key: cord-103511-31njndob authors: Broggi, Achille; Ghosh, Sreya; Sposito, Benedetta; Spreafico, Roberto; Balzarini, Fabio; Lo Cascio, Antonino; Clementi, Nicola; De Santis, Maria; Mancini, Nicasio; Granucci, Francesca; Zanoni, Ivan title: Type III interferons disrupt the lung epithelial barrier upon viral recognition date: 2020-05-05 journal: bioRxiv DOI: 10.1101/2020.05.05.077867 sha: doc_id: 103511 cord_uid: 31njndob Lower respiratory tract infections are a leading cause of mortality driven by infectious agents. RNA viruses such as influenza virus, respiratory syncytial virus and the new pandemic coronavirus SARS-CoV-2 can be highly pathogenic. Clinical and experimental evidence indicate that most severe and lethal cases do not depend on the viral burden and are, instead, characterized by an aberrant immune response. In this work we assessed how the innate immune response contributes to the pathogenesis of RNA virus infections. We demonstrate that type III interferons produced by dendritic cells in the lung in response to viral recognition cause barrier damage and compromise the host tissue tolerance. In particular, type III interferons inhibit tissue repair and lung epithelial cell proliferation, causing susceptibility to lethal bacterial superinfections. Overall, our data give a strong mandate to rethink the pathophysiological roles of this group of interferons and their possible use in the clinical practice against endemic as well as emerging viral infections. 8 efficiently than mice that bear WT stromal cells, and phenocopy Ifnlr1 -/à Ifnlr1 -/chimeras (Fig. 171 3G) . These data further support the direct activity of IFN-λ on epithelial cells. 172 Interestingly, the gene that was most downregulated in Ifnlr1 -/epithelial cells compared to 173 WT cells is the E3 ubiquitin-protein ligase makorin-1 (Mrkn1) (Fig. 3A) , 174 which controls p53 and p21 stability by favoring their proteasomal degradation (48) . Under 175 oxidative stress condition and DNA damage, a hallmark of severe viral infections (49), p53 is 176 stabilized by phosphorylation and p21 degradation, mediated by Mkrn1, favors apoptosis over 177 DNA repair (48). Indeed, Ifnlr1 -/epithelial cells, that express lower levels of Mkrn1, have elevated 178 levels of p21 as measured by flow cytometry (Fig. 3H , I). These data indicate that the capacity of 179 IFN-λ to reduce tissue tolerance stems from its capacity to inhibit tissue repair by directly 180 influencing epithelial cell proliferation and viability. Also, that p21 degradation via Mrkn1 181 upregulation is potently influenced by IFN-λ signaling. 182 RNA viruses can use several strategies to modulate the immune response to their 183 advantage(33, 50), therefore it is crucial to understand the molecular pathways involved in the 184 maintenance of sustained IFN-λ production. Moreover, the difference between mRNA expression 185 and protein levels of interferons suggest that a low abundance cell type with high secretory 186 capacity may be responsible for long term IFN-λ production. We thus investigated the cellular 187 source and molecular pathways that drive IFN-λ production in our model. Early after initial 188 influenza virus infection, IFN-λ is expressed by infected epithelial cells, however, at later time 189 points, DCs from the parenchyma of the lung start to express high levels of the IFN-λ transcript(5). 190 We thus hypothesized that lung DCs are the main producers of IFN-λ and are responsible for the 191 secretion of IFN-λ during viral infections. Accordingly, sorted lung resident dendritic cells express 192 high levels of IFN-λ transcript after 5 days of poly (I:C) treatment, in contrast to epithelial cells, 193 alveolar macrophages and monocytes (Fig. 4A) , which, instead, express inflammatory cytokines (Fig. S8A, B) . Moreover, diphtheria toxin (DT)-mediated depletion of 195 CD11c + cells in CD11c-DT receptor (DTR) mice was sufficient to completely abolish IFN-λ 9 transcript and protein upregulation upon 6 days of poly (I:C) treatment (Fig. 