Dynamic Imaging of T Cell-Parasite Interactions in the Brains of Mice Chronically Infected with Toxoplasma gondii

This information is current as Marie Schaeffer, Seong-Ji Han, Tatyana Chtanova, Giel G. of September 28, 2021. van Dooren, Paul Herzmark, Ying Chen, Badrinath Roysam, Boris Striepen and Ellen A. Robey J Immunol 2009; 182:6379-6393; ; doi: 10.4049/jimmunol.0804307

http://www.jimmunol.org/content/182/10/6379 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2009/05/01/182.10.6379.DC1 Material http://www.jimmunol.org/ References This article cites 70 articles, 25 of which you can access for free at: http://www.jimmunol.org/content/182/10/6379.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 28, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2009 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Dynamic Imaging of T Cell-Parasite Interactions in the Brains of Mice Chronically Infected with Toxoplasma gondii1

Marie Schaeffer,* Seong-Ji Han,* Tatyana Chtanova,* Giel G. van Dooren,† Paul Herzmark,* Ying Chen,‡ Badrinath Roysam,‡ Boris Striepen,2† and Ellen A. Robey2*

The intracellular parasite Toxoplasma gondii can establish persistent infection in the brain of a mammalian host, a standoff that involves the active participation of host CD8 T cells to control infection. CD8 T cells generally protect against intracellular pathogens by local delivery of effector molecules upon recognition of specific pathogen Ags on invaded host cells. However, the interactions between CD8 T cells, T. gondii, and APCs in the brain have not yet been examined. In this study we have used a mouse infection model in conjunction with two-photon microscopy of living brain tissue and confocal microscopy of fixed brain sections to examine the interactions between CD8 T cells, parasites, and APCs from chronically infected mice. We found that Ag-specific

CD8 T cells were recruited to the brains of infected mice and persisted there in the presence of ongoing Ag recognition. Cerebral Downloaded from ,CD8 T cells made transient contacts with granuloma-like structures containing parasites and with individual CD11b؉ APCs including some that did not contain parasites. In contrast, T cells ignored intact Ag-bearing cysts and did not contact astrocytes or neurons, including neurons containing parasites or cysts. Our data represent the first direct observation of the dynamics of T cell-parasite interactions within living tissue and provide a new perspective for understanding immune responses to persistent pathogens in the brain. The Journal of Immunology, 2009, 182: 6379–6393. http://www.jimmunol.org/ mmune responses to pathogens typically take place within host-pathogen interactions in vivo. T. gondii is a protozoan para- highly organized, tightly packed tissue environments, a fact site that infects a wide variety of warm-blooded species, including I that has limited our ability to directly observe interactions mice and humans. During the initial acute phase of the infection, between host immune cells and pathogens. Recently, the use of rapidly dividing tachyzoites hijack host cellular machinery to two-photon microscopy, together with the development of genet- spread throughout the body but are eventually brought under con- ically encoded fluorescent reporters, has made it possible to ex- trol by a strong adaptive immune response (10–13). Some para- amine the behavior of cells within living tissues (1). This approach sites convert into slow-growing bradyzoites, which form cysts in has been particularly informative in the study of the immune sys- brain and muscle that persist for the lifetime of the host and ap- by guest on September 28, 2021 tem and has been extensively used to examine immune responses parently effectively evade destruction by the immune system (14, to model Ags within lymph nodes (2–5). Some recent imaging 15). T. gondii infections in humans are generally asymptomatic, studies have examined innate immune responses during infection although reactivation of intracerebral cysts can lead to toxoplasmic (6–9), but very little is known about the behavior of pathogens in encephalitis in immunocompromised hosts (16). the face of an ongoing adaptive immune response and the cellular Immune responses to pathogens are finely balanced to allow for dynamics of adaptive immune cells, such as T cells, during their control of infection while limiting damage to host tissues. No- interactions with pathogens within infected tissues. where is this balance more crucial than in the brain, a tissue that is Toxoplasma gondii infection in mice provides an excellent ex- essential, fragile, and has a limited capacity to regenerate (17). perimental model system to investigate the cellular dynamics of Control of T. gondii in the brain during chronic infection requires an ongoing adaptive immune response, with CD8 T cells playing a key role (18–22). In many settings effector CD8 T cells can form *Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720; †Center for Tropical and Emerging Global Diseases and Department of lasting contacts leading to the killing of target cells, although dur- Cellular Biology, University of Georgia, Athens, GA 30602; and ‡Department of ing T. gondii infection the production of proinflammatory cyto- Electrical, Computer, and System Engineering, Rensselaer Polytechnic Institute, kines such as IFN-␥ appears to be more important for CD8-medi- Troy, NY 12180 ated protection (22–24). Given that cytokines could act in a Received for publication December 22, 2008. Accepted for publication March 11, 2009. paracrine fashion to stimulate other cell types, such as macro- The costs of publication of this article were defrayed in part by the payment of page phages and astrocytes, to control T. gondii growth (11–13, 25, 26) charges. This article must therefore be hereby marked advertisement in accordance it is conceivable that CD8 T cells could provide protection without with 18 U.S.C. Section 1734 solely to indicate this fact. the need to stably contact invaded host cells. Studies of leukocytes 1 This work was funded by the National Institutes of Health (to B.S. and E.R.), isolated from brains of chronically infected mice have provided Human Frontier Science Program Fellowship (to T.C.), and C. J. Martin Overseas Fellowship 400489 from the Australian National Health and Medical Research Coun- important information about CD8 T cells (24, 27–29); however the cil (to G.v.D.). nature of interactions between CD8 T cells and brain APCs has not 2 Address correspondence and reprint requests to Dr. Ellen A. Robey, Department of yet been explored. Molecular and Cell Biology, 471 Life Sciences Addition, University of California, A related question concerns how CD8 T cells detect the pres- Berkeley, CA 94720. E-mail address: [email protected] or Dr. Boris Striepen, Center for Tropical and Emerging Global Diseases and Department of Cellular Bi- ence of T. gondii in the brain. CD8 T cells typically recognize ology, University of Georgia, Paul D. Coverdell Center, 500 D. W. Brooks Drive, peptides derived from intracellular pathogens presented by MHC Athens, GA 30602. E-mail address: [email protected] class I molecules on infected host cells. Parasites convert into bra- Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 dyzoites within living host cells (30, 31), and cysts within the www.jimmunol.org/cgi/doi/10.4049/jimmunol.0804307 6380 IMAGING T. gondii IN THE BRAIN brains of chronically infected mice reside within intact cells (32). infections in BALB/c mice). Mice were sacrificed between 9 and 80 days However, whether CD8 T cells can recognize and respond directly postinfection. to parasite Ags on the surface of cyst-containing cells is not Ex vivo analysis of tissue samples known. To address these issues, we have used a mouse infection model To determine the number of intracerebral cysts, brains were homogenized and two-photon microscopy to examine the interactions between and resuspended in 10 ml of PBS (Invitrogen). Five hundred microliters of brain suspension was analyzed microscopically for the presence of cysts. fluorescently labeled parasites, Ag-specific T cells, and potential RFP fluorescence from the cysts was used for counting on a Nikon Eclipse APCs within the living brain tissue of chronically infected mice. TS100 upright microscope with a ϫ20 air objective (Nikon ϫ20/1.2; work- We found that Ag-specific CD8 T cells were recruited to the brains ing distance ϭ 3.1 inches) using a mercury lamp (X-cite 120) and a Texas of infected mice and persisted there in the presence of ongoing Ag Red filter. For intracellular IFN-␥ staining, cell suspensions were incubated for 5 h recognition. Cerebral CD8 T cells underwent Ag-dependent tran- at 37°C in the presence of brefeldin A (1 ␮g/ml), and C57BL/6 splenocytes sient arrest near parasites, but seldom formed stable contacts with were pulsed with the OVA peptide (SIINFEKL) for1hat37°C at 2 ϫ 106 APCs and did not form large T cell clusters. Isolated parasites were cells/ml in complete RPMI 1640 medium. After incubation, the cells were primarily found within CD11bϩ APCs, and T cell contacts oc- washed in FACS buffer and permeabilized using a Cytofix/Cytoperm kit curred both with invaded host cells and with host cells that were (BD Pharmingen) and stained before analysis by flow cytometry. The fol- lowing Abs (from eBioscience) were used for flow cytometric analysis: near parasites but did not themselves harbor parasites. We also FITC-conjugated anti-CD45 (clone 30-F11); anti-B220 (clone RA3-6B2); observed T cells making transient contacts with granuloma-like anti-CD4 (L3T4, clone RM4-5); anti-IFN-␥; anti-CD45.2 (clone 104); PE- structures that enclosed parasites. In contrast, T cells ignored intact Cy5-conjugated anti-CD8␣ (Ly-2, clone 53-6.7); PE-Cy5-conjugated

Ag-bearing cysts and did not contact astrocytes or neurons, includ- CD45.1 (A20); PE-Cy5-conjugated anti-CD11b (clone M1/70); and PE- Downloaded from Cy5-conjugated anti-CD11c (clone N418). PE-Texas Red anti-CD8␣ ing neurons containing parasites or cysts. Our data represent the (clone 5H10) was from Caltag Laboratories. Acquisitions were performed first direct observation of the dynamics of T cell-parasite interac- with a Coulter Epics XL-MCL flow cytometer (Beckman-Coulter), and tions within living tissue and provide a new perspective for un- data were analyzed with FlowJo software (Tree Star). derstanding immune responses to persistent pathogens in the brain. Mononuclear cells from the brain were prepared using a protocol adapted from Reichmann et al. (44). Briefly, brains were digested1hat 37°C with 1 mg/ml collagenase type IA (Sigma-Aldrich) and 0.1 mg/ml Materials and Methods DNase I (Roche) in complete RPMI 1640. Brains were dissociated and http://www.jimmunol.org/ Parasites and cell culture filtered through 70-␮m cell strainers (BD Biosciences) before centrifuga- tion for 10 min at 200 ϫ g. Cells were resuspended in 60% Percoll (GE In all experiments, T. gondii of the Prugniaud strain were used. To generate Healthcare), overlaid with 30% Percoll, and centrifuged for 20 min at Prugniaud parasites expressing tandem dimeric tomato red fluorescent pro- 1000 ϫ g. The infiltrating mononuclear cells were collected from the gra- tein (33), we transfected parasites with the pCTR2t vector (7) using stan- dient interface, the remaining RBC were lysed using 0.83% ammonium dard protocols (34). To enrich for parasites stably expressing td-Tomato, chloride (Sigma-Aldrich), and cells were washed twice before analysis. we subjected transfected parasites to several rounds of cell sorting using a Cell suspensions from spleen, lymph nodes, and brain were filtered, MoFlo cytometer (DakoCytomation) with an Enterprise 631 laser tuned to stained in FACS buffer (PBS, 2% FBS, and 2 mM EDTA) and analyzed 488 nm for excitation and an emission filter with a band pass of 570/40 nm. by flow cytometry. For CFSE labeling, splenocytes from OT-1 Rag2- During the final round of cell sorting we deposited parasites into 96-well ␮

