Deliberate attenuation of chikungunya virus by adaptation to heparan sulfate-dependent infectivity: a model for rational arboviral vaccine design

Abstract

Mosquito-borne chikungunya virus (CHIKV) is a positive-sense, single-stranded RNA virus from the genus Alphavirus, family Togaviridae, which causes fever, rash and severe persistent polyarthralgia in humans. Since there are currently no FDA licensed vaccines or antiviral therapies for CHIKV, the development of vaccine candidates is of critical importance. Historically, live-attenuated vaccines (LAVs) for protection against arthropod-borne viruses have been created by blind cell culture passage leading to attenuation of disease, while maintaining immunogenicity. Attenuation may occur via multiple mechanisms. However, all examined arbovirus LAVs have in common the acquisition of positively charged amino acid substitutions in cell-surface attachment proteins that render virus infection partially dependent upon heparan sulfate (HS), a ubiquitously expressed sulfated polysaccharide, and appear to attenuate by retarding dissemination of virus particles in vivo. We previously reported that, like other wild-type Old World alphaviruses, CHIKV strain, La Réunion, (CHIKV-LR), does not depend upon HS for infectivity. To deliberately identify CHIKV attachment protein mutations that could be combined with other attenuating processes in a LAV candidate, we passaged CHIKV-LR on evolutionarily divergent cell-types. A panel of single amino acid substitutions was identified in the E2 glycoprotein of passaged virus populations that were predicted to increase electrostatic potential. Each of these substitutions was made in the CHIKV-LR cDNA clone and comparisons of the mutant viruses revealed surface exposure of the mutated residue on the spike and sensitivity to competition with the HS analog, heparin, to be primary correlates of attenuation in vivo. Furthermore, we have identified a mutation at E2 position 79 as a promising candidate for inclusion in a CHIKV LAV.

Figure 1. Cell culture passage of CHIKV-LR.

CHIKV-LR virus stock was serially passaged 10 times in triplicate parallel series on: A) CHOK1; B) C6/36; or C) pgsA745 cells. After each amplification, the quantity of infectious virions was titered on BHK-21 fibroblasts by standard plaque assay. The fold-increase in titer at passage (P) 5 and P10 were normalized to the titer at P1. Each bar represents one individual passage series.

Figure 2. Effects of positively charged amino acid substitutions in the CHIKV E2 glycoprotein on electrostatic potential in the E1/E2 trimer-heterodimeric spike.

The trimeric CHIKV-LR E1/E2 heterodimer was modeled and APBS was used to determine the electrostatic potential of the CHIKV-LR E2 protein, the location of mutated residues and predicted changes to the electrostatic potential conferred by these mutations. (A) Top view of the CHIKV-LR trimer-heterodimer with E1 shown in gray, E2 domain A in cyan, domain B in green, and domain C in brown. (B–D) The electrostatic potential of the CHIKV-LR trimer-heterodimer where blue indicates positive charge, white indicates neutral charge and red indicates negative charge. B) top view; C) side view I; and D) side view II. (E & F) For ease of viewing, the region of the E2 molecule altered by each mutation is magnified showing only one of the E2 molecules in the trimer-heterodimer. E) Magnification from side view I of the “wing” portion of domain A that contains the E2-55R, E2-79K and E2-82R mutations; F) Magnification of side view II of the acid sensitive region in arch 1 between domains A and B that contains the E2-159R, E2-Δ166E/166K and E2-168K mutations. G–J) The top view of the electrostatic potential of trimetric heterodimer for the E2 mutations in the “wing” portion of domain A: G) CHIKV-LR, H) E2-55R, I) E2-79K and J) E2-82R.

Figure 3. Musculoskeletal disease following subcutaneous infection of CD-1 mice with E2 mutant viruses.

21-1 mice were infected subcutaneously in the left rear footpad with 10⁵ GE (normalized to 10³ PFU of CHIKV-LR) and the metatarsal region (height and width) was measured daily; fold-change over day 0 was calculated based upon pre-infection footpad area. Viruses are color-coded based on disease phenotype: (A–B) hind-limb swelling with little or no attenuation (purple; (A) CHIKV-LR and (B) LR- Δ166E), (C–E) partial attenuation (blue; (C) LR-168K, (D) LR-55R and (E) LR-159R) and (F–H) highly attenuated with little or no hind-limb swelling (green; (F) LR-166K, (G) LR-82R and (H) LR-79K) compared to CHIKV-LR. Individual mice are graphed with the intensity of shading indicating the proportion of mice with overlapping levels of hind-limb swelling. *p<0.001 compared to CHIKV-LR.

Figure 4. Cytokine induction following subcutaneous inoculation of CHIKV E2 mutant viruses.

21-1 mice were infected subcutaneously in the left rear footpad with 10⁵ GE. At 48 h p.i. sera were collected and analyzed for cytokine/chemokine induction using a 20-plex luminex bead kit. A) IP-10; B) MCP-1; C) MIG; D) IL-12; E) IL-5; F) IL-1α. Viruses are color-coded based on MSD severity: little or no attenuation (purple), partial attenuation (blue) or highly attenuated (green) compared to CHIKV-LR. Error bars represent standard deviation. *p<0.05 and **p<0.01 compared to CHIKV-LR.

