Abrogation of mercuric chloride-induced nephritis in the Brown Norway rat by treatment with antibodies against TNFalpha.

HgCl(2) induces an autoimmune disease in the Brown Norway rat characterized by synthesis of autoantibodies (mainly, anti-GBM Abs), severe proteinuria and interstitial nephritis. Also, HgCl(2)- injected rats develop glomerular cell infiltrates consisting of ED1(+) cells (monocyte/macrophage), starting on day 4 and reaching a maximum on day 8. Treatment with anti-TNF-alpha antiserum had preventative effects as it reduced the urinary protein levels to close to the normal range and also blocked the influx of inflammatory cells in the renal glomeruli and interstitium, but circulating anti-GBM and lineal glomerular IgG deposits were unmodified. In addition, whole isolated glomeruli from HgCl(2)-induced nephritis secreted TNF-alpha commencing on day 8, being maximally detected on day 11 and preceding, between 2 to 3 days, the development of proteinuria. The administration of anti-TNF-alpha antiserum or anti-alpha4 integrin mAb completely abrogated the synthesis of TNF-alpha in glomeruli isolated from the respective treated groups of animals, in addition to the proteinuria. Taken together our results confirm that TNF-alpha plays an important role in the induction and development of HgCl(2)-induced nephritis and highlights the pathogenic importance of the local release of TNF in those renal diseases in which prominent glomerular macrophage accumulation is a constant feature.


Introduction
Mercuric chloride (HgCl2) induces an autoimmune disease in the Brown Norway (BN) rat mediated by T-dependent polyclonal B-cell activation 2 resulting in hypergammaglobulinaemia, synthesis of anti-nuclear and anti-glomerular basement membrane (GBM) autoantibodies as well as development of nephritis with glomerular [1][2][3] lineal deposits of Ig and proteinuria.-The h'stological renal lesions consist of a transient influx of mononuclear cells (mainly MHC class IIbeating T-lymphocytes and monocytes) into the renal interstitium and monocytes and CD8 + T lymphocytes into the glomeruli. 4 It has been shown recently that interaction between lymphocytes and endothelial cells (EC) is crucial in the development of this renal disease. 5 Treatment with Abs against the cz4integrin abrogated the development of interstitial nephritis and virtually abolished the anti-GBM Abs production. Consequently, glomerular deposition of anti-GBM Abs was absent and proteinuria was reduced to levels close to the normal range. In contrast, anti-DNA Abs synthesis was unaffected by this treatment, suggesting a selective immunosuppressive role in the anti-z4integrin Ab.
The role of lymphokines in leukocyte recruitment to inflammatory sites has been well documented. The increased expression of counterreceptors for leukocyte adhesion proteins, such as intercellular cell adhesion molecule (ICAM-1), endothelial cell adhesion molecule (ELAM-1) and vascular cell adhesion molecule (VCAM-1) on EC, has been identified in sites of inflammation, 67 including target organs for autoimmunity. Cytokines (especially TNF-, IL-1 and IFN-3,) activate endothelial cells to synthesize and to increase the expression of adhesion molecules and conversely, engagement of these molecules on the surface of such cells can induce and/or mediate cytokine expression. 7'8 Recently, treatment of autoimmune diseases with cytokine agonists and their antagonists has been attempted in order to establish their clinical usefulness in the prevention of these disorders. Inflammatory mediators, such as TNF-cz and IL-13, have been extensively investigated in autoimmune diseases Experimental procedure: Three groups of rats and their potential pro-inflammatory effects were injected s.c., three times a week, over a demonstrated by increasing the severity of the period of 2 weeks with 100 },tg of HgCI2 per glomerular injury in both rats and rabbits in the 100 g/body vv't. 28 Animals from Group I (n 20) anti-GBM antibody mediated model of nephridid not receive any additional treatment. Rats tis. 9'1 In addition, in the renal cortices of lupus belonging to Group II (n 20) received an i.p. nephritic mice an enhancement of mRNA for injection (25000 IU) of anti-TNF-a on days 0, 8 TNF-a and IL-113 has been observed. 1 Also, an and 13 and those from Group III (n 20) augmented production of TNF has been demonreceived an i.p. injection (0.5 mg) of HP2/1 mAb strated in the nephritic glomeruli in anti-GBM on days 0, 8 and 13. A fourth group (Group IV) glomerulonephritis (GN). 12 Moreover, it has (n 20) served as a normal control in which been shown that the administration of anti-TNF rats were injected only with H20 adjusted to the antiserum abrogated the development of nephrosame pH (3.8) as the HgCl2 solution, following 13 15 th r toxic nephritis in the rat.
