Induction of thymocyte positive selection does not convey immediate resistance to negative selection

Abstract

The acquisition of functional competence represents a critical phase during intrathymic development of T cells. Thymocytes reaching this stage represent cells which have been positively selected on the basis of major histocompatibility complex reactivity, but which have also been purged of potentially autoreactive T-cell receptor specificities by negative selection. While the developmental window in which thymocytes are subjected to positive selection is now well defined, the precise developmental timing of negative selection, in relation to positive selection events, is less clear. Moreover, the underlying mechanism allowing single-positive thymocytes to respond to T-cell receptor ligation by activation rather than death, remains controversial. Here we have analysed the developmental timing of negative selection in relation to positive selection, using measurement of thymocyte susceptibility to dendritic cell presentation of the superantigen staphylococcal enterotoxin B (SEB). We show that thymocytes which have received initial positive selection signals, namely CD4⁺ CD8⁺ CD69⁺ thymocytes, like their CD4⁺ CD8⁺ CD69⁻ precursors, are susceptible to negative selection, indicating that induction of positive selection does not convey immediate resistance to negative selection. In contrast, newly generated CD4⁺ CD8⁻ CD69⁺ cells are not only resistant to deletion by SEB, but respond to SEB-mediated T-cell receptor-ligation by activation, indicating that the acquisition of functional competence occurs at the newly generated CD4⁺ CD8⁻ CD69⁺ stage. Finally, by using direct retroviral infection of primary CD4⁺ CD8⁺ thymocytes, we also show that Notch-1 activation in CD4⁺ CD8⁺ thymocytes does not correlate with, nor convey resistance to superantigen-mediated negative selection. Thus, our data suggest that although Notch-1 has been implicated in resistance to thymocyte apoptosis, the acquisition of resistance to negative selection occurs independently of Notch-1 signalling.

Introduction

Specific interactions between the αβT-cell receptor (TCR) and self-peptide–major histocompatibility complex (MHC) complexes represent the basis of an intrathymic screening mechanism which governs the development of a non-autoreactive peripheral T-cell repertoire. This mechanism is dependent upon the operation of both positive and negative selection events at defined stages of thymocyte development. The former selects those thymocytes bearing a TCR of useful specificity, capable of recognizing foreign antigen-derived peptide in the context of self-MHC molecules, while the latter removes thymocytes bearing a TCR with potential for autoreactivity to self-peptide–self-MHC complexes.¹ While these two events are both dependent upon ligation of the TCR by peptide–MHC complexes, these interactions lead to very different outcomes. Positive selection of thymocytes, which requires medium to low avidity TCR ligation, triggers further differentiation and maturation,^2,3 eventually generating a population of phenotypically and functionally mature cells ready for export to the peripheral T-cell pool. In contrast, thymocytes making high avidity TCR–MHC interactions are negatively selected and triggered to undergo apoptosis, thus preventing generation of functional T cells with autoreactive specificities.

The major population of thymocytes within the thymus is that of small cortical cells at the CD4⁺ CD8⁺ stage, and it is this population which is subjected to positive selection events. However, these cells are also targets of negative selection.^1,3 The developmental timing of negative selection in relation to positive selection is poorly defined, and it is not clear whether the window for negative selection encompasses a broad span of thymocyte development, or is more strictly limited to thymocytes at a specific stage of maturation.

