CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming

Article

CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming Graphical Abstract

Authors Anna Brewitz, Sarah Eickhoff, €hling, ..., Marco Colonna, Sabrina Da Ronald N. Germain, € ller Wolfgang Kastenmu

Correspondence [email protected]

In Brief pDCs and XCR1+ dendritic cells are critical for the generation of antiviral CD8+ T cell responses. Brewitz and colleagues demonstrate that primed CD8+ T cells reorganize the intranodal dendritic cell network to optimize pDC and XCR1+ DC cooperativity and thereby enhance CD8+ T cell immunity.

Highlights d

CXCR3 and CCR5 selectively control intranodal pDC migration

d

CD8+ T cells instruct pDC recruitment via CCL3 and CCL4

d

CD8+ T cells directly recruit XCR1+ DCs via XCL1

d

Active colocalization of XCR1+ DCs and pDCs supports DC cooperativity

Brewitz et al., 2017, Immunity 46, 1–15 February 21, 2017 Published by Elsevier Inc. http://dx.doi.org/10.1016/j.immuni.2017.01.003

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

Immunity

Article CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming €hling,1 Thomas Quast,2 Sammy Bedoui,3 Richard A. Kroczek,4 Anna Brewitz,1,14 Sarah Eickhoff,1,14 Sabrina Da Christian Kurts,1 Natalio Garbi,1 Winfried Barchet,5 Matteo Iannacone,6 Frederick Klauschen,7,8 Waldemar Kolanus,2 € ller1,15,* Tsuneyasu Kaisho,9,10,11 Marco Colonna,12 Ronald N. Germain,13 and Wolfgang Kastenmu 1Institute

of Experimental Immunology, University Hospital, University of Bonn, 53127 Bonn, Germany Immunology and Cell Biology, Life and Medical Sciences Institute, University of Bonn, 53115 Bonn, Germany 3Department of Microbiology and Immunology, The University of Melbourne, Peter Doherty Institute for Infection and Immunity, Parkville, VIC 3010, Australia 4Molecular Immunology, Robert-Koch-Institute, 13353 Berlin, Germany 5Institute of Clinical Chemistry and Clinical Pharmacology, University Hospital, University of Bonn, 53127 Bonn, Germany 6Division of Immunology, Transplantation and Infectious Diseases and Experimental Imaging Center, IRCCS San Raffaele Scientific Institute and Vita-Salute San Raffaele University, Via Olgettina 58, Milan 20132, Italy 7Institute of Pathology, Charite €tsmedizin Berlin, Charite´platz 1, 10117 Berlin, Germany ´ Universita 8Einstein Foundation Berlin, 10117 Berlin, Germany 9Department of Immunology, Institute of Advanced Medicine, Wakayama Medical University, Wakayama, Wakayama 641-8509, Japan 10Laboratory for Inflammatory Regulation, RIKEN Center for Integrative Medical Sciences (IMS-RCAI), Yokohama, Kanagawa 230-0045, Japan 11Laboratory for Immune Regulation, World Premier International Research Center Initiative, Immunology Frontier Research Center, Osaka University, Suita, Osaka 565-0871, Japan 12Department of Pathology and Immunology, Washington University School of Medicine, St Louis, MO 63108, USA 13Lymphocyte Biology Section, Laboratory of Systems Biology, National Institute of Allergy and Infectious Diseases, NIH, Bethesda, MD 20892-0421, USA 14Co-first author 15Lead Contact *Correspondence: [email protected] http://dx.doi.org/10.1016/j.immuni.2017.01.003 2Molecular

SUMMARY

INTRODUCTION

Adaptive cellular immunity is initiated by antigenspecific interactions between T lymphocytes and dendritic cells (DCs). Plasmacytoid DCs (pDCs) support antiviral immunity by linking innate and adaptive immune responses. Here we examined pDC spatiotemporal dynamics during viral infection to uncover when, where, and how they exert their functions. We found that pDCs accumulated at sites of CD8+ T cell antigen-driven activation in a CCR5-dependent fashion. Furthermore, activated CD8+ T cells orchestrated the local recruitment of lymph node-resident XCR1 chemokine receptor-expressing DCs via secretion of the XCL1 chemokine. Functionally, this CD8+ T cell-mediated reorganization of the local DC network allowed for the interaction and cooperation of pDCs and XCR1+ DCs, thereby optimizing XCR1+ DC maturation and cross-presentation. These data support a model in which CD8+ T cells upon activation create their own optimal priming microenvironment by recruiting additional DC subsets to the site of initial antigen recognition.

The generation of an adaptive immune response involves multiple myeloid and lymphoid cell populations that act in a highly orchestrated manner to permit optimal cellular interaction and communication (Qi et al., 2014). The critical interaction that initiates an adaptive cell-mediated immune response occurs between antigen-bearing dendritic cells (DCs) and T cells, leading to proliferation and differentiation of the latter. Key inputs driving such responses involve signaling through the T cell receptor (signal 1), costimulatory receptors (signal 2), and receptors for inflammatory cytokines (signal 3), along with the actions of chemokines and other chemoattractants that fine-tune the localization of lymphocytes and DCs within secondary lymphoid organs (Chen and Flies, 2013; Curtsinger and Mescher, 2010; Fooksman et al., 2010; Thelen and Stein, 2008). Besides the interaction between antigen-bearing DCs and T cells, other cell types contribute to optimal cell-mediated responses (Bendelac et al., 2007; Martı´n-Fontecha et al., 2004; Veiga-Parga et al., 2013). Chief among these additional players are plasmacytoid DCs (pDCs) (Yoneyama et al., 2005). pDCs morphologically resemble lymphoid cells rather than myeloid cells and on a single-cell level are known for their ability to produce high amounts of interferon (IFN) type I (Swiecki and Colonna, 2015). For this function, expression of toll-like

Immunity 46, 1–15, February 21, 2017 Published by Elsevier Inc. 1

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

receptor-7 (TLR7) and TLR9 located in endosomal compartments of pDCs are critical. This allows them to sense viral infections irrespective of viral replication within the pDCs. pDC activity enhances adaptive antiviral CD8+ T cell responses and also contributes to innate host defense by inhibiting viral replication during both acute and chronic viral infections (Cervantes-Barragan et al., 2012; Swiecki et al., 2010). Currently, it is unclear how pDCs exert these dichotomous functions (Reizis et al., 2011a). While they are the most prodigious producers of IFN-I on a per cell basis, various other cell types, especially macrophages, can produce IFN-I upon viral infection. Indeed, the total amount of IFN-I in the serum of mice in various experimental systems is unaltered in the presence or absence of pDCs (Barchet et al., 2005; Xu et al., 2015). Moreover, pDCs appear to be only a transient source of IFN-I within the first few days of an infection, with the cells then either being eliminated or inhibited via an IFN-I feedback loop (Swiecki et al., 2011). Some data indicate that pDCs may support anti-microbial resistance by factors other than IFN-I (Ang et al., 2010), while other studies suggest that pDCs can contribute to adaptive immunity by acquiring, processing, and presenting antigens directly to T cells (Di Pucchio et al., 2008; Mourie`s et al., 2008; Villadangos and Young, 2008). Nevertheless, the majority of published work supports the notion that the primary effector function of pDCs lays within their capacity to produce IFN-I (Haeryfar, 2005; Reizis et al., 2011b). Given that many other leukocytes can produce IFN-I, the question then arises as to what the special role of pDC IFN-I production might be. One possibility could be the ability to produce large amounts of IFN-I very rapidly or the capacity to secrete a wider array of IFN-a (Izaguirre et al., 2003). However, using vesicular stomatitis virus (VSV) as model, subcapsular macrophages and not pDCs proved to be the critical early IFN-I source involved in host protection (Iannacone et al., 2010). Therefore, we hypothesized that besides the timing or amount of IFN-I production by pDCs, their motility and capacity for relocalization might allow them to deliver IFN-I at specific sites to promote host defense in a unique manner (Asselin-Paturel et al., 2005). This notion is in line with the numerous chemokine receptors expressed on pDCs that could direct their migrational pattern during infection (Miller et al., 2012). Given this idea, we felt it important to characterize the spatiotemporal dynamics of pDCs during an ongoing immune response. To this end we analyzed pDC migration patterns in the lymph node (LN) using dynamic intravital 2-photon microscopy (IVM). We found that upon viral infection, pDCs migrated either to infected macrophages residing in the subcapsular sinus (SCS) area in a CXCR3-dependent manner or to CD8+ T cell priming sites in a strictly CCR5-dependent manner. We show that pDCs were recruited to sites of functional MHCI antigen presentation by CCL3 and CCL4 chemokines that were produced in the context of productive conventional DC (cDC)-CD8+ T cell interactions. Additionally, CD8+ T cell activation led to a rapid production of XCL1, which in turn attracted LN-resident XCR1+ DCs to CD8+ T cell priming sites. The co-presence of pDCs, XCR1+ DCs, and CD8+ T cells within a confined microenvironment allowed for optimal signal exchange, in particular involving pDCderived IFN-I. This pDC-derived cytokine optimized the maturation of and cross-presentation by XCR1+ DCs, enhancing the 2 Immunity 46, 1–15, February 21, 2017

