A TNF-Regulated Recombinatorial Macrophage Immune Receptor Implicated in Granuloma Formation in Tuberculosis

Macrophages play a central role in host defense against mycobacterial infection and anti- TNF therapy is associated with granuloma disorganization and reactivation of tuberculosis in humans. Here, we provide evidence for the presence of a T cell receptor (TCR) αβ based recombinatorial immune receptor in subpopulations of human and mouse monocytes and macrophages. In vitro, we find that the macrophage-TCRαβ induces the release of CCL2 and modulates phagocytosis. TNF blockade suppresses macrophage-TCRαβ expression. Infection of macrophages from healthy individuals with mycobacteria triggers formation of clusters that express restricted TCR Vβ repertoires. In vivo, TCRαβ bearing macrophages abundantly accumulate at the inner host-pathogen contact zone of caseous granulomas from patients with lung tuberculosis. In chimeric mouse models, deletion of the variable macrophage-TCRαβ or TNF is associated with structurally compromised granulomas of pulmonary tuberculosis even in the presence of intact T cells. These results uncover a TNF-regulated recombinatorial immune receptor in monocytes/macrophages and demonstrate its implication in granuloma formation in tuberculosis.


Introduction
Macrophages are key players in major chronic inflammatory diseases including tuberculosis, atherosclerosis and rheumatoid arthritis. Based on their myeloid origin and professional phagocytic activity they are traditionally regarded as a pillar of innate immunity [1]. Tuberculosis is an infectious disease that in 2008 afflicted more than nine million individuals worldwide and claimed the lives of an estimated 1.3 million patients [2]. The disease is caused by mycobacteria that are efficiently contained by macrophages in highly organized immune structures, the tuberculous granulomas. Ample evidence indicates that the generation and maintenance of tuberculous granulomas require TNF [3,4]. Moreover, reactivation of the disease by therapeutic TNF blockade is associated with disruption of the granuloma architecture that ultimately leads to spreading of the mycobacteria into the surrounding tissue [5].
Within the tuberculous granuloma, cellular immunity to mycobacteria is thought to be solely under the direction of T cells which orchestrate the macrophage host response to the pathogen. However, selective T cell depletion and reconstitution experiments in murine models of tuberculosis point to the involvement of variable host defense mechanisms in the control of mycobacterial infection beyond T cells [6][7][8]. The recent demonstration by our laboratory and others that neutrophils and eosinophils express T cell receptors (TCR) which are generated by V(D)J recombination has provided evidence for the existence of variable immune receptors outside lymphocytes [9][10][11].
These findings and the possibility that variable immune defense mechanisms outside T cells are implicated in the development of the tuberculous granuloma raise the question as to whether macrophages possess a molecular machinery for variable host defense. Here, we report that subpopulations of monocytes and macrophages express a recombinatorial TCRab which is TNF regulated and demonstrate a role of this novel immune receptor in the macrophage host response to mycobacteria and the formation of the tuberculous granuloma.

Subpopulations of monocytes in the circulation and macrophages express the TCRab
To assess the possibility that monocytes, like granulocytes [9][10][11], express the TCRab, we isolated human CD14 + monocytes from healthy donors (n = 12). Expression of the TCRab in peripheral blood monocytes was assessed in MACS-CD14 + purified cells by immunocytochemistry using antibodies to TCRa/TCRb and MHC-II. Utilizing this approach, we consistently detected a ,5% cell fraction that displayed bright TCRab + expression in freshly isolated CD14 + monocytes which showed coexpression of MHC-II ( Figure 1A). Purity of isolated CD14 + monocytes was routinely .99.5% as determined by flow cytometry ( Figure S1A). We next characterized the TCRab expressing monocyte subpopulation in PBMC by flow cytometry. Consistent with immunocytochemistry, CD14 + cells from three normal subjects displayed positive staining for TCRb in a 3-4% subfraction which did not exhibit staining for the T cell marker CD3 ( Figure 1B, Figure S1B). We then determined expression of the TCRab in monocyte-derived macrophages from three healthy donors. For this, monocytes were differentiated into naïve, IFNc activated and IL-4 activated macrophages, respectively, on glass slides for a period of 6 days [12] and stained for TCRab. To quantify the TCRab + populations in adherent macrophages, laser scanning cytometry (LSC) was performed on the slides that were immunostained with Alexa 555-labeled secondary antibodies. LSC analysis demonstrated that a 5% subpopulation of naïve macrophages exhibited high fluorescence indicative of TCRab expression ( Figure 1C, Figure S1C). In contrast, the fraction of TCRab bearing macrophages was significantly higher in the IL-4 (9%) and IFNc activated macrophages (11%). Thus, activation of macrophages induces an increase in the subpopulation of TCRab expressing macrophages.
Immunoblot analysis confirmed the expression of the TCR aand b-chains in monocytes and macrophages (Figures 1D). In line with this, immunogold electron microscopy using an antibody to the TCR a-subunit revealed specific staining in monocyte-derived macrophages ( Figure 1E), which was detectable on the cell surface. Next we looked for evidence of TCRab expressing macrophages in vivo. For this, we performed double-immunostaining for the macrophage marker CD163 and the TCRab in bronchoalveolar lavage (BAL) fluid obtained from three individuals with normal BAL cytology. In fact, we found that a 5-15% subpopulation of alveolar macrophages showed positive staining for the TCRab ( Figure 1F, Figures S1D and E) demonstrating that tissue macrophages are capable of expressing the TCR under physiological conditions.
We then tested whether the TCR is also expressed in macrophages from the mouse. Consistent with our findings in humans, immunocytochemistry demonstrated the presence of the TCRab in a 5-10% subfraction of spleen macrophages from C57BL/6J mice (n = 3) and RT-PCR revealed mRNA expression of both the TCR aand b-constant chains in splenic macrophages ( Figure 1G).
Together, these results reveal that subpopulations of peripheral blood monocytes and in vitro activated monocyte-derived macrophages constitutively express the TCRab. Moreover, they demonstrate the presence of TCRab bearing macrophages in normal human tissue, as exemplified for the lung, and provide evidence for TCRab expression by murine macrophages.