CD45.1 mice were prepared and labeled with 8 M CFSE (Invitrogen) by guest on September 28, 2021 plates to obtain clonal lines. To generate a Prugniaud/tdTomato strain ex- in PBS at 107 cells/ml for 7 min at room temperature. Cells were pressing a secreted form of OVA, we digested the ptubP30OVA/sagCAT washed in medium with 10% FBS. A total of 1–1.5 ϫ 107 cells were vector (35) with HindIII and PacI to remove the CAT cassette. We digested injected in the tail veins of CD45.2 mice infected with OVA-expressing pBSSKϩ SAG1/Ble/SAG1 (a gift from D. Sibley, Washington University, parasites 3–4 wk previously. CFSE dilution in various organs was an- St. Louis, MO) with HindIII and PacI and ligated the excised Ble cassette alyzed by flow cytometry with gating on CD45.1ϩCD8ϩ cells 2–5 days into the P30OVA-expressing vector. This generated the vector pBT/ posttransfer. P30OVA, which expresses secreted P30-OVA from the ␣-tubulin promoter and contains a Ble gene for phleomycin selection in T. gondii. We trans- Generation of hematopoietic chimeric mice fected this construct into Prugniaud/tdTomato parasites and subjected them to two rounds of phleomycin selection as described previously (36) before To generate mice containing labeled OT-1 CD8 T cells, we injected bone cloning into 96-well plates. Clonal lines stably expressing secreted OVA marrow from OT-1 Rag2-GFP mice into newborn C57BL/6 or CD11c- were identified by immunofluorescence assays. Tachyzoites were cultured yellow fluorescent protein (YFP) mice as described (45). This resulted in on monolayers of human foreskin fibroblasts and purified immediately be- mice with ϳ0.2% of splenocytes corresponding to donor cells before in- fore mouse infection as described (7). fection. After infection, the vast majority of GFP-labeled cells of donor origin in the brain corresponded to OT-1 T cells due to the selective Mice and in vivo infections expansion and recruitment of these cells (data not shown). For produc- tion of irradiation chimeras, a 1/10 mixture of Rag2-GFP and CD45.1 All mice were bred and housed in pathogen-free conditions at the Amer- wild-type bone marrow was used to reconstitute sublethally irradiated ican Association of Laboratory Care-approved animal facility at the Life (300 rad, 3 Gy) CD45.2 C57BL/6 mice by i.v. injection. Mice were Science Addition, University of California, Berkeley, CA. All animal ex- infected at least 6 wk after reconstitution. Naive OT-1 Rag2-CFP cells periments were approved by the Animal Care and Use Committee of Uni- (1–1.5 ϫ 107/mouse) were transferred by tail vein injection just prior to Ϫ Ϫ Ϫ Ϫ versity of California Berkeley. C57BL/6, C57BL/6 CD4 / , and CD8␣ / infection (1–5 h; 0.7–1.3% CD8ϩCFPϩ cells in the spleen at the time of Ϫ Ϫ (37, 38), Rag2 / , and BALB/c mice were purchased from The Jackson infection). Flow cytometric analysis of leukocytes from the brains of Laboratory. Ubiquitin-GFP mice (39) were a gift from B. Schaefer (Uni- infected chimeric mice confirmed that Ͼ98% of GFPϩ cells were formed Services University, Bethesda, MD). F5 transgenic mice (40) were CD11bϩ (data not shown). a gift from D. Kioussis (National Institute for Medical Research, London, U.K.). OT-1 Rag2- (41) transgenic mice were from Taconic Farms. In vitro activation of CD8 T cells CD11cYFP transgenic reporter mice (42) and actin promoter cyan fluo- rescent protein (CFP)3 transgenic mice (43) were provided by M. Nussen- CD8 T cells from OT1 or F5 TCR transgenic mice were activated in vitro zweig’s laboratory (The Rockefeller University, New York, NY). Mice based on a protocol from Purbhoo et al. (46). Briefly, splenic leukocytes were bred to generate OT-1 Rag2-GFP, F5 Rag2-CFP, and OT-1 Rag2- from TCR transgenic mice were isolated by dissociation followed by fil- CFP mice. For all experiments, mice were infected with freshly purified tering though 70-␮m cell strainers (BD Biosciences). RBC were lysed tachyzoites i.p. (400 tachyzoites for infections in C57BL/6 and 4000 for using 0.83% ammonium chloride (Sigma-Aldrich). C57BL/6 splenocytes were used as stimulator cells and were resuspended in complete RPMI 1640 medium at 2 ϫ 106/ml and pulsed with peptide (the OVA peptide 3 Abbreviations used in this paper: CFP, cyan fluorescent protein; GFAP, glial fibril- SIINFEKL or the A/NT/60/68 influenza virus nucleoprotein peptide lary acid protein; MAP, microtubule-associated protein; RFP, red fluorescent protein; ASNENMDAM (NP366–374) were used at 100 nM for 1h at 37°C). Stim- YFP, yellow fluorescent protein. ulators were then washed and irradiated at 2000 rad. A total of 2 ϫ 105 The Journal of Immunology 6381

OT-1 or F5 responder cells were mixed with 5 ϫ 105 peptide-pulsed stim- ilar method is used to identify voxels corresponding to the RFP (parasite) ulator cells in a volume of 1 ml in 24-well plates and incubated at 37°C in channel. the presence of 50 U/ml IL-2 (National Cancer Institute Preclinical Re- pository, Frederick, MD) for 5–7 days, until use. For i.v. injection, CTLs Statistical analysis ϫ 7 were washed twice and resuspended in PBS. A total of 1.5–2 10 cells/ Values were expressed as mean Ϯ SEM. Levels of significance were cal- mouse were transferred. culated by the Mann-Whitney U test using the GraphPad Prism program. ءءء Ͻ ءء Ͻ Fluorescence microscopy Differences were considered significant at p 0.05 ( , p 0.005; , p Ͻ 0.0005). Mice were infected with fluorescent parasites, sacrificed at various times after infection, and brains were immediately removed and prepared for Results imaging. Confocal microscopy was performed using 20-␮m frozen brain An experimental system to examine T cell responses to sections and z projections of the entire section are shown unless otherwise Toxoplasma gondii in the brain during chronic infection indicated. For two-photon microscopy, brains were sliced using a vi- bratome in oxygenated artificial cerebrospinal fluid (125 mM NaCl, 2.5 Toxoplasma Ags and the T cell Ag receptors that recognize them mM KCl, 25 mM glucose, 25 mM NaHCO3, 1.25 mM NaH2PO4,2mM are just beginning to be characterized (48, 49). Therefore, to track CaCl2, and 1 mM MgCl2). The 1-mm-thick transverse sections were main- Ag-specific T cell responses we generated an engineered version tained in position using a 1-mm slice-holder (Warner Instruments) while being superfused with oxygenated artificial cerebrospinal fluid medium at of the cyst-forming Prugniaud strain of Toxoplasma that expresses 35°C and imaged by two-photon microscopy using a custom built micro- the model Ag chicken OVA, thus allowing us to use OT-1 OVA- scope with a 20 ϫ 0.95 numerical aperture water dipping objective as specific TCR transgenic mice (41) to track Ag-specific CD8 T cell described previously (7, 47). Imaging volumes (255 ϫ 212 ϫ 30–90 ␮m responses. The OVA peptide was expressed as a protein fusion that or 175 ϫ 142 ϫ 30–90 ␮m) corresponded to regions of the brains at least was secreted into the parasitophorous vacuole, a form that can be Downloaded from 20 ␮m below the cut site and extending up to 200 ␮m below the surface of the slice and were scanned every 13–37 s for 20–40 min. Two-photon recognized in vivo by CD8 T cells (28, 35). We also engineered excitation was achieved using a Spectra-Physics MaiTai laser tuned to 920 the parasites to express a RFP derivative to facilitate imaging or 900 nm. GFP and tdTomato, CFP and GFP, and GFP and YFP emission of the parasites. signals were separated using 560, 495, and 510 dichroic mirrors and col- The course of Toxoplasma infection in mice is highly dependent lected using three nondescanned detectors. For confocal fluorescence microscopy, brains of infected mice were on both the parasite and the mouse strains. We therefore began our

