The JI
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     
 

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Baldwin, K. K.
Right arrow Articles by Davis3, M. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Baldwin, K. K.
Right arrow Articles by Davis3, M. M.
The Journal of Immunology, 1999, 163: 689-698.
Copyright © 1999 by The American Association of Immunologists

Negative Selection of T Cells Occurs Throughout Thymic Development1

Kristin K. Baldwin2,*, Brian P. Trenchak{dagger},{ddagger}, John D. Altman§ and Mark M. Davis3{dagger},{ddagger}

* Program in Immunology, {dagger} Department of Microbiology and Immunology, and {ddagger} Howard Hughes Medical Institute, Stanford University, Stanford, CA 94305; and § Department of Microbiology and Immunology, Emory University, Atlanta, GA 30322


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Thymic positive and negative selections govern the development of a self-MHC-reactive, yet self-tolerant, T cell repertoire. Whether these processes occur independently or sequentially remains controversial. To investigate these issues, we have employed tetrameric peptide-MHC complexes to fluorescently label and monitor polyclonal populations of thymocytes that are specific for moth cytochrome c (MCC)/I-Ek. In TCR ß mice tetramer-positive thymocytes are detectable even in the most immature TCR-expressing cells. In the presence of MCC peptide, thymocytes that bind strongly to MCC/I-Ek tetramers are deleted earlier in development and more extensively than cells that bind weakly. This negative selection of the MCC/I-Ek-specific cells occurs continuously throughout development and before any evidence of positive selection. Thus, positive and negative selections are independent processes that need not occur sequentially.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
The generation of a functional, yet self-tolerant, T cell repertoire is essential to the survival of an organism. To build this T cell repertoire, immature T cells in the thymus rearrange their genomic DNA to generate TCRs of many different sequences and specificities. Of this diverse array of TCR specificities, mature T cells express only those receptors that recognize self MHC molecules. Furthermore, except in cases of autoimmune reactions, mature T cells activate only when confronted with foreign peptides bound to these self MHC molecules. To generate a population of polyclonal cells with such precise recognition properties, each thymocyte must undergo both positive and negative thymic selection.

Positive selection is responsible for ensuring that all T cells recognize self MHC. To become positively selected, an immature thymocyte must express a TCR that can interact with self MHC peptide complexes, generate a signal, and initiate a program of differentiation that includes down-regulation of recombinase-activating gene-1 (RAG1)4 RAG2, HSA, and either CD4 or CD8 as well as transient up-regulation of CD69 (1, 2, 3). The process of selecting cells that react with self MHC peptide complexes, however, leads to a repertoire of mature T cells that includes autoreactive cells (4, 5). Thus, before thymocytes emigrate to the periphery, cells that might react with self MHC peptide complexes to a significant degree must be eliminated. This negative selection also requires the TCR expressed by a thymocyte to interact with MHC peptide complexes. As in positive selection, this binding event generates a signal and induces CD69 expression. Rather than inducing differentiation, this signal causes these autoreactive thymocytes to apoptose and eventually be eliminated from the thymus. Although positive and negative selections each result in different cell fates, these divergent outcomes are both triggered by TCR interactions with self MHC peptide complexes.

How does a thymocyte distinguish between the signals that induce positive selection/differentiation vs negative selection/death? One possibility is that a given thymocyte may respond differently to TCR stimulation by MHC peptide complexes depending on its developmental stage. In this scenario, TCR-mediated signals induce positive selection early in development, while similar signals induce negative selection later, after the threshold for signaling has been raised or after signaling has been coupled to different transcriptional events. In support of this model, CD4+8+ thymocytes contain a subset of TCR+ cells that cannot be deleted by TCR stimulation (6). Further evidence derives from studies of superantigen-mediated negative selection of cells bearing a particular TCR ß-chain. In these studies thymocytes expressing the appropriate Vß were present in the immature CD4+8+ populations, but were absent (or reduced in number) in the more mature CD4+8-/int populations (7). Because the CD8intCD4+ cells were already enriched for the TCR Vß-chains that were positively selected in their particular MHC background, in this system negative selection followed positive selection. Although this study is the only in vivo approach to examine nontransgenic thymocyte selection, the necessity of using superantigens may have biased the results. For example, superantigen-mediated deletion may occur later than peptide-mediated deletion either because the endogenous superantigens interact weakly with most TCRs or because these molecules are not highly expressed in the thymic cortex (8, 9).

Studies using TCR transgenic mice in which most thymocytes express the same TCR suggest an alternative hypothesis. Instead of developmental stage, the specificity and the strength of the TCR-MHC peptide interaction may be the primary determinants of developmental fate. For instance, immature CD4+8+ cells from some TCR transgenic mice are negatively selected early in thymic development (10, 11, 12) or on day 17 of gestation during fetal development, preceding the appearance of positively selected CD8+ T cells by several days (13). Furthermore, when immature thymocytes are removed from nonselecting thymuses and stimulated through their TCR/CD3 complex they can be induced to apoptose in the absence of prior positive selection (14, 15). These studies show that TCR transgenic thymocytes can be deleted early in development and/or ontogeny, but they also show that even under the control of endogenous promoters, transgene-encoded TCR genes are expressed at higher than normal levels in the immature CD4-8- and CD4+8+ cells in the thymus (13).

This aberrant TCR expression may lead to artifactual early deletion in several ways. First, as Berg et al. showed (10), TCR {alpha}ß transgene expression can cause aberrant early CD3 expression that may increase the sensitivity of immature cells to TCR stimulation and lead to abnormally early deletion of cells. In fact, in these studies the expression of both TCR chains, but not TCR ß alone, results in massive deletion of CD4+8+ thymocytes in mice that expressed endogenous superantigens reactive to the transgene-encoded Vß segment. Second, in some in vitro studies, thymocytes removed from intact thymuses have been shown to increase TCR expression (16). Again, aberrantly high TCR levels might lead to nonphysiologically relevant negative selection. Finally, the studies that document early negative selection employ transgenic T cells bearing receptors derived from hybridomas or T cell lines that were selected because they responded vigorously to Ag. Thus these cells are likely to have a reasonably high avidity for the selecting Ag that could lead to unusually efficient or early deletion. Taken together, these arguments indicate that TCR transgenic models of early events in thymic selection are potentially unreliable.

Because of the problems with these model systems, the relative timing of positive and negative selections has remained controversial. In this study we attempt to resolve this controversy by tracking the development of a polyclonal, MHC peptide-specific population of thymocytes in the presence or the absence of the deleting peptide. To identify MHC peptide-specific cells, we have used an MHC tetramer-staining reagent (17, 18) that permits the detection and isolation of Ag-specific T cells even when these cells comprise only a small fraction (0.1%) of the total thymocytes. By crossing mice that are transgenic for a normally regulated TCR ß-chain from an I-Ek+ moth cytochrome c (MCC)-reactive T cell clone (5C.C7) to mice that express a soluble hen egg lysozyme protein modified to include the antigenic MCC88–103 peptide (HELCYT), we are able to monitor positive and negative selection of Ag-specific T cells in vivo, in the absence of early expression of TCR{alpha} or CD3 (19, 20, 21). We show that cells with the highest capacity to bind the negatively selecting ligand are deleted most efficiently, that this deletion occurs throughout thymic development, and that positive selection is not required for negative selection in vivo.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Tetramer production

Soluble {alpha}- and ß-chains of the I-Ek MHC molecule were produced in Escherichia coli (22). The {alpha}-chain has been engineered to contain a short biotinylation motif on the membrane-proximal section of the molecule. After production and folding of the molecule with MCC peptide the complexes were purified by affinity chromatography and specifically biotinylated with the enzyme Bir A in the presence of protease inhibitors (17, 18). The biotinylated molecules were size-purified by fast protein liquid chromatography, then mixed in a 6-fold molar excess with fluorescently labeled streptavidin to form peptide/MHC tetramers. Excess or unconjugated MCC/I-Ek was washed away by centrifugation through a Centricon 100K (Amicon, Beverly, MA).