An important change that is triggered by positive selection is the acquisition of increased resistance to apoptosis. Thus, TCR stimuli that lead to rapid apoptotic death in cortical thymocytes can result in activation in mature T cells.⁴ However, the molecular basis for the increased resistance to apoptosis triggered by positive selection remains unknown. It has previously been demonstrated that Notch-1 signalling plays a role in mediating resistance to the induction of apoptosis in thymocyte cell lines,^5,6 although a recent paper suggests that Notch-1 does not play an essential role in thymocyte positive selection events.⁷ With regard to negative selection, however, the role of Notch-1 remains unclear, since the latter report only studied the role of Notch-1 in resistance to dexamethasone-induced apoptosis, and so the relation of Notch-1 signalling to antigen-induced negative selection is still unknown. Thus, in this report we have studied the developmental timing of negative selection in relation to the induction of positive selection and activation of the Notch-1 signalling pathway. Using a system involving dendritic cell presentation of the superantigen staphylococcal enterotoxin B (SEB) we have investigated thymocyte susceptibility to negative selection at a number of phenotypic stages of development from the CD4⁺ CD8⁺ CD69⁻ stage onwards. Our data indicate that while prepositive selection CD4⁺ CD8⁺ CD69⁻ thymocytes, and also interestingly CD4⁺ CD8⁺ CD69⁺ thymocytes, are still susceptible to negative selection events, resistance to negative selection is acquired at the CD4⁺ CD8⁻ CD69⁺ stage. Moreover, this acquisition of resistance to deletion does not appear to correlate with activation of Notch-1 signalling, nor can constitutive activation of the Notch-1 signalling pathway overcome thymocyte susceptibility to negative selection. Collectively these data indicate that functional competence is achieved early in the CD4⁺ CD8⁻ phase, via a Notch-1-independent mechanism.

Materials and methods

Mice

BALB/c mice, Bcl-2 transgenic (tg) mice (under control of the p56lck promoter)⁸ and MHC-deficient mice (Taconic) were bred and maintained at the Biomedical Sciences Unit, University of Birmingham, UK. Adult mice (4–6 weeks) were used as a source of CD4⁺ CD8⁻ CD69⁺ HSA⁺ (heat stable antigen⁺) thymocytes, and mesenteric and inguinal lymph nodes, while thymocyte subpopulations were isolated from either BALB/c, Bcl-2 tg or MHC-deficient neonatal mice (0–2 days), or adult (4–6 weeks) BALB/c mice.

Antibodies and immunoconjugates

The following antibodies were coated onto anti-rat immunoglobulin G (IgG) or streptavidin-coated Dynabeads (Dynal, Wirral, UK) as appropriate: biotinylated anti-CD69 (clone H1.2f3, Pharmingen, San Diego, CA), anti-CD8 (clone YTS169.4, Seralab, Crawley Down, UK), anti-CD3 (clone KT-3, Serotec, Oxford, UK). Antibodies used for flow cytometric analysis were as follows: phycoerythrin (PE)-conjugated anti-CD4 (GK1.5, Pharmingen), fluorescein isothiocyanate (FITC)-conjugated anti-CD8 (clone 53-6.7, Pharmingen), FITC-conjugated anti-BrdU (clone 3D4, Pharmingen), biotinylated anti-Vβ8 (clone F23.1, Pharmingen), biotinylated anti-CD25 (clone 7D4), biotinylated anti-CD24 (anti-HSA) (clone M1/69). Biotinylated antibodies were detected using a subsequent incubation in streptavidin allophycocyanin (APC) (Pharmingen).

Cell purification

Thymocytes

Methods to isolate thymocyte subsets have been described in detail elsewhere.^9,10 Briefly, pre-selection thymocytes were obtained from either MHC-deficient neonatal mice, which are halted at the CD4⁺ CD8⁺ stage and do not express CD69, due to a lack of exposure to positively selecting MHC molecules, or from Bcl-2 tg mice, as previously described.¹¹ Briefly, thymocytes from neonatal Bcl-2 tg mice were depleted of CD3⁺ cells using anti-CD3-coated beads. This was followed by selection of CD8⁺ cells using anti-CD8-coated beads, which were subsequently removed using Detachabead (Dynal). CD69⁺ cells were prepared by immunomagnetic selection from neonatal BALB/c mice by positive selection using anti-CD69-coated beads, which were subsequently removed by Detachabead (Dynal). CD4⁺ CD8⁺ CD69⁺ or CD4⁺ CD8⁻ CD69⁺ cells were obtained by selection of CD69⁺ cells, followed by either selection or depletion, respectively, of CD8⁺ cells using anti-CD8-coated beads. CD4⁺ CD8⁻ CD69⁺ HSA⁺ thymocytes were purified from adult BALB/c mice by isolation of CD4⁺ CD8⁻ CD69⁺ cells as described, and then selected for HSA⁺ cells by labelling with biotinylated anti-HSA, followed by immunomagnetic selection using streptavidin microbeads (Miltenyi Biotec, Bisley, UK).

