Urinary soluble CD163 is useful as "liquid biopsy" marker in lupus nephritis at both diagnosis and follow-up to predict impending flares

Abstract

Lupus nephritis (LN) diagnosis and follow-up requires noninvasive biomarkers. Therefore, the added value of coupling the urinary soluble (s)CD163/creatinuria ratio with serological markers was evaluated in a real-world clinical practice. To this end, a monocentric and retrospective study was conducted in 139 SLE patients with biopsy-proven nephritis having an active LN (LN-A, n = 63 with a positive SLEDAI-renal score) or inactive (n = 76), as well as 98 non-renal SLE patients. The urinary sCD163/creatinuria ratio outperformed serological markers for predicting LN-A (AUC>0.972; p < 10^-4 with a 100 % specificity threshold fixed at 320 ng/mmol), and for monitoring renal activity allowing prediction of impending flares and remissions in follow-up (AUC = 0.789, p < 10^-4). LN-A patients with an elevated spot proteinuria/creatinuria ratio (p = 8 × 10^-6) and sCD163/creatinuria ratio (p = 10^-3) were at risk for developing end-stage kidney disease but sCD163/creatinuria ratio cannot substitute kidney biopsy to discriminate LN-A from other glomerulonephritis. Among serological markers (n = 14), anti-dsDNA and anti-C1q antibodies (Abs) (AUC>0.750 versus non-LN patients, and AUC>0.640 versus LN-IR patients) best predicted LN-A, and higher levels were retrieved in class III/IV proliferative LN-A. In multivariate logistic regression analysis, the urinary sCD163/creatinuria ratio remained the only statistically significant biomarker to predict LN-A (p < 0.001). In conclusion, and as compared to classical serological markers, the urinary sCD163/creatinuria ratio provides an additional parameter for monitoring LN patients.

Keywords: Biomarkers; End stage kidney disease; Flare; Remission; Urinary soluble CD163; lupus nephritis.

Fig. 1

Serological markers and urinary soluble (s)CD163/creatinuria utility for active lupus nephritis (LN-A). A/B: Anti-C1q and anti-dsDNA autoantibody (Ab) levels in patients with LN-A versus inactive/remission LN (LN-IR) and non-renal patients (non-LN), the number of patients is added in brackets. The dotted lines reflect the manufacturers' cut-offs for anti-C1q Abs (20 arbitrary units [AU]/mL) and anti-dsDNA Abs (10 international units [IU]/mL). C: Area under the curve (AUC) values obtained from receiver operating characteristic (ROC) curves of 11 Abs (targeting C1q, dsDNA, chromatin, SSA 52 kDa, SSA 60 kDa, SSB, Sm, SmRNP, RNP 68, RNP A, and ribosomal antigens) and 3 complement parameters (C3, C4 and CH50) comparing LN-A patients from LN-IR patients (red dots) or from non-renal SLE patients (green dots). D/E: Urinary sCD163/creatinuria (sCD163/Cre) and sCD163/proteinuria (sCD163/Prot) ratio levels in patients with LN-A versus LN-IR and non-LN. The dotted line reflects the 100 % specificity threshold fixed at 320 ng/mmol for urinary sCD163/Cre ratio and at 2.2 ng/g for urinary sCD163/Prot ratio. F: AUC values obtained from ROC curves of urinary biomarkers including spot protein excretion to creatinine ratio (PCR) and sCD163 expressed as a ratio to creatinuria (sCD163/Cre), proteinuria (sCD163/Prot), or not (sCD163 total). LN-A patients are compared to the LN-IR patients (red dots) or to the non-renal SLE patients (green dots). F: ROC curves of the 4 urinary biomarkers. The p values are indicated.

Fig. 2

Proteinuria effect on urinary sCD163 estimation. A: Correlation between spot urinary sCD163/creatinuria (sCD163/Cre) with protein excretion to creatinine ratio (PCR) in patients with active lupus nephritis (LN-A in red, n = 50), inactive/remission LN (LN-IR in blue, n = 60), and non-LN patients (non-LN in green, n = 80). Spearman's rho values are indicated. B: Urinary sCD163/Cre levels in patients with LN-A versus LN-IR and non-LN according or not to the presence of a proteinuria (PCR≥ 1 g/g), p values are indicated when significant. C: Correlation between spot urinary sCD163/proteinuria (sCD163/Prot) with PCR. D: Urinary sCD163/Prot levels in LN-A, LN-IR and non-LN patients according or not to the presence of a proteinuria (PCR≥ 1 g/g).

