Linking Innate and Adaptive Immunity: Human Vγ9Vδ2 T Cells Enhance CD40 Expression and HMGB-1 Secretion

γδ T cells play an important role in regulating the immune response to stress stimuli; however, the mean by which these innate lymphocytes fulfill this function remains poorly defined. The main subset of human peripheral blood γδ T cells responds to nonpeptidic antigens, such as isopentylpyrophosphate (IPP), a metabolite in the mevalonate pathway for both eukaryote and prokaryote cells. IPP-primed γδ T cells significantly augment the inflammatory response mediated by monocytes and αβ T cells to TSST-1, the staphylococcal superantigen that is the major causative agent of toxic shock syndrome. Here we show that the small pool of activated peripheral γδ T cells induces an early upregulation of CD40 on monocytes and the local release of High Mobility Group Box-1 (HMGB-1), the molecule designated as the late mediator of systemic inflammation. This finding provides a new basis for how γδ T cells may serve as influential modulators of both endogenous and exogenous stress stimuli.


Introduction
γδ T cells have a nonredundant role in regulating immune homeostasis. In comparison to αβ T cells, they lack receptor diversity yet demonstrate remarkable plasticity in response to both infectious and noninfectious stimuli. As an extension of their dichotomous nature, γδ T cells have been held liable for both instigating over-zealous immune responses [1][2][3] as well as reigning in inflammation by controlling activated macrophages and promoting wound healing [4][5][6]. In humans, the main subset of peripheral blood γδ T cells usually comprises between 1 and 5% of circulating T cells and bears the canonical Vγ9Vδ2 T cell receptor. These Vγ9Vδ2 T cells are unique to primates, and they respond to endogenous and exogenous stress-induced molecules that are nonpeptidic in nature and preclude the requirement for processing or presentation by classical MHC molecules [7][8][9]. Isopentylpyrophosphate (IPP) is the prototypical phosphoantigen recognized by Vγ9Vδ2 T cells that was first isolated from lysates of mycobacteria [7]. It was later determined to be a metabolite in the mevalonate pathway in all eukaryotic and some prokaryotic cells, including Staphylococcus aureus [10,11]. During the initial phase of infection with low bacterial inocula of S. aureus, Vγ9Vδ2 T cells are activated by endogenous mevalonate metabolites through the accumulation and dephosphorylation of the hydroxymethylglutaryl-coenzyme A reductase, the ratelimiting enzyme of the mevalonate pathway [11]. Therefore, primates have evolved the ability to readily respond to bacterial infection as well as endogenous damage by sensing the dysregulation of the mevalonate pathway, which may be considered a stress-associated "alarmin" for these primordial lymphocytes.
We have shown that stimulating the small pool of human γδ T cells present in PBMC with IPP markedly augmented the early proinflammatory response to Toxic Shock Syndrome Toxin-1 (TSST-1) [3]. TSST-1 is a staphylococcal superantigen that exerts a potent inflammatory response by bridging Vβ2 specific αβ T cell receptors with conserved regions of the MHC class II molecules of antigen presenting cells (APCs) outside the peptide-binding groove [12]. Ferry et al. [13] demonstrated the presence of specific Vβ T cell signatures of various superantigenic toxins, including TSST-1, in the blood of patients with staphylococcal TSS. We previously demonstrated that IPP-activated γδ T cells augment the levels of proinflammatory cytokines within 2 Mediators of Inflammation the first 6 hours post-TSST-1 treatment, thereby changing the kinetics and magnitude of the immune response to the superantigen [3]. We have also shown that High Mobility Group Box Protein-1 (HMGB-1) is released following TSST-1 stimulation [14]. HMGB-1 is a highly conserved nonhistone DNA-binding protein found in virtually all nucleated cells and is regarded as being the late mediator of sepsis [15]. Activated monocytes secrete HMGB-1 after exposure to endotoxin [16], and it is detectable in the serum of septic patients within 24 hours. Importantly, neutralizing the effects of HMGB-1 at this late time point, either by anti-HMGB-1 antibodies or antagonists, rescued mice from sepsis-induced mortality [15]. In contrast to endotoxin, which activates only macrophages to secrete HMGB-1, both T cells and macrophages participate in the nuclear translocation and subsequent release of HMGB-1 into the extracellular environment following TSST-1 stimulation [14].
The first aim of the present study was to evaluate the influence exerted by IPP-primed Vγ9Vδ2 T cells on the change in costimulatory molecule expression that could explain their ability to enhance the immune response to TSST-1. The second objective was to determine the consequence of γδ T cell mediated immune modulation on the expression and secretion of HMGB-1.

