当前位置: 首页 > 医学版 > 期刊论文 > 基础医学 > 免疫学杂志 > 2005年 > 第4期 > 正文
编号:11254163
Distinct Effects of TGF-1 on CD4+ and CD8+ T Cell Survival, Division, and IL-2 Production: A Role for T Cell Intrinsic Smad
http://www.100md.com 免疫学杂志 2005年第4期
     Abstract

    TGF-1 is critical for maintaining T cell homeostasis. Smad3 has been implicated in this regulatory process, yet the cellular targets and molecular details remain poorly understood. In this study, we report that TGF-1 impairs the entry of CD4+ and CD8+ T cells into the cell cycle as well as their progression through subsequent rounds of division, and show that Smad3 is essential for TGF-1 to inhibit TCR-induced division of only CD4+ and not CD8+ T cells. Both CD8+ and CD4+ T cells from Smad3–/– mice were refractory to TGF-1-induced inhibition of IL-2 production, thus demonstrating that not all CD8+ T cell responses to TGF-1 are Smad3 independent. These TGF-1 effects were all T cell intrinsic, as they were reproduced in purified CD4+ and CD8+ T cells. Finally, we found that Smad3 was critical for the survival of CD8+, but not CD4+ T cells following activation ex vivo. The TCR-induced death of Smad3–/– CD8+ T cells was not dependent upon TNF- production. Exogenous TGF-1 partially rescued the CD8+ T cells by signaling through a Smad3-independent pathway. TGF-1 also enhanced survival of TCR-stimulated CD4+CD44high T cells in a Smad3-independent manner. Collectively, these findings firmly establish for the first time that TGF-1 discriminately regulates CD4+ and CD8+ T cell expansion by signaling through distinct intracellular pathways.

    Introduction

    Transforming growth factor-1 plays a critical regulatory role in diverse immune responses ranging from bacterial infection and tumor rejection to immune tolerance and suppression of autoimmune disorders; however, the precise mechanisms of this regulation are poorly defined. The role of TGF-1 as an immune modulator is probably best exemplified by the rapid onset of lethal multiorgan inflammation in TGF-1 knockout and TGF-1 receptor type II (TRII)2 dominant-negative (DN) transgenic mice (1, 2, 3, 4). TGF-1 and its membrane-bound type I and II receptors are ubiquitously expressed, and TGF-1 exerts direct regulatory effects on a broad repertoire of immune cells, including T and B lymphocytes, dendritic and NK cells, macrophages, monocytes, and neutrophils (5, 6). Prolonged survival of TGF-1–/– scid (7), TGF-1–/– Rag1–/– (8), TGF-1–/– Rag2–/– (9), and TGF-1–/– MHC class I–/– mice (10) strongly implicates lymphocytes as primary effectors in the immune disruption in mice lacking TGF-1 or expressing DN TRII. Cell-specific targeted deletion of TGF-1 signaling using promoter-driven DN TRII expression has demonstrated that TGF-1 negatively regulates CD4+ as well as CD8+ T cell expansion in vivo (3, 4). The absence of spontaneous T cell activation in TCR transgenic Rag–/– TRII DN mice suggests that TGF-1 plays a key role in maintaining immunological balance by suppressing Ag-induced T cell responsiveness; yet it remains unclear how these effects are mediated (5).

    TGF-1 binds to a heteromeric transmembrane complex consisting of type I and II serine/threonine kinase receptors to initiate a cascade of signaling events at the cell surface. Smad proteins remain the only known direct TGF-1-signaling effectors; however, Smad-independent signaling pathways have also been described (11, 12, 13). Receptor-activated Smads, i.e., Smad2 and Smad3, heterodimerize with the common Smad4 and translocate into the nucleus, where they, in a cooperative manner with other nuclear cofactors (e.g., transcription factors and chromatin-remodeling regulatory proteins), modulate transcription of target genes. The precise connection of alternative, Smad-independent pathways, e.g., protein phosphatase 2A, rho, and MAPK, to TGF-1 receptor activation and how they contribute to TGF-1 responses are not clear.

    The balance of proliferation and apoptosis is critical to maintain T cell homeostasis. Resolution of immune responses through apoptosis can potentially involve several mechanisms, including direct elimination of activated T cells by programmed cell death, suppression of proinflammatory cytokines (e.g., TNF-, IL-1, and IL-12) (14, 15), and production or activation of local anti-inflammatory cytokines (e.g., IL-10 and TGF-1) (16, 17). The enormous T cell apoptosis storm observed in the TGF-1 knockout mouse illustrates the necessity of TGF-1 for the coordinated clearance of activated apoptotic cells with the release of anti-inflammatory cytokines for restoration of T cell homeostasis (18). In vivo/in vitro studies using the TGF-1 knockout mouse have provided evidence that TGF-1 mediates its T cell antiapoptotic effects through multiple mechanisms, including regulation of Fas and TNF signaling (17, 18, 19, 20).

    TGF-1 is also believed to use multiple signaling pathways to inhibit proliferation. These include direct effects via transcriptional regulation of cell cycle target genes (e.g., cyclins, cyclin-dependent kinases and their inhibitors, and c-myc), as well as indirect effects through modulation of cytokine production (e.g., IL-2 and IL-15) (21, 22, 23, 24, 25, 26, 27, 28). In addition, compelling evidence has been put forth suggesting that TGF-1, at least under certain conditions, imparts its growth-inhibitory effects through regulatory T cells, although it remains unclear exactly how these regulatory effects are mediated (29, 30, 31). Smad3 has been shown to be critical for TGF-1-induced growth inhibition in T cells, but interestingly, not in B cells (32, 33). More recently, using CFSE labeling, we showed that while the ability of TGF-1 to inhibit TCR-induced cell division was greatly reduced in Smad3–/– splenic T cells, its effects are not absent (28). Moreover, we have demonstrated that Smad3 is not essential for inhibition of TGF-1-mediated IL-2-induced proliferation. These Smad3-independent events presumably occur through either Smad2-dependent (e.g., cdc25 or Jak/STAT signaling) or Smad-independent (e.g., protein phosphatase 2A, rho, or MAPK) signaling.

    The incomplete loss of TGF-1-induced inhibition of TCR-mediated proliferation in the Smad3–/– T cells was puzzling and prompted us to further investigate whether this was due to a differential role for Smad3 in CD4+ vs CD8+ T cell proliferation. In addition, because of the ubiquitous nature of TGF- receptors and the complexity of TGF-1 signaling, we compared the responsiveness of Smad3–/– splenocytes with purified CD4+ and CD8+ T cells to determine whether Smad3 mediates its regulatory effects in a T cell intrinsic manner. In this study, we establish two key points regarding the mechanisms whereby TGF-1 regulates T cell expansion. First, the intracellular signaling pathways used by TGF-1 to inhibit CD4+ T cell division are mechanistically distinct from those used to inhibit CD8+ T cell division. Second, Smad3 plays a distinct role in the induction of apoptosis, but not in proliferation, of CD8+ T cells following activation through the TCR.

    Materials and Methods

    Animals

    Smad3–/– mice were provided from the laboratories of S. Wahl, A. Roberts, and J. Letterio. These mice were generated as previously described (32, 34). Wild-type (WT) 129SV x C57BL/6 littermates were used as controls.

    Ex vivo T cell activation

    Mice were sacrificed via cervical dislocation. Spleen and lymph nodes (inguinal, axillary, brachial, cervical, mesenteric, and periaortic) were isolated, and single cell suspensions were generated. RBC were removed from spleen preparations by ammonium chloride potassium lysing buffer (Quality Biologicals). A total of 2 x 105 freshly isolated spleen and/or lymph node T cells, combined as indicated, was stimulated with either 3 μg/ml or 10 μg/ml plate-bound anti-CD3 (145-2C11) (BD Pharmingen) and soluble anti-CD28 ascites (diluted 1/5000) in 200 μl of RPMI 1640:Eagle’s Hanks’ amino acids (1:1) medium containing 10% FCS, 4 mM glutamine, 200 μg/ml penicillin, 200 μg/ml streptomycin, 25 μg/ml gentamicin, and 50 μM 2-ME in Costar 3596, 96-well flat-bottom culture plates (Corning Glass). Reconstituted, active human rTGF-1 (R&D Systems) was added directly to the cell cultures.