Taken toge e, the same procedure described above for the these data suggest that TNF-a and IL-113 can be mercury administration. The dosages and days regarded as potential mediators of glomerular were established on the basis of previous optiinjury, playing an important role in the induction mizing experiments and our appreciation of the and development of renal autoimmune diseases, kinetics of the disease. 5 All animals were sequen-The presence of both TNF-a and VI-a4-integtially bled on different days of the experiment by rin has proved to be necessary for the developtail artery puncture. Four rats from each group ment of proteinuria and accumulation of were killed on days 4, 8, 11, 13 and 23 and inflammatory cells in other rat models of nephrikidneys were processed for study. tis, 5 unlike the requirements found in other models of renal diseases where complement and Proteinuria: Rats were maintained in metabolic neutrophil-induced vascular injury are induced by cages for 24h with free access to food and deposition of immune complexes.
water. Urine samples were taken at regular inter-Using the experimental model of autoimmune vals starting on day 0. Proteinuria was measured renal disease described in this paper, we have by using a Bio-Rad assay (Bio-Rad, Richmond, examined in depth the required presence of CA), according to the manufacturer's protocol.
TNF-a for circulating leukocyte accumulation (in Urine samples were assayed in triplicate. The OD both glomerular and renal interstitium)as well as from each sample was measured in a Titertek in the development of proteinuria.
Anti-GBM and anti-ssDNA Abs assay.. Rat GBM M,ri nd ,hd was isolated, essentially, as described by Bowman et a/. 29 Briefly, glomeruli were obtained from Animals: Brown Norway rats, weighing 180 to normal BN rats by differential sieving and cen-200g, were obtained from IFFA-CREDO (Paris, trifugation of minced kidney cortices. The glo-France) and from our own breeding colony, and merular suspension was sonicated, washed and maintained under standard conditions, lyophilized. The GBM was digested with Type I collagenase (Sigma Chemical Co., St Louis, MO) Antibodies: Rabbit anti-human TNFpolyclonal at 0.7% w/w for I h at 37C. Anti-GBM Abs were antibody, which binds to human TNFand rat measured by ELISA as described previously. 2 TNF-a, 17 was obtained from Genzyme Co. (Cam-Anti-ssDNA Abs were measured by an ELISA bridge, MA). The mouse anti-human HP2/1 mAb developed in our laboratory. 5 All the samples recognizes the VLA-a4-integrin 8 and cross-reacts were assayed in quadruplicate. with the rat VI-4. 9 This mAb blocks the inter-Samples of a serum pool from untreated BN action of both a4131 and t47 integrins with its rats and from BN rats which were treated with 20 21 two known ligands, FN and VCAM-1. The fol-HgCl2 and bled on day 13 of the disease served lowing antibodies were purchased from Serotec as negative and positive controls, respectively.
(Oxford, UK): OX19, which recognizes the CD5 Results were expressed as the percentage of antigen; 22'23 W3/25, which reacts with a subset of binding obtained with samples from positive peripheral T cells expressing the CD4 antigen; 24 control serum.
