HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Krotkova, A. Articles by Eichmann, K. Search for Related Content PubMed PubMed Citation Articles by Krotkova, A. Articles by Eichmann, K. Pubmed/NCBI databases Gene GEO Profiles HomoloGene UniGene

Delayed and Restricted Expression Limits Putative Instructional Opportunities of V1.1/V2 TCR in / Lineage Choice in the Thymus

Max-Planck-Institut für Immunbiologie, Freiburg, Germany

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Development of and T cells depends on productive rearrangement of the appropriate TCR genes and their subsequent expression as proteins. TCR and TCR proteins first appear in DN3 and DN4 thymocytes, respectively. So far, it is not clear whether this is due to a delayed expression of TCR proteins or to a more rapid progression to DN4 of thymocytes expressing TCR. The answer to this question bears on the distinction between instructive and stochastic models of / lineage decision. To study this question, we first monitored initial TCR protein expression in wild-type and TCR transgenic mice in reaggregate thymic organ cultures. A TCR transgene was expressed in nearly all DN3 and DN4 cells, accelerated DN3 to DN4 transition, and strongly diminished the number of cells that express TCR proteins. In contrast, TCR transgenes were expressed only in a fraction of DN4 cells, did not accelerate DN3 to DN4 transition, and did not reduce the number of DN4 cells expressing TCR proteins. The TCR transgene partially inhibited endogenous TCR rearrangements, whereas the TCR transgenes did not inhibit endogenous TCR rearrangements. Second, we analyzed frequencies of productive TCR and TCR V(D)J junctions in DN3 and DN4 subsets. Most importantly, frequencies of productive TCR rearrangements (V5, V1.1, and V2) appeared unselected in DN3. The results suggest a late and restricted expression of the corresponding TCR, severely limiting their putative instructional opportunities in / divergence.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Based on two classes of specific TCRs, T lymphocytes are divided into two major populations, and T cells. The specificity of both types of TCRs is determined by the rearrangement of V(D)J gene segments during T cell ontogeny. Flexibility in this process can endow the TCRs with pronounced diversity. Due to the process of random rearrangements two of three V(D)J recombinations lead to "out-of-frame" products, since they disrupt the reading frame starting at the ATG initiation codon of the V gene segment. The remaining V(D)J junctions produced in this manner (33%) are "in-frame" and can potentially encode functional TCR proteins.

Both and lineages develop within the thymus and arise from a common progenitor population. Early stages of thymocyte maturation occur in the CD3-negative, CD4/CD8-double negative (DN)² population and have been defined by surface phenotype, TCR gene rearrangement status, and developmental potential. DN thymocytes are subdivided into four populations based on the expression of CD25/CD44 surface markers: DN1 (CD44⁺/CD25^–), DN2 (CD44⁺/CD25⁺), DN3 (CD44^–/CD25⁺), DN4 (CD44^–/CD25^–) (1). TCR rearrangements start at the DN2 stage with TCR and TCR genes followed by TCR (2, 3). Most rearrangements at all three loci take place and are completed at the DN3 stage (2, 3). Thereafter, pathways of and development are quite distinct. In T cells the TCR protein combines with the pre-TCR chain (pT) to form the pre-TCR (4, 5) that stimulates proliferation and maturation to the CD4/CD8 double positive (DP) stage where TCR rearrangement occurs (6), mature TCR are synthesized and their repertoire is selected depending on MHC/peptide interactions (7). T cells with TCR develop in two waves using different V genes, the first restricted to the fetus (V3 and V4) and the second beginning prenatally and continuing into adulthood (V2, V5, V1.1, and V1.2; nomenclature for TCR genes as in Ref. 8). Most T cells rarely express CD4 or CD8 coreceptors (9), have no requirement for classical MHC molecules for Ag recognition (10), do not depend on pT expression for maturation (4), and proliferate less than T cells (11).

The developmental stage of / lineage separation is unknown and the mechanisms that govern vs lineage decision in T cell development are still poorly understood. It is clear that successful development of both types of T cells depends on productive rearrangements of the relevant TCR genes that have to be in-frame and without stop codons in their V(D)J junction. However, it is still not known whether the TCRs play a key role in instructing lineage commitment or merely support the development of precursors already committed in other ways. According to instructive models, initial expression of the TCR or the pre-TCR per se is sufficient to direct bipotent precursor cells to become or T cells, respectively. In contrast, stochastic models postulate that / cell fate is predetermined independently of TCR rearrangements.

