Thymocyte Differentiation During Discrete Stages of Postnatal

Analysis of Transcription Factor Expression during Discrete Stages of Postnatal Thymocyte Differentiation

This information is current as Sahba Tabrizifard, Alexandru Olaru, Jason Plotkin, of October 2, 2021. Mohammad Fallahi-Sichani, Ferenc Livak and Howard T. Petrie J Immunol 2004; 173:1094-1102; ; doi: 10.4049/jimmunol.173.2.1094 http://www.jimmunol.org/content/173/2/1094 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2004/07/02/173.2.1094.DC1 Material http://www.jimmunol.org/ References This article cites 64 articles, 34 of which you can access for free at: http://www.jimmunol.org/content/173/2/1094.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on October 2, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2004 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Analysis of Transcription Factor Expression during Discrete Stages of Postnatal Thymocyte Differentiation1

Sahba Tabrizifard,* Alexandru Olaru,† Jason Plotkin,* Mohammad Fallahi-Sichani,* Ferenc Livak,†‡ and Howard T. Petrie2*§

Postnatal T lymphocyte differentiation in the thymus is a multistage process involving serial waves of lineage specification, proliferative expansion, and survival/cell death decisions. Although these are believed to originate from signals derived from various thymic stromal cells, the ultimate consequence of these signals is to induce the transcriptional changes that are definitive of each step. To help to characterize this process, high density microarrays were used to analyze transcription factor gene expression in RNA derived from progenitors at each stage of T lymphopoietic differentiation, and the results were validated by a number of appropriate methods. We find a large number of transcription factors to be expressed in developing T lymphocytes,

including many with known roles in the control of differentiation, proliferation, or cell survival/death decisions in other cell types. Downloaded from Some of these are expressed throughout the developmental process, whereas others change substantially at specific developmental transitions. The latter are particularly interesting, because stage-specific changes make it increasingly likely that the corresponding transcription factors may be involved in stage-specific processes. Overall, the data presented here represent a large resource for gene discovery and for confirmation of results obtained through other methods. The Journal of Immunology, 2004, 173: 1094–1102. http://www.jimmunol.org/ he continuous, lifelong production of new T cells by the mark of T lineage commitment, in the form of rearrangements of thymus is an essential requirement in the maintenance of the TCR loci (10, 11), particularly D␤-J␤ rearrangements (11). T immune function and homeostasis (1). Thymocyte differ- Nonetheless, the thymus generates numerous T lineage cell types, entiation is a stepwise process initiated by bone marrow progeni- including CD4, CD8, ␥␦, NK-T, and T-regulatory cell types (to tors that home to the thymus via the blood. The earliest intrathymic name a few), and it remains unclear at what point these lineages progenitors are defined by the absence of mature T lineage markers diverge. Thus, although DN3 cells are T lineage restricted, addi- (CD4CD8 double-negative; DN3). This population is heteroge- tional lineage specification events continue to be required even neous and can be divided into multiple stages (reviewed in Ref. 2). after this stage. The earliest of these, designated DN stage 1 (DN1), can be iden- Differentiation onward from the DN3 stage appears to occur by guest on October 2, 2021 tified by the expression of CD44 and CD117 and by the absence of only in cells that have generated productive rearrangements at one CD25. DN1 cells are not T lineage restricted and can give rise to or more of the TCR loci (12, 13) and coincides with expression of multiple other lineages under appropriate conditions (3Ð6). After both CD4 and CD8 mature lineage markers (double-positive; DP). an average of ϳ10 days of intrathymic residence (7, 8), DN1 cells The early part of the DP phase (preDP, commonly but inappro- asynchronously leave the perimedullary regions, where they first priately referred to as DN4) is characterized by cells with low enter the thymus (8, 9), and enter the cortex proper (8). This move- surface expression of CD4 and CD8 (14), as well as by the pres- ment corresponds to differentiation into the DN2 stage and is ence of complete, productive V␤-DJ␤ rearrangements (12). The marked by the acquisition of CD25. Cells at the DN2 stage are TCR␤ protein is also expressed on the cell surface at this time (15), more lineage restricted, although they can still give rise to non-T in conjunction with an invariant pre-T ␣-chain (16) and CD3 (14). lineages (3Ð6). In contrast, cells at the next stage of development, Rearrangement of the TCR␣ locus is then initiated (11), and CD4 ϩ low low designated DN3 (CD25 44 117 ), bear the irreversible hall- and CD8 accumulate at high levels on the cell surface. Cytogenesis then ceases for the most part (17, 18), resulting in the production of nondividing, small DP cells (smDP). Further differentiation is *Department of Microbiology and Immunology, University of Miami School of Med- associated with selection of those cells from this pool that express icine, Miami, FL 33101; and †Department of Microbiology and Immunology and ‡Graduate Program in Molecular and Cellular Biology, University of Maryland TCRs of the correct specificity. By this point, each thymic homing School of Medicine, Baltimore, MD 21201 progenitor has undergone ϳ1 million-fold expansion in cell num- Received for publication January 30, 2004. Accepted for publication May 4, 2004. ber (7), representing 20 serial divisions. This proliferation is not The costs of publication of this article were defrayed in part by the payment of page synchronous throughout the differentiation process. The largest charges. This article must therefore be hereby marked advertisement in accordance number of cell divisions occurs at the DN1 stage (7), consistent with 18 U.S.C. Section 1734 solely to indicate this fact. with the relatively long cell cycle times (19, 20) and extended 1 This work was supported by U.S. Public Health Service Grants AI33940 and AI53739 (to H.T.P.) and by funds from Memorial Sloan-Kettering Cancer Center. period of development (8) associated with this stage. DN2 and 2 Address correspondence and reprint requests to Dr. Howard T. Petrie, University of DN3 cells both divide a few times each (7) but cycle at very dif- Miami School of Medicine, P.O. Box 016960 (R-138), Miami, FL 33101. E-mail ferent rates (19, 20), indicating further differences in their genetic address: [email protected] programs. The preDP stage is associated with almost as many cell 3 Abbreviations used in this paper: DN, double-negative; DP, double-positive; preDP, divisions as DN1 cells (7), but in contrast to DN1, cell cycle times early part of the DP phase; smDP, small DP cells; GO, Gene Ontology Consortium; HPRT, hypoxanthine-guanine phosphoribosyltransferase; KLF, Krupple-like factor; appear to be extremely short (19Ð21), given that these cells must PCAF, p300/CBP-associated factor. expand 6- to 8-fold in Ͻ2 days before withdrawal from cycle (7).

