Development of immature CD4−CD8− (double-negative) thymocytes to the CD4+CD8+ (double-positive) stage is linked to productive rearrangement of the TCRβ locus by signals transduced through the pre-TCR. However, the mechanism whereby pre-TCR signaling is initiated remains unclear, in part due to the lack of an in vitro model system amenable to both biochemical and genetic analysis. In this study, we establish the thymic lymphoma Scid.adh as such a model system. Scid.adh responds to Ab engagement of surface IL-2Ra (TAC):CD3ε molecules (a signaling chimera that mimics pre-TCR signaling in vivo) by undergoing changes in gene expression observed following pre-TCR activation in normal thymocytes. These changes include down-regulation of CD25, recombinase-activating gene (RAG)-1, RAG-2, and pTα; and the up-regulation of TCRα germline transcripts. We term this complete set of changes in gene expression, in vitro maturation. Interestingly, Scid.adh undergoes only a subset of these changes in gene expression following Ab engagement of the pre-TCR. Our findings make two important points. First, because TAC:CD3ε stimulation of Scid.adh induces physiologically relevant changes in gene expression, Scid.adh is an excellent cellular system for investigating the molecular requirements for pre-TCR signaling. Second, Ab engagement of CD3ε signaling domains in isolation (TAC:CD3ε) promotes in vitro maturation of Scid.adh, whereas engagement of CD3ε molecules contained within the complete pre-TCR fails to do so. Our current working hypothesis is that CD3ε fails to promote in vitro maturation when in the context of an Ab-engaged pre-TCR because another pre-TCR subunit(s), possibly TCRζ, qualitatively alters the CD3ε signal.
Development of immature thymocytes into immunocompetent T cells is marked by ordered changes in expression of the CD4 and CD8 differentiation Ags (1, 2). Early in this process, thymocytes execute a series of critical gene rearrangements necessary to produce the Ag-binding subunits (β and α) of the TCR. These rearrangement events occur as CD4−CD8− double-negative (DN)3 thymocytes develop through the CD8low immature single-positive (ISP) stage to the CD4+CD8+ double-positive (DP) stage. The transition from the DN to DP stage is governed by a checkpoint termed β selection, which stipulates that only those precursors that have maintained the translational reading frame of the rearranged TCRβ gene are allowed to progress to the DP stage (3). β selection is exerted on a DN thymocyte subpopulation that is HSAhighCD44lowCD25+. Upon expression of the TCRβ protein product, these cells undergo proliferative expansion, down-regulate CD25, progress to the ISP stage by up-regulating CD8, and finally develop into DP thymocytes. Those precursors failing to productively rearrange the TCRβ locus die by apoptosis (3, 4).
The precise mechanism whereby maturation of thymocytes to the DP stage is linked to productive rearrangement of the TCRβ locus is unclear, but appears to involve signaling through a surrogate form of the TCR termed the pre-TCR complex (5, 6). The pre-TCR complex comprises CD3γ, δ, ε, and ζ signaling subunits in association with a heterodimer of TCRβ disulfide linked to an invariant 33-kDa subunit termed pre-Tα (pTα) (7, 8). The ability of the pre-TCR to promote development of DN thymocytes to the DP stage is attenuated in mice lacking any of the receptor subunits (except CD3δ) and in mice with mutations preventing rearrangement of the TCRβ gene locus (e.g., recombinase-activating gene (RAG) deficiency or the scid mutation in DNA-dependent protein kinase) (6, 8, 9, 10, 11, 12, 13). The developmental arrest of RAG-deficient thymocytes can be overcome (i.e, development to the DP stage is restored) by Ab engagment of: 1) chimeric transgenes consisting of the (human IL-2Rα (TAC)) exodomain fused to either the CD3ε or TCRζ cytosolic signaling domains (TAC:CD3ε and TAC:TCRζ) (14); or 2) partial CD3 complexes termed clonotype-independent CD3 (CIC) complexes that are expressed on the surface of DN thymocytes (15, 16, 17). Thus, Ab stimulation of CIC as well as the TAC:CD3ε and TAC:TCRζ chimeras can mimic pre-TCR function.
Most cell surface receptor complexes are activated by ligand engagement, for which Ab stimulation frequently serves as an effective surrogate. However, this does not appear to be true for the pre-TCR. Ab engagement of surface pre-TCR complexes in vivo fails to promote maturation of thymocytes to the DP stage, instead arresting their development at the CD8 ISP stage (18, 19, 20). Moreover, removal of the potential ligand-binding exodomains of TCRβ and pTα does not abrogate the ability of the pre-TCR to support development of thymocytes to the DP stage (21, 22). Taken together, these data suggest that ligand engagement of surface pre-TCR complexes may not be responsible for pre-TCR activation in vivo. Ligand-independent models of pre-TCR triggering have also been proposed, and these can be subdivided into two categories. The first proposes that assembly of TCRβ with the remaining pre-TCR subunits and subsequent deposition of the complete complex on the cell surface are sufficient to trigger pre-TCR signaling, even in the absence of ligand engagement (22). The second proposes that surface expression of pre-TCR complexes may not be necessary and that pre-TCR signaling is instead triggered internally, while the pre-TCR is still en route to the cell surface (23, 24). These models are currently unresolved.
