Distinct Requirements for IL-7 in Development of TCRγδ Cells During Fetal and Adult Life¹

Interleukin-7 is required for development of mature TCRγδ cells. TCRγδ cells are absent from thymus, spleen, skin (dendritic epidermal T cells (DETCs)⁴) and intestinal epithelium (intraepithelial lymphocytes (IEL)) of mice deficient for IL-7, either component of the IL-7R (IL-7Rα or common γ chain, γ_c), or the γ_c signaling molecule Jak3 (1, 2, 3, 4, 5, 6, 7, 8). It has been reported that IL-7^−/− and γ_c^−/− fetal thymus contain a few immature TCRγδ^low heat stable Ag (HSA)^high thymocyte precursors to DETCs (3, 4, 6). However, neither thymus has mature TCRγδ^highHSA^low cells, and no DETCs are detectable in the epidermis of IL-7^−/− or γ_c^−/− mice (3, 9, 10). Thus, generation of mature TCRγδ cells absolutely requires IL-7/IL-7R interactions.

IL-7 is synthesized by thymus stromal cells (10, 11, 12, 13, 14, 15). We have shown that development of thymus-derived TCRγδ cells absolutely requires intrathymic IL-7 but does not require additional IL-7 in peripheral tissues (16). Thus, IL-7 must act on TCRγδ precursors within the thymus, although the molecular events downstream of an IL-7 signal and whether there are multiple stages affected by IL-7 that ultimately result in thymic production and maintenance of mature TCRγδ cells in the periphery have not been completely defined.

In vitro data demonstrate that IL-7 stimulates rearrangement of murine TCRγ genes. Addition of IL-7 to cultures of E14 fetal liver cells, adult BM T cell precursors, or mature TCRαβ HSA^lowCD4⁺ thymocytes yields in-frame, junctionally diverse TCRVγ1.2, TCRVγ2, or TCRVγ4 transcripts (8, 17, 18, 19, 20). In all instances, TCRγ mRNAs are not found in cells cultured without IL-7. Analyses of TCRγ rearrangements in thymocytes from gene-deleted mice (IL-7Rα^−/− (two independently derived lines), γ_c^−/−, or Jak3^−/− mice) also demonstrated that IL-7R signaling facilitates TCRVγ gene rearrangement (2, 3, 7, 8, 21, 22, 23, 24, 25). IL-7R signaling regulates germline transcription and the accessibility of recombinase machinery to the TCRVγ locus (26, 27, 28, 29). These data suggest that IL-7R signaling is required for the initiation of TCRγ gene rearrangement.

In contrast, other groups have reported that PCR analyses of IL-7Rα^−/−, γ_c^−/−, or Jak3^−/− thymocytes reveal that TCRVγ-Jγ gene rearrangements are present but severely reduced in comparison with wild-type thymocytes (6, 30, 31). In the case of IL-7^−/− fetal thymus, it was reported that in addition to being severely reduced in quantity, TCRVγ3 rearrangements are developmentally delayed (30). These data suggest that although γ_c cytokines greatly facilitate TCRVγ rearrangements, they are not absolutely required. If that is true, then γ_c cytokines must also be critically important for later stages of TCRγδ thymocyte development because IL-7^−/−, IL-7Rα^−/−, γ_c^−/−, and Jak3^−/− mice all lack mature TCRγδ cells.

Both chains of the heterodimeric IL-7R are integral components of other cytokine receptors. The α-chain of the IL-7 receptor pairs with a novel chain to make up the thymic stromal lymphopoietin receptor (32, 33). Numerous cytokine receptors that influence T cell development/survival, i.e., IL-2, -4, -7, -9, -15, and -21 use γ_c and Jak3 (34). Therefore, it cannot be assumed that all of the defects observed in IL-7R^−/− mice are due solely to the absence of an IL-7-mediated signal. To definitively determine whether IL-7 is required for TCRγδ gene rearrangement and to elucidate its contribution to later steps in TCRγδ cell maturation, we generated IL-7^−/− mice (1) expressing rearranged G8 TCRγδ transgenes (35, 36), thus bypassing any need for IL-7 during TCRγδ gene rearrangement. Mature TCRγδ^high cells were present in thymus and peripheral tissues of G8 IL-7^−/− mice. Moreover, endogenous TCRVγ chains were expressed on the cell surface of T cells isolated from G8 IL-7^+/− mice but not by T cells isolated from G8 IL-7^−/− mice. These data indicated that IL-7 was absolutely required for protein expression of endogenously rearranged TCRVγ genes. The number of TCRγδ cells isolated from G8 IL-7^−/− animals was drastically reduced when compared with control G8 IL-7^+/− littermates. Decreased production of TCRγδ cells by the thymus only partially accounted for this reduction. Pulse-chase experiments with BrdU revealed that the turnover rate of peripheral TCRγδ cells was higher in G8 IL-7^−/− animals than in G8 IL-7^+/− littermate control animals, suggesting that IL-7 prolonged the life span of mature TCRγδ cells.

