sign in sign up
<<
Home , Lymphocyte, Myeloid tissue, T cell

Human δγ T Cells: A Lymphoid Lineage Cell Capable of Professional Yin Wu, Wutian Wu, Wai Man Wong, Eliot Ward, Adrian J. Thrasher, David Goldblatt, Mohamed Osman, Paul Digard, This information is current as David H. Canaday and Kenth Gustafsson of September 28, 2021. J Immunol 2009; 183:5622-5629; ; doi: 10.4049/jimmunol.0901772 http://www.jimmunol.org/content/183/9/5622 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2009/10/19/183.9.5622.DC1 Material

References This article cites 43 articles, 14 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/183/9/5622.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 28, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of 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

Human ␥␦ T Cells: A Lymphoid Lineage Cell Capable of Professional Phagocytosis1

Yin Wu,2* Wutian Wu,2†‡§ Wai Man Wong,†‡§ Eliot Ward,* Adrian J. Thrasher,* David Goldblatt,¶ Mohamed Osman,* Paul Digard,ʈ David H. Canaday,# and Kenth Gustafsson3*

Professional phagocytosis in mammals is considered to be performed exclusively by myeloid cell types. In this study, we demon- strate, for the first time, that a mammalian subset can operate as a professional . By using confocal microscopy, transmission electron microscopy, and functional Ag presentation assays, we find that freshly isolated human pe- ripheral ␥␦ T cells can phagocytose Escherichia coli and 1 ␮m synthetic beads via Ab opsonization and CD16 (Fc␥RIII), ,leading to Ag processing and presentation on MHC class II. In contrast, other CD16؉ , i.e., CD16؉/CD56؉ NK cells were not capable of such functions. These findings of distinct myeloid characteristics in ␥␦ T cells strongly support the suggestion that ␥␦ T cells are evolutionarily ancient lymphocytes and have implications for our understanding of their role in transitional Downloaded from and the control of infectious diseases and . The Journal of Immunology, 2009, 183: 5622–5629.

nnate immune cells express hard programmed receptors that they constitute only 0.5–10% of total T cells, ␥␦ T cells expand sense molecular structures on microbes and altered self-cells rapidly upon direct encounter with ligands common to I which enable the uptake of these agents as a broad first line and cell markers without the need for conventional defense. Adaptive immune cells evolve clonally within an individ- activation mechanisms. http://www.jimmunol.org/ ual by somatic rearrangement of specific receptors providing a ␥␦ T cells from several mammalian species can surprisingly delayed but tailored response. Cross-talk between the two is cen- present Ags to CD4ϩ ␣␤ T cells on MHC class II (5–8). Brandes tral to an appropriate (1, 2). Myeloid cells such et al. have shown that activated human ␥9␦2 T cells express MHC as , , , and myeloid dendritic class II and can professionally present protein Ags to both naive cells (DCs)4 clearly display innate characteristics, while lymphoid CD4ϩ and CD8ϩ ␣␤ T cells while ␣␤ T cell controls could not (5, lineage B and ␣␤ T cells represent the classical adaptive response. 9). In addition, it seems that activated ␥␦ T cells up-regulate CCR7 ␥␦ T cells, however, display characteristics of both. ␥␦ T cells, and home to local nodes where they interact with ␣␤ T cells while sharing ␣␤ T cell functions, also perform immune surveil- and other APCs (10, 11). The activation profile and behavior of ␥␦ lance of an innate character and are the only major set of - T cells bears an uncanny resemblance to that of the innate myeloid by guest on September 28, 2021 resident T cells (3). Compared with ␣␤ T cells, the ␥␦ TCR rep- lineage DC. However, it remains unknown how ␥␦ T cells acquire ertoire is very restricted and matched to the respective tissue Ag (12). type these cells reside in (4). This restricted repertoire is di- Professional Ag presentation is a hallmark of innate immunity rected against non-MHC-restricted Ags common to pathogens and is often associated with professional phagocytosis. Profes- and stressed self-cells in a fashion reminiscent of pattern recogni- sional phagocytosis, as defined by phagocytosis by myeloid lin- tion receptors on innate myeloid cells. This is despite the fact that eage monocytes, DCs, and neutrophils, is a tissue resident activity, their potential TCR diversity actually exceeds ␣␤ TCRs. Although which allows for sampling of the local environment as part of tissue homeostasis and in readiness for pathogens (13). It is an evolutionary development of unicellular phagocytosis, which in *Molecular Immunology Unit, University College London Institute of Child Health, † ‡ jawed vertebrates was the starting point for the processing and London, U.K. Department of , State Key Laboratory of Brain and Cog- ␥␦ nitive Sciences, §Research Center of Reproduction, Development and Growth, Li Ka presentation of Ags to the adaptive . Given that Shing Faculty of Medicine, University of Hong Kong, Pokfulam, Hong Kong Special T cells are hypothesized to predate the development of other lym- Administrative Region, China; ¶Immunobiology Unit, UCL Institute of Child Health, ʈ phoid lineage B and ␣␤ T cells (14), that they express the phago- London, U.K.; Division of , Department of Pathology, University of Cam- bridge, Tennis Court Road, Cambridge, U.K.; #Division of Infectious Disease, Louis cytic Fc for IgG, CD16 (Fc␥RIII) (15, 16), and their ca- Stokes Cleveland Veterans Affairs Medical Center, Case Western Reserve University, pacity as professional APCs (5, 9), we investigated their capacity Cleveland, Ohio 44106 for phagocytosis. We find in this study that human peripheral Received for publication June 3, 2009. Accepted for publication August 24, 2009. blood ␥␦ T cells can phagocytose both 1-␮m synthetic beads as The costs of publication of this article were defrayed in part by the payment of page well as Escherichia coli via CD16-mediated uptake, and moreover charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. that this is functionally linked to Ag processing and presentation 1 This work was supported by the Jean Shanks Foundation. on MHC class II. 2 These authors contributed equally to this manuscript. 