TARM1 Is a Novel Leukocyte Receptor Complex–Encoded ITAM Receptor That Costimulates Proinflammatory Cytokine Secretion by Macrophages and Neutrophils

The initiation and resolution of immune responses are orchestrated through a complex interplay between activating and inhibitory signals transmitted by an array of cell surface receptors. Many immunoreceptors are found as clusters of closely related genes encoding “paired” cell surface receptors with activating and inhibitory functions and, in some instances, share the same ligands. Families of such paired Ig-like receptors are encoded within the leukocyte receptor complex (LRC) on chromosome 19q13.4 (1). These include the killer Ig-like receptors (KIRs), expressed mainly on NK cells (2), and the leukocyte Ig-like receptors (LILRs), expressed on lymphoid and myeloid cells (3).

The inhibitory members of the LRC gene family contain long cytoplasmic domains with one or more ITIM (4), whereas the activating counterparts have a short cytoplasmic tail and encode a charged amino acid in their transmembrane (TM) domain. This charged amino acid, usually an arginine or a lysine, facilitates the association with an ITAM-bearing signaling adaptor such as the FcRγ (5–8). Examples of such paired receptors include some with antagonistic signaling properties, such as KIR3DL1/KIR3DS1 and LILRB3/LILRA6, which bind MHC class I ligands. The mechanism of signal integration from such receptors and their ability to modulate cellular immune responses are not fully understood.

Although some KIRs and LILRs are known to bind MHC class I ligands, non–MHC ligands have also been identified as targets for several LRC-encoded receptors. For example, the activating LRC receptor osteoclast-associated receptor (OSCAR) interacts with collagen and is involved in bone metabolism (9). The inhibitory LRC receptor leukocyte-associated Ig-like receptor-1 also binds to collagen, an interaction thought to be involved in tumor immune evasion (10).

The concept that multiple activating receptors encoded within the LRC have closely related inhibitory counterparts encoded in close proximity (1) prompted us to investigate whether OSCAR has an inhibitory paralog. We examined the human genomic sequence adjacent to OSCAR and identified a novel gene encoding an orphan receptor related to OSCAR named T cell–interacting, activating receptor on myeloid cells (TARM1). Another related receptor, signal inhibitory receptor on leukocytes 1 (SIRL-1), encoded by the VSTM1 gene (GenBank NM_198481) is located close to TARM1 and has recently been shown to negatively regulate oxidative burst in human phagocytes (11, 12). The amino acid sequences of SIRL-1 and TARM1 are closely related, and they may represent another example of paired receptors that duplicated from a common ancestor and acquired antithetical functions in terms of cellular activation.

Neutrophils have traditionally been viewed as short-lived, terminally differentiated effectors of the innate immune response. However, this view has recently been challenged by emerging evidence that circulating neutrophils may live longer than previously appreciated, can undergo reverse transmigration, and display plasticity and functional and phenotypic heterogeneity (13, 14). There is compelling evidence that neutrophils engage in bidirectional interactions with a variety of immune cells to modulate adaptive immune responses (15, 16). For instance, ex vivo culture of human and murine neutrophils in the presence of IFN-γ, GM-CSF, and IL-3 induces a dendritic cell (DC)–like phenotype, whereby neutrophils become less susceptible to apoptosis while acquiring the ability to prime Ag-specific T cell responses (13, 14, 17–19). Similarly, in the absence of exogenous cytokines, Ag-pulsed neutrophils can present in an MHC class II (MHC II)–dependent manner to Ag-specific T cells and induce their polarization toward a proinflammatory Th1 or Th17 phenotype (20, 21). Additionally, in vivo, neutrophils were found to deliver Ags or live bacilli to the lymph nodes (22, 23) and cross-present to prime naive CD8⁺ T cells (24, 25).

Neutrophils can also inhibit T cell responses through release of soluble factors such as reactive oxygen species (26, 27), arginase I (28–30), thromboxanes (31), and, in mice, production of the anti-inflammatory cytokine IL-10 (32, 33). Cell contact–dependent suppression has also been demonstrated. Activated neutrophils were shown to inhibit lymphocyte proliferation through the programmed death ligand 1/programmed death 1 pathway in vitro (34) and in HIV-1–infected patients (35). CD11b on human neutrophils from patients with acute systemic inflammation was shown to facilitate the formation of immunological synapses with T cells and enabled a local release of H₂O₂ (36).

Because identifying the expression profiles and immunoregulatory properties of polymorphic immunoreceptors, such as those encoded in the LRC, contributes to understanding the delicate balance of immune responses, we investigated the role of TARM1 in relation to these findings. We demonstrate that TARM1 is expressed by human and murine neutrophils and associates with the ITAM-bearing adaptor FcRγ. We also show that cell surface expression of TARM1 is upregulated by certain TLR agonists in neutrophils, macrophages, and DCs and that the concomitant stimulation of TARM1 and TLRs enhances the secretion of proinflammatory cytokines by macrophages and neutrophils. Additionally, an immobilized TARM1-Fc fusion protein inhibits CD4⁺ T cell activation and proliferation in vitro.