Monocytes/macrophages express variable TCR aand brepertoires
We next investigated whether the TCRab expressed by monocytes/macrophages represents a variable receptor. For this, we tested whether CD14 + monocytes and IFNc macrophages have a rearranged TCRb locus. Using a PCR assay based on the protocols established by van Dongen et al. [13] in combination with sequencing of specific amplification products, evidence for TCRb locus genomic Db1RJb (Figure 2A i) and Vb1R Jb (Figure 2A ii) recombination was found in both the monocyte and IFNc macrophage fractions. Length spectratyping of the antigenbinding complementary determining region 3 (CDR3) is a wellestablished method for the assessment of TCR repertoire diversity in defined variable chains [14]. Representative Vb13a CDR3 spectratype analysis in IL-4 or IFNc primed macrophages from healthy individuals (n = 3) revealed monoclonal and oligoclonal repertoires and varied in the same donor depending on IL-4 or IFNc activation ( Figure 2B). Sequencing of the expressed Vb13a CDR3 b clonotypes in one subject (GenBank Acc. No. JF923737-JF923744) indeed revealed marked differences between IL-4 and IFNc polarized macrophages ( Figure 2C). Of note, quantitation of the expressed length variants, respectively, in all three individuals consistently demonstrated increased repertoire TCR Va and Vb diversity in IFNc activated macrophages relative to monocytes and IL-4 macrophages ( Figure S2A).
Evidence for TCRb locus rearrangement in mature CD14 + monocytes and our previous observation of a rearranged TCRab in neutrophils [9] strongly suggested that TCR recombination occurs already at an early stage of myeloid development. To test this possibility, burst-forming unit-erythroid (BFU-E) and granulocyte/macrophage progenitor colonies (CFU-GM) were generated from CD34 + hematopoietic progenitor cells of two normal donors in two independent experiments. TCR Vb mRNA expresssion profiling was performed on 10 randomly selected

Author Summary
Infection with mycobacteria results in a host response which results in the formation of granulomas, highly organized structures characterized by the presence of macrophages, which are considered to rely solely on invariant immune receptors. On the other hand, the presence of variable immune receptors is required for granuloma formation but this process is not solely dependent on T cells. Furthermore, TNF is required for the maintenance of the mycobacterial granuloma structure in humans. We now find evidence for subpopulations of human and mouse macrophages that express variable ab T cell receptors (TCRab). Engagement of the macrophage-TCRab triggers CCL2 release and phagocytosis of baits directed to this receptor is enhanced. TCRab bearing macrophages accumulate in human tuberculosis granulomas and anti-TNF treatment of macrophages results in downregulation of the TCRab, which is associated with caspase 3 cleavage and suppression of TCRj. Anti-TNF treatment reduces mycobacteria induced cluster formation of TCRab positive macrophages, which is in line with reduced granuloma formation in rag1 -/-(T cell rag1 +/+ ) and TNF -/-(T cell TNF +/+ ) chimeric mice. Consequently, both chimeras show reduced CCL2 staining after mycobacterial infection. In summary, we have identified a recombinatorial immunoreceptor in monocytes/macrophages and demonstrate its implication in mycobacterial infection. Fluorescence immunocytochemistry demonstrating that a ,5% subpopulation of human peripheral blood CD14 + /MHC-II + monocytes expresses the TCRab. CD14-MACS purified peripheral blood monocytes were isolated from a healthy donor and double-immunostained with Abs against the TCRab (red) and MHC-II (green). Isotype controls for the anti-TCRab and anti-MHC-II antibodies are shown (right). Scale bars are indicated. Data shown are representative of n = 12 donors. (B) Flow cytometry of peripheral blood mononuclear cells from a healthy individual demonstrates the presence of the TCRb on the surface of a CD14 + monocyte subpopulation (3.5%, red arrows). Staining for TCRb and lineage surface markers are shown. CD14 + monocytes are in pink color, colonies from each individual. CDR3 b length spectratyping revealed expression of single or few rearranged Vb clonotypes in 50% (donor A) and 70% (donor B), respectively, of the CFU-GM analyzed ( Figure 2D, Figure S2B). No TCR gene expression was observed in any of the BFU-E colonies tested (data not shown). The majority of the CFU-GM displayed a monoclonal expression pattern consistent with the clonogenic nature of the myeloblasts and monoblasts in this assay. This indicates that TCRb locus rearrangement and expression of individual-specific Vb repertoires occurs already during the early phase of in vitro myeloid lineage differentiation.
We next sought evidence for TCR aand b-chain variability at the protein level. For this, the TCRab from macrophages of healthy donors was immunoprecipitated and subjected to MALDI-TOF mass spectrometry. Using this proteome profiling approach, we identified two peptides that showed partial sequence identity with known variable TCR a-chain fragments (Ja4, Ja39) and a total of four distinct peptides displaying partial sequence identity with variable TCR b-chain fragments (Jb1.1, Jb1.2, Jb1.4, Jb2.1) ( Figure 2E). Three of the peptides (Ja4, Jb1.4, Jb2.1) spanned VR J and JR C junctions indicating that they originated from rearranged TCR a and b loci. These TCRab proteome profiling results are consistent with the presence of multiple TCR aand b-chain variants in human macrophages.
In mice, TCR Vab mRNA profiling and CDR3 spectratyping of purified peritoneal macrophages confirmed the presence of diverse Va and Vb repertoires in wildtype macrophages. In contrast, no evidence for TCR Vab repertoire expression was found in macrophages from rag1 -/mice which are incapable of rearranging their immune receptor loci and thus lack a variable TCRab ( Figure 2F).
In summary, the combined results from TCRb VDJ rearrangement analyses, Vab CDR3 mRNA expression profiling and mass spectrometric TCRab peptide profiling indicate that the TCRab identified in human and murine macrophages is expressed as a variable receptor.

Macrophage-TCRab engagement induces CCL2 release
Given the presence of the TCRab ligand binding subunits in monocytes/macrophages, we tested whether these cells also express constituents of the TCR signaling pathway. RT-PCR expression profiling revealed expression of all critical components of the TCR signal transduction machinery including CD3f, ZAP70, LAT, Fyn, and Lck, respectively, in monocytes and macrophages from three randomly selected healthy donors ( Figure 3A). To explore whether the TCRab complex is operative in macrophages, we next tested whether specific TCR engagement has an impact on the secretion of effector molecules by IFNc macrophages. Analysis of the secretory pattern of a defined panel of cytokines, chemokines and growth factors (n = 15, Table S1, Figure S3) in response to canonical TCR stimulation with anti-CD3 antibodies revealed enhanced secretion of the major monocyte chemoattractant CCL2 (MCP-1) within 24 h ( Figure 3B). No detectable effect was observed for any of the other studied macrophage effector proteins or the secretory protein CCL5, which served as a marker for potential T cell contamination, indicating that engagement of CD3 dependent macrophage-TCRab signaling triggers selective secretion of CCL2 (Table S1, Figure S3). Consistent with this, anti-CD3 antibodies induced CCL2 gene expression in IFNc macrophages ( Figure 3C). Thus, specific engagement of the TCRab in macrophages induces gene expression and secretion of the monocyte chemoattractant CCL2 indicating that the TCRab signal transduction pathway in macrophages is functional.

