Conceived and designed the experiments: VP MR AS ALM. Performed the experiments: VP ETG MA SD ALM. Analyzed the data: VP ETG MR SD ALM. Contributed reagents/materials/analysis tools: MR AG SD JR AS ALM. Wrote the paper: VP ALM.
The authors have declared that no competing interests exist.
Several studies have implicated viral infection as an important factor in the pathogenesis of IPF and related fibrotic lung disorders. Viruses are thought to cause epithelial cell injury and promote epithelial-mesenchymal transition (EMT), a process whereby differentiated epithelial cells undergo transition to a mesenchymal phenotype, and considered a source of fibroblasts in the setting of lung injury. We have demonstrated an association between the epithelial injury caused by chronic herpes virus infection with the murine γ-herpes virus, MHV68, and lung fibrosis. We hypothesize that EMT in this model of virus-induced pulmonary fibrosis is driven by the expression of the transcription factor Twist.
We conclude that Twist contributes to EMT in the model of virus-induced pulmonary fibrosis. We speculate that in some IPF cases, γ-herpes virus infection with EBV might be a source of injury precipitating EMT through the expression of Twist.
Current understanding of the pathogenesis of Idiopathic Pulmonary Fibrosis (IPF) considers ongoing injury of the lung epithelium as one of the most important drivers of the disease. The cause of this injury remains unknown although chronic herpesvirus infection has been suggested by several studies that show the presence of one or more herpes viruses in the lungs of IPF patients
Recently, we turned our attention to the factors that control fibroblast proliferation in this model of virus-induced lung fibrosis. Fibroblasts from fibrotic lungs can originate locally by proliferation of resident interstitial fibroblasts. They can also originate from circulating fibrocytes derived from bone marrow progenitor cells or from the transdifferentiation of epithelial cells
We set out to investigate the role of EMT in virus-induced lung fibrosis and the factors that influence this process. A crucial role has been identified for TGF-β in the induction of EMT
The mechanisms responsible for fibroblast generation in the setting of virus-induced lung fibrosis are unknown. We believe that EMT is a key event in this setting and hypothesize a role for Twist in this process. In this study, we show evidence that Twist might indeed play a role in EMT and that γ-herpes viruses contribute to EMT by upregulation of Twist.
EBV infection of nasopharyngeal epithelial cells results in EMT and malignant transformation in humans. This process is thought to be regulated by the transcription factor Twist
MLE-15 cells were mock-infected or infected at multiplicity of infection of 0.01 with MHV68 and analyzed 48 h post-infection for Twist expression. (A) Immunostaining of infected cells using an anti-Twist antibody. Twist positive staining (red) was observed in nuclei of MHV68 infected cells but not in mock infected cells. Cells were counterstained with DAPI (blue). Magnification ×1000. Staining is representative of 2 independent experiments. (B) Western blot analysis for Twist in whole cell lysates. The blot was stripped and reprobed with anti-β-actin antibody as loading control. (C) Western blot analysis for Twist, epithelial (Occludin) and mesenchymal (FSP-1, Vimentin) markers in whole cell lysates from uninfected and MHV68 infected MLE-15 cells. Notice the increase in the expression of mesenchymal markers and the decrease in Occludin expression after infection.
To determine whether Twist expression was able to promote EMT, we made stable transfection of MLE-15 cells using Twist cDNA and examined expression of mesenchymal and epithelial cell markers. Control cells maintained a cobblestone-like appearance with polygonal shape of cells and expressed the epithelial marker, E-cadherin, in the membrane compartment as determined by immunostaining. Twist-transfected cells lost the cobblestone appearance, became more scattered, and expressed low levels of E-cadherin in the cytoplasm (
(A) Stable transfection of MLE-15 cells was carried out with Twist1 cDNA. Immunofluorescent staining was performed using anti-Twist (green), anti-E-cadherin (red), and anti-FSP-1 (green) antibodies. Twist positive cells showed loss of E-cadherin staining in the cell membrane. Cells were counterstained with DAPI (blue). Magnification is ×400. Staining is representative of 3 independent experiments. (B) Western blot analysis for Twist, epithelial (E-cadherin) and mesenchymal (Collagen IV and Fibronectin) markers in whole cell lysates from nontransfected and Twist-transfected MLE-15 cells. The blot was stripped and reprobed with anti-β-actin and Twist antibodies as controls. (C) Immunoblot for Twist, E-cadherin (epithelial marker) and Collagen I (mesenchymal marker) from MLE-15 cells transduced with lentiviral constructs expressing control shRNA or Twist1 shRNA as indicated followed by MHV68 infection.