4B, C) , while production remained unaltered (Fig. S8C with the response measured in vivo, TLR7 stimulation did not induce IFN production while it 207 induced upregulation of pro-inflammatory cytokines, and intracellular delivery of poly (I:C) induced 208 high levels of IFN-I but not IFN-λ (Fig.4D, Fig. S9A , B). In agreement with the key role of TLR3, 209 IFN-λ production upon extracellular poly (I:C) encounter was abolished by genetic deletion of the 210 signaling adaptor TRIF (encoded by the gene Ticam1) but not by deletion of the RIG-I/MDA5 211 adaptor MAVS (Mavs) (Fig. 4D) . Conversely, IFN-I production in response to intracellular delivery 212 of poly (I:C) was largely dependent on the signaling adaptor MAVS (Fig. S9A ). Consistent with 213 our previous data, when the RIG-I/MAVS pathway was activated by transfection of the influenza 214 A virus derived pathogen-associated molecular pattern (PAMP) 3-phosphate-hairpin-RNA (3p-215 hpRNA), IFN-I but not IFN-λ, was efficiently induced in a MAVS-dependent manner ( Fig. S10A -216 E, poly (I:C) was used as a control). Finally, inhibition of endosomal acidification by treatment with 217 the pharmacological agent chloroquine abolished IFN-λ induction in response to extracellular poly 218 (I:C), while it preserved IFN-I production upon intracellular poly (I:C) delivery (Fig. S11A, B) . 219 These evidences clearly indicate that TLR3 stimulation potently induces IFN-λ production by DCs 220 in vitro. We, thus, explored the importance of the TLR3-TRIF pathway in vivo under our 221 experimental conditions. Dendritic cells sorted from Ticam1 -/mice treated with poly (I:C) for six 222 days did not express appreciable levels of IFN-λ transcripts while still produced type I interferons 223 ( Fig. 4E, F) . Moreover, poly (I:C) treated Ticam1 -/mice were protected from S. aureus 224 superinfections (Fig. 4G) , and the decrease in bacterial burden correlated with lower IFN-λ 225 transcript levels in the lung, although IFN-I levels remained similar to those of WT mice (Fig. 4H , 226 I). Confirming the crucial role of TLR3 signaling in DCs for IFN-λ production, chimeric mice in 227 which Ticam1 -/bone marrow (BM) cells are transferred in a WT irradiated host (Ticam1 -/ àWT) 228 phenocopied Ticam1 -/animals ( Fig. 4J-L) . 229 The immune system evolved to prevent and resist to pathogen invasion but doing so often 230 threatens host fitness and causes disease in the form of immunopathology (51). RNA viruses are 231 the major cause of most severe lower respiratory tract viral infections (52, 53). While most virus 232 infections manifest as self-limiting upper respiratory tract infections, influenza viruses, SARS-233 CoV, SARS-CoV-2 and MERS-CoV can progress to severe lung disease with potentially lethal 234 outcomes (50, 54, 55) . Although different viruses vary in their virulence and pathogenic potential, 235 the most severe cases of lung RNA viral infections share similar features that suggest an immune 236 pathological etiology. In COVID-19, SARS, MERS and flu, severe symptoms and death occur late 237 after the initial symptoms onset, and after the peak in viral load (56-61) further indicating a central 238 role for an immune etiology of the most severe forms. 239 While IFN-λ is uniquely equipped to induce a gentler immune response that favors viral 240 clearance in the lungs without inducing overt immune activation (1, 3, 62), its impact on epithelial 241 cell biology and its effect on the maintenance of tissue integrity and tolerance to pathogen invasion 242 is incompletely understood. In a system that allowed us to isolate the effect of immune activation 243 from resistance to viral infection, we demonstrate that sustained IFN-λ production in the lung in 244 response to viral PAMPs compromises epithelial barrier function, induces lung pathology and 245 morbidity and predisposes to lethal secondary infections by impairing the capacity of the lungs to 246 tolerate bacterial invasion. Loss of lung barrier tolerance is sufficient to induce lethality upon 247 bacterial challenge independently of bacterial growth (39), and alteration of the repair response 248 11 in the lung can favor bacterial invasion independently from immune cell control (63). In our model 249 immune cell recruitment is not affected by IFN-λ and neutrophils are dispensable for the impaired 250 control of bacterial infections, while IFN-λ signaling on epithelial cells is necessary and sufficient 251 to cause heightened bacterial invasion. 