The transcriptional events that control T cell tolerance to peripheral self Ags are still unknown. In this study, we analyzed the regulation of AP-1- and NF-κB-mediated transcription during in vivo induction of tolerance to a self Ag expressed exclusively on hepatocytes. Naive CD8+Désiré (Des)+ T cells isolated from the Des TCR-transgenic mice that are specific for the H-2Kb class I Ag were transferred into Alb-Kb-transgenic mice that express the H-2Kb Ag on hepatocytes only. Tolerance develops in these mice. We found that the self-reactive CD8+Des+ T cells were transiently activated, then became unresponsive and were further deleted. In contrast to CD8+Des+ T cells activated in vivo with APCs, which express high AP-1 and high NF-κB transcriptional activity, the unresponsive CD8+Des+ T cells expressed no AP-1 and only weak NF-κB transcriptional activity. The differences in NF-κB transcriptional activity correlated with the generation of distinct NF-κB complexes. Indeed, in vivo primed T cells predominantly express p50/p50 and p65/p50 dimers, whereas these p50-containing complexes are barely detectable in tolerant T cells that express p65- and c-Rel-containing complexes. These observations suggest that fine regulation of NF-κB complex formation may determine T cell fate.

Studies over the past few years have indicated a wide variety of mechanisms that lead to peripheral T cell tolerance. Induction of T cell anergy, deletion of the autoreactive T cells usually preceded by a transient phase of activation, down-modulation of the TCR or coreceptor, and immune deviation have been described in several transgenic mice models (1, 2, 3, 4, 5, 6, 7). However, the parameters that determine the mechanisms of peripheral T cell tolerance are still fully unknown.

T cell activation requires two signals, the antigenic signal and a second signal resulting from the interaction of the costimulatory molecules B7.1 and B7.2, expressed by activated APCs, and the CD28 receptor expressed by the T cell (reviewed in Ref. (8)). TCR ligation in the absence of a costimulatory signal is thought to induce T cell unresponsiveness, a mechanism that contributes to peripheral T cell tolerance (9). The signaling cascades triggered by coligation of the TCR and CD28 receptors lead to the activation of several transcription factors including AP-1, NF-AT, and NF-κB, which are thought to be important in the different steps of T cell activation and may participate in T cell survival.

The AP-1 transcription factor is a complex between members of the Fos (cFos, FosB, Fra-1, and Fra-2) and Jun (c-Jun, JunB, and JunD) families of proteins (10). AP-1 transcriptional activity is regulated at the level of both fos and jun gene transcription and by post-transcriptional modifications of the corresponding protein, induced at least in part by CD28-dependent signaling (11, 12). The AP-1 transcription factor regulates the expression of several cytokine genes and may participate in the control of cell cycle (10, 13).

The NF-κB family of transcription factors is composed of NF-κB1 (p50/p105), NF-κB2 (p52/p100), RelA (p65), c-Rel, and RelB, which upon homo- or heterodimerization bind to κB motifs (reviewed in Ref. (14)). NF-κB dimers are retained in the cytoplasm by the inhibitory protein of κB (IκB),3 which, upon activation, is phosphorylated and degraded, thus releasing the NF-κB complexes that can translocate to the nucleus (14). The p65 and p50 molecules and c-Rel appear to be the essential regulators of κB sites in T cells. The p65/p50 and p65/c-Rel heterodimers act as positive regulators of T cell activation and cytokine gene expression (15, 16). The p50 molecule lacks the transactivation domain present in p65 and c-Rel, and p50/p50 homodimers may therefore be negative regulators of κB sites (14). NF-κB also plays a critical role in protecting cells from TNF-α-induced apoptosis (17, 18, 19, 20, 21). Similarly, NF-κB protects mature T cells from activation-induced apoptosis. Indeed, mature CD4+ or CD8+ T cells isolated from a mouse transgenic for a trans-dominant IκBα, which completely blocks the nuclear translocation of p65/p50 and c-Rel complexes, present an increased susceptibility to activation-induced apoptosis (22, 23). NF-κB may also function as a proapoptotic factor for some cell lines and for thymocytes (24).

Studies on the molecular mechanisms of CD4+ T cell anergy suggested that perturbed activation of AP-1 and NF-κB may lead to T cell unresponsiveness (25, 26, 27). Indeed, CD4+ T cells made anergic by repeated injection of the superantigen staphylococcal enterotoxin A have reduced AP-1 and altered NF-κB complexes as compared with activated CD4+ T cells (26, 27). However, how these observations may apply to T cell tolerance to a peripheral self Ag remained to be evaluated.

We therefore established a transfer model to directly analyze AP-1 and NF-κB transcriptional activity as tolerance develops. Naive T cells isolated from the Désiré (Des) TCR-transgenic mice that are specific for the H-2Kb class I Ag were injected into Alb-Kb mice that express the H-2Kb transgene exclusively on hepatocytes (5, 6). Tolerance is established in these mice. We found that self-reactive CD8+ T cells were transiently activated, then became unresponsive and were further deleted. The unresponsive T cells expressed no AP-1 and only weak NF-κB transcriptional activity. This contrasts with T cells primed in vivo under conditions known to induce effector and memory T cells that express high AP-1 and NF-κB transcriptional activity. The differences in transcriptional activity detected in the primed and unresponsive population correlated with differences in the NF-κB complexes generated, suggesting that fine regulation of NF-κB-dependent gene expression may determine the fate of autoreactive T cells.

The Alb-Kb mice are transgenic for the class I H-2Kb molecule driven by the albumin promoter and express H-2Kb on hepatocytes (6). The Des mice are transgenic for a TCR specific for the H-2Kb molecule (5). The AP-1-luciferase (AP-1-Luc)- and NF-κB-Luciferase (NF-κB-Luc)-transgenic mice contain the firefly luciferase gene, controlled by four AP-1 sites or two NF-κB sites, respectively (11, 28). All mice were backcrossed for at least six generations on a B10.BR background. The CBK-transgenic mice express the H-2Kb molecule on all hematopoietic cells (29).

Alb-Kb-transgenic mice or control B10.BR mice were thymectomized at the age of 4–6 wk, rested for 3–4 wk, and γ-irradiated at 6 Gy. After 11–15 days, the mice were injected i.v. with 5–8 × 106 T cells purified from Des-transgenic mice. The low dose of irradiation eliminates most circulating lymphocytes while sparing bone marrow precursors. Reconstitution of the peripheral pool of lymphocytes by endogenous precursors was completed by the time of adoptive transfer. Because the mice were thymectomized, only the B cell compartment but not the T cell compartment was replenished. Indeed, the spleen of such manipulated mice contained, on average, 3% CD4+ T cells and 3% CD8+ T cells and, when adoptively transferred, 2–3% CD8+Des+ T cells (data not shown).

FITC-labeled anti-CD25, -CD44, -CD62L, and -CD69 Ab and PE-labeled anti-CD4 and -CD8 Abs were purchased from BD PharMingen (La Jolla, CA). The Des anti-clonotype and H59.101.2 anti-CD8 Ab were prepared and biotinylated at the Centre d’Immunologie de Marseille-Luminy. For FACS staining, all samples were triple stained for the clonotype, CD4 or CD8, and one activation marker. In brief, 5 × 105 cells were incubated with biotinylated Des Ab, washed two times, and incubated with Streptavin-Tricolor (Caltag, Burlingame, CA) and FITC- and PE-conjugated Ab.

Lymph node T cells used for adoptive transfer were prepared by depletion of B cells using sheep anti-mouse IgG magnetic beads (Dynal, Oslo, Norway). On average, the population recovered contained <5% non-T cells and was composed of <5% CD4+ T cells, of which 50% expressed the Des TCR, and ∼80% CD8+ T cells, of which >90% were Des+. When indicated, the cells were labeled with CFSE (Molecular Probes, Eugene, OR) as previously described (30). Purified CD8+ or Des+ T cells were prepared by MACS (Miltenyi Biotec, Bergisch Gladbach, Germany) separation using either biotinylated anti-CD8 or Des Ab, dichlorotriazinylaminofluorescein-conjugated streptavidin (Immunotech, Marseille, France), and MACS biotin-conjugated beads. Purity of the recovered population was <90%. To isolate liver-infiltrating cells, mice were perfused with PBS. Livers were pressed through nylon mesh and passed over a Ficoll cushion.

T cell-depleted APCs were prepared by Ab-mediated complement lysis using an anti-Thy1.2 Ab (JIJ) and γ-irradiated at 24 Gy. Four million responder spleen cells were stimulated with 2 × 106 APCs. Before stimulation, the representation of CD8+Des+ T cells in the different spleen populations was determined by FACS analysis. This value was used to normalize all samples and calculate the biologic activity of interest for 5 × 105 CD8+Des+ T cells.

