Background
H9N2 influenza virus exists on all continents except Antarctica and is the most common subtype of influenza viruses isolated from poultry (chickens and ducks) in China [
1,
2]. H9N2 viruses have been isolated from pigs and humans with influenza-like illness in Hong Kong and Mainland China [
3‐
9], demonstrating that the H9N2 influenza virus could cross species barriers and expand its host range from birds to mammalians. Recently, a research showed that H9N2 viruses and pandemic H1N1 viruses have high genetic compatibility and they can produce higher pathogenic reassortment in experimental condition [
10]. In addition, H9N2 viruses provide their six inner genes to contribute to the evolution of the H7N9, H10N8 and H5N6 viruses that cause severe human respiratory infections in China [
11‐
13]. All these features indicate that H9N2 virus has a considerable public health threat. Thus, it is valuable to reveal the pathogenesis of H9N2 influenza virus infection and the innate immune responses of host to the H9N2 viruses.
Pathogen invasion could be recognized by pattern recognition receptors (PRRs) and results in production of interferon [
14]. The type I interferon binds to interferon-α/β receptors and activates JAK-STAT signaling pathway, resulting in the expression of hundreds of interferon-stimulated genes (ISGs) [
15]. Products of these genes are mostly antiviral proteins which are essential for mediating the antiviral state of the host. Interferon-induced transmembrane proteins (IFITMs) were identified nearly 30 years ago, and although other functions have been proposed, the primary role of IFITM proteins seems to be antiviral. IFITM proteins are a family of small transmembrane proteins that are induced strongly by interferon, but that are also expressed basally in several cell types and lines [
16,
17]. To date, three IFITM proteins (IFITM1, IFITM2 and IFITM3) have been identified to display broad spectrum antiviral activity in human and mice [
18,
19]. Especially, the IFITM1 and IFITM3 were proven to be more effective in resisting influenza virus infection [
20,
21]. The siRNA-mediated depletion of IFITMs significantly increased H1N1 virus titers in primary lung fibroblasts and in HeLa cells, while overexpression of IFITM1, 2, 3 could improve resistance to the H3N2 influenza in A549 and MDCK cell lines [
22]. Although the antiviral activity of IFITMs has been identified in multiple types of cells, the antiviral activity of IFITM1 has not been reported in endothelial cells infected with H9N2 virus or inoculated with inactivated viral particle.
Generally, airway epithelial cells are considered to be the main target cells of influenza viruses because of expressing two kinds of influenza virus receptors (α-2, 6- and α-2, 3-linked sialic acid receptors) [
23‐
25]. But the more and more evidences indicated that endothelial cells might play an important role in response to influenza virus infection. Most endothelial cells are also showed to express two kinds of influenza virus receptors which provide potential possibility for influenza virus infection in endothelial cells [
26]. H5N1 and H7N9 influenza viruses have been proven to directly infect human lung microvascular endothelial cells and replicate in endothelial cell lines [
27,
28]. Several studies indicated that endothelial cells might be the source of the cytokines and involved in lung injury during influenza virus infection [
29‐
32]. So, it is important to understand the cellular responses in endothelial cell during influenza virus infection. Our previous study revealed that H9N2 virus could infect human umbilical vein endothelial cells (HUVECs) and induce high level expression of several ISGs, especially IFITM1 [
33]. The microarray results showed that both H9N2 virus infection and inactivated viral particle inoculation increase the expression of IFITM1 at transcription level, and the viral particle inoculation induces a higher level (44.75 folds) of IFITM1 than that (8.53 folds) induced by virus infection. However, the level of IFITM1 needs to be quantified by Real-Time PCR and western blot, and little is known about the antiviral activity of IFITM1 induced by H9N2 virus and viral particle in HUVECs. The present study aimed to quantify the expression of IFITM1 induced by H9N2 virus and viral particle and investigate the antiviral state mediated by IFITM1 in HUVECs. And the results showed that both H9N2 virus infection and viral particle inoculation significantly increased the expression of IFITM1 at mRNA and protein levels, and the IFITM1 protein induced by viral particle inoculation significantly enhanced the antiviral state of HUVECs against H9N2 virus infection.
Methods
Experiment protocol
We first quantified the expression levels of IFITM1 in HUVECs induced by H9N2 virus infection or viral particle inoculation using quantitative real-time PCR and western blot. Secondly, we detected levels of interferon-ɑ/β using ELISA kits. Thirdly, we observed the effect of HA and NA on IFITM1 expression, and located the position of inactivated viral particles in HUVECs. Finally, we investigated IFITM1-mediated antiviral response by siRNA and activation plasmid transfection. To compare the expression patterns of IFITM1 between endothelial cells and epithelial cells, all of the above experiments were performed on human epithelial cells at the same time.
