BTN3A Targeting Vγ9Vδ2 T Cells Antimicrobial Activity Against Coxiella burnetii-Infected Cells

Vγ9Vδ2 T cells have been reported to participate to the immune response against infectious diseases such as the Q fever caused by Coxiella burnetii infection. Indeed, the number and proportion of Vγ9Vδ2 T cells are increased during the acute phase of Q fever. Human Vγ9Vδ2 T cell responses are triggered by phosphoantigens (pAgs) produced by pathogens and malignant cells, that are sensed via the membrane receptors butyrophilin-3A1 (BTN3A1) and -2A1 (BTN2A1). Here, by using CRISPR-Cas9 inactivation in THP-1 cells, we show that BTN3A and BTN2A are required to Vγ9Vδ2 T cell response to C. burnetii infection, though not directly involved in the infection process. Furthermore, C. burnetii-infected monocytes display increased BTN3A and BTN2A expression and induce Vγ9Vδ2 T cell activation that can be inhibited by specific antagonist mAb. More importantly, we show that the antimicrobial functions of Vγ9Vδ2 T cells towards C. burnetii are enhanced in the presence of an BTN3A activating antibody. This supports the role of Vγ9Vδ2 T cells in the control of C. burnetii infection and argues in favor of targeting these cells as an alternative treatment strategy for infectious diseases caused by intracellular bacteria.


INTRODUCTION
The role of Vg9Vd2 T cells in the host immune response to bacterial infection is now welldocumented (1). Human Vg9Vd2 T cells, which normally represent 2-5% of peripheral blood T cells, are expanded in infected patients to reach up to ≥50% of the circulating T cells (2,3), as reported for patients undergoing mycobacterial disease, listeriosis, salmonellosis, brucellosis, tularemia, legionellosis and Q fever (4)(5)(6)(7)(8)(9)(10). Furthermore, local expansion of Vg9Vd2 T cells have also been reported in the bronchoalveolar lavage fluids from patients with active pulmonary tuberculosis and in cerebral spinal fluids from patients with bacterial meningitis (11)(12)(13). Two direct antimicrobial actions of Vg9Vd2 T cells against various viruses, protozoa and bacteria were reported, including cytotoxic activity to pathogen-infected cells and a cell-mediated non-cytolytic activity based on cytokine production (1,(14)(15)(16). In vitro studies have shown that Vg9Vd2 T cells are able to effectively kill intracellular pathogens such as M. tuberculosis, L. monocytogenes, and Brucella suis (17)(18)(19)(20)(21).
The butyrophilin 3A1 (BTN3A1) cell surface molecule is involved in cell recognition and the human Vg9Vd2 T cells activation (22,23). Vg9Vd2 T cells are activated by small, phosphorylated nonpeptide antigens, called phosphoantigens (pAgs) (14). The production of these metabolites is increased in tumor or stressed eukaryotic cells, and can be naturally produced by several pathogens (11,24,25). Among the BTN3A isoforms (BTN3A1, BTN3A2, BTN3A3), BTN3A1 is unique in that its intracellular B30.2 domain binds to pAgs (26,27), while its juxtamembrane domain performs a critical function in homodimerization and heterodimerization of BTN3A (28). Conformational changes in the juxta-membrane domain, induced by the binding of pAgs to the B30.2 domain, are involved in Vg9Vd2 T cell activation (29). More recently, BTN2A1 has been identified as a novel actor in pAg sensing by Vg9Vd2 T cells (30)(31)(32). BTN2A1 is a direct ligand for the Vg9 TCR interacting with BTN3A1 to trigger Vg9Vd2 TCR activation (30).
Several evidences highlight the key role of Vg9Vd2 T cells in Q fever, an infectious disease caused by the intracellular bacterium Coxiella burnetii. (1) During the acute phase of the disease, the numbers and proportion of Vg9Vd2 T cells were found increased (2) with a significant increase of the expression of HLA-DR, but not CD25 (10). In this study, we investigated the functional role of Vg9Vd2 T cells and the involvement of BTN3A and BTN2A in host defense against C. burnetii. Here, we observed that C. burnetii infection of healthy monocytes lead to the increase of the expression of these two BTNs. Using a CRISPR-Cas9 knockout model in the THP-1 cell line, we observed that BTN3A and BTN2A are not directly involved in the infection process by C. burnetii but play a role in the host immune response to infection. We reported that infected monocytes induced Vg9Vd2 T cell activation in a BTN3A and BTN2A dependent manner. Finally, the use of a BTN3A activating antibody enhances the antimicrobial functions of Vg9Vd2 T cells against C. burnetii infected cells through the production of cytotoxic molecules and large amounts of IFN-g and TFN-a. Our results highlight the role of Vg9Vd2 T cells in the control of C. burnetii infection and the therapeutic potential of BTN3A activating antibody in infections.