IL-7^−/−-Ly5.1 mice were originally obtained from DNAX Research Institute of Molecular and Cellular Biology (Palo Alto, CA) and were maintained on a C57BL/6 × 129/Ola hybrid background, as previously described (1). A single line of G8 TCRγδ-transgenic mice (35) was originally obtained from Steve Hedrick (University of California, San Diego, CA), and was maintained in our animal facility. IL-7^−/− females and G8 males were crossed to obtain H-2^b/d IL-7^+/− G8^+/− F₁ progeny. Male and female F₁ mice were intercrossed to obtain F₂ progeny. G8 TCRVγ2⁺ cells are deleted by H-2^b; therefore, peripheral blood leukocytes from F₂ animals were screened, and H-2^b/b and H-2^b/d animals were removed from the breeding colony. H-2^d/d animals were screened for IL-7 by Southern blot. F₂ H-2^d/d G8^+/+ IL-7^+/− were bred with H-2^d/d G8^−/− IL-7^−/− mice, and the resultant F₃ H-2^d/d G8^+/− IL-7^+/− or IL-7^−/− animals were analyzed. All G8^+/− mice used in this study were derived from a single line of G8 mice; therefore, G8 transgene copy number and integration site were held constant. All mice were fed sterile food and water and were housed in microisolators under specific pathogen-free conditions. Their welfare was in accordance with institutional and Office of Laboratory Animal Welfare guidelines.

Approximately 10 μg of genomic DNA (prepared from 0.25 inch of tail) were digested with XbaI (Life Technologies, Gaithersburg, MD) and BamHI (New England Biolabs, Beverly, MA) and electrophoresed in a 0.8% Seakem Gold agarose gel (FMC Bioproducts, Rockland, ME). The DNA was transferred by capillary action to a Nytran membrane (0.45 μm, net neutral charge; Schleicher and Schuell, Keene, NH) in 10× standard saline citrate phosphate/EDTA. DNA was UV cross-linked to the membrane (2400 J) and hybridized overnight at 55°C in hybridization buffer (0.1 M Tris-HCl, 5 mM EDTA, 5 mg/ml heparin, 0.1% sodium pyrophosphate, 0.5% Sarkosyl, 10% dextran sulfate, 1 M NaCl, 30% formamide, 0.1 mg/ml sheared salmon sperm, pH 7.5) with a probe containing 3100 bp of intronic IL-7 gene sequence 3′ of exon 5. DNA (∼25 ng) was labeled with [α-³²P]dATP (Amersham Life Sciences, Cleveland, OH) using a Random Primers DNA Labeling System (Life Technologies). Excess, unincorporated [α-³²P]dATP was removed by filtration through a Sephadex G-50 Quick Spin column (Boehringer Mannheim, Indianapolis, IN). IL-7^−/− mice have a neomycin cassette in place of IL-7 exon 4, which results in loss of an XbaI site present in the wild-type gene (1). A 6-kb band represents the wild-type gene, whereas a 14-kb band represents the exon 4-deleted IL-7 gene.

All blots were washed in 2× SSC for 15 min at room temperature, followed by one 30-min and then one 20-min wash in 0.1× SSC plus 0.2% SDS at 65°C. Bands were visualized by exposing membranes to Biomax MS film (Kodak, Rochester, NY) for >5 h with a BioMax TranScreen-HE intensifying screen (Kodak).

Lymphocytes were isolated from thymus, spleen, and lymph nodes using a glass homogenizer and then passed through 100-μm pore size Nitex nylon mesh (Tetko, Kansas City, MO) to remove connective tissue. Splenic RBCs were lysed via two sequential incubations in Tris-ammonium chloride (13 mM Tris, 135 mM NH₄Cl, pH 7.2) for 4 min at 37°C. Before staining for flow cytometric analyses, splenocyte FcR were preblocked with affinity purified mouse IgG (200 μg/ml; Jackson ImmunoResearch Laboratories, West Grove, PA).

Small intestine (gastroduodenal junction to the ileocecal junction) was cut longitudinally, and then into 5-mm pieces, and washed twice with Ca²⁺, Mg²⁺-free HBSS containing 1 mM HEPES and 2.5 mM NaHCO₃ (pH 7.3), and 2% FCS. Washed intestinal pieces were combined, and stirred at 37°C for 20 min in Ca²⁺, Mg²⁺-free HBSS containing 1 mM HEPES and 2.5 mM NaHCO₃ (pH 7.3), with 10% FCS and 1 mM dithioerythritol (Calbiochem, La Jolla, CA). This step was repeated and the cells in the supernatants from both treatments were combined and rapidly filtered through scrubbed nylon wool (NEN, Boston, MA). Cells were then centrifuged in a 44%/67.5% Percoll (Pharmacia, Piscataway, NJ) gradient. Viable cells at the interface were collected and prepared for flow cytometric analysis.