    Cytokine neutralization

    To block IL-2 signaling, anti-IL-2 (S4B6; 20 μg/ml), anti-IL-2R (PC61; 20 μg/ml), and anti-IL-2R (TM-1; 20 μg/ml) were added at the time of activation. To block TNF- signaling, anti-TNF- (TN3-19.12; 20 μg/ml), anti-TNFRI (55R-170; 10 μg/ml), and anti-TNFRII (TR75-32; 20 μg/ml) were added. Survival of the TNF--sensitive WEHI-13VAR and L929 mouse fibroblast cell lines (American Type Culture Collection) and IL-6 production in activated T cells as determined by ELISA (Quantikine-M IL-6; R&D Systems) were used as indicators of TNF- bioavailability. All neutralizing Abs were purchased from BD Pharmingen.

    IL-2 secretion

    The IL-2 PE cytokine secretion assay detection kit (Miltenyi Biotec) was used to identify cells secreting IL-2. Briefly, 2 x 104 T cells in 200 μl were stimulated in Costar 3596, 96-well flat-bottom culture plates (Corning Glass) for 24 h, and then harvested, washed in cold buffer (PBS supplemented with 5% FCS), and resuspended in 100 μl of cold medium (RPMI 1640 supplemented with 5% FCS) containing the IL-2 catch reagent. Following incubation (10 min on ice), 12 ml of prewarmed (37°C) medium was added, and the cells were incubated with continual mixing at 37°C. After 45 min, the cells were washed twice in cold buffer and processed for FACS analysis using IL-2 PE (Miltenyi Biotec), CD4 allophycocyanin (GK1.5), CD8 FITC (RPA-T8), and 7-aminoactinomycin D (7-AAD) (Boehringer Mannheim), and analyzed on a FACSCalibur (BD Immunocytometry Systems). The total amount of IL-2 secreted over the entire 24 h was determined by ELISA (Quantikine-M IL-2; R&D Systems) from cell culture supernatants. For quantifying CD4+ or CD8+ T cell subsets, splenocytes or combined splenocytes and lymph nodes were sorted at the National Institute of Allergy and Infectious Diseases flow facility. Briefly, cells were labeled with a mixture of PE-conjugated Abs to I-Ek (17.3.3S), B220 (RA3/6B2), CD11b (5C6), CD16/32 (2.4G2), and CD8 PE (RPA-T8) or CD4 PE (GK1.5) and negatively sorted to >98% purity for CD4+ or CD8+ T cells, respectively. All Abs were purchased from BD Pharmingen.

    CFSE analyses

    For CFSE labeling, 5 x 106 cells were washed twice in staining buffer (PBS supplemented with 0.1% BSA) and resuspended in 250 μl of staining buffer; 250 μl of freshly diluted 2 μM CFSE (Molecular Probes) was added, and cells were incubated at room temperature for 10 min with frequent mixing. An equal volume of FCS was added, and cells were washed three times with culture medium containing 10% FCS. Cells (2 x 104 T cells in 200 μl) were then activated in Costar 3596, 96-well flat-bottom culture plates (Corning Glass) with plate-bound anti-CD3 + soluble anti-CD28, as described above. The cell division status of cells was determined by measuring CFSE fluorescence of 7-AAD negative-gated cells at 72 h after activation. A total of 106 cells was incubated for 30 min in FACS buffer (PBS plus 1% BSA plus 0.01% sodium azide) with 1 μg/ml anti-CD4 (GK1.5) PerCP and 1 μg/ml anti-CD8 (53.6-7) allophycocyanin. Before labeling with specific Abs, FcRs were blocked with 10 μg/ml 2.4G2 Ab. After two washes, the cells were analyzed on a FACSCalibur. CFSE fluorescence was detected with the FL1 detector (530 ± 30 nm). Data were analyzed using CellQuest (BD Biosciences). Dead cells were excluded by 7-AAD staining, according to the manufacturer’s protocol (BD Biosciences). In some experiments, anti-CD69 (H1.2F3) (BD Biosciences) and forward side scatter were used to monitor cell activation status. Quantitation of CFSE dilution was determined, as previously described (35). Briefly, the TCR-+ (clone H57-597; BD Biosciences) 7-AAD– cells within each division peak (i) were enumerated by calculating the area under the curve (CellQuest software; BD Biosciences). Each of these values was divided by 2i and normalized to 100% to determine the number of input cells contributing to each peak. These input values were then plotted against division number to generate a normal population distribution. The mean division number for the population was determined from a best fit of these data to a Gaussian curve using Prism4 (GraphPad). Unstimulated, CFSE-labeled cells were used to verify the peak corresponding to the undivided population.

    Cell counts and percentage of apoptosis were determined with modifications, as previously described (36). Following ex vivo stimulation, cells were harvested, washed in FACS buffer, and incubated in FACS buffer containing anti-CD4 and anti-CD8 for 30 min at 4°C. Cells were then washed, and annexin V staining was performed in HBSS containing calcium and magnesium for 10 min at 37°C, according to the manufacturer’s protocol (BD Biosciences). Cells were resuspended in a constant volume of FACS buffer containing 7-AAD (0.25 μg/sample). A FACSCalibur (BD Biosciences) was used to acquire cells for analyses for a fixed time (20 s) at a constant flow rate. From these acquisitions, the number of viable cells from duplicate wells was averaged to obtain the percentage of apoptotic (7-AAD+ and/or annexin V+) cells. In some instances, total cell numbers were enumerated using trypan blue and hemocytometer counts.

    Western immunoblotting

    CD4+ or CD8+ splenic or lymph node T cells were purified by negative selection from C57BL/SV129 mice using R&D Systems T cell enrichment kits (R&D Systems). Greater than 95% purity was achieved, as determined by flow cytometry. Whole cell lysates from 5 x 106 splenic or lymph nodal CD4+ or CD8+ T cells were prepared in lysate buffer supplemented with protease inhibitors (Roche Diagnostics), 1 mM DTT, 1 mM PMSF, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 mM sodium orthovanadate for 30 min on ice. Samples were centrifuged at 13,000 x g for 20 min, and the supernatant was retained. Protein was quantified using the bicinchoninic acid assay (Pierce). Aliquoted samples were stored at –80°C. Whole cell splenocyte or lymph node extracts (25 μg/lane) were resolved by 12% SDS-PAGE, transferred to nitrocellulose membranes (Amersham Biosciences), and immunostained with rabbit polyclonal anti-Smad3 (LPC3, 1 μg/ml; Zymed Laboratories) or rabbit polyclonal anti-Smad2 (MHA2, 1 μg/ml; Zymed Laboratories) Abs. Bound Abs were visualized by incubation with HRP-conjugated donkey anti-rabbit IgG (Amersham Biosciences), followed by ECL detection (Pierce). Membranes were stripped and reprobed with the mouse actin AC-40 mAb (1 μg/ml; Sigma-Aldrich) as a loading control.