Cells were washed twice in PBS, then lysed in lysis buffer (luciferase assay system; Promega, Madison, WI), and luciferase activity was developed using the luciferase reagent (Promega). All measurements were done in duplicate. The background measurement was always below the activity found in extracts from nonstimulated cells. Experimental values expressed as relative luminescence units (RLU) were therefore calculated by subtracting the value of unstimulated samples from each sample. To correct for the difference in CD8+Des+ representation in the different samples, the experimental value for 5 × 105 CD8+Des+ T cells was calculated as indicated above.

IL-2 production was measured by bioassay using the CTLL-2 cell line as previously described (31). IFN-γ production was measured by ELISA using the AN18 and biotinylated R46A2 Abs and alkaline phosphatase-conjugated streptavidin (Sigma, St. Louis, MO). The amount of cytokine was expressed as units compared with a standard curve obtained with recombinant cytokines and corrected for the representation of CD8+Des+ T cells in each sample as indicated above.

Total RNA was extracted using the High Pure RNA Isolation kit (Boehringer Mannheim, Indianapolis, IN) according to the manufacturer’s instructions, treated with DNase I (Boehringer Mannheim), and reverse transcribed using oligo(dT)15 and SuperScript II RT (Life Technologies, Grand Island, NY). The cDNA was amplified using specific primers and Taq DNA polymerase (Life Technologies). All samples were normalized on the basis of expression of hypoxanthine phosphoribosyl transferase (HPRT) as previously described (32). Luciferase mRNA was quantified using as forward primer, 5′-CGCGGAATACTTCGAAATGTC-3′, and reverse primer, 5′-CCTTAGGTAACCCAGTAGATCC. Samples were denatured for 1 min at 94°C, and cycling conditions were 94°C for 40 s, 58°C for 20 s, and 72°C for 40 s for 35 cycles. Samples were quantified by phosphor imager after hybridization with a luciferase-specific probe (Fujifilm Bas-1500; Fuji, Tokyo, Japan) and expressed compared with a standard curve obtained with a luciferase plasmid. The amount of luciferase mRNA per 5000 fg HPRT was then calculated. The measurements were always performed in the exponential phase of the reaction, with the standard curve obtained with the plasmid template having an equation with a correlation factor of >0.9. Finally, some samples were reanalyzed by real-time quantitative PCR and gave similar results.

Nuclear and cytoplasmic extracts were obtained as previously described (11). In brief, CD8+ or Des+ T cells were lysed in buffer A (10 mM HEPES (pH 7.6), 10 mM KCL, 0.1 mM EDTA, 0.1 mM EGTA, 0.75 mM spermidine, 0.15 mM spermine, 1 mM DTT, and 0.625% Nonidet P-40) containing protease inhibitors (0.5 mM PMSF and 2 μg/ml aprotenin, leupeptin, pepstatin A, chymostatin, and antipain). Cytoplasmic extracts were collected after centrifugation, and nuclei were lysed in Nuclear lysis buffer (20 mM HEPES (pH 7.6), 0.4 M NaCl, 1 mM EDTA, and 1 mM EGTA) containing protease inhibitors, as described above. For EMSA, the binding reaction was conducted using 2.5–3 μg of nuclear proteins and 105 cpm of 32P-end-labeled double-stranded NF-κB oligonucleotides derived from the mouse κB intronic enhancer (5′-GATCAGAGGGGACTTTCCGAG-3′), a consensus AP-1 (5′-GTCGACGTGAGCGCGC-3′), or consensus OCT1- and 2-binding oligonucleotides (5′-TTCTAGTGATTTGCATTCGACA-3′) in the presence or absence of blocking Ab as previously described (33).

Eight to ten micrograms of the same extracts were run on a 10% polyacrylamide gel, transferred to PolyScreen membrane, and revealed using ECL Plus (Amersham Life Science, Buckinghamshire, U. K.) reagents and HPRT-conjugated protein A or protein G (Amersham). The different blots were quantified by a luminescent image analyzer (LAS-1000 plus; Fuji).

Abs used for supershifts or Western blots were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and were the p65-specific Ab sc-109, the p50-specific Ab sc-1192, the c-Rel-specific Ab sc-70, the anti-Bcl-2 Ab sc-492, the anti-Bcl-xL Ab sc-634, and an actin-specific Ab.

To define the molecular mechanisms of T cell tolerance to a peripheral self Ag, we developed an adoptive transfer model that enabled us to analyze the transcriptional program expressed by T cells in the course of tolerance induction (34). The donor CD8+ T cells isolated from the Des TCR-transgenic mice are specific for the H-2Kb class I Ag and express the Des clonotype (referred to thereafter as CD8+Des+ T cells). The CD8+Des+ T cells were injected into either Alb-Kb-transgenic mice that express the H-2Kb transgene only on hepatocytes (Des→Alb-Kb) or nontransgenic control mice (Des→B10.BR). To determine the fate of autoreactive T cells, we measured the number of the CD8+Des+ T cells in the spleen of Des→B10.BR or Des→Alb-Kb mice at different time points after adoptive transfer. The representation of CD8+Des+ T cells when injected into control B10.BR mice remained fairly constant over the 50-day period of analysis (Fig. 1). On average, the CD8+Des+ T cells represented 3% (1 × 106 cells) of the total spleen population of Des→B10.BR mice. For the first 10 days following transfer, the number of CD8+Des+ T cells in the spleen of Des→Alb-Kb mice was similar to that found in control Des→B10.BR mice (Fig. 1). However, starting around day 11 after injection, the number of CD8+Des+ T cells slowly decreased, and ∼90% of the CD8+Des+ T cells had disappeared from the spleen of Des→Alb-Kb mice by day 49 after transfer (Fig. 1).

FIGURE 1.

Disappearance of CD8+Des+ T cells following transfer into Alb-Kb-transgenic mice. Control B10.BR- (○) or Alb-Kb- (•) transgenic mice were adult thymectomized, irradiated, and i.v. injected with 6–8 × 106 T cells isolated from Des-transgenic mice. At different time points after adoptive transfer, the numbers of CD8+Des+ T cells in the spleens of the recipient mice were determined by FACS staining. For each time point, 10–20 mice were analyzed except for days 22, 28, and 49, on which 3–5 animals were tested. The mean value and SD are shown.

FIGURE 1.

Disappearance of CD8+Des+ T cells following transfer into Alb-Kb-transgenic mice. Control B10.BR- (○) or Alb-Kb- (•) transgenic mice were adult thymectomized, irradiated, and i.v. injected with 6–8 × 106 T cells isolated from Des-transgenic mice. At different time points after adoptive transfer, the numbers of CD8+Des+ T cells in the spleens of the recipient mice were determined by FACS staining. For each time point, 10–20 mice were analyzed except for days 22, 28, and 49, on which 3–5 animals were tested. The mean value and SD are shown.

Close modal

These results suggest that, in the Alb-Kb mice, expression of the H-2Kb Ag by hepatocytes may induce deletion of the CD8+Des+ T cells.

Deletion of autoreactive T cells has been reported in several models of peripheral T cell tolerance. In all cases, deletion was preceded by a transient phase of autoimmunity and proliferation of the autoreactive T cells (2, 3, 4). We therefore determined whether in our model, too, disappearance of the CD8+Des+ T cells from the spleen was preceded by transient activation of the autoreactive T cells.

Histological analysis showed that livers from Des→Alb-Kb mice, but not livers from Des→B10.BR mice, were infiltrated from day 2 to 6 after transfer. However, the infiltration was transient and limited to the portal triads with minor invasion throughout the lobules and no evidence of tissue damage (data not shown). We further characterized the liver-infiltrating cells by FACS staining of mononuclear cells isolated from livers of Des→B10.BR or Des→Alb-Kb mice. Infiltrates from Des→B10.BR liver contained few CD8+Des+ T cells, none of which expressed the early and late activation markers CD69 and CD44, respectively (Fig. 2,A). In contrast, we found an accumulation of CD8+Des+ T cells in the livers of Des→Alb-Kb mice, all of which expressed the activation markers CD69 and CD44 (Fig. 2 A).

FIGURE 2.