fixed with 4% formalin and 10% sucrose in PBS for 1 h, sequentially study by characterizing our infection model. Upon infection with http://www.jimmunol.org/ submerged in 10, 20, or 30% sucrose for 18–24 h each, and frozen over dry a low dose of the engineered Prugniaud parasites, C57BL/6 (B6) ice in OCT. Twenty-micrometer serial sections were generated by cryo- mice developed a chronic progressive disease characterized by Ϫ sectioning (Microm H550; Microm) and stored at 80°C. For CD11b higher leukocyte recruitment and higher cyst burden compared staining, sections were brought to room temperature, fixed with cold 4% paraformaldehyde for 20 min, air dried, and incubated with 10% mouse with BALB/c mice (Fig. 1, A and B), consistent with previous serum in Fc blocking buffer (2.4G2 supernatant) for2hat37°C. The tissue reports using the ME49 Toxoplasma strain (50, 51). The brains of was stained with biotin-conjugated anti-CD11b (R&D Systems) for2hat infected mice contained large numbers of leukocytes that consisted 37°C. For staining using Abs to OVA (Sigma-Aldrich), glial fibrillary acid largely of T cells with CD11bϩ cells making up most of the re- protein (GFAP) (DakoCytomation), microtubule-associated protein (MAP) 2 (Thermo Scientific), and Dolichol lectin (Vector Laboratories), sections mainder (Fig. 1, C and D), which is in line with previous reports were fixed with 100% methanol for 5 min, washed three times in PBS, and (52). In contrast, very few leukocyte-like cells could be isolated by guest on September 28, 2021 blocked with 10% goat serum in Fc blocking buffer for2hat37°C. Goat from the brains of uninfected B6 mice (Fig. 1, A and D). These anti-rabbit Alexa Fluor 647 (Invitrogen) secondary was used for OVA and were predominantly intermediate or negative for the blood cell GFAP staining, and goat anti-mouse IgG-allophycocyanin (Molecular marker CD45 and included CD11bϩCD45int (where “int” is inter- Probes) secondary was used for MAP2 staining. Slides were then washed four times in PBS and incubated with streptavidin-Alexa Fluor 633 (In- mediate) cells that corresponded to resident microglia (Fig. 1, C ϩ vitrogen) for2hat37°C. Sections were then washed four times in PBS and and D) (53). CD11b cells in the brains of infected mice included cover-slipped using Vectashield (Vector Laboratories) mounting medium. cells with surface marker phenotypes associated with microglia Stained sections were visualized on Zeiss 510 Axioplan META NLO up- int ϩ high ϩ Ϫ ϫ ϫ (CD45 CD11b ), macrophages (CD45 CD11b CD11c ), right microscope with a 10 air objective (Plan-Neofluar 10/0.3) and a high ϩ ϫ40 oil objective (Plan-Neofluar ϫ40/1.3 oil; working distance ϭ 0.17 and dendritic cells (CD45 CD11c ) (Fig. 1, C and D, and data mm) using 488-, 543-, and 633-nm laser beams. Images were analyzed not shown), consistent with previous reports from Toxoplasma in- using Imaris Bitplane. fection and other models of brain inflammation (54). To confirm the importance of CD8 T cells in our infection model, we com- Data analysis pared the survival of wild-type, CD4Ϫ/Ϫ, and CD8Ϫ/Ϫ mice to The x, y, and z coordinates of individual cells over time were obtained infection. Survival was strongly dependent on the presence of CD8 using Imaris Bitplane software. Motility parameters were calculated using T cells, whereas the loss of CD4 T cells had only a minor effect on Matlab (code available upon request). The parameters reported here in- clude average speed (defined as pathlength over time), arrest coefficient survival (Fig. 1E), consistent with previous studies pointing to the (percentage of time points of the total time points for the track for which key protective role of CD8 T cells during T. gondii infection the instantaneous speed is Ͻ2 ␮m/min), instantaneous speed (pathlength (18–22). between two time points averaged over three successive time points), and To track Ag-specific CD8 T cell responses in vivo, we generated displacement rate (straight line distance between first and last time points divided by time). To measure the cell density, regions of 80 ϫ 80 ϫ 30–90 mice containing a small number of labeled OT-1 CD8 T cells ␮m (imaging volumes obtained by two-photon microscopy) or regions of infected with a low dose of parasites and analyzed the mice at 80 ϫ 80 ϫ 20 ␮m (imaging volumes obtained by confocal microscopy) various time points postinfection. OT-1 T cells were found in large were centered on parasites or cysts or in areas without red fluorescence and numbers in the brain by day 9 postinfection (ϳ2 ϫ 105 cells/brain) the number of OT-1-GFP T cells was counted using Imaris Bitplane soft- (Fig. 2A). OT-1 T cell recruitment was more rapid than the overall ware and normalized per unit of volume (106 ␮m3). For some data sets a spectral unmixing procedure was used to overcome the spectral cross talk increase in leukocyte number, which did not peak until 3 wk in the image data (Y. Chen, unpublished observations). Briefly, we average postinfection (Fig. 1A). OVA-specific T cells represented ϳ17% all voxel intensities in small spatial neighborhoods (5 voxels wide) to make of total leukocytes and 40% of CD8 T cells recovered from the the procedure robust to imaging noise and compute the Otsu automatic brains of infected mice (Fig. 2B and data not shown). Similar ex- threshold individually for each channel. Voxels for which the brightest channel exceeds the threshold for the GFP channel are labeled T cell vox- pansion of OT-1 T cells was observed in spleen and lymph nodes els. The remaining voxels for which the brightest channel exceeds the of infected mice (Fig. 2B). In contrast, OT-1 T cells did not expand threshold for the YFP channel are considered dendritic cell voxels. A sim- substantially in mice infected with parasites that did not express 6382 IMAGING T. gondii IN THE BRAIN Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 1. Leukocyte recruitment and protective effect of CD8 T cells during T. gondii chronic infection. A and B, C57BL/6 mice show greater leukocyte recruitment to the brain (A) and a higher cyst burden (B) compared with BALB/c following T. gondii infection. The number of brain-derived leukocytes recovered after Percoll isolation from infected C57BL/6 and BALB/c mice as a function of time p.i is shown. Data are represented as mean Ϯ SEM. C, Representative flow cytometry plots of cells isolated from the brains of C57BL/6 mice that were either uninfected or infected 28 days earlier with T. gondii. Percoll-purified leukocyte samples were stained with Abs to CD45 and CD11b. The percentage of the CD45ϩCD11bϪ cells in the uninfected sample varied between uninfected brain samples but was always Ͻ1000 total cells/brain and likely derived from contaminating blood. D, Composition of the brain leukocytes from chronically infected brains of C57BL/6 mice. Percoll-purified leukocytes isolated from the brains of infected C57BL/6 brains were analyzed by flow cytometry for the indicated markers. Data for infected mice are averages from at least four mice analyzed 28–80 days postinfection. Data for uninfected are from one representative experiment based on pooled samples from three perfused mice. E, Survival of C57BL/6 mice to infection is dependent on CD8 T cells. Ten mice for each group (wild type, CD4Ϫ/Ϫ, and CD8␣Ϫ/Ϫ) were infected with 400 tachyzoites i.p. The percentage of mice surviving as a function of time is shown.

OVA or in uninfected mice. OT-1 T cells purified from various response is sufficient to provide substantial protection and that the organs of mice infected with OVA-expressing parasites expressed frequency of Toxoplasma-specific naive T cells is not a limiting IFN-␥ after a short in vitro restimulation with the OVA peptide, factor in this setting. Thus, the transferred OT-1 T cells did not unlike OT-1 T cells from uninfected mice or mice infected with appear to dramatically alter the overall CD8 T cell response to parasites that did not express OVA (Fig. 2C). The majority of infection, and they provided a convenient means for monitoring splenic and cerebral OT-1 T cells from mice infected with OVA- the CD8 response in vivo. expressing parasites lacked expression of programmed death-1 re- To analyze Ag-specific T cell proliferation and trafficking to the ceptor, a marker of T cell exhaustion (55), and expressed high brain, we transferred naive CFSE-labeled, CD45.1 congenically levels of the activation marker CD44, and the lysosomal marker marked OT-1 cells into mice that had been infected with OVA- CD107a (LAMP-1) (data not shown). Thus OT-1 T cells in the expressing T. gondii at least 3 wk earlier. Two days after transfer, brains of infected mice have the phenotype and effector functions donor T cells were not detectable in the brain, although they were of activated CD8 T cells. Importantly, the overall course of infec- readily detectable in the spleen and had already undergone some tion, as measured by cyst burden, leukocyte recruitment to the cell division as indicated by dilution of the CFSE label (Fig. 2, D brain, and survival, was similar whether or not mice received an and E). By day 3 after T cell transfer a small number of OT-1 T OT-1 T cell transfer and whether or not parasites expressed OVA cells could be detected in the brain, and these cells had already (data not shown). This implies that the endogenous CD8 T cell undergone substantial CFSE dilution. By day 5, OT-1 T cells were The Journal of Immunology 6383 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 2. Ag-dependent responses of naive OVA-specific OT-1 CD8 T cells responses during chronic T. gondii infection. A–C, Chimeric mice containing small numbers of naive GFP-labeled OT-1 T cells were infected and cell populations were analyzed at the indicated times by flow cytometry. Data are represented as mean Ϯ SEM. A, Number of OT-1 T cells recovered after Percoll purification of brains from uninfected mice (dashed line), mice infected with parasites expressing OVA (filled circles), or mice infected with parasites without OVA (open circles) as a function of time postinfection. Error bars are from at least three replicates. B, Expansion of OVA-specific T cells in different tissues. Percentages of OT-1 T cells from the indicated samples 24–39 days postinfection are shown. Brain, Percoll isolated brain leukocytes; Spl, spleen; DC, dorsal cervical lymph node; VC, ventral cervical lymph node; ing, inguinal lymph node; MLN; mesenteric lymph node. Error bars are from at least four replicates. Data are the percentage of OT-1 T cells from a live gated population and correspond to ϳ1 ϫ 107 total OT-1 T cells in the spleen, 1.5 ϫ 105 in peripheral lymph nodes, and 8 ϫ 105 in mesenteric lymph node samples. C, OT-1 T cells from infected mice (24–39 days postinfection) produce IFN-␥. Cells from the indicated tissues were stimulated with OVA peptide in vitro for 5 h and then stained for intracellular IFN-␥ and analyzed by flow cytometry. Filled bars are samples from mice infected with OVA-expressing parasites, gray bars are samples from mice infected with parasites without OVA, and open bars are samples from uninfected mice. Percentages of IFN-␥-producing cells among gated OT-1-GFP CD8 T cells are shown. Brain, Percoll isolated brain leukocyte; Spl, spleen; DC, dorsal cervical lymph node; VC, ventral cervical lymph node; ing, inguinal lymph node; MLN, mesenteric lymph node. Error bars are from at least four replicates. D and E, Chronically infected mice (21–48 days postinfection) were injected with naive allelically marked (CD45.1) CFSE-labeled OT-1 T cells, and OT-1 T cells in brain and spleen were quantitated by flow cytometry at the indicted times posttransfer. D, Representative flow cytometry plots of gated on donor derived CD8ϩCD45.1ϩ cells are shown. Percentages of OT-1 cells from the total brain-derived leukocyte population (bottom row) or from the total splenocyte population (top row) are written underneath each plot. Numbers in gates represent percentages of gated OT-1 T cells that had diluted CFSE. E, Compiled data for the indicated times and tissues after transfer. DC, dorsal cervical lymph node; VC, ventral cervical lymph node; ing, inguinal lymph node; MLN, mesenteric lymph node; Spl, spleen. Data are represented as mean Ϯ SEM based on at least three replicates. Data are the percentages of cells that have diluted CFSE as a percentage of the total transferred allelically marked OT-1 T cells (CD45.1ϩ) and corresponds to ϳ2 ϫ 106 OT-1 T cells in the spleen and 1 ϫ 105 in lymph node samples at days 2–5 posttransfer, 400 OT-1 T cells in the brain at day 3, and 8000 OT-1 T cells in the brain at day 5 posttransfer. 6384 IMAGING T. gondii IN THE BRAIN Downloaded from http://www.jimmunol.org/ FIGURE 3. Activated CD8 lymphocytes can enter chronically infected brains in the absence of ongoing antigenic stimulation. A, OT-1 T cells were activated in vitro and then transferred into mice that had been infected 17–34 days earlier with OVA-expressing parasites (filled circles). T cell transfers into mice infected with parasites without OVA (open circles) and infected mice (x) are shown for comparison. OT-1 T cells were quantitated in the brain (left panels) and spleen (right panel) by flow cytometry at the indicated times after transfer. Data are represented as mean Ϯ SEM based on at least three replicates. Data are the percentages of OT-1 T cells as a percentage of total brain leukocytes (upper left panel) or the percentages of OT-1 T cells as a percentage of total splenocytes (right panel). B, Mice were transferred with 2 ϫ 106 in vitro activated OT-1 T cells (filled circles) or left untreated (open circles) and then infected with a high dose (105) of OVA-expressing tachyzoites. Mice were monitored daily and euthanized at the first signs of illness (n ϭ 14 mice for each group). by guest on September 28, 2021 readily detectable in the brain and virtually all had extensively ␮m in brain tissue, and we have yet to observe a cyst or foci of diluted the CFSE label. As expected, there was very little prolif- infection sufficiently close to the skull to be accessible by this eration of OT-1 T cells in mice infected with parasites not ex- approach. pressing OVA or in uninfected mice. Interestingly, when in vitro Parasites exist in the brain as bradyzoites that form tissue cysts activated OT-1 T cells were transferred into chronically infected within infected cells and as individual tachyzoites and bradyzoites, mice, T cells could be detected in the brain as well as spleen by day some of which may have emerged from ruptured cysts (59–61). 2, regardless of whether the parasites expressed OVA, although Ag To examine which of these forms of the parasites T cells recog- was necessary for the retention/expansion of OT-1 cells in the nized, we examined the location of OT-1 T cells relative to OVA- brain Ͼ5 days after T cell transfer (Fig. 3). Together, these data expressing parasites in the brains of chronically infected mice (Fig. suggest that there was ongoing Ag presentation in the periphery of 4). Parasite-containing cysts were readily identifiable by their large chronically infected mice, that naive T cells underwent activation size and typical spheroid shape (Fig. 4A, dashed circle). A fluo- and proliferation in the periphery before they trafficked to the rescent signal was also detected in smaller particles whose size and brain, and that the retention of T cells in the brain required ongoing shape suggested that they correspond to single parasites (Fig. 4, A, Ag recognition. boxed areas labeled 1 and 2, and B, arrowheads). Importantly, the OVA Ag is present in cysts within the cyst wall, identified here by T cells ignored cysts but slowed and accumulated near Dolichol lectin staining (Fig. 4C) as expected, given that the pro- individual parasites moter used to drive OVA expression is active in both bradyzoites To examine the location and behavior of T cells and parasites in and tachyzoites. Time-lapse imaging using two-photon micros- the brain, we used a combination of confocal microscopy of Ab- copy of vibratome-cut brain slices revealed a few examples of stained fixed brain sections and two-photon time-lapse microscopy isolated parasites emerging from disintegrating cysts (Fig. 4B and of living brain tissue. For the two-photon imaging studies we chose supplemental video 1),4 indicating that at least some isolated par- to use vibratome-cut brain slices, a tissue preparation that has been asites derived from ruptured cysts. Although OT-1 T cells were extensively used by neurobiologists and provides a robust system found throughout the brain, we noted that the density of OT-1 T in which cells remain viable and functional for extended time pe- cells was highest in regions containing isolated parasites. In con- riods (56). This approach also has the advantage that is it possible trast, T cell density around T. gondii cysts was similar to the den- to scan relatively large areas of the brain to locate foci of infection, sity in regions that did not contain any visible parasite fluorescence which are relatively rare in infected mice (Ͻ2000 cysts distributed (Fig. 4D). This trend was observed both at early time points (days throughout the entire brain). The alternative approach of intravital imaging of the brain (57, 58) is not a feasible approach for this study, because fluorescent detection is limited to a depth of ϳ200 4 The online version of this article contains supplemental material. The Journal of Immunology 6385 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 4. T cells accumulated near isolated parasites in the brains of infected mice. Mice containing small numbers of GFP-labeled OT-1 T cells were infected with RFP-labeled parasites expressing OVA. A, Representative confocal images of frozen brain sections. Red, T. gondii; green, OT-1-GFP. Left panel shows a field containing a group of intact cysts (dashed circle) (25 days postinfection) Right panel shows a field containing isolated parasites (12 days postinfection) Isolated parasites are shown in enlarged insetstotheright. B, Two time points from a time-lapse sequence showing isolated parasites emerging from an intact cyst. Parasites are in red, OVA-specific OT-1 CD8 T cells are in blue, and GFP-labeled APCs are in green. APCs were labeled by generating irradiation chimeras using mixtures of Rag2Ϫ/Ϫ GFP bone marrow mixed with unlabeled wild-type bone marrow as described in Fig. 9. Projections of 50-␮m imaging volumes are shown; the left panel corresponds to a time point before disruption of the cyst (dashed circle), and the right panel corresponds to a time point after disruption (arrows indicate isolated parasites exiting the cysts). This corresponds to supplemental video 1. C, Cysts contain OVA protein. Representative confocal images of a cyst are shown. Red, T. gondii; blue, OVA staining; green, Dolichol lectin (cyst wall) staining. Left panels show the parasite-only channel (top), OVA staining only (bottom), and right panel shows a color merge. D, Quantification of the OT-1 T cell density. The number of OT-1 T cells in defined volumes located in regions without parasite fluorescence (open circles), regions centered on cysts (shaded circles), or regions centered on isolated parasites (filled circles). T cell density is expressed as the number of T cells per unit of volume (106 ␮m3). “E” indicates early time points 10–16 days postinfection and “L” indicates time points 19–39 days postinfection. Data are compiled from both confocal analysis of fixed brain sections (at least two sections for each time category) and from two-photon microscopy of vibratome-cut brain slices (at least five runs for each category from four different days). A Mann-Whitney U test for the difference in values between early and late time points gave a p value of 0.44 (not significant).

