A Novel Mechanism of Soluble HLA-G Mediated Immune Modulation: Downregulation of T Cell Chemokine Receptor Expression and Impairment of Chemotaxis

Background In recent years, many immunoregulatory functions have been ascribed to soluble HLA-G (sHLA-G). Since chemotaxis is crucial for an efficient immune response, we have investigated for the first time the effects of sHLA-G on chemokine receptor expression and function in different human T cell populations. Methodology/Principal Findings T cell populations isolated from peripheral blood were stimulated in the presence or absence of sHLA-G. Chemokine receptors expression was evaluated by flow cytometry. sHLA-G downregulated expression of i) CCR2, CXCR3 and CXCR5 in CD4+ T cells, ii) CXCR3 in CD8+ T cells, iii) CXCR3 in Th1 clones iv) CXCR3 in TCR Vδ2γ9 T cells, and upregulated CXCR4 expression in TCR Vδ2γ9 T cells. sHLA-G inhibited in vitro chemotaxis of i) CD4+ T cells towards CCL2, CCL8, CXCL10 and CXCL11, ii) CD8+ T cells towards CXCL10 and CXCL11, iii) Th1 clones towards CXCL10, and iv) TCR Vδ2γ9 T cells towards CXCL10 and CXCL11. Downregulation of CXCR3 expression on CD4+ T cells by sHLA-G was partially reverted by adding a blocking antibody against ILT2/CD85j, a receptor for sHLA-G, suggesting that sHLA-G downregulated chemokine receptor expression mainly through the interaction with ILT2/CD85j. Follicular helper T cells (TFH) were isolated from human tonsils and stimulated as described above. sHLA-G impaired CXCR5 expression in TFH and chemotaxis of the latter cells towards CXCL13. Moreover, sHLA-G expression was detected in tonsils by immunohistochemistry, suggesting a role of sHLA-G in local control of TFH cell chemotaxis. Intracellular pathways were investigated by Western Blot analysis on total extracts from CD4+ T cells. Phosphorylation of Stat5, p70 s6k, β-arrestin and SHP2 was modulated by sHLA-G treatment. Conclusions/Significance Our data demonstrated that sHLA-G impairs expression and functionality of different chemokine receptors in T cells. These findings delineate a novel mechanism whereby sHLA-G modulates T cell recruitment in physiological and pathological conditions.


Introduction
The classical HLA molecules, also known as HLA-class Ia molecules, are extremely polymorphic molecules belonging to the immunoglobulin superfamily [1]. HLA-class Ia molecules are widely but not ubiquitously expressed. Each HLA class Ia molecule consists of a single heterotrimer of heavy chain, b2microglobulin and a peptide epitope of eight to ten amino acids embed in the peptide-binding groove of the heavy chain. Most peptide epitopes are derived from proteins that are synthesized in the cell, digested by antigen-processing machinery and loaded into the peptide-binding groove. These peptides are presented to antigen-specific T cells through the interaction with T-cell receptor, leading to the killing of cells that are infected with viruses or intracellular bacteria, or tumor cells [1].
Like HLA class Ia molecules, ''non-classical'' HLA class Ib molecules HLA-E, -F, -G and H can associate with b2microglobulin (b2m) and can present peptides. However, in contrast with HLA-Ia molecules, class Ib molecules are oligomorphic, with only few alleles. So far, limited information are available about the function of HLA-F and HLA-H. HLA-E is able to present peptides derived from the leader sequence of HLA-class Ia molecules and HLA-G, giving a ''self signal'' in cells that express HLA-class I molecules, and may have a role in autoimmunity [2]. HLA-G is expressed not only as a cell surface bound molecule, but also as a soluble moiety in body fluids [3,4,5]. Seven different isoforms encoded by alternative splicing of the same mRNA, that include membrane-bound HLA-G1, HLA-G2, HLA-G3, and HLA-G4 and soluble secreted HLA-G5, HLA-G6, and HLA-G7. The major isoforms of HLA-G present in serum are soluble HLA-G1 and HLA-G5 generated either by shedding or proteolytic cleavage of the membrane bound isoform or by secretion of a soluble isoform [6].
