In immune cells, proinflammatory cytokine gene expression is regulated by glucocorticoids, whereas migration-inhibitory factor (MIF), a pleiotropic cytokine, has the unique property of counteracting the inhibitory effect of glucocorticoids on TNF-α and IL-1β secretion. A few lines of evidence suggest that γδ T cells play an important role in immunoregulation. However, it is unknown whether human γδ T cells participate in regulating MIF secretion, and how γδ T cells, glucocorticoids, and cytokines converge to give a unified physiological response. In this study, we demonstrate that human Vγ2Vδ2 T cells augment MIF secretion. Remarkably, these Vγ2Vδ2 T cells, functioning similarly to MIF in part, counteracted inhibition of dexamethasone on production of IL-1β and TNF-α. SCID mice reconstituted with human PBMC that were mock depleted of Vδ2 T cells and repeatedly infected with lethal dose of Escherichia coli had shorter survival time than those reconstituted with PBMC that were depleted of Vδ2 T cells. Thus, human Vγ2Vδ2 T cells are likely to play broad-spectrum roles in immunoregulation and immunopathology by influencing MIF secretion and the immunomodulatory function of glucocorticoids.
The host response to inflammation and infection involves the production of cytokines. Three pleiotropic cytokines, macrophage migration-inhibitory factor (MIF),3 IL-1β, and TNF-α, have broad biological effects. MIF is a pituitary peptide released during the physiological stress response, a proinflammatory macrophage cytokine secreted after LPS stimulation, and a T cell product expressed as part of the Ag-dependent activation response (1, 2). MIF counteracts the inhibitory effects of glucocorticoids on TNF-α, IL-1β, IL-6, and IL-8 production by monocytes in response to stimulation with LPS in vitro (1). MIF is also involved in broad-spectrum pathophysiological reactions as an inflammatory cytokine. In mouse models, treatment with anti-MIF Ab reduces mortality of septic shock (3), suppresses endotoxin-induced fatal hepatic failure (4), inhibits rheumatoid arthritis (5), and inhibits tumor growth (6, 7). MIF also augments resistance to microbial infection (8), and shows endocrine and enzymatic functions (9, 10).
TNF-α and IL-1β are multifunctional cytokines and play pivotal roles in inflammation and infection. For example, they are considered to be master cytokines in the pathobiology of septicemia and septic shock (11, 12), and in chronic, destructive arthritis (13). TNF-α and IL-1β may contribute to pathogenesis of thyroid autoimmunity (14, 15, 16). They may also mediate host defense responses to neuroinflammation and cell death in neurodegenerative conditions, in particular, multiple sclerosis and Parkinson’s and Alzheimer’s diseases (17, 18, 19, 20, 21, 22). In addition, TNF-α and IL-1β have potent effects in the CNS, resulting in fever, induction of sickness behavior, and activation of the hypothalamic-pituitary-adrenal axis (reviewed in Refs. 23 and 24).
Release of cytokines can be harmful and sometimes lethal to the host (25, 26, 27, 28, 29). Therefore, cytokine production must be reciprocally and finely tuned in vivo, and imbalance of tuning could cause immune system disorders. Glucocorticoids are one of the most potent regulators of cytokine production (30, 31). Glucocorticoids reduce the number of monocytes; lyse immature T cells; block phospholipase A2 activity; down-regulate the synthesis and secretion of IL-1, IL-6, and TNF-α from activated monocytes and macrophages; and inhibit cytokine-induced transcription factors, such as NF-κB and AP1 (32, 33). The crucial role of glucocorticoids has been demonstrated in a number of studies in which following adrenalectomy, challenge with TNF-α or IL-1 at doses that would be well tolerated in adrenal-intact animals proves fatal (34). These lethal effects can be prevented by steroid treatment (34).
