Article

Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells

Graphical Abstract Authors Mauro Gaya, Patricia Barral, Marianne Burbage, ..., Andreas Bruckbauer, Jessica Strid, Facundo D. Batista

Correspondence [email protected] (M.G.), [email protected] (F.D.B.)

In Brief NKT cells are required for the initial formation of germinal centers and production of effective neutralizing antibody responses against viruses.

Highlights d NKT cells promote B cell immunity upon viral infection d NKT cells are primed by lymph-node-resident macrophages d NKT cells produce early IL-4 wave at the follicular borders d Early IL-4 wave is required for efficient seeding of germinal centers

Gaya et al., 2018, Cell 172, 1–17 January 25, 2018 ª 2017 The Authors. Published by Elsevier Inc. https://doi.org/10.1016/j.cell.2017.11.036 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036 Article

Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells

Mauro Gaya,1,2,* Patricia Barral,2,3 Marianne Burbage,2 Shweta Aggarwal,2 Beatriz Montaner,2 Andrew Warren Navia,1,4,5 Malika Aid,6 Carlson Tsui,2 Paula Maldonado,2 Usha Nair,1 Khader Ghneim,7 Padraic G. Fallon,8 Rafick-Pierre Sekaly,7 Dan H. Barouch,1,6 Alex K. Shalek,1,4,5 Andreas Bruckbauer,2 Jessica Strid,9 and Facundo D. Batista1,2,10,11,* 1Ragon Institute of MGH, MIT and Harvard, Cambridge, MA 02139, USA 2The Francis Crick Institute, London NW1A 1AT, UK 3The Peter Gorer Department of Immunobiology, King’s College London, London SE1 9RT, UK 4Institute for Medical Engineering & Science, MIT, Cambridge, MA 02139, USA 5Broad Institute, Cambridge, MA 02142, USA 6Center for Virology and Vaccine Research, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215, USA 7Case Western Reserve University, Cleveland, OH 44106, USA 8Institute of Molecular Medicine, Trinity College Dublin, Dublin, Ireland 9Division of Immunology and Inflammation, Department of Medicine, Imperial College London, London W12 0NN, UK 10Department of Microbiology and Immunobiology & HMS Center for Immune Imaging, Harvard Medical School, Boston, MA 02115, USA 11Lead Contact *Correspondence: [email protected] (M.G.), [email protected] (F.D.B.) https://doi.org/10.1016/j.cell.2017.11.036

SUMMARY through the production of highly specific antibodies. In the past few years, there has been much attention on the generation B cells constitute an essential line of defense from of broadly neutralizing antibodies against highly variable or pathogenic infections through the generation of emerging pathogens, such as influenza, HIV, and Zika virus to class-switched antibody-secreting cells (ASCs) in develop effective vaccines capable of conferring broad and germinal centers. Although this process is known to long-lasting immunity. Therefore, it is crucial to understand be regulated by follicular helper T (TfH) cells, the mech- how viruses elicit B cell immunity during an infection process. anism by which B cells initially seed germinal center Pathogen recognition occurs in secondary lymphoid organs such as the spleen and lymph nodes (Batista and Harwood, reactions remains elusive. We found that NKT cells, a 2009). In the lymph nodes, B cells recognize pathogens on the population of innate-like T lymphocytes, are critical surface of subcapsular sinus (SCS) macrophages, migratory for the induction of B cell immunity upon viral infection. dendritic cells, and follicular dendritic cells (Carrasco and Ba- The positioning of NKT cells at the interfollicular areas tista, 2007; Gaya et al., 2015; Heesters et al., 2013; Junt et al., of lymph nodes facilitates both their direct priming by 2007; Phan et al., 2007; Qi et al., 2006). Following pathogen resident macrophages and the localized delivery of recognition, B cells process and present pathogen-derived pep- innate signals to antigen-experienced B cells. Indeed, tides via major histocompatibility complex (MHC) class II mole- NKT cells secrete an early wave of IL-4 and constitute cules (Amigorena et al., 1994). This results in the recruitment of + up to 70% of the total IL-4-producing cells during the specific CD4 T cells, which help the B cells to differentiate initial stages of infection. Importantly, the requirement into either antibody-secreting plasma cells or germinal center of this innate immunity arm appears to be evolution- (GC) cells. Within germinal centers, B cells undergo class-switch recombination and cycles of proliferation, somatic hypermuta- arily conserved because early NKT and IL-4 gene tion, and affinity-based selection, finally resulting in the produc- signatures also positively correlate with the levels tion of high-affinity antibodies (Victora and Nussenzweig, 2012). of neutralizing antibodies in Zika-virus-infected ma- These processes are strictly dependent on co-stimulatory sig- caques. In conclusion, our data support a model nals, mainly CD40-L, ICOS, and cytokines, such as interferon g wherein a pre-TfH wave of IL-4 secreted by interfollic- (IFN-g), interleukin-4 (IL-4), and IL-21, provided by follicular ular NKT cells triggers the seeding of germinal center helper T (TfH) cells (Weinstein et al., 2016). Because TfH cells cells and serves as an innate link between viral infec- are known to appear later during the immune response, how tion and B cell immunity. B cells initially seed germinal center reactions is still unclear. In addition to pathogen-derived peptides, B cells are able to INTRODUCTION present glycolipid antigens, such as those present on the surface of Sphingomonas species (spp.), Borrelia burgdorferi, and Strep- B cells are key elements of the adaptive immune response tococcus pneumoniae, on CD1d molecules (Barral et al., because they provide protection from pathogenic threats 2008; Kinjo et al., 2011; Leadbetter et al., 2008). Populations of

Cell 172, 1–17, January 25, 2018 ª 2017 The Authors. Published by Elsevier Inc. 1 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

A Uninfected CD1d-/- **** 11 Uninfected 105

4 4.6% 10 0.9% 9.5% CD1d-/-

B cells (%) 5.5 103 + GL7 + 0 Fas 0 Fas 010103 104 5 GL7

B Uninfected CD1d-/- 105 24 Uninfected

104 4.3% 17.9% 18.6% -/-

T cells (%) T CD1d + 12 103 PD-1 + 0 0 CXCR5 010103 104 5 CXCR5 PD-1

C Uninfected CD1d-/- IgD GL7

E D ** ** * 18000 10 0.4 Uninfected Uninfected area + CD1d-/- -/- 9000 CD1d 5 0.2 area /IgD -IAV IgG1 ASC/LN IgG1 -IAV + α structures / section +

GL7 0 0 0 GL7

F Uninfected Wild type/Vaccinia CD1d-/-/Vaccinia ** 18 Uninfected 105 Wild type/Vaccinia 0.3% 14.7% 7.8% 4 10 CD1d-/-/Vaccinia

B cells (%) 9 103 + GL7 + 0 Fas Fas 0 010103 104 5 GL7

G Uninfected Wild type/Vaccinia CD1d-/-/Vaccinia 24 Uninfected 105 Wild type/Vaccinia 4 10 4.7% 19.8% 17.0% -/-

T cells (%) T CD1d /Vaccinia + 12 103 PD-1 + 0 0 CXCR5 CXCR5 010103 104 5 PD-1

(legend on next page)

2 Cell 172, 1–17, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

CD1d-restricted, innate-like T cells, known as natural killer (NK)T process (Figures 1A and 1B; Figures S1A–S1D). We further cells, recognize these lipids and provide cognate help to B cells. compared the spatial organization of germinal centers in wild- This NKT cell-mediated help enhances the production of anti- type and CD1dÀ/À mice by immunohistochemistry. In wild-type bodies by B cells (Bai et al., 2013; Barral et al., 2008; Leadbetter animals, influenza infection induces both the enlargement of et al., 2008). Because there are no reports of lipid antigens orig- IgD+ B cell follicles as well as the formation of follicular GL7+ inating from viruses, it is currently unclear whether NKT cells structures corresponding to germinal centers (Figure 1C). In have a role in inducing antiviral B cell responses. contrast, in NKT cell-deficient animals, germinal center struc- Here we show that, following respiratory viral infection, NKT tures were reduced by day 9 of infection, a defect that cell-deficient mice are impaired in their ability to form germinal was accompanied by a strong decrease in the generation of centers and immunoglobulin G (IgG) class-switched antibody- influenza-specific IgG1 ASCs (Figures 1C–1E). Interestingly, secreting cells (ASCs). In a striking contrast to B cell activation CD1dÀ/À mice also displayed a significant reduction in germinal by glycolipid antigens, CD1d-mediated interactions between center formation when infected with the vaccinia virus Western B cells and NKT cells are dispensable during viral infection. Reserve strain, suggesting that the ability of NKT cells to pro- Instead, NKT cells promote B cell immunity through the early mote B cell responses is a general feature of viral infections (Fig- and localized production of IL-4 at the follicular borders, a pro- ures 1F and 1G). cess that is likely driven by CXCR3 and requires initial priming To address whether NKT cells would also induce B cell immu- by resident macrophages through CD1d and IL-18. Importantly, nity after vaccination with viral antigens, we intranasally immu- this seems to be a conserved process because transcriptomic nized wild-type and CD1dÀ/À mice with hemagglutinin (HA) analysis of Zika-virus-infected macaques shows that early NKT trimmer from the PR8 influenza strain in the presence or absence and IL-4 gene signatures positively correlate with the levels of of two different adjuvants: alpha-galactosylceramide (a-GalCer) Zika-specific neutralizing antibodies. Therefore, our results indi- or poly(I:C). We found that HA-specific germinal centers devel- cate that the production of IL-4 by NKT cells immediately after oped in lung-draining lymph nodes only when HA was adminis- viral infection contributes to a pre-TfH IL-4 wave at the follicular tered together with either of these adjuvants (Figures S1E borders that facilitates the initial seeding of germinal center cells. and S1F). Interestingly, the formation of HA-specific germinal centers and ASCs is completely abrogated in NKT-cell-deficient RESULTS mice when a-GalCer was co-administered, demonstrating that NKT cells are essential to induce B cell immunity when exoge- NKT Cells Promote B Cell Immunity during Viral nous glycolipids are used as adjuvants (Figures S1G and S1H). Infection In addition, when poly(I:C) was co-administered, we observed The role of NKT cells in inducing the humoral response to glyco- a diminution in the number of IgM ASCs and a consistent, lipid-bearing antigens has been established (Bai et al., 2013; although non-significant, reduction in HA-specific germinal cen- Barral et al., 2008; Leadbetter et al., 2008); however, because ters in CD1dÀ/À animals. These results indicate that NKT cells virus-derived lipid antigens are unknown, we sought to investi- may also enhance B cell immunity during intranasal vaccination gate the role of NKT cells in the induction of B cell responses dur- in the absence of exogenous glycolipids (Figures S1I and S1J). ing viral infections. Accordingly, we intranasally infected C57BL/6 wild-type and CD1dÀ/À mice, which lack CD1d and NKT Cells Deliver Non-cognate Help to B Cells to Induce type I and II NKT cells, with 200 plaque-forming units (PFUs) of Antiviral Immunity influenza A virus, Puerto Rico/8/34 (PR8) strain. After 7, 9, and Having shown that CD1dÀ/À mice have diminished B cell re- 21 days of infection, lung-draining lymph nodes were harvested, sponses during influenza and vaccinia virus infection, we sought and adaptive immune response development was assessed to determine the mechanism by which NKT cells help B cells to by flow cytometry. We observed a progressive expansion of mount an antiviral immune response. In the case of bacterium- Fas+GL7+ germinal center cells and CXCR5+PD-1+ TfH cells dur- derived glycolipids, NKT cells differentiate into CXCR5+PD- ing the course of infection in wild-type animals compared with 1+Bcl6+ follicular helper NKT (NKTfH) cells, which engage in uninfected ones (Figures 1A and 1B; Figures S1A–S1D). Interest- CD1d-mediated cognate interactions with B cells, produce ingly, the formation of germinal center cells was significantly IL-21, and support germinal center responses (Chang et al., reduced in NKT cell-deficient mice at early time points, whereas 2011; King et al., 2011). To address whether NKT cells differen- the expansion TfH cells remained similar throughout the infection tiate into NKTfH cells after viral infection, wild-type mice were

Figure 1. NKT Cell-Deficient Mice Exhibit Impaired Antiviral B Cell Responses (A and B) Flow cytometry analysis of mediastinal lymph nodes (LN) from control (black), wild-type (blue), or CD1dÀ/À (red) mice on day 9 of influenza infection. Contour plots display the percentage of (A) germinal center and (B) TfH cells. (C) Confocal microscopy images of mediastinal lymph node sections from mice treated as in (A) and (B), stained with antibodies to IgD (green) and GL7 (magenta). Scale bar, 450 mm. (D) Quantification of the number and follicular area occupied by germinal centers in confocal images shown in (C) by ImageJ. (E) ELISPOT analysis of IgG1 influenza-specific ASCs in mediastinal lymph nodes from wild-type and CD1dÀ/À mice treated as described in (A) and (B). (F and G) Flow cytometry analysis of mediastinal lymph nodes of from control (black), wild-type (blue), or CD1dÀ/À (red) mice on day 9 of vaccinia virus infection. Contour plots display the percentage of (F) germinal center and (G) TfH cells. Data show one representative result from three experiments, where each dot represents one mouse. Data represent mean ± SEM; two-tailed paired Student’s t test, *p < 0.05, **p < 0.01, and ****p < 0.0001.

Cell 172, 1–17, January 25, 2018 3 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

A Gated on B220- LN cells B NKT cells CD4+ T cells 20 105 NKT cells CD4+ T cells 105 NKT cells CD4+ T cells 4 4 1.5% 20.0% 10 10 cells (%) + 10 3 3 10 10 PD-1 +

β 0 0

CXCR5 0 CXCR5 TCR 010103 104 5 010103 104 5 010103 104 5 0369 CD1d-tetramer CD4 PD-1 Days after infection

C *** ** D Targeted CD1d allele GL7 12 16 CD1d-tet Exon 2Lox P Exon 3 Lox P Exon 4 ) -2 m μ

6 cell density 8 + cell number -5 + (10 Exon 2 Lox P Exon 4 CD1d-tet CD1d-tet 0 0 Outside Inside Outside Inside GC GC GC GC

E B cells F CD19-Cre **** G Mb1-Cre **** 5 100 100 10 Cre+ Cre- 100 100 Cre+ Cre- 104

3 50 50 50 50 B cells (%) B cells (%)

10 + + CD1d

0 CD1d Events (% of max) Events (% of max)

CD19 0 0 0 0 3 4 5 3 4 5 3 4 5 01010 10 0 10 10 10 0 10 10 10 - + Cre- Cre+ Cre Cre IgD CD1d CD1d

H CD1d CD19-Cre- CD1d CD19-Cre+ I CD1d CD19-Cre- CD1d CD19-Cre+ 5 20 105 15 10 10.1% 10.2% 104 18.3 % 16.5 % 104 T cells (%) T

3 + 10 10 B cells (%) 7.5 103 + PD-1 0 + GL7 0 +

Fas 0 CXCR5 CXCR5 Fas 0 3 4 5 0 3 4 5 - + 01010 10 Cre- Cre+ 10 10 10 Cre Cre GL7 PD-1

J CD1d Mb1-Cre- CD1d Mb1-Cre+ K CD1d Mb1-Cre- CD1d Mb1-Cre+ 105 12 105 16 8.9% 8.2% 104 15.9 % 14.9 % 104 T cells (%) T 103 +

B cells (%) 6 8 103 + PD-1 0 + GL7 0 + as Fas CXCR5 F 0 0 3 4 5 CXCR5 3 4 5 01010 10 Cre- Cre+ 01010 10 Cre- Cre+ GL7 PD-1

Figure 2. NKT Cells Deliver Non-cognate Help to B Cells during Viral Infection (A and B) Flow cytometry analysis of wild-type mice infected with influenza virus. (A) Gating strategy for NKT and CD4+ T cells. B220+ cells were excluded from the analysis. (legend continued on next page)

4 Cell 172, 1–17, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

intranasally infected with influenza virus, and mediastinal lymph levels of the surface activation marker CD69. We observed nodes were harvested after 0–9 days. Flow cytometry analysis that, although most NKT cells were negative for CD69 in unin- revealed that, although conventional CD4+ T cells gradually ex- fected mice (day 0), the number of CD69+ NKT cells peaked press CXCR5, PD-1, Bcl6, and SLAM and produce high levels 3 days after infection, indicating that activated NKT cells accu- of IL-21 after infection, TCRb+CD1d-tetramer+ NKT cells do mulated early after viral challenge (Figure 3A). Interestingly, influ- not upregulate Bcl6, CXCR5, or PD-1 and express moderate enza infection not only results in an upregulation of activation levels of SLAM and IL-21 (Figures 2A and 2B; Figures markers but also in a robust production of IFN-g and IL-4, with S2A–S2D). These results indicate that NKT cells do not acquire a similar percentage of NKT cells producing either cytokine (Fig- NKTfH cell phenotype after viral infection. ure 3B; Figure S3A). Notably, even though NKT cells produce To further explore the localization of NKT cells within the IFN-g after infection, they represent only 1% of the total IFN-g+ tissues of infected mice, we stained mediastinal lymph nodes cell population in lymph nodes (Figure 3C). In striking contrast, of wild-type mice with R-phycoerythrin-labeled, PBS-57-loaded NKT cells represent up to 50%–70% of the total IL-4+ cells after CD1d-tetramer as reported previously (Lee et al., 2015). Lymph influenza infection (Figure 3C); however, we did not detect NKT node sections were further stained for GL7 and analyzed by cell production of other type 2 cytokines, such as IL-5 and confocal microscopy. We found that tetramer-labeled cells IL-13 (Figure S3B). These observations demonstrate that, localized in the periphery of GL7+ areas, whereas almost no although the population of NKT cells is much smaller than those tetramer-labeled cells were found inside these GL7+ areas (Fig- of conventional T cells, they represent the main source of IL-4 in ure 2C). This observation indicates that NKT cells are excluded lymph nodes at the early stages of influenza infection. from the germinal center after viral infection. Previous studies have shown that TfH cells constitute essen- To directly address whether CD1d-mediated cognate interac- tially all of the IL-4-producing T cells in draining lymph nodes tion between B cells and NKT cells is required for the develop- after infection (King and Mohrs, 2009; Reinhardt et al., 2009; ment of the antiviral B cell response, we generated CD1dflox/flox Yusuf et al., 2010). However, these cells usually accumulate later mice to delete CD1d specifically on B cells. To this end, two during the infection process. To gain insight into the kinetics of loxP sites were inserted, flanking exon 3 of CD1d, and IL-4 production by NKT and TfH cells after influenza infection, CD1dflox/flox mice were then crossed with mice expressing the we took advantage of IL-4 GFP reporter mice, which concomi- Cre recombinase under the promoter of CD19 or Mb1 genes tantly express endogenous levels of IL-4 in addition to GFP (Figure 2D; Figures S2E–S2H). Deletion of CD1d was observed (Mohrs et al., 2001). Flow cytometry analysis revealed an accu- in approximately 80% of B cells in CD1dflox/floxCD19-Cre+ mice mulation of NKT cells shortly after influenza infection, reaching and in 100% of B cells in CD1dflox/floxMb1-Cre+ mice, indicating a maximum number on day 3, and a gradual but continuous that the Mb1-Cre system is a more efficient tool to delete CD1d accumulation of TfH cells, reaching a maximum number on on B cells (Figures 2E–2G). Then, groups of CD1dflox/floxCD19- day 9 (Figure 3D). In agreement with previous studies, we Cre+ and CD1dflox/floxMb1-Cre+ and their corresponding CreÀ observed GFP expression in approximately 20%–40% of NKT littermates were intranasally infected with influenza virus, and cells even on day 0, suggesting that NKT cells already produce mediastinal lymph nodes were harvested after 9 days. The IL-4 mRNAs at steady state (Figure 3E; Lee et al., 2013; Mohrs formation of germinal centers and TfH cells was similar et al., 2005). However, the percentage of GFP+ NKT cells peaks in CD1dflox/flox CD19-CreÀ and CD19-Cre+ mice and in up to 60% to 80% after 3 days of infection, indicating that influ- CD1dflox/flox Mb1-CreÀ and Mb1-Cre+ animals, indicating that enza infection rapidly induces IL-4 production by NKT cells (Fig- CD1d-mediated interactions between B cells and NKT cells are ure 3E). Notably, both steady-state IL-4 producers and those not required for the development of germinal centers during viral that acquire the de novo ability to produce IL-4 can contribute infection (Figures 2H–2K). to the early pool of IL-4 secretors (Figures S3C–S3E). In contrast, the percentage of GFP+ TfH cells increases gradually after infec- An Early NKT Cell Wave of IL-4 Precedes Late IL-4 tion, with around 10% of GFP+ TfH cells on day 3 and 30% of Production by TfH Cells GFP-expressing TfH cells on day 9 of infection (Figure 3E). These One mechanism by which NKT cells might initiate antiviral B cell results indicate that, although NKT cells accumulate and immunity in a non-cognate manner is likely to be via the produc- become IL-4 producers rapidly after infection, TfH cells differen- tion of B cell-modulatory cytokines. To examine the ability of tiate and produce IL-4 later throughout the infection process. NKT cells to produce these molecules, we first assessed their To assess the contribution of NKT and TfH cells to the pool of kinetics of activation after influenza infection by measuring the IL-4-producing cells at different times of infection, we gated on

