Heterogeneity in P-glycoprotein (multidrug resistance) activity among murine peripheral T cells: Correlation with surface phenotype and effector function

P-glycoprotein (P-gly) is the transmembrane efflux pump responsible for multidrug resistance in tumor cells. Functional P-gly activity can be conveniently assessed microfluorometrically using the fluorescent dye rhodamine 123 (Rh123), which is an artificial substrate for the P-gly transporter. Here we assess P-gly activity in subsets of mouse peripheral T lymphocytes using the Rh123 efflux assay. Our data indicate that virtually all CD8⁺ cells extrude Rh123 efficiently, whereas only a subset of CD4⁺ cells exhibit P-gly activity. Correlation of P-gly activity in CD4⁺ cells with the expression of a panel of surface markers revealed that cells bearing an “activated/memory” phenotype (CD45RB^?, CD44^hi, CD62L^?, CD25⁺, CD69⁺) were exclusively found in the fraction that can extrude Rh123. In contrast “naive” phenotype CD4⁺ cells (CD45RB⁺, CD44^lo, CD62L⁺, CD25^?, CD69^?) could be further subdivided into two major subsets based on P-gly activity. In functional studies of sorted cell populations the Rh123-extruding subset of “naive” CD4⁺ cells proliferated more strongly and secreted higher levels of interleukin (IL)-2 than its Rh123-retaining counterpart when activated by a variety of polyclonal stimuli. Furthermore, this subset produced detectable levels of interferon (IFN)-γ upon stimulation but no IL-4 or IL-10. As expected, the Rh123-retaining “naive” subset produced only IL-2 after stimulation, whereas the “memory” subset produced IFN-γ, IL-4 and IL-10 in addition to low levels of IL-2. Collectively, our data indicate that P-gly activity is a novel parameter that can be used to distinguish a subset of “preactivated” CD4⁺ cells that would be considered as naive on the basis of their surface phenotype.     

Downregulated Expression of Ly-6-ThB on Developing T Cells Marks CD4+CD8+ Subset Undergoing Selection in the Thymus

Interaction of TCRs on CD4⁺CD8⁺ immature T cell with MHC-peptide complexes on stromal cells is required for positive and negative selection in the thymus. Identification and characterization of a subpopulation of CD4⁺CD8⁺ thymocytes undergoing selection in the thymus will aid in understanding the mechanisms underlying lineage commitment and thymic selection. Herein, we describe the expression of Ly-6 ThB on developing thymocytes. The majority of CD4⁺CD8⁺ thymocytes express Ly-6 ThB at high levels. Its expression is downregulated in a subset of CD4⁺CD8⁺ thymocytes as well as in mature CD4⁺CD8^- and CD4^-CD8⁺ T cells. More importantly, interaction of TCR/coreceptor with the self-MHC-peptide contributes to the downregulation of ThB expression on developing thymocytes. These findings indicate that downregulation of ThB on CD4⁺CD8⁺ thymocytes identifies a unique subset (CD4+CD8+ThB^neg–low) of thymocytes that has received the initial signals for thymic selection but have not yet downregulated the CD4 and CD8 cell surface expression. In addition, these results also indicate that a high frequency (Ÿ20–40%) of CD4⁺CD8⁺ immature thymocytes receive these initial signals during thymic selection.     

Positive selection of CD4+ T cells by TCR-specific antibodies requires low valency TCR cross-linking: implications for repertoire selection in the thymus

The developmental fate of immature CD4⁺ 8⁺ thymocytes is determined by intrathymic signals transduced by surface TCR complexes. In particular, TCR signals are required for immature CD4⁺ 8⁺ thymocytes to further differentiate into CD4⁺ 8⁻ or CD4⁻ 8⁺ T cells, a process referred to as positive selection. It is generally thought that positive selection results from low affinity TCR interactions with self antigens which engage the relatively few surface TCR complexes that are on immature CD4⁺ 8⁺ thymocytes. However, we now demonstrate with TCR-specific antibodies that positive selection of CD4⁺ T cells requires low valency cross-linking of surface TCR complexes on immature thymocytes. That is, positive selection signals are only generated within a narrow range of TCR cross-linking: cross-linking either too few or too many surface TCR complexes fails to signal positive selection. We interpret these results as indicating that positive selection of CD4⁺ T cells is not signaled by low affinity TCR interactions per se, but rather can be signaled by any combination of TCR affinity and ligand density that induces low valency TCR cross-linking on immature thymocytes.     

