Effects of CTLA4-Fc on glomerular injury in humorally-mediated glomerulonephritis in BALB/c mice

Abstract

The effect of cytotoxic T-lymphocyte-associated molecule 4-immunoglobulin fusion protein (CTLA4-Fc) on humorally-mediated glomerulonephritis was studied in accelerated anti-glomerular basement membrane (anti-GBM) glomerulonephritis induced in BALB/c mice. This strain of mice develops antibody and complement dependent glomerulonephritis under this protocol. Sensitized BALB/c mice developed high levels of circulating autologous antibody titres, intense glomerular deposition of mouse immunoglobulin and complement, significant proteinuria, renal impairment, significant glomerular necrosis and a minor component of crescent formation 10 days after challenge with a nephritogenic antigen (sheep anti-GBM globulin). Early treatment during the primary immune response, or continuous treatment throughout the disease with CTLA4-Fc, significantly suppressed mouse anti-sheep globulin antibody titres in serum, and immunoglobulin and complement deposition in glomeruli. The degree of glomerular necrosis was improved and proteinuria was reduced, particularly in the earlier stages of disease. Late treatment by CTLA4-Fc starting one day after challenge with sheep anti-mouse GBM did not affect antibody production and did not attenuate glomerulonephritis. The low level of crescent formation found in BALB/c mice developing glomerulonephritis was not prevented by the administration of CTLA4-Fc. These results demonstrate that CTLA4-Fc is of benefit in this model of glomerulonephritis by its capacity to attenuate antibody production, without affecting the minor degree of cell-mediated glomerular injury.

INTRODUCTION

Most forms of glomerulonephritis (GN) are dependent on the generation of adaptive immune responses, in which the CD4⁺ cell plays a central role. CD4⁺ cells respond to antigens by providing help for antibody production by B cells, and by effecting injury through the activation of macrophages and induction of cytotoxic CD8⁺ cells. The TcR-MHC complex provides the ‘first signal’ for the development and maintenance of antigen-specific CD4⁺ cells. To respond appropriately in terms of T-cell help, CD4⁺ cells must also interact with antigen-presenting cells and effector cells such as B cells and macrophages via costimulatory molecules providing a ‘second signal’. The two best recognized and described pathways are the CD40, CD154 (CD40L) and the CD28, CD80/86 (B7.1, B7.2) interactions. CD80 and CD86 are expressed on antigen-presenting cells and interact either with CD28, which provides a stimulatory signal, or CTLA4 which has a higher affinity for CD80 and CD86 than CD28, and, in most circumstances, provides a negative signal to the CD4⁺ cell. The development of a longer-acting and soluble form of CTLA4, CTLA4-Fc, which inhibits CD28/B7 signalling [1], has enabled the study of the function of the CD28/CTLA4–CD80/CD86 system in normal and abnormal immune responses.

The glomerulonephritides are a group of diseases characterized by immune glomerular injury. This injury occurs as a result of a number of different processes, including organ-specific or systemic autoimmunity, dysregulated immune responses and often, such as in post-infectious GN, as a result of an otherwise appropriate immune response to infection-related antigens. The variety of pathogenetic mechanisms, Th1/Th2 polarization of the nephritogenic immune responses, the variable participation of immune effectors in glomerular injury and the wide range of clinical course and outcomes in patients with GN, imply that the nature of the host’s immune response to antigenic stimuli is relevant to the patterns and outcomes of immune renal injury [2]. When sheep globulin is planted in glomeruli of sensitized mice, C57BL/6 mice make Th1 predominant responses to sheep globulin, and develop effector CD4⁺ cell dependent and antibody independent crescentic GN with significant T cell, macrophage and fibrin deposition [3,4]. BALB/c mice, when challenged with the same antigen, develop different immune responses and different patterns of injury. They do not make strong Th1 responses but they do develop a high level of antibody, together with significant antibody- and complement-mediated injury but only minor degrees of crescent formation [3,5].

To test the hypothesis, that inhibition of CD80 and CD86 by administration of CTLA4-Fc would diminish antibody-mediated glomerular injury in a non-autoimmune model of injury, GN was studied in BALB/c mice (accelerated nephrotoxic serum nephritis or accelerated ‘anti-glomerular basement membrane’ anti-GBM] GN). BALB/c mice were sensitized to sheep globulin and challenged with sheep anti-mouse GBM globulin. CTLA4-Fc treatment was given at three different time-points: during the induction of the immune response, the effector phase of the injury, and throughout both the induction and the effector phases of the immune response.

