Association between anti-complement factor H antibodies and renal outcome in primary membranous nephropathy

Background: The complement factor H (CFH) regulates activation of the alternative complement pathway. Autoantibodies against CFH are involved in progressive renal dysfunction in cases with primary membranous nephropathy (MN). However, the prevalence and roles of anti-CFH antibodies in the clinical outcome of MN patients remain unclear. Methods: We retrospectively investigated data of 36 Japanese patients with primary MN (23 men, 13 women; mean age: 64.5 [59-72] years) and 18 healthy normal controls (8 men, 10 women; mean age: 31 [27-38] years). Serum anti-CFH antibody titers were measured by enzyme-linked immunosorbent assay. Results: Anti-CFH antibody titers were signicantly higher in MN patients than in normal controls (4.69 [3.69-6.38] AU/mL vs. 0 [0-0] AU/mL, p<0.001). The patients were divided into groups: anti-CFH antibody positive group (n=28) and anti-CFH antibody negative group (n=8). No signicant difference was observed in the remission rate of proteinuria and the incidence of 30% reduction of estimated glomerular ltration rate (eGFR) or 50% elevation of serum creatinine (Cr) levels between both groups. Anti-CFH antibody titer was selected as an independent unfavorable predictor of renal dysfunction by Cox proportional hazards analysis adjusted by age, gender, serum Cr levels, proteinuria (g/gCr), anti-CFH antibody titer, and immunosuppressive therapy (adjusted hazard ratio (HR) 1.344, 95% condence intervals (CI) 1.038 to 1.741, p=0.025 for 30% reduction of eGFR; adjusted HR 1.930, 95% CI 1.108 to 3.363, p=0.020 for 50% elevation of serum Cr). Conclusions: These data suggest that anti-CFH antibodies involved in


Background
Membranous nephropathy (MN) is one of the major causes of nephrotic syndrome in adults.Histologically, reactive changes such as thickening of the glomerular basement membrane, spike formation, and stippling images are observed, and it is characterized by no associated proliferative changes.In adults, approximately 75% of cases are primary, the most common form, and approximately 25% are secondary to various causes such as tumors, infections, autoimmunity, and drugs.The course of MN is variable, and spontaneous remission is observed in about 30% of cases.However, treatment response to immunosuppressive drugs is poor and may progress to end-stage renal disease in approximately 40% of patients in Europe and North America. 1,2In Japan, the renal survival rate of primary MN judged by a requirement for hemodialysis and/or end-stage disease (ESRD) with creatinine (Cr) levels ≥ 3.0 mg/dL was 60.5% of patients after 20 years.In this report, the poor prognostic factors were no immunosuppressive therapy including steroid monotherapy, male, age (60 years or older), renal dysfunction at onset (Cr ≥ 1.5 mg/dL), severe proteinuria at onset, focal glomerulosclerotic lesions, and advanced interstitial brosis. 3he pathogenic mechanism of MN has recently been elucidated from the aspect of autoantibodies.It is known that primary MN is induced by immune complexes composed with autoantigens expressed on the cell surface of glomerular epithelial cells and IgG4 predominant autoantibodies against them.As target autoantigens, thus far, an M-type phospholipase A 2 receptor (PLA2R), thrombospondin type-1 domaincontaining 7A (THSD7A), and neural epidermal growth factor-like 1 protein have been reported, and account for approximately 70%, 1-5%, and 16% in primary MN, respectively. 4,5,6The immune complex formed in the subepithelium locally induces complement activation and the formation of the membrane attack complex (C5b-9), the nal product of the complement pathway.C5b-9 causes podocyte injury and may be involved in proteinuria in MN. 7,8 Complement factors C3 and C5b-9 are deposited along the glomerular tuft with the subepithelial immune complex. 9,10Complement activation may be a major factor in the pathogenesis of MN, although the details of complement pathway involvement have not been clari ed.
Recently, a case was reported in which renal dysfunction progressed, although anti-PLA2R antibody turned negative.In that case, the anti-complement factor H(CFH) antibody titer, an autoantibody against a complement regulatory factor, was observed to increase with renal dysfunction.As such, it was speculated that anti-CFH antibodies may have contributed to the deterioration of renal function in MN. 11 If an immune complex or complement activation is a damaging factor for disease, the complement regulatory factor is a host defense factor that leads to inactivation of activated complement.
Complement activation of MN has been reported to involve alternative pathways, 12 and CFH is a major regulator of alternative pathways.CFH possesses three complement regulatory activities: the protein inhibits formation of the AP C3 convertase, accelerates the dissociation of this complex, and acts as a cofactor for the serine protease factor I. 13 Inhibition of CFH activity by anti-CFH antibodies on the cell surface of podocytes may cause an excessive enhancement of the alternative pathway and contribute to the progression of MN pathology.However, no clinicopathological studies have been conducted thus far to examine this hypothesis in a large cohort.In this study, we investigated whether anti-CFH antibodies are a risk factor for renal function deterioration or proteinuria persistence of MN in Japanese patients, including clinicopathological factors.

