Summary
Properdin is the only positive regulator of the complement system. In this study, we characterize the prevalence, functional consequences and disease associations of autoantibodies against properdin in a cohort of patients with autoimmune disease systemic lupus erythematosus (SLE) suffering from lupus nephritis (LN). We detected autoantibodies against properdin in plasma of 22·5% of the LN patients (16 of 71) by enzyme‐linked immunosorbent assay (ELISA). The binding of these autoantibodies to properdin was dose‐dependent and was validated by surface plasmon resonance. Higher levels of anti‐properdin were related to high levels of anti‐dsDNA and anti‐nuclear antibodies and low concentrations of C3 and C4 in patients, and also with histological signs of LN activity and chronicity. The high negative predictive value (NPV) of anti‐properdin and anti‐dsDNA combination suggested that patients who are negative for both anti‐properdin and anti‐dsDNA will not have severe nephritis. Immunoglobulin G from anti‐properdin‐positive patients’ plasma increased the C3b deposition on late apoptotic cells by flow cytometry. Nevertheless, these IgGs did not modify substantially the binding of properdin to C3b, the C3 convertase C3bBb and the pro‐convertase C3bB, evaluated by surface plasmon resonance. In conclusion, anti‐properdin autoantibodies exist in LN patients. They have weak but relevant functional consequences, which could have pathological significance.
Keywords: complement, lupus nephritis, properdin
Autoantibodies targeting the only positive regulator of the innate immune complement cascade were discovered in over 20% of the patients with lupus nephritis (16/71). Higher levels of anti‐properdin were related to high levels of anti‐dsDNA and ANA and to low concentrations of C3 and C4 in patients. They also correlated with histological signs of lupus nephritis activity and chronicity.
Introduction
Properdin is a plasma glycoprotein, which is the only known positive regulator of the complement by stabilizing C3 (C3bBb) and C5 [(C3b)2‐nBb] convertases of the alternative pathway [1]. Under physiological conditions it is found to form cyclic dimers (P2), trimers (P3) and tetramers (P4), as the convertase‐stabilizing activity of the tetramer is greater than the trimer [2, 3, 4].
Along with its stabilizing role for the C3 convertase, it has been shown that properdin could act as a pattern recognition molecule. Several authors have reported that properdin can recognize structures independently of C3, such as glycosaminoglycans on tubular cells leading to complement activation [5], with microbial surfaces, apoptotic and necrotic cells, providing a platform for C3 convertase assembly [6, 7, 8, 9]. The pattern recognition role remains controversial, as other authors have reported that properdin was only able to bind structures in a C3‐dependent manner [10].
The role of properdin in complement‐mediated diseases is still not clear. Properdin deficiency contributes to infectious and non‐infectious diseases in various models [11, 12, 13]. Moreover, properdin is detected in kidney biopsies and in serum/plasma/urinary samples from patients with various complement‐mediated renal diseases [14]. For example, low serum levels of properdin were detected in patients with membranoproliferative glomerulonephritis and lupus nephritis (LN) but with properdin depositions in glomeruli, implying that low properdin levels may be due to hypercatabolism [15]. Systemic lupus erythematosus (SLE) patients with low plasma levels of C3 also have low plasma levels of properdin [16]. Few studies report isolated cases of anti‐properdin autoantibodies in different pathological contexts. Józsi et al. demonstrated weak antibody positivity to properdin, C3b and factor B – to all components of the convertase – in patients with dense deposit disease (DDD) [17]. Anti‐properdin antibodies were also found in a patient with LN, carrying a heterozygous C3 mutation, together with autoantibodies against other complement alternative pathway proteins – factor I, factor B and C3 [18]. Functional assays showed that all these autoantibodies cause alternative pathway activation, which could contribute to the tissue damage in the patient’s kidney. In sera from patients with membranoproliferative glometulonephritis (MPGN) and DDD, Tanuma et al. found C3 nephritic factor (C3NeF:P), which displayed the properties of properdin and immunoglobulin (Ig)G. The authors consider that C3Nef:P is an immune complex of IgG autoantibody against properdin and properdin [19].
Because LN affects the course of the disease, the quality of the patient’s life and the prognosis of SLE [20, 21, 22], there is an unmet need of more efficient biomarkers for early diagnosis to more precisely evaluate the disease activity, the degree of disease severity and the therapy response. Here we show that autoantibodies against properdin exist in approximately 20% of the LN patients, potentiating its activity. Although probably not a driver of the disease, these autoantibodies may be a contributing factor with pathological relevance for LN.
Materials and methods
Cohort description
Seventy‐one clinically diagnosed SLE patients, according to the American College of Rheumatology (ACR) criteria with biopsy‐proven LN, all from nephrology clinics of University Hospital Tzaritza Ioanna, ISUL, Medical University of Sofia, were enrolled into the study.
LN activity was defined according to the British Islet Lupus Assessment Group (BILAG) renal score [23, 24]. All patients were divided into four BILAG categories, as follows: 23 patients (31·08%) with category A LN, 24 patients (32·43%) with category B LN, eight patients (10·81%) with category C LN and 19 patients (25·68%) with category D LN. There were no patients with category E LN in our cohort.
The patients with biopsy‐proven LN were also distributed according to the LN classification of the International Society of Nephrology (ISN) and the Renal Pathology Society (RPS) [25, 26], as follows: four patients (5·63%) had LN class I, 23 patients (32·39%) had LN class II, seven patients (9·86%) had LN class III, 25 patients (35·21%) had LN class IV, 11 patients (15·49%) had LN class V and one patient (1·41%) had LN class VI.