Pertussis toxin can replace T cell receptor signals that induce positive selection of CD8 T cells

CD4⁺ helper T lymphocytes and CD8⁺ killer T lymphocytes are both generated in the thymus from common precursor cells expressing CD4 and CD8. The development of immature CD4⁺ CD8⁺ thymocytes into mature ‘single-positive’ T cells requires T cell antigen-receptor (TCR)-mediated positive selection signals. Although it is known that the recognition specificity of TCR expressed by CD4⁺ CD8⁺ thymocytes determines their fate to become either CD4⁺ or CD8⁺ T cells, the molecular signals that direct precursor thymocytes to become CD4⁺ and CD8⁺ T cells are unclear. By using ZAP-70^? mutant thymus organ cultures in which T cell development is arrested at the CD4⁺ CD8⁺ thymocyte stage, the present study shows that distinct biochemical treatments can selectively restore the generation of mature CD4⁺ and CD8⁺ T cells, bypassing TCR-induced positive selection signals. The combination of phorbol ester and ionomycin selectively restores the generation of CD4⁺CD8^? TCR^high cells consistent with previous results. On the other hand, we find that the generation of CD4^? CD8⁺ TCR^high cells is selectively induced by pertussis toxin. Interestingly, the signals generated by pertussis toxin, which increase Notch expression, can dominate the signals by phorbol ester and ionomycin, steering thymocyte development to CD8 lineage. These results indicate that distinct biochemical signals replace TCR signals that selectively induce positive selection of CD4⁺ and CD8⁺ T cells, and that biochemical treatment can manipulate the development and choice of CD4⁺ and CD8⁺ T cells.     

Antigen-specific and nonspecific deletion of immature cortical thymocytes caused by antigen injection

Analysis of antigen-induced negative selection of thymocytes in T cell receptor (TCR)-transgenic mice is complicated by the presence of an antigen-responsive peripheral T cell compartment. Our experiments address the question of whether and how peripheral T cell activation can affect immature thymocytes. Following three daily injections of peptide antigen into mice expressing a peptide-specific transgenic TCR and deficient for TAP1, we and others have found profound deletion of the CD4⁺CD8⁺ (DP) thymocyte subset. However, our work shows that even though mature CD8⁺ T cells are inefficiently selected in TAP1-deficient mice, there was a striking degree of peripheral expansion and activation of CD8⁺ peripheral T cells. Furthermore, when cells from TCR-transgenic mice were adoptively transferred, we found that deletion of non-transgenic DP thymocytes occurred in Thy-1-congenic and even more efficiently in TAP1-deficient recipients after repeated peptide injection resulting in peripheral T cell activation. In the adoptive transfer experiments the degree of deletion of immature bystander thymocytes was decreased upon blocking of TNF. These data show that deletion of DP thymocytes can result from excessive peripheral T cell activation and identify TNF as an important effector molecule for this process. When steps are taken to avoid peripheral T cell activation, peptide antigen can induce TCR-mediated thymocyte deletion, presumably in the thymus cortex, since injection of TAP1-deficient TCR-transgenic mice resulted in deletion of immature DP thymocytes prior to detectable peripheral T cell expansion and activation. This effect was not blocked by inhibiting tumor necrosis factor activity. In addition, DP depletion was seen in the absence of peripheral T cell activation when antibody-mediated depletion of CD8⁺ T cells was performed. Our work clearly shows that two mechanisms for deletion of DP thymocytes exist: deletion induced by antigen presentation in the thymus and deletion as a consequence of repeated stimulation of mature peripheral T cells.     

Separately expressed T cell receptor α and β chain transgenes exert opposite effects on T cell differentiation and neoplastic transformation

Two aspects of T cell differentiation in T cell receptor (TCR)-transgenic mice, the generation of an unusual population of CD4^?CD8^?TCR⁺ thymocytes and the absence of γδ cells, have been the focus of extensive investigation. To examine the basis for these phenomena, we investigated the effects of separate expression of a transgenic TCR α chain and a transgenic TCR β chain on thymocyte differentiation. Our data indicate that expression of a transgenic TCR α chain causes thymocytes to differentiate into a CD4^?CD8^?TCR⁺ lineage at an early developmental stage, depleting the number of thymocytes that differentiate into the αβ lineage. Surprisingly, expression of the TCR α chain transgene is also associated with the development of T cell lymphosarcoma. In contrast, expression of the transgenic TCR β chain causes immature T cells to accelerate differentiation into the αβ lineage and thus inhibits the generation of γδ cells. Our observations provide a model for understanding T cell differentiation in TCR-transgenic mice.     