3 Address correspondence and reprint requests to Dr. Kenth Gustafsson, Molecular Materials and Methods Immunology Unit, UCL Institute of Child Health, 30 Guilford Street, London, U.K. Isolation of ␥␦ T cells E-mail address: [email protected] 4 Abbreviations used in this paper: DC, ; LCL, lymphoblastic cell line; Peripheral venous blood from healthy donors was heparinized with 3 IU TEM, transmission electron microscopy; M1, influenza matrix 1 protein; IPP, iso- heparin/ml blood, diluted 1/1 with sterile PBS, and layered over 1/3 vol- pentenyl pyrophosphate; TMR, tetramethylrhodamine; mIL-2, murine IL-2. ume of Ficoll-Paque. Samples were centrifuged for 2200 rpm (Sorvall Leg- end Mach 1.6R) for 30 min at room temperature with no brake. A small Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 volume of plasma was removed and saved for later to block Fc receptors www.jimmunol.org/cgi/doi/10.4049/jimmunol.0901772 The Journal of Immunology 5623 during staining. The PBMC layer was removed and washed in 4°C PBS as A1C5 hybridomas respond only to DRB1*0101 subtype MHC class II before magnetic selection. Fresh ␥␦ T cells and NK cells were isolated presenting the M1 (supplementary Fig. 2 and Fig. 4Cii). from PBMCs using the anti-TCR␥␦ and CD56ϩ/CD16ϩ MicroBead Kits (Miltenyi Biotec), respectively, to a purity of 95% (supplemental Fig. S1).5 Conjugation of M1 to polystyrene beads ␥␦ All steps were performed on ice to minimize nonspecific staining. T cell ␮ preparations were activated by coculture with HLA-DRB1*0701 lympho- Recombinant his-tagged M1 was conjugated to 1 m polystyrene beads blastic cell lines (LCLs), 100 IU/ml human IL-2, and 50 ␮m isopentenyl (Dynabeads TALON 101.01D) according to the manufacturer’s instruc- pyrophosphate, as previously described (5). tions. Beads were washed extensively to remove all unconjugated M1 pro- tein from preparations (Fig. 4B) such that any MHC class II presentation of Incubation with tetramethylrhodamine E. coli and fluorescent M1 were exclusively derived from phagocytosed M1 beads. red latex beads and staining for immunofluorescence Presentation of M1-conjugated beads to A1C5 hybridomas Four ϫ 104 freshly isolated peripheral blood ␥␦ T cells were incubated In brief, 105 ␥␦ T cells were seeded in 100 ␮l of serum-free DMEM in with 20 ␮l of rabbit IgG opsonised tetramethylrhodamine E. coli (Molec- triplicate wells of 96-well round-bottom tissue culture plates. Negative ular Probes BioParticles E. coli E-2862 and Opsonizing Reagent E-2870, control wells with ␥␦ T cells were seeded with 100 ␮l of serum-free reconstituted according to the manufacturer’s instructions) in 100 ␮lof DMEM. Wells then received PBS, nonopsonised M1 beads, or anti-M1 serum-free DMEM in 1 cm2 square wells on polylysine coated slides for 45 rabbit antiserum (19) opsonized M1 beads and were incubated at 37°C/5% min at 37°C. As a control for active phagocytosis, coverslips were incu- CO for 1 h. After 1 h, 105 A1C5 hybridomas (1:1 ratio) were added in 100 bated on ice and no phagocytosis was observed (data not shown). One 2 ␮l to each well. For activation, 50 ␮m IPP, 100 IU/ml recombinant human micrometer fluorescent red latex beads (Sigma-Aldrich, L-2778) were pas- IL-2, and 105 DRB1*0701 LCLs were added to ␥␦ T cell wells and neg- sively coated with BSA and opsonized with rabbit anti-BSA IgG. Opso- ative control wells. Plates were then incubated at 37°C/5% CO for 24–36 nized beads were incubated with ␥␦ T cells as described above. After 45 2 h. Murine IL-2 secretion by A1C5 hybridomas was assessed by ELISA for min, slides were fixed with 4% formaldehyde, permeablized with 0.5% Downloaded from murine IL-2 (R&D Systems Mouse DuoSet DY402). Triton, and blocked with 5% autologous plasma before staining. Slides ␥␦ were stained with mouse anti-human TCR (BD Bioscience 555716) and Inhibiting active phagocytosis and Ag processing with goat anti-mouse Cy5 (Caltag M35011), AlexaFluor 488 phalloidin (In- vitrogen A12379), and DAPI (Vectashield H-1500) before imaging on a cytochalasin D and laser scanning microscope system (TCS NT) (Leica). To inhibit Ag processing, ␥␦ T cells were incubated with 50 ␮g/ml chlo- roquine, an inhibitor of acidification, for one hour before puls- Transmission electron microscopy (TEM) ␥␦ ␮

ing. To block active phagocytosis, T cells were incubated with 10 g/ml http://www.jimmunol.org/ After incubation with E. coli, ␥␦ T cells were washed in PBS and fixed in cytochalasin D diluted in DMSO for one hour before pulsing. ␥␦ T cells 4% formaldehyde and 2.5% gluteraldehyde. Preparation for TEM was done were also pulsed in the equivalent 1/1000 dilution of DMSO to control for following procedures described previously (17). In brief, after fixation ␥␦ any effect the drug solvent might have on the assay. T cells were pelleted, transferred to glass slides, and allowed to dry on a hotplate at 37°C before embedding in 2% agarose solution. Agarose em- CD16 and MHC class II blocking bedded blocks were then fixed in EM fixative (2.5% glutaraldehyde and 2% ␥␦ T cells were blocked with a mouse monoclonal IgG1 anti-CD16 block- formaldehyde in 0.1 M phosphate buffer solution (pH 7.4)) at 4°C for 4 h. ing Ab ( LNK16, Abcam) or nonspecific mouse IgG1 at 10 ␮g/ml for After 4 h, the blocks were washed three times with 0.1 M phosphate buffer 30 min at 37°C/5% CO before pulsing with M1 beads. Likewise, ␥␦ T solution before post fixing with 1% osmium tetroxide solution in wash 2 cells were blocked with an azide-free mouse monoclonal IgG2a anti- buffer at 4°C overnight. Blocks were then dehydrated in graded ethanol,