The presence of anti‐nuclear antibodies (ANA) was detected by indirect immunofluorescence and levels of anti‐dsDNA antibodies were tested by enzyme‐linked immunosorbent assay (ELISA) (U/ml). Pathologically elevated ANA titers (> 1 : 80) were found in 50 (69·4%) patients and pathologically elevated levels of anti‐dsDNA were found in 31 (40·8%) patients.
The C4 and C3 complement components in plasma were measured by immunodiffusion. Reference ranges were from 0·75 to 1·65 g/l for C3 and 0·20 to 0·65 g/l for C4. C3 hypocompletemia was detected in 14 (19·7%, 14 of 66) patients. C4 hypocompletemia was detected in 28 (39·4%, 28 of 71). Both C3 and C4 hypocompletemia were detected in 13 (19·8%).
Seventy‐two healthy volunteers, age‐ and gender‐matched to the patients, were included as a control group. All healthy volunteers were without autoimmune and infectious inflammatory diseases, and without renal, hepatic and hematopoietic dysfunctions.
The study had the approval of the Ethics Review Board of Medical University of Varna (protocol no. 62/04.05.2017) and each patient and healthy volunteer signed a consent form of enrollment.
ELISA for detecting anti‐properdin autoantibodies
ELISA plates (Greiner Bio‐One, Kremsmünster, Austria) were coated with either 20 µg/ml of test antigens or human properdin (Complement Technology, Inc., Tyler, TX, USA) in sodium carbonate buffer (35 mM NaHCO3, 15 mM Na2CO3, pH 9·6, 100 µl/well) overnight at 4◦C. Blocking of the plates was performed by 1% bovine serum albumin (BSA) in phosphate‐buffered saline (PBS) (200 µl/well) for 1 h at 37◦C and washed three times with PBS containing 0·05% Tween‐20 (200 µl/well). Plasmas were diluted 1 : 100 in 0·05% PBS Tween 20 (100 µl/well). After washing, horseradish peroxidase (HRP)‐conjugated anti‐human IgG (Southern Biotech, Birmingham, AL, USA) was applied at a 1 : 1000 dilution in 0·05% PBS Tween 20 (100 µl/well). After washing three times, the color was developed with 0·5 mg/ml o‐phenylenediamine (OPD) (Thermo Scientific, Fremont, CA, USA) (100 µl/well). The reaction was stopped with 2N H2SO4 (50 µl/well) and absorbance at 490 nm was measured using a Synergy 2 ELISA plate reader (BioTek Instruments, Inc., Winooski, VT, USA).
Alternatively, plasma samples were serially diluted, starting from 1 : 50 and applied on coated and blocked plates to evaluate the dose–response of the binding of the anti‐properdin IgG to their antigen.
A sample was considered positive if its optical density exceeded the average of the optical density of the samples from the healthy volunteers [± 3 standard deviations (s.d.)].
IgG purification
IgG was purified from plasma of patients with LN or healthy donors using protein G beads (GE Healthcare, Chicago, IL, USA), as recommended by the manufacturer. The concentration of the IgG was determined by a NanoDrop™ and the purity of IgG by 10% sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS‐PAGE) (Novex; Invitrogen, Life Technologies), followed by Coomassie Blue staining of the gel.
Characterization of the interaction of the anti‐properdin IgG with their antigen by surface plasmon resonance (SPR)
The interaction of IgG with properdin was analyzed in real time using a ProteOn XPR36 SPR equipment (BioRad, Marne‐la‐Coquette, France) and BiaCore2000 (GE Healthcare, Buc, France). Properdin was covalently immobilized to a GLC sensor chip (BioRad) following the manufacturer’s procedure. Alternatively, CM5 chips for BiaCore were used. Protein G‐purified IgG from LN patients or healthy donors was injected for 300 s at six different concentrations (300, 150, 75, 35, 17·5 and 0 µg/ml) diluted in 0·005% PBS Tween or 10 mM Hepes, 145 mM NaCl, 0·005% Tween 20 running buffers. The dissociation was followed for 300 s. Bound protein was regenerated with 1 M NaCl, 50 mM NaOH regeneration buffer.
Effect of anti‐properdin IgG on formation of C3 convertase
The effect of total IgG positive for anti‐properdin autoantibodies on the interaction of properdin with C3b, C3bB and C3bBb was analyzed in real time using the ProteOn XPR36 SPR equipment (BioRad). Properdin was covalently immobilized to a GCL sensor chip (BioRad), following the manufacturer’s procedure. Protein G‐purified IgG from LN patients, positive for anti‐properdin autoantibodies (but negative for anti‐C3b and anti‐FB) and healthy donors were injected for 300 s at dilution 1 : 15 in 10 mM Hepes, 50 mM NaCl, 10 mM MgCl, 0·005% Tween 20, pH 7.4 buffer. The same buffer was separately injected to serve as control. Following 300 s of dissociation, C3b (13 µg/ml), C3b (13 µg/ml) + factor B (10 µg/ml) and C3b (13 µg/ml) + factor B (10 µg/ml) + factor D (0·5 µg/ml) (Complement Technology) were injected for 300 s of association followed by 300 s of dissociation.
