Transcription Factor Runx3 Is Induced by Influenza A Virus and Double-Strand RNA and Mediates Airway Epithelial Cell Apoptosis

Influenza A virus (IAV) targets airway epithelial cells and exploits the host cell machinery to replicate, causing respiratory illness in annual epidemics and pandemics of variable severity. The high rate of antigenic drift (viral mutation) and the putative antigenic shift (reassortant strains) have raised the need to find the host cell inducible factors modulating IAV replication and its pathogenesis to develop more effective antiviral treatment. In this study, we found for the first time that transcription factor Runx3, a developmental regulator and tumor suppressor, was induced by IAV H1N1 and H3N2, viral RNA, a synthetic analog of viral double-stranded RNA (dsRNA) polyinosinic-polycytidylic acid, and type-II interferon-γ (IFNγ) in human airway epithelial cells. Whereas Runx3 was essentially not induced by type-I IFNα and type-III IFNλ, we show that Runx3 induction by IAV infection and viral RNA is mediated through the innate immune receptor MDA5 and the IκB kinase-β−NF-κB pathway. Moreover, we provide substantial evidence indicating that Runx3 plays a crucial role in airway epithelial cell apoptosis induced by IAV infection and dsRNA through the activation of extrinsic and intrinsic apoptosis pathways. Thus, we have identified Runx3 as an inducible and important transcription factor modulating IAV-induced host epithelial cell apoptosis.


Runx3 is strongly induced by dsRNA poly(I:C) in airway epithelial cells. It is believed that the viral
replicative intermediate dsRNA is one of the important components of infecting IAV, and the synthetic dsRNA analog poly(I:C) and viral dsRNA isolated from IAV-infected lungs are each able to induce both the local and systemic cytotoxic effects typical of influenza [29][30][31] . We thus assessed the effect of dsRNA on Runx3 expression. As shown in Fig. 2a, the level of Runx3 protein was strongly induced by low (0.2-1 kb) or high (1.5-8 kb) molecular weight dsRNA poly(I:C) in primary SAECs. Runx3 was also slightly upregulated by transforming growth factor-β (TGFβ ), tumor necrosis factor-α (TNFα ) and Pam3CSK4 (TLR1/2 ligand) in the cells. In contrast, other agonists including R837 (TLR7 agonist), CL075 (TLR8/7 agonist), and lipopolysaccharide (LPS, TLR4 agonist) only had minor effects. It has been shown that primary human airway epithelial cells express all TLRs except TLR8 32 . Similarly, Runx3 protein levels were markedly induced by poly(I:C) in BEAS-2B airway epithelial cells that express all TLRs 32 (Fig. 2b). Runx3 was also induced by poly(I:C) in a dose-dependent manner in 16HBE14o-immortalized normal human bronchial epithelial cells 33 (Supplementary Fig. S2). In contrast, neither Runx1 nor Runx2 was induced by poly(I:C) in BEAS-2B cells ( Supplementary Fig. S1b). We further found that the effect of poly(I:C) on Runx3 induction could be observed at a dose as low as 1 ng/ml and reached a plateau of over 500 ng/ml in BEAS-2B cells (Fig. 2c, left panels) and that Runx3 induction was detected as early as 6 h after poly(I:C) treatment (Fig. 2c, right panels). Moreover, poly(I:C) had a similar ability to induce Runx3 expression when added directly to cell culture medium or delivered into the cells using a transfection reagent lipofectamine 2000 (Fig. 2d). This result is consistent with a previous report that poly(I:C) is internalized and functions via intracellular dsRNA receptors 34 . Additionally, an immunofluorescence analysis revealed that the poly(I:C)-induced Runx3 protein was entirely localized to cell nuclei (Fig. 2e), indicating that Runx3 functions to regulate gene expression in the nucleus. Taken together, our findings demonstrate that Runx3 can be strongly induced by dsRNA poly(I:C) in airway epithelial cells.
We next sought to determine whether Runx3 induction by poly(I:C) was mediated via transcriptional regulation. We found that the level of Runx3 mRNA was markedly induced by low and high molecular weight poly(I:C)s in BEAS-2B cells (Fig. 3a, left panels), which was abolished by pretreatment with a transcription inhibitor, actinomycin D (Fig. 3a, right panels). We further found that the mRNA degradation rates of Runx3 were similar in the control and poly(I:C)-treated cells (Fig. 3b). These data indicate that Runx3 is induced by poly(I:C) through the regulation of transcription but not at the level of mRNA stability.