Intracellular TCR proteins are first detected in DN3 cells, whereas intracellular TCR and - proteins are first detected in DN4 cells (11, 12). Given that TCR and - genes commence rearrangement before TCR (2, 3), it seems unexpected that their expression should be programmed in such a way that TCR genes get expressed after TCR genes. Alternatively, DN3 cells that happen to express a TCR may rapidly down-regulate CD25, whereas DN3 cells that express a pre-TCR down-regulate CD25 with some delay. In this paper, we attempt to distinguish between such alternative mechanisms. On the one hand, we analyzed the expression of clonotypic TCR proteins in DN3/DN4 subsets of TCR transgenic (tg) mice in which all precursor thymocytes possess either a productive TCR gene, or a pair of productive TCR genes, or both. Previous studies on various TCR tg mice (reviewed in Ref. 13) have so far not addressed the initial phase of expression of TCR proteins. We used reaggregate thymic organ cultures (RTOCs) to study newly arising TCR vs TCR expressing thymocytes. This system allows us to monitor a single wave of differentiation together with the differential proliferation of individual subsets. As a second approach we investigated the selection status of rearranged TCR and TCR genes in DN3 and DN4 subsets. As a reference for nonselected populations, TCR and CD3 x p56^lck knockout (KO) mice were included in these analyses. Taken together, our results obtained in RTOC and by V(D)J junction sequencing suggest a late and restricted expression of V2 and V1.1 resulting in limited instructional opportunities of the corresponding TCR in / divergence.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Mice

C57BL/6 (B6) mice, BALB/c mice, TCR-deficient mice (14), mice carrying P14 TCR transgene (15, 16), and double-transgenic mice for P14 TCR and V1.1V6 TCR were bred in the specific pathogen-free animal facilities of the Max Planck Institute (Freiburg, Germany). TCR transgenic mice (V1.1V6) were kindly provided by Dr. P. Pereira, (Pasteur Institute, Paris, France;17). Mice double-deficient for p56^lck and CD3 were described previously (18). Breedings of the TCR and TCR transgenic mice were monitored by DNA-PCR analysis using primers described in original papers (15, 16, 17, 18). All mice were sacrificed at 4- to 6-wk of age.

Antibodies

The following mAbs, unlabeled or labeled with either FITC, PE, APC, or biotin, were purchased from BD PharMingen (San Diego, CA): anti-CD4 (H129.19), anti-CD8 (53-6.7), anti-TCR (H57-597), anti-CD44 (IM7), anti-CD25 (M1/70), anti-NK1.1 (PK136), anti-Ly9.1 (30C7), anti-Thy1.2 (53-2.1), anti-TCRV2 (UC3-10A6), anti-TCR (GL-3), and anti-CD3 (145-2C11). The hybridoma secreting the Ab specific for TCR V1 (2.11) was kindly provided by Dr. P. Pereira, (Paris, France; Ref. 19). Anti-TCRV1 (2.11) mAb was purified and labeled with FITC in our own laboratory. Streptavidin-Red670 was used for biotin-labeled mAb.

Flow cytometry and thymocyte isolation

Thymic lobes were harvested from mice, teased, and ground into single-cell suspensions. DN cells were enriched by Automax (Miltenyi Biotec, Bonn, Germany) depletion of DP thymocytes. DN3 thymocytes positive and negative for intracellular TCR protein were sorted by negative gating of enriched DN cells for CD4, CD8, CD44, CD3, and NK1.1 (all PE), combined with intracellular staining for TCR (APC) and surface staining for CD25 (FITC). To obtain DN4 cells, thymocytes were negatively gated for CD4, CD8, CD44, CD25, CD3, and NK1.1 (all PE) and separated into four subpopulations using intracellular staining for TCR (pooled anti-TCRV2, anti TCRV1.1, anti-TCR all FITC) and TCR (APC). Thymocytes were sorted on a MoFlow Instrument (DakoCytomation, Freiburg, Germany). Nonspecific binding of mAbs to FcRs was reduced by preincubation with supernatant of anti-FcR mAb 2.4G2 before the staining for 20 min. For surface staining, cells were incubated with saturating amounts of directly conjugated mAbs for 20 min, washed twice, and incubated, when necessary, with streptavidin-Red670. Intracellular staining was done on cells fixed with paraformaldehyde and permeabilized with saponin (Sigma-Aldrich, Heidelberg, Germany) as described (20, 21). Controls for cell surface staining used isotype-matched mAbs labeled with the same fluorochrome; controls for intracellular stainings used blocking with an excess of the same unlabeled mAb. Three- and four-color analyses used a FACSCalibur using CellQuest software (BD Biosciences, Mountain View, CA).

