This article requires a subscription to view the full text. If you have a subscription you may use the login form below to view the article. Access to this article can also be purchased.
Abstract
Background and objectives Complement alternative pathway (cAP) activation has recently been recognized as a key pathogenic event in ANCA-associated vasculitis (AAV). cAP dysregulation is also a major determinant of thrombotic microangiopathies (TMA), which can in turn complicate AAV. We explored the prognostic significance of cAP activation and of histologic evidence of TMA in a cohort of patients with renal AAV.
Design, setting, participants, & measurements We studied 46 patients with AAV diagnosed between January 1990 and December 2011 at the Nephrology Unit of Parma University Hospital; 30 of them had undergone renal biopsy. We analyzed serum levels of C3 (sC3) and C4 (sC4) and, for 19 patients who had frozen plasma, plasma Bb and C5b-9 levels. We also reviewed all kidney biopsy specimens, specifically searching for histologic signs of TMA, and performed immunofluorescence or immunohistochemistry for C3d, C4d, Bb and C5b-9.
Results sC3 was below the lower limit of normal in 35% of the patients, whereas C4 was low in only 2%. Patients with low sC3 tended to be older (P=0.04) and to have lower eGFR at diagnosis (P=0.06). The median follow-up was 78 months (interquartile range, 18–135 months); 18 patients reached ESRD (10 of 14 and 8 of 26 in the low and normal sC3 groups, respectively). Death-censored renal survival was lower in the low sC3 group than in the normal sC3 group (log-rank test, P=0.01). Eight of the 30 patients who had undergone biopsy (27%) had histologic signs of TMA; these signs were more frequent in patients with low sC3 (5 of 10 versus 3 of 20; P=0.04). Notably, patients with histologic signs of TMA had a dramatically worse death-censored renal survival than patients without TMA (log-rank test, P=0.01), with ESRD occurring in 8 of 8 patients with TMA versus 8 of 22 patients without TMA.
Conclusions Low sC3 levels and histologic signs of TMA are associated with a poor renal prognosis in patients with AAV.
- complement
- CKD progression
- vasculitis
- anti-neutrophil cytoplasmic antibody-associated vasculitis antibodies, antineutrophil cytoplasmic biopsy kidney failure, chronic thrombotic microangiopathies
Introduction
ANCA-associated vasculitis (AAV) is characterized by pauci-immune necrotizing inflammation of small blood vessels. Pauci-immunity is the hallmark that distinguishes AAV from other immune complex–mediated vasculitis, such as lupus nephritis or Henoch-Schönlein purpura, that usually show glomerular deposits of immunoglobulins and complement fractions (1).
Until recently, it was commonly accepted that serum C3 (sC3) and C4 (sC4) levels are normal in patients with AAV, which probably led to underestimate the pathogenic importance of the complement system in AAV (2). However, previous studies showed that most patients have at least focal complement tissue deposition (3), but the relevance of this observation was underpinned by the important ANCA discovery and the subsequent “pauci-immune/pauci-complement paradigm.” Twenty years later, these findings were reconsidered and prompted subsequent in vivo studies that, through use of murine models of AAV, clearly demonstrated an activation of the terminal part of the complement cascade, particularly of C5, with an engagement of the C5a receptor (4–6). The relevance of the complement system in the pathogenesis of ANCA-associated necrotizing crescentic GN (NCGN) was further supported by immunohistochemical studies that have confirmed the deposition of C3d, factor B, and factor P in glomeruli and small blood vessels (7,8). A prevailing role of complement alternative pathway (cAP) was postulated when significantly higher plasma levels of sC5b9, that marks the terminal complement pathway, and of C3a and Bb were found in active AAV (9).