CD8+Des+ T cells isolated from Des→Alb-Kb mice are activated. A, Liver-infiltrating lymphocytes were isolated 4 days after transfer from four control Des→B10.BR (thin line) or 4 Des→Alb-Kb transgenic (thick line) mice. Cells were triple stained for Des-Tricolor, CD8-PE, and CD69-FITC or CD44-FITC as indicated. Double-positive CD8+Des+ cells were gated (R1) and analyzed for CD69 and CD44 expression. The percentage of double-positive CD8+Des+ cells is indicated in the dot plots. Total recovery from four livers was 2 × 105 cells (3.2 × 103 CD8+Des+) and 2.4 × 106 (2.8 × 104 CD8+Des+) for Des→B10.BR and Des→Alb-Kb mice, respectively. Similar profiles and recoveries were observed at days 5 and 6. B, Control B10.BR or transgenic-Alb-Kb mice were transferred with either unlabeled (Des, two left-side histograms) or CFSE-labeled (Des/CFSE, right-side histograms) T cells isolated from the Des-transgenic mice. The splenocytes were isolated from Des→B10.BR (thin line) or Des→Alb-Kb (thick line) mice at different time points after transfer. Splenocytes from mice that received unlabeled T cells (Des) were triple-stained and analyzed as in A. Splenocytes from mice that received CFSE-labeled T cells (Des/CFSE) were stained with Des-Tricolor and CD8-PE. Double-positive CD8+Des+ cells were gated (R1) and analyzed for the intensity of CFSE staining. C, CD8+ T cells were isolated at day 9 after transfer from Des→B10BR or Des→Alb-Kb mice and stained with Cy-5-conjugated annexin V and propidium iodide (PI) as indicated. The percentage of cells that are alive (annexin VPI), apoptotic (annexin V+PI), or dead (annexin V+PI+) is indicated in the corresponding gate.

FIGURE 2.

CD8+Des+ T cells isolated from Des→Alb-Kb mice are activated. A, Liver-infiltrating lymphocytes were isolated 4 days after transfer from four control Des→B10.BR (thin line) or 4 Des→Alb-Kb transgenic (thick line) mice. Cells were triple stained for Des-Tricolor, CD8-PE, and CD69-FITC or CD44-FITC as indicated. Double-positive CD8+Des+ cells were gated (R1) and analyzed for CD69 and CD44 expression. The percentage of double-positive CD8+Des+ cells is indicated in the dot plots. Total recovery from four livers was 2 × 105 cells (3.2 × 103 CD8+Des+) and 2.4 × 106 (2.8 × 104 CD8+Des+) for Des→B10.BR and Des→Alb-Kb mice, respectively. Similar profiles and recoveries were observed at days 5 and 6. B, Control B10.BR or transgenic-Alb-Kb mice were transferred with either unlabeled (Des, two left-side histograms) or CFSE-labeled (Des/CFSE, right-side histograms) T cells isolated from the Des-transgenic mice. The splenocytes were isolated from Des→B10.BR (thin line) or Des→Alb-Kb (thick line) mice at different time points after transfer. Splenocytes from mice that received unlabeled T cells (Des) were triple-stained and analyzed as in A. Splenocytes from mice that received CFSE-labeled T cells (Des/CFSE) were stained with Des-Tricolor and CD8-PE. Double-positive CD8+Des+ cells were gated (R1) and analyzed for the intensity of CFSE staining. C, CD8+ T cells were isolated at day 9 after transfer from Des→B10BR or Des→Alb-Kb mice and stained with Cy-5-conjugated annexin V and propidium iodide (PI) as indicated. The percentage of cells that are alive (annexin VPI), apoptotic (annexin V+PI), or dead (annexin V+PI+) is indicated in the corresponding gate.

Close modal

Starting at day 5 after transfer, 2–3 days after the beginning of liver infiltration, a fraction of the CD8+Des+ T cells from the spleen of Des→Alb-Kb mice expressed CD44. The percentage of CD8+Des+ T cells that expressed CD44+ further increased thereafter, with most CD8+Des+ T cells being CD44+ by day 6 after transfer (Fig. 2,B). In contrast to liver-infiltrating T cells, the CD8+Des+ T cells recovered from the spleen of Des→Alb-Kb mice did not express the early activation marker CD69 that is only transiently expressed by stimulated T cells (Fig. 2 B). No change in the level of CD25 and CD62 ligand expression by CD8+Des+ T cells isolated from Des→B10.BR or Des→Alb-Kb was detectable (data not shown). The kinetics of CD69 and CD44 expression suggest that, upon injection into Alb-Kb mice, CD8+Des+ T cells rapidly homed to the liver, where they were stimulated by H-2Kb-expressing hepatocytes, and the activated T cells then returned to the spleen.

We further determined whether stimulation by H-2Kb-expressing liver cells induced the proliferation of the CD8+Des+ T cells. T cell division can be measured with the fluorescent dye CFSE, which distributes equally between the two daughter cells at every division, thus resulting in a 2-fold reduction of the fluorescence intensity (30). The Des T cells were CFSE-labeled before transfer into recipient mice. When injected into control B10.BR mice, most CD8+Des+ T cells had not divided, and some divided up to two times within the 12 days of analysis (Fig. 2,B). This is in agreement with the recent observation that peripheral T cells maintain a low rate of division induced by interaction of their TCR with self MHC/peptide complexes (35, 36, 37). In contrast, CD8+Des+ T cells isolated from Des→Alb-Kb mice had divided multiple times (Fig. 2,B). By day 6 after transfer, 50% of the cells had accomplished more than three divisions, and some divided seven times (Fig. 2 B). The CD8+Des+ T cells that accomplished several divisions expressed CD44 and are thus T cells that had likely infiltrated the liver (data not shown).

Therefore, H-2Kb-expressing hepatocytes stimulated and induced proliferation of the CD8+Des+ T cells. Interestingly, despite a high rate of division, the number of CD8+Des+ T cells in the spleen of Des→Alb-Kb mice was not significantly increased as compared with control Des→B10.BR at day 6 or 9 after transfer, suggesting that deletion may already be occurring. Importantly, at no time point did we detect a significant down-modulation of the expression of the TCR or CD8 coreceptor (data not shown). As described in several other models, we did not detect an increase in dying or apoptotic CD8+Des+ T cells in Alb-Kb mice as compared with control B10.BR mice. Indeed, the number of apoptotic CD8+Des+ T cells as detected by annexinV and propidium iodide staining was comparable in control B10.BR mice and transgenic Alb-Kb mice at day 6 and 9 after transfer (Fig. 2 C and data not shown).

The anti-apoptotic molecules Bcl-2 and Bcl-xL play a critical role in the regulation of T cell survival (38). Hence, constitutive expression of Bcl-xL prevents apoptosis in T cells activated in the absence of CD28-dependent costimulatory signal (39). The expression of Bcl-2 or Bcl-xL also prevents the in vivo deletion of mature CD8+ T cells induced by peptide Ag (40). However, neither Bcl-2 nor Bcl-xL prevent the decline of Ag-specific CD8+ T cells after viral infection (40). Therefore, we determined whether deletion of the CD8+Des+ T cells in Alb-Kb mice resulted from abnormal expression of either the Bcl-2 or Bcl-xL protein. The level of expression of Bcl-2 and Bcl-xL proteins was determined by Western blot analysis of cytoplasmic extracts prepared from CD8+Des+ T cells isolated from Des→B10.BR mice, Des→Alb-Kb mice, or Des→B10.BR mice primed in vivo with H-2Kb-expressing APCs (Fig. 3). We found that the level of Bcl-2 and Bcl-xL protein expression was comparable in the different population, indicating that deletion of the CD8+Des+ T cells in Des→Alb-Kb mice is not resulting from reduced expression of either of these two antiapoptotic proteins.

FIGURE 3.

Deletion of CD8+Des+ T cells in Des→Alb-Kb mice does not result from reduced expression of Bcl-2 or Bcl-xL. CD8+ T cells were isolated from either Des→B10.BR or from Des→Alb-Kb mice at day 5 or 9 after transfer. Similarly, CD8+ T cells were isolated from Des→B10.BR mice at day 5 or 9 after priming with H-2Kb-expressing-APCs (Des primed). Eight to ten micrograms of cytoplasmic extracts were used for Western blotting, and the different blots were revealed with Bcl-xL, Bcl-2, and actin-specific Abs as indicated. The blots were quantified, and the relative values corrected for loading differences using the actin control are indicated.

FIGURE 3.