(B) Analysis of the percentage of NKT (blue) and CD4+ T (red) cells that acquire CXCR5 and PD-1 markers after infection. (C) Confocal microscopy analysis of lymph nodes on day 9 of influenza infection incubated with CD1d tetramer (magenta) and GL7 (green). Charts show the number and density of NKT cells inside and outside of germinal centers. Each dot represents a single germinal center. Scale bar, 60 mm. (D) Schematic showing the CD1d locus targeted for deletion with two loxP sites flanking exon 3 of CD1d. (E) Representative contour plot shows the gating strategy for B cells. (F and G) Flow cytometry analysis of CD1d expression in B cells from (F) CD1dflox/floxCD19-Cre or (G) CD1dflox/floxMb1-Cre mice: CreÀ (blue) and Cre+ (red). (H–K) Flow cytometry analysis of germinal center and TfH cell formation in (H and I) CD1dflox/floxCD19-Cre or (J and K) CD1dflox/floxMb1-Cre mice on day 9 of influenza infection: CreÀ (blue dots) and Cre+ (red dots). All data are representative of 3 independent experiments. Unless specified, each dot represents one mouse. Horizontal bars, mean; error bars, SEM; statistical analysis, two tailed Student’s t test, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Cell 172, 1–17, January 25, 2018 5 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

A Lymph node cells NKT cells B NKT cells MFI 50 105 105 d6 1107 NKT cells d5 28.3 % 29.8 % 104 929 104 d4 883 25 3 d3 1182 3 10 10 cells (%) NKT d2 746 IFN-γ+ IL-4+ + β β 0 d1 450 0 d0 380 % Maximum TCR TCR Cytokine 0 010103 104 5 0 103 104 105 010103 104 5 010103 104 5 IFN-γ+ IL-4+ CD1d-tetramer CD69 IFN-γ IL-4

C Lymph node cells IFN-γ+ cells Lymph node cells IL-4+ cells **** + 100 105 105

104 104 10 20.4% 0.5 % 0.2 % 60.0 % 3 103 10 1 0

0 LN population (%) IFN-γ+ cells NKTs IL-4+ cells NKTs NKT cells in cytokine NKT 0.1 SSC SSC + 010103 104 5 010103 104 5 010103 104 5 010103 104 5 IFN-γ IL-4+ IFN-γ CD1d-tetramer IL-4 CD1d-tetramer

D E NKT cells TfH cells

+ + ) 8 4 NKT cells GFP GFP 80 NKT cells TfH cells d9: 65% d9: 29% TfH cells cells (%)

d6: 62% d6: 11% + 4 40 d3: 76% d3: 10%

d0: 40% d0: 5% IL-4 GFP % Maximum Number of cells (x10 0 0 103 104 105 0 3 6 9 0 0 3 6 9 Days after infection IL-4 GFP Days after infection

F Gated on IL-4 GFP+ LN cells Day 0 Day 3 Day 6 Day 9 Early NKT IL-4 Late TfH IL-4

5 80 10 NKT 19.8%NKT 68.3% NKT 32.8% NKT 13.9% NKT cells TfH cells 4

10 LN cells + 40 3 TfH 12.1% TfH 1.5% TfH 42.5% TfH 67.3% 10 104

0 103 0 CD1d Tetramer % of total GFP 0 103 104 105 0 3 6 9 CXCR5 Days after infection

Figure 3. NKT Cells Generate a Pre-TfH Cell Wave of IL-4 (A) Flow cytometry analysis of CD69 levels in the TCRb+ CD1d tetramer+ population after 0–6 days of influenza infection. MFI, mean fluorescence intensity. (B) Representative plots showing the percentage of NKT cells producing IFN-g (blue) and IL-4 (red) after 3 days of influenza infection. (C) Flow cytometry analysis of IFN-g (blue) and IL-4 (red) production on day 3 of influenza infection. Gates were drawn on IFN-g+ and IL-4+ lymph node cells to analyze the percentage of NKT cells inside each cytokine+ population. (D–F) Flow cytometry analysis of NKT (blue) and TfH (red) cells from IL-4 GFP mice infected with influenza virus. (D) Numbers of NKT and TfH cells at different stages of influenza infection. (E) GFP expression in NKT and TfH cells at different times of infection. (F) Percentage of CD1d-tetramer+ NKT cells and CXCR5+ TfH cells inside the IL-4 GFP+ lymph node cell population at different stages of influenza infection. In all charts, each dot represents one mouse. In graphs showing statistical analysis, data are representative of three independent experiments. Horizontal bars, mean; error bars, SEM. Student’s t test, ****p < 0.0001.

6 Cell 172, 1–17, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

A Follicular area B Follicular borders Wild type CD1d-/- ** Wild type CD1d-/- * B220 18 30 CD1d-tet

9 15 cells/follicle cells/border + + CD1d-tet 0 CD1d-tet 0 WT CD1d-/- WT CD1d-/- C D Lymph node Follicular area Follicular borders ** Lymph node cells IL-4 GFP+ cells 100 IL-4 GFP 105 CD1d-tet B220 + - + 104 IL-4 GFP CD1d-tet CD1d-tet cells (%) + 50 103 1.3% 28% 72%

CD1d-tet 0 +

0 SSC

GFP 3 4 5 3 4 5 Follicular Follicular 01010 10 01010 10 area borders IL-4 GFP CD1d-tetramer

E IL-4 GFP+ lymph node cells F CD3e Zbtb16 (PLZF) CXCR6 CXCR3 14 10 12 12 Tet+, Cluster 1 Tet+, Cluster 2 Tet-, Cluster 3 Tet-, Cluster 4 7 5 6 6 tSNE2 Expression level

0 0 0 0 tSNE1 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 Clusters

G Gata3 Tbx21 (T-bet) Rorc (RORγt) H 12 12 10 NKT cells CXCR3 levels ** 105 100 80 IL-4 GFP- IL-4 GFP+ GFP- GFP+ 104 NKT cells NKT 6 6 5 50 + 40 103 0 Expression level % CXCR3 SSC 0 0 010103 104 5 010103 104 5 GFP- GFP+ 0 0 0 IL-4 GFP CXCR3 1 2 3 4 1 2 3 4 1 2 3 4 Clusters

I NKT cells CXCR3- CXCR3+ J 40 * 200 ** 5 5 10 10 CXCR3- CXCR3- CXCR3+ CXCR3+ 104 104 11.1%

35.2% NKTs) 4 20 100 NKT cells NKT 103 103 +

0 0 IL-4 secretors (per 2x10 % IL-4 SSC SSC 0 0 3 4 5 3 4 5 0 10 10 10 01010 10 CXCR3- CXCR3+ CXCR3 IL-4

(legend on next page)

Cell 172, 1–17, January 25, 2018 7 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

TCRb+ IL-4-GFP+ lymph node cells and analyzed the proportion approximately 70% of this population corresponds to NKT cells of this population that was CD1d-tetamer+ (NKT cells) or (CD1d-tet+), whereas the remaining 30% (CD1d-tetÀ) are mostly CXCR5+ (TfH cells). Interestingly, we observed that, 3 days after T cells (TCRb+), excluding a significant contribution of ILC2s to influenza infection, almost 70% of GFP+ cells were NKT cells, the early wave of IL-4 (Figure 4D; Figure S4B). To further study whereas less than 2% were TfH cells (Figure 3F). This trend is these two cell populations, IL-4 GFP+ CD1d-tet+ and IL-4 reversed around 6 days after infection, and, by day 9, less than GFP+ CD1d-tetÀ cells were single cell-sorted and analyzed by 15% of GFP+ cells were NKT cells, whereas almost 70% were RNA sequencing (RNA-seq). After processing (STAR Methods), TfH cells (Figure 3F). These results indicate that, during influenza we retained 222 cells from the CDd1-tet+ (n = 113) and CDd1- infection, there is an early wave of IL-4, in which NKT cells consti- tetÀ (n = 109) populations in our single-cell RNA-seq analysis. tute the main source of this cytokine, and a late wave of IL-4, T-distributed stochastic neighbor embedding (t-SNE) and where TfH cells overcome NKT cells as the main IL-4 producers. shared nearest neighbor (SNN) clustering revealed that single cells from the two subsets (CD1d-tet+ and CD1d-tetÀ) each An Early NKT Cell Wave of IL-4 Occurs at the Follicular bifurcate into two separate clusters, reflecting divergent tran- Borders scriptional states within these populations (Figures 4E; Fig- So far, we have defined the temporal frame of IL-4 production by ure S4C; Table S1). Independent of the cluster, most of these NKT cells during the early stages of influenza infection. To gain cells express CD3e, confirming their T cell origin (Figure 4F). insight into the spatial distribution of these IL-4-producing NKT A feature of the CD1d-tetÀ subset clusters is the expression of cells, we infected wild-type, CD1dÀ/À, and IL-4 GFP reporter CCR7, L-selectin, and S1PR1, mRNAs encoding proteins that mice with influenza virus and harvested mediastinal lymph nodes control T cell migration (Figure S4D). Furthermore, one of the after infection. The lymph nodes were incubated with labeled CD1d-tetÀ clusters is characterized by cells expressing high PBS-57-loaded CD1d-tetramer, and sections were further levels of CD4 and the other one by cells expressing CD8a. Impor- stained against B220 and CD169, a macrophage marker, and tantly, only a small proportion of the cells in these two clusters analyzed by confocal microscopy. Although CD1d-tetramer+ express Bcl-6, ruling out a significant contribution of TfH cells cells were nearly undetectable in uninfected animals, they were to this early IL-4 wave (Figures 3F; Figures S4D and S4E). On observed inside B cell follicles and in direct contact with the other hand, significant numbers of cells from the CD1d-tet+ CD169+ macrophages at the subcapsular sinus and interfollicu- subset express PLZF, a transcription factor that directs the lar areas by day 3 of infection (Figures 4A and 4B; Figure S4A). In NKT cell effector program (Figure 4F). Interestingly, a large pro- contrast, CD1d-tetramer+ cells were almost absent in lymph no- portion of cells from both CD1d-tet+ clusters express the chemo- des from CD1dÀ/À animals, indicating that CD1d-tetramer+ cells kine receptor CXCR6, whereas only one of them expresses are most likely NKT cells (Figures 4A and 4B). Interestingly, CXCR3 (Figure 4F). Both CXCR3+ and CXCR3À clusters express although the great majority of CD1d-tetramer+ cells located GATA-3, whereas the CXCR3+ group specifically expresses inside the B cell follicles do not express GFP, most of the T-bet, and the CXCR3À one expresses RORgt, an observation GFP+ cells appear to be located in the areas surrounding B cell that was further confirmed by flow cytometry (Figures 4G; Fig- follicles (Figure 4C). These results indicate that early IL-4 produc- ures S4F and S4G). tion is restricted to the periphery of the B cell follicles, where CXCR3 has been reported to recruit T cells to the interfollicular antigen-specific B cells relocate to recruit T cell help after zones, the areas where we find most of the IL-4 GFP+ NKT cells activation. after infection (Groom et al., 2012; Sung et al., 2012). Interest- To get insight into the extent of cellular heterogeneity within ingly, flow cytometry analysis of lymph nodes from influenza- the immune cells contributing to the early IL-4 wave, we charac- infected mice revealed that 80% of IL-4 GFP+ NKT cells express terized the IL-4 GFP+ population by flow cytometry 3 days post- CXCR3, whereas only 30% of IL-4 GFPÀ NKT cells express infection with influenza virus. This experiment shows that this marker (Figure 4H). Therefore, we wondered whether

Figure 4. The Early IL-4 Wave Is Localized at the Periphery of B Cell Follicles (A–C) Confocal microscopy analysis of (A and B) wild-type and CD1dÀ/À and (C) IL4-GFP mice on day 3 of influenza infection. Lymph nodes were labeled with CD1d tetramer (magenta) and anti-B220 antibody (green in A and B and white in C). The arrows in (C) indicate CD1d tetramer+ cells expressing IL-4 (IL-4 GFP, green). Scale bars, 300 mm (lymph node) and 60 mm (section). (D) Flow cytometry analysis of IL-4 GFP+ cells in mediastinal lymph nodes on day 3 of influenza infection, showing CD1d-tetÀ and CD1d-tet+ cells. (E) t-SNE plots of CD1d-tetÀ and CD1d-tet+ subsets. The CD1d-tet+ population separates into two clusters (1, dark purple; 2, light purple), as does the CD1d-tetÀ population (3, light green; 4, dark green). (F and G) Expression distribution (violin plots) in each population (horizontal axes) for (F) CD3e, Zbtb16, CXCR6, and CXCR3 and (G) Gata3, Tbx21, and Rorc.

Expression levels are log2(tpm+1). (H) Flow cytometry analysis of CXCR3 levels in IL-4-GFPÀ (red) and GFP+ (blue) NKT cells on day 3 of influenza infection. Lines connect cells from the same mouse. (I) Flow cytometry analysis of IL-4 production by CXCR3À (red) and CXCR3+ (blue) NKT cells on day 3 of influenza infection. Lines connect cells from the same mouse. (J) ELISPOT analysis of IL-4-producing CXCR3À (red) and CXCR3+ (blue) NKT cells sorted on day 3 of influenza infection. Each dot represents pooled NKT cells from 6 mice. In graphs showing statistical analysis, data are representative of at least two independent experiments. Horizontal bars, mean; error bars, SEM. Student’s t test, *p < 0.05, **p < 0.01.

8 Cell 172, 1–17, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

CXCR3À and CXCR3+ NKT cells would have a differential ability TfH cells remained unaltered in the absence of macrophages to produce IL-4 after infection. To this end, we analyzed IL-4 pro- (Figure 5H). These results indicate that CD169+ macrophages duction by both CXCR3À and CXCR3+ NKT cell groups on day 3 are essential for early triggering of NKT cell effector programs of influenza infection by flow cytometry and enzyme-linked im- and the initiation of B cell response following viral infection. munospot (ELISPOT). Strikingly, the number of cells secreting IL-4 in the CXCR3+ NKT cell sub-population is five times higher Induction of NKT Cell-Derived IL-4 by Macrophages than in the CXCR3À group (Figures 4I and 4J). These results sug- Requires CD1d and IL-18 gest that CXCR3 preferentially drives the recruitment of IL-4-pro- Because the depletion of CD169+ macrophages abolished ducing NKT cells to interfollicular areas. the production of IL-4 by NKT cells, we investigated whether Taken together, our results indicate that influenza infection CD1d-mediated interactions between CD169+ macrophages triggers two spatiotemporally divergent waves of IL-4: an early and NKT cells are required for NKT cell activation. To address wave, mainly produced by NKT cells and restricted to the periph- this question, we generated mixed bone marrow chimeras lack- ery of B cell follicles, and a late one, produced by germinal cen- ing CD1d expression specifically on CD169+ macrophages. For ter-resident TfH cells. this, CD169-DTR mice were sub-lethally irradiated and, the following day, adoptively transferred with either 80% CD169- Resident Macrophages Trigger an Early NKT Cell Wave DTR cells + 20% wild-type cells or 80% CD169-DTR cells + of IL-4 and B Cell Immunity 20% CD1dÀ/À cells. After 1, 4, and 7 weeks, mice were treated To define the immune machinery involved in the induction of with diphtheria toxin, and, at week 8 of reconstitution, both this early NKT cell-derived IL-4, we specifically abrogated groups of chimeras were intranasally infected with influenza virus CD1d expression on different immune cell populations by (Figure 6A). Although we observed a robust production of IL-4 in crossing our CD1dflox/flox mice with animals expressing the CD169-DTR/wild-type chimeras by day 3 of infection, NKT cells Cre recombinase under the control of the Mb1, Lyz2, or from CD169-DTR/CD1dÀ/À chimeras were impaired regarding CD11c gene promoters. Flow cytometry analysis showed an this cytokine (Figure 6B). This result demonstrates that close effective deletion of CD1d on Gr-1+CD11b+ neutrophils/ interactions between CD169+ macrophages and NKT cells myeloid-derived suppressor cells (MDSCs) in CD1dflox/flox- through CD1d are necessary to induce IL-4 production by NKT + Lyz2-Cre mice and on MHCII+CD11c+ dendritic cells in cells after viral infection. CD1dflox/exCD11c-Cre+ mice (Figures S5A–S5C). However, The depletion of CD169+ macrophages has a more profound no deletion of CD1d could be obtained in lymph node effect on the induction of IL-4 production by NKT cells than solely CD169+ macrophages in these transgenic lines (Figure S5D). the deletion of CD1d on these cells. Therefore, we wondered Groups of CD1dflox/floxMb1-Cre+, CD1dflox/floxLyz2-Cre+, and whether CD169+ macrophages could enhance IL-4 synthesis CD1dflox/exCD11c-Cre+ and their corresponding CreÀ litter- through the secretion of co-stimulatory cytokines. To address mates were intranasally infected with influenza virus, and this question, we stained lymph nodes on day 2 of influenza mediastinal lymph nodes were harvested after 3 or 9 days. infection with antibodies against CD169, IL-1b, IL-18, and Interestingly, similar percentages of IL-4+ NKT cells (day 3) IL-33. These cytokines are secreted early after infection and and germinal centers and TfH cells (day 9) were observed have been shown to modulate the differentiation and function among these transgenic lines (Figures 2H–2K and 5A–5C; Fig- of innate and adaptive lymphoid cells (Kastenmu¨ ller et al., ures S5E–58H). These results indicate that CD1d-mediated 2012; Sagoo et al., 2016). We detected the accumulation of cognate interactions between NKT cells and B cells, neutro- IL-1b, IL-18, and IL-33 in both subcapsular sinus and medullar phils/MDSCs, or dendritic cells are not required for NKT cell- macrophages, suggesting that these resident macrophages mediated induction of IL-4 production or germinal center are a source of inflammatory cytokines (Figures 6C and 6D; Fig- formation. ure S6A). Interestingly, their production is not the consequence Because we were unable to delete CD1d on CD169+ macro- of direct infection of macrophages because no viral replication phages by using CD11c or Lyz2 Cre lines, we decided to take was detected on them when a recombinant influenza virus a macrophage depletion approach to assess the importance of carrying a GFP reporter gene was used (Manicassamy et al., these macrophages in the induction of IL-4 production by NKT 2010; Figure S6B). cells. Accordingly, mice expressing the diphtheria toxin receptor Next we tested whether macrophage-derived IL-1b, IL-18, or (DTR) under the CD169 promoter, here named CD169-diphtheria IL-33 are required to induce NKT cell production of IL-4. These toxin receptor mice, were treated with PBS or diphtheria toxin. three cytokines are known to signal through the MyD88 adaptor Although PBS injection has no effect on macrophages, a single protein (Adachi et al., 1998; Schmitz et al., 2005); hence, as a injection of diphtheria toxin is sufficient to deplete both subcap- first approach, we compared the ability of wild-type versus sular sinus and medullar macrophages from lymph nodes MyD88À/À NKT cells to produce IL-4 after influenza infection. (Figures 5D and 5E). After 4 days of PBS or diphtheria toxin Our flow cytometry results show that, although a substantial pro- administration, mice were intranasally infected with influenza portion of lymph node NKT cells from wild-type mice produce virus, and mediastinal lymph nodes were harvested 3 or 9 days IL-4, NKT cells from MyD88À/À animals are severely impaired later. Interestingly, as revealed by flow cytometry, the production in the production of this cytokine (Figure 6E). Importantly, their of IL-4 by NKT cells (day 3) as well as the formation of germinal impairment is not due to lack of Toll-like receptor (TLR) signaling centers (day 9) were severely impaired in macrophage-depleted because NKT cells from TLR7À/À mice secrete IL-4 to the same mice (Figures 5F and 5G). Notably however, the development of extent as NKT cells from wild-type animals (Figure 6F). We then

