Complement C4 Prevents Viral Infection through Capsid Inactivation

Abstract

The complement system is vital for anti-microbial defense. In the classical pathway, pathogen-bound antibody recruits the C1 complex (C1qC1r₂C1s₂) that initiates a cleavage cascade involving C2, C3, C4, and C5 and triggering microbial clearance. We demonstrate a C4-dependent antiviral mechanism that is independent of downstream complement components. C4 inhibits human adenovirus infection by directly inactivating the virus capsid. Rapid C4 activation and capsid deposition of cleaved C4b are catalyzed by antibodies via the classical pathway. Capsid-deposited C4b neutralizes infection independent of C2 and C3 but requires C1q antibody engagement. C4b inhibits capsid disassembly, preventing endosomal escape and cytosolic access. C4-deficient mice exhibit heightened viral burdens. Additionally, complement synergizes with the Fc receptor TRIM21 to block transduction by an adenovirus gene therapy vector but is partially restored by Fab virus shielding. These results suggest that the complement system could be altered to prevent virus infection and enhance virus gene therapy efficacy.

Keywords: TRIM21; adenovirus; complement; complement C4; complement-mediated neutralization; gene therapy; host-pathogen; humoral immunity; neutralizing antibodies; non-enveloped virus.

Graphical abstract

Figure 1

Complement Components C1 and C4 Mediate Potent Antibody-Dependent Neutralization of Ad5 (A) Neutralization of Ad5 in 293T-WT and TRIM21 KO cells using 9C12-WT and mutants LALA and P329A (left). The percentage of TRIM21-dependent and -independent neutralization using 9C12-WT at 15 μg/mL and 0.12 μg/mL (middle), as well as the percentage of TRIM21-dependent and -independent neutralization using each antibody at 15 μg/mL (right), is depicted. (B) Relative infection (left) and C1q dependent neutralization (right) in HeLa TRIM21 KO cells using 9C12-WT, LALA, or P329A and the indicated serum. (C) Neutralization of Ad5 in HeLa TRIM21 KO cells using different ratios of 9C12-WT and P329A. (D) Neutralization of Ad5 in HeLa TRIM21 KO cells. Ad5 was incubated with the indicated serum for the depicted amount of time. (E) Neutralization of Ad5 in HeLa TRIM21 KO cells. Ad5 was incubated with NHS or NHS depleted of IgG for the depicted amount of time (left). Western blot showing IgG depletion of NHS (right). Original western blots are included in Figure S6. (F) Neutralization of Ad5 in HeLa TRIM21 KO cells in the presence of human serum depleted of different complement components using 9C12-WT and mutant P329A. (G) Neutralization of Ad5 in the presence of 9C12-WT using 200 ng/mL C1 and the indicated concentrations of C4. Error bars depict the mean ± SEM of nine replicates acquired in three independent experiments (B, C, F, and G); mean ± SEM of six replicates acquired in two independent experiments (A and D); mean ± SD of three replicates acquired in one representative experiment (E).

Figure 2

Activation of the Complement Cascade in Presence of Ad5 and 9C12 Results in Deposition of C4b on the Ad5 Capsid (A) Western blot of C4 (α-chain) cleavage in NHS (top) or C1q-depleted serum (bottom) in the presence of Ad5 and 9C12-WT over 60 min. See also Figure S2E. (B) Experimental set up (left): Ad5-mCherry+9C12-WT were incubated with HBS++/serum for 1 h at 37°C. Ad5-GFP was incubated with HBS++. Ad5-mCherry was added to HeLa TRIM21 KO cells immediately followed by Ad5-GFP. Right: relative infection of Ad5-mCherry and Ad5-GFP. x axis labeling corresponds to the buffer/serum that was incubated with Ad5-mCherry. (C) Western blot of C4 (α-chain) cleavage in NHS, C1q-depleted serum, or C4-depleted serum in the presence of Ad5 and 9C12-WT as indicated. (D) Elisa for C4 using serum (left) or purified protein (right). Ad5+9C12-WT were incubated with the indicated serum or purified protein and pelleted over a sucrose gradient. Error bars depict mean + SD of three replicates acquired in one representative experiment (A and D). Original western blots are included in Figure S6.