The macrophage-TCRab modulates phagocytosis
Next, we investigated whether the TCRab interferes with the phagocytic activity of macrophages. For this, IFNc activated macrophages from two healthy donors were challenged with standardized phagocytosis baits (polystyrene beads, Ø 4.5 mm) for 15 min, 1 h and 10 h, respectively. To induce physical interaction of the baits with the macrophage-TCRab, the beads were coated with anti-TCRa/anti-TCRb antibodies ( Figure 4A). Identical beads coupled to equal amounts of nonspecific IgG antibodies or albumin (irrelevant protein) were used as controls. In addition, macrophages were challenged with albumin-coated beads in the presence of anti-TCRab antibodies that were not physically coupled to baits. As a positive control, baits were targeted to the known mediator of phagocytosis complement receptor 3 (CR3) utilizing antibodies to its subunit CD11b. Using this approach, we observed in both donors that the number of phagocytosing macrophages was significantly increased after 1 h when baits were directed to the TCR (1.4-3.0 fold vs. controls) ( Figure 4B). This increase was already detectable after 15 min, however, did not reach statistical significance for all controls. A similar augmentation of phagocytosis was seen when beads were targeted to the CR3 (1.2-4.5 fold vs. controls). In addition, we found that after 10 h baits directed to the TCR had elevated bead/cell ratios relative to controls (1.3-1.8 fold). These findings suggest that binding of baits to the TCRab facilitates phagocytosis. Consistent with this, we noted that phagocytosis was unaffected when anti-TCRab antibodies were not physically linked to beads ( Figure 4B,C). As expected, we found evidence for close proximity of ingested baits to the TCRab ( Figure 4C).
In light of the observation that the TCR facilitates phagocytosis, we next investigated whether and to which degreee complete CD3 + lymphocytes in blue. (C) (Left) Laser scanning cytometry (LSC) of unstimulated (naïve) monocyte-derived macrophages stained for TCRab (red). Nuclei are counterstained with DAPI (blue). The cytometric analysis shows a subpopulation (5%) with high fluoresence indicative of TCRab positive naïve macrophages (top right) which is highlighted in the histogram below (arrow). (Bottom) LSC of naïve and IL-4 (10 ng/ml) or IFNc (1000 U/ml) stimulated monocyte-derived macrophages, respectively, cultured for 6 days. The percentage of TCRab + cells in each macrophage population is shown for three healthy individuals. (D) Detection of the TCR aand b-chain in CD14 + monocytes and IFNc or IL-4 polarized macrophages by immunoblot. b-actin, loading control. (E) Immunogold electron microscopy demonstrating the presence of the TCR a-subunit on the cell surface of a human IFNc stimulated macrophage (arrows). (Right) isotype control. (F) Immunocytochemical double-staining reveals the presence of the TCRab (green) in alveolar macrophages from a 45 year old male with normal BAL cytology. Shown is a TCRab + alveolar macrophage (top right, arrow) next to a TCRab + T cell (asterisk). The merged image (bottom right) demonstrates that the majority of the cells express the macrophage marker CD163 (red). Giemsa-staining of the BAL cytospin preparation and isotype controls are shown in the left panel. The results are representative of three individuals. Nuclei (blue), DRAQ5. (G) Confocal immunofluorescence microscopy shows TCRab expression in murine macrophages. Spleen macrophages pooled from three normal C57BL/6 J mice were CD11b-MACS purified and immunofluorescence double-staining was performed using the anti-macrophage antibody F4/80 (red) and an anti-mouse TCRb antibody that recognizes a common epitope of the murine TCRab complex (green). The outlined area is shown at a higher magnification. Nuclei (blue) are counterstained with DRAQ5. Isotype controls are shown. (Bottom) RT-PCR demonstrating expression of the murine TCRa and TCRb constant chain genes in CD11b-MACS purified spleen macrophages (MW) from C57Bl6/J mice. Ly6G + neutrophils are shown as positive control. doi:10.1371/journal.ppat.1002375.g001 ablation of the TCR has a negative effect on the macrophage phagocytic capacity. For this, peritoneal macrophages from recombination defective rag1 -/mice (n = 7), which lack the TCRab ( Figure 2F) were incubated with FITC-labeled Mycobacterium bovis Bacille-Calmette-Guérin (BCG). We used mycobacteria as baits because they represent classical macrophage pathogens and, like CCL2, are critically implicated in macrophage-driven granuloma formation [15]. Consistent with the findings in the bead targeting experiments phagocytosis of M. bovis BCG was significantly reduced in rag1 -/macrophages compared to recombination-competent rag1 +/+ control mice (n = 7) after 6 hours of infection ( Figure 4D). Together, these bait targeting and ablation experiments strongly suggest roles of the TCRab in the regulation of macrophage phagocytic activity.

M. bovis BCG infection induces formation of macrophage clusters that express restricted TCR Vb repertoires in vitro
Phagocytosis of mycobacteria by macrophages with subsequent granuloma formation are key features of host defense in tuberculosis infection. Given this and the above results suggesting implication of the macrophage-TCRab in phagocytosis, we next tested whether mycobacterial challenge has an impact on TCR expression in macrophages in vitro. We infected macrophages with FITC-labeled M. bovis BCG and noted that within 4-6 days macrophages routinely formed clusters ( Figure 5). No clusters were formed in the absence of mycobacteria ( Figure 5B). Bacilli frequently formed aggregates which may reflect fragmentation of M. bovis BCG. We found evidence for co-localization of BCG and the TCRab by immunocytochemistry ( Figure 5A i) and immunogold electron microscopy ( Figure 5A ii), however, this phenomenon was rare. Quantitative analysis of BCG-infected IFNc macrophages from two healthy donors revealed a 4-fold increase of the percentage of TCRab + macrophages relative to uninfected controls (8 vs. 32.5) ( Figure 5B, Figure S4A). Consistent with this, we found increased expression of both the TCR a and the b constant chain genes in the macrophages that were infected with BCG. Moreover, CDR3 spectratyping of all 25 TCR Vb chains showed a noticeable but not significant increase in the number of expressed Vb CDR3 length variants (12 vs 18.5) in BCG-infected macrophages ( Figure 5C). Ex situ Vb clonotype analysis of randomly selected BCG/macrophage clusters from both donors revealed consistent expression of highly restricted TCR Vb chain repertoires ( Figure 5D, Figure S4B). In particular, we noted a bias toward the use of the Vb1 chain which was expressed by 10 out of the 11 (91%) BCG/macrophage clusters analyzed. Collectively, these results demonstrate that in vitro M. bovis BCG infection induces formation of TCRab bearing macrophage clusters that express highly restricted Vb repertoires.