Cell Isolation
Blood samples (leucopacks) were obtained from the local French Blood Establishment (Etablissement francais du sang, EFS), which carries out donor inclusions, informed consent, and sample collection. Through a convention established between our laboratory and the EFS (N°7828), buffy coats were obtained and peripheral blood mononuclear cells (PBMCs) were isolated as previously described (33). Monocytes were purified from PBMCs using anti-CD14-conjugated magnetic beads (Miltenyi Biotec, Bergisch Glabach, Germany) and cultured in Roswell Park Memorial Institute-1640 medium (RPMI, Life Technologies, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS, Gibco, Life technologies), 2 mM L-glutamine, 100 U/mL penicillin and 50 µg/mL streptomycin (Life Technologies).
Lentiviral Transduction and CRISPR-Cas9-Mediated BTN3A or BTN2A Knockout For all transductions, THP-1 cells were seeded in 12-wells plates (2.5x10 5 cells/well), and 25 mL of concentrated lentiviral particles were added to the culture. After 24 hours, cells were washed twice in complete medium, and cultured in their regular culture medium for 48 hours. Optimized CRISPR target sequences targeting the three BTN3A gene isoforms and for BTN2A gene inactivation, targeting both BTN2A gene isoforms (sequence available upon request) were cloned into the lentiCRISPR-v2 vector (Addgene #52961). For selection of THP-1 transductants, 1 mg/mL puromycin was added to the culture medium (Supplementary Figure 1).

Bacterial Production
Coxiella burnetii phase I (Nine Mile (NM) strain, RSA493 and Guiana strain, MST17) were cultured in L929 cells for 10 days, as previously described (36). Briefly, infected cells were sonicated and centrifuged at 10,000g for 10 minutes, then washed and stored at -80°C. Bacterial titers were determined using Gimenez staining, and bacterial viability was assessed using the Live/Dead BacLight bacterial viability kit (Molecular Probes, Eugene, OR, USA).
Mycobacterium tuberculosis (H37Rv strain) was cultured in Middelbrook 7H10 (Becton Dickinson, Le Pont de Claix, France) supplemented with 10% oleic acid-albumin-dextrosecatalase (OADC, Becton Dickinson), as previously described (37). Prior to infection, the colonies were resuspended in phosphate buffered saline (PBS, Life Technologies), vigorously vortexed for 10 min using 3 mm sterile glass beads (Sigma-Aldrich) and passed 10 times through a 25 G needle to disperse clustered cells. Calibration was performed at OD 580 nm and confirmed by counting mycobacteria after Ziehl-Neelsen staining.

Cell Infection
Monocytes isolated from healthy donors were infected with C. burnetii strains (50 MOI) or with M. tuberculosis (5 MOI). After 24 hours of infection, the expression of BTN3A and BTN2A were investigated by qRT-PCR and flow cytometry. For co-cultures experiments, monocytes isolated from healthy donors previously infected 24 hours with C. burnetii strains or with M. tuberculosis were co-cultured with autologous Vg9Vd2 T cells (E:T ratio of 1:1). After 4 hours of co-culture, Vg9Vd2 T cell degranulation and cytotoxicity was assessed by flow cytometry and the bacterial load was measured by flow cytometry and qPCR. Finally, the supernatants of the co-cultures were analyzed for the presence of cytokines and cytotoxic molecules by ELISA assay.
The presence of C. burnetii within cells was also assessed by flow cytometry. Briefly, infected cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 (Sigma-Aldrich). After washing, cells were incubated with a rabbit antibody directed against C. burnetii for 30 min and then with an Alexa 647 anti-rabbit antibody (Invitrogen). Data were collected on a BD Canto II instrument (BD Biosciences, Le Pont-de-Claix, France) and analyzed with FlowJo software (FlowJo v10.6.2, Ashland, OR).
For M. tuberculosis infected cells, DNA was extracted from infected cells as follows: aliquots of 150 mL were incubated overnight at 56°C with 150 mL of G2 buffer mixed with 15 mL proteinase K (20 mg/mL). After two cycles of mechanical lysis (45 s), the total DNA was extracted using the EZ1 DNA Tissue Kit (Qiagen). M. tuberculosis DNA detection was performed targeting the M. tuberculosis internal transcribed spacer (ITS) ( Table 1), as previously described (37).