252 Under our experimental conditions, TLR3-TRIF signaling in conventional lung DCs is 253 responsible for the induction of IFN-λ. This is consistent with reports indicating that Tlr3-deficient 254 mice are protected from influenza-induced immune pathology(64). Moreover, TLR3 detects 255 replication intermediates from necrotic cells (35) and is, thus, insensitive to viral immune evasion. 256 This is of particular interest during highly pathogenic human coronavirus infections, whose 257 success in establishing the initial infection is partly due to their ability to dampen TLR7 and MAVS 258 dependent early IFN responses (50) (2020) (available at https://clinicaltrials.gov/ct2/show/NCT04331899). 328 Clin. Microbiol. Rev. 19, 571-582 (2006) . Intratracheal instillations (i.t.) were performed as previously described in (69) Rectal temperature and body weights were monitored daily. Mice were deemed to have reached endpoint at 75% of starting weight or after reaching body temperature of 25°C or lower. To generate mice with hematopoietic-specific deletion of Ifnlr1 or Ticam1, 6-week-old CD45.1+ mice were exposed to lethal whole-body irradiation (950 rads per mouse) and were reconstituted with 5 × 10 6 donor bone marrow cells from 6-week-old wild-type, Ifnlr1 -/or Ticam1 -/mice. Mice were treated with sulfatrim in the drinking water and kept in autoclaved cages for 2 weeks after reconstitution. After 2 weeks, mice were placed in cages with mixed bedding from wild-type, and Ifnlr1 -/or Ticam1 -/mice to replenish the microbiome and were allowed to reconstitute for 2 more weeks. A similar procedure was used to generate bone-marrow chimeras with stromal cellsspecific deletion in Ifnlr1. Here, recipient WT or Ifnlr1 -/mice underwent irradiation and were reconstituted with BM cells derived from CD45.1+ mice similarly as described above. To evaluate the percentage of chimerism, a sample of peripheral blood was taken from chimeric mice after 4 weeks of reconstitution and stained for CD45.1 and CD45.2 (antibodies as identified under 'Reagents and antibodies') and were analyzed by flow cytometry. In order to deplete CD11c + cells, CD11c-DTR mice received 16μg/kg diphtheria toxin (DTx) intravenously starting one day before TLR ligand or saline administration and continuing every other day until day 6 post-treatment to maintain depletion. In vivo depletion of neutrophils was carried out by injecting anti-Ly6G antibody (100μg/mouse) intraperitoneally, starting one day before treatments and then continuing every other day through the duration of the treatment. As controls for no depletion, mice were injected with rat IgG isotype control. To assess lung permeability, treated mice were administered FITC-dextran (10μg/mouse) i.t. before or after S. aureus infection. After 1hr of dextran instillation, blood was collected from the retro-orbital sinus, and the plasma was separated by centrifugation. Leakage of dextran in the bloodstream was measured as FITC fluorescence in the plasma compared to plasma from mocktreated mice. BAL was collected as described in (70) Briefly, the lungs of euthanized mice were lavaged through the trachea with 3ml PBS to collect the BAL. Samples were centrifuged and the supernatants were used for total protein measurement (Pierce BCA Protein Assay, Thermo Fisher Scientific #23227) and LDH quantification (Pierce LDH Cytotoxicity Assay, Thermo Fisher Scientific #C20301). Lungs were excised and used for RNA extraction using TRI Reagent (Zymo Research #R2050-1-200). The left lobe of the lung was weighed and homogenized in 1ml of sterile D.I. water in a Fisherbrand™ Bead Mill 24 Homogenizer. To calculate bacterial load, homogenate was serially diluted and plated on TSB-Agar plates in duplicate. Colonies were counted after 16h incubation, and bacterial burden in the lungs was calculated as CFU normalized to individual lung weight. Cytokines production in the lungs was measured in the supernatants collected after centrifuging the lung homogenates. Lung cells were isolated as described in (71) Briefly, mice were euthanized and perfused. 