To determine whether Twist plays a causal role in the changing epithelial phenotype after infection, we tested whether inhibition of Twist expression would affect MHV68 ability to induce EMT. MLE15 cells were transduced with lentiviral particles expressing Twist shRNA and infected with MHV68 24 h later. Forty-eight hours after MHV68 infection, whole cell lysates were prepared and analyzed by Western blot using anti-Collagen I and E-cadherin antibodies. We found that inhibition of Twist expression was accompanied by persistent E-cadherin expression and lower expression of Collagen compared to cells with knockdown expression of Twist (
Having established that MHV68 infection promotes Twist expression
(A) Expression of Twist was analyzed by Western blot in lung homogenates from IFNγR−/− mice after mock- and MHV68-infection at days 4, 9 and 120 post-infection. Time-dependent increased levels of Twist were found in samples from virus-infected animals. Blots were stripped and reprobed with an anti-β-actin antibody as loading control. (B–D) Detection of Twist expression by IHC staining in lungs of IFNγR−/− mice mock infected (B), and virus-infected at 15 dpi (C), and at 120 dpi (D). Notice the abundant nuclear and cytoplasmic positive staining of lung epithelial cells in the infected mice. Alveolar macrophages were weakly positive for Twist staining (block arrow). (E) Twist detection by IHC staining in lung of C57BL/6 mice infected with MHV68 at day 15 post-infection. Magnification ×400. Staining is representative of 5 different mice at each time point.
Previously, we showed that acute MHV68 infection in IFNγR−/− mice was followed by virus reactivation and the development of progressive and multifocal fibrosis
(A) Immunofluorescent staining of frozen lung sections of MHV68 infected IFNγR−/− mice at day 7. Virus antigen detection was performed using an anti-MHV68 polyclonal antibody (green). Type II cells were detected using anti-pro-SP-C antibody (red). Sections were counterstained with DAPI (blue). Merged image shows yellow cells indicating type II cells that support lytic infection of the virus (arrows). Magnification ×800. Staining is representative of 5 different animals. (B) Immunostaining of frozen lung sections of IFNγR−/− mice at day 7 using a monoclonal antibody against the virus ORF59 (red) and the anti-FSP-1 antibody (green). Merged image shows nuclear localization of viral antigen in FSP-1 positive cells (arrows). Magnification ×400. Staining is representative of 3 different animals. (C) Dual IHC on formalin-fixed paraffin-embedded lung sections of MHV68 infected IFNγR−/− mice at day 120. Epithelial cells expressing mesenchymal cell markers (arrow) were detected using anti-TTF-1 (brown nuclei) and anti-FSP-1 (red cytoplasm) antibodies. Magnification ×400. Box shows area depicted with magnification ×1000. Staining is representative of 5 different animals. (D) Dual immunofluorescence staining in frozen lung sections of MHV68 infected IFNγR−/− mice at day 120 using anti-pro-SP-B (red) and anti-N-cadherin (green) antibodies. Yellow cells indicate type II cells expressing N-cadherin mesenchymal cell marker (arrows). Nuclei were visualized by DAPI staining (blue). Magnification ×200. Staining is representative of 5 different animals.
