Interferon-Λ Enhances Adaptive Mucosal Immunity by Boosting Release of Thymic Stromal Lymphopoietin

Articles https://doi.org/10.1038/s41590-019-0345-x

Interferon-λ enhances adaptive mucosal immunity by boosting release of thymic stromal lymphopoietin

Liang Ye1, Daniel Schnepf1,2, Jan Becker1, Karolina Ebert3, Yakup Tanriver3,4,5, Valentina Bernasconi 6, Hans Henrik Gad7, Rune Hartmann 7, Nils Lycke6 and Peter Staeheli 1,4*

Interferon-λ (IFN-λ) acts on mucosal epithelial cells and thereby confers direct antiviral protection. In contrast, the role of IFN-λ in adaptive immunity is far less clear. Here, we report that mice deficient in IFN-λ signaling exhibited impaired CD8+ T cell and antibody responses after infection with a live-attenuated influenza virus. Virus-induced release of IFN-λ triggered the synthesis of thymic stromal lymphopoietin (TSLP) by M cells in the upper airways that, in turn, stimulated migratory dendritic cells and boosted antigen-dependent germinal center reactions in draining lymph nodes. The IFN-λ–TSLP axis also boosted production of the immunoglobulins IgG1 and IgA after intranasal immunization with influenza virus subunit vaccines and improved survival of mice after challenge with virulent influenza viruses. IFN-λ did not influence the efficacy of vaccines applied by subcutaneous or intraperitoneal routes, indicating that IFN-λ plays a vital role in potentiating adaptive immune responses that initiate at mucosal surfaces.

nterferon-λ (IFN-λ) is an antiviral cytokine produced in response M cells which, in turn, influences germinal center responses by act- to viral infection in a variety of cell types, including airway epithe- ing on migratory dendritic cells (DCs). This previously unknown Ilial cells1,2. IFN-λ acts by binding to a heterodimeric surface recep- IFN-λ–TSLP axis strongly enhances mucosal immunity and confers tor consisting of the ubiquitously expressed IL-10 receptor β-chain enhanced influenza virus resistance. and a second IFN-λ-specific chain designated IFNLR1 (refs. 2–4). IFNLR1 is present on epithelial cells but seemingly absent on most Results immune cells of mice2,3, with the exception of neutrophils5–8. IFN-λ Defective antibody production and weak CD8+ T cell responses in induces an antiviral response in epithelial tissues and thus protects Ifnlr1–/– mice after infection with live-attenuated influenza virus. against infection with viruses that replicate preferentially in the To detect possible impairments in adaptive immunity resulting from respiratory tract9,10 or in the intestine11–13. IFN-λ further protects IFN-λ deficiency, we infected Mx1-WT and Mx1-Ifnlr1–/– mice with against fungal infections of the respiratory tract in mice by activat- a live-attenuated influenza A virus strain, designated hvPR8-ΔNS1. ing neutrophils to produce reactive oxygen species8. These mice carry functional alleles of the IFN-regulated influenza Little is known concerning the role of IFN-λ in adaptive immu- virus resistance gene Mx1 required for revealing the full antiviral nity. The IFN-λ receptor is expressed on B cells from humans7,14 but potential of IFN10. The hvPR8-ΔNS1mutant virus cannot produce not mice6. A study in the human system indicated that IFN-λ is a the NS1 virulence factor and, therefore, fails to suppress the IFN negative regulator of adaptive immunity that can suppress vaccine- response of the host. NS1-deficient influenza viruses are largely induced antibody production by a poorly defined mechanism14. non-pathogenic and trigger strong B and T cell responses in immu- However, another study showed that IFN-λ can stimulate rather nocompetent mice17,18. Mx1-Ifnlr1–/– mice had 8–15-fold reduced than inhibit IgG synthesis by human B cells7. The reasons for these levels of influenza virus hemagglutinin (HA)-specific serum IgG1 discrepant findings remain elusive. In mice, IFN-λ was reported to compared with Mx1-WT mice at 14 and 21 days post-infection with boost CD8+ T cell-mediated immunity by reducing regulatory T cell hvPR8-ΔNS1, whereas serum levels of other IgG subtypes were populations during DNA vaccination15. In another study, IFN-λ was normal (Fig. 1a). HA-specific IgA levels in bronchoalveolar lavage found to enhance neutralizing antibody titers against herpes simplex (BAL) fluids of Mx1-Ifnlr1–/– mice at day 21 post-infection were also virus 2 in mice immunized with a DNA vaccine16, but the mecha- significantly reduced compared with Mx1-WT mice (Fig. 1b). Mx1- nism by which IFN-λ confers adjuvant activity was not resolved. Ifnlr1–/– mice further contained significantly fewer influenza virus We searched for immunomodulatory effects of IFN-λ dur- nucleoprotein-specific CD8+ T cells in the mediastinal lymph nodes ing viral infection of the respiratory tract and found that IFN-λ is at day 7 post-infection with hvPR8-ΔNS1 compared with Mx1-WT required for full-scale production of virus-specific IgG1 and IgA, as mice (Fig. 1c). well as efficient generation of antiviral CD8+ T cells. Mechanistically, In agreement with earlier work6,9,19, we failed to detect functional we show that IFN-λ triggers the synthesis of TSLP in upper airway IFN-λ receptors on naive (Fig. 1d) or antigen-experienced mouse

1Institute of Virology, Medical Center University of Freiburg, Freiburg, Germany. 2Spemann Graduate School of Biology and Medicine, Albert Ludwigs University Freiburg, Freiburg, Germany. 3Institute of Medical Microbiology, Medical Center University of Freiburg, Freiburg, Germany. 4Medical Faculty, University of Freiburg, Freiburg, Germany. 5Department of Internal Medicine IV, Medical Center University of Freiburg, Freiburg, Germany. 6Department of Microbiology and Immunology, University of Gothenburg, Gothenburg, Sweden. 7Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark. *e-mail: [email protected]

Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology 593 Articles NATure IMMunoLoGy

a IgG IgG1 IgG2c IgG2b b IgA )

10 5 5 5 5 WT 1.2 * WT **** –/– –/– 4 4 **** 4 4 Ifnlr1 0.9 Ifnlr1 3 3 3 3 0.6 2 2 2 2 1 1 1 1 0.3 0 0 0 0 0.0 HA-specific Ab (OD) HA-specific Ab (log 7 14 21 7 14 21 7 14 21 7 14 21 21 days post-infection Days post-infection

c –/– d Control WT Ifnlr1 Mock IFN-λ IFN-α 105 2 WT 100 500 4 –/– ** 0.02 0.68 0.29 cell (%) Ifnlr1 80 400

10 + 60 300 103 *

CD8 1 + 40 200 2 10 20 100 0 p-STAT1 MFI Pentamer-PE 0 0

Normalized to mode 0 3 4 5 3 4 0 10 10 10 Pentamer 0 10 10 CD8-FITC p-STAT1-PE

e + + f B220 B cells MFI CD19 B cells MFI –/– WT → WT Ifnlr1 → WT IFN- WT Ifnlr1–/– Ifnlr1–/– Ifnlr1–/– –/– α 155 223 → →

IFN- 158 242 ) λ 6 10

Ifnar1 Mock 155 241 5 *** ** IFN-α 303 546 –/– 4 IFN-λ 201 293

Ifnlr1 Mock 214 296 3 IFN-α 324 507 2 IFN-λ 188 265 WT 1 Mock 174 328

HA-specific lgG1 (log 0 01103 104 105 0 103 104 05 21 days post-infection p-STAT1-PE

Fig. 1 | Defective immune response in Ifnlr1–/– mice after infection with live-attenuated influenza virus is not rescued by hematopoietic cells from wild type mice. a, HA-specific antibodies in sera of Mx1-WT (n = 8) and Mx1-Ifnlr1–/– (n = 7) mice after infection with hvPR8-ΔNS1. The data are representative of three independent experiments. ****P < 0.0001, by two-way ANOVA with Tukey’s multiple-comparison test. b, HA-specific IgA in BAL fluids of Mx1-WT and Mx1-Ifnlr1–/– mice after infection with hvPR8-ΔNS1. Data are representative of two independent experiments with seven mice per group. *P = 0.0117, by unpaired two-tailed Student’s t-test. c, Mx1-WT (n = 5) and Mx1-Ifnlr1–/– (n = 4) mice were infected with hvPR8-ΔNS1 virus. Seven days later, the frequency of influenza virus nucleoprotein-specific CD8+ T cells in mediastinal lymph nodes (mLNs) was measured by flow cytometry using labeled H-2Db/ ASNENMETM pentamer. *P = 0.0157, by unpaired two-tailed Student’s t-test. Data are representative of two independent experiments. d, Isolated spleen cells from naive mice were mock-treated or stimulated in the presence or absence of 1 µg ml−1 IFN-α or IFN-λ2 for 30 min before the levels of phosphorylated STAT1 (p-STAT1) in CD19+ B cells were quantified by flow cytometry. Data are pooled from two independent experiments with six mice per group. **P = 0.0031, by unpaired two-tailed Student’s t-test. e, Mx1-WT (n = 3), Mx1-Ifnar1–/– (n = 3) and Mx1-Ifnlr1–/– (n = 3) mice were vaccinated by intranasal infection with 105 PFU of hvPR8-ΔNS1. At day 21 post-infection, isolated spleen cells were stimulated with 1 µg ml−1 IFN-α or IFN-λ2 for 30 min before levels of p-STAT1 were quantified by flow cytometry. Geometric means of fluorescence intensity (MFI) for p-STAT1 in B220+ or CD19+ B cells are depicted. f, Serum IgG1 levels in BM chimeric mice (n = 6 per group) after infection with hvPR8-ΔNS1. ***P < 0.001 and **P < 0.01, by one-way ANOVA with Tukey’s multiple-comparison test. a–f, Error bars represent s.e.m. centered on the mean.

