Activation and Regulation of Blood Vδ2 T Cells Are Amplified by TREM-1⁺ during Active Pulmonary Tuberculosis

Tuberculosis (TB) is a chronic infectious disease caused by Mycobacteria tuberculosis. In its 2016 global TB report, the World Health Organization estimated that in 2015 there were 10.4 million new TB cases and 1.4 million people died of the disease (1). TB was thus declared as the leading cause of death from infectious diseases, and a great challenge to public health. Although M. tuberculosis infects approximately one third of the global population, only 10% of individuals develop active TB, demonstrating a crucial role of host immunity in the control of M. tuberculosis infection. The containment of M. tuberculosis infection largely depends on host cellular immunity mediated by both APCs and T cells. The inhaled infectious bacilli are phagocytized by APCs, which act as the first line of host defense against infection. APCs not only recognize microbial products of invading pathogen via germ-line encoding pattern-recognition receptors (PRRs) to trigger innate immune response, but more importantly, process Ags of ingested M. tuberculosis and present its peptides to T cells, leading to the induction of adaptive antimycobacterial immune response. Upon activation by the APCs, naive CD4⁺ Th cells differentiate into distinct effector subsets depending on the microenvironmental cytokines present during activation (2, 3). Adaptive immunity to M. tuberculosis infection is characterized by the generation of Ag-specific CD4⁺ Th1 cells (4). Effector Th1 cells recruited to the primary site of infection secrete IFN-γ, which in turn activates macrophages to kill intracellular mycobacteria (5). Although the process of Ag presentation is crucial for CD4⁺ T cell activation and proliferation, little is known about the mechanisms controlling Ag presentation.

Macrophage and dendritic cells (DCs) are essential for Ag presentation during TB infection and have been considered professional APCs. In addition, during infection or inflammation, other cells also display an APC function, which are usually called nonprofessional APCs (6, 7). These cells include fibroblasts, epithelial cells, and endothelial cells. Recently, research identified and characterized the innate-like human Vγ9Vδ2 TCR (also referred to as Vδ2 T cells), a subset of human γδ T cells, which can perform an APC functions (8). In peripheral blood, Vδ2 constitute the majority (>80%) of circulating human γδ T cells (9). It has been reported that Vδ2⁺ T cells specifically recognize the small nonpeptide Ags, mostly derived from microbes or necrotic host cells via their receptors (10, 11). When they encounter pathogens, Vδ2⁺ T cells can efficiently take them up via phagocytic receptors, then degrade them into peptide in a lysosome-dependent manner (12). Peptide-primed Vδ2 T cells can efficiently promote the differentiation of naive CD4⁺ and CD8⁺ T cells into effector cells, and induce robust proliferation of memory T cells (8). Also, Vδ2 T cells can induce maturation of DCs and B cells by differentially producing cytokines, and triggering alloreactive T cell stimulation (13). However, the mechanism(s) by which the Vδ2 T cells execute and control the Ag presentation activity are still not well understood. Recent evidence demonstrates that innate immune PRRs such as TLRs also regulate Vδ2 T cell activation (14, 15). Nevertheless, nothing is known about the expression and function of other PRRs in Vδ2 T cells. It is reported that in myeloid cells, the TLR signaling is closely related to the triggering receptor expressed on myeloid cells (TREM) family. For example, TREM-1 can amplify the TLR signaling (16), whereas TREM-2 suppresses TLR activation and downstream cytokine production during infection and inflammation (17, 18).

TREM-1 is a subset of the TREM family of cell-surface receptors, belonging to the Ig superfamily (19). It is a 30 kDa glycoprotein receptor containing an ectodomain of the Ig-like V-type, a transmembrane part, and a short intracytoplasmic domain (20). It is a well-characterized member of the TREM family, which is expressed broadly on myeloid cells including neutrophils, macrophages, and DCs (21). Its expression on phagocytes leads to a cascade of intracellular events that result in inflammatory effects, and is specifically upregulated by microbial products (22). TREM-1 has recently emerged as an important regulator of diverse cell processes, including the inflammatory response, immune homeostasis, and tumor development. During TREM-1 activation, the secretion of several key proinflammatory mediators, such as IL-1, IL-6, and TNF, is promoted via signal transduction molecule DAP12 (23). In addition, activation of TREM-1 enhances the expression of Ag-presentation molecules, which primarily promote monocyte differentiation into immature DCs (24). Blockade of TREM-1 by administering soluble forms of TREM-1, or RNA silencing in animal models with septic shock, resulted in a significant decrease of inflammatory response after stimulation with microorganisms, which effectively prolonged the survival of the animals (23). Nevertheless, the expression and function of TREM-1 in nonmyeloid cells including Vδ2 T cells remains undefined.