RTOC

RTOC were performed according to Anderson et al. (22) as previously described (12). Stroma cells from day 15 BALB/c embryos (5 x 10⁵) were reaggregated with a similar number of sorted DN3 thymocytes from adult B6 wild-type (WT) or genetically modified mice. Labeling of sorted DN3 cells with CFSE (Molecular Probes, Eugene, OR) was performed as described (12, 23). Analyses used staining protocols as described above and negative gating for Ly9.1, to exclude BALB/c-derived thymocytes.

PCR, cloning, and sequencing of TCR VDJ junctions

DNA was prepared from the sorted populations by the proteinase K-phenol/chloroform method and recovered by ethanol precipitation. DNA was amplified in final volume of 25 ml containing PCR buffer (1 x), 0.5 mM each primer, 200 mM each dNTP (Pharmacia, Freiburg, Germany), and 0.2 U Supertaq (HT Biotechnology, Cambridge, U.K.). Each of the 35 cycles consisted of 20 s at 94°C, 30 s at 55°C (for all TCR primer pairs), and then 45 s at 72°C. PCR was performed using a Peltier Thermal Cycler (MJ Research, Cambridge, MA). TCRV8DJ7 and TCRV5DJ7 junctions were amplified using primers and conditions described in (24, 25, 26). The sequence of the oligonucleotides used as primers for PCR amplification of junctions are as follows: V1.1 (ACACAGCTATACATTGGTACCGGC), V2 (AGAGTTTCTATTATATGTCCT TGC), J1 (TACTATGAGCTTAGTTCCTTCTGC), J4 (TACTACGAGCTTTGTCCCTTTGGC), V5 (GGTCAGTGTCTGGGATGCAGACGC), J1 (CACAGTCACTTGGGTTCCTTGTCC) (nomenclature for TCR genes as in Ref. 27).

PCR products from amplification of TCRV8DJ7 and TCRV5DJ7 junctions were subjected to gel electrophoresis in Tris-acetate-EDTA buffer and bands corresponding to TCRV8DJ1 and TCRV5DJ1 were purified using the Qiagen gel extraction kit (Qiagen, Hilden, Germany). PCR products from amplification of TCRVJ and TCRVDJ junctions were purified using the Qiagen PCR purification kit (Qiagen). All PCR products were cloned into the PCR-XL-TOPO vector (Invitrogen, San Diego, CA) and subjected to automated DNA sequencing using an ABI 310 Genetic Analyser (PerkinElmer, Wellesley, MA). Groups of identical sequences were counted as single sequences. SDs of the percentages of productive junctions were determined using the following equations: variance² = 1/N x p x (1 – p); variance x 1.96 = SD. Where N is the total number of sequences analyzed, p is the proportion of productive rearrangements, and 1 – p is the proportion of nonproductive rearrangements.

Analysis of TCR and TCR gene rearrangements by semiquantitative PCR

Total genomic DNA was isolated from DN3 cells of B6, TCR, and TCR transgenic mice, and from TCR⁺ and TCR^– subsets of B6 DN3 cells, as described above. DNA was amplified using the same PCR conditions as for sequencing. Primers were V8DJ2 (24, 25, 26), V1.1J4, and V2J1 (listed above). DNA concentrations were determined photometrically and then adjusted according to the strength of the signals for the insulin gene that displays as two PCR products using previously described primers (18). After adjustment, PCRs were performed in serial DNA concentrations (60, 40, and 20 ng). PCR products were subjected to gel electrophoresis in Tris-acetate-EDTA buffer with 1.6% agarose and visualized by staining with ethidium bromide as described (26).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Possible mechanisms to account for the sequential detection of TCR and TCR proteins in DN3 and DN4 cells, respectively

