Thymocyte Development in the Absence of Pre-T Cell Receptor Extracellular Immunoglobulin Domains

See allHide authors and affiliations

Science  08 May 1998:
Vol. 280, Issue 5365, pp. 905-908
DOI: 10.1126/science.280.5365.905


Immature thymocytes express a pre–T cell receptor (pre-TCR) composed of the TCRβ chain paired with pre-Tα. Signals from this receptor are essential for passage of thymocytes through a key developmental checkpoint in the thymus. These signals were efficiently delivered in vivo by a truncated form of the murine pre-TCR that lacked all of its extracellular immunoglobulin domains. De novo expression of the truncated pre-TCR or an intact αβTCR was sufficient to activate characteristic TCR signaling pathways in a T cell line. These findings support the view that recognition of an extracellular ligand is not required for pre-TCR function.

Mature αβ T lymphocytes bear on their surface a heterodimeric T cell receptor (TCR) that contains the protein products from rearranged TCRα and TCRβ loci. During development, rearrangement of the TCRβ locus occurs first, allowing immature CD4CD8 [double-negative (DN)] thymocytes to synthesize the TCRβ protein and express it on the surface in association with the pre-Tα (pTα) protein (1,2). Signals from this “pre-TCR” then induce the cells to differentiate into CD4+CD8+ [double-positive (DP)] cells and to undergo a rapid series of cell divisions. In the absence of pre-TCR function, this differentiation is blocked, and thymic cellularity is severely reduced (3-12).

Signaling by the pre-TCR and the mature receptor involves the CD3/ζ chains and depends on the concerted action of Src family and Syk/ZAP-70 tyrosine kinases (4-6). No ligand for the pre-TCR has yet been identified, but one might be expected, given the surface expression of the pre-TCR (11, 13, 14) and its structural resemblance to antigen receptors. Alternatively, surface expression of the TCRβ-pTα heterodimer might be sufficient to initiate the pre-TCR signaling process.

To distinguish between these possibilities, we generated transgenic mice expressing a truncated form of the pre-TCR heterodimer that lacked all of its extracellular immunoglobulin (Ig) domains. Truncated forms of TCRβ and pTα (βT and pTαT, respectively) were constructed (Fig. 1A) (15) that have their normal extracellular Ig domains replaced with Flag or Myc epitope tags but still retain the cysteine residues and transmembrane domains required for heterodimer formation and assembly with the signal-transducing CD3 and ζ chains (16). To confirm that the modifications had not adversely affected assembly of the receptor (17), we transiently transfected the truncated chains, individually or together, by electroporation into a TCR-deficient thymoma (18). Because the cells do not express TCRβ, the expression of pTαTalone did not rescue CD3 on the cell surface (Fig. 1B). In contrast, the introduction of βT alone resulted in a slight but reproducible increase in surface expression of CD3, presumably through the formation of heterodimers with endogenous pTα. The expression of βT and pTαT together resulted in rescue of surface CD3, showing that a pre-TCR complex could be assembled with the ectodomain-deficient heterodimer.

Figure 1

Structure and expression of the extracellular truncation mutants of the pre-TCR. (A) Wild-type and truncated forms of the pre-TCR. The mutant forms of pTα (pTαT) and TCRβ (βT) were truncated on the NH2-terminal side of the disulfide bridge, so that only nine (pTαT) and four (βT) residues of the native sequences were retained. Flag and Myc epitopes were added as indicated, but these were difficult to detect by fluorescence-activated cell sorting (FACS), perhaps because they were obscured by the presence of the CD3 chains. (B) FACS profiles (45) showing surface expression of the truncated pre-TCR after transient transfection (25) into aRag-1 −/−;p53 −/−thymoma (18). Constructs encoding pTαT (thin line), βT (dashed line), or both (thick line) were electroporated into the cells 1 day before staining for surface expression of CD3ε with mAb 145-2C11. A plasmid encoding human CD8 was transfected as a control to establish the amount of background staining (which was indistinguishable from that observed after introduction of pTαT alone) (21).