MATERIALS AND METHODS

Experimental design

Anti-mouse GBM globulin was prepared from serum (adsorbed twice with mouse RBC then precipitated with ammonium sulphate) of a sheep immunized against homogenized and sonicated murine renal cortex in Freund’s complete adjuvant (CFA, Sigma Chemical Co., St. Louis, MO, USA) and later, Freund’s incomplete adjuvant (IFA). Protein G purified mouse CTLA4-Fc was prepared from protein G purified culture supernatant fluid of a transfected NS-1 cell line [1], generously provided by Dr T Strom (Harvard Medical School, Boston, MA, USA). Protein G purified normal mouse IgG was used in control treatment. Each mouse received 100 μg CTLA4-Fc or normal mouse IgG per dose.

Male BALB/c mice between 8 and 10 weeks of age were obtained from Monash University Central Animal Services (Clayton, Victoria, Australia). Mice were sensitized by the injection of a total of 2 mg of sheep globulin in 200 μl of CFA in each flank (day –10); 10 days later (day 0), GN was initiated by intravenous (i.v.) administration of 5 mg of sheep anti-mouse GBM globulin. Glomerular injury was assessed 10 days after administration of anti-GBM antibody (day 10).

The effect of in vivo CTLA4-Fc on the development of anti-GBM GN was studied by administering CTLA4-Fc to mice in three different protocols. The following groups were studied.

1. Control treatment (Ctrl): 8 mice received normal mouse Ig.

2. Continuous treatment (CTLA4 contin): 8 mice received 6 i.v. doses of 100 μg of CTLA4-Fc throughout the disease course, on days −10, −8, −6, −2, +3 and +7.

3. Early treatment (CTLA4 early): 6 mice received 4 i.v. doses of 100 μg CTLA4-Fc over 10 days before administration of sheep anti-mouse GBM globulin, during the initiation of the immune response, on days −10, −8, −6 and −2. No further CTLA4-Fc was administered following the initiation of glomerular injury after challenge with sheep anti-mouse GBM globulin.

4. Late treatment (CTLA4 late): 6 mice received 3 i.v. doses of 100 μg CTLA4-Fc on days 2, 4 and 6, commencing after challenge with i.v. anti-GBM globulin.

Results are expressed as the mean ± s.e.m. The significance of differences between groups was determined by anova, followed by Tukey’s Multiple Comparison Test for paired comparisons (GraphPad Prism, GraphPad Software Inc., San Diego, CA, USA).

Titres of serum anti-sheep globulin immunoglobulin

Titres of mouse anti-sheep globulin immunoglobulin were measured by ELISA on serum collected at the end of experiments in autologous injury, as previously described [6]. Sera from control treated mice with GN were tested separately against each of the three CTLA4-Fc treated groups.

Assessment of histological indices of glomerular injury

Kidney tissue was fixed in Bouin’s fixative, embedded in paraffin, and 3 μm tissue sections cut and stained with periodic acid-Schiff (PAS). A glomerulus was considered to exhibit crescent formation if two or more layers of cells were observed in Bowman’s space. A glomerulus was considered to exhibit necrosis if significant segmental or lobular accumulation of PAS+ material, obliteration of capillary lumina and relative hypocellularity was present. A minimum of 50 glomeruli was assessed to determine the percentage of glomeruli affected by crescent formation or necrosis for each animal.

Glomerular deposition of mouse immunoglobulin and C3

For detection of autologous antibody and C3 in glomeruli, tissue was embedded in Optimal Cutting Temperature Compound, frozen in liquid nitrogen and stored at −70°C. Mouse immunoglobulin was detected on 4 μm cryostat cut sections using FITC-sheep anti-mouse immunoglobulin (Silenus, Hawthorn, Victoria, Australia) at dilutions of 1 in 100 and 1 in 10 000, and C3 using FITC-goat anti-mouse C3 (Cappel, Durham, NC, USA) at dilutions of 1 in 100 and 1 in 5000. Fluorescence intensity was assessed semi-quantitatively (0–3+). Sections in which only some glomeruli were positive were graded as 0·5 (i.e. +/−).