Figure 1. Runx3 is induced by IAV infection in airway epithelial cells.
(a) BEAS-2B cells were treated with control PBS (mock) or infected with active H1N1 PR/8/34 strain at MOI of 1 or inactivated H1N1 viruses (UV light exposure or incubation in 65 o C for 30 min), grown for 24 h, and cell lysates at equal protein amounts were subjected to Western blotting with Runx3 or NP antibodies. (b) BEAS-2B cells were infected without (-) or with ( + ) H1N1 PR/8/34 at various MOIs of 0.05 to 3 for 24 h, and cell lysates at equal protein amounts were subjected to Western blotting with indicated antibodies. (c,d) BEAS-2B cells were infected with H1N1 (MOI = 1) for 0-24 h, and the kinetics of Runx3 mRNA and protein expression were determined by quantitative real-time PCR (c) and Western blot (d), respectively. (e,f) BEAS-2B cells (e) and SAECs (f) were infected without (-) or with ( + ) H3N2 (x:31) A/Aichi/68 or H1N1 PR/8/34 strains at different MOIs for 24 h, and cell lysates at equal protein amounts were subjected to Western blotting with indicated antibodies. (g) BEAS-2B cells were infected with H1N1 at various MOIs of 0.5 to 3 in the presence of control DMSO or cycloheximide (CHX, 5 μ M) for 16 h. Runx3 mRNA level was measured by RT-PCR and normalized to GAPDH, and relative changes (fold) are shown. Data are means ± S.E. (n = 3). *p < 0.05; **p < 0.01 versus the mock-infected cells.
We next assessed the signal pathways for Runx3 induction by poly(I:C) in BEAS-2B cells. We found that Runx3 induction by poly(I:C) was abolished by two different NF-κ B inhibitors piceatannol 38 and BAY11-7082 39 (Fig. 4d). Interestingly, Runx3 expression was attenuated by an IKKβ inhibitor SC514 40 , but not by Amlexanox 41 , an inhibitor of IKKε and TANK-binding kinase 1, or GFX109203X, a general protein kinase C inhibitor. We further found that the poly(I:C)-induced expression of Runx3 was markedly inhibited by knockdown of IKKβ , but not IKKα with two different siRNAs against human IKKβ or IKKα , respectively (Fig. 4e). Moreover, Runx3 induction by poly(I:C) was greatly suppressed by knockdown of NF-κ B p65 (Fig. 4f), but not by silencing of Stat1 42 , an essential transcription factor activated by IFN (Fig. 4g). It has been well documented that TLR3 and RIG1-like receptors trigger the activation of downstream IKKα /β -NF-κ B and TBK1/IKKε -IFN regulatory factor (IRF) pathways 36,[43][44][45] , leading to the production of IFN and pro-inflammatory cytokines. We also found that knockdown of TLR3 and MDA5 inhibited the poly(I:C)-induced phosphorylation (activation) of IKKβ on Ser-177 or NF-κ B p65 on Ser-536 in BEAS-2B cells ( Supplementary Fig. S3). Taken together, these results indicate that MDA5 and TLR3 as well as the IKKβ − NF-κ B pathway are critically involved in Runx3 induction by poly(I:C) in airway epithelial cells (Fig. 4h), although other potential pathways may also be operative.
We next assessed the signal pathways responsible for Runx3 induction by IAV in BEAS-2B cells. We found that similar to Runx3 induction by dsRNA poly(I:C), IAV-induced expression of Runx3 was markedly suppressed by two different NF-κ B inhibitors piceatannol 38 and BAY11-7082 39 and an IKKβ inhibitor SC514 40 (Fig. 5f). Furthermore, the IAV-induced expression of Runx3 was essentially abolished by silencing of NF-κ B p65 and was markedly inhibited by knockdown of the NF-κ B upstream activator IKKβ with two different siRNAs (Fig. 5g). Collectively, our findings indicate that IAV induces Runx3 expression mainly through MDA5 and the IKKβ − NF-κ B pathway in airway epithelial cells (Fig. 5h). Since IAV infection induces the expression and secretion of type-I and type-III IFNs in airway epithelial cells 32 , we assessed the effects of different types of IFNs on Runx3 expression in BEAS-2B cells. Interestingly, we found that Runx3 was induced by type-II IFNγ , but essentially not by type-I IFNα and type-III IFNλ , although they all induced RIG-1 expression at comparable levels (Fig. 5i). This result suggests that autocrine type-I and type-III IFNs may not contribute to Runx3 induction by IAV infection in airway epithelial cells.
