γδ T cells play important but poorly defined roles in pathogen-induced immune responses and in preventing chronic inflammation and pathology. A major obstacle to defining their function is establishing the degree of functional redundancy and heterogeneity among γδ T cells. Using mice deficient in Vγ1+ T cells which are a major component of the γδ T cell response to microbial infection, a specific immunoregulatory role for Vγ1+ T cells in macrophage and γδ T cell homeostasis during infection has been established. By contrast, Vγ1+ T cells play no significant role in pathogen containment or eradication and cannot protect mice from immune-mediated pathology. Pathogen-elicited Vγ1+ T cells also display different functional characteristics at different stages of the host response to infection that involves unique and different populations of Vγ1+ T cells. These findings, therefore, identify distinct and nonoverlapping roles for γδ T cell subsets in infection and establish the complexity and adaptability of a single population of γδ T cells in the host response to infection that is not predetermined, but is, instead, shaped by environmental factors.
The essential role played by γδ T cells in protective immunity to infection has been demonstrated in studies using γδ T cell-deficient mice. These mice are more susceptible than their wild-type (WT)4 counterparts to infection with low doses of viruses (1, 2, 3), bacteria (4, 5, 6), and parasites (7, 8, 9), leading to the development of accelerated and exacerbated inflammatory responses that can result in necrosis, pathology, and even death (reviewed in Ref. 10). In response to microbial infection, γδ T cell responses are staged, accumulating at the site of infection early, before, or concurrent with the involvement of αβ T cells and again later, at the time of pathogen clearance (2, 11, 12, 13, 14, 15, 16, 17). Although reports of the biological role of γδ T cells have been conflicting, models of infectious disease indicate that they may perform different functions at different stages of the immune response. In the early stage of infection γδ T cells appear to be proinflammatory; their absence results in increased pathogen burden (6, 14), and they secrete IFN-γ (3, 8, 17, 18) and up-regulate its production by other lymphocytes and NK cells (19, 20). They can also direct adaptive (αβ T cell) responses (21, 22, 23). During the later stages of the immune response to microbial infection, γδ T cells kill bacteria-elicited, activated macrophages, coincident with or after bacterial clearance (24), indicative of the involvement of γδ T cells in preventing chronic inflammation. A similar staging of the γδ T cell response has also been demonstrated in autoimmune disorders. For example, in a mouse model of collagen-induced arthritis, depletion of γδ T cells before collagen immunization reduces disease severity, putatively through removal of a proinflammatory stimulus, whereas their depletion late in the course of disease increases its severity through a lack of regulatory γδ T cells (25).
What is not clear is whether the pro- and anti-inflammatory functions of γδ T cells are mediated by distinct subpopulations of γδ T cells or by the same cells whose functional phenotype is influenced by the microenvironment in which they are activated. The differential expression of TCR variable (V) genes has been widely used to distinguish between different populations (reviewed in Ref. 26) and has been used to assign specific functions to different subsets of γδ T cells (27, 28). The immune response to certain viral (13, 17) and bacterial (29, 30) infections is dominated by a single subset of γδ T cells expressing TCRs encoded by the GV5S1 (Vγ1; TCR-Vγ nomenclature used is that of Heilig and Tonegawa (31)) and TRADV15 (Vδ6) gene families (32) that accumulate in sites of infection coincident with or after pathogen clearance. The ability to kill Listeria monocytogenes (Lm)-elicited, activated macrophages has also been shown to be a property of the Vγ1/Vδ6+ subset (33) consistent with a role for these cells in macrophage homeostasis and in preventing chronic inflammatory responses after pathogen removal. The degree of functional heterogeneity of this subset is not clear, however, with both pro- and anti-inflammatory activities and with beneficial as well as detrimental effects on host responses having been described (reviewed in Ref. 34). Vγ1+ T cells are prominent during the early stages of Lm infection, in which they have been shown to constitutively express IL-12Rs (30) and to be major producers of IFN-γ (17, 30), consistent with an early proinflammatory role and with acting as a bridge between the innate and adaptive immune systems. However, studies of the overall protective role of Vγ1+ T cells in Lm infection are inconclusive, with in vivo depletion studies resulting in either increased (35) or decreased (36) bacterial numbers, leaving it unclear whether bacterial containment and protection are properties of the Vγ1+ or other γδ T cell subsets.
A major obstacle in defining the function of Vγ1+ T cells is therefore establishing the degree of functional redundancy and heterogeneity of this population during the course of infection. To resolve the conflicting evidence relating to the function of Vγ1+ T cells, TCRVγ1-deficient (Vγ1−/−) mice have been used to determine the requirement for Vγ1+ T cells in the host response to infection and the extent of functional redundancy among γδ T cells during pathogen-induced immune responses. We show in this study that Vγ1+ T cells have an essential, nonredundant function in macrophage homeostasis and may play a novel role in regulating γδ T cell homeostasis. They cannot, however, protect TCRδ−/− mice from immune-mediated tissue injury. In addition, Vγ1+ T cells exhibit a considerable degree of functional plasticity and heterogeneity, and Vγ1+ T cells involved in the early vs late stages of pathogen-induced immune responses comprise nonoverlapping populations that appear to be selectively expanded on the basis of their ability to produce cytokines appropriate to the pervading environmental conditions and the stage of the immune response.