Flow cytometry

Thymocytes were isolated by removing whole thymus into cold PBS and disrupting the tissue by pushing through a metal sieve or plastic mesh. Cells were incubated in red cell lysis buffer (0.74 M ammonium chloride) for 5 min at 37°C, then washed three times in FACS wash buffer (2% FCS/PBS, pH 7.4). Cells were maintained at 4°C for the rest of the procedure, and all washes were performed with FACS wash buffer. Cells were counted with a hemocytometer, and 106 cells were used for each combination of staining reagents. The staining parameters of tetramer preparations varied. Each preparation was titrated on 5C.C7 TCR{alpha}ß transgenic T cells to determine the optimum staining concentration on the day before analysis. The appropriate concentration of tetramer/Av-PE (~50 µg/ml of tetramer/Av-PE) or the equivalent amount of Av-PE alone was added to cells from 2 to 4.5 h before analysis. All other Abs were added 30–50 min before analysis. Dead cells were labeled by incubating cells for 5 min with 10 µg/ml propidium iodide before washing several times. Cells were analyzed by flow cytometry on a customized dual laser Vantage (Becton Dickinson, Mountain View, CA) and analyzed using either FACS Desk or FlowJo software (Beckman Center Shared FACS Facility, Stanford, CA). All analyses excluded small cells, propidium iodide-positive cells, and any cells that bound to CyChrome-conjugated irrelevant Abs (B220-CyChrome). The percentage of tetramer positive cells was computed by dividing the number of tetramer-positive cells by the number of cells found in the next largest subset.

Antibodies

The following Abs were used: anti-B220-CyChrome (RA3-6B2), CD3-FITC (145-2C11), CD69-FITC (H1.2F3), and heat-stable Ag (HSA)/CD24- FITC (M1/69) all from PharMingen (San Diego, CA); Ultralite streptavidin-PE (Molecular Probes, Eugene Oregon); and anti-CD4-allophycocyanine (GK1.5), anti-CD8-BSA-Texas Red, and anti-Vß3-APC (KJ25; produced by members of our laboratory or received as a gift from the Weissman laboratory, and conjugated to fluorophores from Molecular Probes by standard procedures).

Mouse strains

The generation of 5C.C7 TCR ß transgenic and HELCYT transgenic mice has been previously described (19, 20). All mice were bred and maintained in a pathogen-free animal facility (DLAM, Stanford University). Both the HELCYT and 5C.C7 ß x B10.BR strains have been crossed >10 times to the B10.BR background.

Peptides

The antigenic peptides used in folding reactions and for injection of transgenic mice were synthesized in the PAN Facility (Beckman Center, Stanford University) and purified by HPLC. The MCC88–103 peptide used had the sequence ANERADLIAYLKQATK.

Single-cell RT-PCR

Single CD4+V{alpha}11+ cells that were tetramer high, medium, or negative were sorted into RT-PCR buffer as described previously (23). The rearranged TCR V{alpha}11 genes were amplified by RT-PCR using V{alpha}11-specific nested primers. These cDNAs were subcloned, and both strands were sequenced by standard techniques.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
5C.C7 ß-chain transgenic mice exhibit normal thymocyte differentiation

To study the generation of Ag-specific T cells at the early stages of thymocyte development, we examined thymocytes from 5C.C7 TCR ß-chain transgenic mice (ß transgenics). The expression of the TCR ß transgene modifies the normal TCR repertoire such that these mice produce a detectable population of MCC/I-Ek-specific cells without significantly disrupting other developmental processes (10, 19, 20). Three lines of evidence suggest that this is true. First, studies of 5C.C7 ß transgenic fetal thymocyte development show that TCR ß-positive cells appear on the appropriate day of gestation (F. de St. Groth, unpublished observations). Second, CD3 expression in CD4-8- and CD4+8+ cells from TCR ß transgenics resembles expression in wild-type animals (Fig. 1Go) (20). These expression patterns are strikingly different from the aberrantly high level of CD3 expression exhibited by TCR {alpha}ß CD4-8- thymocytes; if anything, CD3 expression is slightly decreased in TCR ß transgenics. Third, TCR ß transgenic mice select a normal thymic repertoire of TCR {alpha}-chains at the appropriate time in development (19, 24). Although the presence of the ß transgene causes the mature T cell repertoire to contain higher than normal numbers of CD4+ cells, this skewing most likely reflects normal thymic selection processes operating on cells that each express a ß-chain that is more reactive with class II MHC than with class I MHC. Thus, unlike the TCR {alpha}ß transgenic model systems used for earlier studies of the timing of positive and negative selections (9, 10, 11, 12, 13), TCR ß transgenic mice provide an accessible, yet physiologically relevant, model to study the thymic development of small, polyclonal populations of Ag-specific T cells.



View larger version (48K):
[in this window]
[in a new window]
  		FIGURE 1. Immature thymocytes from 5C.C7ß transgenic mice develop normally. FACS analyses of 100,000 live thymocytes from a 5C.C7ß transgenic mouse or a nontransgenic B10.Br mouse are displayed as 5% probability plots with outliers (each dot represents one cell). CD4 and CD8 staining were used to identify CD4+CD8+ and CD4-CD8- cells. The forward scatter (size) vs CD3 expression profiles for each of these populations are plotted as indicated. The relative percentage of cells per subset is displayed in the upper right corner of each plot.

 
MHC tetramers identify Ag-specific thymocytes

To identify and follow these Ag-specific thymocytes, we generated a tetrameric MCC/I-Ek staining reagent (17, 18). This reagent consists of four biotinylated MCC/I-Ek complexes that are multimerized by binding them to a fluorescently labeled streptavidin molecule. These tetramers label T cells that express receptors specific for the MCC/I-Ek complex. In a typical experiment a tetramer preparation labels ~95% of total thymocytes from 5C.C.7 TCR {alpha}ß transgenic mice, 4.0% of thymocytes from 5C.C7 ß transgenic mice, and <0.3% of thymocytes from B10.BR nontransgenic mice (Fig. 2GoA). Tetramer staining is TCR specific because incubation with an Ab that binds to the TCR ß-chain reduces the number of tetramer high cells to background levels (Fig. 2GoA). In addition, tetramer staining in ß and TCR-{alpha}ß transgenic mice is significantly higher than background staining with streptavidin-PE (Av-PE) alone and is greatest in the CD4+ cell subset (Fig. 2GoB).



