Abstract
Abstract. The majority of patients with rapidly progressive crescentic glomerulonephritis show histologic features of extensive necrosis and focal and segmental proliferation with fibrin production, but little or absent Ig deposition in the glomerulus. This subcategory of the disease, labeled “pauci-immune” glomerulonephritis, has recently been shown to be associated with the presence of antineutrophil cytoplasmic antibody in the patient's circulation (but not within the glomerulus). The absence of the effectors of humoral immunity at the site of renal injury led to this investigation of the contribution of cell-mediated immunity to the glomerular injury in this form of glomerulonephritis. In 15 patients presenting acutely with pauci-immune glomerulonephritis, CD3-positive T cells (3.7 ± 2.5 [mean ± SD] cells per glomerular cross section, [c/gcs]), CD45RO-positive T cells (2.7 ± 1.9 c/cgs), macrophages (7.3 ± 6.1 c/gcs), fibrin (3+), and endothelial-associated tissue factor were demonstrated to be prominent in glomeruli. These mediators were absent in a group of 12 patients with thin basement membrane disease and only occasionally observed in a group of eight patients with “humorally mediated” (noncrescentic) glomerulonephritis. Thus, in pauci-immune glomerulonephritis, there is the development of significant cell-mediated immunity with activated T cells, macrophages, tissue factor, and fibrin at the site of glomerular injury, suggesting that this glomerular disease is most likely a manifestation of T cell-directed cognate immune injury.
Crescentic glomerulonephritis (GN) is an inflammatory process directed by cognate immune responses (1), which results in severe glomerular injury. In the past, the participation of antibody and complement in human crescentic GN was emphasized, leading to the view that humoral immune effectors, principally immune complexes or anti-glomerular basement membrane (GBM) antibody, were the predominant pathogenic mediators in this severe form of GN (2). More recently, involvement of T cells and macrophages has been recognized, suggesting an additional contribution of cell-mediated immunity (3,4,5).
Delayed-type hypersensitivity (DTH) is a manifestation of cell-mediated immunity that is induced by sensitized T cells that recruit and activate macrophages at the site of antigen challenge. This is associated with prominent local expression of tissue factor (TF). This dysregulated TF expression initiates pathologic coagulation, which is necessary for the full expression of DTH (6). DTH skin reactions in afibrinogenemic patients lack induration, stressing the importance of fibrin to the development of inflammation in this type of immune response (7).
Studies in experimental crescentic GN support an important role for DTH in crescent formation. Transfer of sensitized T cells to animals with planted sensitizing antigens in their glomeruli has been shown to initiate glomerular injury (8). Specific depletion of the essential components of DTH, including T cells (9, 10), macrophages (11), T cell cytokines (12), and fibrin (13,14,15), without altering humoral immune responses, attenuate crescent formation and renal injury. B cell-deficient mice that have normal cell-mediated immunity but cannot produce Ig develop crescentic anti-GBM GN to the same extent as their normal littermates, confirming that crescent formation can occur completely independent of humoral immune responses (16). Glomerular expression of TF is markedly augmented in crescentic GN (17). Antibody inhibition of TF activity or infusion of its natural inhibitor, tissue factor pathway inhibitor, both significantly inhibit glomerular fibrin deposition and crescent formation (18, 19). These studies suggest that in experimental models in which both humoral and cellular responses develop to nephritogenic antigens, cell-mediated effector mechanisms similar to DTH play the major immune role in crescent formation.
It is now recognized that the majority of patients with crescentic GN have little or no antibody or complement deposition in affected glomeruli. The term “immune negative” or “pauci-immune” crescentic GN is now accepted as defining this major subgroup (20). The association of circulating antineutrophil cytoplasmic antibodies (ANCA) (21) with systemic vasculitis has been clearly established in this form of GN (1). Although the relative importance of T cells compared with antibody is difficult to assess in other forms of human GN where both humoral and cell-mediated effectors are present, it is in this group of patients who develop crescentic GN in the absence of glomerular antibody deposition that the potential for a prominent pathogenic role for cell-mediated immunity would seem likely.
Given the persisting controversies over pathogenesis in ANCA-associated pauci-immune GN, we undertook a study of a group of patients with pauci-immune ANCA-associated rapidly progressive crescentic GN to assess the involvement of the effectors of DTH and to clarify the cellular origin of TF.