Materials and Methods
Mice and infection
Male and female mice were used at 6–8 wk of age, with three to five mice per group. C57BL/6 TCRVγ1−/− mice were generated in our laboratory, C57BL/6 TCRδ−/− were obtained from The Jackson Laboratory, C57BL/6 mice were obtained from Harlan Laboratories, and all were housed in the animal facility at University of Leeds. For the generation of TCR-Vγ1−/− mice, fragments of the murine TCR-Vγ1-Jγ4-Cγ4 gene were cloned from a bacterial artificial chromosome murine ES-129/SvJ library, and sequences were checked with those obtained from C57BL/6 genomic DNA. A targeting vector was designed to replace a genomic fragment containing Vγ1, Jγ4, and exon 1 of Cγ4 sequences with a neomycin resistance cassette. The targeting vector was linearized and electroporated into ES cells (Incyte Genetics). After G418 selection, homologous recombinants were identified by Southern blot hybridization using a 5′ probe consisting of an ∼0.3-kb HindII-SacI genomic DNA fragment located ∼0.4 kb 5′ of Vγ2 and a 3′ probe comprising an ∼1.3-kb EcoRV-XbaI genomic DNA fragment encompassing exon 4 of the Cγ4 gene. Two clones heterozygous for the targeted mutation were injected into C57BL/6 blastocysts, which were subsequently transferred into pseudopregnant foster mothers. Chimeric mice were crossed with C57BL/6 mice to produce heterozygous TCR-Vγ1+/− mice. Germline transmission of the mutation was verified by PCR and Southern blot analysis of tail DNA. Heterozygotes were intercrossed to generate homozygous TCR-Vγ1−/− mice. Homozygous and heterozygous mutant mice were backcrossed into C57BL/6 mice more than seven times before use in experiments. WT and TCRδ−/−, and TCRVγ1−/− mice were infected i.p. with 1.5 × 104 CFU of Lm (strain 10403S). Between 2 and 8 days after infection, spleens and livers were weighed and homogenized in distilled water, and 10 μl of the homogenate was plated onto brain-heart infusion agar (Oxoid). Colonies were counted after 24-h incubation at 37°C, and the numbers of CFU per spleen and per gram of tissue were calculated.
Assessment of liver damage
Liver damage in Lm-infected mice was assessed by serum alanine aminotransferase (ALT) levels and histology. Serum ALT was measured on an ADVIA 2400 chemistry system analyzer (Bayer) following the manufacturer’s instructions. For histology, Formalin-fixed, paraffin-embedded, 5-μm tissue sections from the livers of infected and control noninfected mice were stained with H&E and examined with an Axiovert 200M microscope (Zeiss) and Axiovison image analysis software (Imaging Associates).
Cellular phenotypic analysis by flow cytometry
Standard protocols were followed for preparing cell suspensions (37). Briefly, for splenocytes, the tissues were homogenized, contaminating erythrocytes lysed with 0.84% (w/v) ammonium chloride solution, and the cell suspension was passed through a 0.7-μm pore size nylon filter and washed before Ab staining. Small intestinal intraepithelial lymphocytes (iIELs) were isolated from Peyer’s patch-excised small intestines by 37°C incubation in 50 ml of HBSS containing 10% (v/v) FBS, 5 mM EDTA, and 3 mg/ml dithioerythritol, followed by density gradient separation on Percoll (Amersham Biosciences) (38). Peritoneal exudate cells were collected in HBSS containing 10 U/ml heparin. All reagents were purchased from Sigma-Aldrich. For surface staining, the Abs included F(ab′)2 of mAbs specific for TCR-Vγ1 (clone 2.11) (39), TCR Vδ6.3 (clone 17C) (16), and TCR-δ (GL3) and intact Ab specific for TCR-Vγ7 (F2.67). The 2.11 and F2.67 hybridoma cell lines were provided by Dr. P. Pereira (Institut Pasteur, Paris, France). The commercial anti-mouse mAbs used were TCRγδ (GL3), CD3 (145-2C11), TCRVγ4 (Vγ2; UC3-10A6), TCRVγ5 (Vγ3; 536), and F4/80, conjugated to biotin or fluorochromes purchased from Caltag-Medsystems or BD Pharmingen. Streptavidin conjugates of PE, FITC, (Caltag Laboratories), or Alexa Fluor 633 (Molecular Probes) were used as secondary reagents. To block nonspecific Ab binding, cells were incubated with anti-FcR Ab mixture, anti-CD16/32 (20 μg/ml; Caltag Laboratories). Isotype-matched Abs of irrelevant specificity were used to determine the level of nonspecific staining. Stained cells were analyzed on a FACSCalibur flow cytometer using CellQuest software (BD Biosciences).
γδ T cell enrichment and adoptive transfer
γδ T cells were enriched from splenocytes of C57BL/6 mice by positive immunomagnetic selection after labeling with either FITC-F(ab′)2 of anti-TCR-Vγ1 (2.11)- or anti-TCR-δ (GL3)-specific mAbs, followed by anti-FITC magnetic microbeads and selection on MACS LS columns (Miltenyi Biotec). The resulting cell populations routinely contained >80% γδ T cells of >95% viability, and 2–4 × 105 cells were then transferred i.v. into C57BL/6 TCRδ−/− and TCRVγ1−/− mice immediately before bacterial infection.