B cells (Fig. 1e). To determine whether IFN-λ receptor expression intranasally to Mx1-WT mice, only low HA-specific serum IgG on other immune cells might be important for adequate immune titers were observed after one priming and two booster immuniza- responses to live-attenuated influenza virus, we generated bone tions (Fig. 2a). However, when this vaccine was enriched by adding marrow (BM) chimeric mice and measured HA-specific IgG1 levels 1 µg of mouse IFN-λ2 (ref. 20), HA-specific total serum IgG levels in serum after infecting these animals with hvPR8-ΔNS1. Wild-type were about 15-fold increased (Fig. 2a). Analysis of IgG subclasses (WT) mice showed strong IgG1 responses even when their immune revealed that IFN-λ2 enhanced serum IgG1 levels, but did not sig- cells were derived from Ifnlr1–/– mice, and Ifnlr1–/– mice continued nificantly influence the other IgG subclasses (Fig. 2a). In addition, to show weak IgG1 responses even when they were reconstituted IFN-λ2 significantly enhanced HA-specific IgA levels in the BAL with immune cells from wild type animals (Fig. 1f). Thus, IFN-λ fluids (Fig. 2b). The stimulatory effect of IFN-λ2 on HA-specific enhances antibody production through a mechanism that depends IgG1 production was not observed when the Influsplit Tetra vac- on non-hematopoietic cells rather than BM-derived immune cells. cine was applied by either the intraperitoneal (Supplementary Fig. 1a) or the subcutaneous routes (Supplementary Fig. 1b), indicating IFN-λ enhances IgG1 and IgA production after intranasal that IFN-λ-responsive cells in the respiratory tract account for the application of subunit vaccines. To test whether IFN-λ can also IFN-λ-induced stimulation of adaptive immunity. influence antibody production in mice immunized with replica- To determine whether IFN-λ might also improve antibody tion-incompetent subunit vaccines, we used commercial Influsplit responses elicited by vaccine formulation containing a mucosal Tetra influenza vaccine that contains hemagglutinin of four differ- adjuvant, we used the vaccine CTA1-3M2e-DD, designated M2e. ent seasonal influenza A and B viruses as major antigen. When 4 µg This vaccine contains the evolutionarily conserved extracellular of this tetravalent vaccine (1 µg of each HA subtype) was applied domain of the influenza A virus M2 protein fused to the cholera

594 Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology NATure IMMunoLoGy Articles

a b ) ) IgG IgG1 IgG2c lgG2b lgA 10 10 4 4 4 4 HA 3 HA *** 3 3 **** 3 3 HA+|FN-λ **** HA+|FN-λ 2 2 2 2 2 1 1 1 1 1 0 0 0 0 0 HA-specific Ab (log 1st 2nd 1st 2nd 1st 2nd 1st 2nd HA-specific lgA (log

c d ) ) IgG IgG1 IgG2c lgG2b lgA 10 10 5 5 5 5 M2e 4 M2e 4 **** 4 **** 4 4 M2e+|FN-λ 3 **** M2e+|FN-λ 3 3 3 3 2 2 2 2 2 1 1 1 1 1 0 0 0 0 0 M2e-specific Ab (log M2e-specific lgA (log

Fig. 2 | IFN-λ enhances IgG1 and IgA production after intranasal application of influenza subunit vaccines. a,b, Mx1-WT mice were immunized by intranasal application of Influsplit Tetra in the presence (n = 10) or absence (n = 10) of IFN-λ2, and HA-specific IgG titers in sera (a) were determined 10 days after the first and the second booster immunization. IgA levels in BAL fluids (b) were determined 10 days after the second booster immunization. Data are pooled from two independent experiments. Shown are mean ± s.e.m. ***P = 0.0002 (a, left), ****P < 0.0001 (a, right), ****P < 0.0001 (b), by unpaired two-tailed Student’s t-test. c, Mx1-WT mice were immunized by intranasal application of M2e vaccine with (n = 18) or without (n = 17) IFN-λ2, and M2e-specific IgG titers in sera were determined 10 days after a single booster immunization. Data are pooled from three independent experiments. Error bars represent s.e.m. centered on the mean. ****P < 0.0001, by unpaired two-tailed Student’s t-test. d, Mx1-WT mice were immunized by intranasal application of M2e vaccine with (n = 10) or without (n = 11) IFN-λ2, and the IgA levels in BAL fluids were measured 10 days after a single booster immunization. Data are pooled from two independent experiments and shown as mean ± s.e.m. ****P < 0.0001, by unpaired two-tailed Student’s t-test.

toxin A1 ADP-ribosylating enzyme21. M2e vaccine given intra- noted when TSLP receptor-deficient (Tslpr–/–) mice were infected nasally stimulated detectable M2e-specific serum antibodies in with hvPR8-ΔNS1. We next determined whether neutralization Mx1-WT mice after one priming and one single booster immu- of TSLP during the early stages of the immune response would nization (Fig. 2c). When the M2e vaccine was supplemented with reduce the immunostimulatory effect of IFN-λ. When Mx1-WT IFN-λ2, serum levels of M2e-specific IgG1, but not of other IgG mice were treated with a TSLP-specific monoclonal antibody by subclasses, were increased about 25-fold compared with con- the intranasal route during infection with hvPR8-ΔNS1, virus- trol mice (Fig. 2c). Moreover, M2e-specific IgA was found only specific serum IgG1 levels were about eightfold reduced in anti- in the BAL fluids of mice receiving IFN-λ2-enriched antigen TSLP-treated animals compared with isotype-treated controls at (Fig. 2d). No enhancing effect of IFN-λ2 was observed when 14 and 21 days post-infection (Fig. 3f). immunizations with the M2e vaccine were performed via the No IgG1-enhancing effect of IFN-λ was observed in Tslpr–/– mice intraperitoneal (Supplementary Fig. 1c) or the subcutaneous immunized with the M2e (Fig. 3g) or HA-based Influsplit Tetra vac- routes (Supplementary Fig. 1d). Thus, IFN-λ selectively potenti- cine (Fig. 3h). Consistent with these findings, intranasal immuniza- ated adaptive immune responses to antigens that entered the host tion with IFN-λ in Tslpr–/– mice failed to boost the IgA response via mucosal surfaces of the respiratory tract. against the M2e vaccine (Fig. 3i), demonstrating that the enhancing effect of IFN-λ on IgG1 and IgA production depends on TSLP. The enhancing effect of IFN-λ on IgG1 and IgA production depends on TSLP. Since IFN-λ2 strongly enhanced vaccine efficacy Upper airway M cells produce TSLP after exposure to IFN-λ. only when antigen was provided by the intranasal but not by the Epithelial cells can produce TSLP24,26. To determine whether upper intraperitoneal or subcutaneous routes, we hypothesized that IFN-λ airway epithelial cells produced TSLP during intranasal immu- might trigger airway epithelial cells to produce a factor which stim- nization, cells isolated from mouse snouts were stained with a ulates IgG1 and IgA production. Thymic stromal lymphopoietin TSLP-specific antibody and analyzed by flow cytometry. No TSLP- (TSLP) is a cytokine that shifts antibody production toward IgG1 positive cells were detected in the unfractionated CD45– EpCAM+ and IgA synthesis22. It acts on leukocytes by engaging a heterodi- epithelial cell population of Mx1-WT mice infected with hvPR8- meric receptor consisting of the IL-7 α-chain and a second TSLP- ΔNS1 (Supplementary Fig. 2a). Since M cells are particularly specific chain, designated TSLPR23,24. TSLP can boost mucosal important for immunity against inhaled antigens27, we determined immunity by triggering the production of IgA-promoting cytokines whether these specialized epithelial cells might represent an impor- in mucosal CD11c+ dendritic cells25. TSLP was reported to increase tant source of TSLP. Approximately 2% of the CD45– EpCAM+ epi- antibody production in mice on intranasal vaccination with HIV-1 thelial cells isolated from the upper airways of mice were identified glycoprotein22. as NKM16-2-4+ M cells (Supplementary Fig. 3a). In infected mice, To determine whether TSLP plays a role in our system, we approximately 7% of the upper airway M cells expressed detect- first compared the immune-enhancing effect of TSLP with that of able levels of TSLP compared with about 2% in uninfected mice IFN-λ2. When given in combination with the M2e vaccine, TSLP (Fig. 4a). In contrast, M cells from the upper airways of Mx1- stimulated enhanced M2e-specific serum IgG1 levels to a similar Ifnlr1–/– mice contained approximately 2% TSLP+ cells, irrespec- extent as IFN-λ2 (Fig. 3a). The same observation was made when tive of whether the animals were infected with hvPR8-ΔNS1 or not TSLP was given together with the HA-based Influsplit Tetra vaccine (Fig. 4a and Supplementary Fig. 2a). (Fig. 3b). A pronounced defect in producing virus-specific IgG1 When applied into the nostrils of uninfected WT mice, IFN- (Fig. 3c) and nucleoprotein-specific CD8+ T cells (Fig. 3d,e) was λ2 induced TSLP expression in approximately 5% of the upper

Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology 595 Articles NATure IMMunoLoGy

abM2e HA cd –/– M2e+IFN-λ HA+IFN-λ WT ControlWT Tslpr 5 M2e+TSLP HA+TSLP Tslpr –/– 10 4 10 0 0.77 0.29 mLN

) **** 3 5 ) 5 ) 5 10 10

** 10 10 *** *** 2 4 4 4 10 *** 0 3 3 ** 3 0 103104105 5 2 2 2 10

0 15.6 6.18 BAL 104 3 1 1 1 10 102 HA-specific lgG1 (log HA-specific lgG1 (log entamer-PE M2e-specific lgG1 (log

0 0 0 P 0 d14 d21 0 103104105 CD8-FITC efg hi WT IgG control M2e HA M2e –/– Tslpr Anti-TSLP M2e+IFN-λ HA+IFN-λ M2e+IFN-λ ) ) ) 5 2.0 30 5 5 5 ) 10 10 10 **** **** 10 4 4 **** 4 4 1.5 ** **** ****

cells (%) 20 **** + ** 3 3 3 ** *** 3 1.0 CD 8

+ 2 2 2 2 10 0.5 1 1 1 1 entamer M2e-specific lgA (log HA-specific lgG1 (log HA-specific lgG1 (log M2e-specific lgG1 (log P 0.0 0 0 0 0 0 mLN BAL d14 d21 WT Tslpr –/– WT Tslpr –/– WT Tslpr –/–