RNA Isolation and q-RTPCR
Total RNA was extracted from infected cells (2x10 6 cells/well) using the RNeasy Mini Kit (Qiagen) with DNAse I treatment as previously described (38). RNAs quality and quantity were evaluated using a NanoDrop spectrophotometer (Nanodrop Technologies, Wilmington, USA). Reverse transcription was performed using M-MLV Reverse Transcriptase kit (Life Technologies) and oligo(dT) primers. The expression of genes characteristics of M1/M2 macrophage phenotypes, as well as BTN3A isoform genes, was evaluated using real time qPCR, Smart SYBR Green fast Master kit (Roche Diagnostics, Meylan, France) and specific primers ( Table 1). BTN2A levels expression was evaluated using real time qPCR, TaqMan ® Fast Advanced Master Mix (Applied Biosystems, Life Technologies) and specific probes ( Table 1). All qPCRs were performed using a CFX Touch Real-Time PCR Detection System (Bio-Rad, Marnes-la-Coquette, France). Results were normalized by the expression of ACTB or GAPDH housekeeping gene and are expressed as relative expression of investigated genes with 2 -DCt where DCt = Ct target -Ct housekeeping gene as previously described (36).

BTN3A and BTN2A Surface Expression
Cells were suspended in PBS (Life Technologies) containing 1% FBS and 2 mM EDTA (Sigma-Aldrich). Cells were labeled with viability dye (Live/Dead Near IR, Invitrogen), mouse anti-BTN3A (clone 103.2) or anti-BTN2A (clone 7.48) Abs or with the appropriate isotype control (Miltenyi Biotech). After 30 min incubation, primary antibody binding was detected with secondary PE anti-mouse antibody (Invitrogen) and data were collected on a Navios instrument (Beckman Coulter) and analyzed with FlowJo software (FlowJo v10.6.2).

Degranulation Assay
Monocytes were co-cultured with Vg9Vd2 T cells at effector-totarget (E:T) ratio of 1:1 in presence of mouse anti-BTN2A mAb (clone 7.48) or mouse anti-BTN3A mAb (clones 20.1 or 103.2) TABLE 1 | Primers used for the response to infection.

Cytotoxicity Assay
Monocytes were labeled with 10 µM Cell Proliferation Dye eFluor ® 670 (Invitrogen) and then co-cultured with Vg9Vd2 T cells at E:T ratio of 1:1 in presence of mouse anti-BTN3A mAb (clone 20.1) at the indicated concentrations. After 4 hours, cells were stained with CellEvent Caspase-3/7 Green (Invitrogen) to identify dead cells. The cytotoxicity was assessed by flow cytometry as the percentage of Caspase 3/7 + cells in the target cell population (Supplementary Figure 2). Data were collected on a BD Canto II instrument (BD Biosciences) and analyzed with FlowJo software (FlowJo v10.6.2).

Statistical Analysis
Statistical analysis was performed with GraphPad Prism (8.0, La Jolla, CA). After analysis of the distribution of the data with a normality test, the Mann-Whitney U test was used as a nonparametric test and the t test as a parametric test. Hierarchical clustering of gene expression was analyzed using the ClustVis webtool (39). The limit of significance was set up at p<0.05.