Ultrastructural analysis of mouse thymocyte subpopulations

To understand the lineage relationship and to define morphological characteristics of each thymocyte subset, we have performed ultrastructural analysis of highly purified thymocyte subpopulations. By flow cytometry, five subpopulations were sorted based on the expression of CD4 and CD8 and on cell size (forward scatter): large and small CD4⁺8⁺, CD4^?8^?, CD4⁺8^?, and CD4^?8⁺ thymocytes. Small CD4⁺8⁺ thymocytes were the smallest among lymphoid cells, and had a round and smooth cell outline with condensed nuclei, the cytoplasm was scanty and the cell organelles were not developed, suggesting the majority of this subset might be inactive by morphological criteria. CD4⁺8^? thymocytes appeared to be similar to peripheral CD4⁺ T cells. The CD4^?8^? thymocyte subset contained morphologically immature cells in terms of cell size, presence of cell surface villi, and euchromatic appearance of the nucleus. CD4^?8⁺ thymocytes, heterogeneous in cell size, nuclear chromatin contents and amount of cytoplasm, could be divided into two distinct types. Type 1 CD4^?8⁺ thymocytes were intermediate in size, and therefore similar to peripheral mature CD8⁺ T cells. Type 2 CD4^?8⁺ thymocytes were large and irregular in shape (large CD4^?8⁺) with irregular-shaped and euchromatic nuclei. Large CD4^?8⁺ cells were, thus, considered to be at the transitional stage from CD4^?8^? to CD4⁺8⁺. At least two groups of large CD4⁺8⁺ cells were ultrastructurally classified by the nuclear chromatin content. Large CD4⁺8⁺ cells with heterochromatic nuclei were round with a smooth cell membrane, whereas large CD4⁺8⁺ cells with euchromatic nuclei were spherical with projections. Cytological features of heterochromatic large CD4⁺8⁺ cells are similar to those of small CD4⁺8⁺ thymocytes except for cell size. Euchromatic large CD4⁺8⁺ cells could be regarded as active blasts potentially leading to mature cells. Taken together, this is the first report that describes the ultrastructural characteristics of each thymocyte subset highly purified by flow cytometry.     

Identification and cellular distribution of the rat interleukin-2 receptor β chain: induction of the IL-2Rα−β+ phenotype by major histocompatibility complex class I recognition during T cell development in vivo and by T cell receptor stimulation of CD4+8+ immature thymocytes in vitro

A mouse monoclonal antibody (mAb) to the rat interleukin-2 receptor β (IL-2Rβ) chain was generated using IL-2Rβ cDNA-transfected mouse L929 cells for immunization and differential screening. This antibody, called L316, detects a cell surface protein with an apparent molecular mass of about 80 kDa. In peripheral lymphoid organs of young adult rats, IL-2Rβ expression is restricted to T and natural killer (NK) cells, and less than 10% of IL-2Rβ⁺ cells co-express the IL-2Rα chain. IL-2Rβ was detected on all NKRP-1^hi (NK⁻) and NKRP-1^lo cells (T-lineage cells of unknown function), most peripheral γδ T cells and on 30–40% of CD8 and 10% of CD4 αβ T cells. In the adult rat thymus, mAb L316 detects a small subset (about 1%) of predominantly IL-2Rα⁻ cells which express cell surface markers characteristic of mature T lymphocytes and contain a high proportion of CD4⁻8⁻ and CD4⁻8⁺ αβ T cell receptor (TCR)⁺ thymocytes. TCR-V usage suggests that major histocompatibility complex (MHC) class I plays a more important role than MHC class II in the selection of these cells. On immature CD4⁺8⁺ rat thymocytes, IL-2Rβ cell surface expression is readily induced by TCR stimulation in vitro, supporting the idea that in vivo, the IL-2Rβ⁺ phenotype is the result of TCR engagement during thymic selection.     