Fig. 3 | The immune-enhancing effect of IFN-λ depends on TSLP. a, WT mice were immunized by intranasal application of M2e (n = 12) or M2e-enriched with IFN-λ2 (n = 6) or TSLP (n = 12). Ten days after one booster immunization, blood samples were analyzed for M2e-specific IgG1. b, WT mice were immunized with Influsplit Tetra (n = 5) or Influsplit Tetra plus IFN-λ2 (n = 5) or TSLP (n = 6) by the intranasal route. HA-specific IgG1 titers in sera were analyzed 10 days after the second booster immunization. c, HA-specific IgG1 in sera of WT (n = 7) and Tslpr–/– (n = 6) mice infected with hvPR8-ΔNS1. d,e, Flow cytometry contour plots (d) and graphs (e) show the frequency of influenza virus nucleoprotein-specific CD8+ T cells in mLNs (n = 5 mice per group) and BAL cells (n = 4 black symbols, n = 5 blue symbols) of WT and Tslpr–/– mice on day 7 post-infection with hvPR8-ΔNS1. f, HA-specific IgG1 in sera of Mx1-WT mice (n = 10 mice per group) treated with anti-TSLP antibody during infection with hvPR8-ΔNS1. g, WT (n = 17 black symbols, n = 19 red symbols) and Tslpr–/– (n = 16 black symbols, n = 18 red symbols) mice immunized with M2e in the presence of IFN-λ2. M2e-specific IgG1 in sera were analyzed at day 10 after a single booster immunization. h, WT (n = 5) and Tslpr–/– (n = 6) mice were immunized with Influsplit Tetra in the presence or absence of IFN-λ2. HA-specific IgG1 in sera were measured 10 days after the second booster immunization. i, WT (n = 11 black symbols, n = 12 red symbols) and Tslpr–/– (n = 11) mice immunized with M2e in the presence of IFN-λ2. M2e-specific IgA in BAL fluids were analyzed at day 10 after a single booster immunization. a–i, Error bars represent s.e.m. centered on the mean. Data are pooled from two (a,f,i) and three (g) individual experiments. Data are representative of two (c,e,h) or three (b) independent experiments. **P = 0.0015 (a, left), ****P < 0.0001(a, right), **P = 0.0017 (b, left), ***P < 0.0009 (b, right), ***P = 0.0006 (c, left), ***P = 0.0009 (c, right), **P = 0.0047 (e, left), **P = 0.0016 (e, right), ****P < 0.0001 (f,g,i), **P = 0.0013 (h, left), ***P = 0.0002 (h, right), by unpaired two-tailed Student’s t-test. airway M cells compared with, on average, 1.5% in mock-treated IFN-λ promotes TSLP-dependent germinal center reactions. control mice (Fig. 4b and Supplementary Fig. 2b). IFN-λ2-induced Follicular helper T cells (Tfh cells) are critical for B cell expansion expression of the Tslp gene was barely detected by flow cytometry and differentiation in the germinal center28,29, where immunoglobu- in cultured tracheal epithelial cells expressing or lacking the M cell lin class switching and high-affinity maturation of IgG and IgA anti- marker (Supplementary Fig. 4), indicating that the M cells of such bodies occurs29. Intranasal immunization with the M2e vaccine in cultures are not fully functional. the presence of IFN-λ2 led to a twofold increased frequency of PD1+ To corroborate these results, we isolated M cells from the CXCR5+ Tfh cells in lymph nodes of wild type mice compared with upper airways of IFN-λ2-treated mice (gating strategy outlined Tslpr–/– mice, or when the immunizations were performed in the in Supplementary Fig. 3b), and analyzed Tslp gene expression absence of IFN-λ (Fig. 5a). A comparable increase in the frequency by reverse transcription–quantitative PCR (RT–qPCR). M cells of Tfh cells was observed in spleens of WT but not Tslpr–/– mice from IFN-λ-treated mice contained about ninefold higher lev- immunized in the presence of IFN-λ2 (Supplementary Fig. 5a). els of Tslp transcripts than untreated mice (Fig. 4c). As expected, Furthermore, WT mice immunized in the presence of IFN-λ2 Mx1 transcript levels were also strongly enhanced in M cells hosted at least threefold more Fas+ GL7+ germinal center B cells in from IFN-λ2-treated mice compared with untreated controls the lymph nodes than mice immunized in the absence of IFN-λ2 (Fig. 4c). By contrast, IFN-λ2 did not induce Tslp messenger RNA (Fig. 5b). Likewise, approximately four times more IgG1-producing to any great extent in the M cell-depleted epithelial cell fraction, germinal center B cells were present in mice immunized with M2e although transcription of the Mx1 gene was strongly induced and IFN-λ2 compared with mice vaccinated with M2e alone (Fig. 5c). (Supplementary Fig. 2c). Consistent with the notion that immune Such stimulatory effects of IFN-λ on the germinal center B cell cells largely lack IFN-λ receptors, IFN-λ did not induce the Tslp frequency and IgG1 synthesis were not observed in Tslpr–/– mice or Mx1 genes in CD45+ cells (that is, hematopoietic cells) derived (Fig. 5b,c). A similar picture emerged when spleens rather than from the upper airway of mice (Supplementary Fig. 2c). Taken lymph nodes of immunized mice were analyzed (Supplementary together, these results demonstrated that upper respiratory tract Fig. 5b,c). Intranasal treatment of mice with IFN-λ2 in the absence M cells can respond to IFN-λ and represent a major source of of vaccine did not induce such germinal center alterations in drain- TSLP in IFN-λ-exposed mice. ing lymph nodes or spleens (Supplementary Fig. 6), demonstrating

596 Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology NATure IMMunoLoGy Articles

a –/– WT lfnlr1 Mock hvPR8- NS1 Isotype Mock hvPR8-∆NS1 Mock hvPR8-∆NS1 ∆ 5 10 0 1.42 6.08 0.96 1.61 12 ** ** 4 10 8 cells (%) + M

0 + 4 TSLP-PE TSLP 0 –/– 0 WT lfnlr1 50 k 100 k 150 k FSC

b –/– WT lfnlr1 Mock

Isotype Mock IFN-λ Mock IFN-λ IFN-λ 5 10 0 2.04 6.09 1.20 1.56 9 **** **** 4 10 6 cells (%) + M

0 + 3 TSLP-PE TSLP 0 0 WT lfnlr1–/– 50 k 100 k 150 k FSC

c Flow-sorted M+ cells

Tslp Mx1 ) )

10 –3.5 10 –3 * * –4.0 –4 –4.5 –5 –5.0 –5.5 –6 –6.0 –7 Relative to β - actin (log Relative to β - actin (log Mock IFN-λ Mock IFN-λ

Fig. 4 | Upper airway M cells produce TSLP after IFN-λ stimulation. a, Flow cytometry analyzing TSLP in snout M cells (NKM16-2-4+ EpCAM+ CD45−) from Mx1-WT (n = 6 per group) and Mx1-Ifnlr1–/– (n = 6 black symbols, n = 5 red symbols) mice infected with hvPR8-ΔNS1 for 24 hours. The frequency of TSLP-expressing M cells is indicated. The data are representative of two independent experiments and shown as mean ± s.e.m. **P = 0.0012 (left), **P = 0.0053 (right) by unpaired two-tailed Student’s t-test. b, Flow cytometry analyzing TSLP in snout M cells from Mx1-WT (n = 6 black symbols, n = 9 red symbols) and Mx1-Ifnlr1–/– mice (n = 7 per group) treated intranasally with IFN-λ2 for 24 hours. The frequency of TSLP-expressing M cells is indicated. Data are pooled from two independent experiments and expressed as mean ± s.e.m. ****P < 0.0001, by unpaired two-tailed Student’s t-test. c, M cells were purified from snouts of Mx1-WT mice treated intranasally with either saline (mock) or IFN-λ2 for 4 hours. Cells from 12 mice per group were pooled before sorting. Expression of the Tslp and Mx1 genes were assessed by RT–qPCR. Data are pooled from three independent experiments and expressed as mean ± s.e.m. *P = 0.02 (left), *P = 0.04 (right), by unpaired two-tailed Student’s t-test. that the observed changes in immune cell composition are driven Batf3–/– mice, in which CD103+ migratory DCs are defective30, by antigen. In summary, these results demonstrated that IFN-λ fail to mount normal IgA and IgG responses to poly(I:C)-adjuvanted promotes germinal center reactions in draining lymph nodes via a mucosal vaccines31. Accordingly, chimeric mice reconstituted with TSLP-dependent pathway. BM cells from Batf3–/– mice had significantly lower HA-specific serum IgG1 and mucosal IgA levels (Fig. 6a) than control mice if IFN-λ-triggered TSLP acts on CD103+ migratory DCs. To iden- immunized by intranasal infection with hvPR8-ΔNS1, indicat- tify the TSLP-responsive immune cells that improve germinal cen- ing that CD103+ DCs are critically involved in shaping immunity ter reactions in our system, we first generated BM chimeric mice against live-attenuated influenza virus. containing TSLP receptor-deficient B and T cells by reconstituting To determine whether CD103+ cells are responding to IFN-λ- lethally irradiated WT mice with a 1:5 mixture of BM cells from triggered TSLP in our system, we generated BM chimeric mice con- Tslpr–/– and Rag1–/– mice, respectively. In such chimeric mice, B and taining TSLP receptor-deficient CD103+ migratory DCs. To produce T cells originate exclusively from the Tslpr–/– bone marrow, whereas such animals, we reconstituted lethally irradiated WT mice with a the majority of the other hematopoietic cell types (that is, non-B 1:1 mixture of BM cells from Tslpr–/– and Batf3–/– mice. In such chi- and non-T cells) are TSLP signaling-competent, as they are mostly mera, CD103+ migratory DCs are expected to originate exclusively derived from Rag1–/– mice. Seven weeks after BM transplantation, from the Tslpr–/– bone marrow, whereas all other immune cells origi- the chimeric mice were infected intranasally with hvPR8-ΔNS1. nate to similar extents from TSLP-competent and TSLP-defective We found that HA-specific serum IgG1 and IgA levels (Fig. 6a) in bone marrow. When such mixed BM chimeric mice were immu- these chimeric mice were unperturbed, indicating that TSLP sig- nized by intranasal infection with hvPR8-ΔNS1, HA-specific serum naling in T and B cells is not required for the immune-enhancing IgG1 and mucosal IgA levels (Fig. 6a) were significantly lower than effects of IFN-λ. in control mice carrying TSLP receptor-competent CD103+ DCs.

Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology 597 Articles NATure IMMunoLoGy

a WT Tslpr –/–

M2e M2e+IFN-λ M2e M2e+IFN-λ 20 5 M2e 10 6.49 13.7 5.72 6.42 ** ** 4 15 M2e+IFN-λ 10 103 10 102 5 Tfh cells (%) 0 PD1-FITC 0 –/– 0 103 104105 WT Tslpr CXCR5-APC

b –/– WT Tslpr M2e M2e+IFN- M2e M2e+IFN- λ λ 15 5 M2e 10 2.72 7.73 2.61 3.04 ** ** M2e+IFN- 104 10 λ

103 ercp/cy5.5 5 0 GC B cells (%) as-P

F 0 0 103 104 105 WT Tslpr –/– GL7-Pacific Blue

c WT Tslpr –/–

M2e M2e+IFN-λ M2e M2e+IFN-λ 5 4 M2e 10 *** 4 ** M2e+IFN- (%) λ 10 0.58 2.37 0.49 0.52 + s 3 2 10 Fa + 102 lgG 1

lgG1-FITC 0 0 0 103 104 105 WT Tslpr –/– Fas-Percp/cy5.5

Fig. 5 | IFN-λ promotes TSLP-dependent germinal center reactions. a–c, WT (n = 3) and Tslpr-/- (n = 4) mice were intranasally immunized with M2e vaccine in the presence or absence of IFN-λ2. Animals were killed on day 5 post-boosting, and the frequencies of CXCR5+ PD1+ Tfh cells among CD19− CD4+ CD44+ cells (a) and GL7+ Fas+ germinal center (GC) B cells among CD4− B220+ cells (b) in draining lymph nodes were determined. c, Frequency of IgG1+ Fas+ GC B cells among CD4− B220+ cells in draining lymph nodes. Results are representative of two independent experiments and shown as mean ± s.e.m. Each symbol represents an individual animal. **P = 0.0052 (a, left), **P = 0.0017 (a, right), **P = 0.0013 (b, left), **P = 0.0012 (b, right), ***P = 0.0008 (c, left), **P = 0.002 (c, right), by unpaired two-tailed Student’s t-test.

These results revealed that IFN-λ-induced TSLP enhances antibody M2e-specific antibodies mediate influenza resistance by an production by acting on CD103+ migratory DCs. Fcγ receptor-dependent mechanism32. IFN-λ2 significantly We performed flow cytometric analyses to determine whether enhanced the protective ability of the M2e vaccine in Mx1-WT IFN-λ can increase the number of CD103+ DCs in mediastinal mice against a highly virulent variant of PR8/34 (refs. 18,33) lymph nodes of immunized mice. Mx1-WT contained signifi- compared with IFN-λ-free M2e vaccine (Fig. 7b) and reduced cantly higher numbers of CD103+ DCs on day 3 post-infection viral titers in lungs of challenged mice on day 2 post-infection with hvPR8-ΔNS1 than Mx1-Ifnlr1–/– mice, whereas the number of (Supplementary Fig. 7). The observed enhanced influenza virus CD103- CD11b+ cells remained unchanged (Fig. 6b). These results resistance was not due to long-lasting effects of IFN-λ on innate indicated that IFN-λ-induced TSLP boosts immunity by acting on immunity, because animals immunized with IFN-λ-enriched migratory DCs. In agreement with earlier observations in TSLP- mock vaccine lacking viral antigen remained susceptible to treated mice25, we speculate that the increased presence of activated influenza virus infection (Fig. 7c). CD103+ DCs in lymph nodes of mice immunized with IFN-λ- Influenza virus strain A/Udorn/72 is transmissible among mice adjuvanted vaccines triggers stronger antigen-specific germinal on contact34. We used this virus to determine whether enhanced center reactions and stronger antiviral CD8+ T cell responses com- IFN-λ-mediated production of M2e-specific antibodies can reduce pared with mice receiving non-adjuvanted vaccines. virus shedding and transmission from immunized BALB/c to naive DBA/2J mice35. Mice immunized with the M2e vaccine in the pres- Mucosal immunization in the presence of IFN-λ results in ence of IFN-λ had approximately 35-fold lower virus titers in nasal enhanced protection from influenza virus-induced disease. swabs at 48 hours post-infection compared with mice vaccinated We performed a series of influenza virus challenge experiments with the M2e vaccine alone (Fig. 8a). Nine of 12 mice immunized to assess the biological consequences of the observed IFN-λ- with the IFN-λ-free M2e vaccine transmitted the virus to naive mediated vaccine-enhancing effects. Wild type mice immunized DBA/2J animals on co-housing, compared with 1 of 10 mice immu- with IFN-λ-free HA vaccine rapidly lost weight after infection nized with the IFN-λ-complemented M2e vaccine (Fig. 8b). Mice with influenza virus strain PR8/34, and 80% of the infected ani- immunized with M2e vaccine containing TSLP also transmitted mals had to be killed due to severe disease symptoms (Fig. 7a). the virus less frequently to naive contacts compared with mice In contrast, all mice immunized with IFN-λ2-complemented immunized with M2e alone (Supplementary Fig. 8). Taken together, HA vaccine survived this challenge infection with minimal body these results clearly demonstrated that IFN-λ can enhance vaccine- weight loss (Fig. 7a). induced antiviral immunity.

598 Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology NATure IMMunoLoGy Articles

a IgG1IgA 6 1.0 100% WT

) **

10 **** –/– 5 *** 0.8 **** 100% Tslpr 4 **** *** *** 80% Rag1 –/– + 20% Tslpr –/– 0.6 **** **** –/– 3 **** 80% Rag1 + 20% WT 0.4 –/– 2 100% Batf3 –/– –/– 1 0.2 50% Batf3 + 50% Tslpr HA-specific IgA (OD) HA-specific IgG1(log 50% Batf3 –/– + 50% WT 0 0.0

b WT Ifnlr1 –/– 12 80 105 WT (% ) 6.54 2.75 (% ) – 4 ** – NS 10 9 60 Ifnlr1 –/–

3 10 45.4 44.5 6 40 CD11b CD103 + + 2 10 3 20 0 CD103-PE CD103 01103 104 105 0 103 104 05 0 CD11b 0 CD11b-PE-cy7

Fig. 6 | IFN-λ-triggered TSLP acts on CD103+ migratory DCs. a, Lethally irradiated WT mice were grafted with BM cells from the listed mouse strains. Where indicated, mixtures of BM cells from different strains were used. Seven weeks later, the animals were immunized by infection with hvPR8-ΔNS1. Sera and BAL fluids were collected on day 14 post-infection and HA-specific IgG1 and IgA titers were measured by ELISA. Data are shown as mean ± s.e.m. ****P < 0.0001, ***P = 0.0005, ****P < 0.0001, **P = 0.0051, ***P = 0.0005 (left panel, from left to right). Right panel (from left to right), ***P = 0.0001, ****P < 0.0001, ****P < 0.0001, ****P < 0.0001, ****P < 0.0001 (right panel, from left to right), by unpaired two-tailed Student’s t-test. b, Mx1-WT (n = 5) and Mx1-Ifnlr1–/– (n = 5) mice were infected with hvPR8-ΔNS1 before the frequencies of CD103+ CD11b− DCs and CD11b+ CD103− DCs among CD11c+ MHC-II+ DCs were determined at day 3 post-infection. Results are representative of two independent experiments and shown as mean ± s.e.m. Each symbol represents the result from an individual animal. NS, no significant difference. **P = 0.0047 (left), P = 0.1098 (right), by unpaired two-tailed Student’s t-test.

a 110 100 HA

100 *** 80 HA+IFN-λ 90 60 80 40

70 Survival (%) 20 Body weight (%) 60 02468101214 0 2468101214 Days post-infection Days post-infection b 110 100 PBS 100 80 M2e 90 60 M2e+IFN-λ **** 80 40 Survival (%)

Body weight (%) 70 20 60 0246810 12 14 02468101214 Days post-infection Days post-infection

c 110 100 PBS 100 80 IFN-λ 90 60 80 40

70 Survival (% ) 20 Body weight (% ) 60 02468101214 0 2468101214 Days post-infection Days post-infection

Fig. 7 | Mucosal immunization in the presence of IFN-λ results in enhanced protection from influenza virus-induced disease. a, WT mice were immunized by intranasal application of Influsplit Tetra containing (n = 9) or lacking (n = 10) IFN-λ2. Two weeks after the third booster immunization, the animals were infected with PR8/34. Weight loss and survival were monitored daily for 14 days. Data were pooled from two independent experiments. ***P = 0.0005, by log-rank test. b, Mx1-WT mice were intranasally immunized with M2e vaccine in the presence (n = 14) or absence (n = 15) of IFN-λ2. Ten days after a single booster immunization, the animals were challenged with hvPR8. Weight loss and survival were monitored daily for 14 days. Mock- immunized mice (PBS) served as additional controls (n = 10). Data are pooled from two independent experiments. ****P < 0.0001, by log-rank test. c, Mx1-WT mice (n = 5 per group) were mock-immunized in the presence or absence of IFN-λ2. Ten days after mock-booster immunization, the animals were challenged with hvPR8. Weight loss and survival were monitored daily. a–c, Symbols represent mean ± s.e.m. centered on the mean (left).

Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology 599 Articles NATure IMMunoLoGy

a b ** 5 **** 8 PBS 100% 75% 10% M2e 4 6 M2e+IFN-λ 3 4

PFU per ml 2 PFU per ml 10 1 10 2 lo g lo g 0 0 16 24 48 72 Index Index Index Hours post-infection Contact Contact Contact

PBSM2e M2e+IFN-λ

Fig. 8 | IFN-λ-adjuvanted M2e vaccine reduces influenza virus transmission. BALB/c mice were intranasally immunized with M2e vaccine in the presence (n = 5) or absence (n = 6) of IFN-λ2. Mock-immunized animals (PBS) served as additional controls (n = 6). Eight weeks after a single booster immunization, the animals were infected with A/Udorn. Twenty-four hours later, each infected BALB/c mouse was co-housed with one naive DBA/2 J mouse for 4 days. a, Virus excretion was monitored by measuring infectious virus in nasal swabs of infected BALB/c mice at the indicated times post- infection. ****P < 0.0001, by two-way ANOVA with Tukey’s multiple-comparison test. b, Virus transmission to contact mice was assessed by measuring infectious virus in the snouts of immunized BALB/c (Index) and exposed DBA/2 J mice (Contact) on day 4 post co-housing. Data are shown as mean ± s.e.m. The dotted line indicates the detection limit of the assay. **P = 0.0037 by Fisher’s exact test. Each symbol represents an individual animal.

Discussion IFN-λ-induced TSLP stimulates the migration of DCs from the respi- Our work revealed that IFN-λ can regulate adaptive immune pro- ratory tract to the draining nymph nodes. Accordingly, Tfh cells and cesses in the mucosa of the respiratory tract, thereby greatly enhanc- germinal B cells were more abundant in draining lymph nodes and in ing the production of virus-specific CD8+ T cells and antibodies, the spleen of immunized mice possessing an intact IFN-λ–TSLP axis resulting in improved resistance to infection with influenza viruses. compared with Tslpr–/– mice. Thus, activation of the IFN-λ–TSLP axis This activity is mechanistically distinct from the well-documented results in increased antigen presentation by CD103+ DCs and, conse- direct antiviral activity of IFN-λ on epithelial cells of the airways quently, improved adaptive antiviral immune responses. and the intestinal tract12,13,36. Thus, IFN-λ plays decisive roles in Our data demonstrate that IFN-λ has strong adjuvant activity both the innate and the adaptive immune response against viruses when used in conjunction with influenza virus subunit vaccines. that target the respiratory mucosa. This adjuvant activity of IFN-λ was strictly dependent on TSLP and Since most immune cells do not express the IFN-λ receptor2–4, our could be mimicked by applying exogenous TSLP. Importantly, the findings are surprising and counterintuitive at first glance. However, immune-enhancing effect of IFN-λ was observed only when the we found that the observed regulatory effects of IFN-λ on adaptive vaccines were applied by the intranasal route, but not by the subcu- immunity do not result from IFN-λ acting directly on immune cells. taneous or intraperitoneal routes. This selectivity can be explained IFN-λ rather acts indirectly and involves TSLP, a known regulator by assuming that the TSLP-inducing activity of IFN-λ is limited to of adaptive immunity, which is produced by upper airway M cells M cells and possibly additional specialized epithelial cells in the air- in response to exposure to IFN-λ. This previously unknown IFN-λ– ways of mice that express the IFN-λ receptor. TSLP axis represents a remarkable example of a well-orchestrated We found that the adjuvant effect of IFN-λ not only resulted in cross-talk between epithelial cells and the immune system which enhanced antibody titers and CD8+ T cell responses against viral helps fine-tuning of antiviral adaptive immune responses. Thus, our antigens, but also significantly improved protection against influenza study demonstrates that IFN-λ can act as an endogenous adjuvant virus challenge infections. Two previous studies indicated that IFN-λ that dramatically enhances mucosal immune responses. can enhance protective CD8+ T cell responses after DNA vaccination Our study revealed that the ability to regulate adaptive immune by modulating regulatory T cell activity15,16. Since the DNA vaccines responses via TSLP represents a unique property of IFN-λ that can- were applied intramuscularly, it remains unclear how IFN-λ could not be mimicked or compensated for by type I IFN. We found that be beneficial in these cases and whether TSLP is involved. Be that Ifnlr1–/– mice, which possess a fully functional type I IFN system19, as it may, our work confirms and extends previous findings which are unable to produce normal amounts of virus-specific IgG1 and showed that TSLP is a potent mucosal adjuvant22. Our observations IgA antibodies as well as antiviral CD8+ T cells after infection with in mice suggest that formulating mucosal vaccines with either IFN-λ hvPR8-ΔNS1, although this live-attenuated virus vaccine induces or TSLP might greatly improve their protective efficacy. high levels of type I IFN in the respiratory tract of mice18. In agree- Our observations in mice are in conflict with a study which indi- ment with this observation, airway M cells of Ifnlr1–/– mice infected cated that IFN-λ can inhibit rather than promote influenza virus with hvPR8-ΔNS1 produced less TSLP than M cells from wild type vaccine-induced antibody synthesis in humans14. This study showed mice. We currently do not understand why type I IFN is unable to that antibody production in response to intramuscular application of promote TSLP synthesis in upper airway M cells. In human airway standard influenza vaccines is influenced by genetic traits that con- epithelial cells, the Tslp gene is expressed in response to infection trol IFN-λ gene expression. Studies with isolated human lymphocytes with respiratory syncytial virus in a RIG-I helicase-dependent fash- further indicated that IFN-λ can suppress vaccine-induced B cell pro- ion37. The promoter region of the Tslp gene contains several NF-κB liferation and antigen-specific IgG production14, presumably by trig- binding sites but it lacks typical binding sites for known IFN- gering IFN-λ receptors on B cells39. However, a second study found regulated transcription factors38, suggesting that induction of Tslp that IFN-λ stimulates rather than inhibits humoral immune responses gene expression by IFN-λ might involve a non-canonical pathway. in humans by augmenting TLR-mediated activation of B cells7. The The TSLP receptor is expressed by many different immune reasons for these discrepant findings remain unclear. Since mouse B cells23,24. Using mixed bone marrow chimeric mice we demonstrated cells cannot respond to IFN-λ, any regulatory effects of IFN-λ in mice that IFN-λ-induced TSLP must act on CD103+ migratory DCs for are expected to be indirect. It is currently unknown whether IFN-λ the immune-enhancing effect to occur. In Ifnrl1–/– mice infected with can stimulate vaccine-induced antibody production in the upper air- hvPR8-ΔNS1, fewer CD103+ migratory DCs were present in the way of humans via TSLP as observed here for mice. Vaccination trials draining lymph nodes compared with wild type mice, suggesting that in human volunteers are probably required to solve this issue.

600 Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology NATure IMMunoLoGy Articles