Proteinuria and serum creatinine

Urinary protein concentrations were determined by a modified Bradford method [7] on timed urine collections. Mice were housed individually in cages to collect urine over a 24-h period at day 1, 5 and 9 after initiation of disease. Serum creatinine concentrations at the completion of experiments in autologous phase GN (day 10) were measured by the alkaline picric acid method using an autoanalyser.

RESULTS

The effects of CTLA4-Fc on the humoral immune response to sheep globulin

Control treated mice with GN developed high anti-sheep globulin antibody titres in the serum collected on day 10 of disease (Fig. 1). Administration of CTLA4-Fc throughout the disease, including the priming phase and the effector phase, reduced serum antigen-specific antibody levels (Fig. 1a). When the treatment phase was limited only to the priming phase of disease, similar reductions in antibody titres were observed (Fig. 1b). However, delaying CTLA4-Fc injection until the immune responses and GN were established had no effect on serum antibody levels (Fig. 1c).

Fig.1.

Systemic humoral immune responses to sheep globulin in control treated mice (□) and CTLA4-Fc treated mice (▪) showing significant reductions in circulating mouse anti-sheep globulin titres in mice when treatment was commenced during the priming phase of the disease (a: CTLA4 Continuous or b: CTLA4 early). However, there was no difference when treatment was commenced after the systemic immune response against sheep globulin was established (c: CTLA4 late). Each treatment group was tested against the control group in a separate assay. (▾) No GN.

Humoral immune effectors of injury, namely autologous antibody and complement (C3) in glomeruli, were demonstrated by immunofluorescence (Tables 1 and 2). Intense deposition of mouse immunoglobulin and C3 were observed along glomerular capillary loops in control treated mice with GN. In mice in which CTLA4-Fc had been administered throughout the immune response or during only the priming phase, only weak and sparse immune deposition was seen. However, the administration of CTLA4-Fc during the effector phase of disease had no effect on the glomerular deposition of mouse Ig and C3.

Table 1.

Deposition of autologous antibody in glomeruli of mice with GN, showing reduced deposition in mice treated with CTLA4-Fc during the establishment of the immune response

	1 in 150	1 in 300	1 in 600	1 in 1200
Ctrl GN	+ + +	+ +	+	±
CTLA4 continuous	±	−	−	−
CTLA4 early	+	+	±	−
CTLA4 late	+ + +	+ +	+	±

Table 2.

Deposition of C3 in glomeruli of mice with GN, showing reduced deposition in mice treated with CTLA4-Fc during the establishment of the immune response, similar to the pattern observed in assessment of autologous antibody

	1 in 50	1 in 500	1 in 1000	1 in 2000
Ctrl GN	+ + +	+	+	±
CTLA4 continuous	±	−	−	−
CTLA4 early	+	±	−	−
CTLA4 late	+ + +	+	+	±

Effect of CTLA-4 Fc treatment on the development of anti-GBM GN

Significant GN was observed at day 10 in control treated mice, with extensive glomerular segmental or lobular accumulation of PAS+ material, obliteration of capillary lumina and relative hypocellularity, suggestive of necrosis, that involved 74 ± 5% of glomeruli (Figs 2a, 3a,b). Crescent formation was present in 9 ± 1% of glomeruli, consistent with previous data in this model using this strain of mice (Fig. 2b) [3].

Fig.2.

Histological features of glomerular injury in BALB/c mice with GN: areas of necrosis (a) and glomerular crescent formation (b). Control treated mice with GN developed moderately severe GN with most glomeruli affected by at least segmental areas of necrosis (a), but with only a relatively minor degree of glomerular crescent formation (b). Treatment with CTLA4-Fc throughout the course of the disease (CTLA4 contin), or during the priming phase prior to the initiation of glomerular injury by challenge with sheep anti-mouse GBM (CTLA4 early), reduced the degree of glomerular necrosis, with continuous treatment having a greater effect. **P < 0·01 versus control (anova, Tukey’s post test). However, treatment commenced after challenge (CTLA4 late) had no effect on the degree of glomerular necrosis. None of the treatment regimes reduced the proportion of glomeruli affected by crescent formation.

Fig.3.

Histological features of injury in BALB/c mice with GN. Photomicrographs of mice with GN showing severe glomerular injury with significant necrosis in control treated mice with GN (a, b). There was significant protection from injury in mice treated with CTLA4-Fc throughout the course of disease (c, d) or during the initiation of the immune response (e, f). However, delaying treatment until the immune response was established had little effect on the disease (g, h). PAS stain a, c, e, g - low power, b, d, f, h, high power.