Toxin Purification.
Recombinant TSST-1 was purified from culture supernatants of S. aureus strain RN4220 previously transformed to carry the tst gene, using both preparative isoelectric focusing and chromatofocusing [17]. Toxin purity was assessed by silver staining after sodium dodecyl sulfate-polyacrylamide gel electrophoresis on 14% acrylamide gels, and LPS activity was undetectable by the Limulus amoebocyte lysate gelation (sensitivity limit, 10 pg/mL).

Preparation of Cells and Culture Conditions.
Fresh human peripheral blood mononuclear cells (PBMC) from healthy donors were obtained by Ficoll-Paque PLUS (Amersham Biosciences Corp., Piscataway, NJ) density centrifugation, and cultured in 96-well U-bottom plates at 1.5 × 10 6 cells/ml in complete culture medium consisting of RPMI 1640 (StemCell Technologies Inc., Vancouver, BC, Canada), 10% heat-inactivated fetal bovine serum (HyClone Laboratories Inc., Logan, UT), 2 mM L-glutamine (StemCell), 25 mM Hepes buffer (StemCell), and 2 ug/ml of polymyxin B sulphate (Sigma-Aldrich Corp., St. Louis, MO). Depletion of γδ T cells was accomplished by immunomagnetic positive cell selection with an anti-γδ TCR MAb (clone Immu510, BD BioSciences Pharmingen Inc., San Diego, CA) conjugated to an antidextran MAb tetramer (StemCell), which left the rest of the PBMC untouched, using the StemSep protocol (Stem-Cell). Purity of the γδ T cell-depleted PBMC was analyzed by flow cytometry using phycoerythrin-(PE-) conjugated anti-Vδ2 antibody (BD BioSciences Pharmingen). Figure 1 shows the population of Vγ9Vδ2 T cells in PBMCs before (which made up 1.5% of live cells) and after depletion (0.1%).

Treatment of PBMC with IPP and TSST-1.
PBMCs with or without γδ T cell depletion were treated with either 45 μM IPP (Sigma-Aldrich) or left untreated in complete growth medium for 16-18 hours (overnight) at 37 • C, 5% CO 2 prior to treatment with 1 nM TSST-1. For secreted HMGB-1 detection, culture supernatants were microcentrifuged at 800 × g for 5 minutes and frozen at −70 • C until analysis.