developing CD8+ T cell response. These data reveal a biological concept in which T cells actively orchestrate the local development of an optimal priming environment upon their initial successful identification of an antigen-bearing cDC. RESULTS pDCs Maximize CD8+ T Cell Responses To understand how pDCs contribute to antiviral immune responses, we used modified vaccinia virus Ankara (MVA) to study CD8+ T cell responses in the presence or absence of pDCs. This attenuated cytopathic virus strain does not replicate in vivo and therefore induces a highly synchronized immune response in space and time. Additionally, when manipulating the IFN-I system, secondary effects on viral replication and thus antigen abundance that impacts the ensuing CD8+ T cell response may be neglected in this experimental approach. Clec4C is highly expressed by pDCs and Clec4c+/DTR mice provide a means to selectively deplete pDCs by administration of Diphtheria toxin (DTX) to the animals (Figures S1A–S1D; Swiecki et al., 2010). After footpad infection, we found a significant reduction in both the relative and absolute numbers of antigenspecific CD8+ T cells directed against the immuno-dominant epitope B8R20 in the absence of pDCs (Figures 1A–1C). Among the smaller cohort of activated, antigen-specific CD8+ T cells present on d8 after infection in the DTX-treated animals, we detected only modest alterations in CD8+ T cell differentiation using standard markers for various effector subsets (KLRG1 and CD127) (Figures 1D and 1E). Functionally, there were reduced numbers of IFN-g-producing CD8+ T cells after in vitro restimulation with B8R20 peptide (Figures 1F–1H). Among those IFN-g-producing CD8+ T cells, there were no significant changes in polyfunctionality (cells additionally producing TNF-a and/or IL-2) (Figures 1I and 1J). These data indicate that, in accord with earlier studies, pDCs help to promote the magnitude of the CD8+ T cell cohort responding to viral infection, while having a limited impact on the differentiation and functionality of the reduced number of activated T cells. pDCs Are Recruited to Infected Macrophages and CD8+ T Cell Priming Sites To examine whether the specific localization of pDCs contributes to these effects on CD8+ T cell responses and if so, how such behavior compares to the activity and positioning of pDCs involved in mediating innate host defense, we employed SiglechGFP/GFP mice that have fluorescently marked pDCs (Swiecki et al., 2010). IVM was used to visualize and track these cells in the steady state and after infection in the LN. In these mice, GFP is also expressed in medullary macrophages of the LN, but these cells can be easily discriminated from pDCs based on the brightness of the GFP signal (macrophages are dim) and the morphology of the cells (Figure S2A). The MVA used for infection encoded the protein antigen ovalbumin (OVA), allowing us to use OT-I T cell receptor (TCR) transgenic OVA-specific CD8+ T cells to follow the development of an adaptive immune €ller et al., 2013). response at the same time (Kastenmu In the steady state, pDCs were distributed throughout the LN paracortex and the interfollicular area (Figure 2A) with preference around high endothelial venules, similar to human tissue

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

B

DTR 4.28 ± 0.96

B8R multimer

7.56 ± 0.18

10 8 6 4 2 0

WT TE

DP

DN

E

DTR

MP

TE

DP

DN

MP

B8R multimer + (%)

60

5 0

WT DTR

20 0

DTR

DN MP DP

5.97 ± 1.24

H

G 8 IFN- + (%)

CD44

***

TE

WT

3.78 ± 0.93

*

15

6 4 2

10 5

WT DTR

DTR

TNF-

WT DTR

J 60 IFN- + (%)

WT

*

0

0

IFN-

I

10

40

CD127

F

***

WT DTR

IFN- + (x105)

D

15

WT DTR

CD8

KLRG1

C

***

B8R multimer + (x105)

WT

B8R multimer + (%)

A

WT DTR

40 20 0

IL-2

TNF- TNFIL-2

IL-2

Figure 1. pDCs Maximize CD8+ T Cell Responses (A–E) Analysis of B8R20-specific CD8+ T cell responses d8 p.i. (MVA OVA f.p.), comparing WT and Clec4c+/DTR animals treated with DTX (0.1 mg/day; d 2 to d3), showing (A) representative B8R20 multimer staining and (B) relative and (C) absolute numbers of CD8+ T cells specific for B8R20. Relative distribution of B8R20specific memory subsets indicated by the markers KLRG1 and CD127, shown as representative original plots (D) and as relative numbers (E). (F–J) Cytokine production of CD8+ T cells upon in vitro restimulation with B8R20 peptide showing (F) representative plots and (G) relative and (H) absolute numbers of IFN-g-producing CD8+ T cells. Polyfunctionality of CD8+ T cells (gated on IFN-g+) shown as (I) representative plot or (J) relative numbers. Abbreviations are as follows: TE, terminal effector cells; DN, double-negative cells; MP, memory precursor cells; DP, double-positive cells. Data are representative of four independent experiments analyzing at least three mice per group. Error bars indicate the standard deviation. ***p % 0.001, *p % 0.05. See also Figure S1.

Immunity 46, 1–15, February 21, 2017 3

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

Figure 2. pDCs Are Recruited to Infected Macrophages and CD8+ T Cell Priming Sites (A) Immunofluorescent (IF) images of the popliteal lymph node (pLN) from a SiglechGFP/GFP mouse showing the distribution of pDCs in the steady state. (B) Image from intravital microscopy (IVM). (C–E) Steady-state migration analysis of pDCs and OT-I T cells in the interfollicular area showing the speed (C), translated tracks (D), and track displacement (E). (F) IF images of the subcapsular sinus (SCS) area 10 hr p.i. (G) Spatial frequency distribution of pDCs in the LN before and after infection. (H) Image from IVM showing pDCs and infected macrophages 9 hr p.i. (I–K) Migration analysis of pDCs at the SCS showing speed (I), translated tracks (J), and track displacement (K) p.i. (L) IF images of the LN paracortex 10 hr p.i. showing OT-I T cell and pDC co-cluster. (legend continued on next page)

