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 Joachims, M. L. Articles by Thompson, L. F. Search for Related Content PubMed PubMed Citation Articles by Joachims, M. L. Articles by Thompson, L. F. Pubmed/NCBI databases Gene GEO Profiles Nucleotide UniGene

Human and Thymocyte Development: TCR Gene Rearrangements, Intracellular TCR Expression, and Developmental Potential—Differences between Men and Mice^1,2

Michelle L. Joachims^3,*, Jennifer L. Chain^3,*,, Scott W. Hooker^*, Christopher J. Knott-Craig and Linda F. Thompson^4,*,

^* Immunobiology and Cancer Program, Oklahoma Medical Research Foundation, Oklahoma City, OK 73104; Department of Microbiology and Immunology and Department of Surgery, Oklahoma University Health Sciences Center, Oklahoma City, OK 73104

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
To evaluate the role of the TCR in the / lineage choice during human thymocyte development, molecular analyses of the TCR locus in cells and the TCR and loci in cells were undertaken. TCR variable gene segments remained largely in germline configuration in cells, indicating that commitment to the lineage occurred before complete TCR rearrangements in most cases. The few TCR rearrangements detected were primarily out-of-frame, suggesting that productive TCR rearrangements diverted cells away from the lineage. In contrast, in cells, the TCR locus was almost completely rearranged with a random productivity profile; the TCR locus contained primarily nonproductive rearrangements. Productive rearrangements were, however, depleted compared with preselected cells. Productive TCR and rearrangements rarely occurred in the same cell, suggesting that cells developed from cells unable to produce a functional TCR. Intracellular TCR expression correlated with the up-regulation of CD4 and concomitant down-regulation of CD34, and plateaued at the early double positive stage. Surprisingly, however, some early double positive thymocytes retained potential in culture. We present a model for human thymopoiesis which includes development as a default pathway, an instructional role for the TCR in the / lineage choice, and a prolonged developmental window for selection and lineage commitment. Aspects that differ from the mouse are the status of TCR gene rearrangements at the nonexpressed loci, the timing of selection, and maintenance of potential through the early double positive stage of development.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
During thymocyte development, multipotent precursor cells from the bone marrow differentiate into cells of two distinct T cell lineages, and . The successive stages of both murine and human thymocyte development can be broadly categorized as double negative (DN)⁵), double positive (DP) or single positive (SP) according to the expression of the CD4 and CD8 coreceptors. Immature murine DN thymocytes can be separated into four populations (DN I-IV) based on the expression of CD44 and CD25 (1). Immature human thymocytes do not express the same surface markers, however, so the corresponding human DN subpopulations are characterized by the differential expression of CD34, CD38, and CD1a (reviewed in Refs.2, 3, 4). The earliest thymic progenitors are CD34⁺CD38^–CD1a^–, followed by CD34⁺CD38⁺CD1a^–, and CD34⁺CD38⁺CD1a⁺ cells, with CD1a expression correlating with T lineage commitment (5). While murine thymocytes mature via a CD8 immature SP (ISP) intermediate in between the DN and DP stages (6), human thymocytes traverse a CD4 ISP stage following the acquisition of CD1a (7). Human thymocytes have two discrete stages of DP development before surface expression of CD3. The first early DP (EDP) stage is characterized by the expression of CD8, resulting in cells that are CD4⁺CD8^+–CD3^– (8), and the second DP blast stage results from the up-regulation of CD8, yielding CD4⁺CD8⁺⁺CD3^– cells. The process in which the TCR chain pairs with the pre-TCR protein to produce a membrane-localized pre-TCR that signals survival, expansion, and allelic exclusion is referred to as selection, and occurs synchronously in murine thymocytes at the DNIII (CD44^–CD25⁺) to DNIV (CD44^–CD25^–) transition (9, 10). However, the point of selection in human thymocyte development is still controversial. Though one study indicated that selection occurs at the EDP to DP blast transition (11), other studies suggested it begins as early as the CD34⁺CD38⁺CD1a⁺ (4) or CD4 ISP stages (12). In both humans and mice, selection leads to rearrangement of the TCR locus and expression of a complete TCR (2, 13). This is followed by positive and negative selection, and the development of mature CD4 and CD8 SP cells (2, 14).