Abnormal activation of the cAP is also a cardinal feature of thrombotic microangiopathies (TMA), particularly atypical hemolytic uremic syndrome (aHUS). Genetically determined defects of cAP regulatory proteins (e.g., complement factor H, membrane cofactor protein) are responsible for most cases of primary aHUS in adults, but aHUS can also be associated with systemic immune-mediated disorders. In a recent literature review that also included original cases, we showed that aHUS can be associated with several types of vasculitis, particularly AAV, and that patients with aHUS overlapping with AAV usually have a poor outcome (10). Histologic signs of TMA have been encountered in AAV, but there are no systematic studies, and the prognostic importance of these abnormalities is unknown.
In this study, we explored the prognostic significance of complement activation and of histologic evidence of renal TMA in a cohort of patients with AAV.
Materials and Methods
We identified 59 patients with renal AAV, diagnosed between January 1990 and December 2011 at the Nephrology Unit of Parma University Hospital, Parma, Italy. We reviewed all the diagnoses and excluded 13 patients: 12 because they were ANCA negative and 1 because the diagnosis was revised after histologic examination of the renal biopsy specimen.
Clinical data for the remaining 46 patients with renal AAV at the time of diagnosis were collected. eGFR was calculated using the Modification of Diet in Renal Disease study equation (11). All patients met the Chapel Hill Consensus Conference definitions for AAV (12); follow-up data were available for 43 of 46 patients. All patients had renal involvement and 30 of 46 had undergone renal biopsy. All patients were positive at diagnosis for cytoplasmic or perinuclear ANCA and either myeloperoxidase (MPO)-ANCA or proteinase-3 (PR3)–ANCA.
The research complied with the Declaration of Helsinki and was approved by the ethics committee of our hospital. Where possible, written informed consent was retrospectively obtained from participants.
Plasma Complement Evaluation
Plasma samples, collected before immunosuppressive treatment was started, were available in 19 of 46 patients. Plasma samples of 12 age- and sex-matched healthy individuals were also collected.
The blood samples were drawn into EDTA tubes, centrifuged at 2000 rpm for 15 minutes at room temperature within 30 minutes of collection, and stored in aliquots at −80°C until use. At the time of use, the frozen specimens were rapidly thawed at 37°C and immediately transferred onto ice; plasma concentrations of C5b-9 and Bb were determined by ELISAs (Quidel, San Diego, CA) following the manufacturer’s instructions.
Histologic and Immunohistochemical Studies
Renal biopsy specimens were routinely assessed by light microscopy, direct immunofluorescence, and, when sufficient material was available, electron microscopy. For light microscopy, formalin-fixed, paraffin-embedded sections were stained with methenamine silver, periodic acid-Schiff, hematoxylin and eosin, and Masson trichrome. For direct immunofluorescence, the specimens were embedded in optimal cutting temperature (Tissue Tek; Miles Inc., Elkhart, IN) and stored in liquid nitrogen. Samples were cut to 3-µm sections, placed on poly-l-lysine–coated glass and stored at −80°C until use. These sections were rinsed in a 0.01 mol/L PBS, pH 7.4, and incubated for 30 minutes at room temperature with FITC-conjugated rabbit antihuman IgG, IgA, IgM, C3c or fibrinogen antisera (Dako, Copenhagen, Denmark). Negative controls were processed in parallel using PBS or an equivalent concentration of nonimmune rabbit or mouse serum as primary antibody. Samples were observed using a Leitz Diaplan microscope (Leica, Milan, Italy). The deposition of each antiserum was scored as “negative" (diffuse negative, focal slightly positive), “+” (diffuse slightly positive, focal moderately positive), “++” (diffuse moderately positive, focal intensely positive), and “+++” (diffuse intensely positive).
All slides were reviewed by an expert pathologist, who was blinded to the patients’ clinical data; histologic findings were classified into four categories (focal, crescentic, mixed, and sclerotic) following Berden’s classification (13). The pathologist carefully searched for signs of TMA (glomerular microthrombi composed of platelet aggregates and fibrin, endothelial cell swelling, stenosis of capillary lumens, or microthrombi in small renal arteries); TMA was scored as present or absent.