Runx3 plays an important role in airway epithelial cell apoptosis induced by IAV and dsRNA poly(I:C). As shown in Fig. 6a, we determined the detached dead cells from culture with trypan blue using a TC20 automated cell counter and found that the cellular death rate (detached dead cells over total detached and adherent cells) was greatly enhanced by infecting BEAS-2B cells with IAV H1N1. We further found that the IAV-induced cell death was abolished by a general caspase inhibitor Z-VAD-FMK 47 , but was essentially unaffected by a necroptosis inhibitor necrostatin-1 48 . These results indicate that, in agreement with previous reports [10][11][12] , the IAV-induced airway epithelial cell death is mediated by apoptosis. DNA fragmentation represents a characteristic of late stage apoptosis that can be detected by TUNEL assay 49 . TUNEL staining (green color) revealed that infection of BEAS-2B cells with H1N1 induced cell apoptosis, and the apoptotic rate was markedly augmented (2.6-fold) by Runx3 overexpression (Fig. 6b and supplementary Fig. S5). Overexpression of Runx3 also increased the basal cellular apoptotic rate. To understand the molecular mechanisms of Runx3 in mediating apoptosis, we determined the effect of Runx3 on the activation (cleavage) of caspase-3 50 , a primary executioner of apoptosis, and the cleavage of DNA repair enzyme poly (ADP-ribose) polymerase (PARP) 51 , one of the main cleavage targets of caspase-3. We found that H1N1 infection induced the cleavage (activation) of caspase-3 and PARP and that the effects were markedly enhanced by Runx3 overexpression in BEAS-2B cells (Fig. 6c). Similarly, overexpression of Runx3 also augmented the cleavage (activation) of caspase-3 and PARP in primary SAECs (Fig. 6d). To confirm if Runx3 mediates cellular apoptosis in response to IAV, we knocked down Runx3 with two specific siRNAs that target different regions in human Runx3 mRNA and determined the effects on apoptosis by measuring PARP cleavage. As shown in Fig. 6e, the H1N1-induced cleavage of PARP was markedly suppressed by Runx3 knockdown in BEAS-2B cells. These novel findings demonstrate that Runx3 plays a crucial role in airway epithelial cell apoptosis induced by IAV infection.
We then asked whether Runx3 promoted IAV-induced apoptosis by enhancing the apoptosis-inducing effects of viral RNA or dsRNA. We first determined the effect of Runx3 overexpression on cell death induced by total RNA isolated from mock-or IAV-infected cells. As shown in Fig. 7a, viral RNA generated during IAV infection, but not the control cellular RNA, significantly induced cell death and the effect was indeed augmented by Runx3 overexpression. We further found that dsRNA poly(I:C) significantly increased cell death and the cellular death rate was also enhanced by Runx3 overexpression, which was abolished by a general caspase inhibitor Z-VAD-FMK 47 (Fig. 7b). In contrast, overexpression of protein kinase D3 (PKD3), which was used as an experimental control, had a minor effect on the cellular death rate (Fig. 7b) and apoptosis by measuring the cleavage of caspase-3 and PARP (Fig. 7d). Overexpression of Runx3 and PKD3 was verified by Western blots in Fig. 7d (3 rd & 4 th panels). Moreover, TUNEL staining revealed that Runx3 overexpression apparently augmented basal and the poly(I:C)-induced cell apoptotic rate (Fig. 7c). The images of TUNEL staining are shown in supplementary Fig. S6. Consistent with IAV-induced apoptosis, we found that the poly(I:C)-induced cleavage (activation) of caspase-3 and PARP was augmented by Runx3 overexpression (Fig. 7d), but was markedly suppressed by Runx3 knockdown with two different siRNAs in BEAS-2B cells (Fig. 7e). These findings demonstrate that Runx3 positively mediates airway epithelial cell apoptosis induced by IAV viral RNA and dsRNA poly(I:C).