2 ml of warm dispase solution (5mg/ml) were instilled into the lungs followed by 0.5ml of 1% low-melt agarose (Sigma #A9414) at 40°C, and allowed to solidify on ice. Inflated lungs were incubated in dispase solution, for 30' at RT. The lungs were then physically dissociated, incubated 10' with DNAse I 50 μg/ml and filtered through 100μm and 70μm strainers. Red blood cells were lysed using ACK buffer. Single cell suspensions were stained for live/dead using Zombie Red or Zombie Violet, and then with antibodies against surface antigens diluted in PBS + BSA 0.2% for 20 minutes at 4°C. Cells were then washed, fixed with 3.7% paraformaldehyde for 10 minutes at room temperature, washed again and resuspended in PBS + BSA 0.2%. Samples were acquired on a BD LSRFortessa flow cytometer and data were analyzed using FlowJo v.10 software (BD Biosciences). CountBright Absolute Counting Beads (Invitrogen #C36950) were used to quantify absolute cell numbers. Purified RNA was analyzed for gene expression on a CFX384 real-time cycler (Bio-Rad) using a TaqMan RNA-to-CT 1-Step Kit (Thermo Fisher Scientific) or SYBR Green (Bio-Rad). Probes specific for Ifnl2/3, Ifnb1, Il1b, Rsad2, Gapdh were purchased from Thermo Fisher Scientific, and SYBR-Green primers for Rsad2, Cxcl10, Gapdh were purchased from Sigma. Cytokine analyses were carried out using homogenized lung supernatants, and cell supernatants from stimulated FLT3L-DCs. IFNλ2/3 ELISA (R&D Systems DY1789B) and mouse IFNβ, IL1β, IL-6, TNFα ELISA (Invitrogen) were performed according to manufacturer's instructions. Bronchoalveolar lavages (BAL) were obtained from five intensive care unit (ICU)-hospitalized SARS-CoV-2-positive patients. In parallel, five naso-oropharyngeal swabs were collected from both SARS-CoV-2-positive and -negative subjects. Among positive patients, two were hospitalized but without the need of ICU support, whereas three out of five did not require any hospitalization. The negative swabs were obtained from subjects undergoing screening for suspected social contacts with COVID-19 subjects. Swabs were performed by using FLOQSwabs® (COPAN) in UTM® Universal Transport Medium (COPAN). All samples were stored at -80°C until processing. The study involving human participants was reviewed and approved by San Raffaele Hospital IRB in the COVID-19 Biobanking project. The patients provided written informed consent. RNA extraction was performed by using PureLink™ RNA Thermo Fisher Scientific according to manufacturers' instruction. In particular, 500 µL for each BAL and swab analyzed sample were lysed and homogenized in the presence of RNase inhibitors. Then ethanol was added to homogenized samples which were further processed through a PureLink™ Micro Kit Column for RNA binding. After washing, purified total RNA was eluted in 28 µL of RNase-Free Water. System (Invitrogen™) protocol by using 8 µL of RNA extracted from each BAL and swab sample. qRT-PCR analysis for was then carried out for evaluating IL6, IL1B, IFNB1, IFNA2, IFNL1 and IFNL2 expression. All transcripts were tested in triplicate for each sample by using specific primers. GAPDH was also included. Real-time analysis was then performed according to manufacturer instructions by using TaqMan® Fast Advanced Master Mix (Applied Biosystems™ by Thermo Fisher Scientific). Real-Time PCR Analysis was performed on ABI 7900 by Applied Biosystems. Statistical significance for experiments with more than two groups was tested with one-way ANOVA, and Dunnett's multiple-comparison