Materials and Methods
Patients
Fifteen consecutive biopsies were selected on the basis of histopathologic features of focal and segmental necrotizing proliferative crescentic GN with negative or minimal Ig staining and a positive ANCA. All patients had rapidly progressive GN (RPGN), with hematuria, proteinuria, and systemic illness. No patient had any prior presentation with renal disease or had previously received any immunosuppressive therapy. Biopsies were performed before or within 48 h of commencement of immunosuppression.
Biopsies taken at the time of first presentation from two groups of patients with renal disease were used as controls. Biopsies from 12 patients with thin basement membrane disease (TBMD) were included to study renal biopsy tissue with “normal” light microscopic appearances. TBMD was diagnosed in asymptomatic patients presenting with microscopic hematuria, normal light microscopy, negative immunohistology, and a basement membrane thickness of <250 nm on electron microscopy using the intercept method (22). “Normal” renal tissue was also obtained from two potential cadaveric transplant donors in whom surgical harvesting complications prevented the use of this tissue for transplantation. Biopsies from eight patients with noncrescentic GN and prominent glomerular deposition of antibody and complement were used to provide a group with GN traditionally believed (on the basis of clinical and histologic features) to be “humorally mediated” injury (23).
Routine Histologic Assessment of Renal Tissue
Renal biopsy tissue was fixed in 10% neutral-buffered formalin, embedded in paraffin, and stained with periodic acid-Schiff and hematoxylin and eosin to assess light microscopic features. Glomerular neutrophil infiltration was enumerated by their characteristic nuclear morphology. Tissue for immunohistology was frozen in liquid nitrogen cooled isopentane. Cryostat cut sections were stained for IgA, IgG, IgM, C3, C1q, and fibrin using a routine two-layer immunoperoxidase technique. Routine histology was assessed by an experienced histopathologist (J.P.D.). Fibrin staining was scored as described below for TF.
Immunohistologic Assessment of Glomerular Leukocyte Accumulation
Cryostat cut sections (4 μm) of frozen tissue were stained using a three-layer immunoperoxidase technique to demonstrate T cells, T cell subsets, and macrophages within glomeruli as described previously (12). The primary monoclonal mouse antihuman antibodies (anti-CD5 and anti-CD3 as pan T cell markers, anti-CD4 as a T helper cell marker, anti-CD8 as a T cytotoxic/suppressor marker, HAM 56 as a macrophage marker, and anti-CD45RO as a memory T cell marker; Dako, Glostrup, Denmark) were diluted according to the manufacturer's recommendations. The subsequent antibodies were rabbit anti-mouse IgG globulin (Dako) at a dilution of 1:100, followed by a peroxidase-conjugated mouse Ig (Dako) at a dilution of 1:100. Color was developed using diaminobenzidine (Sigma Chemical Co., St. Louis, MO), and sections were counterstained with Harris' hematoxylin (BDH Chemicals, Poole, United Kingdom). Sections of normal human spleen provided a positive control. Glomerular cell numbers in all available glomeruli were assessed using a blinded protocol. Only cells clearly within the Bowman's capsule were counted.
Immunohistologic Assessment of Glomerular TF Expression
TF staining was performed on formalin-fixed paraffin sections. Sections were incubated consecutively with polyclonal rabbit antihuman TF antibody (no. 4502; American Diagnostica, Greenwich, CT), swine anti-rabbit Ig (Dako), then rabbit peroxidase anti-peroxidase complexes (Dako). Color was developed using diaminobenzidine. TF expression was assessed and scored (0 to +3) in a blinded protocol. Only glomeruli cut in or near the equatorial cross section were included. Background staining was scored 0. Glomeruli with staining in the glomerular tuft or Bowman's space covering less than one-third of the glomerular cross section were scored +1, glomeruli with staining covering between one-third and two-thirds of the glomerular cross section were scored +2, and glomeruli in which staining covered greater than two-thirds of the glomerular cross section were scored +3, as described previously (24).