View larger version (32K):
[in this window]
[in a new window]
  		FIGURE 2. Tetramer specificity. A, Representative FACS analyses of whole thymus from 5C.C7 {alpha}ß, B10.Br, and 5C.C7ß+/- anti-Vß3 cells are labeled and displayed as the tetramer staining level (x-axis) vs the relative number of cells (y-axis). Incubation with 10 µg/ml of the anti-Vß3 Ab, KJ25, reduces the staining of most cells (lower panels). The numbers in each panel denote the relative percentage of 100,000 live, B200- class II-, tetramer+ cells that are contained within the gates defined by the bold lines. B, Each panel is a 2% probability plot of 100,000 gated live thymocytes from a ß transgenic mouse stained with MCC/I-Ek tetramer-PE or an equivalent concentration of streptavidin-PE alone, plotted against anti-CD4 Ab staining. Percentages are the number of tetramer-positive cells per quadrant per total live cells.

 
To assess the specificity of the T cells that were stained by the MCC/I-Ek tetramer, we used single-cell RT-PCR and sequencing to determine whether tetramer reactive TCRs resemble the receptors expressed by T cells that are known to bind to MCC/I-Ek complexes or are activated by this ligand (23, 25). Using gates similar to those shown in Fig. 2GoB, thymocytes and lymph node T cells from a 5C.C7 ß transgenic mouse were stained with tetramer as well as Abs against the CD4, CD8, and V{alpha}11 molecules. Nearly all tetramer binding cells also expressed V{alpha}11 (data not shown). Single CD4+V{alpha}11+ cells that were tetramer high, low, or negative were sorted, and the rearranged TCR V{alpha}11 genes were amplified by RT-PCR, subcloned, and sequenced. As shown in Fig. 3Go, tetramer-reactive cells from both lymph node and thymus exhibited the characteristic CDR3 loop length restriction seen in other MCC/I-Ek-reactive cells, while tetramer-negative V{alpha}11-positive cells had a more diverse repertoire of CDR3 loop lengths.



View larger version (28K):
[in this window]
[in a new window]
  		FIGURE 3. CDR3 length is restricted in tetramer-positive cells. The number of amino acids in the predicted CDR3 loops of each of the CD4+8- thymocytes or lymph node T cells was computed and plotted as a histogram. The tetramer high and low cells from Table IGo were pooled in this analysis because all CDR3 lengths were identical (8 aa) in both populations. All identified CDR3 lengths were between 7 and 11 aa.

 
The predicted amino acid sequences in the tetramer high and low groups are quite similar to those in known MCC-reactive cells. We found sequences identical with the 5C.C7 TCR {alpha}-chain in two cells from both the tetramer high and tetramer low sorted cells but not in tetramer-negative cells (Table IGo). In the tetramer high cells the CDR3 loop initiates with the germline-encoded glutamic acid at CDR3 position 1 followed by a small residue and then a serine at position 3. This canonical motif has been found in many peripheral MCC-reactive cells and is highly selected for in the MCC/I-Ek-reactive memory T cell population (23). The middle three residues contain either an N-T-N series or two small residues, such as S, A, or G. Position 99 favors the polar residues lysine or glutamine in 21/22 sequences, and the final position is always occupied by hydrophobic leucine or valine. Several sequences appeared more than once, including one complementarity-determining region-3 amino acid sequence that occurred 11 times (although the nucleotide sequences coding for the amino acids did vary). Thus, the pool of amino acid sequences that can bind well to the tetramer is limited.


View this table:
[in this window]
[in a new window]
  		Table I. V{alpha} sequences from 5C.C7ß transgenic mice1

 
The tetramer low sequences also contained known binding motifs, although their distribution was more diverse, and the motifs were more degenerate. The three sequences that appear in both populations are marked with an asterisk. Examination of the tetramer-negative cells shows that these motifs are not evident in the bulk T cell population. Recent studies using T cell hybridomas have indicated that the tetramer staining level and TCR affinity are linearly related (25). Thus, these high and low tetramer binding populations should consist of groups of cells bearing receptors with different affinities for the MCC/I-Ek ligand. Taken together, the staining controls and sequences show that tetramer binding identifies MCC/I-Ek-specific cells and that the higher binding cells display a more restricted population of receptor sequences than do cells bearing less avid receptor complexes.

Endogenous expression of MCC induces the deletion of MCC/I-Ek-specific thymocytes.

To examine negative selection in this model system, we crossed 5C.C7ß transgenic mice to HELCYT transgenic mice. HELCYT animals express a soluble fusion protein of hen egg lysozyme (HEL) protein and residues 80–103 of the moth cytochrome c protein (19). In the HELCYT x 5C.C7 ß double-transgenic mice (HELCYT/ß), MCC-specific T cells fail to develop (20). Splenocytes and thymocytes were isolated from three age- and sex-matched mice from each treatment group. After gating out dead cells and cells that stained with Abs to B220 and MHC class II Ags, 500,000 events were collected, and tetramer staining was compared with CD4 staining. Although splenocytes have a higher background staining level than thymocytes, the presence of endogenous MCC reduced the number of CD4+ tetramer high cells by 95% and that of tetramer intermediate cells by 87% (Fig. 4Go). The presence of MCC also depleted thymic CD4+tetramer+ cells by 97 and 83%, respectively (Fig. 4Go). Importantly, neither the total number of cells per organ nor the percentage of tetramer-negative cells was affected by the presence of the HELCYT protein. Thus, MCC-specific cells have probably been eliminated in the thymus or depleted in the periphery by the products of the HELCYT transgene, most likely the MCC peptide contained in the sequence 80–103.



View larger version (60K):
[in this window]
[in a new window]
  		FIGURE 4. The HELCYT transgene induces specific deletion of tetramer-binding cells. Two percent probability plots with outliers of 100,000 spleen or 500,000 thymus cells from 5C.C7 ß or HELCYT x 5C.C7 ß mice are displayed as noted. CD4+, tetramer high, intermediate, and negative populations were defined based on the gates used in Fig. 2GoA. The number of cells in each population is shown as the percentage of total live cells analyzed and is printed adjacent to each subset.

 
Negative selection occurs at the earliest stage of TCR expression

To determine the mechanism and timing of the removal of these potentially autoreactive cells, we examined the percentage of tetramer-positive cells present at the stages of thymic development as defined by the expression of the CD4 and CD8 coreceptors (7, 26). Although recent reports indicate that these subsets may have a more complex precursor or lineage relationship than was originally proposed, coreceptor expression provides an initial approximation of thymocyte developmental stage (27, 28, 29, 30). In Fig. 5GoA the gates used to define the various thymic subsets and the tetramer-positive populations are drawn on representative samples from a 5C.C7 TCRß transgenic mouse.