We then turned our attention to the late time points coinciding with the period of fibrosis. Lung tissues from mock- and MHV68-infected IFNγR−/− mice were harvested after 120 days of infection. Double IHC staining of lung tissue was performed to detect epithelial and mesenchymal markers. Lung tissue of mock infected mice expressed only the epithelial marker, thyroid transcription factor-1 (TTF-1), in the nuclei of alveolar epithelial cells. In contrast, lung specimens from MHV68-infected mice demonstrated nuclear staining for TTF-1 as well as diffuse cytoplasmic staining for FSP-1 in alveolar epithelial cells (
After corroborating the occurrence of EMT in animals with virus-induced lung fibrosis, we then determined the relevance of these findings in humans with IPF. Studies with IPF and control tissue samples were approved by the Emory University Institution Review Board. We collected a total of 13 tissue samples from IPF patients who were undergoing lung transplantation; 3 lung control samples (all donor lung specimens at the time of transplantation) and 1 from a COPD patient. All IPF lung specimens had histopathological diagnosis of usual interstitial pneumonia (UIP). We first analyzed for Twist protein expression and found that Twist immunoreactivity was abundant and localized mostly in the nuclei and cytoplasm of epithelial cells overlying fibrotic areas or honeycombing cysts (
(A) Determination of Twist expression by IHC analysis in sections of control/donor, COPD, and IPF lungs. No Twist expression was detected in control lung. Epithelial cells covering fibrotic areas of IPF lung showed abundant nuclear and cytoplasmic Twist positive staining (arrows). COPD lungs showed strong Twist staining in alveolar macrophages but staining was negative in epithelial cells. Upper row pictures have magnification ×100; lower row – magnification ×400. Stainings are representative of 3 independent assays using at least 3 donors, 1 COPD and 13 IPF lung tissues. All 13 IPF samples demonstrated Twist expression. (B) Dual immunofluorescent staining in frozen lung sections of IPF lung using anti-pro-SP-C (green) and anti-Twist (red) antibodies. Numerous type II cells were found expressing Twist in the nuclei (white arrow) and cytoplasm (yellow arrow). Nuclei were visualized by DAPI staining (blue). Magnification ×200. (C) Confocal microscopic image analyses of dual immunofluorescent staining for Twist (red) and pro-SP-C protein (green). Notice nuclear localization of Twist and cytoplasmic pro-SP-C in the same lung epithelial cells. Magnification ×1000. Stainings are representative of 3 independent assays in lung tissues from 3 IPF patients.
To determine whether molecular changes associated with EMT could be detected in lung epithelial cells that expressed Twist, we probed for both epithelial and mesenchymal markers. Analysis of fluorescence and confocal microscope images revealed that Twist positive epithelial cells had abnormal distribution of the epithelial marker, E-cadherin, in the cytoplasm. In contrast, epithelial cells without Twist expression retained immunoreactivity for E-cadherin in the normal site that is within the basolateral domains of cell membranes (
(A) Dual immunofluorescence and (B) confocal microscopic image analyses of frozen lung sections of IPF lung stained with anti-E-cadherin (green) and anti-Twist (red) antibodies. E-cadherin staining was localized in the basolateral membrane in Twist negative cells (closed arrow). Twist positive cells showed weaker E-cadherin staining or cytoplasmic localization (open arrow). Magnification ×400 and ×1000 respectively. (C) Double immunofluorescence and (D) confocal microscropic image analyses of frozen lung sections of IPF lung stained with anti-N-cadherin (green) and anti-Twist (red) antibodies. Notice nuclear Twist positive cells showing N-cadherin (green) positive staining in the cytoplasm (open arrows). Some cells showed Twist positive staining but were negative for N-cadherin (close arrows). Magnification ×200 and ×1000 respectively. Stainings are representative of 3 independent assays in lung tissues from 3 IPF patients.
To determine the frequency of EMT in IPF lungs, dual immunofluorescence microscopy studies were performed on IPF lung tissue sections using epithelial and mesenchymal markers. As shown in
(A) Immunofluorescent staining of lung frozen sections of IPF lung using anti-N-cadherin (green) and anti-pro-SP-B (red) antibodies. Type II cells were detected by Pro-SP-B staining (red, open arrow). Yellow cells indicate type II cells expressing the mesenchymal cell marker N-cadherin (closed arrows). Scattered green cells were observed intercalating red and yellow type II epithelial cells (asterisks). Magnification ×100. Box shows area depicted at magnification ×1000 in (B). Nuclei were visualized by DAPI staining (blue). Stainings are representative of 3 independent assays in lung tissues from 3 IPF patients.