4 Immunity 46, 1–15, February 21, 2017

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

(Colonna et al., 2004). Morphologically, pDCs resembled lymphocytes and migrated actively at an average speed of 5 mm/min (Figures 2B and 2C and Movie S1). Track length and displacement was significantly lower compared to naive CD8+ T cells (Figures 2D and 2E), indicating that pDCs show less extensive tissue scanning as compared to T lymphocytes in LN. Eight hours after infection, we found that pDCs had translocated to two distinct areas of the LN. One fraction of pDCs accumulated at the SCS close to the virus-infected macrophages (Figures 2F and 2G). Here they arrested and interacted with infected SCS macrophages (Figures 2H–2K and Movie S2). The other fraction of pDCs was located around OT-I T cell clusters that formed around a virus-infected cDC in the interfollicular area (Figure 2L). These pDCs directly interacted with the antigen-engaged OT-I T cells and possibly with the infected cDCs, forming a super-cluster of pDCs around the initial CD8+ T cell cluster (Figures 2L–2P and Movie S3). When imaging pDC behavior early after T cell activation, we found evidence of pDC migration toward CD8+ T cell clusters in line with CCR5-dependent chemotaxis and/or chemokinesis and subsequent pDC retention at T cell cluster sites (Figure 2Q). The direct association between pDCs and OT-I T cells might reflect an antigen-specific interaction. To examine this possibility, we transferred labeled peptide-pulsed (OVA257) splenic cDCs and OT-I T cells into pDC reporter mice and visualized their interactions. In this experimental set-up, antigen presentation is restricted to the transferred cDCs. Twelve hours after cDC transfer, we found tight interactions between peptide-pulsed cDCs and OT-I T cells. Importantly, under this condition pDCs still formed clusters around OT-I T cells, suggesting that pDCCD8+ T cell communication does not require antigen-specific interactions (Figure S2B). In summary, we conclude that pDCs show a rapid intranodal relocalization during viral infection. Some pDCs migrate to the SCS to interact with infected macrophages, while another group of pDCs accumulates in the vicinity of CD8+ T cells that are activated by virus-infected, antigen-presenting cDCs. pDCs Migrate to Infected Sites via CXCR3 Chemokine Receptor Given the relocation of pDCs to two distinct sites within infected LN and some evidence of directional migration, we hypothesized that chemokines were playing important roles in promoting these positional changes. Several groups have shown that CXCR3 plays a non-redundant role in controlling lymphocyte migration to the SCS for interactions with infected macrophages (Groom €ller et al., 2013; Sung et al., 2012). Thereet al., 2012; Kastenmu fore, we speculated that CXCR3 might also be critical for pDCs to migrate to this region of the LN. To examine this possibility, we crossed pDC reporter mice to CXCR3-deficient animals and infected these mice with MVA OVA tdTomato. In contrast to WT animals, CXCR3-deficient mice did not show efficient translocation of pDCs to the SCS (Figures 3A and 3B and Movie

S4). Semi-automated quantification supported these visual interpretations (Figure 3C). The few pDCs that were present at the SCS in Cxcr3 / animals interacted with the infected macrophages leading to migrational arrest similar to their WT counterparts (Figures 3D and 3E). In line with a critical role for CXCR3 for pDC migration to the SCS, we found an upregulation of both CXCL9 and CXCL10 after infection with MVA (Figures 3F and 3G). In the context of published work showing that SCS macrophages are primary targets of viral infection and promote viral replication in the LN, these data suggest that CXCR3-mediated pDC recruitment to infected SCS macrophages could contribute to control local viral replication and gene expression (Iannacone et al., 2010; Swiecki et al., 2010). To directly test this notion, we infected WT, Cle4c+/DTR, and CXCR3-deficient animals with replication-competent VSV in the footpad and analyzed the viral titers 24 hr later. If pDCs were depleted, we did not detect differences in the viral titers in the popliteal LN but there was a significant dissemination of virus to the upstream inguinal LN (Figure S3A). In contrast, in CXCR3-deficient animals this enhanced lymphatic viral dissemination was not observed. However, CXCR3-deficient animals had significantly increased pDC numbers in the LNs before infection as compared to control animals, while the migrational speed was unaltered (Figures S3B and S3C). The increased density of pDCs in the CXCR3-deficient mice may place enough of these cells near to the SCS macrophages in the steady state that chemokine-mediated guidance is no longer a limiting feature of the response upon infection. These findings concerning the innate antiviral properties of pDCs left open the question of how pDCs are differentially recruited to CD8+ T cell priming sites and what possible functions they exert in support of T cell activation and adaptive immune responses. CCR5 Guides pDCs to CD8+ T Cell Priming Sites On the assumption that chemokines also direct pDCs toward sites of CD8+ T cell priming by antigen-bearing cDCs, we first sought to determine whether virus-induced inflammatory signals played a role in this colocalization phenomenon or whether the relevant signals arose from the interacting T cells and cDCs. For this purpose, we designed an in vivo system that activates CD8+ T cells yet lacks virus-associated inflammatory cues. Labeled OT-I T cells were transferred into pDC reporter mice that were then immunized with LPS-free OVA. T cell and pDC behavior was then visualized in these animals. Similar to what we observed during viral infection, pDCs colocalized with the arrested OT-I T cells, forming large cellular aggregates (Figures S4A and S4B). This indicated that T cell- or cDC-derived rather than virus-promoted signals might lead to pDC recruitment to T cell priming sites. Besides CXCR3, CCR5 is highly and homogenously expressed on pDCs and plays an important role for pDC transmigration through inflamed vessels (Diacovo et al., 2005; Miller et al., 2012). CCR5 interacts with CCL3, CCL4, and CCL5, chemokines that are produced by activated T cells

(M) Image from IVM showing pDCs co-clustering with OT-I T cells. (N–P) Migration analysis of pDC and OT-I co-cluster 9 hr after infection showing the speed (N), translated tracks (O), and track displacement (P). (Q) Images from IVM showing pDC migration toward an OT-I cluster. Mice were infected with MVA OVA or MVA OVA tdTomato. Data are representative of more than four independent experiments analyzing at least three mice per group. Red bars indicate mean values. Error bars indicate the standard deviation. Scale bars represent 200 mm (A), 50 mm (B, H), 30 mm (F, M), 20 mm (L), 15 mm (Q). ***p % 0.001. See also Figure S2 and Movies S1, S2, and S3.

Immunity 46, 1–15, February 21, 2017 5

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

Figure 3. pDCs Accumulate at the SCS in an CXCR3-Dependent Manner (A and B) IF images showing pDC localization 7–10 hr p.i. in (A) WT and (B) Cxcr3 / mice; dashed line indicates SCS area. (C) Spatial frequency distribution of pDCs comparing WT (SiglechGFP/GFP) and Cxcr3 / (SiglechGFP/GFP) mice 10 hr p.i. (D) Image from IVM showing Cxcr3 / (SiglechGFP/GFP) pDCs 7 hr p.i. (E) Analysis of the mean velocity of pDCs in WT (SiglechGFP/GFP) and Cxcr3 / (SiglechGFP/GFP) mice. (F and G) Quantification of CXCL9 and CXCL10 from WT and Clec4c+/DTR animals treated with DTX using pLN homogenates from mock treated or infected conditions showing dot plots (F) and ELISA (G). Ratios indicate relative intensities adjusted to controls (HSP60). Mice were infected with MVA OVA tdTomato. Data are representative of (A–E) five independent experiments (n R 3), (F) one experiment from pooled samples (n = 9), or (G) two independent experiments (n = 3). Red bars indicate mean values. Error bars indicate the standard deviation. Scale bars represent 150 mm (A, B) or 30 mm (C). ***p % 0.001. Please see also Figure S3 and Movie S4.

(Dorner et al., 2002) and as shown by prior studies, played a key role in cell-cell interactions involved in the generation of cell-mediated immune responses (Castellino et al., 2006; Hugues et al., 2007). Indeed, we found a strong increase in CCL3, CCL4, and CCL5 in LN homogenates 8 hr after infection with MVA OVA (Figures 4A and 4B). To assess the role of CCR5 in intranodal migration of pDCs, we crossed CCR5-defi6 Immunity 46, 1–15, February 21, 2017

cient animals to pDC reporter mice and visualized pDC behavior in vivo upon viral infection using IVM. In CCR5-deficient mice, pDCs were not found at OT-I T cell clusters (Figures 4C and 4D and Movie S5). Similarly to MVA infection, we detected a CCR5-dependent accumulation of pDCs around OT-I T cell clusters in the context of a low-dose VSV OVA infection (Figures S4C–S4E).