C. burnetii Infection Enhances Expression of BTN3A and BTN2A
To assess whether C. burnetii infection affected the expression of BTNs, monocytes from healthy donors were isolated and infected with the reference strain NM1 or with the Guiana strain, described to be more virulent (40,41). After 24 hours of incubation with active or heat-inactivated C. burnetii NM1 strain, increases of transcript expression of both BTN3A1 and BTN3A2 isoforms, but not of BTN3A3 were found. Guiana strain infection enhanced the expression of all three isoforms, similar to M. tuberculosis infection used as control ( Figure 1A) (23). Interestingly, significant differences of BTN3A1 expression were observed between cells infected with active or heatinactivated C. burnetii NM1 (p=0.0374), suggesting that virulence affected BTN3A1 expression. Indeed, inactivated form of C. burnetii are reported to induce a weaker modulation of the expression of the A1 isoform, the essential form for pAg-mediated activation of Vg9Vd2 T cells (23). Significant increase of BTN3A protein expression was found for monocytes infected with C. burnetii NM1 and Guiana strains (p=0.0021 and p=0.0096, respectively) ( Figure 1B). As BTN2A is involved in Vg9Vd2 T-cell activation (31), we also investigated whether C. burnetii infection affected its expression. After 24 hours of infection, BTN2A transcriptional expression for both isoforms (BTN2A1 and BTN2A2) was significantly increased after C. burnetii NM1 and Guiana infection (BTN2A1 p=0.0170 and p=0.0021, respectively; and BTN2A2 p=0.0054 and p=0.0463, respectively) compared to uninfected cells and without significant modulation compared to the heat-inactivated form ( Figure 1C). Regarding BTN2A protein expression, a significant increase was observed for C. burnetii infected monocytes (NM1 strain, p=0.0160; and Guiana strain, p=0.0018) compared to uninfected cells, as observed for M. tuberculosis infection as control ( Figure 1D).
Altogether, like M. tuberculosis infection, C. burnetii infection leads to increased expression of BTN3A and BTN2A in infected cells.
Involvement of BTN3A and BTN2A in C. burnetii Infection Next, we investigated whether BTNs could be involved in the uptake or replication of C. burnetii. For this purpose, we performed a CRISPR-Cas9 knockout of the three BTN3A genes or the two BTN2A genes in the THP-1 cell line. Cells were transduced with a guide targeting either BTN2A1 and 2A2 (BTN2AKO) or BTN3A1, 3A2 and 3A3 (BTN3AKO) isoforms or with an irrelevant CRISPR guide (mock). BTN3AKO, BTN2AKO and mock cells were infected with C. burnetii NM1, and the bacterial load was assessed by qPCR. No differences were observed concerning the bacterial load ( Figure 2A) and replication overtime ( Figure 2B) between BTN3AKO, BTN2AKO and mock cells, suggesting that BTN3A and BTN2A are not directly involved in the process of C. burnetii infection.

Involvement of BTN3A and BTN2A in the Inflammatory Response to C. burnetii Infection
We then investigated the involvement of BTNs in the host immune response following C. burnetii infection in THP-1 cells. As observed in the Figure 3, C. burnetii infection results in modulation of genes characteristics of both pro-inflammatory (TNF, IFNG, IL6, CXCL10, IL1RA and IL1B) and antiinflammatory (IL10, TGFVBand CD163) responses in THP-1 cells. Upon infection with C. burnetii, the hierarchical clustering based on the expression of the above mentioned genes revealed that BTNs expression correlated with the transcriptional response to infection, as depicted by a separate clustering of BTN3AKO/BTN2AKO cells and mock cells ( Figure 3A). Indeed, BTN3AKO and BTN2AKO cells displayed significantly decreased expression of inflammatory genes following C. burnetii infection, in particular that of TNF and IL1B ( Figure 3B). Also, IL6 transcriptional expression appear to be affected by the BTN3A KO (p=0.0862) but not by the BTN2A KO. Furthermore, the expression of IL10 transcript was significantly decreased compared to mock cells (p=0.0435) ( Figure 3B). Consistently, BTN3AKO and BTN2AKO cells presented a significant decrease in TNF and IL-1b release following C. burnetii infection compared to mock cells ( Figure 3C). No significant difference in the levels of antiinflammatory cytokines such as IL-10 and TGF-b was observed.
Taken together, these data reported that both BTN3A and BTN2A are involved in the inflammatory response to C. burnetii infection.