View larger version (49K):
[in this window]
[in a new window]
  		FIGURE 5. Deletion is evident at many stages of thymocyte development. A, Representative FACS analyses of thymocytes isolated from a 5C.C7 ß transgenic mouse. Shown are 5% probability plots of live, B220- cells. The gates drawn on these plots of CD4 vs CD8 (left) were used to define the population of subsets labeled 1–5. The plot of forward scatter vs tetramer-PE (right) shows the gates used to define the tetramer+ populations. Gates were based on background staining levels of B10.Br nontransgenic mice. B, The number of tetramer-positive cells was calculated and divided by the number of cells in the designated subset to generate the percentage of tetramer-positive cells per subset for each of the populations 1–5. The circles with arrows are a schematic of the developmental progression of maturing thymocytes through each of the five subsets. Bars represent the mean of at least five 500,000-event samples taken from each of three separate mice per treatment group. Shaded bars, ß transgenic mice; diagonally hatched bars, HELCYT/ß mice; open bars, B10.Br nontransgenic mice. Error bars define the 95% confidence intervals for the means of each group; nonoverlapping bars are statistically significant differences (p < 0.05).

 
At the earliest stage of thymic development (CD4-8-) most thymocytes have neither completed TCR rearrangement nor begun up-regulation of the CD4 and CD8 coreceptors. Because the absolute number of tetramer-positive cells found in this subset ranged from 0 to 10/half-million cells, the 95% confidence intervals are large. Thus, in the CD4-8- population tetramer-positive cells are not reproducibly present at levels greater than those in B10.BR mice, and the depletion observed in HELCYT/ß mice is not statistically significant (Fig. 5GoB). The first evidence of deletion occurs in the next detectable group of tetramer-positive cells, the CD4int8int subset. In this subset HELCYT/ß mice retain only half as many tetramer-positive cells as ß transgenics. Similarly, depletion of tetramer-positive cells is about 50% in the CD4+8+ subset. In the CD4+8int and CD4+8- subsets, the percentage of tetramer-positive cells per subset increases 5- to 10-fold in the ß transgenic mice (Fig. 5GoB, right panels). This increase suggests that cells that recognize MCC/I-Ek are preferentially selected from the CD4+8+ stage, or alternatively, that these cells have an overall increase in TCR expression. The percentage of tetramer-positive cells in the mature (CD4+8-/int) subsets in HELCYT/ß mice is greatly decreased compared with that in ß transgenics; very few residual tetramer-positive cells remain. Importantly, we observed no effect on tetramer-negative cells; therefore, the HELCYT transgene deletes only cells that have demonstrable avidity for MCC/I-Ek tetramers. In fact, we observed efficient deletion in every thymocyte CD4/CD8 subset that had detectable MCC/I-Ek-reactive TCR expression. Thus, independent of the lineage relationship between these subsets, negative selection must occur in a large percentage of the least mature thymocytes.

The extent and timing of deletion correlate with tetramer binding

To examine the influence of tetramer binding avidity and receptor affinity on the extent and timing of deletion of MCC/I-Ek-specific cells, we analyzed thymocyte development in HELCYT/ß and ß transgenic mice as a function of the tetramer staining level. Tetramer staining of mature T cells bearing TCRs with known affinities for MHC peptide complexes has been shown to reflect the avidity of the TCR tetramer interaction (25). As in Fig. 5Go, thymocytes were first divided into three subsets based on CD4 and CD8 staining, then subdivided by successive levels of tetramer staining (Fig. 6Go, A and B). To ensure that these populations were comparable with respect to CD3 levels, the analysis of immature (CD4+8+/4int8int) cells was restricted to those with CD3 staining levels between 1 and 10 (on the log fluorescence scale Fig. 6GoD). Mature CD4+8- cells were left ungated as their CD3 levels were similar (average CD3 levels range between 4 and 10 on the log fluorescence scale; Fig. 6GoE).



View larger version (46K):
[in this window]
[in a new window]
  		FIGURE 6. The extent of deletion corresponds to the tetramer binding level. A, Gates used to define the CD4-CD8- (1), CD4+CD8+ (2), and CD4+CD8- (3) populations. B, Gates used to define the tetramer staining levels. C, Overlay plot of the percent deletion of cells gated for each tetramer staining subset. The percentage of tetramer-positive cells per subset was computed by finding the percentage of tetramer very low (+/-), low (+), medium (++), high (+++), or negative (neg.) cells in each of the three developmental subsets from every mouse analyzed (three from each of the HELCYT/ß, ß transgenic, and B10.Br treatment groups). The background percentage of tetramer-positive (B10.Br) was subtracted from the HELCYT/ß and ß transgenic numbers. The percent deletion was then computed by dividing the percentage of tetramer-positive cells per subset in HELCYT/ß mice by the average of the same statistic from three ß TCR transgenic mice. The 95% confidence intervals for each population were computed using propagation of error formulas. As deletion was more extensive in the more mature populations, the means of the percent deletion for each population from HELCYT/ß mice were overlaid. Error bars represent 95% confidence intervals for the CD4+ subsets. Black bars represent the least mature CD4int8int subset, dark gray bars are the intermediate CD4+8+ cells, and light gray bars denote the mature CD4+8- population. D, Comparison of the CD3 staining level in mature CD4+ (dark line) vs immature CD4+8+ tetramer-positive cells (light line). The y-axis indicates the relative number of cells. E, Comparison of the CD3 staining profiles of CD4+8- cells with different tetramer staining levels as noted. The y-axis indicates the relative number of cells.

 
We analyzed three 5C.C7ß and three HELCYTß transgenic mice and compared the percentages of tetramer-positive cells per subset for each of the three developmental subsets (Fig. 6GoC). At the earliest stage of development, the CD4int8int cells were ~65% deleted in tetramer high cells, but were only 30% deleted in lower cells. This suggests that cells with higher avidity for tetramer (and presumably higher TCR-MCC/I-Ek affinity) are deleted more efficiently and at an earlier developmental stage than other cells. Similarly, the extent of deletion in the CD4+ 8+ subset was nearly complete in the tetramer high (+++) population, was less complete in the tetramer medium (++) subset and was only 66% in the tetramer low (+) cells. The cells with highest avidity for tetramer (+++) were completely eliminated before they reached the mature CD4+8- stage; the percent deletion was the same in these populations. In contrast, cells with lower avidity for the tetramer (+ and ++) did not become fully deleted until the CD4+8- stage (Fig. 6GoC). The data for the tetramer very low (+/) cells were difficult to assess, as these cells overlap significantly with cells found in avidin-PE alone controls. Thus, deletion of Ag-specific cells occurs more efficiently and earlier when the cells express TCR complexes with high avidity for the I-Ek-MCC complexes.

Negative selection can occur before or during positive selection

These data suggest that negative selection is influenced by TCR specificity and occurs at early developmental stages, but do not permit the relative ordering of positive and negative selections. To determine whether positive selection is required for negative selection in vivo, we examined MCC-mediated deletion in the context of other hallmarks of thymic development, including cell size, CD3 expression, and CD69 expression. Because the expression of CD4 and CD8 may not perfectly reflect the maturation level of an individual thymocyte (30, 31, 32), the CD4+8+ and CD4int8int populations (populations 2 and 3 from Fig. 5Go) were pooled and divided into subpopulations. Although the gates used to define the subpopulations are based on positive and negative staining controls they are necessarily somewhat arbitrary. To control for this we performed analyses using several different gating strategies. Here we describe representative results from robust analyses, defined as those that are not influenced by subtle shifts in the definitions of these populations. By comparing the percentage of tetramer-positive cells (Fig. 7Goa) present in each subset in HELCYT/ß mice to the cells per subset in ß transgenic mice it is possible to determine whether deletion can precede positive selection in vivo.



