W F1 Mice

Journal of Autoimmunity (1996) 9, 41–50

Heparan Sulfate Staining of the Glomerular Basement Membrane in Relation to Circulating Anti-DNA and Anti-Heparan Sulfate Reactivity: a Longitudinal Study in NZB/W F1 Mice Machteld N. Hylkema1, Mieke C. J. van Bruggen2, Ruud van de Lagemaat1, Kees Kramers2, Jo H. M. Berden2 and Ruud J. T. Smeenk1 1

Department of Autoimmune Diseases, Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (C.L.B.), Amsterdam The Netherlands 2 Division of Nephrology, University Hospital St Radboud, Nijmegen, The Netherlands

Received 18 May 1995 Accepted 15 September 1995 Key words: anti-DNA, heparan sulfate, SLE, nefritis, NZB/W mice

Reactivity of serum antibodies with heparan sulfate (HS) has been associated with human and murine lupus nephritis, although the aetiological significance of this association is not clear. Recent work from our laboratories showed that binding of these antibodies to HS could be mediated by histone containing immune complexes. In human lupus nephritis we found a strong decrease in HS staining in the glomerular basement membrane (GBM). The aim of this study was to elucidate the correlation in experimental systemic lupus erythematosus (SLE) between albuminuria, staining of HS in the GBM and anti-DNA and anti-HS reactivity in plasma. We therefore studied NZB/W F1 mice during different stages of glomerular disease and compared them with age matched control NZB/W F1 mice without albuminuria. Anti-DNA and anti-HS reactivity were measured in longitudinally collected plasma samples and correlated with the onset of albuminuria, staining of HS in the glomerular basement membrane and deposition of immunoglobulins (Ig). HS staining was significantly decreased in the glomerular capillary loops of mice with prolonged proteinuria in comparison with age matched control mice (P=0.0013). This decreased HS staining was correlated with increased Ig deposition in the capillary loops (tau= −0.42, P<0.001), albuminuria (tau= −0.508, P<0.001) and a decrease in anti-DNA levels measured in plasma (tau=0.758, P<0.005). Altered anti-HS reactivity in plasma did correlate with increased Ig deposition in the kidney (tau=0.33, P<0.05) but was not correlated with decreased staining of HS in the kidney. In conclusion, our study demonstrates that disappearance of staining of HS in the glomerular capillary loops is associated with albuminuria, increased Ig deposition in the glomerulus and decreased anti-DNA reactivity in plasma. Our findings are compatible with a model in which interaction (‘masking’) of HS with immune complexes consisting of anti-DNA antibodies and nucleosomes takes place. © 1995 Academic Press Limited

Introduction

ing of anti-DNA antibodies to the glomerular basement membrane (GBM) or in-situ immune complex formation initiated by binding of DNA to the GBM or deposition of circulating anti-DNA immune complexes [1, 4, 5]. Recently, nucleosomes have been assigned a role not only in the aetiology of lupus but also in the pathogenesis of lupus nephritis, for several reasons. Firstly, it was recently shown that nucleosomes are a major immunogen in murine lupus, not only for the induction of anti-nucleosome antibodies but also for the generation of anti-DNA and anti-histone antibodies [6]. Secondly, nucleosomes can be detected in

The pathogenesis of glomerulonephritis associated with systemic lupus erythematosus (SLE) is not completely understood. The most widely accepted explanation is based on deposition of immune complexes in the glomerulus, involving anti-DNA antibodies [1–3]. Several mechanisms have been proposed for this intraglomerular deposition of anti-DNA: direct bindCorrespondence to: Dr R. J. T. Smeenk, Department of Autoimmune Diseases, Central Laboratory of the Red Cross Blood Transfusion Service (C.L.B.), Plesmanlaan 125, NL-1066 CX Amsterdam, The Netherlands. 41 0896-8411/96/010041+10 $18.00/0