Dendritic cells

Cells from mesenteric and inguinal lymph nodes were pooled and selected for CD11c⁺ cells, using anti-CD11c microbeads (Miltenyi Biotec). CD11c⁺ cells were further purified by depletion of contaminating B220⁺ cells, using precoated Mouse pan B (B220) Dynabeads (Dynal).

Deletion/stimulation assays

Isolated thymocyte populations were mixed together with isolated dendritic cells, at a ratio of 10:1, by centrifugation. The resultant cell pellet was transferred to the surface of a nucleopore filter in organ culture conditions, in the presence or absence of 10 µg/ml SEB (Toxin Technology, Sarasota, FL). Thymocyte proliferation in cultures was analysed simultaneously with expression of Vβ8. Thus cultures were pulsed with 5 µg/ml 5-bromo 2′-deoxyuridine (BrdU, Sigma Chemical Co., Poole, UK) for the final 24 hr of a 2-day culture period. Cultures were harvested, and cells were labelled for surface expression of Vβ8, with BrdU incorporation being detected as described by Tough and Sprent.¹²

Real-time polymerase chain reaction

Multiplex polymerase chain reaction (PCR) was performed in ABI PRISM 7700 Sequence Detection System (PE Applied Biosystems, Foster City, CA). Relative quantification of gene expression was achieved by using primer pairs to amplify the target genes in the presence of probes labelled with FAM reporter dyes and VIC reporter dye labelled probe for β-actin as endogenous control. Primer pairs and probes were designed using the Taq Man Probe and Primer Design computer programme and data was analysed by the ΔΔC_T method. Primer sequences were as follows: deltex-1: F: CCGTGGTGTGGAACGAGATT, R: ACGTTGTCTAGGTAGCTGGCGT, Probe: CCTCACTGGTCACGGCTACCCCG; β-actin: F: CGTGAAAAGATGACCCAGATCA, R: TGGTACGACCAGAGGCATACAG, Probe: TCAACACCCCAGCCATGTACGTAGCC.

Semi-quantitative PCR

Total RNA was extracted from cells using TRIzol (Life Technologies, Paisley, UK) according to the manufacturer's instructions. RNA samples were treated with RNase-free Dnase1 (Pharmacia Biotech, Uppsala, Sweden) to remove any contaminating genomic DNA. This was followed by reverse transcription (RT). The RT-PCR was performed, with β-actin as a housekeeping gene. The sequences for β-actin and Deltex primers are as follows: β-actin: sense, 5′-GTTACCAACTGGGACGACA-3′; antisense, 5′-TGGCCATCTCCTGCTCGAA-3′: Deltex: sense, 5′-CACTGGCCCTGTCCACCCAGCCTTGGCAGG-3′; antisense, 5′-GAGGCATGTGCCAGGCTAGAGGCAAGGCAA-3′.

RT-PCR involved analysis of samples every three cycles from 18 to 33 (β-actin) or every four cycles from 24 to 44 (Deltex). PCR products were analysed by ethidium bromide agarose gel electrophoresis and identified by fragment size.

Retroviral infection of CD4⁺ CD8⁺ CD69⁻ thymocytes

The cDNA encoding the intracellular domain of human Notch-1, ICN1, cloned into Mig RI, a vector that permits co-expression of cloned cDNA and green fluorescent protein (GFP) from single bicistronic message, was a generous gift from W.S. Pear (Department of Pathology and Laboratory Medicine, University of Pennsylvania Medical Center, Philadelphia, PA). Plasmid DNA was transfected into the ΦNX-A packaging cell line (American Type Culture Collection, Rockville, MD) by electroporation. Retrovirus-containing supernatant was collected following 8-hr incubation of the transfected packaging cells at 32°. Purified CD4⁺ CD8⁺ thymocytes from Bcl-2 tg mice were infected by re-suspending the thymocytes at 7 × 10⁶/ml in retrovirus supernatant with 8 µg/ml polybrene followed by centrifugation for 60 min at 1000 g at room temperature and overnight incubation at 37°.

