Implication of Urinary Complement Factor H in the Progression of Immunoglobulin A Nephropathy

Background After activation, the complement system is involved in the pathogenesis of Immunoglobulin A nephropathy (IgAN). Complement factor H (CFH) is a crucial inhibitory factor of the alternative pathway of the complement system. The study investigated the effects of urinary CFH levels on IgAN progression. Methods A total of 351patients with IgAN participated in this study. They were followed up for an average of 51.8±26.6 months. Renal outcome was defined as a composite endpoint, that included instances of end-stage renal disease (ESRD),≥ 50% decline in estimated glomerular filtration rate (eGFR) or doubling of plasma creatinine levels. Urinary CFH levels were measured by enzyme-linked immunosorbent assay and calculated as the ratio of urinary CFH over creatinine (uCFH/uCr). Results In the whole cohort, uCFH/uCr values were associated with disease progression either as continuous [log(uCFH/uCr)] or categorical traits (dichotomous and quartile variables) after adjusting for eGFR, proteinuria, mean arterial blood pressure, histological grading and immunosuppressive therapy in the Cox proportional hazard model. Kaplan-Meier analysis showed that higher uCFH/uCr values at baseline predicted worse renal outcome during follow-up (log-rank, P<0.001). Receiver operating characteristic curve (ROC) analysis showed that log(uCFH/uCr) had predictive value for renal outcome (area under curve [AUC]=0.745), and the AUC increased to 0.805 after being incorporated into baseline eGFR and proteinuria. In subgroup analysis with eGFR≥60 mL/min/1.73m2, log(uCFH/uCr) had better predictive value (AUC= 0.724, P=0.002) for renal outcome compared to eGFR (AUC = 0.582, P=0.259) and proteinuria (AUC = 0.615, P=0.114). Conclusions Urinary CFH levels are associated with renal function decline and increased urinary CFH levels are a risk factor for progression of IgA nephropathy.


Introduction
Complement factor H (CFH) is one of the crucial inhibitory factors of the alternative pathway of the complement system, acting to accelerate inactivation of C3b [1].Studies suggest that activation of the complement alternative pathway may be involved in Immunoglobulin A nephropathy (IgAN) pathogenesis [2,3,4,5,6,7,8,9,10,11]. Our previous study and results from other studies demonstrated deposition of CFH and C3b in renal tissue of patients with IgAN, and higher levels of serum/urinary CFH were associated with disease activity [12,13,14,15]. However, it has remained unknown whether urinary CFH levels are associated with disease progression.
In this study, we measured urinary CFH levels in our IgAN cohort, analyzed the correlation between these levels and clinical/histological parameters, and evaluated the value of urinary CFH levels for predicting renal outcome.

Subjects
This study was approved by Clinical Research Ethics Committee of Peking University First Hospital. All individuals included in this study signed consent forms so that their information could be stored in the hospital database and used for research.
A total of 351patients with IgAN were included in this study. Patients who had participated in a previous CFH study were not included [15].Clinical parameters, including age, gender, blood pressure, proteinuria, plasma creatinine levels and other physical and biochemical characteristics were collected on the day of renal biopsy and during follow-up. The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation [16]. Renal pathological lesions were graded according to the Hass classification system [17]. Histological grades were divided into groups A (Hass I, II, and III) and B (Hass IV and V) [18] (Table 1).

Measurement of urinary CFH by Enzyme-linked Immunosorbent Assay (ELISA)
patients with IgAN and healthy controls were added to wells in duplicate. After incubation and washing, 1:1000 diluted mouse anti-human CFH (US Biological, Swampscott, MA, USA) was added. Alkaline phosphatase-conjugated goat anti-mouse IgG (Sigma, St. Louis, MO, USA) was added as a second antibody following three washes. The absorbance was measured at 405 nm with an EL312 Bio-Kinetics microplate reader (Bio-Tek Instruments, Inc., Winooski, VT, USA). Commercially available highly purified human factor H was used to establish the standard curve.
The urinary CFH measurements were corrected by urinary creatinine concentration, and the ratios of urinary CFH over urinary creatinine (uCFH/uCr) were used for analysis.

Detection of CFH deposition by immunofluorescence staining
Renal tissue samples were obtained from IgAN renal biopsy specimens. Paraffin-embedded blocks were cut into sections of 4-μm-thick sections and mounted on pre-coated slides. The paraffin section was deparaffinized with xylene and dehydrated with ethanol. Endogenous peroxidase activity was blocked by incubating sections in a 3% H 2 O 2 solution at room temperature for 30 min. Antigen retrieval was performed by incubating the sections in pepsin at 37°C for 1h. After blocking with phosphate buffered saline (PBS) containing 3% bovine serum albumin for 30min, the sections were incubated overnight at 4°C with primary goat anti-human CFH antibody (1:8000, diluted in PBS). Bound antibodies were detected with biotinylated rabbit anti-goat IgG. Diaminobenzidine was used as a chromogen.