Taken together, our work in the mouse model demonstrated that 23. Chappaz, S., Flueck, L., Farr, A. G., Rolink, A. G. & Finke, D. Increased TSLP IFN- plays a decisive role in both innate and adaptive immune availability restores T- and B-cell compartments in adult IL-7 defcient mice. λ Blood 110, 3862–3870 (2007). responses against viruses that target the respiratory mucosa. We 24. Ziegler, S. F. & Artis, D. Sensing the outside world: TSLP regulates barrier found that the enhancing effect of IFN-λ on the adaptive immune immunity. Nat. Immunol. 11, 289–293 (2010). system is indirect and involves IFN-λ-triggered synthesis of TSLP 25. Joo, S. et al. Critical role of TSLP-responsive mucosal dendritic cells in the that acts on migratory DCs. This TSLP-dependent adjuvant activity induction of nasal antigen-specifc IgA response. Mucosal Immunol. 10, of IFN- might be employed to improve mucosal vaccine efficacy in 901–911 (2017). λ 26. Soumelis, V. et al. Human epithelial cells trigger dendritic cell mediated humans and livestock. allergic infammation by producing TSLP. Nat. Immunol. 3, 673–680 (2002). 27. Kim, D. Y. et al. Te airway antigen sampling system: respiratory M Online content cells as an alternative gateway for inhaled antigens. J. Immunol. 186, Any methods, additional references, Nature Research reporting 4253–4262 (2011). summaries, source data, statements of data availability and asso- 28. Vinuesa, C. G., Linterman, M. A., Yu, D. & MacLennan, I. C. M. Follicular helper T cells. Annu. Rev. Immunol. 34, 335–368 (2016). ciated accession codes are available at https://doi.org/10.1038/ 29. Fazilleau, N., Mark, L., McHeyzer-Williams, L. J. & McHeyzer-Williams, M. G. s41590-019-0345-x. Follicular helper T cells: lineage and location. Immunity 30, 324–335 (2009). 30. Grajales-Reyes, G. E. et al. Batf3 maintains autoactivation of Irf8 for Received: 4 September 2018; Accepted: 5 February 2019; commitment of a CD8alpha(+) conventional DC clonogenic progenitor. Published online: 18 March 2019 Nat. Immunol. 16, 708–717 (2015). 31. Takaki, H. et al. Toll-like receptor 3 in nasal CD103(+) dendritic cells is References involved in immunoglobulin A production. Mucosal Immunol. 11, 82–96 (2018). 1. Crotta, S. et al. Type I and type III interferons drive redundant amplifcation 32. Van den Hoecke, S. et al. Hierarchical and redundant roles of activating loops to induce a transcriptional signature in infuenza-infected airway FcgammaRs in protection against infuenza disease by M2e-Specifc IgG1 epithelia. PLoS Pathog. 9, e1003773 (2013). and IgG2a antibodies. J. Virol. 91, e02500–e02516 (2017). 2. Wack, A., Terczynska-Dyla, E. & Hartmann, R. Guarding the frontiers: the 33. Grimm, D. et al. Replication ftness determines high virulence of infuenza biology of type III interferons. Nat. Immunol. 16, 802–809 (2015). A virus in mice carrying functional Mx1 resistance gene. Proc. Natl Acad. Sci. 3. Lazear, H. M., Nice, T. J. & Diamond, M. S. Interferon-lambda: immune USA 104, 6806–6811 (2007). functions at barrier surfaces and beyond. Immunity 43, 15–28 (2015). 34. Klinkhammer, J. et al. IFN-lambda prevents infuenza virus spread from the 4. Sommereyns, C., Paul, S., Staeheli, P. & Michiels, T. IFN-lambda (IFN- upper airways to the lungs and limits virus transmission. Elife 7, e33354 (2018). lambda) is expressed in a tissue-dependent fashion and primarily acts on 35. Kolpe, A., Schepens, B., Ye, L., Staeheli, P. & Saelens, X. Passively transferred epithelial cells in vivo. PLoS Pathog. 4, e1000017 (2008). M2e-specifc monoclonal antibody reduces infuenza A virus transmission in 5. Blazek, K. et al. IFN-lambda resolves infammation via suppression of mice. Antiviral Res. 158, 244–254 (2018). neutrophil infltration and IL-1beta production. J. Exp. Med. 212, 845–853 (2015). 36. Mordstein, M. et al. Interferon-lambda contributes to innate immunity 6. Broggi, A., Tan, Y., Granucci, F. & Zanoni, I. IFN-lambda suppresses of mice against infuenza A virus but not against hepatotropic viruses. intestinal infammation by non-translational regulation of neutrophil PLoS Pathog. 4, e1000151 (2008). function. Nat. Immunol. 18, 1084–1093 (2017). 37. Lee, H. C. et al. Tymic stromal lymphopoietin is induced by respiratory 7. de Groen, R. A., Groothuismink, Z. M., Liu, B. S. & Boonstra, A. IFN-lambda syncytial virus-infected airway epithelial cells and promotes a type 2 response is able to augment TLR-mediated activation and subsequent function of to infection. J. Allergy Clin. Immunol. 130, 1187–1196.e1185 (2012). primary human B cells. J. Leukoc. Biol. 98, 623–630 (2015). 38. Ganti, K. P., Mukherji, A., Surjit, M., Li, M. & Chambon, P. Similarities and 8. Espinoza, V. et al. Type III interferon is a critical regulator of innate diferences in the transcriptional control of expression of the mouse TSLP antifungal immunity. Sci. Immunol. 2, eaan5357 (2017). gene in skin epidermis and intestinal epithelium. Proc. Natl Acad. Sci. USA 9. Galani, I. E. et al. Interferon-lambda mediates non-redundant front-line 114, E951–E960 (2017). antiviral protection against infuenza virus infection without compromising 39. Kelly, A. et al. Immune cell profling of IFN-lambda response shows pDCs host ftness. Immunity 46, 875–890 e876 (2017). express highest level of IFN-lambdaR1 and are directly responsive via the 10. Mordstein, M. et al. Lambda interferon renders epithelial cells of the JAK-STAT pathway. J. Interferon Cytokine Res. 36, 671–680 (2016). respiratory and gastrointestinal tracts resistant to viral infections. J. Virol. 84, 5670–5677 (2010). Acknowledgements 11. Hernandez, P. P. et al. Interferon-lambda and interleukin 22 act synergistically We thank A. Ohnemus for technical support, and H. Pircher, G. Gasteiger and O. Haller for the induction of interferon-stimulated genes and control of rotavirus for helpful discussions and comments on the manuscript. Funding for this work was infection. Nat. Immunol. 16, 698–707 (2015). provided by the European Union’s Seventh Framework Program grant agreement 607690 12. Mahlakoiv, T., Hernandez, P., Gronke, K., Diefenbach, A. & Staeheli, P. (to P.S. and N.L.), the Deutsche Forschungsgemeinschaft grant agreement STA 338/15-1 Leukocyte-derived IFN-alpha/beta and epithelial IFN-lambda constitute a (to P.S.) and TA 436/4-1 (to Y.T.), the Else Kröner-Fresenius Stiftung grant agreement compartmentalized mucosal defense system that restricts enteric virus 2017_EKES.34 (to Y.T.) and the Danish Council for Independent Research, Medical infections. PLoS Pathog. 11, e1004782 (2015). Research grant agreement 11‐107588 (to R.H.). 13. Pott, J. et al. IFN-lambda determines the intestinal epithelial antiviral host defense. Proc. Natl Acad. Sci. USA 108, 7944–7949 (2011). Author contributions 14. Egli, A. et al. IL-28B is a key regulator of B- and T-cell vaccine responses L.Y. designed and performed most of the experiments, analyzed and interpreted the against infuenza. PLoS Pathog. 10, e1004556 (2014). data. D.S. designed and performed experiments, analyzed and interpreted the data. J.B. 15. Morrow, M. P. et al. Comparative ability of IL-12 and IL-28B to regulate Treg performed experiments and analyzed samples. K.E. and D.S. generated bone marrow populations and enhance adaptive cellular immunity. Blood 113, 5868–5877 (2009). chimeric mice. V.B. and H.H.G. provided essential materials. Y.T., R.H. and N.L. analyzed 16. Zhou, Y. et al. Optimized DNA vaccine enhanced by adjuvant IL28B induces and interpreted data, and gave advice. P.S. conceived the project, acquired funding for protective immune responses against herpes simplex virus type 2 in mice. the study, designed experiments and interpreted the data. P.S., L.Y. and D.S. wrote the Viral Immunol. 30, 601–614 (2017). manuscript with input from all authors. 17. Ferko, B. et al. Immunogenicity and protection efcacy of replication-defcient infuenza A viruses with altered NS1 genes. J. Virol. 78, 13037–13045 (2004). 18. Kochs, G. et al. Strong interferon-inducing capacity of a highly virulent Competing interests variant of infuenza A virus strain PR8 with deletions in the NS1 gene. J. Gen. The authors declare no competing interests. Virol. 90, 2990–2994 (2009). 19. Ank, N. et al. An important role for type III interferon (IFN-lambda/IL-28) in TLR-induced antiviral activity. J. Immunol. 180, 2474–2485 (2008). Additional information 20. Dellgren, C., Gad, H. H., Hamming, O. J., Melchjorsen, J. & Hartmann, R. Supplementary information is available for this paper at https://doi.org/10.1038/ Human interferon-lambda3 is a potent member of the type III interferon s41590-019-0345-x. family. Genes Immun. 10, 125–131 (2009). Reprints and permissions information is available at www.nature.com/reprints. 21. Eliasson, D. G. et al. CTA1-M2e-DD: a novel mucosal adjuvant targeted infuenza vaccine. Vaccine 26, 1243–1252 (2008). Correspondence and requests for materials should be addressed to P.S. 22. Van Roey, G. A., Arias, M. A., Tregoning, J. S., Rowe, G. & Shattock, R. J. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in Tymic stromal lymphopoietin (TSLP) acts as a potent mucosal adjuvant for published maps and institutional affiliations. HIV-1 gp140 vaccination in mice. Eur. J. Immunol. 42, 353–363 (2012). © The Author(s), under exclusive licence to Springer Nature America, Inc. 2019

Nature Immunology | VOL 20 | MAY 2019 | 593–601 | www.nature.com/natureimmunology 601 Articles NATure IMMunoLoGy

