Identification of Novel Genes and Transcription Factors Involved in Spleen, Thymus and Immunological Development and Function

ORIGINAL ARTICLE Identification of novel genes and transcription factors involved in spleen, thymus and immunological development and function

BG Hoffman1, KL Williams2, AH Tien1,VLu1, T Ruiz de Algara1, JP-Y Ting2 and CD Helgason1,3 1BC Cancer Agency, Vancouver, BC, Canada; 2University of North Carolina Chapel Hill, Lineberger Comprehensive Cancer Center, Chapel Hill, NC, USA and 3Department of Surgery, Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada

We constructed and analyzed six serial analysis of gene expression (SAGE) libraries to identify genes with previously uncharacterized roles in spleen or thymus development. A total of 625 070 tags were sequenced from the three spleen (embryonic day (E)15.5, E16.5 and adult) and three thymus (E15.5, E18.5 and adult) libraries. These tags corresponded to 83 182 tag types, which mapped unambiguously to 36 133 different genes. Genes over-represented in these libraries, compared to 115 mouse SAGE libraries (www.mouseatlas.org), included genes of known and unknown immunological or developmental relevance. The expression profiles of 11 genes with unknown roles in spleen and thymus development were validated using reverse transcription-qPCR. We further characterized the expression of one of these candidates, RIKEN cDNA 9230105E10 that encodes a murine homolog of Trim5a, in numerous adult tissues and immune cell types. In addition, we demonstrate that transcript levels are upregulated in response to TLR stimulation of plasmacytoid dendritic cells and macrophages. This work provides the first evidence of regulated and cell type-specific expression of this gene. In addition, these observations suggest that the SAGE libraries provide an important resource for further investigations into the molecular mechanisms regulating spleen and thymus organogenesis, as well as the development of immunological competence. Genes and Immunity (2006) 7, 101–112. doi:10.1038/sj.gene.6364270; published online 15 December 2005