    Semiquantitative RT-PCR

    CD4+ or CD8+ splenic or lymph node T cells were purified by negative selection from C57BL/SV129 mice using R&D Systems T cell enrichment kits (R&D Systems). Greater than 95% purity was achieved, as determined by flow cytometry. Total RNA was isolated from 5 x 105 purified cells with the RNeasy mini kits (Qiagen) with Rnase-free DNase (Qiagen) applied directly to the RNeasy columns. Reverse transcription into cDNA was performed using TaqMan (Applied Biosystems) RT-PCR transcription reagents and random hexamers, according to the manufacturer’s recommendations. PCR for Smad2, -3, -4, and -7 mRNA was performed using the following primers derived from previously published sequence data (37): Smad2 (sense, CCCACTCCATTCCAGAAAAC; antisense, GAGCCTGTGTCCATACTTTG); Smad3 (sense, GTTGGACGAGCTGGAGAAG; antisense, GTAGTAGGAGATGGAGCAC); Smad4 (sense, AAGGTGGGGAAAGTGAAAC; antisense, ATGCTTTAGTTCATTCTTGTG); and Smad7 (sense, TCCTGCTGTGCAAAGTGTTC; antisense, AGTAAGGAGGAGGGGGAGAC). Mouse cyclophilin A (sense, CAGACGCCACTGTCGCTTT; antisense, TGTCTTTGGAACTTTGTCTGCA) was used as an internal control for sample normalization. PCR incorporated 5 pmol forward and reverse primer and used the following cycling conditions: 96°C for 60 s, 60°C for 60 s, and 72°C for 60 s. PCR products were analyzed on 1.5% agarose gels.

    The number of cycles was chosen to stop amplification in the exponential phase. PCR product densitometry was evaluated using NIH Image Analysis software v. 1.63. Relative Smad expression was calculated as the ratio of Smad to cyclophilin A expression. For the negative controls, the reverse-transcriptase step was omitted and PCR was performed directly from the RNA.

    Results

    Smad3 is essential for TGF-1 to suppress CD4+ but not CD8+ TCR-induced proliferation

    The ability of TGF-1 to inhibit TCR-induced cell division is largely, but not completely dependent upon Smad3 (28). In this study, we demonstrate that this lack of complete nonresponsiveness of Smad3–/– T cells in the periphery can, at least in part, be explained by a differential responsiveness of CD4+ and CD8+ T cells to TGF-1-inducible Smad3 signaling. Splenocytes from Smad3–/– and WT mice were labeled with CFSE ex vivo and stimulated with plate-bound anti-CD3 + soluble anti-CD28 in the presence or absence of TGF-1 for 72 h, and viable CD4+ and CD8+ T cells were gated for analyses. CD4+ and CD8+ T cells from both WT and Smad3–/– mice divided in an asynchronous manner, and TGF-1, in a concentration-dependent manner, suppressed division of WT CD4+ and WT CD8+ T cells (Fig. 1A). Growth inhibition by TGF-1 was greatly diminished in Smad3–/– CD4+, but not Smad3–/– CD8+, T cells (Fig. 1A). Percentage of precursor frequencies from the CFSE profiles was calculated to quantify TGF-1-induced inhibition of cell cycle entry. Plotting the normalized mean frequencies of the undivided T cells demonstrated that TGF-1 inhibited the entry of both WT CD4+ and WT CD8+ T cells into their first division, and this inhibition was greatly reduced (e.g., 89%, trial 1 (circles); 95%, trial 2 (triangles) with 5 ng/ml (log = 0.7) TGF-1) in Smad3–/– CD4+ T cells (Fig. 1B). In contrast, Smad3–/– CD8+ T cells responded similarly to WT CD8+ T cells; in fact, these cells were slightly more responsive to TGF-1-mediated growth inhibition (Fig. 1B).

    FIGURE 1. Smad3 is essential for TGF-1 to suppress CD4+, but not CD8+, TCR-induced proliferation. Splenocytes from WT or Smad3–/– mice were labeled with CFSE ex vivo and stimulated with plate-bound anti-CD3 and soluble anti-CD28 in the presence of increasing concentrations of exogenous TGF-1, as shown. At 72 h after stimulation, cells were harvested, and gated on CD4+ or CD8+ T cells for analyses, after exclusion of dead cells by 7-AAD. A, CFSE profiles of stimulated CD4+- or CD8+-gated T cells. B, The percentage of undivided CD4+- or CD8+-gated T cells plotted for each concentration of TGF-1. The data are expressed as values normalized to their respective experimental group stimulated without exogenous TGF-1 (shown in D). The data were initially calculated as precursor frequencies using percentages of the total viable CD4+ or CD8+ events within individual peaks of the CFSE profiles. Two separate experiments (three WT and three Smad3–/– mice/experiment) are shown. Circles and triangles designate trials 1 and 2, respectively. The solid and dotted lines represent WT and Smad3–/– T cells, respectively. C, Mean divisions of the proliferating CD4+- and CD8+-gated T cells plotted for each concentration of TGF-1. Values were extrapolated from the percentage of input of each division of the CFSE profile and represent the mean number of divisions, as described in Materials and Methods. The data are expressed as values normalized to their respective experimental group stimulated without exogenous TGF-1 (shown in E).

    To quantify the effect of TGF-1 on the dividing population, mean division numbers were calculated, as described in Materials and Methods. As shown in Fig. 1C, TGF-1 impaired progressive cycling of both CD4+ and CD8+ T cells through subsequent cell divisions, and this TGF-1-mediated suppression was reduced (e.g., 93%, trial 1; 78%, trial 2 with 5 ng/ml TGF-1) in Smad3–/– CD4+ T cells. By comparison, we observed only a modest reduction by TGF-1 in Smad3–/– CD8+ T cells (e.g., 25%, trial 1; 5%, trial 2 with 5 ng/ml TGF-1) (Fig. 1C).

    Compared with WT CD8+ T cells, slightly more (i.e., 13%, trial 1; 14%, trial 2) Smad3–/– CD8+ T cells entered into the cell cycle following activation (Fig. 1D). This modest CD8+-specific effect paralleled a concomitant increase in the mean division number of the dividing population (Fig. 1E). Despite this subtle enhanced commitment to divide, TGF-1 still inhibited CD8+ T cell expansion in a Smad3-independent manner. Peripheral Smad3–/– CD4+ and CD8+ T cells had a mostly resting (CD25–CD69–), naive (CD44lowCD62Lhigh) phenotype, relatively indistinguishable from that of the WT T cells, with no evidence of TCR down-regulation (data not shown). Consistent with these observations, forward side scatter also did not differ between Smad3–/– and WT CD4+ or CD8+ T cells (data not shown). Collectively, these data indicate that the differential responsiveness between Smad3–/– CD8+ and CD4+ T cells to TGF-1 cannot be explained by a difference in their naive status.

    Blocking IL-2 signaling does not allow TGF-1 to inhibit proliferation of Smad3–/– CD4+ T cells

    Because the capacity of TGF-1 to inhibit IL-2 production is impaired in Smad3-deficient T cells (28), the possibility was examined that proliferation of Smad3–/– CD4+ T cells was not inhibited by TGF-1 because they make more IL-2 than WT T cells. However, simultaneous addition of IL-2-, IL-2R-, and IL-2R-neutralizing Abs during stimulation failed to enable TGF-1 to suppress Smad3–/– CD4+ T cell proliferation, as determined by undivided precursor frequency (Fig. 2A) and mean division number (Fig. 2B). Reduction in the mean division number in the absence of TGF-1 (p = 0.044) was used an indicator of Ab neutralization of endogenous IL-2 (Fig. 2B).

    FIGURE 2. Blockade of IL-2 signaling does not enable Smad3–/– CD4+ T cells to respond to TGF-1-induced growth inhibition. T cells obtained from lymph nodes of Smad3–/– mice were labeled with CFSE ex vivo and stimulated with plate-bound anti-CD3 and soluble anti-CD28 in the presence (; ) or absence (; ?) of neutralizing Abs (20 μg/ml) against IL-2, CD25, and CD122 with (; ?) or without (; ) exogenous TGF-1 (5 ng/ml). At 72 h after stimulation, cells were harvested and gated on CD4+ T cells for analyses after exclusion of dead cells by 7-AAD. A, The undivided precursor frequency of CD4+-gated T cells was calculated using percentages of the total viable CD4+ events within individual peaks of the CFSE profiles. B, The mean division number of CD4+-gated T cells. Values were extrapolated from the percentage of input of each division of the CFSE profile and represent the mean number of divisions, as described in Materials and Methods. Three separate experiments are shown. Statistical significance was determined using Student’s t test.

