Plasma C4d Correlates With C4d Deposition in Kidneys and With Treatment Response in Lupus Nephritis Patients

Abstract

Objective: To examine whether C4d plasma levels correlate with treatment response and C4d kidney deposition in systemic lupus erythematosus (SLE) with lupus nephritis (LN).

Methods: C4d plasma levels were analyzed by a unique assay specifically detecting C4d arising from complement activation and C4 plasma levels were quantified with competitive ELISA. SLE patients with LN (71) and active SLE patients without LN (22) plus 145 controls were included. For 52 LN patients samples were available both at baseline and after immunosuppressive treatment. C4d kidney deposition was detected using immunohistochemistry in two matching kidney biopsies of 12 LN patients.

Results: In comparison to population-based controls, plasma C4d levels were significantly increased in SLE patients (0.33 mg/L versus 0.94 mg/ml, p < 0.0001) with significantly higher levels in LN patients (1.02 mg/L) than in non-renal SLE patients (0.57 mg/L, p = 0.004). The C4d/C4 ratio was also significantly higher in LN (11.2) than in non-renal SLE patients (2.5, p = 0.0002). According to ROC curve analysis, C4d was found to be an accurate marker to discriminate LN from non-renal SLE patients (p = 0.004). The C4d/C4 ratio displayed even higher specificity, sensitivity and overall accuracy as marker for LN than C4d and C4 alone. At baseline, C4d levels correlated significantly with urine-albumin to creatinine ratio (r_s = 0.43, p = 0.011) and with renal activity index (r_s = 0.37, p = 0.002). Immunohistochemical staining showed glomerular deposits of C4d in kidney biopsies, which strikingly correlated with plasma C4d levels (r_s = 0.7, p = 0.0002). Plasma C4d declined significantly after treatment in patients that experienced favorable clinical and histopathological response (p < 0.0001), while levels remained mainly unchanged in non-responders.

Conclusion: Plasma C4d discriminates LN from active non-renal SLE, correlates with C4d kidney deposits and appears valuable in monitoring responsiveness to various treatments. The C4d/C4 ratio might be superior to C4d alone.

Keywords: C4d; complement; kidney deposition; lupus nephritis; systemic lupus erythematosus; treatment response.

FIGURE 1

Plasma C4d levels are increased in patients with systemic lupus erythematosus (SLE) and discriminate lupus nephritis (LN) from non-renal SLE patients. (A) C4d levels in control subjects, IgA nephropathy patients (IgAN) and SLE patients without and with lupus nephritis. (B,C) C4 levels and C4d/C4 ratio in non-renal SLE patients and LN patients. Data are presented as medians with 25–75% quantiles plus whiskers, and significance was calculated using Kruskal-Wallis rank-sum and Mann–Whitney U tests. The C4d/C4 ratio for one of the non-renal SLE patients is zero and can thus not be displayed on a logarithmic scale. Dotted lines indicate the upper normal reference range for C4d (0.15 mg/L) and the lower normal reference range for C4 (160 mg/L). (D) Area under the ROC curve analysis showing accuracy of C4d, C3, C4, and C4d/C4 ratio as markers for LN. (E) Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and overall accuracy for C4d, C3, C4, and C4d/C4 ratio as markers for LN. The disease prevalence was set to 50%. Data are presented with 95% CIs. (F) Association of high C4d (>0.67 mg/L), low C3 (<0.77 g/L), low C4 (<154 mg/L) and high C4d/C4 ratios (>4.19) with LN. Significance was calculated using binary logistics, and ORs are indicated with a dot connected to the 95% CI.

FIGURE 2

Correlations of plasma C4d levels with clinical variables. (A,B) Correlations of plasma C4d with C3 (A) and C4 (B) in all 93 SLE patients as well as split up in the 22 non-LN and 71 LN patients at baseline. (C,D) Correlation of plasma C4d with u-ACR, activity and chronicity index as well as creatinine at baseline (C) and post-treatment (D). (A–D) Not all variables were determined for all patients; therefore n is shown in each correlation graph separately. Significant correlation coefficients and p values are bold. Significance was calculated using Spearman’s rank-order correlation test. u-ACR, urine-albumin to creatinine ratio.

FIGURE 3

In lupus nephritis (LN) patients, C4d levels associate with histopathological responsiveness to treatment and plasma C4d levels correlate with C4d deposited in kidney biopsies. (A) C4d levels and (B) C4d/C4 ratio at baseline and post-treatment of LN subgroups. Inactive patients represent ISN class I, II and III C. (C) C4d levels and C4d/C4 ratio at baseline and post-treatment in histopathological responders and non-responders. Each patient sample plus the medians and interquartile ranges are shown. (D) Specificity of C4d staining with anti-C4d-neo monoclonal Ab was determined by immunohistochemistry of Daudi cell pellets incubated with OmCI, ofatumumab and 5% human serum. Heat-inactivated human serum (5%) and 5% factor I depleted sera were applied as negative controls not able to generate C4d. (E) Scoring of C4d levels in kidney biopsies of LN patients. Representative images for each score at 40× magnification are presented. (F) Correlation of C4d plasma levels with C4d deposition in kidney biopsies. PLN samples are shown in black, MLN samples in orange and inactive samples in gray. Significance was calculated using Kruskal-Wallis rank-sum and Mann–Whitney U tests (A,B), Wilcoxon matched-pairs signed rank test (C) and Spearman’s rank-order correlation test (F). PLN, proliferative LN; MLN, membranous LN.

FIGURE 4

In lupus nephritis patients, C4d levels associate with clinical responsiveness to treatment. (A) C4d levels as well as (B) C4d/C4 ratios at baseline and post-treatment in clinical complete, partial and non-responders. Each patient sample plus the medians and interquartile ranges are shown. Significance was calculated using Wilcoxon matched-pairs signed rank test.