Abundant accumulation of TCRab bearing macrophages in human granulomas of tuberculosis
To examine whether the macrophage-TCR is implicated in host defense against mycobacteria in vivo, we next screened for the presence of TCRab expressing macrophages in tuberculous tissue. Lung sections of patients with pulmonary tuberculosis (n = 13, Table S2) were immunostained for TCRaband the macrophage markers CD68 and CD163, respectively. Ten out of 13 patients showed abundant staining for TCRab in well-circumscribed caseous granulomas. Typically, the innermost segment of the epithelioid cell corona, which represents the front line of cellular defense against mycobacteria within tuberculous granulomas, exhibited intense TCRab + staining ( Figure 6A i-iii). Controls showed no staining for CD2 ( Figure 6A iv,v). Quantitative analysis of TCRabCD68 immunofluorescence double-staining revealed that on average 87% of the macrophages expressed the TCRab in this zone (Figures 6A vi, vii). Additional single immunofluorescence staining for CD68 and CD163 confirmed that the predominant cell type in the inner host-pathogen contact zone was macrophages ( Figure 6B, Figure S5A). Consistent with this, T cells and NK cells were typically localized in the peripheral corona zone of caseous granulomas as assessed by CD2 immunostaining ( Figure 6B). We found no evidence for TCRab expressing macrophages in a lymph node from an individual with reactivated M. tuberculosis infection triggered by anti-TNF therapy (adalimumab) ( Figure S5B). In keeping with immunohistochemistry, ex situ clonotype analysis revealed expression of TCR Vb mRNA repertoires in small clusters (20-30 cells) of immunostained CD68 + macrophages that were laser microdissected from the inner epithelioid cell zone ( Figure 6C). Similarly as observed in BCG infected macrophage clusters in vitro ( Figure 5D), we noted that the epithelioid zone macrophages predominantly express the TCR Vb1 chain.

TNF blockade suppresses macrophage-TCRab expression
The cytokine TNF is essential for host defense against mycobacteria and anti-TNF therapy may lead to disorganization of human tuberculous granuloma resulting in reactivation of latent tuberculosis [3,[16][17][18][19]. Since activation of the TCRab in macrophages results in CCL2 release, a key factor in granuloma (PBMC), positive control. HepG2 cells, Ø control. (B) Expression of individual-specific TCR Vb repertoires by monocytes, IL-4 macrophages (blue) and IFNc macrophages (red) from three healthy donors (1-3) representatively shown for Vb13a. A scaled synopsis of the three cell populations is shown at the bottom (g). (C) TCR clonotype analysis by sequencing of the antigen-binding CDR3 loop of Vb13a representatively shown for individual 2. IL-4 (blue) and IFNc activated macrophages (red) express completely different Vb13a clonotypes. The Vb13a chain is not expressed by the monocytes of this individual (cf. B). Colored letters represent deduced amino acid sequences of the newly identified CDR3 b regions (GenBank Acc. No. JF923737-JF923744). D) Expression of rearranged TCR Vb CDR3 clonotypes in granulocyte/macrophage progenitor colonies (CFU-GM) obtained from CD34 + progenitors of two healthy individuals (A and B). Filled boxes indicate positive expression of at least one of the 25 known human TCR Vb chains (xaxis) in a single colony. Colonies are identified by numbering on the y-axis. The repertoires for each of the expressed Vb chains were determined by length variant analysis of the antigen-binding CDR3 b region. The detailed Vb repertoire is representatively shown for colony CFU-GM2 (donor B). The repertoires of additional CFU-GM colonies are summarized in Figure S2B. RT-PCR lineage marker expression profiling documents the monocytic nature of this granulocyte/macrophage progenitor colony. CD2, CD8: T lymphoid markers; MMP25, MPO: granulocyte markers; CD14, CD68, CD163: monocyte markers. (E) Direct mass spectrometric identification of multiple TCR Va-and Vb-chain variants in human macrophages. Protein lysates from IFNc macrophages of a healthy donor were immunoprecipitated using an anti-TCRb antibody and the predicted 58 kD band (boxed) was analyzed by MALDI-TOF mass spectrometry. Peaks 1-6 represent TCR Va-and Vb-specific peptide fragments whose amino acid sequence identities with known TCR Vab-clonotypes are bolded. In three cases (2, 4 and 6), the identified peptides span VR J and JR C junctions (denoted by a gap) indicative of genomic rearrangements in the macrophage TCRa and -b loci. (F) Peritoneal macrophages from C57Bl6/J mice (rag1 +/+ ) but not recombination defective rag1 -/mice express Va (left) and Vb repertoires (right) as evidenced by TCR V-chain mRNA expression profiling (top) and CDR3 spectratyping of representative TCR Va-and Vb-chains (bottom). Peritoneal macrophages were pooled from three rag1 +/+ mice and an equal number of rag1 -/mice, respectively. doi:10.1371/journal.ppat.1002375.g002 formation, we examined whether TNFinhibition has a direct impact on the expression of the macrophage-TCRab. Uninfected or BCG infected macrophages were incubated in the presence of the anti-TNF antibody infliximab (50 mg/ml) or an isotype control antibody (anti-CD20, rituximab). Immunostaining revealed that TNF blockade inhibited TCRab expression relative to controls within 2h ( Figure 7A). Consistent with this, the inhibitory effect could be completely reversed by re-exposure of macrophages to TNFfor 24 h indicating that macrophage-TCRabexpression requires the presence of TNF ( Figure S6A). TNF blockade also had an inhibitory effect on TCR b-chain mRNA expression ( Figure 7B). Immunoblot revealed that TNF blockade not only suppressed expression of the TCRab ligand binding subunit but also that of the f-subunit (CD3) of the TCR signaling complex ( Figure 7C). The latter is essential for TCRab stabilization on the cell surface [20] and its degradation is mediated by caspase 3 [21]. Immunoblot analysis and immunostaining showed that anti-TNF treatment results in an increase in cleaved caspase 3 in uninfected and BCG infected macrophages indicating that TNF blockade induces cleaved caspase 3 ( Figure 7D, Figure S6B). Because TCRf is required for stabilizing TCRab on the cell surface these findings identify TNF as a regulator of macrophage-TCRab expression.

