Phosphoantigen Presentation by Macrophages to Mycobacterium tuberculosis-Reactive Vγ9Vδ2+ T Cells: Modulation by Chloroquine

ABSTRACT Vγ9Vδ2+ T cells (γδ T cells) are activated by Mycobacterium tuberculosis and recognize mycobacterial nonpeptide phosphoantigens. The role of antigen-presenting cells in the processing and presentation of phosphoantigens to Vγ9Vδ2+ T cells is not understood. We analyzed the role of macrophages for activation of γδ T cells by a new synthetic phosphoantigen bromohydrin pyrophosphate (BrHPP) and M. tuberculosis. Macrophages greatly increased γδ T-cell activation by both BrHPP and M. tuberculosis. Fixation of macrophages before infection demonstrated that uptake of M. tuberculosis was required for presentation to γδ T cells. Antigens of M. tuberculosis remained stably associated with macrophage surface and were not removed by paraformaldehyde fixation or washing. Macrophages processed M. tuberculosis for γδ T cells through a brefeldin A-insensitive pathway, suggesting that transport through the endoplasmic reticulum and Golgi complex of a putative presenting molecule is not important in the early processing of M. tuberculosis antigens for γδ T cells. Processing of M. tuberculosis was not eliminated by chloroquine, indicating that processing of γδ antigens is not dependent on acidic pH in the lysosomes. Chloroquine treatment of BrHPP-pulsed macrophages increased activation of γδ T cells. Ammonium chloride treatment of macrophages did not increase reactivity of γδ T cells to BrHPP, indicating that the effect of chloroquine was independent of pH changes in endosomes. Chloroquine, by inhibiting membrane traffic, may increase association and retention of phosphoantigens with cell surface membrane molecules on macrophages.

V␥9V␦2 T cells recognize nonpeptidic compounds with phosphoester structures, collectively called "phosphoantigens." Among phosphoantigens, TUBags were purified from mycobacterial lysates and found inside M. tuberculosis bacilli (15,16,31,41). They are ␥-derivatives of UTP and dTTP (X-␥TTP). Other phosphoantigens include pyrophospho-and phosphomonoesters (isopentenyl pyrophosphate [IPP], monoethylpyrophosphate, 2,3 diphosphoglycerate, glycerol 3-phosphate, tuberculosis antigens 1 and 2 [TUBag1-2], and malaria antigens 1 and 2). IPP was isolated from extracts and culture filtrates of fast-growing mycobacteria (M. fortuitum and M. smegmatis) (34,(49)(50)(51). IPP is a precursor in cholesterol synthesis and in cholesterol derivatives (steroid hormones, vitamin D, bile salts, and lipoproteins) and terpenoids and is conserved between prokaryotic and eukaryotic cells. The basis of selfnonself discrimination between infected and noninfected cells may consist of recognition of metabolic intermediates (IPP precursors) that are produced through biochemical pathways exclusively present in microbes (Rohmer pathway) (7,46). The recently identified 3-formyl-1-butyl pyrophosphate is likely a biosynthetic precursor of mycobacterial IPP, eliciting ␥␦ T-cell activation and targeting responses to infected cells (8). This compound is produced in very small amounts in slow-growing mycobacteria such as M. tuberculosis and accumulates to submicromolar concentrations in culture media from fast-growing mycobacterial species. A synthetic analogue called bromohydrin pyrophosphate (BrHPP) has been developed that mimics the biological properties of natural phosphoantigens, is easily synthesized, and is active at nanomolar concentrations (18).
How ␥␦ TCR recognizes phosphoantigens is not fully understood. Phosphoantigen recognition is TCR dependent, re-quires cell-cell contact, and is not restricted by classical or nonclassical major histocompatibility complex (MHC) (13,19,34,51). Direct binding of ␥␦ TCR to phosphoantigens has not been demonstrated, suggesting the existence of an, as of yet, unknown presenting molecule(s) (31). Phosphoantigens can be recognized without requiring uptake or presentation by professional antigen-presenting cells (APC); however, ␥␦ T-cell responses to intact M. tuberculosis bacilli depend on accessory cells (11,19,47,51). Whether accessory cells, in addition to being a source of costimulatory signals, process and/or present phosphoantigens to ␥␦ T cells is not known. ␥␦ T cells can be activated by phosphoantigens directly (in the absence of APC), but it is not clear that this mechanism is important for responses to M. tuberculosis. Phosphoantigens of M. tuberculosis are found in cytosol of the bacteria and are not secreted. During infection phosphoantigens must traffic from phagosomes to the cell surface by an undefined pathway.