Cell 172, 1–17, January 25, 2018 9 DadE ofclmcocp nlsso CD169 of analysis microscopy Confocal E) and (D fiflez neto:Cre infection: influenza of nalpnl,ec o ersnsoemue aaso n ersnaiersl rmtreeprmns aarpeetmean represent Data experiments. three from result representative 0.0001. one < show ****p Data 0.01, < mouse. **p one test, represents t dot Student’s each panels, dots). (red all DT In and dots) (blue PBS infection: GadH lwctmtyaayi fgria etr()adTH()clsi CD169 in cells (H) TfH and (G) center germinal of analysis cytometry Flow H) and dots). (G (red DT and dots) (blue iue5 eietMcohgsPooeEryI- rdcinb K el n niia elImmunity Cell B CD1d Antiviral (A) and in cells Cells NKT NKT by by production IL-4 Production of IL-4 analysis cytometry Early Flow Promote (A–C) Macrophages Resident 5. Figure 10 F lwctmtyaayi fI- rdcinb K el rmCD169 from cells NKT by production IL-4 300 of bars, analysis Scale cytometry (green). Flow CD169 (F) and (white) B220 to antibodies G E C A laect hsatcei rs s aae l,Iiito fAtvrlBCl muiyRle nInt inl rmSailyPstoe NKT Positioned Spatially from Signals Innate on Relies Immunity Cell B Antiviral of Initiation al., et Gaya https://doi.org/10.1016/j.cell.2017.11.036 as: (2018), press Cell in Cells, article this cite Please 10 10 10

Fas 10 10 10 10 10 10 TCRβ TCRβ 0 0 0 3 4 5 B220 CD169 3 4 5 Cell 3 4 5 GL7 CD1d IL-4 IL-4 CD1d yp oeSbaslrsnsMedulla Subcapsularsinus Lymph Node 010 0 172 CD169 010 10.4 % –7 aur 5 2018 25, January 1–17, , 10 10 10 10.5 % CD11c-Cre 3 3 3 DTR/+ Mb1-Cre 10 10 10 /PBS 9.0% 4 4 4 10 À - bu os n Cre and dots) (blue 5 5 5 - CD1d CD169 CD1d CD169 11.6 % DTR/+ 15.0 % CD11c-Cre DTR/+ Mb1-Cre /DT /DT 1.3% + rddots). (red + DTR/+ +

Fas+ GL7+ B cells (%) IL-4+ NKT cells (%) IL-4+ NKT cells (%) ieatr4dy f()PSo E ihhratxn(T diitain etoswr tie with stained were Sections administration. (DT) toxin diphtheria (E) or PBS (D) of days 4 after mice 10 12 16 20 6 0 0 8 0 Cre Cre m PBS lf)ad60 and (left) m - -

**** Cre Cre DT DTR/+ flox/flox + +

PStetdo CD169 or /PBS-treated b-r,()CD1d (B) Mb1-Cre, m B F H D cne n right). and (center m 10 10 10

CXCR5 TCRβ 10 10 10 TCRβ 10 10 10 0 0 0 DTR/+ 3 4 5 3 4 5 3 4 5 B220 CD169 IL-4 PD-1 IL-4 CD1d PStetdadCD169 and /PBS-treated yp oeSbaslrsnsMedulla Subcapsularsinus Lymph Node 010 010 CD169 010 CD169 10 10 10 2 14.3 % 19.9% 3 DTR/+ 10 flox/flox Lyz2-Cre DTR/+ 3 19.1 % 10 DTR/+ 3 /PBS 10 10 /PBS 4 y2Ce n C CD1d (C) and Lyz2-Cre, 4 4 D-rae ieo a fiflez neto:PBS infection: influenza of 3 day on mice /DT-treated - 5 5 5 CD169 CD1d DTR/+ CD169 CD169 DTR/+ D-rae ieo a finfluenza of 9 day on mice /DT-treated 18.7 % 18.9% PBS Lyz2-Cre DTR/+ DTR/+ 2.6 % /DT /DT flox/ex + D1-r ieo a 3 day on mice CD11c-Cre ±

E;totie paired two-tailed SEM; + + IL-4 NKT cells (%) + + IL-4 NKT cells (%) 40 20

CXCR5 PD-1 T cells (%) 40 20 26 0 13 0 0 Cre PBS PBS -

** Cre DT DT +

Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

A B CD169/Wild type CD169/CD1d-/- ** Irradiation Bone marrow Diphteria toxin Harvest 25 5 CD169-DTR transfer administration infection LNs 10 18.9 % 7.3 % 104 12.5 3 -1 0 7 28 49 56 59 10 NKT cells (%) NKT Days + β 0 i) 80% CD169-DTR IL-4 0 TCR 20% Wild type 0 103 104 105 CD169 CD169 ii) 80% CD169-DTR IL-4 WT CD1d-/- 20% CD1d-/-

C IL-1β IL-18 IL-33 D IL-1β IL-18 IL-33 CD169 CD169 CD169 CD169 CD169 CD169 IL-1β IL-18 IL-33 IL-1β IL-18 IL-33 Medulla Subcapsular sinus

E F Wild type TLR7 -/- Wild type MyD88-/- *** 40 42 105 105 20.4% 5.0% 38.0 % 31.2 % 104 104 20 21 103 103 NKT cells (%) NKT NKT cells (%) NKT + +

β 0 β IL-4 0 IL-4 0 0 TCR TCR -/- 010103 104 5 WT MyD88-/- 010103 104 5 WT TLR7 IL-4 IL-4 G H Wild type IL-1R -/- Wild type IL-18R -/- ** 40 20 105 105 24.2 % 24.8 % 13.9 % 7.0 % 104 104 20 10 103 103 NKT cells (%) NKT NKT cells (%) NKT + +

0 β β 0 IL-4 IL-4 0 0 TCR TCR -/- 3 4 5 WT IL-18R-/- 010103 104 5 WT IL-1R 01010 10 IL-4 IL-4 I J Wild type-CD45.1 IL-33R-/--CD45.2 NKT cells IL-4 secretors 140 50 105 19.2% 20.1% 0 10

4 cells 10 4 70 25 0.1 100 103 NKT cells (%) NKT + IL-4 secretors per 2x10 1 1000 β 0 IL-4 0 TCR 0 00.1110 100 1000 0 103 104 105 WT IL-33R-/- IL-18 (ng/ml) IL-18 (ng/ml) IL-4

Figure 6. Resident Macrophages Promote Early NKT Cell IL-4 Production through CD1d and IL-18 (A) Experimental scheme showing the strategy for the generation of chimeras lacking CD1d specifically on CD169+ macrophages. (B) Flow cytometry analysis of IL-4 production by NKT cells from mice subjected to the experimental scheme show in (A): wild-type chimeras (blue dots) and CD1dÀ/À chimeras (red dots). (legend continued on next page)

Cell 172, 1–17, January 25, 2018 11 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

explored whether signaling through IL-1R, IL-18R, or IL-33R is observed a significant reduction in the formation of germinal involved in this process. We performed influenza infection in centers and in their ability to class-switch to IgG1 in mice treated wild-type, IL-1RÀ/À, or IL-18RÀ/À mice and mixed bone marrow with anti-IL-4 (Figures 7E and 7F; Figure S7A). Importantly, when chimeras transferred with 50% CD45.1 wild-type and 50% we injected anti-IL-4 3 days prior to infection, this reduction was CD45.2 IL-33RÀ/À cells. We found that IL-1R and IL-33R were not significant, suggesting that the blocking effect of this anti- not required for IL-4 production because NKT cells from body only lasts for a few days (Figures S7B). Altogether, these IL-1RÀ/À mice and IL-33RÀ/À experimental chimeras did not results suggest that the early IL-4 wave is necessary to promote show an impairment in IL-4 production compared with their B cell immunity. wild-type counterparts (Figures 6G–6I). In contrast, the percent- Subsequently, to check whether the recovery of the initial age of NKT cells producing IL-4 was significantly reduced in IL-4 wave could rescue the germinal center defect observed IL-18RÀ/À mice, suggesting that IL-18 enhances IL-4 secretion in NKT cell-deficient mice, we administered PBS or IL-4 com- by NKT cells (Figure 6H). Remarkably, MyD88 and IL-18R were plexed with an anti-IL4 antibody (IL-4c) to groups of CD1dÀ/À intrinsically required by NKT cells because MyD88À/À and mice during the initial 2 days of infection. Remarkably, admin- IL-18RÀ/À NKT cells displayed a reduced ability to produce istration of IL-4c for the initial days of infection is sufficient to IL-4 compared with wild-type cells in a mixed bone marrow restore a robust germinal center response and enhance IgG1 chimera setup (Figures S6C–S6E). Furthermore, IL-18 could class-switching in the absence of NKT cells (Figures 7G induce IL-4 production on itself when added ex vivo to primary and 7H). In contrast, administration of more antigen-specific NKT cells (Figure 6J). Altogether, our results indicate that the TfH cell precursors to CD1dÀ/À animals did not rescue the IL-18R-MyD88 axis plays an important role in the production of germinal center defect, indicating that the TfH cells might not IL-4 by NKT cells. compensate for the absence of NKT cells (Figures S7C–S7F). Altogether, our results indicate that the early wave of IL-4 trig- An Early IL-4 Wave Initiates B Cell Immunity to Viral gered after influenza infection is critical for the induction of Infection B cell responses. Because of the central role of NKT cells in producing IL-4 early Previous studies have shown that IL-4 supports the energy after infection, we wondered whether CD1dÀ/À mice would and biosynthetic needs necessary for growth, proliferation, and have reduced numbers of IL-4-secreting cells. To address this effector functions of immune cells (Dufort et al., 2007; Huang question, wild-type and CD1dÀ/À mice were infected with influ- et al., 2016; Ricardo-Gonzalez et al., 2010). Thus, it is feasible enza virus, and the amount of IL-4+ cells was measured on that the early IL-4 wave would induce metabolic reprograming day 3. The number of cells secreting IL-4 was significantly on antigen-experienced B cells to fulfil the high metabolic re- reduced in CD1dÀ/À compared with wild-type animals, indi- quirements associated with the generation of germinal centers. cating that other immune cells cannot compensate for the To test this hypothesis, we initially cultured murine naive B cells absence of NKT cells (Figure 7A). Next, to test whether the or B cells stimulated through their B cell receptor (BCR) impaired B cell responses observed in CD1dÀ/À animals could (anti-IgM) in the presence or absence of IL-4. Interestingly, we be due to reduced IL-4 production, FVB wild-type and IL-4À/À observed that the addition of extracellular IL-4 to our ex vivo mice were infected with influenza virus, and, 8 days later, medi- B cell cultures increases the oxygen consumption rate (OCR) in astinal lymph nodes were harvested. A substantial expansion of both naive and antigen-stimulated lymphocytes (Figure S7G). germinal center and TfH cells was observed in wild-type mice To address whether early IL-4 secretion could induce metabolic compared with uninfected animals (Figures 7B and 7C). Interest- changes on B cells in vivo, we infected groups of mice with influ- ingly, in IL-4À/À mice, there was a significant reduction in the pro- enza virus and treated them with either PBS or anti-IL4 blocking portion of germinal centers, whereas the differentiation of TfH antibody. We isolated lymph node cells on day 3 of infection, cells remained similar to that of wild-type animals (Figures cultured them with anti-IgM and anti-CD3ε for 16 hr, and 7B and 7C). Furthermore, germinal center cells from IL-4À/À measured OCR levels in purified B cells after the incubation. mice have a reduced capacity to class-switch to IgG1 (Fig- We found that B cells from lung-draining lymph nodes have a ure 7D). These results show that IL-4 is necessary for the induc- higher OCR than B cells from non-draining lymph nodes, indi- tion of the B cell response after influenza infection. cating that there is a metabolic reprograming of B cells early after Then we asked whether the early wave of IL-4 was necessary influenza infection (Figure S7H). Interestingly, the ability of B cells for the induction of B cell responses after influenza infection. To to increase their respiratory capacity after viral infection requires address this question, wild-type mice were infected with influ- early IL-4 production because the upregulation of the OCR is enza virus, and a sub-group of them was treated with a blocking diminished when the early IL-4 wave is inhibited (Figure S7H). IL-4 antibody during the first 3 days of infection. By day 8, we These results suggest that the early IL-4 wave triggered after

(C and D) Confocal microscopy analysis on day 2 of influenza infection showing (C) SCS and (D) medullar areas. Sections were stained with antibodies to CD169 (green) and IL-1b/IL-18/IL-33 (magenta). Scale bars, 60 mm. (E–I) Flow cytometry analysis of wild-type and (E) MyD88À/À, (F) TLR7À/À, (G) IL-1RÀ/À, and (H) IL-18RÀ/À mice or (I) IL-33RÀ/À chimeric mice on day 3 of influenza infection. Charts show IL-4+ NKT cells in wild-type (blue dots) or mutant (red dots) populations. (J) ELISPOT analysis of IL-4 production by sorted NKT cells incubated ex vivo with IL-18. Each dot represents one mouse. Data show mean ± SEM and are representative of 3 independent experiments. Statistical analysis: two-tailed Student’s t test, **p < 0.01, ***p < 0.001.

12 Cell 172, 1–17, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

A Wild type CD1d-/- ** B Wild type IL-4-/- ** 5 15 5 5 10 ) 10 3 13.3% 7.8% 104 0.25 % 0.08 % 104

B cells (%) 7.5 2.5 +

3 cells (x10 3 10 + 10 GL7 +

0 IL-4 0 0 Fas 0 SSC Fas -/- -/- WT CD1d 3 4 5 WT IL-4 010103 104 5 01010 10 IL-4 GL7

C Wild type IL-4-/- D Wild type IL-4-/- * 24 5.4 105 105 18.7% 18.0% 4.1% 1.8% 104 104 T cells (%) T + 12 2.7 3 3

10 10 GC cells (%) + PD-1 +

0 0 IgG1 0 0 CXCR5 SSC

CXCR5 -/- 010103 104 5 WT IL-4-/- 010103 104 5 WT IL-4 PD-1 IgG1

α E α F Control -IL4 (d0-d3) Control -IL4 (d0-d3) * 12 * 7.0 105 105 9.7% 4.1% 5.7% 2.3% 104 104 6 B cells (%) 3.5 + 3

3 10 GC cells (%)

10 + GL7 +

0 IgG1 0 SSC Fas 0 0 Fas 010103 104 5 Control α-IL4 010103 104 5 Control α-IL4 IgG1 (d0-d3) GL7 (d0-d3)

-/- G CD1d-/- CD1d-/- + IL-4c (d0-d3) *** H CD1d CD1d-/- + IL-4c (d0-d3) ** 14 44 105 105 2.9% 11.8% 14% 38% 104 104 B cells (%)

+ 7 22 3 3

10 10 GC cells (%) + GL7 + IgG1

0 Fas 0

0 SSC 0 Fas -/- -/- 0 103 104 105 CD1d CD1d 010103 104 5 CD1d-/- CD1d-/- GL7 +IL-4c IgG1 +IL-4c

I Zika virus Harvest LNs Collect Serum K IL-4 infection Gene expression NAbs GRB2 UBE2V2 STK17B NKT AMICA1 SETX TMEM49 ETS1 ST6GAL1 CCR6 VAX2 STK4 ID2 03 7 RBMS1 GPR114 CENPA ITK BCL6 CCL26 ST18 LRRC17 MYCN AKT2 AURKB SOS1 ICAM2 GIMAP4 IL2RB KDM1 A RAMP3 MMP25 JAG1 SDC1 GPX1 J Lymph node day 3 vs. Neutralizing antibodies day 7 BAD TK1 PIK3CA CCNB1 RASA2 CHAD SOX13 KCNK1 GATA3 AVPI1 FAM69 A RNF208 NKT.POPULATIONS NAA1 6 F10 TACC3 IL1R1 BLK ACTN2 PLEKHF1 ITM2 A TOP50. MOST.DEG.NKT1 BATF NPL IL17RE PMF1 CDK1 CYFIP2 EOMES IL4 CDCA8 CD8B IL6R STAT4 RRM1 TOP50. MOST.DEG.NKT2 GMNN RORA TESC CD6 PDE9A WDR76 ASF1B ZMYND1 1 TOP50. MOST.DEG.NKT17 STAT6 IL18R1 TKT IL24 CD2 KLF12 BCL11B HIVEP2 LEF1 IL18 SIGNALING PTPN22 SENP7 MAPK8 IL7R FASLG PRDM1 ISY1 IL4.SUPERPATHWAY.GENECARD SYTL3 CD28 GLB1 LRRFIP1 CD247 IL2RA STAT6 CHIP_SEQ TARGET GENES TMEM43 SKI ITPKB RAB35 CA6 ITGB2 FAM188 A TFH SIGNALING EIF2C2 SAMM50 RAB8B BSCL2 IL18 IKBKE TBC1D2 3 PGM1 SYNGAP1 Genesets normalized enrichment score TFPT RNF168 RGS12 RFX3 -2 +2 KIF13B MED15 Shared genes