Macrophage-T cell coculture
Plastic-adherent peritoneal exudate cells (5 × 105) from day 8 Lm-infected TCRδ−/− mice were incubated with splenocytes (1 × 106) from WT or Vγ1−/− mice on eight-well chamber slides (ICN Pharmaceuticals) and cultured at 37°C in RPMI 1640 (Sigma-Aldrich) with 10% FBS for 1 h as previously described (33). To evaluate macrophage killing by T cells, adherent macrophages were incubated with the Live/Dead cell reagent (Molecular Probes) containing fluorescent dyes that identify the intracellular esterase activity of viable cells (calcein AM) or that are incorporated in the nuclei of dead cells (ethidium bromide homodimer-1) and were examined by UV microscopy using an Axiovert 200M microscope (Zeiss) and Axiovison image analysis software (Imaging Associates). At least 100 live and/or dead cells were counted in four separate fields.
Intracellular cytokines were detected by cytoplasmic staining of splenocytes recovered from Lm-infected WT mice after culture in vitro for 5 h with 10 μg/ml brefeldin A (Sigma-Aldrich). Cells were then surface-stained with anti-CD3, -TCR-γδ, and -TCR-Vγ1 mAbs; fixed in 1% paraformaldehyde; and permeabilized with 0.5% saponin (Sigma-Aldrich) before cytoplasmic staining with PE-conjugated anti-mouse cytokine mAbs to IL-2, IL-4, IL-5, IL-6, IL-10, IFN-γ, and TNF-α (Caltag-Medsystems or BD Pharmingen) or FITC-conjugated polyclonal Abs to MIP-1β and MCP-1 (Sigma-Aldrich). Anti-TGF-β and anti-latency-associated peptide bound to latent TGF-β (anti-LAP) mAbs (R&D Systems) were conjugated to FITC and used for surface staining. PE- and FITC-conjugated isotype-matched mAbs of irrelevant specificity were used to determine levels of nonspecific staining of anti-cytokine mAbs.
TCR-Vγ/Vδ profiling and structural analyses
RNA was extracted from γδ T cell-enriched splenocytes from Lm-infected C57BL/6 mice using Tri-reagent (Sigma-Aldrich) according to the manufacturer’s instructions. RNA was reverse transcribed using an oligo(dT)-primed ImpromII RT kit (Promega), and the resulting cDNA was amplified using Reddy-Mix (Abgene) under the following reaction conditions; denaturation at 94°C, annealing at 55°C, and extension at 72°C for 38 cycles. The primers used (5′-3′) were Vγ1CCGGCAAAAAGCAAAAAAGT, Cγ4AAGGAGACAAAGGTAGGTCCCAGC, Vγ2TTGGTACCGGCAAAAAACAAATCA, Cγ2CAATACACCCTTATGACATCG, Vγ4CTTGCAACCCCTACCCATAT, Vγ5GAGGATCCCGCTTGGAAATGGATGAGA, Vγ6GATCCAAGAGGAAAGGAAAGACGGC, Vγ7GATCCAACTTCGTCAGTTCCACAAC,Cγ1CCACCACTCGTTTCTTTAGG, Vδ1AATAGCAATTCTACTGATGGTGG, Vδ2AGTCCTCAGTCTCTGACAATC, Vδ3CCAGATTCAATGGAAAGTAC, Vδ4GTACAAACAGCAAGGAGGGCAGG, Vδ5CCAGACAGTGGCAAGCGGCACTG, Vδ6TCAAGTCCATCAGCCTTGTC, Cδ1CGAATTCCACAATCTTCTTG, GCTTCTTTGCAGCTCCTTCGTTG (β-actin forward), and TTCTCCATGTCGTCCCAGTTGG (β-actin reverse). Products were visualized on 2% ethidium bromide-stained agarose gels. Structural analyses of PCR-amplified TCR cDNAs was conducted by direct sequencing of pGEM-T (Promega)-cloned cDNAs and by spectratyping (Lark DNA Technologies). For spectratype analysis, 2 μl of PCR product was amplified in a 10-cycle run-off reaction using fluorescently labeled, J-region-specific primers (Jγ1, CTTAGTTCCTTCTGCAAATACC; Jγ4, TACGAGCTTTGTCCCTTTG), and the amplicons were analyzed on sequencing gels using GenescanView4 software (CRIBI).
Differences in mean values between two groups were evaluated by Mann-Whitney U tests using Statistical Package for the Social Sciences software (SPSS).