tests were performed. Two-way ANOVA with Tukey's multiple-comparison test was used to analyze kinetic experiments. Two-way ANOVA with Sidak's multiple-comparison test was used to analyze experiments with 2 grouped variables (i.e. treatment, genotype). Statistical significance for survival curves were evaluated with the Log-rank (Mantel-Cox) test and corrected for multiple comparisons with Bonferroni's correction. To establish the appropriate test, normal distribution and variance similarity were assessed with the D'Agostino-Pearson omnibus normality test using Prism8 (Graphpad) software. When comparisons between only two groups were made, an unpaired two-tailed t-test was used to assess statistical significance. To determine the sample size, calculations were conducted in nQuery Advisor Version 7.0. Primary outcomes for each proposed experiment were selected for the sample size calculation and sample sizes adequate to detect differences with an 80% power were selected. For animal experiments, four to ten mice per group were used, as indicated in the Cells were gated on FSC and SSC to eliminate debris, on FSC-A -FSC-H to select single cells and cells negative for live/dead dye and Lineage markers (CD3, CD19, Ter119). Epithelial cells were gated as CD45 -EpCAM + CD31 -. The EpCAMcells were sorted for immune cells as follows: aMac were gated as CD45 + Ly6g -CD11c hi Siglec-F + , monocytes and monocyte-derived cells (Mo) were gated as CD45 + Ly6g -Siglec-F -Ly6C + , cDCs were gated as CD45 + Ly6g -Siglec-F -Ly6C -CD11c + MHC-II hi . Estimates of the severity of coronavirus 457 disease 2019: a model-based analysis Clinical progression and viral load in a community 462 outbreak of coronavirus-associated SARS pneumonia: A prospective study of the N. T. U. (NTU) C. of M. Hospital Progression in Patients with Severe Acute Respiratory Syndrome Virological assessment 472 of hospitalized patients with COVID-2019 Influenza and rhinovirus viral load and 474 disease severity in upper respiratory tract infections Innate and Adaptive Immune Responses in Patients with Severe Acute Respiratory 481 Materials and Methods Mice C57BL/6J Jax 004509) mice were purchased from Jackson Labs. C57BL/6 IL-28R -/-(Ifnlr1 -/-) mice were provided by Bristol-Myers Squibb. Mice were housed under specific pathogen-free conditions at Boston Children's Hospital. Staphylococcus aureus infections were conducted in the Biosafety Level-2 facility at Boston Children's Hospital. All procedures were approved under the Institutional Animal Care and Use Committee (IACUC) and conducted under the supervision of the department of Animal CD24 (M1/69), MHC-II I-A/I-E (M5/114 R848 (tlr-r848) and 3p-hpRNA (tlrlhprna) were purchased from Invivogen. For in vivo administration of type III IFN, we used polyethylene glycol-conjugated IFN-λ2 (PEG-IFN-λ) (gift from Bristol-Myers Squibb). Diphtheria toxin (unnicked) from Corynebacterium diphtheriae was purchased from Cayman Chemical. Anti-Ly6G antibody, clone 1A8 (BE0075-1) and rat IgG2a isotype control (BE0089) for in vivo administration was purchased from Bioxcell. 2'-Deoxy-5-ethynyl uridine (EdU) was purchased from Carbosynth (NE08701) Epithelial cell proliferation') before being stained with antibodies against cell-surface antigens. Intracellular staining of Ki67 and p21 were carried out using FoxP3 Fix/Perm Buffer set (Biolegend #421403) following the manufacturer's instructions. Epithelial cell proliferation Proliferation of lung After permeabilization cells were washed and incubated with 4 mM Copper sulphate (Millipore-Sigma), 100 mM Sodium ascorbate (Millipore-Sigma) and 5 μM sulfo-Cyanine3-azide (Lumiprobe #A1330) in Tris Buffered Saline (TBS) 100mM, pH 7.6, for 30 min at room temperature. Ion Torrent For targeted transcriptome sequencing, 20 ng of RNA isolated from sorted cells was retro Barcoded libraries were prepared using the Ion AmpliSeq Transcriptome Mouse Gene Expression Kit as per the manufacturer's protocol and sequenced using an Ion S5 system Genes were