Figure 3

C1 and C4 Do Not Prevent Virus Attachment or Internalization Ad5 associated with HeLa TRIM21 KO cells after 30 min of continuous infection in the presence of C1 and C4 (A–C). (A) The viral copy number was determined by qPCR. (B) Western blot for adenovirus and actin. (C) Cells were stained with an anti-human IgG antibody and analyzed by flow cytometry. Histograms (left) and MFI (right) are depicted. (D) Internalization assay of virus after 30 min of continuous infection. Cells were harvested every 30 min, stained with an anti-human IgG antibody, and analyzed by flow cytometry (left). See also Figure S3C. qPCR of viral genomes present at 30 min and 90 min post infection (right). Viral genomes were normalized to the number of viral genomes present at 30 min post infection. Error bars depict the mean + SD of three replicates acquired in one representative experiment (A and C) or mean + SEM of nine replicates acquired in three independent experiments (D).

Figure 4

C1 and C4 Prevent Adenoviral Protein VI Exposure and Endosomal Escape (A) IP of Ad5 with 9C12 after Ad5 was incubated at the indicated temperatures for 30 min. WB: anti-adenovirus. (B) IP after Ad5 and 9C12 were complexed with C1 or C1/C4 and then incubated at 37°C or 49°C for 30 min. WB: anti-adenovirus. (C and D) HeLa TRIM21 KO cells were infected for 30 min in the presence of 9C12 or 9C12+ complement. Error bars depict the mean ± SEM of the indicated number of cells (n) acquired in three independent experiments. Scale bar, 5 μm. (C) Left: Ad5 staining is displayed in green; protein VI staining is depicted in red. Right: quantification of protein VI puncta per cell in the indicated conditions. (D) Left: Ad5 staining is displayed in green; Galectin-3 staining is depicted in red. Right: quantification of Galectin-3 puncta per cell in the indicated conditions. Original western blots are included in Figure S6.

Figure 5

C1 and C4 Prevent Ad5 Entry into the Cytosol and Trafficking to the Nucleus (A) Cytosolic Ad5 staining is displayed in green, total Ad5 staining is displayed in red (left). Percentage of cytosolic Ad5 after infection with Ad5 alone or Ad5+9C12 (middle). Percentage of cytosolic Ad5 after infection with Ad5+9C12 or Ad5+9C12+C1/C4 (right). (B) Ad5 staining is displayed in green; lamin B1 staining is depicted in red (left). Quantification of viruses at the nuclear membrane (right). Scale bar, 5 μm. Error bars depict the mean ± SEM of the indicated number of cells (n) acquired in one representative experiment.

Figure 6

Complement and TRIM21 Mediate Block to Transgene Expression In Vivo (A) Relative light units (RLUs) indicating absolute levels of Ad5 infection in WT and C4 KO animals. (B) Relative infection of Ad5-Luc in WT and C4 KO mice using 9C12-WT. (C) Relative infection of Ad5-Luc in WT and T21 KO mice using 9C12-WT and mutant P329A. (D) SIINFEKL (SL8) specific CD8 T cell frequency in the blood of mice i.v. immunized with Ad5-Ova in the presence of 9C12-WT and mutants P329A and P329A/H433A. (E–G) Competition assays using 9C12 mutants (top) and m9C12 Fab (bottom). Neutralization assay using 1 μg/mL 9C12-WT in HeLa T21 KO cells titrating in P329A (E), in 293T-WT cells titrating in H433A (F), and in 293T-WT cells titrating in P329A/H433A (G). (H) Relative infection of Ad5-Luc in WT mice using 0.5 μg 9C12-WT in the presence of 500 μg m9C12 Fab. (I) Anti-Ad5 antibodies in mouse serum at the indicated days post infection. Groups consisted of 3–9 mice; error bars depict the mean ± SEM (A–D, H, and I). Error bars depict the mean + SD of three replicates acquired in one representative experiment (E–G).