Cellular Importin-α3 Expression Dynamics in the Lung Regulate Antiviral Response Pathways against Influenza A Virus Infection

Abstract

Importin-α adaptor proteins orchestrate dynamic nuclear transport processes involved in cellular homeostasis. Here, we show that importin-α3, one of the main NF-κB transporters, is the most abundantly expressed classical nuclear transport factor in the mammalian respiratory tract. Importin-α3 promoter activity is regulated by TNF-α-induced NF-κB in a concentration-dependent manner. High-level TNF-α-inducing highly pathogenic avian influenza A viruses (HPAIVs) isolated from fatal human cases harboring human-type polymerase signatures (PB2 627K, 701N) significantly downregulate importin-α3 mRNA expression in primary lung cells. Importin-α3 depletion is restored upon back-mutating the HPAIV polymerase into an avian-type signature (PB2 627E, 701D) that can no longer induce high TNF-α levels. Importin-α3-deficient mice show reduced NF-κB-activated antiviral gene expression and increased influenza lethality. Thus, importin-α3 plays a key role in antiviral immunity against influenza. Lifting the bottleneck in importin-α3 availability in the lung might provide a new strategy to combat respiratory virus infections.

Keywords: cytokine storm; immune sensor; influenza; lung; pneumonia.

Graphical abstract

Figure 1

Importin-α Expression in the Mammalian Respiratory Tract (A and B) Importin-α1, -α3, or -α5/7 proteins were stained (red) in human (A) or murine (B) upper (URT, nasal epithelial cells) and lower respiratory tract (LRT, bronchi and alveoli) sections by IHC-P (immunohistochemistry-paraffin protocol) and counterstained with hematoxylin. The importin-α7 antibody cross-reacts with importin-α5. 400× (URT and bronchi in A) or 1000× (alveoli in A and B) original magnification. Scale bar, 10 μm. (C and D) Importin-α mRNA levels in the human (C) and murine (D) URT and LRT determined by qRT-PCR. Relative importin-α1 expression values were set to 1 after normalization against GAPDH (Glycerinaldehyd-3-phosphat-dehydrogenase). Each data point represents an individual sample (n = 6–12). (E and F) Importin-α protein amounts in murine LRT. Western blot analyses using importin-α isoform-specific antibodies and GAPDH adjustment were performed to determine endogenous importin-α protein amounts (α1, α3, α5, α7) in murine organs. For relative quantification (E), standard curves of affinity-purified, FLAG-tagged importin-α proteins were used. Relative importin-α1 protein amount in murine LRT was set to 1. Data shown represent means ± SD (n = 3 biological replicates; technical replicates: n = 1–2 per organ, n = 1–2 western blot analyses). Representative western blots for the endogenous importin-α isoforms and the standard curves are shown in (F). The gaps in (F) depict cropping of the relevant bands run on the same gel. (G–K) Comparison of importin-α mRNA expression levels between murine URT and LRT: α1 (G), α3 (H), α4 (I), α5 (J), or α7 (K). A, alveoli; BE, bronchiolar epithelium; n = 8–12. (L) Schematic partition of the RT with increasing importin-α mRNA levels from URT to LRT. ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001.