HLA-DR (clone L243, BD Biosciences) or nonspecific mouse IgG2a (BD by guest on September 28, 2021 embedded in pure Epon, and kept at 60°C for 3 days. Ninety nanometer Biosciences) for 30 min before addition of responder T cells. ultrathin sections were stained with 8% uranyl acetate and lead citrate before observation under electron microscope. Controlling for PBMC contamination Generation of A1C5 hybridomas In brief, 105 cells from the ␥␦ T cell preparation were used for presentation Ͻ All animal experiments were done in accordance with the guidelines laid of which 5000 (5%) were monocytes. To control for the possible con- ϫ 4 out by the Institutional Animal Care and Use Committee at Case Western tribution of monocytes and the very few peripheral blood DCs, 5 10 Reserve University. Generation of the T cell hybridoma, A1C5 was similar donor PBMCs were used as a source of at least 5000 monocytes (donor to previously published methods (18). In brief, mice transgenic for HLA- content ranged from 13–20%, data not shown) and were incu- DRB1*0101 (Dennis Zaller, Merck Laboratories, New York, NY) were bated with the same Ags and under the same activating conditions in par- ␥␦ injected in the footpad with recombinant influenza matrix (M1) Ag pre- allel with the experiments described above. We also sorted T cells from ␥ pared in CFA according to the manufacturer’s instructions. Poplietal lymph PBMCs on TCR- 9 expression by flow cytometry reducing monocyte con- Ͻ ϫ 4 nodes were isolated 7 days later, restimulated in vitro for 5 days, and the tamination to 0.2% (Fig. 5Bi). Due to lower yields, we used 2.5 10 ␥␦ T cells subsequently fused with the TCR -negative BW1100 thymoma cell highly pure flow cytometry sorted T cells as Ag presenters (Fig. 5Bii). line. Fused cells were selected using medium containing hypoxanthine, aminopterin, and thymidine for 7–9 days and screened for Ag specificity Results with HLA-DRB1*0101 plus EBV-transformed B cells. HLA restriction Confocal microscopic imaging of phagocytosis by ␥␦ T cells was confirmed by the inability of the T cell hybridoma to recognize M1 Ag on non-HLA-DRB1*0101 APCs (supplementary Fig. 2 and Fig. 4Cii), and To assess their phagocytic capacity, ␥␦ T cells were directly iso- by blocking using a mAb against HLA-DR. The influenza M1 epitope was lated from fresh blood by positive anti-TCR␥␦ selection on mag- mapped using overlapping from M1 (BEI Resources) and found to netic microbeads to a purity of 95% (supplemental Fig. S1). Cells be aa127–143 CMGLIYNRMGAVTTESA. were then pulsed with IgG opsonized tetramethylrhodamine Activation of ␥␦ T cells (TMR)-labeled E. coli or IgG opsonized 1 ␮m fluorescent red latex ␥␦ T cells were activated by coculture with LCLs in the presence of 100 beads and stained for immunofluorescence. IU/ml recombinant human IL-2 and 50 ␮m isopentenyl pyrophosphate ␥␦ T cells were identified by positive (Fig. 1A) and often punc- (IPP), essentially as described in previous work (5). Abs used to detect tate (Fig. 1B) anti-TCR␥␦ staining both on the surface and inside CD16, MHC class II, and CD86 were: a monoclonal mouse IgG2 PE- the cell. This punctate staining was consistent with TCR-CD3 clus- conjugated anti-HLA-DR, clone L243 (G46-6) (BD Biosciences); a tering on activated ␣␤ T cells (20) and reports of TCR␥␦ cluster- monoclonal mouse IgG1 PE-conjugated anti-CD86, clone FUN-1 (BD Bio- ␥␦ sciences); and a monoclonal mouse IgG1 tricolor CE-conjugated anti- ing on T cells after TCR ligation (21), and was likely due to CD16, clone 3G8 (Caltag Medsystems). We used allogeneic DRB1*0701 some activation during selection. ␥␦ T cells were also delineated (i.e., non-DRB1*0101) LCLs to activate ␥␦ T cells to avoid a further pu- from the few contaminating monocytes by their distinct morphol- rification and loss of ␥␦ T cells from activation coculture for presentation ogy. ␥␦ T cells were Ͻ10 ␮m across, with round dense nuclei surrounded by a thin round ring of cytoplasm. Monocytes were 5 The online version of this article contains supplementary material. Ͼ10 ␮m, with distinct horseshoe-shaped nuclei surrounded by 5624 HUMAN ␥␦ T CELLS ARE CAPABLE OF PHAGOCYTOSIS

FIGURE 1. Confocal microscopy of E. coli and 1 ␮m latex bead phagocytosis by ␥␦ T cells. Cells incubated with TMR E. coli (red) or 1 ␮m fluorescent beads (red) and stained with anti-TCR␥␦ (magenta), DAPI Downloaded from (blue), and phalloidin (green). Scale bars, 5 ␮m. A, ␥␦ T cell: small round cell body (Ͻ10 ␮m), smooth thin ring of cytoplasm surrounding a relatively large round nucleus and specific anti-TCR␥␦ staining. B, ␥␦ T cell with punctate staining. C, Typical mono- http://www.jimmunol.org/ cyte: large cell (Ͼ10 ␮m), distinct horseshoe nucleus, low nucleus to cytoplasm ratio, and marked lack of specific anti-TCR ␥␦ stain- ing. D, “Hairy” ␥␦ T cell with dendritic pro- cesses. E, Few adsorbed onto the surface of smooth ␥␦ T cells. F, Hairy ␥␦ T cells displayed increased bacterial adsorp- tion. G, Phagocytosis of E. coli by ␥␦ T cell: several bacteria clearly inside cells are stain- ing positive for ␥␦ TCR and have morphol- by guest on September 28, 2021 ogy consistent with ␥␦ T cells. H, Phagocy- tosis of 1 ␮m latex beads by ␥␦ T cell.

