Complex interactions with several arms of the complement system dictate innate and humoral immunity to adenoviral vectors

Abstract

The complement system is known to play critical roles in pathogen identification, initiation of innate immune responses and facilitation of adaptive immune responses. Several studies have suggested that recombinant adenoviruses (rAds) interact with proteins of the complement system within minutes of administration. In this study, we assessed the roles of the alternative (Factor B), classical (C1q and C4) and common (C3) arms of the complement system in the innate and humoral response to systemic rAd administration using mice genetically deficient for each of these functions. Although most plasma cytokines and chemokines induced by Ads appeared to be elicited in a C3-dependent manner, we found that rAd-induced thrombocytopenia was dependent on Factor B and C3, implicating the alternative pathway as responsible for this response. Alteration of the complement-dependent transcriptome response after rAd-induced liver gene expression was also found to be Factor B- and C3-dependent. Ad interactions with the classical and alternative arms of the complement system can also be redundant, as many complement-dependent, Ad-induced innate immune responses appeared to be primarily C3-dependent. We also identified a C3 dependence of Ad-mediated induction of the nuclear factor-kappaB (NF-kappaB) activation pathway. Finally, we confirmed that humoral immune responses to the vector capsid, and the transgene it encodes, are also complement-dependent.

Figure 1

Platelet count assessments in recombinant adenovirus (rAd)-injected C57BL/6 and complement-deficient mice. All mice were injected intravenously with 7.5 × 10¹⁰ viral particle (vp) per mouse. Whole-blood samples were collected at 24 h.p.i., and platelets were enumerated using the Unopette platelet counting system with a hemocytometer. Samples were taken from mock-injected mice and compared with virus-injected animals of the indicated genotype. Bars represent mean±s.d. **P<0.01 was deemed a statistically significant difference. For C57BL/6 mice, N = 14 and N = 10 for mock and virus-injected mice, respectively. For C3-KO, N = 6 and N = 11; for Factor B knockout (FB-KO), N = 16 and N = 11; for C4-KO, N = 6 and N = 8; for C1q-KO, N = 6 and N = 11; for mock and virus-injected mice, respectively.

Figure 2

Plasma cytokine and chemokine elevations after intravenous adenovirus (Ad) injection. C57BL/6, C3-KO, Factor B knockout (FB-KO), C4-KO and C1q-KO mice were injected with 7.5 × 10¹⁰ viral particle (vp) per mouse of rAd5-LacZ. Plasma samples were isolated at 6 h.p.i. and analyzed using a Bio-Plex bead-based enzyme-linked immunosorbent assay (ELISA) assay (see Materials and methods). Bars represent mean±s.d. Student’s t-tests were completed between mock-injected samples (†P<0.05 and ††P<0.01), virus-injected samples (^#P<0.05 and ^##P<0.01) and between mock and virus-injected samples within the same genotype (*P<0.05 and **P<0.01). N = 3 for all samples tested.

Figure 3

The role of C3 in recombinant adenovirus (rAd) vector-induced ERK phosphorylation and IκBα degradation. Wild-type C57BL/6 and C3-KO mice were injected intravenously with rAd5-LacZ (1.5 × 10¹¹ viral particles per mouse). Livers were collected at the indicated time points, snap-frozen and lysates extracted as described in the Materials and methods section. (a) IκBα and tubulin; (b) p-ERK1/2, ERK2 were determined by western blot analysis using LI-COR. Quantification was performed by normalizing the p-ERK1/2 levels to total ERK2 and IκBα levels to tubulin to control for loading (N = 3 for all time points). For each analysis, quantification is shown on the top, and a representative blot on the bottom. Bars represent mean ± s.d. A homoscedastic t-test was used to determine statistical differences between levels in C57BL/6 livers compared with levels in C3-KO mouse livers at each indicated time point (*P<0.05).

Figure 4

Ad neutralizing antibody titers are C3-dependent. C57BL/6, C3-KO, Factor B knockout (FB-KO) and C1q-KO mice were injected with 7.5 × 10¹⁰ viral particle (vp) per mouse of rAd5-GFP (N = 3 for recombinant adenovirus (rAd)-naive animals, N = 4 for all immunized animals). Plasma samples were isolated at 30 d.p.i. and analyzed for neutralizing antibodies using four successive dilutions of plasma (see Materials and methods). Bars represent mean ± s.d. ‘**’ represents a statistical difference, as measured by a homoscedastic t-test, in neutralizing antibody titers found in plasma derived from C57BL/6 animals versus C3-KO animals (P<0.01).

Figure 5

Anti-recombinant adenovirus (rAd)-specific antibodies are dependent on both alternative pathway (AP) and classical pathway (CP) of complement. C57BL/6 (N = 3 naive and N = 7 immunized), C3-KO (N = 4 naive and N = 9 immunized), Factor B knockout (FB-KO) (N = 4 naive and N = 11 immunized) and C1q-KO (N = 3 naive and N = 5 immunized) mice were injected with 7.5 × 10¹⁰ viral particle (vp) per mouse of rAd5-GFP. (a) At 30 d.p.i., plasma was analyzed for anti-rAd capsid-specific total IgG antibodies at the appropriate dilution (1:800). Bars represent mean ± s.d. (b) Plasma was also analyzed for various IgG subclasses. Data points represent mean ± s.d. *P<0.05 and **P<0.01 indicate a statistical difference between C57BL/6 and each respective genotype. (c) The IgG2c/IgG1 ratio, indicative of a Th1/Th2 response, was calculated based on subclass titering.

Figure 6

Anti transgene (GFP)-specific antibodies are C3-dependent. C57BL/6 (N = 3 naive and N = 7 immunized), C3-KO (N = 4 naive and N = 9 immunized), Factor B knockout (FB-KO) (N = 4 naive and N = 11 immunized) and C1q-KO (N = 3 naive and N = 5 immunized) mice were injected with 7.5 × 10¹⁰ viral particle (vp) per mouse of rAd5-GFP. (a) At 30 d.p.i., plasma was analyzed for anti-GFP-specific total IgG antibodies at the appropriate dilution (1:800). Bars represent mean±s.d. (b) Plasma was also analyzed for various IgG subclasses. Data points represent mean±s.d. *P<0.05 and **P<0.01 indicate a statistical difference between C57BL/6 and each respective genotype. (c) The IgG2c/IgG1 ratio, indicative of a Th1/Th2 response, was calculated based on subclass titering.