Deposition of the Membrane Attack Complex in Healthy and Diseased Human Kidneys

Abstract

The membrane attack complex-also known as C5b-9-is the end-product of the classical, lectin, and alternative complement pathways. It is thought to play an important role in the pathogenesis of various kidney diseases by causing cellular injury and tissue inflammation, resulting in sclerosis and fibrosis. These deleterious effects are, consequently, targeted in the development of novel therapies that inhibit the formation of C5b-9, such as eculizumab. To clarify how C5b-9 contributes to kidney disease and to predict which patients benefit from such therapy, knowledge on deposition of C5b-9 in the kidney is essential. Because immunohistochemical staining of C5b-9 has not been routinely conducted and never been compared across studies, we provide a review of studies on deposition of C5b-9 in healthy and diseased human kidneys. We describe techniques to stain deposits and compare the occurrence of deposits in healthy kidneys and in a wide spectrum of kidney diseases, including hypertensive nephropathy, diabetic nephropathy, membranous nephropathy, IgA nephropathy, lupus nephritis, C3 glomerulopathy, and thrombotic microangiopathies such as the atypical hemolytic uremic syndrome, vasculitis, interstitial nephritis, acute tubular necrosis, kidney tumors, and rejection of kidney transplants. We summarize how these deposits are related with other histological lesions and clinical characteristics. We evaluate the prognostic relevance of these deposits in the light of possible treatment with complement inhibitors.

Keywords: C5b-9 (membrane attack complex [MAC]); biopsy; clinicopathological correlation; glomerular disease; histopathology; immunofluorescence; immunohistochemistry; renal.

Figure 1

Deposits of C5b-9 in healthy and diseased human kidneys. Pie charts show the proportion of studies that reported staining of C5b-9 as absent (light) or present (dark). Bar charts show the medians of the proportions of patients reported to exhibit staining. Scatter charts show the median staining intensities in these patients. All charts show data separately for staining in the glomerulus as a whole (glom.), in the mesangium (mes.), along the glomerular capillary wall (cap.), along the tubular basement membrane (tub.), or in the extraglomerular vascular wall (vas.). Error bars show the lowest and highest reported values. Numbers of studies are indicated between brackets. Some studies reported only part of the data shown, explaining differences in the numbers of studies between pie, bar, and scatter charts. Nothing is indicated if the data were never reported. Detailed data per study are listed in Supplementary Table 2 . Membranous nephropathy excludes studies conducted specifically on secondary membranous nephropathy. IgA nephropathy excludes studies conducted specifically on IgA vasculitis with nephritis. Data on these diseases and on glomerular basement membrane diseases, hypertensive nephropathy, interstitial nephritis, acute tubular necrosis, and kidney tumors are only listed in Supplementary Table 2 because of a paucity of data. ANCA: antineutrophil cytoplasmic antibody; MPGN: membranoproliferative glomerulonephritis.

Figure 2

Staining of C5b-9 in a healthy and a diseased kidney. Examples of staining of C5b-9 from our laboratory are shown. (A) In a healthy kidney, staining was present in the vascular pole of the glomerulus and the vascular wall of extraglomerular arteries and focally with less intensity along Bowman’s membrane and the tubular basement membrane. This tissue was obtained with a biopsy from a living donor before kidney transplantation. (B) In a kidney of a patient with aHUS, staining was present along the glomerular capillary wall, in the vascular wall of extraglomerular arteries and focally along Bowman’s membrane and the tubular basement membrane. This tissue was obtained with a clinically indicated biopsy. Both tissues were fixed, paraffin-embedded, and sectioned. After deparaffinization (xylol and ethanol) and antigen retrieval (PBS-0.1% Proteinase XXIV, P8038, Sigma), sections were washed and endogenous peroxidase was blocked (PBS, 0.1% NaN₃, 1% H₂O₂) for 30 min at room temperature. Sections were washed (PBS) and incubated with mouse anti-human C5b-9 (2 µg/ml, aE11, HM2167, Hycult Biotech, Uden, the Netherlands) or an isotype control (mouse IgG2a, 2 µg/ml, X0943, Dako, Jena, Germany) in PBS with 1% BSA over night at room temperature. Next day, slides were washed and incubated with goat anti-mouse horseradish peroxidase (HRP, 5 µg/ml, P0447, Dako) for 30 min at room temperature. Slides were washed and incubated with rabbit anti-goat HRP (2.5 µg/ml, P0449, Dako) for 30 min at room temperature. Slides were washed and developed using NovaRED following protocol (Vector Labs, Peterborough, UK) and counterstained (Mayer’s hematoxylin, 1.09249.0500, Merck, Darmstadt, Germany) for 25 s. Slides were not counterstained with eosin, which explains why tubules may seem dilated. Slides were dried overnight at room temperature before being covered using entellan (1.07961, Merck).

