Endothelial cell tropism is a determinant of H5N1 pathogenesis in mammalian species

Abstract

The cellular and molecular mechanisms underpinning the unusually high virulence of highly pathogenic avian influenza H5N1 viruses in mammalian species remains unknown. Here, we investigated if the cell tropism of H5N1 virus is a determinant of enhanced virulence in mammalian species. We engineered H5N1 viruses with restricted cell tropism through the exploitation of cell type-specific microRNA expression by incorporating microRNA target sites into the viral genome. Restriction of H5N1 replication in endothelial cells via miR-126 ameliorated disease symptoms, prevented systemic viral spread and limited mortality, despite showing similar levels of peak viral replication in the lungs as compared to control virus-infected mice. Similarly, restriction of H5N1 replication in endothelial cells resulted in ameliorated disease symptoms and decreased viral spread in ferrets. Our studies demonstrate that H5N1 infection of endothelial cells results in excessive production of cytokines and reduces endothelial barrier integrity in the lungs, which culminates in vascular leakage and viral pneumonia. Importantly, our studies suggest a need for a combinational therapy that targets viral components, suppresses host immune responses, and improves endothelial barrier integrity for the treatment of highly pathogenic H5N1 virus infections.

Fig 1. Generation and characterization of H5N1 viruses with restricted tropism.

(A) Schematic representation of NP mRNA carrying four copies of miRNA target sites in the 3’UTR. The incorporated target sequences, which are Complementary to the mature miRNA, are shown below. (B) Evaluation of miRNA mediated restriction of NP expression. HEK-293T cells were co-transfected with plasmids expressing specific miRNAs (vector, miR142, miR126) and NP with miRNA target sites, and NP expression was analyzed by western blot. DblT refers to the NP segment carrying two miR-142-3p and two miR-126-3p target sites. Expression of Ku80 is shown as a loading control. (C-D) Evaluation of tropism restricted H5N1 virus replication in human and murine cells. (C) Human epithelial cells (A549, MOI = 0.001), endothelial cells (HMVEC, MOI = 0.01), monocytic leukemia cells (THP-1, MOI = 3), and (D) mouse epithelial cells (LA-4; MOI = 1), endothelial cells (MS1, MOI = 1) and dendritic cells (BMDC, MOI = 0.1) were infected with tropism restricted H5N1 viruses and viral titers at 48 hrs pi (hpi) were determined by plaque assay for all but BMDC. The titers are shown as PFU/mL (mean ± SEM). The limit of detection is 10 PFU/mL. For BMDC, the level of infection at 24hpi was measured by flow cytometry using an anti-H5 antibody. Data is represented as percentage of HA⁺ cells in the CD11c⁺ population (mean ± SEM). Asterisk denotes statistical significance determined by one-way ANOVA in comparison to H5N1-ScrbT group and the values are denoted as *p<0.05, **p<0.01, ***p<0.001 and ns–non significant. Data presented here is a representative of at least three independent experiments performed in triplicate.

Fig 2. Restriction of H5N1 replication in endothelial cells ameliorates disease symptoms in mice.

(A) C57BL/6J mice (n = 10) were intranasally infected with different doses (2, 10, and 25 PFU) of tropism restricted H5N1 viruses and monitored daily for body weight and survival. Top—Body weight loss, shown as relative percentage of day 0 weight (mean ± SEM) and Bottom–Survival. For better display, survival lines are offset by a few percentage points. (B) Viral titers in the lungs of infected mice. C57BL/6J mice (n = 6–10) were infected with 25 PFU and viral loads in the lungs on days 2, 5, and 8 pi were determined by plaque assay (PFU/mL). Each data point represents an individual mouse (mean ± SEM). Asterisk denotes statistical significance determined by one-way ANOVA in comparison to H5N1-ScrbT group and the values are denoted as *p<0.05, **p<0.01, ***p<0.001 and ns–non significant. Data presented here is a pooled average of two independent experiments.

Fig 3. Restriction of cell tropism alters cytokine and chemokine responses in the lungs.

C57BL/6J mice (n = 8) were intranasally infected with a 25 PFU dose of the H5N1 viruses and total RNA from the lungs was isolated on day 5 pi via Trizol extraction. Expression levels of inflammatory genes (IFN-β, MIP-1α, CCL2, IL-6, IL-1β and TNFα) were measured by quantitative PCR analyses. Data is represented as fold expression relative to mock infected mice (mean ± SEM). Asterisk denotes statistical significance determined by one-way ANOVA and the values are denoted as *p<0.05, **p<0.01, ***p<0.001 and ns–non significant. Data presented here is a representative of at least two independent experiments.

