Attenuation of Zika Virus by Passage in Human HeLa Cells

Abstract

Zika virus (ZIKV) is a mosquito-borne Flavivirus. Previous studies have shown that mosquito-transmitted flaviviruses, including yellow fever, Japanese encephalitis, and West Nile viruses, could be attenuated by serial passaging in human HeLa cells. Therefore, it was hypothesized that wild-type ZIKV would also be attenuated after HeLa cell passaging. A human isolate from the recent ZIKV epidemic was subjected to serial HeLa cell passaging, resulting in attenuated in vitro replication in both Vero and A549 cells. Additionally, infection of AG129 mice with 10 plaque forming units (pfu) of wild-type ZIKV led to viremia and mortality at 12 days, whereas infection with 10³ pfu of HeLa-passage 6 (P6) ZIKV led to lower viremia, significant delay in mortality (median survival: 23 days), and increased cytokine and chemokine responses. Genomic sequencing of HeLa-passaged virus identified two amino acid substitutions as early as HeLa-P3: pre-membrane E87K and nonstructural protein 1 R103K. Furthermore, both substitutions were present in virus harvested from HeLa-P6-infected animal tissue. Together, these data show that, similarly to other mosquito-borne flaviviruses, ZIKV is attenuated following passaging in HeLa cells. This strategy can be used to improve understanding of substitutions that contribute to attenuation of ZIKV and be applied to vaccine development across multiple platforms.

Keywords: HeLa cells; Zika virus; Zika virus NS1; Zika virus prM; attenuation; flavivirus.

Figure 1

Passaging of ZIKV PA259249 in HeLa cells led to in vitro attenuation. ZIKV strain PA259249 (a) or FSS13025 (b) was serially passaged in HeLa cells up to six times (P6), and stocks were titrated in Vero cells. Titration limit of detection: 50 pfu (10^1.7 pfu). (c) Focus morphology of passaged ZIKV PA259249; (d) stocks of WT Zika PA259249 virus serially passaged in HeLa cells, four, five, or six times (WT, HeLa-P4, -P5, and -P6, respectively) were used to infect Vero cells. Duplicate samples were harvested from cell supernatants at 7 hours, 1, 2, or 3 days p.i. and titrated in Vero cells. All titers decreased on days 4 and 5 (not shown). Dashed line: limit of detection: 100 ffu/ml (10^2.0 ffu/ml). (e) Stocks of WT Zika strain PA259249 virus HeLa-P4 and -P5 were used to infect A549 cells; duplicate samples harvested from cell supernatants at 1, 2, or 3 days p.i. were titrated in Vero cells. Dashed line: limit of detection: 50 ffu/ml (10^1.7 pfu/ml). Mock-infected cells had no detectable titers in Vero or A549 cells (not shown).

Figure 2

ZIKV passaged in HeLa cells was attenuated and generated an immune response in AG129 mice. (a) Adult AG129 mice were inoculated via the i.p. route with 10⁰ or 10¹ pfu of WT ZIKV or with 10³ pfu of HeLa-P6 virus, and monitored for morbidity and mortality for 30 days. Statistics were determined using Log-rank (Mantel–Cox) tests. WT ZIKV 10⁰ pfu n = 4; WT ZIKV 10¹ pfu n = 9; HeLa-P6 10³ pfu n = 9; (b) AG129 mice were bled 3 days p.i. to determine acute infection viremia levels. Lines indicate the mean titer. Statistics were determined using one-way ANOVA and Tukey’s post-test; n.s.: not significant. WT ZIKV 10⁰ pfu n = 4, WT ZIKV 10¹ pfu n = 9, Hela-P6 10³ pfu n = 9; (c) At time of death, infected animal brain and testis samples were harvested to determine viral loads. Lines indicate means. Statistics were determined using t tests. WT ZIKV 10¹ pfu days 11–12, brain n = 3, testis n = 3; HeLa-P6 10³ pfu brain n = 6, testis n = 3; (d) Neutralizing antibody titers present in terminal sera from euthanized mice infected with WT ZIKV 10¹ pfu (n = 2, 11–12 days p.i.), HeLa-P6 10³ pfu (n = 5, days 16–27 p.i.) or naïve (n = 1) mice were determined using a FRNT₅₀ assay in Vero cells. Endpoint titers are plotted in a log₂ scale to reflect the 2-fold dilution series. Dashed line: limit of titer detection was 1:20. Results between WT ZIKV and HeLa-P6 were not significantly different by Mann–Whitney test.

Figure 3

Passaging of ZIKV in HeLa cells increased circulating cytokines in AG129 mice. Mice were inoculated with WT ZIKV (10⁰ or 10¹ pfu) or 10³ pfu HeLa-P6, and the circulating cytokine response was measured at 3 days p.i. Cytokine and chemokine levels increased following infection with HeLa-P6 compared to WT ZIKV. WT ZIKV 10⁰ pfu n = 4, WT ZIKV 10¹ pfu n = 9, Hela-P6 10³ pfu n = 9. Kruskal–Wallis tests with Dunn’s multiple comparisons post-tests were used for statistical analysis. Bars represent the mean +/- standard error. * p < 0.05, ** p < 0.01, ** p < 0.001.

Figure 4

Location of amino acid substitutions incurred after HeLa cell passage. (a) ZIKV genome organization showing the non-synonymous nucleotide changes occurring after in vitro HeLa cell passaging at positions 732 and 2797 (dark blue vertical lines), and after in vivo mouse passaging of HeLa-P6 at 1181 and 2242 (light blue vertical lines). (b) Cartoon representation of the location of prM amino acid 87 in the pr peptide domain, which is cleaved by furin protease during maturation. (c) Top view of the NS1 protein dimer structure [9] shows the location of residue 103 (green spheres) in the wing domain (yellow). (d) Side view of of the cryo-EM [10] of three M–E proteins showing E 68 in cyan in domain 2 and E 422 in magenta in the stem region. One M and one E protein are shown in color (red—E domain 1, yellow—E domain 2, green—fusion loop, dark blue—E domain 3, marine blue—E stem and transmembrane, orange—M transmembrane).