Female NOD mice were bred and maintained under barrier conditions in the Biological Services facility of the Department of Pathology, University of Cambridge. Female C57BL/6 mice were purchased at 6 wk of age from Harlan Laboratories (Loughborough, U.K.) and maintained under barrier conditions at the Department of Veterinary Medicine, University of Cambridge. Animals received standard laboratory food and water ad libitum. All animal experiments were approved by the Ethical Review Committee of the University of Cambridge. All human studies were approved by the local Research Ethics Committee and complied with the Declaration of Helsinki. Informed written consent was obtained from healthy adult volunteers to purify circulating polymorphonuclear neutrophils (ethical approval UK06/Q0108/281).

cDNAs encoding the full coding sequence of human TARM1 (GenBank, DQ479398) and murine TARM1 (DQ973493) were amplified by RT-PCR from total RNA of bone marrow (BM) and spleen, respectively, using the following primers: human TARM1 forward primer 5′-ACTCTGGGAGGGCTAAGGAG-3′ was specific to exon 1 (5′ untranslated region [UTR]) and reverse primer 5′-GAATGCAGTCCAGCAGGTTG-3′ was specific to exon 5 (3′ UTR). Murine TARM1 forward primer 5′-AGACCTGCTGAAGACCTTTG-3′ was specific to exon 1 (5′ UTR) and reverse primer 5′-AGGGTTTATTTGGAGACAGC-3′ was specific to exon 5 (3′ UTR).

Total RNA was extracted from tissues of 8- to 10-wk-old C57BL/6 female mice with TRIzol reagent (Invitrogen) following the manufacturer’s instructions. cDNA was synthesized from 2 μg total RNA using oligo deoxythymine [oligo(dT)] primer and SuperScript III (Invitrogen). PCR screening was performed using the following primers: forward primer 5′-AGACCTGCTGAAGACCTTTG-3′ was specific to 5′ UTR region of TARM1 and reverse primer 5′-TTCAACCAGGAAGCCTCCCACTATTA-3′ was specific to exon 6. Mouse Gapdh was used as a reference gene with the following primers: forward 5′-GCAGTGCCAGCCTCGTCC-3′ and reverse 5′-TGAGGTCAATGAAGGGGTCGT-3′. Human total RNA master panel II was purchased from Clontech (catalog no. 636643). cDNA was synthesized from 2 μg total RNA using oligo(dT) primer and SuperScript III (Invitrogen). Forward primer 5′-CACAAGGGGAGATGGGTCAC-3′ was specific to the junction of exons 2 and 3; reverse primer 5′-AGCCCCGGTTCAAGATGGAG-3′ was specific to exon 5. Human GAPDH was used as a reference gene with the following primers: forward 5′-GAAGGTGAAGGTCGGAGTC-3′ and reverse 5′-CATCACGCCACAGTTTCCC-3′.

Mouse tissues were harvested at indicated time points following Salmonella infection and stored in RNAlater (Qiagen) at −20°C until further processing. Total RNA was extracted using an RNeasy kit (Qiagen) and cDNA was synthesized from 2.5 μg total RNA using oligo(dT) primer and SuperScript III (Invitrogen). Quantitative PCR (qPCR) was performed using GoTaq qPCR master mix (Promega) according to the manufacturer’s instructions on an ABI 7500 Fast real-time PCR system. Forward primer sequence 5′-TCTGTGATAGACAACCATCTGCCTC-3′ was designed to span the junction of exons 4 and 5. Reverse primer sequence 5′-ACACCGACCCGGATGAGATT-3′ was specific to exon 6. Gapdh was used as a reference gene using Gapdh QuantiTect primer mix (Qiagen, catalog no. QT01658692). Primer amplification efficiencies (E = 1.9 for both TARM1 and Gapdh) were calculated from the slope of a standard curve prepared with a 2-fold serial dilution of splenic cDNA. TARM1 gene expression levels were calculated as fold change over the levels detected in control animals using the fold change = 2^−∆∆CT method (37). Primer specificity was verified by melt curve analysis and amplicon sequencing.

DNA sequences encoding the extracellular portions of murine (aa 16–255) and human (aa 16–233) TARM1 were cloned into the mammalian expression vector Signal pIg Plus consisting of the CD33 signal sequence and a human IgG1 Fc tail. The plasmids were transfected into HEK293 cells and stable cell lines were established. Cells were maintained in DMEM with 2% low IgG FCS (Invitrogen). Culture medium containing the secreted Fc fusion proteins was collected and TARM1-Fc was purified on columns packed with protein A–agarose FastFlow beads (Sigma-Aldrich). Proteins were secreted predominantly as dimers as assessed by Western blot.

Monoclonal mouse anti-human TARM1 and rat anti-mouse TARM1 Abs were raised against the ectodomain of mouse and human TARM1 proteins using TARM1-Fc as immunogens. The performance and specificity of several Ab clones were validated by flow cytometry and Western blot against transfected and primary cells.