Deletion of CD8+Des+ T cells in Des→Alb-Kb mice does not result from reduced expression of Bcl-2 or Bcl-xL. CD8+ T cells were isolated from either Des→B10.BR or from Des→Alb-Kb mice at day 5 or 9 after transfer. Similarly, CD8+ T cells were isolated from Des→B10.BR mice at day 5 or 9 after priming with H-2Kb-expressing-APCs (Des primed). Eight to ten micrograms of cytoplasmic extracts were used for Western blotting, and the different blots were revealed with Bcl-xL, Bcl-2, and actin-specific Abs as indicated. The blots were quantified, and the relative values corrected for loading differences using the actin control are indicated.

Close modal

Despite the proliferative response of the CD8+Des+ T cells in Des→Alb-Kb mice, we detected no sign of liver damage or autoimmune disease (data not shown), suggesting that activation of the CD8+Des+ T cells in the Des→Alb-Kb mice was transient, followed by T cell tolerance. One would predict, then, that the CD8+Des+ T cells recolonizing the spleen should be tolerant and thus unresponsive to in vitro restimulation with H-2Kb-expressing APCs. Indeed, splenocytes from Alb-Kb mice responded very poorly to H-2Kb-expressing APCs in vitro (Fig. 4). Proliferation (data not shown) and IL-2 and IFN-γ production by in vitro stimulated CD8+Des+ T cells isolated from Des→Alb-Kb was reduced by day 5 and almost completely abrogated by day 11 after transfer (Fig. 4). Interestingly, as the percentage of CD44+CD8+Des+ T cells (reflecting the percentage of cells that had been stimulated by H-2Kb-expressing hepatocytes) increased, the unresponsiveness augmented (compare Figs. 2 B and 4). We could not detect IL-4 production by these unresponsive T cells (data not shown).

FIGURE 4.

Partial unresponsiveness of CD8+Des+ T cells isolated from Des→Alb-Kb mice. Splenocytes from Des→B10.BR- or Des→Alb-Kb-transgenic mice were restimulated in vitro with B10.BR or H-2Kb-expressing CBK splenocytes. Groups of three and five mice were tested at days 5 and 11 after transfer, respectively. A total of 4 × 106 splenocytes (8–12 × 104 CD8+Des+) were stimulated in vitro, and supernatants were collected at 24 h and 48 h after activation to measure IL-2 production and γIFN production, respectively. Because the representation of CD8+Des+ T cells in the different populations analyzed varied slightly, we normalized all the responses to 1 × 105 CD8+Des+ T cells. Thus, the values presented are these corrected values. The percentage in the different graphs refers to the reduction of the response. Similar results were obtained at 4 (three experiments), 5 (two experiments), 6 (one experiment), and 10–12 (four experiments) days.

FIGURE 4.

Partial unresponsiveness of CD8+Des+ T cells isolated from Des→Alb-Kb mice. Splenocytes from Des→B10.BR- or Des→Alb-Kb-transgenic mice were restimulated in vitro with B10.BR or H-2Kb-expressing CBK splenocytes. Groups of three and five mice were tested at days 5 and 11 after transfer, respectively. A total of 4 × 106 splenocytes (8–12 × 104 CD8+Des+) were stimulated in vitro, and supernatants were collected at 24 h and 48 h after activation to measure IL-2 production and γIFN production, respectively. Because the representation of CD8+Des+ T cells in the different populations analyzed varied slightly, we normalized all the responses to 1 × 105 CD8+Des+ T cells. Thus, the values presented are these corrected values. The percentage in the different graphs refers to the reduction of the response. Similar results were obtained at 4 (three experiments), 5 (two experiments), 6 (one experiment), and 10–12 (four experiments) days.

Close modal

The CD8+Des+ T cells isolated from Des→Alb-Kb mice appear unresponsive to further in vitro stimulation. Therefore, these T cells will be considered as tolerant by contrast with the CD8+Des+ T cells isolated from control Des→B10.BR mice, which will be considered as naive.

The AP-1 and NF-κB transcription factors participate in the expression of several cytokine genes and are two known molecular targets of CD4+ T cell anergy induced in vitro by stimulation in the absence of CD28 costimulation or in vivo by repeated injection of superantigen (25, 26). To understand the molecular mechanisms of unresponsiveness in the CD8+Des+ T cell population, we examined the AP-1 and NF-κB transcriptional activity induced by in vitro restimulation of these cells. To directly measure transcriptional activity, we used two different lines of transgenic mice that express the firefly luciferase gene under the control of either AP-1 (AP-1-Luc) (11) or NF-κB sites (NF-κB-Luc) (28). As previously reported, luciferase activity in the AP-1-Luc-transgenic mice directly reflects AP-1 transcriptional activity (11, 41). Likewise, luciferase activity in the NF-κB-Luc-transgenic mice directly reflects NF-κB transcriptional activity. Indeed, luciferase activity in the NF-κB-Luc mice is only detected in tissues that demonstrate NF-κB DNA binding activity (28). Furthermore, luciferase reporter activity is constituvely expressed in B cells and subpopulations of thymocytes but not in T cells, although it can be induced in resting T cells by known inducers of NF-κB ((28, 42, 43, 44) and Fig. 5).

FIGURE 5.

Regulation of NF-κB transcriptional activity in B and T cells. A, The luciferase activity of 5 × 106 naive B or T cells from NF-κB-Luc was analyzed. B, CD4+ T cells were isolated from NF-κB-Luc single-transgenic mice and stimulated with either PMA (5 ng/ml) and ionomycin (250 ng/ml) (PI) or plate-bound anti-CD3 (5 μg/ml) in the absence or presence of soluble anti-CD28 (1 μg/ml), as indicated. At the indicated time after stimulation, the cells were harvested and luciferase activity measured. Syngeneic response (subtracted) did not exceed 25 RLU.

FIGURE 5.

Regulation of NF-κB transcriptional activity in B and T cells. A, The luciferase activity of 5 × 106 naive B or T cells from NF-κB-Luc was analyzed. B, CD4+ T cells were isolated from NF-κB-Luc single-transgenic mice and stimulated with either PMA (5 ng/ml) and ionomycin (250 ng/ml) (PI) or plate-bound anti-CD3 (5 μg/ml) in the absence or presence of soluble anti-CD28 (1 μg/ml), as indicated. At the indicated time after stimulation, the cells were harvested and luciferase activity measured. Syngeneic response (subtracted) did not exceed 25 RLU.

Close modal

Mice transgenic for the Des TCR were thus crossed with either the AP-1-Luc (AP-1 × Des) or NF-κB-Luc (NF-κB × Des) transgenic mice, and T cells from double-transgenic mice were injected into either B10.BR or Alb-Kb mice. The in vitro stimulation of naive CD8+Des+ T cells isolated from Des→B10.BR mice with H-2Kb-expressing APCs induced both AP-1 and NF-κB transcriptional activity (Fig. 6). The AP-1 and NF-κB transcriptional activity induced by in vitro stimulation of the tolerant CD8+Des+ T cells isolated from Des→Alb-Kb mice was reduced as compared with that of naive CD8+Des+ T cells (Fig. 6). The overall reduction in AP-1 and NF-κB transcriptional activity was comparable to the reduction of cytokine production (compare Figs. 4 and 6).

FIGURE 6.

Tolerized T cells present a reduced AP-1 and NF-κB transcriptional activity after in vitro restimulation. Des→B10.BR (▨) or Des→Alb-Kb (▪) mice were injected with Des+ T cells isolated from AP-1 × Des (A) or NF-κB × Des (B) double-transgenic mice. At day 5 after transfer, 12 × 106 splenocytes were stimulated with syngeneic B10.BR APCs, H-2Kb-expressing CBK APCs, or Con A in the presence of B10.BR APCs. At the indicated time after stimulation, the cells were harvested and luciferase activity read. The percentage of CD8+Des+ T was determined before activation. Using this value, the RLU for 5 × 105 CD8+Des+ T cells were calculated and are shown in the figure. Syngeneic response (subtracted) did not exceed 25 RLU. The percentage in the different graphs refers to the reduction of the response. One experiment of two for AP-1 activity and four for NF-κB activity is shown.

FIGURE 6.

Tolerized T cells present a reduced AP-1 and NF-κB transcriptional activity after in vitro restimulation. Des→B10.BR (▨) or Des→Alb-Kb (▪) mice were injected with Des+ T cells isolated from AP-1 × Des (A) or NF-κB × Des (B) double-transgenic mice. At day 5 after transfer, 12 × 106 splenocytes were stimulated with syngeneic B10.BR APCs, H-2Kb-expressing CBK APCs, or Con A in the presence of B10.BR APCs. At the indicated time after stimulation, the cells were harvested and luciferase activity read. The percentage of CD8+Des+ T was determined before activation. Using this value, the RLU for 5 × 105 CD8+Des+ T cells were calculated and are shown in the figure. Syngeneic response (subtracted) did not exceed 25 RLU. The percentage in the different graphs refers to the reduction of the response. One experiment of two for AP-1 activity and four for NF-κB activity is shown.