    The differential inhibitory effect of TGF-1 on Smad3–/– CD4+ and CD8+ T cells is not the result of altered Smad expression

    To test the possibility that Smad3-independent effects were mediated through compensatory Smad-dependent pathways, e.g., overexpression of the co-Smad, Smad2, or reduced expression of the anti-Smad, Smad7, we examined Smad steady state mRNA and protein expression in resting as well as activated CD4+ and CD8+ T cells. As illustrated in Fig. 3A, steady state Smad2, -3, -4, or -7 mRNA expression was similar between CD4+ and CD8+ lymph node T cells, either before or after TCR-induced stimulation with or without exogenous TGF-1. The CD4+ to CD8+ ratio for Smad mRNA expression is summarized in Fig. 3B. Total Smad2 and Smad3 protein expression was also similar between CD4+ and CD8+ T cells (Fig. 3C). These data collectively suggest that the differential responsiveness between Smad3–/– CD8+ and CD4+ T cells to TGF-1 cannot be explained by a difference in their Smad expression levels.

    FIGURE 3. Smad expression is similar between CD4+ and CD8+ T cells. CD4+ or CD8+ lymph node T cells were purified by negative selection from WT mice using R&D Systems T cell enrichment kits according to the manufacturer’s instructions. A, Cells were stimulated in the presence (5 ng/ml) or absence of TGF-1 for 2, 4, or 8 h before isolation of total RNA. Cyclophilin A (CyPA) was used as a normalization control. The cyclophilin A-normalized Smad mRNA expression values were calculated, as indicated. B, The CD4+:CD8+ ratios for Smad-2, -3, -4, and -7 steady state mRNA expression were calculated using the cyclophilin A-normalized values. C, Analyses of Smad protein expression. CD4+ or CD8+ splenic or lymph node T cells were purified by negative selection from C57BL/SV129 mice. Whole cell lysates were isolated for Western immunoblotting. -Actin-normalized Smad expression values were calculated as indicated.

    Smad3 functions as a T cell intrinsic factor to regulate CD4+ T cell division

    Soluble factors secreted into the microenvironment and cell-cell contact signaling are two factors that markedly impact T cell fate. To determine whether the observed TGF-1-induced antiproliferative effect was T cell intrinsic or due to microenvironment perturbation, purified CD4+ and CD8+ T cells from lymph nodes and spleens of WT and Smad3–/– mice were stimulated for 72 h with plate-bound anti-CD3 + soluble anti-CD28 in the presence or absence of TGF-1. Viable CD4+ and CD8+ T cells were gated for FACS analyses. Percentage of precursor frequencies was determined from the CFSE profiles to quantify the inhibitory effect of TGF-1 on cell division (Fig. 4A).

    FIGURE 4. Smad3 functions as a T cell intrinsic factor to regulate CD4+ T cell division. Spleen and lymph nodes were obtained from WT or Smad3–/– mice. CD4+ and CD8+ T cells were purified by negative selection and FACS sorting as described in Materials and Methods, labeled with CFSE, and stimulated with 3 μg/ml plate-bound anti-CD3 and soluble anti-CD28 in the presence or absence of 2.5 ng/ml TGF-1. At 72 h after stimulation, cells were harvested and gated on 7-AAD– CD4+ or CD8+ T cells for analyses. A, CFSE profiles of purified CD4+ or CD8+ T cells from WT and Smad3–/– mice. B and C, Percentages of undivided CD4+ and CD8+ T cells from WT (?; ) and Smad3–/– (; ) mice observed in the presence or absence of TGF-1, as shown in A. The values were calculated using precursor frequencies and are reported as a percentage of the total viable CD4+ or CD8+ T cells recovered. Data are reported as values normalized to the percentage of undivided cells that were stimulated in the absence of 2.5 ng/ml exogenous TGF-1 (shown in parentheses). Two separate experiments (three WT and three Smad3–/– mice/experiment) are shown. Circles and triangles designate trials 1 and 2, respectively. The solid and dotted lines represent WT and Smad3–/– T cells, respectively. D and E, Mean divisions of the proliferating CD4+- and CD8+-gated T cells from WT (?; ) and Smad3–/– (; ) mice observed in the presence and absence of TGF-1, as shown in A. Values were extrapolated from the percentage of input of cells from each division and represent the mean number of divisions, as described in Materials and Methods. The data are expressed as values normalized to their respective experimental group stimulated without exogenous TGF-1 (shown in parentheses). Two separate experiments (three WT and three Smad3–/– mice/experiment) are shown. Circles and triangles designate trials 1 and 2, respectively.

    TGF-1 increased the percentage of purified WT CD4+ and WT CD8+ T cells that failed to undergo a first division (Fig. 4, B and C), and this effect on CD4+ T cells was either abrogated (trial 1) or greatly reduced (2.1-fold, trial 2) in Smad3–/– T cells (Fig. 4B). In contrast, Smad3–/– CD8+ T cells were only marginally affected (Fig. 4C). Conversely, plotting the mean division number of the divided cells illustrated that TGF-1 suppressed the continuous progression of purified WT CD4+ and WT CD8+ T cells through multiple rounds of division (Fig. 4, D and E). This inhibition was greatly impaired in Smad3–/– CD4+ T cells (93%, trial 1; 58%, trial 2) (Fig. 4D); however, Smad3–/– CD8+ T cells responded similarly to WT CD8+ T cells (Fig. 4E). Overall, these results confirm the conclusions reached with unsorted cell populations and demonstrate that TGF-1 can inhibit TCR-induced CD4+ and CD8+ T cell division in a manner that is independent of signals from non-T cells.

    TGF-1 suppresses TCR-induced proliferation of CD44lowCD4+ T cells and enhances the survival of CD44highCD4+ T cells

    Under certain conditions, TGF-1 augments rather than suppresses T cell expansion (38, 39). Although the mechanisms underlying this bimodal effect of TGF-1 are not completely understood, Sung et al. (40) have recently shown that TGF-1 augments survival and in vitro proliferation of naive CD4+ T cells that are activated under strong anti-CD3 stimulation conditions in the presence of anti-CD28 costimulation. To determine whether TGF-1 also enhanced the survival of naive Smad3–/– CD4+ T cells, we purified CD44lowCD4+ from CD44highCD4+ T cells from lymph nodes and spleens of WT and Smad3–/– mice and assessed survival as well as cell division using 7-AAD/annexin V and CFSE staining, respectively (Fig. 5). Exogenous TGF-1 did not affect (p = 0.09) the survival of WT CD44lowCD4+ T cells (Fig. 5A), despite a 30% increase (p < 0.001) in undivided precursor frequency (Fig. 5B) and a 2-fold reduction (p < 0.001) in mean division number (Fig. 5C). Survival of Smad3–/– CD44lowCD4+ T cells was also not significantly affected by TGF-1 (p = 0.17) (Fig. 5A); however, the antiproliferative effects were abrogated similar to the results observed with the unfractionated Smad3–/– CD4+ T cells (Figs. 1 and 4). In striking contrast, TGF-1 augmented the survival of both Smad3–/– CD44high and WT CD44highCD4+ T cells (Fig. 5D), but interestingly, did not inhibit proliferation of either population (Fig. 5, E and F). These results demonstrate that the major effect of TGF-1 is to inhibit the proliferation of naive (CD44low) CD4+ T cells, and it does so in a Smad3-dependent manner, without significantly affecting survival. The prosurvival effect of TGF-1 is predominantly on the previously activated, i.e., CD44high, CD4+ T cells, and this effect is largely Smad3 independent.