Keywords: SAGE; Trim5a; transcription factor

Introduction developing spleen can no longer be seen at this time in Hox11À/À, capsulinÀ/À and Bapx1À/À mice,1–3,5 confirm- Comparatively little is known about the molecular ing the critical role of these transcription factors in spleen mechanisms regulating the development of the spleen organogenesis. The first hematopoietic cells begin to and thymus. Considerably more attention has been appear in the spleen anlage, and between E13.5 and focused on elucidating the mechanisms regulating approximately E16.5, the spleen is a site of hematopoiesis lymphoid cell development and function, although it is in the fetus. Shortly after E16.5 hematopoiesis migrates to likely that many genes critical to these processes remain the bone marrow and lymphoid precursors begin to to be discovered. infiltrate the spleen. Thereafter, the final organization of Development of the mouse spleen begins just before the spleen is established with the formation of the white embryonic day 11.5 (E11.5) with condensation of the pulp, containing B-cell follicles, T cells as well as mesenchyme underlying the dorsal mesogastrium lymphocytes and antigen-presenting dendritic cells epithelium.1 The genetic mechanisms underlying the (DCs) in the marginal zone, and the red pulp, which is generation of these initial splenic precursors are essen- composed of a complex system of blood vessels that tially unknown. By E11.5, the expression of several allow the removal of old or damaged red blood cells.6,7 transcription factors essential to spleen organogenesis Thymus development begins as early as E10.5 with the including Hox11, Nkx2.5, capsulin/Pod-1, Wt1 and Bapx1 proliferation of cells in the posterior part of the third can be detected.1–3 The molecular hierarchy regulating pharyngeal pouch.8,9 Subsequently, thymic precursor expression of these transcription factors, and the path- cells migrate and fuse at the midline. Interactions ways they function within, are only beginning to be between the thymic epithelium and the surrounding established.1,4 By E13.5, the splenic primordium is visible neural crest-derived mesenchyme play a key role in as a ridge of cells on the dorsal stomach. A normally expansion of thymic precursors.9,10 Several studies have highlighted the importance of transcription factors such as Foxn1, Hoxa3, Pax1 and Pax9 in early thymus Correspondence: Dr CD Helgason, British Columbia Cancer development.8,9,11–17 These studies point to a hierarchy Agency, 675 West 10th Avenue, Vancouver, B.C., V5Z 1L3, Canada. E-mail: [email protected] of transcription factor expression, with Foxn1 and Hoxa3 Received 19 September 2005; revised 12 October 2005; accepted 13 playing a role in early thymic development and Pax9 and October 2005; published online 15 December 2005 Pax1 becoming involved subsequently.8,9 At roughly the SAGE analysis of spleen and thymus development BG Hoffman et al 102 same time that Foxn1 expression becomes critical to candidate uncharacterized genes, a mouse homolog of thymic development (E11.5), hematopoietic precursors the primate Trim5a HIV-1 restriction factor, was further begin to invade the thymic primordium. Interestingly, assessed in numerous adult tissues as well as various Foxn1À/À mice lack expression of the chemokines TECK immune cell types including macrophages, DCs and and SDF-1, suggesting that these molecules are targets of various T-cell populations. Our results demonstrate Foxn1 and that they are critical in attracting prothymo- regulated expression of this transcript in macrophages cytes to the thymus.8 At E11.5, the thymic epithelium is and Flt-3L-derived DC populations in response to TLR not fully differentiated and is unable to support T-cell stimulation. The libraries created and described herein development. By E14.5, concomitant with maturation of provide a unique and valuable resource for the identi- the thymic epithelium, T-cell precursors become fully fication of developmentally and immunologically rele- committed to the T-cell lineage at least in part via the vant transcripts in the spleen and thymus. Moreover, we Notch signaling pathway.18–23 Thymocyte interactions provide the first evidence that expression of an im- with the thymic epithelium and mesenchyme are critical munologically relevant transcript identified in these to the proper development of both the thymus and the libraries is differentially regulated. thymocytes. Lineage-committed T cells drive the estab- lishment of the appropriate architecture of cortical thymic epithelial cells. Subsequently, the final maturation of T cells induces the expansion and organization of Results medullary thymic epithelial cells. Overview of the libraries Serial analysis of gene expression (SAGE) analysis, like A total of 625 070 tags were sequenced to generate the six microarrays, provides a quantitative analysis of gene LongSAGE spleen and thymus libraries (available at expression profiles. However, it has the additional www.mouseatlas.org or http://cgap.nci.nih.gov/SAGE). advantage that it permits the identification of novel These include three spleen libraries from: E15.5 when the transcripts, which has been improved by the develop- spleen is predominantly erythropoietic, E16.5 when ment of a LongSAGE protocol, which generates 21 mers lymphocytes begin to appear in this tissue and adult that enhances the efficiency and confidence of tag-to- (84 days post-natal (dpn)) when the spleen is fully gene mapping. SAGE also has the added benefit that the mature and functional. Similarly, we generated three data are digital and thus easily shared among investiga- thymus libraries from: E15.5 when a mature thymic tors and compared across different experiments and epithelium is present and T-cell precursors become fully tissues.24,25 committed, E18.5 when mature T cells are first estab- The analysis of expression patterns associated with lished and from 84 dpn when the thymus is fully mature different stages of spleen and thymus development is and functional. A summary of the libraries is shown in expected to further facilitate our understanding of the Table 1. genetic cascade required for their development. In Analysis of these libraries revealed expression of addition, such studies are likely to provide insights into 83 182 different tag types after stringent quality selection: the numerous transcription factors, cell surface proteins 47 727 in the spleen libraries (58% unique to the spleen and signaling pathway components involved in the libraries as compared to the thymus libraries) and 55 282 maturation and function of lymphocytes. We thus in the thymus libraries (64% unique). Of these tags, 45% created six SAGE libraries representing critical stages of (37 202) mapped to known transcripts using the Refseq, spleen (E15.5 and E16.5) and thymus (E15.5 and E18.5) MGC and Ensembl transcript databases, 23% (19 214) to development, as well as libraries from the corresponding known genes using the Ensembl genes database, 12% adult tissues. Here, we describe the analyses of these (9626) to the genome using the Golden path database and libraries focusing on the identification of transcription 0.1% (100) to the mitochondrial genome or to noncoding factors, signaling pathway members and uncharacter- genes using the Genebank database. The remaining 20% ized genes. We next validated the expression of seven (17 040) of the tags could not be mapped using any of transcription factors and four uncharacterized genes these databases. Eighty-nine percent (15 166) of these with unknown roles in spleen/thymus/lymphocyte unmapped tags were found at a single count, implying development or function. The expression of one of the that many may have been generated by sequencing, PCR,