The physiological role of this molecule is to establish immune tolerance at the maternal-fetal interface, abrogating the activity of maternal NK cells against fetal tissue [7]. HLA-G can also present peptides, but it is still unclear whether these peptides are important for host defense against pathogens or they act to stabilize the surface expression of HLA-G [2].
Many immunoregulatory functions have been described in the last years for HLA-G molecules, in particular on T cells, B cells, NK cells and antigen presenting cells. HLA-G molecules induce apoptosis [4], inhibit cell proliferation [8], cytotoxicity [5] and differentiation [9], and modulate cytokine release [10].
HLA-G binds four receptors, i.e. ILT2 (immunoglobulin (Ig)like transcript 2)/CD85j, ILT4 (Ig-like transcript 4)/CD85d, KIR2D4L (killer inhibitory receptor)/CD158d. ILT2 is broadly expressed by lymphoid and myeloid cells (T and B cells, NK cells, dendritic cells, monocytes, macrophages), whereas ILT4 is myeloid-specific (monocytes, macrophages, dendritic cells), and KIR2DL4 is expressed only on NK cells. ILT2 and ILT4 interact with HLA-G, but also with classical HLA class I molecules (with lower affinity), whereas KIR2DL4 can interact only with HLA-G [11]. All these receptors are expressed at low levels by resting cells and are up-regulated on activated cells or in pathological condition, i.e. viral infection [12].
Moreover, LeMaoult et al. have demonstrated that HLA-G can up-regulate the expression of its own receptors, since these receptors were found to be overexpressed in the same pathologies in which HLA-G is upregulated (i.e. AIDS, tumors, autoimmune diseases) [13].
CD160 a glycosylphosphatidylinositol-anchored member of the immunoglobulin superfamily, have been recently described as a receptor for HLA-G. Fons et al. have demonstrated that interaction between soluble HLA-G and CD160 on endothelial cells lead to apoptosis of the latter cells and inhibition of angiogenesis [14]. However, CD160 is expressed not only on endothelial cells, but also on NK cells, NKT cells, T cells [14].
Chemotaxis of immune effector cells is a crucial event for the recirculation of these cells between lymphoid organs and inflamed tissues [15]. Chemotaxis is mediated by the interaction of chemotactic cytokines (chemokines) with their receptors. Chemokines are highly basic proteins of 70-125 amino acids with molecular masses ranging from 6 to 14 kDa and 20-95% sequence identity each other, with 4 highly conserved cysteine residues. Four chemokines subfamilies have been defined according to the number of amino acids between the first two cysteine residues: CC chemokines, CXC chemokines, XC chemokines and CX3C chemokines. Chemokine receptors are heptahelical G protein-coupled receptors, with a single polypetide chain of 340-380 aminoacids spanning 7 times the surface membrane and with 25-80% sequence identity each other. All receptors display an extracellular acidic N-terminal domain, a serine/threonine-rich intracellular C-terminal domain, and two disulfide bonds inbetween the N-terminal domain and the second extracellular loop [16].
Intracellular signalling through G-proteins leads to extension of lamellipodia through cytoskeletal restructuring, shape changes and firm adhesion (chemotaxis), release of oxygen radicals, histamine and cytotoxic granules from granulocytes, and finally modulation of gene transcription. A given chemokine receptor can be inactivated at least through two different mechanisms: i) strong or prolonged chemokine stimulation, leading to uncoupling of Gproteins, binding of arrestin, and internalization of the receptor, or ii) stimulation with an unrelated chemoattractant resulting in the activation of PKA and PKC, phosporylation of the receptor and its inactivation without internalization [17]. Chemokine receptor expression can be modulated by pro-inflammatory (i.e. LPS, TNF, IL-1, IFN-c) and anti-inflammatory (glucocorticoid hormones, IL-10) stimuli, and by cytokines such as IL-4 and IL-13, leading to a modulation of immune cells recruitment at sites of infection and inflammation [18].