Cells culture
Human umbilical vein endothelial cells (HUVECs, CRL-1730, ATCC), human bronchus epithelial cells (BEAS-2Bs, CRL-9609, ATCC) were used for detecting the expression of IFITM1 and secretion of interferon-ɑ/β. Madin Darby canine kidney cells (MDCK, CCL-34, ATCC) were used for plaque assay. These cell lines were cultured in DMEM (Gibco) supplemented with 10% FBS (Gibco), 100 U/mL of penicillin G and 100 μg/mL of streptomycin (Gibco) at 37 °C in a 5% CO2 incubator. The cells in passages 5~8 were seeded into six-well plates and cultured for 24 h before each experiment.
Virus, virus infection and virus inactivation
The H9N2 virus used in this study was A/Chicken/Hebei/4/2008 (H9N2) (Ck/HB/4/08). The complete genome sequences of the virus are available from GenBank under accession numbers FJ499463–FJ49947033. The viruses were propagated in 9-day-old embryonated eggs from specific-pathogen-free (SPF) hens at 37 °C for 60 to 72 h. Virus titers were determined by plaque assay. Our previous study demonstrated that exogenous trypsin was required for the efficient replication of the H9N2 virus in HUVECs [
33]. Thus, 0.25 μg/mL exogenousl-1-tosylamide-2-phenylethyl chloromethyl ketone (TPCK)-treated trypsin was added to the medium for all experiments. To inspect the interaction between viral particles and HUVECs, we created inactivated H9N2 viral particles by using 0.094% β-propionolactone (SERVA Electrophoresis) according to previous description [
34], and plaque assay was used to evaluate whether viral replicative capacity was completely destroyed.
Quantitative real-time PCR analysis
To quantify the levels of IFITM1 induced by H9N2 virus and viral particle, HUVECs or BEAS-2Bs were divided into three groups: control group (inoculation with virus-free media), H9N2 virus group (infection with virus at a multiplicity of infection (MOI) of 5) and viral particle group (inoculation with viral particles at a MOI of 5). To detect the effect of HA and NA (Sino Biological) on the mRNA expression of IFITM1, HUVECs or BEAS-2Bs were divided into control group, low concentration group, high concentration group. Cells in control group were inoculated with free media, cells in low and high concentration groups were respectively treated with 0.1 μg/mL and 1 μg/mL HA or NA [
35,
36]. RNA in each group used for Real-Time PCR analysis was extracted using TRIzol reagent (Invitrogen) at 6, 12 and 24 h after treatment. cDNA was synthesized using random hexamer and a superscript III reverse transcriptase kit (Invitrogen). Real-time PCR was performed on the cDNA for RNA quantitation using Ex TaqMix (Takara Bio) and Eva green (Biotium) for IFITM1. GAPDH was amplified in parallel with the target genes as an endogenous control and all samples were analyzed in triplicate. The fold changes of specific mRNA from different groups compared to control group. The primer sequences were used as follows: GAPDH, F (5′ to 3′): ACAACTTT GGTATCGTGGAAGGAC and R (5′ to 3′): AGGGATGATGTTCTGGAGAGCC, IFITM1, F (5′ to 3′): ACTGAAACGACAGGGGAAAG and R (5′ to 3′): GAACAGGGACCAGACGACAT.
Western blot analysis
Cells in each group used for western blot analysis were collected at 12, 24 36 h after treatment and lysed with RIPA lysis buffer (Cell Signaling Technology). The cell lysates were centrifuged and the resultant supernatants were resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto polyvinylidene difluoride (PVDF) membranes (Roche). The membranes were blocked with 5% skim milk and probed with a monoclonal antibody to β-actin (Santa Cruz Biotechnology), IFITM1 (Sigma-Aldrich). After a further incubation with peroxidase (HRP) conjugated secondary antibody (ORIGENE). Proteins were visualized using enhanced chemiluminescence. The relative protein level of IFITM1 to β-actin was analyzed by Image J software.
ELISA assay
ELISA assay was used to quantify the levels of interferon-ɑ/β. Supernatant in each group was collected at 6, 12 and 24 h postinfection, and analyzed by ELISA kits (R & D Systems). According to previous description [
37,
38], cells in positive control group were treated with poly I: C (Sigma) at the concentration of 10 μg/mL.