Close modal

Therefore, the inability of unresponsive CD8+Des+ T cells to produce cytokines upon TCR signaling in vitro correlates with a reduced AP-1 and NF-κB transcriptional activity, further suggesting that unresponsive CD8+Des+ T cells have a defect on the TCR signaling pathway.

In our model, hepatocytes stimulated the CD8+Des+ T cells. However, activation was transient, and the stimulated T cells became unresponsive and likely died. This contrasts with T cells stimulated by professional APCs that develop into effector cells and long-lived memory cells. To understand how T cell fate may be controlled in these two situations, we examined the regulation of AP-1 and NF-κB transcription factors as tolerance or immunity developed in vivo.

The tolerant T cells were generated as described above by transfer, into Alb-Kb mice, of CD8+Des+ T cells isolated from either AP-1 × Des- or NF-κB × Des- transgenic mice. In parallel, in vivo primed CD8+Des+ cells were generated by immunizing AP-1 × Des or NF-κB × Des double-transgenic mice with H-2Kb-expressing APCs, a protocol known to induce effector and memory CTLs. The AP-1 and NF-κB transcriptional activity was analyzed at different time points thereafter. We found that in vivo primed CD8+Des+ T cells exhibited a low AP-1 and NF-κB transcriptional activity, with maximal level of luciferase activity 24 h after priming (data not shown). However, in vivo primed T cells had a 2- to 3-fold lower AP-1 and NF-κB transcriptional activity as compared with in vitro activated T cells. Not surprisingly, then, when CD8+Des+ T cells isolated from Des→Alb-Kb or Des→B10.BR mice were analyzed for in vivo AP-1 or NF-κB transcriptional activity, no signal could be detected using this enzymatic assay. Thus, we used RT-PCR to measure luciferase mRNA expression that also reflects AP-1 or NF-κB transcriptional activity but should be more sensitive than enzymatic assays. Significant levels of AP-1 and NF-κB transcriptional activity were detected at 13 h after activation (Fig. 7 A). The AP-1 transactivation declined thereafter, whereas the NF-κB transcriptional activity was long lasting and still clearly detectable at 37 h after in vivo priming. Thus, in vivo activation of CD8+Des+ T cells is associated with an induction of AP-1 and NF-κB transcriptional activity.

FIGURE 7.

AP-1 and NF-κB transcriptional activity expressed by in vivo primed or tolerized T cells. A, AP-1 × Des (▨) or NF-κB × Des (▪) double-transgenic mice were primed by i.v. injection of H-2Kb-expressing B cell blast (primed) or left unimmunized (unprimed). At different time points after immunization, CD8+ T cells were isolated and luciferase mRNA measured by semiquantitative RT-PCR. B, Control B10.BR or Alb-Kb-transgenic mice were injected with Des+ T cells isolated from AP-1 × Des (○) or NF-κB × Des (•) double-transgenic mice. At different time points after transfer, CD8+ T cells were isolated and luciferase mRNA expression was determined by RT-PCR. All samples were normalized on the basis of the amount of HPRT mRNA detected by RT-PCR and corrected for the number of CD8+Des+ T cells. The results are expressed as a ratio of (A) the amount of luciferase mRNA expressed by T cells isolated from primed mice divided by the amount expressed by T cells isolated from unprimed mice and (B) the amount of luciferase mRNA expressed by T cells isolated from experimental Des→Alb-Kb mice divided by the amount expressed by T cells isolated from control Des→B10.BR mice. For each sample, the reverse transcription and the PCR were run in duplicates and the average of the four measurements used for calculation. Each symbol represents one experimental group.

FIGURE 7.

AP-1 and NF-κB transcriptional activity expressed by in vivo primed or tolerized T cells. A, AP-1 × Des (▨) or NF-κB × Des (▪) double-transgenic mice were primed by i.v. injection of H-2Kb-expressing B cell blast (primed) or left unimmunized (unprimed). At different time points after immunization, CD8+ T cells were isolated and luciferase mRNA measured by semiquantitative RT-PCR. B, Control B10.BR or Alb-Kb-transgenic mice were injected with Des+ T cells isolated from AP-1 × Des (○) or NF-κB × Des (•) double-transgenic mice. At different time points after transfer, CD8+ T cells were isolated and luciferase mRNA expression was determined by RT-PCR. All samples were normalized on the basis of the amount of HPRT mRNA detected by RT-PCR and corrected for the number of CD8+Des+ T cells. The results are expressed as a ratio of (A) the amount of luciferase mRNA expressed by T cells isolated from primed mice divided by the amount expressed by T cells isolated from unprimed mice and (B) the amount of luciferase mRNA expressed by T cells isolated from experimental Des→Alb-Kb mice divided by the amount expressed by T cells isolated from control Des→B10.BR mice. For each sample, the reverse transcription and the PCR were run in duplicates and the average of the four measurements used for calculation. Each symbol represents one experimental group.

Close modal

Using the same approach, we compared the level of AP-1 and NF-κB transcriptional activity expressed by CD8+Des+ T cells that were isolated at different time points from Des→Alb-Kb (tolerant) or Des→B10.BR (naive) mice. The tolerant T cells did not show detectable levels of AP-1 transcriptional activity (Fig. 7,B). The very weak AP-1 signal detected at day 6 after transfer was not sustained at day 11 or 12 after transfer, when most CD8+Des+ T cells isolated from Des→Alb-Kb mice are tolerant. Similarly, at the early time point when no tolerant CD8+Des+ T were found in the spleen of Des→Alb-Kb, T cells isolated from Des→B10.BR or Des→Alb-Kb mice expressed no significant NF-κB transcriptional activity (Fig. 7,B, day 4). However, tolerant T cells showed a low level of NF-κB transcriptional activity. Indeed, NF-κB transcriptional activity was low at day 5 after transfer, when only 15–20% of the CD8+Des+ T cells were CD44+ (tolerant) and increased thereafter as the fraction of tolerant CD8+Des+ T cells increased (Fig. 7,B). However, the maximal level of NF-κB transcriptional activity expressed by tolerant T cells was extremely low and remained 10- to 20-fold lower than the NF-κB transcriptional activity expressed by in vivo primed CD8+Des+ T cells (compare Fig. 7, A and B).

Therefore, T cell activation and survival correlate with induction of AP-1 and NF-κB transcriptional activity, whereas deletion of the T cell correlates with weak NF-κB and undetectable AP-1 transcriptional activity.

We further examined how the differences in AP-1 and NF-κB transcriptional activity detected in the different CD8+Des+ T cell populations may correlate with differences in the AP-1 and NF-κB complexes generated. We performed EMSA on nuclear extracts prepared from the different CD8+Des+ T cells analyzed above. To define the AP-1 and NF-κB complexes induced by in vivo priming, we isolated CD8+ T cells from Des-transgenic mice either primed 5 h earlier with H-2Kb-expressing APCs or left unprimed. To analyze the complexes induced by a tolerogenic signal, we isolated the CD8+ T cells from either Des→B10.BR or Des→Alb-Kb mice 9 days after adoptive transfer, when most CD8+Des+ T cells isolated from Des→Alb-Kb mice are CD44+ and unresponsive (Figs. 1, 2, and 4).

In vivo priming led to an increase in AP-1 complexes detected by EMSA (Fig. 8,A). Naive and tolerant T cells, that did not show AP-1 transactivation, had no detectable AP-1 complexes (Fig. 8 A).

FIGURE 8.

Tolerant T cells express unusual NF-κB complexes. Des-transgenic mice were primed as described in Fig. 6 (primed) or left untreated (unprimed). Five hours later, the CD8+ T cells were isolated and nuclear extracts prepared. CD8+ T cells were isolated from either Des→B10.BR or Des→Alb-Kb mice at day 9 after transfer and nuclear extracts prepared. Three micrograms of nuclear extracts were used for EMSA (A, B, and C) or supershift experiments (D and E) using an AP-1- (A), NF-κB- (B, D, and E), or consensus OCT1 and 2- (C) specific probe. The AP-1 complex present in the “Des-primed” cells was competed with a cold AP-1 oligonucleotide (A, cold). The specificity of the Ab used in the supershifts is indicated above each lane (D and E). Supershifts were run normally (E) to exemplify the supershifted complexes or for a shorter period (D) to exemplify the lower mobility complexes and the inhibitions of the corresponding complexes. ns, Nonspecific. ◂ indicates DNA binding complexes and ◃, the supershifted complexes. Similar results were obtained in three independent experiments.