large irregular cytoplasms giving a high cytoplasm-to-nucleus ra- examples of such cells are shown in supplemental Fig. S3 and tio (Fig. 1C). The majority of ␥␦ T cells observed displayed typical S4, of which the stack in Fig. S3 corresponds to the cell dis- lymphocyte morphology but up to 10% had distinct plasma mem- played in Fig. 1G. Given recent reports of teleost fish and Xe- brane dendritic processes resembling pseudopods formed by nopus dendritic-like B cells capable of phagocytosis (22), we phagocytic cells (Fig. 1D). Otherwise, these cells were of the same speculate that the dendritic-like population of ␥␦ T cells were size and had the same nuclear morphology and TCR␥␦ staining as responsible. These results were repeated with cells from mul- other ␥␦ T cells. When pulsed with IgG-opsonized TMR E. coli, tiple donors and also by studying phagocytosis with IgG-opso- these “hairy” ␥␦ T cells had more E. coli adsorbed onto their nized 1-␮m latex beads (Fig. 1H). surface than smooth ␥␦ T cells (Fig. 1, E and F). We found up to ␥␦ 10% of ␥␦ T cells with internalized E. coli (Fig. 1G). To confirm Confirmation of phagocytosis by T cells by TEM that the E. coli were indeed inside the cells, “stacks” of a minimum To confirm these cells as ␥␦ T cells, we performed TEM. Consis- of nine images were acquired from different levels of the cell. Two tent with confocal findings, the cells were Ͻ10 ␮m (even smaller The Journal of Immunology 5625 Downloaded from

FIGURE 2. TEM showing detailed morphology and anti-TCR␥␦ stain- ing. A, TEM section of a typical ␥␦ T cell characterized by a small cell body, a large round nucleus surrounded by a thin ring of cytoplasm with http://www.jimmunol.org/ few organelles and presence of anti-TCR␥␦ microbeads both on the surface and inside endosomes (black arrowheads). B, A contaminating monocyte characterized by its large size, horseshoe shaped nucleus, large irregular FIGURE 3. TEM: phagocytosis of E. coli by ␥␦ T cells. Ai, E. coli cytoplasm, and lack of anti-TCR␥␦ microbeads. Several bacteria could be (arrows) adsorption to a hairy ␥␦ T cell with pseudopod-like dendritic seen inside the monocyte (arrows). C, A contaminating charac- processes. Aii, Enlargement of square area in (Ai) showing tight adher- terized by its multilobed nucleus, abundance of electron dense specific ence of E. coli to the plasma membrane and presence of anti-TCR␥␦ granules (white arrowheads) and lack of anti-TCR␥␦ microbeads. Several microbeads (arrowheads). Bi,A␥␦ T cell phagocytosing E. coli with bacteria could be seen inside the neutrophil. D, TEM of E. coli: 0.5–1.0 ␮m bacteria clearly visible both inside and on the surface in phagocytic in diameter with electron dense nucleic acid core surrounded by an arti- ruffles. Several phagocytic vacuoles can be seen. Bii, Enlargement of by guest on September 28, 2021 factual vacuole and thin electron dense cell wall. square area in (Bi) showing bacteria and anti-TCR␥␦ microbeads in detail. Ci,A␥␦ T cell with a swollen cytoplasm full of phagocytic vacuoles. Cii, Enlargement of square area in (Ci) showing bacteria and anti-TCR␥␦ beads in detail. Some large and empty vacuoles are sug- from TEM preparation associated shrinkage) with typical lympho- gestive of phagosomal degradation. cyte morphology. Also seen were punctate collections of highly electron dense bodies up to 25 nm in diameter both on the cell surface and in endocytic compartments (Fig. 2A) representing the ␥␦ magnetic microbeads used to select ␥␦ T cells which together with Phagocytosis by T cells is functionally linked to, and can be morphology distinguish ␥␦ T cells from other cell types. The ␥␦ T assayed by, Ag presentation on MHC class II cells observed in this study and the pattern of ␥␦ TCR staining, Although it is intriguing that human ␥␦ T cells can phagocytose, including clustering in endosomes, are consistent with previous do they, like neutrophils, simply ingest and remove offensive ma- electron micrograph studies of these cells (23). Few contaminating terial? In light of their ability to up-regulate MHC class II and monocytes and only one neutrophil were seen with both displaying present Ags to ␣␤ T cells (5), we hypothesized that phagocytosis markedly different morphology (Fig. 2, B and C, respectively) and and Ag presentation are functionally linked in ␥␦ T cells. To test lack of electron dense anti-TCR␥␦ microbeads. Bacteria were this hypothesis and as an additional assay of phagocytosis, we identified by their size (0.5–1 ␮m) and distinct morphology (Fig. conjugated whole influenza A matrix (M1) protein onto 1 ␮m 2D). Again, a small proportion of ␥␦ T cells had formed pseudo- polystyrene beads (M1 beads) and incubated these with freshly pod-like protrusions around adsorbed bacteria (Fig. 3A). Bacteria isolated nonactivated ␥␦ T cells. Surface MHC class II loading of could be seen inside cells with distinct lymphocyte morphology M1 was detected by A1C5 murine T cell hybridomas and positive anti-TCR␥␦ microbead staining at a frequency of which secrete murine IL-2 (mIL-2) in recognition of processed M1 5–10% between sections (Fig. 3, B and C). Although these ␥␦ T peptide (18). When pulsed with anti-M1 rabbit antiserum opso- cells with phagocytosed E. coli had a somewhat increased cyto- nized M1 beads, the human peripheral blood ␥␦ T cells could plasmic volume reminiscent of monocytes, this was most likely clearly process and present M1-derived peptides on surface MHC due to increased swelling after ingestion. Moreover, these cells class II as detected by the mIL-2 ELISA (Fig. 4A). An equal num- retained lymphocyte nuclear morphology and had microbeads ber of unpulsed ␥␦ T cells, as well as similarly purified freshly present at the cell surface as well as intracellularly, whereas isolated human CD16ϩ/CD56ϩ NK cells, were used as negative none of the monocytes or the neutrophil carried any microbeads de- controls. Pulsed NK cells did not result in a significantly different spite extensive uptake of bacteria (Fig. 2, A–C). We speculate that the presentation of M1 peptide to A1C5 hybridomas compared with aforementioned dendritic population of ␥␦ T cells are responsible for unpulsed NK cells. To demonstrate active phagocytic uptake, ␥␦ T phagocytosis. cells were incubated with cytochalasinD1hbefore Ag pulsing. To 5626 HUMAN ␥␦ T CELLS ARE CAPABLE OF PHAGOCYTOSIS Downloaded