In this study, we explored the expression and function of TREM-1 on Vδ2 T cells (hereafter referred to as TREM-1⁺Vδ2 T cells) during TB. We demonstrated that a substantial proportion of TREM-1⁺Vδ2 T cells were induced in the peripheral blood of patients with active TB. Furthermore, the TREM-1⁺Vδ2 T cell subset had a high expression of costimulatory molecules and classical MHC class II (MHC-II) Ag-presenting molecules. They displayed highly effective Ag-presentation activity in vitro for both memory and naive CD4 αβ T cells. Of note, these cells also acted as phagocytes and had a strong ability to kill intracellular mycobacteria, therefore may be useful as a therapeutic target.

The cohort study was approved by the institutional review boards of Guangzhou blood center, Guangzhou Chest Hospital, and Sun Yat-sen University prior to commencement of the study. All participants provided written informed consent for their participation in the study (only adult participants were enrolled).

Whole blood from registered healthy blood donors (n = 44), was collected from Guangzhou blood center (Guangzhou, China). The active TB patients were recruited from Guangzhou Chest Hospital (Guangzhou, China). A total of n = 105 patients were screened and selected as active pulmonary TB. Eligible participants were adults over 18 y of age, who were serologically confirmed HIV-negative. All patients were TB treatment-naive at the time of recruitment. The exclusion criteria included patients with a known pulmonary disease (pneumonia, cancer, and those under chemotherapy or TB therapy). Detailed clinical characteristics and laboratory information are shown in Table I. All study participants in this cohort were recruited at the same time.

The cell staining procedure used in this study was previously described by Zhan et al. (25) with slight modifications. For intracellular cytokine staining, cells were restimulated for 4–6 h with 50 ng/ml PMA (Sigma, MO), 1 μg/ml ionomycin (Sigma-Aldrich), and 3 μg/ml brefeldin A (BFA) (eBioscience, CA). Intracellular cytokines were stained using the intracellular fixation/permeabilization buffer set (eBioscience, CA). Flow cytometric analysis was performed on a FACSCanto II (BD, NJ), and data were analyzed using FlowJo software (Tree Star). Abs were purchased from BD Biosciences, eBioscience, or BioLegend (CA) except where otherwise indicated. Anti-human: CD3 (Clone UCHT1; BD), CD3 (Clone UCHT1; BioLegend), Vδ2 TCR (Clone B6; BD), Vδ2 TCR (Clone B6; BioLegend), TCRγδ (Clone B1; BD), TCRαβ (Clone IP26; eBioscience), CD4 (Clone L200; BD), CD8 (Clone RPA-T8; BD), CD45RA (Clone HI100; BD), CD45RO (Clone UCHL1; BD), CD27 (Clone M-T271; BD), CD86 (Clone IT2.2; eBioscience), CD80 (Clone 2D10.4; eBioscience), HLA-DR (Clone LN3; eBioscience), TNF (Clone MAb11; eBioscience), TNF (Clone MAb11; BD), IFN-γ (Clone 4S.B3; eBioscience), IFN-γ (Clone B27; BD), granzyme B (Clone GB11; BD), Foxp3 (Clone 236A/E7; eBioscience), CD69 (Clone FN50; eBioscience), CCR7 (Clone 150503; BD), CD62L (Clone DREG-56; BD), CD44 (Clone G44-26; BD), CD25 (Clone BC96; eBioscience), TREM-1 (Clone 198015, conjugated to PE; R&D Systems). The gating strategy for TREM-1⁺Vδ2 T cells staining is shown in Supplemental Fig. 1A.

Human PBMCs were used to isolate γδ T cells, CD14^high cells, and bulk CD4⁺ αβ T cells by positive or negative selection using the magnetic cell sorting system from BD Biosciences. CD3⁺ T cell subsets were purified by positive selection using anti-human CD3 magnetic particles (BD). Human TREM-1⁺Vδ2 T cells were sorted with a phenotype as CD3⁺TREM-1⁺Vδ2 TCR⁺ in FACSAria (BD Biosciences). Human naive CD4⁺ T cells were sorted by negative selection using a human naive CD4⁺ T cell enrichment set (BD Biosciences). Purity of isolated cells was confirmed by analysis showing more than 95% were in the CD3⁺CD4⁺CD45RA⁺ subset. The purity of sorted Vδ2 T cells, and TREM-1⁺ or TREM-1⁻ Vδ2 T cell was analyzed by gating CD3 versus Vδ2 TCR, followed by analyzing Vδ2 TCR-positive cells versus TREM-1 using flow cytometry (Supplemental Fig. 1B, 1C).

CD14^high cell subsets were purified by positive selection using anti-human CD14 magnetic particles (BD). DCs were obtained from CD14⁺ monocytes after being cultured with GM-CSF and IL-4 for 5 d as previously reported (8).