Demonstration of the Cellular Localization of TF
The cellular localization of TF was demonstrated using combined immunoperoxidase and immunoalkaline phosphatase histology on frozen cryostat cut tissue. Sections were incubated consecutively with polyclonal rabbit antihuman TF antibody, biotin-labeled goat anti-rabbit Ig (Dako), and streptavidin-alkaline phosphatase complexes (Dako). Fast Blue (Dako) was used for color development. Sections were then incubated with horseradish peroxidase-conjugated monoclonal antihuman macrophage antibody (anti-CD68; Dako) or horseradish peroxidase-conjugated monoclonal anti-CD34 (Dako) to demonstrate endothelial cells. This was followed by rabbit anti-mouse Ig (Dako) and mouse peroxidase anti-peroxidase complexes (Dako). Diaminobenzidine was used for color development and methyl-green was used as a counter stain.
Results
Patients with Pauci-Immune GN
Clinical Presentation. The patient's ages ranged from 36 to 87 yr (mean ± SD, 58 ± 15 yr) with a male predominance (10:5) (Table 1). All patients presented with nonspecific systemic symptoms including arthralgia, arthritis, fever, loss of weight, myalgia, and anorexia with a duration of 3 days to 14 mo. Skin involvement with a typical vasculitic rash was noted in three patients. Six of the 15 patients had significant upper or lower respiratory tract involvement or both. Three patients had sinusitis, three had a cough, and two had minor hemoptysis in the absence of significant pulmonary infiltrate on chest x-ray. Patients presented with renal impairment in 14 cases and in one case with nephritic syndrome and an initially normal creatinine, which subsequently rose within 48 h of the biopsy (patient 13). No patient was dialysis-dependent upon presentation, although two patients subsequently required short-term dialysis before regaining dialysis independence after treatment. Four of the 15 patients had a prior history of sustained hypertension (duration 2 to 23 yr), but none had a past history of vasculitis.
Clinical features of the patients with pauci-immune GNa
Laboratory Findings. At presentation, serum creatinine was elevated in 14 of 15 patients with a range from 84 to 500 μmol/L (mean 287 ± 42 μmol/L). All patients had microscopic hematuria (mean 525,000 ± 125,600 red blood cells/ml). Proteinuria ranged from 0.1 to 3.5 g/24 h (mean 1.6 ± 0.5 g/24 h). Anti-GBM serology was negative in all patients. Anti-nuclear antibody titer was less than 1:40 in all but one patient whose titer was 1:160 with negative double-stranded DNA. All patients had positive ANCA serology. Seven patients had a cytoplasmic ANCA pattern and four of these had respiratory tract involvement. A perinuclear ANCA pattern was present in eight patients but only two had respiratory tract involvement. Thus, four of six patients (67%) with respiratory tract involvement had a cytoplasmic ANCA pattern, whereas six of nine (67%) with no respiratory tract involvement had a perinuclear ANCA pattern.
Routine Histologic Features. All patients had proliferative, focal, and segmental necrotizing crescentic GN on light microscopy (Table 2, Figure 1B), with histologic severity similar to previous reports (1). Crescents were found in 30 to 80% of glomeruli (mean 53 ± 13%) and ranged in appearance from predominantly cellular to fibrous crescents with glomerular scarring (Figure 1A). Obsolete glomeruli ranged from 5 to 35% (mean 13 ± 4%). No granulomatous lesions were detected. Neutrophil infiltration was 2.8 ± 1.7 cells per cross-sectional glomerulus (c/gcs), consistent with previous findings (25) (Figure 1D). Staining was negative for Ig and complement (IgG, IgM, IgA, C3, and C1q) in 13 of 15 patients. Sparse granular glomerular deposition of C3 was observed in one patient and IgG and IgM in another.
Histologic features of the patients with pauci-immune GN
Histologic features and cell-mediated immunity effector staining in pauci-immune glomerulonephritis (GN). (A) Fibrocellular crescent (periodic acid-Schiff [PAS]). (B) Fibrin-related antigen deposition in a glomerulus with necrotizing lesions. (C) Prominent glomerular macrophage infiltration (arrows). (D) Infrequent neutrophils (arrows; PAS). (E) Prominent infiltration with CD4-positive T cells. (F) Many T cells (arrows) expressed CD45RO, a memory T cell marker. Immunoperoxidase stain, unless otherwise stated. Magnification, ×200.
