Poliovirus intrahost evolution is required to overcome tissue-specific innate immune responses

Abstract

RNA viruses, such as poliovirus, have a great evolutionary capacity, allowing them to quickly adapt and overcome challenges encountered during infection. Here we show that poliovirus infection in immune-competent mice requires adaptation to tissue-specific innate immune microenvironments. The ability of the virus to establish robust infection and virulence correlates with its evolutionary capacity. We further identify a region in the multi-functional poliovirus protein 2B as a hotspot for the accumulation of minor alleles that facilitate a more effective suppression of the interferon response. We propose that population genetic dynamics enables poliovirus spread between tissues through optimization of the genetic composition of low frequency variants, which together cooperate to circumvent tissue-specific challenges. Thus, intrahost virus evolution determines pathogenesis, allowing a dynamic regulation of viral functions required to overcome barriers to infection.RNA viruses, such as polioviruses, have a great evolutionary capacity and can adapt quickly during infection. Here, the authors show that poliovirus infection in mice requires adaptation to innate immune microenvironments encountered in different tissues.

Fig. 1

RNA viral evolution capacity is required to overcome the type I interferon (IFN) response. a The structure of RNA dependent RNA polymerase (RdRp) of poliovirus. G64S (G) confers higher replication fidelity and D79H (D) reduces recombination rate^–. b Mutation rates associated of wild type (WT) and G64S/D79H (GD) poliovirus. Mutation rates are shown for G to A, C to U and U to A substitutions, based upon the frequency of lethal mutations (P < 0.05, seven passages, n = 7, mean ± s.d.). c Recombination rates for WT and GD poliovirus. For additional details see Supplementary Fig. 1. d Percentage survival of Tg21 mice or IFNAR^−/− mice (n = 10 for each cohort) infected with virus by I.P. inoculation route. Mice were inoculated with 10⁸ plaque-forming units (PFU) per Tg21 mouse or with 10⁴ PFU per IFNAR^−/− mouse (n = 10 for each cohort). e 50% lethal dose (LD50) by intra-peritoneal (I.P.) and intra-muscular (I.M.) route (n = 10 for each cohort)

Fig. 2

The tissue distribution of the virus strains in Tg21 and IFNAR^−/− mice. a Experimental schematic for determining the virus titer in individual tissues. b, c Virus titers in tissue collected from WT or GD infected mice by I.P. inoculation route. Tg21 (10⁶ PFU per mouse) or IFNAR^−/− (10⁴ PFU per mouse). Data are shown as logarithm, mean ± s.d., PFU per gram of tissue (n = 5, the number of mice is five for each time point and each group). Student’s t-test, n.s. indicates P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Limited detection level is 20 PFU per gram tissue

Fig. 3

Tissue-specific transcriptional responses induced by poliovirus infection. a Line plot showing the gene expression profiles for 341 genes in liver and kidney from Tg21 mice infected with WT poliovirus at 10⁷ PFU by I.P. inoculation route. (n = 3, the number of mice is three for each time point and each group). Gene expression profiles across all experimental conditions were clustered to yield 5 clusters (Methods). Each gene’s trajectory is shaded by its cluster and the mean trajectory for each cluster is shown as a thick line. b Multidimensional scaling plot of the transcriptional response trajectories in liver and kidney. Each point represents the mean transcriptional response of three mice. Shaded ovals are drawn to emphasize the distinct trajectories of liver and kidney responses in infected mice. c Scatter plot showing the relative induction of genes in liver and kidney at 1 (left) and 3 (right) days postinfection. Genes are colored by cluster. d A multidimensional scaling plot of the coregulatory relationships between genes. Genes are arranged by the similarity of their expression response (determined by Euclidian distance) and colored by cluster. e Known innate immune and antiviral genes group together in clusters 1 and 5