10–16) postinfection when the number of OT-1 T cells in the brain plateau (Figs. 1A and 2A). These data suggested that isolated par- had reached a plateau but total leukocyte numbers were still in- asites, but not intact cysts, were visible to CD8 T cells. creasing, as well as at later time points (days 19–39) postinfection The accumulation of CD8 T cells around isolated parasites sug- when overall leukocyte recruitment to the brain had also reached a gested that they might be detecting Ag in these areas. If this were 6386 IMAGING T. gondii IN THE BRAIN Downloaded from

FIGURE 5. T cells slowed and arrested near isolated parasites and during ongoing antigenic recognition. A, OT-1 T cell migration in brain slices from mice infected with OVA-expressing parasites. Graphs show speeds and arrest coefficients for individually tracked T cells from imaging volumes (typically 255 ϫ 212 ϫ 60 ␮m) that contained isolated parasites, intact cysts only, or no visible parasites as indicated. Data were compiled from six runs with isolated parasites, five runs with intact cysts, and seven runs without visible parasites. The samples were analyzed 10–39 days postinfection and no significant /p Ͻ 0.0005, indicating significant http://www.jimmunol.org ,ءءء .(differences in speed or arrest coefficients were observed between early vs late time points (data not shown differences. Panels to the right show projections of three-dimensional imaging volumes representing a single time point from a time-lapse series obtained by two-photon microscopy. Left panel shows a run containing isolated parasites (red, indicated by arrows in enlarged inset). Rightmost panel shows a sample containing an intact cyst and no isolated parasites. The lines show the paths taken by individual OT-1 T cells over a 13-min time period for T cells around parasites (orange lines) or around cysts only (white lines). CD11cYFP reporter-expressing cells are shown in green. This corresponds to supplemental video 2. The impact of proximity to parasites within individual imaging volumes is presented in Fig. 6. B, T cell slowing and stopping induced by ongoing Ag recognition. In vitro activated OT-1 or F5 T cells were transferred into mice infected with OVA-expressing parasites and brains imaged 4–6 days posttransfer. As an additional control, OT-1-GFP blasts were transferred into mice infected with parasites not expressing OVA. Right panels show projections of representative imaging volume representing single time points from a time-lapse series obtained by two-photon microscopy. The lines show

the paths taken by individual T cells over a 13-min time period with Ag-specific cell tracks in orange and the tracks of control T cells (F5 and OT-1 in by guest on September 28, 2021 mice infected with ϪOVA parasites) in white. Arrows indicate the positions of individual RFP-labeled parasites (red). This corresponds to supplemental videos 3 and 4. the case we would expect T cells to slow and stop in these areas, activated OT-1 T cells provided some protection against a high as has been observed for Ag recognition by T cells in other tissues dose of OVA-expressing parasites, confirming that the transferred (2, 3). To examine this question, we performed time-lapse imaging T cells were functional in vivo (Fig. 3B). We activated T cells from using two-photon microscopy (Fig. 5). Overall, the speed of OT-1 OT-1 and F5 TCR transgenic mice in vitro and then compared T cells (5–6 ␮m/min) was substantially lower than that previously their behavior in the brain following transfer into mice that had reported for naive CD8 T cells in lymph nodes, but in line with that been infected 2–4 wk earlier with OVA-expressing parasites. As of previous reports of effector T cell migration within tissues (62, an additional comparison, we also transferred in vitro activated 63). Importantly, OT-1 T cells in imaging volumes that contained OT-1 T cells into mice that had been infected with parasites that isolated parasites showed reduced speed compared with T cells in did not express OVA. Using both approaches, we showed that imaging volumes with only cysts or with no visible parasite fluo- Ag-specific T cells moved more slowly and arrested more fre- rescence. (Fig. 5A, supplemental Fig. S1A, and supplemental video quently than nonspecific T cells (Fig. 5B, supplemental Fig. S1B, 2. Moreover, the arrest coefficient, defined as the proportion of and supplemental videos 3 and 4. Importantly, these differences time that T cells spend not moving, was significantly increased were observed even when OT-1 and F5 T cells were tracked within around isolated parasites (Fig. 5A), which is consistent with the the same imaging volumes (Fig. 5B, lanes labeled “co-transfer”). hypothesis that T cells were actively recognizing Ag in these areas. These data suggest that the slowing and stopping of Ag-specific T The relationship between proximity to parasites and T cell motility cells around isolated parasites were not due solely to changes in could also be observed within a single 255 ϫ 212 ϫ 60-␮m im- the tissue environment but rather were due to local Ag recognition. aging volume with T cells that were relatively close to parasites showing reduced speeds and increased stopping relative to T cells T cell contacts with granuloma-like structures and individual in the field that were more distant from parasites (Fig. 6). APCs in the brain To test whether Ag recognition was required to induce T cell The Ag-dependent slowing and stopping of CD8 T cells in the slowing and stopping in the brains of chronically infected mice, we brains of chronically infected mice was consistent with their in- compared OT-1 T cells to T cells expressing an irrelevant TCR teraction with Ag-bearing cells. To characterize potential APCs, (F5) (40). For this experiment, we took advantage of our obser- we examined fixed brain sections by confocal microscopy using vation that in vitro activated T cells trafficked to the brain of chron- Abs to CD11b, a marker expressed by monocyte/myeloid lineage ically infected mice and persisted there for up to 5 days in the cells and microglia that has been previously shown to be associ- absence of ongoing antigenic stimulation (Fig. 3A). Transfer of ated with parasites in the brains of chronically infected mice (64). The Journal of Immunology 6387 Downloaded from

FIGURE 6. Relationship between T cell motility and distance to parasites. A, Left-hand panel shows a projection of a three-dimensional imaging volume ϫ ϫ ␮

(255 212 60 m) obtained by two-photon microscopy showing OT-1 T cells (green) in brain slices from mice infected with OVA-expressing parasites http://www.jimmunol.org/ (red). Center panel is an enlarged inset from a single plane showing the isolated parasites. Rightmost panel shows the tracks of individual T cells with a dashed box indicating a 100 ϫ 100 ϫ 60-micron volume centered around the fluorescent signal from the parasites. Tracks for T cells that spent most of the time inside the volume are in green, and tracks of T cells that spend most of the time outside the volume are shown in white. Parasite tracks are shown in red. Data are from four of the runs plotted in Fig. 5A. B, Average speeds and arrest coefficients for individually tracked T cells that were within (filled indicate ءءء circles) or outside of (open circles) a volume of 100 ϫ 100 ϫ 60 ␮m centered on the parasites. Red horizontal bars indicate the mean and significant differences ( p Ͻ 0.0005).