To prepare DETC suspensions for fluorescence flow cytometric analysis abdominal and back skin was shaven with a straight razor, excess fat and blood vessels were removed, and the skin was cut into 1-cm-wide strips. Pieces were incubated epidermal side up in 0.3% trypsin (type XI; Sigma-Aldrich, St. Louis, MO) in 0.17% glucose, 0.88% NaCl, and 0.04% KCl (pH 7.6) at 37°C for 1.5 h. Epidermal sheets were separated from the underlying dermis by scraping and placed in fresh 0.3% trypsin with 0.01% DNase (ICN Nutritional Biochemicals, Cleveland, OH) at 37°C for 10 min with shaking. An equal volume of cold MEM with 10% FCS, 0.01% DNase, 100 U/ml penicillin, and 100 mg/ml streptomycin was added to inactivate the trypsin. Clumps of stratum corneum were removed by filtering the suspension through Nitex. The filtrate was then centrifuged at 1000 rpm for 10 min at 4°C in a Sorvall RTH-750. Pellets were resuspended in 4 ml of MEM with FCS, penicillin, and streptomycin and then underlaid with an equal volume of Histopaque 1083 (Gallard Schlesinger Chemical Manufacturing, Carle Place, NY). Gradients were centrifuged at 1200 rpm for 20 min at room temperature. Epidermal cells harvested at the interface (IEC) were washed once with MEM plus FCS, penicillin, and streptomycin and then counted. Viability was assessed via trypan blue exclusion. Before staining, IEC were cultured overnight in RPMI 1640 supplemented with 10% FBS, 25 mM HEPES, 20 mM l-glutamine, 10 mM sodium pyruvate, 30 mM 2-ME, nonessential amino acids, and penicillin-streptomycin, to allow reexpression of trypsin-sensitive epitopes. IEC suspensions were stained as described below.

Mice were provided with water ad libitum supplemented with 0.8 mg/ml BrdU (Sigma) for 7 days as described by Tough and Sprent (37). Where indicated, following the 1-wk BrdU labeling period, mice were returned to normal, unsupplemented water for an additional 14 days and then analyzed.

The following mAbs were used: anti-Thy-1.2-FITC or -PE (53-2.1); anti-CD3ε-FITC (145-2C11) (38) or anti-CD3ε-biotin (500A2); anti-TCRγδ-PE or -biotin (GL3) (39); anti-TCRVγ5-FITC or -biotin (GL1) (39); anti-Vγ2-FITC or -biotin (UC3-10A6); anti-TCRVγ1 (2.11) was a generous gift of P. Pereira (40); anti-TCRVγ3-FITC or -biotin (F536) (41); anti-TCRVδ4-FITC (GL2) (39); anti-TCRαβ-FITC, -Cy-Chrome, or -PE (H57.597); anti-CD8α-FITC (3.168) (42) or anti-CD8α-PE (53-6.7); anti-CD8β-FITC or -biotin (H35-17-2); anti-CD4-FITC (RM4–4) or anti-CD4-PE (GK1.5) (Becton Dickinson Collaborative Technologies, Bedford, MA) or anti-CD4-TriColor (CT-CD4) (Caltag Laboratories, South San Francisco, CA); anti-CD44-PE (IM7); anti-CD25-FITC (7D4); anti-HSA (CD24)-PE (M1/69); anti-CD62L-PE (Mel-14); anti-H-2K^b-PE (AF6-88.5); anti-H-2K^d-FITC (SF1-1.1); anti-BrdU-FITC (BD Biosciences, San Jose, CA). All mAbs were obtained from BD PharMingen (San Diego, CA) unless otherwise noted. Biotin-conjugated Abs were visualized with Streptavidin Red 670 (Life Technologies), streptavidin-PE, or streptavidin-Cy5 (Jackson ImmunoResearch Laboratories). Relative fluorescence intensities were measured with a FACScan or FACSCalibur (BD Biosciences).

A single-cell suspension of lymphocytes in PBS containing 0.2% BSA and 0.1% NaN₃ (PBS-BSA-NaN₃) was incubated with properly diluted mAb at 4°C for 20 min. After staining, cells were washed twice with PBS-BSA-NaN₃, and relative fluorescence intensities were measured by fluorescence flow cytometry. Fluorescence intensity is presented on a 4-decade log scale. A minimum of 10,000 cells within the forward scatter vs side scatter lymphocyte gate were analyzed in each sample.