Phenotypic characteristics of Vγ1−/− mice
Vγ1-deficient (Vγ1−/−) mice were generated to define Vγ1+ T cell function and determine the extent of functional redundancy among γδ T cells during pathogen-induced immune responses. Mice homozygous for the Vγ1 null allele were created by mutating the Vγ1 gene as well as the Jγ4 and exons 1 and 2 of the Cγ4 gene to prevent any rearrangement and the possibility of pairing of other Vγ genes with an otherwise intact Cγ4 gene. Founder animals heterozygous for the mutation were bred to obtain homozygous mice, which were identified by Southern blotting (Fig. 1,A). The absence of Vγ1 mRNA expression in Vγ1−/− spleen and thymus confirmed that these mice were deficient in TCRVγ1-expressing T cells (Fig. 1,B). Homozygous mice were backcrossed at least seven generations onto the C57BL/6 background before experimentation. Homozygous mice were healthy, able to mate, and had a normal life span. The only observed physiological differences between Vγ1−/− and WT mice was an enlargement of the primary lymphoid organs in adult (>6 wk old) Vγ1−/− mice resulting in an ∼1.5-fold increase in the cellularity of the thymus (average of 1.1 × 108 in WT and 1.6 × 108 in Vγ1−/− mice) and spleen (average of 4 × 107 in WT and 6 × 107 in Vγ1−/− mice). This increase in cellularity was not attributable to any one population of leukocytes and resulted in increased numbers of γδ T cells in these tissues. A comparison of the splenic TCR-Vγ and -Vδ mRNA profiles in noninfected adult WT and Vγ1−/− mice showed no qualitative differences in the profile of expression of Vγ2-, Vγ4-, and Vδ6-encoded TCRs in the absence of Vγ1+ T cells (Fig. 1 B). The expression of Vδ6 in these mice, which is the major δ-chain paired with Vγ1 in WT mice (33, 40), shows that Vγ1 is not required for the expression of this δ-chain and that it can be used by other Vγ receptors, such as Vγ2 and Vγ4 (41).
In WT mice, Vγ1+ T cells are a major γδ T cell subset in the spleen (20–50%) and among iIELs (15–30%) (39). The effect of Vγ1+ T cell deficiency on the splenic and IEL γδ T cell repertoires was assessed by comparing the distribution and number of γδ T cells between adult WT and Vγ1−/− mice. In the spleen, the percentages of γδ T cells were similar (Fig. 1,C), although their absolute numbers were increased by 30–40% due to the enlargement of the spleen in Vγ1−/− mice, indicating that Vγ1+ T cell deficiency in Vγ1−/− was compensated for by the expansion of other γδ T cell subsets. Based on the TCRVγ mRNA profiles (Fig. 1,B), these were mainly Vγ4+ and Vγ2+ T cells. A similar pattern was seen among iIELs, which were ∼1.5-fold higher in number in Vγ1−/− compared with WT animals, with an increased frequency of γδ+ T cells from 30% in the WT to 52% (Fig. 1,C). There was also a corresponding increase in the size of the major, Vγ7+, iIEL subset, which accounted for almost 90% of all TCRγδ+ iIELs (Fig. 1 D). These findings suggest that Vγ1+ T cells may exert some influence on iIEL homeostasis in the small intestine.
γδ T cell recruitment during Lm infection is regulated by Vγ1+ T cells
The γδ T cell response to Lm infection in C57BL/6 mice is biphasic, with increased numbers of cells seen early (day 2), before the αβ T cell response, and later (days 6–8), coincident with or after bacterial clearance (16, 42). The γδ T cells present at both the early and late stages of the response are dominated by a single subset that coexpresses Vγ1 and Vδ6.3 (Vγ1/Vδ6) TCRs (16). The γδ T cell response to infection in the absence of this subset was determined by comparing splenocyte TCR-Vγ and -Vδ receptor expression in both the early and late responses in WT and Vγ1−/− mice.
Up to 2 days after infection, TCR-Vγ and -Vδ expression in both strains of mice was similar (Fig. 2 A), with the exception of a lack of Vγ1 expression in Vγ1−/− mice. There was also no change in TCR-Vδ expression from the noninfected state, in which only TCR-Vδ6 mRNA was detected (data not shown). Unexpectedly, TCR-Vγ7 mRNA, which was not detected in the spleen of noninfected mice of either strain, was present in the spleen of both Vγ1−/− and WT mice on day 2 after infection.