from FIGURE 5. Ag presentation of phagocytosed M1 beads is directly attributable to ␥␦ T cells and is enhanced by Ab opsonization and ac- FIGURE 4. Phagocytosis and Ag presentation in ␥␦ T cells are func- tivation of ␥␦ T cells. Ai and Aii, ␥␦ T cells (dark gray bars) were pulsed tionally linked, a functional assay of phagocytosis. A, Unpulsed CD16ϩ with naked M1 beads or anti-M1 opsonized beads (“IgϩM1 beads”) NK T cells (white) and ␥␦ T cells (dark gray) were very poor at activating with or without activation and presentation was measured by A1C5 T A1C5 hybridomas. CD16ϩ NK T cells pulsed with anti-M1 opsonized M1 cell hybridoma mIL-2 production. As negative controls (white bars), the beads did not display significantly better presentation (p Ͼ 0.05) compared same preparations devoid of ␥␦ T cells were used. To control for mono- http://www.jimmunol.org/ with unpulsed. Pulsed ␥␦ T cells however, induced vigorous activation of cyte contamination, parallel negative controls were spiked with 5 ϫ 104 M1 peptide specific A1C5 hybridomas which was inhibited by chloroquine PBMCs containing an excess number of monocytes (light gray bars). Activation of ␥␦ T cells. Bi, ␥␦ T cells were sorted on TCR␥9by ,ءء Chq) and cytochalasin D (CyD). DMSO drug solvent control had no effect) on A1C5 activation by pulsed ␥␦ T cells. B, ␥␦ T cells incubated with flow cytometry reducing CD14ϩ monocyte contamination to Ͻ0.2%. supernatants from M1 bead preparations were unable to activate A1C5 Bii, Sorted cells were pulsed with M1 beads, activated and cocultured hybridomas so any Ag presentation must have been derived from phago- with A1C5 hybridomas (gray bar). The same activation coculture with cytosis of beads. The mean Ϯ SD of triplicate Ag presentations is plotted. M1 beads plus A1C5 hybridomas but devoid of ␥␦ T cells (white bar) Ci, A blocking mAb (L243) against HLA-DR is able to completely block was used as a negative control. In the virtual absence of monocyte any detectable presentation to the HLA-DR1*0101-restricted M1 Ag-spe- contamination, ␥␦ T cells still retained Ag presenting capacity. The cific A1C5 hybridoma, whereas an control is not. Cii, HLA- mean Ϯ SD of triplicate Ag presentations is plotted. by guest on September 28, 2021 DRB*0102 and DRB*0101 ␥␦ T cells were pulsed with whole heat-inac- tivated influenza A/PR/8/34 virus and used to present to A1C5 hybridomas. M1 peptide was only presented efficiently by DRB1*0101 ␥␦ T cells (gray shade) and not by DRB1*0102 ␥␦ T cells (white). The mean Ϯ SD of ␥␦ T cells are unable to support presentation of this epitope to triplicate Ag presentations is plotted. this hybridoma (Fig. 4Cii), as are HLA-DRB1*0701 LCL cells (supplementary Fig. 2). In these experiments, whole heat-inac- tivated nonopsonized influenza A/PR/8/34 virus was used as a show that the Ag presentation was dependent on active Ag pro- source of M1 Ag for the presentation. cessing in the ␥␦ T cells, chloroquine was added to inhibit acidi- ␥␦ fication of the Ag processing and loading compartments. ␥␦ T cells Monocyte contamination in T cell preparations does not were also pulsed in the equivalent dilution of DMSO to control for account for Ag presentation of phagocytosed M1 beads any effect that the drug solvent might have on the assay. The ad- Previous studies of MHC class II Ag presentation by human ␥␦ dition of cytochalasin D and chloroquine reduced presentation of T cells have not directly addressed the role of contaminating M1 peptide to background levels (Fig. 4A), demonstrating that the cell types such as monocytes in cell preparations (5). Neither M1-conjugated bead uptake by ␥␦ T cells was dependent on active have human ␥9␦2 T cells been directly shown to be responsible actin reorganization compatible with phagocytosis, and on pH-de- for the Ag presentation. To account for contaminating PBMCs, pendent degradation and processing for MHC class II loading. To especially monocytes, parallel experiments were performed us- confirm that the only source of M1 protein available was in the ing PBMCs containing a number of monocytes in excess to the form conjugated to M1 beads and hence required phagocytic up- potential contaminating population (5 ϫ 103), but devoid of ␥␦ take, ␥␦ T cells were also incubated with centrifuged supernatant T cells and these confirmed that ␥␦ T cells were responsible for from M1 beads. In the absence of M1 beads, ␥␦ T cells incubated the majority of the presentation seen (Fig. 5, Ai and Aii). Fur- in these supernatants were unable to activate the A1C5 hybridoma thermore, ␥␦ T cells were sorted on TCR-␥9 expression by flow (Fig. 4B). To conclusively show that the presentation of the M1 cytometry and thereby reducing monocyte contamination to peptide was indeed mediated by MHC class II, and specifically Ͻ0.2% (Fig. 5Bi). These ␥␦ T cell preparations, used at 2.5 ϫ by HLA-DR molecules, we blocked the presentation by incu- 104, still retained their capacity to present phagocytosed Ag bation of the ␥␦ T cells with a well characterized mAb (clone (Fig. 5Bii). Moreover, the fact that these cells were purified L243) against human HLA-DR (24) (Fig. 4Ci). This blocking with anti-TCR-␥9, indicates that the V␥9␦2 subtype, or a sub- resulted in an almost complete absence of presentation whereas type thereof, was responsible for the phagocytosis. This of incubation with isotype control Abs did not. In addition, we course does not preclude the possibility that also other subtypes showed that human MHC class II HLA-DRB1*0102-positive are capable of similar uptake and Ag presentation. The Journal of Immunology 5627