Anti-BrdU staining was done as described by Tough and Sprent (37). Briefly, cells were stained with mAb conjugated to fluorochromes detected in FL2 (PE), and either FL3 (Red-670, Tricolor, or Cy-Chrome) or FL4 (APC or Cy5) as above, washed in PBS, fixed in 70% ethanol for 30 min, and then permeabilized overnight in 1% paraformaldehyde, 0.01% Tween at 4°C. Following two washes in PBS, samples were incubated with 50 U/ml DNase (DNase I from bovine pancreas; Roche Molecular Biochemicals, Mannheim, Germany) in 0.9% NaCl (pH 5.0), for 10 min at 37°C, washed once with PBS-BSA-NaN₃ and once with PBS, and then stained with FITC-conjugated anti-BrdU mAb (1/10) for 30 min at room temperature. Following washing with PBS/BSA/NaN₃, samples were resuspended in PBS and analyzed immediately by fluorescence flow cytometry.

All two-tailed Student t tests were conducted using InStat Instant Biostatistics (GraphPad Software, San Diego, CA). Error bars represent the SEM.

IL-7^−/− mice lack mature TCRγδ cells (4). To determine whether fully rearranged TCRγδ transgenes were able to bypass the need for IL-7 during TCRγδ cell development, we looked for TCRγδ lymphocytes in the peripheral lymphoid tissues of G8 TCRγδ-transgenic IL-7^−/− (G8 IL-7^−/−) mice. As expected, TCRγδ cells were absent from nontransgenic IL-7^−/− animals and present in lymph node, spleen, and thymus of G8 TCRγδ-transgenic IL-7^+/− (G8 IL-7^+/−) mice. TCRγδ cells were also present in lymph node, spleen, small intestinal IEL, and peripheral blood of G8 IL-7^−/− mice (Fig. 1 and data not shown). This demonstrated that expression of TCRγδ transgenes overcame the absence of IL-7 in generation of TCRγδ lineage cells. Moreover, because the G8 TCRVγ2 transgene was under control of its endogenous regulatory regions, IL-7 was not absolutely required for either transcription or cell surface expression of the productively rearranged TCRγ gene.

Introduction of TCRγδ transgenes increases the number of TCRγδ cells in wild-type TCRγδ-transgenic mice; however, TCRαβ allelic exclusion is not complete. Variable numbers of TCRαβ cells are present in G8 mice (35). This is analogous to results from nine independently derived lines of TCRVγ4/Vδ1-transgenic mice that were found to differ significantly in their ability to exclude TCRαβ expression (43). Some G8 TCRγδ-transgenic T cells express dual TCRαβ/TCRγδ and CD4, suggesting that the TCRγδ transgene is expressed in cells that belong to the TCRαβ lineage (36). Lymphocytes from G8 IL-7^+/− and G8 IL-7^−/− mice were analyzed for dual TCRαβ/TCRγδ and CD4⁺ or CD8β⁺ cells, an example of which is shown (Fig. 2). The percentage of TCRαβ and TCRαβ/TCRγδ cells varied between individual mice and did not correlate with age, sex, or presence of IL-7. Interestingly, CD4⁺ or CD8β⁺ cells consistently expressed lower levels of TCRVγ2, presumably due to intracellular competition between TCRαβ and TCRVγ2 for CD3 chains that facilitate transit to the cell surface. The developmental requirements of TCRαβ/TCRγδ cells likely reflected those of TCRαβ cells, rather than TCRγδ cells; therefore, cells coexpressing TCRαβ were excluded, and all subsequent analyses were done on TCRVγ2⁺TCRαβ⁻ cells.

The ability of TCRγδ transgenes to preempt the requirement for IL-7 in the thymus suggested that IL-7 was critical for a stage of T cell development that preceded expression of a TCRγδ. This was consistent with either IL-7 directly stimulating TCRγ gene rearrangement in normal mice or an indirect effect, i.e., premature expression of TCRγδ providing a survival signal to TCRγδ precursors. To begin to dissect this, the expression of endogenous TCRγ genes was examined. TCRγ genes are not strictly allelically excluded at the level of gene rearrangement (44). Therefore, we reasoned that if IL-7 merely provided a survival signal to TCRγδ precursors, then both G8 IL-7^+/− and G8 IL-7^−/− mice would have some TCRγδ cells that expressed endogenous TCRγ chains on their cell surface. However, if IL-7 was required to initiate TCRγ rearrangements, then endogenous TCRγ could be present only on the surface of T cells from G8 IL-7^+/− mice and not on T cells from G8 IL-7^−/− mice.

