Direct activation of the alternative complement pathway by SARS-CoV-2 spike proteins is blocked by factor D inhibition

Abstract

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a highly contagious respiratory virus that can lead to venous/arterial thrombosis, stroke, renal failure, myocardial infarction, thrombocytopenia, and other end-organ damage. Animal models demonstrating end-organ protection in C3-deficient mice and evidence of complement activation in humans have led to the hypothesis that SARS-CoV-2 triggers complement-mediated endothelial damage, but the mechanism is unclear. Here, we demonstrate that the SARS-CoV-2 spike protein (subunit 1 and 2), but not the N protein, directly activates the alternative pathway of complement (APC). Complement-dependent killing using the modified Ham test is blocked by either C5 or factor D inhibition. C3 fragments and C5b-9 are deposited on TF1PIGAnull target cells, and complement factor Bb is increased in the supernatant from spike protein-treated cells. C5 inhibition prevents the accumulation of C5b-9 on cells, but not C3c; however, factor D inhibition prevents both C3c and C5b-9 accumulation. Addition of factor H mitigates the complement attack. In conclusion, SARS-CoV-2 spike proteins convert nonactivator surfaces to activator surfaces by preventing the inactivation of the cell-surface APC convertase. APC activation may explain many of the clinical manifestations (microangiopathy, thrombocytopenia, renal injury, and thrombophilia) of COVID-19 that are also observed in other complement-driven diseases such as atypical hemolytic uremic syndrome and catastrophic antiphospholipid antibody syndrome. C5 inhibition prevents accumulation of C5b-9 in vitro but does not prevent upstream complement activation in response to SARS-CoV-2 spike proteins.

Graphical abstract

Figure 1.

SARS-CoV-2 spike proteins induce complement-mediated cell killing that can be blocked by complement inhibitors in the mHam assay. TF1PIGAnull cells were treated with 20% NHS preincubated with diluted SARS-CoV-2 spike protein subunit 1 (S1), subunit 2 (S2), N proteins (N), and HCoV-OC43 S proteins (2.5 µg/mL to 20 µg/mL), and then measured for cell killing. Complement-mediated cell killing (%) was markedly increased in a dose-dependent manner with addition of S1 (A) and S2 (B). Increasing the concentration of N protein (C) and HCoV-OC43 S protein (D) did not increase the cell death (%) from baseline NHS level. Complement inhibition with 1 µM factor D inhibitor or 50 µg of anti-C5 monoclonal antibody completely blocked the cell killing induced by 20 µg/mL S1 (E) and S2 (F). The dotted line at 20% nonviable cells was established as a threshold for a positive mHam based on a receiver operative curve. All experiments were repeated at least 3 times. Statistical significance was calculated between each CoV protein–treated group and the NHS-treated group (*P < .05, **P < .01, ***P < .001, ****P < .0001). Anti-C5Ab indicates anti-C5 monoclonal antibody; FD Inh, factor D inhibitor (ACH145951); NHS(H), heat-inactivated NHS; NS, not significant.

Figure 2.

SARS-CoV-2 spike proteins induce C5b-9 deposition on the cell surface mainly through the alternative pathway. Flow cytometry demonstrated C5b-9 deposition on TF1PIGAnull cells after adding NHS preincubated with SARS-CoV-2 S1, S2, N, and HCoV-OC43 S proteins (2.5 µg/mL to 20 µg/mL) in either all pathway buffer (GVB⁺⁺ pH 7.4) or APC-specific buffer (GVB⁰ 10 mM MgEGTA pH 6.4). (A) SARS-CoV-2 S1, S2, and HCoV-OC43 S proteins elevated C5b-9 deposition in a dose-dependent manner in GVB⁺⁺ buffer, whereas N protein did not increase C5b-9 from the baseline NHS level. (B) Both SARS-CoV-2 S1 and S2 led to marked increase of C5b-9 depositions in APC-specific buffer. By contrast, SARS-CoV-2 N and HCoV-OC43 S proteins showed minimal C5b-9 increase in APC-specific buffer. All experiments were repeated 6 times. Statistical significance was calculated between each 20 µg/mL CoV protein-treated group and the NHS-treated group (*P < .05, **P < .01, ***P < .001, ****P < .0001).

Figure 3.

C5 and factor D inhibition block complement activation induced by SARS-CoV-2 spike proteins. Flow cytometry demonstrated C5b-9 depositions induced by 20 µg/mL SARS-CoV-2 S1 (A) and S2 (C) were completely blocked in the presence of 1 µM factor D inhibitor (ACH145951) or 50 µg of anti-C5 antibody. C3c depositions induced by 20 µg/mL S1 (B) and S2 (D) were significantly reduced by factor D inhibitor but not by anti-C5 antibody. (E) A representative flow cytometry analysis demonstrated that 1 µM factor D inhibitor (dark blue curve) completely blocked the C5b-9 (left panel) and C3c deposition (right panel) triggered by 20 µg/mL S1 proteins (red curve). C5 inhibition with 50 µg of anti-C5 antibody (cyan curve) prevented S1-induced C5b-9 deposition, but not C3c accumulation. All experiments were repeated 3 times. Unpaired Student t test P values indicate statistical significance (*P < .05, **P < .01).

Figure 4.

Serum level of factor Bb increases upon NHS activation by SARS-CoV-2 spike proteins in the presence of cells. Serum Bb concentration was measured by enzyme-linked immunosorbent assay (ELISA) after incubation of SARS-CoV-2 S1 and S2 with NHS in the presence and absence of TF1PIGAnull cells in APC-specific buffer. (A) S1 and S2 proteins significantly increased the Bb concentrations in serum when cells were present. Incubating NHS with S1 and S2 in the absence of cells did not significantly elevate serum Bb concentration from the baseline NHS level. (B) Increased Bb level in the cellular phase was completely blocked by 1 µM factor D inhibitor but not by anti-C5 antibody. All experiments were repeated 3 times. Unpaired Student t test P values indicate statistical significance (*P < .05, **P < .01, NS, not significant).

Figure 5.

SARS-CoV-2 spike proteins bind to HS and interfere with factor H function. Coincubating 5 µg/mL His-tagged SARS-CoV-2 spike proteins with 2 mg/mL HS solution completely blocked the binding of S1 subunit (A) and S2 subunit (B) to TF1PIGAnull cells. Supplementing NHS with 10 µg of purified factor H protein significantly inhibited C5b-9 (C) and C3c deposition (D) triggered by 20 µg/mL S1 in APC-specific buffer. Addition of 10 µg of purified factor H also significantly inhibited the S2-induced C5b-9 deposition (E) and C3c deposition (F) in APC-specific buffer. All experiments were repeated at least 3 times. Unpaired Student t test P values indicate statistical significance (*P < .05, **P < .01, ***P < .001, ****P < .0001). FH, purified factor H protein; NaHS, HS sodium solution.

Figure 6.

Proposed model for SARS-CoV-2 induced APC activation. Under normal conditions, factor H binds to HS on the cell surface and interacts with C3b, which facilitates factor I cleavage and deactivation of C3b. Upon infection, SARS-CoV-2 spike protein binds to HS on the cell surface and interferes with factor H function, which facilitates factor B binding to C3b and cleavage by factor D. AT III, antithrombin III; CFH, factor H; ecSOD, extracellular superoxide dismutase; FGF2, fibroblast growth factor 2; HB-EGF, heparin-binding epidermal growth factor; TGF-β, transforming growth factor β; VEGF, vascular endothelial growth factor.