Glomerular Cell-Mediated Immunity Effectors. In patients with pauci-immune GN, the predominant intraglomerular effector cells were macrophages (7.3 ± 6.1 c/gcs) (Table 3, Figure 1C). T cells were also prominent in their glomerular lesions: CD3-positive cells, 3.7 ± 2.5 c/gcs (Figure 1E); CD5, 5.9 ± 4.7 c/gcs; CD4, 1.8 ± 1.7 c/gcs; CD8, 1.8 ± 1.4 c/gcs. The memory T cell marker CD45RO was expressed on 2.7 ± 1.9 c/gcs (Figure 1F). Glomerular expression of TF was increased in patients with crescentic GN. Staining of the glomerular epithelial and mesangial cells was observed in all glomeruli. The increase in TF staining was shown to be predominantly on endothelial cells and macrophages (Figure 2, C and D). Some TF was not cell-associated, indicating shedding from cell surfaces. Extensive and strong (+3) fibrin staining was seen in all but one patient, whose fibrin staining was graded as +2 (Figure 1B).
Intraglomerular inflammatory cells in the patients with pauci-immune GN
Tissue factor localization by immunohistochemistry. (A) Tissue factor staining in a patient with pauci-immune GN demonstrating diffuse and marked tissue factor expression (immunoperoxidase stain). Magnification, ×200. (B) In patients with thin basement membrane disease, tissue factor staining of the endothelial cells (arrows) was negative, whereas staining was observed on epithelial and mesangial cells (immunoperoxidase stain). Magnification, ×200. (C) Tissue factor and macrophage colocalization in a patient with pauci-immune GN. Magnification, ×200. Tissue factor is stained blue (immunoalkalinephosphatase), macrophages are stained brown (immunoperoxidase), macrophages expressing tissue factor are stained dark purple (an example is indicated by an arrow). (D) Tissue factor and endothelial cell colocalization (immunoperoxidase, single staining brown, dual staining with tissue factor, dark purple). Inset shows a single crescentic glomerulus with compressed glomerular tuft (boxed area), fibrosis, and necrosis. Magnification, ×100. The boxed area has been magnified ×700 in the main picture. Endothelial cell nuclei stained green (methyl green counterstain), endothelial cell membrane stained brown (immunoperoxidase), tissue factor stained blue (immunoalkalinephosphatase), and endothelial cells expressing tissue factor stained dark purple (an example is indicated by an arrow).
Control Groups
Patients with TBMD. These patients ranged in age from 26 to 64 yr (mean 46 ± 8 yr) and had a male predominance (8:4) similar to that of the pauci-immune GN patients. Microscopic hematuria was present with a mean of 230,000 red blood cells/ml. Their protein excretion ranged from 0.1 to 0.4 g/24 h (mean 0.3 ± 0.1 g/24 h), and their serum creatinine ranged from 95 to 121 μmol/L (mean 112 ± 13 μmol/L). Obsolete glomeruli ranged from 2 to 8% (mean 3.5 ± 1.4%). Their mean basement membrane thickness was 207 ± 15 nm.
Inflammatory cell infiltration in glomeruli was not detected in patients with TBMD (Table 4). TF expression was observed on glomerular epithelial and mesangial cells but not endothelial cells (Figure 2B). Fibrin staining was absent in these biopsies. Renal tissue from the two potential transplant donors showed similar findings with regard to leukocyte infiltration, TF expression, and fibrin deposition.
Intraglomerular inflammatory cells in the patients with TBMDa
Patients with “Humorally Mediated” GN. These patients ranged in age from 18 to 58 yr (mean 39 ± 6 yr) with a female predominance (5:3). Six of the patients had hypertension. Serum creatinine was elevated in five patients with a range of 98 to 220 μmol/L (mean 141 ± 15 μmol/L) but no patient was dialysis-dependent. All patients had microscopic hematuria. Proteinuria ranged from 0.6 to 10.0 g/24 h (mean 2.9 ± 3.1 g/24 h). All patients had negative anti-GBM and ANCA serology. Their clinical diagnoses and relevant laboratory findings are listed in Table 5.
Clinical details in the “humorally mediated” GN patientsa
All patients had nongranulomatous GN on light microscopy. Obsolete glomeruli ranged from 5 to 28% (mean 9.4 ± %). There was prominent deposition of Ig and complement in all patients in this group, whereas fibrin staining was graded from 0 to +1 (Table 6). Glomeruli showed a sparse inflammatory cell infiltrate comprised predominantly of macrophages (1.0 ± 0.2 c/gcs). Occasional T cells (CD3-positive cells 0.3 ± 0.1 c/gcs) and neutrophils (0.16 ± 0.04 c/gcs) were observed. Glomerular TF expression was similar to that observed in patients with TBMD.