A more precise understanding of how pre-TCR signals are triggered and initiate thymocyte differentiation has been hindered by the lack of an in vitro system amenable to both biochemical and genetic analysis. Such in vitro systems have been indispensable tools in the study of T cell activation (25), T cell development from the DP to single-positive stage (26, 27, 28), and B cell development (29, 30, 31, 32). In this study, we establish the Scid.adh thymic lymphoma as an in vitro model system with which to investigate proximal pre-TCR signal transduction and associated changes in gene expression. Scid.adh is phenotypically similar to an immature thymocyte just before β selection (HSAhighCD44lowCD25+) and exhibits the ability to differentiate in vitro. In response to stimulation with the TAC:CD3ε chimera, which promotes maturation of normal DN thymocytes to the DP stage in vivo (14), Scid.adh down-regulates mRNAs encoding pTα, RAG-1, and RAG-2 (33, 34, 35), and down-regulates surface expression of CD25, one of the earliest phenotypic indicators of pre-TCR signaling in vivo (36, 37). In addition, TAC:CD3ε stimulation of Scid.adh induces expression of germline transcripts from the TCRα locus associated with the maturation of DN thymocytes to the DP stage (34, 38). We term this collective set of changes in gene expression, “in vitro maturation”. Importantly, Ab engagement of pre-TCR complexes expressed on the surface of Scid.adh induces only a subset of the changes in gene expression defined as in vitro maturation, despite the fact that pre-TCR complexes contain CD3ε subunits. The inability of Ab engagement of CD3ε to fully induce in vitro maturation despite the presence of CD3ε within the pre-TCR complex suggests that another pre-TCR subunit(s) may alter the CD3ε signal and prevent it from coupling to the requisite downstream signaling cascades. Moreover, the failure of Ab engagement of surface pre-TCR complexes to fully induce in vitro maturation of Scid.adh is not consistent with ligand-dependent models of pre-TCR activation.
Materials and Methods
Cell lines and Abs
Scid.adh, SL-12, and SL343 are all spontaneous thymic lymphomas isolated from mice bearing the scid mutation (scid mice) and adapted to growth in culture. SL-12 (TCR−) (39) and SL-343 (TCR−) (40) were obtained from Dr. M. Bosma (Fox Chase Cancer Center, Philadelphia, PA). SL-12β.12 was generated by electroporation of pXS-2B4β into SL-12, as previously described (7). The DP α/βTCR+ thymic lymphoma VL3-3 M2 was obtained from Dr. C. Guidos (Hospital for Sick Children, Toronto, Canada) (41). The retroviral producer line φ2 (NIH 3T3) was obtained from Dr. P. Tsichlis (Thomas Jefferson University, Philadelphia, PA) (42). The anti-TAC mAb-producing hybridoma hd245/332 was obtained from the American Type Culture Collection (Manassas, VA) with the permission of Dr. T. Waldman (National Institute of Child Health and Human Development (NICHD), National Cancer Institute, Bethesda, MD) (43). All lines were maintained in RPMI supplemented, as previously described (44). The following mAbs were used to stimulate cells in culture: anti-TCRβ (H57-597) (45); anti-TAC (hd245/332) (43). The following fluorochrome-conjugated mAbs (PharMingen, San Diego, CA) were used in flow cytometry: anti-CD5-PE (53-7.3), anti-CD25-FITC (7D4), anti-TAC-FITC (M-A251), anti-TCRβ-FITC (H57-597), and anti-human CD3-FITC (Leu4) (UCHT1). The following Abs were used in immunoprecipitations: anti-TCRβ (H57-597), anti-CD3γ/ε (7D6), anti-CD3δ (R9), and anti-TCRζ (551). Anti-pTα Ab used in the recapture assay has been previously described (7).
Cell stimulation and flow cytometry
All stimulations were conducted on plate-bound Ab at 37°C. Cells were harvested at early- to mid-log phase, counted, and plated in 2 ml of RPMI (2–3 × 105/well) in 24-well tissue culture plates that had been precoated overnight at 37°C with either anti-TCRβ or anti-TAC at 5 μg/ml in PBS. After the indicated time of stimulation, cells were removed from the wells by repeated pipetting. For each sample, 106 cells were washed in staining buffer (0.1% BSA/0.1% sodium azide/HBSS), stained for 30 min at 4°C with the indicated fluorochrome-conjugated mAb, washed twice in staining buffer, and analyzed by flow cytometry using a FACScan cytometer (Becton Dickinson, San Jose, CA) and CELLQUEST software (Becton Dickinson). Dead cells were excluded using the vital dye propidium iodide. For RT-PCR analysis, stimulated and unstimulated cells were prepared for flow cytometry as above and isolated using a dual laser/dye laser flow cytometer (FACStarPlus; Becton Dickinson).