© 1996 Academic Press Limited

M. N. Hylkema et al.

42

plasma of SLE patients [7, 8] and thirdly, in lupus mice, the autoantibody response is initially directed against nucleosomal antigens while anti-DNA antibodies are produced later in the disease [9]. In addition, it has been reported that histones, which constitute the nucleosome together with DNA, have a high affinity for the GBM [10]. In view of the above and our in-vivo data, in which we found that anti-DNA antibodies can bind to the GBM after renal perfusion with histone and DNA [11] or anti-nucleosomal antibodies via nucleosomes [12], we currently hold nucleosome containing complexes responsible for the onset of glomerulonephritis [13, 14]. Reactivity of serum antibodies with heparan sulfate (HS) in vitro [15–18] and in vivo [19] has been associated with human and murine lupus nephritis. Recently, we showed that this HS reactivity was mediated by histone containing immune complexes [20, 21]. HS is the glycosaminoglycan side-chain of heparan sulfate proteoglycan (HSPG) which is an intrinsic constituent of the GBM. HS, as a negatively charged molecule, is responsible for the charge dependent permeability of the GBM, but also has an important function in the integrity of its sizeselective barrier. Therefore, binding of immune complexes to HS in the GBM may be an important event in lupus nephritis, by interference with physiological functions and/or by triggering local inflammatory reactions. The aim of this study was to relate anti-DNA and anti-HS reactivity in plasma with onset of albuminuria and staining for HS and immunoglobulin (Ig) deposition in the kidney. We therefore studied NZB/W F1 mice, a murine model for SLE, with different stages of renal disease and compared these with age matched control mice of the same strain without albuminuria. Anti-DNA and anti-HS reactivity in plasma was longitudinally measured in relation to development of nephritis. The deposition of Ig and staining of HS was determined with direct and indirect immunofluorescence techniques and correlated with the magnitude of albuminuria and the anti-DNA and anti-HS reactivity in plasma. Our data support a mechanism wherein anti-DNA/nucleosome immune complexes interfere with HS in the GBM.

Materials and Methods Mice Female NZB/W F1 mice (n=65) were obtained from Harlan CPB, Zeist (The Netherlands) and kept in a conventional environment. Plasma samples were taken from the retroorbital plexus once every two or three weeks, starting at 7–8 weeks of age. The plasma samples were stored at −20°C until use. Weekly measurements of urinary albumin excretions were performed with Albustix (Bayer Diagnostics, Gorinchem, The Netherlands). Based on the magnitude of albuminuria, 31 of these mice were divided into 4 groups for histological examination.

Group 1: age matched control mice (n=11) without albuminuria (Albustix <300 µg/ml). The mice of this group covered all ages of the mice in the other groups. Group 2: mice (n=9) with a period of albuminuria lasting for less than 1 week (Albustix between 1,000 and 3,000 µg/ml). Group 3: mice (n=6) with a period of albuminuria lasting for 1 week to 4 weeks (Albustix >1,000 µg/ml). Group 4: mice (n=5) which had albuminuria lasting for more than 4 weeks to 14 weeks (Albustix >1,000 µg/ml). From all mice of each group, 18-hour urine was collected using a metabolic cage 1 day before the mice were sacrificed. The kidneys were perfused with saline before removal and immediately snap frozen in liquid nitrogen for immunofluorescence (IF). Urinary albumin excretion was measured by single radial immunodiffusion [22]. Results from these measurements correlated well with Albustix findings.

Immunofluorescence techniques In order to study deposition of mouse Ig, direct immunofluorescence was performed on 2 µm cryostat sections of kidneys of the selected mice by incubating these sections with F(ab)2FITC labelled sheep antimouse Ig (Cappel, Organon Technika NV, Turnhout, Belgium) diluted 1:750 in phosphate buffered saline (PBS, 0.01 M sodium phosphate, pH 7.4, 0.14 M NaCl), containing 1% bovine serum albumin (BSA). Staining for the presence of HSPG was performed using indirect IF with a goat anti-human HSPG-core antiserum [23], diluted 1:200. As a secondary antibody, FITC labelled rabbit anti-goat Ig (de Beer Med BV, Hilvarenbeek, The Netherlands), diluted 1:500, was used. HS was stained with a mouse anti-rat HS monoclonal antibody (mAb, JM-403 [24]). Because this mAb was biotinylated, the sections were first treated with an avidin-biotin blocking kit (Vector Laboratories Inc., Burlingame, CA) before incubation with JM-403 in a dilution of 1:60. To develop the sections, FITC labelled extravidin (Sigma, St Louis, USA) diluted 1:400 was used. The anti-human and anti-rat antibodies used in this study all crossreact with mouse HS(PG) constituents. After staining, the sections were embedded in aquamount (BDH Ltd, Poole, UK) and examined with a Zeiss fluorescence microscope. Blinded sections were examined by four independent investigators. Glomeruli were scored for capillary loop staining semi-quantitatively on a 0–4 scale, in which a grade of 4 was given for the brightest possible staining and a grade of 0 in the absence of capillary loop staining. In a control experiment an excellent interobserver variability was established.