(legend on next page)

Cell 172, 1–17, January 25, 2018 13 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

infection could be regulating seeding of germinal center cells synthetic glycosphingolipid, results in strong protection against through the induction of metabolic changes in antigen-specific subsequent challenges with various strains of influenza virus B cells. (Artiaga et al., 2016; Galli et al., 2007). However, whether NKT cells could play a role in the induction of adaptive immune re- Early NKT Cell and IL-4 Signatures in Lymph Nodes from sponses in the absence of exogenous glycolipids was not clear. Zika-Infected Macaques Here we showed that NKT cells are required for the early seeding We were interested to find out whether the early role for NKT cells of germinal center B cells and the production of class-switched and IL-4 in the initiation of antiviral B cell immunity might extend antibody-secreting cells in response to influenza infection. to other species. To address this question, we analyzed tran- Furthermore, we found that NKT cells can enhance B cell immu- scriptomic data from lymph nodes of rhesus macaques that nity to influenza-derived proteins when administered together were subcutaneously infected with Zika virus (Aid et al., 2017). with TLR adjuvants. Our findings are in line with previous studies We assessed the transcriptomic profiles of lymph nodes ob- showing that NKT cells enhance the activity of CD8+ T cells after tained on day 3 or 14 post-infection to identify pathways that influenza infection even in the absence of exogenous glycolipids correlated with Zika-specific neutralizing antibody titers on day (De Santo et al., 2008; Ishikawa et al., 2010). Therefore, the 7 and day 14, respectively (Figure 7I; Figure S7I). Pathway importance of NKT cells in the induction of adaptive immunity enrichment analysis revealed a significant and positive correla- goes beyond the presence of glycolipid antigens. tion of NKT cells (normalized enrichment score [NES] = 1.48, How do NKT cells promote B cell responses during infection? false discovery rate [FDR] = 0.06), IL-4 (NES = 1.43, FDR = Previous work showed that, in response to bacterium-derived 0.02), IL-18 (NES = 1.42, FDR = 0.05), and STAT6 signaling glycolipids, NKT cells differentiate into NKTfH cells that engage (NES = 1.83, FDR = 0.02) signatures as early as day 3 of infection in prolonged cognate interactions with B cells. These CD1d- with neutralizing antibody titers at day 7 (Figure 7J). Furthermore, mediated interactions rapidly induce germinal center and extra- gene-interacting network inference showed the high intercon- follicular plasma cell formation, leading to the production of high nectivity of all these pathways, highlighted by co-expression of titers of IgM and early class-switched antibodies (Bai et al., 2013; their leading genes with a central role of IL-4 in connecting all Barral et al., 2008; Chang et al., 2011; King et al., 2011; Leadbet- of these genes (Figure 7K). Importantly, no significant correlation ter et al., 2008). However, we found that, upon influenza virus between TfH cell signatures (FDR > 0.20) and neutralizing anti- infection, NKT cells do not differentiate into NKTfH cells, nor bodies was found at this early time point (Figure 7J). In contrast, they localize in germinal centers. Furthermore, CD1d-mediated a positive correlation between TfH cell signatures and neutral- cognate interactions between B cells and NKT cells are dispens- izing antibodies was observed on day 14 of infection (NES = able for the induction of antiviral B cell responses. Therefore, 1.23, FDR = 0.13) (Figures S7J and S7K). These correlates are even though NKT cells promote B cell immunity in response to in line with our previous findings, pointing out to a possible role both bacterium-derived glycolipids and viral infection, the mech- for NKT cells at the early stages of B cell immunity, with TfH cells anism behind this regulation varies depending on the absence or only appearing later on the infection process. Furthermore, these presence of pathogen-derived glycolipids on the surface of the analyses indicate that an early pre-TfH wave of NKT cells and infectious agent. IL-4 might be conserved upon influenza and Zika virus infection Our results show that, during the initial stages of a viral infec- in mice and primates, respectively. tion, NKT cells become the main source of IL-4 in draining lymph nodes. Flow cytometry and single-cell RNA-seq analysis re- DISCUSSION vealed that around 70% of the IL-4-producing cells are NKT cells, whereas the remaining ones are mainly T cells, excluding Several studies have focused on the use of NKT cell agonists a significant contribution by ILC2s and TfH cells to the IL-4 pro- as vaccine adjuvants to boost NKT cell activity and immune duction at early stages of infection. This finding was unexpected responses. In particular, immunization with inactivated influenza because previous studies reported that TfH cells constitute virus or virus-derived proteins in combination with a-GalCer, a essentially all of the IL-4 producing cells in lymph nodes in

Figure 7. A Conserved Early IL-4 Wave Is Critical for Induction of B Cell Immunity (A) Flow cytometry analysis on day 3 of influenza infection. Graphs show the number of IL-4+ lymph node cells in wild-type (blue dots) and CD1dÀ/À (red dots) animals. (B–D) Flow cytometry analysis on day 8 of influenza infection. Graphs show the percentage of (B) germinal center, (C) TfH, and (D) IgG1+ cells in wild-type (blue dots) and IL-4À/À mice (red dots). (E and F) Flow cytometry analysis on day 8 of influenza infection of wild-type mice treated with PBS or blocking a-IL-4 antibody during the initial 3 days of infection. Graphs show (E) germinal center and (F) IgG1+ cells in wild-type (blue dots) and a-IL-4 treated mice (red dots). (G and H) Flow cytometry analysis on day 8 of influenza infection of CD1dÀ/À mice treated with PBS or IL-4 complex (IL-4c) during the initial 3 days of infection. Graphs show (G) germinal center cells and (H) IgG1+ cells in wild-type (blue dots) and IL-4c-treated mice (red dots). (I) Experimental scheme showing the Zika virus infection strategy. (J) Linear regression model of gene signatures for NKT cells, TfH cells, IL-4, STAT6, and IL-18 from harvested lymph nodes on day 3 with neutralizing antibodies measured in plasma on day 7. Red squares, positive correlation; blue squares, negative correlation; white squares, no significant correlation. (K) Gene-interacting network inference representing interconnectivity of NKT cells, IL-4, STAT6, and IL-18 pathways. In (A)–(H), each dot represents a single mouse. Data are representative of three independent experiments. Statistical analysis: two-tailed Student’s t test, *p < 0.05, **p < 0.01, ***p < 0.001.

14 Cell 172, 1–17, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

response to infection with different pathogens (King and Mohrs, NKT cell-IL-4 gene signatures and the serum levels of Zika 2009; Reinhardt et al., 2009; Yusuf et al., 2010). However, this virus-neutralizing antibodies. We believe our findings raise the previously unreported NKT cell-mediated wave of IL-4 produc- prospect of harnessing the interplay between macrophages tion occurs at the early stages of infection, even before the and NKT cells and the early production of IL-4 at follicular appearance of TfH cells or germinal centers. Furthermore, this borders in future approaches for the design of new vaccination phenomenon is restricted to the follicular borders, where B cells strategies. migrate immediately after pathogen encounter. The spatiotem- porally restricted production of IL-4 by NKT cells appears to STAR+METHODS instruct antigen-activated B cells at the periphery of B cell folli- cles early after infection for the subsequent seeding of germinal Detailed methods are provided in the online version of this paper centers. Interestingly, a recent report showed that TfH cells only and include the following: switch to IL-4 production at the late stages of the infection pro- cess (Weinstein et al., 2016). Therefore, it will be important to d KEY RESOURCES TABLE discern the contribution of the early NKT cell IL-4 wave at the d CONTACT FOR REAGENT AND RESOURCE SHARING follicular borders versus the late TfH cell IL-4 wave at germinal d EXPERIMENTAL MODEL AND SUBJECT DETAILS centers to the different processes involved in the development B Mice of B cell immunity. B Infections and injections The positioning of NKT cells at the follicular borders during the B Flow Cytometry early stages of infection appears to be at the core of the NKT cell B Immunohistochemistry ability to initiate B cell immunity. Single-cell RNA-seq and flow B qRT-PCR cytometry analysis revealed that IL-4-producing NKT cells pref- B ELISPOT erentially express CXCR3, a chemokine receptor used by T cell B Extracellular flux assay subsets to relocate to the interfollicular areas of the lymph B scRNA-Seq Library Preparation node (Groom et al., 2012; Sung et al., 2012). However, whether B Single-cell RNA-seq Data Analysis NKT cells acquire CXCR3 after activation or whether a popula- B Transcriptomic analysis of samples from Zika infected tion of NKT cells already expressing CXCR3 expands after infec- macaques tion remains to be elucidated. Interestingly, Th2 cells have also d QUANTIFICATION AND STATISTICAL ANALYSIS been shown to locate and exert their functions at the periphery d DATA AND SOFTWARE AVAILABILITY of B cell follicles (Turqueti-Neves et al., 2014). Therefore, we believe our study reveals a new layer of regulation of B cell immu- SUPPLEMENTAL INFORMATION nity at the initial stages of the infection process, when B cells relocate to follicular borders and recruit T cell help. Supplemental Information includes seven figures and one table and can be found with this article online at https://doi.org/10.1016/j.cell.2017.11.036. The early production of IL-4 by NKT cells is mainly triggered by CD169+ macrophages and is mediated by CD1d interactions AUTHOR CONTRIBUTIONS and IL-18 secretion. Interestingly, a great number of NKT- macrophage contacts can be observed at the subcapsular sinus M.G. designed and performed experiments, generated CD1d-floxed mice, and interfollicular areas as early as 2 days post-influenza infec- analyzed the data, and wrote the manuscript. P.B. performed some initial tion. Our findings are in line with the notion that NKT cell activa- experiments, assist with the setup of in vivo infections, and contributed to sci- tion after infection is the result of a combination of both CD1d entific discussions. A.W.N. and A.K.S. performed and analyzed single-cell and cytokine signaling, with CD1d helping to position the cells RNA-seq data. M.B., C.T., S.A., and P.M. performed experiments and contrib- uted to scientific discussions. M.A. and D.H.B. analyzed transcriptomic data and cytokines inducing the effector functions of NKT cells from Zika-virus-infected macaques and contributed to scientific discussions. (Brennan et al., 2013). Furthermore, previous studies by others K.G. and R.-P.S. contributed to scientific discussions regarding transcriptomic and us have shown that these macrophages can present bacte- analyses. P.G.F. provided IL-33RÀ/À bone marrow. B.M. contributed to the rium-derived glycolipids on CD1d to mediate early NKT cell generation of CD1d-floxed animals and provided technical support. A.B. expansion and that IL-18 can enhance NKT cell activity in analyzed data and provided technical support for microscopy. U.N. edited À/À response to lipopolysaccharide (LPS), when co-administered the manuscript. J.S. provided IL-4 mice and contributed to scientific dis- with a-GalCer or in autoimmune models (Barral et al., 2010; Ha¨ g- cussions. F.D.B. conceived the project, supervised the work, and wrote the manuscript. glo¨ f et al., 2016; Leite-De-Moraes et al., 2001; Nagarajan and

Kronenberg, 2007). Therefore, it seems that the induction of ACKNOWLEDGMENTS diverse NKT cell effector functions by CD169+ macrophages and IL-18 is a conserved mechanism observed upon infection We thank the flow cytometry facility for technical support and the biological with various pathogens and in autoimmune disorders. resource unit for the breeding of animals and assistance with the generation In this study, we have unveiled a new layer of regulation of of CD1d-floxed animals (at the Francis Crick and Ragon Institutes). We thank B cell immunity at the early stages of viral infection that involves Kathrin Kirsch for the maintenance of mouse colonies. We thank Neil Rogers and Caetano Reis e Sousa for the provision of MyD88À/À, TLR7À/À, CD11c- a tightly regulated NKT cell-IL-4-B cell axis. Interestingly, this Cre, and Lyz2-Cre mice and for scientific discussions. We thank Michael phenomenon seems to be conserved upon viral infection in other Reth for the Mb1-Cre strain. We thank Mark Wilson for the IL-4-GFP reporter species because our transcriptomic analysis of Zika-virus-in- mice. We thank Andrew McKenzie for the IL-33RÀ/À bone marrow. We thank fected macaques revealed a positive correlation between early Samuel W. Kazer and Sanjay M. Prakadan for assistance with RNA-seq

Cell 172, 1–17, January 25, 2018 15 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

analysis. We thank Adolfo Garcı´a-Sastre for the NS1-GFP virus. We thank Paul myeloid-derived suppressor cells in mice and humans. J. Clin. Invest. 118, Thomas for the PR8-OTII virus. We thank Daniel Lingwood for the HA trimer. 4036–4048. We thank the NIH Tetramer Core Facility for the provision of labeled CD1d tet- Dufort, F.J., Bleiman, B.F., Gumina, M.R., Blair, D., Wagner, D.J., Roberts, ramers. We thank all members of the Lymphocyte Interaction Laboratory for M.F., Abu-Amer, Y., and Chiles, T.C. (2007). Cutting edge: IL-4-mediated pro- critical reading of the manuscript and scientific discussions. This work was tection of primary B lymphocytes from apoptosis via Stat6-dependent regula- supported by the Francis Crick Institute, which receives its core funding tion of glycolytic metabolism. J. Immunol. 179, 4953–4957. from Cancer Research UK (FC001035), the United Kingdom Medical Research Galli, G., Pittoni, P., Tonti, E., Malzone, C., Uematsu, Y., Tortoli, M., Maione, D., Council (FC001035), the Wellcome Trust (FC001035), the Center for HIV/AIDS Volpini, G., Finco, O., Nuti, S., et al. (2007). Invariant NKT cells sustain specific Vaccine Immunology and Immunogen Discovery of the NIH (UM1AI100663), B cell responses and memory. Proc. Natl. Acad. Sci. USA 104, 3984–3989. the Phillip T. and Susan M. Ragon Institute Foundation, and Bill and Melinda Gates Foundation Innovation Award 228966 (to F.D.B.). Gaya, M., Castello, A., Montaner, B., Rogers, N., Reis e Sousa, C., Bruckba- uer, A., and Batista, F.D. (2015). Host response. Inflammation-induced disrup- tion of SCS macrophages impairs B cell responses to secondary infection. DECLARATION OF INTERESTS Science 347, 667–672.

The authors declare no competing interests. Groom, J.R., Richmond, J., Murooka, T.T., Sorensen, E.W., Sung, J.H., Bank- ert, K., von Andrian, U.H., Moon, J.J., Mempel, T.R., and Luster, A.D. (2012). Received: December 5, 2016 CXCR3 chemokine receptor-ligand interactions in the lymph node optimize Revised: September 11, 2017 CD4+ T helper 1 cell differentiation. Immunity 37, 1091–1103. Accepted: November 20, 2017 Ha¨ gglo¨ f, T., Sedimbi, S.K., Yates, J.L., Parsa, R., Salas, B.H., Harris, R.A., Published: December 14, 2017 Leadbetter, E.A., and Karlsson, M.C.I. (2016). Neutrophils license iNKT cells to regulate self-reactive mouse B cell responses. Nat. Immunol. 17, REFERENCES 1407–1414. Heesters, B.A., Chatterjee, P., Kim, Y.-A., Gonzalez, S.F., Kuligowski, M.P., Adachi, O., Kawai, T., Takeda, K., Matsumoto, M., Tsutsui, H., Sakagami, M., Kirchhausen, T., and Carroll, M.C. (2013). Endocytosis and recycling of Nakanishi, K., and Akira, S. (1998). Targeted disruption of the MyD88 gene re- immune complexes by follicular dendritic cells enhances B cell antigen binding sults in loss of IL-1- and IL-18-mediated function. Immunity 9, 143–150. and activation. Immunity 38, 1164–1175. Aid, M., Abbink, P., Larocca, R.A., Boyd, M., Nityanandam, R., Nanayakkara, Huang, S.C.-C., Smith, A.M., Everts, B., Colonna, M., Pearce, E.L., Schilling, O., Martinot, A.J., Moseley, E.T., Blass, E., Borducchi, E.N., et al. (2017). Zika J.D., and Pearce, E.J. (2016). Metabolic Reprogramming Mediated by the Virus Persistence in the Central Nervous System and Lymph Nodes of Rhesus mTORC2-IRF4 Signaling Axis Is Essential for Macrophage Alternative Activa- Monkeys. Cell 169, 610–620.e14. tion. Immunity 45, 817–830. Amigorena, S., Drake, J.R., Webster, P., and Mellman, I. (1994). Transient Ishikawa, H., Tanaka, K., Kutsukake, E., Fukui, T., Sasaki, H., Hata, A., Noda, accumulation of new class II MHC molecules in a novel endocytic compart- S., and Matsumoto, T. (2010). IFN-g production downstream of NKT cell acti- ment in B lymphocytes. Nature 369, 113–120. vation in mice infected with influenza virus enhances the cytolytic activities of Artiaga, B.L., Yang, G., Hackmann, T.J., Liu, Q., Richt, J.A., Salek-Ardakani, both NK cells and viral antigen-specific CD8+ T cells. Virology 407, 325–332. S., Castleman, W.L., Lednicky, J.A., and Driver, J.P. (2016). a-Galactosylcer- Junt, T., Moseman, E.A., Iannacone, M., Massberg, S., Lang, P.A., Boes, M., amide protects swine against influenza infection when administered as a vac- Fink, K., Henrickson, S.E., Shayakhmetov, D.M., Di Paolo, N.C., et al. (2007). cine adjuvant. Sci. Rep. 6, 23593. Subcapsular sinus macrophages in lymph nodes clear lymph-borne viruses Bai, L., Deng, S., Reboulet, R., Mathew, R., Teyton, L., Savage, P.B., and and present them to antiviral B cells. Nature 450, 110–114. Bendelac, A. (2013). Natural killer T (NKT)-B-cell interactions promote pro- Kastenmu¨ ller, W., Torabi-Parizi, P., Subramanian, N., La¨ mmermann, T., and longed antibody responses and long-term memory to pneumococcal capsular Germain, R.N. (2012). A spatially-organized multicellular innate immune polysaccharides. Proc. Natl. Acad. Sci. USA 110, 16097–16102. response in lymph nodes limits systemic pathogen spread. Cell 150, Barral, P., Eckl-Dorna, J., Harwood, N.E., De Santo, C., Salio, M., Illarionov, P., 1235–1248. Besra, G.S., Cerundolo, V., and Batista, F.D. (2008). B cell receptor-mediated King, I.L., and Mohrs, M. (2009). IL-4-producing CD4+ T cells in reactive lymph uptake of CD1d-restricted antigen augments antibody responses by recruiting nodes during helminth infection are T follicular helper cells. J. Exp. Med. 206, invariant NKT cell help in vivo. Proc. Natl. Acad. Sci. USA 105, 8345–8350. 1001–1007. Barral, P., Polzella, P., Bruckbauer, A., van Rooijen, N., Besra, G.S., Cerun- King, I.L., Fortier, A., Tighe, M., Dibble, J., Watts, G.F.M., Veerapen, N., Haber- dolo, V., and Batista, F.D. (2010). CD169(+) macrophages present lipid anti- man, A.M., Besra, G.S., Mohrs, M., Brenner, M.B., and Leadbetter, E.A. (2011). gens to mediate early activation of iNKT cells in lymph nodes. Nat. Immunol. Invariant natural killer T cells direct B cell responses to cognate lipid antigen in 11, 303–312. an IL-21-dependent manner. Nat. Immunol. 13, 44–50. Batista, F.D., and Harwood, N.E. (2009). The who, how and where of antigen Kinjo, Y., Illarionov, P., Vela, J.L., Pei, B., Girardi, E., Li, X., Li, Y., Imamura, M., presentation to B cells. Nat. Rev. Immunol. 9, 15–27. Kaneko, Y., Okawara, A., et al. (2011). Invariant natural killer T cells recognize Brennan, P.J., Brigl, M., and Brenner, M.B. (2013). Invariant natural killer glycolipids from pathogenic Gram-positive bacteria. Nat. Immunol. 12, T cells: an innate activation scheme linked to diverse effector functions. Nat. 966–974. Rev. Immunol. 13, 101–117. Leadbetter, E.A., Brigl, M., Illarionov, P., Cohen, N., Luteran, M.C., Pillai, S., Carrasco, Y.R., and Batista, F.D. (2007). B cells acquire particulate antigen in a Besra, G.S., and Brenner, M.B. (2008). NK T cells provide lipid antigen-specific macrophage-rich area at the boundary between the follicle and the subcapsu- cognate help for B cells. Proc. Natl. Acad. Sci. USA 105, 8339–8344. lar sinus of the lymph node. Immunity 27, 160–171. Lee, Y.J., Holzapfel, K.L., Zhu, J., Jameson, S.C., and Hogquist, K.A. (2013). Chang, P.-P., Barral, P., Fitch, J., Pratama, A., Ma, C.S., Kallies, A., Hogan, Steady-state production of IL-4 modulates immunity in mouse strains and is J.J., Cerundolo, V., Tangye, S.G., Bittman, R., et al. (2011). Identification of determined by lineage diversity of iNKT cells. Nat. Immunol. 14, 1146–1154. Bcl-6-dependent follicular helper NKT cells that provide cognate help for Lee, Y.J., Wang, H., Starrett, G.J., Phuong, V., Jameson, S.C., and Hogquist, B cell responses. Nat. Immunol. 13, 35–43. K.A. (2015). Tissue-Specific Distribution of iNKT Cells Impacts Their Cytokine De Santo, C., Salio, M., Masri, S.H., Lee, L.Y.-H., Dong, T., Speak, A.O., Por- Response. Immunity 43, 566–578. ubsky, S., Booth, S., Veerapen, N., Besra, G.S., et al. (2008). Invariant NKT Leite-De-Moraes, M.C., Hameg, A., Pacilio, M., Koezuka, Y., Taniguchi, M., cells reduce the immunosuppressive activity of influenza A virus-induced Van Kaer, L., Schneider, E., Dy, M., and Herbelin, A. (2001). IL-18 enhances