Retroviral gene transfer
A φ2 retroviral packaging cell line expressing the MFG-2B4 β retroviral vector was obtained from Dr. L. Spain (Wistar Institute, Philadelphia, PA) (46). The murine stem cell virus (MSCVneo) expression vector was obtained from Dr. R. Hawley (University of Toronto, Toronto, Canada). pXS-TAC and pXS-TAC:CD3ε were obtained from Dr. J. Bonifacino (NICHD, National Institute of Health, Bethesda, MD). The TAC:CD3ε chimera contains the TAC exo- and transmembrane domains linked to the cytoplasmic tail of CD3ε, thus eliminating the motifs in CD3ε that promote assembly with other TCR components (47). The TAC and TAC:CD3ε inserts were shuttled from pXS through the PCR cloning vector pCR2.1 (Invitrogen, SanDiego, CA) into MSCVneo by PCR using standard methodology (48). Briefly, the coding regions of TAC and TAC:CD3ε were amplified using oligonucleotide primers that flank the translational start and stop sites, and that were appended with linkers encoding the HpaI (5′) and XhoI (3′) restriction sites. The amplified inserts were verified by sequencing and directionally subcloned into MSCVneo using standard methodology. MSCVneo, MSCVneo-TAC, and MSCVneo-TAC:CD3ε were separately transfected into the φ2 retroviral packaging line using lipofectamine, according to manufacturer’s specifications (Life Technologies, Grand Island, NY). Stable viral-producing cells were selected in 500 μg/ml G418. The specified thymic lymphoma lines were retrovirally infected by coculture. Briefly, in a single well of a six-well culture dish, 3 × 105 lymphoma cells and 106 φ2 retroviral producer cells were seeded in 3 ml RPMI containing 4 μg/ml polybrene, and incubated for 48 h at 37°C. Infected thymic lymphomas were removed from the producer monolayer and selected in 500 μg/ml G418. Infected thymic lymphoma cells expressing homogenously high levels of TCRβ, TAC, or TAC:CD3ε were isolated by flow cytometry as above.
Semiquantitative RT-PCR and Southern blot analysis
For each treatment condition, a total of 5 × 106 thymic lymphoma cells was sorted into microfuge tubes by flow cytometry. Total RNA was isolated into a final volume of 50 μl of RNase-free H2O using the RNeasy RNA purification system (Qiagen, Valencia, CA). All RNA samples were treated with DNase I (Life Technologies) before performing first strand cDNA synthesis (1 μg RNA/reaction) using random primers and the Superscript preamplification system (Life Technologies). Titrated amounts of cDNA were amplified by PCR (19 cycles) using primers specific for: β-actin, RAG-1, RAG-2, pTα, and TCR-Cα (see Table I). Forty percent of each PCR reaction was resolved on a 1.5% agarose electrophoresis gel and blotted onto Nytran membranes (Scheicher & Schuell, Keene, NH), which were rinsed in phosphate buffer (111 mM Na2HPO4, 87 mM NaH2PO4-H20, pH 6.8) and allowed to air dry overnight. Prehybridization of membranes was conducted for 4 h at 67°C in pre-hyb solution (5× SSC, 5× Denhardt solution, 0.5% SDS, 1 mM EDTA, 0.1 mg/ml salmon sperm DNA). cDNA probes were labeled with [32P]dCTP using the Prime-It II random primer labeling kit (50 ng probe/reaction; Stratagene, La Jolla, CA) and purified on NucTrap probe purification columns (Stratagene). Labeled probe was added directly to membranes in pre-hyb solution (106cpm/ml), and probe hybridization was allowed to proceed overnight at 67°C. Blots were washed sequentially in 2× and 1× SSC/0.5% SDS before quantification using a Fuji phospor imager and Fuji MacBas V2.2 software (Fuji Photo Film, Tokyo, Japan).
|Gene .||5′ Oligonucleotide .||3′ Oligonucleotide .||Fragment Size (bp) .|
|Gene .||5′ Oligonucleotide .||3′ Oligonucleotide .||Fragment Size (bp) .|
The annealing temperature for specific primer pairs are as follows: β-actin and pTα, 54°C; RAG-1, RAG-2, and TCR-Cα, 60°C.
Biotin labeling, immunoprecipitation, and immunoblotting
Scid.adh cells were biotin labeled, as described, after which cell viability was consistently >98% (49). Cell lysis, immunoprecipitation, recapture assays, electrophoresis, and immunoblotting were as described previously (7). Briefly, digitonin extracts of biotin-labeled Scid.adh cells were exhaustively cleared of biotin-labeled proteins by sequential passage over streptavidin-agarose (Pierce, Rockford, IL). The SDS-eluted proteins were resolved by SDS-PAGE, blotted onto Immobilon polyvinylidene difluoride membranes (NEN Lifescience Products, Boston, MA), and probed with anti-CD3ε Ab (HMT3-1). Bound Ab was visualized with HRP-conjugated protein A (HRP-protein A; Kirkegard and Perry, Gaithersburg, MD), followed by chemiluminescence (Renaissance, NEN Lifescience Products).