We could envisage three different models to account for the sequential detection of clonotypic TCR proteins in thymic development. In model A, DN3 cells are already precommitted for either the or lineages, and -committed cells down-regulate CD25 more rapidly than cells committed to the lineage. Model B assumes that DN3 cells are still uncommitted. DN3 cells with productively rearranged TCR genes express the corresponding TCR and immediately shut down CD25. In contrast, DN3 cells with a productive TCR rearrangement express the pre-TCR and down-regulate CD25 with some delay. Both models, therefore, allow expression of TCR proteins to take place before, or simultaneously with, the pre-TCR. Model C assumes that TCR gene expression is programmed in such a way that TCR genes are expressed before and TCR genes are expressed after down-regulation of CD25. Only in model C is there a true delay between TCR gene rearrangements and expression. Models A–C make different predictions with respect to the proportions and kinetics of appearance of DN3/DN4 cells that express clonotypic TCR proteins in WT and TCR tg mice, and with respect to the selection of productive TCR V(D)J genes before and after TCR protein expression.

Analysis of thymocyte development from the DN3 stage onwards using RTOCs

The kinetics of emergence of thymocytes that express TCR/ proteins is best analyzed in a single wave of differentiation and proliferation in vitro, thus avoiding the dynamic complexities of the steady state as well as possible competitive effects in vivo. We have previously shown that purified DN3 cells develop in RTOC into four subsets of DN4 cells that intracellulary express only TCR protein (⁺/^–, both TCR and proteins (^–/⁺), all three clonotypic TCR proteins (⁺/⁺), or none of the three (^–/^–) (12). The DN4 subsets generated in RTOC are essentially similar to that seen in vivo (11, 12). To determine the number of cell divisions in this system, DN3 thymocytes were labeled with CSFE, placed into RTOC at day 0, and analyzed daily until day 6 as shown in Fig. 1. DN4 cells as well as the first DP cells were generated by day 2. DN3 cells have disappeared and most DN4 cells have progressed into the DP population by day 6. CSFE patterns suggest that cells leave the DN3 stage after 0–3 (days 2 and 3) or 0–4 (days 4–6) divisions. Cells exit the DN4 stage after a minimum of 1 (day 2) and a maximum of 6 cell divisions (days 4–6). Thus, the DN4 stage in our RTOC system lasts for 1–3 cell divisions. To reach the DP stage from DN3, cells undergo between 1 and 2 (day 2) and 6 and 7 (day 6) divisions.

View larger version (47K): [in this window] [in a new window]   FIGURE 1. Proliferation of thymocytes developing from DN3 cells in RTOC. DN3 cells of B6 mice were sorted, labeled with CSFE, and placed into RTOC, as described in Materials and Methods. After the indicated number of days, cultures were harvested and analyzed as indicated in the figure. Gating strategies were as described in Materials and Methods. Events above the vertical bars in the CSFE histograms indicate cells that have not divided. Numbers of cell divisions are indicated.

View larger version (20K): [in this window] [in a new window]   FIGURE 2. Proliferation of DN4 cell subsets derived from DN3 cells in RTOC and defined by TCR protein expression. DN3 cells of TCR tg mice were sorted, labeled with CSFE, and placed into RTOC, as described in Fig. 1. On day 6, cultures were harvested and analyzed by 4-color flow cytometry using a pool of mAb to CD4, CD8, CD44, CD25, and NK1.1, for negative gating and intracellular staining (IC) for TCR (pooled anti-TCRV2, anti TCRV1.1, and anti-TCR) and TCR. Two parameter dot plots are shown as indicated for three DN4 subsets gated according to decreasing CSFE levels (see Fig. 1). Percentages of subsets are given in the insets.