Statistical analyses
Baseline features were reported as mean±SD, median (inter-quartile range [IQR]) for continuous variables, and ratios/percentages for categorical data. One-way post-hoc ANOVA using Bonferroni multiple comparison, Chi-square test, Mann-Whitney U test, and Kruskal-Wallis test was used to test for differences between groups. Spearman correlation coefficient was used to determine the correlation between urinary CFH levels and clinical variables.
A Cox proportional hazards model was used to analyze the association between urinary CFH levels and renal outcome in the whole group, as well as in subgroups with eGFR60 and eGFR<60mL/min/1.73m 2 . Both log (uCFH/uCr) and uCFH/uCrby quartiles and median were analyzed in the Cox models. The lowest group in quartiles or median was defined as the reference group in each analysis. Hazard ratio was tested in unadjusted and three adjusted Cox models. Survival curves were derived from the Kaplan-Meier analysis and compared by logrank test. Receiver operating characteristic (ROC) analysis was used to assess the predictive power of log(uCFH/uCr).
Immunofluorescence staining of CFH showed increased CFH deposition in the glomeruli of patients with IgAN compared to normal controls (S1 Fig). Urinary CFH levels associated with renal survival in the whole cohort Twenty-seven patients (7.7%) developed ESRD, 10 patients (2.8%) had a 50%eGFR decline, and 3 patients (0.85%) had doubled plasma creatinine levels during follow-up (51.8±26.6 months).
Patients were divided into four groups by quartiles of uCFH/uCr; patients in the first quartile had the lowest uCFH/uCr values(0.0-13.0ng/mg).Baseline data showed that patients in the third and fourth quartiles had more proteinuria, lower eGFR, diffuse histological lesions, and higher ESRD (Table 2).Univariate analyses revealed that clinical features, including uCFH/uCr, mean arterial pressure (MAP), eGFR, proteinuria, histological grading, and immunosuppressive therapy, were associated with renal outcome(S1 Table).We therefore performed unadjusted and adjusted models in further multivariate Cox's regression analyses. Model I was adjusted for eGFR, proteinuria and MAP. Histological grading was added to Model II based on Model I, and immunosuppressive therapy was added to the model III based on model II. In all four Cox models, uCFH/uCr was associated with disease progression as continuous and categorical traits. When the lowest quartile uCFH/uCr was selected as a reference group, the risk of kidney function decline increased in the third and fourth quartile (Table 3). Kaplan-Meier curve analysis showed that higher uCFH/uCr ratios predicted worse renal outcome during follow-up (log rank, P< 0.001) (Fig 2A).
We also analyzed uCFH/uCr by median in multivariate Cox regression. Briefly, we set up the same models described above and adjusted for the same parameters as in the whole cohort. Higher levels of urinary CFH were also associated with poor renal outcome as continuous and categorical traits (Table 3, Fig 2B) Then we separately calculated the area under curve (AUC) for log(uCFH/uCr), eGFR and proteinuria. The AUC for eGFR and proteinuria were 0.757and 0.705, respectively. Log(uCFH/uCr) was not more predictive than eGFR or proteinuria (AUC = 0.745). However the AUC increased to 0.805 after we incorporated log(uCFH/uCr), eGFR and proteinuria ( Fig  3A and 3B).  Predictive value of urinary CFH levels in the subgroup with eGFR60 mL/min/1.73m 2 We also explored whether urinary CFH levels predicted renal outcome in subgroups in early stages of CKD (eGFR60 mL/min/1.73m 2 ).
Baseline data for the subgroup are listed in Table 1. Fewer patients reached the composite end point compared with the whole group. Only 17patients reached the composite endpoint (ESRD, eGFR decline 50% of baseline, or doubling of plasma creatinine levels).
Because of limited number of patients that reached the composite endpoint, we only analyzed median uCFH/uCr in multivariate Cox regression. Briefly, we set up the same models and adjusted the same parameters as for the whole cohort. Higher levels of urinary CFH were also associated with poor renal outcome as continuous and categorical traits (Table 4).
Kaplan-Meier analysis also showed that patients with higher uCFH/uCr values had obvious lower renal survival (log rank, P< 0.001) (Fig 2C).
ROC curve analysis (Fig 3C), showed that log(uCFH/uCr) had a better predictive value (AUC = 0.724, P = 0.002) for renal outcome at early stage (eGFR60 mL/min/1.73m 2 ) compared to baseline eGFR (AUC = 0.582, P = 0.259) and proteinuria (AUC = 0.615, P = 0.114). Urinary CFH levels predicted renal survival in the IgAN cohort. The survival curve was derived from Kaplan-Meier analysis and compared by log-rank test. A. Urinary CFH levels and renal survival in the whole cohort by quartiles. Patients were divided into four groups according to uCFH/uCr quartiles; the first quartiles had the lowest uCFH/uCr values (0.0-13.0ng/mg). Patients in the third and fourth quartiles had more frequently progressed to composite endpoints. B. Median urinary CFH levels and renal survival in the whole cohort. All individuals were divided into two groups by median uCFH/uCr values. The patients with higher uCFH/uCr had increased risk of progressing to the composite endpoint. C. Median urinary CFH levels and renal survival in the subgroup of patients with eGFR60mL/min/1.73m 2 . Individuals with eGFR60mL/ min/1.73m 2 were divided into two groups by median uCFH/uCr. The patients with higher uCFH/uCrhad increased risk of progress to a composite endpoint. D. Median urinary CFH levels and renal survival in the subgroup of patients with eGFR<60mL/min/1.73m 2 . doi:10.1371/journal.pone.0126812.g002 We also investigated whether urinary CFH levels predicted renal outcome in a subgroup of patients with eGFR<60 mL/min/1.73m 2 .Urinary CFH levels were not associated with renal outcome in the 70 patients in this subgroup (Table 4 and Figs 2D and 3D).