We and others have demonstrated that more than half of IPF patients show evidence of chronic EBV infection in their lungs
The level of Twist protein in 6 IPF lung samples (5 EBV positive, 1 EBV negative), and 3 EBV negative donor specimen was assessed using Western blot analysis. Twist expression was detected at different levels in all 5 EBV positive IPF samples. In contrast, Twist expression was barely detectable in the EBV negative IPF lung and weakly positive in control (donor) lung samples (
Western blot analysis for the expression of Twist was performed in whole cell lysates of control donors and IPF lungs. The blot was stripped and reprobed with anti-β-actin antibody as a loading control. Same lung samples were tested for the presence of EBV genome [(+) ≥103 EBV DNA copies per 106 cells] by real time PCR.
Although EMT has been implicated in lung fibrosis, the mechanisms underlying EMT remain unclear. In this report, using an
Microarray analysis of IPF tissue identified an IPF-specific gene expression signature characterized by the up-regulation of genes involved in tissue remodeling and enriched with genes associated with lung development
Twist regulates EMT, cell movement, and tissue reorganization during early embryogenesis. Similarly, Twist expression in tumor cells is associated with EMT and increased cell motility, suggesting that Twist may contribute to metastasis. Suppression of Twist expression in highly aggressive 4T1 mammary carcinoma cells specifically inhibits their ability to metastasize from the mammary gland to the lung
Recently, several studies demonstrated EMT induction in the setting of viral infection. Li et al. showed positive expression of hepatitis C virus core protein associated with decreased expression of E-cadherin and α-catenin in conjunction with increased expression of N-cadherin, vimentin, and fibronectin in tissues from cholangiocarcinoma
Strong evidence for EMT
Our data suggest that Twist might contribute to the EMT process in pulmonary fibrosis. Furthermore, they suggest that γ-herpes virus might be, in some IPF cases, the source of injury precipitating EMT. We demonstrated that biopsies with relatively high Twist expression were positive for EBV genome. Several other studies suggest a potential role for EBV in viral epithelial injury and potential EMT development. In a viral cell line model, TGFβ1 was shown to be induced in epithelial cells following EBV lytic phase induction. TGFβ1 expression was promoted by the EBV early genes Rta and Zta
Although our studies suggest that virus-induced Twist activation might represent an important mechanism for EMT in IPF, further studies will be needed to test the true role of Twist in EMT
In summary, our studies, suggest that EMT may be a cellular mechanism of fibrogenesis in the lung associated with virus-induced epithelial injury. Twist is a well known master and a transcriptional regulator of EMT during embryogenesis and metastasis. The abundant expression of Twist in alveolar epithelial cells is likely to contribute to EMT and an important source of fibroblasts in IPF lungs. The identification of the downstream effector pathways that are activated during EMT holds the promise of revealing new diagnostic markers of early stages of pulmonary fibrosis and, quite possibly, novel targets for anti-fibrotic therapeutics.
Antibodies to detect MHV68, polyclonal anti-MHV68 and chicken monoclonal anti-MHV68 ORF59, were a generous gift from Dr. Samuel Speck (Department of Microbiology/Immunology, Emory University)
MLE-15 cells were obtained from Dr. Brigham Willis (University of Texas) and were maintained in HITES medium (RPMI 1640 (GIBCO, Gaithersburg, MD) supplemented with 2% fetal bovine serum (Thermo Scientific HyClone), 100 U/ml penicillin, 100 µg/ml streptomycin, 1% insulin-transferrin-sodium selenite, 5 µg/ml transferrin, 10 nM hydrocortisone, 10 nM β-estradiol, 2 mM glutamine, and 2 mM HEPES).
The animal experiments were reviewed and approved by the Emory University Institutional Animal Care and Use Committee before the experiments were conducted. Animals were infected as described before
Lung tissues were obtained from 13 IPF patients with histological evidence of UIP, and from 4 control subjects (organ donors, COPD). Samples were immediately frozen for double immunofluorescence staining or placed in 4% (w/v) paraformaldehyde for IHC. The study protocol was approved by the Emory University IRB (protocol 2005–361). Informed consent was obtained in written form from each subject.
