Developing a platform system for gene delivery: amplifying virus-like particles (AVLP) as an influenza vaccine

Abstract

Delivery of a gene of interest to target cells is highly desirable for translational medicine, such as gene therapy, regenerative medicine, vaccine development, and studies of gene function. Parainfluenza virus 5 (PIV5), a paramyxovirus with a negative-sense RNA genome, normally infects cells without causing obvious cytopathic effect, and it can infect many cell types. To exploit these features of PIV5, we established a system generating self-amplifying, virus-like particles (AVLP). Using enhanced green fluorescent protein (EGFP) as a reporter, AVLP encoding EGFP (AVLP-EGFP) successfully delivered and expressed the EGFP gene in primary human cells, including stem cells, airway epithelial cells, monocytes, and T cells. To demonstrate the application of this system for vaccine development, we generated AVLPs to express the HA and M1 antigens from the influenza A virus strain H5N1 (AVLP-H5 and AVLP-M1H5). Immunization of mice with AVLP-H5 and AVLP-M1H5 generated robust antibody and cellular immune responses. Vaccination with a single dose of AVLP-H5 and M1H5 completely protected mice against lethal H5N1 challenge, suggesting that the AVLP-based system is a promising platform for delivery of desirable genes.

Fig. 1

Generation of PIV5 AVLP constructs. a Schematics of PIV5, AVLP–EGFP, AVLP–H5, AVLP–M1H5, AVLP–HT–EGFP, and AVLP–HTL–EGFP. Leader and trailer sequences are important for PIV5 RNA synthesis. NP nucleoprotein, P phosphoprotein, V V protein, M matrix protein, SH small hydrophobic protein, F fusion protein, HN Hemagglutinin-neuraminidase protein, L RNA-dependent RNA polymerase, Hyg hygromycin, EGFP enhanced green fluorescent protein, H5 HA of H5N1, M1 M1 of H5N1, TK thymidine kinase, nL nano-luciferase. b Schematics of AVLP generation work flow. BHK21 cells were transfected with plasmids expressing PIV5 P, NP, L, and T7 polymerase together with PIV5 AVLP plasmid expressing EGFP. The transfected cells were then selected with hygromycin. BHK (AVLP–EGFP) cell clones containing PIV5 AVLP genome were expanded under the selection of Hygromycin. BHK(AVLP-EGFP) cells were transfected with plasmids expressing PIV5 F, HN, and M. Particles carrying the PIV5 AVLP genome packaged by PIV5 HN, F, and M were generated in the supernatants of transfected cells. The particles were designated as AVLP–EGFP and used to infect Vero cells. The infected Vero cells were developed into stable cell lines carrying PIV5 AVLP genome [Vero (AVLP–EGFP)] under selection of Hygromycin. Vero (AVLP–EGFP) cells were transfected with plasmids expressing PIV5 F, HN, and M, and then AVLP–EGFP particles were produced in the supernatants of transfected cells. H.W. created this figure

Fig. 2

Infection of AVLP–EGFP to primary and continuous cell lines. a EGFP expression in Vero cells. b EGFP expression in human MSC cells. c EGFP expression in human primary epithelial cells. d. EGFP expression in human PBMCs. H.W. created this figure

Fig. 3

Examination of killing efficiency by GCV in established stable cell lines and monitoring of the expression kinetics of AVLP–HTL–EGFP in vivo. a Effects of GCV on AVLP-containing cells. BHK21 stable cell line containing AVLP–, AVLP–HT–, or AVLP–HTL–EGFP were treated with GCV at various concentrations. The images were taken at 3 days postinfection using a fluorescence microscope. b Detection of AVLP in vivo. BALB/c mice were infected with AVLP–HTL–EGFP and monitored for expression of nL using an IVIS camera. H.W. created this figure

Fig. 4

Characterization of Vero (AVLP–H5) cell line and Vero cells infected with AVLP–H5. a Detection of H5 and PIV5 V/P expression in Vero (AVLP–H5) cells by IFA. The cells were fixed and stained with anti-H5 or anti-PIV5 V/P antibodies followed by staining with FITC- conjugated secondary antibody. DAPI staining was performed with ProLong® Gold Antifade Mountant. b Detection of H5, PIV5 V/P, and NP expression in cells by western blotting. The Vero (AVLP–H5) cell samples were stained with anti-H5, anti-PIV5 V/P, or anti-PIV5 NP antibodies. Vero cell samples were used as a negative control. Samples derived from the same experiment and gels/blots were processed in parallel. c Surface expression of H5. Two days after infection with AVLP–H5 or PIV5–H5, Vero cells were fixed and treated with triton or PBS. The cells were stained with anti-H5 antibody and ProLong^® Gold Antifade Mountant as in a. d Analysis of particles. The concentrated AVLP-H5 and PIV5 particles were treated with anti-PIV5 HN antibody and then secondary antibody labeled with gold particles. The samples were examined using an electron microscope. e H5 incorporation. The purified AVLP-H5 particles were subjected to WB analysis with H5 specific antibody. Samples derived from the same experiment and gels/blots were processed in parallel. H.W. created this figure

Fig. 5

A single dose immunization of AVLP-H5 protects against H5N1 virus challenge. Mice in a group of 15 were vaccinated with a single dose of PBS, PIV5–H5 4 × 10⁵ PFU per mouse), or AVLP–H5 intranasally at 4 × 10⁴ or 4 × 10⁵ AP. At 25 days post vaccination, mice were challenged with 10 LD₅₀ of H5N1. Weight loss a and survival b were monitored for 10 days following challenge. Weight loss is graphed as an average percentage of the original weight (the day of challenge; ± SEM). On day 4 post-challenge, a subset of animals was humanely euthanized, and tissues were collected for lung virus titers c. The limit of detection in c was 1.2 log₁₀ TCID₅₀/ml. S.M.T. created this figure

Fig. 6

A single dose immunization of AVLP–M1H5 protects against H5N1 virus challenge. a Detection of M1 and H5 in cell media. Supernatants from BHK (AVLP–H5) and BHK (AVLP–M1H5) cell cultures were purified and subjected to WB. Samples derived from the same experiment and gels/blots were processed in parallel. b and c ELISPOT assay for IFN-γ. Mice in a group of 5 were vaccinated with a single dose of PBS, AVLP–H5 (1 × 10⁵ AP), or AVLP–M1H5 intranasally (1 × 10⁵, 1 × 10⁴ or 1 × 10³ AP). At day 21 post-prime immunization, mice were sacrificed, and spleens were collected. Splenocytes were stimulated with PBS, H5, or M1 recombinant protein. Results are presented as the mean number of IFN-γ-producing cells per 10⁶ splenocytes. Differences were evaluated by one-way ANOVA. d and e morbidity and mortality after challenge. H.W. created figures a and b. S.M.T. created figures c and d