The numerous complex changes that accompany lymphopoiesis diately scanned on a Hewlett Packard GeneArray Scanner (Hewlitt Pack- in the postnatal thymus are initiated by signals from stromal cells ard, Palo Alto, CA). that constitute the thymic microenvironment (reviewed in Ref. 22). Identification of genes expressed at each stage of development Ultimately, however, cellular differentiation is defined by changes in the expression patterns of the required genes. Gene expression The methods for analysis of the chip image were those recommended by can be controlled in a variety of ways, but a major regulatory the manufacturer (Affymetrix). Two primary parameters are of interest for each gene on the chip, namely, the absolute call (present, absent, or mar- mechanism is the modulation of transcription, mediated by nuclear ginal) and the signal intensity. Calculation of these values was achieved as transcription factors. Transcription factors bind to specific cis-reg- follows. The difference between probe pair (matched/mismatched) inten- ulatory DNA elements (promoters, enhancers, and silencers) to sities was used to generate a discrimination score. If the discrimination ␶ induce or repress gene expression (23). Transcription factors score was higher than a default predefined threshold ( 0.015), the gene was assigned a call of present; if it was lower, a call of absent or marginal present within a cell act in regulatory networks to establish various was assigned. The one-sided Wilcoxon signed rank test was then used to patterns of gene expression, and changes in these patterns define determine a weighted mean of the discrimination scores for each probe set cellular differentiation (24). Consequently, the identification of (i.e., for each gene). These results were then used to generate confidence transcriptional regulatory networks present in precursor cells vs limits that the combined score was different from the default threshold (␶). When the confidence interval indicated that the weighted mean was sig- their progeny may be used to reveal the mechanisms that drive ␶ Ͻ nificantly different from ( pnull 0.04), the gene product was designated differentiation. Recent studies have examined gene expression dur- as being present. A list of genes present in each stage was then prepared, ing hemato- and lymphopoiesis (25, 26) but have not focused on and for these the signal intensity value was extracted. Signal intensity specific subsets of genes in specific developmental stages. To this represents a quantitative assessment of the relative level of expression for end, we have performed gene expression analyses in progenitor each individual gene. This is determined by taking the log2 of the difference in intensity between each matched/mismatched probe pair (negative values Downloaded from thymocyte stages associated with the lineage commitment and pro- are not used) and then calculating a weighted mean using the one-step liferation processes, using high density microarrays. Here we sum- Tukey biweight estimate. This mean value is then expressed relative to the marize the results with regard to transcription factors that are ex- mean of all genes found to be present on the chip, set to an arbitrary value Ͻ pressed at different progenitor stages, with particular emphasis on (in this case, 500). Thus, genes that are present ( pnull 0.04) and have a those that change with respect to various developmental transi- signal intensity of 1000 would be categorized as being expressed at roughly twice the mean level for all genes in that sample. tions. This analysis reveals novel patterns of expression of numerous transcription factors, including many not previously known to Identification of differentially expressed genes and transcription http://www.jimmunol.org/ be involved in the thymocyte differentiation process. We discuss factors some examples of these with respect to their known roles in other The list of all genes present at one or more stage of differentiation (de- cell types and potential corresponding function in thymocyte pro- scribed above) was merged, and the signal intensities of adjacent stages of genitors, and/or their relationships to upstream or downstream differentiation (e.g., DN1 vs DN2, DN2 vs DN3, etc.) were compared with components known to be required by developing thymocytes. In identify genes that underwent changes of Ͼ2-fold at each developmental addition, the data provided here constitute a rich informational transition. This list was then annotated by submitting a batch query to the Affymetrix NetAffx website (https://www.affymetrix.com/analysis/query/ resource for interrogation and validation of transcriptional patterns interactive_query.jsp) using the probe set identification numbers corre- in developing lymphocytes.

sponding to these differentially expressed genes. This annotated list was by guest on October 2, 2021 then filtered based on Gene Ontology (GO) Consortium designations re- lated to transcription factor activity. Specifically, the following designa- Materials and Methods tions were used: GO biological process numbers 16481 (negative regula- Cell sorting and RNA extraction tion of transcription), 45941 (positive regulation of transcription), 6350 (transcription), 6351 (transcription, DNA dependent), 6355 (regulation of Thymocytes were isolated from 4- to 5-wk-old male C57BL/6 mice (The transcription), 6366 (transcription from polII promoter), 6367 transcription Jackson Laboratory, Bar Harbor, ME). Early lymphopoietic progenitors factor from polI promoter; GO cell component numbers 5634 (nucleus) and were isolated by first depleting small thymocytes (DP cells) using an iso- 5667 (transcription factor complex); GO molecular function num- osmotic gradient of 13.4% Opti-Prep (Greiner Bio-One, Longwood, FL) in bers16563 (transcription activator), 16564 (transcriptional repressor), 3676 HBSS, followed by staining with a lineage mixture of Abs (CD3, CD4, (nucleic acid binding), 3677 (DNA binding), 3700 (transcription factor), CD8, CD11b, Gr1, TER-119) and depletion with paramagnetic beads. Pop- 3712 (transcription cofactor), 3713 (transcription coactivator), and 3714 ulations were identified as follows: DN1, CD24ϩ25Ϫ44ϩ; DN2, ϩ ϩ ϩ ϩ ϩ low ϩ Ϫ low (transcription corepressor. Irrelevant genes emerging from these categories CD24 25 44 ; DN3, CD24 25 44 ; preDP, CD24 25 44 . Addi- (e.g., RNA polymerases, histones, etc.) were manually removed from Ј Ј Ϫ tional sort discriminators were 4 ,6 -diamidino-2-phenylindole (viable), the list. low side scatter, and singlets (forward scatter pulse area vs width). Small DP thymocytes were identified as being forward scatter low as well as Real-time RT-PCR analysis expressing high levels of CD4 and CD8. Only sorted populations that were Ͼ99% pure were used as template for microarray analysis. RNA was ex- cDNA was synthesized from DNase-treated RNA from sorted thymocyte tracted using StrataPrep total RNA kits as recommended by the manufac- subsets (prepared as described above), using random hexamer primers and turer (Stratagene, La Jolla, CA). Overall, purified cells from 16 different Superscript II reverse transcriptase (InVitrogen). Real time, quantitative cell sorting days, each including 8Ð10 thymuses, were pooled to make a PCRs were performed in duplicate. A standard curve consisting of four minimum of 5 ␮g of RNA template for further analysis. 5-fold dilutions of thymus or brain cDNA was included in each experiment for each primer pair. PCRs were performed for 40 cycles on an ABI 7900 Probe synthesis and chip hybridization Sequence Analyzer (Applied Biosystems, Foster City, CA) using a SybrGreen master mix according to the manufacturer’s instructions. After RNA was not amplified before use as a template to generate labeled probes. the final cycle, automatic melting curve analysis was performed to ensure All procedures were conducted by the Molecular Genomics Core Facility that only single PCR products were included in the data analysis. The final at Memorial Sloan-Kettering Cancer Center (New York, NY). Biotin-la- product was also analyzed by agarose gel electrophoresis to verify the size beled cRNA probes were prepared as recommended by the chip manufac- and integrity of the product and to ensure that it corresponded to the single turer (Affymetrix, Santa Clara, CA). Briefly, 5 ␮g of total cellular RNA peak seen with melting curve analysis. Quantitative analysis was per- were used to generate double-stranded cDNAs with oligodeoxythymidine formed using SDS2.0 software (Applied Biosystems). The minimum num-

primers and SuperScript reverse transcription reagents (InVitrogen, Carls- ber of cycles required for detection above a threshold (cycle threshold, Ct) bad, CA). The resulting dsDNA was used to prepare biotinylated cRNA was used to calculate the absolute cDNA amount in each sample by ex-

probes with a BioArray High Yield RNA Transcript kit (Affymetrix). Bi- trapolating Ct values onto the four-point standard curve. Expression of a otin-labeled cRNA was fragmented and hybridized to the MG-U74A array housekeeping gene (hypoxanthine-guanine phosphoribosyltransferase; for16hat45¡C as recommended by the manufacturer. Washing was per- HPRT) was likewise calculated for each sample. The relative level of ex- formed using an automated ﬂuidics workstation, and the array was imme- pression for each transcription factor gene was then calculated by dividing 1096 TRANSCRIPTIONAL CONTROL OF THYMOCYTE DEVELOPMENT Downloaded from