Where paraffin-embedded tissue blocks were available, additional 3-μm sections were cut to perform immunohistochemistry for the detection of C4d, C5b9, and Bb. The expression of these proteins was tested using the following primary antibodies: anti-human C4d monoclonal antibody (Ventana Medical System, reference 760–4803, dilution 1:500), anti-C5b9 polyclonal antibody (Abcam, Cambridge, United Kingdom; reference ab55811, dilution 1:50), and anti-human factor Bb monoclonal antibody (Quidel, San Diego, CA; dilution 1:20). For antigen retrieval, the histologic sections devoted to C4d staining were treated with EDTA buffer (pH of 8; Bio Optica, Milan, Italy; reference 15-M820) for 40 minutes at 90°C, those for C5b9 with citrate buffer (pH of 6; Bio Optica; reference 15-M103) for 40 minutes at 90°C, and those for Bb by means of proteinase K (Dako, Carpenteria, CA; reference S3020, ready to use) for 6 minutes at room temperature. The Kit Advance HRP (Dako; reference K4069) was used in all of the preceding three cases as a non–biotin-based, immunohistochemical visualization system. These reactions were subsequently revealed by 3,3-diaminobenzidine (Dako; reference K3468) and were mildly counterstained with Harris hematoxylin (Bio Optica; reference W01030708).
Statistical Analyses
Data are reported as mean±SD or median (range), unless otherwise stated. Two-sample differences in continuous and categorical variables were tested by Mann-Whitney test and Fisher exact test or Pearson chi-square analysis, respectively. We used the Kaplan-Meier method to estimate the crude probability of each outcome associated with low sC3 or histologic evidence of TMA. The following outcomes were analyzed: death-censored ESRD, death, and the combined endpoint of ESRD or death, with patients death-censored at the time of ESRD and the time at risk starting from the day of diagnosis; we used the log-rank test to test the equality of survivor functions. We used Cox regression models to adjust the relation between sC3, TMA, and outcome for patient’s age. We additionally adjusted these models for renal function at diagnosis. P values from Cox regression models were computed using the likelihood ratio test.
A two-tailed P value <0.05 was considered to represent a statistically significant difference. All statistical analyses were performed with Stata software, version 13 (Stata Corp., College Station, TX).
Results
Demographic and Clinical Data at Disease Onset
The main characteristics of the patients at disease onset are reported in Table 1. Of the 46 patients, 29 (63%) were men. Mean age at diagnosis was 64.7±15 years and median eGFR upon admission was 11.7 ml/min per 1.73 m2 (range, 2.4–45 ml/min per 1.73 m2). The mean Birmingham Vasculitis Activity Score was 18.2±6.5. Mean arterial pressure was 100.5±7.8 mmHg.
Main demographic and clinical characteristics of the patients with renal ANCA-associated vasculitis
Neither AAV type (microscopic polyangiitis/granulomatosis with polyangiitis) nor ANCA specificity (MPO/PR3) differed in terms of mortality or renal survival (data not shown). Data on treatment were not available for 9 patients. The remaining 37 patients were treated with conventional immunosuppressive approaches, consisting of methylprednisolone pulses followed by oral prednisone, usually combined with cyclophosphamide (35 patients) or methotrexate (2 patients). Plasma exchange was performed in 4 patients. The median follow-up was 78 months (range, 2–215 months); at last follow-up, 20 (43%) patients were dialysis dependent and 14 (30%) were dead.
Serum Complement Evaluation
Levels of sC3 and sC4 were available in 40 of the 46 patients at the time of diagnosis. As shown in Table 1, mean sC3 level on admission was 113.3±6.2 mg/dl (normal range, 90–180 mg/dl); 14 of 40 patients (35%) had low sC3 levels (Supplemental Figure 1).
