Lineage Commitment in the Thymus: Only the Most Differentiated (TCR^hibcl-2^hi) Subset of CD4⁺CD8⁺Thymocytes Has Selectively Terminated CD4 or CD8 Synthesis

C57BL/6 (B6) and TCRα^−/− mice (14) were obtained from Jackson Laboratories (Bar Harbor, Maine) and were housed and bred in a specific pathogen-free facility. Thymuses were dissected from mice between 6 and 10 wk of age and single cell suspensions were made by gently teasing dissected lobes with forceps and filtering over nylon.

Thymocytes were stained and sorted as previously described (12). In brief, 5 × 10⁵ cells were distributed in wells of a 96-well plate (Nunclon, Roskike, Denmark), pelleted, and stained for 30 min at 4°C with saturating concentrations of PE-conjugated anti-CD4 mAb (GK1.5; Becton Dickinson & Co., Mountain View, CA), FITC-conjugated anti-CD8 mAb (53-6-7; Becton Dickinson), and, where indicated, biotinylated anti-CD5 mAb (53.7.3; Becton Dickinson & Co.), biotinylated anti-CD69 mAb (H1.2F3; PharMingen, San Diego, CA) or biotinylated anti-TCRβ mAb (H57-597; PharMingen) in 30 μl final volume of staining medium (0.5% BSA, 0.5% NaN₃ in HBSS). PE-conjugated, FITC-conjugated, and biotinylated antihuman CD3 mAb (Leu 4; Becton Dickinson & Co.) were used as controls for nonspecific staining. All staining was also performed in the presence of anti-FcReceptor (FcR) antibody (2.4G2; Becton Dickinson & Co.) to block nonspecific antibody binding to FcRs. Cells were washed twice with 150 μl of staining medium and if a second step were required, were stained again in 30 μl final volume of saturating concentrations of either Texas-red streptavidin or RED670-streptavidin (GIBCO BRL, Gaithersburg, MD) for 5 min at 4°C. Cells were washed twice again and analyzed by flow cytometry. For sorting, cells were stained at a final concentration of 10⁸/ml in 12 × 75 mm tubes (no. 2058; FALCON, Becton Dickinson & Co.) using saturating amounts of PE-conjugated anti-CD4 mAb, FITC-conjugated anti-CD8 mAb and biotinylated anti-CD5 mAb. After washing two times with staining medium, cells were resuspended at a concentration of 10⁸/ml and stained with saturating concentrations of Texas red (TR)-streptavidin (GIBCO BRL). After sorting, pronase treatment, and culture cells were typically restained with the same antibody combinations for analysis. In those cases when it was necessary to stain for an additional marker, we restained with PE anti-CD4 and FITC anti-CD8 as usual, but used distinct biotinylated antibodies (biotinylated anti-TCRβ mAb (H57-597, PharMingen) or antimurine bcl-2 mAb (3F11, PharMingen) followed by a biotinylated anti-hamster mAb (G94-56, PharMingen) in combination with RED670-conjugated streptavidin to avoid confusion with any residual anti-CD5/TR-streptavidin conjugates. Staining with anti-bcl-2 required that the cells be permeabilized by treatment with 0.03% saponin and we followed the method described by Veis et al. (15). Flow cytometry was performed on a Becton Dickinson FACStar^® Plus and analyzed using software designed by the Division of Computer Research and Technology at the National Institutes of Health except in cases where RED-670 was used as a second step, when cells were analyzed on a Becton Dickinson FACScan^® with Cell Quest software. Dead cells were excluded electronically by gating on forward scatter and propidium iodide intensity when analyzed on the FACStar^® Plus and by gating on forward and side scatter when analyzed on the FACScan^®.

The coreceptor reexpression assay was performed as previously described (12). In brief, sorted cells were washed two times with PBS and resuspended at 1–5 × 10⁶/ml with or without 0.04% Pronase (Calbiochem Novabiochem, La Jolla, CA) in PBS. They were incubated for 15 min at 37°C and then pelleted and incubated with fresh pronase for another 10 min at 37°C. Cells were subsequently washed three times with complete medium then distributed in 0.5-ml vol into 24-well plates at a final concentration of 0.5–2.0 ×10⁶ per ml. Cells were incubated overnight (12–16 h) at 4°C or 37°C, stained, and analyzed.

Because CD4 and CD8 surface proteins can survive for hours on the surface of DP thymocytes, DP thymocytes can selectively terminate synthesis of one or the other coreceptor molecule without an obvious change in surface phenotype. To detect the coreceptor molecules that DP thymocytes are actively synthesizing, we have utilized the coreceptor reexpression assay (Fig. 1). In brief, thymocytes are stripped of surface CD4 and CD8 proteins by extracellular treatment with low concentrations of the protease pronase, washed, and then cultured overnight at 37°C in medium alone to allow surface reexpression. Surface reexpression of coreceptor molecules in this assay has previously been shown to require new transcription and active protein synthesis and so provides a convenient demonstration of the coreceptor molecules that individual DP thymocytes are actively synthesizing (12). In the present study we utilized the coreceptor reexpression assay to determine if any relationship existed between the selective termination of coreceptor synthesis and developmental maturity. As well-characterized markers of developmental maturity we utilized surface expression of CD5 and CD69 (9, 16–21).