CD3ε signals induce maturation of the Scid.adh thymic lymphoma in vitro
Ab stimulation of a transgenic TAC:CD3ε chimera elicits signals capable of substituting for the pre-TCR complex, as evidenced by the ability of such signals to promote development of pre-TCR-deficient DN thymocytes to the DP stage (14). In contrast, Ab triggering of the pre-TCR complex itself arrests development at the CD8 ISP stage (18, 19, 20). As a first step toward understanding why Ab engagement of the pre-TCR fails to promote development to the DP stage, we sought to establish an in vitro system in which to study pre-TCR function. A series of scid thymic lymphomas was evaluated to identify cell lines that exhibit changes in gene expression in response to TAC:CD3ε stimulation similar to those observed following pre-TCR signaling (triggered by productive rearrangement of TCRβ) in normal thymocytes in vivo. We reasoned that studying proximal signaling events in a cell line in which physiologically relevant gene expression changes are induced is critical, because those changes in gene expression attest to the relevance of the proximal signals. Scid thymic lymphomas were chosen because scid thymocytes are unable to mediate V(D)J recombination, and so their development is arrested just before β selection at the HSA+CD44−CD25+ stage (50). Consequently, thymic lymphomas from scid mice, being phenotypically similar to their nontransformed counterparts, might retain the ability to respond to pre-TCR signals in a physiologically relevant manner. We identified such a cell line, Scid.adh. The Scid.adh lymphoma, although CD4−CD8+, resembles most closely the phenotype of a DN thymocyte in that it is HSAhighCD44lowCD25+ (Fig. 1,A), expresses message for pTα, and does not express a functional TCRβ gene, as confirmed by both biochemical and genetic analysis (data not shown). After 24 h of stimulation with plate-bound anti-TAC mAb, surface expression of the activation Ag CD5 was markedly increased on TAC:CD3ε-expressing Scid.adh cells (Fig. 1,B), as was expression of other differentiation Ags including CD27 and the costimulatory molecule CD28; however, surface expression of the CD4 and CD8 coreceptor molecules was not altered (Fig. 1,B). More importantly, TAC:CD3ε stimulation sharply down-regulated the surface expression of CD25 (Fig. 1 B), which is a hallmark of pre-TCR activation in vivo (8, 51). It should be noted that CD25 down-regulation was only observed after stimulation of Scid.adh lymphoma cells, whereas TAC:CD3ε stimulation failed to do so in three other lymphomas tested: 1) SL12-β12, DN, pre-TCR+; 2) SL-343, DP, pre-TCR−; and 3) VL3-3 M2, DP, αβTCR+ (data not shown).
In addition to down-modulating CD25, pre-TCR signaling of normal thymocytes in vivo (or anti-CD3ε stimulation of RAG-deficient DN thymocytes) greatly reduces the abundance of pTα and RAG mRNA (33, 34, 35), and up-regulates sterile TCRα transcripts, which precedes the initiation of gene rearrangement at the TCRα locus (34, 35, 38). To determine whether TAC:CD3ε stimulation alters gene expression in Scid.adh in a similar way, TAC:CD3ε-expressing Scid.adh cells were cultured on plate-bound anti-TAC mAb for 24 h, following which those that had fully (CD25low, Fig. 2,A) or partially (CD25int; Fig. 2,A) down-modulated CD25 were isolated using flow cytometry. The effect of TAC:CD3ε stimulation on expression of pTα, RAG-1, RAG-2, and TCR-Cα mRNA in these populations was assessed using RT-PCR (Fig. 2,B and Table II). In TAC:CD3ε-stimulated CD25low Scid.adh cells, mRNA encoding full-length pTαa was decreased 30-fold, while that of RAG-1 and RAG-2 were each decreased ∼20-fold. Moreover, the CD25low cells exhibited a 30-fold increase in germline TCR-Cα transcripts as well as a 40-fold increase in the ISP stage-specific transcript Jα49 (data not shown) (38). Curiously, while the up-regulation of TCR-Cα mRNA in CD25int Scid.adh cells was comparable with that in CD25low cells, the down-regulation of pTαa and RAG-1/2 mRNA in the CD25int cells was modest (Table II). Thus, the extent to which RAG and pTα mRNA levels were down-regulated correlates tightly with the extent to which CD25 surface expression was reduced, but not with the extent of TCR-Cα induction. Taken together, the changes in surface phenotype and in molecular markers induced following TAC:CD3ε stimulation of Scid.adh are consistent with the way a normal thymocyte responds to β selection.
|.||TAC:CD3ε CD25+ Unstimulated .||TAC:CD3ε CD25low Stimulated .||TAC:CD3ε CD25int Stimulated .||TCRβ CD25+ Unstimulated .||TCRβ CD25+ Stimulated .|
|.||TAC:CD3ε CD25+ Unstimulated .||TAC:CD3ε CD25low Stimulated .||TAC:CD3ε CD25int Stimulated .||TCRβ CD25+ Unstimulated .||TCRβ CD25+ Stimulated .|
The numerical values shown for each specific cDNA were obtained from semiquantitative RT-PCR analyses carried out on both unstimulated and stimulated Scid.adh cells as described in Figs. 2 and 4. All signals were quantitated and normalized to β-actin for each sample using a Fuji PhosphoImager and Fuji MacBAS V2.2 software. These values are expressed as fold changes relative to expression of the particular mRNA in unstimulated Scid.adh-TAC:CD3ε control. The values shown are derived from one of three separate experiments, all of which yielded similar results.