Transgenic mice were generated that expressed either βT or pTαT in thymocytes under the control of the Lck proximal promoter (19). For analysis of thymocyte development in the absence of endogenous TCR gene rearrangements, these mice were interbred and back-crossed to mice with no recombinase activating gene–1 (Rag-1) (10). Most thymocytes in wild-type mice were CD4+CD8+ DP cells with a minor fraction of cells expressing either CD4 or CD8 (mature single-positive cells) (Fig.2). In contrast, theRag-1 −/− thymus was blocked at an early developmental stage and contained no DP or mature thymocytes. The Rag-1 −/− developmental block was partially corrected in mice that expressed the βTtransgene; most of the thymocytes in these mice were CD4+CD8+, and their thymic cellularity was 10-fold greater than in nontransgenic Rag-1 −/−mice (Fig. 2). However, these mice still had only 10% of the number of thymocytes in wild-type mice. The DP cells also retained CD25 expression, a characteristic that has previously been associated with impaired development at this stage (3, 20) (Fig.3A). Although surface expression of CD3, and thus of the pre-TCR, was undetectable on these thymocytes (Fig.3B), intraperitoneal injection of a monoclonal antibody (mAb) to the NH2-terminal epitope tag on βT or of an mAb to CD3ε restored both the differentiation and expansion of the DP compartment (21-23). Thus, small amounts of the truncated TCRβ were, indeed, expressed at the cell surface in association with CD3.

Figure 2

Rescue of CD4+- CD8+ thymocyte development inRag-1 −/− mice expressing truncated pre-TCRs. FACS profiles show the relative expression of CD4 and CD8 for representative wild-type (WT),Rag-1 −/−, and transgenic mice expressing the indicated mutant pre-TCRs. Thymocytes were stained with mAbs specific for CD4 and CD8 and then analyzed by flow cytometry (45). The FACS profiles are representative of data acquired from multiple experiments; the associated figures for thymic cellularity are averages generated from figures for the indicated number of mice between the ages of 5 to 8 weeks. Older mice showed the same phenotypes as those presented here but were not used for calculating averages (21).

Figure 3

Cell surface phenotype of thymocytes from transgenic mice expressing the indicated mutant pre-TCRs. Relative expressions of CD25 (A) and CD3ε (B) on unseparated thymocytes from the following mice: Rag-1 −/−(dashed line), βT;Rag-1 −/−(dotted line), pTαTT;Rag-1 −/−(thick solid line), and C57BL/6–wild type (thin solid line).

Rag-1−/− mice with the pTαT chain had no DP thymocytes (Fig. 2), as was expected given the absence of endogenous TCRβ rearrangement in these mice (10, 12). In contrast, Rag-1−/−mice expressing both the βT and pTαTtransgenes had DP cells that were phenotypically similar to their wild-type counterparts in abundance, forward scatter, and expression of surface molecules, such as CD25 (Fig. 3A) and CD2 (21). A minor population of large thymocytes in these mice expressed low but detectable amounts of surface CD3, similar to what has been previously reported for cells that express the pre-TCR (11, 13); almost all of these were DP cells (21). CD25+ DN cells had no detectable pre-TCR by flow cytometry. Thus, the observed developmental effects were not caused by pre-TCR overexpression.

We also confirmed the function of the truncated pre-TCR by transfecting the pTαT and βT transgenes intoRag-1 −/− embryonic stem (ES) cells (19,24) and making chimeric mice by microinjection intoRag-1 −/− blastocysts. Almost all of the ES cell–derived thymocytes (Ly9.1+) in the chimeric mice were CD25 DP cells (21). Thus, as in the transgenic mice, the truncated pre-TCR was proficient in promoting differentiation from the DN to DP stages.