Anti-DNA and anti-HS ELISA To measure anti-dsDNA and anti-HS reactivity, we used ELISAs in which we coated photobiotinylated

GBM-heparan sulfate and lupus nephritis

43

10,000

10,000 A

B

a-DNA (units)

10

a-HS (units); Ig (mg/ml)

10

1

100 10 10

50

0

150 200 100 Age (days)

250

1

1 300

10,000

250

1 300

D

100

10 10

a-DNA (units)

100

100

1000 a-HS (units); Ig (mg/ml)

1000 a-DNA (units)

150 200 100 Age (days)

10,000 C

1

50

0

100 10

a-HS (units); Ig (mg/ml)

a-DNA (units)

100

100

1000

a-HS (units); Ig (mg/ml)

100

1000

10

0

50

150 200 100 Age (days)

250

1 300

1

0

50

150 200 100 Age (days)

250

1 300

Figure 1. Representative examples of anti-DNA (e) and anti-HS titers (C) and Ig levels in individual NZB/W F1 mice belonging to the different groups. (A) A group 1, age matched control mouse without albuminuria; (B) a group 2 mouse with a short period of albuminuria; (C) a group 3 mouse and (D) a group 4 mouse, both with prolonged albuminuria. The onset of albuminuria is marked by an arrow. Note that all antibody patterns are drawn to the same scales (left scale for anti-DNA and right scale for anti-HS reactivity and Ig concentration) to emphasize the point that the mice show various levels of antibody responses.

antigens. Both ELISAs were described previously [25, 26]. Briefly, streptavidin (1 µg/ml, Sigma, St Louis, MO) diluted in distilled water was coated overnight to microELISA plates (96 well, Nunc Maxisorb). After washing three times with PBS containing 0.02% Tween 20 and three times with distilled water, the plates were coated overnight with photobiotinylated dsDNA or HS (1 µg/ml in PBS). After washing the plates again, they were blocked with PBS containing 10% normal goat serum (NGS) and 0.02% Tween for 2 h at room temperature. The coated wells were incubated with mouse plasma, serially diluted in PBS containing 0.02% Tween 20 (PT) and 10% NGS for 2 h

at room temperature and washed again. In order to detect bound Ig, the plates were incubated with a rat monoclonal anti-mouse light chain kappa antibody complexed to horseradish peroxidase (CLB-RM-19-E, prepared in our institute), diluted 1:1,000 in PT. The plates were washed again and developed with 3,5,3′,5′-tetramethylbenzidin (Merck) at 100 µg/ml in 0.11 M sodium acetate (pH 5.5) containing 0.003% H2O2. Colour development was stopped after 15 min with 2 M H2SO4, and absorption was measured in a Titertek Multiscan at 450 nm. On each plate serial dilutions of a relevant mouse plasma sample (for anti-dsDNA) or a monoclonal anti-HS antibody

M. N. Hylkema et al.

44

4

100

4000 A

80

2404

60

1606

40

808

20

3

IF score

3202

2

1

10 –18

–10 –14 –4 –2 0 2 Weeks before/after onset albuminuria

4

0

0

1

2

3

4

Group

Figure 3. Staining of HS ( ) and HSPG ( ) and deposition of Ig ( ) (mean±standard deviation) in the glomerular capillary loops of mice of the different groups. Staining of HS and HSPG and Ig deposition was scored semiquantitatively in IF on a 0–4 scale. Only mice with prolonged albuminuria (group 4) showed significantly decreased staining of HS compared with age matched control mice without proteinuria (P<0.002). Staining of HSPG remains unaltered, irrespective of the absence or presence of albuminuria. Regarding Ig deposition, albuminuric mice (mice in groups 2, 3 and 4) showed significantly increased Ig deposition in the capillary loops compared with age matched control mice (group 1), respectively P<0.01, P<0.01 and P<0.001. There were no significant differences for the various parameters between groups 2, 3 and 4.

110 B 88

66

44

22

0 –18

–10 –14 –4 –2 0 2 Weeks before/after onset albuminuria

4

Figure 2. Representative examples of anti-DNA (•) and anti-HS titers (●) and Ig levels (e) (mean±standard deviation) of all NZB/W F1 mice, expressed as units (A) or as percentages of the highest peak reactivity for each individual mouse (B). Timepoint 0 is the onset of albuminuria. Note that the anti-DNA response is depicted on the left scale and the anti-HS reactivity and Ig concentration on the right scale.

(JM-403) as a positive control, arbitrarily set at respectively 5,000 U/ml and 100,000 U/ml, and a normal mouse plasma sample as a negative control were run to correct for inter-assay variation.

ELISA for detection of total mouse Ig in plasma Plates were coated for 2 h at room temperature with a poloclonal goat-anti-mouse Ig antiserum (CLB-GM17, prepared in this institute, diluted 1:1,500 in PBS). After washing three times with PBS, the plates were incubated for 2 h with mouse plasma, serially diluted in PBS containing 0.02% Tween 20 and 0.1% gelatin (PTG). In order to detect bound Ig, the plates were incubated with GM-17 complexed to horseradish peroxidase (CLB-GM-17-E, prepared in this institute,

diluted 1:2,000 in PTG). The plates were then washed and developed in the same way as described for the anti-DNA ELISA. On each plate serial dilutions of normal mouse serum, containing 10 mg/ml Ig, were included to correct for inter-assay variation.