Results

Positive selection does not convey immediate resistance to negative selection

Conflicting evidence exists regarding whether thymocytes which have initiated positive selection are still susceptible to deletion.^9,13–15 Thus it is still unclear whether the induction of positive selection conveys immediate resistance to negative selection, or whether a window exists following the initiation of positive selection where thymocytes are still susceptible to deletion. To address this controversy, in an initial set of experiments we compared the responses of a number of freshly isolated thymocyte populations to dendritic cell presentation of the bacterial superantigen SEB, a well-defined model of negative selection.¹⁶ To study a range of thymocyte populations which encompass positive and negative selection events, we isolated preselection CD4⁺ CD8⁺ cells known to be at a prepositive selection stage by virtue of a CD69⁻ phenotype. We also obtained purified CD69⁺ thymocytes, in which CD69 expression is an indicator of the initiation of positive selection events,^17,18 and CD4⁺ CD8⁻ CD69⁺ cells, representing a population of newly generated CD4⁺ CD8⁻ thymocytes.

MHC class II presentation of SEB is known to target specifically, amongst others, TCRs using the Vβ8 segment in their β-chain. Thus we have analysed here the SEB-reactive Vβ8⁺ fraction of thymocytes. As expected, CD4⁺ CD8⁺ CD69⁻ thymocytes cultured overnight with dendritic cells in the presence of 10 µg/ml SEB showed a specific reduction in numbers of Vβ8⁺ CD4⁺ CD8⁺ cells (Fig. 1a, b). In fact, approximately 45% of Vβ8⁺ CD4⁺ CD8⁺ thymocytes were deleted in comparison to the number of Vβ8⁺CD4⁺ CD8⁺ cells recovered from culture in the absence of SEB (Fig. 1c). Interestingly, culture of CD69⁺ thymocytes in the same culture system also showed evidence of deletion of Vβ8⁺ CD4⁺ CD8⁺ cells, with approximately 55% of Vβ8⁺ CD4⁺ CD8⁺ CD69⁺ cells deleted by SEB (Fig. 1f). In contrast, however, CD4⁺ CD8⁻ CD69⁺ thymocytes, representing newly generated T cells and the direct descendants of CD4⁺ CD8⁺ CD69⁺ cells, were not deleted by SEB, with similar numbers of Vβ8⁺ CD4⁺ CD8⁻ CD69⁺ thymocytes recovered from cultures whether SEB was absent or present in the culture medium (Fig. 1d–f). It is important to note here that we have also looked at the response of a thymocyte population expressing a Vβ which does not interact with SEB. It has previously been demonstrated that Vβ2⁺ thymocytes are not SEB-reactive,¹⁹ and indeed, in these experiments, while culture with SEB led to deletion of CD4⁺ CD8⁺ thymocytes expressing Vβ8, CD4⁺ CD8⁺ Vβ2⁺ thymocytes remained unaffected by culture in the presence of SEB (data not shown).

Figure 1.

Induction of positive selection does not convey immediate resistance to negative selection. CD4⁺ CD8⁺ CD69⁻ thymocytes (a, b) from MHC-deficient neonatal mice, or CD69⁺ (d, e) thymocytes isolated from wild-type neonatal mice were placed in culture with freshly isolated dendritic cells at a ratio of 10:1, in the absence (a, d) or presence (b, e) of SEB at 10 µg/ml. Following a 24-hr culture period, thymocytes were harvested and analysed for expression of CD4, CD8 and Vβ8. In the experiment shown, 1 × 10⁶ thymocytes were placed in each culture, with cell recoveries as follows: 9 × 10⁵, 8 × 10⁵, 4·7 × 10⁵ and 5·5 × 10⁵ for cultures (a), (b), (d) and (e), respectively. From this the total number of Vβ8⁺ CD4⁺ CD8⁺ cells was calculated, as shown for the CD4⁺ CD8⁺ CD69⁻ cultures in (c), with 1·1 × 10⁵ and 6 × 10⁴ cells recovered from culture in the absence or presence of SEB (f), respectively. Similarly, from the CD69⁺ cultures, 4·6 × 10⁴ and 2·1 × 10⁴ Vβ8⁺ CD4⁺ CD8⁺ thymocytes were recovered from culture in the absence or presence of SEB Similar results were obtained in three separate experiments.