In the brains of chronically infected mice we observed spherical, cinity of parasites, nor did we observe any clear examples of tightly packed aggregates of CD11b-expressing cells that may cor- GFAPϩ astrocytes containing parasite fluorescence (Fig. 7B and by guest on September 28, 2021 respond to the “nodules” of leukocytes described previously (32, data not shown). Although we did observe a small number of neu- 59) (Fig. 7Ai and Aii). Isolated parasites were often found within rons containing parasites, these were rare and were not contacted CD11b aggregates (Fig. 7Ai, arrowheads), but they were also by OT-1 T cells (Fig. 7C and data not shown). Previous electron found outside of the aggregates within nearby isolated CD11bϩ microscopy studies indicated that cysts could be found within in- cells (Fig. 7Aii, arrowheads). Approximately one-third (44/122) of tact neurons (32), which is compatible with our observation that isolated parasites appeared to be inside a CD11bϩ cell or aggre- 25% of cysts (4/16) were surrounded by MAP2 staining (Fig. 7D gate. Moreover, the majority of the CD11b aggregates observed and data not shown). These data suggest that astrocytes and neu- contained one or more isolated parasites (20 aggregates with a total rons are not major APC types in the brain, although neurons can of 28 parasites from 12 different sections) and aggregates of harbor parasites and cysts. CD11bϩ cells were not seen around cysts (data not shown). This To allow for dynamic imaging of potential APCs in the brain, is consistent with previous light microscopy studies (32, 59) and we made use of mice bearing a dendritic cell reporter consisting of implies that, like T cells, CD11bϩ cells in the brains of chronically YFP under the control of the CD11c promoter (42). YFP cells infected mice responded to isolated parasites rather than to intact increased in number in the brains of chronically infected mice in cysts. CD11bϩ aggregates were most prominent at late time points parallel with the overall increase in leukocyte number, reaching a postinfection (19 days postinfection), whereas at earlier times plateau of ϳ5 ϫ 104 reporter expressing cells per brain at 3 wk postinfection parasite fluorescence was more often seen associated postinfection (Fig. 1 and data not shown). Flow cytometric anal- with isolated CD11bϩ cells (Fig. 7Aiii, boxed area labeled 1). In ysis showed that the majority of YFP cells expressed CD11b, al- these areas we also observed individual T cells contacting isolated though only ϳ25% expressed surface CD11c (data not shown), CD11bϩ cells, including those that did and did not contain visible consistent with previous reports (42). Confocal analysis of brain parasite fluorescence (Fig. 7Aiii, boxed areas labeled 1 and 2). sections showed that a subset of cells within the CD11b aggregates Ag-specific T cells were also found in contact with the aggregates also expressed the CD11cYFP reporter (Fig. 8A). Thus, the of CD11bϩ cells and sometimes inside them (Fig. 7A), in line with CD11cYFP reporter marks a subpopulation of cells within the higher density of T cells in areas with isolated parasites (Fig. CD11bϩ aggregates and could serve as a marker for identifying 4A). Interestingly, T cells within CD11bϩ aggregates did not pref- aggregates during dynamic imaging. Indeed, two-photon micros- erentially cluster around the isolated parasites, but rather appeared copy of brains slices from infected CD11cYFP reporter mice re- to be evenly distributed throughout the entire aggregate (Fig. 7A). vealed spherical aggregates of YFP-bright cells containing isolated To examine other potential APCs, we also stained brain sections parasites and surrounded by Ag-specific T cells (Fig. 8B and sup- from infected mice with Abs specific for an astrocyte marker, plemental video 5). Isolated parasites within aggregates were gen- GFAP, and a neural marker, MAP2. We did not see any indication erally not enclosed by reporter bright cells (Fig. 8B, inset 1), al- of OT-1 T cells clustering around astrocytes that were in the vi- though we occasionally observed individual reporter bright cells 6388 IMAGING T. gondii IN THE BRAIN Downloaded from http://www.jimmunol.org/

FIGURE 7. Potential APCs in the brains of infected mice. Mice containing small numbers of GFP-labeled OT-1 T cells were infected with RFP-labeled parasites expressing OVA and brains were imaged at the indicated times postinfection. A, Confocal analysis of fixed brain sections showing CD11bϩ cells (blue), parasites (red), and OT-1 T cells (green). i, Left panel shows a field containing an intact cyst (dashed circle) and an aggregate of CD11bϩ cells (white box) containing isolated parasites and surrounded by OT-1 T cells (day 25 postinfection). Middle panel shows an enlarged view of the boxed area without a CD11b signal and with arrowheads indicating the position of isolated parasites. Right panel shows a single optical section from the boxed area. Only one of the isolated parasites is visible in this z-plane. ii, This example shows parasites inside CD11bϩ cells just outside a nodule of CD11bϩ cells (day 25 postinfection). Arrows indicate the position of parasites (red). iii, Representative example of field with OT-1 T cells interacting with individual CD11bϩ by guest on September 28, 2021 cells (day 12 postinfection). Boxes show T cells interacting with CD11bϩ cells containing parasites (boxed area labeled 1), or not (boxed area labeled 2) with enlarged views of single optical sections shown in the right-hand panels. Arrows indicate the position of parasites (red). B, Confocal analysis of a fixed brain section showing GFAPϩ cells (blue), parasites (red), and OT-1 T cells (green) (day 31 postinfection). C and D, Confocal analysis of fixed brain sections showing MAP2ϩ cells (blue), parasites (red), and OT-1 T cells (green) (day 31 postinfection). C, A field containing a MAP2ϩ cell containing parasite fluorescence (boxed area) with an enlarged views of the boxed area to the right. D, A field containing a cyst surrounded by MAP2 staining. Right panel shows the same view with the MAP2 staining only.

forming intimate associations with cysts (supplemental video 6). be distributed throughout the granuloma-like structures rather than Time-lapse imaging of brain slices from CD11cYFP reporter mice being localized to individual parasite-containing cells. revealed that OT-1 T cells migrated actively around and through Because most parasites were found in CD11bϩ but not CD11c aggregates of reporter-positive cells, whereas the reporter-positive YFP reporter-positive cells (Fig. 8B and data not shown), we de- cells were largely sessile (Fig. 8B, right panel, and supplemental veloped an alternative approach for dynamic imaging of potential video 5). These structures, which are composed of balls of static APCs in the brain. We generated hematopoietic chimeras recon- myeloid lineage cells enclosing isolated parasites and are sur- stituted with mixtures of bone marrow from Rag2Ϫ/Ϫ GFP-labeled rounded by the motile T cells described here, were strikingly sim- donor mice and bone marrow from unlabeled wild-type mice. The ilar to granulomas observed in the liver upon infection of mice rationale behind this approach was 2-fold. First, because Rag2Ϫ/Ϫ with the bacterium Mycobacterium bovis (8) and may serve to bone marrow cannot give rise to T or B cells, the majority of contain parasite growth during the chronic phase of infection. GFP-labeled leukocytes in the brains of infected mice should be To examine more closely the interactions of T cells, parasites, CD11bϩ cells, based on our flow cytometric analysis (Fig. 1C and and reporter-expressing cells within the granulomas, we performed data not shown). In addition, wild-type unlabeled bone marrow spectral unmixing of data sets and quantitated T cell motility inside donor cells would give rise to unlabeled T and B cells, providing and outside of the granulomas (Fig. 8, C and D, and supplemental a more normal immune response and making it easier to visualize video 7). OT-1 T cells within granulomas migrated more slowly individual GFP -labeled CD11bϩ cells. Reconstituted mice were and arrested more frequently compared with T cells that were just transferred with naive OT-1 CFP-labeled T cells and infected with outside of the granuloma (Fig. 8D), which is consistent with the OVA-expressing parasites. Brains were examined at relatively hypothesis that Ag recognition by T cells took place within these early times postinfection (10–13 days), times at which interactions structures. Interestingly, T cells within the granuloma that were with individual CD11b cells were prominent based on immuno- closest to the parasites did not show a further decrease in speed or fluorescence analysis of fixed brain sections (Fig. 7 and data not an increase in stopping (Fig. 8D). These data suggest that Ag may shown). Two-photon microscopy of vibratome-cut brain slices The Journal of Immunology 6389 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 8. CD8 T cell dynamics near parasite-containing granulomas. A, Confocal image of a frozen brain section from an infected CD11cYFP reporter mouse 30 days postinfection stained for CD11b. Boxed areas labeled 1 and 2 are shown as enlarged insets of regions containing parasite fluorescence. Arrowheads indicate the positions of parasites. B, Two-photon image of CD11c-YFP reporter-expressing cells (yellow), OT-1 T cells (cyan), and parasites (red) of vibratome-cut brain (day 10 postinfection) A projection of an 80-␮m imaging volume is shown. Boxed area 1 shows an area in which individual parasites are outside of reporter-bright cells. Boxed area 2 shows an area outside of the aggregate containing T cells migrating around an isolated parasite. Middle panels show enlarged views of single optical sections within the boxed areas with arrowheads indicating the isolated parasites (red). Right panel shows representative tracks of OT-1 T cells (blue), CD11c-YFP reporter cells (green), and parasites (red) from the same run. This corresponds to supplemental video 5. C, Two-photon image after spectral unmixing. Sample is similar to B, except the mouse was infected 12 days before analysis. Left panel shows signal from reporter expressing cells (blue), parasites (red), and OT-1 T cells (green) from a projection of a single time frame of the run 60-␮m thick. Right panel shows the tracks of individual OT-1 T cells in white, and the parasite signal is in red. This corresponds to supplemental video 6. D,T cell tracks from the run shown in C were categorized by whether most of the time points were inside or outside of the granuloma (indicated by blue dashed oval in C) and whether they were close to parasites (indicated by the dashed box in C). Plots show the average speed (left) and arrest coefficient (right). .p Ͻ 0.0005, indicating significant differences ,ءءء .Dots represent individual tracks and red horizontal lines indicate the mean revealed numerous GFP-labeled cells and CFP-labeled T cells in imaging (Fig. 9, B and D, and supplemental videos 9 and 10), regions containing isolated parasites as well as occasional exam- whereas in other examples T cells remained in contact with a sin- ples of GFPϩ cells apparently engulfing parasites (Fig. 9A and gle labeled cell during an imaging run (Fig. 9D and supplemental supplemental video 8), which is consistent with a response to re- video 11). Only five of 31 of the contacts observed occurred with cent parasite release by cyst rupture or egress from an infected cell. GFP cells that contained visible parasite fluorescence (Fig. 9C, To examine the duration of T cell contacts with potential APCs hatched bars). Although most contacts were brief (average dura- in regions containing parasites, we tracked individual T cells and tion 7.8 min; Fig. 9C), they coincided with a transient reduction in scored their interactions with GFP-labeled cells over time (Fig. 9, speed (Fig. 9D and supplemental videos 9 and 10). Compiled data B–D, and supplemental videos 9–11). In some cases T cells inter- for multiple T cell contacts over several runs (supplemental Fig. acted with several different GFP-labeled cells during the course of S1C) showed that the average speed of T cells for all time points 6390 IMAGING T. gondii IN THE BRAIN Downloaded from http://www.jimmunol.org/