TCRVγ usage varies with anatomic location. Most peripheral blood and splenic TCRγδ cells express TCRVγ2, the same V region encoded by the transgene. Therefore, we analyzed IEL from the small intestine or skin of G8 IL-7^−/− mice or G8 IL-7^+/− littermates for surface expression of their characteristic TCRVγ chains. TCRVγ5 is the predominant TCRVγ region used by small intestinal IEL (40, 45, 46), and TCRVγ3 is used exclusively by DETC in murine skin (47, 48). Although the vast majority of TCRγδ IEL in the small intestine or skin of G8 mice were TCRVγ2⁺, it was not the only population of TCRγδ cells in G8 IL-7^+/− mice (Fig. 3,A). A small population of small intestinal IEL (∼2%) isolated from G8 IL-7^+/− mice expressed other TCRVγ, either alone (TCRVγ5 or TCRVγ1), or dual TCRVγ2/TCRVγ5 or TCRVγ2/TCRVγ1 (Fig. 3,A). A larger percentage of DETC (30–40%) isolated from G8 IL-7^+/− mice expressed other TCRVγ, either exclusively TCRVγ3⁺ or dual TCRVγ2/TCRVγ3 (Fig. 3,B, top and middle). This result was consistent with another TCRVγ2-transgenic mouse line, KN6, that has TCRγδ⁺ DETC that do not express the transgenic TCRVγ2/Vδ5 (49). T cells expressing endogenous tissue-characteristic TCRVγ were absent in IEL and DETC of IL-7^−/− mice (Fig. 3). Even after 3 wk of expansion in vitro with Con A plus IL-2, 10% of G8 IL-7^+/− DETC expressed TCRVγ3, whereas G8 IL-7^−/− DETC expressed only TCRVγ2⁺ (Fig. 3 B, bottom). Thus, IL-7 had a direct effect on TCRVγ gene expression within TCRγδ cells that ultimately resided in intestinal or skin epithelium.

The presence of TCRγδ cells in peripheral lymphoid tissues of G8 IL-7^−/− mice demonstrated that by directing surface expression of TCRγδ, TCRγδ transgenes bypassed at least the earliest requirement for IL-7 during TCRγδ cell development. In vitro, IL-7 enhances proliferation and survival of TCRγδ cells (50, 51, 52). Therefore, we analyzed the number of TCRγδ cells in G8 IL-7^−/− mice. The density of DETC in the skin of 8.5-wk-old mice was not significantly different between G8 IL-7^−/− and G8 IL-7^+/− littermates (Fig. 4). However, the absolute number of TCRVγ2⁺ cells in all other peripheral lymphoid tissues of adult G8 IL-7^−/− mice was significantly reduced when compared with age-matched G8 IL-7^+/− control mice (Fig. 4). Such decreases could have resulted from reduced production, proliferation, or life span of TCRγδ cells. The contribution of these three nonmutually exclusive possibilities was evaluated.

Most G8 TCRγδ cells are thymus derived (53). The absolute number of TCRVγ2⁺TCRαβ⁻ thymocytes isolated from G8 IL-7^+/− and G8 IL-7^−/− mice of various ages was determined (Fig. 5). The number of TCRVγ2 thymocytes isolated from fetal or neonatal G8 IL-7^−/− mice was 10-fold less than from G8 IL-7^+/− littermates. However, after 3 wk of age, there was a significant decrease in the number of TCRVγ2 thymocytes isolated from G8 IL-7^−/− animals. The age-dependent onset of the decline, along with its gradual nature, suggested that the reduction was due to a change in the developmental requirements of fetal vs adult thymocytes, rather then a developmental arrest of immature TCR⁺ thymocytes. Accordingly, the ratio of HSA^high to HSA^low TCRγδ cells was the same in thymocytes isolated from G8 IL-7^−/− and G8 IL-7^+/− mice at all time points analyzed (data not shown). Thus, a reduction in the number of TCRγδ cells produced by the adult thymus partially explained the decrease in peripheral TCRγδ cells isolated from G8 IL-7^−/− mice.

The phenotype of splenic and lymph node TCRVγ2 cells was assessed in G8 IL-7^−/− mice and their G8 IL-7^+/− littermates. The percentages of TCRVγ2 cells expressing Thy-1, CD62 ligand, CD45RB, or HSA were not significantly different (p ≥ 0.3) between G8 IL-7^−/− and G8 IL-7^+/− mice (Fig. 6). Although some difference in expression of CD44 was observed (p = 0.067), the biological significance of changes in expression of this particular phenotypic marker when accompanied by negligible changes in the others is uncertain. In addition, the paradigm of assignment of activation status as applied to TCRαβ cells may not be as relevant for TCRγδ cells (53). Thus, our conclusion was that although drastically reduced numbers of TCRγδ cells were present in the periphery of G8 IL-7^−/− mice, few, if any, significant phenotypic differences were detected. This suggested that similar populations of TCRγδ cells developed in both the presence and the absence of IL-7.

To investigate the effects of IL-7 on TCRγδ cell proliferation in vivo, G8 IL-7^−/− and G8 IL-7^+/− mice were provided with BrdU-supplemented water for 1 wk. BrdU is a thymidine analog that is incorporated into the DNA of dividing cells (37). After 7 days of continuous BrdU administration, TCRVγ2 lymph node cells isolated from either G8 IL-7^+/− or G8 IL-7^−/− mice were 30–40% BrdU⁺. However, the mean fluorescence intensity (MFI) of BrdU staining in TCRγδ cells isolated from G8 IL-7^−/− mice was higher than that in G8 IL-7^+/− lymph node cells (Fig. 7). Somewhat different results were obtained with TCRγδ IEL. In contrast to lymph node TCRγδ cells, after 7 days of BrdU treatment, the percentage of G8 IL-7^−/− IEL that were BrdU⁺ was approximately twice that of G8 IL-7^+/− IEL (mean, 33 and 15%, respectively). As in lymph nodes, the MFI of G8 IL-7^−/− IEL was greater than that of G8 IL-7^+/− IEL (Fig. 7). Thus, within the IEL compartment, IL-7 influenced both the fraction of cells dividing, as well as the levels of BrdU incorporated into individual cells.

