Identification of plasma proteins binding oxidized phospholipids using pull-down proteomics and OxLDL masking assay

Abstract

Oxidized phospholipids (OxPLs) are increasingly recognized as toxic and proinflammatory mediators, which raises interest in the mechanisms of their detoxification. Circulating OxPLs are bound and neutralized by plasma proteins, including both antibodies and non-immunoglobulin proteins. The latter group of proteins is essentially not investigated because only three OxPC-binding plasma proteins are currently known. The goal of this work was to characterize a broad spectrum of plasma proteins selectively binding OxPLs. Using pull-down-proteomic analysis, we found about 150 non-immunoglobulin proteins preferentially binding oxidized 1-palmitoyl-2-arachidonoyl-sn-glycero-phosphatidylcholine (OxPAPC) as compared to non-oxidized PAPC. To test if candidate proteins indeed can form a barrier isolating OxPLs from recognition by other proteins, we applied an immune masking assay. Oxidized LDL (OxLDL) immobilized in multiwell plates was used as a carrier of OxPLs, while mAbs recognizing OxPC or OxPE were used as "detectors" showing if OxPLs on the surface of OxLDL are physically accessible to external binding partners. Using an orthogonal combination of pull-down and masking assays we confirmed that previously described OxPL-binding proteins (non-fractionated IgM, CFH, and Apo-M) indeed can bind to and mask OxPC and OxPE on liposomes and OxLDL. Furthermore, we identified additional plasma proteins selectively binding and masking OxPC including Apo-D, Apo-H, pulmonary surfactant-associated protein B, and antithrombin-III. We hypothesize that in addition to circulating antibodies, multiple non-immunoglobulin plasma proteins can also bind OxPLs and modulate their recognition by innate and adaptive immunity.

Keywords: LDL/Oxidation/antioxidants; OxPL defense; antibodies; glycosaminoglycan; oxidation specific epitopes; oxidized lipids; phospholipases; phospholipids/metabolism.

Fig. 1

Protective action of human plasma on effects of OxPAPC. A: OxPAPC-induced IL-8 secretion was attenuated by human plasma. HUVECtert cells were incubated with increasing concentrations of OxPAPC in serum-free medium or in medium containing 1% pooled human plasma for 6 h, followed by analysis of IL-8 in supernatants by ELISA. B: Protective effect of human plasma on toxicity of OxPAPC. The cells were incubated with the indicated concentrations of OxPAPC either in serum-free medium, or in the presence of 1% pooled human plasma for 24 h. Metabolic activity of cells was quantified using the XTT assay.

Fig. 2

A number of OxPAPC-binding plasma proteins identified by pull-down proteomics also bind to heparin. A: Silver staining of plasma proteins enriched by three independent pull-down experiments using OxPAPC-liposomes, which were incubated with human plasma (diluted 1:2 in PBS) for 30 min at 37°C followed by intensive washes. Captured proteins were eluted consecutively using 8 M Urea, 1% CHAPS and Laemmli buffer. Following tryptic digest, samples were analyzed by LC-MS/MS and DIA-NN. Details of these procedures are described in the “Methods” section. B: Results from 3 independent experiments presented as volcano plots showing log2 fold-change difference (OxPAPC – PAPC) of the three different elution conditions (“UREA”, “CHAPS” and “SDS”) plotted against statistical significance determined by multiple t-tests using Perseus software. FDR = 0.05, s0 = 0.5 (0.2 for SDS). C: Functional enrichment analysis of protein hits from OxPAPC pull-downs. D: Binding to OxPAPC liposomes is not a property of high-abundant proteins. Abundance of proteins enriched via OxPAPC pull-down (right) were mapped against abundance of whole human plasma proteome (left). The dataset was derived from PeptideAtlas. E: Overlap of hits significantly enriched with two formats of pull-down: i) using OxPAPC-liposomes vs. ii) OxPAPC-coated 96-well plates. Technical details are described in the “Methods” section. F: Overlap of protein hits from pull-down using heparin-Sepharose and liposomes containing OxPAPC. G: Functional enrichment analysis of heparin-binding proteins.