Glomerular cell-mediated effectors and antibody deposition in the “humorally mediated” GN patientsa
Discussion
This study demonstrates the prominence of the effectors of cell-mediated immunity in glomeruli of patients with pauci-immune crescentic GN. These patients all had clinical diagnoses of RPGN (26) with systemic features on the basis of their clinical, laboratory, and histologic findings. The spectrum clinical presentation, disease severity, and ANCA pattern in this group of patients are similar to that in previous large studies of this disease (20, 27). In all patients, DTH mediators were upregulated in glomeruli, with a significant influx of T cells, including memory T cells, suggesting cognate immunity. The cell numbers in this study are conservative estimates due to the strict morphologic inclusion criteria and are slightly lower than the numbers reported in studies that included granulomatous lesions (25). In contrast, patients with “humorally mediated” GN did not show prominent glomerular accumulation of these DTH effectors. The prominence of DTH effectors, in combination with the absence of antibody and complement in pauci-immune GN, suggests an important pathogenic role for cell-mediated immunity.
Previous studies have found individual components of DTH in the glomeruli of patients with RPGN (including the ANCA subgroup) but have not comprehensively documented their simultaneous involvement. Rastaldi et al. demonstrated upregulation of vascular cell adhesion molecule-1 in association with T cells and macrophages in proliferative GN (28). Noronha et al. demonstrated intraglomerular production of tumor necrosis factor-α and interleukin-1β mRNA by mononuclear cells (29). Other studies have also shown increased serum markers of T cell activation, including soluble interleukin-2R (30), soluble CD4, and soluble CD8 in ANCA-associated GN (31). Serum endothelial leukocyte adhesion molecule-1 levels have been suggested as a marker of disease activity in patients with pauci-immune GN (32). Treatment with monoclonal anti-CD4 and anti-CD52 antibodies provided significant benefits in patients with RPGN refractory to other treatments, further supporting an important pathogenic role for cell-mediated immunity in this disease (33).
TF is the major in vivo initiator of the extrinsic pathway of coagulation and is normally anatomically sequestered from the blood (34). TF is expressed in normal glomeruli (on mesangial cells and epithelial cells) but not on glomerular endothelial cells. This localization of staining was also demonstrated in patients with thin basement membrane disease in the current study. TF is markedly upregulated in crescentic human (35) and experimental GN (17). Macrophages are the most abundant inflammatory cells observed in glomeruli in pauci-immune GN and have been shown to be responsible for augmentation of glomerular TF and fibrin deposition in experimental crescentic GN (11). In pauci-immune GN, it is likely that macrophages play a pivotal role in augmentation of glomerular TF expression, both expressing (and shedding) TF themselves and inducing TF expression on intrinsic glomerular cells. Human endothelial cells have been demonstrated to express TF after stimulation with tumor necrosis factor-α in vitro (36). Endothelial cell expression of TF has been demonstrated in endotoxic shock in baboons (37) but not in human disease. The current study provides the first evidence for in vivo expression of TF by endothelial cells in human disease.
The pathogenic role of ANCA in pauci-immune crescentic GN is controversial. There is a strong association with the ANCA titer and disease activity, although there are some patients that continue to have high titers during disease quiescence. Although ANCA is present in the circulation, it is not detectable in glomeruli at the time of ongoing renal damage. A small subgroup of patients (up to 13%) with pauci-immune crescentic GN have negative ANCA serology (27), indicating that ANCA is not essential for the development of this glomerular lesion, therefore casting doubt on its pathogenic significance. It is possible that ANCA may not have a primary pathogenic role, but may be involved in disease progression following exposure of ANCA antigens. This could explain the correlation between circulating ANCA titers and disease activity.