Effect of anti‐properdin IgG on C3 activation fragments and properdin deposition on apoptotic cells
Human endothelial cells from an umbilical vein (HUVEC) were used to study the possibility of anti‐properdin to modulate the opsonization of apoptotic cells by the complement. The cells were characterized as late apoptotic cells after detection of phosphatidylserine on the cell membrane (binding to anexin V), impaired membrane permeability (permeability for propidium iodide) and DNA fragmentation [4′,6‐diamidino‐2‐phenylindole (DAPI)]. The apoptotic cells were incubated with 1 : 10 diluted human sera mixed with IgG from LN patients with high levels of anti‐properdin or anti‐C3 autoantibodies. The dilution buffer contained 10 mM ethylene glycol tetraacetic acid (EGTA) and 7 mM MgCl2 to allow activation only of the alternative pathway. Also, IgG from patients with anti‐properdin positivity, but negative for anti‐C1q, anti‐C3b, anti‐FB and anti‐FH were selected to minimize the confounding effects of other autoantibodies. EGTA‐Mg allows activation of the alternative complement pathway by inhibition of classical and lectin pathways. After incubation for 30 min at 37°C the cells were labeled with mouse anti‐C3c antibody (Quidel, San Diego, CA, USA) or anti‐properdin antibody (Quidel), diluted 1 : 50, followed by Alexa Fluor 555‐conjugated anti‐mouse IgG antibody (1 : 100) (Thermo Fisher). The cells were analyzed by fluorescence activated cell sorter (FACS) on an LSRII machine (BD Biosciences, San Jose, CA, USA) and FlowJo software (TreeStar, Inc., Ashland, OR, USA.
Effect of anti‐properdin on alternative pathway activation in serum
IgG from LN patients, positive for anti‐properdin autoantibodies and healthy donors were added to normal human serum and incubated 1 h at 37°C to test their capacity to activate complement in fluid phase. The released Ba fragment was measured as an indicator for the formation of a C3 convertase. The test was performed according to the manufacturer’s procedure (MicroVue Ba kit; Quidel).
Prediction of the antigenic determinants
The B cell epitopes on human properdin were predicted using IEBD server B Cell Epitope Prediction Tools (prediction of linear epitopes from protein sequence; http://tools.iedb.org/main/bcell/). The crystal structure of the properdin monomer [1] was used as input file. The predicted antigenic determinants were visualized using PyMOL.
Statistical analysis
Statistical analysis was carried out using software GraphPad Prism version 6.01. Quantitative data were expressed as mean ± s.d. For comparison between groups of patients and healthy volunteers, the Mann–Whitney U‐test for continuous variables for two‐group comparison was used. Fisher’s exact test was also used for data analysis. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Spearman’s correlation was used to analyze the relativeness of the study parameters. Statistical significance was considered as P < 0·05.
Results
Anti‐properdin autoantibodies in patients with LN
The screening by ELISA revealed a presence of autoantibodies against properdin in patients with LN; 22·53% (16 of 71) of the patients were seropositive for anti‐properdin IgG autoantibodies. For positivity cut‐off the average value was taken of the autoantibodies determined against properdin in healthy volunteers ± s.d. (Fig. 1a). The IgG binding from positive patients to properdin was dose‐dependent (Fig. 1b). Detection of anti‐properdin antibodies by ELISA was validated by SPR. Purified IgG from three LN patients positive for anti‐properdin antibodies (P33, P35 and P38) showed the presence and specificity of the binding of purified IgG from patients to immobilized properdin (Fig. 1c–e). The SPR analyses revealed a slow off‐rate and lack of complete dissociation of the complexes, suggesting stable interaction.
Fig. 1.
Detection of anti‐properdin autoantibodies: (a) levels of anti‐properdin in 74 patients with LN and 72 healthy volunteers. (b) Dose‐dependent enzyme‐linked immunosorbent assay (ELISA) with anti‐properdin‐positive patients. (c) Surface plasmon resonance (SPR) sensograms of binding of purified immunoglobulin (Ig)G from seropositive patient 33 (P33), (d) from seropositive patient 35 (P35) and (e) from seropositive patient 38 (P38) to properdin. (f) SPR sensograms of binding of purified IgG from healthy volunteer (K1) and (g) from healthy volunteer (K2) to properdin.
Four of 15 (26·67%) men and 12 of 56 (21·43%) women were positive for anti‐properdin. The presence of elevated levels of anti‐properdin antibodies was not related to the patients’ sex (P = 0·112, data not shown). There was no correlation between the patients’ age (r = −0·022, P = 0·854) or duration of LN (r = −0·083, P = 0·498) and levels of anti‐properdin (data not shown).
Association of the anti‐properdin IgG with markers of disease activity
The association between levels of anti‐properdin and proteinuria, urinary sediment and renal function was investigated.
The median level of proteinuria in anti‐properdin seropositive patients was 1·56 g/24 h (from 0·05 to 15·72), and 0·32 g/24 h (from 0·02 to 8·73) in antibody‐negative anti‐properdin patients. Positive anti‐properdin patients showed a trend towards higher proteinuria than negative patients (P = 0·056, Fig. 2b), but there was no correlation between anti‐properdin and proteinuria (r = 0·165, P = 0·170, data not shown) and anti‐properdin and estimated glomerular filtration rate (eGFR) (r = −0·225, P = 0·059, Fig. 2c). Also, the presence of anti‐properdin did not determine the presence of active urinary sediment (more than eight erythrocytes/μl, more than eight leukocytes/μl or cellular casts in a non‐centrifuged urine sample) (P = 0·931, data not shown).
Fig. 2.