MLE15 epithelial cells (7×104 cells/well in 4-well chamber slides), were infected with MHV68 with a multiplicity of infection (MOI) of 0.01, fixed 48 hours post-infection with 4% paraformaldehyde (Sigma-Aldrich), and permeabilized in 0.02% Triton X-100 (Sigma-Aldrich). Similarly, cells were infected in 6 well plates and whole cell extracts were prepared.
MLE15 cells were transfected by electroporation with Twist cDNA (plko.1 vector, Origene) and pcNeo DNA in a 5∶1 ratio. After electroporation, cells were washed and cultured with HITES medium. Positive selection was done using G418 (400 µg/ml) after the second day of transfection.
MLE-15 cells were transduced with lentivirus expressing mouse Twist1 shRNA and non-target control shRNA (MISSION Sigma-Aldrich) at an MOI of 2. After 24 h, cells were infected with MHV68 at an MOI of 0.05. The cells were washed with PBS and harvested for protein isolation at 48 hours post-MHV68 infection.
Real time PCR reactions were Taqman-based (Qiagen, Valencia, CA). DNA extractions were carried out using the Qiagen DNeasy Blood & Tissue Kit (Valencia, CA). EBV PCR targeted the BamH1W viral genome segment and was performed as described before using frozen human lung tissue
Human and murine lungs were placed in 4% paraformaldehyde and processed for paraffin embedding. Sections (5 µm) were cut, mounted on the slides, subjected to the antigen retrieval in a decloaking chamber (BioCare medical). Endogenous peroxidase activity was quenched with 3% peroxide for 5 min. Immune complexes were visualized with biotinylated secondary antibody and 3,3′-diaminobenzidine tetrahydrochloride using streptavidin-biotin complex method. Double immunohistochemistry was carried out using EnVision G//2 doublestain system, Rabbit/Mouse (DAB+/Permanent Red) kit (DAKO Cytomation, Denmark).
Frozen human and murine lungs were cut (5 µm) and fixed in acetone/methanol (1∶1). After washing in several changes of PBS, tissue sections on glass slides and MLE15 cells on chamber slides were incubated with primary antibodies in antibody diluent with background reducing component (Dako Cytomation, Denmark) overnight at 4°C. The indirect immunofluorescence assay was performed by incubation with secondary antibodies conjugated to Alexa Fluor 594 and/or Oregon Green 488 (Invitrogen). Nuclei were visualized by 4,6-diamidino-2-phenylindole staining (DAPI, Sigma-Aldrich, Saint Louis, MO). Slides were analyzed under fluorescent microscope Olympus BX4 equipped with Olympus SN 1H045294-H camera and Zeiss Axioplan 2 imaging LSM 510 META confocal microscope. Semi-quantitative evaluation of single positive cells (pro-SP-C) and double positive cells (pro-SP-C+N-cadherin) was perfomed using the Axiovision morphometric software. Five representative pictures at 40× magnification were taken from the left lung of mice at day 120 post-infection (n = 5) and cells were counted. Pictures were taken from areas without big airways or blood vessels.
MLE15 cells and human and murine lung tissue specimens were homogenized in extraction buffer (50 mM HEPES, 250 mM NaCl, 5 mM EDTA, 0.1% NP-40, 1 mM PMSF,1 mM DTT supplemented with Protease inhibitor cocktail; BD Biosciences, San Diego, CA). Whole proteins were extracted by centrifugation (12,000x g) for 10 min at 4°C. Samples containing 40 µg of protein were separated by electrophoresis on SDS 10–20% gradient polyacrylamide gels. Proteins were transferred to PVDF membrane (Millipore, Bedford, MA). Membrane was blocked in 5% skim dry milk and probed overnight at 4°C in TBS-Tween buffer (0.02 M Tris [pH 7.6], 0.1 M sodium chloride, 0.05% Tween 20) with primary antibodies. Membranes were washed in TBS-Tween buffer and incubated with secondary antibodies for 1 h at RT. After washing membranes in TBS-Tween buffer again, they were visualized with SuperSignal West Pico Chemiluminescent Substrate (Pierce Biotechnology, Rockford, IL).
We thank to Dr. Seth Force, Emory University Thoracic Surgery, for his support with human tissue collection, and to Jianguo Xu and Melinda Dorson for excellent technical assistance.