Patients
We retrospectively enrolled 36 Japanese MN patients who were admitted to Kanazawa Medical University Hospital.We followed these patients for at least 6 months (median [IQR], 69.0 [41.5-99.5]months).
Diagnosis was con rmed in all patients by percutaneous needle renal biopsy.To detect secondary causes of MN, clinical workup, including detailed medical history and physical examination, serological analysis (tumor markers, autoantibodies, and infectious disease), and computer tomography (CT) scans, were conducted.The patients were treated non-randomly, depending on the judgment of the doctor in charge of each case.Serum anti-CFH antibodies were also measured in 18 healthy volunteers.The study protocol was approved by the Clinical Study Ethics Review Board of Kanazawa Medical University.
Verbal/written informed consent was obtained from all patients prior to study inclusion (Clinical Study Ethics Review Board of Kanazawa Medical University, Approval No. I425).This study was conducted according to the principles of the Declaration of Helsinki.

Measurement of serum anti CHF antibodies
Circulating antibodies to CFH were detected by using enzyme-linked immunosorbent assay (ELISA) plates coated with human complement CFH (Vidia Vestec, Czech Republic).Plates were incubated with the sera of patients and controls and revealed with horse radish peroxidase (HRP)-conjugated anti-human IgG antibody for total IgG detection according to the manufacturer's instructions.The measurable range of this ELISA kit was 3.9-250.0AU/mL.Here, the data of < 3.9 AU/mL were judged arbitrary 0 AU/mL.

Laboratory And Pathological Examinations
We summarized the methods in short.Detailed methods were described in the previous paper 14 .

Laboratory Examinations
Laboratory data such as serum creatinine and albumin levels, and urinary protein were evaluated in the standard methods.The estimated glomerular ltration rate (eGFR) was calculated by the 3-variable GFRestimating equation for Japanese individuals (194 × SCr − 1.094 × age − 0.287 × 0.739, if female).
Ehrenreich and Churg classi cation for the ultrastructural staging of MN were assessed by electron microscopic study.

Endpoint And De nition
We measured the urine protein:creatinine ratio (uPCR) and evaluated the clinical status according to Japanese clinical categories as follows: the nephrotic state, that is, the presence of heavy proteinuria (> 3.5 g/gCr) and hypoalbuminemia (< 3.0 g/dL); incomplete remission type 2 (ICR2), that is, mean proteinuria of 1.0-3.5 g/gCr accompanied by an improvement of serum albumin levels (> 3.0 g/dL); incomplete remission type 1 (ICR1), that is, mean proteinuria of 0.3-1.0g/gCr with normal serum albumin levels (> 3.0 g/dL); and complete remission (CR), that is, proteinuria of < 0.3 g/gCr with normal serum albumin levels. 15Renal function was evaluated by means of serum Cr and eGFR.Renal dysfunction was de ned as 30% reduction of eGFR and 50% and 100% elevation of serum Cr from baseline.

Statistical analysis
Continuous measures and ranking scales were summarized using medians (25-75% IQR), whereas categorical measures were summarized using counts and percentages.The Mann-Whitney U-test was used to assess the differences between both groups.Fisher's exact tests and chi-square tests were utilized to compare proportions.The Kaplan-Meier life-table method and Cox proportional hazards analyses were used to evaluate predictors of clinical outcomes.A p value < 0.05 was regarded as signi cant.