Plaque assay
The plaque assay was used to detect the virus titers propagated in eggs and cell culture supernatant. Plaque assay was performed on MDCK cells as described previously [
39]. Briefly, MDCK cells were seeded into 6-well plates. Confluent monolayers were washed with phosphate buffered saline (PBS) and infected with H9N2 virus or cell culture supernatants. The inoculum was discarded after incubation for 1 h and the remaining cells were washed with PBS. An overlay consisting of a mixture of 1.6% agarose (Lonza) and double-strength DMEM with 0.25 μg/mL TPCK-trypsin (Worthington) was added to the above cells and incubated at 37 °C for 72 h. Plaques were stained with 0.1% crystal violet and counted.
siRNA transfection
For interference assay, cells were transfected with control siRNA or IFITM1 specific siRNA using Lipofectamine 3000 transfection reagent kit (Invitrogen). IFITM1 specific and control siRNA used in the study were purchased from Santa Cruz Biotechnology, and the siRNA for human cells is a pool of 3 target-specific 19–25 nt siRNAs designed to knock down gene expression. According to instructions of products, we firstly prepared mixtures (0.2 μM siRNA and 7.5 μL Lipofectamine 3000) of siRNA and Lipofectamine 3000 transfection reagents using Opti-MEM Media. To interfere the expression of IFITM1, cells were infected with H9N2 virus or viral particle and incubated for 1 h, then cells were covered with mixtures for 36 h at 37 °C in a 5% CO2 incubator. The interference effect of IFITM1 specific siRNA was detected by western blot at 36 h postinfection.
IFITM1 CRISPR activation plasmid transfection
For overexpression assay, cells were transfected with control plasmid or IFITM1 CRISPR activation plasmid (Santa Cruz Biotechnology) using Lipofectamine 3000 transfection reagent kit (Invitrogen). The IFITM1 CRISPER activating plasmid is a synergy activation medium (SAM) transcriptional activation system designed to specifically upregulate gene expression. According to instructions of products, we firstly prepared mixtures (2 μg plasmid and 7.5 μL Lipofectamine 3000) of plasmid and Lipofectamine 3000 transfection reagents using Opti-MEM Media. Then cells were covered with the mixtures for 36 h at 37 °C in a 5% CO2 incubator. The overexpression level of IFITM1 was detected by western blot at 36 h after transfection.
Immunofluorescence
Immunofluorescence was used to locate the position of viral particles in cells. According to previous description [
40], HUVECs or BEAS-2Bs were directly fixed in 2% paraformaldehyde, permeabilized with 0.5% Triton X-100. After blocking with phosphate-buffered saline (PBS) containing 5% bovine serum albumin (BSA) for 1 h at 37 °C. Cells were labeled with anti-nucleoprotein (NP) antibody A-3 (Aviva Systems Biology) overnight, followed by stained with FITC conjugated anti-rabbit secondary antibody (ORIGENE) for 1 h at 37 °C. The images were captured using OLYMPUS fluorescence microscopy.
Statistical analysis
The data were expressed as means ± standard deviations (SD). All the statistical tests were performed using GraphPad Prism software (version 6.0). Statistical significance of differences were determined using the Student’s t-test or one-way analysis of variance (ANOVA). P < 0.05 was considered statistically significant.
Discussion
The expression of ISGs is an early response of host to virus infection, and their products confers host the antiviral state which inhibits the entry process or replication of invading virus. To date, multiple proteins translated by ISGs have been validated as antiviral proteins, such as protein kinase R (PKR), myxovirus-resistance proteins, ISG15, Schlafen 11 and so on. IFITM proteins are recently identified antiviral factors that play critical roles in the intrinsic and interferon-mediated control of virus infection [
20,
22]. Since initially identified by a siRNA screen for factors that restrict influenza virus replication, more and more researches revealed the inhibition effect of IFITMs on enveloped viruses through affecting the interaction between virus envelope proteins and endosomal or lysosomal [
43]. Recently, a study demonstrated that depletion of IFITM1 with siRNA increased titers of H1N1 virus in primary lung fibroblast cells and in HeLa cell line, overexpression of IFITM1 resisted H3N2 virus infection in A549 and MDCK cell lines [
22]. In this study, our results showed that overexpression of IFITM1 significantly reduced the virus titers in HUVECs and BEAS-2Bs (Fig.
8). The present data reaffirmed the restriction effect of IFITM1 on influenza virus infection. To investigate whether IFITM1 induced by H9N2 virus infection and viral particle inoculation mediated the antiviral response in HUVECs, we detected the virus titers in HUVECs transfected with IFITM1 specific siRNA or control siRNA. The results showed that IFITM1 induced by H9N2 virus infection could not enhanced the antiviral state in HUVECs. In contrast, IFITM1 induced by viral particle inoculation significantly enhance the antiviral response in HUVECs (Fig.