Emergence of DN4 cells expressing intracellular TCR and/or TCR proteins in RTOC from DN3 cells of WT and TCR tg mice

RTOC experiments were set up with purified DN3 cells from WT, TCR tg (V8), TCR tg (V1.1,V6), and TCR/ double-tg mice. V1.1 represents one of the two V genes frequently expressed in DN4 cells of WT mice (11). Transcripts of the V1.1 tg are seen in DN2 and DN3 cells, whereas comparable levels of endogenous V1.1 mRNA are first seen in DN3 (P. Pereira, unpublished observations). Transcripts of the TCR tg are present in DN3 cells at WT V8 levels (Ref. 26 , and our unpublished observations). TCR tg protein levels are lower than that of endogenous TCR genes, but sufficient for their allelic exclusion (Ref. 26 ; see also Fig. 4A). Cultures were analyzed daily on days 2–6 using intracellular staining for TCR proteins in DN3 and DN4 populations as shown for selected days in Fig. 3A. Absolute cell numbers were calculated for the four DN4 subsets (⁺/^–, ⁺/⁺, ^–/⁺, and ^–/^–)at all time points and are represented in Fig. 3B.

View larger version (33K): [in this window] [in a new window]   FIGURE 4. Semiquantitative PCR of rearranged TCRV1.1, TCRV2, and TCRV8 genes in DN3 cells of B6 (WT), TCR tg, and TCR tg mice (A), and of TCRV1.1 and TCRV2 genes in isolated TCR+ and TCR– DN3 cells of B6 mice (B). DNA was isolated, adjusted to similar concentrations, subjected to PCR amplification at the indicated concentrations using primers specific for the TCR genes and for the insulin gene, and subjected to agarose electrophoresis, as described in Materials and Methods. The V1.1 PCR amplifies the transgene in TCR tg thymocytes but endogenous V1.1 rearrangements in B6 and TCR tg thymocytes. The V2 and V8 PCRs amplify only endogenous rearrangements in all three strains. The six V8 bands represent rearrangements to J1.1-6 from top to bottom; the appearance of only some of the bands is an indication for small numbers of rearrangements in the sample.

View larger version (34K): [in this window] [in a new window]   FIGURE 3. Expression of intracellular TCR and TCR proteins in DN3 and DN4 cells of B6, TCR tg, TCR tg, and TCR/ tg mice developed in RTOC. DN3 cells were isolated from the four different strains of mice and placed into RTOC, as described in Materials and Methods. At days 2–6, cultures were harvested and analyzed by four-color flow cytometry using a pool of mAb to CD4, CD8, CD44, and NK1.1 for negative gating, anti-CD25, and intracellular staining (IC) for TCR (pooled anti-TCRV2, anti TCRV1.1, and anti-TCR) and TCR. A, Two-parameter dot plots and single parameter histograms are shown as indicated on day 3 (TCR tg and TCR/ tg) or day 4 (B6 and TCR tg). Quadrants differ as experiments were done on different occasions. The blocking control using unlabeled pooled anti- mAb in excess is included in bold in the histograms. Percentages of subsets after substraction of control values are shown in the insets, absolute cell numbers recovered per lobe are given on top in bold, absolute numbers of DN3 and DN4 cells above each frame. B, Absolute cell numbers were calculated for the four DN4 subsets on each day of RTOC. B6, ; TCR tg, ; TCR tg, ; and TCR/ tg, •.

With respect to the timing of appearance, ⁺/^– and ^–/⁺ DN4 cells emerged with strikingly parallel kinetics (Fig. 3B). In WT mice, both populations increased until day 4 of RTOC, while in TCR tg mice, they increased until day 3 followed by a plateau until day 4. In addition, there was no acceleration in appearance of the ⁺ DN4 populations in the TCR tg compared with WT mice. In contrast, the TCR tg accelerated the appearance of ⁺/^– and ⁺/⁺ DN4 cells such that most of these cells were generated before day 2, with some further increase in ⁺/^– cells until day 3. These results are inconsistent with models A and B which predict that cells expressing endogenous or tg TCR proteins exit more rapidly from the DN3 population than cells expressing TCR. TCR and - transgenes also had nonreciprocal influences on the sizes of DN4 subsets. TCR tg mice showed significantly increased absolute numbers of total ⁺ DN4 cells while the ⁺/^– DN4 population was not consistently decreased in these mice. Moreover, similar absolute numbers of DN4 cells expressed the TCR tg in the presence and absence of TCR tg. This suggests that the TCR transgenes are unable to divert significant numbers of DN3 cells to develop into ⁺ DN4 cells and is, therefore, inconsistent with predictions by model B. In contrast, the increase in ⁺/^– DN4 cells of the TCR tg mouse was associated with a strongly diminished population of total ⁺ DN4 cells (sum of ^–/⁺ and ⁺/⁺), whereas in TCR/ double-tg mice, the number of total ⁺ DN4 cells was similar to that in TCR tg mice. For explanation, we tested the hypothesis that the TCR transgene had an inhibitory effect on endogenous TCR rearrangements. Indeed, as shown in Fig. 4A, the TCR transgene partially inhibited endogenous V1.1 to J4 and V2 to J1 rearrangements in comparison to WT mice, thus accounting for the reduction of the total ⁺ DN4 population in TCR tg, but not in TCR/ double-tg mice. Yet, in TCR tg mice, rearrangements at the TCR locus were not inhibited. Consistent with model C, these data indicate that the TCR transgene is expressed when TCR rearrangements are still ongoing, while the TCR transgenes are expressed after completion of TCR rearrangement.