Since and cells derive from a common precursor (15, 16), the issue of when the two lineages diverge during development is of considerable interest. In both mice and humans, TCR rearranges first, followed closely by TCR and then TCR, with TCR rearrangement occurring later at the DP stage (4, 12, 17, 18). In mice, , , and rearrangements are completed while thymocytes are still in the DN stage, and thymocytes derive mainly from DN cells (1, 19). However, in humans, a substantial fraction of complete TCR and rearrangements and the significant onset of complete TCR rearrangements do not occur until the CD4 ISP stage (4, 12). Thus, it is likely that cells diverge from the main pathway of thymocyte development at a later stage in humans than in mice. Indeed, human CD4 ISP cells have the capacity to develop into cells as shown by retroviral overexpression studies (20), but the point in development when potential is lost has not been reported.

As TCR gene rearrangements undoubtedly impact the process of / lineage commitment, several models have been proposed to explain their role in this process (for reviews, see Refs.21, 22, 23, 24). The instructive model asserts that the TCR plays a primary role in determining lineage fates. In this model, / T cell precursors are bipotent before TCR gene rearrangements, but formation of a functional TCR instructs the cell to develop as a cell, while expression of the pre-TCR complex directs the cell to become an cell. This model is supported by data showing depletion of in-frame and rearrangements in murine cells (15, 17, 25, 26). The stochastic or "separate lineage" model asserts that / lineage commitment is independent of, and probably precedes, TCR gene rearrangements (27, 28, 29). Rearrangements of the TCR , , and/or genes occur in each developing thymocyte, but only cells that make productive rearrangements of the TCR genes that match the cell fate predetermined by other factors are selected to survive. Support for this model comes from studies of mice unable to assemble a pre-TCR in which lineage cells develop using the TCR (30, 31, 32). Factors such as IL-7R expression (33) or Notch signaling (34, 35, 36, 37) may also play roles in directing thymocytes into either the or lineage. Finally, the newer signal strength model postulates that either type of TCR can direct development into both the and lineages, but the strength of signal is what determines lineage choice (38, 39). Strong signals promote development, while weaker signaling leads to commitment. This model is supported by previously unreconciled data indicating that a given TCR can promote cross-lineage development; i.e., the TCR allowing the development of DP lineage cells in TCR^–/– mice (30, 31, 32) and the TCR promoting the development of DN lineage cells in TCR transgenic mice (40, 41).

Considerable insight into the contribution of the TCR to the / lineage decision can be gleaned from an analysis of the relevant TCR loci in human thymocytes (4, 12, 42). In this study, we present analyses of the extent and productivity of rearrangement of the locus in thymocytes and of the and loci in thymocytes. We are the first group to use single-cell PCR to assess the productivity of both and rearrangements in single primary thymocytes. Our data reveal significant differences between humans and mice that impact the role of the TCR in the / lineage choice. Furthermore, we determine the developmental potential and expression of intracellular TCR in successive phenotypically defined thymocyte populations. These findings are integrated into a model for human and thymocyte development and discussed in relationship to each of the previously proposed models for / lineage divergence. Our results illustrate unique features of human thymopoiesis including depletion of in-frame rearrangements in thymocytes, a prolonged developmental window of selection, and preservation of potential through the EDP stage of differentiation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Abs, cells, and cell isolations