FIGURE 8.

Tolerant T cells express unusual NF-κB complexes. Des-transgenic mice were primed as described in Fig. 6 (primed) or left untreated (unprimed). Five hours later, the CD8+ T cells were isolated and nuclear extracts prepared. CD8+ T cells were isolated from either Des→B10.BR or Des→Alb-Kb mice at day 9 after transfer and nuclear extracts prepared. Three micrograms of nuclear extracts were used for EMSA (A, B, and C) or supershift experiments (D and E) using an AP-1- (A), NF-κB- (B, D, and E), or consensus OCT1 and 2- (C) specific probe. The AP-1 complex present in the “Des-primed” cells was competed with a cold AP-1 oligonucleotide (A, cold). The specificity of the Ab used in the supershifts is indicated above each lane (D and E). Supershifts were run normally (E) to exemplify the supershifted complexes or for a shorter period (D) to exemplify the lower mobility complexes and the inhibitions of the corresponding complexes. ns, Nonspecific. ◂ indicates DNA binding complexes and ◃, the supershifted complexes. Similar results were obtained in three independent experiments.

Close modal

Low levels of two defined NF-κB complexes were present in unprimed T cells, and they were strongly up-regulated in extracts isolated from primed T cells (Fig. 8,B, complexes N1 and N2). Variable levels of these NF-κB complexes were present in the nuclei of CD8+ T cells isolated from either Des→B10.BR or Des→Alb-Kb mice. Mainly complex N1 was barely detectable in some extracts prepared from the tolerant population (Fig. 8,D and data not shown). Interestingly, an additional complex was highly abundant in the tolerant CD8+ T cells isolated from Des→Alb-Kb mice (Fig. 8 B, complex N3).

To determine the composition of the different NF-κB complexes present in primed or tolerant CD8+Des+ T cells, we performed EMSA in the presence of Abs specific for the different NF-κB members. Supershift experiments indicated that the NF-κB complexes in primed CD8+Des+ T cells contain p65 and p50 and may contain c-Rel (Fig. 8, D and E). Therefore, complex N1 likely corresponds to the previously reported p65/p50 dimers, and complex N2 contains p50/p50 and, potentially, p65/c-Rel dimers (Fig. 8,D). The lower band of the N3 complexes present in the tolerant CD8+Des+ T cells isolated from Des→Alb-Kb mice was weakly inhibited by p50-specific Ab and completely inhibited by a p65-specific Ab but not with an anti-c-Rel Ab (Fig. 8, D and E, complexes IV). The upper band of these complexes was neither inhibited nor supershifted by p65- or c-Rel-specific Abs alone, but disappeared almost completely when a mix of anti-p65 and anti-c-Rel Abs was included in the binding reaction, suggesting that it may contain p65/c-Rel heterodimer (Fig. 8, D and E, complexes III). This band was also inhibited to some extent by p50-specific Abs and may therefore contain p50 homodimers (Fig. 8 D). The migration pattern of complexes III and IV is rather different from that expected from p65-containing complexes. This may reflect additional posttranscriptional modifications (45, 46, 47).

We found that the primed and tolerant T cells that show high and weak NF-κB transcriptional activity, respectively, also express different NF-κB complexes. Indeed, the complexes detected in primed CD8+ T cells are mainly p65/p50 and p50/p50 dimers. These classical p50-containing complexes are barely detectable in the tolerant population that expresses predominantly p65- and c-Rel-containing complexes that may correspond to homo- or heterodimer with c-Rel.

As an approach to understand the in vivo regulation of the fate of autoreactive T cells, we developed an adoptive transfer protocol in which both the cellular and molecular mechanisms of CD8+ T cell tolerance could be defined. We analyzed the response of CD8+ T cells to a self Ag, the class I alloantigen H-2Kb exclusively expressed by liver cells. Adoptive transfer of mature CD8+Des+ T cells into Alb-Kb mice led to autoimmunity, as evidenced by the liver infiltration observed by day 2 after transfer. However, the autoimmune response was mild and transient and tolerance followed. Indeed, we found no evidence of liver damage, and liver infiltration cleared by day 6 after transfer. Clearly, hepatocytes can stimulate the H-2Kb-specific T cells, as evidenced by the induction of CD69 and CD44 expression and the proliferation of the CD8+Des+ T cells. However, despite extensive proliferation, the number of CD8+Des+ T cells does not increase, and deletion of the self-reactive CD8+Des+ T cells is evident by day 10 after transfer. Therefore, tolerance in this model is likely resulting from an abortive activation by hepatocytes of the H-2Kb-specific T cells leading to T cell death, as previously suggested by Bertolino and collaborators (3, 48, 49).

It is largely believed that naive T cells do not generally have access to tissue. Our results suggest that the liver is an exception to this rule. Indeed, in our model liver, infiltration is observed as early as 2 days after adoptive transfer, wheras no activated T cells are yet detected in the spleen or lymph nodes of Alb-Kb mice. Furthermore, activated CD8+Des+ T cells appear in the spleen of Alb-Kb mice by day 6 after transfer, when liver infiltration has almost completely cleared. Finally, in contrast to liver-infiltrating CD8+Des+ T cells, the CD8+Des+ T cells found in the spleen do not express CD69, a marker that is only transiently expressed by activated T cells. Altogether, these results strongly suggest that naive CD8+Des+ T cells gain access to the liver, are activated by H-2Kb-expressing liver cells, and then recirculate to the spleen.

The mechanism of T cell deletion in Des→Alb-Kb is still unknown. It is possible that T cell death directly results from TCR stimulation in the absence of CD28-dependent costimulatory signals. Lack of expression of the antiapoptotic molecule Bcl-xL is thought, in this case, to cause deletion of the responding T cells (39). However, both Bcl-2 and Bcl-xL are normally expressed in the tolerant CD8+Des+ T cells isolated from Des→Alb-Kb. An alternative explanation comes from the recent observation that the long-term maintenance of naive or memory T cells depends on survival signals induced by TCR engagement (35, 36, 37, 50, 51). Interestingly, hepatocyte-stimulated CD8+Des+ T cells are unresponsive. Indeed, the CD8+Des+ T cells do not proliferate, produce cytokines, nor transactivate AP-1 or NF-κB upon in vitro restimulation. Therefore, the tolerized CD8+Des+ T cells present a generalized defect on the TCR signaling pathway. Deletion of the CD8+Des+ T cells in Des→Alb-Kb mice may therefore result from a lack of survival signals due to a desensitization of their TCR. Further studies of the exact downstream events induced by TCR engagement that regulate T cell homeostasis are clearly required to resolve that issue.

In this study, we compared the transcriptional program expressed by three different T cell populations: 1) unstimulated, naive T cells (Des→B10BR) that have a life span of >50 days, 2) T cells that have been primed in vivo and would develop under those conditions into effectors and memory cells, and 3) tolerant CD8+ T cells that are unresponsive and further deleted. We found that prone-to-die tolerant CD8+Des+ T cells diverge from long-lived naive and primed T cells mainly at the NF-κB signaling pathway. Indeed, hepatocyte-stimulated T cells show no detectable AP-1 and only a weak NF-κB transcriptional activity. This contrasts with primed T cells, which demonstrate high AP-1 and NF-κB transactivation, or naive T cells, which do not have detectable AP-1 and NF-κB transcriptional activity. As naive CD8+Des+ T cells, tolerant cells show no detectable AP-1 complexes and nuclear localization of JunD in the absence of Fos family members (data not shown). Priming induces an increase in nuclear JunD and all Fos family members (data not shown). Biochemical studies further revealed an intriguing difference in the NF-κB complexes expressed by tolerant CD8+Des+ T cells as compared with naive or primed CD8+Des+ T cells. The NF-κB complexes present in naive and primed T cells are mainly composed of p65/p50 (N1 complex) and p50/p50 (N2 complex) dimers; priming only increases the relative representation of these complexes and may induce low levels of c-Rel-containing complexes. Likewise, at day 5 or 10 after in vivo priming, CD8+Des+ T cells express the N1 and N2 complexes only at levels comparable to that of naive T cells (data not shown). Although the NF-κB complexes present in naive and primed T cells contain substantial levels of p50, this NF-κB member is almost undetectable in the complexes present in the tolerant CD8+Des+ T cells that contain p65 and c-Rel. Although the exact stoichiometry of the NF-κB complexes cannot be clearly established by supershift experiments, these results suggest that the NF-κB complexes detected in the tolerant CD8+Des+ may correspond to homo- or heterodimers of p65 with c-Rel. How the change in NF-κB complexes is regulated in the tolerant CD8+Des+ T cells still remains to be determined; however, it may result from defective TCR signaling or signaling through other cell surface receptors.