Statistical analysis with anti‐properdin autoantibodies: (a) levels of anti‐properdin autoantibodies in patients divided on the complex clinical laboratory estimation for lupus nephritis (LN) activity according the British Islet Lupus Assessment Group (BILAG) renal score. (b) Comparative analysis of proteinuria in anti‐properdin‐positive and anti‐properdin‐negative patients in cross‐section analysis. (c) Correlation between levels of anti‐properdin antibodies and estimated glomerular filtration rate (eGFR). (d) Levels of anti‐properdin in patients with LN depending on the presence or absence of pathological anti‐nuclear antibodies (ANA), (e) anti‐dsDNA, (f) C3 hypocomplementemia, (g) C4 hypocomplementemia and (h) anti‐C1q; + shows the presence of increased ANA, anti‐dsDNA and anti‐C1q; – shows the reference values of ANA, anti‐dsDNA and anti‐C1q in the samples. The two groups in every graphic were compared via Mann–Whitney U‐test and non‐parametric t‐test. (i) Correlation between levels of anti‐properdin and levels of ANA, (j) anti‐dsDNA, (k) C3, (l) C4 and (m) anti‐C1q in patients with LN. In order to estimate the correlations between anti‐properdin with every immunological marker, a non‐parametric Spearman’s analysis was used.
The median levels of anti‐properdin antibodies in patients positive for anti‐dsDNA levels (0·232 ± 0·292) were higher than the median level of anti‐properdin in patients negative for anti‐dsDNA (0·070 ± 0·134) (P = 0·013, Fig. 2e). Nine of 14 (64·3%) patients were positive for anti‐properdin and anti‐dsDNA. Thirty‐one of 47 (66.0%) patients were negative for anti‐properdin and anti‐dsDNA. We found a trend towards an association of the serological status of anti‐properdin with that of anti‐dsDNA (P = 0·064).
Patients with low C3 levels had a higher median level of anti‐properdin (0·243 ± 0·253) in comparison to patients with reference levels of C3 (0·108 ± 0·204, P = 0·008, Fig. 2f). Six of 14 (42·9%) patients had increased anti‐properdin levels and low levels of C3 and six of 53 (11·3%) patients had reference levels of anti‐properdin and C3. The presence of pathologically increased statistically significant anti‐properdin determines the presence of C3 hypocomplementemia with a relative risk of 3·79 (95% CI =1·44–9·95, P = 0·013).
Moderate correlations between anti‐properdin and ANA titers (r = 307, P = 0·020, Fig. 2i) and between anti‐properdin and anti‐dsDNA (r = 309, P = 0·017, Fig. 2j) were established. Weak and negative correlations between anti‐properdin and C3 (r = −256, P = 0·036, Fig. 2k) and anti‐properdin and C4 levels (r = −270, P = 0·034, Fig. 2l) were found.
For positive anti‐properdin patients, according to the BILAG renal score, six of 21 (28.6%) were in category A, six of 24 (25·0%) in category B, two of eight (25·0%) were in category and two of 18 (11·1%) in category D (Fig. 2a). Statistically significant differences in the levels of anti‐properdin in different categories of BILAG renal score have not been established (Fig. 2a).
Category A BILAG patients had a higher anti‐properdin titers in comparison to patients in other BILAG categories (Fig. 2a). The significance of anti‐properdin alone to identify patients in BILAG category A or in a group with other markers of LN activity was evaluated (Table 1). Anti‐C1q alone and in combination with anti‐dsDNA or in combination with anti‐dsDNA and levels of complements C3 and C4 showed significant specificity to identify patients with BILAG category A (Table 1). Although anti‐properdin alone could not be used for identification, BILAG category A patients (P = 0·275), in combination with anti‐dsDNA, could significantly increase sensitivity (70·7%) and negative predictive value (NPV) (89·8%), but decrease the specificity (57·3%) in the identification of patients in BILAG category A in comparison with anti‐C1q and anti‐dsDNA together (sensitivity 38·1%, specificity 91·4%, Table 1).
Table 1.
Significance of anti‐properdin antibodies alone and in combination with conventional markers for LN activity for determination of category A, according BILAG renal score
| BILAG A versus BILAG B, C, D | Specificity, % | NPV, % | Sensitivity, % | PPV, % | P |
|---|---|---|---|---|---|
| Anti‐C1q | 87·9 (188/214) | 85·5 (188/219) | 40·4 (21/52) | 44·7 (21/47) | 0·000 |
| Anti‐properdin | 18·7 (40/214) | 75·5 (40/53) | 74·5 (38/51) | 17·9 (38/212) | 0·275 |
| Anti‐C1q and anti‐properdin | 88·3 (189/214) | 85·1 (189/222) | 35·3 (18/51) | 41·9 (18/43) | 0·000 |
| Anti‐C1q and anti‐dsDNA | 91·4 (169/185) | 86·7 (169/195) | 38·1 (16/42) | 50·0 (16/32) | 0·000 |
| Anti‐properdin and anti‐dsDNA | 57·3 (106/185) | 89·8 (106/118) | 70·7 (29/41) | 26·9 (29/108) | 0·001 |
| Anti‐C1q, anti‐dsDNA and C3 and C4 | 97·8 (176/180) | 84·2 (176/209) | 19·5 (8/41) | 66·7 (8/12) | 0·000 |
| Anti‐properdin, anti‐dsDNA and C3 and C4 | 94·4 (170/180) | 85·9 (170/198) | 30·0 (12/40) | 54·5 (12/22) | 0·000 |
LN = lupus nephritis; BILAG = British Islet Lupus Assessment Group; NPV = negative predictive value; PPV = positive predictive value.
Comparative analyses between levels of anti‐properdin were made in the groups of patients with and without histological signs of LN activity and chronicity (Table 2). High levels of anti‐properdin significantly associated with renal histological lesions, such as subendothelial immune deposits of the ‘wire loop’ type (P = 0·009, Table 2), cellular (P = 0·009, Table 2) and fibrous crescents (P = 0·008, Table 2). A statistically significant correlation between levels of anti‐properdin and histological activity and chronicity indices was not found (r = 0·175, P = 0·190 and r = 0·094, P = 0·482, data not shown).