The M. tuberculosis purified protein derivative (PPD) was added to TREM-1⁺ or TREM-1⁻ Vδ2 T cells that were sorted from fresh peripheral blood PBMCs for 24 h. Then, for 24 h PPD was added to immature DCs that were matured during the final 8 h with LPS. These APCs were irradiated (γδ T cells 10–12 Gy and DCs 40 Gy), extensively washed, and used to coculture with responder cells. Naive CD4 T cells, CD4 T cells or bulk CD3 T cells (responder cells) were stained with 1 uM of CFSE (Invitrogen) and cultured in 96-well flat-bottom plates (1 × 10⁵ cells per well). APCs and responder cells were added at a ratio of 1:10. We calculated the percentage of divided responder T cells by gating on CD4⁺ cells.

To study the effect of chloroquine (CQ) or BFA on Ag processing, both TREM-1⁺ and TREM-1⁻ Vδ2 T cells were incubated together with PPD in the presence of 80 μM CQ or 1 μg/ml BFA. Following washing and irradiation, CQ- or BFA-treated γδ T cells were tested for induction of proliferation responses. MHC-II restriction was studied by preincubating TREM-1⁺ or TREM-1⁻ Vδ2 T cells with a blocking Ab against HLA-DR (Clone L243; BioLegend) or control IgG2a before mixing with responder cells.

TREM-1⁺ and TREM-1⁻ Vδ2 T cells were sorted from fresh peripheral blood PBMCs, then challenged with dextran labeled with FITC (1 μM; Invitrogen) or M. bovis bacillus Calmette-Guérin (BCG) labeled with Texas Red at a multiplicity of infection (MOI) of 10. After 30 min of incubation, cells were washed three times with cold PBS, and centrifuged to remove extracellular bacteria. Thereafter, cells were collected and analyzed using a FACSCanto II (BD Biosciences).

Immunofluorescence staining was performed as previously described (26). TREM-1⁺ or TREM-1⁻ Vδ2 T cells grown on cover slips were infected with Texas Red–labeled M. bovis BCG at an MOI of 25 for 1 h. Then cells were fixed with 4% paraformaldehyde followed by membrane permeabilization using 0.2% Triton X-100. After blockade with 5% BSA, cells were sequentially incubated with Lyso Tracker (Invitrogen). Thereafter, slices were visualized with a confocal microscope (LSM710; Zeiss Axiovert).

TREM-1⁺ and TREM-1⁻ Vδ2 T cells were sorted from fresh PBMCs, before they were challenged with M. bovis BCG at an MOI of 25. The number of internalized and killed bacteria was assessed following 1 h of incubation at 37°C in 5% CO₂. After incubation, cells were washed with PBS three times and lysed with 0.1% Triton-X, and bacterial CFU were determined by plate count. The phagocytosis efficiency was calculated by CFU data at 1 h postinfection and normalized to the control group.

To assess the intracellular killing capacity of phagocytic cells, a second series of assays was run in parallel to determine the number of viable bacteria following 12, 24, and 48 h of incubation. Cells were lysed with 0.1% Triton-X and then serially diluted, spread onto 7H10 agar plates, and incubated at 37°C in 5% CO₂ for 2 wk. Live bacteria were counted and results are presented as the percentage of live bacteria 1 h postinfection for phagocytosis.

Total RNA was extracted from cultured cells with TRIzol (Invitrogen) according to the manufacturer’s instructions. The methods for cDNA generation and real-time PCR were used as previously described (27).

Data analyses were performed in GraphPad Prism 5.0 Software (San Diego, CA). Statistical significance was determined with Kruskal–Wallis or Mann–Whitney U test for nonparametric tests, and with ANOVA or Student t test analyses for parametric tests. Data are shown as mean ± SD unless otherwise. A p value < 0.05 was regarded as statistically significant.

To determine whether TREM-1 is expressed in lymphocytes, we analyzed the expression level of TREM-1 in CD4⁺/CD8⁺ T cells and CD19⁺ B cells by flow cytometry in human PBMC from healthy participants and active pulmonary TB patients. TREM-1⁺ cells had no expression in CD4⁺/CD8⁺ T cells and CD19⁺ B cells from healthy controls and active TB patients (Fig. 1A). We analyzed the expression level of TREM-1 in Vδ2 T cells by flow cytometry in human PBMC from healthy participants (n = 44) and active pulmonary TB patients (n = 105) (Table I). TREM-1⁺ cells were gated on CD3 and Vδ2 TCR-labeled cells. We found that TREM-1 was highly expressed in human peripheral blood Vδ2 T cells of active TB patients when compared with the healthy controls (Fig. 1B). Next, we analyzed the percentage of TREM-1⁺Vδ2 T cells. We found that circulating Vδ2 T cells in active TB patients showed a high frequency of TREM-1⁺Vδ2 T cells (42%, 42.34 ± 2.40). In contrast, TREM-1⁺Vδ2 T cells represented ∼7% (7.20 ± 0.80) of all circulating Vδ2 T cells in healthy donors (Fig. 1B). Furthermore, when purified Vδ2 T cells from healthy control participants were infected with M. bovis BCG or stimulated with IFN-γ or TNF, we found slight increase in the expression of TREM-1 in these cells when compared with the unstimulated cells (data not shown). Together, these results demonstrate that TREM-1 is expressed on circulating Vδ2 T cells and increased in active TB patients.

