Figure 1.
Anti-properdin antibodies show high avidity towards human properdin.
The binding strengths of newly generated mAb 1340 as well as commercially available mAb A233 and mAb A235 were tested in an indirect ELISA. MAbs were serially diluted (1,000–0.1 ng/mL) and properdin binding was detected with a peroxidase conjugated anti-mouse IgG antibody. MAb 1340 (EC50 25 ng/mL), mAb A233 (EC50 11 ng/mL) and mAb A235 (EC50 15 ng/mL) showed comparable binding strengths towards immobilized human properdin. Shown are means (n = 9±s.e.m.) out of three independent experiments each with three replicates. After background subtraction, data were normalized to 1,000 ng/mL mAb 1340 reactivity against properdin (set to 100%).
Figure 2.
Monoclonal antibodies specifically detect properdin in human serum.
(A) Purified human properdin, human serum, recombinant mouse properdin, human embryonic kidney (HEK) cell supernatant and fetal calf serum were immobilized on an ELISA plate. In house mAb 1340 (1 µg/mL) and commercial mAb A235 (1 µg/mL) detected only purified human properdin or human serum, respectively. MAb A233 (1 µg/mL) detected all antigens tested. Shown are the respective means (± s.e.m.) for two independent experiments. After background subtraction data were normalized to mAb 1340 reactivity against properdin (set to 100%). (B) Human properdin was isolated from human serum by immunoprecipitation (IP) using mAb 1340 (left lane). The precipitated proteins and purified control properdin (right lane) were separated by non-reducing, denatured SDS-PAGE. Western blot detection was performed with mAb 1340 and a peroxidase conjugated anti-mouse IgG antibody. The generated mAb precipitated and detected properdin monomer (∼55 kDa), dimer (∼110 kDa) and trimer (∼165 kDa), respectively.
Figure 3.
MAb 1340 precipitates properdin, complement factor 3, clusterin and immunoglobulins from human serum.
(A) Proteins from human serum binding to mAb 1340 were isolated by coimmunoprecipitation. After size separation in an SDS-PAGE, protein fractions were identified using Coomassie staining and combined LC-MS/MS analysis. MAb 1340 precipitates complement factor C3(H2O) (>170 kDa) as well as C3-fragments (<170 kDa), clusterin (53 kDa), immunoglobulins (150 kDa, murine IgG (mIgG), human IgG (hIgG)) and properdin (55 kDa). C3 fragments detected at different molecular sizes, included: C3 β chain, C3c alpha chain fragments 1 and 2, C3dg at >170 kDa; C3 β chain, C3c alpha chain fragments 1, C3dg at <170 kDa and C3 β chain at 68 kDa. The confidence index for each identified protein is given by the molecular weight search (MOWSE) score. (B) Exemplary MS/MS-spectrum matching the properdin peptide YPPTVSMVEGQGEK (amino acids 409–422) is shown. Detected C-terminal ions (y-ions) are annotated within the peptide sequence.
Figure 4.
MAb 1340 inhibits function of the alternative complement system.
(A) Effect of mAb 1340 on the alternative complement system was tested in a hemolysis assay. Human normal serum was preincubated with different antibodies (0.03–50 µg/mL). Lysis of sheep erythrocytes was measured at 414 nm after addition of the serum-mAb mixtures. Lysis of the erythrocytes was reduced after adding increasing concentrations of mAb anti-properdin (mAb 1340 (blue), mAb A235 (green)). Other complement system specific antibodies resulted in an inhibition of the complement system at higher concentrations (mAb anti-Ba (grey), mAb anti-C5 (red)). We observed lysis of sheep erythrocytes after incubation with a non-specific mAb anti-botulinum neurotoxin (BoNT, orange, isotype control). (B) Inhibition of NHS-associated E. coli lysis by mAb was tested. E. coli were cultivated in media with EGTA containing NHS and mAb in different concentrations (0.1–33 µg/mL). Percentages of viable E. coli cells were calculated in comparison to an untreated control. The E. coli lysis assay showed comparable inhibitory results with the hemolysis assay. MAb A235 showed a decrease in viable E. coli at concentrations above 1 µg/mL. (C) Blocking activity of different anti-complement mAb (6.6 µg/mL, anti-BoNT as isotype control) for EGTA containing NHS (AP alternative pathway, black columns) and NHS (CP classical pathway, white columns) mediated lysis of E. coli was analyzed. Anti-properdin mAb (mAb 1340, mAb A235) inhibited the alternative pathway but not the classical pathway. MAb anti-C5 blocked the lysis of E. coli mediated by alternative and classical pathway. For all figures means out of two independent experiments (n = 4±s.e.m., except for 4B concentrations 0.08, 0.52, 2.6, 12.1 µg/mL n = 2±s.e.m.) are shown.