FIGURE 9. T cell contacts with potential APCs in the brain. Mice were irradiated and reconstituted with Rag2Ϫ/Ϫ GFP bone marrow together with unlabeled bone marrow to label a subset of the potential APCs in the brain. Reconstituted mice were transferred with a small number of CFP-labeled OT-1 T cells and infected with RFP OVA-expressing parasites and analyzed 10–13 days postinfection. A, Example of a GFP-labeled cell engulfing a parasite. Panels show projections of 20-␮m imaging volumes from a time-lapse series obtained by two-photon microscopy. OVA-specific, CFP-labeled OT-1 CD8 T cells are in blue, GFP-labeled APCs are in green, and parasites are in red; left panel corresponds to a time point before engulfment (arrowhead indicates by guest on September 28, 2021 the parasite) and middle and right panels correspond to time points during and just after engulfment (circled). The corresponds to supplemental video 8. B, Examples of T cells contacting GFP-labeled cells. Red, T gondii; green, GFP; blue, OT-1-CFP. Portions of the tracks spent in contact with a GFP-labeled cell are in pink, and portions of the track spent not making visible contact are in yellow. Left panel shows a T cell making two successive contacts with two different GFP-labeled cells. This corresponds to supplemental video 9. Middle panel represents a T cell making two successive contacts with two different GFP labeled cells, one of which contains parasite fluorescence (shown in inset). This corresponds to supplemental video 10. Right panel shows a T cell that remains in contact with the same GFP-labeled cell for the length of the run (18 min). This corresponds to supplemental video 11. C, T cell-APC contact durations. Plot summarizes the number of OT-1 cell contacts with GFP cells for the indicated durations. T cell contacts with GFP-labeled cells containing visible parasites are indicated by the shaded portions of bars. Data are pooled from three runs. D, Decrease in T cell speed correlates with contacts with GFP-labeled cells. Plots of the instantaneous speed as a function of time for individual T cells are shown. Shaded areas correspond to time points in which T cells made a visible contact with a GFP-labeled cell. Arrowheads indicate the time points shown on still images.

in which visible contacts occurred was 3.39 ␮m/min, a value that represent an early response to cyst rupture preceding granuloma was similar to the speed of Ag-specific T cells near isolated par- formation. asites in experiments in which APCs were not visualized (4.43 ␮m/min; supplemental Fig. S1A). To estimate the basal migration rate of T cells in the brain, we considered the speed of T cells Discussion during time points that did not have a visible contact with an APC The control of the parasite T. gondii during chronic infection in the (7.07 ␮m/min), the speed of Ag-nonspecific T cells (7.6 ␮m/min; brain depends on CD8 T cells, but the nature of Ag presentation to supplemental Fig. S1B), and the speed of T cells in areas that did CD8 T cells and how their response protects against parasites in not contain parasite fluorescence (6.35 ␮m/min; Fig. S1A, open the brain is not clear. In this study we have directly visualized T circles). Although each of the values is subject to caveats, alto- cells, parasites, and potential APCs in living brain tissue to address gether these data suggest that the basal migration rate of T cells in these questions. We found that CD8 T cells ignored intact Ag- the brain in the absence of Ag recognition is ϳ7 ␮m/min and that bearing cysts but slowed and transiently arrested near isolated par- frequent transient contacts with Ag-bearing cells lead to average T asites. We provided evidence that the slowing of T cells near par- cell speeds of ϳ5 ␮m/min. Together, these data indicate that T asites was Ag specific and could occur through interactions with cells formed relatively brief contacts with APCs that were in the granuloma-like structures as well as contacts with individual vicinity of parasites, only some of which contained intact parasites. APCs, some of which did not contain parasites. These results pro- These contacts may serve to either stimulate T cell effector re- vide an important new framework for understanding immune re- sponses or allow T cells to activate macrophages or both and may sponses to persistent infections in the brain. The Journal of Immunology 6391

The issue of which host cells can present T. gondii Ags to CD8 aggregates of CD11bϩ cells near isolated parasites, but not intact T cells has been somewhat of a puzzle. CD8 T cells recognize cysts, consistent with previous reports based on light and electron peptide fragments of Ag bound to MHC class I on the surface of microscopy (32, 59). Interestingly, in those earlier studies exam- another cell, and these peptides are typically derived from the cy- ples were noted in which a small numbers of macrophages sur- tosol of a host cell that has been invaded by an intracellular patho- rounded a cyst that was within a host cell with a disrupted mem- gen. Although T. gondii is an intracellular pathogen, it resides brane, suggesting an early macrophage response to cyst rupture within a parasitophorous vacuole that prevents parasite proteins (59). This may correspond to the occasional examples that we from entering the cytosol (65), and thus it is unclear how parasite observed of CD11c FP reporter-expressing cells or GFP-labeled Ags gain access to the class I presentation pathway. In some set- APCs making intimate contact with intact cysts (supplemental vid- tings class I MHC Ag presentation can occur by an alternative eos 1 and 6). Together, these observations suggest that cyst rupture “cross-presentation” pathway in which Ag is produced in one cell may be initially detected by APCs of the brain followed later by a but presented by a different host cell, and it seemed plausible that larger influx of macrophages and T cells, leading eventually to this pathway might operate during T. gondii infection (66). How- granuloma formation. Ag within cysts may contribute to this re- ever, only invaded but not bystander dendritic cells isolated from sponse by being released during cyst rupture and then taken up and spleens of acutely infected mice present parasite-encoded Ags to T cross-presented by APCs within the granuloma. cells, and Ag presentation requires components of the classical Prolonged T cell arrest and cluster formation are often corre- class I presentation pathway, indicating that direct presentation by lated with T cell priming in lymph nodes and have also been as- invaded host cells is important, at least during acute infection (35, sociated with T cell effector responses in tissues (2, 63, 69–71). In 49, 67). Thus, it was surprising that CD8 T cells in the brain made contrast, in the current study we observed no large clusters of CD8 Downloaded from contacts not only with cells containing parasites but also with T cells, and T cells made only transient contacts with APCs. The APCs that were close to parasites but did not themselves contain motility of T cells and the nature of their interactions with APCs parasites. Moreover, many T cell contacts occurred with tightly is influenced by many factors, including the differentiation state of ϩ packed aggregates of CD11b cells and with T cells distributed T cells, the tissue environment, the amount of Ag, the nature of the evenly throughout the aggregate and did not preferentially arrest APC, and the presence of immune-suppressive mechanisms such near the parasites within these structures. These results suggest as regulatory T cells (reviewed in Ref. 3). The lack of T cell stop- http://www.jimmunol.org/ ϩ that CD8 T cells were detecting Ag on the surface of CD11b ping and stable cluster formation in our study may reflect limiting cells that did not contain parasites and suggest that cross-pre- levels of Ag presentation due to immune evasion by the parasites sentation pathways may operate in the brain during chronic in- and/or the presence of immune regulatory mechanisms to limit T fection. Alternatively, APCs may have initially harbored live cell activation in the brain. parasites but eventually degraded the parasites (68), leading to Regarding the functional consequences of these transient inter- the loss of the fluorescent signal. In either case, this distinct mode actions, it is worth noting that while stable contacts often correlate of Ag presentation may reflect the unique ability of the environ- with T cell activation, transient contacts are sufficient to activate T ment of the brain to limit Ag presentation. Moreover, it may be cells in a three-dimensional collagen matrix, and transient contacts by guest on September 28, 2021 more accurate to think of CD8 T cells detecting Ag on aggregates have also been associated with effector responses in vivo (63, 71, of APCs rather than on individual cells. 72). Although we cannot be sure that all of the contacts that we The concept of APCs acting as aggregates of cells rather than observe are a direct result of Ag recognition, several lines of ev- individual cells during their response to T. gondii is also relevant idence suggest that they do reflect functional interactions. These to the issue of immune effector mechanisms. CD8 T cells can include the Ag dependence of T cell motility changes (Fig. 5B), protect either by direct killing of target cells or by the release of evidence that Ag is required for T cell retention in the brain (Fig. cytokines, both of which can be triggered when effector CD8 T 3), evidence that cerebral OT-1 T cells produce the protective cy- cells recognize Ag on a target cell. Although killing requires direct tokine IFN-␥ (Fig. 2C), and indications from our study and others contact between a CD8 T cell and an invaded target cell to provide that CD8 T cells play a protective role in this setting (Figs. 1E and effective protection, this need not be the case for cytokine release. 3B and Refs. 18–22). Transient serial interactions with Ag-bearing Given that the major mode of protection by CD8 T cells during T. cells would likely be sufficient to sustain the activation states of gondii infection is cytokine production rather than direct killing effector CD8 T cells and trigger the local release of cytokines near (22–24), interaction with a group of APCs, only some of which granulomas and foci of infection in the brain. Future studies using contain parasites, would be compatible with the delivery of effec- real-time markers for TCR signaling events and cytokine release, tive protection. Moreover, the physical “walling off” of parasites combined with two-photon imaging, should eventually provide a within the granuloma could provide a means of controlling parasite more detailed and complete picture of T cell effector function in ␥ spread, perhaps in conjunction with IFN- -inducible GTPases the brain and other sites during infection. (26), while avoiding a more potentially damaging effector mech- In summary, the current study provides evidence for a surprising anism such as inducible NO synthase production. The cytokines mode of Ag recognition in the brain in which T cells ignore Ag- produced by CD8 T cells that surround the granulomas could serve bearing cysts but make transient contacts with APCs and granu- to activate macrophages within these structures to make these pro- loma-like structures near foci of infection. This mode of interac- tective mechanisms more effective. tion likely reflects immune evasion by parasites as well as the In contrast to the response to individual parasites, we saw no ability of the unique environment of the brain to limit Ag presen- evidence that CD8 T cells responded to cysts. This is despite the tation and modulate immune responses and is compatible with a fact that cysts express the model Ag OVA and reside within intact mode of protection in which CD8 T cells provide cytokines to host cells that could potentially present Ag to T cells (32). One APCs rather than directly kill target cells. These data provide a possible explanation is that many cysts are found within neurons new perspective on CD8 T cell responses to chronic infection in (Fig. 7D and Ref. 32), cells that express low levels of class I MHC. the brain. In addition, bradyzoites within brain cysts are surrounded by a cyst wall and, thus, very little parasite material may escape into the Note added in proof. After this manuscript was submitted, an- cytoplasm of the cyst-containing host cell. We also observed large other study was published describing T cell migration in the brains 6392 IMAGING T. gondii IN THE BRAIN of mice chronically infected with Toxoplasma gondii (Wilson, 21. Brown, C. R., and R. McLeod. 1990. Class I MHC genes and CD8ϩ T cells E. H., T. H. Harris, P. Mrass, B. John, E. D. Tait, G. F. Wu, M. determine cyst number in Toxoplasma gondii infection. J. Immunol. 145: 3438–3441. Pepper, E. J. Wherry, F. Dzierzinski, D. Roos, et al. 2009. Behav- 22. Suzuki, Y., and J. S. Remington. 1990. The effect of anti-IFN-␥ antibody on the ior of parasite-specific effector CD8ϩ T cells in the brain and vi- protective effect of Lyt-2ϩ immune T cells against toxoplasmosis in mice. J. Im- sualization of a kinesis-associated system of reticular fibers. Im- munol. 144: 1954–1956. 23. Denkers, E. Y., G. Yap, T. Scharton-Kersten, H. Charest, B. A. Butcher, munity 30: 300–311). P. Caspar, S. Heiny, and A. Sher. 1997. Perforin-mediated cytolysis plays a limited role in host resistance to Toxoplasma gondii. J. Immunol. 159: 1903–1908. Acknowlegments 24. Wang, X., H. Kang, T. Kikuchi, and Y. Suzuki. 2004. Gamma interferon pro- We thank Dr. Ena Ladi for help with motility calculations, Shiao Chan for duction, but not perforin-mediated cytolytic activity, of T cells is required for technical assistance, Holly Aaron for help with confocal microscopy, Hec- prevention of toxoplasmic encephalitis in BALB/c mice genetically resistant to the disease. Infect. Immun. 72: 4432–4438. tor Nolla and Alma Valeros of the University of California (UC) Berkeley 25. Robben, P. M., M. LaRegina, W. A. Kuziel, and L. D. Sibley. 2005. Recruitment Cancer Research Laboratory flow cytometry facility and Julie Nelson of of Gr-1ϩ monocytes is essential for control of acute toxoplasmosis. J. Exp. Med. the Center for Tropical and Emerging Global Diseases flow cytometry 201: 1761–1769. facility for help with flow cytometry, Drs. Nicolas Blanchard, Philippe 26. Zhao, Z., B. Fux, M. Goodwin, I. R. Dunay, D. Strong, B. C. Miller, K. Cadwell, Bousso, and Janine Coombes for useful suggestions and proofreading of M. A. Delgado, M. Ponpuak, K. G. Green, et al. 2008. Autophagosome-inde- pendent essential function for the autophagy protein Atg5 in cellular immunity to the manuscript, the Boothroyd Laboratory for providing nonvirulent Prug- intracellular pathogens. Cell Host Microbe 4: 458–469. niaud isolate, the National Cancer Institute Preclinical Repository for pro- 27. Schluter, D., T. Meyer, L. Y. Kwok, M. Montesinos-Rongen, S. Lutjen, viding IL-2, the UC Berkeley Committee for Protection of Human Subjects A. Strack, M. L. Schmitz, and M. Deckert. 2002. Phenotype and regulation of (CPHS) mouse/virus core for providing mouse strains, and the UC Berke- persistent intracerebral T cells in murine Toxoplasma encephalitis. J. Immunol.