γδ T cells comprise only 2–5% of CD3+ cells in human peripheral blood. Approximately 70% of these γδ T cells coexpress Vγ2 and Vδ2 TCR chains, and are uniformly reactive to nonpeptide organophosphate and alkylamine Ag, without CD1 or MHC restriction (35, 36). Several lines of evidence suggest that these cells play an important role in immunoregulation (37, 38, 39, 40, 41, 42, 43). γδ T cells prime macrophages to produce TNF-α in response to LPS stimulation (43), and augment production of macrophage-derived NO (44). Depletion of γδ T cells results in a decrease of TNF-α, IL-1β, IFN-γ, and IL-6 gene expression in the spinal cord of mice with autoimmune encephalomyelitis (45). However, the roles of γδ T cells in the immunoregulation of cytokine production are still poorly understood. It is unknown whether human γδ T cells participate in regulating MIF secretion, and how γδ T cells, glucocorticoids, and cytokines converge to give a unified physiological response. In this study, we demonstrate that Vγ2Vδ2 T cells augment MIF secretion and counteract the inhibitory effect of glucocorticoids on production of IL-1β and TNF-α, suggesting that Vγ2Vδ2 T cells are involved in immunoregulation. In an in vivo hu-SCID model (SCID mice reconstituted with human PBMC), human Vγ2Vδ2 T cells were found to participate in the pathogenesis of septic shock.
Materials and Methods
Ab and Ag reagents
mAb ascites against T cell Ags used were as follows: control mAb (P3), pan-γδ TCR (anti-TCRδ1), Vδ1 (A13), Vδ1/Jδ1 (δTCS1), Vδ2 (BB3), Vγ2 (7A5), and CD3 (OKT3). The specificity of these mAbs is reviewed in Porcelli et al. (46). Other reagents were purchased as follows: FITC-conjugated F(ab′)2 goat anti-mouse IgM and IgG (catalog number AMI4708; BioSource International, Camarillo, CA); isobutylamine (IBA; catalog number I-3634; Sigma-Aldrich, St. Louis, MO); pamidronate (Novartis, East Hanover, NJ); PE-conjugated mouse anti-human TNF-α (catalog number 18630D; BD PharMingen, San Diego, CA), and anti-human IL-1β (catalog number IC201P; R&D Systems, Minneapolis, MN); ELISA human TNF-α set (catolog number 555212; BD PharMingen), and IL-1β set (catolog number 2687KI; BD PharMingen); anti-human MIF mAb (catalog number MAB289; R&D Systems); biotinylated anti-human MIF Ab (catalog number BAF289; R&D Systems); human rMIF (catalog number 289-MF; R&D Systems); LPS (catalog 201, from Escherichia coli 0111:B4; List Biological Laboratories, Campbell, CA); and Limulus amebocyte lysate (catalog number GS003; Associates of Cape Cod, Falmouth, MA).
Human PBMC obtained from random healthy donor leukopacks (Dana-Farber Cancer Institute, Boston, MA) were isolated by Ficoll-Hypaque centrifugation (Pharmacia, Piscataway, NJ). PBMC were cryopreserved in FBS containing 10% DMSO at −196°C until use.
Depletion of Vγ2Vδ2 T cells
Depletion of Vγ2Vδ2 T cells was performed by use of mouse anti-human Vδ2 Ab (BB3), or P3, an isotype-matched mock control, and goat anti-mouse IgG Dynabeads M-450 (catalog number 110.06; Dynal Biotech, Oslo, Norway), according to the manufacturer’s instructions. For most depletions, P3, an isotype-matched control mAb, was substituted for the anti-Vδ2 mAb. For some depletions, anti-human Vδ1 Ab (A13) was taken as an alternative control and substituted for the anti-Vδ2 mAb. Vδ2 T cells constituted 70–90% of total γδ T cells for the PBMC used in our experiments. Over 95% of Vδ2 T cells, confirmed by surface marker staining and analysis of flow cytometry, were depleted. Since Vδ2 T cell constituted only about 2% of CD3+ cells in our experiments, depletion of Vδ2 T cells from PBMC with a specific Vδ2 T cell Ab (BB3) did not show significant influence on percentages of CD14+ and other cells when analyzed by flow cytometer (data not shown). We screened several donors by two-color fluorescence and found that 100% of Vδ2-bearing T cells also expressed Vγ2. Although the Vδ2 TCR chain paired with Vγ1 or Vγ3 has been described, they are extremely rare, and there is no evidence that they respond to nonpeptide Ags. Thus, the likelihood that we are studying a population other than Vγ2Vδ2+ T cells is very remote.