Fig. 3

Anti-C1q/dsDNA antibodies (Ab) levels but not urinary sCD163/creatinuria levels discriminate proliferative from membranous active lupus nephritis (LN-A). A: Area under the curve (AUC) values obtained from receiver operating characteristic (ROC) curve of 11 IgG autoantibodies (targeting C1q, dsDNA, chromatin, SSA 52 kDa, SSA 60 kDa, SSB, Sm, SmRNP, RNP 68, RNP A, and ribosomal antigens), 3 complement parameters (C3, C4 and CH50), and 2 urinary parameters (spot protein to creatinuria ratio [PCR], and urinary sCD163 normalized to creatinuria) comparing proliferative LN-A (class III/IV ± V) from membranous LN-A (isolated class V). Biomarkers presenting good (0.7 ≤ AUC<0.89, p < 0.05) performance to discriminate proliferative LN-A are presented: anti-dsDNA Abs (B), anti-chromatin Abs (C), anti-C1q Abs (D), and anti-Sm Abs (E). The dotted line reflects the histological cut-offs for anti-dsDNA Abs (60 international units [IU]/mL, 100 % sensitivity) and anti-C1q Abs (40 arbitrary units [AU]/mL, 100 % sensitivity), the number of patients is added in brackets, ROC curves are included for anti-dsDNA and anti-C1q Abs, and p values are indicated when significant. F: Spearman's correlation matrix between the 16 biomarkers to test their associations in LN-A.

Fig. 4

Serological markers and even more urinary sCD163/creatinuria (sCD163/Cre) levels are correlated with renal disease activity (SLEDAI-R) in lupus nephritis (LN) patients. A: Weak correlation between anti-C1q Abs and SLEDAI-R in a cross-sectional analysis (n = 122), r and p values are reported. B: Very good correlation between urinary sCD163/Cre levels and SLEDAI-R in a cross-sectional analysis (n = 117). C: Spearman's absolute rho [abs (rho)] values obtained between SLEDAI-R and urinary biomarkers (spot protein to creatinuria ratio [PCR], and urinary sCD163/Cre) or serological biomarkers (IgG anti-dsDNA Abs, anti-chromatin Abs, anti-C1q Abs, and complement C3, C4, CH50). D: Changes (delta) in anti-C1q Abs correlated weakly in LN patients (n = 92) with the evolution of disease activity at follow-up (delta SLEDAI-R). E: in contrast, delta-urinary sCD163/Cre levels presented a good correlation with delta SLEDAI-R (n = 61). F: Spearman's absolute rho (abs (rho)) values obtained between delta SLEDAI-R and changes (delta) in urinary and serological biomarkers during LN follow-up. According to the variations in SLEDAI-R, delta SLEDAI-R is considered to reflect therapeutic response/remission (negative delta value), an inactive/stable disease (unchanged delta value), and flare (positive delta value).

Fig. 5

Urinary sCD163/creatinuria (sCD163/Cre) levels but not serological markers are weakly correlated with the evolution of the renal function (eGFR) in lupus nephritis (LN) patients. A-B: In cross-sectional analysis, a weak correlation was found between eGFR and spot protein to creatinuria ratio (PCR, n = 139) (A), and with urinary sCD163/Cre ratio levels (n = 117) (B). C: Spearman's absolute rho [abs (rho)] values obtained between eGFR and urinary biomarkers or serological biomarkers. D-F: No significant correlation (p < 0.05) between urinary and serological biomarkers with delta eGFR during LN follow-up. According to the variations in eGFR, delta eGFR is considered to reflect end stage kidney disease evolution (ESKD, negative delta value), a stable evolution (unchanged delta value), and kidney repair with time (positive delta value).