FIGURE 1. Strategy for bulk identiﬁcation of transcription factors present during early thymocyte differentiation.

its cDNA value by the corresponding HPRT cDNA value. Primer se- http://www.jimmunol.org/ quences were as follows. Bcl6: forward, CCGTACCCCTGTGAAATCTG; reverse, AAGTCGCAGTTGGCTTTTGT. Ezh1: forward, AAAACGGAA GCGCCATGCTA; reverse, TGTGCACTGAGGGGGAAGTG. Ezh2: forward, TTCGTGCCCTTGTGTGATAG; reverse, AGCATGGACACTGTT TGGTG. Egr1: forward, CAGGAGTGATGAACGCAAGA; reverse, TGGGGATGGGTAAGAAGAGA. Klf7: forward, CAAGTGCTCATGG GAAGGA; reverse, TGGTCAGACCTGGAGAAACA. Klf3: forward, AGAACCATCCTTCCGTCATC; reverse, GGTGCATTTGTACGGCTTTT. Irf7: forward, CCTCTTGCTTCAGGTTCTGC; reverse, GCTGCATAG GGTTCCTCGTA. Irf5: forward, TTCCAGAAGGGCCAGACTAAT; reverse, TGACATCAGGCCATTCTTCTC. BhlhB2: forward, AGCCGTG by guest on October 2, 2021 GACTTGAAAGAGA; reverse, TGGATGACTGGCACACAGTT. c-Fos: forward, ATCCTTGGAGCCAGTCAAGA; reverse, TCCCAGTCTGCTG CATAGAA. c-Jun: forward, TAACAGTGGGTGCCAACTCA; reverse, CGCAACCAGTCAAGTTCTCA. Rorc: forward, GCCCACCATATTC FIGURE 2. Transcription factors that are differentially expressed during CAATACCT; reverse, TAAGTTGGCCGTCAGTGCTAT; Nab1: forward, the DN1 to DN2 transition. Values on the horizontal axis indicate fold GACCCACACAAAGAGGAGGA; reverse, GGGCATTGTCCTTCA changes in transcription factor expression between stages, as indicated. CACAC. Nab2: forward, GAACCAGAGATGGTGCGAAT; reverse, Only transcription factors with a Ͼ2-fold change are shown. Values for TTCCGGATCTCCTCTTCCTT. l2ra: forward, AACGGCACCATC those factors that changed Ͼ10-fold are indicated next to the corresponding CTAAACTG; reverse, CTGTGTTGGCTTCTGCATGT. Hprt: forward, ϱ ATCAGTCAACGGGGGACATA; reverse, TTGCGCTCATCTTA bar. Fold changes depicted as infinite ( ) indicate that the corresponding GGCTTT. factor was only present in one of the two stages being compared; these genes are listed in Table I. Results Changes in transcription factor expression during inthrathymic lymphoid progenitor development from this list that are associated with transcriptional regulation. Ͼ To reveal new candidates involved in transcriptional control of T This final list of transcription factors that undergo 2-fold changes lymphocyte developmental programs, we performed microarray in expression levels at defined developmental transitions was used analysis using Affymetrix high density mouse arrays (U74A). to generate the data shown in Figs. 2Ð5. In addition, genes that are Ͼ RNA was extracted from purified populations of intrathymic lym- expressed throughout all subsets but do not change 2-fold at any 4 phopoietic progenitors, including DN1, DN2, DN3, and preDP individual transition are presented in Supplemental Table I. Some Ͼ cells, as well as small, noncycling DP cells as ␣␤ lineage com- of these may change 2-fold over multiple phases of the transition mitted, postmitotic controls. The strategy for identification of tran- from DN1 to smDP. In the interest of space, we have not attempted scription factors that might be important during specific develop- to add the numerous additional figures that would be required to mental transitions is illustrated in Fig. 1. First, a list of all genes display such changes. For those interested, such findings can be present (see Materials and Methods) in at least one progenitor extracted from Supplemental Table II, which provides the signal stage was prepared. Next, this list was filtered to provide a list of levels and present/absent calls for all genes described in all figures genes that changed more than 2-fold on transition from one stage and tables of this article. of differentiation to another. This list of all genes that change at Because of the complexities involved in attempting to describe one or more developmental transitions was then annotated by a gene products in terms of their biological processes or molecular batch query of the Affymetrix database. GO designations (www. geneontology.org) were then used to specifically identify genes 4 The on-line version of this article contains supplemental material. The Journal of Immunology 1097