HEK293T cells were transfected with full-length Flag-tagged TARM1 using FuGENE HD (Promega). Total cell lysates were prepared with Nonidet P-40 lysis buffer (1% Nonidet P-40, 0.15 M NaCl, 0.05 M Tris-HCl [pH 8]) containing 1 mM AEBSF (Sigma-Aldrich) for 25 min on ice and then centrifuged at 14,000 × g and 4°C for 15 min. Supernatants were collected and N-linked carbohydrates were removed by incubation with peptide-N-glycosidase F (PNGase F) or endoglycosidase H (Endo H) (New England BioLabs) for 4 h at 37°C, according to the manufacturer’s instructions. Following the treatment, protein mobility shift was examined by SDS-PAGE and Western blotting.

HEK293 cells stably expressing hemagglutinin (HA)-tagged Dap10, Dap12, and FcRγ adaptors were transiently transfected with full-length Flag-tagged human TARM1 and cell lysates were prepared with TBS lysis buffer (1% Triton X-100, 0.15 M NaCl, 0.02 M Tris-HCl [pH 8]) containing 0.5 mM AEBSF (Sigma-Aldrich) and ProteoBlock (Fermentas) for 25 min on ice. Lysates were centrifuged at 10,000 × g and 4°C for 5 min. Supernatants were collected and incubated with monoclonal anti-TARM1 Ab for 1 h at 4°C on a rotor. Protein A–agarose FastFlow beads (Sigma-Aldrich) were added and incubated on a rotor for 1 h at 4°C. The agarose beads were then washed three times with TBS lysis buffer and twice in TBS. Proteins were eluted from the beads by heating at 70°C for 10 min in lithium dodecyl sulfate sample buffer and separated by SDS-PAGE. Coimmunoprecipitated proteins were analyzed by Western blotting.

Cell lysates were denatured in lithium dodecyl sulfate sample buffer (Invitrogen), separated on 10 or 12% Tris-glycine gels, and transferred to an Immobilon-P polyvinylidene difluoride membrane (Millipore). Blocking and immunoblotting steps were carried out in PBS with 0.05% Tween 20 and 5% milk. Detection was performed with mouse anti-human TARM1 and goat anti-mouse HRP Ab (Invitrogen), or rat anti-HA high-affinity Ab (Roche) and anti-rat HRP mouse-adsorbed Ab (AbD Serotec). All blots were developed with ECL Western blotting reagent (GE Healthcare).

To induce sterile inflammation, C57BL/6 mice were injected with ultrapure LPS (3 μg/mouse) (InvivoGen) i.p., and tissues were harvested for immunophenotyping 24 h later. Systemic infection was induced with an aroA attenuated strain of S. enterica serovar Typhimurium, SL3261. Bacteria were grown overnight in Luria–Bertani broth until stationary phase and C57BL/6 mice were injected i.v. with 1 × 10⁶ CFU diluted in PBS. At indicated times following infection, mice were euthanized and TARM1 expression in cells and tissues was analyzed by flow cytometry, qPCR, and confocal microscopy.

Spleen fragments from control or Salmonella-infected mice were harvested at 2 wk postinfection (wpi) and fixed in 4% paraformaldehyde, paraffin embedded, and stored until analysis. Tissue sections were deparaffinized, and heat-induced epitope retrieval was performed in citrate buffer (pH 5.5). Sections were sequentially stained with anti-TARM1 and goat F(ab′)₂ anti-rat FITC (AbD Serotec) secondary Ab. Nuclei were visualized with DAPI. Multiple images of each section were taken to ensure accurate representation.

A single-cell suspension was prepared from spleen, lymph nodes, liver, BM, and peritoneal exudate cells (PECs) of mice using standard procedures. Spleens and lymph nodes were dissociated by gently passing the tissues through a 70-μm cell strainer (BD Falcon) followed by hypotonic RBC lysis. BM was extracted from femurs and tibias by flushing with 5 ml ice-cold PBS. To prepare total liver cells, livers were treated with 1 mg/ml collagenase D (Roche) for 30 min at 37°C and then passed through a 70-μm cell strainer. Total liver cell suspension was washed with PBS–2.5% FCS, and the pellet was resuspended in 33% Percoll (GE Healthcare) solution. The suspension was centrifuged for 20 min at 2100 rpm at room temperature and the pellet was washed in PBS–2.5% FCS followed by hypotonic RBC lysis. PECs were isolated by three successive washes of the peritoneal cavity with 4 ml ice-cold PBS.

For flow cytometric analysis of whole blood, 10 ml was collected from healthy volunteers, and leukocyte-rich plasma was prepared by sedimentation of RBCs using HetaSep (StemCell Technologies) according to the manufacturer’s instructions. Circulating neutrophils were purified from healthy volunteers using dextran sedimentation and discontinuous plasma-Percoll gradients as previously described (38). Briefly, 40 ml blood was collected into 50-ml tubes (containing 4 ml 3.8% [w/v] sodium citrate) using a 19-gauge needle and centrifuged for 20 min at 300 × g. The platelet-rich plasma upper layer was removed for later use. RBCs were removed from the leukocyte/erythrocyte pellet by dextran sedimentation. The upper leukocyte-rich layer was removed and centrifuged at 275 × g for 5 min. The cell pellet was resuspended in platelet-poor plasma, and neutrophils were isolated by discontinuous plasma-Percoll gradient centrifugation. Neutrophil purity as determined by cytospins was routinely >95% using this method.