The aim of the present study was to investigate the role of macrophages as APC in the processing and presentation of natural and synthetic phosphoantigen to ␥␦ T cells. M. tuberculosis bacilli and IPP were used as sources of natural phosphoantigens, and BrHPP was used as a synthetic analogue. Phosphoantigens and M. tuberculosis were compared in terms of dependence on macrophages for ␥␦ T-cell activation. The effect of intracellular antigen-processing inhibitors in early processing of M. tuberculosis antigens for ␥␦ T cells was assessed also. Phosphoantigens did not require intracellular processing but remained stably associated on the surface of macrophages in the presence of chloroquine, suggesting an interaction between BrHPP and cell membrane molecules. In addition, intracellularly derived antigens of M. tuberculosis were stably associated on the surface of macrophages and were not removed after paraformaldehyde fixation and extensive washings. Our results support a model in which phosphoantigens associate with host molecules on the surface of macrophages and membrane traffic regulates the availability of these phosphoantigens for ␥␦ T-cell recognition.
BrHPP was kindly provided by Christian Belmant, Innate Pharma, Marseille, France. IPP was purchased from Sigma Chemical Co, St. Louis, Mo. Purified protein derivative (PPD) from M. tuberculosis was obtained from Wyeth-Lederle Vaccines, Pearl River, N.Y.
Bacteria and bacterial lysate. M. tuberculosis H37Ra was cultured in Middlebrook 7H9 with albumin dextrose catalase enrichment and frozen stocks prepared as described previously (11,24). Bacterial counts and viability were performed by light microscopy and by counting CFU on 7H10 medium. M. tuberculosis-H37Ra stocks were tested periodically for viability and with an M. tuberculosis complex-specific DNA probe (AccuProbe; Gen-Probe, San Diego, Calif.) to assure purity of M. tuberculosis stocks. Before use in T-cell assays, mycobacteria were washed three times in RPMI 1640, sonicated for 40 s, passed multiple times through a 25-gauge needle to disrupt clumps, and diluted in non-heat-inactivated serum-containing medium.
M. tuberculosis H37Ra lysate was obtained by sonicating and washing mycobacteria on ice (three 3-min sonications) and then passing them through a French press (three times). Lysate of M. tuberculosis was centrifuged for 2 h at 145,000 ϫ g at 4°C. Soluble material was harvested and protein content was determined by Bio-Rad assay (Bio-Rad Laboratories, Richmond, Calif.). Ali-quots (5 mg/ml protein) were stored at Ϫ80°C. For ␥␦ T-cell stimulation, dilutions of stock material were prepared and normalized according to protein content.
Isolation of PBMC, monocytes, and macrophages. Peripheral blood mononuclear cells (PBMC) were isolated by density gradient centrifugation over sodium diatrizoate-Hypaque, and monocytes were obtained by adherence from PBMC as previously described (52). PBMC were isolated from healthy tuberculin-positive persons (18 to 45 years old). They were selected for consistency of ␥␦ T-cell expansion (20 to 60% ␥␦ TCR ϩ T cells) after stimulation with live M. tuberculosis for 7 to 10 days.
For preparing monocyte-derived macrophages, PBMC were incubated on plastic tissue culture dishes precoated with pooled human serum; nonadherent cells were removed, and adherent cells were collected by scraping with a plastic policeman. Monocytes were resuspended in culture media containing RPMI 1640 supplemented with 10% pooled human serum, 20 mM HEPES, 2 mM L-glutamine, 100 U/ml penicillin, and 100 g/ml streptomycin. Cells were placed in flat-bottom 96-well plates and allowed to differentiate without addition of growth factors for 7 to 10 days at 37°C.
Purification of M. tuberculosis-activated CD4 ؉ and ␥␦ T-cell populations. PBMC stimulated with live M. tuberculosis for 10 days were used to obtain CD4 ϩ and ␥␦ T cells by positive selection. Viable cells were harvested by density sedimentation on sodium diatrizoate-Hypaque gradients. CD4 ϩ and ␥␦ T-cell subsets were purified by positive selection with magnetic beads coated with antibodies (Militenyi Biotec, Gladbach, Germany). For CD4 ϩ T-cell enrichment cells were incubated with beads conjugated to monoclonal mouse anti-human CD4 (Leu-3a). For ␥␦ T-cell purification, cells were first incubated with a haptenmodified anti-␥␦ TCR antibody, followed by treatment with FITC-conjugated antihapten microbeads. Purity was checked by fluorescence-activated cell sorting. One cycle of selection was sufficient to obtain Ͼ95% T-cell purity (Ͼ95% V␦2-positive cells). Cells were rested for 48 h in medium containing IL-2 (50 U/ml) before use in functional assays.