    FIGURE 5. TGF-1 suppresses TCR-induced proliferation of CD44lowCD4+ T cells and enhances the survival of CD44highCD4+ T cells. Splenocytes were obtained from WT or Smad3–/– mice. CD44lowCD4+ and CD44highCD4+ T cells were purified by FACS sorting, labeled with CFSE ex vivo, and stimulated with plate-bound anti-CD3 and soluble anti-CD28 in the presence (; ?) or absence (; ) of exogenous TGF-1, as shown. At 72 h after stimulation, cells were harvested and gated on T cells for analyses, after exclusion of dead cells by 7-AAD. A and D, The percentage of apoptotic 7-AAD+-gated CD44low and CD44high T cells, respectively. B and E, The undivided precursor frequency (UPF) 7-AAD–-gated CD44low and CD44high T cells, respectively. C and F, The mean division number (MD#) of 7-AAD–-gated CD44low and CD44high T cells, respectively. The results from either two or three separate experiments are shown. Statistical significance was compared using Student’s t test.

    Smad3 deficiency compromises survival of CD8+, but not CD4+ T cells following activation ex vivo, and exogenous TGF-1 prevents this death in a Smad3-independent manner

    To more thoroughly investigate whether TGF-1 uses Smad3 signaling to rescue CD4+ and CD8+ T cells from apoptosis, we quantified the percentages of cells undergoing apoptosis and enumerated total cell recoveries over a concentration range of TGF-1. The total (live + dead) number and the percentage of dead (7-AAD+) cells were determined, as described in Materials and Methods, and illustrated in Fig. 6. Regardless of the presence or absence of exogenous TGF-1, Smad3 deficiency did not markedly alter the percentage of unfractionated apoptotic CD4+ T cells present at 72 h after T cell stimulation (Fig. 6, A and D). In the absence of TGF-1, activation of WT CD4+ and Smad3–/– CD4+ T cells resulted in similar total and viable cell recoveries (Fig. 6, B and C). Addition of exogenous TGF-1 attenuated both total and viable WT CD4+ T cell recovery by 50% (Fig. 6, E and F). In contrast, TGF-1 did not inhibit recovery of Smad3–/– CD4+ T cells (Fig. 6, E and F). This dichotomy reflects the differential effect of TGF-1 on WT and Smad3–/– CD4+ T cell expansion described in Figs. 1 and 4.

    FIGURE 6. Smad3 deficiency compromises survival of CD8+, but not CD4+, T cells following activation ex vivo, and exogenous TGF-1 prevents this death in a Smad3-independent manner. A and G, Percentage of apoptosis. The percentages of 7-AAD+ CD4+ and 7-AAD+ CD8+ WT and Smad3–/– T cells were determined by FACS analyses at 72 h after stimulation. B and H, Total (live + dead) CD4+ and CD8+ cell counts were enumerated at 72 h after stimulation, as described in Materials and Methods. C and I, Recovery of viable cells. The percentage of the viable cells recovered at 72 h after activation was determined by exclusion of 7-AAD+ CD4+ and 7-AAD+ CD8+ T cells. D and J, The percentages of 7-AAD+ CD4+ and 7-AAD+ CD8+ WT and Smad3–/– T cells were determined by FACS analyses at 72 h after stimulation and plotted for each concentration of TGF-1. The data are expressed as values normalized to their respective experimental group stimulated without exogenous TGF-1, shown in A and G. E and K, Total (live + dead) CD4+ and CD8+ cell counts were enumerated at 72 h after stimulation. Values are plotted against the concentration of TGF-1. The data are expressed as values normalized to their respective experimental group stimulated without exogenous TGF-1, shown in B and H. F and L, The percentage of the viable cells recovered at 72 h after activation was determined. Values are plotted for each concentration of TGF-1. The data are expressed as values normalized to their respective experimental group stimulated without exogenous TGF-1 shown in C and I.

    A different situation was observed for CD8+ T cells. Although the percentage of apoptotic WT CD8+ T cells was similar to that observed for WT CD4+ T cells, both in the presence and absence of exogenous TGF-1, Smad3–/– CD8+ T cells were markedly (i.e., 2.3- to 2.4-fold) more susceptible to death following activation (compare Fig. 6, A and G). Addition of exogenous TGF-1 reduced Smad3–/– CD8+ T cell death in a concentration-dependent manner (Fig. 6J), as indicated by a reduction in the percentage of apoptotic CD8+ T cells from 54 to 33% (trial 1) or 60 to 36% (trial 2) at the highest concentration of TGF-1 tested. In comparison, 25% of the activated WT CD8+ T cells were apoptotic, and this did not change appreciably with the addition of exogenous TGF-1 (Fig. 6, G and J). Notably, the total (live + dead) recoveries of Smad3–/– CD8+ and WT CD8+ T cells were also similar, i.e., 96% (trial 1) and 82% (trial 2) (Fig. 6H), indicating that the majority of the dying cells were still detectable. Consistent with this, the total (78–100%, WT; 74–108%, Smad3–/–) and annexin V– 7-AAD– (71–78%, WT; 74–83%, Smad3–/–) CD8+ T cell recoveries at 24 h after stimulation were indistinguishable between Smad3–/– and WT CD8+ T cells, indicating that the loss of the Smad3–/– CD8+ T cells was not an early event. Finally, addition of TGF-1 over a log concentration range produced a comparable inhibition of total CD8+ T cell recovery for Smad3–/– and WT CD8+ T cells (Fig. 6K). Collectively, these data indicate that TGF-1 blocks CD8+ T cell proliferation and apoptosis in a Smad3-independent manner (Fig. 1). This combined effect of TGF-1 on Smad3–/– CD8+ T cells is evident from a reduction in viable cell recovery (Fig. 6, I and L). The antiapoptotic effects of TGF-1 slightly augmented the total number of viable Smad3–/– CD8+ T cells recovered compared with that of WT CD8+ T cells (Fig. 6L). The subtleness of this difference, however, suggests that TGF-1 principally impairs CD8+ T cell proliferation (Fig. 1).

    Neutralization of TNF- signaling does not ameliorate Smad3–/– CD8+ T cell death

    The delayed onset of death in the Smad3–/– CD8+ T cells prompted us to investigate whether this phenomenon was mediated through TNF--dependent killing. TNF- production was up-regulated in both WT and Smad3–/– splenocyte cultures at 60 h post-T cell activation (Fig. 7A), a time point that coincided with the delayed onset of Smad3–/– CD8+ T cell death (kinetic data not shown). The amount of TNF- produced was comparable at 72 h (p = 0.12) between WT (723 ± 23 pg/ml) and Smad3–/– cells (812 ± 39 pg/ml) (Fig. 7A). TGF-1 inhibited TNF- production in both WT and Smad3–/– cultures; however, the magnitude of the suppression was greater for the WT cells (64.6%; p < 0.0001) than for Smad3–/– cells (38.7%; p = 0.003) (Fig. 7A). These results suggest that both Smad3-dependent and -independent mechanisms mediate this inhibition. In addition, neither the rate (Fig. 7B) nor the magnitude (Fig. 7C) of TNFRII surface expression differed between Smad3–/– and WT CD8+ splenic T cells following anti-CD3 + anti-CD28 stimulation. TGF-1 also had little or no inhibitory affect on either Smad3–/– or WT CD8+ T cell TNFRII expression (percentage, rate, or magnitude) (Fig. 7, B and C). Similar results were obtained with lymph node CD8+ T cells (data not shown). Most importantly, eradication of TNF signaling using neutralizing mAbs against TNF-, TNFRI, and TNFRII did not impact either TCR-mediated Smad3–/– CD8+ T cell death or the ability of TGF-1 to rescue this death in either splenic (Fig. 7D) or lymph node (data not shown) CD8+ T cells. Successful TNF blockade with the neutralizing Abs under these conditions was assessed as a reduction in IL-6 production (p = 0.003) (Fig. 7E) and impaired survival of the TNF--sensitive WEHI-13VAR and L929 mouse fibroblast cell lines (data not shown). Collectively, these results show that TNF- is not the cytokine responsible for the increase in cell death seen in activated CD8+ Smad3–/– T cells.