We first investigated the phenotype of Vδ2 T cells expressing TREM-1, where we found both TREM-1⁻Vδ2 and TREM-1⁺Vδ2 T cells expressed memory marker CD45RO as shown in the dot plot (Fig. 2A). However, the percentage of CD45RO in TREM-1–expressing Vδ2 T cells was significantly increased in the TREM-1⁺Vδ2 T cell population compared with the TREM-1⁻Vδ2 T cell population (Fig. 2A). In addition, a highly significant level of CD25 was observed in the TREM-1⁺Vδ2 T cell population compared with TREM-1⁻Vδ2 T cells (Fig. 2A). Because Vδ2 T cells have been shown to have an Ag cross-presenting ability, suggesting that they express APC surface markers (12), we next explored whether TREM-1⁺Vδ2 T cells display any features of professional APCs, then further examined peripheral human blood Vδ2 T cells of human subjects. An upregulation of the Ag-presenting MHC-II surface molecule (HLA D locus-related protein) HLA-DR was observed on TREM-1⁺Vδ2 T cells compared to TREM-1⁻Vδ2 T cells. Because intracellular HLA-DR is required for the MHC-II–peptide complex, we tested the total HLA-DR expression in TREM-1⁺Vδ2 and TREM-1⁻Vδ2 T cell populations of active TB patients, by staining for both surface and intracellular molecules. Surface versus intracellular staining revealed de novo that the production of HLA-DR was significantly higher (60.56 ± 2.00) in the TREM-1⁺Vδ2 T cell population compared with the TREM-1⁻Vδ2 T cell population (18.81 ± 1.88) (Fig. 2C, 2E). Moreover, a wide array of T cell costimulatory molecules, and a receptor that plays an important role in the organization of the primary immune response, the lymph node –homing receptor CCR7 were also upregulated in TREM-1⁺Vδ2 T cells compared with TREM-1⁻Vδ2 T cells (Fig. 2C). The mean fluorescence intensity (MFI) of HLA-DR was significantly upregulated in TREM-1⁺Vδ2 T cells compared to TREM-1⁻Vδ2 T cells (Fig. 2D). The MFI of the relative contribution of costimulation to the efficiency of Ag presentation function of TREM-1⁺Vδ2 T cells was analyzed. We found that CD40, CD80, and CD86 MFI were significantly upregulated in TREM-1⁺Vδ2 T cells compared to their TREM-1⁻Vδ2 T cell counterparts (Fig. 2D). We also analyzed the MFI of lymph node–homing receptor CCR7, which was similarly more upregulated with the HLA-DR molecule in TREM-1⁺Vδ2 T cells compared with TREM-1⁻Vδ2 T cells.

Altogether, these results indicate that TREM-1⁺Vδ2 T cells could be a subset of activated APCs that during active TB, if activated, can express significant amounts of Ag-presentation and costimulation molecules.

Because TREM-1⁺Vδ2 T cells display an APC-like phenotype, we assessed the direct role of TREM-1 in the regulation of APC molecules. Vδ2 T cells sorted from circulating blood were cultured in plates coated with anti–TREM-1 agonistic mAb or isotype control (rat IgG1), and the expression of surface markers (HLA-DR, CD40, CD80, CD86, and CCR7) was assessed by flow cytometry. Vδ2 T cells that were activated by TREM-1 mAb expressed higher levels of HLA-DR, CD40, CD80, and CD86 than those treated with IgG (Fig. 3A, 3B). The expression level of CCR7 was unchanged (Fig. 3A, 3B). Next, we cultured PBMCs in plates coated with TREM-1/Fc fusion protein or isotype control (human IgG) in the presence of LPS stimulation. We found that Ag presentation in the TREM-1/Fc–treated PBMCs was affected as seen by significant decreases in the expression of HLA-DR, CD40, CD80, and CD86. Furthermore, the treatment of PBMC with TREM-1/Fc did not affect the CCR7 expression (Fig. 3C, 3D). To further explore the APC function of TREM-1⁺Vδ2 T cells, we analyzed the expression of CD11b, CD18, and CD54 in TREM-1⁺Vδ2 T cells. Flow cytometry analysis demonstrates a significant increase in the expression of CD11b, CD18, and CD54 on TREM-1⁺Vδ2 T cells when compared with TREM-1⁻Vδ2 cells (Supplemental Fig. 2). Together, these data show that the Ag presentation molecules in Vδ2 T cells are directly promoted by TREM-1.