Since TCR and TCR rearrangements overlap to some extent in WT mice, it was possible that endogenous TCR proteins would also inhibit endogenous TCR rearrangements by pre-TCRinduced down-regulation of RAG1 and RAG2. However, only a small degree of inhibition of TCR rearrangements is seen in TCR⁺ DN3 cells compared with TCR^– DN3 cells of WT mice (Fig. 4B), implying that most TCR genes finish rearrangement before endogenous TCR genes are expressed. Thus, suppression of TCR rearrangements by productive TCR genes may have a limited quantitative role in normal mice.

TCR gene selection status in DN3/DN4 populations of WT, TCR KO, and CD3 x p56^lck KO mice

Before expression of clonotypic TCR proteins, thymocyte subpopulations should be unselected as reflected by frequencies of productive TCR gene rearrangements near 33%. Upon expression of TCR proteins and TCR-dependent further developmental progression, the remaining thymocyte population becomes depleted of productive rearrangements. Thymocytes of KO mice with compromised TCR should maintain nonselected frequencies. Frequencies of productive TCR V(D)J gene rearrangements were analyzed in purified ^– or ⁺ subsets of DN3 thymocytes that do not express TCR proteins and in the four DN4 subsets purified according to intracellular expression of TCR and TCR proteins. As a reference for nonselected frequencies, we included TCR KO mice which are unable to generate a TCR (14), and CD3 x p56^lck KO mice which do not produce a functional CD3 complex (18). Sequencing analysis of TCR rearrangements included V1.1J4 and V2J1 gene segments, together accounting for >80% of all TCR chains expressed in DN4 cells (11) and V5J1 rearrangements.

In DN3 cells of WT mice, productive junctions of V5(D)J1 and of V1.1J4 appeared to occur at nonselected frequencies, i.e., not significantly different from 33% (Tables I and II). Proportions of productive V2J1 rearrangements in WT DN3 cells were between 9 and 17% (Table III). This is due to an in-frame stop codon at the 3' end of the V2 gene that needs to be deleted in order for the rearrangement to be productive, and the observed frequencies are therefore, consistent with absence of selection. TCR expression does not seem to influence the TCR selection status at this stage. No statistically significant differences were detected in comparison to the results on the two KO strains that define the range of results in the absence of selection. Together, TCR rearrangements appear to be nonselected in DN3 cells. This is clearly inconsistent with model B that predicts depletion of productive TCR rearrangements in DN3 cells.

View this table: [in this window] [in a new window]   Table I. Proportions of productive TCR V5 rearrangements in DN3 and DN4 subsets

View this table: [in this window] [in a new window]   Table II. Proportions of productive TCR V1.1 rearrangements in DN3 and DN4 subsets

View this table: [in this window] [in a new window]   Table III. Proportions of productive TCR V2 rearrangements in DN3 and DN4 subsets

TCR gene selection status in DN3/DN4 populations of WT, TCR KO, and CD3 x p56^lck KO mice