Hepatic and nutritional measures of patients with low sC3 were similar to those of patients with normal sC3: the two groups did not differ for serum albumin (P=0.37), body mass index (P=0.14), pseudocholinesterase (P=0.09), total cholesterol (P=0.21), or triglycerides (P=0.15). These findings make it unlikely that low sC3 was due to deficient synthesis.
Mean sC4 was 32.1±14.1 mg/dl (normal range, 10–40 mg/dl); sC4 was low in only one patient (2%). Compared with patients with normal sC3, those with low sC3 tended to be older and to have lower eGFR at diagnosis (P=0.04 and P=0.06, respectively); neither AAV type (microscopic polyangiitis/granulomatosis with polyangiitis) nor ANCA specificity (MPO-ANCA/PR3-ANCA) was associated with the presence of low sC3 (Table 2).
Comparison of clinical and laboratory features and outcomes of patients with and without reduced serum C3
As shown in Figure 1, A and B, low sC3 was associated with a markedly lower death-censored renal survival and a non–statistically significant lower patient survival (P=0.01 and P=0.11, respectively). After adjustment for age at diagnosis, low sC3 was associated with an approximately 3-fold higher hazard of death-censored ESRD, as well as of the hazard for the combined endpoint of death or ESRD (hazard ratio [HR] for death-censored ESRD, 3.13 [95% confidence interval (CI), 1.02 to 9.58], P=0.04; HR for ESRD or death, 2.72 [95% CI, 1.01 to 7.29]; P=0.04) (Table 3). After additional adjustment for eGFR at diagnosis, the HRs for death-censored ESRD and ESRD or death were 2.77 (95% CI, 0.91 to 8.46; P=0.06) and 2.54 (95% CI, 0.95 to 6.79; P=0.06), respectively.
Markedly lower death-censored renal survival in patients with low serum C3 (sC3) versus normal sC3 and in those with versus without histologic signs of thrombotic microangiopathy (TMA). Comparison of death-censored renal survival and patient survival in patients with low serum C3 (sC3) versus normal sC3 (A and B, respectively) and in those with versus those without histologic signs of TMA (C and D, respectively). Two patients had no follow-up data and therefore were not included in the survival analyses. The Kaplan-Meier method was used to calculate survival and the log-rank test was used to compare it between the groups.
Risk of ESRD or death with respect to serum C3 levels and histologic evidence of thrombotic microangiopathy
Because sC3 reduction usually reflects cAP activation, we measured in the available sera obtained at diagnosis (19 patients) the levels of sBb and sC5b-9, whose elevation also reflects cAP activation (especially for sBb, while sC5b-9 is the end-product of all complement pathways); both sBb and sC5b-9 were significantly higher in patients with AAV than in healthy controls (P=0.02 and P<0.001, respectively) (Figure 2).
Significantly higher serum levels of Bb and C5b-9 in 19 patients with active ANCA-associated renal vasculitis compared with 12 healthy controls.
Histologic and Immunohistochemical Studies
Thirty of the 46 patients underwent kidney biopsy. The histologic and immunohistochemical data along with serum levels of complement fractions are reported in Table 2, where patients are divided into two groups according to their (normal or low) sC3 levels.
We identified, using Berden’s classification (13), 3 patients in the focal class, 19 in the crescentic class, 5 in the mixed class, and 3 in the sclerotic class. We found no significant correlation between these histopathologic classes and renal prognosis (data not shown).
We identified 8 of 30 (27%) patients with histologic findings of TMA, which coexisted with typical AAV findings (Figure 3); 2 (6%) of the patients with TMA had hemolysis at presentation, indicating full-blown aHUS associated with AAV. TMA signs tended to be more frequent in patients with low sC3; in fact, they were detected in 5 of 10 (50%) patients with low sC3 and in 3 of 20 (15%) patients with normal sC3 (P=0.04) (Supplemental Figure 1).
Renal biopsy images showing signs of thrombotic microangiopathy associated with crescentic glomerulonephritis. Thrombotic microangiopathy signs were found in small renal arterioles (A, B, and D, arrows) and in the glomeruli (C, arrow) (A and B, Masson trichrome; C and D, silver methenamine) (original magnification, 340 A–D).