Table 1 Summary of the LongSAGE spleen and thymus libraries

Library name Stage Biological significance Tags sequenced Total tagsa Total tag types % Replicate ditags

Spleen libraries SM055 E15.5 Spleen becomes predominantly erythropoietic 128 520 95 829 28 571 11.72 SM011 E16.5 Lymphocytes appear 60 592 41 994 15 074 12.30 SM012 Adult Fully functional adult organ 86 372 62 017 20 561 13.70

Thymus libraries SM057 E15.5 Mature thymic epithelium present, 110 638 83 792 25 386 10.49 T-cell precursors become fully committed SM022 E18.5 Presence of mature T cells 108 120 64 734 23 011 6.21 SM024 Adult Fully functional adult organ 130 828 105 418 32 166 5.42

aRefers to tag number after removal of replicate ditags and after filtering for tag quality (using a 95% cutoff).

Genes and Immunity SAGE analysis of spleen and thymus development BG Hoffman et al 103 or other errors. Regardless, some of these tags likely Forty-four transcriptional regulators were identified as represent valid, novel, transcripts expressed in these being over-represented in the spleen libraries as com- tissues since 196 of them were found at a significant level pared to our meta-library (a pool of the 115 libraries (total count X7). The 56 416 tags that mapped to gene or discussed above) (Supplementary Table 2). The tran- transcript databases represented 36 133 different genes, scriptional regulators identified with the highest sig- ESTs or other transcripts, suggesting that many (30%) nificance factor (the mean count in the libraries of had alternative transcript termination sites. interest multiplied by the mean ratio determined in We searched the libraries for members of the Wnt, those libraries versus all libraries) included Tcf21 MAPK, TGF-b, FGF, BMP, Notch, Shh and Jak-STAT (Capsulin),1 Pax-5 that plays a key role in B-cell develop- pathways and found components of each in our libraries, ment,34 FoxP1 that is associated with poor outcome in including several with no known role in spleen, thymus patients with diffuse large B-cell lymphomas35,36 and or immune cell development/function (data not shown). Nkx2.3 that, when disrupted, produces mice with We next used gene ontology (GO), which provides a architectural defects in their spleens, as well as defects controlled vocabulary to describe gene and gene product in B- and T-cell maturation.37 Interestingly, several attributes,26 to classify the genes identified. An asso- transcription factors with uncharacterized roles in ciated GO term was obtained for 15 921 (44%) of the immune cell or spleen development and function were genes. Of these, 8851 were annotated at the third level also very highly ranked using this method, including under molecular function that contains the classification Deaf1 that plays a role in neural tube closure and skeletal terms transcription factor, cofactor, transcriptional re- patterning,38 the metal responsive transcription factor pressor or transcriptional activator activity, which we Mtf2 that is related to Mtf1, which has been shown to be were particularly interested in since they are critical involved in regulating lymphocyte cell numbers in the downstream mediators of signaling pathways. Moreover, hematopoietic system,39 and Nr2f2 (COUP-TFII) that there is a great deal of interest in developing ‘network’ plays important roles in regulating glucose homeostasis40 interactions critical to developmental processes and and in repressing Notch signaling.41 identification of such factors will contribute to the Seventy-five transcriptional regulators were identified generation of accurate network maps. Eight hundred as over-represented in the thymus libraries as compared and seventy-one of the genes in our libraries were to our meta-library (Supplementary Table 2). Those with associated with one of these terms (henceforth, collec- high significance factors included Pax1,13,15,16 c-Myb42–44 tively referred to as transcriptional regulators). Included Satb145,46 and Tcf7 (Tcf1)47,48 that all have important roles in this list were Pax-1,13,15,16 Pou2f1 (Oct-1),27 Pou6f1,28 in T-cell development, as well as Foxn1 (nude) that is FoxP329–32 and Tcf21 (capsulin).1 Of greater interest are critical for thymus development.11,12,49 Several transcrip- the numerous identified transcriptional regulators not tion factors with uncharacterized roles in the thymus previously implicated in these processes. We further were also present, including the metal response element identified 21 796 (representing 38% of the mapped 56 416 binding transcription factor Mtf2 (discussed above), tags) as Riken transcripts, ESTs, other uncharacterized Dmrt2 that was thought to be specifically expressed in genes and transcripts, or genes for unnamed and somites and thus involved in somitogenesis50,51 and uncharacterized proteins. Tead3 that regulates the cardiac Troponin T promoter.52 Eighty-one and 128 tags from the spleen/meta-library or thymus/meta-library comparisons, respectively, Identification of signaling pathway components, transcription mapped to Riken transcripts or to other uncharacterized factors and novel genes over-represented in the spleen and genes (data not shown). Of these, several had a high thymus libraries degree of specificity (mean ratio 45) and a significant Transcripts that are significantly over-represented in a count level (47), and thus make interesting candidates tissue are more likely to play a significant role in that for functional studies to determine their roles in spleen, tissue and thus make interesting candidates for further thymus and immune cell development/function. analysis. To identify such genes/transcripts, we first identified tags over-represented (Pp0.05 using Audic- Identification of signaling pathway components, transcription Claverie statistics33) in the spleen and thymus libraries factors and novel genes over-represented in the developmental compared with 115 other available Mouse Atlas Long- spleen and thymus libraries SAGE libraries. This revealed 1857 over-represented tags To further identify signaling pathway components, in the spleen and 2814 in the thymus. We then used these transcription factors and uncharacterized genes that play tags to identify signaling pathway genes, as described significant roles in the development of the spleen, above, over-represented in the spleen or thymus thymus and/or immune cells, we identified tags over- libraries. As expected, we found numerous members of represented in the embryonic spleen and thymus the Jak-STAT pathway over-represented in both the libraries in comparison to the adult spleen and thymus spleen and thymus libraries, including Stat1 (with two libraries, respectively. In the spleen, we identified 712 different tags on both lists) and Stat4 (Supplementary tags over-represented in the two developmental libraries Table 1). Components of the Notch (Hes6), MAPK (Pp0.05 using Audic-Claverie statistics33). Included in (Mapk1, Map4k1) and the noncanonical Wnt/Ca2 þ this list were representatives of the MAPK (Raf1), Notch (Nfatc1, Nfatc3) pathways were also present. Of these, (Notch4), TGF-b (Smad5, Tgfb1) and Wnt (Ctnnb1) signal- Stat5B and Hes6 were found to be over-represented in the ing pathways (Supplementary Table 3). Twenty-five of thymus libraries as compared to the spleen libraries, the tags expressed at higher levels in the developing whereas Jak1 and Frzd were found over-represented in spleen unambiguously mapped to transcription regula- the spleen libraries when compared to the thymus tors using GO analysis. This list included Pbx1,53 Tcf21 libraries. (capsulin)1 and Fus (Pigpen),53 which have been shown to