Statistical methods For statistical analysis the Mann–Whitney test was used for comparison of different groups and Kendall’s rank correlation test for paired observations. Values of P<0.05 were considered significant.

Results Anti-DNA and anti-HS reactivity versus Ig concentration in plasma of individual mice Longitudinally collected plasma of all NZB/W F1 mice were tested for anti-DNA and anti-HS reactivity and Ig concentrations. Figure 1 shows examples of the antibody patterns of selected mice of the different groups. An age matched control mouse (Figure 1A, mouse from group 1), which had no sign of albuminuria, shows a significant rise in anti-DNA reactivity and anti-HS reactivity over time. The Ig concentration increases parallel to the anti-DNA response. Figure 1B features a mouse from group 2 showing that the anti-DNA reactivity peaks 2 weeks before the onset of

GBM-heparan sulfate and lupus nephritis

45

Figure 4. Representative examples of staining of HS (A, B and C) and HSPG (D) in the glomerular capillary loops. (A) Age matched control mouse without albuminuria (group 1); (B) mouse with a short period of albuminuria (group 2) and (C) mouse with prolonged albuminuria (group 4). (D) Representative example of staining of HSPG in the GBM of mice of all different groups. Note that HS staining of the capillary loops almost disappeared in mice with prolonged proteinuria; staining of HSPG remains unaltered in all groups. A=500×, B=500×, C=500×, D=500×.

proteinuria. At that point, the Ig concentration also reaches its highest level. The anti-HS reactivity shows an early significant increase which is continued with a more gentle rise. Figures 1C and D show examples of mice with prolonged albuminuria (mice from group 3 and 4). In both figures it is seen that the anti-DNA response peaks just before the onset of albuminuria in contrast to the anti-HS reactivity which reaches its highest point after the onset of albuminuria. The Ig concentration decreases quickly after the onset of proteinuria.

Average levels of anti-DNA and anti-HS reactivity versus Ig concentration in plasma of all mice The mean and standard deviation of Ig concentrations and plasma reactivities for DNA or HS of all NZB/W F1 mice which developed proteinuria are shown in Figure 2A. The mean anti-DNA level rises significantly (P<0.03) from 4 weeks until 2 weeks before the onset of albuminuria and drops (although not significantly) at time zero (point at which the mice develop albuminuria), while anti-HS increases steadily from 14 weeks until 2 weeks before the onset of albuminuria,

and rises significantly (P=0.04) 2 weeks after the onset of albuminuria. The average Ig concentration increases slightly from 14 weeks before the onset of albuminuria to the point where the mice start developing albuminuria and then decreases to the level of young NZB/W F1 mice. Because of large mouse to mouse variability, antiDNA and anti-HS antibody levels were converted from units to percentages of the highest level for each individual mouse, to normalize the titers and correct for the large inter-individual differences. Figure 2B shows that after the normalization it is more evident that anti-DNA peaks before and anti-HS after the onset of albuminuria. The evolution of Ig concentration now shows more clearly that the mice develop the highest Ig concentration during the 4 weeks before the onset of albuminuria with a significant drop (P<0.02) thereafter.

Immunofluorescence studies of the kidney The expression of HS in the capillary loops was studied in kidneys of all mice in the different groups. From Figure 3 it can be derived that HS

46

M. N. Hylkema et al.

Figure 5. Representative examples of Ig deposition of age matched control mice without albuminuria (group 1, A), of mice with a short period of albuminuria (group 2, B) and of mice with prolonged albuminuria (group 4, C). Note that mice from group 1 show only Ig deposition within the mesangium, whereas mice with albuminuria have Ig deposits along the capillary wall. This is most pronounced in mice with prolonged albuminuria. A=500×, B=500×, C=500×.

staining was significantly decreased in mice with prolonged albuminuria (group 4, P<0.002) in comparison with age matched control mice (group 1). This is in contrast to the staining for the core protein of HSPG to which HS is bound, since there was no difference among the different groups (Figures 3 and 4D). Representative examples of staining for HS in kidney sections of mice of group 1, 2 and 4 are shown in Figure 4A–C. Studying deposition of Ig in the glomerular capillary loops, we observed that although in all mice deposition of Ig occurred, comparison of mice having albuminuria (groups 2, 3 and 4) with age matched control mice (group 1) demonstrated a significant increase in Ig deposition. This phenomenon was most pronounced in mice with prolonged albuminuria (group 4, P=0.001). Figure 5 shows representative examples of Ig deposition in group 1 mice (Figure 5A), which is confined to the mesangium, in group 2 mice (Figure 5B) in which Ig depositions in the capillary loops start to appear, and in group 4 mice (Figure 5C) where large amounts of Ig deposition in the capillary loops are seen.