View larger version (41K):
[in this window]
[in a new window]
  		FIGURE 7. Negative selection can precede or coincide with positive selection. Top, a–c are 5% probability plots of the CD4int/+CD8int/+ (populations 2 and 3 from Fig. 4Go) from a 5C.C7ß transgenic mouse with outliers (a) or without outliers (b and c). The gates used to define each population analyzed in d–h are drawn on each plot. Bottom, Results from three 5C.C7ß (gray bars), four HELCYTß (diagonally filled bars), or four nontransgenic (white bars) mice per treatment group were averaged. Error bars denote the 95% confidence intervals of the means. Each bar represents the mean number of tetramer+ cells per subset as a percentage of the number of total cells in each subset. No other CD4int/+CD8int/+ subsets from 5C.C7ß mice contained more tetramer-positive cells than are present in nontransgenic mice.

 
Cell size is one marker of pre-positive selection thymocytes. Separation of thymocytes by centrifugal elutriation has shown that CD4+8+ large cells are enriched for cells that have not yet undergone positive selection, while many CD4+8+ small cells are presumably on the way to death by neglect (7, 32). In other studies, however, small cortical thymocytes have been shown to be capable of being positively selected (28, 30, 33), although perhaps less efficiently than larger cells (31). By using the forward scatter parameter in flow cytometric analyses to define large and small subsets, we found that the larger cells are enriched for tetramer-positive cells in ß transgenics. Furthermore, HELCYT/ß mice delete these cells almost to background levels (Fig. 7Go, a and d). In addition, although small CD4+8+ tetramer-positive cells are present at or near background levels, the few detectable tetramer-positive cells in ß transgenics are reduced to background levels in HELCYT/ß mice. Thus, negative selection can efficiently remove potentially autoreactive cells from both large and small CD4+8+/CD4 int8int thymocyte populations. Although the small cells may include cells that have not yet been positively selected, they also seem to include cells that have already failed positive selection and are on the pathway to death by neglect (32). Because such a small percentage of these cells stains with tetramer, the small cells that express low levels of other coreceptors and are indistinguishable from dying cells have been excluded from further analyses.

A well-studied marker of thymic subpopulations is the CD3 expression level. For instance, in one study cells that had undergone positive selection and were CD3highCD4+8int were the earliest detectable deleted subset, while CD3lowCD4+8int cells were unaffected by the presence of a deleting Ag (7). In the present study although the CD4+8+ and CD3lowCD4+8int precursors and the CD3highCD4+8int cells are efficiently deleted (Fig. 5Go and data not shown), the CD3lowCD4+8+ subset might still be partly resistant to deletion. By examining the percentage of tetramer-positive cells in three mice from each treatment group we found that while ß transgenic mice have clearly detectable populations of tetramer-positive cells in both the CD3high and large CD3low subsets, these populations are significantly reduced in HELCYT/ß mice (Fig. 7Go, b, e, and f). The number of tetramer-positive cells in the small CD3low subset did not differ significantly from background staining of B10.Br mice. These data suggest that MCC peptide-mediated deletion occurs at the earliest detectable stage of TCR/CD3 expression and during or before positive selection.

Perhaps the most rigorous way to identify a pre-positive selection population is to identify cells that have not yet received TCR-mediated signals. To distinguish between cells that have received such signals and those that have not, we examined the appearance of the T cell activation marker CD69. In the thymus CD69 is up-regulated rapidly and transiently on CD4+8+ and CD4+8- thymocytes when they have undergone either positive or negative selection (1, 34). The percentage of tetramer-positive cells present in the CD69lowCD4+8+/CD4int8int population (composed of cells that have not yet received TCR-mediated signals for positive selection) in HELCYT/ß mice was reduced to background levels (Fig. 7Go, c, g, and h). Deletion appeared equally efficient in the CD69lowCD4+8+ and CD4int8int populations (data not shown). Therefore, these preselected cells can be efficiently deleted. Again, the number of tetramer-positive cells in the small CD69 low populations was similar in B10.Br, HELCYT/ß, and ß transgenic mice (data not shown). Given our present ability to define pre-positive selection subsets, these data show that an appreciable amount of negative selection can precede or coincide with positive selection.

Negative selection also occurs later in thymocyte development

Although the increasing efficiency of deletion in later stages of development suggests that deletion does occur at more than one developmental time point, other studies of TCR transgenic mice have suggested that in vivo, deletion occurs within a restricted developmental window (14, 35, 36). To determine whether deletion also occurs at later stages of development in HELCYT/ß transgenic mice, we reasoned that if all deletion occurs at or before the immature CD4+8+ stage, the ratio of mature tetramer-positive cells to immature tetramer-positive CD4+8+ cells should remain constant in the presence or the absence of deleting Ag. Instead, as is shown in Table IIGo, the HELCYT/ß mice have a decreased ratio of mature/immature tetramer positive thymocytes. In addition, deletion induced by a single administration of peptide Ag either 14 or 26 h before removal of the thymocytes also results in a decrease in this ratio compared with that in ß transgenics. The increased ratio of the injected mice compared with the HELCYT/ß mice suggests either that the injection of peptide produces less efficacious complexes, or that there is a more mature population that is resistant to deletion by injected peptide.


View this table:
[in this window]
[in a new window]
  		Table II. The ratio of mature to immature tetramer-positive cells1

 
On the basis of other studies we postulated that if they existed, these resistant cells might be derived from a mature thymic subset that has lost its sensitivity to Ag-mediated deletion, perhaps in preparation for emigration. Others have shown that immature recent cortical emigrants, which are HSA positive, retain their sensitivity to Ag, while more mature HSAlowCD4+8- cells are not susceptible to deletion, perhaps because HSA is required for deletion (37, 38). Mature CD4+8-cells were identified (Fig. 8Goa) and subdivided into subsets based on three markers of differentiation: CD8, HSA, and CD69. As CD4+8- thymocytes mature they down-regulate all three markers (although they first up-regulate CD69 if given a TCR-mediated signal) (1, 2, 39). To determine whether any of these markers identified a particularly resistant thymic population, we calculated the percent deletion in each treatment group for each of nine different subsets (Fig. 8Go, a, c, and e). In eight of these, the presence of the HELCYT Ag induced more complete deletion than did peptide injection. The percent deletion was relatively constant in all CD8 subsets, indicating that these subsets are functionally similar with respect to negative selection (Fig. 8Gob). In contrast, the HSA+ subsets, although equally deleted in HELCYT/ß mice, showed a differential susceptibility to injected peptide Ag. In contrast to other studies, the HSAlow mature cells are most efficiently deleted (Fig. 8Go, c and d), suggesting that these cells are not resistant to deletion and that HSA is not required for the induction of negative selection. Similarly, CD69low cells are more efficiently deleted than the presumably less mature CD69int cells (Fig. 8Go, e and f). As the time for progression from CD4+8+ to CD4+8- mature thymocytes is thought to be 3 days (40), it is unlikely that all of this deletion could be accounted for by precursor depletion. Instead, these data suggest that even putatively mature populations of thymocytes can undergo Ag-induced cell death, perhaps by mechanisms similar to those that regulate activation-induced cell death of mature peripheral T cells.