Genes and Immunity ee n Immunity and Genes 104

Table 2 Transcription factor tags over-represented in the developmental spleen or thymus libraries as compared to the adult libraries

Sequence ID Gene Tag positiona TPH at E15.5 TPH at E16.5 TPH at 84 dpn Mean Mean in Mean Significancec (SM055- 128520) (SM011-60592) (SM012-86372) other librariesb ratio

Spleen libraries GATATTTTTTTTTGGGG 26252107 Fus 1 85 40 38 54.2 19.8 2.7 148.4 ACAAAATAACACTGTTG 27503724 Nr2f2 1 46 8 0 18.1 4.9 3.7 66.8 GCCCAGCTATCCTTAGC ENSMUSG00000004328 Hif3a 7 26 5 0 10.5 2.0 5.4 56.2 CAAAATACATTTATTTG 16741183 Maged1 1 60 40 2 33.9 32.0 1.1 36.0 AATGACTGCAAGCCTGG 31566109 Tcf21 1 22 10 0 10.6 3.5 3.0 32.2 CTCGAATAAAAATGTAG 38173725 Zfhx1b 1 13 15 3 10.5 3.5 3.0 32.0 TTTTTGAACAAAAACTT 28878995 Ctnnb1 1 27 23 10 20.3 18.2 1.1 22.5 development thymus and spleen of analysis SAGE TCCCCCACACAGTCCCC 35193058 Tgfb1i4 1 34 28 3 21.9 24.3 0.9 19.8 TCGCTGAGGCTTCTACC 6671613 Barx1 1 12 3 0 5.0 1.3 3.7 18.5 TGGGTTAGACCAGGTTA 19354232 Pfdn1 1 20 20 6 15.3 13.0 1.2 17.9 GGATGATGAAAACCTGG 14198161 Nr2f6 1 19 12 2 10.8 8.0 1.4 14.7 GATAATAAAGCTAATTT 6754873 Notch4 1 4 7 0 3.5 0.9 4.0 14.0 TGAGTGAAGAGATGTCA 44890607 Sox12 11 9 7 0 5.1 2.0 2.6 13.0 GAGCAGGAGCCTGCCCT 38148662 Hdac5 1 9 7 1 5.4 3.4 1.6 8.7 AATAAACTTTGCTTTCC ENSMUSG00000038718 Pbx1 4 9 5 1 4.9 3.1 1.6 7.7 TCTGTAGCCCAGTGCCC 28386178 Thra 1 13 13 0 8.8 10.5 0.8 7.4

TGTTTAATGTGATTATT 28436932 Asb4 1 8 3 0 3.7 2.2 1.7 6.2 Hoffman BG AATGCTTGATTTGTCAA 13277791 Rbbp7 1 9 7 1 5.4 5.2 1.0 5.6 ACCATTACTGCATTGTC ENSMUSG00000024140 Epas1 257 9 2 0 3.4 2.2 1.6 5.4 CTGCTGACATATTCTGG 27734179 Tgif2 1 8 3 0 3.7 2.6 1.4 5.3

AAATGCAATAAATCTTG 18043552 Nfyc 1 6 5 1 4.1 3.3 1.2 5.1 al et TTTATGAAAATAAATAG 13529442 Nmyc1 1 8 2 0 3.1 2.6 1.2 3.7

Sequence ID Gene Tag position TPH at E15.5 TPH at E18.5 TPH at 84 dpn Mean Mean in Mean Significance (SM057-110638) (SM022-108120) (SM024-130828) other libraries ratio