16 Cell 172, 1–17, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

IL-4 production by ligand-activated NKT lymphocytes: a pro-Th2 effect of IL- Schmitz, J., Owyang, A., Oldham, E., Song, Y., Murphy, E., McClanahan, T.K., 18 exerted through NKT cells. J. Immunol. 166, 945–951. Zurawski, G., Moshrefi, M., Qin, J., Li, X., et al. (2005). IL-33, an interleukin-1- Manicassamy, B., Manicassamy, S., Belicha-Villanueva, A., Pisanelli, G., Pu- like cytokine that signals via the IL-1 receptor-related protein ST2 and induces lendran, B., and Garcı´a-Sastre, A. (2010). Analysis of in vivo dynamics of influ- T helper type 2-associated cytokines. Immunity 23, 479–490. enza virus infection in mice using a GFP reporter virus. Proc. Natl. Acad. Sci. Sung, J.H., Zhang, H., Moseman, E.A., Alvarez, D., Iannacone, M., Henrickson, USA 107, 11531–11536. S.E., de la Torre, J.C., Groom, J.R., Luster, A.D., and von Andrian, U.H. (2012). Mohrs, M., Shinkai, K., Mohrs, K., and Locksley, R.M. (2001). Analysis of type 2 Chemokine guidance of central memory T cells is critical for antiviral recall immunity in vivo with a bicistronic IL-4 reporter. Immunity 15, 303–311. responses in lymph nodes. Cell 150, 1249–1263. Thomas, P.G., Brown, S.A., Yue, W., So, J., Webby, R.J., and Doherty, P.C. Mohrs, K., Wakil, A.E., Killeen, N., Locksley, R.M., and Mohrs, M. (2005). (2006). An unexpected antibody response to an engineered influenza virus A two-step process for cytokine production revealed by IL-4 dual-reporter modifies CD8+ T cell responses. Proc. Natl. Acad. Sci. USA 103, 2764–2769. mice. Immunity 23, 419–429. Tirosh, I., Izar, B., Prakadan, S.M., Wadsworth, M.H., 2nd, Treacy, D., Trom- Nagarajan, N.A., and Kronenberg, M. (2007). Invariant NKT cells amplify the betta, J.J., Rotem, A., Rodman, C., Lian, C., Murphy, G., et al. (2016). Dissect- innate immune response to lipopolysaccharide. J. Immunol. 178, 2706–2713. ing the multicellular ecosystem of metastatic melanoma by single-cell Phan, T.G., Grigorova, I., Okada, T., and Cyster, J.G. (2007). Subcapsular RNA-seq. Science 352, 189–196. encounter and complement-dependent transport of immune complexes by Townsend, M.J., Fallon, P.G., Matthews, D.J., Jolin, H.E., and McKenzie, A.N. lymph node B cells. Nat. Immunol. 8, 992–1000. (2000). T1/ST2-deficient mice demonstrate the importance of T1/ST2 in devel- Qi, H., Egen, J.G., Huang, A.Y.C., and Germain, R.N. (2006). Extrafollicular oping primary T helper cell type 2 responses. J. Exp. Med. 191, 1069–1076. activation of lymph node B cells by antigen-bearing dendritic cells. Science Trombetta, J.J., Gennert, D., Lu, D., Satija, R., Shalek, A.K., and Regev, A. 312, 1672–1676. (2014). Preparation of Single-Cell RNA-Seq Libraries for Next Generation Reinhardt, R.L., Liang, H.-E., and Locksley, R.M. (2009). Cytokine-secreting Sequencing. Curr. Protoc. Mol. Biol. 107, 4.22.1–4.22.17. follicular T cells shape the antibody repertoire. Nat. Immunol. 10, 385–393. Turqueti-Neves, A., Otte, M., Prazeres da Costa, O., Ho¨ pken, U.E., Lipp, M., Ricardo-Gonzalez, R.R., Red Eagle, A., Odegaard, J.I., Jouihan, H., Morel, Buch, T., and Voehringer, D. (2014). B-cell-intrinsic STAT6 signaling controls C.R., Heredia, J.E., Mukundan, L., Wu, D., Locksley, R.M., and Chawla, A. germinal center formation. Eur. J. Immunol. 44, 2130–2138. (2010). IL-4/STAT6 immune axis regulates peripheral nutrient metabolism Victora, G.D., and Nussenzweig, M.C. (2012). Germinal centers. Annu. Rev. and insulin sensitivity. Proc. Natl. Acad. Sci. USA 107, 22617–22622. Immunol. 30, 429–457. Sagoo, P., Garcia, Z., Breart, B., Lemaıˆtre, F., Michonneau, D., Albert, M.L., Weinstein, J.S., Herman, E.I., Lainez, B., Licona-Limo´ n, P., Esplugues, E., Levy, Y., and Bousso, P. (2016). In vivo imaging of inflammasome activation Flavell, R., and Craft, J. (2016). TFH cells progressively differentiate to regulate reveals a subcapsular macrophage burst response that mobilizes innate and the germinal center response. Nat. Immunol. 17, 1197–1205. adaptive immunity. Nat. Med. 22, 64–71. Yusuf, I., Kageyama, R., Monticelli, L., Johnston, R.J., Ditoro, D., Hansen, K., Satija, R., Farrell, J.A., Gennert, D., Schier, A.F., and Regev, A. (2015). Spatial Barnett, B., and Crotty, S. (2010). Germinal center T follicular helper cell IL-4 reconstruction of single-cell gene expression data. Nat. Biotechnol. 33, production is dependent on signaling lymphocytic activation molecule recep- 495–502. tor (CD150). J. Immunol. 185, 190–202.

Cell 172, 1–17, January 25, 2018 17 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies PE-Cy7 Rat Anti-Mouse CD19, Clone 1D3 ThermoFisher Cat# 25-0193-82, RRID:AB_657663 Scientific Alexa Fluor 488 Mouse Anti-Mouse CD95 (APO-1/ Fas), ThermoFisher Cat# 53-0951-82, RRID:AB_10671269 Clone 15A7 Scientific Alexa Fluor 647 Rat Anti-Mouse/Human GL-7 Biolegend Cat# 144606, RRID:AB_2562185 (T and B cell activation Marker), Clone GL7 PE Rat Anti-Mouse CD4, Clone GK1.5 Biolegend Cat# 100408, RRID:AB_312693 APC Rat Anti-Mouse CD4, Clone GK1.5 Biolegend Cat# 100412, RRID:AB_312697 Biotin Rat Anti-Mouse CD185 (CXCR5), Clone SPRCL5 ThermoFisher Cat# 13-7185-82, RRID:AB_2572800 Scientific PerCP eFluor 710 Rat Anti-Mouse CD185 (CXCR5), ThermoFisher Cat# 46-7185-82, RRID:AB_2573837 Clone SPRCL5 Scientific FITC Hamster Anti-Mouse CD279 (PD-1), Clone J43 ThermoFisher Cat# 11-9985-82, RRID:AB_465472 Scientific Pacific Blue Rat Anti-Mouse/Human CD45R (B220), Biolegend Cat# 103227, RRID:AB_492876 Clone RA3-6B2 Brilliant Violet 650 Rat Anti-Mouse CD45R (B220), Biolegend Cat# 103241, RRID:AB_11204069 Clone RA3-6B2 APC Rat Anti-Mouse MHC Class II (I-A/I-E), Clone ThermoFisher Cat# 17-5321-82, RRID:AB_469455 M5/114.15.2 Scientific PerCP/Cy5.5 Armenian Hamster Anti-Mouse CD11c, Biolegend Cat# 117328, RRID:AB_2129641 Clone N418 eFluor 450 Armenian Hamster Anti-Mouse CD11c, ThermoFisher Cat# 48-0114-82, RRID:AB_1548654 Clone N418 Scientific APC Rat Anti-Mouse CD11b, Clone M1/70 BD Biosciences Cat# 553312, RRID:AB_398535 eFluor 450 Rat Anti-Mouse CD11b, Clone M1/70 ThermoFisher Cat# 48-0112-82, RRID:AB_1582236 Scientific eFluor 450 Rat Anti-Mouse CD3, Clone 17A2 ThermoFisher Cat# 48-0032-82, RRID:AB_1272193 Scientific eFluor 450 Rat Anti-Mouse CD5, Clone 53-7.3 ThermoFisher Cat# 48-0051-82, RRID:AB_1603250 Scientific PE Rat Anti-Mouse CD127 (IL-7Ra), Clone A7R34 Biolegend Cat# 135010, RRID:AB_1937251 FITC Armenian Hamster Anti-Mouse TCRb, Clone ThermoFisher Cat# 11-5961-82 RRID:AB_465323 H57-597 Scientific APC Armenian Hamster Anti-Mouse TCRb, Clone ThermoFisher Cat# 17-5961-82 RRID:AB_469481 H57-597 Scientific PE PBS57-loaded Mouse CD1d tetramer NIH Tetramer N/A Core Facility PE-Cy7 Armenian Hamster Anti-Mouse CD69, ThermoFisher Cat# 25-0691-82, RRID:AB_469637 Clone H1.2F3 Scientific PE Rat anti-Mouse CD1d, Clone 1B1 BD Biosciences Cat# 553846, RRID:AB_2073521 Pacific Blue Rat Anti-Mouse Ly6G-Ly6C (Gr-1), Biolegend Cat# 108430, RRID:AB_893556 Clone RB6-8C5 V450 Rat Anti-Mouse IgG1, Clone A85-1 BD Biosciences Cat# 562107, RRID:AB_10894002 PerCP/Cy5.5 Mouse anti-Mouse CD45.1, Clone A20 ThermoFisher Cat# 45-0453-82, RRID:AB_1107003 Scientific Pacific Blue Mouse anti-Mouse CD45.2, Clone 104 Biolegend Cat# 109820, RRID:AB_492872 (Continued on next page)

e1 Cell 172, 1–17.e1–e7, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

Continued REAGENT or RESOURCE SOURCE IDENTIFIER PerCP/Cy5.5 Rat anti-Mouse CD45, Clone 30-F11 Biolegend Cat# 103132, RRID:AB_893340 PE-Cy7 Armenian Hamster Anti-Mouse CD183 Biolegend Cat# 126516, RRID:AB_2245493 (CXCR3), Clone CXCR3-173 APC Armenian Hamster Anti-Mouse CD183 Biolegend Cat# 126512, RRID:AB_1088993 (CXCR3), Clone CXCR3-173 Alexa Fluor 488 Rat Anti-Mouse CD150 (SLAM), Biolegend Cat# 115916, RRID:AB_528744 Clone TC15-12F12.2 APC Rat Anti-Mouse CD8a, Clone 53-6.7 Biolegend Cat# 100712, RRID:AB_312751 Purified Rat Anti-Mouse CD16/32, Clone 93 Biolegend Cat# 101302, RRID:AB_312801 Alexa Fluor 647 Mouse Anti-Mouse GATA-3, Biolegend Cat# 653810, RRID:AB_2563217 Clone 16E10A23 APC Rat Anti-Mouse RORgt, Clone B2D ThermoFisher Cat# 17-6981-82, RRID:AB_2573254 Scientific PE-Cy7 Mouse Anti-Mouse T-bet, Clone 4B10 ThermoFisher Cat# 25-5825-82, RRID:AB_11042699 Scientific Alexa Fluor 647 Mouse Anti-Mouse Bcl-6, BD Biosciences Cat# 561525, RRID:AB_10898007 Clone K112-91 PE-Cy7 Rat Anti-Mouse IFN-g, Clone XMG1.2 ThermoFisher Cat# 25-7311-41, RRID:AB_1257211 Scientific APC Rat Anti-Mouse IL-4, Clone 11B11 ThermoFisher Cat# 17-7041-82, RRID:AB_469494 Scientific InVivoMab Rat Anti-Mouse IL-4, Clone 11B11 BioXCell Cat# BE0045, RRID:AB_1107707 Alexa Fluor 488 Rat Anti-Mouse CD169 (Siglec-1), Produced in the lab N/A Clone SER-4 FITC Rat Anti-Mouse IgD, Clone 11-26c.2a BD Biosciences Cat# 553439, RRID:AB_394859 Alexa Fluor 647 Rat Anti-Mouse CD207 (Langerin), Dendritics Cat# DDX0362, RRID:AB_1148742 Clone 929F3.01 Purified Goat Anti-Mouse IL-1b, Polyclonal R&D Systems Cat# AF-401-NA, RRID:AB_416684 Purified Rabbit Anti-Mouse IL-18, Polyclonal Abcam Cat# ab71495, RRID:AB_1209302 Purified Goat Anti-Mouse IL-33, Polyclonal R&D Systems Cat# AF3626, RRID:AB_884269 Alexa Fluor 555 Goat anti-Rabbit IgG (H + L) ThermoFisher Cat# A32732 RRID:AB_2633281 Scientific Alexa Fluor 555 Donkey anti-Goat IgG (H + L) ThermoFisher Cat# A-21432, RRID:AB_141788 Scientific Purified Goat anti-R-Phycoerythrin, Polyclonal Novus Biologicals Cat# NB100-78573, RRID:AB_1085348 Biotin Goat Anti-Mouse IgM, Human adsorbed Southern Biotech Cat# 102008, RRID:AB_616726 Biotin Goat Anti-Mouse IgG1, Human adsorbed Southern Biotech Cat# 107008, RRID:AB_609714 Purified Armenian Hamster anti-Mouse CD3ε, Biolegend Cat# 100302, RRID:AB_312667 Clone 145-2C11 Bacterial and Virus Strains Influenza virus A/Puerto Rico/8/1934 (PR8) H1N1 Caetano Reis e N/A Sousa’s lab PR8-OTII Thomas et al., 2006 N/A Influenza NS1-GFP Manicassamy N/A et al., 2010 Vaccinia virus Western Reserve Caetano Reis e N/A Sousa’s lab Chemicals, Peptides, and Recombinant Proteins PE Streptavidin Biolegend Cat# 405204 eFluor 450 Streptavidin ThermoFisher Cat# 48-4317 Scientific (Continued on next page)

Cell 172, 1–17.e1–e7, January 25, 2018 e2 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

Continued REAGENT or RESOURCE SOURCE IDENTIFIER PR8 Hemagglutinin trimer From Daniel N/A Lingwood Poly I:C HMW Invivogen Cat# 31852-29-6 Alpha Gal-Cer KRN7000 Enzo Life Sciences Cat# BML-SL232 Diphtheria Toxin from Corynebacterium diphtheriae Sigma Cat# D0564 Recombinant Murine Interleukin 4 Prepotech Cat# 214-14 LIVE/DEAD Fixable Blue ThermoFisher Cat# L23105 Scientific Cell Stimulation Cocktail plus protein transport inhibitor ThermoFisher Cat# 00-4975-03 Scientific Influenza A H1N1 (A/Puerto Rico/8/34) Hemagglutinin Sino Biological Cat# 11684-V08H Streptavidin-Alkaline Phosphatase Sigma Cat# S2890 Recombinant Mouse Interleukin 18 R&D Systems Cat# 9139-IL