View larger version (51K):
[in this window]
[in a new window]
  		FIGURE 8. Comparison of extent of deletion in more and less mature CD4+ populations. Three criteria were used to define immature and mature CD4+ populations: CD8 (a and b), HSA (c and d), and CD69 (e–g). In c–g, only the CD4+CD8neg/low populations were analyzed. In a and b, CD8 staining was used to define three populations, and the percentage of tetramer-positive cells/subset was computed for each treatment group. One mouse from each of the following treatment groups was sacrificed: B10.Br nontransgenic, HELCYTß, 5C.C7 ß plus PBS, 5C.C7 ß injected with MCC88–103 26 h before analysis, and 5C.C7 ß injected with MCC88–103 14 h before analysis. The percent deletion in each subset was computed by subtracting the background (B10.BR) percentage of tetramer+ cells from each sample, dividing the number of tetramer-positive cells per sample by the number of tetramer+ cells in the 5C.C7ß sample, and subtracting that percentage from 100. Results from HELCYTß (diagonal lines), 5C.C7ß plus MCC 26 h before analysis (dark gray), or 5C.C7ß plus MCC 14 h before analysis (light gray) are plotted in b, d, and f. g, The fold increase in the percentage of CD69 very high cells in each deleting group.

 
Further evidence for active deletion of these cells derives from the appearance of a new population of CD4+8-CD69high cells. This population is a minor component of tetramer positive CD4+8- cells from a PBS-injected ß transgenic mouse, but is increased in both sets of MCC-injected animals (Fig. 8Goe). In HELCYT/ß mice the percentage of these CD69high cells that are tetramer positive is indistinguishable from that in PBS-injected ß transgenics. In contrast, MCC-injected mice exhibit a 200-fold increase in the relative number of CD69high cells in the mature CD4+8- compartment 14 h after injection and a 30-fold increase over that in PBS-injected mice 26 h after injection (Fig. 8Gog). These cells are likely to be the cells that have recently received a strong TCR-mediated signal but have not yet fully executed their apoptotic program or that have not yet been eliminated by phagocytosis. One explanation for this result is that the injection of peptide may produce extensive simultaneous deletion that temporarily overwhelms the phagocytic cells found in the thymus. We conclude that negative selection by endogenous MHC peptide complexes can occur throughout thymic development.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
By employing fluorescent MCC/I-Ek tetramers to quantitate the relative number of Ag-specific cells present in the major TCR+ thymocyte population, we have found that developmental stage is not critical for the negative selection of thymocytes. Specifically, we have found evidence for peptide-mediated deletion at the earliest stages of thymic development as defined by CD4/8 coreceptor expression (Fig. 4Go) and CD3 expression (Fig. 6Go). In addition, the tetramer-positive large CD4+8+ cells, which are enriched for cells still subject to positive selection, and the tetramer-positive small CD4+8+ cells, which have been recently shown to contain pre-positively selected cells, are each deleted in HELCYT/ß transgenic mice (Fig. 6Go). Finally, immature cells that have not yet up-regulated CD69 in response to the TCR-mediated signaling involved in positive selection are also deleted in HELCTY/ß mice (Fig. 6Go). By injecting mice with peptide just 14 h before sacrifice, we also find that deletion can occur in later stages of thymic development.

Recent studies have suggested that positive selection may also occur at an earlier stage than was previously thought (30), and that immature thymic subsets are likely to have more complex lineage relationships than those outlined in this paper (29, 31). Thus, the early subsets that we have analyzed may contain some positively selected cells. To address this possibility, we performed additional analyses on smaller subsets defined by CD4, CD8, size, and either CD69 or CD3 and were unable to identify a population of tetramer-positive cells that was not deleted in the presence of the HELCYT transgene. The overwhelming extent of the HELCYT-induced negative selection observed in every thymocyte subset that contains detectable tetramer-positive cells suggests that if there is a detectable population of pre-positively selected cells, they are most likely eliminated in the HELCYT/ß transgenic mice.

Although it is formally possible that in this study, each negatively selected cell receives a positively selecting signal just before deletion or as a component of its negatively selecting signal, several observations argue against these models. When immature CD4+8+/4int8int thymocytes from either TCR {alpha}ß transgenic or nontransgenic mice are removed from positively selecting thymuses and cultured, very few of these immature thymocytes become CD4+8- cells. Instead, unless they are cultured in the presence of thymic stromal cells bearing the correct MHC peptide complexes, they remain at the CD4+8+ stage and eventually die of neglect (32, 41, 42). Thus, even in the presence of a rearranged TCR and expression of the appropriate MHC molecules in vivo, the majority (>90%) of large CD4+8+/4int8int cells have not yet been positively selected. Because we observe >50% deletion in each early thymocyte subset, the deleted cells must include some cells that have not undergone positive selection.

These observations argue strongly against the hypothesis that all T cells in the thymus are first positively selected in the cortex, then negatively selected in the thymic medulla (7). Interestingly, despite evidence showing that cortical thymic epithelial cells are poor or inefficient mediators of negative selection (4, 43), we observed efficient negative selection in cell populations that reside primarily in the thymic cortex. One possible explanation for this is that the endogenous expression of the soluble HELCYT protein permits unusually high expression of deleting complexes on cortical epithelial cells. This possibility is weakened by the fact that cultured thymic epithelial cells from HELCYT mice fail to induce negative selection or enhanced calcium signaling in 5C.C7 TCR transgenic cells in vitro unless exogenous MCC peptide is added to the culture medium (K. K. Baldwin, manuscript in preparation) (20). Alternatively, cortical thymocytes may encounter interdigitating bone marrow-derived cells or subcapsular APCs in the cortex that capture, process, and present the circulating HELCYT protein and subsequently induce deletion of MCC/I-Ek-specific cells. In either case, these data show that early deletion of Ag-specific cells can occur even when TCR expression is developmentally normal, and the population of interest is a minor component of the entire T cell repertoire.

The fact that some Ag-specific cells are eliminated at later stages of thymic development (Fig. 8Go and Table IIGo) indicates that early deletion is incomplete and that, as others have shown, medullary deletion is necessary to ensure the complete elimination of autoreactive cells (4). An interesting question with relevance to autoimmunity is: how do these tetramer-specific cells escape early deletion? One possibility is that they do not encounter the appropriate APCs until the thymocytes enter the deleting cell-rich medulla. Alternatively, the escaping cells may have lower expression levels of coreceptors that enhance deletion (although none of the coreceptors we examined correlated with deletion). Finally, the TCRs expressed by the resistant thymocytes may have a lower avidity for the deleting MHC peptide complexes. Our data suggest that TCR affinity and/or thymocyte avidity for tetramer can play a role in the timing of deletion. Thymocytes that have the highest tetramer binding ability are deleted somewhat more efficiently overall than tetramer low cells (~90 vs 77%). Similarly, higher binding cells are already 66% deleted by the CD4int8int stage compared with 40% in the less avid tetramer-specific population (Fig. 6GoC). Thus, the cells most likely to initiate an autoimmune response are eliminated most efficiently and earliest in the thymus.