During the late response to Lm infection, striking differences were seen in TCR-Vγ and -Vδ expression (Fig. 2,B). In WT mice, the profile of TCR-Vγ expression was similar to that on day 2 after infection, although TCR-Vγ7 mRNA was no longer detected. By contrast, in Vγ1−/− mice, TCR-Vγ5 and -Vγ6 mRNA were detected in addition to TCR-Vγ7 mRNA, which was unexpected because Vγ5 and Vγ6 expressions normally define monomorphic, oligoclonal populations restricted to the skin (Vγ5+ dendritic epidermal T cells (DETC)) (43) and mucosal epithelia of the reproductive tract (Vγ6) (44). There were also differences in TCR-Vδ expression during the late stage of the response to Lm infection. Although TCR-Vδ2 and -Vδ3 mRNA were present in both WT and Vγ1−/− mice, additional TCR-Vδ1 and -Vδ5 mRNA were detectable in Vγ1−/− mice. The presence of these unusual subsets of γδ T cells in Lm-infected Vγ1−/− mice was confirmed by Ab staining and flow cytometry. The spleen of Vγ1−/− mice at 8 days after infection contained significant proportions of Vγ5+ (∼25%) and Vγ7+ (∼20%) γδ T cells (Fig. 2,C), consistent with the TCR-Vγ mRNA analysis (Fig. 2,B). Abs specific for the Vγ6 receptor chain were not available. By contrast, the levels of staining of splenocytes of WT mice and TCRVγ1−/− mice reconstituted with Vγ1+ T cells immediately before infection with anti-Vγ5/7 Abs (<5%) were no higher than those seen with isotype-matched control Abs. The size of the Vγ4+ subset in both strains of mice was, however, comparable (35% in Vγ1−/− and 40% in WT). To attempt to establish the origin and relationship of Vγ5+ and Vγ7+ T cells in the spleens of Lm-infected Vγ1−/− mice with, respectively, DETC and iIELs of WT mice, TCR-CDR3 spectratype analysis was conducted on RT-PCR-amplified TCRVγ5/7 cDNAs of skin and small intestinal RNA from noninfected adult C57BL/6 mice and compared with that of day 8 Lm-infected Vγ1−/− mice. The spectratype profiles in Fig. 2 D show that the Vγ5 and Vγ7 TCRS from Vγ1−/− mice were similar to those expressed in the skin and small intestine of WT mice, with the same peak predominating among Vγ5 and Vγ7 receptors of WT and Vγ1−/− mice. The profiles seen in Vγ1−/− samples could be distinguished from those in WT mice by the presence of one or more minor peaks absent in WT mice (Vγ5) or increased representation of individual peaks common to both WT and Vγ1−/− mice (Vγ7). These profiles suggested a commonality in the structure and possibly origin of Vγ5+ and Vγ7+ T cells in the spleens of Lm-infected Vγ1−/− and the DETC and IELs of WT mice. However, the finding that only ∼10% of Vγ5+ T cells in Lm-infected Vγ1−/− mice were reactive with the DETC anti-TCRVγ5/Vδ1 clonotype Ab, 17D.1 (45) (E. M. Andrew and S. R. Carding, unpublished observations) suggests that the majority of these cells are not skin derived, and/or that the Vγ5 receptor chain expressed in the spleens of Vγ1−/− mice pairs with other Vδ-chains. Collectively, these results suggest that Vγ1+ T cells are immunoregulatory and involved in the homeostatic regulation of other γδ T cell subsets in response to infection.
Vγ1+ T cells are both necessary and sufficient for γδ T cell involvement in macrophage homeostasis
Previously we have shown that γδ and Vγ1/Vδ6+ T cells kill populations of activated macrophages (33, 42). Peak cytocidal activity occurs during the late γδ T cell response coincident with the appearance of large numbers of terminally differentiated macrophages and with bacteria clearance. Vγ1−/− mice were used therefore to establish whether macrophage killing is a nonredundant function of Vγ1+ T cells and what the consequences of their absence is on the response and fate of macrophages during infection. Similar to those in Lm-infected TCRδ−/− mice, the percentage and absolute number of macrophages were increased in Vγ1−/− mice in the primary sites of infection, with a 3- to 4-fold increase in the spleen (Fig. 3,A) and a 2-fold or more increase in the peritoneal cavity (Fig. 3,B) late during the course of infection. This increase in macrophage numbers in TCRVγ1−/− mice was similar to or greater than that seen in TCRδ−/− mice and was reduced to levels comparable with those in WT individuals by reconstitution with Vγ1+ T cells before infection. Most significant was the absence of any macrophage cytotoxic activity among splenocytes from Vγ1−/− mice (Fig. 3 C). As shown in this study and previously (33), macrophage cytotoxicity was only evident among effector cells obtained from either noninfected or Lm-infected mice with intact γδ T cell populations (WT and TCRβ−/− mice) and was restricted to the macrophage-adherent Vγ1+ subset of γδ T cells. Thus, Vγ1+ T cells are both necessary and sufficient to kill Lm-activated macrophages and play an important and nonredundant role in macrophage homeostasis during pathogen-induced immune responses.
Vγ1+ T cells do not protect mice from immune-mediated tissue injury
Exaggerated inflammation and extensive tissue (liver) necrosis are hallmark features of microbial and Lm infection in γδ T cell-deficient mice (4, 5). Although the identity and means by which specific populations of γδ T cells normally protect the host from immune-mediated tissue injury as a consequence of infection have not been identified, the reported anti-inflammatory and antibacterial activities of Vγ1+ T cells (24, 36) suggest that they may perform such a role. This possibility was investigated further in this study by determining whether Vγ1+ T cells contribute to pathogen containment and eradication and/or can prevent the liver necrosis seen in Lm-infected TCRδ−/− mice.
Vγ1−/− mice were able to contain bacterial infection to the same extent as WT animals (Fig. 4 A). With the exception of day 2 after infection, there was no significant difference in bacterial numbers in the livers of Vγ1−/− mice and WT mice throughout infection, although there was a fairly wide distribution in bacterial numbers between individuals in the same group on all days after infection. At 2 days after infection, there were significantly more bacteria in the livers of Vγ1−/− compared with TCRδ−/− mice (p = 0.03). By 4 days after infection, however, which is before or coincident with the onset of liver necrosis, the pattern had reversed, with TCRδ−/− mice containing significantly (p = 0.05) more bacteria than either Vγ1−/− or WT mice. Because the total weight of the liver of each mouse strain was similar and did not change throughout the course of infection, the bacterial CFU per gram values obtained were representative of the total bacterial load in the liver. The difference observed in bacterial burden between Vγ1−/− and TCRδ−/− mice may reflect the slower migration of bacteria in TCRδ−/− mice from the initial site of infection (peritoneal cavity) to the liver and then more rapid bacterial growth due to a complete lack of γδ T cells.