Ab engagement of surface pre-TCR complexes promotes only partial in vitro maturation of Scid.adh
While Ab engagement of surface TAC:CD3ε molecules is able to mimic pre-TCR function in vivo and promote development of DN thymocytes to the DP stage (14), engagement of surface pre-TCR complexes using anti-TCRβ mAb fails to do so, instead arresting thymocytes at the CD8 ISP stage (18, 19, 20). Consequently, we asked whether anti-TCRβ stimulation of surface pre-TCR complexes on Scid.adh would induce in vitro maturation. Because of the scid defect in V(D)J recombination, Scid.adh is unable to rearrange its endogenous TCRβ locus (39). To facilitate expression of surface pre-TCR complexes, Scid.adh cells were retrovirally transduced with a cDNA encoding the 2B4 TCRβ subunit (Scid.adh-TCRβ; Figs. 3 and 4). The TCRβ cDNA induced Scid.adh to express TCRβ on the cell surface, as measured by flow cytometry (Fig. 4). Moreover, biochemical analysis of surface biotin-labeled Scid.adh-TCRβ cells revealed that pTα-TCRβ heterodimers were coprecipitated with Ab to CD3γε, CD3δ, and TCRζ, indicating that the pre-TCR complexes expressed by these cells comprised all of the known pre-TCR subunits (TCRβ, pTα, CD3γδε and ζ) (Fig. 3). Stimulation of the pre-TCR-expressing Scid.adh cells with plate-bound anti-TCRβ mAb induced a substantial increase in expression of the activation Ag CD5 (Fig. 4), verifying that the anti-TCRβ mAb did indeed trigger the pre-TCR. However, anti-TCRβ-induced signals caused only a marginal down-modulation of CD25 relative to that caused by TAC:CD3ε stimulation (Fig. 4). Furthermore, the inability of Ab engagement of the pre-TCR to down-modulate CD25 was not unique to anti-TCRβ mAb, as stimulation of the pre-TCR complex with anti-CD3ε mAb also failed to down-modulate CD25 (data not shown). Because anti-TCRβ stimulation of Scid.adh-TCRβ cells did not significantly alter CD25 expression, we wanted to determine how anti-TCRβ stimulation affected other genes whose expression is altered during in vitro maturation: pTα, RAG-1, RAG-2, and TCR-Cα (Fig. 2,B). Semiquantitative RT-PCR analysis revealed that anti-TCRβ stimulation of Scid.adh-TCRβ cells markedly increased the levels of TCR-Cα transcript (∼70-fold) (Fig. 5 and Table II) as well as the ISP stage-specific transcript Jα49 by 50-fold (data not shown) (38); however, the induction of TCR-Cα levels is not well correlated with CD25 down-modulation even in TAC:CD3ε-stimulated cells (compare CD25low to CD25int; Table II). In agreement with its failure to down-modulate CD25, anti-TCRβ stimulation was unable to significantly down-modulate pTα or RAG mRNA levels (Fig. 5 and Table II). Taken together, these data demonstrate that Ab engagement of the complete pre-TCR complex is unable to induce in Scid.adh the complete set of changes in gene expression defined as in vitro maturation. More specifically, these data demonstrate that Scid.adh maturation can be induced by CD3ε signals when CD3ε is engaged in isolation as a TAC:CD3ε chimera, but not when CD3ε is engaged in the context of the complete pre-TCR complex. Furthermore, the inability of pre-TCR ligation to fully promote in vitro maturation does not appear to be specific to the Vβ3 TCRβ transgene used above, as Ab engagement of Vβ8-containing pre-TCR complexes expressed in Scid.adh also fails to induce in vitro maturation (data not shown).