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

(legend on next page)

Immunity 46, 1–15, February 21, 2017 7

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

When we visualized pDCs after MVA infection in CXCR3-deficient mice, focusing on the CD8+ T cell priming sites, we found a significantly enhanced recruitment of pDCs that formed groups of up to about 60 cells interacting with the OT-I T cell cluster (Figures 4C, 4E, and 4G and Movie S6). However, this enhanced recruitment may be simply based on the increased numbers of pDCs present in CXCR3-deficient mice (Figure S3B). The steady-state distribution and migrational speed of pDCs in skin-draining LNs was unaltered in the absence or presence of CCR5 or CXCR3 (Figures S3C and S4F). We have previously shown that CD4+ and CD8+ T cells are initially activated separately on distinct cDCs (Eickhoff et al., 2015). Therefore, we wanted to address whether pDCs are relocalized to and interacted with activated CD4+ T cells and if so whether this process was also CCR5 dependent. To this end we transferred TCR transgenic CD4+ T cells (SMARTA) into pDC reporter mice and infected them with MVA GP (glycoprotein from LCMV). Eight hours after infection, we observed CD4+ T cell clusters in the paracortex, in deeper areas of the LN than observed for their CD8+ T cell counterparts. Importantly, we did not detect large pDC superclusters around activated CD4+ T cells, arguing that pDCs are more effectively recruited toward activated CD8+ T cells (Figure S4G). However, CD4+ T cell clusters consisted of fewer T cells than CD8+ T cell clusters and some pDCs did accumulate adjacent to CD4+ T cell clusters (Figures S4H and S4I). Therefore, quantitative (number of cells producing a chemokine) rather than qualitative (selective chemokine production by CD8+ versus CD4+ T cells) differences may be the basis for the observed differences. To gain further insight into this issue, we stimulated purified naive CD4+ and CD8+ T cells in the presence of plate-bound anti-CD3 and anti-CD28 antibodies and probed the supernatant for CCL3 using ELISA. Eight hours after activation, we detected a robust amount of CCL3 in the supernatant of CD8+ T cell cultures (Figure 4H). In contrast, we failed to detect CCL3 in CD4+ T cell cultures. Importantly, the extent of activation as measured by CD69 upregulation was comparable in both cultures (Figure 4H). These results are in line with a predominant recruitment of pDCs to CD8+ rather than to CD4+ T cell priming sites based on selective production of CCL3 by CD8+ T cells. Prior studies showed that the interaction of antigen-specific CD4+ T cells with antigen-bearing cDCs led to local generation of a combination of CCL3 and CCL4, suggesting that T cell induction of chemokine production by cDCs could account for the less robust but measurable in vivo recruit-

ment of pDCs to antigen-driven CD4+ T cell clusters (Castellino et al., 2006). To further examine whether CCL3 and CCL4 were the critical chemokines leading to the recruitment of pDCs to CD8+ T cell activation sites, we blocked CCL3 and CCL4 signals in vivo using neutralizing antibodies (Castellino et al., 2006) and analyzed the migratory behavior of pDCs upon viral infection in the presence of OT-I T cells. Such treatment significantly reduced the numbers of pDCs at CD8+ T cell priming sites as compared to what was observed in animals treated with isotype control antibodies (Figures 4I–4K). Overall, our results are in line with the interpretation that CCR5 instructs specific recruitment of pDCs toward CD8+ T cell priming sites, where CCL3 and CCL4 are secreted by activated CD8+ T cells and likely their interacting cDCs. IFN-I Signaling Involving cDCs Is Critical for Optimization of the Antiviral CD8+ T Cell Response Having established how pDCs are recruited to CD8+ T cell priming sites, we next examined how the presence of pDCs contributed to enhancing the CD8+ T cell response. The central function of pDCs appears to be their capacity to produce high amounts of IFN-I upon viral infection (Haeryfar, 2005). However, after depletion of pDCs, we did not observe a reduction in the total IFN-a response in the draining LN 8 hr after infection (Figure 5A). To assess whether IFN-I was critical for pDC function in our experimental system, we crossed Clec4c+/DTR mice with Ifnar1 / animals. When analyzing the immune response on d8 after infection with MVA OVA, we found no significant differences regarding the antigen-specific CD8+ T cell response when comparing Ifnar1 / animals that did or did not have a pDC population (Figures 5B–5E and S5A–S5D). The loss of the augmenting effect of pDCs in Ifnar1 / mice indicated that IFN-I is critical for pDC function in our experimental model. Together with our results that pDCs do not substantially alter the total amount of IFN-a in the LN (Figure 5A), we conclude that pDC-derived IFN-I acts locally on the interacting cellular partners. To identify the cell population(s) on which IFN-I is acting to optimize CD8+ T cell priming, we first analyzed mice that specifically lack IFNAR1 on T cells (Cd4-cre 3 Ifnar1flox/flox) and compared them to their WT littermates. In four independent experiments, we consistently observed a slightly reduced antigen-specific CD8+ T cell response among the IFNAR1-negative T cells (Figures 5F, 5H, and S5E–S5H). However, the statistical analyses of the pooled datasets did not show significant differences (Figures 5G and

Figure 4. pDCs Are Recruited to CD8+ T Cell Priming Sites via CCR5 (A and B) Dot blot detecting CCL3, CCL4, and CCL5 (A) and ELISA detecting CCL3 (B) using LN homogenates from mock treated and infected (8 hr) animals. Ratio indicates relative intensities normalized to controls (HSP60). (C) Quantification of pDCs in proximity to OT-I T cell clusters in WT, Ccr5 / , and Cxcr3 / animals (all mice SiglechGFP/GFP) 10 hr p.i. (D) IF image from Ccr5 / (SiglechGFP/GFP) mice showing pDCs and OT-I T cells 10 hr p.i. (E) IF image from Cxcr3 / (SiglechGFP/GFP) mice showing pDCs and OT-I T cells 10 hr p.i. (F and G) Image from IVM of pLN showing OT-I T cell cluster and pDCs from Ccr5 / (SiglechGFP/GFP) (F) and Cxcr3 / (SiglechGFP/GFP) mice (G). (H) CCL3 ELISA from supernatants of activated (aCD3 and aCD28, 8 hr) or non-activated, purified naive (CD44lo) CD4+ and CD8+ T cells. CD69 indicates extent of T cell activation within cultures. (I–K) IF images and quantification of OT-I T cells and pDC localization 8 hr p.i., comparing isotype- with aCCL3- and aCCL4-treated animals. Mice were infected with MVA OVA. Data are representative of (A) one experiment from pooled samples (n = 9), of (C–G and I–K) at least three independent experiments (n = 3), or of (B, H) two independent experiments (n = 3). Scale bars represent 150 mm and 15 mm (D), 150 mm and 20 mm (E), 20 mm (F and G), 150 mm and 10 mm (J), or 200 mm and 10 mm (K). Red bars indicate mean values. Error bars indicate the standard deviation. ***p % 0.001, *p % 0.05, ns indicates non-significant. Please see also Figure S4 and Movies S5 and S6.

8 Immunity 46, 1–15, February 21, 2017

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

A

B

C

D

E

F

G

H

I

J

K

L

M

Figure 5. pDCs Function and Target Cells of IFN-I (A) IFN-a ELISA from LN homogenates from DTX-treated WT and Clec4c+/DTR mice treated 8 hr p.i. (B–M) Analysis of immune response on d8 p.i., comparing Ifnar / and Ifnar / 3 Clec4c+/DTR animals (B–E), CD4 Cre 3 Ifnarflox/flox with littermates (F–I), and Itgax-Cre 3 Ifnarflox/flox with littermates (J–M). (B, F, J) Representative FACS plots of IFN-g-producing CD8+ T cells. (C, G, K) Absolute numbers of IFN-g-producing CD8+ T cells. (D, H, L) Representative FACS plots showing TNF-a and IL-2 production (gated on CD8+ and IFN-g+). (E, I, M) Quantitative analysis of TNF-a- and IL-2-producing CD8+ T cells (gated on CD8+, IFN-g+). Mice were infected with MVA OVA. Data are representative of (A) three (n = 3) independent and (B–M) four (n = 4) independent experiments analyzing at least four animals per group. Pooled data from four independent experiments (n = 16) shown in (G) and (I). Error bars indicate the standard deviation. **p % 0.01, *p % 0.05. Please see also Figure S5.