Since Vβ8⁺ CD4⁺ CD8⁻ CD69⁺ thymocytes showed no evidence of deletion following overnight exposure to dendritic cell presentation of SEB, we analysed Vβ8⁺ CD4⁺ CD8⁻ CD69⁺ thymocytes cultured in the absence or presence of SEB for evidence of activation. Interestingly, following a 24-hr culture period, we found that Vβ8⁺ CD4⁺ CD8⁻ CD69⁺ thymocytes recovered from cultures in which SEB was present showed significant induction of CD25 expression, with 81% of cells having a CD25⁺ phenotype (Fig. 2b), in contrast to the 4% of Vβ8⁺CD4⁺ CD8⁻ CD69⁺ thymocytes found to express CD25 following culture in the absence of SEB (Fig. 2a). This evidence for activation induced by dendritic cell presentation of SEB was underlined further by assessing proliferation in Vβ8⁺ CD4⁺ CD8⁻ CD69⁺ thymocytes recovered from cultures after a 48-hr culture period. Thus, 80% of Vβ8⁺ CD4⁺ CD8⁻ CD69⁺ thymocytes cultured in the presence of SEB were found to be proliferating, as assessed by BrdU incorporation (Fig. 2d), while Vβ8⁺ CD4⁺ CD8⁻ CD69⁺ thymocytes recovered from culture in the absence of SEB were found to be out of cell cycle (Fig. 2c).

Figure 2.

CD4⁺ CD8⁻ CD69⁺ thymocytes show evidence of activation in response to dendritic cell presentation of SEB. CD4⁺ CD8⁻ CD69⁺ thymocytes from wild-type neonatal mice were placed in culture with freshly isolated dendritic cells at a ratio of 10:1 in the absence (a, c) or presence (b, d) of SEB. Following a 24-hr culture period (a, b), thymocytes were harvested and analysed for expression of Vβ8 and CD25. In the experiment shown here, 1 × 10⁶ thymocytes were placed in each culture, with recoveries of 4 × 10⁵ and 3·4 × 10⁵ in the absence or presence of SEB, respectively. Of these, 1 × 10⁴ (a) and 3·3 × 10⁴ (b) were of a Vβ8⁺ CD25⁺ phenotype. Similarly, the remaining cultures (c, d) were pulsed with BrdU after 24 hr, harvested after a further 24-hr culture period, and analysed for BrdU incorporation and Vβ8 expression. Total cell recoveries were 4·3 × 10⁵ (c) and 5·8 × 10⁵ (d), with the number of proliferating Vβ8⁺ CD4⁺ CD8⁻ cells at 0·4 × 10⁴ and 9·2 × 10⁴ in the absence or presence of SEB, respectively. Similar results were obtained from three separate experiments.

In the adult thymus, it has been postulated that newly generated CD4⁺ CD8⁻ HSA^hi medullary thymocytes are still sensitive to negative selection.¹³ However, these results are based upon the induction of negative selection by injection of anti-TCR antibodies, and the system we use here, of dendritic cell presentation of SEB is perhaps more physiologically relevant. Therefore, we analysed HSA expression on our sorted CD4⁺ CD8⁻ CD69⁺ neonatal thymocytes, and compared this with HSA expression on CD4⁺ CD8⁻ CD69⁺ adult thymocytes. Interestingly, we found that CD4⁺ CD8⁻ CD69⁺ neonatal thymocytes were uniform in their high expression of HSA (Fig. 3a). In contrast, CD4⁺ CD8⁻ CD69⁺ adult thymocytes were heterogeneous for HSA, with 20% of cells being of an HSA⁻ phenotype (Fig. 3b). We therefore compared the response of neonatal and adult CD4⁺ CD8⁻ CD69⁺ HSA^hi thymocytes to dendritic cell presentation of SEB. Interestingly, we found that CD4⁺ CD8⁻ CD69⁺ HSA^hi thymocytes isolated from either neonatal or adult mice both responded to stimulation with SEB not by deletion but by activation. Indeed, the majority of both neonatal and adult CD4⁺ CD8⁻ CD69⁺ HSA^hi thymocytes were found to express CD25 after a 24-hr culture period (Fig. 3c), and showed evidence of proliferation by BrdU incorporation after a 48-hr culture period (Fig. 3d).