T cells, it also increased their expression of MHC class II pre- sentation markers (Fig. 6A), which may in part explain this finding. Furthermore, activation may also have increased the expression of other phagocytic receptors. To directly address the role of CD16, we investigated the ability of a murine IgG1 anti-CD16 blocking Ab (clone LNK16) to block phagocytosis. ␥␦ T cells were pulsed with M1 beads in the pres- ence and absence of opsonization. Opsonization increased Ag pre- sentation to the A1C5 hybridoma as described before (Fig. 6C). This enhancement was partially abrogated by preincubation of ␥␦ T cells with anti-CD16 blocking Ab. Murine IgG1 does not bind to FIGURE 6. CD16 is involved in phagocytosis and subsequent Ag pre- sentation by ␥␦ T cells. A, CD16 was expressed on the majority of freshly human CD16 via Fc (25). To confirm this, a murine IgG1 isotype isolated ␥␦ T cells (gray shade) but lost after overnight activation with IPP control was used at the same concentration and was unable to (solid line). Activated ␥␦ T cells (solid lines) also up-regulated Ag pre- block anti-M1 enhancement of Ag presentation. Previous studies sentation markers including HLA-DR and CD86 compared with freshly have looked at CD16 on ␥␦ T cells in light of Ab dependent cell isolated ␥␦ T cells (gray shade). B, CD16 high freshly isolated ␥␦ T cells (25, 27) or as a regulatory receptor (28). In this study, we show that CD16 engagement with Ig Fc plays an important role (ءء␦␥) were pulsed with opsonized M1 beads and subsequently activated Freshly isolated ␥␦ T cells from the same preparation were in phagocytosis and/or efficient Ag processing and presentation by .(ء␦␥) or not also directly activated overnight resulting in a CD16 low population, which human ␥␦ T cells, although other receptors are also likely to play Downloaded from ءءء␦␥ was then subsequently pulsed with opsonized M1 beads ( ). Surpris- a role. ingly, the CD16 low IPP-activated population of ␥␦ T cells were able to process and present M1 peptides to A1C5 hybridomas better than CD16 high M1 pulsed ␥␦ T cells. C, Ig opsonization enhancement of Ag pre- Discussion sentation was partially abrogated by blockade with mouse IgG1 ␣CD16. The ability to phagocytose large foreign bodies and process and Nonspecific mouse IgG1 at the same concentration had no effect. The present associated Ags on MHC class II is a hallmark of innate mean Ϯ SD of triplicate Ag presentations is plotted. myeloid lineage cells such as monocytes and DCs and has never http://www.jimmunol.org/ before been described in mammalian lymphocytes. We have shown in this study that freshly isolated ␥␦ T cells are able to Activation and CD16 engagement increases phagocytosis and phagocytose both E. coli and 1 ␮m synthetic beads via the CD16 ␥␦ Ag presentation by T cells receptor. Only a proportion of ␥␦ T cells observed had phagocy- Freshly isolated ␥␦ T cells display high expression of the phago- tosed and we suspect that this is due to differences in their acti- cytic receptor CD16/Fc␥RIII (16, 25). Upon activation, these cells vation state or perhaps different subpopulations of V␥9␦2 T cells lose expression of CD16 while up-regulating MHC class II, CD80, present. Furthermore, we have shown that phagocytosis leads to and CD86 (5, 16, 25) (Fig. 6A). The prolifically phagocytic DCs Ag processing and presentation on MHC class II of peptides also express CD16 and likewise upon activation DCs also lose clearly derived from the phagocytosed complexes. The use of both by guest on September 28, 2021 CD16 and up-regulate MHC class II, CD80, and CD86 to become confocal microscopy and TEM combined with anti-TCR-␥␦ stain- potent presenters of phagocytosed Ags (26). We therefore tested ing ensured the accuracy of the observations despite the presence whether CD16 engagement through opsonizing anti-M1 Abs of a low degree of monocyte contamination. Previous TEM studies and/or activation of ␥␦ T cells improved phagocytosis and subse- of ␥␦ T cells also reported very similar structural observations quent Ag presentation. ␥␦ T cells were pulsed with M1 beads in (29). The monocyte contamination was also of concern for the the presence or absence of anti-M1 opsonization and then were Ag presentation experiments. To ensure that the ␥␦ T cells were either IPP activated or not. To assess the role of CD16, ␥␦ T cells responsible for the Ag presentation of phagocytosed bead- were also pulsed with nonopsonized M1 beads. bound material, we used monocytes in exceeding amounts in ␥␦ T cells pulsed with nonopsonized M1 beads showed poor control experiments with resulting lower Ag presentations (Fig. MHC class II presentation of M1 peptide to A1C5 hybridomas. ␥␦ 5A). As an additional control experiment, we showed that T cells pulsed with opsonized M1 beads and ␥␦ T cells pulsed with CD16ϩ/CD56ϩ NK cells are unable to present M1 Ag to the nonopsonized M1 beads but which were subsequently activated by same T cell hybridoma (Fig. 4). A previous study by Kang et al. IPP were both able to present Ag to A1C5 hybridomas with com- (30) indicated that CD16ϩ NK cells may be able to phagocy- parable potency. Ig opsonization and IPP activation potentiated Ag tose. However, because these cells were only defined as NK presentation markedly, to a level greater than the sum of the two cells on the basis of CD16 positivity, we suggest that they may alone (Fig. 5A). well have been ␥␦ T cells. The TEM images obtained by Kang To further characterize the role of CD16 in phagocytosis and Ag et al. (30) showed cells with very similar ultrastructure to that presentation in ␥␦ T cells, we compared the ability of CD16 high of ␥␦ T cells reported in this study and previously. freshly isolated ␥␦ T cells and CD16 low overnight IPP-activated It has been proposed that T cells were the first lymphocytes to ␥␦ T cells to phagocytose M1 beads and present M1 peptide to acquire variable-diversity-joining-type receptors following their A1C5 hybridomas. CD16 high freshly isolated ␥␦ T cells were evolution from myeloid and lympho- (31). Recent pulsed with anti-M1 opsonized M1 beads and were subsequently work showed that the earliest thymic progenitors for T cells still either activated with IPP or not. Freshly isolated ␥␦ T cells from retain myeloid lineage potential, whereas progenitors do not the same preparation were also directly activated overnight with (32, 33). In addition, phylogenetic studies strongly favor a scenario IPP. These CD16 low activated ␥␦ T cells were then pulsed in which the ␥␦ TCR evolved before both the ␣␤ TCR and B cell with opsonized M1 beads. Despite lower expression of CD16, immunoglobulins (14). Because of their unique display of both IPP preactivated ␥␦ T cells pulsed with opsonized M1 beads myeloid and lymphoid lineage features, including phagocytosis, were able to activate A1C5 hybridomas more efficiently than we speculate that the predecessors of ␥␦ T cells represent the pro- CD16 high ␥␦ T cells that were pulsed first and then activated posed proto-T cell evolutionary intermediate from which lymphoid (Fig. 6B). Although activation reduced CD16 expression on ␥␦ cells arose from myeloid precursors (14). If this is correct, then it 5628 HUMAN ␥␦ T CELLS ARE CAPABLE OF PHAGOCYTOSIS is perhaps not surprising that these cells are capable of phagocy- Disclosures tosis and professional Ag presentation. It is interesting to note that The authors have no financial conflict of interest. B lymphocytes in teleost fish and Xenopus have also been shown to phagocytose microbes (22), although it was not shown whether References this can lead to Ag presentation. We speculate that the ability to 1. Hoebe, K., E. Janssen, and B. Beutler. 