The TMA group showed a lower eGFR at presentation (median, 9.3 ml/min per 1.73 m2 [range, 2.4–23.5 ml/min per 1.73 m2] versus 12.7 ml/min per 1.73 m2 [range, 23.7–45.0 ml/min per 1.73 m2]; P=0.05). No significant differences were found when TMA-positive and TMA-negative patients were compared with respect to other variables, such as age at onset, sex distribution, proteinuria, type of AAV, or ANCA specificity (data not shown). In addition, treatment approaches were similar in patients with and without TMA: Cyclophosphamide and steroids were used in 5 (71%) patients in the former group and in 17 (89%) in the latter group; likewise, plasma exchange was performed in 1 (14%) TMA-positive and 2 (10%) TMA-negative patients. The presence of TMA at renal biopsy portended a dismal renal prognosis, although it was not associated with patient survival (Figure 1, C and D). Similar to what observed with low sC3, after adjustment for age at diagnosis, TMA was associated with an approximately 2.5- to 3-fold higher hazard of ESRD, although there was no association with the hazard of death (HR for death-censored ESRD, 3.04 [95% CI, 0.99 to 9.30], P=0.05; HR for ESRD or death, 2.47 [95% CI, 0.89 to 6.84]; P=0.09) (Table 3). After additional adjustment for eGFR at diagnosis, the HRs for death-censored ESRD and for ESRD or death were 1.99 (95% CI, 0.65 to 6.08; P=0.23) and 1.96 (95% CI, 0.67 to 5.69; P=0.22), respectively. The immunohistochemical examination of renal biopsy specimens showed strong C3c staining in 11 of 27 patients (41%) and positive discrete glomerular deposition of C4d in 8 patients and of C5b-9 in 5 cases (20%), while Bb was usually negative (Figure 4). We could not find any correlation between C4d, C5b-9, and Bb immunohistochemical positivity and abnormalities in their respective circulating counterparts (sC3, sC4, sC5b-9, and Bb) (Table 4); no association or correlation was identified between tissue deposition of complement fractions and eGFR, proteinuria, or Birmingham Vasculitis Activity Score , and the same was true for ESRD or death (data not shown).
Immunohistochemical and immunogluorescence analysis of renal biopsies showing significant C5b-9, C4d and C3c deposition but almost absent Bb deposition in the glomeruli. Immunohistochemical (A–C) and direct immunofluorescence analysis (D) of glomerular deposition of complement fractions C5b-9, C4d, Bb, and C3c in patients with ANCA-associated renal vasculitis.
Histologic and immunohistochemical findings and serum levels of complement fractions in the 30 patients who underwent renal biopsy divided into low and normal serum C3 groups
Discussion
A growing body of evidence has recently accumulated about the pathogenic role of the complement system in AAV (1). In 2007, Xiao et al. (4) demonstrated how infusion of ANCA antibodies in wild-type mice could induce glomerular lesions typical of pauci-immune NCGN; interestingly, no lesion was observed when ANCA were injected in mice knocked out for C5 or complement factor B, whereas RPGN fully developed in C4-knockout mice; because C5 belongs to the terminal part of the complement cascade and complement factor B belongs to the cAP, while C4 is a key molecule of the classic pathway, these findings clearly showed that cAP plays a central role in experimental AAV (4). In keeping with this, Gou et al. showed that patients with active AAV have serologic signs of activation of the cAP (augmented serum levels of activated complement proteins, such as C3a, C5a, sC5b9, and Bb), which correlated with acute-phase reactants, number of crescents, and severity of kidney histology (9). The same group demonstrated that urinary levels of Bb correlated with the severity of renal injury (14). Moreover, in a cohort of 112 patients with AAV who underwent kidney biopsy, 37 patients showed C3 deposition, and these patients had more severe signs of renal involvement (higher proteinuria and a lower eGFR) (8).