Abs used were as follows: FITC anti-CD1a, allophycocyanin anti-CD34, PE anti-TCR, FITC anti-CD8, and PE anti-TCR obtained from BD Pharmingen; PE and PE-Texas Red anti-CD8, PE-Cy5 and allophycocyanin anti-CD4, PE anti-CD34, FITC and PE anti-TCR, and FITC and allophycocyanin anti-CD3 obtained from Caltag Laboratories; purified or PE-labeled anti-CD8 obtained from Serotec; purified or PE anti-TCRC1 obtained from Ancell; and the relevant isotype control Abs. Purified anti-CD8 was labeled with Alexa Fluor 488 (Molecular Probes). Human thymus was obtained from cardiac surgeries on infants and children at Children’s Hospital in Oklahoma City, OK under protocols approved by the Institutional Review Boards of both the University of Oklahoma and the Oklahoma Medical Research Foundation. Single-cell suspensions were made by forcing thymic tissue though a 70-µm nylon filter. Human thymocytes were prepared using the Miltenyi isolation kit (Miltenyi Biotec), followed by sorting to >98% purity on a MoFlo Cell sorter (Cytomation). Human thymic cells were prepared by sorting anti-human TCR-stained thymocytes to a purity of at least 98%. Human CD34⁺ thymocytes were enriched with anti-CD34 magnetic beads (Dynal Biotech), and subsequently stained with Abs to CD34, CD1a, CD4, and CD8. CD34⁺CD4^–CD8^–CD1a^– and CD34⁺CD4^–CD8^–CD1a⁺ cells were then sorted to >98% purity. CD4 ISPs were isolated using a two-step procedure. Thymocyte suspensions were first depleted of CD3⁺ cells using EasySep CD3 magnetic beads (Stem Cell Technologies), followed by depletion of CD8⁺ cells using EasySep CD8 beads. CD8-depleted cells were then stained for CD4, CD8, and CD3 expression and CD4 ISPs were sorted as CD4⁺CD3^–CD8^– cells. EDPs were isolated from the fraction of cells bound to EasySep CD8 beads by staining with Abs to CD8, CD8, CD3, and CD4, then sorting for CD4⁺CD8^+–CD3^– cells. HeLa cells, used as a source of germline control DNA in PCR experiments, were grown in DMEM containing 10% FCS (HyClone), 2 mM glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin (Invitrogen Life Technologies).

Quantitation of the extent of TCR and locus rearrangements by real-time PCR

Genomic DNA was prepared using the Puregene kit (Gentra Systems). Primers and probes for real-time PCRs have been described previously (43) and were purchased from Applied Biosystems and Sigma Genosys. Real-time PCRs for the quantitation of germline TCR and TCR loci were performed as previously described (43) using a strategy to detect a germline DNA amplicon which is deleted upon TCR VDJ or VJ recombination, similar to that described for quantitative Southern blotting (17, 18, 44, 45). The amount of germline DNA detected was normalized to an amplicon not deleted during TCR rearrangements, and the values were used to calculate the percentage of germline DNA remaining in each population. The percentage of germline DNA at the V locus in cells was corrected for the presence of TCR excision circles (-TRECs) (43).

Amplification and sequencing of TCR gene rearrangements

TCR gene rearrangements from functional V regions were amplified using the primers listed in Tables I and II; the TCR gene nomenclature is that of the International ImMunoGeneTics database (IMGT; http://imgt.cines.fr). TCR rearrangements were amplified by multiplex PCR from 200 ng of thymocyte genomic DNA using 0.2 µg of forward and reverse primer (Table II), and 1 U of JumpStart TaqDNA Polymerase (Sigma-Aldrich) in a 30-µl reaction. Cycling conditions were: 94°C for 5 min, 35 cycles of 94°C, 30 s; 60°C, 30 s; 72°C, 30 s, with a final elongation of 5 min at 72°C. TCR gene rearrangements were amplified from 200 ng of thymocyte genomic DNA using 1 µM forward and reverse primers (Table II) and 1 U of JumpStart TaqDNA Polymerase in a 25-µl reaction. Cycling conditions were as above, using 55°C annealing for J2.7 reactions and 62°C annealing for all other reactions, followed by an extension at 72°C for 2.5 min. PCR products were gel-purified, cloned, and sequenced. Unique sequences without a premature stop codon and with a preserved amino acid joining sequence were counted as in-frame.

View this table: [in this window] [in a new window]   Table II. Primers for population sequencing of TCR and TCRa