It has been shown that both extrinsic (death receptor/caspase-8) and intrinsic (mitochondrial/caspase-9) apoptosis pathways are activated by dsRNA poly(I:C) and are involved in the poly(I:C)-induced apoptosis of endothelial and lung cancer cells 52,53 . In agreement with previous reports, we found that treatment of BEAS-2B cells with poly(I:C) triggered the activation of caspase-8 and caspase-9, resulting in the appearance of cleaved fragments p41/43 or p35/37, respectively (Fig. 8a). We further found that the poly(I:C)-induced activation of caspase-8 and caspase-9 were markedly suppressed by knockdown of Runx3 with two different siRNAs, indicating that Runx3 is required for the activation of caspase-8 and caspase-9 apoptosis pathways. We then utilized specific peptide inhibitors for caspase-8 or caspase-9 and assessed the involvement of these apoptosis pathways in poly(I:C)or H1N1-induced apoptosis in BEAS-2B cells overexpressing Runx3 or control vector. As shown in Fig. 8b,c, the poly(I:C)-and H1N1-induced cellular death was abolished by a general caspase inhibitor Z-VAD-FMK 47 and the specific caspase-8 inhibitor Z-IETD-FMK, but was partially inhibited by the specific caspace-9 inhibitor Z-LEHD-FMK. Similarly, the enhanced cell death rates induced by poly(I:C) and H1N1 in Runx3-overexpressing cells were also abolished by the general caspase inhibitor Z-VAD-FMK and the caspase-8 inhibitor, and partially attenuated by the caspase-9 inhibitor. Collectively, these results indicate that Runx3 is required for the activation of both extrinsic caspase-8 and intrinsic caspsas-9 apoptosis pathways that may promote airway epithelial cell apoptosis induced by dsRNA poly(I:C) and IAV infection. Discussion IAV infects all age groups and poses a global health and economic threat. The emergence of viral resistance to current anti-influenza drugs and putative pandemic strains has raised the need to find the host cell inducible factors modulating IAV replication and pathogenesis for antiviral treatment [7][8][9] . In the present study, we report a novel finding that induction of transcription factor Runx3 is a host airway epithelial cell response to IAV infection, viral RNA, dsRNA poly(I:C), and type-II IFNγ and that Runx3 plays a crucial role in the epithelial cell apoptosis induced by IAV and dsRNA. We further show that Runx3 induction by IAV infection and viral RNA is mediated through the innate immune receptor MDA5 and IKKβ − NF-κ B pathway. Our findings identify Runx3 as an inducible and important transcription factor that can modulate the IAV-induced host airway epithelial cell apoptosis.
Runx3 is involved in neurogenesis, thymopoiesis, gastrointestinal and lung development [19][20][21][22][23] . Recent studies also indicate that Runx3 can function as a tumor suppressor for a variety of cancers of gastric, breast, pancreatic, liver, lung and colon origins 24 . However, little is known about the regulation of Runx3 expression. The first finding from this study is the induction of Runx3 by IAV infection, viral RNA, dsRNA poly(I:C) and IFNγ in airway epithelial cells. This is the first report that Runx3 can be upregulated in epithelial cells as far as we know. The induction of Runx3 by IAV required active virus replication in host cells; and Runx3 was induced in a dose-dependent manner by H1N1 PR/8/34 and H3N2 (x:31) A/Aichi/68 strains. However, unlike the H1N1 strain, the H3N2 (x:31) A/ Aichi/68 strain at high doses inhibited Runx3 expression. This may be because the nonstructural protein 1 of H1N1 PR8/34 strain lacks a functional binding domain to polyadenylation stimulating factor 30 to inhibit host gene expression 54 . We also found that Runx3 mRNA level was significantly increased by H1N1 infection in the presence of a protein synthesis inhibitor cycloheximide, and that viral RNA generated during IAV infection, but not cellular RNA, contributed to Runx3 induction. Furthermore, we have demonstrated that Runx3 is strongly induced by dsRNA poly(I:C) through the regulation of transcription but not at the level of mRNA stability. Hence, our findings indicate that Runx3 is induced by IAV infection likely through viral replicative RNA or dsRNA in airway epithelial cells. It has been shown IAV infection results in accumulation of viral replicative intermediate dsRNAs, which can be recognized by innate immune receptors, such as membrane-bound TLR3 and cytosolic RIG1-like receptors including RIG-1, MDA5, and LGP2 [35][36][37] . TLR3 localizes to both the cell membrane surface and intracellular endosomal compartments of epithelial cells 32,55 and mediates inflammatory cytokine production by IAV 56 . RIG-1 preferentially binds to short (< 300 bp) dsRNAs that have blunt ends and a 5′ -triphosphate moiety 37 and plays an essential role in type-I IFN production and