Surface expression of CD5 and CD69 is markedly heterogeneous among DP thymocytes, with DP thymocytes contained within cell populations expressing low, intermediate, and high levels of each marker (Fig. 2, a and b). More importantly, surface CD5 and CD69 expression is directly correlated with developmental maturity such that the least differentiated DP thymocytes express low amounts of CD5 and CD69 and the most differentiated DP thymocytes express high levels of CD5 and CD69 (17, 19, 20). Consequently, to determine if termination of coreceptor synthesis occurred in a developmentally restricted subpopulation of DP thymocytes, we performed the coreceptor reexpression assay on thymocytes that were electronically purified by 3-color cell sorting into DP thymocytes expressing low/intermediate versus high levels of surface CD5 and CD69 (Fig. 3, a and b). All DP thymocytes expressing low/intermediate levels of CD5 or CD69 either synthesized both coreceptor molecules or synthesized no coreceptor molecule (Fig. 3, a and b, upper panels); essentially none of these cells selectively synthesized only one coreceptor molecule (Fig. 3, a and b). In contrast, DP thymocytes expressing high levels of surface CD5 or CD69 did contain cells that were selectively synthesizing only one coreceptor molecule (Fig. 3, a and b, lower panels). DP thymocytes selectively synthesizing CD8 usually outnumbered thymocytes selectively synthesizing CD4. We suspect this is because thymocytes that have selectively terminated coreceptor synthesis lose surface CD8 expression more quickly than surface CD4 expression, a possibility consistent with observations by other investigators (22, 23). We also observed a significant number of cells that have terminated or diminished synthesis of both CD4 and CD8. The fate of these cells is not yet known but they may represent precursors of mature DN T cells. Most importantly, these data indicate that cells selectively synthesizing only one coreceptor molecule (either CD4 or CD8) are not present among the vast majority of DP thymocytes, but are only present among DP thymocytes expressing high surface levels of CD5 and CD69.

Even though all DP thymocytes selectively synthesizing only one coreceptor molecule were CD5^hiCD69^hi, most of the cells that were CD5^hiCD69^hi actively synthesized both coreceptor molecules (Fig. 3,a). Consequently, we wished to determine if markers existed that could uniquely identify CD5^hiCD69^hi DP thymocytes that had selectively terminated synthesis of one coreceptor molecule and distinguish them from the majority of CD5^hiCD69^hi DP thymocytes that continued to synthesize both coreceptors. As a first attempt we wished to quantitatively compare CD5 and CD69 surface expression on these different DP subpopulations. DP thymocytes that had undergone the coreceptor reexpression assay could be subsequently assessed for surface CD5 but not CD69 expression because CD5 is resistant to pronase treatment whereas CD69 is not. We found that surface CD5 expression was essentially equivalent on all CD5^hi DP thymocytes, regardless of whether they were actively synthesizing one or both coreceptor molecules (Fig. 3 c) and therefore could not be used to distinguish the subpopulations.

However, we observed a clear difference when we examined the small subset of CD5^hiCD69^hi DP thymocytes that had selectively terminated synthesis of one or the other coreceptor molecule for expression of two other differentiative markers, surface TCR and internal bcl-2 proteins (24–31) (Fig. 4). CD5^hi DP thymocytes that were selectively synthesizing only one coreceptor molecule (either CD4 or CD8) expressed significantly higher levels of both surface TCRβ and internal bcl-2 protein than CD5^hi DP thymocytes that continued to synthesize both coreceptor molecules (Fig. 4). Thus, DP thymocytes that have selectively terminated synthesis of one coreceptor molecule (either CD4 or CD8) are exclusively contained within a small subpopulation of DP thymocytes that are not only CD5^hiCD69^hi, but are also TCR^hibcl-2^hi.

These results demonstrate that selective termination of synthesis of one coreceptor molecule is a stringently regulated event during development which is evident only within the small, most differentiated subset of DP thymocytes defined as CD5^hiCD69^hiTCR^hibcl-2^hi. In addition, these results identify three subsets of DP thymocytes that differ in their state of coreceptor synthesis: (a) CD5^lo CD69^lo DP thymocytes that actively synthesize both coreceptor molecules; (b) CD5^hiCD69^hiTCR^lobcl2^lo DP thymocytes that synthesize both coreceptor molecules; and (c) CD5^hiCD69^hiTCR^hibcl2^hi DP thymocytes that actively synthesize only one coreceptor molecule having terminated expression of the other (Table 1). We propose there may be a precursor/progeny relationship among these subpopulations such that they define the progression of DP thymocytes from uncommitted to precommitted to lineagecommitted thymocytes.