Table 2.
Comparative analysis between levels of anti‐properdin in groups of patients with and without histological signs of LN activity and chronicity
| Histological features | Anti‐factor P median (from–to) | P‐value | |
|---|---|---|---|
| Presence | Absence | ||
| Endocapillary proliferation | 0·048 (0·000–0·942) | 0·054 (0·000–0·326) | 0·797 |
| ‘Wire loop’ deposits | 0·238 (0·000–0·942) | 0·047 (0·000–0·788) | 0·009 |
| Fibrinoid necrosis/karyorrhexis | 0·076 (0·000–0·666) | 0·048 (0·000–0·942) | 0·655 |
| Cellular crescents | 0·124 (0·073–0·666) | 0·043 (0·000–0·942) | 0·009 |
| Interstitial inflammation | 0·061 (0·000–0·942) | 0·051 (0·000–0·846) | 0·740 |
| Glomerular sclerosis | 0·053 (0·000–0·788) | 0·061 (0·000–0·942) | 0·697 |
| Fibrous crescents | 0·124 (0·073–0·788) | 0·043 (0·000–0·942) | 0·008 |
| Tubular atrophy | 0·078 (0·000–0·942) | 0·045 (0·000–0·846) | 0·349 |
| Interstitial fibrosis | 0·064 (0·000–0·942) | 0·048 (0·000–0·846) | 0·829 |
LN = lupus nephritis.
Functional consequences of anti‐properdin IgG
Anti‐properdin autoantibody‐positive LN patients who showed dose–response reactivity were used for functional analysis. The presence of IgGs from patients positive for anti‐properdin showed very weak effects on the capability of properdin to bind C3b, C3bB and C3bBb. There was a weak increase in properdin binding to C3b (Fig. 3a,d,g) and pro‐convertase (C3b + factor B) (b,h) in patients 33 and 38 in the presence of anti‐properdin antibodies. These effects were weak and inconsistent among the tests and patients and hence could not be considered to affect the stabilizing function of properdin.
Fig. 3.
Functional analysis with anti‐properdin autoantibodies: (a) surface plasmon resonance (SPR) sensograms for the effect of purified immunoglobulin (Ig)G from patient 33 (P33), anti‐properdin antibody‐positive, and from a healthy volunteer (K1) on properdin binding with C3b, (b) with C3b + FB (pro‐convertase) and (c) with C3b + FB + FD (convertase) in real time. (d) The effect of purified IgG from patient 35 (P35), anti‐properdin‐positive, and from healthy volunteer (K1) on properdin binding with C3b, (e) with C3b + FB (pro‐convertase) and (f) with C3b + FB + FD (convertase) in real time. (g) The effect of purified IgG from patient 38 (P38), anti‐properdin‐positive and from a healthy volunteer (K1) on properdin binding with C3b, (h) C3b + FB (pro‐convertase) and (i) and with C3b + FB + FD (convertase) in real time. Properdin is immobilized on a SPR chip and then exposed to IgG for anti‐properdin‐positive patients (P33, P35 and P38) and IgG from a healthy volunteer (K1), followed by C3b (a,d,g), C3b + FB (b,e,h) and C3b + FB + FD (c,f,i.) addition. (j) Histogram of fluorescence activated cell sorter (FACS) analysis of C3 deposition in the presence of purified IgG from patients p (P9) and (k) purified IgG from patient 35 (P35), both positive for anti‐properdin. (l) Histogram of FACS analysis of C3 deposition in the presence of purified IgG from patients 32 (P32) and (m) purified IgG from patient 17 (P17), both positive for anti‐C3. All patients are compared with a control sample (K85).
The functional effect of anti‐properdin containing IgG on the activation of the alternative pathway in serum was measured by the release of Ba. No significant difference was detected between levels of Ba fragments in IgGs from LN patients in comparison with the IgGs from healthy volunteers (data not shown).
To explore the capacity of anti‐properdin‐positive IgG to activate complement on dying cells, purified IgGs from positive patients and healthy volunteers were incubated with late apoptotic cells in alternative‐pathway favoring conditions. In two patients (P9 and P35, Fig. 3j,k), increased deposition of C3b was detected on late apoptotic cells. Deposition of C3 fragments was not observed in the other two patients positive for anti‐properdin (P33 and P38, data not shown). They were the same patients in whom anti‐properdin antibodies weakly increased the binding of properdin to C3b (Fig. 3a,g) and pro‐convertase (Fig. 3b,h). IgGs isolated from two patients, who were negative for anti‐properdin but positive for anti‐C3 (P32 and P17), also increased deposition of C3b on late apoptotic cells (Fig. 3l,m).
Purified IgGs from the same patients (P9, P17, P32, P33, P35, P38), as well IgGs from healthy volunteers (K85, K3, K2 and collective K), were studied for their effect on properdin deposition on late apoptotic cells. The presence of patients’ or healthy donors’ IgG did not affect the deposition of properdin on late apoptotic cells (date not shown).
Prediction of epitopes of anti‐properdin
Anti‐properdin epitopes were predicted using the IEBD server (http://tools.iedb.org/bcell/; (Supporting information, Fig. S1). The majority of the binding epitopes were outside the ‘vertex’ formed at the junction of two monomers and responsible for the convertase binding. They were located in the linker regions between different vertices.