NK1.1+ CD4+ CD8- thymocytes with specific lymphokine secretion

CD4+8^? or CD4^?8⁺ thymocytes have been regarded as direct progenitors of peripheral T cells. However, recently, we have found a novel NK1.1⁺ subpopulation with skewed T cell antigen receptor (TcR) Vβ family among heat-stable antigen negative (HSA^?) CD4⁺8^? thymocytes. In the present study, we show that these NK1.1⁺ CD4+8^? thymocytes, which represent a different lineage from the major NK1.1^? CD4⁺8^? thymocytes or CD4⁺ lymph node T cells, vigorously secrete interleukin (IL)-4 and interfron (IFN)-γ upon stimulation with immobilized anti-TcR-αβ antibody. On the other hand, neither NK1.1^? CD4+8^?thymocytes nor CD4⁺ lymph node T cells produced substantial amounts of these lymphokines. A similar pattern of lymphokine secretion was observed with the NK1.1⁺ CD4⁺ T cells obtained from bone marrow. The present findings elucidate the recent observations that HSA^? CD4⁺8^? thymocytes secrete a variety of lymphokines including IFN-γ, IL-4, IL-5 and IL-10 before the CD4⁺8^? thymocytes are exported from thymus. Our evidence indicates that NK1.1⁺ CD4⁺8^? thymocytes are totally responsible for the specific lymphokine secretions observed in the HSA- CD4⁺8^? thymocytes.     

Expression of human NKRP1A by CD34+ immature thymocytes: NKRP1A-mediated regulation of proliferation and cytolytic activity

In this study, we show that NKRP1A is expressed and functions on a subset of immature human thymocytes. We took advantage of the monoclonal antibody (mAb) 191B8 that was obtained by immunizing mice with cultured human thymocytes characterized by an immature surface phenotype [CD2^? CD3^? CD4^? CD8^? stem cell factor receptor (SCFR)⁺] and expressing cytoplasmic CD3? chain. The 191B8 antibody homogeneously reacted with the immunizing population but not with most unfractionated thymocytes. It stained a minor population of resting immature thymocytes co-expressing CD34, SCFR, or both. Following culture of the CD34⁺ or CD34^? fractions of CD2^? CD3^? CD4^? CD8^? purified immature thymocytes with recombinant interleukin-2 (rIL-2), the 191B8-defined antigen was expressed on virtually all cells even when 191B8⁺ cells were removed from the starting population. On the other hand, no 191B8⁺ cells were detected in fresh or cultured thymocytes expressing a more mature phenotype. Biochemical analysis of 191B8 mAb-reactive molecules revealed, under non-reducing conditions, two bands displaying apparent molecular masses of 80 and 44 kDa and a single band of 44 kDa under reducing conditions. Digestion with proteases indicated that the 80-kDa form represented a homodimeric form of two 44-kDa molecules, while deglycosylation with N-glycanase suggested the existence of four N-glycosylation sites. Transfection of COS7 or NIH3T3 cells with hNKRP1A cDNA showed that the 191B8 mAb recognized NKRP1A as shown by both immunofluorescence analysis and immu-noprecipitation experiments. Functional studies showed that the 191B8/NKRP1A molecule mediated strong inhibition of the cytolytic activity of culturd CD2^? CD3^? immature thymocytes against a panel of tumor target cells. More importantly, 191B8 mAb induced proliferation of CD2^? CD3^? fresh thymocytes which was not increased by rIL-2. Thus, we propose that NKRP1A molecules, which are expressed in highly immature thymocytes, may play a regulatory role in their growth and function.     

The variable occurrence of CD45RA on CD8+ thymocytes correlates with the presence of Mtv sequences; its expression on other thymocytes is rare

While all thymocytes express CD45, only a small fraction (<3% in mice) bear the high molecular weight isoform, CD45RA. It has been suggested that these cells alone constitute the generative thymocyte lineage and should, therefore, occur within every ontogenic subset. To test this, we determined CD45RA expression among thymocyte subsets defined by CD4 and CD8. In some strains, exemplified by C57BL/Icrf, very few (< 0.2%) Tthymocytes expressed CD45RA and were mostly CD4^?CD8^? or CD4⁺CD8⁺, inconsistent with them constituting the generativelineage. In others, exemplified by BALB/c, CD45RA was expressed on up to 3% of T thymocytes, which were mostly CD4^?CD8⁺. The limited occurrence of CD4^?CD8⁺CD45RA⁺ thymocytes suggests that they are a nonconstitutive subset playing a role in the thymus of only some strains. Their occurrence correlates with that of Mtv proviruses within the mouse genome; however, their T cell receptor V_β repertoire is diverse, suggesting they are not uniquely selected by Mtv superantigens. We propose that they may be mature CD8 T cells, possibly responsible for introducing viral superantigens into the thymus.     