Single-cell PCR analysis of TCR and TCR rearrangements

Unfractionated thymocytes were stained with PE anti- TCR and single TCR⁺ cells were sorted into 96-well plates containing 20 µl of 0.2 µg/µl Proteinase K (Amresco) in 1x JumpStart TaqPCR buffer. Plates were incubated at 50°C for 50 min for Proteinase K digestion, followed by 95°C for 10 min. The first-round PCR used 15 ng each of the forward and reverse "outside" primers in Table I ( and ), and 1 U of JumpStart TaqDNA Polymerase in a 50-µl reaction under the following cycling conditions: 94°C, 3 min, 36 cycles of 94°C, 20 s; 60°C, 40 s; 72°C, 30 s with a final elongation of 5 min at 72°C. The first round PCR products (8 µl) were dephosphorylated with shrimp alkaline phosphatase (SAP; Roche Applied Science) and treated with Exonuclease I (ExoI; Epicentre) to remove leftover primers in a 10-µl reaction of 3.6x SAP buffer (25 mM Tris-HCL, 1 mM MgCl₂, 0.1 mM ZnCl₂, 50% glycerol, pH 7.6), 0.5U SAP, and 1 U of ExoI, incubated at 37°C for 60 s, followed by 85°C for 15 min. A second round of nested PCR for TCR was then performed using 2.5 µl of the SAP/ExoI-treated products from each cell with 15 ng each of the forward and reverse "inside primers" (Table I), and 1 U of JumpStart TaqDNA Polymerase in a 30-µl reaction under cycling conditions of 94°C for 3 min, 36 cycles of 94°C, 20 s; 60°C, 20 s; 72°C, 20 s with a final elongation of 5 min at 72°C. Products were analyzed by electrophoresis on 2.5% agarose gels, and detectable TCR rearrangements were gel-purified, cloned, and sequenced. For cells with an in-frame TCR, TCR rearrangements were amplified using "inside" primers (Table I), cloned, and sequenced as described for TCR.

Intracellular TCR expression and cell cycle analysis

The following populations of thymocytes were assessed for the expression of intracellular TCR (TCR^ic): DN CD34⁺CD1a^–, DN CD34⁺CD1a⁺, CD34⁺CD4⁺ISP, CD34^–CD4⁺ISP, EDP, CD3^–DP blasts, and CD3⁺DP. Enriched populations of DN and CD4 ISPs were obtained by depleting CD3- and CD8-expressing cells with EasySep beads as described above. Enriched populations of EDPs were obtained from cells binding to CD8 beads during the above depletions. TCR^ic expression on DP cells was analyzed on unfractionated thymocytes. Staining for TCR^ic expression was performed essentially as described previously by fixing cells with 1% formaldehyde and permeabilizing with 0.5% saponin (46). Purified anti-TCRC1 was added at the same time as other Abs to cell surface proteins to block surface staining of TCR. Samples were analyzed using an LSR II flow cytometer and CellQuest software (BD Biosciences). Isotype controls were used to set gates, and the percentages of TCR^ic-expressing cells were obtained by subtracting the percentages of cells stained intracellularly with an isotype control Ab.

The following populations of cells were analyzed for cell cycle status by staining with propidium iodide (PI): CD34⁺CD4⁺TCR^ic-, CD34⁺CD4⁺TCR^ic+, CD34^–CD4⁺TCR^ic–, and CD34^–CD4⁺TCR^ic+. These populations were sorted to >98% purity from CD3^–CD8^– cells prepared as described above. Sorted cells were resuspended in 285 µl of a 50:50 mixture of FCS and PBS and then fixed with 715 µl of cold 70% ethanol added dropwise with gentle vortexing. Fixed cells were resuspended in 1 ml of PBS containing PI (50 µg/ml) and RNase A (500 µg/ml; Qiagen) and incubated for 20 min at 37°C. Cells were then cooled on ice and analyzed for DNA content with a FACScan flow cytometer (BD Biosciences) and CellQuest software.

Chimeric human/mouse fetal thymic organ cultures (hu/mo FTOC)

Hu/mo FTOC was performed essentially as described (47). Reconstituted deoxyguanosine-treated murine fetal thymic lobes were incubated for up to 3 wk in Yssel’s medium (48) supplemented with 2% human AB serum and 5% FCS. Upon harvest, cells were counted, stained with Abs to human TCR and TCR, and analyzed using a FACSCalibur flow cytometer and CellQuest software (BD Biosciences).