Figure 2

Molecular Function and Regulation of Importin-α3 (A) Schematic overview of importin-α3-mediated NF-κB p65 nuclear translocation. (B) Pull-down of NF-κB from lysates of untreated or TNF-α-treated A549 cells with Sepharose-immobilized GST-importin-α1 and -α3. Expression of GST-importin-α isoforms was verified with Coomassie staining (upper panel), and bound NF-κB p105, p65, and p50 were detected via western blot (lower panel). GAPDH protein detection served as loading control. One representative experiment is shown out of three independent experiments. (C) Effect of TNF-α treatment on expression and localization of NF-κB p65 and importin-α3 in WT MEFs (n = 3). (D and E) Nuclear fraction (NF; control p84) and cytoplasmic fraction (CF; control GAPDH) localization of NF-κB p65 in α3^−/− MEFs or WT controls thereof (α3^+/+) after TNF-α treatment using cell fractionation assay (D) or immunofluorescence (E) as well as quantifications thereof (n = 3–5). Scale bar, 20 μm. (F and G) After seeding, cells were serum starved for 24 h. Subsequently, cells were control treated (w/o) or treated with TNF-α. Endogenous importin-α3 and NF-κB p65 protein levels were detected and quantified for whole-cell lysates (WCLs), CFs and NFs of WT MEFs (n = 3) (F), and WCLs of HSAEpCs (n = 3) (G). Control-treated samples of each fraction were set to 100%. (H) NF-κB binding sites in the promoter element of the human importin-α3 gene in A549 cells (Raskatov et al., 2012). (I–K) Effect of TNF-α treatment on importin-α3 promoter activity in HSAEpC (n = 2) (I), MEFs (n = 3) (J), and A549 cells (n = 6) (K) upon control (Ctr) or importin-α3 promoter reporter construct (α3) transfection. Relative importin-α3 promoter activity in control-treated samples was set to 100%. (L) Importin-α3 mRNA levels in mock or TNF-α-treated MEF NEMO^−/− cells (n = 6). Relative expression values of importin-α3 in samples were normalized to GAPDH, and importin-α1 mRNA expression was set to 1. Each data point represents an individual sample. (M) MEF NEMO^−/− cells were transfected with empty plasmid (Mock) or NEMO and untreated or treated with 50 ng TNF-α. Importin-α3 and NEMO protein expression levels were detected (n = 3). GAPDH protein detection served as loading control. (N) MEF NEMO^−/− cells were treated with 10 ng TNF-α, IL-6, IL-10, IFN-γ and TGF-β and importin-α3 mRNA levels were measured as described above (n = 3) (O) Importin-α3 gene expression is controlled by TNF-α-activated NF-κB in a dose-dependent manner. Low amounts of TNF-α induce activation of NF-κB and high expression of importin-α3 (upper panel). In contrast, high amounts of TNF-α induce high activation of NF-κB and low expression of importin-α3 (lower panel). Data shown represent means ± SD of at least three independent biological experiments. ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001.

Figure 3

HPAIV Replication Kinetics in Importin-α3^−/− Cells (A) TNF-α induction in MEFs infected with avian-type SC35-PB2_701D or human-type SC35M-PB2_701N H7N7 HPAIV (MOI = 0.1) measured by ELISA (n = 3). (B–E) Protein extracts thereof were analyzed and quantified by western blot at 48 h p.i. Importin-α3 and viral NP protein were detected (B and D) and quantified (C and E) with GAPDH (B and C) or NP adjustment (D and E) (n = 4). Dashed lines in (D) depict cropping of the relevant bands run on the same gel. (F and G) Additionally, importin-α1, -α4, -α5, and -α7 and viral NP protein were detected and quantified with GAPDH adjustment to confirm that observed effects were specific for importin-α3. Control-treated, uninfected samples or SC35-PB2_701D-infected samples were set to 100%. Cross-reactivity of importin-α antibodies with NP resulted in a band (#) in (B) and (F). (H–K) Growth kinetics of avian-type SC35-PB2_701D (H and I) and human-type SC35M-PB2_701N (J and K) H7N7 HPAIV in importin-α^−/− MEFs (n = 3). WT (black), importin-α3^−/− (α3^−/−, red), and importin-α7^−/− (α7^−/−, green) MEFs were infected with low (H and J) or high (I and K) multiplicities of infection (MOI). Virus titers given in plaque-forming units (p.f.u.) were determined by plaque assay on MDCKII cells (0, 24, 48, 72, and 96 h p.i.). (L) H1299 α3^+/+ (black) and α3^−/− (red) cells were infected with human-type SC35M-PB2_701N HPAIV (MOI = 0.001) and replication kinetics measured in the presence or absence of TNF-α (n = 3). Virus titers were determined by plaque assay on MDCKII cells (0, 24, 48 h p.i.). Data shown represent means of relative importin-α amounts or logarithmic virus titers ± SD of at least three independent biological experiments (*p ˂ 0.05; ^∗∗∗p < 0.001).