Detection of endotoxin
Endotoxin levels in the Abs and reagents were assessed with the Limulus amebocyte lysate kit according to the procedures recommended by manufacturer. The detection sensitivity of the assay was 0.03 EU/ml. Endotoxin levels in P3 (isotype-matched mock control Ab) and BB3 (Vδ2 Ab) Abs were all negative at the concentration used for depletion of Vδ2 T cells.
Stimulation of PBMC with dead bacteria and LPS in vitro
To determine whether γδ T cells, under a physiological condition, play a role in regulating monocytes to produce IL-1β in presence of dexamethasone, neither LPS nor dead E. coli was applied in the experiments (Fig. 2), whereas, to determine whether γδ T cells regulate PBMC to produce and secrete IL-1β and TNF-α in a pathological circumstance, LPS or dead E. coli was used in the experiments (Figs. 3 and 4), as described below. PBMC were washed twice after 16-h preincubation with 1 mM IBA or medium in presence or absence of dexamethasone. LPS (final concentration: 1 μg/ml) or dead E. coli (inactivated at 56°C for 2 h, final concentration: 5 × 105 CFU/ml) was added to each well. The culture supernatants at the indicated time points were collected for analysis of MIF, IL-1β, or TNF-α by ELISA.
Detection of human IL-1β, TNF-α, and MIF in the tissue culture supernatant
Human IL-1β, TNF-α, and MIF ELISA were performed according to the procedures recommended by the manufacturers. The detection limit of the assay was 15.6 pg/ml for IL-1β, 7.8 pg/ml for TNF-α, and 60 pg/ml for MIF.
Intracellular cytokine staining
Human PBMC were cultured in RPMI medium with inclusion or exclusion of 10 nM dexamethasone in the absence of LPS (Fig. 2), or in the presence of LPS (Fig. 4). Four hours before staining with cellular surface marker, monensin (GolgiStop; BD PharMingen) that enhanced intracellular cytokine accumulation was added in the medium. Cells were washed with PBS and stained with surface marker AlexaFluor-conjugated 488 IgG control Ab, pan-TCRδ1 (Abs purified and conjugated by our laboratory), or FITC-conjugated anti-CD14. After two washes, cells were fixed with 2% formaldehyde in PBS and permeabilized with 0.5% (w/v) saponin (BD PharMingen). Intracellular TNF-α and IL-1β were stained with PE-conjugated Ab in saponin buffer. After two washes, cells were resuspended in PBS and analyzed using FACS flow cytometer (BD Biosciences, Mountain View, CA) and Flowjo software (Tree Star, San Carlos, CA). Vδ2 T cells themselves did not produce intracellular IL-1β, TNF-α, and IFN-γ after 16 to 24 h culture in the medium containing 10% FBS or 1 μg/ml LPS (47) (L. Wang, unpublished observation).