So far, no information is available on possible effects of sHLA-G molecules (and other HLA-I molecules) on chemotaxis of immune effector cells.
We have investigated this issue on different T cell populations, not only on classical CD4 + and CD8 + T cells, but also on TCRcd + T cells and on follicular helper T cells (T FH ).
TCRcd T cells circulate in peripheral blood (3-5% of PBMNC) and express a peculiar TCR composed of c and d chains, that recognize phosphoantigens. Most circulating TCRcd T cells show the Vd2c9 rearrangement [19]. T FH are a small subset of TCR ab T cells present in peripheral blood [20] and secondary lymphoid organs [21]. These cells, that express CXCR5 and ICOS, gain access to the germinal center of secondary lymphoid follicles where they exert a potent helper function for B cell differentiation to antibody secreting cells through IL-21 production [22].
Our data describe a novel immunoregulatory property of sHLA-G molecules, based on the inhibition of chemokine-driven migration of different T cell populations together with downregulation of the expression and function of selected chemokine receptors on these cells.

Soluble HLA-G impairs the expression of different chemokine receptors in different T cell populations
We first evaluated by flow cytometry the expression of a panel of CC-and CXC-chemokine receptors in the two major subsets of circulating T lymphocytes, CD4 + and CD8 + T cells. These receptors were selected on the ground of their physiological relevance [23]. Cells were polyclonally stimulated with anti-CD3 mAb, anti-CD3/CD28 coated beads or PHA, in the presence or absence of sHLA-G. All these stimuli produced superimposable results. Therefore, from now onwards, all the experiments described will be those performed using anti-CD3 mAb. The effect of sHLA-G on chemokine receptor expression was lower on resting T cells (not shown).
As shown in Figure  Next, we investigated the effects of sHLA-G on three polarized subsets of CD4+ T cells, i.e. Th1, Th2 and Th17 cells. These subsets were identified by intracellular staining and flow cytometric analysis according to cytokine production profiles (not shown). In particular, Th1 and Th2 cell clones were .95% IFN-c + /IL-4 2 and .95% IFN-c 2 /IL-4 + , respectively. Bulk Th17 cells enriched from PBMNC were .50% IL17 + . Chemokine receptor expression in the latter cells was analyzed by flow cytometry gating on cells expressing IL-17A. All these T cell subsets were stimulated in the presence or absence of sHLA-G, and the expression of distinctive chemokine receptors was assessed by flow cytometry [24].
Among Th1-associated chemokine receptors, sHLA-G significantly downregulated the expression of CXCR3 (mean MRFI6SD: 19,58619 vs 3,0360,9 p = 0.03) in Th1 clones, whereas CCR5 and CXCR6 were not affected (Fig. 2, panel A). The expression of three Th2-associated chemokine receptors, CCR3, CCR4 and CCR8 was not affected by sHLA-G in Th2 clones (Fig. 2, panel B). In Th17 T cells, sHLA-G treatment did not modulate expression of CCR6 or CCR7, two Th17 associated chemokine receptors (Fig. 2, panel C). Mean results from five different experiments 6SD are shown in Figs. 2D, E and F. We next investigated modulation of a wide panel of CC-and CXC-chemokine receptors in TCR Vd2c9 T cells by sHLA-G (Fig. 2

Chemotaxis of different T cell populations is dampened by soluble HLA-G treatment
The above T cell populations but Th2 and Th17 cells were next subjected to in vitro migration assays using ligands of the chemokine receptors modulated by sHLA-G treatment.

sHLA-G is expressed in secondary lymphoid follicles and impairs in vitro chemotaxis of T FH cells through downregulation of CXCR5 expression
All T cell populations here investigated recirculate from peripheral blood to secondary lymphoid organs. sHLA-G is secreted by different cell types, such as monocytes, dendritic cells and endothelial cells and is detected in sera from normal donors [25].