FIGURE 4. Transcription factors that are differentially expressed during the DN3 to preDP transition. Values on the horizontal axis indicate fold FIGURE 3. Transcription factors that are differentially expressed during changes in transcription factor expression between stages, as indicated. Downloaded from the DN2 to DN3 transition. Values on the horizontal axis indicate fold Only transcription factors with a Ͼ2-fold change are shown. Values for changes in transcription factor expression between stages, as indicated. those factors that changed more than 10-fold are indicated next to the Only transcription factors with a Ͼ2-fold change are shown. Values for corresponding bar. Fold changes depicted as infinite (ϱ) indicate that the those factors that changed Ͼ10-fold are indicated next to the corresponding corresponding factor was only present in one of the two stages being com- bar. Fold changes depicted as infinite (ϱ) indicate that the corresponding pared; these genes are listed in Table I. factor was only present in one of the two stages being compared; these genes are listed in Table I. http://www.jimmunol.org/ cells was not amplified before probe synthesis. Purification of suf- ficient RNA template from rare DN1 and DN2 progenitors was functions, not all proteins generally considered to be transcription therefore quite costly and very time consuming, making replicate factors are necessarily revealed when filtered based on GO anno- microarray analysis an unattractive option. Therefore, to validate tations. However, at the time of this writing, every effort has been microarray results, two other types of confirmatory analyses were made to ensure that all relevant genes are included, including by performed. In the first, microarray results for genes with well- manual inspection of gene lists as described in Materials and documented changes in expression patterns during thymocyte dif- Methods. It is also important that neither c-fos nor c-jun were ferentiation were examined. A small panel of these is shown in by guest on October 2, 2021 identified by microarray analysis, either as genes that change dur- Fig. 8. Consistent with what is already known (27), CD3⑀ is not ing the differentiation process (Figs. 2Ð5) or as common genes expressed at statistically significant levels in DN1 cells but in- (Supplemental Table I). This is not because these genes are not creases in expression thereafter. Also consistent with previously expressed in progenitor cells (Fig. 6), but rather they result from published results (28), the invariant preT␣ component of the pre- poor probe design and performance on the U74A chip; i.e., signal TCR (16) is highest on DN3 cells and lowest on DN1 and DN2 was detected in the matched oligos but was not significantly dif- cells. Likewise, the RAG-2 and c-kit expression patterns revealed ferent in the single-base mismatch. Despite such occasional prob- by microarray analysis very closely parallel those previously re- lems, the microarray analysis does provide, for the most part, reported by others (29, 30). CD25, which marks intermediate stages liable and informative results, as described in further detail below. of DN differentiation (30Ð32), is found at significant levels only on A few caveats regarding these data are warranted. First, all pop- DN2 and DN3 cells by microarray analysis, whereas CD4 is up- ulations of progenitor thymocytes contain cells at different stages regulated only at the DP stage. We have also reported other con- of cell cycle (19, 20), including smaller (G1-S phases) and more sistencies between these microarray results and other methods of blastic (G2-M phases) cells, and the relative proportions of such analysis (33). In general, microarray results for genes known to be cells may influence the gene expression outcomes for any individ- differentially regulated during T cell differentiation were remark- ual phase. Likewise, because differentiation in the steady state thy- ably consistent with what would be expected, suggesting that most mus is not synchronous, each subset likely includes a proportion of transcription factors identified by this method (Figs. 2Ð5) are likely cells that have already received signals to differentiate to the next to be valid candidates for further investigation. stage, as well as cells that have not yet received such signals, and the relative proportions of these again may contribute to the gene Validation of microarray results by real-time, quantitative expression patterns attributed to any single subset. Finally, without RT-PCR additional verification (such as by RT-PCR; see Validation of mi- To specifically validate microarray data regarding potential tran- croarray results by analysis of known genes for selected genes in scriptional regulators of thymocyte differentiation, real time RT- this category), it may be risky to generalize the fold changes and/or PCR was performed for 12 selected transcription factors identified absolute expression levels described here. Rather, we believe that by microarray screening, as well as 1 additional control gene our results should be used as guidelines for more critical assess- (CD25). The results of these analyses are shown in Fig. 6. Can- ments of gene expression and/or for validation of existing results. didates for confirmation were selected to span the variety of patterns and expression levels found, i.e., genes that went down dur- Validation of microarray results by analysis of known genes ing differentiation, up during differentiation, or various To most accurately assess gene expression patterns during intra- combinations thereof, as well as genes that were relatively abun- thymic T lymphocyte development, RNA extracted from purified dant vs those that were present but less abundant. In general, the 1098 TRANSCRIPTIONAL CONTROL OF THYMOCYTE DEVELOPMENT

FIGURE 5. Transcription factors that are differentially expressed during the preDP to smDP transition. Values on the horizontal axis indicate fold changes in transcription factor expression between stages, as indicated. Only Downloaded from transcription factors with a Ͼ2-fold change are shown. Values for those factors that changed Ͼ10-fold are indicated next to the corresponding bar. Fold changes depicted as inﬁnite (ϱ) indicate that the corresponding factor was only present in one of the two stages being com- http://www.jimmunol.org/ pared; these genes are listed in Table I. by guest on October 2, 2021

patterns of gene expression revealed by real time PCR were re- thymic microenvironment (reviewed in Ref. 22). The variety of markably consistent with those found by microarray analysis. The signals assimilated by individual precursor cells are subsequently only significant exception to this was Ezh2, which appears to re- manifest as changes in transcriptional activity, many of which are, main high or increase slightly at the DN2, DN3, and preDP stages in turn, a consequence of changes in transcription factor expres- by microarray results, while decreasing over the same period by sion. Thus, the identification of transcription factors that are ex- real time PCR. In other cases, such as Irf7, the patterns of expres- pressed by all early thymocytes (Supplemental Table I), and es- sion were similar, but the overall fold changes were slightly dif- pecially those that fluctuate during the differentiation process ferent between real time and microarray results. Bhlhb2 and Nab2 (Figs. 2Ð5), represents an essential step in characterizing the biol- exhibited similar trends by both microarray and real time PCR but ogy of the lymphopoietic process. The advent of high density mi- differed slightly at a single transition (DN1/DN2 or DN2/DN3, croarray analysis, as presented here, represents a high throughput respectively). Overall, however, both the patterns and relative fold method for accomplishing this goal. In particular, we have focused changes appear to be very similar for both types of analysis. To- on the lymphopoietic stages of intrathymic differentiation, i.e., gether with the comparisons described earlier, these findings sug- those stages that derive from non-self-renewing blood-borne pro- gest that, in general, the microarray results reflect a fairly true genitors and that generate the raw materials for positive and neg- assessment of gene expression and have the potential to reveal ative selection (i.e., smDP cells). novel transcriptional candidates important in control of the intra- To accurately assess the expression of transcription factor genes thymic differentiation process. in early thymocyte progenitors, RNA was isolated directly from cells and used without further amplification. Instead, numerous Discussion samples were pooled from multiple days of purification and cell The differentiation of T lymphocytes in the postnatal thymus is sorting to make adequate RNA (5 ␮g) for the labeling and hybrid- initiated by signals originating from stromal elements that form the ization processes. Analysis of predicted patterns of expression (see The Journal of Immunology 1099

FIGURE 6. Analysis of c-fos and c-jun by real time quantitative PCR. As described in the text, the U74A array gave poor results for two transcription factors known to be expressed in thymocytes, namely c-fos and c-jun. Visual examination of these probe images (not shown) revealed high levels of hybridization to mismatched oligonucleotides, indicating that the absent calls resulting from Wilcoxon signed rank analysis were a result of poor probe design in these cases. To rectify this, PCR analysis was performed, and as shown here, c-fos and c-jun are expressed in early DN progenitors. The y-axis values represent relative fos or jun expression, compared with that of a housekeeping gene (HPRT), nominally deﬁned as equal to 1.