5I). Although these results indicate some contribution of direct IFN-I signaling in CD8+ T cells to the effects mediated by pDCs, this mechanism can only partially account for the differences we observed when removing the pDC compartment (Figure 1). Therefore, we speculated that pDC-derived IFN-I might additionally act on cDCs to optimize their function. To address

this hypothesis, we analyzed mice that specifically lack IFNAR1 on DCs (Itgax-cre 3 Ifnar1flox/flox). In such animals, there was a substantially reduced CD8+ T cell response on d8 after immunization in comparison to their WT littermates (Figures 5J–5M and S5I–S5L). This result pointed to a critical role of IFN-I in optimizing DC functionality. Immunity 46, 1–15, February 21, 2017 9

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

A

D

B

C

E

F

G

Figure 6. pDCs Optimize XCR1+ DC Maturation and Cross-presentation (A) Phenotypic analysis of GFP-positive splenic DCs and their maturation comparing Ifnar1flox/flox (WT) and Itagx-Cre 3 Ifnar1flox/flox mice 8 hr p.i. (MVA OVA GFP i.v.). (B and C) Representative histograms and quantitative analysis of splenic DC subsets comparing Ifnarflox/flox (WT) and Itgax-Cre 3 Ifnar1flox/flox mice 36 hr p.i. (MVA OVA i.v.). (D and E) Representative histograms and quantitative analysis of splenic DC subsets comparing DTX-treated WT and Clec4c+/DTR animals (36 hr MVA OVA i.v.). (F and G) Proliferation profile and quantitative analysis of OT-I T cells in pLN 3 days after transfer into infected (60 hr MVA OVA f.p.), DTX-treated WT, Clec4c+/DTR, and Ccr5 / animals (all groups SiglechGFP/GFP). Data are representative of at least three independent experiments (n R 3). Error bars indicate the standard deviation. ***p % 0.001, **p % 0.01, *p % 0.05, ns indicates non-significant. Please see also Figure S6.

pDCs Locally Cooperate with Cross-presenting XCR1+ DCs via IFN-I Given the physical proximity among pDCs, CD8+ T cells, and infected cDCs within multicellular clusters, we first analyzed the expression of costimulatory molecules on infected cDCs in the presence or absence of IFN-I signaling. We infected ItgaxCre 3 Ifnar1flox/flox and control littermates i.v. with MVA GFP and analyzed the expression of CD80, CD86, and CD40 on infected GFP+ cDCs 8 hr later. These analyses did not reveal any 10 Immunity 46, 1–15, February 21, 2017

significant differences in the expression of costimulatory molecules on infected DCs in the presence or absence of IFN-I signaling (Figure 6A). To test whether the IFN-I might act on non-infected cDCs, we isolated splenic cDCs 36 hr after infection, identified them by surface marker expression (XCR1 versus CD11b), and analyzed the expression pattern of costimulatory molecules. This revealed a striking reduction in CD40, CD80, and in particular CD86 expression on XCR1+ DCs but not CD11b DCs, if IFNAR1 was absent on cDCs (Figures 6B and

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

Figure 7. Activated CD8+ T Cells Recruit XCR1+ DCs via XCL1 (A–E) IF images showing OT-I T cell and XCR1+ DC localization in (A) the steady state or (B–E) 10 hr p.i. (MVA OVA) in (B, C) WT (XCR1+/venus) and (D) XCR1deficient animals (XCR1venus/venus) or (E) using Xcl1 / OT-I T cells. (F) IF images showing OT-I T cell cluster and XCL1 staining 10 hr p.i. (legend continued on next page)

Immunity 46, 1–15, February 21, 2017 11

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

6C). To determine whether this finding was related to pDC function, we depleted pDCs, infected the animals, and analyzed the expression of costimulatory markers on splenic DCs 36 hr later. Again, we detected a significant reduction of CD80 and CD86 expression on XCR1+ DCs in pDC-depleted as compared to non-depleted control animals (Figures 6D and 6E). These data indicated that pDC-derived IFN-I acted selectively on XCR1+ DCs rather than CD11b+ DCs. It is well established that XCR1+ DCs are of particular importance for cross-presentation of viral antigens and also serve as a platform that allows interaction with both antigen-specific CD4+ and CD8+ T cells for optimal T cell programming and memory formation in the context of viral infections (Bachem et al., 2012; Eickhoff et al., 2015; Hor et al., 2015). Since MVA does not replicate and therefore heavily depends on cross-presentation for CD8+ T cell priming (Gasteiger et al., 2007), we investigated the role of XCR1+ DCs in MVA-induced CD8+ T cell responses. Indeed, the CD8+ T cell response directed against the immunodominant viral epitope (B8R20) was strongly impaired if XCR1+ DCs were depleted, confirming the critical role of this DC subset for generating functional CD8+ T cell responses (Figures S6A–S6H). To further analyze whether XCR1+ DC functionality was enhanced by pDCs, we probed the capacity of XCR1+ DCs to drive OT-I T cell proliferation in the presence or absence of pDCs. As we previously demonstrated, the initial activation of CD8+ T cells after vaccinia virus infection is independent of cross-presenting XCR1+ DCs but is driven by infected DCs, predominantly the CD11b DC subset (Eickhoff et al., 2015). To discriminate between these two populations (infected DCs versus cross-presenting XCR1+ DCs), we transferred labeled OT-I T cells 60 hr after infection and analyzed the proliferation profile 3 days later. At this time point after infection, infected cDCs had largely died and been cleared, so antigen presentation is dominated by cross-presenting XCR1+ DCs (Eickhoff et al., 2015; Gasteiger et al., 2007). As expected, depletion of XCR1+ DCs largely abrogated OT-I proliferation in vivo in this time frame as compared to control animals (Figure S6I and S6J). In a similar experimental set-up using dyelabeled OT-I GFP T cells, pDC depletion also led to a significantly reduced OT-I proliferation in vivo as compared to non-depleted animals (Figures 6F and 6G). To look for a link between these data and the requirement for CCR5 in directing pDCs to CD8 priming sites, we analyzed the proliferation of WT OT-I T cells in CCR5-deficient hosts. Similar to pDC-depleted animals, OT-I T cells proliferated significantly less in CCR5-deficient hosts (Figures 6F and 6G). Taken together, these results indicate that pDC-derived IFN-I directly or indirectly activates XCR1+ DCs to optimize the expression of costimulatory molecules and their ability to activate CD8+ T cells via cross-presentation. Activated CD8+ T Cells Recruit XCR1+ DCs via XCL1 To better understand where and when pDCs provided XCR1+ DCs with IFN-I-dependent stimulation, we analyzed the migration and localization of XCR1+ DCs after infection with MVA. In the steady state, XCR1+ DCs were dispersed throughout the par-

acortex and the interfollicular area of the LN, with some XCR1+ DCs located within the subcapsular sinus (Figure 7A; Kitano et al., 2016). Importantly, after infection we found that XCR1+ DCs accumulated around OT-I T cell clusters in a pattern similar to that of recruited pDCs, creating a new microenvironment around the initial site of CD8+ T cell engagement with antigenpresenting cDCs (Figures 7B, 7C, 7F, and 7G and Movie S7). Upon activation XCL1 is rapidly expressed by CD8+ T cells (Dorner et al., 2002), so we speculated that this chemokine could be critically involved in the reorganization of the XCR1+ DC network observed after infection. To address this, we transferred OT-I cells into XCR1venus/venus mice that lacked functional XCR1 expression, infected them with MVA OVA, and analyzed LN sections 10 hr later. As anticipated, we did not observe a substantial cluster formation of XCR1+ DCs in XCR1-deficient animals (Figures 7D, 7F, and 7G). To test whether CD8+ T cell-derived XCL1 is critical for XCR1+ DC recruitment, we transferred Xcl1 / OT-I T cells into XCR1+/venus mice and analyzed the XCR1+ DC localization 10 hr after infection. We observed a significant reduction in the spatial correlation between XCR1+ DCs and clustered Xcl1 / OT-I T cells as compared to WT OT-I T cells (Figures 7E–7G). To further investigate the migrational behavior of XCR1+ DCs, we visualized this cell population in vivo shortly after T cell cluster formation. Using this approach, we found in vivo evidence for direct migration of XCR1+ DCs toward clustered OT-I T cells, supporting the notion of chemokine-based attraction of XCR1+ DCs (Figure 7H and Movie S8). In summary, we conclude that XCR1+ DCs are recruited to CD8+ T cell priming sites allowing them to interact with pDCs but also with infected cDCs that are localized in the middle of the cluster of T cells. This positioning and exposure to IFN-I provides XCR1+ DCs with a microenvironment that provides local access to cell-associated viral proteins once the initially infected DC undergoes cell death, thereby optimizing cross-presentation, maturation, and productive second phase interactions with already activated CD8+ T cells. DISCUSSION In this study, we analyzed the spatiotemporal dynamics of pDCs to address when and where pDC-derived IFN-I supports innate and adaptive immunity during viral infection. We found that pDCs are directed toward two different sites in the LN as a consequence of selective and possibly competing chemokine receptor engagement. In particular, CXCL9 and CXCL10 interaction with CXCR3 directs pDCs to subcapsular and medullary macrophages. It seems intuitive that this translocation to sites of viral infection may be essential to suppress viral replication, although we did not find direct evidence to support this notion. Whether CXCR3-mediated pDC migration serves another function or indeed supports the antiviral function of pDCs but was masked in our experimental setting due to an increase in pDC density within the LN needs to be further tested.