called expressed (n=11,294) if they had average log2 expression of 2 or greater in either WT or Ifnlr1 -/-. Differentially expressed genes (DEGs) between WT and Ifnlr1 -/-were selected by thresholding on fold change (+/-1.5) and p value (0.005). In heatmaps, DEGs were Z-scaled and clustered (Euclidean distance, Ward linkage). Pathway analysis was performed with the R package hypeR Cell culture FLT3L-DCs were differentiated from bone marrow cells in Iscove's Modified Dulbecco's Media (IMDM; Thermo Fisher Scientific), supplemented with 30% B16-FLT3L derived supernatant and 10% fetal bovine serum (FBS) for Where indicated poly (I:C) stimulated cells were pre-treated with 10μg/ml chloroquine for 5 minutes prior to stimulations. qRT-PCR and ELISA RNA was isolated from cell cultures using a GeneJET RNA Purification Kit (Thermo Fisher Scientific #K0731) according to manufacturer's instructions. RNA was extracted from excised lungs by homogenizing them in 1ml of TRI Reagent. RNA was isolated from TRI Reagent samples using phenol-chloroform extraction or column-based extraction systems (Direct-zol RNA Microprep and Miniprep, Zymo Research #R2061 and #R2051) according to the manufacturer's protocol Flow Cytometric Isolation of Primary Murine Type II Alveolar Epithelial Cells for Functional and Molecular Studies WT mice were treated daily with i.t. 0.5 mg/kg poly (I:C), 0.5 mg/kg R848 or saline for 6 days and infected i.t. with 0.5 x 10 8 CFU of S. aureus at day 6. (A) Body temperature, (B) total protein in the BAL and **P < 0.01 and ***P < 0.001 (One-way ANOVA). Each mouse represents one point A-H) WT mice were treated daily for 1, 3, 5 or 6 days with i.t. 0.5 mg/kg poly (I:C) or 6 days of saline, and infected with i.t. 0.5 x 10 8 CFU of S. aureus for 12h. Total lung homogenates were analyzed by qPCR for Bacterial burden was evaluated in total lung homogenate **P < 0.01 and ***P < 0.001 (One-way ANOVA compared to Saline treatment). Each mouse represents one point WT and Ifnlr1 -/-mice were treated daily with i.t. 0.5 mg/kg poly (I:C) for 6 days and infected with i.t. 0.5 x 10 8 CFU of S. aureus for 12h. (A) Weight change, (B) total protein in the BAL **P < 0.01 and ***P < 0.001 (Two-way ANOVA). Each mouse represents one point Lung resident DCs are the primary producers of IFN-β upon poly (I:C) treatment 5 mg/kg poly (I:C) or saline for 6 days were sorted for epithelial cells (EC), resident DC (resDC), monocyte-derived DC (moDC), and alveolar macrophages (aMac) and assessed for (A) Il1b and (B) Ifnb1 relative mRNA expressions. CD11c-DTR mice depleted for CD11c + cells in vivo by DTx injections were treated daily with i.t. 0.5 mg/kg poly (I:C) or saline for 6 days. Total lung lysates of the treated mice were analyzed for (C) Ifnb1 relative mRNA expression, and (D) IFN-β protein expression by ELISA 05, **P < 0.01 and ***P < 0.001 (Two-way ANOVA) RIG-I or TLR7 ligands. FLT3L-DCs from WT, Ticam1 -/-or Mavs -/-mice were treated with 50μg/ml poly (I:C), 1μg/ml transfected poly (I:C) or 50μg/ml R848 for 3h. Ifnb1 (A), and Il1b (B) relative mRNA expressions were evaluated by qPCR 05, **P < 0.01 and ***P < 0.001 (Two-way ANOVA) Mean and SEM of 3 independent experiments is depicted FLT3L-DCs upregulate IFN-λ uniquely upon activation of TLR3 signaling and not in response to the RIG-I specific ligand 3p-hpRNA. FLT3L-DCs from WT Mavs -/-mice were treated with 50μg/ml poly (I:C), or 1μg/ml transfected 3p-hpRNA for 3h or 6h Ifnb1 (B), and Il1b (C) relative mRNA expressions were evaluated by qPCR after 3h and IFN-β (E) levels in the supernatants were evaluated by ELISA after 6h The endosomal TLR inhibitor Chloroquine inhibits poly (I:C) dependent IFN-λ expression in FLT3L-DCs. FLT3L-DCs from WT mice were treated with 50μg/ml poly (I:C), or 1μg/ml transfected poly (I:C) for 3h in the presence or absence of 10μg/ml Chloroquine We thank Dr. JC Kagan for discussion, help and support. Funding: IZ is supported by NIH grant 506 1R01AI121066, 1R01DK115217, and NIAID-DAIT-NIHAI201700100. AB is supported by CCFA