Discussion
Our study showed that anti‐properdin IgG were present in 22·5% of patients with LN, and correlated with some clinical parameters. These antibodies were specific and enhanced the deposit of C3 activation fragments on apoptotic cells.
Properdin stabilizes the alternative pathway C3 and C5 convertases and hence it is tempting to speculate that autoantibodies binding to it may enhance the stabilization capacity, increasing the half‐life of these otherwise labile enzymes. Such enhanced stabilization could result in complement overactivation and pathological consequences similar to the C3Nef found in C3 glomerulopathies [27]. LN is a hallmark of a disease with complement overactivation in the kidney, and indeed we discovered that more than 20% of the patients in our cohort were positive for anti‐properdin IgGs. Anti‐properdin IgG titres showed a trend towards negative correlation with eGFR, suggesting a possible association with renal damage. Moreover, higher levels of anti‐properdin IgG were related to high levels of anti‐dsDNA and ANA and low concentrations of C3 and C4. The correlation with C3 complement consumption could be related either to the overall autoimmunity status, where the anti‐properdin IgGs are simply an epiphenomenon, or could indicate functional relevance.
To understand whether anti‐properdin IgGs affect the functions of properdin as complement regulator, their functional consequences were characterized. The anti‐properdin IgG formed stable complexes with its target. Except to stabilize the C3bBb, it is reported that properdin binds to C3b, promoting it subsequent association with factor B [28]. We found that in some patients anti‐properdin‐positive IgG weakly increased the binding of properdin to C3b and to pro‐convertase (C3bB), and did not affect the alternative complement pathway C3 convertase, unlike C3NeF, which reacted with C3 convertase and stabilized it [19]. In a case report of anti‐properdin‐positive IgG activated complement in serum [18], we did not detect fluid‐phase complement activation by anti‐properdin IgG, contrary to autoantibodies against other alternative pathway components, such as anti‐factor B, anti‐C3b or anti‐FH [29, 30, 31, 32]. A possible explanation for the weak or absent effect of anti‐properdin IgG could be the usage of low pH elution buffer for IgG purification, which may have a dramatic effect on the biological activity of IgGs and their antigen‐binding behavior [33]. Nevertheless, purified IgG showed a strong and dose‐dependent interaction with properdin by SPR, suggesting preserved binding capacity. Another possibility is that the epitopes of anti‐properdin IgG are outside the C3‐convertase binding region of properdin. Although the in‐silico prediction of antigenic determinants requires experimental validation and has to be taken with caution, predicted epitopes showed a higher density of antigenic determinants outside the C3bBb‐binding area. This data ensemble suggests that the effect of these antibodies may be indirect, and probably affects other functions.
It has been reported that properdin binds specifically to late, but not to early, apoptotic cells, and this occurs independently of C3b [8]. We performed an analysis with late apoptotic cells in order to understand whether anti‐properdin affects the C3b and properdin deposition. Anti‐properdin IgG did not contribute to properdin deposition on late apoptotic cells in all studied patients. Nevertheless, we found that anti‐properdin increased the C3b deposition in two of four tested patients to similar levels as anti‐C3b IgG from LN patients, which have overt functional consequences [30]. Those two patients were the same in whom anti‐properdin slightly increased binding of properdin to C3b and pro‐convertase, but not to the convertase. The C3 levels in both patients were in the reference range. This suggests that in those patients there was no excessive consumption of C3 followed by increased C3b deposition. Taken together, these results suggest that anti‐properdin IgG could contribute to the complement overactivation in a subgroup of patients, but that this is not a general phenomenon, and the functional consequences of these autoantibodies are somewhat weak.
Further, we explored whether the anti‐properdin positivity could serve as a biomarker in combination with other characteristics of LN patients to predict flares and severity. Anti‐C1q are the parameters correlated more often with the renal flares in LN. It is known that they are associated with LN activity and severity with renal histological lesions [34, 35, 36, 37, 38, 39, 40]. Anti‐C1q are positively associated with BILAG renal score [41] as well as with systemic lupus erythematosus disease activity index (SLEDAI) score [42]. The combination of anti‐C1q and anti‐dsDNA was reported as a stronger marker for renal involvement and increased specificity for the identification of LN activity. Julkunen et al. found that anti‐C1q and complement C3 and C4 are more effective markers for lupus nephritis activity than anti‐dsDNA, and that anti‐dsDNA and complement C3 and C4 were more effective than anti‐C1q to evaluate the overall and non‐renal activity of SLE [43]. In our study, anti‐C1q alone and in combination with anti‐dsDNA and in combination with anti‐dsDNA and serum levels of C3 and C4 could significantly increase the specificity, but decreased sensitivity for the identification of patients in category A, according to the BILAG renal score. These findings confirmed established trends in the study by Chi et al., who evaluated the role of anti‐C1q alone and in combination with other serological markers to identified patients with active LN [35]. Anti‐properdin alone could not be a determinant for the high category of LN, according to the BILAG renal score, but in combination with anti‐dsDNA anti‐properdin could significantly increase sensitivity and NPV in the identification of patients in BILAG category A. The high NPV of anti‐properdin and anti‐dsDNA combination suggested that patients will not have severe nephritis in the absence of anti‐properdin and anti‐dsDNA. Although anti‐properdin did not associate with more active and severe LN, they were significantly associated with renal flares. We found that pathological high levels of anti‐properdin were associated with some renal histological lesions, such as ‘wire loop’ deposits, fibrous and cellular crescents.
In conclusion, we found the presence of anti‐properdin autoantibodies in the patients’ sera with LN, which correlates with clinical parameters and affects properdin function in a subgroup of patients. Although probably not a driver of the disease, these autoantibodies may be a contributing factor with pathological relevance for LN.