(G) Analysis of 3D spatial correlation between OT-I T cells and XCR1+ DCs. (H) IVM images of XCR1+/venus mice showing directed XCR1+ DC migration toward an OT-I T cell cluster. Scale bars represent 250 mm (B), 150 mm (D, E), 100 mm (A), 20 mm (C, H), or 10 mm (F). Data are representative of at least three independent experiments (n = 3). Error bars indicate the standard deviation. ***p % 0.001. Please see also Movies S7 and S8.

12 Immunity 46, 1–15, February 21, 2017

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

The second area of pDC accumulation involves the priming sites of CD8+ T cells in the LN, which is the focus of this study. The critical chemokines are CCL3 and CCL4 that are produced in the context of antigen-dependent CD8+ T cell activation and engage CCR5 to direct pDCs toward such initial sites of CD8+ T cell priming. The resulting accumulation of pDCs around activated CD8+ T cells is likely a combination of chemotactic and adhesive cues. In this study, we have provided evidence for direct interaction between pDCs and activated CD8+ T cells, which does not require cognate antigen presentation by pDCs. Similarly, pDCs may interact with the infected cDCs that activated naive CD8+ T cells initially. This may be critical for pDCs to sense viral infections and lead to release of IFN-I, given that we did not find any evidence for direct infection of pDCs. Besides pDCs, we also detected a rapid recruitment of XCR1+ DCs toward the activated XCL1-producing CD8+ T cells, creating a new microenvironment around the initial site of CD8+ T cell activation. Within this microenvironment, pDC-derived IFN-I can act on CD8+ T cells as a signal 3 cytokine and perhaps more importantly on XCR1+ DCs to optimize their maturation, costimulatory capacity, and ability to cross-present viral antigens. This in turn contributes to maximizing the fraction of the antigen-specific CD8+ T cell repertoire recruited into the immune response and/or the extent of proliferation of that cohort of CD8+ T cells encountering TCR ligand. Taking these findings together, we propose a concept in which T cells that could locate rare antigen-bearing DCs early during an infection become the nucleus of a new microenvironment that is actively created by signals emitted in response to the initial CD8+ T cell-antigen-presenting cell (APC) interaction. Thus, rather than having to continue to search for the cells necessary to optimize their response, this mechanism permits each CD8+ T cell that locates a useful APC to accumulate needed accessory cells at this existing site of activation, supporting a robust response without the risk of missing the right partner cells during further T cell migration within the lymphoid tissue. XCR1+ DCs, beyond their critical function to cross-present viral antigens, also serve as an essential platform to transmit CD4+ T cell-dependent signals to CD8+ T cells (Eickhoff et al., 2015; Hor et al., 2015). Chemokines, particularly CCL3 and CCL4, optimize cellular encounters between cDCs, CD4+, and CD8+ T cells (Castellino et al., 2006; Hugues et al., 2007). In our previous work we have shown that CD4+ T cell-cDC interactions lead to CCL3 and CCL4 production, which is critical to recruitment of activated CD8+ T cells for the delivery of helper signals (Castellino et al., 2006). Given our data, the likely cellular source of CCL3 in those prior experiments was cDCs rather than CD4+ T cells. Several lines of evidence argue that activated rather than naive CD4+ and CD8+ T cells are exchanging helper signals (Eickhoff et al., 2015; Jusforgues-Saklani et al., 2008). Activated T cells typically have short-term rather than long-term interactions with cDCs (Mempel et al., 2004; Miller et al., 2004; Stoll et al., 2002). Therefore, it makes sense to induce chemokine production by licensed DCs to optimize CD8+ T cell recruitment for the delivery of help rather than having the source of these chemokines be CD4+ T cells that may have disengaged the licensed DCs before CD8+ T cell recruitment. This is in line with a consecutive rather than simultaneous interaction model for the delivery of helper signals for CD8+ T cells (Ridge et al., 1998).

An important aspect of our study is the unraveling of how and where pDCs cooperate with cDCs to support adaptive immunity. Our data indicate pDC-XCR1+ DCs cooperativity early after viral infection within T cell priming hubs. This is in line with previous studies demonstrating that pDCs provide IFN-I only early after viral infection (Swiecki et al., 2010). In our study, pDCs do not seem to present antigen, but do promote cross-presentation by colocalized XCR1+ DCs. Additionally, they optimize the maturation of these DC subsets as shown by an upregulation of CD86, which is the critical costimulatory molecule for priming CD8+ T cells in the context of vaccinia virus infections (Salek-Ardakani et al., 2009). It is important to point out that while in our model IFN-I is critical, pDCs might also support adaptive immune responses through the production of other mediators such as IL-12 (Asselin-Paturel et al., 2001). The importance of pDC and XCR1+ DC relocalization in carrying out their specific functions as shown here provides strong evidence that the capacity of pDCs and XCR1+ DCs to rapidly migrate within lymphoid tissues is a key aspect of their physiological function. Our data further support a concept in which the initial CD8+ T cell-APC interaction is critical for the reorganization of the DC network, allowing for optimal cellular interactions and cooperativity to maximize CD8+ T cell priming. With our information about the spatiotemporal behavior of pDCs and XCR1+ DCs in LN during viral infections and the dissection of the molecular signals guiding such migration, we provide a framework for considering how to harness these cell populations to maximize cellular immunity in the context of vaccination. EXPERIMENTAL PROCEDURES Animals Mice were purchased from Jackson or Janvier Labs or maintained at in-house facilities. All mice were maintained in specific-pathogen-free conditions at an Association for Assessment and Accreditation of Laboratory Animal Care-accredited animal facility. All procedures were approved by the NIAID Animal Care and Use Committee (NIH) and the North Rhine-Westphalia State Environment Agency (LUA NRW), respectively. For details on mouse strains, see Supplemental Information. Treatment of Mice For depletion of pDCs, transgenic mice and control littermates were treated with 0.1 mg DTX i.p. (Merck Millipore) on d 2, d 1, d0, d1, d2, and d3 (for analysis at d8), on d 2, d–1, and d0 (for analysis 8 hr p.i.), and on d 2, d 1, d0, and d1 i.p. (for analysis 36 hr p.i.). For depletion of XCR1+ DCs, transgenic mice and control littermates were treated with 0.5 mg DTX i.p. (Merck Millipore) on d 1, d0, and d1. For chemokine neutralization, chemokine blocking antibodies or isotype matched controls were injected i.v. at the time of infection (R&D; CCL3: AF450-NA, MAB450-100; CCL4: AF-451-NA, MAB451-100; Isotype matched controls: AB-108-C, MAB006) (Castellino et al., 2006). For immunization with LPS-free OVA (Hyglos), 50 mg were injected into the footpad, and pLNs were harvested 8 hr after injection. Viral Infections 105 PFU VSV OVA or 106–108 IU MVA OVA GFP, MVA OVA tdTomato, MVA OVA, or MVA WT were diluted in PBS and injected in the footpad (foothock [Kamala, 2007]) or intravenously. Adoptive T Cell Transfer OT-I, SMARTA, or polyclonal CD8+ T cells were sorted using a MACS CD4+ or CD8+ T cell negative selection kit (Miltenyi) combined with biotinylated antiCD44 (IM7, BD Biosciences). 2–4 3 106 cells were transferred i.v.