Discussion
In this study, we measured urinary CFH levels in an IgAN cohort and found that higher urinary CFH levels were associated with lower eGFR, heavier proteinuria and diffuse pathological changes, similar to our previous study findings [15]. We also observed in the current study that baseline urinary CFH levels predicted IgAN progression as either continuous or categorical traits, in the whole cohort as well as the subgroup with eGFR60mL/min/1.73m 2 . ROC curve analysis demonstrated that high levels of urinary CFH had a similar predictive value for renal outcome with baseline eGFR and proteinuria in the whole cohort. Subgroup analysis of patients with normal eGFR 60 mL/min/1.73m 2 , revealed that urinary CFH showed a better predictive value compared with eGFR or proteinuria. In the subgroup with eGFR<60 mL/min/ 1.73m 2 , urinary CFH level were not associated with renal outcome, and renal survival was determined by eGFR.
We identified that baseline urinary CFH levels, a correlate of the inflammatory process, also predicted IgAN progression. A recent study also found that decreased circulating C3 levels and mesangial C3 deposition was associated with renal outcome in patients with IgAN [4]. Both studies supported the crucial roles of complements during the disease process, and measurement of complement factors in fluid or renal tissue could be a good marker for disease activity, prognosis, and even inform treatment with immunosuppressive agents. However detection of such markers in urine is easier, less influenced by other factors and can be repeated during extended patient follow up. IgAN repeats IgA deposition, complement activation, inflammatory status and fibrosis during its progression to ESRD. CFH, in its role as a complement-related factor, may be a good marker and alternative to repeated renal biopsy to guide treatment. However only baseline urinary CFH levels were measured in this study. Future studies should measure the variation of urinary CFH levels during patient follow-up and evaluate the effectiveness of immunosuppressive treatments to decrease urinary CFH levels and thus, improve renal outcome.
In previous studies, impaired renal function was identified as the most powerful predictive factor of renal outcome [22,23,24,25,26]. However, patients with IgAN and impaired renal function are usually in a later disease stage, and are less likely to receive appropriate and timely treatment. Thus, it is necessary to identify biomarkers that can distinguish IgAN patients with poor prognosis at early stages of IgAN, when eGFR is normal. In this study, we also observed that higher urinary CFH levels increased the risk of renal function decline in the subgroup of patients with normal eGFR. ROC analysis showed that log(uCFH/uCr) levels in IgAN had better predictive value compared with eGFR or proteinuria when eGFR60 mL/min/1.73m 2 . Therefore in addition to proteinuria, higher urinary levels of CFH may be of concerned in early stage of IgAN, and urinary CFH levels may be better indicators of inflammation status than the presence of proteinuria.
Although 351 patients were involved in this study, they were all from a single center. Since IgAN is a renal disease with slow progression, and individuals undergoing renal biopsy were usually in the early stages, few patients reached the composite endpoint in this study, especially in the subgroup with eGFR> = 60mL/min/1.73m 2 , which might result in false-positive or false-negative results. A longer follow-up time and lager cohort from additional centers are needed to validate these findings.
In conclusion, we found urinary CFH levels to be independently associated with progression of IgAN, and urinary CFH may be a valuable factor to predict progress of IgAN. These findings suggest that complement activation may be involved in the IgAN pathogenesis.