C. burnetii Infection Leads to Vg9Vd2 Cells Activation in a BTN3A and BTN2-Dependent Manner
Since BTNs appeared to be over-expressed in monocytes following C. burnetii infection, we hypothesized that it could enhance the Vg9Vd2 T cell activation. After 4 hours of co-culture with C. burnetii infected monocytes, Vg9Vd2 T cell displayed enhanced degranulation as depicted by increased membrane expression of CD107, which also increased with the titer of bacteria used for monocytes infection ( Figure 4A). We then investigated whether Vg9Vd2 T cell activation by C. burnetiiinfected cells was dependent on BTNs by using anti-BTN3A antagonist (clone 103.2) (26) and anti-BTN2A antagonist (clone 7.48) (30) antibodies. Both antibodies led to significant inhibition of Vg9Vd2 T cell degranulation against cells infected with C. burnetii NMI or Guiana strains, or M. tuberculosis as positive control, suggesting that both BTNs are involved in Vg9Vd T cell activation in an infectious context ( Figures 4B, C) as it was previously shown for malignant cells (30). Taken together, C. burnetii infection leads to Vg9Vd2 T cell activation in a BTN3A and BTN2A dependent manner.
We next hypothesized that Vg9Vd2 T cell activation towards C. burnetii-infected cells could be enhanced by an humanized BTN3A agonist antibody (clone 20.1) (26) that activates Vg9Vd2 T cells. As illustrated in the Figure 4D, we observed that the BTN3A activating antibody leads to increased expression of CD107 ( Figure 4D) and the cytotoxic activity ( Figure 4E) of Vg9Vd2 T cells towards C. burnetii infected monocytes as observed for M. tuberculosis after 4 hours of co-culture. A similar effect was observed for all C. burnetii strains, to the same extent as M. tuberculosis, suggesting that the 20.1 antibody can induce Vg9Vd2 T cell activation even towards virulent bacteria. These data show that targeting Vg9Vd2 T cells with the 20.1 antibody leads to the activation of their cytotoxicity against C. burnetii-infected cells.

Anti-BTN3A Agonist Antibody Increases Antimicrobial Activity of Vg9Vd2 T Cells
Since the anti-BTN3A agonist antibody (clone 20.1) increases Vg9Vd2 T cell activation, we wondered whether it was able to boost their antimicrobial activity. For this purpose, monocytes were infected with C. burnetii NM1 for 24 hours and then cocultured with Vg9Vd2 T cells for 4 hours in presence of 20.1 antibody (0, 0.1, 1 or 10 µg/ml) and the bacterial load was measured by flow cytometry and qRT-PCR. First, Vg9Vd2 T cells lead to a significant reduction of C. burnetii load from 5.10 7 to 6.10 6 in monocytes in the presence of Vg9Vd2 T lymphocytes (p=0.0021) ( Figures 5A, B), as observed for M. tuberculosis infection (Supplementary Figure 3). BTN3A activating antibody resulted in a dose-dependent decrease in C. burnetii load in monocytes, reaching from 6.10 6 to 4.2.10 6 (0 vs. 10 µg/ml,  p=0.0501) (Figures 5A, B). This effect is similar to that observed in the case of M. tuberculosis, where the 20.1 antibody resulted in a decrease in the bacterial load in monocytes (0 vs. 10 µg/ml, p=0.0158) (Supplementary Figure 3). Altogether, BTN3A activating antibody increases the antimicrobial activity of Vg9Vd2 T lymphocytes against monocytes infected with C. burnetii.

Anti-BTN3A Agonist Antibody Increases the Secretion of Cytokines and Cytotoxic Molecules by Vg9Vd2 T Cells
Since treatment with the anti-BTN3A agonist antibody leads to bacterial load reduction, we investigated whether this could be related to the secretion of cytokines and cytotoxic molecules, which are strongly produced by activated Vg9Vd2 T cells (17)(18)(19)(20)(21). Indeed,