BM was flushed from femurs and tibias of 8- to 12-wk-old female C57BL/6 mice, dispersed by passage through a 70-μm cell strainer, and seeded into 100-mm cell culture dishes at 1 × 10⁶ cells/ml in complete IMDM (Life Technologies) containing 10% FBS (Life Technologies), 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine (all from Life Technologies), and 50 μM 2-ME (Sigma-Aldrich). For BM-derived macrophage (BMM) differentiation, the medium was supplemented with 10% l-cell conditioned medium as a source of M-CSF. For BM-derived DC (BMDC) differentiation, the cells were allowed to adhere for 30 min at 37°C and the nonadherent cells were reseeded into new plates in complete IMDM supplemented with 20 ng/ml GM-CSF and 10 ng/ml IL-4 (both from PeproTech). Cells were harvested on day 7 and the cellular phenotype was confirmed by flow cytometry using CD11b and F4/80 markers for BMMs and CD11c and CD80 for BMDCs.

BM was flushed from femurs and tibias of C57BL/6 mice and the total granulocyte population was sorted on a MoFlo cell sorter (DakoCytomation) using side scatter (SSC) and forward scatter (FSC) characteristics. Subsequent staining with anti-CD11b and Gr-1 markers showed that >94% of sorted cells were neutrophils.

For the analysis of TARM1 cell surface expression, BMMs and BMDCs were stimulated for 24 h with the following TLR agonists (all from Apotech): 100 ng/ml Pam3CSK4 (TLR1/2 agonist), 50 μg/ml poly(I:C) (TLR3 agonist), 100 ng/ml LPS from Escherichia coli R515 (TLR4 agonist), 50 ng/ml flagellin from S. Typhimurium (TLR5 agonist), 80 ng/ml MALP-2 (TLR2/6 agonist), 1 μg/ml imiquimod (TLR7 agonist), 2 μg/ml CpG oligodeoxynucleotide 1585 (TLR9 agonist), and 250 ng/ml profilin from Toxoplasma gondii (TLR11 agonist). For analysis of cytokine secretion, 1 × 10⁵ BMMs or 1.3 × 10⁵ sorted neutrophils were cultured in 96-well tissue culture plates coated with goat anti-rat capture Ab and rat anti-TARM1 crosslinking Ab or a rat IgG2a isotype control Ab (all at 10 μg/ml). Neutrophils were stimulated in the presence or absence of 10 ng/ml ultrapure LPS from E. coli K12 (InvivoGen) for 8 h at 37°C. BMMs were stimulated in the presence or absence of either 0.5 μg/ml Pam3CSK4, 0.5 μg/ml poly(I:C), 1 ng/ml LPS, or 10 μM imiquimod for 16 h at 37°C. Tissue culture supernatants were collected and cytokine production was determined using the cytometric bead array mouse inflammation kit (Becton Dickinson). DuoSet sandwich ELISA (R&D Systems) was used to determine IL-2 secretion by stimulated mouse T cells.

For flow cytometric cell phenotyping, 0.5–1 × 10⁶ cells were stained per well. The following Abs were used: goat anti-mouse IgG Alexa Fluor 647 secondary Ab (Molecular Probes) and goat anti-rat IgG (H+L) PE mouse adsorbed secondary Ab (SouthernBiotech). Anti–CD3-FITC (2C11), CD19-allophycocyanin (1D3), CD11b-allophycocyanin (M1/70), CD11c (N418), Gr-1 (RB6-8C5), CD25 (PC61), and CD69 (H1.2F3) were from eBioscience; Ly6G-FITC (1A8) and Ly6C-allophycocyanin (HK1.4) were from BioLegend; and Annexin V^FITC was from BD Pharmingen. Flow cytometry was performed on BD FACScan with Cytek DxP three-laser setup, and data analysis was carried out using FlowJo 10 software (Tree Star).

For CFSE dilution assays, CD4⁺ T cells were purified from naive NOD female mice using EasySep positive selection magnetic beads (StemCell Technologies) to 93–98% purity as assessed by flow cytometric analysis. Cells were loaded with 1 μM CFSE (Molecular Probes) in PBS for 15 min at 37°C, washed, and seeded into 96-well flat-bottom plates coated, in two layers, with first 10 μg/ml anti-human IgG Fc-specific Ab (ImmunoReagents) and 1 μg/ml anti-CD3 (clone 145-2C11, eBioscience) together with 0.25 μg/ml anti-CD28 (clone 37.51, Miltenyi Biotec) overnight at 4°C, washed, and then coated with 10 μg/ml mouse TARM1-Fc or human IgG1 control. A total of 1.5 × 10⁵ T cells were added to each well in a volume of 200 μl RPMI 1640 with 10% FCS, penicillin/streptomycin, and 50 μM 2-ME. CFSE dilution was assessed after 90 h of culture. For metabolic MTT assays, CD4⁺ T cells were purified using EasySep negative selection magnetic beads (StemCell Technologies) and activated as described above, but excluding anti-CD28 Ab. Cells were cultured for 20, 40, and 80 h. Two hours prior to each time point, MTT (Sigma-Aldrich) stock solution was added directly to the wells to give a 500 μg/ml final concentration, and following 2 h incubation at 37°C, DMSO was used to dissolve formazan crystals and absorbance (570 nm) was measured on a Synergy HT plate reader. Human T cells were isolated from peripheral blood of healthy donors (n = 2) using CD4⁺ positive selection magnetic beads (Miltenyi Biotec) to 95–98% purity and activated with plate-bound anti-CD3 (1.2 μg/ml, clone OKT3) in the presence of plate-bound human TARM1-Fc (10 μg/ml) or hIgG1 (10 μg/ml). Recombinant human IL-2 (50 U/ml) was added to some wells as indicated. T cells were cultured for 3 d and the expression of CD25 and CD69 was analyzed by flow cytometry.