Discussion
This is the rst cohort study to examine the clinical signi cance of anti-CFH antibodies in Japanese patients with primary MN, and we obtained three new ndings.First, the anti-CFH antibody positive rate was signi cantly higher in patients with MN than control patients.Second, elevated anti-CFH antibody titers were a risk factor for renal dysfunction.Third, no correlation was observed between the amount of anti-CFH antibody titer and the amount of urinary protein.
Here, it was revealed that primary MN cases had a higher anti-CFH antibody positive rate and a signi cantly higher antibody titer than the normal control group.As for the involvement of anti-CFH antibodies in renal diseases, atypical hemolytic-uremic syndrome (aHUS) is well known.Anti-CFH antibodies have been found in approximately 10-20% of aHUS cases.Anti-CFH antibody-associated aHUS cases have a poor prognosis and are prone to recurrence, similar to un-associated aHUS cases, with approximately 50% reportedly progressing to ESRD. 16,17,18Although the appearance of anti-CFH antibodies in aHUS is reportedly mostly associated with genetic abnormalities, some are con rmed not to have genetic abnormalities, and have been observed to emerge due to an acquired factor. 19MN is predicted to emerge from an acquired factor depending on the age of onset, but as with other autoimmune diseases, the mechanism of development is unclear.Moreover, it was unclear in this study whether the appearance of anti-CFH antibodies occurred after the onset of MN or were retained before the onset of MN.Since one positive case was observed even in the normal control group, it is possible that there are cases with anti-CFH antibodies as a background pathological condition, that may be involved in the onset of MN itself.In order to clarify these, a large-scale prospective study, including normal controls, is needed.
This study revealed that anti-CFH antibody titer was an independent unfavorable predictor of renal dysfunction in MN cases.Thus far, the only report on the association between MN and anti-CFH antibodies is the case report wherein the progression of renal dysfunction was observed with the increase in anti-CFH antibody titre. 11Since an effect on renal dysfunction was observed, it is speculated that the anti-CFH antibodies had a stronger effect on glomerular and/or interstitial injuries.It was also reported from a study on aHUS that C5b-9 injures glomerular endothelial cells. 20As such, the renal injury caused by anti-CFH antibodies in this study may have involved microvascular injury, such as aHUS.In mouse models of nephrotic syndrome, IgA nephropathy, and lupus nephritis, C5b-9 deposition has been shown to be involved in renal tubulo-interstitial injury. 21,22,23Here, no difference was found in the tubulo-interstitial lesions and glomerular function at baseline between the anti-CFH antibody positive and negative patients.Further study with a larger scale is needed to elucidate the mechanism of renal injury in primary MN caused by anti-CFH antibodies.
In this study, anti-CFH antibody titer was not associated with the levels of proteinuria and clinical remission in primary MN cases.However, it was well established that C5b-9 causes podocyte injury and is involved in proteinuria in MN.Furthermore, glomerular deposition of CFB was reported in the histological ndings of MN, 9 and MN was also found in cases of C4 de ciency, 24 suggesting that an alternative pathway is involved in MN complement activation.CFH, a major regulator of the alternative pathway, is a single-chain liquid protein with a molecular weight of 150 kDa and is composed of 20 short consensus repeats (SCR).The binding site to C3b is the C-terminal domain (SCR19-20). 25It has been reported that anti-CFH antibodies suppress the regulatory function by binding to SCR15-20, a part of the CFH binding site. 26These accumulated ndings lead us to speculate that CFH may have an impact on proteinuria.However, our data did not support this speculation.A potential explanation could be that low titer anti-CFH antibodies could not su ciently reach the podocyte condition in vivo.Meanwhile, in this study, anti-CFH antibody titer was not associated with the levels of proteinuria and clinical remission in primary MN cases.Hence, it is likely that the low titer of anti-CFH antibodies may not be involved in podocyte injury.
The antibody titers obtained in this study were low compared to previous case reports on aHUS. 11In the case of aHUS, the cut-off value for anti-CFH antibody titer is reported to be 100 AU/mL. 17aHUS with a genetic abnormality or anti-CFH antibodies causes abnormal activation of an alternative pathway triggered by infection and in ammation, resulting in thrombocytopenia and renal injury due to thrombosis and hemolysis in microvessels.As the antibody titer in MN reported in this study was not as high as that of aHUS, the clinical ndings on blood vessels may be milder than that of aHUS.However, the circulating anti-CFH antibody titer was negatively correlated with proteinuria and was elevated after remission in some cases (data not shown).Anti-CFH antibodies may result in loss of urine, because two-thirds of the cases at baseline had nephrotic syndrome and anti-CFH antibody negative cases showed lower serum IgG levels (additional le 1, p = 0.057).Thus, a long-term observation is required to de ne the exact relation between circulating anti-CFH antibodies and proteinuria.It is necessary for future studies to examine with greater detail the effect of different antibody titers on vascular endothelium, glomerular, and tubulo-interstitial injuries.
The notable point of this study was that the non-invasive and simple test of anti-CFH antibody titer was able to predict poor prognosis of MN renal function at an early stage.If the baseline anti-CFH antibody titer can be used to determine the risk of developing end-stage renal failure, an appropriate treatment method can be promptly selected.
There are two main limitations in this study.First, because this was a single-center study with a small sample size, the generalizability of the results is limited.Secondly, this study showed an epidemiological causal relationship between anti-CFH antibody titer and prognosis of renal function, although the mechanism by which anti-CFH antibodies causes renal dysfunction is still unknown.Therefore, it is necessary to carry out a future follow-up study that will target a large number of MN patients at multiple centers, and to conduct basic research on the mechanism of anti-CFH antibodies causing renal dysfunction.
Hence, this study suggested that anti-CFH antibody titer was an independent unfavorable predictor of renal dysfunction in primary MN patients.It signi es that in the future, anti-CFH antibody titer could be a biomarker of renal prognosis in primary MN.

Figures
Figures

Figure 1 Comparison
Figure 1

Table 2
Characteristics of primary membranous nephropathy patients with or without anti-CFH antibodies.

Table 3
Predictors of eGFR 30% reduction in patients with primary membranous nephropathy eGFR, estimated glomerular ltration rate; HR, hazard ratio; CI, con dence interval.Data are the HR, 95% CI and p-value from Cox proportional hazard regression analyses.CFH, complement factor H; Cr, creatinine, uPCR, urinary protein to creatinine ratio.

Table 4
Predictors of serum creatinine 50% elevation in patients with primary membranous nephropathy HR, hazard ratio; CI, con dence interval.Data are the HR, 95% CI and p-value from Cox proportional hazard regression analyses.CFH, complement factor H; Cr, creatinine, uPCR, urinary protein to creatinine ratio.