There was no evidence of any liver lesion in the Vγ1−/− mice during Lm infection (Fig. 4,C and data not shown). This was substantiated by serum ALT measurements, which, as seen in infected TCRδ−/− mice (Fig. 4,B), increase as a result of liver damage (46), but remained unchanged in infected Vγ1−/− and WT mice (Fig. 4,B). The inability of Vγ1+ T cells to prevent immune-mediated liver injury during listeriosis in TCRδ−/− mice was confirmed in γδ T cell adoptive transfer experiments. The adoptive transfer of enriched Vγ1+ T cells (>80% TCRVγ1+) into TCRδ−/− recipients before Lm infection failed to prevent the development of necrotic liver lesions (Fig. 4,D). In contrast, γδ (Vγ1−) T cells from Vγ1−/− mice were able to provide protection to TCRδ−/− mice from developing liver necrosis after Lm infection, and they also slightly improved bacterial containment (Fig. 4 E). These findings show that the function of Vγ1+ T cells in pathogen-induced immune responses is restricted to the homeostatic control of macrophage and perhaps γδ T cell activity, and that they play no role in preventing chronic inflammation and tissue injury.
Vγ1+ T cells show stage-dependent functional differences during Lm infection
That γδ T cells are functionally heterogeneous and their involvement in certain pathogen-induced immune responses is staged are well established (3, 6, 11, 12, 14, 16, 17, 18, 21, 47, 48). Important questions, however, are: what is the extent of functional plasticity among cells that express TCRs encoded by the same Vγ/δ gene segments, and are the same or different populations of T cells involved in different stages of the host response to infection? To address these issues, the functional attributes and structure of the CDR3 regions of the TCRs expressed by Vγ1+ T cells during the early and late stages of the immune response to Lm infection in WT mice were compared.
Intracellular cytokine staining and flow cytometry were used to profile the cytokine response of Vγ1+ T cells at 2 and 8 days after infection, corresponding to the peaks of their response to Lm in C57BL/6 mice (Fig. 5,A) and mirroring the waves of γδ T cells that form the early and late responses (10, 16). At both stages of the immune response to Lm infection, Vγ1+ T cells were the major source of the cytokines produced by γδ T cells as a whole; with the exception of IL-6 production at 2 days after infection, the frequency of Vγ1+ cells that produced any of the cytokines tested was at least equivalent to and in some cases higher than that produced by all other (Vγ1−) γδ T cells (Fig. 5, C and D). In general, the highest number of cytokine-producing Vγ1+ T cells was present during the late stage (day 8) of the immune response (Fig. 5, C vs D), with significantly more (p < 0.01) Vγ1+ T cells producing the anti-inflammatory cytokines, TGF-β, LAP, and IL-10, during the late vs the early stage of the host response to infection (Fig. 5,B). There were also significant increases in the frequency of Vγ1+ T cells producing IL-6 and IL-2 during the late stage of the immune response to Lm (Fig. 5,B) and in IL-5 production by Vγ1+ at the early stage of the response (Fig. 5,C). In contrast, during the early stage of the anti-Lm response, Vγ1+ T cells produced only one cytokine, IL-4, at significantly higher levels (p ≤ 0.01) than at the late stage of the response (Fig. 5,B). In the early stages the production of IL-4, which is a strong promoter of IL-12 (49) and of neutrophil recruitment to the liver (50), suggests a more proinflammatory role for Vγ1+ T cells early in infection. There were essentially no differences in the frequency of cells producing the other cytokines and chemokines examined (IL-5, IFN-γ, TNF-α, MCP-1, and MIP-1β), which were all produced to a similar extent by Vγ1+ T cells at both time points after infection. On day 8 after infection, the vast majority of LAP- and TGF-β-positive γδ T cells were Vγ1+ (Fig. 5 D). Because this was detected on the cell surface, it is not clear whether it was produced by the cells or was bound to the cell surface, although high levels of TGF-β mRNA expression were detected among FACS-purified, late-stage Vγ1+ T cells by microarray analysis (data not shown). The presence of TGF-β in both its latent and active forms, in conjunction with the production by these cells of IL-10 and IL-2, is reminiscent of the functional phenotype of αβ CD4+ T regulatory cells, suggestive of a role for Vγ1+ T cells in down-modulating the late-stage αβ T cell response either directly or perhaps via induction of CD4+CD25+ regulatory T cells (51). This analysis of cytokine production by Vγ1+ T cells clearly demonstrates their functional heterogeneity, which is appropriate to the stage of infection at which they become involved.