Loss of macrophage-TCR is associated with granuloma disorganization and reduced CCL2 expression in murine tuberculosis
It is well-established that abrogation of TNF results in defective tuberculous granulomas [7,8,22]. Given this and our finding that BCG induces TCRab expression, we next tested whether TNF blockade affects BCG triggered macrophage cluster formation in vitro. We found that treatment with the anti-TNF antibody infliximab significantly reduced the number and size of human macrophage clusters that formed during infection with M. bovis BCG in vitro ( Figure 8A). The presence of the TCRab in mouse macrophages offers the possibility to study the role of the TNF/ TCRab/ CCL2 regulatory axis we identified in human macrophages in tuberculosis in vivo. For this, we determined whether deletion of the macrophage-TCR impacts granuloma formation and macrophage CCL2 release in a murine model of M. tuberculosis infection. Wildtype (wt) mice (n = 5) and TCR deficient rag1 -/mice (n = 7) that were reconstituted with wt CD3 + T cells were infected via aerosol with ,100 M. tuberculosis bacilli [7]. The latter chimeric rag1 -/-(T cell wt) mice develop lung tuberculosis in the presence of intact T cells but absence of TCR bearing macrophages and T lymphocytes are routinely detectable in the tuberculomas [8]. Four weeks post adoptive T cell transfer and M. tuberculosis infection all wt mice displayed compact, well-delineated granulomatous lesions in their lungs that were predominantly composed of macrophages and lymphocytes ( Figure 8B i,ii). Rag1 -/-(T cell wt) chimeras lacking the TCR in their macrophages developed granulomatous foci containing macrophages and lymphocytes as seen in the wildtype mice. However, these lesions were generally diffuse ( Figure 8B iii,iv) and on average 1.5 fold larger than those of control mice indicating that rag dependent mechanisms outside the T cell system are required for proper granuloma formation in murine tuberculosis ( Figure 8C). Importantly, immunostaining revealed abundant CCL2 staining in the tuberculous lesions of wt mice ( Figure 8B i,ii) but CCL2 was routinely near absent in the chimeras lacking the macrophage-TCR ( Figure 8B iii,iv). Collectively, these results demonstrate in vivo that ablation of the variable macrophage immune receptor in murine lung tuberculosis is associated with suppression of CCL2 and defective granulomas formation.
We finally tested the impact of TNF abrogation on CCL2 expression and tuberculoma formation in TNF deficient mice that were reconstituted with wildtype CD3 + T cells (TNF -/-(T cell wt) mice). Infection with M. tuberculosis infection in these chimeras resulted in disorganized tuberculous granulomas in which CCL2 was consistently absent (Figure 8B v,vi). Similarly as in the rag1 -/-(T cell wt) chimeras, the granulomatous foci in TNF -/-(T cell wt) mice displayed a consistent increase in size (1.7 fold vs.  (Table S1) were determined in the supernatant by multiplex cytokine assay. The near absence of the T cell secretory protein CCL5 documents that macrophages were virtually free of T lymphocytes (bottom). Macrophages were collected from two healthy donors (ind 1, ind 2). +, CD3 + T cells (positive control). *p,0.05. (C) TCR engagement upregulates CCL2 gene expression in macrophages as assessed by quantitative RT-PCR (qPCR). The results shown represent the qPCR analysis of IFNc macrophages from three healthy donors that were stimulated with anti-CD3 antibodies for 24 h. doi:10.1371/journal.ppat.1002375.g003 controls) ( Figure 8C). This together with our observation that TNF blockade inhibits expression of the macrophage-TCR strongly suggests that an intact TNF/TCR macrophage pathway is required for CCL2 production in the tuberculous granuloma. Consistent with this, we found no CCL2 expression in the tuberculous lymphnode of a patient who received therapeutic anti-TNF treatment (adalimumab) ( Figure S7).

Discussion
In this study, we report the existence of an as yet unrecognized recombinatorial TCRab based immune receptor in monocytes/ macrophages (macrophage-TCRab) and provide evidence for its implication in a major infectious disease -tuberculosis. We find that monocytes and macrophages have rearranged Vb gene loci and consistently express diverse TCR Vab repertoires and show that macrophage-TCRab engagement induces the release of the monocyte chemoattractant CCL2 and demonstrate that the expression of the novel variable immune receptor depends on TNF. Furthermore, blockade of TNF in macrophages results in caspase 3 cleavage, TCRf degradation and TCRab downregulation.
Traditionally, variable immune receptors are thought to be restricted to cells of lymphoid origin [1]. However, recent work from our laboratories and others demonstrating TCRab and TCRcd expression in neutrophils and eosinophils [9][10][11], antigen receptor gene rearrangement in thymic granulocytes of mice [23], and NK cells which display immunological memory in response to viral infection [24] challenge this longstanding concept. These findings, which identify lymphoid features in myeloid cells, and the reciprocal demonstration that lymphocyte progenitors retain myeloid potential [25] and B cells in primitive vertebrates possess phagocytotic capabilities [26] extend the current concept of the strict dichotomy of the vertebrate immune system into nonspecific/ myeloid and antigen-specific/ lymphoid immunity [1] and suggest a closer kinship between both lineages than commonly appreciated. Combined in vitro evidence demonstrates a role for the macrophage-TCRab in the modulation of phagocytosis and the macrophage response during the initial phase of mycobacterial infection. In vivo, we find that the host/pathogen interface of human pulmonary tuberculomas is characterized by the massive presence of TCRab bearing macrophage-derived epithelioid cells and deletion of the macrophage-TCR in murine tuberculosis is associated with disorganized granuloma structure. Murine studies have previously suggested the possibility that variable immune receptors oustide the realm of T cells are involved in the control of mycobacterial infection [6][7][8]. Our demonstration that a TNFdependent variable macrophage immune receptor is part of the host defense against mycobacteria provides evidence for this concept. In particular, our results suggest that macrophage based antigen-specific host response mechanisms are operative in the development of the tuberculous granuloma and thus add an all new aspect to the current understanding of tuberculoma formation ( Figure 9). Furthermore, since it is well established that TNF is indispensable for the proper formation of tuberculous granulomas [7,8,22,27], the finding that TNF blockade leads to suppression of the macrophage-TCRab provides a novel potential mechanism that may underly the reactivation of tuberculosis during therapeutic anti-TNF treatment [5].
TNF dependent regulation of the macrophage-TCR implies the involvement of the TNF receptor 1 and/or TNF receptor 2 signaling pathways in the control of expression of the novel immune receptor. At this point, it is unclear at which level the TNF receptor signaling cascade interferes with the macrophage-TCR machinery. It will be challenging to identify signaling components that connect TNF to the TCR in macrophages. Also, it will be interesting to see if TCR activation has an impact on TNF expression.
Given the near-ubiquitous presence of macrophages and their involvement in acute and chronic inflammation, it is likely that the macrophage-TCRab is implicated in further pathologies beyond tuberculosis. Candidate diseases include chronic inflammatory diseases such as atherosclerosis and autoimmune disorders such as rheumatoid arthritis. Efforts are underway to further explore the biological function of the recombinatorial macrophage receptor in the immune response.