Disclosures
The authors have no conflicts of interest to declare.
Supporting information
Fig S1. Prediction of the B cell epitopes of properdin: A. Epitopes, predicted by the IEDB server http://tools.iedb.org/bcell/. B. Visualizaiton of the predicted peptides (red) on the surface of a properdin monomer (green)
Acknowledgements
M. R., G. M., D. I. and M. D. carried out sample preparation and performed experiments. V. V. and V. L. were responsible for the selection and diagnosis of LN patients. M. R., L. R. and G. M. wrote the manuscript. M. R., L. R. and V. V. conceived and designed the study and edited the manuscript. All authors have read and approved the final manuscript. This work was supported by the Bulgarian National Science Fund (DNTS/France 01/11/09.05.2017).
References
- 1. Pedersen DV, Gadeberg TAF, Thomas C et al Structural basis for properdin oligomerization and convertase stimulation in the human complement system. Front Immunol 2019; 10:2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. van den Bos RM, Pearce NM, Granneman J, Brondijk THC, Gros P. Insights into enhanced complement activation by structures of properdin and its complex with the C‐terminal domain of C3b. Front Immunol 2019; 10:2097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Pangburn MK. Analysis of the natural polymeric forms of human properdin and their functions in complement activation. J Immunol 1989; 142:202–7. [PubMed] [Google Scholar]
- 4. Smith CA, Pangburn MK, Vogel CW, Muller‐Eberhard HJ. Molecular architecture of human properdin, a positive regulator of the alternative pathway of complement. J Biol Chem 1984; 259:4582–8. [PubMed] [Google Scholar]
- 5. Zaferani A, Vivès RR, van der Pol P et al Identification of tubular heparan sulfate as a docking platform for the alternative complement component properdin in proteinuric renal disease. J Biol Chem 2011; 286:5359–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Kemper C, Atkinson JP, Hourcade DE. Properdin: emerging roles of a pattern‐recognition molecule. Annu Rev Immunol 2010; 28:131–55. [DOI] [PubMed] [Google Scholar]
- 7. Spitzer D, Mitchell LM, Atkinson JP, Hourcade DE. Properdin can initiate complement activation by binding specific target surfaces and providing a platform for de novo convertase assembly. J Immunol 2007; 179:2600–8. [DOI] [PubMed] [Google Scholar]
- 8. Xu W, Berger SP, Trouw LA et al Properdin binds to late apoptotic and necrotic cells independently of C3b and regulates alternative pathway complement activation. J Immunol 2008; 180:7613–21. [DOI] [PubMed] [Google Scholar]
- 9. Gaarkeuken H, Siezenga MA, Zuidwijk K et al Complement activation by tubular cells is mediated by properdin binding. Am J Physiol Renal Physiol 2008; 295:F1397–403. [DOI] [PubMed] [Google Scholar]
- 10. Harboe M, Johnson C, Nymo S et al Properdin binding to complement activating surfaces depends on initial C3b deposition. Proc Natl Acad Sci USA 2017; 114:E534–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Al‐Rayahi IA, Browning MJ, Stover C. Tumour cell conditioned medium reveals greater M2 skewing of macrophages in the absence of properdin. Immun Inflamm Dis 2017; 5:68–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Dupont A, Mohamed F, Salehen N et al Septicaemia models using Streptococcus pneumoniae and Listeria monocytogenes: understanding the role of complement properdin. Med Microbiol Immunol 2014; 203:257–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Steiner T, Francescut L, Byrne S et al Protective role for properdin in progression of experimental murine atherosclerosis. PLOS ONE 2014; 9:e92404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. van Essen MF, Ruben JM, de Vries APJ, van Kooten C. Role of properdin in complement‐mediated kidney diseases. Nephrol Dial Transplant 2019; 34:742–50. [DOI] [PubMed] [Google Scholar]
- 15. Ziegler JB, Rosen FS, Alper CA, Grupe W, Lepow IH. Metabolism of properdin in normal subjects and patients with renal disease. J Clin Invest 1975; 56:761–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Perrin LH, Lambert PH, Miescher PA. Properdin levels in systemic lupus erythematosus and membranoproliferative glomerulonephritis. Clin Exp Immunol 1974; 16:575–81. [PMC free article] [PubMed] [Google Scholar]
- 17. Józsi M, Reuter S, Nozal P et al Autoantibodies to complement components in C3 glomerulopathy and atypical hemolytic uremic syndrome. Immunol Lett 2014; 160:163–71. [DOI] [PubMed] [Google Scholar]
- 18. Nozal P, Garrido S, Martínez‐Ara J. Case report: lupus nephritis with autoantibodies to complement alternative pathway proteins and C3 gene mutation. BMC Nephrol 2015; 16:40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Tanuma Y, Ohi H, Hatano M. Two types of C3 nephritic factor: properdin‐dependent C3NeF and properdin‐independent C3NeF. Clin Immunol Immunopathol 1990; 56:226–38. [DOI] [PubMed] [Google Scholar]
- 20. Appel GB, D’Agati VD. Lupus nephritis‐pathology and pathogenesis In Wallace DJ, Hahn BH, eds. Dubois’ Lupus Erythematosus. 7th ed Philadelphia: Lippincott Williams & Wilkins; 2007:1094–112. [Google Scholar]
- 21. Gerald BA, Radhakrishnan J, D’Agati VD et al Systemic lupus erythematosus In Taal M, Chertow G, Marsden P, Skorecki K, Yu A, Brenner B, eds. Brener & Rector’s The Kidney. 9th ed Philadelphia: Elsevier Saunders; 2012:1193–208. [Google Scholar]