We performed a bioinformatics analysis of open reading frames encoded at the centromeric end of the human LRC on chromosome 19q13.4 and identified a putative gene TARM1 (GenBank NM_001135686) closely related to VSTM1 and OSCAR (Fig. 1A). We also identified the analogous murine TARM1 gene, which closely resembles its human ortholog both in sequence and exon structure and is encoded within the syntenic region of the murine chromosome 7A1 (Fig. 1B). The TARM1 gene spans ∼11.4 kb of human genomic sequence encoding a polypeptide of 271 aa and ∼13.4 kb of mouse sequence encoding a polypeptide of 288 aa. Protein alignment of human and mouse orthologs indicated ∼47% sequence identity and 62% sequence similarity. Both are type Ia membrane proteins composed of a 16-aa hydrophobic signal peptide followed by two extracellular Ig-like domains, a TM region with a conserved arginine, and a short cytoplasmic tail devoid of signaling motifs.

A BLAST search of the National Center for Biotechnology Information human and mouse sequence databases using nucleotide and protein sequences of TARM1 revealed that it is most closely related to members of the Ig superfamily (IgSF) encoded within the LRC, in particular SIRL-1 and OSCAR. The single Ig V domain of SIRL-1 is ∼48% identical and ∼58% similar to the first Ig-like domain of TARM1. The ectodomains of OSCAR and TARM1 proteins are ∼36% identical and ∼46% similar in their amino acid sequences.

We amplified a full-length transcript of TARM1 and examined its expression in normal human and mouse tissues by RT-PCR. TARM1 mRNA was expressed in lymphoid organs thymus, spleen, and mesenteric lymph nodes, with most transcript detected in the BM. TARM1 expression was also detected in nonlymphoid tissues, lung, and uterus (Fig. 1C).

To characterize TARM1 further we raised specific mAbs to the human and mouse TARM1 receptors for use in flow cytometry and Western blotting. Ab clones mAb13 (rat anti-mouse TARM1) and mAb4 (mouse anti-human TARM1) stained TARM1-transfected HEK293T cells but not mock-transfected cells, as determined by flow cytometry (Fig. 2A). Clones mAb2 (rat anti-mouse TARM1) and mAb74 (mouse anti-human TARM1) detected Flag-tagged TARM1 in transfected cell lysates by Western blot (Fig. 2B). Surprisingly, the anti-Flag Ab failed to detect the higher molecular mass (M_r) band detected by the TARM1-specific Abs in both human and mouse blots. We speculated that this discrepancy may be due to the steric occlusion of the Flag epitope caused by TARM1 glycosylation.

Two N-linked glycosylation sites were predicted for both human and mouse TARM1 polypeptides. To determine the glycosylation state, HEK293T cells were transiently transfected with Flag-tagged full-length human TARM1 and the total cell lysates were treated with either PNGase F or Endo H. The protein mobility shift was examined by SDS-PAGE (Fig. 2C). The predicted M_r of the unglycosylated TARM1 protein is ∼29.5 kDa. Immunoblotting of untreated or mock-treated protein with anti-TARM1 mAb74 showed two bands of ∼40 and ∼34 kDa, indicating the presence of posttranslational protein modification. Treatment with Endo H resulted in a conversion of only the 34-kDa band to ∼30 kDa (Fig. 2C, left panel). This suggests that the immature, Endo H–sensitive form of TARM1 protein is recognized by both the anti-Flag and the anti-TARM1 Abs; however, the fully glycosylated higher M_r TARM1 species may not be recognized by anti-Flag Ab. Treatment with PNGase F resulted in a single TARM1 that migrated at ∼30 kDa, which represented the fully deglycosylated form (Fig. 2C, right panel).

An arginine residue is embedded within the TM region of mouse and human TARM1. In the OSCAR TM, this residue mediates association with FcRγ, resulting in increased cell surface expression and transduction of activating ITAM signals (6, 8, 39, 40). To examine whether TARM1, similar to OSCAR, associates with an ITAM-containing adaptor, we generated HEK293 cell lines stably expressing HA epitope–tagged FcRγ (HA-FcRγ), DAP12 (HA-DAP12), and DAP10 (HA-DAP10) adaptors, respectively. The cell lines were transiently transfected with a full-length Flag-tagged human TARM1 and used in coimmunoprecipitation experiments. Only FcRγ, but not DAP10 or DAP12, coimmunoprecipitated with TARM1 (Fig. 2D). These data demonstrate that TARM1 can specifically associate with the ITAM-containing adaptor FcRγ, consistent with an activating function (6, 8, 39, 40).