DISCUSSION
An alteration of circulating Vg9Vd2 T cells has been observed in Q fever patients (10). During acute phase of the disease, the proportion of Vg9Vd2 T cells is significantly increased in patients (16% vs. 4% in healthy donors) (10), indicating the involvement of these cells in the acute immune response to C. burnetii. Since human Vg9Vd2 T cell responses are triggered via an interaction with the BTN2A1/BTN3A1 complex, we first assessed whether their expression was modulated following C. burnetii infection. We found that in vitro infection of monocytes with C. burnetii induced a significant increase in the Values represent mean ± standard deviation. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.
transcriptomic and plasma membrane expression of these two BTNs. This increase is similar to that observed with M. tuberculosis and between C. burnetii strains, suggesting that the aggressiveness of the bacteria appears to have limited impact on BTN expression. Similarly, increased expression of these two BTNs has recently been described in red blood cells infected by Plasmodium falciparum (42). Our team has recently shown higher expression of BTN3A, but not BTN2A, following SARS-CoV-2 infection of myeloid cells and lung cell lines (submitted manuscript). This may suggest different mechanisms depending on the pathogen. Using CRISPR-Cas9 gene inactivation in the THP-1 cell line, we found that BTN3A and BTN2A are not directly involved in the infection process of cells by C. burnetii but play a role in the cellular immune response to infection. Indeed, THP-1 cells inactivated for BTN3A or BTN2A show a repressed inflammatory response following C. burnetii infection, with a significant decrease in TNF and IL1B gene expression. These results suggest that higher expression of these two molecules on monocytes could favor responses to C. burnetii infection.
The fact that both BTN3A and BTN2A, essential for Vg9Vd2 T cell activation, are more expressed following C. burnetii infection could enhance their activation and antibacterial activity. Using a Vg9Vd2 T cell/infected monocyte co-culture model, we observed that monocytes infected with C. burnetii strains of different aggressiveness resulted in similar degranulation of Vg9Vd2 T cells. Several studies have confirmed that the activation of Vg9Vd2 T cells is dependent on BTN3A during infections. Indeed, the anti-BTN3A antagonist antibody 103.2 was able to inhibit the degranulation of Vg9Vd2 T cells when they were co-cultured with cells infected with M. bovis (BCG), L. monocytogenes, P. falciparum or Epstein-Barr virus (23,(42)(43)(44). In our study, similar results are obtained with the 103.2 antibody but are also observed with an anti-BTN2A antagonist antibody (clone 7.48), underlining the importance of these two BTNs in the activation of Vg9Vd2 T cells.
Next, we evaluated the effect of BTN3A on antibacterial activity. For this purpose, we used the anti-BTN3A agonist antibody 20.1 to treat Vg9Vd2 T cell/C. burnetii-infected monocyte co-cultures. Our results show that the 20.1 mAb increases the antibacterial activity of Vg9Vd2 T cells leading to a decreased intracellular load of C. burnetii. In our study, Vg9Vd2 T cells, whose cytotoxic activity is enhanced by the 20.1 mAb, were both able to kill C. burnetii-infected monocytes through the production of lytic granules (granulysin, perforin, granzymes) and at the same time produce large amounts of IFNg and TFN-a. These cytokines play an essential role in protection against intracellular bacteria by activating the antimicrobial machinery of phagocytes. Indeed, IFN-g induces C. burnetii killing by promoting apoptosis of infected monocytes (36,45), and TNF-a shows an essential role in the control of C. burnetii infection like for other pathogens including M. tuberculosis or L. monocytogenes (46,47). These data extend previous studies as human Vg9Vd2 T cells have already been shown to effectively kill intracellular pathogens, such as M. tuberculosis, L. monocytogenes and B. suis, through the secretion of IFN-g, TNF-a and cytotoxic molecules such as granzymes, perforin and granulysin (17)(18)(19)(20)(21)48). Some studies have also reported that NKG2D contributed to the anti-infective activity of Vg9Vd2 T cells against Brucella sp. and M. tuberculosis (49,50). In contrast, in other studies on M. tuberculosis or L. monocytogenes, NKG2D was not involved (20,43). These discrepancies may be due to the different expression of NKG2D ligands between infections and between cell populations. Diverse functions of NKG2D ligands could have an impact on the anti-infective activity of Vg9Vd2 T cells.