Figure 3.

CD4⁺ CD8⁻ CD69⁺ HSA^hi thymocytes from neonatal and adult mice show evidence of activation by SEB. CD69⁺ neonatal (a) or adult (b) thymocytes were isolated from wild-type mice, and analysed for expression of CD4, CD8 and HSA. Gating on CD4⁺ CD8⁻ thymocytes revealed the HSA profile of CD4⁺ CD8⁻ CD69⁺ cells from neonatal and adult thymus (a and b, respectively). Neonatal CD4⁺ CD8⁻ CD69⁺ thymocytes (all HSA^hi) and adult CD4⁺ CD8⁻ CD69⁺ thymocytes sorted for expression of HSA, were placed in culture with dendritic cells, at a ratio of 10:1 for 2 days, with BrdU added to cultures after 24 hr. Thymocytes were subsequently harvested and analysed for expression of Vβ8 and CD25 (c), or Vβ8 and BrdU incorporation (d). In the experiment shown, 1 × 10⁶ thymocytes were placed into each culture, with recoveries of 4·3 × 10⁵, 5·8 × 10⁵ (neonatal, − or + SEB), 2·8 × 10⁵ and 4·1 × 10⁵ (Adult, − or + SEB). The number of Vβ8⁺ CD25⁺ cells (c) or Vβ8⁺ BrdU⁺ cells (d) were then calculated. This experiment is representative of three separate experiments.

Notch-1 signalling does not convey resistance to SEB induced negative selection

Mechanisms regulating negative selection and acquisition of resistance to deletion are poorly defined. It has been suggested that Notch-1 signalling plays a role in conveying resistance to apoptosis induced by dexamethasone, or by treatment with anti-TCR antibodies.^5,6 To see if Notch-1 activation correlates with phases of susceptibility to negative selection as defined above, we looked for evidence of Notch-1 activation in thymocyte subsets. As before, we isolated CD4⁺ CD8⁺ CD69⁻, CD4⁺ CD8⁺ CD69⁺, and CD4⁺ CD8⁻ CD69⁺ thymocytes from neonatal mice, and used expression of Deltex, a downstream signalling molecule in the Notch-1 signalling pathway, as evidence of Notch-1 activation.⁶ Deltex expression within these populations was analysed by real-time PCR to allow comparative analysis of mRNA (Fig. 4). While no evidence of Deltex mRNA was found in CD4⁺ CD8⁺ CD69⁻ thymocytes, which is in agreement with a previous report,⁶ we find that as positive selection progressed through the CD4⁺ CD8⁺ CD69⁺ stage to a CD4⁺ CD8⁻ CD69⁺ phenotype, an increase in Deltex expression was observed. Thus, these data indicate that the Notch-1 signalling pathway is activated upon the induction of positive selection, prior to the acquisition of resistance to negative selection, and persists into the post-positive selection CD4⁺ CD8⁻ CD69⁺ stage.

Figure 4.

Measurement of Notch-1 activation in thymocyte subsets by analysis of Deltex expression. Pre-positive selection CD4⁺ CD8⁺ CD69⁻ thymocytes were isolated from MHC-deficient neonatal mice, and populations of CD4⁺ CD8⁺ CD69⁺ or CD4⁺ CD8⁻ CD69⁺ cells were prepared from neonatal BALB/c mice, as described. Real-time PCR analysis was performed to allow relative quantitation of Deltex mRNA levels.