Immunity 46, 1–15, February 21, 2017 13

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

In Vivo Proliferation Assay OT-I GFP or OT-I tdTomato T cells were MACS sorted (CD8 negative selection kit, Miltenyi), labeled with Cell Proliferation Dye eFluor 670 (eBioscience), and transferred into recipient mice 60 hr after MVA OVA infection. 72 hr after transfer, pLNs were harvested and single cell suspensions were analyzed by FACS. ELISA and Chemokine Array ELISAs and chemokine array of whole LN homogenates were performed according to manufacturer’s instructions (R&D Systems and pbl Assay Science). Preparation of Peptide-Pulsed DCs For transfer of peptide-pulsed DCs, splenic DCs from CD11c-YFP mice were harvested, digested with collagenase and DNase for 30 min, and enriched using a CD11c positive seletion kit (Miltenyi). DCs were incubated in the presence of SIINFEKL peptide (1 mg/mL) and 5 pg/mL LPS at 37 C for 1 hr. After incubation, 6.25 3 105 DCs were injected into footpad. OT-I T cells were transferred i.v. 18 hr after injection of peptide-pulsed DCs. 10 hr after T cell injection, pLNs were harvested and prepared for immunohistological analyses. Flow Cytometry For analysis, LNs and spleens were harvested and single-cell suspensions were generated. For details on antibodies see Supplemental Information. Immunofluorescence Staining PLP-fixed, frozen tissues were were cut, stained, mounted, and acquired on a 710 confocal microscope (Carl Zeiss Microimaging). For details on antibodies see Supplemental Information. Intravital Two-Photon Imaging Mice were anesthetized, popliteal LNs were exposed, and intravital microscopy was performed using a protocol modified from a previous report €ller et al., 2013). Raw imaging data were processed using a (Kastenmu semi-automated approach (Imaris/Bitplane). For details see Supplemental Information.

Received: July 3, 2016 Revised: November 30, 2016 Accepted: December 22, 2016 Published: February 9, 2017 REFERENCES Ang, D.K., Oates, C.V., Schuelein, R., Kelly, M., Sansom, F.M., Bourges, D., Boon, L., Hertzog, P.J., Hartland, E.L., and van Driel, I.R. (2010). Cutting edge: pulmonary Legionella pneumophila is controlled by plasmacytoid dendritic cells but not type I IFN. J. Immunol. 184, 5429–5433. Asselin-Paturel, C., Boonstra, A., Dalod, M., Durand, I., Yessaad, N., DezutterDambuyant, C., Vicari, A., O’Garra, A., Biron, C., Brie`re, F., and Trinchieri, G. (2001). Mouse type I IFN-producing cells are immature APCs with plasmacytoid morphology. Nat. Immunol. 2, 1144–1150. Asselin-Paturel, C., Brizard, G., Chemin, K., Boonstra, A., O’Garra, A., Vicari, A., and Trinchieri, G. (2005). Type I interferon dependence of plasmacytoid dendritic cell activation and migration. J. Exp. Med. 201, 1157–1167. €ttler, S., Mora, A., Zhou, X., Hegemann, A., Bachem, A., Hartung, E., Gu Plantinga, M., Mazzini, E., Stoitzner, P., Gurka, S., et al. (2012). Expression of XCR1 characterizes the Batf3-dependent lineage of dendritic cells capable of antigen cross-presentation. Front. Immunol. 3, 214. Barchet, W., Krug, A., Cella, M., Newby, C., Fischer, J.A., Dzionek, A., Pekosz, A., and Colonna, M. (2005). Dendritic cells respond to influenza virus through TLR7- and PKR-independent pathways. Eur. J. Immunol. 35, 236–242. Bendelac, A., Savage, P.B., and Teyton, L. (2007). The biology of NKT cells. Annu. Rev. Immunol. 25, 297–336. Castellino, F., Huang, A.Y., Altan-Bonnet, G., Stoll, S., Scheinecker, C., and Germain, R.N. (2006). Chemokines enhance immunity by guiding naive CD8+ T cells to sites of CD4+ T cell-dendritic cell interaction. Nature 440, 890–895. Cervantes-Barragan, L., Lewis, K.L., Firner, S., Thiel, V., Hugues, S., Reith, W., Ludewig, B., and Reizis, B. (2012). Plasmacytoid dendritic cells control T-cell response to chronic viral infection. Proc. Natl. Acad. Sci. USA 109, 3012–3017.

Statistical Analysis Student’s t test (two-tailed) was used for the statistical analysis of differences between two groups with normal distribution.

Chen, L., and Flies, D.B. (2013). Molecular mechanisms of T cell co-stimulation and co-inhibition. Nat. Rev. Immunol. 13, 227–242.

SUPPLEMENTAL INFORMATION

Colonna, M., Trinchieri, G., and Liu, Y.J. (2004). Plasmacytoid dendritic cells in immunity. Nat. Immunol. 5, 1219–1226.

Supplemental Information includes six figures and eight movies and can be found with this article online at http://dx.doi.org/10.1016/j.immuni.2017.01.003. AUTHOR CONTRIBUTIONS A.B., S.E., and S.D. planned and performed experiments and analyzed and interpreted the data. N.G. and W. Kolanus analyzed data and designed experiments. T.K., C.K., R.A.K., T.Q., M.I., S.B., W.B., and M.C. were involved in study design and provided critical reagents. F.K. analyzed imaging data. €ller conceptualized the study and analyzed the R.N.G. and W. Kastenmu €ller, and A.B. wrote the manuscript. data. R.N.G., W. Kastenmu ACKNOWLEDGMENTS We would like to thank S. Ebbinghaus and S. Rathmann for technical assistance and D.H. Busch for kindly providing MHCI multimers. This research was supported by the Intramural Research Program, NIAID, NIH. W. Kas€ller, N.G., C.K., and W. Kolanus are members of the DFG Excellence tenmu Cluster ImmunoSensation in Bonn, Germany, and are supported by grant €ller are supported by the DFG Graduate SFB704. S.D., S.B., and W. Kastenmu €ller is supported by NRWprogram 2168/1 (Bo&MeRang). W. Kastenmu €ckkehrerprogramm of the German state of Northrhine-Westfalia. T.K. was Ru supported by the Kishimoto Foundation and a Grant-in-Aid for Scientific Research (B) from Japan Society for the Promotion of Science (JSPS, 26293106). F.K. is supported by the Einstein Foundation Berlin (JF-Klauschen-2014).

14 Immunity 46, 1–15, February 21, 2017

Curtsinger, J.M., and Mescher, M.F. (2010). Inflammatory cytokines as a third signal for T cell activation. Curr. Opin. Immunol. 22, 333–340. Di Pucchio, T., Chatterjee, B., Smed-So¨rensen, A., Clayton, S., Palazzo, A., Montes, M., Xue, Y., Mellman, I., Banchereau, J., and Connolly, J.E. (2008). Direct proteasome-independent cross-presentation of viral antigen by plasmacytoid dendritic cells on major histocompatibility complex class I. Nat. Immunol. 9, 551–557. Diacovo, T.G., Blasius, A.L., Mak, T.W., Cella, M., and Colonna, M. (2005). Adhesive mechanisms governing interferon-producing cell recruitment into lymph nodes. J. Exp. Med. 202, 687–696. Dorner, B.G., Scheffold, A., Rolph, M.S., Huser, M.B., Kaufmann, S.H., Radbruch, A., Flesch, I.E., and Kroczek, R.A. (2002). MIP-1alpha, MIP1beta, RANTES, and ATAC/lymphotactin function together with IFN-gamma as type 1 cytokines. Proc. Natl. Acad. Sci. USA 99, 6181–6186. Eickhoff, S., Brewitz, A., Gerner, M.Y., Klauschen, F., Komander, K., Hemmi, €ller, W. (2015). H., Garbi, N., Kaisho, T., Germain, R.N., and Kastenmu Robust anti-viral immunity requires multiple distinct T cell-dendritic cell interactions. Cell 162, 1322–1337. Fooksman, D.R., Vardhana, S., Vasiliver-Shamis, G., Liese, J., Blair, D.A., Waite, J., Sacrista´n, C., Victora, G.D., Zanin-Zhorov, A., and Dustin, M.L. (2010). Functional anatomy of T cell activation and synapse formation. Annu. Rev. Immunol. 28, 79–105. Gasteiger, G., Kastenmuller, W., Ljapoci, R., Sutter, G., and Drexler, I. (2007). Cross-priming of cytotoxic T cells dictates antigen requisites for modified vaccinia virus Ankara vector vaccines. J. Virol. 81, 11925–11936.