Isolation and culturing of monocytes/macrophages
Human monocytes/macrophages were isolated as previously reported [28] and purified by CD14 + Magnetic Cell Sorting (MACS) (Miltenyi Biotec, Bergisch Gladbach, Germany) Monocyte cell purity was routinely .99.5% as assessed by flow  (Table S2). Immunofluorescence double-staining for the TCRab (red) and the macrophage marker CD68 indicates that the TCRab + cells are macrophages (vi, yellow, merged). (vii) Mean percentage of CD68 + /TCRab + cells in the inner epithelioid cell corona of circumscribed caseous tuberculous granulomas from 8 different patients. Percentages of double positive cells in this zone (CD68 + /TCRab + cells : total number of nuclei) were determined from three sections of each patient. (B) Single immunostaining for the markers CD68 and CD163 demonstrates that macrophages represent the vast majority of cells in the epithelioid cell corona of the caseous tuberculous granulomas (i-iii). Staining for CD2 reveals that T cells and natural killer cells are predominantly located in the outer segment of the corona (iv). A focal accumulation of lymphocytes (L) located in the peripheral corona zone is shown in the right box (i,v). Note positive staining for CD2 (vi). Bars in all images span the inner host-pathogen contact zone. N, necrotic core. Isotype controls are shown in Figure S5A. ii-iv,vi: 40x. (C) Laser microdissection of a 20-30 cell cluster of CD68 + /TCRab + double-immunostained macrophages (encircled) from the inner host-pathogen contact zone of a caseous tuberculous granuloma. TCRab + , red; CD68 + , green; scale bar, 75 mm. Dissected cells are highlighted in yellow (left). Ex situ Vb CDR3 spectratype analysis of the excised cells reveals expression of restricted TCR Vb1 repertoires. doi:10.1371/journal.ppat.1002375.g006 cytometry ( Figure S1A) using the following antibodies: anti-CD14-FITC (DAKO, Glastrup, Denmark), anti-CD3-PE (BD Biosciences, Franklin Lakes, USA), anti-CD2-PE (BD Biosciences), mouse IgG1-FITC Isotype control and mouse IgG1-PE Isotype control. CD14 + monocytes were cultivated for 6 days in X-VIVO 10 serum-free and endotoxin-free medium (Cambrex, Verviers, Belgium) at a concentration of 5610 5 cells/ml in the presence of IFNc (1000 U/ml) (PeproTech, Rocky Hill, USA) or IL-4 (10 ng/ ml) (Tebu Bio, Frankfurt, Germany) to induce differentiation into Th1 and Th2 polarized macrophages. X-VIVO 10 is an optimal medium for cultivating human monocyte-derived macrophages [12,28] which does not contain exogenous growth factors or artificial stimulators of cellular proliferation. Burst-forming uniterythroid (BFU-E) and colony-forming units containing granulocytes and macrophages (CFU-GM) were generated from human CD34 progenitor cells as previously described [29].

Ethics statement
Infected granuloma tissue was obtained from the Department of Pathology, Universitä tsmedizin Göttingen, Georg-August-Univer-sity. Bronchial alveolar lavage fluid samples were excess material derived from three patients in the course of their medical care. The use of these specimens and mononuclear cells from healthy probands was approved by the Ethics Committee of the Faculty of Medicine Mannheim, University of Heidelberg, Germany and Ethics Committee of the Medical University of Göttingen, Germany. (Permit Number: 2007-254N-MA and 27/6/11). All patient samples were stricly anonymized. In accordance with the Declaration of Helsinki [30] no written informed consent was provided by study participants and/or their legal guardians.
The study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Mouse procedures performed in this study were conducted at the Centenary Institute, after protocol review and approval by the University of Sydney Animal Ethics Committee (K75/3-2004/3/ 3878) and at the animal facility of the Klinikum Göttingen Georg-August-Universitä t (Göttingen, Germany) according to the Deutsche Tierschutzgesetz (LAVES Niedersachsen A-008/09), after protocol review and approval by the University of Göttingen.  ) and BCG infected IFNc macrophages following TNF blockade (infliximab). Controls, untreated macrophages; +TNF, macrophages stimulated with TNF (10 ng/ml). Bar graphs (mean 6 SEM) represent quantitative analyses of qPCR or densitometry of the immunoblots from two independent experiments. Representative immunoblots are shown. Whole cell lysates from IFNc macrophages were immunoblotted using antibodies to TCRf, cleaved caspase 3 and b-actin, respectively. Quantitative densitometrical analyses of the immunoblots was normalized to b-actin. doi:10.1371/journal.ppat.1002375.g007 All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering by the attending veterinarian.
Animal models C57BL/6 mice and rag1 null mutant mice (rag1 -/-) were bred and maintained at the animal facility, Klinikum Göttingen Georg-August-Universitä t (Göttingen, Germany). Macrophages from the spleen were collected and purified by CD11b-MACS for further analyses. Thioglycollate-elicited mouse peritoneal macrophages were collected according to standard protocols. Mice used in the M. tuberculosis infection experiments were maintained under specific pathogen-free conditions in the Animal Facility of the Centenary Institute of Cancer Medicine and Cell Biology (Newtown, Australia). Genetically modified chimeric mice were generated by transfer of 1610 6 purified wildtype CD3 + T cells into TCR deficient rag1 -/mice and TNF null mice, respectively (n = 7 in each group) following whole body irradiation with 5 Gy as previously described [7]. At the time of T cell reconstitution wildtype, rag1 -/-(CD3 rag1 +/+ ) and TNF -/-(CD3 TNF +/+ ) chimeric mice, respectively, were infected via aerosol with , 100 M. tuberculosis bacilli (H37Rv) utilizing a Middlebrook airborne infection apparatus. After four weeks all mice were sacrificed and paraffin embedded lung sections were used for immunohistochemical analyses.