To determine whether there is a causal relationship between Notch-1 signalling and resistance to negative selection, we isolated a population of preselection CD4⁺ CD8⁺ CD69⁻ thymocytes and retrovirally infected them with a GFP-associated retroviral vector containing cDNA for the intracellular domain of human Notch-1,²⁰ resulting in a constitutively active form of Notch-1 (IC-Notch-1). Due to the lengthy manipulation procedure required prior to incorporation into deletion assays, CD4⁺ CD8⁺ CD69⁻ thymocytes were, for these latter experiments, isolated from Bcl-2 tg mice, thymocytes from which have a longer lifespan than wild-type thymocytes. Thus, Bcl-2 tg CD4⁺ CD8⁺ CD69⁻ thymocytes were infected with control (GFP only) or IC-Notch-1-containing vectors, and placed in culture with dendritic cells in the absence or presence of SEB, to see if constitutive activation of the Notch-1 signalling pathway affected their susceptibility to superantigen-mediated negative selection. Detection of GFP by flow cytometry enabled us to analyse specifically those thymocytes which had been successfully infected with the relevant virus (Fig. 5a,b), revealing an infection rate of 20–30%. Thus, thymocytes infected with the control GFP vector or IC-Notch-1 vector were cultured with dendritic cells in the absence or presence of SEB. Thymocytes used in these cultures required exposure to SEB for a longer culture period than wild-type thymocytes in order to see significant SEB-mediated deletion in the Vβ8⁺ population, perhaps as a result of expression of the Bcl-2 transgene. However, following a 2-day culture period, subsequent analysis within the GFP⁺ population revealed that Vβ8⁺ thymocytes infected with the control GFP virus showed evidence of deletion in response to dendritic cell presentation of SEB (Fig. 5c). This level of deletion was similar to that observed in non-infected CD4⁺ CD8⁺ CD69⁻ thymocytes in earlier experiments (Fig. 1c), measured following a 1-day culture period, thus indicating that retroviral infection, per se, did not influence this assay, although total cell recovery following a 2-day culture period (Fig. 5) was lower than the total cell recovery after 1 day in culture (Fig. 1). This is likely to be due to the limited lifespan of the CD4⁺ CD8⁺ thymocyte input population, but an increase in non-specific thymocyte death by an increased culture period is unlikely to mask any SEB-specific cell death, since background cell death does not affect the relative proportions of different Vβ subsets (data not shown).

Figure 5.

Expression of a constitutively active form of Notch-1 does not protect CD4⁺ CD8⁺ thymocytes from superantigen-mediated cell death. CD4⁺ CD8⁺ CD69⁻ thymocytes from Bcl-2 tg neonatal mice, giving them an extended lifespan to allow for complex and lengthy manipulation procedures, were retrovirally transfected with a GFP-associated vector containing cDNA for IC-Notch-1 (b) or a control GFP-associated vector (a). After overnight incubation with the retrovirus, control infected and IC-Notch-1 infected CD4⁺ CD8⁺ cells were analysed for GFP expression by flow cytometry (a, b) and for expression of Deltex by semiquantitative PCR (d), and placed in culture with dendritic cells, at a ratio of 10:1, in the absence or presence of SEB at 10 µg/ml. Following a 48-hr culture period, thymocytes were harvested and analysed for expression of GFP and Vβ8. 1 × 10⁶ CD4⁺ CD8⁺ thymocytes were placed in each culture, with recoveries of 3·1 × 10⁵, 1·9 × 10⁵ cells (control virus, without and with SEB, respectively), 2·2 × 10⁵ and 1·5 × 10⁵ (IC-Notch virus without and with SEB, respectively). Thus, this enabled calculation of numbers of transfected (GFP⁺) Vβ8⁺ cells for each culture