	Results

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
The extent of TCR rearrangements in human thymocytes

To investigate the role of the TCR in human vs lineage commitment, the extent of complete TCRV gene rearrangements was analyzed in thymocytes using quantitative real-time PCR. This assay (43) quantitates the amount of germline DNA in between the most 3' V segment and D1 remaining in a given sample, using a strategy analogous to that of quantitative Southern blotting previously used in analyses of murine thymocytes (17, 18, 44, 45) and includes a correction for the contribution of -TRECs to the germline signal. In the case of thymocyte DNA, only 4.6 ± 2.4% (mean ± SD) of the germline TCR signal was attributed to -TRECs (data not shown). Because only complete VDJ rearrangements enable a cell to produce a functional TCR chain, immature DJ rearrangements were not evaluated. Therefore, the percentage of germline DNA refers to the proportion of the locus whose variable genes have not rearranged. Fig. 1 shows the percentages of the TCR locus rearranged in six (8.7 ± 6%) and two (58 ± 4%) thymocyte isolates. The data indicate that most cells had not undergone any complete TCR rearrangements. The observed extent of TCR rearrangement in thymocytes (55–60%) was slightly less than the value predicted for an allelically excluded thymocyte population (70%; Ref.45), perhaps because some rearrangements occurred by inversion rather than by deletion.

View larger version (16K): [in this window] [in a new window]   FIGURE 1. The TCRV variable gene segments remain largely in germline configuration in human thymocytes. Genomic DNA was isolated from purified human and thymocytes and the percent of the V locus remaining in germline configuration was determined by real-time PCR as described in Materials and Methods. For the DNA samples, these percentages were corrected for the contribution of -TRECs. These values were subtracted from 100 to give the percent of the locus rearranged. The percent of TCRV rearranged in each DNA sample is shown as the mean ± SD from six replicates.

View larger version (25K): [in this window] [in a new window]   FIGURE 2. The TCR locus is highly rearranged in human and thymocytes. Genomic DNA was isolated from purified human thymic and cells and the percent of the locus remaining in germline configuration was determined by real-time PCR as described in Materials and Methods. These percentages were subtracted from 100 to give the percent of the locus rearranged, which is shown as the mean ± SD for six replicates.

To assess the potential of the rearranged TCR genes to generate expressed TCR proteins, the productivity of TCR and TCR rearrangements was analyzed in and thymocytes, respectively. TCR gene rearrangements were amplified by PCR of genomic DNA using primers specific for V- and J-gene segments (Table II), then cloned and sequenced. As shown in Table III, productive TCR rearrangements using either the J1 or the J2 cluster were underrepresented in all three isolates of cells relative to the value expected if such rearrangements were to occur randomly (33%), suggesting that the small percentage of cells with TCR rearrangements may have undergone some type of selection, either through lineage commitment or survival. Interestingly, gene rearrangements using the J2 cluster, the more commonly used cluster in human peripheral blood T cells (49), were statistically more likely to be nonproductive, compared with those using the J1 cluster (p = 0.0002, by ² analysis). In contrast, in two independent thymocyte isolates, approximately one-third of the TCR gene rearrangements were in-frame (30%; Table IV), suggesting that the presence of an in-frame rearrangement did not preclude development.

View this table: [in this window] [in a new window]   Table III. Productivity of TCR gene rearrangements in human thymocytes

View this table: [in this window] [in a new window]   Table IV. Productivity of TCR gene rearrangements in human thymocytesa

TCR and rearrangements in single thymocytes

To determine whether cells expressing an TCR ever had the chance to make a functional TCR and to become a cell earlier in development, TCR and rearrangements were cloned and sequenced in single thymocytes. Because the locus is deleted during TCR rearrangement and -TRECs are lost with cell division, we expected that many of the cells would not contain detectable sequences. Therefore, to avoid sequencing rearrangements from many cells that would not have detectable rearrangements, we sequenced first. For those cells with detectable sequences, we then cloned and sequenced rearrangements. Out of a total of 1100 single cells analyzed, 134 cells had detectable rearrangements (Fig. 3, lower panel). Only 19 (14%) of these sequences were in-frame (Table V). Assuming that the productivity of the rearrangements that were deleted and lost is similar to that of the sequences we obtained, these data suggest that between 14% (if one allele per cell is rearranged) and 26% (if both alleles are rearranged) of thymocytes had at least one allele rearranged in-frame. From the 19 cells containing in-frame TCR rearrangements, we were able to amplify and sequence 21 of 38 alleles (Fig. 3, upper panel), only 3 of which (14%) were productively rearranged. If both alleles are rearranged in almost all cells as suggested by our real-time PCR data, this frequency of in-frame alleles should lead to 26% of cells with at least one allele in-frame. Therefore, the percentage of thymocytes with rearrangements of both and in-frame would be predicted to be between 3.6% (14% of 26%) and 6.8% (26% of 26%), depending upon the average number of alleles rearranged per cell. Thus, our analyses indicate that the vast majority of thymocytes could not have expressed a TCR at some point earlier in development. It is impossible to know whether those few cells with in-frame rearrangements of both and could have produced a functional TCR, as not all combinations of and chains can make a functional receptor (50).