antiviral activity 46 . On the other hand, MDA5 preferentially binds to long dsRNA (> 1000 bp) with no end specificity and thus can bind poly(I:C) and long-scale viral replicative intermediate dsRNA 37 . Interestingly, we found that Runx3 induction by IAV infection and viral RNA was primarily mediated by MDA5. Whereas Runx3 induction by direct addition of the synthetic dsRNA analog poly(I:C) was mediated by MDA5 and TLR3. A previous study shows that poly(I:C) is internalized and functions via intracellular dsRNA receptors 34 . Consistent with this report, we found that poly(I:C) had a similar ability to induce Runx3 expression when added directly to cell culture medium or delivered into the cells. These results suggest that the internalized poly(I:C) may activate cytosolic MDA5 and endosomal TLR3 to induce Runx3 expression, while long-scale dsRNA derived from virus replication may only activate MDA5 for Runx3 induction. They also suggest that the entry of long scale viral dsRNA into the endosome through autophagic processes may be limited. Moreover, we have provided substantial evidence indicating that IKKβ − NF-κ B pathway is critically involved in Runx3 induction by IAV infection and dsRNA poly(I:C) in airway epithelial cells. Additionally, we found that Runx3 was induced by type-II IFNγ but essentially not by type-I IFNα and type-III IFNλ , and that silencing of Stat1 42 that is an essential transcription factor downstream of IFN did not affect Runx3 expression by poly(I:C). These findings indicate that Runx3 induction by IAV infection or dsRNA in airway epithelial cells is independent of autocrine type-I and type-III IFNs. Sequence analysis reveals that Runx3 promoter area (+ 1 to −1424) contains three sites of NF-κ B core binding sequence (GGGRNWYYCC) and one gamma activation sequence (TTC(N 2-4 )GAA).
Another important finding from this study is that Runx3 plays a crucial role in airway epithelial cell apoptosis induced by IAV infection and dsRNA. It has been shown that IAV infection induces cell apoptosis in vitro and in animal tissues and that the apoptotic cells along with the replicative virus within the cell are phagocytosed, digested and killed by macrophages and neutrophils, hence cellular apoptosis is regarded as a host defense mechanism to effectively clear the virus-infected cells to prevent spread of the virus [10][11][12]57 . Under this scenario, Runx3 may function as a host antiviral factor as Runx3 promotes airway epithelial cell apoptosis in response to IAV infection and dsRNA. However apoptosis is a double-edged sword for IAV infection. Too much or uncontrolled apoptosis could cause pulmonary tissue damage and lung dysfunction, which contributes to the disease morbidity and mortality, so upregulation of Runx3 may enhance the IAV-induced pathogenesis and the disease severity by promoting host cell apoptosis and tissue injury. The role of Runx3 in lung epithelial cell response and disease progression to IAV infection in vivo merits further investigation.
In summary, we have identified Runx3 as an inducible and crucial transcription factor that can modulate airway epithelial cell apoptosis in response to IAV infection, viral RNA and dsRNA poly(I:C). Our study also opens new fields of investigation of the upregulation of tumor suppressor Runx3 in carcinoma cells or in other forms of acute lung injury and repair, in which lung epithelial cell apoptosis is generally a prominent feature.
Cell culture and treatment. SAECs were obtained from American Type Culture Collection (ATCC) and Lonza, cultured in airway epithelial cell growth medium (ATCC) and used for experiments within 3 passages. BEAS-2B cells were obtained from ATCC and cultured in airway epithelial cell growth medium. 16HBE14o-human bronchial epithelial cells 33 were kindly provided by Dr. Dieter Gruenert (University of California at San Francisco) and cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum. For agonist treatment, epithelial cells were washed twice with Hanks' balanced salt solution and then incubated in bronchial or small airway epithelial basal medium (BEBM or SABM, Lonza) or DMEM serum-free medium with indicated agonists.
Virus infection. BEAS-2B or SAECs grown until subconfluence or pretreated were washed twice with RPMI-1640, infected with IAV H1N1 PR/8/34 or H3N2 (x:31) A/Aichi/68 strains (Charles River) in BEBM at an appropriate MOI for 1 h at 37 o C, then incubated in airway epithelial cell growth medium without aspirating the viruses for 1 h to facilitate IAV entry into cells. The cells were washed once with RPMI-1640 and incubated in BEBM or SABM for 24 h. For adenoviral vector-mediated expression of Runx3, BEAS-2B cells were infected with recombinant adenovirus expressing control vector, human Runx3 or PKD3 (SignaGen Laboratories) at a MOI of 5.