To investigate the effects of IL-7 on the life span of TCRγδ cells in vivo, G8 IL-7^−/− and G8 IL-7^+/− mice were fed BrdU-supplemented water for 1 wk and then returned to unsupplemented water for the next 2 wk. For the purposes of these analyses, the definition of the life span of a cell was the time between when it arose from a dividing precursor until it either divided or died (53). As BrdU⁺ cells continue to divide during a 2-wk chase period, BrdU that incorporates during the week of labeling becomes diluted and thus reduces the MFI of BrdU staining. As a measure of the life span of TCRγδ cells, the percentage of BrdU^high cells remaining after 2 wk without BrdU was compared with the percentage of BrdU^high cells present after 7 days of continuous BrdU administration (Fig. 8). During the 2-wk chase period, G8 IL-7^−/− mice incurred a greater loss of BrdU^high cells. An example is shown in Fig. 8. In lymph nodes, 9% of G8 IL-7^+/− and 3% of G8 IL-7^−/− TCRVγ2⁺ cells remained BrdU^high after 2 wk of unsupplemented water. This corresponded to ∼30% (9 of 31) and ∼5% (3 of 58) of the day 7 values. Thus, the percentage of BrdU^high cells lost during the 2-wk chase period was greater in G8 IL-7^−/− mice (70% vs 95%). The same was true for TCRVγ2⁺ IEL. In the example shown, 9% of G8 IL-7^+/− and 14% of G8 IL-7^−/− TCRVγ2⁺ cells remained BrdU^high, corresponding to ∼65% (9 of 14) and ∼40% (14 of 36) of the day 7 values. This indicated that G8 IL-7^−/− mice lost nearly twice as many BrdU^high IEL during the 2-wk chase period (35% vs 60%). Thus, IL-7 prolonged the life span of TCRγδ cells. Taken together, both decreased thymic production of new TCRγδ cells and decreased life span of mature TCRγδ cells in the periphery contributed to the 100-fold decrease in peripheral TCRγδ cells in adult G8 IL-7^−/− mice.

Expression of G8 TCRγδ transgenes bypassed the absolute requirement for IL-7 in TCRγδ cell development. Mature HSA^lowTCRγδ⁺TCRαβ⁻ cells were present in the thymus and peripheral tissues of G8 IL-7^−/− mice, albeit in reduced numbers when compared with G8 IL-7^+/− littermates. The reduction in TCRγδ cell number resulted from both decreased de novo generation of T cells in the adult thymus and decreased life span of peripheral TCRγδ cells. T cells expressing nontransgenic TCRγ, or dual transgenic/endogenous TCRγ chains were present in G8 IL-7^+/− mice but absent from G8 IL-7^−/− mice. These data allowed a number of inferences to be made about the role of IL-7 in TCRγδ cell development.

Because productively rearranged TCRγδ genes were sufficient to bypass the developmental blockade, IL-7 must have been required for rearrangement of TCRVγ genes, either directly by influencing rearrangement itself or indirectly by stimulating survival of precursors to a developmental stage at which they could attempt TCRγ rearrangement. The latter was less likely because TCRγ and TCRδ loci undergo rearrangement concurrently beginning at the pro-T cell stage (CD44⁺CD25⁺) (54, 55) and TCRδ rearrangements are normal in IL-7R^−/− mice (6, 8, 21, 22, 23). Moreover, although the thymus cellularity of E14 IL-7R^−/−, IL7^−/−, or Jak3^−/− embryos is drastically reduced, the relative percentages of CD44/CD25 triple-negative subsets are normal, indicating that immature thymocytes differentiate normally (56).

We showed that T cells expressing nontransgenic or dual transgenic/endogenous TCRγ proteins on the cell surface were present in G8 IL-7^+/− mice but absent in G8 IL-7^−/− mice. These data strongly suggested that IL-7 stimulated rearrangement of endogenous TCRγ genes and, in the case of IEL, were consistent with the absence of TCRVγ5 rearrangements in IL-7Rα^−/−, γ_c^−/−, or Jak3^−/− thymocytes (6, 7, 8, 21, 22, 23, 31). Moreover, they are the first cell surface protein expression data showing that murine TCRγ loci are not strictly allelically excluded. Because TCRδ genes are also allelically included (44), any given TCRγδ cell has the potential to express four different TCR on its surface simultaneously. Perhaps the expression of TCRVγ genes in ordered waves (40, 54, 57, 58) is a means of limiting the number of different TCR expressed on individual cells.