Despite the differences in Ag stimulation, the response of CD8+Des+ T cells to H-2Kb-expressing hepatocytes and the in vivo responses of CD4+ T cells to superantigens have similar features: both stimuli induce proliferation of the responding cells and subsequent unresponsiveness. However, these responses diverge in that tolerant CD8+Des+ T cells were further deleted, whereas superantigen-stimulated CD4+ T cells were anergic and long-lived. In vivo anergized CD4+ T cells showed no AP-1 transactivation and show a failure of NF-κB activation linked to reduced nuclear localization of RelA and c-Rel (26). When comparing these two different studies, it is tempting to speculate that the difference in NF-κB complexes’ regulation may determine the fate of the responding T cell’s deletion or survival. However, this is at odds with recent studies that suggest that NF-κB is an essential regulator of T cell survival (22, 23). This was shown in transgenic mice expressing in T cells a trans-dominant IκBα that completely blocks the nuclear translocation of p65/p50 complexes, whereas the regulation of p50 homodimers remained normal (22, 23). One possible explanation to reconcile these observations is that the different NF-κB complexes have different roles in T cell activation and homeostasis. In correlation, it has been shown that the affinity for κB sites and the regulation of gene expression depends on the nature of the NF-κB complexes (reviewed in Ref. (52)). Furthermore, targeted disruption of individual members of the NF-κB family suggested that the different proteins have distinct biological functions (15, 16, 53, 54, 55, 56). Finally, it remains possible that, in the tolerant CD8+Des+ T cells, the p65-containing complexes (N3 complexes) are further regulated by post-trancriptional modification, as suggested by their altered migration in EMSA. These modifications may alter the transcriptional activity of these complexes, a possibility that still needs to be investigated. Further characterization of the NF-κB complexes detected in the tolerant CD8+Des+ T cell population may help to determine to what extent κB-dependent gene expression participates in T cell survival.

We thank R. A. Flavell and S. Ghosh for providing us the AP-1- and NF-κB-Luc-transgenic mice, B. Arnold and G. Hämmerling for the Désiré and Alb-Kb mice, and M. Pophillat for animal care. We are grateful to C. Boyer, M. Buferne, A. Guimezanes, B. Malissen, and L. Leserman for critical experimental advice, helpful discussion, and critical reading of the manuscript.

1

This work was supported by institutional grants (to S.G. and A.-M.S.-V.) from Institut National de la Santé et de la Recherche Médicale and Centre National de la Recherche Scientifique, grants from Association pour la Recherche sur le Cancer, Ligue Nationale Contre le Cancer (Axe Immunologie des Tumeurs), Ligue Nationale Contre le Cancer (Comité des Bouches du Rhône) and Groupement des Entreprises Françaises dans la Lutte Contre le Cancer, and Grant R29AI42138 from the National Institutes of Health (to M.R.).

3

Abbreviations used in this paper: IκB, inhibitor of κB; Des, Désiré; AP-1-Luc, AP-1-luciferase; NF-κB-Luc, NF-κB luciferase; RLU, relative luminescence units; HPRT, hypoxanthine phosphoribosyl transferase; PI, propidium iodide.