CD3-induced apoptosis of CD4+CD8+ thymocytes in the absence of clonotypic T cell antigen receptor

Clonal selection of T cells mediated through the T cell antigen receptor (TCR) mostly occurs at the CD4⁺CD8⁺ double positive thymocyte stage. Immature CD4⁺CD8⁺ thymocytes expressing self-reactive TCR are induced to die upon clonotypic engagement of TCR by self antigens. CD3 engagement by antibody of the surface TCR-CD3 complex is known to induce apoptosis of CD4⁺CD8⁺ thymocytes, a process that is generally thought to represent antigen-induced negative selection in the thymus. The present study shows that the CD3-induced apoptosis of CD4⁺CD8⁺ thymocytes can occur even in TCRα^? mutant mice which do not express the TCRαβ/CD3 antigen receptor. Anti-CD3 antibody induces death of CD4⁺CD8⁺ thymocytes in TCRα^? mice either in cell cultures or upon administration in vivo. Interestingly, most surface CD3 chains expressed on CD4⁺CD8⁺ thymocytes from TCRα^? mice are not associated with clonotypic TCR chains, including TCRβ. Thus, apoptosis of CD4⁺CD8⁺ thymocytes appear to be induced through the CD3 complex even in the absence of clonotypic antigen receptor chains. These results shed light on previously unknown functions of the clonotype-independent CD3 complex expressed on CD4⁺CD8⁺ thymocytes, and suggest its function as an apoptotic receptor inducing elimination of developing thymocytes.     

CD4+CD8− thymocytes that express L-selectin protect rats from diabetes upon adoptive transfer

Previous studies have shown that insulin-dependent diabetes can be induced in normal PVG.RT1^u rats by a protocol of adult thymectomy and irradiation. The injection of CD4⁺ T cells from non-irradiated syngeneic donors prevents the onset of disease in approximately 50 % of pre-diabetic recipients but all rats are protected if a particular subset of CD4⁺ cells is transferred. These protective cells express TCRαβ and have a memory phenotype, being CD45RC^low RT6⁺. Further studies have demonstrated that the transfer of CD4⁺CD8⁻ thymocytes, like that of unfractionated CD4⁺ peripheral T cells, also protects approximately half of recipients from diabetes suggesting that, as with the peripheral T cells, a functional heterogeneity may exist amongst CD4⁺CD8⁻ thymocytes. In this study, we show that L-selectin is expressed by 50–60 % of all CD4⁺CD8⁻ thymocytes from 6-week-old rats. Adoptive transfer of these populations into thymectomized and irradiated rats revealed that the protection from diabetes observed by CD4⁺CD8⁻ thymocytes was mediated almost entirely by the L-selectin⁺ subset. Cells with this phenotype were also able to mediate both humoral and cell mediated responses, providing primed B cells with help for secondary antibody responses and mediating local graft-versus-host reactions. L-selectin⁻ CD4⁺CD8⁻ thymocytes failed to mediate these responses. These data indicate that CD4⁺CD8⁻ thymocytes must mature to the stage of L-selectin expression, before they can mediate normal T cell function. The implications of these results are discussed with respect to the possible role of murine NK1.1⁺ thymocytes in the control of autoimmunity.     

Central tolerance: Clonal deletion or clonal arrest?

Studies in various experimental animals have shown that developing T cells with specificity for self antigens can be prevented from maturation at an early stage of development. While several in vitro and in vivo experiments have shown that the mechanism of silencing autospecific T cells is the deletion of immature CD4⁺8⁺ thymocytes other experiments were interpreted to indicate that tolerance could also result from developmental arrest of more immature CD4^?8⁺ thymocytes not involving cell death. Here we show that immature CD4^?8⁺ cells when confronted with T cell receptor ligands in vitro neither survive nor differentiate into cells which cannot be deleted, indicating that clonal elimination rather than developmental arrest is the mechanism of central tolerance of all immature T cells.     