AffiniPure F(ab’)2 Fragment Goat Anti-Mouse IgM, m Jackson Cat# 115-006-020 Chain Specific ImmunoResearch 3-Amino-9-ethylcarbazole Sigma Cat# A5754 Critical Commercial Assays Foxp3 Transcription Factor Staining Buffer Set ThermoFisher Cat# 00-5523-00 Scientific Cytofix/Cytoperm BD Biosciences Cat# 554722 RNeasy Micro Kit QIAGEN Cat# 74004 SuperScript III first-strand system ThermoFisher Cat# 18080051 Scientific SYBR green PCR master mix ThermoFisher Cat# 4309155 Scientific Mouse IL-4 ELISPOT kit BD Biosciences Cat# 551017 Pan B cell Isolation Kit, mouse Miltenyi Cat# 130-095-813 Seahorse XF Cell Mito Stress Test Kit Agilent Cat# 103015-100 Maxima Reverse Transcriptase ThermoFisher Cat# EP0741 Scientific Deposited Data Single-Cell RNA-Seq data This study GEO: GSE103753 Experimental Models: Cell Lines Dog (female), MDCK cells Caetano Reis e N/A Sousa’s lab Chicken, DF-1 cells Caetano Reis e N/A Sousa’s lab Experimental Models: Organisms/Strains Mouse: Wild type (C57BL/6J) The Jackson Strain Code: JAX 000664 Laboratory Mouse: Wild type (FVB/NCrl) Charles Rivers Strain Code: 207 Laboratory Mouse: B6.SJL-Ptprca Pepcb/BoyJ (CD45.1) The Jackson Strain Code: JAX 002014 Laboratory Mouse: B6;129-Siglec1 < tm1(HBEGF)Mtka > (CD169-DTR) Riken BioResource Strain Code: RBRC04395 Center Mouse, C.Cg-Cd19tm1(cre)Cgn Ighb/J (CD19-Cre) The Jackson Strain Code: JAX 004126 Laboratory Mouse, B6.129S6-Del(3Cd1d2-Cd1d1)1Sbp/J The Jackson Strain Code: JAX 008881 (CD1d KO) Laboratory (Continued on next page)

e3 Cell 172, 1–17.e1–e7, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

Continued REAGENT or RESOURCE SOURCE IDENTIFIER Mouse, B6.129S7-Il1r1tm1Imx/J (IL-1R KO) The Jackson Strain Code: JAX 003245 Laboratory Mouse, B6.129P2-Il18r1tm1Aki/J (IL-18R KO) The Jackson Strain Code: JAX 004131 Laboratory Mouse, B6.129P2-Lyz2tm1(cre)Ifo/J (Lyz2-Cre) The Jackson Strain Code: JAX 004781 Laboratory Mouse, B6.Cg-Tg(Itgax-cre)1-1Reiz/J (CD11c-Cre) The Jackson Strain Code: JAX 008068 Laboratory Mouse, B6.129S1-Tlr7tm1Flv/J (TLR7 KO) The Jackson Strain Code: JAX 008380 Laboratory Mouse, C57BL/6-Tg(TcraTcrb)425Cbn/Crl (OT-II) The Jackson Strain Code: JAX 004194 Laboratory Mouse, B6.129P2(SJL)-Myd88tm1.1Defr/J (MyD88 KO) The Jackson Strain Code: JAX 009088 Laboratory Mouse, C.129-Il4tm1Lky/J (IL-4/GFP-enhanced transcript, 4Get). The Jackson Strain Code: JAX 004190 Backcrossed to B6 background for 10 generations. Laboratory Mouse, B6.C(Cg)-Cd79atm1(cre)Reth/EhobJ (Mb1-Cre) The Jackson Strain Code: JAX 020505 Laboratory Mouse, B6.129P2-Il4tm1Cgn/J (IL-4 KO). Backcrossed The Jackson Strain Code: JAX 002253 to FVB/N background for 10 generations. Laboratory Mouse, Il1rl1tm1Anjm (IL-33R KO) Townsend MGI: 2386675 et al., 2000 Mouse, CD1d flox/flox CD19-Cre This study N/A Mouse, CD1d flox/flox Mb1-Cre This study N/A Mouse, CD1d flox/flox Lyz2-Cre This study N/A Mouse, CD1d flox/flox CD11c-Cre This study N/A Mouse, B6.Cg-Tg(Pgk1-flpo)10Sykr/J (FLPo) The Jackson Strain Code: JAX 011065 Laboratory Oligonucleotides Primer: HPRT Forward GCCCTTGACTATAATGAGTACTTCAGG This study N/A Primer: HPRT Reverse TTCAACTTGCGCTCATCTTAGG This study N/A Primer: IL-4 Forward ACGAGGTCACAGGAGAAGGGA This study N/A Primer: IL-4 Reverse AGCCCTACAGACGAGCTCACTC This study N/A Primer: IL-5 Forward TGACAAGCAATGAGACGATGAGG This study N/A Primer: IL-5 Reverse ACCCCCACGGACAGTTTGATTC This study N/A Primer: IL-13 Forward CCTCTGACCCTTAAGGAGCTTAT This study N/A Primer: IL-13 Reverse CGTTGCACAGGGGAGTCTT This study N/A Primer: IL-21 Forward TCAGCTCCACAAGATGTAAAGGG This study N/A Primer: IL-21 Reverse GGGCCACGAGGTCAATGAT This study N/A Software and Algorithms Diva BD Biosciences http://www.bdbiosciences.com/us/ instruments/clinical/software/flow- cytometry-acquisition/bd-facsdiva- software/m/333333/overview FlowJo (version 10.1) FlowJo, LLC https://www.flowjo.com Zen Zeizz https://www.zeiss.com/microscopy/ int/downloads/zen.html ImageJ (version 1.50c) NIH Image https://imagej.nih.gov/ij/ Imaris (version 8.2.0) Bitplane http://www.bitplane.com/ GraphPad Prism (version 7.0) GraphPad https://www.graphpad.com/ Software Inc scientific-software/prism/ (Continued on next page)

Cell 172, 1–17.e1–e7, January 25, 2018 e4 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

Continued REAGENT or RESOURCE SOURCE IDENTIFIER ELISPOT reader with immunospot software Cellular Technology Ltd N/A InDesign CS5.5 Adobe http://www.adobe.com/products/ indesign.html R (Version 3.3.2), Seurat package (version 1.4.0.16) R Core Team https://www.r-project.org/ GSEA Broad Institute http://software.broadinstitute.org/ gsea/index.jsp Gene functional annotation and network inference http://cytoscape.org/ http://genemania.org/

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to, and will be fulfilled by, the Lead Contact, Facundo D. Batista ([email protected]).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mice 8-10 week old wild-type C57BL/6 and FVB/N mice were obtained from Jackson and Charles River Laboratories. Congenic CD45.1 mice were from the Francis Crick Institute, UK. Siglec1DTR/+ (CD169-DTR) mice were obtained from the Riken BioResource Center, Japan. CD19-Cre, CD1dÀ/À, IL-1RÀ/À and IL-18RÀ/À mice were obtained from Jackson Laboratories, USA. Lyz2-Cre, Cd11c-Cre, TLR7À/À, OT-II and Myd88À/À mice were obtained from Dr. Caetano Reis e Sousa, Francis Crick Institute, UK. IL-4/GFP-enhanced transcript (4Get) mice were obtained from Dr. Mark Wilson, Francis Crick Institute, UK. FVB/N IL-4À/À mice were obtained from Dr. Jessica Strid, Imperial College London, UK. Mb1-Cre mice were obtained from Dr. Michael Reth, Max Planck Institute, Germany. IL-33RÀ/À bone marrows (Townsend et al., 2000) were obtained from Andrew McKenzie, MRC laboratory of Molecular Biology, UK and Padraic Fallon, Trinity College Dublin, Ireland. For the generation of CD1d conditional mice, we used homologous recombination to insert loxP sites flanking exon 3 of cd1d1 gene in B6 embryonic stem (ES) cells. Identification of ES cell clones with a single correct insertion of the targeted locus in the genome was performed through long-range PCR and Southern blot. Transgenic mice derived from these ES cells were crossed with B6 mice expressing the FLP recombinase to remove the frt-flanked PGK-neomycin sequence used for selection of ES cells that have inserted the targeted vector into their genome. Mice carrying the flox allele were crossed with mice expressing the Cre recombinase under the promoters of CD19, Mb1, Lyz2 and CD11c genes. Finally, mice were crossed again with the same floxed mice to make them homo- zygous for the floxed allele. For the generation of mixed bone marrow chimeras, CD45.1 C57BL/6 or CD169-DTR of 6-8 weeks of age were irradiated with 2 doses of 500 rad, 4 hours apart. One day later, bone marrow cells were injected i.v. in recipient animals (2x106 cells in total). Exper- imental animals were kept on acidified water for one week prior and 3 weeks post irradiation treatment. Chimeras were used after 8 weeks of reconstitution. Mice were bred and maintained at the animal facilities of the Francis Crick and Ragon Institutes. All mice were maintained in indi- vidually ventilated cages and were used at the age of 8 to 12 weeks. Up to five mice per cage were housed in individually ventilated cages in a 12 hr light/dark cycle, with room temperature at 22 C (19 C-23 C change). They were fed with autoclaved standard pellet chow and reverse osmosis water. All cages contained 5 mm of aspen chip and tissue wiles for bedding and a mouse house for envi- ronmental enrichment. Littermates (males or females) were randomly assigned to experimental groups. Generally between 4 to 8 mice were used per experimental group. All experiments were approved by the Animal Ethics Committee of Cancer Research UK and the United Kingdom Home Office, and by the Institutional Animal Care and Use Committee of the United States.

Infections and injections Mice were anesthetized i.p. with Ketamine/Xylazine (100 mg/kg body) and intranasally infected with 200 PFU of Influenza virus A/Puerto Rico/8/1934 (PR8) H1N1 strain or 100 PFU of PR8-OTII virus (Thomas et al., 2006)or103 PFU of Influenza NS1-GFP virus (Manicassamy et al., 2010) or 200 PFU of Vaccinia virus Western Reserve strain in 20 ml of PBS. Influenza virus was amplified on MDCK cells and VACV on DF-1 cells. Purification of viral particles was performed in a sucrose 30% cushion at 25, 000 RPM for 2 hours in an SW32Ti rotor. For protein immunizations, 5 mg of PR8 Hemagglutinin (HA) trimmer, kindly provided by Daniel Lingwood, was intranasally administered together with 10 mg of Poly I:C (Invivogen) or 2 mgofa-GalCer (Enzo) in 20 ml of PBS. For depletion of CD169+ macrophages, mice were i.p. injected with 500 ng of Diphtheria toxin (Sigma) in 200 ml of PBS. For in vivo IL-4 blocking, mice were i.p. injected with 1.5 mg of anti-IL4 (11B11, BioXCell) in 200 ml of PBS at different days of infection. For the administration of IL-4 complex (IL-4c), 5 mg of recombinant mouse IL-4 (PrepoTech) was complexed to 25 mg of anti-mouse IL-4 (BioXCell, 11B11),

e5 Cell 172, 1–17.e1–e7, January 25, 2018 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

diluted in 200 ml of PBS and administered i.p. on days 0, 1 and 2 of infection. For adoptive transfers, sorted NKT cells (8.105), OT-II CD4+ isolated CD4+ T cells (2.104) or bone marrow cells (2.106) were resuspended in 200 ml of PBS and injected in the tail vein.

Flow Cytometry For labeling of surface markers, single cell suspensions from lymph nodes or lungs were incubated with anti-mouse CD16/32 (BD Biosciences) and LIVE/DEAD Fixable blue (Life Technologies) in PBS 2% FCS for 20 minutes on ice to block nonspecific antibody binding and exclude dead cells. Cells were then washed and stained for 20 minutes on ice with the indicated anti-mouse antibodies, PBS57-loaded CD1d tetramer (kindly provided by NIH tetramer core facility) and/or biotinylated HA (Sino biological). When using biotinylated antibodies or proteins, cells suspensions were washed following antibody labeling and incubated for 20 minutes on ice with labeled streptavidin. If labeling of transcription factors was performed, cells were fixed and permeabilized with Foxp3 tran- scription factor staining buffer set (eBioscience) according to manufacturer instructions and incubated for 30 min with anti-mouse antibodies. Finally, cells were either resuspended in 400ml of FACS buffer and analyzed on a Fortessa cytometer (BD Biosciences) or in RPMI media and used for cell sorting on a FACSAria II (BD, Bioscience). For analysis of cytokine production, lymph node single cell suspensions were incubated 4 hours at 37C in RPMI complete medium supplemented with 1X Cell stimulation cocktail plus protein transport inhibitor (eBioscience). After surface staining, cells were fixed and permeabilized with Cytofix/Cytoperm (BD Biosciences) according to manufacturer instructions and incubated for 30 min with anti-mouse antibodies. Finally, cells were resuspended in 400ml of FACS buffer and data were collected on a Fortessa cytometer (BD Biosciences) and analyzed with FlowJo (TreeStar).

Immunohistochemistry For surface staining, cryostat sections (10mm thick) of lymph nodes were dried, fixed in 4% PFA for 10 minutes, blocked with PBS 5% BSA and then incubated with the following anti-mouse antibodies in 1% BSA for 1 hour: B220 Pacific Blue (RA3-6 B2, Biolegend), CD169 AF488 (Ser-4, ATCC), IgD FITC (11-26c.2a, BD Biosciences), Langerin AF647 (929F3.01, Dendritics) and GL-7 AF647 (GL7, Biolegend). For intracellular staining, sections were permeabilized with PBS Triton 0.3% for 5 minutes before blocking and then incu- bated with anti-mouse IL-1b (polyclonal, R&D), IL-18 (polyclonal, Abcam) or IL-33 (polyclonal, R&D), followed by incubation with AF555 anti-goat or anti-rabbit IgG (Invitrogen). CD1d tetramer staining was performed as previously described (Lee et al., 2015). Briefly, lymph nodes were incubated overnight with PE-labeled PBS-57 loaded CD1d tetramer in 2% FCS at 4C. Next day, lymph nodes were washed and fixed with 4% PFA for 1 hour and snap frozen. 10mm sections were blocked with 5% BSA and stained with anti-PE antibody (polyclonal, Novus Biologicals) followed by goat anti-rabbit AF555 (Life Technology). Imaging was carried out on a LSM 780 (Zeiss) inverted confocal microscope using a Plan-Apochromat 40x NA 1.3 oil immersion objective. qRT-PCR NKT cells, TfH cells and conventional CD4+ T cells from mediastinal lymph nodes were sorted at day 3 and 9 of infection using the gating strategy shown in Figure S2A. RNA was extracted from sorted cells using the RNeasy micro Kit (QIAGEN). Between 103 and 104 cells were used in each condition. mRNA was reversely transcribed into cDNA using random hexamers from the SuperScript III first-strand system (Thermofisher). cDNA was amplified using custom primer sets (Sigma) and SYBR green PCR master mix (Thermofisher) on a 7500 real-time PCR system (Applied Biosystems). The relative mRNA abundance of target genes was determined by means of the DDCT method, using HPRT as house keeping gene.

ELISPOT To measure influenza-specific ASCs, Enzyme-linked immunosorbent spot (ELISPOT) multiscreen filtration plates (Millipore) were activated with absolute ethanol, washed with PBS and coated overnight at 4C with 1 mg/ml of PR8 virus or 2 mg/ml of PR8 HA (Sino Biological) diluted in PBS. Plates were subsequently blocked for 1 hour with complete medium and incubated for 24 hours at 37C with serial dilutions of lymph node single cell suspensions. Plates were washed with PBS 0.01% Tween-20 and incubated for 1 hour with 1mg/ml of biotinylated anti-IgM or IgG1 (Southern Biotech) diluted in PBS 1% BSA. Then, plates were washed and incubated for 1 hour with 1mg/ml Streptavidin-Alkaline Phosphatase (Sigma). Finally plates were washed and developed with 3-amino-9-ethyl-carbazole (Sigma). To measure IL-4 production after influenza infection, NKT cells were sorted from mediastinal lymph nodes (TCRb+ CD1d-tetramer+ CXCR3+/À) and incubated in RPMI complete medium supplemented with 1X Cell stimulation cocktail (eBioscience) for 32 hours at a density of 2.104 cells/well. To measure IL-4 production in the presence of IL-18, NKT cells were sorted from spleen (TCRb+ CD1d- tetramer+) and incubated with increasing concentrations of mouse recombinant IL-18 (R&D systems) for 32 hours at a density of 2.104 cells/well. IL-4 secreting cells were detected using the mouse IL-4 ELISPOT kit from BD Biosciences following manufacturer’s instructions.

Extracellular flux assay Splenic naive B cells were purified using Pan B cell Isolation Kit (Miltenyi) and stimulated with 10 ng/ml IL-4 (PrepoTech) and/or

5 mg/ml F(ab’)2 anti-IgM (Jackson ImmunoResearch) for 16 hours. Lymph node cells were obtained after 3 days of influenza infection, cultured with 5 mg/ml F(ab’)2 anti-IgM and 3 mg/ml anti-CD3ε (Biolegend, 145-2C11) and B cells were purified after 16 hours. B cells

Cell 172, 1–17.e1–e7, January 25, 2018 e6 Please cite this article in press as: Gaya et al., Initiation of Antiviral B Cell Immunity Relies on Innate Signals from Spatially Positioned NKT Cells, Cell (2018), https://doi.org/10.1016/j.cell.2017.11.036

were resuspended in Seahorse medium supplemented with 11mM glucose (Sigma Aldrich), 2 mM glutamax (Life Technologies) and 1 mM Sodium pyruvate (Life Technologies). Cells were settled on 96-well assay plate (Seahorse Bioscience) coated with poly-lysine (Sigma Aldrich) for 30 minutes. Data was recorded with the XF96 Extracellular Flux analyzer. Oligomycin A (3 mM), FCCP (5 mM) and Rotenone (5 mM) (all from Agilent) were added sequentially at indicated time points to generate the OCR profile. scRNA-Seq Library Preparation Plate-based scRNA-Seq libraries were prepared as previously described (Trombetta et al., 2014). Briefly, cells were FACS sorted into 96-well plates containing 10 mL of RLT lysis buffer (QIAGEN) supplemented with 1% b-mercaptoethanol (Aldrich) in each well. Two plates of GFP+CD1d-tetramer+ and two plates of the GFP+CD1d-tetramer- cells were sorted. Smart-Seq2 whole transcriptome amplification (WTA) was performed as per (Trombetta et al., 2014) with Maxima Reverse Transcriptase used in place of SuperScript II (Life Technologies). Libraries were sequenced on an Illumina NextSeq500, with 30 base-pair paired-end reads, to an average depth of 670k ± 30k reads per cell.

Single-cell RNA-seq Data Analysis Sequenced reads were aligned to the UCSC mm10 mouse transcriptome and gene expression was calculated using RSEM as pre- viously described (Tirosh et al., 2016). Cells with greater than 1,000 genes mapped and with more than 10,000 transcripts detected were used for downstream phenotypic analysis. Genes expressed in fewer than 15 cells were filtered out from further analysis, leav- ing 222 cells and 9,271 genes. Expression data was log transformed (Log2(tpm+1)) before further analysis. Clustering and differential expression were performed using the Seurat package (version 1.4.0.16) in R (version 3.3.2) (Satija et al., 2015). Briefly, we identified 708 variable genes with log-mean expression values greater than 0.05 and dispersion (variance/mean) greater than 0.8. A principal components analysis was then run over variable genes, and the first five principal components were selected for downstream ana- lyses based on a combination of Q-Q plots and the standard deviation of PCs as determined by an elbow plot (calculated using the Jackstraw algorithm in Seurat). tSNE and SNN clustering (FindClusters function in Seurat with k.param = 20 & resolution = 0.6) were used to depict and cluster the cells in each analysis, respectively. Genes differentially expressed between identified clusters were calculated using the bimodal distribution option. Differentially expressed markers were removed if identified in less than 20 percent of cells in a given cluster or if the average log fold change between clusters was less than 0.5. Heatmaps depict the 25 most signif- icantly differentially expressed genes (p values < 0.002) in each cluster (full data available in Table S1). Violin plots were generated over curated NKT genes.

