Levels of Influenza A Virus Defective Viral Genomes Determine Pathogenesis in the BALB/c Mouse Model

Abstract

Defective viral genomes (DVGs), which are generated by the viral polymerase in error during RNA replication, can trigger innate immunity and are implicated in altering the clinical outcome of infection. Here, we investigated the impact of DVGs on innate immunity and pathogenicity in a BALB/c mouse model of influenza virus infection. We generated stocks of influenza viruses containing the internal genes of an H5N1 virus that contained different levels of DVGs (indicated by different genome-to-PFU ratios). In lung epithelial cells, the high-DVG stock was immunostimulatory at early time points postinfection. DVGs were amplified during virus replication in myeloid immune cells and triggered proinflammatory cytokine production. In the mouse model, infection with the different virus stocks produced divergent outcomes. The high-DVG stock induced an early type I interferon (IFN) response that limited viral replication in the lungs, resulting in minimal weight loss. In contrast, the virus stock with low levels of DVGs replicated to high titers and amplified DVGs over time, resulting in elevated levels of proinflammatory cytokines accompanied by rapid weight loss and increased morbidity and mortality. Our results suggest that the timing and levels of immunostimulatory DVGs generated during infection contribute to H5N1 pathogenesis. IMPORTANCE Mammalian infections with highly pathogenic avian influenza viruses (HPAIVs) cause severe disease associated with excessive proinflammatory cytokine production. Aberrant replication products, such as defective viral genomes (DVGs), can stimulate the antiviral response, and cytokine induction is associated with their emergence in vivo. We show that stocks of a recombinant virus containing HPAIV internal genes that differ in their amounts of DVGs have vastly diverse outcomes in a mouse model. The high-DVG stock resulted in extremely mild disease due to suppression of viral replication. Conversely, the stock that contained low DVGs but rapidly accumulated DVGs over the course of infection led to severe disease. Therefore, the timing of DVG amplification and proinflammatory cytokine production impact disease outcome, and these findings demonstrate that not all DVG generation reduces viral virulence. This study also emphasizes the crucial requirement to examine the quality of virus preparations regarding DVG content to ensure reproducible research.

Keywords: defective viral genomes; influenza; pathogenesis.

FIG 1

Virus stocks grown at different MOIs contain different amounts of infectious viral particles. (a) Schematic showing the generation of the virus stocks used in this study. (b) Ratio of genome copy number/mL to PFU/mL. Mean PFU/mL was determined for each virus stock (n = 3), and this value was used to calculate the ratio. RNA was extracted from viral stocks, and a one-step RT-qPCR performed using primers and probe for the M gene to calculate the M gene copy number/mL. Data points show ratios calculated from RT-qPCR using RNA obtained from at least two independent extractions. Bars and error bars represent mean values ± SD. (c) Infectious-to-total (I/T) particle ratios in virus stocks.

FIG 2

Characterization of DVGs identified in virus stocks. (a) Gel electrophoresis of RT-PCR products of DVGs and full genomes from viral stocks using segment-specific terminal primers for the polymerase genes. (b) DVGs identified in viral stocks through Sanger sequencing of TOPO-cloned RT-PCR products for PB1, PB2, and PA segments. All bands not corresponding to the full-length genomes (presumed to be DVGs) from the 6:2 Tky/05 and 7:1 Tky/05 LOW virus stocks were cloned and sequenced. One PCR product from each polymerase segment (band sizes are 479 nt for PB1, 425 nt for PB2, and 388 nt for PA) was also cloned and sequenced from the 7:1 Tky/05 and 7:1 Tky/05 HIGH virus stocks, and these were identical in sequence to those present in the 7:1 Tky/05 LOW stock (marked by asterisks).