Thymus libraries CAAAACTGTAGGTAATG ENSMUSG00000026923 Notch1 4 24 17 11 17.2 2.1 8.2 141.3 TGGTGGACCTGTGTAAA 13277896 Irf1 1 16 10 8 11.4 1.1 10.5 119.5 GACTGAGCCAGGAGAGC 20913926 Hypothetical protein 1 5 2 0 2.4 0.1 22.0 53.3 LOC68701 CAATAAAATTTAAAAAT 60334797 Trp63 1 11 4 1 5.1 0.6 8.3 42.6 GTGCTCTTGATGTACAG ENSMUSG00000002249 Tead3 3 8 15 2 8.4 2.2 3.9 32.8 CCTCAAACAGCAGCCTC ENSMUSG00000003809 Gcdh 12 0 6 0 1.8 0.2 11.2 20.7 AGAATGGCAGACACTGA 47124392 Nrarp 1 5 6 0 3.7 0.8 4.4 16.2 TGTAAGTTGATGTTTGT 12805060 Tfam 1 5 12 2 6.0 2.4 2.6 15.4 CAAAATACATTTATTTG 16741183 Maged1 1 32 28 6 21.8 33.9 0.6 14.0 CACTGCATTTGTATATA 17390875 Hes1 1 5 5 1 3.3 1.0 3.5 11.5 RIKEN cDNA CAGGTGCCATACTGAGG 29748017 2700067D09 1 10 2 1 4.2 2.0 2.1 8.9 AAAATGTGGTTCTTGCA 55991527 Hey1 1 5 4 0 3.0 1.0 2.9 8.9 TGGGTTAGACCAGGTTA 19354232 Pfdn1 1 14 11 6 10.6 13.5 0.8 8.2 GCTAAGGCTTAGCAGCA 27369498 Ankfy1 1 2 4 0 1.8 0.7 2.8 5.1 AATGATGCCTTTGGTGA 37231552 Trim28 1 14 3 4 6.7 10.4 0.6 4.3 GGGTCCTCTTTTCTCAG 16924210 Meis2 1 4 2 0 1.8 3.3 0.6 1.0 CTCATCCGGCAGGGTAT 12805366 Hsbp1 1 2 6 0 2.5 6.3 0.4 1.0

