Bacterial superglue enables easy development of efficient virus-like particle based vaccines

Abstract

Background: Virus-like particles (VLPs) represent a significant advance in the development of subunit vaccines, combining high safety and efficacy. Their particulate nature and dense repetitive subunit organization makes them ideal scaffolds for display of vaccine antigens. Traditional approaches for VLP-based antigen display require labor-intensive trial-and-error optimization, and often fail to generate dense antigen display. Here we utilize the split-intein (SpyTag/SpyCatcher) conjugation system to generate stable isopeptide bound antigen-VLP complexes by simply mixing of the antigen and VLP components.

Results: Genetic fusion of SpyTag or SpyCatcher to the N-terminus and/or C-terminus of the Acinetobacter phage AP205 capsid protein resulted in formation of stable, nonaggregated VLPs expressing one SpyCatcher, one SpyTag or two SpyTags per capsid protein. Mixing of spy-VLPs with eleven different vaccine antigens fused to SpyCatcher or SpyTag resulted in formation of antigen-VLP complexes with coupling efficiencies (% occupancy of total VLP binding sites) ranging from 22-88 %. In mice, spy-VLP vaccines presenting the malaria proteins Pfs25 or VAR2CSA markedly increased antibody titer, affinity, longevity and functional efficacy compared to corresponding vaccines employing monomeric proteins. The spy-VLP vaccines also effectively broke B cell self-tolerance and induced potent and durable antibody responses upon vaccination with cancer or allergy-associated self-antigens (PD-L1, CTLA-4 and IL-5).

Conclusions: The spy-VLP system constitutes a versatile and rapid method to develop highly immunogenic VLP-based vaccines. Our data provide proof-of-concept for the technology's ability to present complex vaccine antigens to the immune system and elicit robust functional antibody responses as well as to efficiently break B cell self-tolerance. The spy-VLP-system may serve as a generic tool for the cost-effective development of effective VLP-vaccines against both infectious- and non-communicable diseases and could facilitate rapid and unbiased screening of vaccine candidate antigens.

Fig. 1

The spy-VLP antigen display platform. a Three types of spy expressing VLPs were constructed by genetic fusion of SpyTag or SpyCatcher to the virus-like particle (VLP)-forming AP205 capsid protein. (1) “SpyTag-VLP” had the SpyTag fused to the N-terminus of the AP205 capsid protein and present 180 potential SpyCatcher-antigen binding motifs (2) “2xSpyTag-VLP” had SpyTag fused to both the N- and C-terminus of the AP205 capsid protein and present 360 potential SpyCatcher-antigen binding motifs; (3) “SpyCatcher-VLP” had SpyCatcher fused to the N-terminus of the AP205 capsid protein and present 180 potential Spytag-antigen binding motifs. b Transmission electron microscopy (TEM) images showing the SpyTag-VLP, 2xSpyTag-VLP and SpyCatcher-VLP. Purified spy-VLP samples were placed on carbon, adsorbed to a grid and negatively stained with 2 % phosphotungstic acid. Scale bar 50 nm. Images show uniform, non-aggregated particles of approximately 30 nm (SpyTag-VLP and 2xSpyTag-VLP) and 42 nm (SpyCatcher-VLP). c Reduced SDS-PAGE gels loaded with VLP vaccines demonstrating that vaccine proteins had formed covalent bonds to the AP205 capsid protein. Left panel shows that mixing of SpyTag-VLPs with SpyCatcher-IL-5 resulted in three protein bands corresponding to the size of an antigen-VLP capsid protein conjugate (48 kDa) (top), uncoupled vaccine antigen (33 kDa) (middle) and unconjugated SpyTag-VLP capsid protein (16.5 kDa) (bottom). The middle panel shows that mixing of 2xSpyTag-VLP with SpyCatcher-IL-5 resulted in four protein bands representing; a conjugate of two vaccine antigens bound to each end of a 2xSpyTag-VLP capsid protein (83 kDa), a conjugate of the 2xSpyTag-VLP capsid protein and a single vaccine antigen (48 kDa), uncoupled vaccine antigen (33 kDa) and unconjugated 2xSpyTag-VLP capsid protein (18.5 kDa). The left panel shows that mixing of SpyCatcher-VLP with PD-L1-SpyTag resulted in three protein bands representing; an antigen-VLP capsid protein conjugate (50 kDa), uncoupled vaccine antigen (33 kDa) and unconjugated SpyCatcher-VLP capsid protein (27 kDa)