FIG 3

Growth kinetics, accumulation of DVGs, and type I IFN induction in infected cells. (a) Growth kinetics of 6:2 Tky/05, 7:1 Tky/05 HIGH, and 7:1 Tky/05 LOW viruses in MDCK and A549 cells. MOIs used for infections are displayed and were normalized by PFU. At designated time points, supernatants were taken and plaque assays performed to determine viral titers. Data are shown as mean values ± standard deviations (SD) (n = 3). Dotted lines represent the limit of detection. (b) A549 cells were infected using the viruses whose growth kinetics are shown in panel a at an MOI of 1 or mock infected. At 2, 6, and 24 h.p.i., total RNA was extracted, and RT-PCR was performed using terminal primers to detect full genomes and DVGs (n = 3). The gels each show the results for one representative well. (c) A549 cells were infected using the viruses whose growth kinetics are shown in panel a at an MOI of 1 or mock infected. At 2, 6, and 24 h.p.i., total RNA was extracted, and RT-qPCR was performed using primers for IFNB. Relative expression was calculated by the 2^−ΔΔCT method and normalized to the results for the GAPDH housekeeping gene, and data points represent the fold increase over the value for the mock-infected cells. Bars and error bars represent the mean values from 2 or 3 biological replicates ± SD. (d) IFN-β protein levels in A549 cell supernatants at 24 h.p.i. as determined by ELISA. Bars and error bars represent mean values ± SD (n = 3). Dotted line represents the limit of detection. (e) A549 IFN-β luciferase reporter cells were infected with an MOI of 1 of the same virus stocks whose growth kinetics are shown in panel a or mock infected, and at 2, 6, and 24 h.p.i., cells were lysed and luciferase expression measured. Bars and error bars represent mean values ± SD (n = 3). Variance among groups was calculated by two-way analysis of variance (ANOVA) with Tukey’s post hoc test for multiple comparisons. *, P < 0.05; ***, P < 0.001; ****, P < 0.0001.

FIG 4

DVGs replicate in BMDMs and trigger proinflammatory cytokines. (a) BMDMs were infected at an MOI of 1 with the 6:2 Tky/05, 7:1 Tky/05 HIGH, and 7:1 Tky/05 LOW virus stocks or mock infected (n = 3). At 2, 6, and 24 h.p.i., total RNA was extracted, and one-step RT-qPCR was performed using primers and probe for the M gene. Horizontal bars represent the mean values. (b) Gel electrophoresis of RT-PCR products using segment-specific terminal primers to detect full-length polymerase genes and DVGs from infected BMDMs. The gels each show the results for one representative well. (c and d) RT-qPCR for mRNA Il6 and Tnf gene expression was performed on infected BMDMs. Relative expression was calculated by the 2^−ΔΔCT method and normalized to the value for the Gapdh housekeeping gene, and data points represent the fold increase over the value for the mock-infected cells. Bars and error bars represent the mean values ± SD (n = 3). Variance among groups was calculated by two-way ANOVA with Tukey’s post hoc test for multiple comparisons. (e) BMDMs were infected at an MOI of 10 with the 6:2 Tky/05 and 7:1 Tky/05 LOW virus stocks or mock infected (n = 4). At 24 h.p.i., IFN-α protein levels in supernatants were determined by ELISA. Dotted line represents the limit of detection. Bars and error bars represent the mean values ± SD. Variance among the groups was determined by one-way ANOVA with Tukey’s post hoc test for multiple comparisons. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