Coexpression of both truncated molecules, but not βTalone, in Rag-1 −/− mice allowed efficient development of thymocytes to the DP stage (Fig. 2). We also observed that the fully truncated heterodimer prevented the efficient emergence of single-positive thymocytes and expression of αβ TCRs in mice expressing Rag-1, whereas βT alone did not (Fig. 4A). This blockage of αβ TCR expression resulted from suppression of rearrangements at the endogenous TCRβ locus (Fig. 4B). Thus, pre-TCRs containing endogenous full-length pTα paired with βT were not efficient in either promoting the differentiation of DP cells or mediating allelic exclusion; that is, endogenous pTα did not play a substantial role in promoting the development of thymocytes that expressed the truncated heterodimer.

Figure 4

Allelic exclusion mediated by the fully truncated pre-TCR. (A) Relative expression of TCRβ on thymocytes from the following mice: Rag-1 −/− (dashed line), βT;Rag-1 +/− (dotted line), pTαTT;Rag-1 +/−(thick solid line), and C57BL/6–wild type (WT) (thin solid line). (B) Allelic exclusion at the endogenous TCRβ locus in thymocytes from mice of the indicated genotypes. Genomic DNA was isolated from thymocytes, and rearrangements at the TCRβ locus were detected with a semiquantitative PCR assay and primers specific for Vβ11 and Jβ2, as previously described (46). PCR primers that amplify a region of the IgM constant domain were used as a control. PCR products were detected by Southern blot with primers that hybridize in the amplified regions. Consistent results were obtained with a similar PCR assay with primers specific for Vβ12 and Jβ2 or by Southern blot with a probe from the intron upstream of Dβ1 (47).

Immunofluorescence experiments indicated that the truncated pre-TCR was distributed on the surface of transfected cells in a manner indistinguishable from that of a wild-type TCR (21). We also employed an in vitro assay to examine signaling from the truncated receptor and to determine whether the truncations had caused it to be constitutively activated (25). To do so, we transiently transfected either the truncated pre-TCR or a wild-type αβ TCR into a TCRβ-deficient Jurkat T cell line (26). Receptor-dependent signaling was then monitored with a luciferase assay (27) that reports on the relative transcriptional activity of the nuclear factor of activated T cells (NFAT) (28). To increase the sensitivity of the system, we cotransfected the cells with an activated form of ras, which provided some (29) but not all (30) of the signal needed to activate the NFAT cassette.

Transfection of plasmids encoding either an irrelevant protein (cre recombinase) or βT alone into the TCRβ-deficient Jurkat line did not induce changes in NFAT activity, as detected with the luciferase reporter assay (Fig. 5A), and did not reconstitute appreciable amounts of surface CD3 (Fig. 5B). In contrast, ligand-independent signal transduction was detected within hours of transfection of either TCRβ or pTαTT. The signal was transient and had subsided by 40 to 50 hours, despite sustained expression of the receptors (Fig. 5B). The unmodified αβ receptor reproducibly signaled more robustly than the truncated receptor even though the former was less abundant on the cell surface. Thus, the truncations had not resulted in a constitutively activated signaling phenotype. Further support for this view came from control experiments in which the same cells were transfected with a constitutively active mutant of Lck , a kinase essential for pre-TCR function (5-8), or a constitutively ligated CD8/ζ chimeric receptor (31), which is ligated by endogenous major histocompatibility complex class I molecules on the Jurkat cells. In both cases, by 7 hours after transfection, the luciferase signal was already roughly 500 times greater than the response of the transfected TCR (21). The ligand-independent signal could also be observed in the TCR-deficient mouse thymoma (as used in Fig. 1) (18, 21). Finally, when comparable amounts were expressed, both the αβ and pTαTT receptors responded equivalently to ligation by mAb to CD3ε (Fig. 5C).