Transcriptomic analysis of samples from Zika infected macaques Genes signatures for NKT, NKT subpopulations, IL4, IL18, TfH, STAT6 and B cell signaling were compiled from different public repository (http://software.broadinstitute.org/gsea/msigdb/, http://pathcards.genecards.org/, PMID:16698943, PMCID:PMC3528072). We correlated the expression of these signatures in lymph nodes from ZIKA virus infected Rhesus monkeys to neutralizing antibodies measured in plasma from the same animals on day 7 and day 14 (Aid et al., 2017). We used a linear regres- sion analysis (R language) to identify genes that correlated significantly to neutralizing antibodies or viral loads using a p value cut-off of 0.05. Significant genes were then used to perform a genesets enrichment analysis using the above signatures (http://software. broadinstitute.org/gsea/index.jsp). Significant genes were used to infer gene interacting networks using GeneMania (http://genemania.org/) and Cytoscape (http://cytoscape.org/).

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical parameters including the exact value of n with the description of what n represents, the mean, the SEM and the p value are reported in the Figures and the Figure Legends. Statistical analyses were performed using Prism (GraphPad Software), two-tailed t test, p values < 0.05 were considered significant. In figures, asterisks stand for: *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

DATA AND SOFTWARE AVAILABILITY

The single-cell RNA-seq data reported in this paper have been deposited in the NCBI Gene Expression Omnibus (GEO) database under accession number GEO: GSE103753.

e7 Cell 172, 1–17.e1–e7, January 25, 2018 Supplemental Figures

A Day 7 B Day 7 CD1d-/- * CD1d-/- 105 5 105 14 5.1% 4.9% 104 3.6% 1.7% 104 T cells (%) T + B cells (%) 2.5

+ 7 103 103 PD-1 + GL7

+ 0 0 Fas CXCR5 Fas 0 0 010103 104 5 WT CD1d-/- 010103 104 5 CXCR5 WT CD1d-/- GL7 PD-1

C Day 21 D Day 21 CD1d-/- CD1d-/- 105 13 105 40 10.8% 10.6% 104 104 20.8% 20.3% T cells (%) T + B cells (%) 6.5

+ 20 103 103 PD-1 + GL7 0 + 0 Fas CXCR5 Fas 0 0 010103 104 5 WT CD1d-/- 010103 104 5 CXCR5 WT CD1d-/- GL7 PD-1

E HA trimer HA trimer + α-Galcer HA trimer + Poly I:C * * 105 34 HA trimer 0.5% 16.0% 8.6% HA trimer + α-Galcer 104 HA trimer + Poly I:C 3 10 B cells (%) 17 + 0 GL7 + Fas 010103 104 5 Fas 0 GL7

F Non-immunized α-Galcer Poly I:C 34 Non-immunized 105 0.8% 0.0% 0.0% α-Galcer 104 Poly I:C

B cells (%) 17 103 + GL7

0 +

Fas 0 Fas 010103 104 5 GL7

G H Wild type/ HA + α-Galcer CD1d-/-/ HA + α-Galcer * 104 * 104 * 46 Wild type 105 Wild type 25.4% 0.5% CD1d-/- CD1d-/- 104 2 2 10 10 B cells (%) 23 103 + GL7 + -HA IgG1 ASC/LN IgG1 -HA -HA IgM ASC/LN IgM -HA 0 α α 0 0 10 Fas 0

10 Fas 010103 104 5 GL7

-/-

) J Wild type/ HA + Poly I:C CD1d / HA + Poly I:C I * 2 120 25 10 Wild type 105 Wild type 6.4% 4.4% CD1d-/- CD1d-/- 104

60 12 B cells (%) 5 103 + GL7 +

-HA IgM ASC/LN IgM -HA 0 α Fas

-HA IgG1 ASC/LN (x10 IgG1 -HA 0 0 0 Fas α 010103 104 5 GL7

(legend on next page) Figure S1. NKT Cell-Deficient Mice Display Early Impairment in B Cell Immunity, Related to Figure 1 (A–D) Flow cytometry analysis of mediastinal lymph node cells from wild type or CD1d/ mice that were intranasally infected with 200 PFU of influenza virus for (A-B) 7 or (C-D) 21 days. Representative contour plots display the percentage of (A-C) B220+ cells bearing biomarkers of germinal cell activity, Fas+GL7+, and (B–D) CD4+ T cells bearing biomarkers of TfH cells, CXCR5+PD-1+. Dot plots show quantification of (A-C) germinal center and (B-D) TfH cells. (E and F) Flow cytometry analysis of mediastinal lymph nodes harvested at day 10 from wild type mice that were intranasally challenged with (E) 10 mgofHA trimmer or (F) PBS in the presence or absence of poly I:C or a-GalCer. Representative contour plots display the percentage of HA-specific B220+ cells differ- entiated into Fas+GL7+ germinal center cells. (G and I) ELISPOT analysis of IgG1 HA-specific ASCs in mediastinal lymph nodes from wild type and CD1d/ mice treated as in figure S1E. (H and J) Flow cytometry analysis of HA-specific germinal center cells in mediastinal lymph nodes from wild type and CD1d/ mice treated as in figure S1E. In all quantification charts, each dot represents one mouse. Data are representative from two experiments. Statistical analysis two-tailed Student’s t test; *p < 0.05. A Gating strategy B Day 3 Day 9 105 105 105 1024 IL-21 IL-21

NKT cells 4 T cells 104 104 10 TfH cells +

103 103 103 32 β

0 Fold Increase 0 0 CD4+ T cells TCR CXCR5 CD4 3 4 5 / CD4 NKT-TfH 0 010103 104 5 010103 104 5 01010 10 NKT TfH NKT TfH CD1d-Tetramer FSC PD-1

C **** 240 + 240 **** + 105 CD4 T cells 105 CD4 T cells NKT cells NKT cells 104 104 TfH cells TfH cells TfH 120 120 103 103 TfH Bcl-6 MFI Bcl-6 MFI

0 0 + CD4+ NKT CD4 NKT Bcl-6 0 Bcl-6 0 010103 104 5 010103 104 5 CD1d-Tetramer CD1d-Tetramer D *** *** ** *** 7000 + 2800 + 105 CD4 T cells 105 CD4 T cells TfH NKT cells TfH NKT cells 104 104 NKT TfH cells TfH cells 3500 NKT 1400 103 103

+ SLAM MFI 0 CD4 SLAM MFI 0 CD4+ SLAM 0 SLAM 0 010103 104 5 010103 104 5 CD1d-Tetramer CD1d-Tetramer

E H Southern blot strategy Neomycine probe Wild type locus Clones 1 2 3 Sac I 10.6 kb Sac I

Southern probe CD1d (exon 3)

Targeted locus

Sac I 7.3 kb Sac I Sac I

Southern probe lox P site CD1d (exon 3) Neor FRT site

Neomycine probe

7.1 kb 6.6 kb Primer 1 Primer 2 Primer 3 Primer 4

F Long range PCR strategy G Southern blot strategy Primers 1-2 Primers 3-4 Southern probe Kb Kb Kb 8 8 10 6 6 8 5 5 6

Clones 1 2 3 1 2 3 Clones 1 2 3

(legend on next page) Figure S2. Expression Analysis of TfH Markers by NKT Cells and Strategy for the Generation of CD1d-Floxed ESCs, Related to Figure 2 (A) Flow cytometry gating strategy for the analysis of NKT cells (TCRb+ CD1d tetramer+), CD4+ T cells (TCRb+ CD1d tetramer- CD4+) and TfH cells (TCRb+ CD1d tetramer- CD4+ CXCR5+ PD-1+) in mediastinal lymph nodes at day 9 of influenza infection. (B) RT-qPCR analysis of IL-21 expression at day 3 and 9 of influenza infection by sorted NKT and TfH cells compared to CD4+ cells. (C and D) Representative contour plots show the expression of (C) Bcl-6 and (D) SLAM by NKT, TfH and CD4+ T cells at day 3 and 9 of infection. Quantification on the right charts shows the mean fluorescence intensity for (C) Bcl-6 and (D) SLAM. (E) Scheme showing the wild type and targeted CD1d locus containing two loxP sites flanking the exon 3 of CD1d as well as two FRT sites for the removal of the Neomycin resistance cassette. (F) Specific fragments amplified by long-range PCR for the 50 end (7.1kb) and 30 end (6.6kb) of the clones that correctly performed homologous recombination. (G) DNA from three positive clones was digested with SacI restriction enzyme and Southern blot performed using a labeled Southern probe. The upper band corresponds to the wild type locus (10.6kb) and the lower band to the targeted locus (7.3kb). (H) Southern blot of the three positive clones performed with a labeled Southern probe against the Neomycin cassette. Orange arrow indicates the single 5.2kb band recognized by this probe, indicating single insertion of the construct. The clone 2, in red, was chosen for injections. In all quantification charts, each dot represents one mouse. Data are representative from three experiments. Statistical analysis two-tailed Student’s t test; **p < 0.01, ***p < 0.001 and ****p < 0.0001. A B C **** Groups Day -1: Adoptive transfer 106 2400 80 1 IL4-GFP+ CD45.2 No adoptive transfer T T cells 104 + IL4-GFP- CD45.1 40 2 No adoptive transfer cells / LN 1200 + NKT cells / LN NKT 2 + 3 IL4-GFP- CD45.1 + IL4-GFP+ CD45.2 NKTs 10 Fold Increase N.D. NKT/CD4

Cytokine 0 - - 0 4 IL4-GFP CD45.1 + IL4-GFP CD45.2 NKTs

10 Cytokine 0 + + + + IFN-γ IL-4 IFN-γ IL-4 IL-4 IL-5 IL-13

D Lymph Nodes - Day 3 Group 1: IL-4 GFP+ CD45.2 mice Group 2: IL-4 GFP- CD45.1 mice

5 5 105 10 10 99.0% 58.1% 0% 0% NKT cells 104 104 104

103 103 103

β 0 0 0 CD45.2 TCR CD45.2 010103 104 5 010103 104 5 010103 104 5 010103 104 5 010103 104 5 CD1d-Tetramer CD45.1 IL-4 GFP CD45.1 IL-4 GFP

Group 3: CD45.1 mice + IL-4 GFP+ CD45.2 NKT cells Group 4: CD45.1 mice + IL-4 GFP- CD45.2 NKT cells

5 105 10 0.12% 75.0% 0.11% 33.3% 104 104

103 103

0 0 CD45.2 CD45.2 010103 104 5 010103 104 5 010103 104 5 010103 104 5 CD45.1 IL-4 GFP CD45.1 IL-4 GFP

E Lungs - Day 3 Group 1: IL-4 GFP+ CD45.2 mice Group 2: IL-4 GFP- CD45.1 mice

5 5 105 10 10 99.5% 56.7% 0% 0% NKT cells 104 104 104

103 103 103 β 0 0 0 CD45.2 TCR CD45.2 010103 104 5 010103 104 5 010103 104 5 010103 104 5 010103 104 5 CD1d-Tetramer CD45.1 IL-4 GFP CD45.1 IL-4 GFP

Group 3: CD45.1 mice + IL-4 GFP+ CD45.2 NKT cells Group 4: CD45.1 mice + IL-4 GFP- CD45.2 NKT cells

5 105 10 0.26% 87.5% 0.27% 22.2% 104 104

103 103

0 0 CD45.2 CD45.2 010103 104 5 010103 104 5 010103 104 5 010103 104 5 CD45.1 IL-4 GFP CD45.1 IL-4 GFP

(legend on next page) Figure S3. Analysis of the NKT Cell Source of IL-4, Related to Figure 3 (A) Quantification of IFN-g+ (blue) and IL-4+ (red) producing lymph node (left chart) or NKT cells (right chart) in mediastinal lymph nodes after 3 days of influenza infection. (B) RT-qPCR analysis of IL-4, IL-5 and IL-13 expression at day 3 of influenza infection by sorted NKT cells compared to CD4+ T cells. Each dot represents one mouse. (C) Adoptive transfer strategy for the analysis of the source of IL-4+ NKT cells after influenza infection. (D and E) Flow cytometry analysis at day 3 of influenza infection of (D) mediastinal lymph node or (E) lung NKT cells from mice that were treated as in (C) to detect IL4-GFP expression. Data are representative from two experiments. (legend on next page) Figure S4. Single-Cell RNA-Seq Analysis of IL-4+ Lymph Node Cells, Related to Figure 4 (A) Confocal microscopy of mediastinal lymph nodes sections at day 3 of influenza infection. Lymph nodes were labeled with PBS-57 loaded CD1d tetramer (magenta) and antibodies to B220 (white) and CD169 (green). IF, interfollicular. Scale bars, 300 mm (left); 60 mm (right). Chart on the right shows the quantification of NKT cells-macrophages contacts observed in the different areas of the lymph node at day 0, 1, 2 and 3 of infection in mediastinal lymph nodes. Data are representative from three experiments. (B) Flow cytometry analysis of IL4-GFP+ CD1d-tet- cells in mediastinal lymph nodes at day 3 of influenza infection showing TCRb- and TCRb + cells. (C) Single-cell RNA-seq heatmap showing the expression of 25 significant marker genes (Table S1) for the CD1d-tet+ clusters (columns 1 and 2) and CD1d-tet- clusters (columns 3 and 4). Single cells and genes were determined on the basis of unsupervised hierarchical clustering (Table S1).

(D) Expression distribution (violin plots) in each population (horizontal axes) for Ccr7, Sell, S1pr1, CD4, CD8a and Bcl6. Expression levels are Log2(tpm+1). (E) Flow cytometry analysis of IL4-GFP+ TCRb+ CD1d-tet- cells in mediastinal lymph nodes at day 3 of influenza infection showing CD4+ and CD8a + cells. (F) Representative contour plot show the expression of GATA-3 by NKT cells, ILC2s (Live/Dead- B220- CD3- CD5- CD11b- CD11c- CD127+ CD45+ GATA-3+) and CD4+ T cells at day 3 of infection. Quantification on the right chart shows the mean fluorescence intensity for GATA-3. (G) Flow cytometry analysis of RORgt+ and T-bet+ NKT cells in mediastinal lymph nodes at day 3 of influenza infection. Histogram shows the levels of CXCR3 in RORgt+ and T-bet+ populations. Dot graphs show the quantification of CXCR3+ NKT cells in the RORgt+ (red dots) and T-bet+ populations (blue dots), with lines connecting NKT cells from the same mouse. In graphs showing statistical analysis, data are representative of at least two independent experiments. Statistical analysis two-tailed Student’s t test; **p < 0.01, ***p < 0.001 and ****p < 0.0001. A Neutrophils/MDSCs Lyz2-Cre **** B 100 105 100 100 CD1d Lyz2-Cre- CD1d Lyz2-Cre+ - 4 Cre+ Cre 10 CD1d-/- 50 50 50 103 NFs/MDSCs (%) + 0 Events (% of max) Events (% of max) CD1d Gr-1 0 0 0 010103 104 5 0 103 104 105 Cre- Cre+ 0 103 104 105 CD11b CD1d CD1d

C Dendritic cells CD11c-Cre **** 105 100 100 Cre+ Cre- 104

3 50 DCs (%) 50 10 +

0 CD1d Events (% of max)

MHCII 0 0 3 4 5 01010 10 3 4 5 - 0 10 10 10 Cre Cre+ CD11c CD1d

D Wild type CD1d-/- CD1d Lyz2-Cre+ CD1d CD11c-Cre+

CD169

CD1d

E CD1d Lyz2-Cre- CD1d Lyz2-Cre+ F CD1d Lyz2-Cre- CD1d Lyz2-Cre+ 50 105 12 105 7.6% 7.9% .26.8 25.2 % 104 104 T cells (%) T +

B cells (%) 6 25 3 103 + 10 PD-1 +

GL7 0 0 + CXCR5 Fas 0 Fas

0 CXCR5 3 4 5 0 103 104 105 01010 10 Cre- Cre+ Cre- Cre+ GL7 PD-1

+ G CD1d CD11c-Cre- CD1d CD11c-Cre+ * H CD1d CD11c-Cre- CD1d CD11c-Cre 20 105 12 105 6.6% 8.2% 13.6% 11.4 % 104 104 T cells (%) T + 10 B cells (%) 6 3 103 + 10 PD-1 +

GL7 0 0 + CXCR5

Fas 0 CXCR5 Fas 0 3 4 5 3 4 5 0 10 10 10 - + 01010 10 Cre- Cre+ Cre Cre GL7 PD-1

(legend on next page) Figure S5. Conditional Targeting of CD1d in the Myeloid Lineage, Related to Figure 5 (A–C) Flow cytometry plots showing the gating strategy for (A) Gr-1+CD11b+ neutrophils (NFs) and myeloid-derived suppressor cells (MDSCs) and (C) MHCII+CD11c+ dendritic cells (DCs). Red and blue histograms depict the levels of CD1d in the gated populations in (A) CD1dflox/floxLyz2-Cre- versus Cre+ and (B) CD1d/ mice or (C) CD1dflox/exCD11c-Cre- versus Cre+ mice. These quantitative differences are depicted via dot plots; blue and red dots show CD1d+ cells in the Cre- and Cre+ animals, respectively. (D) Confocal microscopy of mediastinal lymph nodes sections from wild type, CD1d/, CD1dflox/floxLyz2-Cre+ and CD1dflox/exCD11c-Cre mice. Sections were stained with antibodies to CD169 (green) and CD1d (magenta). Scale bars, 60 mm. (E–H) Flow cytometry analysis of mediastinal lymph node cells at day 9 of influenza infection. Representative contour plots show the percentage of B cells and T cells from (E-F) CD1dflox/floxLyz2-Cre- and Cre+ or (G-H) CD1dflox/exCD11c-Cre- and Cre+ mice that acquire (E and G) germinal center markers, Fas and GL7, or (F and H) TfH cell markers, CXCR5 and PD-1. Quantifications are shown in dot plots, Cre- mice (blue dots) and Cre+ mice (red dots). In all panels, each dot represents a single mouse. Horizontal bars – mean; error bars – SEM. Data are representative of three independent experiments. Statistical analysis, two tailed Student’s t test, *p < 0.05, ****p < 0.0001. SCS Medulla A 30 80

IL-1β IL-1β + + IL-18 IL-18 IL-33 IL-33 CD169 CD169 + 15 + 40 cells / section cells / section Cytokine Cytokine 0 0

B ** 20 CD169 GFP Langerin GFP CD169+ cells B220 B220 B220 B220 Langerin+ cells

10 cells/section + GFP 0

C Mixed bone marrow chimeras D Mixed bone marrow chimeras

-/- Wild type-CD45.1 Wild type-CD45.2 Wild type-CD45.1 MyD88 -CD45.2 ** 14 14 105 105 14.0% 13.1% 10.8% 4.9% 104 104 7 7 103 103 NKT cells (%) NKT NKT cells (%) NKT + + β β 0 0 IL-4 IL-4 TCR TCR 0 0 0 103 104 105 WT WT 010103 104 5 WT MyD88/- IL-4 IL-4