T FH cells display an unique homing pattern to secondary lymphoid organs where they are attracted to the germinal centres of secondary lymphoid follicles by a gradient of CXCL13 [26].
We evaluated by flow cytometry the expression of two distinctive chemokine receptors, i.e. CXCR5 and CCR7, in T FH cells stimulated in presence or absence of sHLA-G. As shown in To investigate the physiological relevance of our findings, we next evaluated the expression of HLA-G in tonsil tissue sections by immunohistochemical staining with an anti HLA-G1/G5 mAb. As shown in Fig. 4, panel D and E, numerous HLA-G1/G5 + cells were detected in germinal centres and sub-epithelial areas (arrows), whereas such cells were virtually absent from the follicular mantle of secondary lymphoid follicles. These findings suggest that trafficking of T FH cells in germinal centres may be physiologically modulated by sHLA-G. The inhibitory receptor ILT2/CD85j is involved in sHLA-G mediated inhibition of T cell chemotaxis The expression of three receptors for sHLA-G on the T cell surface was next investigated by flow cytometry. As shown in The expression of the sHLA-G receptors investigated was not modulated by sHLA-G treatment (data not shown).
To assess the role of ILT2/CD85j in the inhibition of T cell chemotaxis by sHLA-G, we performed five experiments in which CD4 + T cells were stimulated, and blocking antibodies against ILT2/CD85j or ILT4/CD85b, respectively, or isotypic control were added. CXCR3 expression was evaluated by flow cytometry as read-out of the experiment since this receptor was consistently downregulated by sHLA-G in different T cell populations.
As shown in Fig. 5, panel C, inhibition of CXCR3 expression on CD4 + T cells by sHLA-G was significantly reverted by adding a blocking mAb against ILT2/CD85j, as compared with isotypic control (% of inhibition 31,46 vs 47,55, p = 0,0052); conversely, no effect was observed in the presence of a blocking mAb against ILT4/CD85b. These data indicated that sHLA-G inhibited T cell chemokine receptor expression by interacting mainly with ILT2/CD85j, although other sHLA-G receptor(s) are likely involved.

Pathways involved in sHLA-G driven intracellular signaling in T cells
The signaling pathways modulated by sHLA-G in stimulated CD4 + T cells was next investigated by Western-blot analysis of a panel of housekeeping and phosphorylated (p) proteins involved in HLA-G/ILT2 signaling [27].  A representative experiment out of three performed is shown in Fig. 6, panel A. Quantification of proteins have been performed by densitometry autoradiography films. Results are summarized in Table 1.
p-Stat5 and p-p70 s6k, but not Stat5 and p70 s6k, respectively, were strongly downregulated by sHLA-G treatment. Up-regulation of p-SHP-2 and p-b-arrestin, but not of the corresponding housekeeping proteins, was detected in sHLA-G treated CD4 + T cells (Fig. 6, panel A). sHLA-G treatment decreased levels of p-PKC-d and increased those of the corresponding housekeeping protein (Fig. 6, panel A).
Expression of GSK-3b and mTOR, but not that of their phosphorylated forms, was downregulated by sHLA-G. Expression of either p-Stat3 or Stat3 was unaffected by sHLA-G (Fig. 6,  panel A).
Finally, no modulation of AKT (MRFI 189,97 vs 157) and p-AKT (MRFI 1,41 vs 1,25) expression by sHLA-G treatment was detected by flow cytometry. A representative experiment out of three performed is shown in Fig. 6, panel B.
Our data indicated that multiple intracellular signaling pathways in T lymphocytes can be affected by sHLA-G treatment.