Engraftment and infection of SCID mice
Homozygous C.B-Igh-1b/Gbms-Prkdc(SCID)-Lyst(beige)N7 (SCID) male mice, 5–6 wk old, were purchased from Taconic Farms (Germantown, NY) and maintained in microisolator cages. Animals were fed autoclaved food and water, and all manipulations were performed under laminar flow. The mice were weighed and randomly distributed into groups of 5–14 animals with equal mean body weight. E. coli (ATCC 25922) was grown in Luria-Bertani (LB) broth at 37°C until the culture reached early stationary phase. E. coli was aliquoted (1 ml/vial) and stored in LB broth containing 10% glycerol at −80°C until use. Before infection, E. coli were washed once with 30 ml PBS and plated on LB agar to determine CFU. The SCID mice were injected i.p. with 0.5 ml RPMI medium containing 3 × 107 human PBMC and 0.5 ml PBS containing 1–5 × 107 E. coli under aseptic conditions. For septic shock model, the mice were inoculated with E. coli twice, 1 × 107 as first dose and 5 × 107 as second dose. The interval between two bacterial infections was 20 h. The animals were observed four times per day for their survival time.
Values were expressed as means ± SEM of the respective test or control group. Statistical significance between control and test groups was calculated by Student’s t test (two-tailed) and among groups by analysis of variations. Survival time was analyzed by Kaplan-Meier. Data were representative of two to four experiments.
Human Vγ2Vδ2 T cells up-regulate MIF secretion
MIF plays a critical role in the systemic inflammatory response by counter-regulating the inhibitory effect of circulating glucocorticoids on immune cell activation and proinflammatory cytokine production (1). To determine whether Vγ2Vδ2 T cells regulate MIF secretion, human PBMC were depleted or mock depleted of Vδ2 T cells and subsequently stimulated with LPS. The culture supernatant was analyzed for MIF titer by ELISA. PBMC mock depleted of Vδ2 T cells secreted up to 2-fold more MIF than those depleted of Vδ2 T cells either in the absence (Fig. 1,a) or presence (Fig. 1,b) of dexamethasone, suggesting that optimal MIF secretion is dependent on presence of Vγ2Vδ2 T cells. Addition of 10 nM dexamethasone to the culture, in agreement with others (1, 2), inhibited MIF secretion (Fig. 1,b). As an additional control, PBMC depleted of Vδ1 T cells had greater levels of MIF secretion compared with Vδ2-depleted PBMC, after exposure to LPS (data not shown). Optimum augmentation of MIF secretion by Vγ2Vδ2 T cells occurred at 3–6 h after stimulation, with a major loss of titer by 9 h after stimulation (Fig. 1 c).
Human Vγ2Vδ2 T cells counteract the inhibitory effect of dexamethasone on IL-1β production
Since Vγ2Vδ2 T cells augmented the release of MIF (Fig. 1), we speculated that these cells might also influence production and secretion of IL-1β. As monocytes are a major source of IL-1β, we first analyzed intracellular IL-1β production of monocytes at a physiologic concentration of dexamethasone in the absence of LPS stimulation, by use of two-color flow cytometry. In absence of dexamethasone and LPS, 41.3 and 42.6% of monocytes produced IL-1β in the absence or presence of Vγ2Vδ2 T cells, respectively. In contrast, when PBMC were cultivated in 10 nM dexamethasone (equivalent to a physiological level of bioactive cortisol in normal human blood), monocytes reduced their intracellular IL-1β production by up to 30-fold in the absence of Vδ2 T cells (Fig. 2). This low level of intracellular IL-1β staining of monocytes in the absence of Vδ2 T cells was not due to cytokine release to the extracellular medium, since there was no detectable IL-1β in the culture supernatant in the absence of LPS or dead E. coli stimulation (Fig. 3 a). These data suggest that Vγ2Vδ2 T cells counteract the inhibitory effect of glucocorticoids on IL-1β production by monocytes under physiologic conditions, resulting in an increased storage pool of intracellular IL-1β. If PBMC were stimulated with LPS for 1–6 h, >90% of CD14+ monocytes produced intracellular IL-1β and TNF-α in the absence of dexamethasone, and 30–60% in the presence of dexamethasone, dependent on the concentration of dexamethasone added. As a result, any potential difference in intracellular cytokine production in the presence or absence of γδ T cells was veiled by this overwhelming cytokine production during short-term (1–6 h) exposure to LPS.