Cells were analyzed on a FACScan (Becton and Dickinson) with the CellQuest software. Cells were gated in a two-parameter plot of 90°versus forward angle scatter. The gate for lymphocytes or monocytes was set widely. Five thousand events were recorded for each cell surface marker. The cutoff lines for positive and negative fluorescence were set manually based on the distribution of cells stained with FITC and PE-conjugated isotypic control antibodies alone and were kept constant within each experiment. The percentage reported for a given surface marker represents the proportion of gated cells with a positive signal less the percentage of cells staining positive with isotypic control antibody alone.
Proliferation and IFN-␥ assays. PBMC or prestimulated and positively selected CD4 ϩ and ␥␦ T cells from different donors (5 ϫ 10 4 cells per 200 l-well) were cocultured with or without autologous macrophages or phorbol 12-myris- twice with RPMI before addition of 0.2 M lysine. After 20 min, lysine was discarded and plates were washed four times with RPMI. Alternatively, noninfected macrophages were fixed and antigen pulse was performed for 4 h. After infection and pulse were carried out, macrophages were washed extensively and 5 ϫ 10 4 purified CD4 ϩ or ␥␦ T cells were added to the wells. On day 3 supernatants were harvested for IFN-␥ measurement by ELISA and cells were pulsed with 1 Ci of [ 3 H]thymidine for 12 to 16 h before harvesting and proliferation was determined as described above.
For brefeldin A (BFA) and chloroquine treatment, cells were pretreated with BFA (Boehringer Mannheim Biochemical, Indianapolis, Ind.) at 5 g/ml, chloroquine (Sigma-Aldrich) at 5 to 500 M, or ammonium chloride at 5 to 40 mM for 30 min before addition of the antigens. Inhibitors were present during the entire period of incubation (4 h) with antigens. APC cells were washed and fixed with paraformaldehyde, and purified CD4 ϩ or ␥␦ T cells (5 ϫ 10 4 cells/well) were added. After 3 days, IFN-␥ was measured in culture supernatants and T-cell proliferation was assessed by [ 3 H]thymidine incorporation.
Statistical analysis. Statistical analysis was determined by Student's t test, and a P of Ͻ0.05 was considered significant.

BrHPP expands ␥␦ T cells from peripheral blood at nanomolar concentrations and requires IL-2 for proliferation and IFN-␥ secretion.
First the ability of BrHPP to expand resting ␥␦ T cells from peripheral blood was compared to T-cell expansion by IPP and M. tuberculosis. PBMC from healthy individuals (10 6 cells/ml; 6 ml of culture) were stimulated with live M. tuberculosis bacilli (1:1 or 10:1 M. tuberculosis-to-macrophage ratio), IPP (100 nM to 10 M), or BrHPP (10 nM to 1 M) in the presence of IL-2 (50 U/ml). After 10 days, viable cells were harvested, counted, and analyzed by two-color flow cytometry for CD3, CD4, CD8, and ␥␦ TCR expression. Results are expressed as absolute number of ␥␦ TCR ϩ T cells at the end of culture (Fig 1A). ␥␦ T cells expanded from 3 to 5% at baseline to 20 to 60% after stimulation with M. tuberculosis or phosphoantigens (n ϭ 5). As shown in Fig. 1A, M. tuberculosis induced the greatest expansion of V␥9V␦2 T cells after 10 days of stimulation. When IPP and BrHPP were compared, it was observed that BrHPP induced the same degree of ␥␦ expansion as IPP but at 100-fold lower concentrations. While threshold concentrations of these antigens for ␥␦ expansion varied from donor to donor, BrHPP consistently had a lower threshold dose than IPP (i.e., when BrHPP caused a significant ␥␦ T-cell expansion at 100 nM, IPP threshold was 1 M). Neither BrHPP nor IPP caused expansion of CD4 ϩ or CD8 ϩ T cells (data not shown). Responses to BrHPP were completely eliminated by pretreatment with calf intestinal alkaline phosphatase (Promega, Madison, Wis.) confirming that the pyrophosphate was responsible for ␥␦ T-cell activation (data not shown). Thus, BrHPP on a molar basis had greater activating capability for ␥␦ T cells than IPP and thus is suitable for studies of antigen presentation.