As an independent test of our conclusion that selective termination of coreceptor synthesis is a developmentally regulated event that only occurs in the most differentiated subset of CD5^hi DP thymocytes, we examined DP thymocyte populations from TCRα^−/− mice that are unable to assemble and express surface TCRαβ complexes and, consequently, lack CD5^hi DP thymocytes (Fig. 5,a). In this experiment, we attempted to increase our ability to detect DP thymocytes synthesizing only one coreceptor molecule by performing the coreceptor reexpression assay on purified populations of CD4⁺8^lo and CD4^lo8⁺ transitional DP thymocytes that may be developmentally more advanced than the bulk of DP thymocytes (Fig. 5,b). Even so, we failed to detect any DP thymocytes in TCRα^−/− mice that were selectively synthesizing only one coreceptor molecule (Fig. 5 b). Thus, surface expression of TCRαβ/ CD3 complexes and the signals they transduce are required for generation of CD5^hi DP thymocytes and for selective termination of a coreceptor molecule.

Having confirmed that selective termination of coreceptor synthesis occurs only in CD5^hi DP thymocytes, we wished to examine the developmental relationship between CD5^hi DP thymocytes that continue to synthesize both coreceptor molecules and those that have selectively terminated coreceptor synthesis. In particular, if CD5^hi DP thymocytes that continue to synthesize both coreceptor molecules are the precursors of those that have selectively terminated synthesis of one coreceptor molecule, it would be expected that these populations would appear sequentially in development. To examine this possibility, we performed the coreceptor reexpression assay on CD5^hi DP thymocytes from newborn mice less than 12 h old (Fig. 6). Interestingly, newborn mice contained CD5^hi DP thymocytes at a frequency comparable to that of adult mice (10–15% of all DP cells). However, essentially all neonatal CD5^hi DP thymocytes actively synthesized both coreceptor molecules; very few, if any, had terminated synthesis of either or both coreceptor molecules (Fig. 6). Thus, CD5^hi DP thymocytes that synthesize both coreceptor molecules appear earlier in development than CD5^hi DP thymocytes that synthesize only one coreceptor molecule, consistent with (but not proving) a precursor/progeny relationship.

Lineage commitment is a developmental process by which positively selected DP thymocytes are induced to differentiate into either CD4⁺ or CD8⁺ T cells. Identification of the intrathymic signals that induce lineage commitment would be significantly facilitated if the molecular consequences of lineage commitment could be recognized as soon as they occurred in DP thymocytes. We have proposed that lineage commitment be molecularly defined as the selective termination of synthesis of one or the other coreceptor molecule in a DP thymocyte expressing both coreceptors on its surface. Lineage commitment by any definition must require that DP thymocytes shift from synthesizing both coreceptor molecules to synthesizing only one. However, the validity of defining lineage commitment in DP thymocytes by coreceptor synthesis has been questioned because of the possibility that selective termination of coreceptor synthesis may occur randomly in DP thymocytes and, therefore, may be unrelated to their differentiation state.

The present study, however, demonstrates that selective termination of CD4 or CD8 synthesis as determined by the coreceptor reexpression assay does not occur randomly in DP thymocytes, but, instead is a highly regulated event which has occurred exclusively within a developmentally discrete subpopulation of DP thymocytes that is CD5^hiCD69^hi TCR^hibcl-2^hi. DP thymocytes expressing high levels of CD69, TCR and bcl-2 have each been shown by other investigators to have undergone positive selection (19–20, 24–30) and TCR^hi subpopulations of DP thymocytes, specifically, have been shown to contain the immediate precursors of SP thymocytes (23–26). That termination of coreceptor synthesis was only evident within that population of thymocytes which has undergone positive selection is compelling evidence that selective termination of coreceptor synthesis is not a capricious event in DP thymocytes but is a molecular indicator of lineage commitment. Importantly, selective termination of coreceptor synthesis as detected by the coreceptor reexpression assay, unlike other markers identifying lineage committed DP thymocytes, reveals the T cell lineage to which individual DP thymocytes are committed since it reveals the coreceptor molecule that committed thymocytes continue to synthesize.