View larger version (75K): [in this window] [in a new window]   FIGURE 3. Single-cell PCR analysis of TCR and rearrangements in human thymocytes. Genomic DNA was isolated from single human thymocytes and nested PCRs performed to detect either TCR or TCR gene rearrangements as described in Materials and Methods. A representative agarose gel of PCR products from 10 individual thymocytes is shown. M, m.w. markers.

View this table: [in this window] [in a new window]   Table V. Productivity of TCR gene rearrangements in single thymocytesa

Intracellular TCR expression in developing human thymocytes and evidence for selection

To correlate the status of TCR gene rearrangements in developing thymocytes with expression of TCR protein, a comprehensive analysis of intracellular TCR (TCR^ic) expression was performed. Cells were enriched for early thymocyte populations as described in Materials and Methods, and stained for TCR^ic in combination with markers to define each subpopulation. Representative histograms are shown in Fig. 4A and the percentages of TCR^ic+ cells from individual experiments are shown in Fig. 4B. TCR^ic expression was virtually undetectable in the earliest DN CD34⁺CD1a^– population and only small percentages of cells were TCR^ic+ at the next CD34⁺CD1a⁺ stage. These findings agree with previously published data indicating that complete VDJ rearrangements are not readily detectable until the CD4 ISP stage (4, 12). However, modest to significant levels of TCR^ic were detectable in the CD4 ISP subsets, with higher percentages of TCR^ic-expressing cells correlating with the down-regulation of surface CD34. On average, 85% of EDPs displayed a high level of TCR^ic expression. CD3^–/low and CD3⁺ DP cells (expressing CD4 and CD8) showed only slightly increased percentages of TCR^ic+ cells when compared with EDPs. These data indicate that by the time the cells reach the EDP stage, the majority are either in the process of, or have undergone, selection. Fig. 4B shows that the major increase in TCR^ic expression begins in the CD4 ISP stage, and reaches a maximum by the EDP stage. It is also interesting to note the degree to which individuals vary in the timing and extent of TCR^ic expression (Fig. 4B).

View larger version (41K): [in this window] [in a new window]   FIGURE 4. Intracellular TCR expression and cell cycle status in human thymocyte subpopulations. Thymocyte populations were enriched from neonatal thymus and stained for intracellular TCR expression as described in Materials and Methods. A, Cells were electronically gated on each subpopulation. DNs were defined as CD4–CD8–CD34+CD1a– or CD1a+. ISPs were gated as CD3–CD8–CD4+CD34+ or CD34–. EDPs were defined as CD3–CD8–CD4+CD8+. DPs were gated as CD4+CD8+ and CD3– was used to define the blast population. The filled histograms indicate the PE-TCRic staining in each cell population, while the isotype control (PE-IgG2a) is depicted by the black line tracing. The percentage of cells expressing TCRic after subtracting the isotype control percentage is indicated in each histogram. B, Individual values for the percentages of TCRic-expressing cells in various thymocyte subpopulations from up to 10 different thymi are shown. The mean %TCRic+± SD for each population is plotted in the line graph. C, CD34+/–CD4 ISP TCRic+/– cells were sorted and analyzed for cell cycle status by staining with PI. The percentages of cells in the (S + G2-M) phases of the cell cycle are shown.

Cells through the EDP stage can give rise to thymocytes in hu/mo FTOC

The and developmental potential of CD34⁺CD1a⁺, CD4 ISP, and EDP thymocytes was assessed in hu/mo FTOC. After 1–3 wk, the cells were harvested and stained with Abs to and TCR. All three populations generated large numbers of TCR-expressing cells (Fig. 5B). Fig. 5A shows representative histograms of surface TCR expression on gated populations of (DN + CD8 SP) cells, fractions enriched for T cells. The percentage of TCR⁺ cells in this subpopulation is highest in cultures initiated with CD34⁺CD1a⁺ thymocytes, but significant levels of TCR⁺ cells were also observed in cultures derived from both CD4 ISP and EDP cells. Fig. 5B shows the absolute numbers of TCR⁺ and TCR⁺ cells generated from each subpopulation in four separate experiments. Although the absolute numbers of TCR⁺ cells declined as the input cells were more mature, the EDP population had not lost its ability to yield cells even though an average of 85% of EDP cells expressed TCR^ic.