Both early and continuous treatment significantly reduced the degree and incidence of necrosis in glomeruli (Figs 2a, 3c–f; P < 0·001 versus control). Late treatment failed to attenuate the histological manifestations of disease in glomeruli. There was no significant difference compared with the control group (Figs 2a, 3g,h). The relatively low level of crescent formation found in control treated mice was not affected by CTLA4-Fc treatment at any time-point (Fig. 2b). Significant proteinuria was found in control mice after initiation of anti-GBM GN from day 1 to day 10 (Fig. 4a). Late treatment by CTLA4-Fc did not attenuate the severity of proteinuria. Proteinuria was reduced by early and continuous treatment in the early phase of disease, but it was increasing by the last timed collection (Fig. 4a). Control treated mice had developed renal impairment, with an elevated serum creatinine at day 10 of disease (Fig. 4b). Early or continuous treatment attenuated renal impairment compared with the control treated mice (Fig. 4b), although late treatment with CTLA4-Fc did not prevent the decline in renal function.

Fig.4.

Proteinuria (a) and renal function (b) in BALB/c mice with GN. (a) Control treated mice with GN developed significant proteinuria (□). There was no effect of delayed treatment with CTLA4-Fc (diamonds ▾). Proteinuria in the early stages of autologous injury was reduced in mice treated with CTLA4-Fc either throughout the course of the disease (CTLA4 contin, ▪) or during only the priming phase (CTLA4-Fc early, ▴), with a trend to a greater reduction in mice treated throughout the disease. The dotted line represents the urinary protein excretion of normal mice without GN. *P < 0·05 versus control, **P < 0·01 versus control (anova, Tukey’s post test). (b) Renal function, showing a raised serum creatinine (impaired renal function) in control treated mice. The dotted line represents the normal serum creatinine. There was a significant reduction in serum creatinine present in mice treated with CTLA4-Fc either throughout the course of disease (CTLA4 contin) or early in the immune response (CTLA4 early). *P < 0·05 versus control, **P < 0·01 versus control (anova, Tukey’s post test).

DISCUSSION

In the current studies, the model of glomerular injury used is not dependent on an autoimmune response to nephritogenic antigens [8]. Mice are sensitized to sheep globulin and challenged by planting this antigen in glomeruli. Injury results from the immune response to sheep globulin as a foreign antigen which, in BALB/c mice, is predominantly humorally-mediated, associated with high levels of antigen-specific antibody, complement-dependent and with only a relatively low number of CD4⁺ cells and macrophages, and sparse fibrin deposition in glomeruli [3,5]. Glomerular crescent formation is relatively infrequent and is the only significant feature of the disease that is effector CD4⁺-cell dependent. In contrast, severe renal injury in C57BL/6 mice is effector CD4⁺-dependent [3,5], antibody-dependent [4] and dependent on the development of Th1 nephritogenic immune response to sheep globulin [3]. Glomerular lesions in C57BL/6 mice have features of delayed-type hypersensitivity.

The current studies demonstrate that CTLA4-Fc can, by inhibiting CD28–CD80/CD86 interactions in anti-GBM GN in BALB/c mice, reduce humoral responses and glomerular injury, but only if the treatment is commenced prior to glomerular injury. At least some of the relatively minor degree of glomerular crescent formation present in BALB/c mice (compared with C57BL/6 mice) is effector CD4⁺-dependent [3] and was not inhibited by CTLA4-Fc administration, even when it continued through-out the priming and effector phases, consistent with results in C57BL/6 mice [9]. The absence of an effect on renal injury in delayed treatment is consistent with the lack of effect on serum antibody levels or on the glomerular deposition of antibody or complement. Whether prolonged treatment with CTLA4-Fc would diminish renal injury in this model of injury is uncertain, and is likely to depend on the capacity of CTLA4-Fc to diminish the production of antigen-specific antibody in the context of an established immune response, which is not characterized by systemic or organ-specific autoimmunity.