Fig. 8) confirmed the usefulness of this approach, as the vast ma- Downloaded from jority of genes for which expression patterns are known were cor- roborated by this analysis (additional data not shown). Further, real time quantitative PCR analysis for randomly selected transcription factors identified by the present studies also confirmed the validity of the results (Fig. 6). In addition to the validations described above, comparison of FIGURE 8. Validation of a panel of microarray results by quantitative http://www.jimmunol.org/ transcription factors already known to be involved in the T cell real time RT-PCR. A panel of 12 transcription factors were selected to span differentiation process correlated extremely well with gene expres- the variety of expression patterns and expression levels revealed by mi- sion results by microarray. For instance. forced expression of Sfpi1 croarray analysis, and these were further analyzed by quantitative real time f (PU.1) caused a block at the DN2/DN3 transition (34), indicating RT-PCR. Expression patterns found by RT-PCR ( ) are shown on the left that down-regulation of this transcription factor is required for this axis; units are a ratio of cDNA levels calculated from the standard curve (see Materials and Methods) for the gene of interest compared with that of transition to occur normally. Consistent with this, our data show a housekeeping gene (HPRT). Values shown for RT-PCR are the mean of that DN1 and DN2 cells express Sfpi1, but it is absent after DN3 two separate experiments. Expression patterns found by microarray (E) are (Table I). Deficiency in Gfi1 leads to developmental arrest at the shown on the right axis relative to all genes expressed, where mean gene

DN2 stage (35), a stage at which our data indicate that it is sub- expression levels are set at 500 (see Materials and Methods). In general, by guest on October 2, 2021 patterns observed by RT-PCR very closely parallel those found by microarray, conﬁrming the validity of microarray data.

stantially up-regulated (Fig. 2). In Tcf1 mutant mice (36), differentiation is impaired just before the DP stage, a stage at which our data indicate that Tcf1 is up-regulated (Table I). Consistencies were also noted between our data and published ﬁndings for Egrs (37) c-Myb (38), GATA-3 (39), Lmo1 (40), Heb (41), Runx1 (42), and many others. Overall, these consistencies together with the documentation provided by known genes (Fig. 8) and quantitative PCR (Fig. 6) indicate that the additional novel transcription factors revealed by this study are very likely to be authentic regulators of thymocyte differentiation. Although the sheer volume of factors revealed by these studies precludes a discussion of all of them, a few examples, together with their potential relevance, are discussed below.

Regulation of proliferation during early thymic lymphopoiesis Bone marrow thymocyte progenitors expand 1 million-fold on en- try into the thymus but ultimately withdraw from the cell cycle before maturation and export to the periphery (7). Expansion at later intrathymic stages (i.e., at the DN/DP transition) is associated FIGURE 7. Analysis of expression of genes known to be differentially with expression of components of the TCR (43). However, the regulated during thymocyte progenitor differentiation. As discussed in the regulation of proliferation at earlier stages is still largely unchar- text, microarray results (shown here) correlate extremely well with known gene expression patterns in developing thymocytes. For instance, CD3⑀ is acterized. Of particular interest is the down-regulation of prolifer- very low in DN1 cells and increases thereafter, as does CD4 and RAG-2, ative status that occurs at the transition from DN2 to DN3 (19, 20). whereas CD25 marks mainly the DN2 and DN3 progenitor stages, and c-kit Evaluation of our microarray data revealed a profound decrease in the DN1 and DN2 stages. Values on the y-axis are global expression units Lyl1 associated with this transition (Table I). Breakpoint translo- (mean ϭ 500; see text for an explanation). cation fusing Lyl1 to the TCR␤ locus is a cytogenetic abnormality 1100 TRANSCRIPTIONAL CONTROL OF THYMOCYTE DEVELOPMENT

Table I. Transcription factor genes that turn on or off at speciﬁc developmental transitionsa

DN1/DN2 Transition DN2/DN3 Transition DN3/preDP Transition PreDP/smDP Transition

Turned OFF Turned ON Turned OFF Turned ON Turned OFF Turned ON Turned OFF Turned ON

1190001I08Rik 4833408P15Rik 2610110L04Rik Creb1 1110037D23Rik 2610110L04Rik 2610110L04Rik 1110037D23Rik 4921515A04Rik Atm 9130025P16Rik Egr1 4833408P15Rik C630022N07Rik Bard1 4930566A11Rik 4930565N07Rik Bard1 9130211I03Rik Esr1 Atm D930018N21Rik Bn51t-pending Arnt 4930565N07Rik Bn51t-pending Bhlhb2 Hes5 Creb1 Deaf1 Gcn512 Atm 4930566A11Rik Dedd C630022N07Rik Idb3 Ddit3 Dlx6 Hand1 C2ta Ankrd2 Dlx6 Creg Kcnh2 Esr1 DP1 Hes1 Clock Atbf1 Ets2 Ctbp2 Lef1 Gata1 Dri1 Hoxb8 Creb1 Atf3 Fhl2 Deaf1 Nab2 Hes5 Egr2 Hoxd3 D11Ertd714e Batf Hivep2 Dedd Nr1d1 Hhex Ets1 Htr5a Ddit3 Bcl3 Hoxa7 Dlx6 Nr3c1 Madh7 Hand1 Irx3 Dedd C2ta Ikbkg Elk3 Nr4a1 Nfe2 Hoxb7 Klf9 Dusp16 Cebpa Mad3 Gfi1b Pax6 Nmyc1 Hoxb8 Mcmd6 E2f3 Cebpd Mllt10 Hoxa6 Pbx3 Nr1d1 Hoxd3 Nfix E2f5 Clock Mybl2 Hoxa7 Pou6f1 Nr1h2 Irx3 Nr1h4 Elf4 Dri1 Myst2 Hoxb8 Rbpsuhl Nr2c1 Mcmd6 Pou2af1 Elk3 Egr1 Nfatc2 Hoxd3 Sox15 Nr3c1 Mllt10 Pou2f3 Gata6 Eif2ak3 Nfkbie Irf5 Spib Pax6 Nr1h4 Ptrf Hhex Fos Nr1h2 Lmo1 Tceb2 Pbx3 Pou2af1 Rbpsuhl Hoxb8 Downloaded from Gata6 Pknox1 Lyl1 Trp63 Per1 Pou2f3 Sox3 Irf6 Hipk2 Pla2g6 Mafg Trpc4 Sox15 Prdm1 Tmpo Junb Hlx Rfx2 Mllt10 Zfp105 Spib Ptrf Zfp239 Mef2b Hmga2 Rfxank Nfatc2 Zfp46 Tbx6 Rorc Mllt2h Hoxa2 Sap18 Nfil3 Zfp60 Tcf1 Sox3 Mrg2 Hoxa4 Sox3 Nfkbie Zfy2 Trpc4 Zfp26 Mtf1

Hoxa9 Sox7 Pou2af1 Zfp29 Zfp422 Nfatc3 http://www.jimmunol.org/ Idb3 Tal1 Rab25 Zfp385 Znfn1a1 Nﬁl3 Irf4 Tcf20 Rfx2 Zfp46 Npas1 Irf7 Tle4 Sfpi1 Zfp60 Nr1d1 Irf7 Zfp26 Sox3 Zfp61 Nr21 Junb Zfp260 Sox7 Zfp99 Nr3c1 Klf4 Zfp358 Tal1 Zfy2 Per1 Klf5 Zfp97 Tle4 Phf1 Lef1 Zfp26 Pparbp Lisch7-pending Zfp288 Rara Lmo2 Zfp358 Rfx2 Lmyc1 Zfp422 Rnf4 by guest on October 2, 2021 Nab2 Rora Nr4a1 Sp4 Nr6a1 Tbx6 Pax5 Tcf1 Pax6 Tle1 Pbx1 Zfp276 Pbx3 Zfp288 Prdm1 Zfp29 Rbpsuhl Zfp358 Relb Zfp36 Spib Zfp385 Tceb2 Zfp46 Tnfrsf5 Zfp61 Zfp105 Zfp99 Zfp36 Zfp46

a These genes are designated as inﬁnite (Ϫ) fold changes in Figs. 2Ð5. A categorical summary of this table (i.e., by transcription factor family) can be found in Supplemental Table III.