Since intracellular cytokine staining reflects a snapshot of accumulated cytokine by each individual cell, and only secreted cytokines affect the immunologic response, we then assessed IL-1β levels in the supernatant of PBMC in the presence of 10 nM dexamethasone. In the absence of stimulation with LPS or dead bacteria, PBMC secreted only trace amounts of IL-1β, whereas, after stimulation with LPS or dead bacteria, PBMC that were mock depleted of Vδ2 T cells secreted up to 2-fold more IL-1β than PBMC depleted of Vδ2 T cells (Fig. 3,a). If the cells were cultivated in the medium containing a natural Vγ2Vδ2 T cell-specific Ag, IBA (48, 49), and then stimulated with LPS, PBMC that were mock depleted of Vδ2 T cells secreted up to 4-fold more IL-1β than PBMC depleted of Vδ2 T cells (Fig. 3,a). Replacement of LPS with dead E. coli (Fig. 3 a), or replacement of IBA with the pharmaceutical Vγ2Vδ2 T cell-specific Ag, pamidronate (data not shown), yielded similar results. These data suggest that Vγ2Vδ2 T cell-specific Ags activate γδ T cells, which then augment the ability of PBMC to secrete IL-1β upon exposure to LPS or dead bacteria.
To determine whether MIF could reverse the inhibition of glucocorticoids on IL-1β secretion, we added 1 ng MIF to the cultures containing 10 nM dexamethasone and 1 μg LPS. In this combination, PBMC depleted of Vδ2 T cells resumed their IL-1β secretion nearly to the levels of PBMC that were mock depleted of Vδ2 T cells (Fig. 3 b). Whereas addition of exogenous MIF did not significantly augment IL-1β secretion from the PBMC that were mock depleted of Vδ2 T cells, addition of 1 ng/ml MIF restored IL-1β secretion from Vδ2 T cell-depleted cultures to the levels similar to those achieved in the presence of Vγ2Vδ2 T cells.
Human Vγ2Vδ2 T cells counteract the inhibition by dexamethasone of TNF-α production and secretion
Since Vγ2Vδ2 T cells counteracted the inhibitory effect of glucocorticoids on IL-1β production (Fig. 2), we speculated that these γδ T cells might also influence production and secretion of TNF-α. As monocytes are a major source of TNF-α, we first analyzed the effect of 10 nM dexamethasone on the intracellular TNF-α production of monocytes. In the absence of Vδ2 T cells, only 0.045% of CD14+ cells produced intracellular TNF-α, whereas in presence of Vδ2 T cells, 1.27% of CD14+ cells produced intracellular TNF-α (Fig. 2). We then determined the TNF-α production of monocytes in response to stimulation with LPS or heat-killed E. coli. Similar to the results obtained with IL-1β (data not shown), monocytes produced up to 2-fold more intracellular TNF-α in the presence than in the absence of Vγ2Vδ2 T cells after 18-h exposure to LPS or heat-killed E. coli (Fig. 4 a). These data suggest that human Vγ2Vδ2 T cells, similar to mouse γδ T cells (43), up-regulate intracellular TNF-α production by human monocytes in response to stimulation with LPS or dead bacteria.
We further assessed TNF-α secretion from PBMC in response to stimulation with LPS. Human PBMC depleted of Vδ2 T cells secreted up to 2-fold less TNF-α as compared with PBMC that were mock depleted of Vδ2 T cells, when stimulated with LPS (Fig. 4 b). Since Vγ2Vδ2 T cells themselves did not produce and secrete TNF-α in response to stimulation with LPS (47), these data suggest that optimal TNF-α secretion by PBMC is dependent on presence of Vγ2Vδ2 T cells.