2004. The interface between innate and phagocytose was retained after lymphocyte evolution from ances- adaptive immunity. Nat. Immunol. 5: 971–974. ␥␦ 2. Crozat, K., E. Vivier, and M. Dalod. 2009. between components of the tral myeloid lineages in B cells in some animal groups and in : promoting anti-microbial defenses and avoiding immu- T cells in mammals further complicating simple interpretations of nopathologies. Immunol. Rev. 227: 129–149. how the evolved (34). Further studies of 3. Hayday, A., and R. Tigelaar. 2003. Immunoregulation in the tissues by ␥␦ T cells. Nat. Rev. Immunol. 3: 233–242. lymphocyte phagocytosis in different representative species should 4. Thedrez, A., C. Sabourin, J. Gertner, M. C. Devilder, S. Allain-Maillet, shed further light on this. J. J. Fournie, E. Scotet, and M. Bonneville. 2007. Self/non-self discrimination by ␥␦ The functional consequences of these findings are far-reaching. human T cells: simple solutions for a complex issue? Immunol. Rev. 215: 123–135. It is clear that Ab-assisted opsonization and uptake into DCs in- 5. Brandes, M., K. Willimann, and B. Moser. 2005. Professional -presenta- volves the low affinity receptor for IgG CD16 (Fc ␥RIII) (35). tion function by human ␥␦ T cells. Science 309: 264–268. 6. Cheng, L., Y. Cui, H. Shao, G. Han, L. Zhu, Y. Huang, R. L. O’Brien, Furthermore, this can lead to efficient and functionally important W. K. Born, H. J. Kaplan, and D. Sun. 2008. Mouse ␥␦ T cells are capable of enhancement of MHC class II presentation (36, 37). Judging by expressing MHC class II molecules, and of functioning as antigen-presenting anti-CD16 blocking experiments (Fig. 6C), a major part of the cells. J. Neuroimmunol. 203: 3–11. 7. Collins, R. A., D. Werling, S. E. Duggan, A. P. Bland, K. R. Parsons, and ␥␦ ϩ receptor-mediated phagocytic uptake into T cells appears to C. J. Howard. 1998. ␥␦ T cells present antigen to CD4 ␣␤ T cells. J. Leukocyte also be CD16 mediated. Other receptors are also likely to be in- Biol. 63: 707–714. Downloaded from volved, as we found CD16 low IPP-activated ␥␦ T cells were 8. Takamatsu, H. H., M. S. Denyer, and T. E. Wileman. 2002. A sub-population of circulating porcine ␥␦ T cells can act as professional antigen presenting cells. Vet. capable of phagocytic uptake and Ag presentation of M1 beads Immunol. Immunopathol. 87: 223–224. (Fig. 6B). We speculate that such uptake, apart from leading to Ag 9. Brandes, M., K. Willimann, G. Bioley, N. Levy, M. Eberl, M. Luo, R. Tampe, ␥␦ F. Levy, P. Romero, and B. Moser. 2009. Cross-presenting human ␥␦ T cells presentation, represents an important part of how T cells sense induce robust CD8ϩ ␣␤ T cell responses. Proc. Natl. Acad. Sci. USA 106: and react to microbes. The tissue resident distribution of some 2307–2312. 10. Caccamo, N., L. Battistini, M. Bonneville, F. Poccia, J. J. Fournie, S. Meraviglia, subsets of ␥␦ T cells (epithelial, including mucosal, tissues) would http://www.jimmunol.org/ G. Borsellino, R. A. Kroczek, C. La Mendola, E. Scotet, et al. 2006. CXCR5 make them ideal for this purpose. TLR and related “sensing of identifies a subset of V␥9V␦2 T cells which secrete IL-4 and IL-10 and help B nonself” molecules in endosomes and in the cytoplasm are known cells for production. J. Immunol. 177: 5290–5295. to be present and responsive in ␥␦ T cells (38, 39), and could 11. Devilder, M. C., S. Maillet, I. Bouyge-Moreau, E. Donnadieu, M. Bonneville, and E. Scotet. 2006. Potentiation of antigen-stimulated V ␥ 9V ␦ 2 T cell therefore represent an important regulating step in the encounter production by immature dendritic cells (DC) and reciprocal effect on DC matu- with an endocytosed or phagocytosed . It was recently ration. J. Immunol. 176: 1386–1393. 12. Moser, B., and M. Eberl. 2007. ␥␦ T cells: novel initiators of adaptive immunity. demonstrated that mycobacteria activates and expands Immunol. Rev. 215: 89–102. certain subsets of V␥9␦2 ␥␦ T cells that then develop greater Ag 13. Rabinovitch, M. 1995. Professional and non-professional : an intro- specificity by TCR selection reminiscent of ␣␤ T cells forming duction. Trends Cell Biol. 5: 85–87.

14. Richards, M. H., and J. L. Nelson. 2000. The evolution of vertebrate antigen by guest on September 28, 2021 true memory (40). It would be interesting to address the possibility receptors: a phylogenetic approach. Mol. Biol. Evol. 17: 146–155. that such specific V␥9␦2 ␥␦ T cell expansion involved an initial 15. Nagarajan, S., S. Chesla, L. Cobern, P. Anderson, C. Zhu, and P. Selvaraj. 1995. ␥ phagocytosis event. It is also likely that eukaryotic cells or parts Ligand binding and phagocytosis by CD16 ( III) isoforms: phago- cytic signaling by associated ␨ and ␥ subunits in Chinese hamster cells. thereof, including autologous cells, can be phagocytosed by ␥␦ T J. Biol. Chem. 270: 25762–25770. cells. This would have consequences for tissue renewal, repair, and 16. Bodman-Smith, M. D., A. Anand, V. Durand, P. Y. Youinou, and P. M. Lydyard. 2000. Decreased expression of Fc␥RIII (CD16) by ␥␦ T cells in patients with wound healing where apoptotic or necrotic cells need to be re- . Immunology 99: 498–503. moved in a controlled, noninflammatory fashion. Additionally and 17. Mi, S., B. Hu, K. Hahm, Y. Luo, E. S. Kam Hui, Q. Yuan, W. M. Wong, importantly, ␥␦ T cells are well known to take an active part in L. Wang, H. Su, T. H. Chu, et al. 2007. LINGO-1 antagonist promotes remyelination and axonal integrity in MOG-induced experimental autoimmune antitumor immune responses, and their use in immunotherapeutic encephalomyelitis. Nat. Med. 13: 1228–1233. cancer treatments is part of several clinical trials (41, 42). Strid 18. Canaday, D. H., A. Gehring, E. G. Leonard, B. Eilertson, J. R. Schreiber, et al. (43) recently showed that murine intraepithelial ␥␦ T cells C. V. Harding, and W. H. Boom. 2003. T-cell hybridomas from HLA-transgenic mice as tools for analysis of human . J. Immunol. Methods 281: respond to and inhibit carcinogenesis in response to specific mo- 129–142. lecular signals and Peng et al. demonstrated that ␥␦ T cells within 19. Amorim, M. J., E. K. Read, R. M. Dalton, L. Medcalf, and P. Digard. 