    FIGURE 7. Neutralization of TNF signaling does not ameliorate Smad3–/– CD8+ T cell death. A, Splenic T cell suspensions from WT (; ?) or Smad3–/– (; ) mice were activated ex vivo (3 μg/ml plate-bound anti-CD3 + soluble anti-CD28) in the presence (?; ) (5 ng/ml) or absence (; ) of exogenous TGF-1. Cell culture supernatants were collected at 0, 24, 60, and 72 h, and TNF- production was determined by ELISA. Data are reported as the mean ± SEM for triplicate determinations. B, The percentages of 7-AAD– CD8+-gated TNFRIIhigh WT and Smad3–/– splenic T cells were determined by FACS analyses. C, The MFI of the TNFRIIhigh 7-AAD– CD8+-gated WT and Smad3–/– T cells were determined by FACS analyses. D, Smad3–/– splenic T cells were activated ex vivo (3 μg/ml plate-bound anti-CD3 + soluble anti-CD28) in the presence (; ) or absence (?; ) of TNF-blocking Abs with or without exogenous TGF-1. The percentage of 7-AAD– annexin V– CD8+ WT and Smad3–/– lymph node T cells was determined by FACS analyses at 72 h postactivation. E, Supernatants were collected from Smad3–/– splenocytes following activation (3 μg/ml plate-bound anti-CD3 + soluble anti-CD28) for 48 h in the presence (; ) or absence (?; ) of TNF-blocking Abs with (; ) or without (; ?) exogenous TGF-1. IL-6 production was determined by ELISA.

    TGF-1 inhibits the percentage of total CD4+ and CD8+ IL-2 producers in a Smad3-dependent manner

    We have previously reported that Smad3 is essential for TGF-1 to suppress IL-2 production in anti-CD3 + anti-CD28-stimulated splenocytes. In this study, we expand upon this initial observation to investigate the role of Smad3 in regulating IL-2 production in CD4+ and CD8+ T cells. Splenocytes from Smad3–/– and WT mice were stimulated with plate-bound anti-CD3 + soluble anti-CD28 in the presence or absence of TGF-1 for 24 h and processed for IL-2 secretion, as described in Materials and Methods. Viable CD4+ and CD8+ T cells were gated for analyses. IL-2 secretion was not detected in unstimulated WT or Smad3–/– T cells (Fig. 8A). Plotting IL-2 production vs forward scatter demonstrated that, for the most part, IL-2 secretion occurred before T cell blastogenesis (Fig. 8A). In the absence of TGF-1, increasing the concentration of anti-CD3 from 3 to 10 μg/ml increased the percentage of WT CD4+ IL-2 secretors from 21 to 29% (1.4-fold, trial 1) or from 31 to 39% (1.2-fold, trial 2) and the percentage of WT CD8+ IL-2 secretors from 6 to 10% (1.7-fold, trial 1) and from 4.8 to 7.7% (1.6-fold, trial 2) (Fig. 8B), but had little effect on the amount of IL-2 produced per cell (Fig. 8, C and E). Smad3 deficiency increased anti-CD3 + anti-CD28-induced IL-2 mean fluorescence intensity (MFI) by 1.5- to 2.0-fold (Fig. 8, C and E), and sometimes decreased the percentage of IL-2 producers (Fig. 8, B and D). Addition of TGF-1 consistently reduced the percentage of both WT CD4+ and WT CD8+ IL-2 producers (Fig. 8, B and D), but not IL-2 MFI (Fig. 8, C and E). Most importantly, this inhibition was abrogated in both Smad3–/– CD8+ and Smad3–/– CD4+ T cells. These results suggest that the mechanism by which TGF-1 inhibits CD8+ T cell division can be dissociated from its effect on IL-2 production.

    FIGURE 8. TGF-1 inhibits the percentage of total CD4+ and CD8+ IL-2 producers in a Smad3-dependent manner. Splenocytes from WT or Smad3–/– mice were stimulated with plate-bound anti-CD3 (either 3 or 10 μg/ml, as indicated) and soluble anti-CD28 in the presence or absence of 2.5 ng/ml exogenous TGF-1 for 24 h. A, IL-2 production was determined by an IL-2 secretion assay, as described in Materials and Methods. CD4+- or CD8+-gated cells, after exclusion of dead cells by 7-AAD, were used for analyses. Data are plotted as forward scatter vs IL-2. B and D, The percentages of CD4+ and CD8+ T cells that secrete IL-2 were determined in two separate experiments and shown as trial 1 (B) and trial 2 (D). The percentage of IL-2 producers was determined from the gate shown in A, which represents trial 1. C and E, The IL-2 MFI of the IL-2+ cells from B and D are shown in C and E.

    Smad3 functions as a T cell intrinsic factor to mediate TGF-1-induced inhibition of IL-2 production

    To determine whether Smad3 in the target T cells themselves mediates the TGF-1-induced inhibition of IL-2 production, we purified CD4+ and CD8+ T cells from lymph nodes and spleens of WT and Smad3–/– mice, stimulated them for 24 h with plate-bound anti-CD3 + soluble anti-CD28 in the presence or absence of TGF-1, and quantified IL-2 secretion. Viable CD4+ or CD8+ T cells were gated for FACS analyses (Fig. 9A). The percentage of CD4+ IL-2 producers was comparable between total splenocyte and purified CD4+ T cell preparations (compare Fig. 9, B and D with Fig. 8, B and D). Fewer IL-2 producers (Fig. 9, B and D) with decreased MFI (Fig. 9, C and E) were observed in the purified WT and Smad3–/– CD8+ T cells. This was mirrored by low yields of IL-2 from purified CD8+ T cells as quantified by ELISA (Fig. 10). Smad3–/– CD4+ and CD8+ T cells also generated slightly fewer IL-2-producing cells (Fig. 9, B and D), but the MFI (Fig. 9E) was unaffected. The reduction in the percentage of CD8+ IL-2 producers may be a consequence of insufficient CD4+ T cell help (41). Regardless of the magnitude of IL-2 produced in response to anti-CD3 + anti-CD28 stimulation, TGF-1 suppressed IL-2 in purified WT CD4+ and WT CD8+ T cells, and this inhibition was abrogated in both Smad3–/– CD4+ and Smad3–/– CD8+ T cells. Thus, TGF-1 suppresses IL-2 production in a Smad3-dependent, T cell autonomous manner in both cell types.

    FIGURE 9. Smad3 functions as a T cell intrinsic factor to mediate TGF-1-induced inhibition of IL-2 production. Spleens and lymph nodes were obtained from WT or Smad3–/– mice. CD4+ and CD8+ T cells were purified by negative selection with FACS sorting and stimulated with plate-bound anti-CD3 (either 3 or 10 μg/ml, as indicated) and soluble anti-CD28 in the presence or absence of exogenous TGF-1 (2.5 ng/ml) for 24 h. A, IL-2 production was determined by an IL-2 secretion assay. CD4+ or CD8+ cells, after exclusion of dead cells by 7-AAD, were used for analyses. Data are plotted as CD4 or CD8 expression vs IL-2. B and D, The percentages of CD4+ and CD8+ T cells that secrete IL-2 were determined in two separate experiments and are shown as trial 1 (B) and trial 2 (D). The percentages of IL-2 producers were determined from the gate (upper right quadrant) shown in A, which represents trial 1. C and E, The IL-2 MFI of the IL-2+ cells from B and D are shown in C and E.

    FIGURE 10. IL-2 secretion as quantified by ELISA. At 48 h after anti-CD3 (3 or 10 μg/ml) + anti-CD28 activation in the presence (2.5 ng/ml) or absence of exogenous TGF-1, supernatants were collected from WT or Smad3–/– cultures, as shown in Fig. 9, B and D. Data are reported as the mean ± SEM for triplicate determinations. One of two representative experiments is shown.