Fig. 4

Genetic composition of poliovirus promotes intrahost evolution through counteracting tissue specific interferon responses. a Tissue-specific patterns of diversity in organ-resident populations. Virus was isolated at 3 days postinfection in Tg21 mice (n = 3, the number of mice is three for each time point and each group). Liver (LV), kidney (KD), and spleen (SP). Diversity is represented as Shannon’s Entropy. b The genetic structures of in vivo adapted populations. Multidimensional scaling correlation. In vivo adaptation leads to tissue-specific genetic structures (colored ovals) in WT populations from LV, KD and SP, but not in GD populations (grey oval). c Correlation plot of diversity between tissue-specific adapted populations between viral strains. d Model highlighting how efficient recombination and mutation leads to rapid adaptation during intrahost spread. G64S/D79H(GD), engineered to be deficient in recombination and diversity, is limited in its ability to rapidly adapt to changing selection pressures

Fig. 5

The 2B coding region of the poliovirus genome is a hotspot for the accumulation of mutations following cell culture adaptation. a High-density genetic screen procedure to identify adaptive mutations on cell culture conditions. WT, GD, or WT were passaged in HeLa cells or HeLa cells pre-treated with interferon β (IFN β). Using circular resequencing (CirSeq), the mutation composition, allele frequency distribution, and fitness of specific alleles were determined. b Scatter plot showing the position and fitness of identified adaptive mutations. Adaptive alleles with fitness values >2 standard deviations (s.d.s) above neutrality are shown. Alleles with fitness > 25 s.d.s above neutrality are labeled. The details of the specific mutations labeled in the plot are shown in Supplementary Table 1. The 2B coding region of the genome is shaded grey. c Allele frequency spectrums for 4 independently passaged of poliovirus (PV) populations. Alleles are ranked by frequency and plotted to show the relative diversity of passaged populations. All populations shown are from passage 6 of parallel passage experiments. d Interferon (IFN) sensitivity assay for specific 2B mutations identified in the passage screen. One-tailed Student’s t-test. *P ≤ 0.05, **P ≤ 0.01, n = 3. Data were presented as logarithm(mean ± s.d.), PFU per ml. n = 3, three replicates for each time, for each viral strain. e,f Comparison of immune signaling in MG63 cells infected by WT or 2B mutants at m.o.i=20. e Supernatant was collected at the indicated times to measure IFN β induction by ELISA. IFN β level < 2.3 pg /ml is undetectable. f mRNA expression level of IL6, IL8, RIG-I, TNFα measured by qRT-PCR at postinfected 5 h by 2B mutants and WT. Data shows mean ± s.d., * P < 0.05, ** P < 0.01, n = 3, Student’s t-test. g Growth kinetics of poliovirus carrying 2B mutations identified in HeLa cells. Viruses replication was examined for 2B mutants and WT by one-step growth curves in HeLa cells (m.o.i=10). (Data are shown as mean ± s.d., ***P < 0.0001, n = 3, three replicates for each time, for each viral strain, Student’s t-test). h. SEAP reporter assay to assess the ability of 2B mutants to inhibit protein secretion activity. One-tailed Student’s t-test. *P ≤ 0.05, n = 3, three replicates for each time, for each viral strain

Fig. 6

Adaptive 2B mutants increased virus replication and pathogenesis immune-competent mice. a, b The adaptive 2B mutants induce less IFN response in primary murine embryo fibroblast cells (MEFs). a MEFs were infected with WT and 2B mutants at m.o.i=0.1. supernatant was collected to measure IFNβ by ELISA at 72 h postinfection. RNA copy number of poliovirus genome was measured by qRT-PCR on MEFs at postinfection 72 h. b mRNA expression level of IRF7 and ISG56 was measured by qRT-PCR on MEFs at postinfection 72 h (n = 3, three replicates for each time, for each viral strain. Student’s t-test). n.s. indicates P > 0.05; *P ≤ 0.05, **P ≤ 0.01. c Virus tissue distribution. Viral titers in kidney, muscle and brain of Tg21 mice inoculated with 10⁶ PFU of WT, 2B I2V, 2B I6V, and 2B I6M administered I.P. route. Data were presented as logarithm(mean ± s.d.), PFU per ml. n = 4, the number of the mice is four per time point per viral strain. Limited detection level is 20 PFU per gram tissue. d Survival curve of Tg21 mice Tg21 mice were injected with 10⁷ PFU WT, 2B I2V, 2B I6V, or 2B I6M administered by I.P. route. (n = 10, the number of the mice is ten per group)