We next analyzed whether hyper-phosphorylation of SHP-2 phosphatase was a crucial event in sHLA-G induced downregulation of chemokine receptors. Again, CXCR3 expression was analysed by flow cytometry as read-out of the experiments. Fig. 6, panel C, shows a representative experiment out of the three performed in which downregulation of CXCR3 expression in CD4 + T cells induced by sHLA-G was completely reverted when CD4 + T cells were pre-treated with PHPS1, a specific inhibitor of SHP-2 [28] (mean MRFI6SD: CTR 17,560,5; HLA-G 2.4660,5; HLA-G + PHPS1 5,3960,53; p = 0,05).

Discussion
In this paper, we describe a novel mechanism of immune modulation operated by sHLA-G molecules. This mechanism involves phenotypic downregulation of some chemokine receptors CD4 + T cells orchestrate adaptive immune responses by promoting B cell activation and differentiation to antibody secreting cells, maturation of cytotoxic T cells, activation of NK cells and macrophages. Activated CD4 + T cells are attracted by chemokine gradients to inflammatory sites where they migrate across endothelial vessels [29]. This latter process involves CD4 + T cell expression of different adhesion molecules, such as selectins and integrins, and different chemokine receptors [29]. Under polarizing antigenic stimulation, activated naïve CD4 + T cells differentiate into discrete subsets with distinctive patterns of cytokine production and effector functions [24]. These subsets display different chemotactic properties in vitro and in vivo related to the expression of distinctive sets of chemokine receptors [30].
In particular, i) Th1 T cells express CXCR3, CCR5 and CXCR6 and accumulate in inflammatory sites following gradients of chemokines such as CCL2 and CXCL10 [31]; ii) Th2 T cells express CCR3, CCR4 and CCR8 and accumulate in inflammatory sites following gradients of chemokines such as CCL17 and CCL22 [31] and iii) Th17 cells express CCR6 and CCR7 and are attracted by CCL20 and CCL19 or CCL21, respectively [32].
Here we show that CXCR3 expression was strongly downregulated by sHLA-G in total CD4+ T cells and Th1 cell clones, and accordingly chemotaxis of these cell fractions to the CXCR3 ligands CXCL10 and CXCL11 was significantly impaired by sHLA-G. In contrast, sHLA-G treatment did not affect CCR3, CCR4 or CCR8 expression by Th2 cell clones and CCR6 or CCR7 expression by Th17 cells.
Failure of sHLA-G to modulate Th2 cell associated chemokine receptors is consistent with the recent demonstration of high levels of sHLA-G in sera from allergic patients supporting the hypothesis that sHLA-G contributes to maintain Th2-polarized immune responses [33].
Th17 cells play a pivotal pathogenic role in mouse models of human rheumatoid arthritis [34], multiple sclerosis [35] and in patients with Crohn disease [36], in which CCR6 is crucial for Th17 cell attraction to sites of inflammation. Here we show for the first time that sHLA-G does not alter CCR6 expression in Th17 cells, suggesting that CCR6 driven chemotaxis is not amenable to inhibition mediated by sHLA-G.
We have here addressed also the effects of sHLA-G molecules on two populations of cytotoxic T cells, i.e. CTL and TCR Vd2c9 T cells, both of which control infections mediated by intracellular pathogens. In particular, CTL are specialized in the protection from viral infections [37], while TCR Vd2c9 T cells are mainly involved in protection from mycobacterial infections [38]. In analogy to that observed with CD4 + T cells and Th1 cell clones, both CTL and TCR Vd2c9 T cells showed strong downregulation of CXCR3 by sHLA-G associated with inhibition of chemotaxis to CXCL10 and CXCL11. The finding that CXCR4 expression was up-regulated by sHLA-G in TCR Vd2c9 T cells without an increase of the migration of the latter cells towards CXCL12 was not surprising, since several examples in the literature have shown that modulation of chemokine receptor expression is not always coupled with changes in receptor function [39].