Linomide Enhances Apoptosis in CD4+CD8+ Thymocytes

Linomide, a quinoline-3-carboxamide, has a pleiotropic immune modulating capacity and inhibits development as well as progression of disease in animal models of autoimmunity. Linomide treatment of mice resulted in a dramatic, dose-dependent decrease of the thymic cell number shortly after the start of administration. Flow cytometric analysis revealed that the major thymocyte subset, the early immature type CD4⁺CD8⁺ thymocytes, were reduced in number by 75%, mature CD4⁺CD8^? or CD4^?CD8⁺ thymocytes were less sensitive to treatment. The polyclonal T cell activator Con A (Concanavalin A) was used together with IL-2 to evaluate the potential proliferative responsiveness of ex vivo thymocytes. Thymocytes from mice treated with Linomide exhibited a more vigorous proliferation than control cultures. An effect shown to not only be due to the enrichment of mature thymocytes in the cultures from Linomide treated animals, but also when purified, mature thymocytes (CD4⁺CD8^? and CD4^?CD8⁺) were cultured with Con A and IL-2, these cells responded with a significantly enhanced proliferation. In vivo Linomide treatment did not result in increased plasma concentrations of corticosterone and treatment of adrenalectomized mice resulted in a reduction of thymocytes which was comparable to the effect in intact mice, indicating that glucocorticoids (GC) are not major mediators of Linomide-induced thymocyte deletion. In addition to this, and supporting a glucocorticoid independent mode of action, Linomide treatment of thymocytes in vitro resulted in a significant increase in the number of apoptotic cells, specifically in the CD4⁺CD8⁺ subset, implicating apopotosis as one component in the course of thymocyte reduction. In addition to this, in vivo treatment with Linomide resulted in an identical pattern to that seen in vitro in that there was significantly increased apoptosis only in the CD4⁺CD8⁺. These data indicate that Linomide modifies thymocyte development using a glucocorticoid independent pathway and results in the increased apoptosis of the CD4⁺CD8⁺ subset.     

Calcium-dependent lectin activity with novel distribution on thymocyte subsets

In this study we use carbohydrate probes to search for novel cell surface lectins in the immune system. Many carbohydrate binding proteins are members of the C-type lectin superfamily, whose ligand binding is dependent on calcium. To identify potential new members of this superfamily, fluorescein-conjugated carbohydrate polymers were used to probe for calcium-dependent cell surface binding. This approach offers advantages over the use of monoclonal antibody probes since only carbohydrate binding proteins would be identified. We have identified a cell surface lectin, referred to as thy-lec, detected by the probe fucoidan-fuorescein isothiocyanate. This calcium-dependent lectin has a novel distribution on thymocyte subsets. It is present on the surface of immature CD4⁺8⁺ and on large, cycling CD4^?8^? cells and CD8⁺4^? cells, but not on small, mature phenotype CD8⁺4^? or CD4⁺8^? thymocytes. This lectin is not found on mature T cells or other leukocytes in lymph nodes, spleen or bone marrow. It is proposed that this novel cell surface has a function in the maturation of T cells in the thymus.     

Discrimination between maintenance- and differentiation-inducing signals during initial and intermediate stages of positive selection

As well as signaling through the αβ T cell receptor complex, positive selection of immature CD4⁺ 8⁺ thymocytes involves additional ill-defined accessory interactions provided by thymic epithelial cells. Here, we have isolated CD4⁺ 8⁺ thymocytes at a pre-positive selection stage of development (TCR^? CD69^? 4⁺ 8⁺ cells), or after initiation of positive selection (CD69⁺ 4⁺ 8⁺ cells), from mice where the normal lifespan of thymocytes is extended by the presence of a bcl-2 transgene, to allow us to discriminate between requirements for maintenance and differentiation signals during positive selection. We find that MHC class II⁺ thymic epithelial cells drive positive selection of TCR^? CD69^? 4⁺ 8⁺ bcl-2 tg thymocytes to the CD4⁺ and CD8⁺ stage, while no such mature subsets are observed when thymocytes are cultured alone or with major histocompatibility complex (MHC) class II⁺ salivary epithelial cells. However, CD4⁺ 8⁺ cells remain in such cultures in considerable numbers, and retain the potential for positive selection if re-cultured with thymic epithelium, suggesting that thymic epithelial cells provide specific differentiation-inducing signals for positive selection. In contrast, intermediate CD69⁺ 4⁺ 8⁺ thymocytes show some capacity for phenotypic conversion in the absence of thymic stromal cells although strikingly the single-positive CD4⁺ and CD8⁺ cells generated are not functionally competent. Finally, we show that prior culture of thymic epithelial cells under monolayer conditions abrogates their ability to support the initiation of positive selection, suggesting that the epithelial cell molecules necessary for the provision of differentiation signals during positive selection are down-regulated under such conditions.