View larger version (38K): [in this window] [in a new window]   FIGURE 5. developmental potential of early human thymocyte subpopulations in hu/mo FTOC. Human thymocyte subpopulations (CD34+CD1a+, CD4 ISP, and EDP) were isolated and placed into hu/mo FTOC as described in Materials and Methods. A, Cells harvested from 3-wk cultures were stained with PE mouse anti-TCR or PE mouse IgG1 as a control. The histograms represent the TCR expression on gated (DN + CD8 SP) cells. The gray lines represent isotype control staining. B, The absolute numbers of either (left graph) or thymocytes (right graph) generated from various subpopulations of thymocytes are shown for four independent experiments. Experiment 4 involved only a 1-wk culture of EDP cells rather than 3 wk as done for all the others. Note the difference in scale between the two graphs.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

VDJ

View larger version (21K): [in this window] [in a new window]   FIGURE 6. Model for human and thymocyte development. Human and thymocytes develop from a common precursor. developmental potential decreases with increasing maturation, but is retained until at least the EDP stage. Decreasing potential is paralleled by increasing proportions of cells expressing TCRic. Thus, cells undergo selection and commitment to the lineage as they progress through several phenotypic stages (see Discussion for detailed explanation).

VDJ

Our findings for human thymocytes are quite different. First, a much smaller percentage of the locus showed complete rearrangements than in the mouse (8.7% on average, Fig. 1). Since our results were corrected for the presence of -TRECs (43), this cannot be the reason for the higher percentage of germline DNA in our experiments. We did find extensive DJ rearrangements in human cells (data not shown), indicating that early on, at least part of the TCR locus was accessible to the recombination machinery. Our data are consistent with those of Couedel et al. (53) who showed that cloned peripheral blood T cells contained a low percentage of complete VDJ rearrangements. Second, and more importantly, the few complete rearrangements detected were predominately out-of-frame (Table III), suggesting an instructional role for the pre-TCR in diversion away from the lineage.

In one of the few studies with primary human cells available to compare with ours, Margolis et al. (42) concluded that cells from individual thymi followed different pathways of development, depending upon the timing of and gene rearrangements. In some thymi, the locus was rearranged in cells to the same extent as in cells. Furthermore, in one of nine thymi, cells had predominately in-frame rearrangements. The discrepancies between their results and ours are difficult to reconcile, but may be related to methodology, as they used PCR spectrotyping to evaluate both the extent and productivity of rearrangements, rather than quantitative real-time PCR and DNA sequencing. We may have failed to see thymocytes with predominately in-frame rearrangements, as this was an infrequent finding in their studies (one of nine samples). In any case, as there were very few TCR VDJ rearrangements in any of our six preparations of thymocytes, we conclude that most human thymocytes become lineage-committed before complete VDJ gene rearrangements begin. Whether the few cells with complete in-frame rearrangements die by apoptosis or are diverted to the lineage is currently unknown, making the impact of productively rearranged TCR genes on development of interest for future studies.