Figure 5.
(A) A model of mAb 1340 secondary structure shows the heavy (blue) and light (red) variable chain. The binding cleft is determined by six CDR loops (marked are overlapping amino acids from Table 1) either in the heavy chain H1 (orchid), H2 (salmon), H3 (magenta) or in the light chain L1 (dark green), L2 (chartreuse), L3 (springgreen). The distance between the chains varies between 11.66–24.64 Å. Modeling was performed with Rosetta. (B) Spherical display of mAb 1340 light (light grey) and heavy (dark grey) variable chain. The predicted amino acids for antigen contact are highlighted. The coloring of the residues is according to their contact probability value (the higher the probability the deeper the color). Modeling was performed with proABC.
Table 1.
Single letter amino acid sequence of complementarity-determining regions (CDR) of mAb 1340.
Figure 6.
A sandwich ELISA detected human properdin from blood samples highly sensitively, specifically and reproducibly.
(A) Serial dilutions of human normal plasma and serum pools (1∶10–1∶156,250) were used as a reference curve for properdin detection from human samples in a sandwich ELISA. Detection of unknown samples was reliable in a plasma and serum dilution range from 1∶100 to 1∶1,000. The EC50 of the plasma curve was at 1∶800 and for serum at 1∶400, respectively. The mean value of three independent measurements with duplicates is shown. The reference curves were representable for all performed experiments. (B) Recovery rates of 100 ng/mL properdin from properdin-depleted NHP and NHS are shown. Plasma matrix did not interfere with properdin recovery. Serum dilution 1∶50 and higher gives a recovery rate between 75–125%. (C) The described sandwich ELISA specifically detected human properdin only either purified or from blood samples. Properdin from mouse, rat or calf serum showed no signal in an ELISA with immobilized antigens. Other negative controls such as human blood depleted from properdin, human normal urine or bovine serum albumin were not detected. Two independent experiments with duplicates were performed. (D) Intra-assay and inter-assay coefficient of variation for different plasma and serum concentrations in the sandwich ELISA displayed high reproducibility. Serial dilutions of a human normal plasma and serum pool (1∶10–1∶156,250) were determined. The sandwich ELISA shows CV values below 25% for plasma dilutions from 1∶10 to 1∶10,000 and for serum dilutions 1∶10 to 1∶1,000, respectively.
Figure 7.
Systemic properdin concentrations in patient serum compared to healthy controls.
Serum samples of healthy blood donors (controls, n = 26), patients with age-related macular degeneration (AMD, n = 20), systemic lupus erythematosus (SLE, n = 6), connective tissue diseases (CTD, n = 10), polymyalgia rheumatica (PR, n = 31), rheumatoid arthritis (RA, n = 38), spondyloarthritis (SPA, n = 40) and systemic sclerosis (SSc, n = 16) were diluted in PBS (1∶700). Properdin amount in serum samples was compared to a positive control (PC, NHS pool 1∶700, ratio on y-axis) using the described ELISA for human properdin (see Figure 6). Patient and control groups showed a properdin amount in serum between 66–155% of the positive control. This corresponded to 13–30 µg/mL properdin. There was no significant difference in properdin concentration between the cohorts (two-tailed, paired t-test, P>0.001).