2007. Nuclear export of influenza A virus mRNAs requires ongoing RNA polymerase cancer tissue can operate as T regulatory cells suppressing an an- II activity. Traffic 8: 1–11. titumor response (39). Interestingly, this suppression could, how- 20. Krummel, M. F., M. D. Sjaastad, C. Wulfing, and M. M. Davis. 2000. Differential ever, be overcome by engagement of TLR8 within the ␥␦ T cells. clustering of CD4 and CD3␨ during T cell recognition. Science 289: 1349–1352. ␥␦ 21. Chen, Y., L. Shao, Z. Ali, J. Cai, and Z. W. Chen. 2008. NSOM/QD-based If apoptotic or necrotic tumor cells can be ingested by T cells, nanoscale immunofluorescence imaging of antigen-specific T-cell receptor re- this is likely to have an impact on both their own interactions with sponses during an in vivo clonal V␥2V␦2 T-cell expansion. Blood 111: tumors as well as their ability to interact with and instruct other 4220–4232. 22. Li, J., D. R. Barreda, Y. A. Zhang, H. Boshra, A. E. Gelman, S. Lapatra, L. Tort, arms of antitumor responses. The uptake of particulates by ␥␦ T and J. O. Sunyer. 2006. B lymphocytes from early vertebrates have potent phago- cells also has a potential impact on the development of new vac- cytic and microbicidal abilities. Nat. Immunol. 7: 1116–1124. 23. Sanderson, R. J., R. T. Shepperdson, A. E. Vatter, and D. W. Talmage. 1977. cination strategies, for infectious disease as well as in autoimmu- Isolation and enumeration of peripheral blood monocytes. J. Immunol. 118: nity and cancer , because it may represent an im- 1409–1414. portant initial step in the categorization of immune complexes as 24. Robbins, P. A., E. L. Evans, A. H. Ding, N. L. Warner, and F. M. Brodsky. 1987. Monoclonal that distinguish between class II (HLA-DP, DQ, self or nonself. and DR) in 14 haplotypes. Hum. Immunol. 18: 301–313. 25. Braakman, E., J. G. van de Winkel, B. A. van Krimpen, M. Jansze, and R. L. Bolhuis. 1992. CD16 on human ␥␦T lymphocytes: expression, function, and specificity for mouse IgG isotypes. Cell Immunol. 143: 97–107. Acknowledgments 26. Gehring, S., S. H. Gregory, P. Wintermeyer, M. San Martin, C. Aloman, and We thank Ayadh Eddaoudi and Joan Sinclair for technical help with flow J. R. Wands. 2008. Generation and characterization of an immunogenic dendritic cell population. J. Immunol. Methods 332: 18–30. cytometry; Siobhan Burns for technical help with confocal microscopy; 27. Tokuyama, H., T. Hagi, S. R. Mattarollo, J. Morley, Q. Wang, H. Fai-So, and Prof. Adrian Hayday and Dr. Helen Karlsson for critical discussions. F. Moriyasu, M. Nieda, and A. J. Nicol. 2008. V ␥ 9V␦ 2 T cell cytotoxicity The Journal of Immunology 5629

against tumor cells is enhanced by drugs– and 36. Herrada, A. A., F. J. Contreras, J. A. Tobar, R. Pacheco, and A. M. Kalergis. trastuzumab. Int. J. Cancer 122: 2526–2534. 2007. -induced enhancement of bacterial 28. Lafont, V., J. Liautard, J. P. Liautard, and J. Favero. 2001. Production of TNF-␣ requires Fc␥ receptor III expression on dendritic cells. Proc. Natl. Acad. Sci. USA by human V ␥ 9V ␦ 2 T cells via engagement of Fc ␥ RIIIA, the low affinity type 104: 13402–13407. 3 receptor for the Fc portion of IgG, expressed upon TCR activation by nonpep- 37. Rafiq, K., A. Bergtold, and R. Clynes. 2002. Immune complex-mediated antigen tidic antigen. J. Immunol. 166: 7190–7199. presentation induces tumor immunity. J. Clin. Invest. 110: 71–79. 29. Lundqvist, C., V. Baranov, S. Teglund, S. Hammarstrom, and M. L. Hammarstrom. 38. Kabelitz, D., and R. Medzhitov. 2007. Innate immunity–cross-talk with adaptive ␥␦ 1994. Cytokine profile and ultrastructure of intraepithelial T cells in chronically immunity through pattern recognition receptors and . Curr. Opin. Im- inflamed human gingiva suggest a cytotoxic effector function. J. Immunol. 153: munol. 19: 1–3. 2302–2312. 39. Peng, G., H. Y. Wang, W. Peng, Y. Kiniwa, K. H. Seo, and R. F. Wang. 2007. 30. Kang, Y. H., M. Carl, L. P. Watson, and L. Yaffe. 1987. Immunoultrastructural Tumor-infiltrating ␥␦ T cells suppress T and dendritic cell function via mecha- studies of human NK cells: II, effector-target cell binding and phagocytosis. Anat. nisms controlled by a unique toll-like receptor signaling pathway. Immunity 27: Rec. 217: 290–304. 334–348. 31. Boehm, T., and C. C. Bleul. 2007. The evolutionary history of lymphoid organs. 40. Spencer, C. T., G. Abate, A. Blazevic, and D. F. Hoft. 2008. Only a subset of Nat. Immunol. 8: 131–135. ␥ ␦ 32. Bell, J. J., and A. Bhandoola. 2008. The earliest thymic progenitors for T cells phosphoantigen-responsive 9 2 T cells mediate protective immu- J. Immunol. possess myeloid lineage potential. Nature 452: 764–767. nity. 181: 4471–4484. 33. Wada, H., K. Masuda, R. Satoh, K. Kakugawa, T. Ikawa, Y. Katsura, and 41. Caccamo, N., S. Meraviglia, F. Scarpa, C. La Mendola, D. Santini, H. Kawamoto. 2008. Adult T-cell progenitors retain myeloid potential. Nature C. T. Bonanno, G. Misiano, F. Dieli, and A. Salerno. 2008. Aminobisphospho- ␥␦ 452: 768–772. nate-activated T cells in immunotherapy of cancer: doubts no more. Expert 34. Kawamoto, H., and Y. Katsura. 2009. A new paradigm for hematopoietic cell Opin Biol. Ther. 8: 875–883. lineages: revision of the classical concept of the myeloid-lymphoid dichotomy. 42. Kabelitz, D., D. Wesch, and W. He. 2007. Perspectives of ␥␦ T cells in tumor Trends Immunol. 30: 193–200. immunology. Cancer Res. 67: 5–8. 35. Woelbing, F., S. L. Kostka, K. Moelle, Y. Belkaid, C. Sunderkoetter, S. Verbeek, 43. Strid, J., S. J. Roberts, R. B. Filler, J. M. Lewis, B. Y. Kwong, W. Schpero, A. Waisman, A. P. Nigg, J. Knop, M. C. Udey, and E. von Stebut. 2006. Uptake D. H. Kaplan, A. C. Hayday, and M. Girardi. 2008. Acute upregulation of an of by dendritic cells is mediated by Fc␥ receptors and facili- NKG2D ligand promotes rapid reorganization of a local immune compartment Downloaded from tates acquisition of protective immunity. J. Exp. Med. 203: 177–188. with pleiotropic effects on carcinogenesis. Nat. Immunol. 9: 146–154. http://www.jimmunol.org/ by guest on September 28, 2021