Figure 4

High Importin-α3 Homology across Species and Its Role in HPAIV Interspecies Transmission from Birds to Humans (A–F) Importin-α homologies across animal kingdoms. Homolgies for importin-α1 (A), -α3 (B), -α4 (C), -α5 (D), -α6 (E), and -α7 (F). Pairwise amino acid sequence identities were calculated for importin-α across species. (G and H) Primary human bronchial (HBEpC; G) or primary human small airway epithelial cells (HSAEpC; H) were infected with clinical fatal case A/Thailand/1(KAN-1)/2004 (H5N1, MOI = 1) or A/Netherlands/219/2003 (H7N7, MOI = 10) isolates harboring the original human-type signatures (PB2_701N or PB2_627K, respectively) or their recombinant avian-type counterparts (PB2_701D or PB2_627E, respectively). Importin-α3 was quantified in NP-adjusted western blots of HBEpC (H) and HSAEpC (I) infected with H5N1 (MOI = 0.1) or H7N7 (MOI = 1). Importin-α3/NP ratios of human-type virus infected samples at 24 h p.i. were set to 100%. Data shown represent means ± SD of at least three independent biological experiments. ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001.

Figure 5

Importin-α3 Expression and Downregulation in Murine Lungs (A–F) WT mice were control treated or infected with 6 × 10⁴ p.f.u. of H7N7 recombinant viruses. Importin-α mRNA expression levels (α1, α3, α4, α5, α7) on day 3 p.i. in lungs upon control treatment (A) or infection with SC35-PB2_701D (B), SC35M-PB2_701N (C), SC35-PB2_701N (D), SC35-NP_319K (E), or SC35-PB2_701N-NP_319K (F) viruses. Relative importin-α1 expression values were set to 1 after normalization against GAPDH. Each data point represents an individual sample (n = 5, ^∗p < 0.05). (G and H) Immunohistochemical staining of importin-α3 (G) and viral antigen (H) on day 3 p.i. in the LRT upon infection with H7N7 HPAIV. 400× original magnification. Scale bar, 10 μm. Immunohistochemical staining of virus antigen (red-brown) in the URT and LRT in α3^−/− mice or their WT litters α3^+/+ infected with 6 × 10⁴ p.f.u. of SC35-PB2_701D or SC35M-PB2_701N viruses on day 3 p.i. Control mice received PBS.

Figure 6

Genes Responsive to HPAIV Infection in Lungs of WT and Importin-α3^−/− Mice WT and importin-α3^−/− (α3^−/−) mice were infected with SC35-PB2_701D (H7N7) or SC35M-PB2_701N (H7N7). Control mice received PBS. (A) Heatmap depicting the row Z scores of the normalized expression values of 143 differentially expressed genes (p value ≤ 0.01 and |log ₂FC| ≥ 1.5) that were clustered according to their expression profiles in lungs on day 3 p.i. (B) Heatmap depicting row Z scores of the normalized expression values of 79 genes with a previously described role in innate immunity (Engels et al., 2017). (C) TNF-α expression levels detected by transcriptome analysis on day 3 p.i. in lungs of infected α3^+/+ mice relative to the maximum expression observed. Control was set to 100%. (D) TNF-α protein levels measured by ELISA on day 1 p.i. in α3^+/+ and α3^−/− mice. Data show average results obtained from pooled lung samples of three WT mice per virus strain (^∗p < 0.05; ^∗∗∗p < 0.001). (E and F) Relative mRNA-expression levels of antiviral genes under an NF-κB promoter determined by qRT-PCR 3 days p.i. in control (E) and infected lungs (F) of α3^+/+ and α3^−/− mice. Relative expression values of α3^+/+ for each gene normalized to HPRT were set to 1. Data shown represent mean ± SD (n = 5; ^∗∗p < 0.01; ^∗∗∗p < 0.001). (G and H) Survival in α3^+/+ and α3^−/− mice infected with avian- (G) or human-type (H) H7N7 HPAIV. Data shown represent means ± SD (n = 9 for SC35 and n = 4 for SC53M).