The disappearance of HS correlated with increased Ig deposition in the capillary loops (tau= −0.42, P<0.001, Figure 6) and with albuminuria (tau= −0.508, P<0.001). There was also a correlation between Ig deposits in the glomerular capillary loops and albuminuria (tau=0.361, P<0.025).

Renal histology in comparison with anti-DNA and anti-HS reactivity in plasma Plasma was obtained from mice within the different groups just before their kidneys were taken for IF studies. In these samples, anti-DNA and anti-HS levels as well as Ig concentrations were measured and expressed as percentages of the highest peak reactivity of that individual mouse (Figure 7). As can be seen, the percentage anti-DNA reactivity is significantly lower in mice with prolonged albuminuria (group 4, P=0.001, Figure 7A) in comparison with mice of group 2. This decrease is correlated with an increase in Ig deposition (r=0.53, P<0.025) and a decrease in HS staining in the glomerular capillary loops (r=0.758, P<0.005). Regarding the anti-HS reactivity of the

GBM-heparan sulfate and lupus nephritis

47

4

120 100

3 Percentage

Intensity of HS staining

A

2

80 60 40

1

20 0

1

2 3 Intensity of Ig deposition

4

0

Figure 6. Correlation between decreased staining of GBM-HS and increased Ig deposition in the glomerular capillary loops (tau= −0.42, P<0.001).

2

3

4

2

3

4

2

3

4

120 B

Percentage

100

plasma of the different groups there is also a tendency, thought just not significant, toward decreased anti-HS reactivity in the mice of group 4 (P=0.05, Figure 7B). This decrease in anti-HS reactivity in plasma is also correlated with an increase in Ig deposition (r=0.33, P<0.05) but not with decreased staining of HS in the kidney. Regarding the Ig concentrations in plasma, mice with prolonged albuminuria (group 4) have significantly lower Ig concentrations than mice from group 2 (P=0.001) and this is correlated with increased Ig deposition (tau= −0.57, P<0.005) and with decreased staining of HS (tau=0.47, P<0.05) in the kidney (Figure 7C).

80 60 40 20 0

120 C

Discussion Percentage

In this study, a histological and serological analysis was performed to elucidate the correlation in experimental lupus nephritis between albuminuria, staining of HS in the kidney and anti-DNA and anti-HS reactivity in plasma. Loss or masking of HS may play an important role in the immunopathogenesis of lupus nephritis since HS has a major function in the maintenance of the charge-selective and size-selective barrier of the GBM [27, 28]. Our results show that decreased staining of HS within the GBM is correlated with an increase in urinary albumin excretion and increased Ig deposition in the glomerular capillary loops. These data are in accordance with findings of Van den Born et al. [24]. Using the same anti-HS monoclonal antibody, they observed complete disappearance of HS staining in the GBM in 12 out of 13 renal biopsies of patients with lupus nephritis. In the past, reduction or neutralization of HS or HSassociated anionic sites has been held responsible for proteinuria in congenital nephrotic syndrome [29]. Our data suggest that the same applies to murine lupus nephritis. Whether loss of HS staining is due to ‘masking’ of HS by immunoglobulins or immune complexes or to an actual decrease in HS cannot be concluded from these experiments. A decrease in GBM-HS can be the result of either reduced synthesis

100 80 60 40 20 0

Figure 7. Reactivity against DNA (A) and HS (B) and the amount of Ig (C) in plasma, expressed as percentages of the highest peak reactivity of that individual mouse. Plasma samples were taken at the time kidneys were collected. Note the tendency of decreased Ig concentrations and anti-DNA and anti-HS reactivity in mice with prolonged albuminuria (group 4) in comparison with mice which had a short period of albuminuria (groups 2 and 3), although for anti-HS reactivity this was not significant. Decrease in anti-DNA reactivity of group 4 in comparison to groups 2 and 3 was significant (resp. P=0.001 and P=0.004), but there was no significant difference between groups 3 and 2. With regard to the decrease in Ig concentration, mice from group 4 showed a significant decrease in comparison to mice from group 2 (P=0.001), but not to group 3. There was no significant difference between groups 3 and 2.