sHLA-G is detected in sera from normal donors and is physiologically produced by myeloid cells such as monocytes and dendritic cells, as well as by activated endothelial cells [25]. Furthermore, serum sHLA-G concentration is modulated in different pathological conditions. In particular, serum sHLA-G levels increase in patients with solid and haematological tumors, allergy and viral infections, and decrease in patients with autoimmune disorders [25]. Moreover, the concentration of sHLA-G tested in our experiments in vitro (100 ng/ml) was observed in sera from patients with solid and haematological malignancies, whereas concentration of sHLA-G in sera healthy subjects is in general below 20 ng/ml [40,41]. Therefore, our results may have patho- physiological relevance since disease-related fluctuations in sHLA-G serum levels may translate into modulation of T cell migration to inflamed tissues. In this connection, we have demonstrated that CXCR3 is the main target of sHLA-G mediated inhibition in different T cell populations. CXCR3, together with CCR5, controls Th1 cell migration [42] and is therefore involved in a broad spectrum of functional activities of these cells. Th1 cells collaborate with TCR Vd2c9 T cells in the control of mycobacterial infections [43]. Therefore, it is tempting to speculate that sHLA-G, by dampening CXCR3 mediated chemotaxis, modulates the recruitment of these cell populations to granulomatous mycobacterial lesions, perhaps limiting excessive tissue damage. Likewise, in the course of viral infections, attraction of CTL to infected tissues by CXCR3 ligands may be modulated by sHLA-G whose serum levels are increased in patients with different types of viral infection.
Another implication of our results deals with anti-tumor immune responses in which both CTL and TCR cd T cells play important roles. Attraction of these cells to the tumor mass is operated in part by CXCR3 ligands expressed in variable amounts in the tumor microenvironment by tumor cells or stromal cells [44]. Since sHLA-G levels are strongly increased in many malignancies, dampened recruitment of CTL and TCR cd T cells to the tumor site is likely to take place in vivo and provide a novel mechanism of sHLA-G related immunosuppression.
Several examples in literature support our data. It has been reported that CXCR3 is downregulated in several pathological conditions, in which serum sHLA-G levels are increased, such as allergy [45], T cell lymphoma [46]and multiple sclerosis [47]. Conversely, CXCR3 has been found to be upregulated in diseased associated with decerased levels of serum sHLA-G, such as rheumatoid arthritis and lupus [48].
Here we show for the first time that CXCR5 expression and CXCL13 driven chemotaxis were strongly downregulated by sHLA-G in T FH cells, a subset of CD4 + cells coexpressing CXCR5 and ICOS that circulate from peripheral blood to secondary lymphoid organs, that may be developmentally related to Th1, Th2 or Th17 cells [49]. T FH cells exert a potent helper function for centrocyte differentiation to plasma cells in the light zone of the GC, where we detected expression of sHLA-G molecules that was also observed in the subepithelial area of tonsil.
Taken together, these findings support the conclusion that migration of T FH cells in the light zone of the germinal centre may be modulated by a gradient of sHLA-G and this can in turn interfere in T FH -dependent plasma cell differentiation. Based upon recent studies of dynamic imaging of GC cell trafficking in mice, it has been proposed that positive selection of GC B cells depends not only on recognition of antigen presented by follicular dendritic cells, but also on competition for help provided by T FH cells [50]. Thus, in principle, modulation of CXCR5 expression and function by sHLA-G may impact indirectly on GC B cell positive selection.
sHLA-G mediated inhibition of T cell chemokine receptor expression and chemotaxis was found to depend mainly on the inhibitory receptor ILT2/CD85j. The signalling pathway initiated by sHLA-G interaction with ILT2/CD85j involved modulation of phosphorylation of SHP-2, Stat5 and p70 s6k. sHLA-G induced over-phosphorylation of SHP-2 together with reduced phosphorylation of Stat5. SHP-2 is a tyrosine phosphatase that upon phosphorylation by ligand-stimulated inhibitory receptors such as ILT2/CD85j and ILT4/CD85b can de-phosphorylate and inactivate Stat5, leading to reduced transcription of several [51] genes including some implicated in cell motility [52]. Indeed, we demonstrated that SHP-2 plays a pivotal role in sHLA-G induced downregulation of T cell CXCR3 using the specific SHP-2 inhibitor PHPS1 [28].