FIG 5

Tky/05 virus stocks containing different levels of DVGs have diverse infection outcomes. (a) Six- to 8-week-old female BALB/c mice (15 per group) were mock infected or infected intranasally with 10⁵ PFU of 6:2 Tky/05 and 7:1 Tky/05 virus in 25-μL volumes. Five mice per group were culled at 48 h, 72 h, and 7 days postinfection. Weight loss was monitored daily. Data represent the mean values ± SD. Differences between the 6:2 Tky/05- and 7:1 Tky/05 virus-infected mice were analyzed by two-way ANOVA with Tukey’s post hoc test for multiple comparisons. (b) Virus titers in homogenized lung tissues were determined by plaque assay at 48 and 72 h postinfection. Horizontal bars represent the mean values. (c) Six- to 8-week-old female BALB/c mice (25 per group) were mock infected (PBS) or infected intranasally with 10⁵ PFU of 6:2 Tky/05, 7:1 Tky/05 HIGH, or 7:1 Tky/05 LOW virus in 35-μL volumes. Five mice per group were culled at 6 h, 24 h, 48 h, 96 h, and 14 days postinfection. Weight loss was monitored daily. Data represent the mean values ± SD. Dark blue asterisks indicate statistically significant differences between mice infected with 6:2 Tky/05 and 7:1 Tky/05 LOW viruses, gray asterisks indicate statistically significant differences between mice infected with 7:1 Tky/05 HIGH and 7:1 Tky/05 LOW viruses, and teal asterisks indicate statistically significant differences between mice infected with 7:1 Tky/05 HIGH and 6:2 Tky/05 viruses. Dotted line represents the severity limit. (d) Survival curves of infected mice. All mice were culled when they lost ≥20% of their original body weight (day zero). (e) Virus titers in homogenized lung tissues were determined by plaque assay at 6, 24, 48, and 96 h.p.i. (n = 5 per group). Horizontal bars and error bars represent the mean values ± SD. (f) RNA was extracted from the homogenized lungs, and one-step RT-qPCR performed using primers and probe for the M gene (n = 5 per group). Horizontal bars and error bars represent mean values ± SD. (g) BAL fluid was obtained at 6, 24, and 48 h.p.i., and cell counts calculated (n = 5 per group). Horizontal bars and error bars represent the mean values ± SD. Variance among groups was calculated by two-way ANOVA with Tukey’s post hoc test for multiple comparisons. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

FIG 6

Levels of DVGs in the Tky/05 virus stocks impact cytokine production. (a) Total RNA was extracted from murine lungs, and RT-qPCR for Ifna5 gene expression was performed. Relative expression was calculated by the 2^−ΔΔCT method and normalized to the value for the Gapdh housekeeping gene, and data points show the fold increase over the value for the mock-infected cells. Bars and error bars represent mean values ± SD (n = 5). Variance among groups was calculated by two-way ANOVA with Tukey’s post hoc test for multiple comparisons. (b to d) IFN-α, TNF, and IL-1B protein levels in murine lung homogenates at 6, 24, 48, and 96 h.p.i. as determined by ELISA. Horizontal bars represent the mean values (n = 5). (e to j) IFN-γ, MCP-1, IL-6, IP-10, TNF, and MIP-1β protein levels in the BAL fluid at 48 h.p.i. Horizontal bars represents the mean values (n = 5). Variance among groups was calculated by one-way ANOVA with Tukey’s post hoc test for multiple comparisons. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

FIG 7

Number of DVG junction reads per viral read in infected murine lungs. RNA was extracted from the lungs of the infected mice, sequenced, and analyzed by ViReMa. The junction reads for PB1 (a), PB2 (b), and PA (c) were divided by the total number of viral reads per sample. Each data point represents the value for 1 mouse. Horizontal bars represent the mean values.

FIG 8

DVGs cloned from the 7:1 Tky/05 virus are immunostimulatory. (a) Gel electrophoresis showing RT-PCR products of full-length (FL) genome or DVGs using segment-specific terminal primers for the polymerase genes. RNA from HEK293T cells transfected with 3P (PB1, PB2, and PA) and NP expression plasmids, reporter plasmids, and DVG or FL genome template was extracted 24 h posttransfection. The gel was spliced to remove wells for clarity and labeling alignment with panel b. (b) IFN-β promoter activation induced by the replication of either full-length segment or DVG. Bars and error bars represent mean values ± SD (n = 3, representative of 3 independent experiments). Differences between each DVG and the full-length segment it was derived from were calculated by the two-sided unpaired t test. *, P < 0.05; **, P < 0.01.