In vitro studies have demonstrated that ANCA antibodies can activate leukocytes and enhance leukocyte-mediated endothelial cell injury. In vitro, ANCA can activate monocytes and upregulate TF expression on endothelial cells (38); however, there is no evidence that circulating monocytes are similarly activated in patients with ANCA. Mayet et al. have shown that human umbilical vein endothelial cells can express proteinase 3 (39). However, King et al. demonstrated the absence of proteinase 3 on the glomerular endothelial and epithelial cells in patients with ANCA-positive RPGN (40). T cell proliferation in response to proteinase 3 and to neutrophil azurophilic granules is augmented in patients with pauci-immune GN, but this also occurs in 33% of normal individuals (41). It is possible that ANCA locally activates neutrophils in glomeruli of patients with pauci-immune crescentic GN. However, the reason why this response is restricted to glomeruli when ANCA circulate systemically has not been explained.
Experimental models of ANCA-associated GN have been reported. However, in these models ANCA in itself is not pathogenic (42). Sensitization to neutrophil enzymes followed by additional injurious agents including intrarenal infusion of H2O2 and neutrophil enzymes (43) or a subnephritogenic dose of anti-GBM antibody (44) were required to induce glomerular injury. These models show glomerular deposition of Ig and complement, but not T cell accumulation. These pathologic features do not correspond with the key features in human pauci-immune GN, i.e., the absence of antibody and the prominence of T cells.
In conclusion, the current study demonstrates prominent glomerular involvement of mediators of DTH in patients with pauci-immune GN. In this disease, in which local humoral immune mediators are usually absent, this provides strong evidence for an important pathogenic role for cell-mediated immunity.
Acknowledgments
Acknowledgments
This work was supported by grants from the National Health and Medical Research Council of Australia and the Australian Kidney Foundation.
Footnotes
-
American Society of Nephrology
- © 1999 American Society of Nephrology
Jump to section
More in this TOC Section
Cited By...
- Tolerogenic Dendritic Cells Attenuate Experimental Autoimmune Antimyeloperoxidase Glomerulonephritis
- Mast Cell Stabilization Ameliorates Autoimmune Anti-Myeloperoxidase Glomerulonephritis
- M2 Macrophage Infiltrates in the Early Stages of ANCA-Associated Pauci-Immune Necrotizing GN
- ANCAs Are Also Antimonocyte Cytoplasmic Autoantibodies
- The immunodominant myeloperoxidase T-cell epitope induces local cell-mediated injury in antimyeloperoxidase glomerulonephritis
- Basic and Translational Concepts of Immune-Mediated Glomerular Diseases
- Th1 and Th17 Cells Induce Proliferative Glomerulonephritis
- Anti-Neutrophil Cytoplasmic Antibodies and Effector CD4+ Cells Play Nonredundant Roles in Anti-Myeloperoxidase Crescentic Glomerulonephritis
- Toll-Like Receptor 2 Agonists Exacerbate Accelerated Nephrotoxic Nephritis
- Anti-Neutrophil Cytoplasm-Associated Glomerulonephritis
- T Cells in Crescentic Glomerulonephritis
- IL-12p40 and IL-18 in Crescentic Glomerulonephritis: IL-12p40 is the Key Th1-Defining Cytokine Chain, Whereas IL-18 Promotes Local Inflammation and Leukocyte Recruitment
- An IL-12-Independent Role for CD40-CD154 in Mediating Effector Responses: Studies in Cell-Mediated Glomerulonephritis and Dermal Delayed-Type Hypersensitivity
- Intrinsic Renal Cell Expression of CD40 Directs Th1 Effectors Inducing Experimental Crescentic Glomerulonephritis
- T Cells and Minimal Change Disease
- The Requirement for Granulocyte-Macrophage Colony-Stimulating Factor and Granulocyte Colony-Stimulating Factor in Leukocyte-Mediated Immune Glomerular Injury
- Human Glomerulonephritis Accompanied by Active Cellular Infiltrates Shows Effector T Cells in Urine
- Glomerulonephritis Induced by Recombinant Collagen IV{alpha}3 Chain Noncollagen Domain 1 Is Not Associated with Glomerular Basement Membrane Antibody: A Potential T Cell-Mediated Mechanism
- Interleukin-12 from Intrinsic Cells Is an Effector of Renal Injury in Crescentic Glomerulonephritis
- Expression of the Chemokine Monocyte Chemoattractant Protein-1 and Its Receptor Chemokine Receptor 2 in Human Crescentic Glomerulonephritis
- Clinical Utility of Testing for Antineutrophil Cytoplasmic Antibodies