IL-7R signaling has been shown to directly influence TCRγ rearrangement. IL-7 affects accessibility of the TCRVγ locus to the recombinase machinery. IL-7-mediated activation of STAT5 (28, 33, 59) up-regulates the sterile transcripts that precede the appearance of TCRγ gene rearrangements. Both TCRγ enhancers and 5′-HsA, a newly described regulatory region upstream of TCRVγ2 that is required for consistent rearrangement of TCRVγ2 transgenes, have STAT5 binding motifs (60, 61, 62). STAT5 activated in response to IL-7 stimulation leads to sterile TCRγ transcripts in IL-7-dependent pre-B cell lines, and a constitutively active form of STAT5A restores TCRV-Jγ rearrangements in IL-7Rα^−/− thymocytes (7). Also, it has been shown that IL-7 renders TCRγ loci accessible to the recombinase machinery by modifying histone acetylation (26, 28, 29). Like many silenced genes, TCRVγ loci are highly methylated in IL-7Rα^−/− thymocytes, and pretreatment of IL-7Rα^−/− precursors with a histone acetylase (TSA) restores TCRV-Jγ rearrangements in FTOC (23).

IL-7 was not absolutely required for either terminal differentiation steps within the thymus or survival of mature TCRγδ cells. TCRγδ⁺TCRαβ⁻ cells were present in the thymus and peripheral tissues of G8 IL-7^−/− mice. DETC develop exclusively in fetal thymus, from fetal precursors; thus, DETC in the skin of adult animals are the progeny of cells that arose during the fetal period (47, 48). The presence of TCRγδ DETC in adult G8 IL-7^−/− mice indicated that once a TCRγδ was expressed, TCRγδ cells developed normally in fetal thymus, and then survived >8 wk in the absence of IL-7. This is in agreement with the pattern of expression of IL-7Rα in that precursors to DETC in the fetal thymus express IL-7Rα, but mature TCRγδ cells in the skin do not (L. Puddington, J. M. Lewis, and R. E. Tigelaar, unpublished observations). The ability of TCRγδ cells to survive in the periphery without IL-7 was consistent with the presence of splenic TCRγδ cells in TCRVγ2-transgenic IL-7Rα^−/− mice (8) and the results of our earlier thymus grafting experiments, in which IL-7⁺ thymus graft-derived TCRγδ cells were found in spleen and IEL of IL-7^−/− hosts up to 10 wk postgrafting (16).

The developmental requirements of fetal vs adult thymocytes differed in their dependence upon IL-7. Fetal thymus-derived DETC (47, 48) are absent from the skin of non-TCR-transgenic IL-7^−/− mice (10). A TCRγδ transgene restored a normal density of TCRVγ2⁺ DETC to the skin of G8 IL-7^−/− mice. Similar results were obtained in the skin of TCRVγ3/Vδ1 (the canonical fetal DETC-type TCR)-transgenic IL-7Rα^−/− mice (63). This indicated that in addition to a role in the maintenance of T cell progenitors (56), fetal TCRγδ thymocytes were dependent on IL-7 for rearrangement of TCR. In contrast, beginning at ∼4 wk of age, the number of TCRVγ2 thymocytes sharply declined, until very few TCRVγδ thymocytes matured in the thymus of adult G8 IL-7^−/− animals. This correlated with the transition from the first to the second wave of lymphoid precursor cells that seed fetal thymus (64), suggesting that TCRγδ thymocytes developing later in ontogeny also required IL-7 for survival and/or proliferation. These results were consistent with the paucity of TCRγδ cells found in adult TCRVγ3/Vδ1-transgenic IL-7Rα^−/− mice, TCRVγ1-transgenic γ_c^−/− mice, or G8 TCRγδ-transgenic Jak3^−/− mice (6, 31, 63). Moreover, a similar conclusion was reached studying TCRαβ development in IL-7Rα^−/− mice, i.e., survival of adult, but not fetal, CD25⁺ double-negative thymocytes is IL-7Rα dependent (65).

The concept that fetal and adult lymphocyte progenitor cells are inherently different was first suggested by the differential usage and junctional diversity of TCRVγ gene segments in fetal vs adult thymocytes (46, 47, 48). It has now become a common finding in developmental immunology and is exemplified by mice deficient in cytokines or chemokines, i.e., stem cell factor, IL-7, stromal cell-derived factor-1 (65, 66, 67, 68, 69, 70), transcription factors, i.e., Ikaros, T cell factor-1 (71, 72), adhesion molecules, i.e., α₄ integrins (73), or receptors for hormones, i.e., estrogen (74). In light of this, it is important to reevaluate the conclusions reached in previous studies where the effects of age were not considered and to better control for age-related parameters in future studies.