Productive V8DJ1 and V5DJ1 rearrangements were strongly depleted from TCR^– DN3 cells and fully enriched in TCR⁺ DN3 cells, including that of the KO mice (Table IV). These data corroborate our conclusion from TCR tg mice that virtually all DN3 cells are permissive for expression of productively rearranged TCR genes. In DN4 cells, in-frame rearrangements of the TCR genes were, as expected, enriched in both TCR⁺ populations and were strongly depleted in the ^–/⁺ population. However, intriguingly, the ^–/^– subset was found to be positively selected for in-frame TCR rearrangements. Similarly, TCR rearrangements were highly positively selected in DN4 cells of TCR KO mice including the TCR^– subset. We think that this is not at variance with our conclusion from TCR tg mice that most DN4 cells are able to express productive TCR rearrangements as TCR protein. We have previously shown that ^–/^– DN4 cells contain apoptotic cells (12). Therefore, we suggest that the ^–/^– subset consists to a large extent of initially -selected cells that fail to further develop due to other reasons (see Discussion). TCR rearrangements were nonselected in the DN4 population of CD3 x p56^lck double-deficient mice, suggesting that the DN4 cells generated in these mice fail to get -selected owing to compromised CD3 signaling.

View this table: [in this window] [in a new window]   Table IV. Proportions of productive TCR V rearrangements in DN3 and DN4 subsets

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

It should be pointed out that the failure to detect intracellular or cell surface TCR proteins by FACS does not exclude the expression of low amounts of these proteins and the presence of functional TCRs at the cell surface. Therefore, we performed a second series of experiments to examine the existence of functional TCR in DN3, independently of the detection of TCR proteins. To this end, we determined the proportions of in-frame V(D)J rearrangements of and loci in DN3 and DN4 subpopulations. Before functional pre-TCR or TCR expression at the cell surface, the frequency of in-frame TCR gene rearrangements should reflect the random rearrangement process in the absence of selection. As most TCR rearrangements are completed in DN3 cells (2, 3), delayed expression of TCR proteins up to the DN4 stage would result in random TCR and - rearrangements in DN3 cells. Alternatively, if expression of TCR proteins takes place in DN3 and immediately drives maturation of these cells to DN4, the DN3 subset should be depleted of productive TCR rearrangements. Our data indicate that the DN3 subset contains nondepleted frequencies of productive TCR rearrangements. The analysis is complicated by the relatively small differences to be expected between random and depleted frequencies on the one hand, and the considerable statistical variation in the case of low proportions of productive genes on the other. Nevertheless, we obtain consistent results for three different V gene segments in a considerable number of independently sorted and analyzed populations, isolated from WT mice and from mice with compromised TCR. Therefore, we are confident that our results are inconsistent with model B that predicts depletion of productive TCR rearrangements from DN3 cells. Taken together, our results from RTOC and junction sequencing lend strong support to model C that postulates a true temporal delay in the expression of TCR proteins compared with TCR (Table V).

It is important to point out that these results apply to the two V genes that predominate in DN4 cells, and may therefore, not be valid for the other two V genes rearranged in adult thymus (Vg5 and Vg1.2) and rarely expressed in DN4 cells (30, 31, 32), or for V3 and V4 that are rearranged prenatally. Indeed, preliminary evidence from our lab suggests that productive V5J1 and V1.2J2 segments may be underrepresented in DN3 cells (G. Turchinovich, K. Eichmann, and A. Krotkova, manuscript in preparation). Therefore, it is possible that most cells expressing these V genes exit the population at DN3 or even earlier without appearing in DN4. Moreover, it is not excluded that some cells expressing any of the four V genes bypass the DN3 subset. However, as most TCR and - rearrangements take place in DN3 (2, 3), we think that this putative maturation pathway may be a minor one.

Sequencing analyses of DN4 thymocytes were performed on four populations subdivided according to the expression of intracellular TCR and TCR proteins. As expected, the ⁺/^– and ⁺/⁺ populations were significantly enriched for productive TCR rearrangements, although this was much more pronounced for the TCR than for the TCR genes. This could be due to lack of allelic exclusion in the loci (33, 34) and the complexity of the locus, which allows any cell with a productively rearranged allele to carry up to five nonproductively rearranged alleles (27, 35, 36). In TCR^– DN4 populations, independently of TCR expression, frequencies of productive TCR and TCR V(D)J rearrangements were not significantly different from that in DN3 and DN3/4 cells of WT and TCR and CD3 x p56^lck KO mice, respectively.