169: 315–322. Downloaded from ley CPHS imaging core for access to equipment. 28. Kwok, L. Y., S. Lutjen, S. Soltek, D. Soldati, D. Busch, M. Deckert, and D. Schluter. 2003. The induction and kinetics of antigen-specific CD8 T cells are Disclosures defined by the stage specificity and compartmentalization of the antigen in murine toxoplasmosis. J. Immunol. 170: 1949–1957. The authors have no financial conflict of interest. 29. Lutjen, S., S. Soltek, S. Virna, M. Deckert, and D. Schluter. 2006. Organ- and disease-stage-specific regulation of Toxoplasma gondii-specific CD8-T-cell re- sponses by CD4 T cells. Infect. Immun. 74: 5790–5801. References 30. Soete, M., B. Fortier, D. Camus, and J. F. Dubremetz. 1993. Toxoplasma gondii: 1. Yuste, R. 2005. Fluorescence microscopy today. Nat. Methods. 2: 902–904. kinetics of bradyzoite-tachyzoite interconversion in vitro. Exp. Parasitol. 76: http://www.jimmunol.org/ 2. Cahalan, M. D., and I. Parker. 2008. Choreography of cell motility and interac- 259–264. tion dynamics imaged by two-photon microscopy in lymphoid organs. Annu. Rev. 31. Dzierszinski, F., M. Nishi, L. Ouko, and D. S. Roos. 2004. Dynamics of Toxo- Immunol. 26: 585–626. plasma gondii differentiation. Eukaryot. Cell 3: 992–1003. 3. Celli, S., Z. Garcia, H. Beuneu, and P. Bousso. 2008. Decoding the dynamics of 32. Ferguson, D. J., J. Huskinson-Mark, F. G. Araujo, and J. S. Remington. 1994. A T cell-dendritic cell interactions in vivo. Immunol. Rev. 221: 182–187. morphological study of chronic cerebral toxoplasmosis in mice: comparison of 4. Germain, R. N., M. J. Miller, M. L. Dustin, and M. C. Nussenzweig. 2006. four different strains of Toxoplasma gondii. Parasitol. Res. 80: 493–501. Dynamic imaging of the immune system: progress, pitfalls and promise. Nat. Rev. 33. Shaner, N. C., R. E. Campbell, P. A. Steinbach, B. N. Giepmans, A. E. Palmer, Immunol. 6: 497–507. and R. Y. Tsien. 2004. Improved monomeric red, orange and yellow fluorescent 5. Sumen, C., T. R. Mempel, I. B. Mazo, and U. H. von Andrian. 2004. Intravital proteins derived from Discosoma sp. red fluorescent protein. Nat. Biotechnol. 22: microscopy: visualizing immunity in context. Immunity 21: 315–329. 1567–1572. 6. Peters, N. C., J. G. Egen, N. Secundino, A. Debrabant, N. Kimblin, S. Kamhawi,