OX8, which recognizes a stable determinant of the rat CD8 heterodimer; 25 OX6 mAb, which Isolation and culture of glomeruli: Harvested recognizes the rat MHC class II antigens (RT1kidneys from the four experimental groups of B);iand ED1 mAb, which reacts with rat monorats were decapsulated and the renal medulla cytes and macrophages. 2r removed. The dissected cortex was then minced with a razor blade in Hanks' balanced salt solution (HBSS) and seived through decreasing pore size (250, 150 and 751.tm) as described previously. Glomeruli were finally collected on the top of a 75 I.tm sieve, with the suspension containing more than 95% of glomeruli free of tubular fragments. Preparations were suspended at a final concentration of 5 000 glomeruli/ml in RPMI 1640 (Whittaker Bioproducts, Walkerville, MD) containing 10% (v/v) FCS (Flow), ampicillin (100 l.tg/ml) and streptomycin (100 l.tg/ml), and then incubated in 24-well plates (Nunc, Roskilde, Denmark) (5000 glomeruli/well) at 37C in a 5% CO2 atmosphere. Supernatants were harvested at 24h, centrifuged at 100g and stored at -20C until assayed for the release of TNF.
All samples were screened for the presence of endotoxin by using the Limulus amoebocyte lysate (L.A.L.) assay (Whittaker). Briefly, dilutions were performed from standard endotoxin (0.5 to 0.03 E.U./ml, 1 E.U. 0.1 ng/ml of endotoxin) as positive control as well as for our samples (using serial dilutions, 1:2 to 1:64). Non-pyrogenic water was used as negative control. All samples were assayed in quadruplicate. One hundred l.tl of L.A.L. (sensibility, 0.1 E.U./ml) was added to each sample and then were incubated at 37C for i h.
The endotoxin concentration was obtained as the product of the lysate sensibility per the 'limit point dilution' (the last dilution showing reaction with the L.A.L.). The limit of detection in the assay was 10 pg/ml, with all the samples screened showing less than 10pg/ml. TNF<, assay: The TNF-a activity was determined by standard MTT method on L929 cells (ATCC, Rockville, MD) in the presence of actinomycin D (ll.tg/ml). Briefly, 3 x 10 4 cells/well were resuspended in DMEM (Whittaker) containing 10% (v/v) FCS, ampicillin (100 l.tg/ml), streptomycin (100 l.tg/ml), and glutamine (10mM) and further incubated overnight on fiat-bottomed, 96well plates (Nunc) at 37C. Dilutions of supernatants from cultured glomeruli were added in triplicate. Also, half-los dilution ,of hrTNF (specific activity, 2 x 10 to 2 10 9 units/ml) (Genzyme) ranging from 200t.tg/ml to i pg/ml was added to some wells to produce a standard curve. A 251.tl aliquot of MTT stock solution (5 mg/ml) was added to each of the wells and incubated for 4 h at 37C. The supernatants were removed by careful aspiration and 200 I.tl of a 1:1 DMSO and ethanol mixture to solubilize the crystals was then added to each well. The plates were shaken gently and OD (to 595nm) was determined using a Titertek Multiskan Plus.
For neutralization studies, anti-TNF-a Ab or control sera were added simultaneously to wells containing 400pg/ml of hrTNF-a (producing 50% lysis of 3 x 104 L929 cells, in our experiments) or test samples, with a final concentration of 50% (v/v). This Ab neutralizes 2000U of rTNF-a. 31 No TNF-a activity was found on L929 cells in those wells incubated with anti-TNF-a Abs.
Concentrations of TNF-a were extrapolated from a standard curve with known hrTNF-a dilutions. Results are expressed as pg/ml of TNF.