Figure 3

Technical aspects of staining of C5b-9. (A) Glomerular staining intensity of C5b-9 is shown in relation with those of its individual components C5, C6, and C9 in kidney biopsies of patients with IgA nephropathy (n = 18). Antibody anti-MAC-neo was used for staining of C5b-9. We plotted previously published individual data (52). (B) Glomerular staining intensities of C5 and C9 are compared in kidney biopsies of patients with IgA nephropathy (n = 15). We plotted previously published individual data (76). (C) Staining intensities of C5b-9 in the mesangium (mes.) and along the capillary wall (cap.) are shown for first and repeat biopsies with the time between both biopsies in patients with lupus nephritis (n = 8) who responded or did not respond to immunosuppressive treatment. Antibody aE11 was used for staining. We plotted previously published individual data (18).

Figure 4

Deposits of C5b-9 in diabetic nephropathy. (A) Presence of C5b-9 in the mesangium (mes.), along the glomerular capillary wall (cap.), in glomerular hili, and in the extraglomerular vascular wall (vas.) is compared between patients without diabetes or kidney disease (n = 41), patients with diabetes who had no nephropathy (n = 58), and patients with diabetic nephropathy (n = 101). (B) Presence of C5b-9 in the glomerulus is compared between patients with different classes of diabetic nephropathy (n = 101) according to the classification of the Renal Pathology Society (146). Patients with diabetes but without diabetic nephropathy (n = 58) are indicated as class 0. Differences between classes were tested with Spearman’s correlation. We reproduced both panels without adaptations from their previous publication under the CC BY-NC-ND license (27), ^© International Society of Nephrology. (C) Presence of C5b-9 in the glomerulus is compared between patients with diabetes type 1 (n = 17) and type 2 (n = 120). It was different between diabetes types 1 and 2, both among patients without and with diabetic nephropathy, as tested with Fisher’s exact test (both p < 0.05). The antibody used for staining in these three panels was unspecified. We plotted previously published data (27).

Figure 5

Deposits of C5b-9 in primary membranous nephropathy: examples of correlations with histological lesions and clinical characteristics. (A) The extent of staining of C5b-9 in the capillary wall is shown in relation with mesangial hypercellularity scored on scale from 0–3 in children with idiopathic membranous nephropathy (n = 16). The antibody used for staining was unspecified. Relations were tested with Spearman’s correlation (ρ). We plotted previously published individual data (87). (B) The extent of staining of C5b-9 in tubules is shown in relation with the number of leukocytes in the interstitium in patients with idiopathic membranous nephropathy (n = 27). (C) The extent of staining of C5b-9 in tubules is shown in relation with serum creatinine at the time of biopsy in patients with idiopathic membranous nephropathy (n = 27). Antibody aE11 was used for staining in both panels. Relations were tested with Pearson’s correlation (r). We reproduced both panels without adaptations from their previous publication with permission (112), ^© European Renal Association–European Dialysis and Transplant Association. (D) The extent of staining of C5b-9 in the capillary wall is shown in relation with the clinical outcome after 14-171 months of treatment in children with idiopathic membranous nephropathy (n = 16). The antibody used for staining was unspecified. The hazard ratio (HR) for kidney failure, proteinuria, or hematuria as compared with normal outcomes is given as estimated with Cox’s regression. We plotted previously published individual data (87).