The structure of the Vγ1/Vδ6 TCRs expressed during the early and late stages of the immune response to Lm was examined by spectratyping and DNA sequencing of RT-PCR-amplified and cloned cDNAs. Spectratype analyses of Vγ1-TCRs expressed before and during the early and late stages of infection were strikingly similar, with all of the prominent peaks representing productive rearrangements of Vγ1-Jγ4-Cγ4 receptor chains (Fig. 6,A). In addition, a single peak corresponding to 231 bp was predominant in each of the three samples, representing ∼50% of all productive rearrangements (Fig. 6,D). Although these spectratype profiles imply commonality in the structure of the Vγ1 TCRs expressed at different times during the response to infection, sequencing of individual Vγ1 and Vδ6 TCR cDNAs showed very little or no similarity in the CDR3 junctional regions of these TCRs (Fig. 6, B and C). Although there was evidence of restricted gene segment usage by Vδ6 (TRDV15-1 and 15-2)-Dδ(2)-Jδ(1) receptors and that some TCR sequences were over-represented at the different time points analyzed, particularly among Vγ1 receptors expressed before infection, there was minimal or no overlap between the Vγ1 and Vδ6 TCR sequences. These findings imply the existence of a large number of Vγ1+ T cell clones involved in the early and later stages of the host response to infection. Although Vδ6-encoded receptors do not exclusively pair with Vγ1 TCRs (41), the differences in all Vδ6 sequences analyzed makes it likely that those that do pair with Vγ1 are also different. A striking feature of the Vδ6 TCR analysis was the change in the overall charge of the CDR3 regions of the TCR cDNAs from being uniformly positive before infection to almost exclusively negative after infection, suggestive of the selective expansion of Vδ6 TCRs with particular structural features and presumably Ag specificities.
The ability of γδ T cells to interact with and alter the activity of other immune cells provides them with the opportunity and means to influence the course and outcome of inflammatory immune responses. Understanding how they do this, however, is made difficult by the conflicting effector functions that have been ascribed to pathogen-elicited γδ T cells, and whether different populations of γδ T cells perform specific functions at different stages of the immune response. The response of Vγ1+ T cells, a ubiquitous and motile population of γδ T cells, to infection with Lm is staged, occurring both before initiation of the αβ T cell response and after the bulk of the infection has been cleared (10). What controls this staged involvement and what the relationship is between early- and late-responding T cells are not fully understood, although both pro- and anti-inflammatory mechanisms that may be beneficial or deleterious to the host have been implicated (reviewed in Ref. 34). Using Vγ1-deficient mice, we have shown that they possess immunoregulatory properties and have obtained evidence for the segregation of function among γδ T cells in response to infection. Our findings demonstrate that Vγ1+ T cells are both necessary and sufficient for macrophage homeostasis and the elimination of activated macrophages, and identify an unexpected role in γδ T cell homeostasis during infection. Protection from immune-mediated pathology, however, was a property of another, as yet unidentified, γδ T cell subset(s). Our findings also provide evidence of functional diversity among a single subset of γδ T cells. Based upon their cytokine profiles, the early- and late-responding Vγ1+ T cells are functionally diverse, which appears to be attributable to the activation and staged involvement of specific Vγ1+ clones distinguished by the expression of structurally distinct TCRs. Despite this structural diversity, all the Vγ1 TCRs expressed share some important features. In particular, there is a striking conservation of the distribution of the size of Vγ1 CRD3 regions, which are virtually identical regardless of the stage of infection at which they are expressed. All TCRs, regardless of their CDR3 sequences and the stage of infection at which they are expressed, confer the ability to interact with activated macrophages. This could therefore be interpreted as evidence of strong selective pressures acting to maintain a repertoire of structurally constrained TCRs that confer functional diversity and enable Vγ1+ T cells to adapt and contribute to different stages of the immune response to infection. The selection of Vγ1+ T cells’ effector functions executed during the early and late stages of the immune response may therefore be influenced and directed by the prevailing microenvironmental conditions at the time of activation.
Previously it was shown that Vγ1/Vδ6.3+ T cells kill activated macrophages during the late stage of Lm infection (33, 42), an activity that is dependent on specific activation of the macrophages, but not of the Vγ1/Vδ6.3+ T cells, and is TCR mediated (33) and Fas-Fas ligand dependent (52). This study extends these findings by demonstrating that macrophage cytocidal activity is unique to Vγ1+ T cells and that these T cells are essential for eliminating activated macrophages and regulating activated macrophage numbers. The readiness of Vγ1/Vδ6.3+ T cells from noninfected mice to kill activated macrophages may be explained by their constitutive state of activation; in the spleen, they have been shown to constitutively express IL-12Rs (30) and markers of a memory/activation phenotype (53), consistent with their being poised for a rapid response. The inability to detect pathogen specificity among responding γδ T cells, their reactivity with macrophages activated in response to different infectious and noninfectious stimuli (33), and the reactivity of γδ T cell clones and hybridomas with unknown autologous Ags (reviewed in Ref. 10) suggest that Vγ1/Vδ6.3+ T cells are innately autoreactive (54) and that activated macrophages express an array of (self) Ags that can be recognized by large numbers of structurally diverse Vγ1/Vδ6 TCRs.