The second important difference between humans and mice is the status of the TCR locus in thymocytes. We found almost complete rearrangement (97%) of the TCR locus in these cells, and the rearrangements showed a random productivity profile (Fig. 2, Table IV). Although the TCR locus has been reported to be highly rearranged in lineage leukemic cells (54) and alloreactive T cell clones (55), ours is the first assessment of the extent of TCR gene rearrangement in normal primary human thymocytes. The TCR locus is also highly rearranged in murine cells, though the exact extent is difficult to discern from the literature, as the organization of the murine TCR locus precludes a straightforward analysis analogous to ours (17, 56, 57). Several reports indicated that murine TCR rearrangements are likely subject to allelic exclusion (58, 59), while our data and those of others suggest a lack of allelic exclusion at the human TCR locus (54, 55, 60). Furthermore, thymocytes from mice were selectively depleted of in-frame TCR gene rearrangements (15, 17, 25), suggesting that cells derived from thymocytes unable to productively rearrange their locus. In contrast, our data showed a random productivity profile of TCR rearrangements in cells, suggesting at first glance, that TCR rearrangements do not influence lineage commitment. However, this would be surprising given the strong selection against cells that can express a TCR during development (see below). One possibility is that this selection is mediated at the level of - and -chain pairing due to constraints placed on which V chains can pair with V proteins to produce a functional TCR. Precedence for this mechanism is found in studies of mouse cells (50). A second explanation for this apparent contradiction is that V-J recombination could be biased toward productive rearrangements due to microhomology-based joining (61) and that the experimental value of 30% in-frame TCR rearrangements could actually reflect a selection against in-frame rearrangements. To test this possibility, we analyzed TCR rearrangements in CD34⁺CD4 ISP cells, a population unlikely to be committed to the lineage due to its low percentage of TCR^ic-expressing cells. Preliminary results revealed that these cells contained 40% productive TCR rearrangements (61 of 151 sequences), a finding that by 2 analysis is significantly different (p = 0.048) from that for thymocytes (30%, Table IV). However, this value is not significantly different than a random distribution (33%, p = 0.18 by 2 analysis), raising the question of whether in-frame rearrangements are truly depleted in thymocytes. Further experimentation will be required to determine whether human TCR rearrangements are subject to microhomology domain biases and to fully understand the role of these gene rearrangements in the vs lineage decision.

The finding that one-third of the TCR rearrangements in thymocytes were in-frame raised the possibility that significant proportions of cells might have the potential to express a functional TCR. To address this issue and to assess the impact of TCR rearrangements on the entry of progenitors into the potential pool, we analyzed the productivity of over 100 TCR rearrangements in sorted single thymocytes. TCR rearrangements were significantly less productive (14%) than would be predicted by random chance (33%) (Table V), indicating a selective mechanism for the depletion of cells with in-frame rearrangements during human thymocyte development. A similar situation occurs during murine development (15, 26). To address the issue of whether any thymocytes had the potential to express a functional TCR, we analyzed the productivity of TCR gene rearrangements in those cells with productive rearrangements. Since at least 93% of human thymocytes had nonproductive rearrangements at the and/or locus, we conclude that the vast majority of cells that develop into thymocytes are those that could not express a functional TCR. As in the mouse, these data indicate an important role for the expression of a functional TCR in the vs lineage decision.

A third significant difference between human and murine thymocyte development is a prolonged window of development through which selection and lineage commitment occur simultaneously. Murine thymocytes can develop into both and thymocytes through the DNIII stage, but show greatly reduced potential in the DNIV compartment as assessed by culture of sorted thymocytes in FTOC (1). Thus, in the mouse, it is well-accepted that coexpression of CD4 and CD8 marks commitment to the lineage. In contrast, we show here for the first time, that developmental potential persists into the later phases of human thymocyte development until at least the CD3^–CD4⁺CD8^+– (EDP) stage (Fig. 5). Even though 85% of EDP cells expressed TCR^ic (Fig. 4), a small percentage with this cell surface phenotype remained uncommitted in terms of lineage decision, and still produced cells in hu/mo FTOC (Fig. 5). The potential in the next developmental stage (DP blasts) was greatly diminished, as cells cannot be identified with certainty in hu/mo FTOC initiated with this population (data not shown). potential correlated inversely with the observed expression of TCR^ic in human thymocyte populations. DN CD34⁺CD1a⁺ cells were the first population to have a small percentage of cells expressing TCR^ic (Fig. 4). These data are consistent with recent work by Dik et al. (4) who also placed the onset of complete TCR rearrangements at this stage. However, pre-T protein expression is not appreciably detectable before the CD4 ISP stage (62), so it is possible that TCR^ic expression is not synchronously linked to the expression of pre-TCR and selection. We did find significant percentages of TCR^ic+ cells at the CD4 ISP stage, consistent with the findings of Blom et al. (12), especially in the subpopulation that had down-regulated CD34 expression (Fig. 4). In fact, the TCR^ic+ cells in the DN CD34⁺CD1a⁺ fraction had significantly lower CD34 expression than the remainder of the population (data not shown), suggesting that down-regulation of CD34 correlates with the onset of complete TCR gene rearrangements and/or selection. Populations of CD4 ISP cells expressing TCR^ic