As described previously, ␥␦ T-cell activation by natural phosphoantigens (IPP) depends not only on TCR signaling but also on additional signals provided by T-cell growth factors such as IL-2 (53). We confirmed that proliferation and IFN-␥ secretion by ␥␦ T cells in response to the synthetic phosphoantigen BrHPP depended on IL-2. Without IL-2 minimal proliferation and no IFN-␥, secretion was observed even at high phosphoantigen concentrations (Fig. 1B and C).
Optimal responses of ␥␦ T-cell lines to BrHPP require APC. To study the role of APC on ␥␦ T-cell activation, ␥␦ T cells were first expanded from PBMC of different donors with M. were observed with M. tuberculosis and with BrHPP in the absence of macrophages, but the addition of macrophages increased ␥␦ T-cell responses three-to fourfold (Fig. 2). Thus, although M. tuberculosis and phosphoantigens can somewhat activate ␥␦ T cells directly in absence of macrophages, optimal responses are obtained in the presence of macrophages. Experiments conducted with resting ␥␦ T cells purified from freshly isolated PBMC demonstrated that activation requirements for resting ␥␦ T cells were the same as those observed for in vitro-activated ␥␦ T cells. Antigen-presenting cells greatly increased resting ␥␦ T-cell responses (data not shown). In Fig. 3 increased ␥␦ T-cell responses in the presence of macrophages was not dependent on matching of MHC, since ␥␦ T-cell lines and macrophages used in these experiments did not share MHC class I (MHC-I) or MHC-II haplotypes. In addition, ␥␦ T-cell line responses to antigen-pulsed matched and mismatched macrophages from different donors were sim-ilar, confirming the lack of self MHC restriction (data not shown). Thus, ␥␦ T-cell responses to the new synthetic phosphoantigen BrHPP resemble those observed to natural phosphoantigens or M. tuberculosis bacilli both in terms of the ability of APC to enhance ␥␦ T-cell activation and the lack of self MHC restriction.
Uptake of M. tuberculosis by macrophages generates antigens that are stably presented to ␥␦ T cells. The role of antigen uptake and intracellular processing was first assessed by means of pulse fixation experiments. In these experiments, macrophages were infected or pulsed with antigen before or after fixation with paraformaldehyde. Cells were extensively washed after antigen pulse and therefore there was no free (soluble) antigen present at the time of T-cell addition. As shown in Fig. 4 macrophages fixed after infection with M. tuberculosis were able to activate ␥␦ T cells, although at much lower levels than nonfixed macrophages ( Fig. 2A). In contrast macrophages exposed to M. tuberculosis after fixation did not activate ␥␦ T cells. Responses to soluble antigens such as M. tuberculosis lysate and BrHPP pulsed onto the macrophages before fixation were low but detectable. These experiments suggest that uptake of M. tuberculosis and subsequent degra- dation inside APC results in antigens for ␥␦ T cells that are stably associated with the cell surface.

Effect of BFA and chloroquine on M. tuberculosis and phosphoantigen processing by human macrophages.
To determine if M. tuberculosis and phosphoantigens differ in cellular processing requirements for ␥␦ T-cell activation, macrophages were treated with BFA or chloroquine before (30 min) and during antigen pulse (4 h). BFA treatment had no effect on early processing of M. tuberculosis and did not affect phosphoantigen presentation to ␥␦ T cells (n ϭ 4). Control experiments demonstrated that BFA inhibited MHC-II processing of tetanus toxoid (TT) for a tetanus toxoid-specific hybridoma (data not shown). These results indicated that unlike classical MHC-I or MHC-II antigen-processing pathways, anterograde transport through the endoplasmic reticulum and Golgi had no role in antigen processing for ␥␦ T cells. Chloroquine did not eliminate M. tuberculosis processing and presentation for ␥␦ T cells. Surprisingly, chloroquine treatment of phosphoantigen pulsed-macrophages consistently increased activation of ␥␦ T cells (Fig. 5). These findings suggest either that intralysosomal pH variation by chloroquine does not affect generation of ␥␦Ϫactivating antigens in M. tuberculosis-infected macrophages or that potential down-regulation of processing by chloroquine is obscured by up-regulation of phosphoantigen presentation.