34. Striepen, B., and D. Soldati. 2007. Genetic manipulation of Toxoplasma gondii. by guest on September 28, 2021 P. Lawyer, M. P. Fay, R. N. Germain, and D. Sacks. 2008. In vivo imaging In Toxoplasma gondii: The Model Apicomplexan. Perspective and Methods. reveals an essential role for neutrophils in leishmaniasis transmitted by sand flies. L. D. Weiss and K. Kim, eds. Elsevier, pp. 391–415. Science 321: 970–974. 35. Gubbels, M. J., B. Striepen, N. Shastri, M. Turkoz, and E. A. Robey. 2005. Class 7. Chtanova, T., M. Schaeffer, S. J. Han, G. G. van Dooren, M. Nollmann, I major histocompatibility complex presentation of antigens that escape from the P. Herzmark, S. W. Chan, H. Satija, K. Camfield, H. Aaron, et al. 2008. Dy- parasitophorous vacuole of Toxoplasma gondii. Infect. Immun. 73: 703–711. namics of neutrophil migration in lymph nodes during infection. Immunity 29: 36. Messina, M., I. Niesman, C. Mercier, and L. D. Sibley. 1995. Stable DNA trans- 487–496. formation of Toxoplasma gondii using phleomycin selection. Gene 165: 213–217. 8. Egen, J. G., A. G. Rothfuchs, C. G. Feng, N. Winter, A. Sher, and R. N. Germain. 2008. Macrophage and T cell dynamics during the development and disintegra- 37. Rahemtulla, A., W. P. Fung-Leung, M. W. Schilham, T. M. Kundig, tion of mycobacterial granulomas. Immunity 28: 271–284. S. R. Sambhara, A. Narendran, A. Arabian, A. Wakeham, C. J. Paige, R. M. Zinkernagel, et al. 1991. Normal development and function of CD8ϩ cells, 9. Frevert, U., S. Engelmann, S. Zougbede, J. Stange, B. Ng, K. Matuschewski, but markedly decreased helper cell activity in mice lacking CD4. Nature 353: L. Liebes, and H. Yee. 2005. Intravital observation of Plasmodium berghei sporo- 180–183. zoite infection of the liver. PLoS Biol. 3: e192. 10. Plattner, F., and D. Soldati-Favre. 2008. Hijacking of host cellular functions by 38. Fung-Leung, W. P., M. W. Schilham, A. Rahemtulla, T. M. Kundig, the apicomplexa. Annu. Rev. Microbiol. 62: 471–487. M. Vollenweider, J. Potter, W. van Ewijk, and T. W. Mak. 1991. CD8 is needed 11. Denkers, E. Y., and R. T. Gazzinelli. 1998. Regulation and function of T-cell- for development of cytotoxic T cells but not helper T cells. Cell 65: 443–449. 39. Schaefer, B. C., M. L. Schaefer, J. W. Kappler, P. Marrack, and R. M. Kedl. mediated immunity during Toxoplasma gondii infection. Clin. Microbiol. Rev. ϩ 11: 569–588. 2001. Observation of antigen-dependent CD8 T-cell/dendritic cell interactions 12. Lieberman, L. A., and C. A. Hunter. 2002. The role of cytokines and their sig- in vivo. Cell. Immunol. 214: 110–122. naling pathways in the regulation of immunity to Toxoplasma gondii. Int. Rev. 40. Mamalaki, C., J. Elliott, T. Norton, N. Yannoutsos, A. R. Townsend, P. Chandler, Immunol. 21: 373–403. E. Simpson, and D. Kioussis. 1993. Positive and negative selection in transgenic 13. Yap, G. S., and A. Sher. 1999. Cell-mediated immunity to Toxoplasma gondii: mice expressing a T-cell receptor specific for influenza nucleoprotein and endog- initiation, regulation and effector function. Immunobiology 201: 240–247. enous superantigen. Dev. Immunol. 3: 159–174. 14. Schluter, D., J. Lohler, M. Deckert, H. Hof, and G. Schwendemann. 1991. Tox- 41. Hogquist, K. A., S. C. Jameson, W. R. Heath, J. L. Howard, M. J. Bevan, and oplasma encephalitis of immunocompetent and nude mice: immunohistochemical F. R. Carbone. 1994. T cell receptor antagonist peptides induce positive selection. characterisation of Toxoplasma antigen, infiltrates and major histocompatibility Cell 76: 17–27. complex gene products. J. Neuroimmunol. 31: 185–198. 42. Lindquist, R. L., G. Shakhar, D. Dudziak, H. Wardemann, T. Eisenreich, 15. Weiss, L. M., and K. Kim. 2007. Bradyzoite development. In Toxoplasma gondii: M. L. Dustin, and M. C. Nussenzweig. 2004. Visualizing dendritic cell networks The Model Apicomplexan. Academic Press, London, p. 341–361. in vivo. Nat. Immunol. 5: 1243–1250. 16. Luft, B. J., and J. S. Remington. 1992. Toxoplasmic encephalitis in AIDS. Clin. 43. Hadjantonakis, A. K., S. Macmaster, and A. Nagy. 2002. Embryonic stem cells Infect. Dis. 15: 211–222. and mice expressing different GFP variants for multiple non-invasive reporter 17. Galea, I., I. Bechmann, and V. H. Perry. 2007. What is immune privilege (not)? usage within a single animal. BMC Biotechnol. 2: 11. Trends Immunol. 28: 12–18. 44. Reichmann, G., E. N. Villegas, L. Craig, R. Peach, and C. A. Hunter. 1999. The 18. Suzuki, Y., and J. S. Remington. 1988. Dual regulation of resistance against CD28/B7 interaction is not required for resistance to Toxoplasma gondii in the Toxoplasma gondii infection by Lyt-2ϩ and Lyt-1ϩ, L3T4ϩ T cells in mice. brain but contributes to the development of immunopathology. J. Immunol. 163: J. Immunol. 140: 3943–3946. 3354–3362. 19. Parker, S. J., C. W. Roberts, and J. Alexander. 1991. CD8ϩ T cells are the major 45. Witt, C. M., S. Raychaudhuri, B. Schaefer, A. K. Chakraborty, and E. A. Robey. lymphocyte subpopulation involved in the protective immune response to Toxo- 2005. Directed migration of positively selected thymocytes visualized in real plasma gondii in mice. Clin. Exp. Immunol. 84: 207–212. time. PLoS Biol. 3: e160. 20. Gazzinelli, R., Y. Xu, S. Hieny, A. Cheever, and A. Sher. 1992. Simultaneous 46. Purbhoo, M. A., D. J. Irvine, J. B. Huppa, and M. M. Davis. 2004. T cell killing depletion of CD4ϩ and CD8ϩ T lymphocytes is required to reactivate chronic does not require the formation of a stable mature immunological synapse. [Pub- infection with Toxoplasma gondii. J. Immunol. 149: 175–180. lished erratum appears in 2004 Nat. Immunol. 5: 658.] Nat. Immunol. 5: 524–530. The Journal of Immunology 6393

47. Ladi, E., P. Herzmark, and E. A. Robey. 2007. In situ imaging of the mouse 61. Takashima, Y., K. Suzuki, X. Xuan, Y. Nishikawa, A. Unno, and K. Kitoh. 2008. 5hymus using 2-photon microscopy. J. Vis. Exp. 11: 652. Detection of the initial site of Toxoplasma gondii reactivation in brain tissue. Int. 48. Yarovinsky, F., H. Kanzler, S. Hieny, R. L. Coffman, and A. Sher. 2006. Toll-like J. Parasitol. 38: 601–607. ϩ receptor recognition regulates immunodominance in an antimicrobial CD4 T 62. Kawakami, N., U. V. Nagerl, F. Odoardi, T. Bonhoeffer, H. Wekerle, and cell response. Immunity 25: 655–664. A. Flugel. 2005. Live imaging of effector cell trafficking and autoantigen recog- 49. Blanchard, N., F. Gonzalez, M. Schaeffer, N. T. Joncker, T. Cheng, A. J. Shastri, nition within the unfolding autoimmune encephalomyelitis lesion. J. Exp. Med. E. A. Robey, and N. Shastri. 2008. Immunodominant, protective response to the 201: 1805–1814. parasite Toxoplasma gondii requires antigen processing in the endoplasmic re- 63. Boissonnas, A., L. Fetler, I. S. Zeelenberg, S. Hugues, and S. Amigorena. 2007. ticulum. Nat. Immunol. 9: 937–944. In vivo imaging of cytotoxic T cell infiltration and elimination of a solid tumor. 50. Suzuki, Y., K. Joh, M. A. Orellana, F. K. Conley, and J. S. Remington. 1991. A J. Exp. Med. 204: 345–356. gene(s) within the H-2D region determines the development of toxoplasmic en- 64. Courret, N., S. Darche, P. Sonigo, G. Milon, D. Buzoni-Gatel, and I. Tardieux. cephalitis in mice. Immunology 74: 732–739. 2006. CD11c- and CD11b-expressing mouse leukocytes transport single Toxo- 51. Brown, C. R., C. A. Hunter, R. G. Estes, E. Beckmann, J. Forman, C. David, plasma gondii tachyzoites to the brain. Blood 107: 309–316. J. S. Remington, and R. McLeod. 1995. Definitive identification of a gene that 65. Schwab, J. C., C. J. Beckers, and K. A. Joiner. 1994. The parasitophorous vacuole confers resistance against Toxoplasma cyst burden and encephalitis. Immunology membrane surrounding intracellular Toxoplasma gondii functions as a molecular 85: 419–428. sieve. Proc. Natl. Acad. Sci. USA 91: 509–513. 52. Schluter, D., D. Bertsch, K. Frei, S. B. Hubers, O. D. Wiestler, H. Hof, 66. Luder, C. G., and F. Seeber. 2001. Toxoplasma gondii and MHC-restricted an- A. Fontana, and M. Deckert-Schluter. 1998. Interferon-␥ antagonizes transform- tigen presentation: on degradation, transport and modulation. Int. J. Parasitol. 31: ing growth factor-␤2-mediated immunosuppression in murine Toxoplasma en- 1355–1369. cephalitis. J. Neuroimmunol. 81: 38–48. 67. Dzierszinski, F., M. Pepper, J. S. Stumhofer, D. F. LaRosa, E. H. Wilson, 53. Sedgwick, J. D., S. Schwender, H. Imrich, R. Dorries, G. W. Butcher, and L. A. Turka, S. K. Halonen, C. A. Hunter, and D. S. Roos. 2007. Presentation of V. ter Meulen. 1991. Isolation and direct characterization of resident microglial Toxoplasma gondii antigens via the endogenous major histocompatibility com- cells from the normal and inflamed central nervous system. Proc. Natl. Acad. Sci. plex class I pathway in nonprofessional and professional antigen-presenting cells. USA 88: 7438–7442. Infect. Immun. 75: 5200–5209. 54. Fischer, H. G., and G. Reichmann. 2001. Brain dendritic cells and macrophages/ microglia in central nervous system inflammation. J. Immunol. 166: 2717–2726. 68. Andrade, R. M., M. Wessendarp, M. J. Gubbels, B. Striepen, and C. S. Subauste. Downloaded from 55. Urbani, S., B. Amadei, D. Tola, M. Massari, S. Schivazappa, G. Missale, and 2006. CD40 induces macrophage anti-Toxoplasma gondii activity by triggering C. Ferrari. 2006. PD-1 expression in acute hepatitis C virus (HCV) infection is autophagy-dependent fusion of pathogen-containing vacuoles and lysosomes. associated with HCV-specific CD8 exhaustion. J. Virol. 80: 11398–11403. J. Clin. Invest. 116: 2366–2377. 69. Breart, B., F. Lemaitre, S. Celli, and P. Bousso. 2008. Two-photon imaging of 56. Gahwiler, B. H., M. Capogna, D. Debanne, R. A. McKinney, and ϩ S. M. Thompson. 1997. Organotypic slice cultures: a technique has come of age. intratumoral CD8 T cell cytotoxic activity during adoptive T cell therapy in Trends Neurosci. 20: 471–477. mice. J. Clin. Invest. 118: 1390–1397. 57. Davalos, D., J. Grutzendler, G. Yang, J. V. Kim, Y. Zuo, S. Jung, D. R. Littman, 70. Mrass, P., H. Takano, L. G. Ng, S. Daxini, M. O. Lasaro, A. Iparraguirre, L. L. Cavanagh, U. H. von Andrian, H. C. Ertl, P. G. Haydon, and W. Weninger. M. L. Dustin, and W. B. Gan. 2005. ATP mediates rapid microglial response to http://www.jimmunol.org/ local brain injury in vivo. Nat. Neurosci. 8: 752–758. 2006. Random migration precedes stable target cell interactions of tumor-infil- 58. Nimmerjahn, A., F. Kirchhoff, and F. Helmchen. 2005. Resting microglial cells trating T cells. J. Exp. Med. 203: 2749–2761. are highly dynamic surveillants of brain parenchyma in vivo. Science 308: 71. Matheu, M. P., C. Beeton, A. Garcia, V. Chi, S. Rangaraju, O. Safrina, 1314–1318. K. Monaghan, M. I. Uemura, D. Li, S. Pal, et al. 2008. Imaging of effector 59. Ferguson, D. J., W. M. Hutchison, and E. Pettersen. 1989. Tissue cyst rupture in memory T dells during a delayed-type hypersensitivity reaction and suppression mice chronically infected with Toxoplasma gondii. An immunocytochemical and by Kv1.3 channel block. Immunity 29: 602–614. ultrastructural study. Parasitol. Res. 75: 599–603. 72. Gunzer, M., A. Schafer, S. Borgmann, S. Grabbe, K. S. Zanker, E. B. Brocker, 60. Odaert, H., M. Soete, B. Fortier, D. Camus, and J. F. Dubremetz. 1996. Stage E. Kampgen, and P. Friedl. 2000. Antigen presentation in extracellular matrix: conversion of Toxoplasma gondii in mouse brain during infection and immu- interactions of T cells with dendritic cells are dynamic, short lived, and sequen- nodepression. Parasitol. Res. 82: 28–31. tial. Immunity 13: 323–332. by guest on September 28, 2021