    Discussion

    Despite the underlying importance of TGF-1 for maintaining immunological balance, the precise mechanisms by which it inhibits T cell expansion remain poorly understood. Elucidation of these pathways will provide important new understandings for the molecular basis of T cell homeostasis. Although Smad3 has emerged as an important regulator of cell proliferation, we have recently demonstrated that Smad3 signaling is not essential for all TGF-1-induced T cell growth arrest (28). In this study, we expand upon these initial findings to describe the diversification of TGF-1 effects on the expansion of CD4+ and CD8+ T cells. We show that TGF-1 uses different intracellular signaling pathways to discriminately regulate CD4+ and CD8+ T cells. Specifically, we demonstrate that TGF-1 inhibits TCR-induced naive (i.e., CD44low) CD4+ T cell proliferation almost exclusively through Smad3 signaling. In contrast, inhibition of TCR-induced CD8+ T cell proliferation by TGF-1 is largely Smad3 independent. We also assessed the role of Smad3 on CD4+ and CD8+ T cell survival, and show that in the absence of Smad3, CD8+ but not CD4+ T cells are extremely sensitive to TCR-induced apoptosis. Interestingly, exogenous TGF-1 protected against this apoptotic death, and did so in a Smad3-independent manner. TGF-1 also enhanced the survival of CD44highCD4+ T cells in a Smad3-independent manner. In contrast, IL-2 production by both cell types was inhibited by TGF-1 in a Smad3-dependent manner. These findings may have significant clinical implications for targeting cell type-specific inflammatory and autoimmune disorders.

    Our study is the first to use CFSE profiling to directly assess cell division separately for both CD4+ and CD8+ T cells from Smad3–/– mice. These data illustrate that TGF-1 has the capacity to inhibit TCR-induced T cell proliferation through at least two distinct molecular mechanisms. This finding is consistent with earlier observations showing that anti-Ig- and LPS-stimulated B cells as well as IL-2-stimulated T cells do not require Smad3 for TGF-1-induced growth inhibition (28, 32, 33). Also consistent with our observed cell type-specific, Smad3-mediated responses, CD4+ and CD8+ T cells differentially respond to several other signaling pathways, including STAT4 (42), p38 (43), and JNK (44). The regulatory effects of TGF-1 are quite complex, and although not completely understood, it has been proposed that the overall responsiveness to TGF-1 is mediated through cooperative cross talk among numerous intracellular signaling pathways, including STAT and MAPK (11, 45). Additional studies will deduce whether STAT and MAPK signaling contribute to the differential responsiveness of CD4+ and CD8+ T cells to TGF-1-induced growth arrest.

    Our observation that purified CD4+ and CD8+ T cells respond to TGF-1-induced growth arrest in a manner similar to unsorted T cells establishes that Smad3 within CD4+ T cells is the key mediator of their intrinsic responsiveness to TGF-1. These data eliminate the possibility that the differential responsiveness of CD4+ and CD8+ T cells is mediated indirectly through APCs. Furthermore, we have verified that CD8+ T cells can signal through a TGF-1-inducible, Smad3-dependent pathway, as demonstrated by inhibition of IL-2 production in purified T cells. Studies to determine whether the Smad3 independence extends to Ag-specific CTL killing are ongoing. Thus, the Smad3-independent pathway is selectively used for only some biological functions, e.g., preventing TCR-induced proliferation of CD8+ T cells (Figs. 1 and 4), while the Smad3-dependent pathway is selectively used for others, e.g., suppression of IL-2 production (Figs. 8–10).

    In addition to its effects on proliferation and IL-2 production, TGF-1 regulates T cell survival (Figs. 5 and 6). Examination of CD4+ and CD8+ T cell apoptosis in these studies highlighted a role for Smad3 in CD8+ T cell survival. We found that TCR-activated Smad3–/– CD8+ T cells were very sensitive to apoptosis (Fig. 6). These observations are consistent with the selective increase in CD8+ T cell apoptosis seen in the TGF-1 knockout mice (18). TCR-induced apoptosis can result through either Fas (Apo1/CD95)/Fas ligand (FasL) or TNF-dependent pathways. Fas is important for the apoptosis of most CD4+ T cells; however, its role in the depletion of CD8+ T cells seems to be less significant, as evidenced by normal deletion of TCR transgenic CD8+ T cells in Fas-deficient lpr mice (46, 47, 48). Blockage of TNF receptor has been shown to suppress TCR-induced death of CD8+ T cells (49). A delay in the onset of TNF-mediated death relative to FasL-mediated death has also been shown (49). Moreover, TGF-1 protects CD4+ and CD8+ T cells from FasL-mediated apoptosis, but selectively protects only CD8+ T cells from TNF receptor-induced death (18). Our observation that the total (live + dead) cell recoveries of Smad3–/– CD8+ and WT CD8+ T cells were similar 3 days after TCR-induced stimulation, despite the marked difference in the percentage of apoptotic cells, was consistent with a delayed TNF-dependent pathway. A possible TNF-mediated death effect was further supported by the observation that total cell recovery and percentage of apoptosis were comparable between Smad3–/– CD8+ and WT CD8+ T cells through the first 48 h following activation (data not shown). Thus, it was surprising that neutralization of TNF signaling did not block Smad3–/– CD8+ T cell death in our system. Nonetheless, alternative mechanisms are also conceivable. For example, Smad3–/– CD8+ T cells may be more susceptible to IL-2-induced degradation of cellular FLIP (50). Furthermore, the slow onset of the death phenomenon is consistent with cytokine deprivation. The best-characterized CD8+ T cell trophic factors are IL-7 and IL-15, whose receptors share the common -chain (51). Lymphoid cells do not produce these cytokines, so their availability will be constrained in vitro. Moreover, competition between IL-2 and IL-7 or IL-15 for common -chain binding may also contribute to impaired IL-7- and IL-15-mediated survival. In addition, IL-2 has also been shown to down-regulate IL-7R (52). IL-12 signaling also has been implicated for CD8+ T cell survival ex vivo (53). Regardless of the mechanism of late CD8+ T cell apoptosis in Smad3–/– mice, the prevention by TGF-1 is another clear example of a Smad3-independent signaling pathway in these cells. This same signaling pathway may also be used by preactivated CD4+ T cells to prevent apoptosis.

    We have yet to define exactly why TGF-1 uses a different intracellular signaling pathway to inhibit proliferation of TCR-stimulated CD4+ T cells than it uses to inhibit proliferation of TCR-induced CD8+ T cells, IL-2-stimulated CD4+ T cells, and anti-Ig- or LPS-stimulated B cells. Our data, nonetheless, clearly suggest a significant diversification in the response of different lymphocytes to TGF-1. Conceptually, the simplest explanation for the relevance of this observation would be to provide diversified negative feedback regulation to modulate adaptive immune responses in a cell type-specific manner. For example, TGF-1 may function to block or delay Ag-specific CD4+ T cell expansion immediately following Ag presentation; however, once Th cells have become primed to secrete proinflammatory cytokines, they may resist direct TGF-1-induced growth inhibition due to the up-regulation of the antagonistic Smad7. TGF-1 also enhances preactivated Th cell survival, and this occurs in a Smad3-independent manner (Fig. 5). These two effects would thus augment the amount of available CD4+ help. However, if CD8+ T cells need to maximize effector function for pathogen clearance at the expense of cell division, then they might circumvent the Smad7-negative feedback mechanism via a Smad3-independent pathway to facilitate TGF-1-induced antiproliferative and antiapoptotic signaling. Additional information concerning the intracellular mechanisms whereby TGF-1 regulates lymphocyte expansion should provide further insight not only into normal T cell homeostasis, but also into potential dysregulation that leads to inflammatory disease or autoimmunity.

    Acknowledgments

    We are grateful to the many individuals who contributed to these studies. We thank Dr. Norbert Kaminski for valuable discussions and comments; Drs. Sharon Wahl, Anita Roberts, and John Letterio for their gracious donation of the mice; Kevin Holmes, Tom Moyer, and Calvin Eigsti for support with the flow cytometry sorting experiments; Mario Anzono, Nancy McCartney-Francis, and George McGrady for technical assistance; and Drs. Ronald Gress and Wanjun Chen for their critical review of the manuscript.

    Footnotes

    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.