associated with ϳ20% of human T cell acute lymphocytic leuke- Transcriptional control of Notch signaling mias (44, 45), suggesting a role for Lyl1 in T cell in proliferation. The mammalian homologs of the Drosophila hairy-enhancer of TCR␤ gene rearrangements also initiate at the DN2/DN3 transition (11, 30), further strengthening the association and suggesting that split (Hes) genes encode at least seven basic helix-loop-helix tran- mistakes in the somatic (VDJ) recombination process in develop- scription factors (Hes1Ð7). Hes gene products are major transcrip- ing T cells may contribute to aberrant proliferation and oncogenic tional mediators of Notch signaling (reviewed in Ref. 47). Notch transformation. Lyl1 is highly expressed early in T cell develop- plays an important role in T lymphopoiesis (reviewed in Ref. 48), ment (DN1 and DN2) but is completely repressed in DN3 cells and regulation of Hes genes can therefore impact T cell differen- (Table I), consistent with RT-PCR results of others (46). Thus, our tiation by regulating Notch signals. For instance, inactivation of data indicate the possibility that down-regulation of Lyl1 may rep- either Notch1 or Hes1 results in similar arrest at early stages of resent a critical step in repressing cell cycle activity during VDJ intrathymic T cell differentiation (49, 50). In the present data, Hes1 recombination in developing lymphocytes. was expressed in the DN1Ð3 stages, significantly down-regulated The Journal of Immunology 1101 upon the transition to preDP, and lost in smDP thymocytes. The ex- References pression of Hes1 thus strongly parallels published data regarding pat- 1. Haynes, B. F., and L. P. Hale. 1999. Thymic function, aging, and AIDS. Hosp. terns of Notch1 expression (51) and signaling during the DN to DP Pract. (Off. Ed.) 34:59. Ͻ 2. Ceredig, R., and T. Rolink. 2002. A positive look at double-negative thymocytes. transition (52). There is also a modest, 2-fold, up-regulation of Hes1 Nat. Rev. Immunol. 2:888. during the DN1 to DN2 transition (not shown), likewise in accordance 3. Manz, M. G., D. Traver, T. Miyamoto, I. L. Weissman, and K. Akashi. 2001. with published data (53), again corroborating the validity of microar- Dendritic cell potentials of early lymphoid and myeloid progenitors. Blood 97:3333. ray results presented here. However, our data also provide new in- 4. Michie, A. M., J. R. Carlyle, T. M. Schmitt, B. Ljutic, S. K. Cho, Q. Fong, and sights regarding the regulation of Notch signaling in T cell develop- J. C. Zuniga-Pflucker. 2000. Clonal characterization of a bipotent T cell and NK ment. Hes1 acts as a transcriptional corepressor upon interaction with cell progenitor in the mouse fetal thymus. J. Immunol. 164:1730. 5. Ikawa, T., H. Kawamoto, S. Fujimoto, and Y. Katsura. 1999. Commitment of members of the Groucho family of proteins (reviewed in Ref. 54). common T/natural killer (NK) progenitors to unipotent T and NK progenitors in One member of this family, Groucho-5 (also known as amino-termi- the murine fetal thymus revealed by a single progenitor assay. J. Exp. Med. 190:1617. nal enhancer of split, Aes), is specifically up-regulated on transition to 6. Ardavin, C., L. Wu, C. L. Li, and K. Shortman. 1993. Thymic dendritic cells and the DN3 stage and down-regulated again in preDP cells, mirroring the T cells develop simultaneously in the thymus from a common precursor popu- expression peaks for Hes1 (Table I). Of interest, Groucho-5 is again lation. Nature 362:761. 7. Shortman, K., M. Egerton, G. J. Spangrude, and R. Scollay. 1990. The generation up-regulated at the smDP stage (Table I). Thus, down-regulation of and fate of thymocytes. Semin. Immunol. 2:3. Groucho-5 appears to correlate specifically with the DN to DP tran- 8. Porritt, H. E., K. Gordon, and H. T. Petrie. 2003. Kinetics of steady-state differ- sition, where Notch activity is known to be required (55). The up- entiation and mapping of intrathymic-signaling environments by stem cell trans- plantation in nonirradiated mice. J. Exp. Med. 198:957. regulation of Groucho-5 at the smDP stage, where Hes1 is down- 9. Lind, E. F., S. E. Prockop, H. E. Porritt, and H. T. Petrie. 2001. Mapping pre- regulated (Table I) may seem counterintuitive. However, Groucho cursor movement through the postnatal thymus reveals specific microenviron- ments supporting defined stages of early lymphoid development. J. Exp. Med. Downloaded from proteins also interact with a large variety of transcription factors other 194:127. than Hes (54). Further, we find Hes6 (Supplemental Table I) and 10. Livak, F., M. Tourigny, D. G. Schatz, and H. T. Petrie. 1999. Characterization of Notch3 (data not shown) are both expressed at the smDP stage, sug- TCR gene rearrangements during adult murine T cell development. J. Immunol. 162:2575. gesting that Groucho-5 may alternatively interact with these. It is also 11. Petrie, H. T., F. Livak, D. Burtrum, and S. Mazel. 1995. T cell receptor gene possible that interaction of Groucho-5 with Hes6 (56) could interfere recombination patterns and mechanisms: cell death, rescue, and T cell produc- with the activity of Hes1, thus promoting terminal differentiation in tion. J. Exp. Med. 182:121. 12. Dudley, E. C., H. T. Petrie, L. M. Shah, M. J. Owen, and A. C. Hayday. 1994. http://www.jimmunol.org/ DP cells similar to its role in the nervous system (57). T cell receptor ␤ chain gene rearrangement and selection during thymocyte development in adult mice. Immunity 1:83. 13. Burtrum, D. B., S. Kim, E. C. Dudley, A. C. Hayday, and H. T. Petrie. 1996. TCR ␣␤ ␥␦ ␤ Cell cycle arrest and cell death or survival at the DP stage gene recombination and - lineage divergence: productive TCR- rearrangement is neither exclusive nor preclusive of ␥␦ cell development. J. Immunol. Another cluster of transcriptional regulatory complexes that ap- 157:4293. 