Fig 4. H5N1 infection of endothelial cells increases vascular leakage and causes pronounced damage in the lungs.

(A) Evaluation of vascular leakage in H5N1 infected lungs. C57BL/6J mice (n = 5) were intranasally infected with 25 PFU of the H5N1 viruses. On day 7 pi, Evans Blue dye was injected into mice via retro-orbital route. After 1h, mice were euthanized and the levels of Evans Blue dye in the BALF were measured. Each data point represents the concentration of Evans Blue dye in the BALF of individual mice (mean ± SEM). (B) Evaluation of infection of various cell populations in vivo. C57BL/6J mice (n = 3) were infected with 25 PFU of the H5N1 viruses. On day 3 pi, lungs were harvested and analyzed for H5N1 infection of the various cell compartments by flow cytometry. Cell surface markers CD45 and CD31 were used to distinguish endothelial cells (CD45^-, CD31⁺), hematopoietic cells (CD45⁺) and non-hematopoietic cells (CD45^-), and surface expression of viral HA was used to define infected cells (HA⁺). Data is represented as the mean (± SEM). Asterisk denotes statistical significance determined by one-way ANOVA and the values are denoted as *p<0.05, **p<0.01, ***p<0.001 and ns–non significant. Data presented here is a representative of at least two independent experiments. (C) Histopathological analysis of murine lungs by hematoxylin and eosin staining. C57BL/6J mice (n = 5) were intranasally infected with 25 PFU of the H5N1 viruses and on day 5 post-infection, the lungs were isolated and analyzed by H&E staining. The areas of infection and inflammation are indicated and outlined in black. B-bronchiole, V-blood vessel.

Fig 5. Endothelial cell infection induced damage is necessary for extrapulmonary spread of H5N1 virus.

(A) Viral titers in different organs. C57BL/6J mice (n = 5) were intranasally infected with 25 PFU of the H5N1 viruses and on day 8 pi, different organs were harvested for viral titer determination. Each data point represents an individual mouse (PFU/mL). (B) Viral replication in the brain upon intracerebellar injection. C57BL/6J mice (n = 3–5) were injected with 100 PFU of the H5N1 viruses in the cerebellum and on day 3 pi whole brains were isolated for viral titer determination. Each data point represents the viral titer of an individual mouse (PFU/mL; mean ± SEM). The limit of detection is 10 PFU/mL. Asterisk denotes statistical significance determined by one-way ANOVA and the values are denoted as *p<0.05, **p<0.01, ***p<0.001 and ns–non significant. Data presented here is a representative of at least two independent experiments.

Fig 6. Restriction of H5N1 replication in endothelial cells ameliorates disease symptoms in ferrets.

Five-month old male finch ferrets that tested seronegative for circulating influenza viruses were intranasally infected with a 3500 PFU dose (n = 6/group) of the H5N1 viruses, and monitored for weight loss and clinical signs of infection. (A) Weight loss and survival of infected ferrets. On day 5 pi, H5N1-ScrbT infected ferrets showed severe signs of neuronal infection and were humanely euthanized. (B) Viral loads in the nasal washes. Nasal washes were performed on day 1 and 3 pi and viral titers were determined by plaque assay. (C) Viral replication at different sites in the respiratory tract and brain. On day 4 pi, three ferrets from each group were euthanized and different organs were harvested for viral titer determination. Each data point represents the viral titer of an individual ferret. The limit of detection is 10 PFU/mL. Asterisk denotes statistical significance determined by one-way ANOVA and the values are denoted as *P<0.05, **P<0.01, ***P<0.001 and ns–non significant. Ferret experiment was performed once.

Fig 7. Restriction of low pathogenic H5N1 replication in endothelial cells does not alter disease outcome in mice.

(A) Body weight loss and survival of infected mice. C57BL/6J mice (n = 5) were intranasally infected with 100 PFU of tropism restricted low pathogenic H5N1 viruses and monitored daily for body weight and survival. Left—Body weight loss, shown as relative percentage of day 0 weight (mean ± SEM) and Right–Survival. For better display, survival lines are offset by a few percentage points. (B) Viral titers in the lungs of infected mice. C57BL/6J mice (n = 5–6) were infected with 100 PFU and viral loads in the lungs on days 3, 6, and 8 pi were determined by plaque assay (PFU/mL). Each data point represents an individual mouse (mean ± SEM). Asterisk denotes statistical significance determined by one-way ANOVA in comparison to the H5N1-ScrbT group and the values are denoted as *p<0.05, **p<0.01, ***p<0.001 and ns–non significant. Data presented here is a representative of two independent experiments.