Methods dilutions resulting in OD values which were two- (for HA-specific antibody) or Mice. C57BL/6 (designated WT), BALB/c and DBA/2J mice were purchased from fourfold (for M2e-specific antibody) above background. Janvier Laboratories. B6.A2G-Mx1 mice carrying intact Mx1 alleles (designated Mx1-WT), congenic B6.A2G-Mx1-Ifnαr1–/– mice lacking functional IFN-α Collection of BAL fluids. Bronchoalveolar lavage was performed by inoculating receptors (designated Mx1-Ifnαr1–/–) and B6.A2G-Mx1-Ifnlr1–/– mice lacking 500 μl of PBS into the lungs via the trachea using a peripheral venous catheter (BD functional IFN-λ receptors (designated Mx1-Ifnlr1–/–) were described before10. Venflon), and flushing the tissue twice. BAL fluids were collected and centrifuged Tslpr–/– mice40 were provided by D. Finke and Batf3–/– mice30 were provided by at 1,500 r.p.m. for 5 min. The BAL fluid supernatants were stored at −80°C and W. Kastenmüller. Rag1–/– mice were maintained locally. With few exceptions, all further used to analyze IgA levels. BAL cells were used for analyzing influenza experiments were also performed with female mice. In the exceptional cases when virus nucleoprotein-specific CD8+ T cells by FACS. male mice were included, the sex ratios in the control groups were kept similar. Mice were bred and kept under specifc-pathogen-free conditions in the local Isolation of snout epithelial cells. Mice were treated intranasally with 2 µg IFN-λ2 5 animal facility. All experiments with mice were carried out in accordance with the in a volume of 20 µl or infected with 10 PFU of hvPR8-ΔNS1 in a volume of 20 µl. guidelines of the Federation for Laboratory Animal Science Associations and the Approximately 24 h later, animals were killed and the snouts were collected. The national animal welfare body. Experiments were in compliance with the German tissue was mechanically disrupted and resulting tissue fragments were digested 1 animal protection law and were approved by the animal welfare committee of the with five cycles of trypsin treatment (1 mg ml− ) at 37 °C on a shaker for 7–10 min Regierungspräsidium Freiburg (permits G-16/177 and G-17/74). each. Released cells were transferred to DMEM medium (Gibco, Life Technologies) containing 10% FCS and were then passed through a 70-µm cell strainer. Cells Immunizations. Mice carrying functional Mx1 alleles were immunized by were washed in FACS buffer (1% BSA, 0.1% NaN3 in PBS) and used for antibody intranasal infection with 105 plaque-forming units (PFU) of live-attenuated staining. influenza virus strain hvPR8-ΔNS1 (ref. 18) in a volume of 40 µl. Mice lacking functional Mx1 alleles were immunized by intranasal infection with 5 × 104 PFU of Primary mouse tracheal epithelial cell culture. The preparation of primary 34 hvPR8-ΔNS1. Titers of HA-specific IgG subtypes were measured by ELISA on days mouse airway epithelial cells was performed as previously described . Briefly, 7, 14 and 21 post-infection. cells were isolated from the tracheae of Mx1-WT mice by enzymatic treatment For intranasal immunization with Influsplit Tetra vaccine (GSK), mice were and seeded onto a 0.4-μm pore size clear polyester membrane (Corning) coated anesthetized with a ketamine/xylazine mixture and treated with 30 µl of diluted with a collagen solution. At confluency, the medium was removed from the upper vaccine (containing 1 µg of HA derived from the H1N1 influenza virus strain A/ chamber to establish an air–liquid interface. Fully differentiated, 7–10-day-old −1 Michigan/45/2015) in the presence or absence of either 1 µg IFN-λ2 (ref. 20) or post-air–liquid interface cultures were treated with 100 ng ml of IFN-λ2 for 24 h 1 µg TSLP. Booster immunizations were performed 10 and 20 days later. In some before staining and analysis by flow cytometry. experiments, mice were immunized by the intraperitoneal or the subcutaneous routes with 1 µg of Influsplit Tetra in the presence or absence of 1 µg IFN-λ2. Flow cytometry. Single cell suspensions of spleen and lymph nodes were generated Booster immunizations were performed 10 and 20 days later. Blood samples were by mechanical dissociation. Single cell suspensions from snouts (prepared collected on day 10 after the first and second booster immunization for analysis of according to the protocol above) were treated with ACK lysis buffer (0.15 mM HA-specific antibodies by ELISA. BAL fluid samples were collected at day 10 after NH4Cl, 10 mM KHCO3, 0.1 mM EDTA–Na2 in Milli-Q-H2O) to remove red the second booster immunization for analysis of HA-specific IgA titers. blood cells. Cells were suspended in FACS buffer (1% BSA, 0.1% NaN3 in PBS) For intranasal immunization with CTA1-3M2e-DD (M2e), mice were and incubated with anti-mouse CD16/32 (eBiosciences) for 20 min on ice to anesthetized with a ketamine/xylazine mixture and treated with 1 µg of M2e21 in block non-specific antibody binding. For surface staining, cells were incubated the presence or absence of 1 µg IFN-λ2 or 1 µg TSLP (555-TS, R&D Systems) in with cocktails of anti-mouse and LIVE/DEAD fixable near-IR dead cell dye a total volume of 30 µl. A booster immunization was performed ten days later. (Invitrogen) for 30 min on ice. For intracellular staining, cells were fixed and Ten days after boosting, blood and BAL fluids were collected for analysis of M2e- permeabilized with Cytofix/Cytoperm (BD Biosciences, 554714) according to specific antibodies by ELISA. In some experiments, mice were immunized via the the manufacturer’s instructions and subsequently incubated with anti-mouse intraperitoneal or the subcutaneous routes with 1 µg of M2e in the presence or TSLP (BAF555, R&D Systems) or IgG1 (A85-1, BD Biosciences). PE-conjugated absence of 1 µg IFN-λ2 in a total volume of 200 µl. streptavidin (BioLegend) was used as a secondary antibody for detecting TSLP. To detect phosphorylated STAT1, spleen cells were collected and fixed with fix buffer Treatment of mice with TSLP-neutralizing antibody. Mx1-WT mice were I (557870, BD Biosciences) for 10 min. After washing, cells were permeabilized intranasally infected with of 105 PFU of hvPR8-ΔNS1 in the presence of 25 µg of with perm buffer III (558050, BD Biosciences) for 30 min and then stained with a TSLP-specific monoclonal antibody (MAB555, R&D Systems) or 25 µg of anti- p-STAT1 antibody (562069, BD Biosciences) for 1 h at room temperature. Finally, IgG isotype control (MAB006, R&D Systems). On day 2 post-infection, animals cells were passed through a 70-µm cell strainer. Data were acquired on a BD were treated with the same amounts of anti-TSLP or IgG isotype control by the LSRFortessa cell analyzer (BD Biosciences) and analyzed with FlowJo software intranasal route. Titers of virus-specific serum IgG1 were determined by ELISA on (TreeStar). days 14 and 21 post-infection. Reagents for flow cytometry. The following fluorescence-labeled antibodies Bone marrow chimeras. To generate IFN-λ receptor-chimeric mice, Mx1-WT and reactive to mouse antigens were used: anti-CD4 (RM4-5, BioLegend), anti- Mx1-Ifnlr1–/– recipient mice were lethally irradiated (2 × 5.5 Gy, 4 h interval) and CD19 (1D3, BioLegend), anti-B220 (RA3-6B2, eBioscience), anti-CD16/32 (93, reconstituted with 107 donor BM cells from Mx1-WT or Mx1-Ifnlr1–/– donor mice, BioLegend), anti-CD45 (30-F11, BioLegend), anti-EpCAM (G8.8, BioLegend), respectively. Eight weeks post-transplant, mice were immunized by intranasal anti-NKM16-2-4 (NKM16-2-4, Miltenyi Biotec), anti-PD1 (29F.1A12, BioLegend), infection with 105 PFU of hvPR8-ΔNS1 in a total volume of 40 µl. Sera were anti-CXCR5 (L138D7, BioLegend), anti-CD44 (IM7, BioLegend), anti-FAS collected at 21 days post-infection. (SA367H8, BioLegend), anti-GL7 (GL7, BioLegend), anti-TSLP (BAF555, To generate mixed chimeric mice containing TSLP receptor-deficient T and R&D Systems), anti-IgG1 (A85-1, BD Biosciences), anti-p-STAT1 (562069, BD B cells, WT recipient mice were lethally irradiated (2 × 5.5 Gy, 4 h interval) and Biosciences), anti-CD8a (53-6.7, BioLegend), anti-CD11c (N418, BioLegend), reconstituted with mixtures (107 cells in total) of BM cells from Rag1–/– (80%) and anti-CD103 (2E7, BioLegend), anti-CD11b (M1/70, Biolegend), anti-MHC-II Tslpr–/– (20%) mice. To generate mixed chimeric mice containing TSLP receptor- (M5/114.15.2, Biolegend) and anti-CD3 (17A2, BioLegend). Fluorescent-labeled b deficient migratory DCs, lethally irradiated WT recipient mice were reconstituted influenza virus nucleoprotein-specific pentamer (H-2D /ASNENMETM) was with mixtures (107 cells in total) of BM cells from Batf3–/– (50%) and Tslpr–/– (50%) purchased from Proimmune. mice. Mice were allowed to reconstitute for seven weeks before immunization by intranasal infection with 5 × 104 PFU of hvPR8-ΔNS1. Sera and BAL fluids were Cell sorting and RT–qPCR. For sorting M+ cells, M− cells and CD45+ leukocytes, collected on day 14 post-infection. snout tissues from IFN-λ2-treated or control Mx1-WT mice were cut into small pieces and subjected to five cycles of trypsin digestion (1 mg ml−1) at 37°C for Measuring antigen-specific antibodies. M2e-specific and HA-specific antibodies 7–10 min each. Released cells were transferred to DMEM medium (Gibco, in serum and BAL fluids from individual mice were determined by ELISA. To do Life Technologies) containing 10% FCS and then passed through a 70-µm cell −1 this, microtiter plates (MaxiSorp, Nunc) were coated with M2e peptide (2 µg ml , strainer. Cells were washed in FACS buffer (1% BSA, 0.1% NaN3 in PBS), and GenScript) or inactivated PR8 virus at 4°C overnight. After washing four times then incubated with anti-CD16/CD32 (93, BioLegend) to block Fc receptors. with 0.1% Tween-20, the plates were blocked for 2 h with 5% BSA in PBS. Diluted After washing, cells were further incubated with fluorescence-conjugated anti- serum samples were added and incubated at room temperature for 1 h. Bound NKM16-2-4 (NKM16-2-4, Miltenyi Biotec), anti-EpCAM (G8.8, BioLegend) antibodies were detected using horseradish peroxidase-labeled antibodies directed and anti-CD45 (30-F11, BioLegend) for 30 min on ice. After washing and against either total IgG (62-6520, Invitrogen), IgG1 (A10551, Invitrogen), IgG2c passeing through a 70-µm cell strainer, M+ cells (CD45− EpCAM+ NKM16-2- (1078-05, Southern Biotech), IgG2b (M32407, Invitrogen) or IgA (626720, 4+), M− cells (CD45− EpCAM+ NKM16-2-4−) and leukocytes (CD45+ EpCAM−) Invitrogen), followed by incubation with tetramethylbenzidine substrate (Life were sorted (to purity of >93%, >98% and >99%, respectively) with a BD Technologies) for 10–30 min. The reaction was stopped by adding 0.5 M H2SO4 and FACSAria Fusion cell sorter (BD Biosciences). RNA was extracted with RNeasy absorbance was measured at 450 nm. The endpoint titers were defined as highest Plus Mini Kit (Qiagen) according to the manufacturer’s instructions and then

Nature Immunology | www.nature.com/natureimmunology NATure IMMunoLoGy Articles reverse-transcribed into complementary DNA using the Thermoscript RT–PCR for 20 s each. Snout and lung samples were then centrifuged for 10 min at system, following the manufacturer’s instructions (Invitrogen). RT–qPCR was 9,000 r.p.m. Resulting supernatants and nasal swab eluates were subjected to performed on an ABI-Prism 7900 sequence detection system with SYBR Green serial tenfold dilutions in OptiMEM (Gibco, Life Technologies) supplemented Master mix (Applied Biosystems) or TaqMan Universal Master Mix using with 0.3% BSA (Sigma). Samples (450 μl) were placed on confluent MDCK primers against Tslp (forward primer: 5′-TTCACTCCCCGACAAAACAT-3′, cells in 12-well plates, and incubated for 1–2 h at room temperature before reverse primer: 5′-TGTGAGGTTTGATTCAGGCA-3′, and probe: the inoculum was removed and the cells were overlaid with 1.5% Avicel (FMC 5′-[6FAM]TGACCACTGCCCAGGCTACCCT[BHQ1]-3′), Mx1 BioPolymer) in DMEM (Gibco, Life Technologies) supplemented with 10% BSA (forward primer: 5′-CTGAGATGACCCAGCACCTG-3′, reverse (Sigma), glutamine and penicillin–streptomycin for 72 h. Cells were then fixed primer: 5′-GCTGCACTTACTGGTGTCCT-3′, and probe: 5′-[6FAM] with 3.7% formaldehyde and stained with 1% crystal violet for at least 15 min. AGCCTACTACCAGGAGT[BHQ1]-3′), and β-Actin (Qiagen, catalog no. Plaques were counted and used to calculate virus titers, defined as PFU per ml QT00095242). The relative expression of Tslp and Mx1 genes was calculated by of tissue lysate. the 2-ΔCt method relative to the expression of housekeeping gene β-actin. Statistical analysis. Statistical analyses were performed using GraphPad Prism Influenza virus challenge infections. To evaluate antiviral protection induced by 6 software. Data are represented as mean and s.e.m. Statistical significance was the Influsplit Tetra, vaccinated WT mice were challenged by intranasal infection determined by unpaired two-tailed Student’s t-test, log-rank test, Fisher’s exact with 3 × 103 PFU of influenza virus strain PR8/34 in a volume of 40 µl. To evaluate test, one-way analysis of variance (ANOVA) with Tukey’s multiple-comparison antiviral protection induced by the M2e vaccine, immunized Mx1-WT mice were test, or two-way ANOVA with Tukey’s multiple-comparison test, as specified in the challenged with 103 PFU (in 40 µl volume) of a highly virulent variant of influenza relevant figure legends. virus strain PR8/34 (designated hvPR8)33 by the intranasal route. Mice were monitored for weight loss during the following 14 days and were killed when the Reporting Summary. Further information on research design is available in the body weight dropped below 75% of the starting value. Nature Research Reporting Summary linked to this article. To evaluate the effect of immunization on virus transmission, BALB/c mice were immunized with 1 μg of the M2e vaccine in the presence or absence of 1 µg IFN-λ2 or 1 µg TSLP by the intranasal route for three times, 10 days apart. Eight weeks after Data availability the second booster immunization, mice were challenged by intranasal infection with The data that support the findings of this study are available from the 3 × 104 PFU of influenza virus strain A/Udorn/72 (H3N2) in a total of 20 µl. At 24 h corresponding author upon request. post-infection, each of these animals (designated index) was co-housed individually with a naive DBA/2J (designated contact) mouse for four days35. Viral titers in nasal swabs of index mice and snouts of contact mice were determined by plaque assay. References Viral titrations. Snouts and lungs were homogenized in 800 μl cold PBS using 40. West, E. E. et al. A TSLP-complement axis mediates neutrophil killing of FastPrep spheres and homogenizer (MP Biomedicals), three cycles at 6.5 m s−1 methicillin-resistant Staphylococcus aureus. Sci. Immunol. 1, eaaf8471 (2016).