M. N. Hylkema et al.

48

or increased degradation. Concerning the latter possibility, it has been described that HS can be released from proteoglycans on endothelial cells during inflammation by the enzyme heparanase [30]. There are, however, several studies in the literature which support the idea that HS is masked. Potential candidates for the masking of HS by Ig are crossreactive anti-DNA antibodies [4, 15, 31] and cationic anti-DNA antibodies [32, 33]. Deposition of anti-DNA/ nucleosome immune complexes in the GBM could also lead to masking of HS since in murine lupus glomerular immune deposits have been described which may contain histones [34]. In addition, histones have recently been identified in biopsies of patients with lupus nephritis [35]. Our data favour the possibility that HS is masked, since the decrease in HS staining in the GBM is correlated with an increase in Ig deposition. Furthermore, decrease in anti-DNA, and to a lesser extent anti-HS reactivity and Ig concentration, in plasma is also correlated with decreased HS staining. These latter data and the observation that the anti-DNA response rises significantly just before the onset of albuminuria suggests a pathogenic role for anti-DNA antibodies in lupus nephritis which is in agreement with studies done by Swaak et al. [36, 37] and Terborg et al. [38]. The finding that the decrease in anti-HS reactivity in plasma in this study is not correlated with the decrease in anti-DNA in plasma may be explained by the possibility that anti-HS reactivity is partly caused by nucleosome containing immune complexes formed by other autoantibodies than antiDNA, for instance anti-nucleosome or anti-histone antibodies. Another explanation for the finding that the pattern of anti-HS reactivity in plasma seems not to correlate with anti-DNA reactivity could be that after onset of nephritis, release of nucleosomes, because of the renal inflammation, leads to the formation of anti-DNA/nucleosome immune complexes [7] which are both HS and DNA reactive. This means that after onset of nephritis the anti-HS titer rises whereas there is no significant alteration of anti-DNA titer. The findings in these mice are in line with recent work from our laboratory performed on human lupus nephritis in which we found that anti-HS reactivity is higher during disease exacerbations with renal manifestations [18, 25]. A discrepancy with studies on human lupus nephritis is that in patients, anti-DNA and anti-HS reactivity always decrease significantly and concomitantly after the onset of nephritis, which was not observed in our murine study. The most logical explanation for this is that the significant decline in antibody levels is due to immunosuppressive treatment. Studying the average anti-DNA and anti-HS response of all NZB/W F1 mice, large mouse to mouse variability was observed, resulting in high standard deviations. Despite being inbred, these mice displayed large differences in antibody titers and the age of onset of albuminuria. This high variability among mice is not unique to the NZB/W F1 strain since such a phenomenon has also been described for

two other murine SLE models (MRL/lpr and BXSB, [9]). The observation that staining of the core protein of HS-proteoglycan (HSPG) was not altered in kidneys of mice with prolonged proteinuria suggests that Ig or immune complexes bind primarily to the negatively charged side chain of HSPG, namely HS. These findings are in line with experiments recently done by Kashihara et al. [39] in which they administered a poloclonal antibody to the core protein of HSPG to NZB/W F1 mice. They observed no competition for the binding to the GBM by the endogenous anti-DNA antibodies. This underlines that anti-DNA antibodies bind to different parts of the molecule. Still, animals injected with anti-HSPG developed proteinuria earlier than control NZB/W F1 mice, suggesting an additional physiological role for HSPG-core protein in sustaining the glomerular permselectivity. Taken together, these data in our view favour the concept that anti-DNA antibodies do interact with HS in the GBM during lupus nephritis. This binding could well be mediated by nucleosomal material complexed to the anti-DNA antibody.

Acknowledgements The authors thank Irma Molenaar and Theo Jansen Hendriks for excellent technical assistance and Drs Aarden and Meilof for stimulating discussions. This work was financially supported by the Netherlands Rheumatism Foundation (grant 90/CR/287) and the Dutch Kidney Foundation (C91.1081). Dr J. H. M. Berden was supported by the EC Biomed I program (BMH1-CT92-1766). The anti-HS mAb JM-403 was a kind gift of Dr J. van den Born, Division of Nephrology, University Hospital Nijmegen, The Netherlands.

References 1. Koffler D., Agnello V., Thoburn R., Kunkel H.G. 1971. Systemic Lupus Erythematosus: prototype of immune complex nephritis in man. J. Exp. Med. 134: 169S–179S 2. Izui S., Lambert P.H., Miescher P.A. 1976. In vitro demonstration of a particular affinity of glomerular basement membrane and collagen: a possible basis for a local formation of DNA/anti-DNA complexes in Systemic Lupus Erythematosus. J. Exp. Med. 144: 428–441 3. Mannik M. 1992. Immune complexes. In Systemic Lupus Erythematosus. R.G. Lahita, ed. John Wiley, New York. pp. 327–341 4. Raz E., Brezis M., Rosenmann E., Eilat D. 1989. Anti-DNA antibodies bind directly to renal antigens and induce kidney dysfunction in the isolated perfused rat kidney. J. Immunol. 142: 3076–3082 5. Couser W.G., Salant D.J. 1980. In situ immune complex formation and glomerular injury. Kidney Int. 17: 1–13 6. Mohan C., Adams S., Stanik V., Datta S.K. 1993. Nucleosome: a major immunogen for pathogenic autoantibody-inducing T cells of lupus. J. Exp. Med. 177: 1367–1381 7. Fournié G.J. 1988. Circulating DNA and lupus nephritis. Kidney Int. 33: 487–497