The present study has identified three distinct subpopulations of DP thymocytes that are distinguishable by surface phenotype and lineage commitment status: (a) CD5^lo CD69^lo DP thymocytes that have not selectively terminated synthesis of either CD4 or CD8 coreceptor molecules, (b) CD5^hiCD69^hiTCR^lobcl-2^lo DP thymocytes that continue to synthesize both coreceptor molecules, and (c) CD5^hiCD69^hiTCR^hibcl-2^hi DP thymocytes that actively synthesize only one or the other coreceptor molecule, but not both. A striking and unanticipated finding of the present study is that all DP thymocytes that have selectively terminated either CD4 or CD8 synthesis are CD5^hiCD69^hiTCR^hibcl-2^hi and, furthermore, that essentially all CD5^hiCD69^hiTCR^hibcl-2^hi DP thymocytes have selectively terminated either CD4 or CD8. This precise overlap suggests that intrathymic signals induce the immediate precursors of these lineage committed DP thymocytes (presumed to be CD5^hiCD69^hiTCR^lobcl-2^lo DP thymocytes) to coordinately: (a) upregulate TCR expression, (b) increase bcl-2 content, and (c) selectively terminate synthesis of one coreceptor molecule. Consequently, CD5^hiCD69^hiTCR^intbcl-2^lo DP thymocytes synthesizing both coreceptor molecules appear to be the targets of lineage commitment signals in the thymus. A summary of these findings is presented in Table 1.

We think that a precursor/progeny relationship exists among all three DP thymocyte subpopulations which is consistent with their sequential appearance during development, although it is difficult to prove such precursor/progeny relationships conclusively. From a lineage commitment perspective, we consider CD5^loCD69^lo DP thymocytes to be uncommitted; CD5^hiCD69^hiTCR^lobcl-2^lo DP thymocytes to be potential targets of lineage commitment signals and so are precommitted; and CD5^hiCD69^hiTCR^hibcl-2^hi DP thymocytes to be committed. The differentiation of CD5^lo CD69^lo DP thymocytes into CD5^hiCD69^hiTCR^lobcl-2^lo DP thymocytes presumably requires intrathymic signals transduced by surface TCR/CD3 complexes as CD5 and CD69 upregulation is mediated by TCR/CD3 signaling (17, 20, 32). Notably, the present study indicates that such intrathymic TCR/CD3 signals are not sufficient to selectively terminate synthesis of either CD4 or CD8 coreceptor molecules, as CD5^hiCD69^hiTCR^lobcl-2^lo DP thymocytes continue to synthesize both coreceptor molecules. Rather, we think additional signals are required in CD5^hi CD69^hiTCR^lobcl-2^lo DP thymocytes to increase TCR expression, to increase bcl-2 expression, and to selectively terminate either CD4 or CD8 coreceptor synthesis. It is important to note, however, that the present data indicate that TCR/CD3 signals are necessary for selective termination of CD4 or CD8 synthesis. Whether these signals are generated by TCR engagement by antigen or by ligand engagement of non-clonotypic receptors that signal via the TCR/CD3 signaling complex, such as Thy-1 (33), is not known.

In the present study the coreceptor reexpression assay has been performed on DP thymocytes, whereas we have previously utilized the coreceptor assay on transitional populations of DP thymocytes that have quantitatively downregulated surface expression of one or the other coreceptor molecule (12). The results of our previous studies indicated that termination of CD4 synthesis was more stringently regulated during development than termination of CD8 synthesis, and we advanced a model, the asymmetric commitment model, to explain our findings. Specifically, we proposed that DP thymocytes progress to a stage in development (termed the commitment checkpoint) when they are assessed for the presence or absence of a CD8-commitment signal. The presence of a CD8-commitment signal results in selective termination of CD4 synthesis (CD8commitment), whereas the absence of a CD8-commitment signal results by default in selective termination of CD8 synthesis (CD4-commitment). The present study provides a phenotype that can be assigned to each of the commitment stages in the asymmetric commitment model (Fig. 7): CD5^loCD69^lo DP thymocytes are uncommitted cells that are preselection; CD5^hiCD69^hiTCR^lobcl-2^lo DP thymocytes are cells at the commitment checkpoint and so are precommitment; CD5^hiCD69^hiTCR^hibcl-2^hi DP thymocytes have committed to either the CD4 or CD8 T cell lineages (Fig. 7). However, the present finding that all lineage-committed DP thymocytes, even those that have selectively terminated CD8 synthesis, are CD5^hiCD69^hi indicates that commitment to either lineage occurs only in TCR/CD3 signaled DP thymocytes. Hence, we would like to clarify the asymmetric commitment model to state that: (a) DP thymocytes that fail to receive TCR/CD3 signals remain CD5^loCD69^lo and die from neglect as uncommitted DP thymocytes; and (b) CD4-commitment occurs in CD5^hiCD69^hi DP thymocytes that have not received CD8-commitment signals.

In conclusion, the present study demonstrates that selective termination of CD4 or CD8 coreceptor synthesis in DP thymocytes is a highly regulated developmental process that occurs in strict parallel with other events that define positive selection. It is anticipated that the molecular definition of lineage commitment validated in this study will enhance identification of the intrathymic signals that induce lineage commitment in developing thymocytes.