- 22. Appel GB, D’Agati VD. Renal involvement in systemic lupus erythematosus In: Massry S, Glassock R, eds. Textbook of Kidney Disease. St Louis, MO: Williams & Wilkins 2001; 2000:787–97. [Google Scholar]
- 23. Hay EM, Bacon PA, Gordon C et al The BILAG index: a reliable and valid instrument for measuring clinical disease activity in systemic lupus erythematosus. Q J Med 1993; 86:447–58. [PubMed] [Google Scholar]
- 24. Isenberg DA, Rahman A, Allen E et al BILAG 2004. Development and initial validation of an updated version of the British Isles Lupus Assessment Group's disease activity index for patients with systemic lupus erythematosus. Rheumatology 2005; 44:902‐6s. [DOI] [PubMed] [Google Scholar]
- 25. Weening JJ, D’Agati VD, Appel GB et al The classification of glomerulonephritis in systemic lupus erythematosus revisited. J Am Soc Nephrol 2004; 15:241–50. [DOI] [PubMed] [Google Scholar]
- 26. Markowitz GS, D’Agati VD. Classification of lupus nephritis. Curr Opin Nephrol Hypertens 2009; 18:220–5. [DOI] [PubMed] [Google Scholar]
- 27. Marinozzi MC, Chauvet S, Le Quintrec M et al C5 nephritic factors drive the biological phenotype of C3 glomerulopathies. Kidney Int 2017; 92:1232–41. [DOI] [PubMed] [Google Scholar]
- 28. Hourcade DE. Properdin and complement activation: a fresh perspective. Curr Drug Targets 2008; 9:158–64. [DOI] [PubMed] [Google Scholar]
- 29. Marinozzi MC, Roumenina LT, Chauvet S et al Anti‐factor B and anti‐C3b autoantibodies in C3 glomerulopathy and Ig‐associated membranoproliferative GN. J Am Soc Nephrol 2017; 28:1603–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Vasilev VV, Noe R, Dragon‐Durey MA et al Functional characterization of autoantibodies against complement component C3 in patients with lupus nephritis. J Biol Chem 2015; 290:25343–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Vasilev VV, Radanova M, Lazarov VJ, Dragon‐Durey MA, Fremeaux‐Bacchi V, Roumenina LT. Autoantibodies against C3b‐functional consequences and disease relevance. Front Immunol 2019; 10:64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Blanc C, Togarsimalemath SK, Chauvet S et al Anti‐factor H autoantibodies in C3 glomerulopathies and in atypical hemolytic uremic syndrome: one target, two diseases. J Immunol 2015; 194:5129–38. [DOI] [PubMed] [Google Scholar]
- 33. Djoumerska‐Alexieva IK, Dimitrov JD, Voynova EN, Lacroix‐Desmazes S, Kaveri SV, Vassilev TL. Exposure of IgG to an acidic environment results in molecular modifications and in enhanced protective activity in sepsis. FEBS J 2010; 277:3039–50. [DOI] [PubMed] [Google Scholar]
- 34. Moroni G, Quaglini S, Radice A et al The value of a panel of autoantibodies for predicting the activity of lupus nephritis at time of renal biopsy. J Immunol Res 2015; 2015:106904. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Chi S, Yu Y, Shi J et al Antibodies against C1q are a valuable serological marker for identification of systemic lupus erythematosus patients with active lupus nephritis. Dis Markers 2015; 2015:450351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Tan Y, Song D, Wu LH, Yu F, Zhao MH. Serum levels and renal deposition of C1q complement component and its antibodies reflect disease activity of lupus nephritis. BMC Nephrol 2013; 14:63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Matrat A, Veysseyre‐Balter C, Trolliet P et al Simultaneous detection of anti‐C1q and anti‐double stranded DNA autoantibodies in lupus nephritis: predictive value for renal flares. Lupus 2011; 20:28–34. [DOI] [PubMed] [Google Scholar]
- 38. Meyer OC, Nicaise‐Roland P, Cadoudal N et al Anti‐C1q antibodies antedate patent active glomerulonephritis in patients with systemic lupus erythematosus. Arthritis Res Ther 2009; 11:R87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Orbai AM, Truedsson L, Sturfelt G et al Anti‐C1q antibodies in systemic lupus erythematosus. Lupus 2015; 24:42–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Akhter E, Burlingame RW, Seaman AL, Magder L, Petri M. Anti‐C1q antibodies have higher correlation with flares of lupus nephritis than other serum markers. Lupus 2011; 20:1267–74. [DOI] [PubMed] [Google Scholar]
- 41. Marto N, Bertolaccini M, Calabuig E, Hughes G, Khamashta M. Anti‐C1q antibodies in nephritis: correlation between titres and renal disease activity and positive predictive value in systemic lupus erythematosus. Ann Rheum Dis 2005; 64:444–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Bock M, Heijnen I, Trendelenburg M. Anti‐C1q antibodies as a follow‐up marker in SLE patients. PLOS ONE 2015; 10:e0123572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Julkunen H, Ekblom‐Kullberg S, Miettinen A. Nonrenal and renal activity of systemic lupus erythematosus: a comparison of two anti‐C1q and five anti‐dsDNA assays and complement C3 and C4. Rheumatol Int 2012; 32:2445–51. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Fig S1. Prediction of the B cell epitopes of properdin: A. Epitopes, predicted by the IEDB server http://tools.iedb.org/bcell/. B. Visualizaiton of the predicted peptides (red) on the surface of a properdin monomer (green)