Having established that retroviral infection of thymocytes did not influence this deletion assay, we looked at thymocytes which had been infected with the IC-Notch-1 vector. Surprisingly, we found that thymocytes infected with the GFP/IC-Notch-1 virus also showed considerable evidence of deletion following culture with SEB, with 58% of Vβ8⁺ GFP/IC-Notch-1 infected cells being deleted as compared to the 42% of Vβ8⁺ GFP control cells (Fig. 5c). From these results it therefore cannot be said that the protective effects of the Bcl-2 transgene in any way obscured protection conveyed by IC-Notch-1, since we saw specific deletion of Vβ8⁺ thymocytes by SEB even following infection of thymocytes with IC-Notch-1. Importantly, using Deltex expression as a marker of Notch-1 signalling,⁶ Fig. 5(d) shows that infection of primary CD4⁺ CD8⁺ thymocytes with the IC-Notch-1 construct leads to up-regulation of Deltex expression, indicative of activation of the Notch-1 signalling pathway by IC-Notch-1. Thus the lack of resistance to deletion observed in IC-Notch-1 infected cells is not merely attributable to the lack of activation of Notch-1 signalling. Collectively then, these data suggest that introduction of IC-Notch-1 into primary CD4⁺ CD8⁺ thymocytes is capable of activating the Notch-1 signalling cascade, but fails to convey resistance to superantigen-mediated negative selection.

Discussion

In this study we have addressed mechanisms regulating the transition from a functionally immature thymocyte to a functionally mature T cell. We have shown firstly that induction of positive selection does not convey immediate resistance to negative selection. Moreover, thymocytes which have initiated positive selection as identified by a CD69⁺ phenotype, are also known to have up-regulated levels of TCR expression. Therefore the continued susceptibility of CD4⁺ CD8⁺ CD69⁺ thymocytes to deletion at a stage where they have increased avidity for TCR–MHC interactions represents a biologically important point in development for the screening and removal of potentially autoreactive T cells. Interestingly, our results also indicate that, once thymocytes reach the CD4⁺ CD8⁻ CD69⁺ stage, not only are they resistant to deletion, but they are capable of responding to the same stimulus by activation, as shown by CD25 expression and by BrdU incorporation. This is in contrast to previous observations by Kishimoto and Sprent,¹³ who found that medullary CD4⁺ CD8⁻ thymocytes, from adult mice, expressing a semi-mature HSA^hi phenotype are still susceptible to apoptosis induced by anti-TCR antibody. However, the discrepancy between these and our data cannot be attributed to differences in the adult and neonatal thymocyte populations, since we show here that CD4⁺ CD8⁻ HSA^hi thymocytes from both adult and neonatal mice respond to dendritic cell presentation of SEB by activation. Thus, it may be that variation in TCR stimuli, such as between superantigen and anti-TCR antibody, causes activation of different responses.

By analysing activation of Notch-1, a molecule purported to play a role in resistance to negative selection,⁶ we have also shown that activation of Notch-1 in CD4⁺ CD8⁺ thymocytes does not correlate with, nor convey resistance to, negative selection. Thus, our results suggest that although Notch-1 has been shown to convey resistance to a variety of triggers of apoptosis, including non-TCR-mediated apoptosis induced by steroid treatment,⁶ it is not the sole factor in regulating the change in susceptibility of thymocytes to TCR-mediated negative selection signals. On a similar point, the differences between our results using primary thymocytes and superantigen, and those reported by Jehn et al.⁵ using thymocyte lines with anti-TCR antibody may additionally reflect intrinsic differences between cell lines and primary thymocytes and methods of TCR stimulation. Indeed, these discrepancies may indicate that whilst Notch-1 signalling can confer resistance to apoptosis induced by anti-TCR antibodies alone, it is insufficient to overcome the more physiological stimulus of TCR ligation in association with co-stimulatory signals provided by professional antigen-presenting cells, as studied here. In summary, our results show that there is a developmental overlap between positive and negative selection events, with resistance to negative selection being acquired at the CD4⁺ CD8⁻ CD69⁺ stage but not immediately following the initiation of positive selection. Moreover, the acquisition of this resistance to deletion appears to occur independently of Notch-1 signalling, suggesting that additional, and as yet undefined, cell interactions may be required for this process.