GBM-heparan sulfate and lupus nephritis

8. Rumore P.M., Steinman C.R. 1990. Endogenous circulating DNA in systemic lupus erythematosus. Occurrence as multimeric complexes bound to histone. J. Clin. Invest. 86: 69–74 9. Burlingame R.W., Rubin R.L., Balderas R.S., Theofilopoulos A.N. 1993. Genesis and evolution of antichromatin autoantibodies in murine lupus implicates T-dependent immunization with self antigen. J. Clin. Invest. 91: 1687–1696 10. Schmiedeke T.M.J., Stöckl F.W., Weber R., Sugisaki Y., Batsford S.R., Vogt A. 1989. Histones have high affinity for the glomerular basement membrane. Relevance for immune complex formation in lupus nephritis. J. Exp. Med. 169: 1879–1894 11. Termaat R.M., Assmann K.J.M., Dijkman H.B.P.M., Van Gompel F., Smeenk R., Berden J.H.M. 1992. Anti-DNA antibodies can bind to the glomerulus via two distinct mechanisms. Kidney Int. 42: 1363–1371 12. Kramers C., Hylkema M.N., Van Bruggen M.C.J., Van de Lagemaat R., Dijkman H.B.P.M., Assmann K.J.M., Smeenk R.J.T., Berden J.H.M. 1994. Anti-nucleosome antibodies complexed to nucleosomal antigens show anti-DNA reactivity and bind to rat glomerular basement membrane in vivo. J. Clin. Invest. 94: 568–577 13. Brinkman K., Termaat R.-M., Berden J.H.M., Smeenk R.J.T. 1990. Anti-DNA antibodies and lupus nephritis: the complexity of crossreactivity. Immunol. Today 11: 232–234 14. Kramers C., Hylkema M.N., Termaat R.-M., Brinkman K., Smeenk R.J.T., Berden J.H.M. 1993. Histones in lupus nephritis. Exp. Nephrol. 1: 224–228 15. Eilat D. 1986. Cross-reactions of anti-DNA antibodies and the central dogma of lupus nephritis. Immunol. Today 6: 123–127 16. Faaber P., Rijke G.P.M., Van de Putte L.B.A., Capel P.J.A., Berden J.H.M. 1986. Crossreactivity of human and murine anti-DNA antibodies with heparan sulphate: the major glycosaminoglycan in glomerular basement membrane. J. Clin. Invest. 77: 1824–1830 17. Termaat R.M., Brinkman K., Nossent J.C., Swaak A.J.G., Smeenk R.J.T., Berden J.H.M. 1990. Anti-heparan sulphate reactivity in sera from patients with systemic lupus erythematosus with renal or non-renal manifestations. Clin. Exp. Immunol. 82: 268–274 18. Kramers C., Termaat R.M., ter Borg E.J., Van Bruggen M.C.J., Kallenberg C.G.M., Berden J.H.M. 1993. Higher anti-heparan sulphate reactivity during systemic lupus erythematosus (SLE) disease exacerbations with renal manifestations; A long term prospective analysis. Clin. Exp. Immunol. 93: 34–38 19. Neparstek Y., Ben-Yehuda A., Madaio M.P., Bar-Tana R., Schuger L., Pizov G., Neeman Z., Cohen I.R. 1990. Binding of anti-DNA antibodies and inhibition of glomerulonephritis in MRL-lpr/lpr mice by heparin. Arthritis Rheum. 33: 1554–1559 20. Termaat R.-M., Brinkman K., van Gompel F., van de Heuvel L.P.W.J., Veerkamp J.H., Smeenk R.J.T., Berden J.H.M. 1990. Crossreactivity of monoclonal anti-DNA antibodies with heparan sulfate is mediated via bound DNA/histone complexes. J. Autoimmunity 3: 531–538 21. Hylkema M.N., van der Zwet I., van de Lagemaat R., Kramers C., van Bruggen M.C.J., Swaak A.J.G., Berden J.H.M., Smeenk R.J.T. 1995. No evidence for an independent role of anti-heparan sulfate reactivity