Numerous signalling pathways converge on p70 s6k that, upon phosphorylation, participate in the control of cell cycle progression [53]. We also demonstrated that sHLA-G induced overphosphorylation of b-arrestin, that in its phosphorylated form binds chemokine receptors and promotes their internalization, thus preventing further interactions with their ligands [54]. Surprisingly, sHLA-G modulated expression of the housekeeping but not the phosphorylated mTOR, GSK-3b and PKC-d proteins.
The functional significance of these findings is unknown.
The main findings of this study have been summarized in Figure 7.
In conclusion, we have shown that sHLA-G performs a novel immunomodulatory effect on T cells, downregulating the expression and function of chemokine receptors. Such inhibition can affect the recirculation of T lymphocytes between peripheral blood, secondary lymphoid organs and inflamed tissues, leading to important repercussions on their effector functions.

Cell isolation and culture
The study was approved by the Ethical Committee of the G. Gaslini Institute, Genoa, Italy. Surgically removed tonsils and normal peripheral blood (PB) samples were obtained following written informed consent. Mononuclear cells (MNC) were isolated by Ficoll-Hypaque (Sigma) density gradient.
T cell clones were generated by cloning at limiting dilution as previously described [56].
Th1 T cell clones were obtained from normal donors' PBMNC after stimuation with Candida Albicans bodies for 6 days as reported [57]. Th2 T cell clones were obtained from PBMNC of donors with allergic sensistization to Dermatophagoides Pteronissinus (DP) after stimulation with DP extract (Stallergenes, France) 20 mg/ml for 6 days as reported [57]. After stimulation, T cell blasts were separated using PERCOLL (GE Healthcare) density gradient, and were then expanded in RPMI 10% FBS +200 U/ml IL-2 (Chiron, Milano, Italy). Th17 T cells were obtained from normal donors' PBMNC using Th17 expansion kit (Myltenyi Biotec) following manufacturer's protocol.
T FH cells were isolated from human tonsils mononuclear cells by immunomagnetic selection of ICOS + T cells, using anti-ICOS Histogram summarized data obtained in Western Blot analysis on CD4 T cells stimuated in the presence or absence of sHLA-G. Densitometry of autoradiography films was performed and relative density (RD) was calculated for control (ctr) and sHLA-G as follows: (density of specific band/ density of GAPDH band). Change induced by sHLA-G was calculated as follows: (RD ctr-RD sHLA-G). % fold change induced by sHLA-G treatment was calculated as follows: [(RD sHLA-G -RD ctr)-1/RD ctr 6100]. doi:10.1371/journal.pone.0011763.g007 mAb (Santa Cruz Biotechnology, CA, USA) and anti-mouse IgG1 microbeads (Myltenyi Biotec). These fractions contained .95% of CD4 + /ICOS + /CXCR5 + T FH cells.
TCR TCRcd T cells circulate in peripheral blood (3-5% of PBMNC) and express a peculiar TCR composed of c and d chains, that recognize phosphoantigens. Most circulating TCRcd T cells show the Vd2c9 rearrangement [17]. TCR Vd2c9 T cells were obtained by stimulating PBMNC from normal donors with 5 uM n-bisphosphonate zoledronate (Novartis) for 7 days, as described [58]. After stimulation, 92% of these cells were TCR Vd2c9 + T cells.
All these T cell populations were stimulated in vitro for 48 h in RPMI 10% FBS at 37uC and 5% CO 2 with anti-CD3 mAb (OKT3, coated O.N. on 96 well plates) in the presence or absence of sHLA-G (100 ng/ml) before being subjected to flow cytometric analysis or in vitro migration assay.