To determine whether TREM-1⁺Vδ2 T cells induce primary CD4 T cell responses, we used M. tuberculosis PPD in Ag presentation assays. TREM-1⁺Vδ2 T cells or monocyte-derived DCs were loaded with PPD for 24 h, washed, and then cocultured with autologous CFSE-labeled responder T cells for 5 d. The percentage of proliferated responder cells was assessed by the reduction in CFSE signals. Purified naive CD4⁺ T cells cocultured with TREM-1⁺Vδ2 T cells showed clear proliferation responses (Fig. 4A). All naive T cells stimulated by TREM-1⁺Vδ2 T cells were able to upregulate CD45RO expression (Fig. 4A). When assessing monocyte-derived DCs, a type of professional APC used as positive control, we found that responder cells induced by TREM-1⁺Vδ2 T cells had a low proliferating index compared with the DCs (Fig. 4A, 4C). As expected, the APC-dependent responses were also obtained with bulk CD3⁺ T cells as responder cells (Fig. 4A). Next, we compared the capacity of TREM-1⁺ and TREM-1⁻Vδ2 T cells in inducing CD4 T cell responses. Limited proliferation and lower CD45RO expression were observed in purified naive CD4⁺ T cells of TREM-1⁻Vδ2 T cells when compared with TREM-1⁺Vδ2 T cells (Fig. 4B, 4C). Consistently, TREM-1⁻Vδ2 T cells induced less proliferation response of bulk CD3 T cells when compared with TREM-1⁺Vδ2 T cells (Fig. 4B). We also tested the proliferation of TREM-1⁻Vδ2 and TREM-1⁺Vδ2 T cells in vitro loaded with PPD in different concentrations, and we found that TREM-1⁺Vδ2 T cells induced the proliferation of responder CD4 T cells to a greater extent than their TREM-1⁻Vδ2 T cell counterparts (Fig. 4D).

We continued with the examination of proliferation-naive CD4 T cells by sorting Vδ2 T cells from circulating blood and culturing in plates coated with anti–TREM-1 agonistic mAb or isotype control (rat IgG1) in the presence of PPD for 24 h. Next, cultured cells were washed and then cocultured with autologous CFSE-labeled CD4 T responder cells for 5 d. Vδ2 T cells stimulated with TREM-1 mAb substantially increased proliferation of responder cells compared with those coated with isotype control (Fig. 4E, 4F); Vδ2 T cells treated with TREM-1/Fc fusion protein to block TREM-1 signals partially inhibited the capacity to stimulate naive CD4 T cells or CD4⁺ T cells when compared with the human IgG-coated plate (Fig. 4G, 4H).

Published studies have shown that during the primary immune response, naive CD4⁺ T cells differentiate into Th1 cells with the capability to produce IFN-γ (28, 29). Because IFN-γ–producing Th1 cells play a crucial role in modulating the immune response during TB, we thus examined whether naive CD4⁺ T cells are differentiated into effector Th1 cells in response to PPD-presenting TREM-1⁺Vδ2 T cells. At an APC/responder ratio of 1:10, naive CD4 T cells stimulated by TREM-1⁺Vδ2 T cells for 7 d displayed a Th1-polarized response, with 12% IFN-γ⁺ cells, which was significantly higher than that of TREM-1⁻Vδ2 T cells, but considerably lower than the Th1-polarized cells stimulated by DCs (Fig. 4I). Next, we treated PBMC from active TB patients with (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate to assess the expression of proinflammatory cytokines; we found an increased percentage of TNF and IFN-γ in TREM-1⁺Vδ2 T cells compared with TREM-1⁻Vδ2 T cells (Supplemental Fig. 3A). Furthermore, we evaluated whether the proliferation of CD4 T cells depends on Ag presentation or cytokine expression. The data show that blocking TNF or IFN-γ had little effect on the proliferation of CD4⁺ T cells (Supplemental Fig. 3B–D), suggesting that the proliferation of CD4⁺ T cells by TREM-1⁺Vδ2 T cells depends on Ag presentation, not cytokines. In addition, we found that few CD25⁺Foxp3⁺ regulatory T cells were generated after coculturing naive CD4 T cells with either TREM-1⁺Vδ2 T cells or DCs (data not shown). Overall, these data demonstrate that TREM-1⁺Vδ2 T cells could efficiently induce the proliferation and differentiation of naive CD4 T cells in vitro just like DCs, which was due to Ag presentation as demonstrated by the flow cytometry data.

Thus, TREM-1 signaling enhances Ag cross-presentation activity of Vδ2 T cells. Data further indicate that TREM-1⁺Vδ2 T cells induce Th1 cell differentiation, which may promote cellular immunity to M. tuberculosis infection.