Fluorescent Microscopy Analysis.
Surface expression of HMGB-1 on PBMC by fluorescent microscopy has been previously described in detail [14]. In brief, PBMCs were cultured (37 • C, 5% CO 2 ) directly on sterile cover slips to allow attachment of adherent cells overnight before stimulation with 1 nM of TSST-1. For HMGB-1 expression, cells were incubated with rabbit anti-human HMGB-1 IgG as previously described [14], diluted in blocking buffer (PBS, 3% FBS), followed by a secondary Alexa488-conjugated goat antirabbit IgG antibody. Cells were subsequently fixed using Cytofix buffer (Pharmingen). Fluorescently labelled PBMCs were then stained with Hoechst 3342 nuclear dye (Molecular Probes) according to manufacturer's directions, and mounted on slides using Prolong Anti-Fade reagent (Molecular Probes). Cells were visualized with an AxioPlan II fluorescence microscope equipped with a CCD camera using Northern Eclipse software (Epix) for acquisition of images. Images were taken with the 63 x oil immersion objective lens, and Adobe Photoshop 6.0 software was used for image layout.
2.6. Western Blot Analysis of Secreted HMGB-1. PBMC culture supernatants were concentrated (∼10-fold) from the original volume of 1.5 ml using Amicon Ultra centrifugal filters with a molecular weight cutoff of 10 kDa (Millipore). The supernatants then had been further prepared using the SDS-PAGE Clean-up kit (Amersham) according to the manufacturer's directions prior to running on a 12% polyacrylamide gel. Western blotting was performed by semidry transfer of proteins (Trans-Blot SD Semi-Dry Electrophoretic Transfer Cell, BioRad) onto an Immobilon-P PVDF membrane (Millipore) which was blocked for 1 hour at room temperature with 1% BSA, 0.5% Tween in TBS Mediators of Inflammation prior to overnight incubation (at 4 • C) with rabbit polyclonal anti-HMGB-1 antibody. The membrane was subsequently incubated with anti-rabbit IgG horse-radish-peroxidase-(HRP-) conjugated secondary antibody for 1 hour at room temperature on a shaker. HMGB-1 detection was performed using Super Signal substrate (Pierce) and developed as well as analyzed using the Alpha Innotech 3400 Gel Documentation system (Alpha Innotech).

Statistical Analysis.
Statistical analysis was performed using Prism 4.0 software package (GraphPad Software, Inc., San Diego, CA). Student t-tests (one-tail) were used to assess the significance of differences in HMGB-1 and costimulatory molecule expression between IPP treated and untreated PBMC at different times post-TSST-1 stimulation. The effects of IPP over time on co-stimulatory molecule expression by APC (CD40, CD80, CD86, HLA-DR) and αβ T cells (CD25, CD28) were further examined by 2way ANOVA. Differences were considered significant if the probability of the null hypothesis was less than five percent (P < .05). Each experiment represents 3 to 4 healthy subjects unless otherwise noted.

IPP-Primed γδ T Cells Induce an Early
Upregulation of CD40 following TSST-1 Stimulation. Following overnight treatment with IPP, CD40 expression was upregulated in comparison to those that were untreated (IPP + TSST-1 versus TSST-1 alone; P < .05, Student's t-test) (Figure 2(a)). The elevated CD40 expression was observed up to 6 hours postsuperantigen stimulation, after which time CD40 was downregulated or internalized. We previously demonstrated that CD40L (CD154) is expressed on T cells within 3 hours following exposure to TSST-1 [18], therefore, the downregulation of CD40 beyond 6 hours is likely a consequence of CD40 binding to CD40L on TSST-1-activated T cells.
We verified that up-regulation of CD40 expression in populations exposed to IPP was solely attributed to the presence of activated γδ T cells, since removal of γδ T cells completely abrogated the effect of IPP (Figure 2(b)). The effect of IPP-activated γδ T cells over time on CD40 expression following TSST-1 stimulation was further verified using 2-way ANOVA (IPP x time interaction, P < .05).
In contrast to CD40, the expression of CD86 was downregulated early in IPP treated PBMC at 3 hours following TSST-1 stimulation (IPP + TSST-1 versus TSST-1 alone; P < 0.05, Student's t-test) (Figure 2(a)). Again, this effect was abrogated following depletion of γδ T cells (Figure 2(b)). Two-way ANOVA confirmed an interaction effect (IPP x time interaction, P < .05). There was a similar trend for decreased expression of HLA-DR (the MHC class II molecule to which TSST-1 binds on APC) and CD80 by IPP treatment, but the differences were not statistically significant.