Kidney tissue processing: On days 13 and 23, rat kidneys (n 4) from the groups I, II and IV were processed for histological and immunohistochemistry studies. For light microscopy, 3 l.tm paraffin-embedded kidney sections were stained with periodic acid-Schiffs (PAS). For immunohistochemistry studies, pieces of renal tissue were snap-frozen in isopentane precooled in liquid nitrogen, and stored at -70C until used. Direct immunofluorescence studies were performed on ether/ethanol-fixed serial cryostat sections, by using FITC-conjugated rabbit anti-rat IgG (Serotec), as described previously. 2 In addition, the glomerular cell infiltrates were characterized in the kidneys of rats (n 4) injected with mercury and in two control animals on days 4, 8, 13 and 23. These tissue kidneys were stained with an indirect immunophosphatase method (APAAP) using a panel of mAbs. The specificity of these mAbs was assessed by using normal serum, normal mouse IgG, and hybridomainduced ascitic fluids containing unrelated Abs. Positive controls of the reagents were sections of normal rat spleen. These studies were performed by using a conventional light microscopy objective ( x 63), as described previously. 34 Statistical analysis: The results are given as mean _ _ _ S.D. Values obtained from the levels of proteinuria and ELISA results were analysed using the Student's t-test. For TNF assay, statistical analysis was performed by using a Wilcoxon rank-sum method for non-parametric significance testing.

Results
Effect of anti-TNF<, treatment on the proteinuria: As shown in Fig. 1, rats belonging to the Group I developed proteinuria in two different phases. A first short phase, which occurred immediately after the first injection of HgCl2, due to the direct effect of mercury on tubular renal cells. This first phase was followed by a second phase starting on day 11   integrin Abs (HP2/1 mAb). 5 In both situations, the first phase of proteinuria was unaffected.
Effect of anti-TNF<x treatment on anti-GBM and anti-ssDNA synthesis.. Increased anti-GBM Ab concentration in the serum from rats injected with merculT (Group I, Fig. 2A) was detected by ELISA from day 8, with the mammal concentration being observed on day 13 of the disease. After day 13, serum levels of anti-GBM Abs started to decline as also occurred with the proteinuria. The serum levels of anti-ssDNA Abs in this same Group I (Fig. 2B) showed a significant increase, first observed on day 4, but their kinetics of secretion were different from the anti-GBM Abs production. In contrast to that found in BN rats treated with HP2/1 mAbs, which showed a significant (p<0.001) reduction in the anti-GBM Abs serum levels, 5 the group of rats treated with anti-TNF-z Abs (Group II) presented circulating anti-GBM Abs levels similar to those found in Group I (HgCl2-injected rats) ( Fig. 2A). Histopathology and immunofluorescence studies on kidneys: Renal tissues from rats treated with I'gCl2 (Group I) and HgCl2 plus anti-TNF-a Abs (Group II) (n 4 rats/group), were examined by light microscopy on days 13 and 23 of the disease (Fig. 3A and B). As reported previously, '5 kidney tissues from BN rats injected with mercury presented a severe interstitial mononuclear cell infiltrate. The inflammatory cells were preferentially located in the perivascular regions of the renal interstitium. However, in the same group of rats treated with anti-TNF-a Abs, cell infiltrates were not observed in the renal interstitium (Fig. 3B). Kidneys from HiOinjected rats showed a normal renal histology.
Phenotypic analysis of glomerular cell infiltrates are given in Table 1. We found an impor-   infiltrates, also bear the Ia + antigen cell marker (Table 1). On the other hand, when rat kidney tissues were examined by direct immunofluorescence studies, rat IgG showed a positive linear pattern deposition along the GBM in both HgCli-injected and anti-TNF<z-treated rats (but not control rats), at day 13 of the disease (Fig. 3C and D).
Kinetics of TNF<x secretion by culturing glomeruli: TNF-ez concentration from glomerular cultured supernatants was measured after 24h incubation by using a bioassay. As shown in Fig. 4A, the TNF-z levels in supernatants from cultured HgCla-injected rat (Group I) glomeruli started to increase before day 8 of the disease, reached a peak on day 11 (p < 0.01), and then it declined by day 12. Figure 4A illustrates both the kinetics of TNF-ez production and the urinary protein excretion. Interestingly, TNF-0t production preceded the development of proteinuria by 2-3 days. When TNF<z production was analysed in glomeruli isolated from rats belonging to both Groups II (anti-TNF<z treated) and III (anti<z4 treated), no significant levels of this cytokine was detected ( Fig. 4B and C), in accordance with the low urinary protein levels.