Figure 5

Ligand-independent signaling by TCRs. (A) Luciferase activity after transfection of a TCRβ-deficient Jurkat cell line with plasmids encoding cre recombinase (control; open squares), βT (open triangles), the Jurkat TCRβ chain (closed squares), or pTαTT (closed triangles), all under the transcriptional control of the EF-1α promoter (25, 30). An NFAT-luciferase reporter plasmid (44) and a plasmid encoding v-Ras were included in all transfections. Transfection efficiency was equivalent between samples, as determined by comparing responses to ionomycin alone. (B) Surface CD3ε expression was assessed at 14.5 hours after transfection in TCRβ-deficient Jurkat cells transfected with cre recombinase (shaded histogram, with next-to-highest peak), βT (thin solid line and highest peak), TCRβ (dotted line), or pTαTT (thick solid line). The inset figure plots the percentage of CD3ε+ cells over time (line designations as in the accompanying histogram). The data are representative of three such extended kinetic experiments (10 time points) and are consistent with the results from many experiments analyzing NFAT activity at various single time points after transfection. (C) TCRβ-deficient Jurkat cells were cotransfected with the indicated constructs together with the v-Ras and NFAT-luciferase plasmids. At 7 hours after electroporation, cells were either left untreated (black bars) or incubated with anti-CD3 mAb Leu 4 (4 μg/ml) (hatched bars). The cells were incubated at 37°C for 5 hours and then lysed and analyzed for luciferase activity. Transfection efficiency was assessed by comparing responses to ionomycin alone.

Our data show that a TCRβ;pTα heterodimer lacking all of its extracellular Ig domains is fully capable of delivering the signal that initiates the differentiation and proliferation of immature DN thymocytes. DP thymocytes developed inRag-1 −/− mice expressing a TCRβ that lacks the variable domain but retains the constant domain (32). Our truncated pre-TCR removed the potential for interactions between any of the Ig domains of the TCRβ;pTα heterodimer and putative extracellular ligands. Moreover, the expression of either a truncated pre-TCR or a wild-type αβ TCR was sufficient to activate TCR signaling pathways in vitro. Thus, ligand binding by the extracellular portion of the TCRβ;pTα heterodimer is probably not essential for the pre-TCR signal to be generated in vivo.

If the extracellular domain of the pTα molecule is not involved in ligand binding, then its primary purpose may be to provide the necessary structural stability for assembly with the TCRβ and the CD3/ζ chains. In this respect, a pre-TCR composed of a truncated β chain and the endogenous pre-TCRα chain was inefficiently expressed at the cell surface (Figs. 1B and 3B), perhaps because it was improperly assembled or unstable, and consequently was retained in the endoplasmic reticulum. The coexpression of a pTαT chain with βT efficiently rescued both cell surface CD3 and the developmental transition; therefore, the fully truncated heterodimer may be more stable structurally than an asymmetrical heterodimer.

Ligand-independent signaling in response to de novo expression of a cell surface receptor has a precedent in the receptor-dependent but ligand-independent activation of heterotrimeric G proteins in lipid vesicles reconstituted with α-adrenergic receptors (33). Low-level signaling may be a general response to the acquisition of a cell surface receptor and the coincident redistribution of signaling enzymes and their substrates (34). The DN to DP transition of thymocytes is a unique developmental checkpoint that may depend on this type of low-grade signaling activity. Unlike most other developmental checkpoints, this one tests for the productive outcome of an autonomous genomic rearrangement process and therefore need not necessarily derive its differentiation cue from an extracellular interaction.

The pre-B cell receptor is essential for a stage of B cell development analogous to that of the thymic cells studied here, the transition from pro- to pre-B cells (35). Pre-B cells can develop in the absence of the IgM variable domain and independently of association with surrogate light chains (36, 37). Thus, like the DN to DP transition, the pro- to pre-B cell transition also may be regulated by ligand-independent signaling from a precursor form of the antigen receptor. Finally, the survival of peripheral lymphocytes depends on signals provided by cell surface antigen receptors; for B cells (38), but probably not T cells (39), these signals might be generated in a ligand-independent fashion.

  • * To whom correspondence should be addressed. E-mail: nigel{at}


View Abstract

Navigate This Article