TSST-1, but Not IPP-Primed γδ T Cells, Activate αβ T Cells.
In contrast to the direct affect of IPP on monocyte function, there was little notable direct influence on the activation state of αβ T cells as evaluated by the expression of CD25 or CD28, which remained largely unchanged in response to IPP pretreatment ( Figure 3). The upregulation of CD25 on αβ T cells at 24 hours was solely dependent on TSST-1 stimulation regardless of IPP, as shown by the comparison of IPP + TSST-1 versus IPP alone (P = .001, Student's t-test, Figure 3). Two-way ANOVA confirmed this effect (TSST-1 effect, P < .01; TSST-1 x time interaction, P < .005).  To determine the influence of TSST-1 on IPP-activated γδ T cells, comparisons were made between IPP + TSST-1 versus IPP alone. Overnight IPP treatment significantly up-regulated CD40 expression at 3hours and 6hours after TSST-1 stimulation ( * P < .05, student's t test), and down-regulated CD86 expression at 3h after TSST-1 stimulation ( * P < .05). No other significant differences in receptor expression were observed for the other co-stimulatory molecules.

IPP-Primed
HMGB-1 to the cell surface where it serves a function for process extension, migration, and cell-cell interaction [19]. Extracellular cell-surface expressed HMGB-1 has been called amphoterin to differentiate it from intracellular HMGB-1 found in the nucleus [19]. Given the direct influence of γδ T cells on the activation, maturation, and regulation of cells of monocytic lineage [20,21], we wanted to determine how IPP stimulation would influence the surface manifestation of HMGB-1 on cultured monocytes present in PBMC. This change in surface expressed HMGB-1 was assessed by flow cytometry, which enabled us to both confirm and quantify the release of HMGB-1 from differentiated monocytes following TSST-1 stimulation. Figure 4(b) displays the difference between the various treatment conditions on the loss of cell-surface expressed HMGB-1 assessed by flow cytometry. IPP-pretreated PBMCs that were subsequently stimulated with TSST-1 released more HMGB-1 from the cell surface (i.e., less HMGB-1 expression on the cell surface) than those cells that were left untreated with IPP prior to superantigen stimulation (IPP + TSST-1 versus TSST-1 alone, P < .05, paired Student's t-test). The actual secretion of HMGB-1 into the extracellular milieu from TSST-1 stimulated PBMC was further verified by Western blot of cell culture supernatants (Figure 4(c)), which again substantiated that IPP pretreatment followed by TSST-1 stimulation led to its greater release than PBMC stimulated with TSST-1 without IPP-activation of γδ T cells (IPP + TSST-1 versus TSST-1 alone, P < .05, paired Student's t-test).