␥␦ losis lysate or the natural phosphoantigen IPP was used as antigen. Up-regulation of phosphoantigen presentation by chloroquine was seen at concentrations of chloroquine higher than 5 M and at concentrations of BrHPP less than or equal to 10 M (IPP less than or equal to 100 M), suggesting a dose-dependent mechanism in which a saturation state is reached at high antigen concentrations. The effect of chloroquine on processing and presentation of a soluble antigen (PPD) for MHC-II-restricted CD4 ϩ T cells was investigated also. Macrophages from the same donor were treated with chloroquine and used to present antigens to CD4 ϩ or ␥␦ T cells. Figure 7 demonstrates that processing and presentation of PPD, especially at high antigen concentrations, was readily inhibited by chloroquine. In contrast, BrHPP responses were significantly increased when cells were treated with 50 M chloroquine. Parallel results were obtained when [ 3 H]thymidine incorporation (proliferation) was used as a measure of ␥␦ T-cell or CD4 ϩ T-cell activation (data not shown).
Next, we compared chloroquine with another lysosomotrophic agent, ammonium chloride. Figure 8 demonstrates the effect of chloroquine and ammonium chloride treatment on BrHPP-pulsed macrophages. Chloroquine-treated BrHPPpulsed macrophages exhibited up-regulated stimulatory capability for ␥␦ T cells in a dose-dependent manner compared to non-chloroquine-treated macrophages. In contrast, a wide range of ammonium chloride concentrations (5 to 40 mM) failed to affect ␥␦ T-cell responses to BrHPP, indicating that the effect is independent of pH changes in the lysosomes.

DISCUSSION
The most-common ␥␦ T-cell subset in adult humans, V␥9V␦2 T cells, can be stimulated by a variety of pathogenic microorganisms (36). The common feature of V␥9V␦2 T-cellstimulating compounds is the presence of phosphate groups. Among these antigens, TUBags and IPP were the first ones to be identified. How phosphoantigens are recognized by ␥␦ TCR is not understood. Although intracellular processing may not be required, cell surface presentation and possibly extracellular processing may play a role in optimal recognition of ligands by ␥␦ TCR (7). However, the existence of presenting molecules and requirements for cell surface processing of phosphoantigens remain to be established.
The present study compared responses to a new synthetic phosphoantigen, BrHPP; the natural phosphoantigen IPP; and M. tuberculosis bacilli-particularly focused on requirements for antigen processing and presentation. First the ability of a new synthetic phosphoantigen, BrHPP, to expand and stimulate V␥9V␦2 T cells was investigated. BrHPP was more potent on a molar basis than IPP in stimulation of V␥9V␦2 T cells. M. tuberculosis was the most-potent stimulator of ␥␦ T cells compared to either phosphoantigen. Consistent with previous reports, V␥9V␦2 T-cell responses to phosphoantigen were dependent on exogenous IL-2 (53).
Our studies demonstrate that ␥␦ T-cell responses to both M. tuberculosis and to soluble antigens were markedly increased in the presence of APC. Most studies have suggested that enhanced ␥␦ T-cell responses in the presence of APC are primarily due to costimulatory activity rather than antigen processing and presentation. The large increases in ␥␦ T-cell responses observed in our study in the presence of APC suggest additional mechanisms besides costimulation, such as intra-and/or extracellular processing or association to a receptor or presenting molecule resulting in increased affinity of the epitope or allowing multivalent interaction with the V␥9V␦2 TCR. The APC-enhancing effect on ␥␦ T-cell responses was dependent on APC density (data not shown) and type of APC (i.e., THP-1 increased ␥␦ T-cell responses more than autologous or heterologous macrophages [ Fig. 2 and 3]). In addition, responses with live macrophages in continuous presence of the antigen were higher than responses to fixed and/or pulsed macrophages. The fact that ␥␦ T cells can respond to antigens in the absence of macrophages is consistent with a model of T-cell-to-T-cell presentation.
Pulse-fixation experiments demonstrated that viable cells were required for M. tuberculosis presentation to ␥␦ T cells. These results suggest that intact mycobacteria need intracellular processing for phosphoantigens to become available for recognition on the surface of APC. Two inhibitors of intracellular processing were tested: BFA and chloroquine. BFA inhibits transport from the endoplasmic reticulum to the trans-Golgi network (22). BFA blocks the supply of nascent MHC-I and MHC-II molecules to the endogenous and endocytic pathways, respectively, and inhibits the formation of most peptide-MHC complexes in the endoplasmic reticulum or in late endocytic compartment (2,38). Previous work by our group focused on effects of BFA on ongoing antigen processing (i.e., after 12 h of pulse with the antigen) and used cytotoxicity as a measure of ␥␦ T-cell activation. Our present approach differs kinetically from the previous one in that BFA was used to study initial processing (i.e., occurring at the time of antigen uptake) and ␥␦ T-cell activation was assessed by proliferation and IFN-␥ production. In our present work BFA did not inhibit initial processing of M. tuberculosis or soluble phosphoantigens. In contrast, previous studies demonstrated that BFA partially inhibited ongoing processing of M. tuberculosis in unfixed monocytes in a CTL assay (4). This may indicate that a different pool of presenting molecules is used in initial processing which does not involve de novo synthesis and trafficking through endoplasmic reticulum to the trans-Golgi network. Our present experimental approach indicates that nascent presenting molecules are not required for initial processing and presentation of M. tuberculosis antigens to ␥␦ T cells.