Nature Immunology | www.nature.com/natureimmunology nature research | reporting summary

Corresponding author(s): Peter Staeheli

Last updated by author(s): Jan 22, 2019 Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Research policies, see Authors & Referees and the Editorial Policy Checklist.

Statistics For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section. n/a Confirmed The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one- or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section. A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)

For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable.

For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated

Our web collection on statistics for biologists contains articles on many of the points above. Software and code Policy information about availability of computer code Data collection BD LSRFortessa (BD Biosciences), BD FACSAria™ (BD Biosciences) and ABI-Prism 7900 (Thermo Fisher Scientific).

Data analysis FlowJo (Treestar Inc., VX), Prism (GraphPad Inc., V6) and SDSv2.4 (Applied Biosystems) For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information. Data Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: - Accession codes, unique identifiers, or web links for publicly available datasets - A list of figures that have associated raw data - A description of any restrictions on data availability

Data are available from the corresponding author upon reasonable request. October 2018

Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

1 nature research | reporting summary Life sciences study design All studies must disclose on these points even when the disclosure is negative. Sample size No sample size calculation was performed as no information on the experimental outcome was available at the time when this study was designed. Group sizes of 5-6 mice were chosen for experiments with 2-3 times replication, because we realized that these setups were sufficiently large to identify differences between groups.

Data exclusions No data was excluded from any results.

Replication All experiments were replicated successfully. A sufficiently high number of animals was used in each group to demonstrate statistical significance, as indicated in the figure legends.

Randomization Mice were randomly allocated for immunization, infection and treatment. All animals were bred and kept under specific-pathogen-free conditions in the local animal facility, thus they have similar baseline immune condition.

Blinding Investigators were not blinded to genotypes or treatment groups prior to data collection as non-subjective measures were used (i.e. antibody titer, viral load or body weight) to describe observations.

Reporting for specific materials, systems and methods We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response. Materials & experimental systems Methods n/a Involved in the study n/a Involved in the study Antibodies ChIP-seq Eukaryotic cell lines Flow cytometry Palaeontology MRI-based neuroimaging Animals and other organisms Human research participants Clinical data

Antibodies Antibodies used The following fluorescence-labeled antibodies were used for flow cytometry: CD4 (1:200, AF700, clone RM4-5, #4324114, eBioscience), CD19 (1:200, PerCP/5.5, clone 1D3/CD19, #152406, BioLegend), B220 (1:200, PE, clone RA3-6B2, #103208, eBioscience), CD16/32 (1:100, clone 93, #101320, BioLegend), CD45 (1:200, FITC, clone 30-F11, #103108, BioLegend), CD45 (1:200, PE, clone 30-F11, #103108, BioLegend), CD45 (1:200, AF700, clone 30-F11, #103108, BioLegend), CD45 (1:200, APC, clone 30-F11, #103108, BioLegend), CD45 (1:200, PerCP/5.5, clone 30-F11, #103108, BioLegend), CD45 (1:200, BV421, clone 30-F11, #103108, BioLegend), CD45 (1:200, APC-Cy7, clone 30-F11, #103108, BioLegend), EpCAM (1:100, AF488, clone G8.8, #118210, BioLegend), EpCAM (1,100, BV421, clone G8.8, #118225, BioLegend), NKM16-2-4 (1:50, APC, clone NKM16-2-4, # 30102149, Miltenyi Biotec), PD1 (1:100, FITC, clone 29F.1A12, #135214, BioLegend), CXCR5 (1:100, APC, clone L138D7, #145506, BioLegend), CD44 (1.100, BV421, clone IM7, # 103040, BioLegend), FAS (1:100, PerCP/5.5, clone SA367H8, #152610, BioLegend), GL7 (1:100, PB, clone GL7, #144614, BioLegend), TSLP (1:30, biotinylated, #BAF555, R&D Systems), October 2018 Streptavidin (1:500, PE, #405204, Biolegend) IgG1 (1.100, FITC, clone A85-1, #553533, BD Biosciences). P-STAT1 (1:30, PE, clone 4a, #562069, BD Biosciences) CD8a (1:100, FITC, clone 53-6.7, #100706, BioLegend) CD11c (1:100, AF700; clone N418, #117320, BioLegend) CD11b (1:100, PE-Cy7, clone M1/70, #101216, Biolegend) CD103 (1:100, PE, clone 2E7, #121406, Biolegend) MHC-II (1:100, I-A/I-E), BV421, clone M5/114.15.2, #107632, BioLegend) CD3 (1.100, APC, clone 17A2, #100236, BioLegend)

2 CD3 (1:100, PE-Cy7, clone 17A2, #100220, BioLegend)

nature research | reporting summary The following horseradish peroxidase-labeled antibodies were used for ELISA: IgG (1:2,000, #62-6520, Invitrogen) IgG1 (1:2,000, #A10551, Invitrogen) IgG2c (1:7,000, #1078-05, Southern Biotech) IgA (1:2,000, #626720, Invitrogen)

For in vivo neutralization of TSLP, anti-TSLP (R&D Systems MAB555, 25 μg per treatment) and IgG isotype control (R&D Systems MAB006, 25 μg per treatment) were used.

Validation All antibodies used in this study were purchased from commercial sources and were validated by the vendors. Validation data are available on the manufacturer's websites. We further verified specificity by using isotype control antibodies when performing experiments.

Eukaryotic cell lines Policy information about cell lines Cell line source(s) Madin-Darby Canine Kidney (MDCK) cell lines were purchased from ATCC.

Authentication MDCK cell lines were not authenticated.

Mycoplasma contamination MDCK cell lines were not tested for Mycoplasma contamination.

Commonly misidentified lines No commonly misidentified cell lines were used in this study. (See ICLAC register)

Animals and other organisms Policy information about studies involving animals; ARRIVE guidelines recommended for reporting animal research Laboratory animals C57BL/6J wildtype and the following mutant mice having C57BL/6J genetic background were used: B6.A2G-Mx1 carrying intact Mx1 alleles (designated Mx1-WT), B6.A2G-Mx1-Ifnαr1-/- lacking functional IFN-α receptors (designated Mx1-Ifnαr1-/-), B6.A2G- Mx1-Ifnlr1-/- lacking functional IFN-λ receptors (designated Mx1-Ifnlr1-/-), Tslpr-/- mice, and Batf3-/- mice. In addition, BALB/c and DBA/2J mice were used for virus transmission studies. Male and female mice of 6-12 weeks of age were used.

Wild animals Our study did not involve wild animals.

Field-collected samples Our study did not involve samples collected from the field.

Ethics oversight All experiments with mice were carried out in accordance with the guidelines of the Federation for Laboratory Animal Science Associations (FELASA) and the national animal welfare body. Experiments were in compliance with the German animal protection law and were approved by the animal welfare committee of the Regierungspräsidium Freiburg (permits G-16/177 and G-17/74). Note that full information on the approval of the study protocol must also be provided in the manuscript.

Flow Cytometry Plots Confirm that: The axis labels state the marker and fluorochrome used (e.g. CD4-FITC). The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers). All plots are contour plots with outliers or pseudocolor plots. A numerical value for number of cells or percentage (with statistics) is provided.

Methodology

Sample preparation The details of sample preparation for flow cytometry are demonstrated in the Methods section. October 2018

Instrument We used a BD LSRFortessa machine (BD Biosciences; 4 lasers - blue, red, violet and Yellow-Green; 14 fluorescent parameters) for analysing cells. For cell sorting, we used a BD FACSAria™ Fusion cell sorter machine (BD Biosciences)

Software We used the software FACSDiva 6.2 and FACSD 8.0.1 (BD Biosciences) to collect the data and the software FlowJo (Treestar Inc., VX) to analyze the data.

3 After cell sorting, we obtained between 3x10e5 and 1x10e6 cells from each target population (which were used for RNA

Cell population abundance nature research | reporting summary extraction). Post-sort analysis showed >93% to purity of target subpopulations.

Gating strategy For cell sorting, we first selected cells based on FSC-A x SSC-A; then, we selected the singlets based on FSC-W x FSC-A; SSC-W x SSC-A and FSC x Live-Dead gating. Afterwards, we selected CD45 single positive cells and EpCAM single postive cells. M positve cells were further gated on EpCAM positive cells using NKM16-2-4 maker, as indicated in the Fig. S5. For Tfh and GC B cells gating, among single and live cells, Tfh cells were gated as: CD19-CD4+CD44+CXCR5+PD1+; GC B cells were gated as: CD4-B220 +GL7+Fas+. Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information. October 2018