Flow cytometry
PBMC were isolated from freshly collected whole blood or buffy coats from healthy individuals by Ficoll density gradient centrifugation as previously described [9]. Cells were first fixed with 4% paraformaldehyde and then blocked with normal horse IFNc macrophages were infected with M. bovis BCG in the presence of the anti-TNF antibody infliximab (50 mg/ml) or an equal amount of an isotype antibody (anti-CD20) for the indicated timepoints. Bar graphs represent total numbers of small (,100000 pixels) and large macrophage clusters (100000-500000 pixels) that were formed. The area (number of pixels) of the macrophage clusters was determined from electronic images. The results (mean 6 SEM) are based on the analysis of three independent donors. ** p,0.01, ***p,0.001. (B) Lung sections of wildtype (wt) mice demonstrating formation of compact, well-circumscribed tuberculous lesions after four weeks infection with M. tuberculosis (i). A higher magnification of the boxed granuloma is shown in (ii). In contrast, chimeric rag1 -/-(T cell wt) mice that lack the macrophage-TCR but have intact T cells develop disorganized tuberculous lesions that are consistently larger and more diffuse than those of wildtypes (iii, iv higher magnification). All sections were immunostained for CCL2. Note intense CCL2 staining (brown) in the tuberculous lesions of wt mice but near absence of CCL2 in the macrophage-TCR deficient rag1 -/-(T cell wt) chimeras (i-iv). Chimeric tnf -/-(T cell wt) mice with systemic deletion of TNF but wildtype T cells display disorganized granulomas characterized by absence of CCL2 (v, vi) similarly as observed in macrophage-TCR deficient mice (iii, iv). All mice were infected via aerosol with , 100 M. tuberculosis bacilli at the time of adoptive T cell transfer. The lung sections shown are representative of 5-7 mice in each experimental group. Scale bars are indicated. (C) Increased size of tuberculous granulomas in chimeric rag1 -/-(T cell wt) and tnf -/-(T cell wt) mice relative to wildtype controls. Shown is the mean percentage of the lung area covered by granulomatous foci infection with M. tuberculosis. Data were calculated from scanned lung cross sections and are based on the analysis of five mice in each group. * p,0.05. doi:10.1371/journal.ppat.1002375.g008 serum (Jackson Immuno Research Laboratories). TCR staining was performed using the TCRbF1 antibody (clone 8A3), which was also utilized in the immunocytochemistry and immunoblot experiments. This framework antibody recognizes TCRab dimers. Mouse IgG1 and IgG2b antibodies serve as isotype controls and CD235a (purified) as a nonsense control antibody (all BD Bioscience). Antibody binding was detected with a secondary goat-anti-mouse FITC-labeled antibody (BD Bioscience). After saturating with normal mouse serum (Jackson Immuno Research Laboratories) the cells were stained with the following directly labeled antibodies: CD45 APC-H7, CD3 PerCP-Cy5.5, CD8-PE and CD14-APC (BD Bioscience). Analyses were performed on a FACS Canto using the FACS Diva software (BD Bioscience).

Laser scanning cytometry (LSC)
Monocytes/macrophages were stained with mouse anti-human TCRaF1/TCRbF1 antibodies (Thermo Scientific, Waltham, USA) in combination with Alexa Fluor 555 goat anti-mouse IgG (Invitrogen, Carlsbad, USA) and analyzed on an iCYS laser scanning cytometer (Compucyte, Cambridge, MA, USA). For assessment of apoptosis, annexin V/ PI staining was used. Contouring of cells was achieved by nuclear staining with DAPI. Photomultiplier tube settings for voltage, offset and gain were optimized ( Figure S1C). Data were acquired and analyzed using the iCYS cytometric acquisition and analysis software (CompuCyte). For statistical analysis, the entire area of the microscopic slide was scanned and for every event pictures of each channel (red, blue, and scatter) were recorded and merged within a gallery. The count settings selected for signal area of DAPI staining were within the range 15-150 mm 2 in order to exclude artefacts such as debris.

Protein isolation and immunoblotting
Modified RIPA buffer was used to extract whole cell lysates. SDS-PAGE and transfer to nitrocellulose membranes was conducted utilising the NuPAGE protein electrophoresis systems (Invitrogen). The primary monoclonal mouse anti-human TCRaF1/TCRbF1 antibodies (Thermo Scientific) and the polyclonoal antibodies for cleaved caspase 3 (ASP175) or CD3 zeta (Abcam, Cambridge, UK) were used. Mouse monoclonal antibodies to b-actin (Abcam) were used as loading control.

Protein identification by MALDI-TOF mass spectrometry
Immunoprecipitated TCRab bands were separated by SDS-PAGE and visualized by silver staining. The predicted 58kD single band was excised and destained. An in-gel digestion with trypsin was performed according to a standard protocol [31]. After o/n incubation at 37uC, peptides were extracted with a C18 affinity chromatography (ZipTip, Millipore) and eluted with 0.5% formic acid in 1:1 (v/v) water: acetonitrile. The eluate (1 ml) was mixed with saturated solution (9 ml) of a-cyano-4-hydroxycinnamic acid in 50% CAN and 0.1% TFA, spotted onto a steel target and the droplet was air dried prior to MS-analysis. Peptide mass fingerprinting was performed on a MALDI-TOF-MS (Autoflex II, Bruker Daltonics) operating in the reflector mode. The MS spectra for peaks in the range of 1-3.5 kDa were generated by summarizing 350 laser shots (50 laser shots at 7 different spot positions). Spectra were analyzed using the flexanalysis software (Bruker) [31,32]. Analysis of the MS spectra was performed utilizing the BioTools software (Bruker) in combination with an integrated online link to the Mascot database (www.matrixscience.com).
TCRb locus recombination assay DNA from 10 6 monocytes, IFNc-macrophages, HepG2 (negative controls) and PBMC (positive controls) was isolated using the Wizard Genomic DNA purification Kit (Promega). Screening for Db R Jb and Vb R Jb rearrangements at the TCRb locus was performed by PCR utilizing a modified non-multiplex approach according to the protocols by van Dongen et al. [13] and confirmed by sequencing. The customized primers used can be requested by the authors.