Even so, TCR avidity as measured by tetramer binding ability is probably not the sole determinant of the timing of negative selection. A large number of the tetramer low (but specific) cells are eliminated even at the earliest developmental stage, and a few cells that stain brightly with the tetramer persist throughout thymic development (Figs. 5Go and 6Go). These cells are of particular interest as they may represent potentially autoreactive cells. Some tetramer high cells are detectable in the periphery of HELCYT/ß double-transgenic mice as well (Fig. 4Go), yet the mice are healthy and exhibit no obvious autoimmunity. Thus, these self-reactive cells may somehow be rendered unresponsive by peripheral tolerance mechanisms such as anergy. Alternatively, as has been shown in other systems, these autoreactive cells may be induced to undergo Ag-induced cell death in the periphery (reviewed in Ref. 44). Using the peptide/MHC tetramer to track these cells and follow them during the course of immune challenge with MCC or related Ags may provide further insight into the mechanisms of thymic and peripheral tolerance induction.


   Acknowledgments
 
We thank Irving Weissman, Yueh-hsiu Chien, Aude Fahrer, and Lawren Wu for critical reading of the manuscript and helpful discussions. We thank Michael McHeyzer-Williams and Felix Baker for assistance with the single-cell PCR and flow cytometry. We thank Koichi Akashi for helpful discussions and Abs. We also thank Nelida Prado, Dan Lyons, Pete Savage, Jay Boniface, Ines Gutgemann, and Loan Nguyen for assistance with tetramer preparations and staining; Suzanne Ybarra for sequencing; and Paul Fallon for assistance with flow cytometry.


   Footnotes
 
1 This work was supported by a grant (to M.M.D.) from the National Institutes of Health (AI22511). K.K.B. was supported by a Howard Hughes Medical Institute predoctoral fellowship, and J.D.A. was supported by an American Cancer Society fellowship. Back

2 Current address: Howard Hughes Medical Institute/Center for Neurobiology and Behavior, Columbia College of Physicians and Surgeons, New York, NY 10032 Back

3 Address correspondence and reprint requests to Dr. Mark M. Davis, Howard Hughes Medical Institute, Stanford University, 279 Campus Drive, Stanford, CA 94305-5323. E-mail address: Back

4 Abbreviations used in this paper: RAG, recombinase-activating gene; MCC, moth cytochrome c; HSA, heat-stable Ag; HEL, hen egg lysozyme. Back