To determine whether MIF could reverse the inhibition of glucocorticoids on TNF-α secretion, we added 1 ng MIF to the cultures containing 1 μM dexamethasone and 1 μg LPS. In this combination, PBMC depleted of Vδ2 T cells resumed their TNF-α secretion nearly to the levels of PBMC that were mock depleted of Vδ2 T cells (Fig. 4 c). Whereas addition of exogenous MIF did not significantly augment TNF-α secretion from the PBMC that were mock depleted of Vδ2 T cells, addition of 1 ng/ml MIF restored TNF-α secretion from Vδ2 T cell-depleted cultures to levels similar to those achieved in the presence of Vγ2Vδ2 T cells.
Taken together, the above data suggest that Vγ2Vδ2 T cells augment MIF secretion, counteract the inhibitory effect of glucocorticoids on production and secretion of IL-1β and TNF-α, and influence the interaction of MIF and glucocorticoids on cytokine secretion by PBMC.
Vγ2Vδ2 T cells play a role in septic shock in vivo
Since MIF, TNF-α, and IL-1β have been shown to be critical mediators of septic shock (3, 11), we suspected that Vγ2Vδ2 T cells might play a role in septic shock. To determine whether human Vγ2Vδ2 T cells play a role in bacterial infection in vivo, we reconstituted SCID mice with human PBMC that were either mock depleted or depleted of γδ T cells, and challenged these mice with E. coli. Six SCID mice i.p. challenged with 1 × 107 E. coli all died of infection within 2 days postinfection. In contrast, five mice reconstituted with human PBMC and subsequently challenged with the same dose of E. coli all survived, suggesting that human PBMC play a crucial role against bacterial infection, and that residual mouse immune cells have negligible effects. This hu-SCID model enables us to further investigate the roles of Vγ2Vδ2 T cells in sepsis and septic shock. SCID mice were reconstituted with human PBMC that were depleted or mock depleted of Vδ2 T cells, and subsequently infected with 1 × 107 CFU E. coli. After 20 h, the mice received a second dose of E. coli (5 × 107 CFU). In marked contrast to the experiments using low-dose bacterial infection (50), the mice receiving PBMC depleted of Vδ2 T cells had a longer survival time than those receiving PBMC that were mock depleted of Vδ2 T cells (p = 0.0428, Fig. 5), suggesting that Vγ2Vδ2 T cells play a role in septic shock.
The immunoregulatory function of Vγ2Vδ2 T cells is still poorly understood. In this study, we found that Vγ2Vδ2 T cells up-regulated MIF secretion (Fig. 1). Optimal production and secretion of IL-1β and TNF-α were dependent on presence of Vγ2Vδ2 T cells (Figs. 2, 3, and 4). Furthermore, Vγ2Vδ2 T cells counteracted the inhibition by glucocorticoids of IL-1β and TNF-α production (Figs. 2, 3 b, and 4c).
Recent studies suggest that physiological levels of glucocorticoids are immunomodulatory rather than solely immunosuppressive, resulting in a shift of cytokine production from a Th1- to a Th2-type pattern (30). Interruptions of this loop at any level, such as genetic, surgery, infection, or pharmacological treatments, can cause host susceptibility to infections and inflammatory diseases (30). We speculate that overactivation of glucocorticoid regulation, as occurred in absence of Vγ2Vδ2 T cells, might affect severity of infectious disease. Furthermore, in absence of Vγ2Vδ2 T cells, MIF, TNF-α, and IL-1β, three pleiotropic and important cytokines for host resistance to bacterial infection (8, 51, 52, 53, 54, 55, 56, 57), had reduced production or secretion. The relative overactivity of glucocorticoids and reduced levels of cytokine production and secretion might account for the severity of bacterial infection in vivo. Consistent with this hypothesis, SCID mice reconstituted with human PBMC depleted of Vδ2 T cells and subsequently infected with either E. coli, Morganella morganii, or Staphylococcus aureus had much higher mortality and bacterial load than those reconstituted with human PBMC that were mock depleted of Vδ2 T cells (50).