Studies have demonstrated that Ags and their proteins are broken down into peptides within acidified lysosomes (30–32). To examine whether acidification of endosomes and lysosomes was involved in the MHC-II pathway process of TREM-1⁺Vδ2 T cells, we treated TREM-1⁺Vδ2 T cells with CQ to block endosomal and lysosomal acidification in the presence of PPD Ag. Indeed, CQ efficiently prevented responder cell proliferation induced by either TREM-1⁺ or TREM-1⁻Vδ2 T cells, suggesting that endosomal and lysosomal acidification are essential for intracellular processing of PPD in TREM-1⁺Vδ2 T cells (Fig. 5A, 5B). Next, we tested the ability of BFA, which selectively targets the trans-Golgi network, and thus inhibits the classical protein export-dependent MHC-II pathway (12). We found that treatment of TREM-1⁺Vδ2 T cells with BFA resulted in the impairment of TREM-1⁺Vδ2 T cells to export Ag during the response induction to CD4⁺ T cells (Fig. 5A, 5B). Furthermore, we used a neutralizing Ab HLA-DR to completely block the cross-presentation mechanisms in TREM-1⁺Vδ2 or TREM-1⁻Vδ2 T cells. Results also demonstrate a complete failure in the processing of exogenous peptide by TREM-1⁺Vδ2 or TREM-1⁻Vδ2 T cells after neutralization of lysosome acidity (Fig. 5C, 5D). Data suggest that the Ag-presentation mechanism in TREM-1–expressing Vδ2 T cells requires lysosomes, and involves MHC-II molecules.

Phagocytosis is the process by which cells recognize and engulf solid particles to form internal vesicles known as phagosomes. This process is vital to host defense mechanisms as well as to tissue homeostasis, repair, and morphogenetic remodeling (33, 34). To explore the mechanism by which TREM-1⁺Vδ2 T cells take up Ag, we compared the expression of phagocytic receptors in TREM-1⁺Vδ2 T cells with TREM-1⁻Vδ2 T cells and monocytes. We found that TREM-1⁺Vδ2 T cells displayed higher expression of the phagocytic receptor CD16 (also called FcγRIII) than their TREM-1⁻Vδ2 T cell counterparts (Fig. 6A). Receptors involved in the phagocytosis were also more abundantly expressed in TREM-1⁺Vδ2 T cells than in TREM-1⁻Vδ2 T cells (Fig. 6A). However, when comparing TREM-1⁺Vδ2 T cells with the monocytes, no significant difference was observed (Fig. 6B).

Next, we stimulated the Vδ2 T cells with either IgG or TREM-1 agonistic mAb. CD16 expression was significantly increased in Vδ2 T cells that were stimulated with anti–TREM-1 (Fig. 6B). We wanted to know whether inhibiting TREM-1 signals with TREM-1/Fc–blocking peptide affects the phagocytosis process. We found reduced expression of CD16 in TREM-1/Fc–treated Vδ2 T cells, suggesting that TREM-1 is involved in the regulation of phagocytic receptors (Fig. 6C). We further examined the phagocytosis process of dextran and M. bovis BCG in Vδ2 T cells after being activated by TREM-1 mAb. The flow cytometry data revealed ingestion of FITC-labeled dextran and live Texas Red–labeled M. bovis BCG were significantly (p < 0.05) increased after treatment with TREM-1 mAb as compared with isotype controls (Fig. 6D, 6E). Again, when we treated the sorted TREM-1⁺Vδ2 T cells and TREM-1⁻Vδ2 T cells with either dextran or M. bovis BCG, surprisingly we observed that ∼90% of the TREM-1⁺Vδ2 T cells from active TB patients were able to ingest FITC-labeled dextran, and only 20% of TREM-1⁻Vδ2 T cells were phagocytic (Fig. 6F). Interestingly, TREM-1⁺Vδ2 T cells ingested more M. bovis BCG than did TREM-1⁻Vδ2 T cells, where over 45% of TREM-1⁺Vδ2 T cells with associated fluorescence contained internalized bacteria; whereas TREM-1⁻Vδ2 T cells ingested a similar percentage to with dextran (Fig. 6G). Next, we examined whether TREM-1 affects the formation of lysophagosomes. TREM-1⁺ and TREM-1⁻Vδ2 T cells were infected with Texas Red–labeled M. bovis BCG for 1 h, and then stained with Lyso Tracker, a fluorescent acidotropic probe for lysosomes. Confocal microscopy was applied to visualize the formation of M. bovis BCG containing lysophagosomes. The results showed increased colocalization of Texas Red–labeled M. bovis BCG (red) and lysosomes (green) in TREM-1⁺Vδ2 T cells, demonstrating increased number of lysophagosomes containing M. bovis BCG in TREM-1⁺Vδ2 T cells (Fig. 7A). To further explore the survival rate and mycobactericidal activity of TREM-1⁺Vδ2 T cells, we assessed the intracellular live M. bovis BCG in TREM-1⁺ Vδ2 T cells. The bacterial survival rate is presented as a percentage of live bacteria in either TREM-1⁺Vδ2 T cells or TREM-1⁻Vδ2 T cells at different time points. As we expected, due to the different phagocytic capability of each cell type, the number of live bacteria at 1 h postphagocytosis in TREM-1⁺Vδ2 T cells was ∼80% higher than in their TREM-1⁻Vδ2 T cell counterparts (Fig. 7B). The survival of intracellular bacteria in TREM-1⁺Vδ2 T cells or TREM-1⁻Vδ2 T cells decreased in a time-dependent manner. After 48 h, <20% of the phagocytized bacteria in TREM-1⁺Vδ2 T cells survived (Fig. 7B). Subsequently, we analyzed the killing rate of intracellular bacterial at 24 h postinfection, and found that TREM-1⁺Vδ2 T cells killed 65% of intracellular bacteria, whereas TREM-1⁻Vδ2 T cells killed 40% (Fig. 7C). Taken together, these data indicated that TREM-1⁺ promoted phagocytic ability in Vδ2 T cells by upregulating CD16 expression, and suggest that TREM-1⁺Vδ2 T cells have potent microbicidal abilities.