Productive V(D)J gene rearrangements were expectedly enriched in DN3 and DN4 populations that express TCR proteins and were strongly underrepresented in ⁺/^– DN4 cells. This finding supports our suggestion based on experiments with TCR tg mice that most DN3 and DN4 cells express TCR proteins if they harbor a functionally rearranged TCR gene. However, some DN4 cells fail to express any TCR proteins and this population was strongly enriched for in-frame VDJ junctions, indicating that the cells have matured to DN4 in the context of -selection. The same enrichment was also seen in the corresponding DN4 subset of TCR KO mice. As we have previously shown, this population contains apoptotic cells and is the endpoint of unsuccessful pre-T cell development (12). We can imagine two nonmutually exclusive mechanisms to explain the presence of productive TCR rearrangements in this subset. First, this population might contain lineage-committed cells that have productive TCR, but no productive TCR gene rearrangements, as postulated by stochastic models. Second, some TCR chains may give rise to pre-TCRs that may not be able to deliver the signals required by DN4 cells for survival and further maturation (12). In either case, such cells might become apoptotic, terminate all TCR protein expression, and thus appear in the ^–– DN4 population.

The data reported here severely constrain putative opportunities of lineage instruction by TCR containing V2 and V1.1 chains. Indeed, in instructive models, it would be difficult to explain the existence of T cells expressing these V genes together with a productive TCR rearrangement. Nevertheless, we observe such cells among ⁺/⁺ DN4 cells (this paper) and peripheral T cells (data not shown) of WT and TCRVg1.1 tg and double-tg mice. Therefore, such cells must be precommitted to the lineage rather than instructed by a TCR. However, our data do not generally exclude instructive mechanisms. For example, it has been suggested that pre-TCR signals inhibit lineage development to interpret the increases in the number of T cells in mice deficient for the pre-TCR (pT and TCR KO mice; Refs. 4 , 37 and 38). Such a mechanism is not at variance with the results presented here.

A sequential expression model based on our data is outlined schematically in Fig. 5. As our data suggest TCR-independent -lineage commitment for at least some of the DN4 cells, we think that some of the DN4 cells are -lineage committed as well. Random TCR gene rearrangements in DN2/DN3 generate four types of DN3 cells: Cells with a productive TCR gene, cells with a pair of productive TCR genes, cells with all three TCR genes productively rearranged, and cells with no productive TCR gene rearrangement. TCR proteins are expressed in all DN3 cells that possess a productive TCR gene, whereas TCR proteins are not expressed in DN3. All DN3 cells mature to DN4, i.e., down-regulate CD25 with similar kinetics, but cells expressing TCR proteins divide with a faster rate than cells that do not. After entering the DN4 stage, all cells are still permissive for TCR gene expression but only -lineage-committed DN4 cells are permissive for TCR gene expression. As a consequence, DN4 cells with three productively rearranged TCR genes (⁺⁺) express both TCR and TCR proteins if they are lineage committed or only TCR if they lack -lineage commitment. These cells are the source of mature and T cells, respectively, that harbor productive rearrangements of the opposite lineage. Cells that possess productive genes for only one of the two TCRs (^+– or ^–+) express the corresponding TCR protein(s). However, in a certain proportion of these cells, there seems to be a mismatch between lineage commitment and productively rearranged TCR gene(s). In this case, or in case of other developmental impairments, the cells terminate production of all TCR proteins and die. It will be of interest to determine the onset of lineage commitment in this sequence, which could be as late as at the DN3 to DN4 transition or the DN4 stage itself.

View larger version (17K): [in this window] [in a new window]   FIGURE 5. Lineage divergence with sequential TCR and TCR protein expression in DN3 and DN4 cells, respectively. Original dot plots of TCRIC vs TCRIC protein expression in DN2-DN4 populations are shown. The proposed fate of the DN4 subsets is indicated. The developmental progression is shown by arrows. Arrow 1, TCR-independent maturation of DN2 cells to DN3. Arrow 2, DN3 cells with a productive TCR gene express TCR protein, including cells with productive TCR genes. Arrows 3 and 4, All DN3 cells mature to DN4, TCR+ cells (arrow 3) divide more often than TCR– cells (arrow 4). Arrow 5, TCR