Coengagement of pre-TCR and TAC:CD3ε complexes on the cell surface does not disrupt TAC:CD3ε-driven in vitro maturation of Scid.adh
There are three potential explanations as to why in vitro maturation of Scid.adh can be fully induced by signals from isolated CD3ε molecules, but only partially by signals from CD3ε molecules within the pre-TCR complex. First, the number of CD3ε subunits expressed on the surface of Scid.adh as part of the pre-TCR complex may be significantly lower than the number expressed on the surface in the form of the TAC:CD3ε chimera, resulting in a signal that falls below the threshold required to induce in vitro maturation. Second, expression of TCRβ in Scid.adh may render the cell refractory to induction of in vitro maturation. Third, the association of CD3ε with the other pre-TCR subunits may alter its ability to signal maturation. We tested these possibilities by introducing a TCRβ cDNA into cells already expressing TAC:CD3ε. First, to assess the ratio of surface CD3ε expressed as TAC:CD3ε to that expressed in a pre-TCR-associated form, surface biotin-labeled proteins were isolated using avidin-agarose and then blotted with anti-CD3ε Ab. Because the amount of immunoreactivity found in the TAC:CD3ε form was approximately equivalent to that of the CD3ε monomer, we conclude that the expression levels of TAC:CD3ε and pre-TCR-associated CD3ε are equivalent (Fig. 6). When these dual-expressing Scid.adh cells were stimulated with anti-TCRβ mAb, they were unable to down-modulate CD25 (Fig. 7). Consequently, differences in expression level do not explain the inability of the pre-TCR to fully induce in vitro maturation. Second, to assess whether expression of TCRβ renders Scid.adh unable to undergo in vitro maturation, we stimulated the dual-expressing Scid.adh line with anti-TAC mAb. Similar to cells expressing only the TAC:CD3ε chimera, TAC:CD3ε stimulation induced in vitro maturation of Scid.adh as defined by down-modulation of CD25, indicating that expression of TCRβ does not render Scid.adh unable to undergo in vitro maturation (Fig. 7). The remaining possibility that association of CD3ε with the other subunits of the pre-TCR alters its signaling ability could be manifested in the following two ways: 1) CD3ε transduces a dominant signal that arrests in vitro maturation; or 2) CD3ε transduces a limited signal that neither promotes development nor interferes with induction of development by another stimulus. To distinguish between these possibilities, we simultaneously stimulated Scid.adh via TAC:CD3ε and the pre-TCR. In response to costimulation with both anti-TAC and anti-TCRβ Ab, Scid.adh down-modulated CD25 expression, indicating that coengagement of the pre-TCR was not able to disrupt in vitro maturation of Scid.adh in response to TAC:CD3ε stimulation (Fig. 7). This suggests that the failure of CD3ε molecules to induce in vitro maturation within the context of the pre-TCR is not due to transduction of a dominant-negative signal. Rather, it appears that Ab engagement of the pre-TCR transduces a limited signal that is incompetent to induce full in vitro maturation. Taken together, these data serve to underscore three important points. First, of the thymic lymphomas examined, Scid.adh was the only one whose responses to TAC:CD3ε stimulation paralleled those of normal thymocytes following β selection in vivo. Indeed, TAC:CD3ε stimulation, which mimics pre-TCR function in vivo by driving development of DN thymocytes to the DP stage (14), induces down-modulation of CD25, pTα, RAG-1, and RAG-2 in Scid.adh in vitro. Second, while engagement of isolated CD3ε molecules (i.e., the TAC:CD3ε chimera) is capable of inducing the complete set of changes in gene expression associated with in vitro maturation of Scid.adh, engagement of surface pre-TCR complexes is not, despite the fact that the pre-TCR also contains CD3ε. This suggests that another pre-TCR component(s) may alter the ability of CD3ε to promote maturation when it is engaged while in the context of the complete pre-TCR complex. Third, coligation of the intact pre-TCR with TAC:CD3ε does not disrupt the ability of TAC:CD3ε to induce in vitro maturation, suggesting that Ab engagement of the pre-TCR does not induce a dominant-negative signal.
Productive rearrangement of the gene segments comprising the TCRβ locus enables assembly of a surrogate receptor complex termed the pre-TCR, which transduces signals that promote development of DN thymocytes to the DP stage (51). However, the mechanism whereby pre-TCR signaling is initiated in vivo remains unclear. Moreover, our ability to gain insight into the molecular events that underlie pre-TCR triggering has been hampered by the lack of an established in vitro system amenable to genetic as well as biochemical analysis. We report in this study the identification of a thymic lymphoma, Scid.adh, whose response to pre-TCR signals parallels that of normal thymocytes in vivo. Stimulation with the signaling chimera TAC:CD3ε, which is able to mimic pre-TCR function in vivo (14), is able to drive in vitro maturation of Scid.adh, as defined by the following effects: 1) down-regulation of CD25 surface expression; 2) down-regulation of mRNA encoding pTα, RAG-1, and RAG-2; and 3) up-regulation of TCRα germline transcripts. All of the effects listed above have been shown to occur following pre-TCR activation (or a stimulus mimicking pre-TCR activation) in normal thymocytes in vivo (1, 33, 34, 35, 38, 51). Interestingly, however, Ab engagement of pre-TCR complexes expressed on the surface of Scid.adh fails to induce the complete set of gene expression changes associated with in vitro maturation. It is particularly curious that Ab engagement of surface pre-TCR complexes fails to fully drive in vitro maturation of Scid.adh because pre-TCR complexes contain the CD3ε subunit that, when engaged in the form of an isolated TAC:CD3ε signaling chimera, is quite capable of inducing Scid.adh maturation.