Figure 5. Positive charge CHIKV E2 mutations increase specific infectivity and infectivity is dependent on ionic interactions.

Specific infectivity (GE/PFU) of A) CHIKV-LR WT; and B) the mutants CHIKV E2 mutants were evaluated on different cell types. For the CHIKV E2 mutants, fold-change in specific infectivity was normalized to CHIKV-LR. C) CHIKV E2 mutants were evaluated for dependency on ionic interaction for infectivity by interacting virus in either RPMI1640 (basal salt concentration of 103 mM NaCl) or RPMI1640 with increasing salt concentration and MC3T3-E1 cells were then infected for a standard plaque assay. Percent infectivity was normalized to 103 mM NaCl, which was set to 100%. Viruses are color-coded based on control groups (orange) or MSD severity: little or no attenuation (purple), partial attenuation (blue) or highly attenuated (green) compared to CHIKV-LR. Error bars represent standard deviation. *p<0.01 and **p<0.001 compared to CHIKV-LR (B) or 103 mM NaCl (C).

Figure 6. Soluble heparin can block infectivity and alter plaque size of several of the positive charge CHIKV E2 mutants on MC3T3-E1 osteoblasts.

A) Soluble heparin was used to block infectivity of the virus by competing with the ability of the virus to attach to cells. Virus was incubated with 200 µg of either soluble heparin or BSA, infection was synchronized and overlay was added for a standard plaque assay. Percent heparin blocking was normalized to BSA control. B & C) Standard plaque assay on MC3T3-E1 was used to determine plaque size of the different CHIKV viruses. C) Plaque sizes were quantified by measuring the size of the plaques using a caliper. D) Virus was titered using a standard plaque assay on MC3T3-E1 cells in the presence or absence of 200 µg/mL of soluble heparin in the overlay. Plaque sizes were quantified by measuring the size of the plaques using a caliper. Viruses are color-coded based on control groups (orange) or MSD severity: little or no attenuation (purple), partial attenuation (blue), or highly attenuated (green) compared to CHIKV-LR. Error bars represent standard deviation. *p<0.001 compared to CHIKV-LR (B) or 103 mM NaCl (C).

Figure 7. HS-dependent infectivity of the CHIKV E2 mutants by removal of HS on MC3T3-E1 cells.

CHIKV E2 mutants were evaluated for dependency on HS for infectivity on MC3T3-E1 cells by digestion of HS with different heparinases (Hep). MC3T3-E1 cell surfaces were mock-treated, or digested with various concentrations of A) Hep I; B) Hep II; or C) Hep III and then infected with virus for a standard plaque assay on MC3T3-E1 cells. Percent infection was normalized to mock treatment which was set to 100%. Viruses are color-coded based on control groups (orange) or MSD severity: little or no attenuation (purple), partial attenuation (blue), highly attenuated (green) compared to CHIKV-LR. Error bars represent standard deviation. For all heparinases at all concentrations, there was no significant difference in infectivity for SINV-TR339, CHIKV-LR and LR-Δ166E compared to the no Hep control. For all of the CHIKV E2 mutants, with the exception of LR-159R at 0.5 U/mL of Hep III, had a significant decrease in infectivity compared to the mock-treated control (p<0.01).

Figure 8. Viral load following subcutaneous inoculation of CHIKV E2 mutants.

21-1 mice were infected subcutaneously in the left rear footpad with 10⁵ GE. At 48 h p.i., A) footpad; B) quadriceps muscle; C) knee joint; D) popliteal lymph node (LN); E) serum; and F) spleen were harvested and analyzed for viral load by standard plaque assay. The limit of detection (LD) per g, mL, or LN is indicated. Viruses are color-coded based on MSD severity: little or no attenuation (purple), partial attenuation (blue) or highly attenuated (green) compared to CHIKV-LR. Error bars represent standard deviation. *p<0.05, **p<0.01 and ***p<0.001 compared to CHIKV-LR.

Figure 9. Protective immunity and neutralizing antibody elicited by LR-82R and LR-79K.

21-1 mice were infected with 10⁵ GE of CHIKV-LR WT, LR-82R, or LR-79K subcutaneously in the rear footpad. At 21 d p.i., mice were bled and challenged with 10³ PFU of CHIKV-LR. A) Hind-limb swelling post-challenge was determined by measuring the metatarsal region (height and width) daily and fold change was calculated based on pre-challenge footpad area. The legend indicates the virus used for the primary infection. Individual mice are graphed with the intensity of shading indicating the proportion of mice with overlapping levels of hind-limb swelling. B) Neutralizing antibody levels were determined by interacting a CHIKV-LR luciferase-expressing virus with sera and then infecting cells overnight. Lysates were then harvested and analyzed for luciferase signal. Levels of neutralizing antibody were determined as fold reduction in luciferase signal compared to mock sera. Error bars represent standard deviation.