The host cell’s ability to engulf M. tuberculosis and present it to the T cells is perhaps the most vital determinant of innate and adaptive immune response efficiency during disease. The major innate effectors cells include professional phagocytes (monocytes, macrophages, neutrophils, and DCs) and NK cells (35). These cells display a variety of surface receptors and molecules that recognize either pathogens or endogenous molecules that are expressed during tissue damage (35). γδ T cells represent a distinct subset of T cells characterized by TCRs, and possess unique structural and Ag-binding characteristics (36–39). However, studies have demonstrated increased (40, 41) and decreased (42–47) numbers of γδ T cells in patients with active TB. The γδ T cell subset, which has been recently proposed as a nonclassical APC, recognizes peptide Ags derived from microbes or necrotic tissues (10, 11). Stimulation of Vδ2 T cells by nonpeptidic phosphoantigens from bacilli activates the APC function to increase stimulation of restricted T cell responses. Vδ2 T cells play a role in the induction of professional APC function and initiation of adaptive immune processes (8). Human blood Vδ2 T cells have also been reported to rapidly and substantially expand in response to microbial infections (10, 11, 38, 39). A potent base for first line of defense and manipulation of the adaptive response against pathogens has been reported, whereby Vδ2 T cells and DCs nurture the activation of each other after being activated by TLRs (14).

Previous studies have reported that TLRs are expressed in activated Vδ2 T cells, and can act directly in combination with TCR stimulation to enhance the production of cytokines and chemokines (48). However, the function of other PRRs in Vδ2 T cells remains unknown. In the current study, we showed that TREM-1, a receptor found in myeloid cells, is widely expressed in human Vδ2 T cells and subsequently enhances their professional Ag-presentation function. In this study, we found that TREM-1⁺ but not TREM-1⁻ was able to exercise the above reported function in Vδ2 T cells. TREM-1⁺ expression was found to be lower in Vδ2 T cells from healthy donors, but significantly higher in patients with active TB. Interestingly, the TREM-1⁺Vδ2 T cells obtained from TB patients can be considered as having an activated status due to the high expression levels of CD45RO and CCR7. Also, when Vδ2 T cells obtained from healthy controls were infected or stimulated ex vivo with BCG, IFN-γ, or TNF, we found that these cells had a slight increase in TREM-1 expression in all the tested conditions.

Studies have shown that during infectious diseases, Vδ2 T cells are activated and undergo proliferation, cytokine/chemokine secretion, as well as cell-surface marker expression (10, 49– 51). Our findings demonstrate that human peripheral blood TREM-1⁺Vδ2 T cells but not TREM-1⁻Vδ2 T cells had an increased expression of the Ag-presentation molecule HLA-DR, and costimulation molecules including CD40, CD80, and CD86 during active TB. Using the PPD Ag-presentation model, again TREM-1⁺Vδ2 T cells efficiently induced proliferation and differentiation of CD4 T cells. Although the effects observed in TREM-1⁺Vδ2 T cells are slightly weaker than in DCs, they were much stronger than in TREM-1⁻Vδ2 T cells. Surprisingly, the Ag-presentation activity in TREM-1⁺Vδ2 T cells was dependent also on lysosomal acidification and de novo synthesized MHC-II molecules. Because the short cytoplasmic tail of TREM-1 lacks signaling motifs, the initiation of TREM-1 signaling depends on its association with adaptor molecule DAP12, which contains an ITAM (52). Intracellular pathways elicited by TREM-1 result in the production of cytokines and chemokines and upregulation of costimulatory molecules, which may be regulated by different pathways, including Akt, ERK1/2, and NF-κB (21, 53). However, cross-linking of TREM-1 with agonist Ab did not enhance the production of IFN-γ and TNF in Vδ2 T cells of patients with TB (data not shown). This may indicate that soluble cytokines may not be involved in the Ag-presentation activity of TREM-1⁺Vδ2 T cells. Treatment of PBMC with the agonist TREM-1 mAb promoted the Ag-presentation capacity of Vδ2 T cells, which was associated with the upregulation of costimulation molecules. Bouchon et al. (19) reported that TREM-1 signaling increased the expression of costimulatory molecules such as CD86 and MHC-II.