Fig. 2

Antigen-specific IgG levels in mice after immunization with soluble or spy-VLP displayed malaria antigens. (a, upper panels) Antigen-specific IgG levels (OD Elisa) in serum from mice (n = 5 per group) immunized with a Pfs25 2xSpyTag-VLP vaccine (filled circles) or with a control vaccine consisting of soluble Pfs25 mixed with untagged AP205 VLPs (open squares). Both vaccines were formulated using aluminum hydroxide adjuvant (Statens Serum Institut, Copenhagen, Denmark). Mice were immunized on days 0, 21 and 42 and serum was collected on the indicated days after first immunization. Differences in median endpoint titers between vaccination groups were analyzed using Mann–Whitney Rank Sum test; day 14 (P < 0.01), day 35 (P < 0.01), day 56 (P < 0.01) and day 212 (P = 0.03). (a, lower panels) Similar results for the VAR2CSA based vaccines (VAR2CSA SpyCactcher-VLP and soluble VAR2CSA plus untagged AP205 VLP), which were formulated without extrinsic adjuvant. Statistical analysis; day 14 (P = 0.03), day 35 (P = 0.09), day 56 (P = 0.09) and day 137 (P = 0.03). b Antibody avidity was assessed on days 35 and 56 in serum samples from mice vaccinated with the Pfs25 or VAR2CSA vaccines. Avidity index values were determined by measuring the resistance of antibody-antigen complexes to 8 M urea. The avidity index was calculated as the ratio of the mean ELISA OD₄₉₀ value of urea-treated wells to PBS control wells multiplied by 100. Mann–Whitney Rank Sum test was used for statistical comparisons. c The distribution of IgG1, IgG2a and IgG2b relative to the total vaccine-induced IgG response in mice (n = 5 per group) following Pfs25 or VAR2CSA immunization. Anti-Pfs25 and anti-VAR2CSA sera (left) were obtained on days 98 and 88, respectively. Mann–Whitney Rank Sum test was used for statistical comparisons

Fig. 3

Functional activity of spy-VLP vaccine-induced humoral responses. a Transmission reducing activity (TRA) of anti-Pfs25 (day 56) sera following immunization of mice with the Pfs25 spy-VLP vaccine or with the control vaccine (soluble Pfs25 + untagged AP205 VLP) as described in Fig. 2a. Y-axis shows the number of oocysts identified in the midgut of each of 20 A. stephensi mosquitoes. Pre-immunization mouse serum was used as additional negative control. Mann–Whitney rank sum test was used for statistical comparisons. b Binding between VAR2CSA expressing infected erythrocytes and CSA in the presence of different concentrations of serum from immunized mice. Binding in the presence of serum from non-immunized mice was set to 100 %. Serum pools were from groups of 5 mice immunized with the VAR2CSA spy-VLP vaccine (black circle) or with control vaccine (empty square). The EC50 value for the spy-VLP vaccinated mice was 8.8 fold higher [3.213–14.41] than the value from mice receiving the control vaccine

Fig. 4

Breakage of self-tolerance. IgG autoantibody responses measured by standard ELISA. a, b C57BL/6 mice (n = 10 per group) were immunized with a PD-L1 (a) or CTLA-4 (b) SpyCatcher-VLP vaccine (filled circles) or with a control vaccine (n = 3 per group) consisting of similar amounts of spy-antigen mixed with untagged AP205 VLPs (open squares). Both vaccines were formulated using aluminum hydroxide adjuvant (Statens Serum Institut, Copenhagen, Denmark). Mice were immunized with a dose of 5 µg antigen on days 0, 21 and 42 and serum was collected on day 56 after first immunization. Median endpoint titers were compared for the PD-L1 vaccination groups (P = 0.01) and the CTLA-4 vaccination groups (P = 0.01) using Mann–Whitney Rank Sum test. c, d BALB/c mice (n = 4) were immunized with an IL-5 SpyTag-VLP vaccine or a control vaccine (soluble IL-5 + untagged AP205 VLP) (n = 5) which were both formulated with aluminum hydroxide (Statens Serum Institut, Copenhagen, Denmark). Mice were immunized on days 0, 21 and 42 with antigen doses of 5, 2.5 and 2.5 µg, respectively, and serum was collected on days 56 (c) and 112 (d). Median endpoint titers for the two vaccination groups were compared using Mann–Whitney Rank Sum test; day 56 (P = 0.2) and day 122 (P = 0.2)