Figure 6

Deposits of C5b-9 in IgA nephropathy: examples of correlations with histological lesions and clinical characteristics. (A) Staining intensity of C5b-9 in the glomerulus is shown in relation with histological patterns in patients with IgA nephropathy (n = 18). Antibody anti-MAC-neo was used for staining. Differences between staining intensities were tested with Fisher’s exact test. We plotted previously published individual data (52). (B) The extent of staining of C5b-9 in tubules is shown in relation with the number of interstitial monocytes and macrophages in patients with IgA nephropathy (n = 18). Antibody aE11 was used for staining. We reproduced this panel without adaptations from its previous publication with permission (110), ^© European Renal Association–European Dialysis and Transplant Association. (C) Staining intensity of C5b-9 in the glomerulus is shown in relation with serum creatinine at the time of biopsy in patients with IgA nephropathy (n = 14). Antibody aE11 was used for staining. We plotted previously published individual data (128). (D) The extent of staining of C5b-9 in tubules is shown in relation with serum creatinine at the time of biopsy in patients with IgA nephropathy (n = 18). Antibody aE11 was used for staining. We reproduced this panel without adaptations from its previous publication with permission (110), ^© European Renal Association–European Dialysis and Transplant Association. Relations were tested with Pearson’s correlation (r). (E) Staining intensity of C5b-9 in the glomerulus is shown in relation with proteinuria at the time of biopsy in patients with IgA nephropathy (n = 18). Antibody anti-MAC-neo was used for staining. Differences between staining intensities were tested with Fisher’s exact test. The relatively small number of patients may explain why proteinuria <1 g/d was not observed in the group with a staining intensity of ++. We plotted previously published individual data (52).

Figure 7

Deposits of C5b-9 in lupus nephritis: examples of correlations with histological lesions and clinical characteristics. The extent of staining of C5b-9 in tubules is shown in relation with the extent of peritubular inflammation in patients with lupus nephritis class II, III, IV, or V (n = 22). The antibody indicated as Kolb 1975 in Table 2 was used for staining. The relation was tested with Pearson’s correlation (r). We reproduced this panel without adaptations from its previous publication with permission (95), ^© Rockefeller University Press.

Figure 8

Deposits of C5b-9 in hypertension-associated thrombotic microangiopathy: examples of correlations with histological lesions and clinical characteristics. Staining intensity of C5b-9 along the glomerular capillary wall (cap.) and in the extraglomerular vascular wall (vas.) is shown in relation with (A) proteinuria and (B) plasma complement activity (CH50) at the time of biopsy in patients with hypertension-associated thrombotic microangiopathy (n = 6). The antibody used for staining was unspecified. Relations were tested with Pearson’s correlation (r). We plotted previously published individual data (192).

Figure 9

Deposits of C5b-9 in ANCA-associated vasculitis: examples of correlations with histological lesions and clinical characteristics. (A) The average staining intensity of C5b-9 in glomeruli is shown in relation with the percentage of glomeruli with mild mesangial hypercellularity and that with cellular crescents in patients with myeloperoxidase antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (n = 7). The antibody used for staining was unspecified. The correlation coefficient for mild mesangial hypercellularity is given; that for cellular crescents was nonsignificant. We plotted previously published individual data (116). (B) Staining intensity of C5b-9 in the glomerulus is shown in relation with histological patterns in patients with renal ANCA-associated vasculitis (n = 25). Antibody ab55811 was used for staining. Differences between staining intensities were tested with Fisher’s exact test. We plotted previously published individual data (53). (C) The average staining intensity of C5b-9 in glomeruli is shown in relation with serum creatinine at the time of biopsy in patients with myeloperoxidase antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (n = 7). The antibody used for staining was unspecified. We plotted previously published individual data (116). (D) The average staining intensity of C5b-9 in glomeruli is shown in relation with serum creatinine at the time of biopsy in patients with ANCA-negative pauci-immune crescentic glomerulonephritis (n = 12). The antibody used for staining was unspecified. We plotted previously published individual data (117). Relations were tested with Pearson’s correlation (r).