According to flow cytometry analysis, our data show that CD4αβ T cell response was induced in vitro by activated Vδ2 T cells, which is similar to those observations found in mature DCs in vitro (8). We demonstrate that TREM-1⁺Vδ2 T cells have a weaker capacity than mature DCs to induce proliferation of CD4 T cells, as well as Th1 polarization. This may be attributed to the relatively lower HLA-DR expression found in TREM-1⁺Vδ2 T cells than in DCs. Therefore, these findings suggest that TREM-1⁺–expressing Vδ2 T cells generated from patients with active TB were not equivalent to the in vitro activated Vδ2 T cells. Furthermore, TREM-1⁻Vδ2 T cells displayed lower expressions of HLA-DR and costimulation molecules than TREM-1⁺Vδ2 T cells, and have less capability to induce proliferation and differentiation of CD4 T cells. Previous studies suggested the absence of APC function in resting Vδ2 T cells (8). This indicates that the Vδ2 T cells analyzed in our TREM-1⁻Vδ2 T cell cohort were activated rather than resting Vδ2 T cells. Recent studies have shown that TREM-2–Fc fusion proteins impede TREM-2 recognition of apolipoproteins, which appear to impair ligand binding (54). Because this is a member of the TREM family, it is therefore worthwhile mentioning that the use of TREM-1–Fc is interesting and valuable, but not overly physiological. Also, it is unclear what effect the fusion of an Fc fragment with TREM-1 may have receptor (such as Fc receptor) cross-linking and stimulation on the same cell. In this study, unspecific Fc-binding was blocked by use of IgG, and the TREM-1–Fc fusion protein did not affect the CCR7 expression. To prevent the unspecific binding and blockage of CD16, which may prevent its detection in flow cytometry analysis, the unconjugated Abs for IgG and live/dead detection were included in the mixture staining.

As well-known nonclassical professional APCs, γδ T cells could take up and degrade exogenous soluble protein for peptide loading on MHC-I, and induce CD8 T cell response through Ag cross-presentation (12). Compared to monocyte-derived DCs, γδ T cells are more efficient in Ag cross-presentation, which has been suggested to be induced by delayed intracellular protein degradation and endosomal acidification (55). Therefore, we hypothesized that TREM-1⁺Vδ2 T cells may also induce a CD8 T cell response by Ag cross-presentation, which needs further investigation. In mammals, phagocytosis is accomplished mainly by professional phagocytes, but recent research showed that the earliest thymic progenitors for T cells still retain myeloid lineage potential (56). In this study, we demonstrate that TREM-1⁺Vδ2 T cells from patients with active TB are capable of ingesting large particles like M. bovis BCG, and were also involved in killing bacteria. Moreover, BCG vaccination of human adults has been shown to increase Vδ2 T cell responses (57), and Vδ2 T cells in vitro are activated by M. tuberculosis (58). The prolifically phagocytic γδ T cells express CD16, which is believed to mediate phagocytosis in humans (59). In this study, we find little or no expression of CD16 in the TREM-1⁻Vδ2 T cell population, whereas high expression was observed in TREM-1⁺Vδ2 T cells and monocytes. This suggests that TREM-1⁺Vδ2 T cells may be involved in the innate phase of immune responses. On the one hand, phagocytosis leads to Ag processing and presentation on MHC-II of peptides derived from the phagocytized complexes. On the other hand, phagocytosis in TREM-1⁺Vδ2 T cells leads to elimination of invading pathogens. TB and HIV coinfections have been reported to be associated with decreased Vδ2 T cell circulation and function (46). Depletion of these cells may contribute to significant further dissemination of M. tuberculosis in HIV-positive patients. Whether TB and HIV coinfection share a common mechanism causing TREM-1⁺Vδ2 T cell depletion still needs to be investigated.

In conclusion, we have attained a better understanding of the molecular mechanisms driving phagocytosis, as well as mycobacterial killing by Vδ2 T cells. Our findings shed new light on the activation and regulation of Vδ2 T cells in the phase of immune responses. Because some requirement for the right care for TB requires safe-acting drugs, TREM-1⁺ on Vδ2 T cells may serve as a therapeutic target.

We thank Lianqiang Feng and all members of the department of immunology at Zhongshan School of Medicine, Sun Yat-sen University for technical support.

The authors have no financial conflicts of interest.