In our study, more than one third of the patients with active renal AAV had reduced sC3 levels; this subset of patients had a dramatically worse renal prognosis, an observation that is in line with findings of a recently published study (15).
When we looked at renal deposition of complement fractions, we observed a substantial glomerular deposition of C3c and C4d and a moderate deposition of C5b-9. Data about C4d renal deposition in AAV are scarce. A previous small study on seven ANCA-positive patients (7) documented C4d negativity in NCGN, but later C4d was found to coexist with C3d in renal biopsy specimens from patients with ANCA-negative pauci-immune NCGN (16). In this study, the activation of the complement lectin pathway was hypothesized as superimposed on the activation of the cAP via the production of C3b. We failed to detect Bb deposition, although this may be due to technical reasons because we used paraffin-embedded slides rather than frozen tissue. The discrepancies between our immunohistochemistry results and those of previous studies may be due to the relatively small cohort sizes.
Overall, complement deposition in the kidney did not correlate with sC3 reduction and did not predict renal outcome. Although this may also be due to the limited sample size, it is also possible that, as has been hypothesized in other glomerulonephritides (10), complement activation would produce renal damage in AAV through two different pathways. The first may be triggered by severe renal inflammation also mediated by local complement activation, with an intense activity of chemotaxins, such as C5a and local activation of both complement pathways. The second would play an aggravating role through a systemic hyperactivation of the cAP; this would be unveiled by sC3 reduction, and the mechanisms leading to this may be related to genetically determined or autoantibody-mediated defects in cAP regulatory proteins.
We evaluated the presence of histologic and clinical signs of TMA in patients with AAV. Notably, in our case series 25% of the patients with AAV whose renal biopsy specimen was available for review had histologic signs of TMA associated with a classic pauci-immune NCGN; these patients had lower eGFR at presentation than those without TMA, and almost all of them developed fulminant ESRD unresponsive to conventional treatment. The difference in renal survival between TMA-positive and TMA-negative patients was striking. In a recent review (10), we identified >30 published cases of AAV associated with clinically overt TMA; most of them developed ESRD or died despite standard therapy.
We acknowledge that the association between low sC3/TMA and renal failure was no longer statistically significant after adjustment for eGFR. However, eGFR should be regarded as an intermediate variable in the causal pathway between sC3/TMA and ESRD. Therefore, the point estimates of the eGFR-adjusted HR might be biased toward the null value.
Abnormalities in the regulation of the complement system, and particularly defects in the proteins that physiologically control the activity of the cAP, are responsible for most cases of aHUS, a form of TMA that can be associated with autoimmune diseases, such as lupus nephritis or other GN (10). Moreover, a correlation between C4d staining and TMA in lupus nephritis was previously described (17).
In conclusion, a subgroup of patients with AAV show persistent systemic cAP dysregulation with consequent sC3 reduction, and some have associated signs of renal TMA. Both sC3 reduction and histologic evidence of TMA are associated with a worse renal prognosis, probably because this phenotype confers resistance to common immunosuppressive approaches. Evidence supporting a role for cAP activation in AAV, as emerged from our data, may provide the rationale for new therapeutic options for AAV, such as the C5a receptor inhibitor CCX168 (18), which is being tested in an ongoing trial (ClinicalTrials.gov identifier NCT01363388).
Disclosures
None.
Acknowledgments
The authors gratefully acknowledge Ms. Gabriella Becchi, Department of Biomedical, Biotechnological, and Translational Sciences (S.Bi.Bi.T.), Unit of Pathology, University of Parma, Parma, Italy, for her technical assistance with immunohistochemical analysis.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
This article contains supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.00120115/-/DCSupplemental.
- Received January 5, 2015.
- Accepted August 18, 2015.
- Copyright © 2015 by the American Society of Nephrology
References
Jump to section
More in this TOC Section
Original Articles
Cited By...
- No citing articles found.