Discussion
The autoimmune disease induced in the BN rats by the injection of HgC12 is characterized by the synthesis of autoantibodies (mainly, anti-GBM Abs) due to a polyclonal B cell activation. The renal lesions consist of rat Ig deposition on the glomerular basement membrane with an influx of mononuclear cells and the development of proteinuria. [1][2][3][4] The administration of anti-TNF-ot antiserum to HgCli-injected rats significantly reduced the urinary protein levels but the synthesis of anti-GBM Abs as well as glomerular deposits of Ig remained unaffected. In addition, no infiltrating interstitial cells were found in this group of rats. These findings suggest that TNF<z plays an important role in the development of this renal injury, as has been demonstrated previously in other different experimental models of nephritis. 14 '5 The elevated serum production of autoantibodies observed after anti-TNF-a treatment strongly suggests that this antiserum acts mainly at a local level. More evidence supporting this local anti-TNF<, Abs effect was obtained when we studied the production of TNF-a from isolated nephritic glomeruli. The glomeruli from HgC12-treated rats secreted TNF-a between days 4-8 to the day 11 of the disease. Also, the studies carried out on the glomerular cell infiltrates in HgCli-injected rats demonstrated an influx of monocytes on day 4 of the disease with the highest number of infiltrating glomerular cells having been observed on day 8. On the other hand, an increase in the number of CD8 + lymphocytes was seen in the nephritic glomeruli on day 13. Moreover, the treatment with the anti-a4 integrin mAb to HgCli-injected rats completely abrogated the secretion of TNF-0t from rat glomeruli. The 0t4 integrin, an adhesion molecule expressed by almost all leukocytes, interacts with VCAM-1 (an endothelial pro-inflammatory inducible cell molecule), as well as with the alternative spliced form (CS-1) of FN, playing a central role in mediating leukocyte adhesion, extravasation, and migration to sites of inflammation. 5 '36 As has been previously demonstrated, 5 the administration of anti-a4 integrin mAb to HgCli-injected rats blocked the influx of circulating leukocytes into the renal interstitium and the accumulation of monocytes in the renal glomeruli. These data might suggest that infiltrating glomerular monocytes are the major source of glomerular TNF. The results presented here are in agreement with those previously reported by Tipping et al. 12 demonstrating the association between the glomerular monocyte infiltration and glomerular injury in anti-GBM nephritis. Nevertheless, the initial source of TNF in the glomerulus is still not entirely understood. It is well documented that resident glomerular cells can be responsible for TNF secretion. The stimulation of mesangial cells with LPS induces the synthesis of TNF-ot, 7 and the in vivo administration of LPS also induces glomerular TNF-a mRNA expression in the absence of leukocyte cells infiltration. 8 Also, it has been shown that stimulated resident glomerular macrophages have the faculty to release cytokines such as IL-8, GM-CSF and TNF-a, among others. 39 It is well known that inorganic mercury remains a major environmental toxin that alters cell calcium homeostasis 4 and mitochondrial functions. 41 In concentrations commonly used in experimental models the Hg 2+ acts as an ionophore as well as Cu+. 42 '3 In addition, the administration of HgC12 causes the activation of circulating lymphocytes, resulting in a polyclonal B-cell activation 2 and also producing a direct toxic effect on tubular renal cells. It is possible that mercury also has the potential to induce the activation of resident glomerular cells, thus initiating the secretion of cytokines as TNF-a. This local secretion of cytokines, in addition to stimulating chemotaxis, can induce the recruitment of circulating monocytes throughout the VtA-4/ VCAM-1 cell adhesion pathway. The mechanism proposed could be supported based on the low levels of TNF-0t in the supernatants of cultured glomeruli and on the absence of renal tissue cell infiltrates after mercury-injected rats were treated with the anti-4 integrin.