A similar decline in thymocyte number was not reported in 2- to 5-wk-old TCRVγ2-transgenic IL-7Rα^−/− mice (8). In that study, it is not clear that dual TCRαβ/γδ cells (see Fig. 2) were gated out during FACS analysis of TCRVγ2-transgenic IL-7Rα^−/− cells. Inclusion of dual TCRαβ/TCRVγ2 cells in the absolute cell numbers could have masked the onset of TCRVγ2⁺ thymocyte decline that would have only just begun in 4- to 5-wk-old IL-7Rα^−/− mice. Another possibility is the TCRVγ2-transgenic IL-7Rα^−/− mice were on a pure B6 background, whereas the data presented here were obtained from the study of mice on a mixed background (B6 × 129/Ola × BALB/c). Background genes can modulate the IL-7 dependence of developing thymocytes, as exemplified by exon 3 IL-7Rα^−/− mice. On a mixed B6 × 129/J background, ∼70% of mice have thymocytes that are completely arrested at the double-negative stage (24, 65), whereas on a pure B6 background, IL-7Rα^−/− TCRαβ thymocyte development progresses through the single positive stage in all mice (30). Similarly, IL-7^−/− mice on a FVB/N background have a more severe phenotype than IL-7^−/− mice on a B6 × 129/Ola or a 129/Ola background (K. Laky, U. von Freeden Jeffry, B. E. Rich, R. Murray, and L. Puddington, unpublished observations).

In lymph node or small intestinal IEL, similar or greater percentages of TCRγδ cells were proliferating in IL-7^−/− mice than IL-7^+/− mice. Presumably, the same was also true for TCRγδ cells in the skin because TCRVγ2⁺ cells were 10-fold reduced in E15-E18 fetal thymus, but not in the skin of G8 IL-7^−/− mice. The stimulus for that proliferation was unclear, but it was not IL-7, and presumably it was not Ag driven, because all G8 IL-7^−/− T cells express a single, clonotypic TCR. It is possible that IL-2 or IL-15 cytokine signaling via IL-2Rβ could have provided this survival signal. Whereas expression of Vγ3/Vδ1 transgenes rescues development of fetal thymocytes and DETC in IL-7Rα^−/− mice, it is not able to rescue DETC in IL-2Rβ^−/− mice (63). In G8 IL-7^−/− mice of all ages, the number of peripheral TCRγδ cells was reduced ∼100-fold. It is possible that cells were merely undergoing homeostatic proliferation in an attempt to fill the available space in IL-7^−/− mice. Such a space-filling model was first suggested by experiments in which mature thoracic duct lymphocytes were adoptively transferred to lymphocyte-deficient scid or nude mice and observed to expand in a manner inversely proportional to the size of the inoculum, before being maintained at a steady state number (75, 76, 77, 78). If this were the case, it would represent a difference between TCRγδ and TCRαβ cells because IL-7 has been found to regulate of homeostatic proliferation of TCRαβ cells (79).

Similar percentages of lymph node T cells proliferated in the presence or absence of IL-7. In contrast, the percentage of IEL dividing in G8 IL-7^−/− mice was greater than in G8 IL-7^+/− mice. This reemphasizes the differential regulation of TCRγδ cell homeostasis in the small intestinal IEL compartment vs peripheral tissues. We have previously shown that T cell requirements for stem cell factor/c-Kit interaction differ at these two anatomic locations. In W/W^v and Sl/Sl^d mice, both TCRαβ and TCRγδ IEL homeostasis is disrupted in the absence of stem cell factor-c-Kit interactions, whereas peripheral T cell populations are unaffected (66, 67).

In the absence of IL-7 the life span of TCRγδ cells was shortened. The pulse chase experiments described here did not allow discrimination between cell death and dilution of label due to division. However, we favor an explanation that includes at least a component of the latter. During the 7-day pulse period, BrdU⁺ cells in G8 IL-7^−/− animals had higher MFIs, results that could suggest that cells were proliferating at a higher rate in the absence of IL-7. It is possible that they continued to do so during the following 14-day chase period. Moreover, we have previously shown that IL-7 is not required for survival of mature TCRγδ cells in the periphery (Ref.16 and see above).

In summary, we have used a TCRγδ-transgenic IL-7^−/− model system to evaluate the requirements of TCRγδ cells for IL-7 in the thymus and peripheral lymphoid tissues of mice ranging from E15 through 35 wk of age. With regard to thymocytes, IL-7 was required for endogenous TCRγ gene rearrangements in both fetal and adult thymus, and adult thymocytes also required IL-7 for survival. With regard to peripheral TCRγδ cells, IL-7 was not required for the survival or proliferation of TCRγδ cells, although it prolonged the life span of individual TCRγδ cells.

We acknowledge DNAX Research Institute of Molecular and Cellular Biology for their gift of IL-7^−/− mice; Pablo Pereira for mAb against TCRVγ1; Elizabeth G. Lingenheld for technical assistance; and Leo Lefrançois, David Tough, and B. J. Fowlkes for helpful discussions of the data.