49

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

apart from anti-DNA in lupus nephritis. Clin. Exp. Immunol. 101: 55–65 Mancini G., Carbonara A.O., Heremans J.F. 1965. Immunochemical quanti- tation of antigens by single radial immunodiffusion. Immunochemistry 2: 235–254 van den Heuvel, L.P.W.J., van den Born J., van de Velden T.J.A.M., Veerkamp J.H., Monnens L.A.H., Schroder C.H., Berden J.H.M. 1990. Isolation and partial characterization of heparan sulfate proteoglycan from human glomerular basement membrane. Biochem. J. 264: 457–465 van den Born J., van den Heuvel L.P.W.J., Bakker M.A.H., Veerkamp J.H., Assman K.J.M., Weening J.J., Berden J.H.M. 1993. Distribution of GBM heparan sulfate proteoglycan core protein and side chains in human glomerular diseases. Kidney Int. 43: 454–463 Hylkema M.N., Huygen H., Kramers C., van der Wal Th.J., de Jong J., van Bruggen M.C.J., Swaak A.J.G., Berden J.H.M., Smeenk R.J.T. 1994. Clinical evaluation of a modified ELISA, using photobiotinylated DNA, for the detection of anti-DNA antibodies. J. Immunol. Methods 170: 93–102 Hylkema M.N., Kramers C., Van der Wal T.J., Van Bruggen M.C.J., Swaak A.J.G., Berden J.H.M., Smeenk R.J.T. 1994. A new ELISA for the detection of anti-heparan sulfate reactivity, using photobiotinylated antigen. J. Immunol. Methods 176: 33–43 Kanwar Y.S., Linker A., Farquhar M.G. 1980. Increased permeability of the glomerular basement membrane to ferritin after removal of glycosaminoglycans (heparan sulphate) by enzyme digestion. J. Cell. Biol. 86: 688 Deen W.M. 1982. The glomerular barrier to macromolecules: theoretical and experimental considerations. In Nephrotic Syndrome. Contemporary issues in nephrology. B. M. Brenner and J. H. Stein, eds. Churchill Livingstone, New York. Vernier R.L., Klein D.J., Sisson S.P., Mahan M.D., Oegama T.R., Brown D.M. 1983. Heparan sulphate-rich anionic sites in human glomerular basement membrane: decreased concentration in congenital nephritic syndrome. N. Engl. J. Med. 309: 1001 Ihrcke N.S., Wrenshall L.E., Lindman B.J., Platt J.L. 1993. Role of heparan sulfate in immune system–blood vessel interactions. Immunol. Today 14: 500–505 Madaio M.P., Carlson J., Cataldo J., Ucci A., Migliorini P., Pankewycz O. 1987. Murine monoclonal anti-DNA antibodies bind directly to glomerular antigens and form immune deposits. J. Immunol. 138: 2883–2889 Suzuki N., Harada T., Mizushima Y., Sakane T. 1993. Possible pathogenic role of cationic anti-DNA autoantibodies in the development of nephritis in patients with systemic lupus erythematosus. J. Immunol. 151: 1128–1136 Tsao B.P., Ebling F.M., Roman C., Panosian-Sahakian N., Calame K., Hahn B.H. 1990. Structural characteristics of the variable regions of immunoglobulin genes encoding a pathogenic autoantibody in murine lupus. J. Clin. Invest. 85: 530–540 Schmiedeke T., Stoeckl F., Muller S., Sugisaki Y., Batsford S., Woitas R., Vogt A. 1992. Glomerular immune deposits in murine lupus models may contain histones. Clin. Exp. Immunol. 90: 453–458 Stöckl F., Muller S., Batsford S., Schmiedeke T., Waldherr R., Andrassy K., Sugisaki Y., Nakabayashi K.,

50

Nagasawa T., Rodriguez-Iturbe B., Donini U., Vogt A. 1994. A role for histones and ubiquitin in lupus nephritis. Clin. Nephrol. 41: 10–17 36. Swaak A.J.G., Aarden L.A., Statius van Eps L.W., Feltkamp T.E.W. 1979. Anti-dsDNA and complement profiles as prognostic guides in systemic lupus erythematosus. Arthritis Rheum. 22: 226–235 37. Swaak A.J.G., Groenwold J., Aarden L.A., Statius van Eps L.W., Feltkamp T.E.W. 1982. Prognostic value of anti-dsDNA in SLE. Ann. Rheum. Dis. 41: 388–395

M. N. Hylkema et al.

38. ter Borg E.J., Horst G., Hummel E.J., Limburg P.C., Kallenberg C.G.M. 1990. Measurement of increases in anti-double-stranded DNA antibody levels as a predictor of disease exacerbation in systemic lupus erythematosus: A long-term, prospective study. Arthritis Rheum. 33: 634–654 39. Kashihara N., Makino H., Szekanecz Z., Waltenbaugh C.R., Kanwar Y.S. 1992. Nephritogenicity of anti-proteoglycan antibodies in experimental murine lupus nephritis. Lab. Invest. 67: 752–760