Received for publication February 1, 1999. Accepted for publication April 30, 1999.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Bendelac, A., P. Matzinger, R. A. Seder, W. E. Auul, R. H. Schwartz. 1992. Activation events during thymic selection. J. Exp. Med. 175:3.
  2. Brandle, D., S. Muller, C. Muller, H. Hengartner, H. Pircher. 1994. Regulation of RAG-1 and CD69 expression in the thymus during positive and negative selection. Eur. J. Immunol. 24:145.[Medline]
  3. Wilson, A., W. Held, H. R. MacDonald. 1994. Two waves of recombinase gene expression in developing thymocytes. J. Exp. Med. 179:1355.[Abstract/Free Full Text]
  4. Laufer, T. M., J. DeKoning, J. S. Markowitz, D. Lo, L. H. Glimcher. 1996. Unopposed positive selection and autoreactivity in mice expressing class II MHC only on thymic cortex. Nature 383:81.[Medline]
  5. Bevan, M. J.. 1997. In thymic selection, peptide diversity gives and takes away. Immunity 7:175.[Medline]
  6. Finkel, T. H., J. C. Cambier, R. T. Kubo, W. K. Born, P. Marrack, J. W. Kappler. 1989. The thymus has two functionally distinct populations of immature {alpha}ß+ T cells: one population is deleted by ligation of {alpha}ß TCR. Cell 58:1047.[Medline]
  7. Guidos, C. J., J. S. Danska, C. G. Fathman, I. L. Weissman. 1990. T cell receptor-mediated negative selection of autoreactive T lymphocyte precursors occurs after commitment to the CD4 or CD8 lineages. J. Exp. Med. 172:835.[Abstract/Free Full Text]
  8. Speiser, D. E., R. K. Lees, H. Hengartner, R. M. Zinkernagel, H. R. MacDonald. 1989. Positive and negative selection of T cell receptor Vß domains controlled by distinct cell populations in the thymus. J. Exp. Med. 170:2165.[Abstract/Free Full Text]
  9. Pircher, H., K. Burki, R. Lang, H. Hengartner, R. Zinkernagel. 1989. Tolerance induction in double specific T-cell receptor transgenic mice varies with antigen. Nature 342:559.[Medline]
  10. Berg, L. J., B. Fazekas de St. Groth, A. M. Pullen, M. M. Davis. 1989. Phenotypic differences between {alpha}ß versus ß T cell receptor transgenic mice undergoing negative selection. Nature 340:559.[Medline]
  11. Kisielow, P., H. Bluthmann, U. Staerz, M. Steinmetz, H. von Boehmer. 1988. Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4+8+ thymocytes. Nature 333:742.[Medline]
  12. Sha, W. C., C. A. Nelson, R. D. Newberry, D. M. Kranz, J. H. Russell, D. Y. Loh. 1988. Positive and negative selection of an antigen receptor on T cells in transgenic mice. Nature 336:73.[Medline]
  13. Teh, H. S., H. Kishi, B. Scott, P. Borgulya, H. von Boehmer, P. Kiselow. 1990. Early deletion and late positive selection of T cells expressing a male specific receptor in T cell receptor transgenic mice. Dev. Immunol. 1:1.[Medline]
  14. Spain, L. M., L. J. Berg. 1992. Developmental regulation of thymocyte susceptibility to deletion by "self"-peptide. J. Exp. Med. 176:213.[Abstract/Free Full Text]
  15. Takahama, Y., E. W. Shores, A. S. Singer. 1992. Negative selection of precursor thymocytes before their differentiation into CD4+CD8+ cells. Science 258:653.[Abstract/Free Full Text]
  16. Cibotti, R., J. Punt, K. Dash, S. O. Sharrow, A. S. Singer. 1997. Surface molecules that drive T cell development in vitro in the absence of thymic epithelium and in the absence of lineage-specific signals. Immunity 6:245.[Medline]
  17. Altman, J. D., P. A. H. Moss, P. J. R. Goulder, D. H. Barouch, M. G. McHeyzer-Williams, J. I. Bell, A. J. McMichael, M. M. Davis. 1996. Phenotypic analysis of antigen-specific T lymphocytes. Science 274:94.[Abstract/Free Full Text]
  18. Gütgemann, I., A. M. Fahrer, J. D. Altman, M. M. Davis, Y. H. Chien. 1998. Induction of rapid T cell activation and tolerance by systemic presentation of an orally administered antigen. Immunity 8:667.[Medline]
  19. Fazekas de St. Groth, B., P. A. Patten, W. Y. Ho, E. P. Rock, and M. M. Davis. 1993. An analysis of T cell receptor-ligand interaction using a transgenic antigen model for T cell tolerance and T cell receptor mutagenesis. Mol. Mech. Immunol. Self-Recognition. 123.
  20. Ho, W. Y. W. 1995. Transgenic mouse models of T cell tolerance and in vitro T cell-B cell collaboration. Doctoral Dissertation, Stanford University, Palo Alto, CA.
  21. Fink, P.J., L. A. Matis, D. L. McElligott, M. Bookman, S. M. Hedrick. 1986. Correlations between T-cell specificity and the structure of the antigen receptor. Nature 321:219.[Medline]
  22. Altman, J. D., P. A. Reay, M. M. Davis. 1993. Formation of functional peptide complexes of class II major histocompatibility complex proteins from subunits produced in Escherichia coli. Proc. Natl. Acad. Sci. USA 90:10330.[Abstract/Free Full Text]
  23. McHeyzer-Williams, M. G., M. M. Davis. 1995. Antigen-specific development of primary and memory T cells in vivo. Science 268:106.[Abstract/Free Full Text]
  24. Sant’Angelo, D., P.G. Waterbury, B. E. Cohen, W. D. Martin, L. H. Van Kaer, A. C. Hayday, C. A. Janeway. 1997. The imprint of intrathymic self-peptides on the mature T cell receptor repertoire. Immunity 4:517.
  25. Crawford, F., H. Kozono, J. White, P. Marrack, J. Kappler. 1998. Detection of antigen-specific T cells with multivalent soluble class II MHC covalent peptide complexes. Immunity 8:675.[Medline]
  26. Egerton, M., R. Scollay, K. Shortman. 1990. Kinetics of mature T cell development in the thymus. Proc. Natl. Acad. Sci. USA 87:2579.[Abstract/Free Full Text]
  27. Kydd, R., K. Lundberg, D. Vremec, A. W. Harris, K. Shortman. 1995. Intermediate steps in thymic positive selection: generation of CD4-8+ T cells in culture from CD4+CD8+, CD4int8+, and CD4+8int thymocytes with up-regulated levels of TCR-CD3. J. Immunol. 155:3806.[Abstract]
  28. Lundberg, K., K. Shortman. 1994. Small cortical thymocytes are subject to positive selection. J. Exp. Med. 179:1475.[Abstract/Free Full Text]
  29. Lucas, B., R. N. Germain. 1996. Unexpectedly complex regulation of CD4/CD8 coreceptor expression supports a revised model for CD4+CD8+ thymocyte differentiation. Immunity 5:461.[Medline]
  30. Sant’Angelo, D., B. Lucas, P. G. Waterbury, B. Cohen, T. Brabb, J. Goverman, R. Germain, C. A. Janeway. 1998. A molecular map of T cell development. Immunity 9:179.[Medline]
  31. Akashi, K., M. Kondo, I. L. Weissman. 1997. Two distinct pathways of positive selection of thymocytes. Proc. Natl. Acad. Sci. USA 95:2486.[Abstract/Free Full Text]
  32. Akashi, K., I. L. Weissman. 1996. The c-kit+ maturation pathway in mouse thymic T cell lineages and selection. Immunity 5:147.[Medline]
  33. Swat, W., L. Ignatowicz, P. Kisielow. 1991. Detection of apoptosis of immature CD4+8+ thymocytes by flow cytometry. J. Immunol. Methods 137:79.[Medline]
  34. Ziegler, S. F., S. D. Levin, L. Johnson, N. G. Copeland, D. J. Gilbert, N. A. Jenkins, E. Baker, G. R. Sutherland, A. L. Feldhaus, F. Ramsdell. 1994. The mouse CD69 gene. Structure, expression, and mapping to the MHC locus. J. Immunol. 152:1228.[Abstract]
  35. Zal, T., A. Volkmann, B. Stockinger. 1994. Mechanisms of tolerance induction in major histocompatibility complex class II-restricted T cells specific for a blood-borne self-antigen. J. Exp. Med. 180:2089.[Abstract/Free Full Text]
  36. Douek, D. C., K. T. Corley, T. Zal, A. Mellor, P. J. Dyson, D. M. Altmann. 1996. Negative selection by endogenous antigen and superantigen occurs at multiple thymic sites. Int. Immunol. 8:1413.[Abstract/Free Full Text]
  37. Hitsumoto, Y., D. S. Song, M. Okada, F. Hamada, S. Saheki, N. Takeuchi. 1996. Enhancement of CD3-mediated thymocyte apoptosis by the cross-linkage of heat-stable antigen. Immunology 89:200.[Medline]
  38. Kishimoto, H., J. Sprent. 1997. Negative selection in the thymus includes semimature T cells. J. Exp. Med. 185:263.[Abstract/Free Full Text]
  39. Kishimoto, H., C. D. Surh, J. Sprent. 1995. Upregulation of surface markers on dying thymocytes. J. Exp. Med. 181:649.[Abstract/Free Full Text]
  40. Shortman, K., D. Vremec, M. Egerton. 1991. The kinetics of T cell antigen receptor expression by subgroups of CD4+8+ thymocytes: delineation of CD4+8+3(2+) thymocytes as post-selection intermediates leading to mature T cells. J. Exp. Med. 173:323.[Abstract/Free Full Text]
  41. Baldwin, K. K., P. A. Reay, L. Wu, A. Farr, M. M. Davis. 1999. A T cell receptor-specific blockade of positive selection. J. Exp. Med. 189:13.[Abstract/Free Full Text]
  42. Sen-Majumdar, A., M. Lieberman, S. Alpert, I. L. Wiessman, M. Small. 1992. Differentiation of CD3-4-8- thymocytes in short-term thymic stromal cell culture. J. Exp. Med. 176:543.[Abstract/Free Full Text]
  43. Lorenz, R. G., P. M. Allen. 1989. Thymic cortical epithelial cells lack full capacity for antigen presentation. Nature 340:557.[Medline]
  44. Kabelitz D, T., T. Pohl, K. Pechhold. 1993. Activation-induced cell death (apoptosis) of mature peripheral T lymphocytes. Immunol. Today 7:338.



This article has been cited by other articles:


Home page
 BloodHome page
A. C. M. Davalos-Misslitz, T. Worbs, S. Willenzon, G. Bernhardt, and R. Forster
Impaired responsiveness to T-cell receptor stimulation and defective negative selection of thymocytes in CCR7-deficient mice
Blood, December 15, 2007; 110(13): 4351 - 4359.
[Abstract] [Full Text] [PDF]

Home page
 J. Immunol.Home page
F. R. Santori, Z. Popmihajlov, V. P. Badovinac, C. Smith, S. Radoja, J. T. Harty, and S. Vukmanovic
TCRbeta Chain That Forms Peptide-Independent Alloreactive TCR Transfers Reduced Reactivity with Irrelevant Peptide/MHC Complex
J. Immunol., May 15, 2007; 178(10): 6109 - 6114.
[Abstract] [Full Text] [PDF]

Home page
 Ann. N. Y. Acad. Sci.Home page
V. GEENEN
Thymus-Dependent T Cell Tolerance of Neuroendocrine Functions: Principles, Reflections, and Implications for Tolerogenic/Negative Self-Vaccination
Ann. N.Y. Acad. Sci., November 1, 2006; 1088(1): 284 - 296.
[Abstract] [Full Text] [PDF]