Antibodies and flow cytometry
The following mAbs were used: anti-CXCR4 PE (clone #12G5), Cells were stained with fluorochrome-conjugated mAbs or with isotype and fluorochrome-matched control antibodies, and were run on a FACSCalibur (Becton Dickinson). 10 4 events were acquired and analyzed using the CellQuest software (Becton Dickinson).
The specific SHP-2 phosphatase inhibitor PHPS1 [28] was purchased from Sigma Aldrich. Cells were cultured for 2 h at 37uC in the presence or absence of PHPS1 (10 mM), and then stimulated as described above. Agonistic mAb (clone # F278) [59] specific for ILT2/CD85j was kindly gifted by Dr. Daniela Pende (IST, Genoa, Italy). Cells were stimulated with anti-CD3 mAb or PHA for 72 h, in the presence or absence of 5 ml/10 6 cells of anti-ILT2/CD85j monoclonal antibody.

Chemotaxis
Chemotaxis was investigated using 5 mm pore-size transwell plates (Costar, Cambridge, MA) as reported [60]. Five 610 5 cells were dispensed in the upper chamber, whereas chemokines or medium alone were added to the lower chamber. CCL2, CCL8, CCL21, CXCL10, CXCL11 (Immunotools, Friesoythe; Germany) CXCL12 and CXCL13 (Abnova, Heidelberg, Germany) were tested at 300 ng/mL [60] following preliminary titration experiments. Plates were incubated 2 h at 37uC. Migrated cells were collected and counted, and migration index was calculated as follows: (nu of migrated cells/nu of dispensed cells*100). Migration index obtained with medium alone was subtracted from each value.

Immunohistochemical staining of human tonsil
Immunohistochemical staining of tissue sections was performed using the Envision System HRP mouse (Dako, Glostrup, Den-mark). Briefly, 5 mm thick sections were cut from formalin fixed, paraffin embedded blocks, deparaffinized with xylene and rehydrated by passages through decreasing concentrations of ethanol (from 100% to 80%). Endogenous peroxidase activity was blocked by a 30 min incubation at room temperature with methanol containing 3% H 2 O 2 . Tissue sections were then incubated at 98uC for 40 min in citrate buffer (pH 6.0) for antigen retrieval (ChemMate, Dako). After rinsing in Optimax TM Wash Buffer (Menarini Diagnostics, Firenze, Italy), tissue sections were incubated 1 hour at room temperature with optimal amounts of anti-HLA-G (1:25 MEM-G2, Exbio) or isotype control (mouse IgG1, Invitrogen). Tissue sections were washed twice in Optimax TM Wash Buffer and incubated for 30 min at room temperature with Dako Envision System horse radish peroxidase (HRP) Mouse. After washing in Optimax TM Wash Buffer, peroxidase activity was detected by incubating tissue sections for 6-10 min at room temperature with Dako Liquid DAB Substrate Chromogen System (Dako). Tissue sections were counterstained with Mayer's hematoxylin (Sigma).

Western blot
Western blot analysis was performed on total extracts from 30*10 6 of CD4+ T cells stimulated as described above in presence or absence of sHLA-G (100 ng/ml). Protein extracts were obtained using Cell Extraction Buffer buffer (BioSource International, CA, USA) plus protease inhibitor cocktail (Sigma). Protein quantification was performed by BCA assay (Sigma).
Peroxidase-conjugated goat anti-mouse and anti-rabbit polyclonal antisera were used as secondary reagents (Upstate/ Millipore, MA, USA and Santa Cruz Biotechnology, respectively).
Immune complexes were visualized with the use of a Supersignal West Pico Chemiluminescent Substrate (Pierce, Rockford, USA) according to the manufacturer's instructions, and were normalized to internal controls (a rabbit monoclonal antibody against glyceraldehyde-3-phosphate dehydrogenase (Cell signaling). Protein levels were quantified by scanning densitometry of the autoradiography films using VersaDoc 3000 Gel Imaging System (BioRad, CA, USA) and normalized over (ratio) the housekeeping protein levels.