TCR CDR3 length spectratyping and qPCR
RNA from all monocyte/macrophage populations was prepared with TRI Reagent (Sigma) and transcribed into cDNA using the Reverse Transcription System (Promega). RT-PCR expression profiling of components of the TCR machinery and size spectratyping of the antigen binding TCR Va/Vb CDR3 regions were performed as previously reported [9]. Vab spectratypes of the human TCR CDR3 regions were assessed on a CEQ TM 8000 Genetic Analysis System (Beckman Coulter) using the D4-labeled primers D4-GCAGACAGACTTGTCACTGG (TCRa) and D4-TTGGGTGTGGGAGATCTCTGC (TCRb), respectively. To determine the detailed CDR3 clonotypes for Vb13a, specific RT-PCR amplification products were cloned into a pCR-TOPO vector (TOPO TA Cloning Kit, Invitrogen) using standard protocols. The cDNA sequences of the Vb13a CDR3 regions were analyzed from at least 10 randomly picked clones. qPCR for the TCRb constant chain and CCL2 was conducted using the IQ SYBR Supermix (Biorad, Hercules, USA). B2MG and GAPDH were used as housekeeping genes. Purity of monocyte/macrophage RNA was confirmed by PCR amplification of the leukocyte lineage markers CD2, CD8, CD14, CD64, CD68, CD163, MMP25 and MPO, respectively (exemplified in Figure 2D). Authenticity of all relevant PCR products was confirmed by sequencing. PCR runs were repeated at least twice. The sequences of additional PCR primers used in this study can be requested from the authors.

Immunoelectron microscopy
IFNc stimulated macrophages were incubated in the presence or absence of BCG for 6 h and subsequently fixed in 4 % formaldehyde and 0.2 % glutaraldehyde (0.1 M phosphate buffer). After washing the cells were scraped from the dish in 0.1 M phosphate buffer containing 1% gelatin, spun down and resuspended in 10% gelatin (0.1 M phosphate buffer) at 37uC. The cooled gelatin pellets were cut in small blocks, infiltrated in 2.3 M sucrose in 0.1 M phosphate buffer and mounted onto aluminum pins for ultramicrotomy before shock freezing. Ultrathin cryosections were picked in a 1:1 mixture of 2 % methylcellulose and 2.3 M sucrose. For immuno-labeling sections were incubated with monoclonal antibodies specific for TCRaF1 (Thermo Scientific) and rabbit anti-mouse IgG antiserum (Rockland, Gilbertsville, PA, USA) followed by protein A-gold (10 nm). Alternatively, the primary mouse antibody was detected utilizing goat anti-mouse antibodies conjugated to gold (Aurion, Wageningen, The Netherlands). Sections were analyzed using a LEO EM912 Omega transmission electron microscope (Zeiss, Oberkochen, Germany) and digital micrographs were obtained with an on-axis 204862048-CCD camera (Proscan, Scheuring, Germany).

Infection with M. bovis BCG and quantitation of BCG phagocytosis
Attenuated Mycobacterium bovis BCG (Bacillus Calmette-Guérin, BCG-medac) bacteria, strain RIVM, were re-constituted with the supplied 0.9% NaCl solvent (BCG-Medac, Hamburg, Germany) and used for infection of macrophages (MW:BCG = 1:10). For quantitation of phagocytosed bacilli, BCG mycobacteria were immunostained with a FITC-labeled antibody to M. tuberculosis (Acris). BCG phagocytosis was quantitated from a total of 20 randomly selected fluorescence microscopy images using the NIH image J software. The phagocytotic index (PI) was calculated as (percentage of macrophages containing at least one bacterium) x (mean area of bacterial staining per cell).

Assessment of BCG/ macrophage cluster areas
The areas (number of pixels) of the M. bovis BCG infected macrophage clusters were determined from electronic images using the image J software. From each individual 30 clusters were analyzed.

Clonotype analysis of M. bovis BCG infected macrophages
Isolated CD14 + cells were allowed to differentiate into macrophages for 6 days on semisolid agarose plates (0.4% agarose in X-VIVO 10 medium) in the presence of IFNc (1000 U/ml) and M. bovis BCG. Single foci were collected using a 1 ml syringe containing 100 ml PBS. cDNA was prepared using the RNeasy MircoKit and Sensiscript RT Kit (Qiagen, Hilden, Germany) and subjected to TCR Vb CDR3 spectratyping.

Cytokine analysis
For selective activation of the TCR, freshly obtained CD14 + monocytes from two healthy donors were allowed to differentiate into macrophages in the presence of IFNc. After 6 days IFNcpolarized macrophages were washed twice with X-VIVO-10 medium (Cambrex) and co-stimulated with an endotoxin-free mouse anti-human CD3 (1 mg/ml, Beckman Coulter) monoclonal antibody as previously reported [9]. The culture supernatants were collected at various timepoints and the release of a selected panel of cytokines, chemokines and growth factors (n = 15) ( Table S1) was assessed utilizing a customized Luminex-based multiplex Procarta Cytokine Assay Kit (Multimetrix, Heidelberg, Germany). All measurements were conducted at least in duplicate. In addition, CCL2 (Thermo Scientific) and CCL5 (Qiagen) in the supernatants were quantified using specific ELISAs.

Bead phagocytosis assay
For phagocytosis of baits targeted to the TCRab, IFNcpolarized macrophages were co-incubated with polystyrene bead baits (Ø 4.5 mm, Invitrogen) coated with anti-TCRa/anti-TCRb, anti-CD11b (BD Biosciences), nonspecific IgG isotype control antibodies or albumin for 15 min, 1 h and 10 h, respectively, at 37uC (MW:beads = 1:1, 5 mg protein/10 7 beads). In addition, phagocytosis of albumin-coated beads was assessed in the presence of uncoupled anti-TCRa/anti-TCRb antibodies. Quantitation of phagocytosed beads was conducted by bright field microscopy of at least 12 randomly selected fields of vision performed by two independent observers.
Murine proteins (Swiss-Prot). TCRa: M64239, TCRb: M26057; M26053, Rag1: AAP76401, F4/80: Q3S4B0, TNFa:P06804.  Figure S3 Effect of CD3 mediated TCR activation on cytokine release from macrophages. Aliquots of 56105 IFNc macrophages were incubated with soluble antibodies to CD3, isotype control antibodies (I) or in the absence of antibodies (-) for the indicated timepoints as in Figure 3B. Cytokines were determined in the supernatant by multiplex cytokine assay. The results are summarized in Table S1. Macrophages were collected from two healthy donors (ind 1, ind 2). (TIF)  Figure 6B. Bars in both images span the inner hostpathogen contact zone. N, necrotic caseous core. 40x. (B) Immunofluorescence microscopy of two tuberculous granulomas present in the lymph node of a patient with therapeutic anti-TNF treatment (adalimumab). Paraffin sections of the granulomas were double-stained for TCRab Alexa 555, red) and the macrophage marker CD68 (FITC, green). Nuclei are DAPI-stained (blue). Shown are merged images. Scale bars are indicated. Note the consistent absence of TCR bearing macrophages (TCRab+/CD68+, yellow) within the granulomas. The AFB (Acid.Fast Bacilli) test was used to confirm active infection with acid fast mycobacteria in the patient.