1
Morahan, G., J. Allison, J. F. A. P. Miller.
1989
. Tolerance of class I histocompatability antigens expressed extrathymically.
Nature
339
:
622
2
Kurts, C., W. R. Heath, H. Kosaka, J. F. Miller, F. R. Carbone.
1998
. The peripheral deletion of autoreactive CD8+ T cells induced by cross-presentation of self-antigens involves signaling through CD95 (Fas, Apo-1).
J. Exp. Med.
188
:
415
3
Bertolino, P., W. R. Heath, C. L. Hardy, G. Morahan, J. F. A. P. Miller.
1995
. Peripheral deletion of autoreactive CD8+ T cells in transgenic mice expressing H-2Kb in the liver.
Eur. J. Immunol.
25
:
1932
4
Fields, L. E., D. Y. Loh.
1992
. Organ injury associated with extrathymic induction of immune tolerance in doubly transgenic mice.
Proc. Natl. Acad. Sci. USA
89
:
5730
5
Schönrich, G., U. Kalinke, F. Momburg, M. Malissen, A.-M. Schimtt-Verhulst, B. Malissen, G. J. Hammerling, B. Arnold.
1991
. Down-regulation of T cell receptors on self-reactive T cells as a novel mechanism for extrathymic tolerance induction.
Cell
65
:
293
6
Schönrich, G., F. Momburg, M. Malissen, A. M. Schmitt-Verhulst, B. Malissen, G. J. Hammerling, B. Arnold.
1992
. Distinct mechanisms of extrathymic T cell tolerance due to differential expression of self antigen.
Int. Immunol.
4
:
581
7
Scott, B., R. Liblau, S. Degermann, L. A. Marconi, L. Ogata, A. J. Caton, H. O. McDevitt, D. Lo.
1994
. A role for non-MHC genetic polymorphism in susceptibility to spontaneous autoimmunity.
Immunity
1
:
73
8
Boussiotis, V. A., G. J. Freeman, J. G. Gribben, L. M. Nadler.
1996
. The role of B7-1/B7-2:CD28/CLTA-4 pathways in the prevention of anergy, induction of productive immunity and down-regulation of the immune response.
Immunol. Rev.
153
:
5
9
Guerder, S., R. A. Flavell.
1995
. Costimulation in tolerance and autoimmunity.
Int. Rev. Immunol.
13
:
135
10
Jain, J., C. Loh, A. Rao.
1995
. Transcriptional regulation of the IL-2 gene.
Curr. Opin. Immunol.
7
:
333
11
Rincòn, M., R. A. Flavell.
1994
. AP-1 transcriptional activity requires both T-cell receptor-mediated and co-stimulatory signals in primary T lymphocytes.
EMBO J.
13
:
4370
12
Su, B., E. Jacinto, M. Hibi, T. Kalunki, M. Karin, Y. Ben-Neriah.
1994
. JNK is involved in signal integration during costimulation of T lymphocytes.
Cell
77
:
727
13
Wisdom, R., R. S. Johnson, C. Moore.
1999
. c-Jun regulates cell cycle progression and apoptosis by distinct mechanisms.
EMBO J.
18
:
188
14
Ghosh, S., M. J. May, E. B. Kopp.
1998
. NF-κB and Rel proteins: evolutionarily conserved mediators of immune responses.
Annu. Rev. Immunol.
16
:
225
15
Kontgen, F., R. J. Grumont, A. Strasser, D. Metcalf, R. Li, D. Tarlinton, S. Gerondakis.
1995
. Mice lacking the c-Rel proto-oncogene exhibit defects in lymphocyte proliferation, humoral immunity, and interleukin-2 expression.
Genes Dev.
9
:
1965
16
Doi, T. S., T. Takahashi, O. Taguchi, T. Azuma, Y. Obata.
1997
. NF-κB RelA-deficient lymphocytes: normal development of T cells and B cells, impaired production of IgA and IgG1 and reduced proliferative responses.
J. Exp. Med.
185
:
953
17
Beg, A. A., W. C. Sha, R. T. Bronson, S. Ghosh, D. Baltimore.
1995
. Embryonic lethality and liver degeneration in mice lacking the RelA component of NF-κB.
Nature
376
:
167
18
Beg, A. A., D. Baltimore.
1996
. An essential role for NF-κB in preventing TNF-α-induced cell death.
Science
274
:
782
19
Van Antwerp, D., S. Martin, T. Kafri, D. Green, I. Verma.
1996
. Suppression of TNF-α-induced apoptosis by NF-κB.
Science
274
:
787
20
Wang, C., M. Mayo, A. Baldwin, Jr.
1996
. TNF- and cancer therapy-induced apoptosis: potentiation by inhibition of NF-κB.
Science
274
:
784
21
Liu, Z.-G., H. Hsu, D. Goeddel, M. Karin.
1996
. Dissection of TNF receptor 1 effector functions: JNK activation is not linked to apoptosis while NF-κB activation prevents cell death.
Cell
87
:
565
22
Boothby, M., A. Mora, D. Schere, J. Brockman, D. Ballard.
1997
. Perturbation of the T lymphocyte lineage In transgenic mice expressing a constitutive repressor of nuclear factor (NF)-κB.
J. Exp. Med.
185
:
1897
23
Ferreira, V., N. Sidenius, N. Tarantino, P. Hubert, L. Chatenoud, F. Blasi, M. Korner.
1999
. In vivo inhibition of NF-κB in T-lineage cells leads to a dramatic decrease in cell proliferation and cytokine production and to increased cell apoptosis in response to mitogenic stimuli, but not to abnormal thymopoiesis.
J. Immunol.
162
:
6442
24
Hettmann, T., J. DiDonato, M. Karin, J. M. Leiden.
1999
. An essential role for nuclear factor κB in promoting double positive thymocyte apoptosis.
J. Exp. Med.
189
:
145
25
Kang, S.-M., B. Beverly, A.-C. Tran, K. Brorson, R. H. Schwartz, M. J. Lenardo.
1992
. Transactivation by AP-1 is a molecular target of T-cell clonal anergy.
Science
257
:
1134
26
Sundstedt, A., M. Sigvardsson, T. Leanderson, G. Hedlung, T. Kalland, M. Dohlsten.
1996
. In vivo anergized CD4+ T cells express perturbed AP-1 and NF-κB transcription factors.
Proc. Natl. Acad. Sci. USA
93
:
979
27
Sundstedt, A., M. Dohlsten.
1998
. In vivo anergized CD4+ T cells have defective expression and function of the activating protein-1 transcription factor.
J. Immunol.
161
:
5930
28
Voll, R. E., E. Jimi, R. J. Phillips, D. F. Barber, M. Rincon, A. C. Hayday, R. A. Flavell, S. Ghosh.
2000
. NF-κB activation by the pre-T cell receptor serves as a selective survival signal in T lymphocyte development.
Immunity
13
:
677
29
Yeoman, H., A. L. Mellor.
1992
. Tolerance and MHC restriction in transgenic mice expressing a MHC class I gene in erythroid cells.
Int. Immunol.
4
:
59
30
Lyons, A. B., C. R. Parish.
1994
. Determination of lymphocyte division by flow cytometry.
J. Immunol. Methods
171
:
131
31
Guerder, S., D. E. Picarella, P. S. Linsley, R. A. Flavell.
1994
. Costimulator B7-1 confers antigen-presenting-cell function to parenchymal tissue and in conjunction with tumor necrosis factor α leads to autoimmunity in transgenic mice.
Proc. Natl. Acad. Sci. USA
91
:
5138
32
Reiner, S. L., S. Zheng, D. B. Corry, R. M. Locksley.
1993
. Constructing polycompetitor cDNAs for quantitative PCR.
J. Immunol. Methods
165
:
37
33
Phillips, R. J., S. Ghosh.
1997
. Regulation of IκB β in WEHI 231 mature B cells.
Mol. Cell. Biol.
17
:
4390
34
Kearney, E., K. Pape, D. Loh, M. Jenkins.
1994
. Visualization of peptide-specific cell immunity and peripheral tolerance induction in vivo.
Immunity
1
:
327
35
Goldrath, A. W., M. J. Bevan.
1999
. Low-affinity ligands for the TCR drive proliferation of mature CD8+ T cells in lymphopenic hosts.
Immunity
11
:
183
36
Ernst, B., D. S. Lee, J. M. Chang, J. Sprent, C. D. Surh.
1999
. The peptide ligands mediating positive selection in the thymus control T cell survival and homeostatic proliferation in the periphery.
Immunity
11
:
173
37
Viret, C., F. S. Wong, C. A. Janeway, Jr.
1999
. Designing and maintaining the mature TCR repertoire: the continuum of self-peptide:self-MHC complex recognition.
Immunity
10
:
559
38
Sprent, J..
1997
. Immunological memory.
Curr. Opin. Immunol.
9
:
371
39
Dahl, A. M., C. Klein, P. G. Andres, C. A. London, M. P. Lodge, R. C. Mulligan, A. K. Abbas.
2000
. Expression of bcl-XL restores cell survival, but not proliferation off effector differentiation, in CD28-deficient T lymphocytes.
J. Exp. Med.
191
:
2031
40
Petschner, F., C. Zimmerman, A. Strasser, D. Grillot, G. Nunez, H. Pircher.
1998
. Constitutive expression of Bcl-xL or Bcl-2 prevents peptide antigen-induced T cell deletion but does not influence T cell homeostasis after a viral infection.
Eur. J. Immunol.
28
:
560
41
Rincon, M., B. Dérijard, C.-W. Chow, R. Davis, R. Flavell.
1997
. Reprogramming the signalling requirement for AP-1 (activator protein-1) activation during differentiation of precursor CD4+ T-cells into effector Th1 and Th2 cells.
Genes Funct.
1
:
51
42
Fraser, J. H., M. Rincon, K. D. McCoy, G. Le Gros.
1999
. CTLA4 ligation attenuates AP-1, NFAT and NF-κB activity in activated T cells.
Eur. J. Immunol.
29
:
838
43
Millet, I., R. J. Phillips, R. S. Sherwin, S. Ghosh, R. E. Voll, R. A. Flavell, A. Vignery, M. Rincon.
2000
. Inhibition of NF-κB activity and enhancement of apoptosis by the neuropeptide calcitonin gene-related peptide.
J. Biol. Chem.
275
:
15114
44
Auphan, N., S. Ghosh, R. A. Flavell, A. M. Schmitt-Verhulst.
1999
. Differential requirements for NF-κB and AP-1 trans-activation in response to minimal TCR engagement by a partial agonist in naive CD8 T cells.
J. Immunol.
163
:
5219
45
Zhong, H., R. E. Voll, S. Ghosh.
1998
. Phosphorylation of NF-κB p65 by PKA stimulates transcriptional activity by promoting a novel bivalent interaction with the coactivator CBP/p300.
Mol. Cell
1
:
661
46
Levkau, B., M. Scatena, C. M. Giachelli, R. Ross, E. W. Raines.
1999
. Apoptosis overrides survival signals through a caspase-mediated dominant-negative NF-κB loop.
Nat. Cell Biol.
1
:
227
47
Ravi, R., A. Bedi, E. J. Fuchs.
1998
. CD95 (Fas)-induced caspase-mediated proteolysis of NF-κB.
Cancer Res.
58
:
882
48
Bertolino, P., M. C. Trescol-Biemont, C. Rabourdin-Combe.
1998
. Hepatocytes induce functional activation of naive CD8+ T lymphocytes but fail to promote survival.
Eur. J. Immunol.
28
:
221
49
Bertolino, P., M. C. Trescol-Biemont, J. Thomas, B. F. de St. Groth, M. Pihlgren, J. Marvel, C. Rabourdin-Combe.
1999
. Death by neglect as a deletional mechanism of peripheral tolerance.
Int. Immunol.
11
:
1225
50
Tanchot, C., F. A. Lemonnier, B. Perarnau, A. A. Freitas, B. Rocha.
1997
. Differential requirements for survival and proliferation of CD8 naive or memory T cells.
Science
276
:
2057
51
Kirberg, J., A. Berns, H. von Boehmer.
1997
. Peripheral T cell survival requires continual ligation of the T cell receptor to major histocompatibility complex-encoded molecules.
J. Exp. Med.
186
:
1269
52
Siebenlist, U., G. Franzoso, K. Brown.
1994
. Structure, regulation and function of NF-κB.
Annu. Rev. Cell Biol.
10
:
405
53
Weih, F., D. Carrasco, S. K. Durham, D. S. Barton, C. A. Rizzo, R. P. Ryseck, S. A. Lira, R. Bravo.
1995
. Multiorgan inflammation and hematopoietic abnormalities in mice with a targeted disruption of RelB, a member of the NF-κB/Rel family.
Cell
80
:
331
54
Carrasco, D., J. Cheng, A. Lewin, G. Warr, H. Yang, C. Rizzo, F. Rosas, C. Snapper, R. Bravo.
1998
. Multiple hemopoietic defects and lymphoid hyperplasia in mice lacking the transcriptional activation domain of the c-Rel protein.
J. Exp. Med.
187
:
973
55
Sha, W. C., H. C. Liou, E. I. Tuomanen, D. Baltimore.
1995
. Targeted disruption of the p50 subunit of NF-κB leads to multifocal defects in immune responses.
Cell
80
:
321
56
Snapper, C. M., P. Zelazowski, F. R. Rosas, M. R. Kehry, M. Tian, D. Baltimore, W. C. Sha.
1996
. B cells from p50/NF-κB knockout mice have selective defects in proliferation, differentiation, germ-line CH transcription, and Ig class switching.
J. Immunol.
156
:
183