The absence in Lm-infected Vγ1−/− mice of the immune-mediated liver necrosis that is a hallmark feature of listeriosis in TCRδ−/− mice and the inability to transfer this protection to TCRδ−/− mice with splenic Vγ1+ T cells provide strong evidence for nonoverlapping function of γδ T cell subsets during pathogen-induced immune responses. This does not, however, entirely exclude the possibility that in the absence of the Vγ1+ subset of γδ T cells any protective effect they might have is compensated for by other γδ T cells. It is likely, however, that populations of γδ T cells other than Vγ1+ T cells contained within the Vγ1− fraction of splenic γδ T cells (Fig. 4) are responsible for providing protection against immune cell-mediated tissue injury. Although the mechanism(s) of tissue injury in infected TCRδ−/− mice remains unclear, the phenotype of Lm-infected Vγ1−/− mice excludes the involvement of activated macrophages that occurs as a result of a breakdown in macrophage homeostasis due to the absence of Vγ1+ T cells. A requirement for the presence of CD4+ or CD8+ T cells in the development of inflammatory lesions, however, has been demonstrated in Ab depletion studies (23). In contrast to previous studies in which Ab-mediated depletion of Vγ1+ T cells enhanced bacterial clearance (36), we found no evidence for a requirement for Vγ1+ T cells to limit or control bacterial growth in the primary sites of Lm infection beyond the first 2 days of infection. This apparent discrepancy may be related to differences in the efficiency and specificity of the Abs used for cell ablation vs gene targeting and differences in the strain of Lm used (regarding virulence and infectious dose), the route and site of infection, and the strain of mouse used, all of which influence the magnitude and kinetics of γδ T cell and Vγ1+ T cell responses (16, 34).
An unexpected role for Vγ1+ T cells in the homeostatic regulation of other γδ T cell populations in response to infection has been identified in this study. The appearance in the spleen of Lm-infected Vγ1−/− mice of Vγ5+ and Vγ7+ T cells bearing TCRs that structurally resemble those of Vγ5+ DETC and Vγ7+ iIEL suggests a role for Vγ1+ T cells in the homeostatic regulation of γδ T cells. The fact that reconstitution of Vγ1+ T cells in TCRVγ1−/− mice restores normal immune cell regulation (splenic γδ T cell repertoires and macrophage homeostasis) effectively defines Vγ1+ T cells as an immunoregulatory T cell subset (26). The presence of invariant Vγ6 TCRs (D. J. Newton and S. R. Carding, unpublished observations) in the spleens of Lm-infected Vγ1−/− mice was also unexpected. Although it was not possible to establish the origin and relationship of the cells in the periphery of infected Vγ1−/− mice and the epithelial tissues of WT animals, they may be distinct. The vast majority (∼90%) of Vγ5+ T cells in Lm-infected Vγ1−/− mice were not reactive with the DETC anti-TCRVγ5/Vδ1 clonotype Ab, 17D.1 (45) (E. M. Andrew and S. R. Carding, unpublished observations), suggesting that the majority of these T cells are not skin derived and that the Vγ5 receptor chain expressed in the spleen of Vγ1−/− mice may pair with other Vδ-chains. The coincident expression of Vδ4- and Vδ5-chains in addition to Vδ1 in the spleens of infected Vγ1−/−, but not WT, mice suggests that some of the Vγ5+ (and Vγ6+) cells may use these alternative Vδ-chains (41), and may confer specificities and functions on these cells different from their epithelia-associated counterparts. At this time it is not clear what, if any, is the functional significance of the unusual Vγ5+, Vγ6+, and Vγ7+ T cells in the spleens of Vγ1−/− mice late in the course of infection. The inherent anti-inflammatory and cytotoxic activities of mucosa-associated γδ T cells that use the same Vγ-encoded TCRs and their appearance late in the course of infection suggest that they might play a similar role in the spleens of infected Vγ1−/− mice, compensating, perhaps, for the absence of anti-inflammatory Vγ1+ T cells. What stimulates these nonresident T cells to locate to the spleen, and the regulatory mechanism by which Vγ1+ T cells control or prevent this, are not clear, although they may be attracted by chemotactic signals, such as MCP-1 (55), from activated macrophages that have not been down-modulated in the absence of anti-inflammatory Vγ1+ T cells. Future studies addressing the novel homeostatic properties of Vγ1+ T cells should resolve these issues.
Analysis of the cytokine profile of Vγ1+ T cells accumulating during the early and late stages of the immune response to Lm demonstrates that this population of γδ T cells is functionally heterogeneous. Differences in Vγ1 (and Vδ6) CDR3 sequences expressed at the early and late stages of the immune response to Lm implies the existence of unique V1γ+ T cell clones at both time points that could confer and account for the different functional characteristics of early- and late-responding Vγ1+ T cells. This functional plasticity among pathogen-elicited Vγ1/Vδ6 T cells does not fit well, however, with the idea that γδ T cell function is determined or dictated by TCRVγ usage (27, 28). Indeed, although this may be more applicable to the restricted and invariant epithelia-associated populations of γδ T cells, the data presented in this study argue for the functional heterogeneity of systemic Vγ1+ T cell responses to infection are most likely influenced by their microenvironment and conditions of activation and, therefore, by the stage of infection.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported in part by National Institutes of Health Grant AI45993 (to P.S. and S.R.C.) and The Wellcome Trust (to S.R.C.).
Abbreviations used in this paper: WT, wild type; ALT, alanine aminotransferase; DETC, dendritic epidermal T cell; iIEL, small intestinal intraepithelial lymphocyte; LAP, latency-associated peptide; Lm, Listeria monocytogenes.