Models attempting to explain how pre-TCR signaling is initiated can be divided into two categories: ligand dependent and ligand independent. Currently, there are several pieces of evidence that are inconsistent with a ligand-dependent model of pre-TCR triggering. First, analysis of the efficiency with which TCRβ transgenic DN precursors differentiate to the DP stage suggests that this transition is not constrained by a limiting number of intrathymic “niches” or extracellular ligands, as is true of the Ag-driven selection events that promote maturation of DP thymocytes to the CD4+ or CD8+ single-positive stage (52). Second, the absence of peptide-bearing major histocompatibility Ags (the ligand of the TCRαβ complex) does not disrupt pre-TCR function (53). Third, the ability of the pre-TCR to promote maturation of DN thymocytes to the DP stage does not require the potential ligand-binding exodomains of either pTα or TCRβ (21, 22, 54). Finally, as mentioned above, Ab engagement of surface pre-TCR complexes not only blocks development at the CD8 ISP stage (18, 19, 20), but also fails to induce in vitro maturation of Scid.adh. Taken together, these observations are most consistent with ligand-independent triggering of pre-TCR signals. Ligand-independent pre-TCR activation might be triggered by pre-TCR complexes at the cell surface or alternatively, during assembly or intracellular transport. At present, there is no clear evidence to distinguish these possibilities.
Although Ab engagement of surface pre-TCR complexes fails to fully induce in vitro maturation of Scid.adh or complete maturation of normal thymocytes to the DP stage, there is an important difference in these responses. In vitro, Ab engagement of complete pre-TCR complexes on Scid.adh fails to down-regulate CD25 or RAG expression, whereas these molecules are down-regulated following anti-TCRβ stimulation of normal thymocytes in vivo (19, 20). One possible explanation for this discrepancy is that Ab-induced pre-TCR signals in vivo may be augmented by costimulatory interactions between the thymocyte and the thymic stroma, which might mask the inadequacy of the pre-TCR signals. Indeed, we have recently found that Ab engagement of the pre-TCR complex on Scid.adh more effectively down-modulates CD25 expression when coengaged with CD2 (data not shown). Importantly, CD2 is expressed by thymocytes at the stage of β selection (55, 56) and has been shown to act synergistically with αβTCR signals to promote later stages of thymocyte development (57). Thus, to gain insight into the precise role that pre-TCR signaling plays in promoting development, it becomes necessary to separate the contribution of pre-TCR signals from that of stromal costimulation. It is curious that Ab stimulation of surface pre-TCR complexes (comprising known signaling subunits CD3γδε and TCRζ) is not able to mimic pre-TCR function, because Ab ligation of isolated CD3 subunits (i.e., TAC:CD3ε or CIC) is able to do so (14, 15, 16, 17). The mechanistic basis for this difference is currently unclear, but is likely to be related to the content of signaling subunits in each of the complexes. TAC:CD3ε contains only the cytoplasmic tail of CD3ε and its single immunoreceptor tyrosine-based activation motif (ITAM), whereas each pre-TCR complex contains multiple ITAMs from four different signaling subunits, and potentially other motifs in the cytoplasmic tail of pTα (58). Indeed, association of CD3ε with the signaling subunits in the pre-TCR may alter the resultant signal either quantitatively or qualitatively. Regarding quantitative effects, we have excluded the trivial explanation in Scid.adh that the level of expression of the isolated TAC:CD3ε differs from that of CD3ε in the context of the pre-TCR complex (Fig. 6). In addition, we have determined that Ab-induced ligation of either TAC:CD3ε or pre-TCR in Scid.adh results in an equivalent level of phosphorylation of the tyrosine kinase ZAP70 (manuscript in preparation). Moreover, the induction of CD5 expression following ligation of the pre-TCR complex exceeds that induced following TAC:CD3ε stimulation (compare Figs. 1 A and 4). Taken together, these data indicate that the signal intensity produced by Ab engagement of the pre-TCR is at least as great as that induced by TAC:CD3ε stimulation. Yet, pre-TCR ligation fails to fully drive in vitro maturation, irrespective of the level at which the pre-TCR complex is expressed on the cell surface (data not shown). Thus, a more likely explanation for the inability of Ab-stimulated pre-TCRs to drive in vitro maturation of Scid.adh is that the resultant signal is qualitatively distinct from that produced by TAC:CD3ε.
If association of CD3ε with other pre-TCR subunits qualitatively alters the resultant signal, which subunit(s) is responsible for this effect? We propose that TCRζ is the most attractive candidate for four reasons. First, the binding of phosphorylated CD3ε with ZAP70 following Ab stimulation is enhanced in TCR complexes lacking ζ or its cytoplasmic signaling domain (59). Second, Ab engagement of CIC (heterodimers of CD3γε and CD3δε that lack TCRζ) promotes development of DN thymocytes all the way to the DP stage (15, 16, 17). Third, while Ab engagement of the pre-TCR complex arrests thymocyte development at the CD8 ISP stage (18, 19, 20), Ab stimulation of the pre-TCR complex expressed on thymocytes in TCRζ-deficient mice appears to promote development to the DP stage (60). Fourth, it was demonstrated recently that productive stimulation of the αβTCR complex results in complete phosphorylation of TCRζ, while stimuli producing unresponsiveness are unable to induce complete phosphorylation of TCRζ (61). These differences in phosphorylation may be a determining factor in the nature of the resultant signal (61). For all of these reasons, we think that the TCRζ subunit is the most likely candidate to be responsible for limiting the ability of CD3ε molecules in Ab-engaged pre-TCR complexes to promote maturation. Interestingly, it would appear that this proposed inhibitory effect of TCRζ can occur only from within the context of the pre-TCR complex, as a TAC:TCRζ chimera can drive thymocyte development as efficiently as TAC:CD3ε (14).