High expression of mouse TARM1 mRNA was detected in the BM of healthy C57BL/6 mice by RT-PCR analysis. Constitutive cell surface expression of TARM1 protein was confirmed by flow cytometry and was only observed within the BM and not at peripheral sites (Fig. 3A). Closer inspection revealed that TARM1 expression was restricted to granulocytes with a CD11b⁺Gr-1⁺ cell surface phenotype (Fig. 3B).

TARM1 mRNA expression could be further enhanced by i.p. injection of low-dose (150 μg/kg) LPS (data not shown). To assess whether this increase in TARM1 transcript correlated with the increase in cell surface protein expression, we repeated the low-dose LPS treatment and examined TARM1 expression in several organs 24 h following i.p. injection. LPS treatment caused a strong upregulation of TARM1 cell surface expression within the CD11b⁺Gr-1⁺ population (Fig. 3B). Microscopy of TARM1⁺ cells isolated from the BM of LPS-treated animals revealed a heterogeneous population composed of mature and immature neutrophils, as determined by cell morphology (Fig. 3B, inset).

Interestingly, following i.p. LPS administration, PECs contained a population with a FSC^intSSC^hi phenotype comprised predominantly of TARM1⁺ cells (Fig. 3C). This population was not present in PECs from control animals. The anti–Gr-1 (RB6-8C5) Ab recognizes epitopes present in both Ly6C and Ly6G. We therefore used Ly6C- and Ly6G-specific Abs for further phenotypic characterization of TARM1⁺ cells. We found that TARM1⁺ cells in the BM coexpressed classical markers of the granulo-monocytic lineage, such as CD11b, Ly6G, and Ly6C, but were negative for markers of B cells (CD19), T cells (CD3), and DC (CD11c) (Fig. 4A). Twenty-four hours following LPS injection, TARM1⁺ cells infiltrating the peritoneum formed two distinct populations based on their FSC characteristics (Fig. 4B). The predominant population in gate 1 (Fig. 4B) was CD11b⁺Ly6C^intLy6G^hiMHC II⁻. The population in gate 2 was CD11b⁺Ly6C^hiLy6G^loMHC II⁺. These phenotypic characteristics are consistent with granulocytic origin of the population in gate 1, and monocytic origin of the population in gate 2. Because TARM1⁺ cells did not express DC/macrophage marker CD11c in either the BM or the peritoneum, the MHC II⁺ cells in gate 2 are most likely inflammatory monocytes. These cells were shown to be rapidly recruited to the sites of inflammation where they can give rise to proinflammatory DCs and macrophages (41). To investigate whether DCs and macrophages express TARM1 under inflammatory conditions, we generated BMDCs and BMMs and stimulated the cells with a panel of TLR agonists (Fig. 5). Unstimulated BMDCs (Fig. 5A), but not BMMs (Fig. 5B), expressed low levels of cell surface TARM1. The cell surface expression of TARM1 was upregulated to varying degrees following stimulation of BMDCs and BMMs with agonists for TLR-1/2, -3, -4, -2/6, and -7. In contrast, TARM1 expression remained unchanged following stimulation with agonists for TLR-5, -9, or -11 (Fig. 5A, 5B). Upregulation of Tarm1 transcript by BMMs and BMDCs following 24 h LPS stimulation was confirmed by semiquantitative RT-PCR (Fig. 5C). These results show that DCs and macrophages upregulate TARM1 at the transcriptional and translational level under specific proinflammatory conditions.

We also examined TARM1 protein expression in human peripheral blood (Fig. 6). As in mice, human TARM1 was detected on circulating granulocytes, but the expression varied greatly among donors (n = 15) (Fig. 6A). Western blot analysis of purified peripheral blood neutrophils confirmed the presence of TARM1 protein within these cells but not in neutrophil-depleted mononuclear cell fraction (Fig. 6B).

These results indicate that in healthy mice TARM1 expression is restricted to the BM CD11b⁺Ly6C⁺Ly6G⁺ mature and immature neutrophils. The inflammatory response was strongly associated with increased TARM1 expression on BM neutrophils and with homing of TARM1⁺ cells to peripheral sites. Monocytes infiltrating the site of inflammation also expressed cell surface TARM1. As in mice, human TARM1 was also expressed on granulocytes, suggesting evolutionary conservation of TARM1 function.

To study the effects of a systemic immune challenge on TARM1 expression, we administered an aroA attenuated strain of S. Typhimurium SL3261 i.v. to C57BL/6 mice. Flow cytometric analysis of spleens at 1, 2, 3, and 4 wpi showed an influx of TARM1⁺ granulocytes as evidenced by the accumulation of the FSC^int/hiSSC^hi population (Fig. 7A). TARM1 receptor density correlated positively with Ly6G expression. TARM1⁺ cells reached ∼10% of total splenocytes by 1 wpi and declined between 3 and 4 wpi (Fig. 7B). qPCR analysis of Tarm1 mRNA expression in total splenocytes of infected mice indicated an ∼6-fold upregulation over control animals at 24 h following infection, with a peak expression of ∼22-fold at 1 wpi (Fig. 7C). The presence of infiltrating TARM1⁺ cells was also confirmed by confocal microscopy 2 wpi in the spleens of infected but not control animals (Fig. 7D). A strong upregulation of TARM1 expression on the cell surface of CD11b⁺Ly6C⁺ and CD11b⁺Ly6G⁺ cells was observed at the same time point after infection in the BM (Supplemental Fig. 1). Thus, systemic infection with Salmonella resulted in upregulation of TARM1 cell surface expression on granulocytes and their migration to the spleen.