14. Petrie, H. T., M. Pearse, R. Scollay, and K. Shortman. 1990. Development of pears to be up-regulated at the transition to smDP (Fig. 5) includes immature thymocytes: initiation of CD3, CD4, and CD8 acquisition parallels the Krupple-like factors (KLFs), the CBP/p300 transcriptional co- down-regulation of the interleukin 2 receptor ␣ chain. Eur. J. Immunol. 20:2813. activator Cited2, and p300/CBP-associated factor (PCAF). A num- 15. Petrie, H. T., R. Scollay, and K. Shortman. 1992. Commitment to the T cell receptor-␣␤ or -␥␦ lineages can occur just prior to the onset of CD4 and CD8 ber of KLFs are up-regulated in smDP cells, including Klf3, 7, and expression among immature thymocytes. Eur. J. Immunol. 22:2185. by guest on October 2, 2021 13 (Fig. 5). KLFs are heavily implicated in cell differentiation, cell 16. Groettrup, M., K. Ungewiss, O. Azogui, R. Palacios, M. J. Owen, A. C. Hayday, and H. von Boehmer. 1993. A novel disulfide-linked heterodimer on pre-T cells cycle regulation, and cell survival/death decisions (reviewed in consists of the T cell receptor ␤ chain and a 33 kd glycoprotein. Cell 75:283. Ref. 58), all of which are characteristic of smDP cells (7). KLFs 17. Egerton, M., K. Shortman, and R. Scollay. 1990. The kinetics of immature mu- have already been shown to have various roles in cells of the T rine thymocyte development in vivo. Int. Immunol. 2:501. 18. Penit, C., F. Vasseur, and M. Papiernik. 1988. In vivo dynamics of CD4Ϫ8Ϫ lineage, including differentiation, cell cycle withdrawal, and sur- thymocytes: proliferation, renewal and differentiation of different cell subsets vival (for reviews, see Refs. 59 and 60). One particularly interest- studied by DNA biosynthetic labeling and surface antigen detection. Eur. J. Im- ing factor in this family is KLF13. KLF13 has a role in the dif- munol. 18:1343. 19. Tourigny, M. R., S. Mazel, D. B. Burtrum, and H. T. Petrie. 1997. T cell receptor ferentiation of effector T lymphocytes (61, 62). KLF13 (TCR)-␤ gene recombination: dissociation from cell cycle regulation and devel- transcriptional activity is regulated by the transcriptional coacti- opmental progression during T cell ontogeny. J. Exp. Med. 185:1549. 20. Penit, C., B. Lucas, and F. Vasseur. 1995. Cell expansion and growth arrest vators CBP/p300 and PCAF (63), and PCAF and Cited2 are up- phases during the transition from precursor (CD4Ϫ8Ϫ) to immature (CD4ϩ8ϩ) regulated simultaneously with KLF13 at the smDP stage (Fig. 5). thymocytes in normal and genetically modified mice. J. Immunol. 154:5103. Another factor, Klf7, has been implicated in exit from cell cycle by 21. Hoffman, E. S., L. Passoni, T. Crompton, T. M. Leu, D. G. Schatz, A. Koff, M. J. Owen, and A. C. Hayday. 1996. Productive T-cell receptor ␤-chain gene inducing expression of cell cycle inhibitors (64), many of which rearrangement: coincident regulation of cell cycle and clonality during develop- we find to be up-regulated in smDP cells (not shown). Thus, our ment in vivo. Genes Dev. 10:948. data indicate that the activity of KLFs, and in particular KLF7 22. Petrie, H. T. 2002. Role of thymic organ structure and stromal composition in steady-state postnatal T-cell production. Immunol. Rev. 189:8. and/or 13 and their regulatory partners, may represent good can- 23. Orphanides, G., and D. Reinberg. 2002. A unified theory of gene expression. Cell didates for further characterization of the poorly understood net- 108:439. 24. Davidson, E. H., J. P. Rast, P. Oliveri, A. Ransick, C. Calestani, C. H. Yuh, works that control cell cycle arrest, homeostasis, and cell survival/ T. Minokawa, G. Amore, V. Hinman, C. Arenas-Mena, et al. 2002. A genomic death of smDP thymocytes. regulatory network for development. Science 295:1669. Overall, the data presented here provide numerous novel obser- 25. Park, I. K., Y. He, F. Lin, O. D. Laerum, Q. Tian, R. Bumgarner, C. A. Klug, K. Li, C. Kuhr, M. J. Doyle, et al. 2002. Differential gene expression profiling of vations regarding the transcriptional control of thymic T lympho- adult murine hematopoietic stem cells. Blood 99:488. cyte differentiation. Our confidence in the data is such that a num- 26. Hoffmann, R., L. Bruno, T. Seidl, A. Rolink, and F. Melchers. 2003. Rules for ber of these are already being followed up in our laboratories, gene usage inferred from a comparison of large-scale gene expression profiles of T and B lymphocyte development. J. Immunol. 170:1339. using transgenic and knockout approaches. However, the sheer 27. Wilson, A., and H. R. MacDonald. 1995. Expression of genes encoding the pre- volume of factors found is daunting, and thus we believe that these TCR and CD3 complex during thymus development. Int. Immunol. 7:1659. 28. Wilson, A., J. P. de Villartay, and H. R. MacDonald. 1996. T cell receptor ␦ gene data provide a rich resource for gene discovery by others. In ad- rearrangement and T early ␣ (TEA) expression in immature ␣␤ lineage thymo- dition, these findings may be used to confirm gene expression anal- cytes: implications for ␣␤/␥␦ lineage commitment. Immunity 4:37. ysis derived from other methods, such as RT-PCR and to break 29. Wilson, A., W. Held, and H. R. MacDonald. 1994. Two waves of recombinase gene expression in developing thymocytes. J. Exp. Med. 179:1355. those findings down into discrete subset results when whole thy- 30. Godfrey, D. I., J. Kennedy, T. Suda, and A. Zlotnik. 1993. A developmental mocytes have been used as template. pathway involving four phenotypically and functionally distinct subsets of 1102 TRANSCRIPTIONAL CONTROL OF THYMOCYTE DEVELOPMENT