E Mixed bone marrow chimeras

-/- Wild type-CD45.1 IL-18R -CD45.2 * 22 105 15.8% 10.4% 104 11 103 NKT cells (%) NKT +

β 0 IL-4

TCR 0 010103 104 5 WT IL-18R-/- IL-4

Figure S6. Intrinsic Role of MyD88 and IL-18R in NKT Cell Production of IL-4, Related to Figure 6 (A) Quantification of IL-1b+, IL-18+ and IL-33+ cells in the SCS and medullar areas from the images shown in Figures 6C and 6D. (B) Confocal microscopy of mediastinal lymph nodes from wild type mice that were infected with 200 PFU of PR8 NS1-GFP and sacrificed after 3 days. Sections were stained with antibodies to CD169 (red), B220 (gray) and Langerin (magenta). Scale bars, 60 mm. Quantification of GFP+ cells is shown in the right chart. (C–E) Flow cytometry analysis of chimeric mice that were generated by transferring a 1:1 bone marrow mixture of wild type (CD45.1+) and (C) wild type (CD45.2+), (D) MyD88/ (CD45.2+) or (E) IL-18R/ (CD45.2+) to sub-lethally g-irradiated CD45.1 recipient mice. Eight weeks after bone marrow transplant, mice were infected with 200 PFU of influenza virus and mediastinal lymph nodes were harvested after further 3 days. Flow cytometry panels show the production of IL-4 by CD45.1+ and CD45.2+ NKT cells in the different sets of chimeras. Dot graphs show the quantification of IL-4+ NKT cells in the CD45.1+ population (blue dots) and the CD45.2+ population (red dots), with lines connecting the two NKT cell subpopulation coming from the same chimera. In all panels, each dot represents a single mouse. Horizontal bars – mean; error bars – SEM. Data are representative of two independent experiments. Statistical analysis, two tailed Student’s t test, *p < 0.05, **p < 0.01. A Uninfected α * 0.4 IgD Uninfected

GL7 area + Control α-IL4 (d0-d3) 0.2 area /IgD +

GL7 0

B Control α-IL4 (day -3) C D Lymph node 8 105 105 Endogenous OT-II 7.8% 6.4% i.v OT-II T cells Harvest mLNs 104 104

B cells (%) 4 + 103 -1 0 7 103

GL7 Days 0 + 0 Fas

0 CD4 Fas 010103 104 5 Control α-IL4 010103 104 5 GL7 (day -3) CD45.1 E F Endogenous T cells OT-II T cells ** Wild type + OT-II CD1d-/- + OT-II * 5 100 2.4 10 105 2.1% 0.9% 104 4 T cells (%) T 10 cells (%) + 13% 84% + 50 1.2 103 103 PD-1 +

0 CXCR5 0 + 0

CXCR5 0 GL7 CXCR5 CXCR5 3 4 5 01010 10 Endogenous OT-II 0 103 104 105 WT CD1d-/- PD-1 GL7

G Oligomycin FCCP Rotenone H Oligomycin FCCP Rotenone **** **** 120 110 **** * 60 60

60 55 30 30 OCR (pmol/min) OCR (pmol/min) Max OCR (pmol/min)

0 Max OCR (pmol/min) 0 0 0 0306090120 No stimulus 0306090120 Non-draining LN Minutes + IL-4 Minutes Mediastinal LN α-IgM Mediastinal LN / α-IL-4 α-IgM + IL-4

I Zika virus Harvest LNs and collect serum K IL1R1 infection Gene expression and NAbs 0.015429415 SMOX ABI3BP JAG1 Shared genes 0.011072036 NKT cells 0.0100437090.011657506 0.010829483 EMB 0.01519685 S100A4 SERINC20.016368505 0.01318893 DAPK2 HSD11B1 PRELID2 0.010493873 0.0162913240.0270938910.0140824690.011003553 SDC1 0.013271378 IL12RB2 0.015700813 LAX10.011594475 GLT25D2 0.010854919 0.0214360130.025380895 IL2RB 0.015635138 014 0.016046662TNFRSF210.011596566 0.021871060.024041943 0.015477868 IGF1R 0.0118996 RGS1 0.015140184 0.012223331 0.01298504 0.012155684 IL-18 MMP25 0.021018473 0.0106151250.0103104850.011974632 ITGB5 LY6E PPM1J0.0122959940.010576856ITGB4 SEPP1 0.012392153 0.014356517 0.012749335 0.0126660950.0155471340.0148451560.015786204 0.010053013KLRD1 0.012121715 Lymph node day 14 vs. Neutralizing antibodies day 14 IL17RE0.014697292 0.0138834380.017748201STOM 0.021536583 0.014257311 J NSA2 0.015470931 0.019345129 DOPEY2 0.011824749 0.01838226 JUN 0.019416662 CXCR3 0.0119885230.013870903TNS4 DLG5 0.010203674 CHAD0.012169804 NOP58 0.010741385 0.011468454 0.024453746 0.01119987 0.011534583 0.012416903 0.0124849360.018724404 MYCN LEF1 NKT.POPULATIONS NR1H2 0.0264291520.010380626 0.013281825 REXO4 0.0188848260.018384848 0.0180182890.010821869NPL 0.0247662330.0123720960.012952557 NKG7 0.0108033880.016697038 0.0183807370.014397028 FGL20.014366682 0.0244089180.0198466030.014546911 0.011141850.0168531310.012966084 0.0126803810.0162420010.0153589850.0144780710.015070294 LZTFL1 WDR76 0.015073859 0.0219151430.0133361510.0122168170.015909238RAMP3 0.012319515 0.022630535 PIAS2 0.021525236 0.0156227560.0117974410.0178805530.0127865390.015282566 0.02527271 0.012307171 0.0179296360.016033484 0.0120169310.0214925370.01246645 0.01084820.012140085 0.0289434050.011852172 KCNK1 0.012918495 TOP50. MOST.DEG.NKT1 0.0194918810.0150899860.0129546310.0180274550.011699469 0.012230240.013973099 SAMM50 0.0188610130.0118581720.0165161790.022146184 0.015702274 0.014758236 0.011489145 0.021689225IL23R0.013410385 0.012160562 0.023279125 0.0130068360.0143854120.0101141420.0141889030.0112003540.014758769 0.011820130.013282981 0.014965060.012266212 HAUS3 0.015487644 0.0144429640.0121525310.014263560.012382553 0.010942082 0.012470361 0.016195277 0.0120676390.012702811 0.025139615 0.010597174 ACD 0.022872396 0.0109117040.01923993 0.0131176650.015236114 0.025105948 0.010548177 B cells TOP50. MOST.DEG.NKT2 CREBBP 0.015949557 0.011752614 0.017569305 BLK0.015386345 0.011562778 0.0175981220.0130232360.015751053 0.022727065 0.019486673 0.011779791 0.017603973 0.019264285 0.0173534060.01193442 0.017691877 0.013915279 0.0128353890.010221882 0.010677717 0.013202892 0.0119975140.01148795 0.013382702 C8A 0.0153599120.0127961790.018407391 MAPK9 RASA2 0.016460529 0.010473339 0.0152534690.010072585 0.0131673380.0103460370.017676152 0.0195683020.0190089620.019858915 0.014969677 0.013622384 0.012216643 0.015445159 0.0175319560.018375628 0.011319747 TOP50. MOST.DEG.NKT17 IL-4 0.020095417 0.0177993220.017631747STAT40.010765362 0.0148937640.0122152130.010998044 0.0105764620.010426386FGF180.0118459510.010140981IKBKB 0.012525561 0.011709638 0.015953310.022537433 SYK 0.01975779 0.010710160.016110955 CDH17 0.015109989 0.0136195850.011542165 FCER1G0.024360010.010239518CD8B 0.0183863470.019401101 0.012266147 0.016478315 0.012903953 0.019934522 0.01932949 0.0147567410.011317985 0.010143301 0.012060402 0.0106490920.01794713 0.0163672530.017516598 0.016142515 0.011454908 BCL11B0.012196074 0.012017687 0.0160105230.01586429 0.0135583990.0110140570.018996222 0.017370477HIPK2 0.0156867540.018122507 0.0154724110.0180859270.0110263720.014927189 0.0110675370.014162485FGFR3 0.012560469 0.014118372 0.0121230060.016973529 0.020435147 0.0121084630.010864384 0.011351636 0.014697029 0.0174610970.016408242 0.0182606690.0167561270.0121629020.016515503 0.014413077 0.011005486 0.010123569 0.013609584IL7R0.013506545 0.0153415960.0201090160.0146428980.0131417130.0100166780.0139359880.01288066 IL10 0.022851929 0.017502153 0.0153076210.014057010.0138260340.014295010.013029292 0.0114115510.0198111740.011899780.014387260.015154514 0.0111798660.017454084 ITGB1 IL18 SIGNALING SOS1 0.014178259 0.01924158 0.0100226040.0186701550.028544348 0.013778259IL6R 0.0111331870.0120715740.014570290.0177713580.011157078 0.018555671 PTPN6 DOK2 0.0135043130.0154190940.0269621440.020613749 IL18R1 0.0179832060.0170941930.01050905 ID2 0.0100537730.015855901 0.012062198 0.0145326680.018380629 0.016881403 0.010650929 0.021557692 0.0152425110.0122658690.014572250.01909311 IL4 0.0124631350.02075929 0.014598424FOS 0.0222833580.0134308520.020470127 0.010140702 CD46 0.0142620810.0211269820.0238756910.0168478640.0122345560.0118075130.0148021550.015290955 0.0110525490.0108041540.01646053 BLNK 0.01796581 0.020615323 0.0116153880.020834110.015919045 0.0165290850.018186785 0.022675749 0.015857125 CR1 CD280.0162525140.0175384730.012511780.014564360.0101601250.0205415230.020005240.015617610.010308822 0.0124958740.0214755370.022714070.0271552610.013313063 IL4.SUPERPATHWAY.GENECARD 0.019055948 0.02225820.012674276 0.0100786540.013444564CD247 0.015290309 0.0161888920.0112485830.0108177570.012516972 0.0119189260.016241718IKBKG 0.017606745 0.0121280190.0118290930.0127854440.0123757390.011458408 MALT1 0.023192078 0.010252288 0.0102315670.012519578 0.0246365960.0102151940.012631252 0.0199668970.016800420.011609847 RBPJ GTF3A HTT0.011128846 0.0119721740.013337469 0.027142730.0190816260.0151539430.0114608280.012768193 0.01639617 0.013472381 NAA15 0.018284597 0.0102424750.0155352070.0143856760.0180947670.022541858 0.023352617 0.010160251 0.0102237220.021056995 0.013588131 AKT20.0160408520.018218767 0.0134785930.0161377860.0112265290.0110380580.010315209 0.0100707490.0102257590.011549573 0.011943626CBLB 0.0190432180.019026786 ARHGAP25 0.0175725740.0165998660.0167524260.013262890.023438638 0.010816079 0.017233074 PLCL2 0.021150390.0129591250.019742930.02288302 0.0191901070.0253615680.014396441 0.0103918660.0107868070.0202729270.0121206730.018868322 MFNG STAT6 CHIP_SEQ TARGET GENES SOCS5 0.015965710.025012394 0.017860660.019275704ST6GAL10.0101642270.012217976 CD3G PIAS4 0.013408039 0.0121182130.0134563340.0138299240.013662432AKT30.0164948220.0114867040.0112565140.018102817ICOS 0.0134850890.011639956 0.021956917 0.011012814 0.0204111620.0125751640.010278405 0.019865047 0.0112822020.01194442 0.016143924 0.017440299 IL2RG0.0196675120.0130201460.0135638420.020390745 0.018879987 0.012880244 0.015273024 0.010432524 LTA SELP 0.0118505830.0146428920.015034821 0.011716519 0.0190567080.024837721 0.010674533 0.0239258280.0119013930.016336827 0.0127270010.018463304 0.0151973260.012124317 0.0109716210.0152359920.0151559390.0165183790.011014584 0.0233187530.01215963 TFH SIGNALING 0.010752411 EPAS10.012762003RBMS1 0.011232376 0.014245229 IL160.0170149840.0142017810.0118021490.012606612 ABL1 0.0111486870.017110558 0.0198804660.0110369120.0203122310.014719022 0.0146499930.0106632680.012528501 C7 0.0107111430.0205395150.0133769610.020390470.0167215150.012134356DPP40.0164813980.0112458460.0156291050.0219840560.01192374LY960.0117334020.012453869 0.017748993 C2 SETX 0.0182214270.0136363620.0101620550.021171153NSMCE10.015989935 0.0105913 0.0153086630.010500984 0.012606218 0.012605351 JAK2 0.011163474 0.024552554SOCS30.0143294860.01341601 0.012395194 ILK MTOR 0.0232386680.0177785960.018055838 0.0163617170.011438731 0.0123226480.010138639 0.014466414 0.0169634540.0116503040.0207860950.0153233340.0207018090.012649393 0.011216347 0.013150808 0.0120050430.0118528830.010127493 0.0139728930.011313702 0.013766988 0.0117301750.01141884 0.010904187 0.0191180350.019449813 0.0128526060.017696970.0143922330.014355966 0.014572669 0.011243172 0.0108474530.027871920.013832759 DARS2 F13A1 0.014820142 0.013721234 0.014805953BCL20.0222390440.013660352 0.0137369160.0258443080.01077254ITGAL 0.010508782 0.020039102 0.023584234 0.017119745 0.016666163 0.013598777 STK40.0135089840.013023692 0.0105908190.0114460120.017499158 0.018833889CD37 0.01040281 0.0139914960.0148177310.0111252190.022896403 0.0164220130.0137092320.0123520590.0102831720.0168711 CTLA4 SIGNALING BY BCR 0.0170229730.016328994 0.0136315230.0201634520.012944256 0.0236028130.0100550120.012904036 0.0131838730.0216673310.0107288080.0139752580.012071137 0.0169122 0.014223583 0.0172687750.0124033850.014885793 0.0124414350.012638848 0.010519140.0197128450.016470721 0.0169963820.011987714 0.0110179040.02045104 0.012936182 0.0187703450.012672219 0.0133708950.0109940260.012570350.01305955 0.0111325 0.010141686 0.010110741 0.011000860.015244135ITK 0.012497327 0.0106477170.0184526390.017339904 0.013265601 0.010056509 0.016414134 0.011991788 0.019989913PPP3CA PDZD2 0.01257054 0.022676373 0.0105340980.012417934 0.0141636080.0113356990.010221333PRKCH0.0147033220.01361251 B_CELL_DIFFERENTIATION 0.0162500930.019746821 0.015395921 0.022504530.012559289 0.020646857 0.012650777 0.015612876 0.0110094770.0115091130.0148951730.01735941 0.0223699530.0194265060.0136112160.017876709 0.0154067110.0130920940.012618517LAT2 0.011079663 0.0129453940.013177319 0.01068719 0.028772091 0.013262910.01084340.0127856710.015565435 SENP7 0.018641843 0.0119974840.016253738 0.016628686 0.015906308 PLCG1 0.014832935 0.012140101 0.021482557NFAT50.0114222380.0137965930.012712552PIK3R1 0.02210078 0.013721783 GO_B_CELL_PROLIFERATION PIK3R5 0.010092102SDK10.0195135230.018580282 0.0151379650.013783992 THADA 0.01341504 0.014050107 0.024038987RAC1 ELMO1 KIAA0748 0.020789623 0.013748328 0.0101325350.011343884 0.021527624 0.019629660.012073907 0.0102873610.0227205250.0119254660.017174827 0.010576494 0.020213980.02306794 0.010089644 CD84 GO_B_CELL_ACTIVATION NUP210 0.013633448 0.0107499580.0146534620.013000086PTPRC ISY1 CARD110.012786892PRKCD 0.018419405ATP2A1 0.0171570720.022594852 0.017203087 MYH10 0.015307196IL24 0.0119200260.014008874 0.010614986 POLR1A 0.016009180.024862885 EOMES NFATC2 0.016881710.013279185 TfH cells 0.011927539 PRKD3 0.0132440640.012809452 STAT6 CA6 0.012110713 0.01189715LCP2 0.0110563570.01206414NFKBIB Genesets normalized enrichment score 0.012805958 0.013091532 NFATC3 0.010461881 LRRC16A RORA HIPK10.013081491 CUBN 0.019292485 0.012091244 TMEM68 TXK 0.017498158TLR1 DDX26B 0.010950093 VAV1 RAB7A ZCCHC6 -2 0 +2 STK17B

(legend on next page) Figure S7. Importance of Early IL-4 Production on B Cell Immunity, Related to Figure 7 (A) Representative confocal microscopy images of mediastinal lymph node sections from mice that were treated as in Figure 7E. Sections were stained with antibodies to the B lymphocyte marker, IgD (green) and the germinal center marker, GL7 (magenta). Scale bar, 450 mm. Quantification of the germinal center area in confocal images by ImageJ is shown in the right chart. (B) Flow cytometry analysis at day 8 of mediastinal lymph nodes from wild type mice that were treated with PBS or anti-IL4 blocking antibody on day 3, and infected with 200 PFU of influenza virus at day 0. Contour plots show Fas+GL7+ germinal center cells. The responses are quantified in right charts. (C) Experimental scheme showing the strategy for the generation of antigen-specific TfH cell precursors. (D–F) Flow cytometry analysis of mediastinal lymph node cells from mice treated as in (C). Representative contour plots display (D) the gating strategy used to differentiate endogenous CD4+ T cells from adoptively transferred OT-II cells, (E) the percentage of endogenous and exogenous cells bearing the TfH cell markers PD-1 and CXCR5 and (F) the percentage of B cells that acquire GL7 marker. Quantifications are shown on the right charts. (G and H) OCR at baseline and after sequential treatment with oligomycin, FCCP and Rotenone of B cells that were (G) stimulated ex vivo with anti-IgM and/or IL-4 or (H) isolated from lymph node cells of influenza-infected mice (day 3). Bar charts show the respiratory capacity obtained as (OCR after FCCP)- (basal OCR). Each dot represents an individual measurement. (I) Experimental scheme showing Zika virus infection strategy and time points at which lymph nodes and serum were collected from macaques. (J) Linear regression model of gene signatures for NKT cells, TfH cells, B cells, IL-4, STAT6 and IL-18 from harvested lymph nodes at day 14 with neutralizing antibodies measured in plasma at day 17. Red squares represent positive correlation, blue squares represent negative correlation and white squares represent no significant correlation. (K) Gene interacting network inference using GeneMANIA software representing interconnectivity of NKT, TfH cells, B cells, IL-4, STAT6 and IL-18 pathways and the co-expression of their leading genes. Unless stated, each dot represents a single mouse. Horizontal bars – mean; error bars – SEM. Data are representative of two independent experiments. Statistical analysis, two tailed Student’s t test, *p < 0.05, **p < 0.01 and ****p < 0.0001.