Fig. 3

Heparin inhibits the OxLDL-masking activity of human blood plasma. OxLDL-coated 96-well plates were incubated for 1 h at 37°C with PBSE/BSA (10 mg/ml) with or without human plasma (1:20 dilution) and glycosaminoglycans, followed by washing and incubation with mAb 509 and anti-mouse IgM-HRP conjugate. O-phenylendiamine (OPD) and peroxide were used as substrates. Optical density of samples treated without plasma was taken as 100%. ∗P < 0.05, ∗∗P < 0.01 and ∗∗∗P < 0.001. A: Blood plasma pre-treatment reduces interaction of mAb 509 with immobilized OxLDL. The data show ‘OxLDL-masking’-effect measured in triplicates in 40 independent experiments. Statistical significance was determined by one-way ANOVA and Tukey’s post-hoc test. B: Blood plasma pre-treatment reduces interaction of mAb E06 with immobilized OxLDL. The data show ‘OxLDL-masking’-effect measured in triplicates in 15 independent experiments. Statistical significance was determined by one-way ANOVA and Tukey’s post-hoc test. C: Blood plasma treatment does not detach OxLDL from plates. OxLDL immobilized in ELISA plates was detected using anti-apoB100 primary antibody. Statistical significance was determined by one-way ANOVA and Tukey’s post-hoc test. D: Masking activity is inhibited by glycosaminoglycans (GAGs). The inhibitory activity of GAGs correlates with their negative charge and uronic acid contents. All GAGs were applied at 10 μg/ml. Hyaluronic acid (HA), chondroitin sulfate (CS), heparan sulfate (HS), dermatan sulfate (DS) and heparin were dissolved in PBSE/BSA (10 mg/ml) with or without pooled plasma (1:20) and incubated with plates containing immobilized OxLDL. Statistical significance was determined by two-way ANOVA with Tukey’s post-hoc test. E: The OxLDL-masking activity is adsorbed on heparin-Sepharose. Plasma has been separated into heparin-binding and flow-through fractions by chromatography. Note that the flow-through fraction completely lost the OxLDL-masking activity, which was recovered in NaCl-eluate. Statistical significance was determined by one-way ANOVA and Tukey’s post-hoc test.

Fig. 4

The masking assay is independent on the enzymatic activity of PAF-AH. A: The effect of recombinant PAF-AH in the masking assay was studied according to the standard procedure using incubation with PBSE/BSA (10 mg/ml) as a control or with the same solution containing indicated concentrations of recombinant PAF-AH and its inhibitor darapladib. The data are presented as mean values ± standard deviations (SD), with experiments performed in sextuplicate. Statistical significance was determined by one-way ANOVA and Tukey’s post-hoc test. ∗P < 0.05, ∗∗P < 0.01 and ∗∗∗P < 0.001. B: Inhibition of endogenous PAF-AH has minimal effect on the masking activity of plasma. OxLDL was incubated with plasma (1:20 in PBSE/BSA (10 mg/ml)) either in absence or presence of darapladib. Even at a concentration 400-fold above the reported IC50, darapladib was not able to decrease the masking effect of plasma significantly. C: The masking reaction is minimally influenced by low temperature. The masking assay was performed according to the standard procedure except that incubation of plates with plasma was performed at 37° or 4°C. Data are presented as means ± SD of triplicate measurements. Note that temperature reduction had a minimal or no effect on the masking activity.

Fig. 5

Removal of endogenous immunoglobulins reduces plasma masking capacity against OxLDL recognition by mAbs 509 and E06. A: Effect of IgG-, IgM- or IgA-removal on the masking capacity of blood plasma. mAb 509-masking assay was performed under standard conditions using control plasma or plasma depleted from IgG, IgA and IgM. The data were obtained in 12 independent experiments and are presented as means ± SD. Statistical significance was determined by one-way ANOVA and Tukey’s post hoc test. The Western blot demonstrates the efficiency of the immunoglobulin depletion procedures. B: Experiment with immunoglobulin-depleted plasma samples was performed by the standard masking assay but using mAb E06 recognizing another class of OxPLs (see “Methods”). The data are representative of 5 independent experiments and are presented as means ± SD. Statistical significance was determined by one-way ANOVA and Tukey’s post-hoc test. The Western blot demonstrates the efficiency of the immunoglobulin depletion procedures. C: IgM that was adsorbed on columns, was eluted by changing to acidic pH followed by neutralization. Eluted IgM was tested in a standard mAb 509-masking assay at concentration corresponding to plasma diluted 1:20. Absence of IgM in flow-through and presence in eluate was confirmed by western blotting. D: Masking activity of commercially purchased IgM at concentrations in the range of 20-fold diluted human plasma. Data are presented as means ± SD. Statistical significance was determined by one-way ANOVA and Tukey’s post hoc test. E: Non-immunoglobulin heparin-binding proteins play an important role in the mechanisms of the masking effect. The masking activity data and anti-IgM western blotting show that the eluate from heparin-Sepharose containing heparin-binding proteins demonstrated strong masking activity in spite of the lack of IgM.

Fig. 6

Complement factors C1q, C3 and factor B do not have masking activity. A: Masking assay using mAb 509 was performed according to the standard protocol described in “Methods” using pure complement factor H at 25 μg/ml. B: The masking experiment was performed using factors C1q, factor B and C3 (7, 20 and 130 μg/ml, respectively) alone or in combination. The incubation was performed in the presence of non-saturating dilution of plasma. The presence of low plasma concentration allows to detect i) intrinsic activities of complement factors and ii) combined activities of complement with immunoglobulins in case if complement and immunoglobulins are active only in combination with each other.

Fig. 7

Analysis of selected proteins enriched with OxPAPC-liposome pull-down for mAb 509-masking activity. Several hits from the pull-down experiments that were commercially available were tested in a standard masking assay using mAb 509. Incubation of immobilized OxLDL with recombinant or purified proteins was performed in PBSE/BSA (10 mg/ml) buffer containing physiological concentrations of the following proteins: Apo-D (10 μg/ml), Apo-H (20 μg/ml), Apo-M (5 μg/ml), PAF-AH (0.25 μg/ml), SP-B (5 μg/ml), ATIII (7.5 μg/ml).