CD3ϪCD4ϪCD8Ϫ triple-negative adult mouse thymocytes defined by CD44 and 46. Ferrando, A. A., S. Herblot, T. Palomero, M. Hansen, T. Hoang, E. A. Fox, and CD25 expression. J. Immunol. 150:4244. A. T. Look. 2004. Biallelic transcriptional activation of oncogenic transcription 31. Pearse, M., L. Wu, M. Egerton, A. Wilson, K. Shortman, and R. Scollay. 1989. factors in T-cell acute lymphoblastic leukemia. Blood 103:1909. A murine early thymocyte developmental sequence is marked by transient ex- 47. Artavanis-Tsakonas, S., M. D. Rand, and R. J. Lake. 1999. Notch signaling: cell pression of the interleukin 2 receptor. Proc. Natl. Acad. Sci. USA 86:1614. fate control and signal integration in development. Science 284:770. 32. Shimonkevitz, R. P., L. A. Husmann, M. J. Bevan, and I. N. Crispe. 1987. Tran- 48. Guidos, C. J. 2002. Notch signaling in lymphocyte development. Semin. Immu- sient expression of IL-2 receptor precedes the differentiation of immature thy- nol. 14:395. mocytes. Nature 329:157. 49. Radtke, F., A. Wilson, G. Stark, M. Bauer, M. J. van, H. R. MacDonald, and 33. Plotkin, J., S. E. Prockop, A. Lepique, and H. T. Petrie. 2003. Critical role for M. Aguet. 1999. Deficient T cell fate specification in mice with an induced CXCR4 signaling in progenitor localization and T cell differentiation in the post- inactivation of Notch1. Immunity 10:547. natal thymus. J. Immunol. 171:4521. 50. Tomita, K., M. Hattori, E. Nakamura, S. Nakanishi, N. Minato, and 34. Anderson, M. K., A. H. Weiss, G. Hernandez-Hoyos, C. J. Dionne, and R. Kageyama. 1999. The bHLH gene Hes1 is essential for expansion of early T E. V. Rothenberg. 2002. Constitutive expression of PU. 1 in fetal hematopoietic cell precursors. Genes Dev. 13:1203. progenitors blocks T cell development at the pro-T cell stage. Immunity 16:285. 51. Hasserjian, R. P., J. C. Aster, F. Davi, D. S. Weinberg, and J. Sklar. 1996. 35. Yucel, R., H. Karsunky, L. Klein-Hitpass, and T. Moroy. 2003. The transcrip- Modulated expression of notch1 during thymocyte development. Blood 88:970. tional repressor Gfi1 affects development of early, uncommitted c-Kitϩ T cell 52. Deftos, M. L., E. Huang, E. W. Ojala, K. A. Forbush, and M. J. Bevan. 2000. progenitors and CD4/CD8 lineage decision in the thymus. J. Exp. Med. 197:831. Notch1 signaling promotes the maturation of CD4 and CD8 SP thymocytes. 36. Verbeek, S., D. Izon, F. Hofhuis, E. Robanus-Maandag, H. te Riele, Immunity 13:73. M. van de Wetering, M. Oosterwegel, A. Wilson, H. R. MacDonald, and 53. Reizis, B., and P. Leder. 2002. Direct induction of T lymphocyte-specific gene H. Clevers. 1995. An HMG-box-containing T-cell factor required for thymocyte expression by the mammalian Notch signaling pathway. Genes Dev. 16:295. differentiation. Nature 374:70. 54. Fisher, A. L., and M. Caudy. 1998. Groucho proteins: transcriptional corepressors for specific subsets of DNA-binding transcription factors in vertebrates and in- 37. Carleton, M., M. C. Haks, S. A. Smeele, A. Jones, S. M. Belkowski, vertebrates. Genes Dev. 12:1931. M. A. Berger, P. Linsley, A. M. Kruisbeek, and D. L. Wiest. 2002. Early growth Ϫ Ϫ 55. Huang, E. Y., A. M. Gallegos, S. M. Richards, S. M. Lehar, and M. J. Bevan. response transcription factors are required for development of CD4 CD8 thy- ϩ ϩ 2003. Surface expression of Notch1 on thymocytes: correlation with the double- mocytes to the CD4 CD8 stage. J. Immunol. 168:1649. negative to double-positive transition. J. Immunol. 171:2296. 38. Allen, R. D., 3rd, T. P. Bender, and G. Siu. 1999. c-Myb is essential for early T Downloaded from 56. Gao, X., T. Chandra, M. O. Gratton, I. Quelo, J. Prud’homme, S. Stifani, and cell development. Genes Dev. 13:1073. R. St.-Arnaud. 2001. HES6 acts as a transcriptional repressor in myoblasts and 39. Hendriks, R. W., M. C. Nawijn, J. D. Engel, H. van Doominck, F. Grosveld, and can induce the myogenic differentiation program. J. Cell Biol. 154:1161. A. Karis. 1999. Expression of the transcription factor GATA-3 is required for the 57. Bae, S., Y. Bessho, M. Hojo, and R. Kageyama. 2000. The bHLH gene Hes6, an development of the earliest T cell progenitors and correlates with stages of cel- inhibitor of Hes1, promotes neuronal differentiation. Development 127:2933. lular proliferation in the thymus. Eur. J. Immunol. 29:1912. 58. Dang, D. T., J. Pevsner, and V. W. Yang. 2000. The biology of the mammalian 40. Chervinsky, D. S., X. F. Zhao, D. H. Lam, M. Ellsworth, K. W. Gross, and Kruppel-like family of transcription factors. Int. J. Biochem. Cell Biol. 32:1103. P. D. Aplan. 1999. Disordered T-cell development and T-cell malignancies in 59. Yusuf, I., and D. A. Fruman. 2003. Regulation of quiescence in lymphocytes. SCL LMO1 double-transgenic mice: parallels with E2A-deficient mice. Mol. Trends Immunol. 24:380. http://www.jimmunol.org/ Cell. Biol. 19:5025. 60. Song, A., T. Nikolcheva, and A. M. Krensky. 2000. Transcriptional regulation of 41. Barndt, R., M. F. Dai, and Y. Zhuang. 1999. A novel role for HEB downstream RANTES expression in T lymphocytes. Immunol. Rev. 177:236. ␣␤ or parallel to the pre-TCR signaling pathway during thymopoiesis. J. Immu- 61. Song, A., A. Patel, K. Thamatrakoln, C. Liu, D. Feng, C. Clayberger, and nol. 163:3331. A. M. Krensky. 2002. Functional domains and DNA-binding sequences of 42. Woolf, E., C. Xiao, O. Fainaru, J. Lotem, D. Rosen, V. Negreanu, Y. Bernstein, RFLAT-1/KLF13, a Kruppel-like transcription factor of activated T lymphocytes. D. Goldenberg, O. Brenner, G. Berke, D. Levanon, and Y. Groner. 2003. Runx3 J. Biol. Chem. 277:30055. and Runx1 are required for CD8 T cell development during thymopoiesis. Proc. 62. Nikolcheva, T., S. Pyronnet, S. Y. Chou, N. Sonenberg, A. Song, C. Clayberger, Natl. Acad. Sci. USA 100:7731. and A. M. Krensky. 2002. A translational rheostat for RFLAT-1 regulates RAN- 43. Fehling, H. J., and H. von Boehmer. 1997. Early ␣␤ T cell development in the TES expression in T lymphocytes. J. Clin. Invest. 110:119. thymus of normal and genetically altered mice. Curr. Opin. Immunol. 9:263. 63. Song, C. Z., K. Keller, Y. Chen, and G. Stamatoyannopoulos. 2003. Functional 44. Saltman, D. L., J. D. Mellentin, S. D. Smith, and M. L. Cleary. 1990. Mapping interplay between CBP and PCAF in acetylation and regulation of transcription

of translocation breakpoints on the short arm of chromosome 19 in acute leuke- factor KLF13 activity. J. Mol. Biol. 329:207. by guest on October 2, 2021 mias by in situ hybridization. Genes Chromosomes Cancer 2:259. 64. Laub, F., R. Aldabe, V. J. Friedrich, S. Ohnishi, T. Yoshida, and F. Ramirez. 45. Ferrando, A. A., and A. T. Look. 2003. Gene expression proﬁling in T-cell acute 2001. Developmental expression of mouse Kruppel-like transcription factor lymphoblastic leukemia. Semin. Hematol. 40:274. KLF7 suggests a potential role in neurogenesis. Dev. Biol. 233:305.