ITAM receptors such as OSCAR and TREM-1 have been shown to regulate cytokine secretion (42) and proinflammatory responses (43, 44) of myeloid cells. Because TARM1 expression was regulated by TLR ligands, we examined whether TARM1 ligation in vitro might influence the secretion of proinflammatory cytokines by BMMs and neutrophils. BMMs or primary mouse BM neutrophils were cultured in tissue culture plates coated with either TARM1-specific or isotype control mAbs in the presence or absence of TLR agonists (Fig. 8). Cross-linking of TARM1 in the absence of TLR agonists did not stimulate cytokine secretion by either BMMs or neutrophils (Fig. 8). However, the concomitant stimulation of TARM1 and TLR-1/2, -3, -4, and -7 enhanced secretion of the proinflammatory cytokines TNF-α and IL-6 by BMMs (Fig. 8A). Similarly, concomitant stimulation of TARM1 and TLR4 enhanced the secretion of TNF-α and IL-6 by neutrophils (Fig. 8B). These results show that TARM1 cooperates with TLR stimulation to enhance the secretion of proinflammatory cytokines by macrophages and neutrophils.

Numerous studies have demonstrated a role for neutrophils in the regulation of T cell responses. We therefore examined the immunomodulatory effects of TARM1 on anti-CD3/-CD28–induced CD4⁺ T cell activation and proliferation. For this purpose, we generated soluble mouse TARM1-Fc fusion protein comprised of the ectodomain of TARM1 fused to the Fc portion of human IgG1.

When purified primary mouse CD4⁺ T cells were activated with plate-bound anti-CD3 and anti-CD28 in the presence of plate-bound TARM1-Fc fusion protein, T cell proliferation was inhibited as measured by CFSE dilution assay (Fig. 9A). Cells proliferated normally when control IgG1 was used instead of TARM1-Fc. Because cellular activation and proliferation are accompanied by an increase in cellular metabolism, we used a tetrazolium salt MTT as a redox indicator in a metabolic proliferation assay. This assay allows multiple measurements to be carried out throughout the course of cell stimulation and enables the study of T cell activation dynamics. We used MTT to measure TARM1-Fc–mediated inhibition of CD4⁺ T cell activation during the course of 4 d and observed a significant inhibition as early as 20 h (Fig. 9B). The inhibition was accompanied by a significantly reduced IL-2 secretion (Fig. 9C). The proliferation and IL-2 secretion were not inhibited when human IgG1 replaced TARM1-Fc. The inhibition could not be attributed to cytotoxicity, as TARM1-Fc had no impact on T cell viability (Fig. 9D). The decreased live cell count seen in the presence of TARM1-Fc is most likely due to the T cell apoptosis as a consequence of the absence of stimulation. In support of this, the T cell activation markers CD25 and CD69 were lower on T cells stimulated in the presence of TARM1-Fc (Fig. 9E). Immobilized human TARM1-Fc also inhibited the activation of human CD4⁺ T cells compared with control human IgG1 (Supplemental Fig. 2). Collagen was identified as the ligand for the closest TARM1 homolog OSCAR (9). However, we did not detect specific binding of human TARM1-Fc to collagens I–V (data not shown). These results show that the TARM1 receptor can interact with an as yet unidentified ligand on T cells to inhibit T cell activation and proliferation.

We report the identification and characterization of a novel LRC-encoded receptor, TARM1. Similar to many other LRC members, TARM1 has two Ig-like folds and thus belongs to the broader IgSF. Similar to the activating LRC receptors, TARM1 lacks signaling motifs in its short cytoplasmic tail and, instead, associates with FcRγ via a conserved arginine in its TM, suggesting an activating function (4). Phylogenetic analysis shows that human TARM1 is most closely related to an inhibitory receptor SIRL-1 (11) encoded by the neighboring VSTM-1 gene. SIRL-1 expression is restricted to phagocytes such as neutrophils and monocytes, where it plays a role in the negative regulation of the oxidative burst (12) and can prevent the pathogenic release of neutrophil extracellular traps in cells from systemic lupus erythematosus patients (45).