TPH ¼ tags per hundred thousand; dpn ¼ days post-natal. a Tag position for transcript based resources (i.e. Refseq, MGC, and so on) is calculated by assigning the 30 most CATG in the transcript as position 1 tag and the next 30 most CATG the position 2 tag, and so on. For Ensembl genes, the position 1 tag is from the 30 most CATG on the + strand in the known gene sequence on the genome, including introns, UTR, and so on. The À1 position tag would conversely be from the 30 most CATG on the À strand. b Refers to the mean counts in 115 Non-spleen/Thymus LongSAGE Libraries available at www.mouseatlas.org. c The mean count in the spleen or thymus libraries multiplied by the mean ratio determined in those libraries as compared to 115 other Mouse Altas LongSAGE libraries. SAGE analysis of spleen and thymus development BG Hoffman et al 105 play important roles in B-cell development and spleen RT-qPCR analysis of the cell-type specificity and regulated organogenesis (Table 2). Six of the genes on this list expression of RIKEN 9230105E10 (Tcf21, Nr2f2, Zfhx1b, Fus, Hif3a and Barx1) were also Although the analysis of our libraries revealed the identified as over-represented in the spleen libraries as expression of numerous interesting and potentially compared to the other Mouse Atlas LongSAGE libraries developmentally or immunologically significant tran- (Supplementary Table 2). Uncharacterized transcripts scripts, we focused our further work on RIKEN accounted for twenty-four of the over-represented tags in 9230105E10 that encodes a homolog of the primate the developmental spleen libraries (Supplementary Table HIV-1 restriction factor Trim5a, which has been shown 4). This list included transcripts with extremely high to confer resistance to HIV-1 in Rhesus monkeys.56,57 significance factors, including hypothetical protein MNCb- Although expression of this transcript did not change 5680 that had a significance factor of 69.9. during development, understanding its expression pat- Within the two developing thymus libraries, 767 tags tern is of considerable immunological relevance. were over-represented, including components of the Although previous data suggest that this gene is Notch (Hes1, Notch1), TGF-b (Smad5), Wnt (Fzd7) and ubiquitously expressed,58 the data are not convincing FGF (Fgfbp1) signaling pathways. Twenty tags expressed and immune cell types were not analyzed. at higher levels in the developing thymus mapped to First, we assessed the expression of RIKEN 9230105E10 transcriptional regulators, including Irf1 that acts as an in various adult tissues and cell lines (Figure 2a and b). activator of interferon-g and -b expression,54,55 as well as Although its expression was detected in each of the Notch1 and Hes1 (Table 2), which are both members of samples analyzed, with the exception of the fibroblast the Notch signaling pathway. Seven of these (Tfam, cell line NIH3T3, far higher levels (410 times) were Notch1, Tead3, Gcdh, Nrarp, hypothetical protein LOC68701 noted in samples with an increased concentration of and Trp63) were also over-represented in the thymus as immune cells, such as broncheolar lavage and peripheral compared to the other Mouse Atlas libraries. In the blood mononuclear cells, and in several of the immune developing thymus, 31 of the over-represented tags cell-like cell lines including the macrophage-like cell line mapped to uncharacterized transcripts (Supplementary WEHI-3 and the B-cell-like cell line K46. We next Table 4). As observed in the spleen, many of these determined the expression pattern of RIKEN had extremely high significance factors including 9230105E10 in numerous purified immune cell types. 2010001M09Rik protein (129.7), hypothetical protein RT-qPCR analysis revealed no significant differences in LOC66167 (98.9) and RIKEN cDNA D930015E06 (94.6). expression in the various cell types analyzed. However, expression in these cell types was substantially higher Validation of transcription factor and uncharacterized gene (410 times) compared to the levels found in primary expression by RT-qPCR tissues (Figure 2c). Reverse transcription (RT)-qPCR was used to confirm, Although Trim5a has been shown to play a role in for selected candidates, the expression patterns sug- resistance to HIV-1, it is not known whether its gested by our SAGE analysis. Seven transcription factors expression is regulated at the transcriptional level. CpG and four uncharacterized genes were selected to: (A) activates signaling through TLR9, which is highly validate the expression of interesting transcripts that expressed on plasmacytoid DCs, as well as B cells, but have not been implicated in spleen or thymus develop- not myeloid DC.59,60 LPS, on the other hand, signals ment, particularly transcriptional regulators with low through TLR4 that is expressed on various immune cells count numbers and (B) ensure that a wide range of including macrophages, B cells and myeloid DC, but not expression profiles were included (Table 3). plasmacytoid DC.59–64 Thus, we sought to determine if In general, the two techniques provided similar Trim5a expression is regulated by TLR activation in results. For example, tags for Tbx6 were only detected different cell types. FACS-sorted Flt-3L-derived B220À in the two adult libraries (at a count of one in the adult myeloid and B220 þ plasmacytoid DC populations were spleen and two in the adult thymus) and the RT-qPCR stimulated with CpG, whereas phenol-purified LPS65 results (Figure 1) confirmed this. The expression patterns was used to stimulate thioglycollate-induced peritoneal observed in the SAGE libraries for Scml4, Dmrt2 and macrophages and primary spleen B cells. Interestingly, Sp100 were also confirmed. Klhl6 was found at a CpG stimulation resulted in a significant increase in relatively constant level in all libraries by both techni- Trim5a transcript levels, but only in the plasmacytoid ques. Since Zfp7 and Trim34 were detected at a single population, which is consistent with the restricted count only in the adult spleen and E18.5 thymus expression of TLR9 to this population59 (Figure 3a). The libraries, respectively, it was not possible to determine LPS-stimulated macrophages also exhibited an increase if there were statistical differences in their expression (two to threefold) in RIKEN 9230105E10 transcript levels patterns using SAGE. RT-qPCR analysis of Zfp7 expres- (Figure 3b). Of note, this increase in expression was not sion revealed relatively constant levels at all stages observed in the B cells. assessed, whereas Trim34 showed expression in the E18.5 thymus library as well as the adult spleen and thymus libraries. Unknown (MGC: 62947) was detected predomi- Discussion nantly in the developing thymus libraries using both SAGE and RT-qPCR, whereas Riken 1810037B05 was The molecular mechanisms regulating embryonic devel- detected at a very high level only in the adult spleen. opment of the mouse spleen and thymus, as well as cDNA sequence BC027373 and RIKEN 9230105E10 were associated immune cells, are poorly understood. We thus detected at a similar level in all libraries by RT-qPCR and generated and analyzed six LongSAGE libraries – three by SAGE, although cDNA sequence BC027373 was from the spleen and three from the thymus – as a means detected at a very low level by SAGE. to identify additional factors that may be critical to these