In GN, glomerular injury can occur via a number of different mediator systems. Injury mediated by the glomerular deposition of antibody, and its consequent interactions with either the complement system or FcR, is responsible for glomerular injury in a number of human and experimental forms of GN. In contrast, in at least some proliferative and/or crescentic forms of GN, cell-mediated injury plays an important role in the development of severe glomerular injury. As GN is a collection of a number of different diseases, the nature of the immune responses that lead to glomerular injury varies. Some forms of GN are characterized by autoimmunity, either systemic such as SLE, or organ specific such as anti-GBM GN, while others are a consequence of an immune response against exogenous antigens that affects the kidney, such as most forms of infection-related GN. While it is possible (and in fact likely) that there are features common to some or many forms of GN, the diverse nature of GN implies that immune responses that lead to GN may be vulnerable in different ways and therefore, that careful exploration of the role of costimulatory signals in different models of GN is necessary.

The role of interactions between CD28 and either CD80 or CD86, and, in particular, the potential inhibitory function of CTLA4, has been studied in several forms of experimental GN. In lupus prone mice, where renal injury occurs as a result of the development of systemic autoimmunity and a combination of humoral and, to some degree, cell-mediated mechanisms, CTLA4-Fc has successfully blocked the development of disease and of renal injury [10], with treatment having an effect even when delayed. However, unless combined with anti-CD154 (CD40L) antibodies, a short course early in the disease is not enough to prevent future development of disease [11]. In rat autoimmune anti-GBM GN, where organ-specific autoimmunity affects the kidney, CTLA4-Fc diminished GN [12,13] including, in one study [12], when treatment was delayed until 14 days after the administration of the autoantigen. In chronic graft versus host disease, where an allogeneic response results in immune complex formation and chronic glomerular injury that is Th2 driven, CTLA4-Fc was able to diminish injury when given 7 days after the initiation of disease [14].

Given these findings and considering them in the context of our studies, whether CTLA4-Fc is able to successfully inhibit humoral and cellular immune responses would seem to depend on a number factors that have implications for the translation of studies in animals to human disease. The timing of the administration of CTLA4-Fc is important. In the model used for the current studies (where the nephritogenic antigen is a planted non-self antigen), by day 0, an immune response against sheep globulin is strong and well established, and abrogation of ongoing CD4⁺ help by a depleting monoclonal antibody does not affect antibody levels [3]. The lack of effect of CTLA4-Fc treatment commenced at day 0 in the current studies is consistent with this finding, in contrast to delayed treatment with CTLA4-Fc in an autoimmune anti-GBM rat model [12].

The intensity or nature of the immune response in GN is important, and it may be easier to inhibit a (possibly more subtle) autoimmune response [10,12] than an immune response against a foreign antigen, as used in the current studies. Why it is seemingly easier to inhibit crescentic renal injury in autoimmune anti-GBM GN than planted antigen GN (the current studies in BALB/c mice and their counterparts in C57BL/6 mice [9]) is not clear. There may be a minor degree of crescent formation in this model that is T-cell-independent, as CD4⁺ depletion in the effector phase in BALB/c mice has previously reduced, but not abrogated, glomerular crescent formation [9]. In autoimmune diseases, CTLA4 has been demonstrated to be important in maintaining tolerance [15], and data on a role for a subset of CD25⁺ CD4⁺ cells in negatively regulating autoimmune processes suggest that constitutive expression of CTLA4 by these cells may be important in their function [16]. It is possible that in autoimmune GN, part of the mechanism of the effect of CTLA4-Fc (even in delayed treatment) is due to CTLA4 enhancing the function of these CD25⁺ CD4⁺ regulatory cells. Lastly, our current data and previous data [9] suggest that it may be easier to inhibit nephritogenic humoral responses than cell-mediated immune responses.

In summary, these studies demonstrate that in humorally-mediated GN induced by a foreign antigen, limiting antibody-mediated injury using CTLA4-Fc treatment can attenuate GN. However, the minimal crescent formation that occurs is not affected by reducing the antibody response. They support previous work in the anti-GBM ‘planted antigen’ model by (i) demonstrating that most of the injury in BALB/c mice is humorally-mediated [3,5] and (ii) by showing that in nephritogenic immune responses to foreign antigens, it is easier to inhibit humoral than cellular responses, as was the case in C57BL/6 mice using either combined anti-B7.1 and anti-B7.2 antibody treatment, or CTLA4-Fc therapy. They demonstrate that in this model, not involving autoimmune responses or a major contribution for cell-mediated injury, CTLA4-Fc can be effective, but timing of administration is crucial. Finally, they imply that the minor cell-mediated component of injury is less vulnerable to inhibition by engagement of CTLA4.