Chloroquine is a weak base amine that inhibits both MHC-II and MHC-I antigen-processing pathways (endosomal processing pathways) (21,32). Ammonia and weakly basic alkylamines increase the endosomal and lysosomal pH and thus decrease degradation of endocytosed proteins, affecting ligand-receptor dissociation and receptor recycling (40,45). Some authors have reported that apart from effects on endocytic compartments, certain Golgi functions, i.e., transport of lipoproteins or antibody secretion and/or glycosylation, may be affected by changes in pH caused by chloroquine or ammonia (inhibition of MHC-I processing) (43,48). The effect of ammonia is primarily due to increased pH, whereas chloroquine has a variety of additional effects. These include direct inhibition of protein-  (44,54). Our results suggest that chloroquine modified the cell membrane, increasing the ability of phosphoantigen-pulsed macrophages to stimulate the ␥␦ TCR cells, whereas ammonium chloride had no effect. The fact that we did not see the same effect with ammonium chloride may be attributable to the more restricted effects of ammonia on pH. The enhancement effect of chloroquine was specific for phosphoantigens and ␥␦ T cells since it was not observed on PPDpulsed macrophages and CD4 ϩ T-cell responses. Since chloroquine alone does not stimulate ␥␦ T cells and is washed away before addition of T cells, the possibility that is binding directly to the ␥␦ TCR, increasing the effect phosphoantigens, is low. We hypothesize that the effect of chloroquine on presentation of phosphoantigens is due to modification of the APC membrane that results in increased association of phosphoantigens with cell surface molecules (putative receptors). It is tempting to speculate that chloroquine decreases the degradation or turnover of receptors by altering the dynamics of the endocytic compartment or changing membrane fluidity. Alternatively, chloroquine may be inducing expression of a new molecule on APC. This newly synthesized molecule may interact with ␥␦ T cells rendering them more sensitive to the activating effects of phosphoantigens. Other authors have reported that phosphoantigens cannot be pulsed onto APC (34,47). These studies did not use macrophages as APC, but use Epstein-Barr virustransformed B-cell lines, syngeneic and allogeneic PBMC, B lymphoma cells, fibroblasts, and V␥9V␦2 T-cell clones. Furthermore, in these previous reports the APCs were not fixed after antigen pulse. Although the presence of antigen on cell surface has not been formally demonstrated, the ability of phosphoantigen-pulsed chloroquine-treated macrophages to stimulate M. tuberculosis-reactive ␥␦ T cells after extensive washing and fixation suggests that the antigen was in fact stably associated with the surface of the APC and presented to the ␥␦ TCR. The fact that prefixed macrophages could not be pulsed with phosphoantigens indicates that the interaction of antigen with macrophage membrane is less stable than the MHCpeptide interaction (classical MHC presentation of peptides) or that putative presenting molecules become unavailable or unreceptive after fixation. Recent reports of the human ␥␦ TCR crystal structure and docking of phosphoantigens into a V␥9V␦2 model do not provide a mechanism for ␥␦ T-cell activation by direct interaction of small phosphoantigen and the ␥␦ TCR. These studies leave open the possibility of an antigen-presenting molecule (3,35). Our results present indirect evidence for an interaction between ␥␦ TCR, phosphoantigens, and cell membrane components that promote ␥␦ T-cell activation. Chloroquine inhibition of membrane turnover may allow greater retention of phosphoantigens at the cell surface and hence greater ␥␦ TCRantigen contact.
In summary BrHPP triggers ␥␦ T-cell responses more efficiently than natural phosphoantigens (IPP), and the response is increased in the presence of APC and chloroquine. Although further studies are required to understand the mechanism, the ability of phosphoantigen-pulsed chloroquine-treated macrophages to stimulate ␥␦ T cells may be an important tool in understanding the interaction between ␥␦ T cells and microbial pathogens such as M. tuberculosis.