronically infected CBA/Ca mice is sensitive interactions in syndecan-1-deficient mice involve heparan sulfate- to treatment with a monoclonal antibody that depletes cutaneous dependent and -independent steps. Curr Eye Res 2005; 30: 417–422. effector cells. Parasite Immunol 1987; 9: 667–682.

Kidney International (2007) 72, 1204–1215 1215