Discussion
Overnight treatment with IPP resulted in a notable upregulation of CD40 on monocytes in the presence of γδ T cells. We previously demonstrated that impeding the 6 Mediators of Inflammation  Figure 3: Modulation of CD25 and CD28 on αβ T cells by IPP-activated γδ T cells. PBMCs were cultured overnight in the presence of IPP or absence of IPP before stimulation with 1 nM of TSST-1 (a). Data (mean ± SEM) for expression of CD25 and CD28 were shown after various time intervals following TSST-1 stimulation, and the numbers of healthy donors studied for each coreceptor were as indicated. Baseline values were obtained from PBMC treated with RPMI growth medium alone. To determine the influence of IPP-activated γδ T cells on receptor expression, comparisons were made between IPP + TSST-1 (dashed red lines) versus TSST-1 alone (solid black line) and significant differences were denoted by ( * ). To confirm that any observed difference in receptor expression was due to the presence of activated γδ T cells, a cognate group of PBMCs was depleted of γδ T cells (<0.1%) and studied in parallel experiments (b). To determine the influence of TSST-1 on IPP-activated γδ T cells, comparisons were made between IPP + TSST-1 versus IPP alone, and significant differences were denoted by ( * * ). There was no additional effect of IPP-activated γδ T cells on CD25 and CD28 expression following TSST-1 stimulation; however, CD25 expression on αβ T cells following TSST-1 stimulation in the presence of IPP-activated γδ T cell was significantly higher at 24 hour than with IPP alone ( * * P = .001, student's t-test).
CD40-CD40L interaction with CD40 monoclonal antibodies (but not the CD86 or CD80 interaction) completely blocks TSST-1-induced inflammation [18]. The early up-regulation of CD40 on monocytes by IPP-primed γδ T cells in the current study provides a basis for the prior observation that TNFα and IFNγ levels increased 3-to 4-fold during the early phase of the immune response to TSST-1 in PBMC containing activated γδ T cells [3]. This early augmentation was also abrogated after depletion of γδ T cells [3]. This study also demonstrates that the exaggerated response to TSST-1 by activated γδ T cells is through the direct modulation of the responsiveness of monocytes and not αβ T cells since we found that the upregulation of CD25 on αβ T cells was affected only by stimulation with TSST-1 and not IPP. The regulation of CD40 expression by activated γδ T cells appears to parallel the mechanism of linking innate to adaptive immunity by glycolipid responsive Vα14 + NK T cells stimulated with α-galactosylceramide. It was found that the rapid release of TNFα and IFNγ to ovalalbumin, acquisition of CD4+ and CD8+ T cell immunity, and maturation of dendritic cells were all dependent on the CD40-CD40L interaction induced by the priming of Vα14+ NK T cells by α-galactosylceramide [22]. In contrast to CD40, CD86 expression was found to be reduced in PBMC treated with IPP overnight. This monocytic phenotype may be reflective of a state described as "immature dendritic cells" in which cycling of the HLA-DR molecule is rapid and levels of CD86 are relatively low [23]. Taken    Mediators of Inflammation these studies suggest that innate T lymphocytes responsive to specific "alarmins" influence the ability of APC to effectively stimulate cells of adaptive immunity, and the CD40-CD40L interaction likely plays a pivotal role in this process. This scenario is in agreement with recent demonstrations that γδ T cells induce early dendritic cell maturation and share an intimate relationship with APC [21,24,25]. Eberl et al. [26] further provided evidence that human γδ T cells are activated by microbial mevalonate metabolites and interact with monocytes and local macrophages to drive the acute inflammation in bacterial infections. Thus, γδ T cells provide a direct link between innate and adaptive immunity. It should be noted, however, that these studies primarily used either expanded γδ T cell lines or a large number of purified γδ T cells in coculture conditions. The conditions used in this study may be more reflective of the physiological state in which a small number of γδ T cells present in peripheral blood would be able to immediately respond to their stressassociated antigens, such as IPP. This early event would preclude the possibility of any expansion of γδ T cells, but it would lead to the efficient priming of APC.
This study also provides the first demonstration that IPPprimed human peripheral blood γδ T cells influence the secretion of HMGB-1 by macrophages. This finding is similar to the work by Semino et al. [27] who demonstrated that NK cells reciprocally activate autologous dendritic cells, and this interaction was also mediated through the secretion of HMGB-1. It is feasible that the enhanced HMGB-1 secretion observed in cells containing activated γδ T cells could be in response to an increase in apoptosis [28] due to the fact that IPP-stimulated γδ T cells are particularly susceptible to activation-induced cell death [29,30] as a self-limiting precaution. During apoptosis, HMGB-1 is normally tightly bound to nucleosomes [28] as opposed to necrosis where HMGB-1 is released from cells [31]. However, work by Urbonaviciute et al. [28] showed that biologically active HMGB-1 may also be released from apoptotic cells. One of the main ligands for extracellular HMGB-1 is the receptor for advanced glycation end (RAGE) products, a member of the Ig superfamily. The expression of RAGE increases following activation on both dendritic cells and antigen stimulated T cells [32]. The release of HMGB-1 and its subsequent binding to RAGE is required for the clonal expansion, survival, and functional polarization of naive T cells [33].
Both dendritic cell maturation and Th1 polarization are events that have been attributed to the activity of γδ T cells and HMGB-1, respectively [21,[34][35][36]. Our findings connect these two previously unrelated concepts and brings forth the suggestion that γδ T cells mediate this effect and link innate and adaptive immunity by influencing the early upregulation of CD40 expression and release of HMGB-1 from APC, thereby priming the immune response to subsequent stimuli.