Our data suggest that targeting Vg9Vd2 T cells to activate their cytotoxic functions may be considered a promising strategy for the treatment wide range of pathogens like for C. burnetii. Indeed, alterations in the phenotype and/or functions of Vg9Vd2 T cells have been reported in several infections usually caused by intracellular pathogens. For example, in patients with active tuberculosis, a progressive loss of effector function of circulating Vg9Vd2 T cells has been reported, leading to decreased IFN-g production and granulysin expression (51,52). This alteration was correlated with disease progression (53, 54), suggesting that a high level of bacteria can lead to chronic stimulation of Vg9Vd2 T cells that would result in their apoptosis and/or senescence. Targeting Vg9Vd2 T cells in the context of persistent infections could therefore be an attractive strategy. Future phenotypic and functional analyses of Vg9Vd2 T cells from patients with Q fever will allow to determine whether their capacity is altered.
Recently, a novel approach has been developed to expand and activate Vg9Vd2 T cells besides pAgs. This strategy is based on the development of a new class of molecules called immunoantibiotics, in particular the inhibitor IspH (55). IspH, an enzyme of the isoprenoid synthesis pathway, is essential for the survival of most Gram-negative bacteria and the absence of B A FIGURE 6 | Anti-BTN3A agonist antibody increases the secretion of cytokines and cytotoxic molecules in Vg9Vd2 T cell/infected-monocyte co-cultures. Monocytes isolated from healthy donors (n = 4) previously infected 24 hours with C. burnetii NM1 (50 MOI) or with M. tuberculosis (5 MOI) were co-cultured with autologous Vg9Vd2 T cells (E:T ratio of 1:1) in the presence of anti-BTN3A antibody (clone 20.1) (0-10 µg/ml). After 4 hours of co-culture, the culture supernatants were analyzed for the presence of cytokines (A, left panel) and cytotoxic molecules (B, right panel) by ELISA assay. Data were analyzed using a normality test and a parametric t test. Values represent mean ± standard deviation. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.
IspH causes an accumulation of its substrate HMBPP, which in turn activates Vg9Vd2 T cells. Another approach would be to target specifically the ligands expressed on the surface of stressed infected cells, such as BTN3A, which will vehicle activation and cytotoxicity of Vg9Vd2 T cells (56). This is the case in a trial in cancer patients where the approach is to activate Vg9Vd2 T cells by targeting BTN3A (NCT04243499, ImCheck Therapeutics, Marseille, France) (57,58).
In addition, we have also explored the effect of the 20.1 mAb in the case of SARS-CoV-2 infection. By activating the Vg9Vd2 T cells, 20.1 mAb may affect intracellular SARS-CoV-2 replication in vitro in infected cells (submitted manuscript). Future studies should be conducted to elucidate the detailed mechanisms of protective Vg9Vd2 T cell activation and how precisely BTN3A is involved in infections. These results highlight that the BTN3A agonist antibody could represent powerful therapeutic tool in infections to overcome the imbalances in immune responses observed in some patients and open new perspectives in Vg9Vd2 T-cell-based immunotherapies in infectious diseases.
In summary, this study provided further insight into the role of Vg9Vd2 T cells in infections with intracellular bacteria. We demonstrated that C. burnetii infection results in modulation of BTN3A and BTN2A co-receptor expression, allowing activation of Vg9Vd2 T cells. We report for the first time the role of a BTN3A agonist antibody in the control of intracellular bacterial infection. The latter boosts the cytotoxic functions of Vg9Vd2 T cells in vitro such as their degranulation, the production of TNFa and IFN-g, and killing activity leading to a better clearance of C. burnetii load of infected target cells. These results may facilitate new approaches to the treatment of persistent bacterial infections by enhancing Vg9Vd2 T cell responses in presence of infected cells.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
A convention No.7828 was established between our laboratory and the Etablissement Francais du Sang (Marseille, France). The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
LG, MG, and MF performed the experiments and analyzed the data. SM, CC, EF, L.M, J-LM and DO supervised the work. LG, SM, J-LM, and DO participated in the writing of the paper. All the authors read and approved the final manuscript.

FUNDING
LG was supported by a Cifre fellowship from ImCheck Therapeutics. Supplementary Figure 3. | Anti-BTN3A agonist antibody increases antimicrobial activity of Vg9Vd2 T cells towards M. tuberculosis infected monocytes. Monocytes isolated from healthy donors (n=4) previously infected 24 hours with M. tuberculosis (5 MOI) were co-cultured with autologous Vg9Vd2 T cells (E:T ratio of 1:1) in the presence of anti-BTN3A antibody (clone 20.1) (0-10 µg/ml). After 4 hours of coculture, M. tuberculosis load was measured by qPCR. Data were analyzed using a normality test and a Mann-Whitney U test. Values represent mean ± standard deviation. *p < 0.05, **p < 0.01 and ***p < 0.001.