    1 Address correspondence and reprint requests to Dr. Susan C. McKarns, Laboratory of Cellular and Molecular Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Building 4, Room 111, MSC-0420, 4 Center Drive, Bethesda, MD 20892. E-mail address: smckarns@niaid.nih.gov

    2 Abbreviations used in this paper: TRII, TGF-1 receptor type II; 7-AAD, 7-aminoactinomycin D; DN, dominant negative; FasL, Fas ligand; MFI, mean fluorescence intensity; wt, wild type.

    Received for publication June 8, 2004. Accepted for publication December 10, 2004.

    References

    Kulkarni, A. B., C. G. Huh, D. Becker, A. Geiser, M. Lyght, K. C. Flanders, A. B. Roberts, M. B. Sporn, J. M. Ward, S. Karlsson. 1993. Transforming growth factor 1 null mutation in mice causes excessive inflammatory response and early death. Proc. Natl. Acad. Sci. USA 90:770.

    Larsson, J., M. J. Goumans, L. J. Sjostrand, M. A. van Rooijen, D. Ward, P. Leveen, X. Xu, P. ten Dijke, C. L. Mummery, S. Karlsson. 2001. Abnormal angiogenesis but intact hematopoietic potential in TGF- type I receptor-deficient mice. EMBO J. 20:1663.

    Gorelik, L., R. A. Flavell. 2000. Abrogation of TGF signaling in T cells leads to spontaneous T cell differentiation and autoimmune disease. Immunity 12:171.

    Lucas, P. J., S. J. Kim, S. J. Melby, R. E. Gress. 2000. Disruption of T cell homeostasis in mice expressing a T cell-specific dominant negative transforming growth factor II receptor. J. Exp. Med. 191:1187.

    Gorelik, L., R. A. Flavell. 2002. Transforming growth factor- in T-cell biology. Nat. Rev. Immunol. 2:46.

    Wahl, S. M., J. Swisher, N. McCartney-Francis, W. Chen. 2004. TGF-: the perpetrator of immune suppression by regulatory T cells and suicidal T cells. J. Leukocyte. Biol.

    Diebold, R. J., M. J. Eis, M. Yin, I. Ormsby, G. P. Boivin, B. J. Darrow, J. E. Saffitz, T. Doetschman. 1995. Early-onset multifocal inflammation in the transforming growth factor 1-null mouse is lymphocyte mediated. Proc. Natl. Acad. Sci. USA 92:12215.

    Schultz Jel, J., S. A. Witt, B. J. Glascock, M. L. Nieman, P. J. Reiser, S. L. Nix, T. R. Kimball, T. Doetschman. 2002. TGF-1 mediates the hypertrophic cardiomyocyte growth induced by angiotensin II. J. Clin. Invest. 109:787.

    Engle, S. J., J. B. Hoying, G. P. Boivin, I. Ormsby, P. S. Gartside, T. Doetschman. 1999. Transforming growth factor 1 suppresses nonmetastatic colon cancer at an early stage of tumorigenesis. Cancer Res. 59:3379.

    Kobayashi, S., K. Yoshida, J. M. Ward, J. J. Letterio, G. Longenecker, L. Yaswen, B. Mittleman, E. Mozes, A. B. Roberts, S. Karlsson, A. B. Kulkarni. 1999. 2-microglobulin-deficient background ameliorates lethal phenotype of the TGF-1 null mouse. J. Immunol. 163:4013.

    Roberts, A. B.. 2002. The ever-increasing complexity of TGF- signaling. Cytokine Growth Factor Rev. 13:3

    Derynck, R., Y. E. Zhang. 2003. Smad-dependent and Smad-independent pathways in TGF- family signalling. Nature 425:577.

    Byfield, S. D., A. B. Roberts. 2004. Lateral signaling enhances TGF- response complexity. Trends Cell Biol. 14:107.

    Zhang, H. M., J. N. Rao, X. Guo, L. Liu, T. Zou, D. J. Turner, J. Y. Wang. 2004. Akt kinase activation blocks apoptosis in intestinal epithelial cells by inhibiting caspase-3 after polyamine depletion. J. Biol. Chem. 279:22539.

    Yu, J. J., C. S. Tripp, J. H. Russell. 2003. Regulation and phenotype of an innate Th1 cell: role of cytokines and the p38 kinase pathway. J. Immunol. 171:6112.

    Bzowska, M., K. Guzik, K. Barczyk, M. Ernst, H. D. Flad, J. Pryjma. 2002. Increased IL-10 production during spontaneous apoptosis of monocytes. Eur. J. Immunol. 32:2011.

    Chen, W., M. E. Frank, W. Jin, S. M. Wahl. 2001. TGF- released by apoptotic T cells contributes to an immunosuppressive milieu. Immunity 14:715.

    Chen, W., W. Jin, H. Tian, P. Sicurello, M. Frank, J. M. Orenstein, S. M. Wahl. 2001. Requirement for transforming growth factor 1 in controlling T cell apoptosis. J. Exp. Med. 194:439.

    Genestier, L., S. Kasibhatla, T. Brunner, D. R. Green. 1999. Transforming growth factor 1 inhibits Fas ligand expression and subsequent activation-induced cell death in T cells via down-regulation of c-Myc. J. Exp. Med. 189:231.

    Wahl, S. M., J. M. Orenstein, W. Chen. 2000. TGF- influences the life and death decisions of T lymphocytes. Cytokine Growth Factor Rev. 11:71.

    Brabletz, T., I. Pfeuffer, E. Schorr, F. Siebelt, T. Wirth, E. Serfling. 1993. Transforming growth factor and cyclosporin A inhibit the inducible activity of the interleukin-2 gene in T cells through a noncanonical octamer-binding site. Mol. Cell. Biol. 13:1155.

    Hannon, G. J., D. Beach. 1994. p15INK4B is a potential effector of TGF--induced cell cycle arrest. Nature 371:257.

    Polyak, K., J. Y. Kato, M. J. Solomon, C. J. Sherr, J. Massague, J. M. Roberts, A. Koff. 1994. p27Kip1, a cyclin-Cdk inhibitor, links transforming growth factor- and contact inhibition to cell cycle arrest. Genes Dev. 8:9.

    Landesman, Y., F. Bringold, D. D. Milne, D. W. Meek. 1997. Modifications of p53 protein and accumulation of p21 and gadd45 mRNA in TGF-1 growth inhibited cells. Cell. Signal. 9:291.

    Seoane, J., C. Pouponnot, P. Staller, M. Schader, M. Eilers, J. Massague. 2001. TGF influences Myc, Miz-1 and Smad to control the CDK inhibitor p15INK4b. Nat. Cell Biol. 3:400.

    Campbell, J. D. M., G. Cook, S. E. Robertson, A. Fraser, K. S. Boyd, J. A. Gracie, I. M. Franklin. 2001. Suppression of IL-2-induced T cell proliferation and phosphorylation of STAT3 and STAT5 by tumor-derived TGF- is reversed by IL-15. J. Immunol. 167:553.

    Koff, A., M. Ohtsuki, K. Polyak, J. M. Roberts, J. Massague. 1993. Negative regulation of G1 in mammalian cells: inhibition of cyclin E-dependent kinase by TGF-. Science 260:536.

    McKarns, S. C., R. H. Schwartz, N. E. Kaminski. 2004. Smad3 Is Essential for TGF-1 to suppress IL-2 production and TCR-induced proliferation, but not IL-2-induced proliferation. J. Immunol. 172:4275.

    Green, E. A., L. Gorelik, C. M. McGregor, E. H. Tran, R. A. Flavell. 2003. CD4+CD25+ T regulatory cells control anti-islet CD8+ T cells through TGF--TGF- receptor interactions in type 1 diabetes. Proc. Natl. Acad. Sci. USA 100:10878.

    Belghith, M., J. A. Bluestone, S. Barriot, J. Megret, J. F. Bach, L. Chatenoud. 2003. TGF--dependent mechanisms mediate restoration of self-tolerance induced by antibodies to CD3 in overt autoimmune diabetes. Nat. Med. 9:1202.(Susan C. McKarns1 and Ron)