7). Taking into account the above results, we hypothesized that virus replication preceded the expression of IFITM1 in initially infected cells. A previous study demonstrated that IFITM3 efficiently restricted influenza virus and IFITM1 modestly restricted influenza virus, and the expression patterns of IFITMs are likely to be an independent determinant of viral tropism [
20]. Our previous microarray results showed that H9N2 virus and viral particle do not induce the expression of IFITM3 in HUVECs [
33]. We speculated that this might be a unique response of HUVECs to H9N2 virus infection or viral particle inoculation. Taken together, our data may offer further insight into the innate immune response of endothelial cells to influenza virus infection.
In the present study, the level of IFITM1 induced by viral particle inoculation was higher than that induced by H9N2 virus infection in HUVECs (Fig.
1). The observation differed to what we generally expected, so it is worthy of thinking the sense of IFITM1 expression upregulated by viral particle. Moreover, a recent study indicated that H9N2 influenza virus infection induces the expression of IFITM1 in lung, heart and liver in BALB/c mice [
44]. In vivo, the basal lamina with an average thickness of 1 μm is the only structure which separates epithelial cells and endothelial cells [
45]. Death of infected epithelial cells creates gaps and the released virus particles readily access to endothelial cells [
46,
47]. Thus, the released virus particles may stimulate the antiviral response in endothelial cells. In addition, cytokines produced by epithelial cells could further activate neighboring endothelial cells during influenza virus infection [
48]. So it is conceivable that interferon-α/β released by infected epithelial cells could readily induce the expression of IFITM proteins in endothelial cells. Taken together, we may consider to enhance the antiviral state of host by stimulating the respiratory tract with inactivated viral particles.
Previous studies indicated that H5N1 virus infection up-regulate the expression of type I interferon in human pulmonary microvascular endothelial cells and induce high levels of interferon-β in HUVECs [
28,
49]. However, our results showed that expression of IFITM1 induced by H9N2 virus and viral particle independently of interferon-α/β in HUVECs (Fig.
2). Although a recent study demonstrated that human IFITM3 and mouse IFITM3 are induced by cytokines of the gp130 family (such as IL-6) [
50], we failed to find a possible mechanism involved in IFITM1 expression. Thus, we investigated the effects of HA and NA proteins on the expression of IFITM1 in HUVECs. The results showed that treatment with HA or NA could not upregulate the levels of interferon-α/β and IFITM1 (Figs.
3,
4). Virus RNA activates the cellular antiviral response has been widely reported. In the present study, we did not investigate the effect of RNA of H9N2 virus on the antiviral response in endothelial cells. However, our results showed that H9N2 virus could infect and replicate in HUVECs with invalid effect on antiviral response of HUVECs. The results indicated that RNA of H9N2 virus may play an insignificant role in the expression of IFITM1 induced by H9N2 virus infection or viral particle inoculation. We speculated that the induction of IFITM1 expression may depend on the interaction between viral particles and cellular factors. Moreover, a previous study demonstrated that interactions between cellular factors and envelope glycoprotein B of replication defective human cytomegalovirus may induce ISGs expression [
51]. Then we verified whether the viral particles were taken up by cells using immunofluorescence. The results showed that viral particles could enter the HUVECs (Fig.
5), suggesting that cellular interaction might be involved in the induction of IFITM1 in HUVECs. Obviously, more work needs to be done to explore this induction mechanism, and data of the present study might considerably narrow the range of possible mechanisms.
Generally, human influenza virus and avian influenza virus prefer to infect epithelial cells expressing α-2, 6- and α-2, 3-linked sialic acid receptors [
24,
25]. To investigate whether these phenomena were unique responses to H9N2 virus and viral particle in HUVECs, we performed the same experiments on epithelial cells synchronously. Our results showed that endothelial cells and epithelial cells shared different features. Consistent with a previous study [
52], H9N2 virus infection and viral particle inoculation elevated the levels of type I interferon in BEAS-2Bs (Fig.
2). In addition, contrary to endothelial cells, H9N2 virus infection induced a higher level of IFITM1 in BEAS-2Bs (Fig.
1). These data suggested that H9N2 virus or viral particle may stimulate the innate immune response via different ways in vivo.
It is of course that further studies are necessary to go beyond our present results. For example, further investigation needs to be done for revealing the precise mechanism of distinct expression patterns of IFITM1 in epithelial cells and endothelial cells. In particular, whether inactivated viral particle inoculation increases IFITM1 expression in endothelial cells in vivo and enhances the antiviral state of host.