Our flow cytometric analysis of human peripheral blood showed that, similarly to SIRL-1, human TARM1 is also expressed on neutrophils but not monocytes from healthy individuals, although the expression level varied greatly among donors. The expression of murine TARM1 is restricted to the BM CD11b⁺Ly6G⁺Ly6C⁺ granulocyte precursors and mature neutrophils under steady-state conditions. LPS challenge, or systemic infection with S. Typhimurium, induced a strong upregulation of cell surface TARM1 expression by BM CD11b⁺Ly6G⁺Ly6C⁺ cells and their migration to the site of inflammation. Following LPS administration i.p., TARM1⁺ cells homed rapidly to the peritoneum, whereas systemic Salmonella infection induced an accumulation of TARM1⁺ cells in the spleen. Because the proinflammatory environment has been shown to increase neutrophil lifespan (13, 14), and inflammation upregulates TARM1 expression by neutrophils, increased TARM1 expression (TARM^hi) may be a marker for neutrophils with enhanced in vivo lifespans.

More than 50% of TARM1⁺ cells that migrated to the peritoneum upregulated MHC II expression and had a CD11b⁺Ly6C^hiLy6G^lo phenotype but lacked the DC/macrophage marker CD11c. Given the rapid recruitment to the peritoneum following LPS injection, the CD11b⁺Ly6C^hiLy6G^loMHC II⁺TARM1⁺ cells likely represent inflammatory monocytes. These cells were shown to rapidly migrate to the sites of inflammation where they can give rise to proinflammatory DCs and macrophages (41). Indeed, TARM1 expression could be induced on BMMs and BMDCs following stimulation with ligands for TLR-1/2, -3, -4, -2/6, and -7, but not TLR-5, -9, or -11 agonists. Further studies are required to determine whether human TARM1 can be expressed by additional immune cell types other than neutrophils (e.g., monocytes, macrophages, or DCs) and whether human TARM1 cell surface expression is modulated by TLR agonists and other proinflammatory stimuli.

Interestingly, the amino acid sequence of the single IgV domain of SIRL-1 is ∼48% identical to the first Ig-like domain of TARM1. However, owing to its two ITIMs, SIRL-1 exerts inhibitory effects on neutrophil function (12, 45), whereas TARM1 associates with ITAM-bearing adaptor FcRγ, suggesting an activating function (4). Indeed, TARM1 engagement enhances the secretion of proinflammatory cytokines by murine BMMs and BM neutrophils stimulated with TLR ligands, such as LPS. Therefore, TARM1 and the related SIRL-1 may be examples of paired receptors. The LRC contains several sets of closely related, paired receptors with activating and inhibitory functions such as the leukocyte-associated Ig-like receptors, KIRs, and LILRs. The integration of opposing signals delivered by these receptors may balance the modulation of the innate and adaptive immune cell responses (1).

A large body of work suggests that both human and mouse neutrophils can be activated by, and may exert immunoregulatory effects on, T cells (17, 20, 46–48). Notably, T cells from the synovial fluid of rheumatoid arthritis patients were shown to activate neutrophils at the site of inflammation (47). In patients with renal cell carcinoma, neutrophils were shown to suppress T cell responses through arginase I release (30). The secretion of IL-10 (32, 33, 49) and programmed death ligand 1 expression (50) have been proposed as the major components of the immunosuppressive mechanisms employed by neutrophils during bacterial infections. We demonstrate that the TARM1 receptor ectodomain is capable, at least in in vitro assay, of potently inhibiting anti-CD3/-CD28–induced CD4⁺ T cell activation and proliferation. Bearing in mind that TARM1 cell surface expression is rapidly upregulated by BMMs and BMDCs and by neutrophils and inflammatory monocytes in vivo during immune challenge, and TARM1⁺ cells home to the sites of inflammation, TARM1 may constitute a previously uncharacterized negative regulator of T cell activation expressed by neutrophils and inflammatory monocytes, macrophages, and DCs to modulate the early stages of T cell responses.

Our data indicate that the TARM1 receptor ectodomain interacts with an as yet unidentified ligand on T cells that can mediate the inhibition of T cell activation. In contrast, TARM1 receptor stimulation on macrophages and neutrophils costimulated the secretion of proinflammatory cytokines induced by TLR ligands, such as LPS. These data suggest that TARM1 receptor interaction with an unidentified molecule on T cells may result in a bidirectional signaling between macrophages and neutrophils expressing TARM1 and T cells expressing an unidentified inhibitory molecule (Supplemental Fig. 3). Bidirectional signaling has been previously observed between myeloid cells and innate lymphocytes, such as NK cells, as exemplified by the interaction between AICL and NKp80 (51). Similarly to TARM1, the cell surface expression of AICL is also regulated by TLR ligands. The AICL/NKp80 bidirectional signaling interaction resulted in the activation of both NK cells and monocytes, respectively. The identification of the inhibitory TARM1 ligand expressed by T cells would be an interesting subject for further research.

In conclusion, our data show that TARM1 expression is a hallmark of neutrophils and their precursors in the steady-state and that TARM1 is upregulated by mature, activated neutrophils, monocytes, macrophages, and DCs under specific inflammatory conditions. TLR-specific proinflammatory cytokine release by BMMs and primary BM neutrophils was costimulated by TARM1 engagement, and the ability of TARM1-Fc to suppress T cell activation and proliferation in vitro likely points to a role for TARM1-mediated bidirectional signaling in the regulation of immune cell function.

We are grateful to Prof. Jim Kaufman for critically evaluating this manuscript.

The authors have no financial conflicts of interest.