Vγ2Vδ2 T cells remarkably counteract the inhibitory effect of glucocorticoids on IL-β production, and slightly on TNF-α production (Fig. 2). Interestingly, the increased IL-1β in presence of Vγ2Vδ2 T cells was intracellularly stored in monocytes, but not extracellularly secreted in the absence of LPS or bacterial stimulation. Once exposed to LPS or dead bacteria, monocytes began to secrete IL-1β and TNF-α, which was up-regulated by Vγ2Vδ2 T cells (Figs. 3 and 4). The physiological significance of this finding might be that, with assistance from Vγ2Vδ2 T cells, immune cells accumulate a high level of intracellular cytokines ready to be secreted in case of infection, making the immune response earlier and stronger. Since the prognosis of infection is dependent on the speed of immune system reaction and pathogen proliferation, early response to bacterial infection is crucial for the immune system to eliminate pathogens. Therefore, in the absence of Vγ2Vδ2 T cells, host resistance to bacterial infection would be impaired.
On the other hand, MIF, IL-β, and TNF-α play important roles in immunopathology. Vγ2Vδ2 T cells are capable of augmenting MIF secretion and counteracting the inhibitory effect of glucocorticoids on production and secretion of IL-1β and TNF-α, thereby accounting for clinical symptoms. One piece of supportive evidence is that patients treated with the pharmaceutical Vγ2Vδ2 T cell-specific Ag, pamidronate, had fever and influenza-like symptoms (58). This and other aminobisphosphonate drugs have been widely used to inhibit osteoclastic bone resorption. As potent Ags, aminobisphosphonates stimulate Vγ2Vδ2 T cells in a TCR-dependent, MHC- and CD1-unrestricted manner (48, 49). Influenza-like symptoms occurred in these patients most likely due to the release of cytokines, since TNF-α and IL-1β have been demonstrated to have potent effects in the CNS, resulting in fever and sickness behavior (23, 24). These cytokines, besides being directly produced from activated Vγ2Vδ2 T cells, were optimally generated by other immune cells in a Vγ2Vδ2 T cell-dependent manner.
The interactions of Vγ2Vδ2 T cells and glucocorticoids implicate these T cells in a broad spectrum of immunoregulatory effects. It is likely that Vγ2Vδ2 T cells participate in pathophysiological reactions wherever glucocorticoids are involved, such as infection, inflammatory, autoimmune, and allergic illnesses, including rheumatoid arthritis, systemic lupus erythematosus, Sjögren’s syndrome, allergic asthma, and atopic skin disease (30, 31, 59). Since activated Vγ2Vδ2 T cells reduce the inhibitory effect of glucocorticoids on production and secretion of TNF-α and IL-1β, any natural Vγ2Vδ2 T cell-specific Ags, such as nonpeptide alkylamine (35), or organophosphate Ags (36), and drugs that activate Vγ2Vδ2 T cells should be considered for their interference with anti-inflammatory treatment regimens and disease outcomes.
Under extreme situations, overwhelming secretion of MIF, TNF-α, or IL-1β plays a critical role in the pathogenesis of septic shock (3, 11, 12), resulting in tissue damage, multiple organ failure, and even death. Since Vγ2Vδ2 T cells augment MIF secretion and counteract the inhibitory effect of glucocorticoids on production and secretion of TNF-α and IL-1β, we suspect that these cells might play a role in septic shock. To determine whether human γδ T cells mediate antibacterial effects in vivo, we developed a hu-SCID infective model, which has proven to be a powerful method for the study of human cells and tissues (60, 61, 62, 63, 64). We found that SCID mice inoculated with either human PBMC or Salmonella typhi all survived, whereas SCID mice receiving both PBMC and S. typhi were all dead within 10 days postinfection (L. Wang, unpublished observations), dependent on the dose of inoculated human PBMC and bacteria. These data suggest that septic shock resulted from interaction of human immune cells and bacteria. We therefore reconstituted SCID mice with human PBMC that were either mock depleted or depleted of γδ T cells, then repeatedly challenged these mice with lethal doses of E. coli. The mice receiving PBMC depleted of Vδ2 T cells had a longer survival time than those receiving PBMC mock depleted of Vδ2 T cells (Fig. 5), suggesting that human Vγ2Vδ2 T cells play a role in pathogenesis of sepsis.