Please cite this article in press as: Brewitz et al., CD8+ T Cells Orchestrate pDC-XCR1+ Dendritic Cell Spatial and Functional Cooperativity to Optimize Priming, Immunity (2017), http://dx.doi.org/10.1016/j.immuni.2017.01.003

Groom, J.R., Richmond, J., Murooka, T.T., Sorensen, E.W., Sung, J.H., Bankert, K., von Andrian, U.H., Moon, J.J., Mempel, T.R., and Luster, A.D. (2012). CXCR3 chemokine receptor-ligand interactions in the lymph node optimize CD4+ T helper 1 cell differentiation. Immunity 37, 1091–1103. Haeryfar, S.M. (2005). The importance of being a pDC in antiviral immunity: the IFN mission versus Ag presentation? Trends Immunol. 26, 311–317. Hor, J.L., Whitney, P.G., Zaid, A., Brooks, A.G., Heath, W.R., and Mueller, S.N. (2015). Spatiotemporally distinct interactions with dendritic cell subsets facilitates CD4+ and CD8+ T cell activation to localized viral infection. Immunity 43, 554–565. Hugues, S., Scholer, A., Boissonnas, A., Nussbaum, A., Combadie`re, C., Amigorena, S., and Fetler, L. (2007). Dynamic imaging of chemokine-dependent CD8+ T cell help for CD8+ T cell responses. Nat. Immunol. 8, 921–930. Iannacone, M., Moseman, E.A., Tonti, E., Bosurgi, L., Junt, T., Henrickson, S.E., Whelan, S.P., Guidotti, L.G., and von Andrian, U.H. (2010). Subcapsular sinus macrophages prevent CNS invasion on peripheral infection with a neurotropic virus. Nature 465, 1079–1083. Izaguirre, A., Barnes, B.J., Amrute, S., Yeow, W.S., Megjugorac, N., Dai, J., Feng, D., Chung, E., Pitha, P.M., and Fitzgerald-Bocarsly, P. (2003). Comparative analysis of IRF and IFN-alpha expression in human plasmacytoid and monocyte-derived dendritic cells. J. Leukoc. Biol. 74, 1125–1138.

Mourie`s, J., Moron, G., Schlecht, G., Escriou, N., Dadaglio, G., and Leclerc, C. (2008). Plasmacytoid dendritic cells efficiently cross-prime naive T cells in vivo after TLR activation. Blood 112, 3713–3722. €ller, W., and Germain, R.N. (2014). Spatiotemporal basis of Qi, H., Kastenmu innate and adaptive immunity in secondary lymphoid tissue. Annu. Rev. Cell Dev. Biol. 30, 141–167. Reizis, B., Bunin, A., Ghosh, H.S., Lewis, K.L., and Sisirak, V. (2011a). Plasmacytoid dendritic cells: recent progress and open questions. Annu. Rev. Immunol. 29, 163–183. Reizis, B., Colonna, M., Trinchieri, G., Barrat, F., and Gilliet, M. (2011b). Plasmacytoid dendritic cells: one-trick ponies or workhorses of the immune system? Nat. Rev. Immunol. 11, 558–565. Ridge, J.P., Di Rosa, F., and Matzinger, P. (1998). A conditioned dendritic cell can be a temporal bridge between a CD4+ T-helper and a T-killer cell. Nature 393, 474–478. Salek-Ardakani, S., Arens, R., Flynn, R., Sette, A., Schoenberger, S.P., and Croft, M. (2009). Preferential use of B7.2 and not B7.1 in priming of vaccinia virus-specific CD8 T cells. J. Immunol. 182, 2909–2918. Stoll, S., Delon, J., Brotz, T.M., and Germain, R.N. (2002). Dynamic imaging of T cell-dendritic cell interactions in lymph nodes. Science 296, 1873–1876.

Jusforgues-Saklani, H., Uhl, M., Blache`re, N., Lemaıˆtre, F., Lantz, O., Bousso, P., Braun, D., Moon, J.J., and Albert, M.L. (2008). Antigen persistence is required for dendritic cell licensing and CD8+ T cell cross-priming. J. Immunol. 181, 3067–3076.

Sung, J.H., Zhang, H., Moseman, E.A., Alvarez, D., Iannacone, M., Henrickson, S.E., de la Torre, J.C., Groom, J.R., Luster, A.D., and von Andrian, U.H. (2012). Chemokine guidance of central memory T cells is critical for antiviral recall responses in lymph nodes. Cell 150, 1249–1263.

Kamala, T. (2007). Hock immunization: a humane alternative to mouse footpad injections. J. Immunol. Methods 328, 204–214.

Swiecki, M., and Colonna, M. (2015). The multifaceted biology of plasmacytoid dendritic cells. Nat. Rev. Immunol. 15, 471–485.

€ller, W., Brandes, M., Wang, Z., Herz, J., Egen, J.G., and Germain, Kastenmu R.N. (2013). Peripheral prepositioning and local CXCL9 chemokine-mediated guidance orchestrate rapid memory CD8+ T cell responses in the lymph node. Immunity 38, 502–513.

Swiecki, M., Gilfillan, S., Vermi, W., Wang, Y., and Colonna, M. (2010). Plasmacytoid dendritic cell ablation impacts early interferon responses and antiviral NK and CD8(+) T cell accrual. Immunity 33, 955–966.

Kitano, M., Yamazaki, C., Takumi, A., Ikeno, T., Hemmi, H., Takahashi, N., Shimizu, K., Fraser, S.E., Hoshino, K., Kaisho, T., and Okada, T. (2016). Imaging of the cross-presenting dendritic cell subsets in the skin-draining lymph node. Proc. Natl. Acad. Sci. USA 113, 1044–1049. Martı´n-Fontecha, A., Thomsen, L.L., Brett, S., Gerard, C., Lipp, M., Lanzavecchia, A., and Sallusto, F. (2004). Induced recruitment of NK cells to lymph nodes provides IFN-gamma for T(H)1 priming. Nat. Immunol. 5, 1260–1265. Mempel, T.R., Henrickson, S.E., and Von Andrian, U.H. (2004). T-cell priming by dendritic cells in lymph nodes occurs in three distinct phases. Nature 427, 154–159.

Swiecki, M., Wang, Y., Vermi, W., Gilfillan, S., Schreiber, R.D., and Colonna, M. (2011). Type I interferon negatively controls plasmacytoid dendritic cell numbers in vivo. J. Exp. Med. 208, 2367–2374. Thelen, M., and Stein, J.V. (2008). How chemokines invite leukocytes to dance. Nat. Immunol. 9, 953–959. Veiga-Parga, T., Sehrawat, S., and Rouse, B.T. (2013). Role of regulatory T cells during virus infection. Immunol. Rev. 255, 182–196. Villadangos, J.A., and Young, L. (2008). Antigen-presentation properties of plasmacytoid dendritic cells. Immunity 29, 352–361.

Miller, M.J., Safrina, O., Parker, I., and Cahalan, M.D. (2004). Imaging the single cell dynamics of CD4+ T cell activation by dendritic cells in lymph nodes. J. Exp. Med. 200, 847–856.

Xu, R.H., Wong, E.B., Rubio, D., Roscoe, F., Ma, X., Nair, S., Remakus, S., Schwendener, R., John, S., Shlomchik, M., and Sigal, L.J. (2015). Sequential activation of two pathogen-sensing pathways required for type I interferon expression and resistance to an acute DNA virus infection. Immunity 43, 1148–1159.

Miller, J.C., Brown, B.D., Shay, T., Gautier, E.L., Jojic, V., Cohain, A., Pandey, G., Leboeuf, M., Elpek, K.G., Helft, J., et al.; Immunological Genome Consortium (2012). Deciphering the transcriptional network of the dendritic cell lineage. Nat. Immunol. 13, 888–899.

Yoneyama, H., Matsuno, K., Toda, E., Nishiwaki, T., Matsuo, N., Nakano, A., Narumi, S., Lu, B., Gerard, C., Ishikawa, S., and Matsushima, K. (2005). Plasmacytoid DCs help lymph node DCs to induce anti-HSV CTLs. J. Exp. Med. 202, 425–435.

Immunity 46, 1–15, February 21, 2017 15