It is paradoxical that the presence of TCRζ in Ab-engaged pre-TCR complexes might be inhibitory because TCRζ is required for the pre-TCR to promote development of thymocytes to the DP stage (10, 62, 63). However, this apparent paradox might be explained in the following manner. Currently, the preponderance of evidence suggests that pre-TCR signaling is not induced by ligand engagement. Thus, it is possible that the requirement of TCRζ for pre-TCR function may be either to stabilize pre-TCR structure or alternatively to promote transport of the pre-TCR to the cell surface, where its signaling is triggered in response to nonspecific, low affinity interactions with thymic stroma. Consistent with this viewpoint, truncated TCRζ transgenes that lack ITAMs are able to compensate for TCRζ deficiency and restore development to the DP stage (64). In contrast, engagement of surface pre-TCR complexes by high affinity ligands (e.g., anti-TCRβ Ab) could induce complete TCRζ phosphorylation and signal amplification, possibly perceived by the developing thymocyte as autoreactivity and resulting in developmental arrest. In the absence of TCRζ, even such high affinity ligands would be unable to induce the same extent of signal amplification. If the Ab-engaged pre-TCR complexes are inducing inhibitory signals, at least in Scid.adh, those inhibitory signals cannot disrupt the productive signals induced by TAC:CD3ε. Experiments are in progress to assess the role of TCRζ phosphorylation in pre-TCR function.
Progress in understanding how pre-TCR signaling is triggered in vivo has been hampered by the lack of an in vitro model system in which pre-TCR function can be studied both genetically and biochemically. Our analysis of the Scid.adh thymic lymphoma represents the first description of an in vitro tumor model system that responds to stimuli that mimic pre-TCR function in vivo (i.e., TAC:CD3ε) by altering gene expression in a fashion that closely parallels that of normal thymocytes undergoing β selection in vivo. In fact, the chief utility of Scid.adh is that the induced changes in gene expression are physiologically relevant, providing confidence that the proximal signaling events are also representative of those which occur in normal thymocytes. It should be noted that while the in vitro maturation of Scid.adh does closely parallel the behavior of normal thymocytes, Scid.adh differs from normal thymocytes in that Scid.adh does not up-regulate CD4 to become DP following TAC:CD3ε stimulation. Another tumor model system that has been used to study proximal pre-TCR signaling events is the scid thymic lymphoma SCB.29 (65), but in that study the authors did not attempt to determine whether the proximal pre-TCR signals produced physiologically relevant changes in gene expression. An earlier study, however, did demonstrate that Ab engagement of pre-TCR complexes expressed on SCB.29 resulted in CD25 up-regulation, rather than the down-regulation that results from pre-TCR signaling in normal thymocytes (66). Consequently, the appropriateness of SCB.29 as a model system to study pre-TCR signaling and its downstream effects on gene expression remains to be established.
Using the Scid.adh model system, we intend to dissect the signaling pathways that lie downstream of the pre-TCR and to investigate the way in which these pathways are linked to the changes in gene expression responsible for promoting thymocyte development. By comparing the signaling events triggered by a stimulus that fully induces in vitro maturation of Scid.adh (Ab engagement of TAC:CD3ε) with one that does so only partially (Ab engagement of the complete pre-TCR), we should gain insight into the molecular requirements for productive pre-TCR signaling.
We thank Dr. L. Spain for kindly providing the ψ2 -MFG-2B4β-expressing retroviral producer cell line, Dr. R. Hawley for the retroviral vector MSCVneo, Dr. T. Waldman for the anti-TAC hybridoma (hd245/332), and Dr. J. Bonifacino for TAC and TAC:CD3ε cDNA. We thank Drs. K. Campbell, D. Kappes, G. Koretzky, J. Clements, J. Punt, L. Spain, and A. Singer for critical review of this manuscript.
This research was supported by grants from the National Institutes of Health (R29 CA-73656 and Core Grant CA-06927) and from the Human Frontier Science Program (3752-01). M.C. was supported by National Institutes of Health Training Grant CA-09035-23 and a postdoctoral fellowship from the Cancer Research Institute. M.A.B. was supported by National Institutes of Health Training Grant AI-07492 and by a fellowship from the Arthritis Foundation.
Abbreviations used in this paper: DN, double negative; CIC, clonotype-independent CD3; DP, double positive; HSA, heat stable Ag; ISP, immature single positive; ITAM, immunoreceptor tyrosine-based activation motif; MSCV, murine stem cell virus; RAG, recombinase-activating gene; (TAC), IL-2Rα.