The mechanism by which Vγ2Vδ2 T cells counteract inhibition of glucocorticoids on IL-1β and TNF-α production is unclear. Monocytes constitutively secrete MIF, which could be augmented by exposure to LPS (1, 2). T cells produce MIF as part of the Ag-dependent activation response (1, 2). In our experiments, Vδ2 T cells mediating immune regulation were observed at 16 h, but not 6 h post-in vitro culture, suggesting that immune regulation of these monocytes required considerable engagement time. In addition, we had experienced technical difficulties (e.g., a very poor ratio of signal-noise) in distinguishing a small proportion of γδ T cells (about 2% of CD3+ cells in our experiments) from whole PBMC in production of intracellular MIF. Therefore, we were unable to correlate MIF that may have been produced by a small proportion of γδ T cells at a certain time point, to an inhibitory effect of glucocorticoids on IL-1β and TNF-α production. However, in the presence of γδ T cells, MIF secretion by PBMC was augmented (Fig. 1), and the inhibitory effect of glucocorticoids on IL-1β and TNF-α production was counteracted ( Figs. 2–4).
We speculate that the major function of Vδ2 T cells is to up-regulate αβ T cells and monocytes to produce more MIF, whereas Vδ2 T cell-produced MIF might constitute only a small proportion of the MIF in the supernatant. In addition, since the use of anti-MIF Ab could not significantly abrogate the enhanced cytokine secretion seen in the presence of Vγ2Vδ2 T cells (data not shown), it is possible that, besides augmentation of MIF secretion, Vγ2Vδ2 T cells counteract glucocorticoid activity by other unknown pathways. This possibility is supported by the fact that addition of exogenous MIF only partially restored the deficit in IL-1β and TNF-α production associated with γδ T cell depletion (Figs. 3 b and 4c). This work demonstrates that γδ T cells augment MIF secretion and counteract the inhibitory effect of glucocorticoids on IL-1β and TNF-α production. However, further studies are needed to determine the exact mechanisms by which this counteraction occurs.
γδ T cells did not produce IL-1β, TNF-α, and IFN-γ in response to the stimulation with LPS or dead E. coli (47). However, if these γδ T cells were pretreated with IBA, a Vδ2 T cell-specific Ag secreted by live bacteria, then exposed to LPS, they started to produce cytokines and to expand afterward (Kamath et al., manuscript in preparation). It is possible that human γδ T cells, like murine cells, express Toll-like receptors (TLRs) (65), or most likely up-regulate TLR expression on monocytes, which are involved in LPS recognition. Recently, it was found that MIF regulated monocyte immune responses through modulation of TLR4 (66). Although we were unable to detect TLR2 and TLR4 on γδ T cells by flow cytometry, more sensitive methods, such as RT-PCR used for detecting TLR expression on mouse γδ T cells (65), or microarray, might be alternative approaches for this purpose.
In conclusion, since Vγ2Vδ2 T cells augment MIF secretion and counteract inhibition of glucocorticoids on production of IL-1β and TNF-α, we speculate that these cells play more broad-spectrum roles in immunoregulation by influencing the glucocorticoid immunomodulatory loop.
This research was supported by grants from the National Institutes of Health and the Arthritis Foundation.
Abbreviations used in this paper: MIF, migration-inhibitory factor; hu-SCID, SCID mice reconstituted with human PBMC; IBA, isobutylamine; LB, Luria-Bertani; TLR, Toll-like receptor.