Genes and Immunity SAGE analysis of spleen and thymus development BG Hoffman et al 106 b processes. These libraries were created from unfractio- nated tissues, rather than specific subpopulations, mak- ing them a unique resource that is not restricted to the

Mean in identification of factors involved in a specific process or other libraries cell type. Our bioinformatic analyses of these libraries showed that many (36%) mapped to uncharacterized transcripts, or to the genome in locations where thymus Mean in transcripts have not previously been identified. This suggests that these libraries can identify and describe the

spleen expression patterns of relatively uncharacterized tran- Mean in scripts. Our RT-qPCR experiments validated the expression

most CATG the position 2 tag, and so on. profiles of 11 transcripts (seven transcription factors and 0

1 position tag would conversely be from the four uncharacterized genes) and showed that our À libraries are a useful tool for the accurate detection and TPH at 84 dpn (SM024-130828) quantification of transcripts. Of note, the expression of transcripts detected even at a single count in our SAGE data was always confirmed by our RT-qPCR analysis. However, deeper sequencing of these libraries would improve confidence in the expression of low abundance TPH at E18.5

(SM022-108120) transcripts and provide a more robust detection of rare transcripts.66 One of our main goals was to identify key regulators of developmental or immunological processes in the spleen or thymus, as well as genes highly expressed or specific TPH at E15.5

(SM057-110638) to these tissues whose functional significance is not yet known. To achieve these objectives, we compared the two data sets to a meta-library composed of 115 other Mouse Atlas SAGE Libraries. This data set was chosen owing to the compatibility of the library types (all TPH at 84 dpn

(SM012-86372) LongSAGE libraries), the comparable depth of sequen-

most CATG in the transcript as position 1 tag and the next 3 cing of all of the libraries and the fact that the 0 other libraries in this data set broadly represent numerous normal developing and adult tissues. We also isolated tags over-represented in the developing TPH at E16.5

(SM011-60592) libraries as compared to the adult libraries in order to identify transcripts with roles specifically in the development of these tissues or immune cells. From each of these analyses, we identified known key transcriptional regulators of spleen, thymus and/

TPH at E15.5 or immune cell development or function, validating (SM055- 128520) our approach. More importantly, it suggests that a the other transcriptional regulators identified may 11111 08 0 0 61 01 01 01 01 0 0 0 1 0 0 0 0 1 0 1 0 0 12 0 0 2 0 1 2 0 0 3 3 4 3 38 0 2 0 3 1 0 46 5 7 1 0 0 0 0 1 0 34 2 1 13 1 0 2 0.0 1 0 0.3 0 0.00 6.3 14 0.3 0.7 0.7 2.7 12 8.0 0.3 0 0.0 0.7 1.2 0 0.1 0.00 5 0.2 1.0 1.2 31.4 12.7 0.1 0.1 0.8 4.7 0.0 2.0 0.8 0.3 1.8 0.2 0.4 0.1 0.3 10 0 0 1 0 0 0 0.3 0.0 0.0 Tag also be involved in these processes. Numerous unchar- position acterized transcripts were also identified from these comparisons. We further analyzed the expression of one of the genes validated, RIKEN cDNA 9230105E10, owing to its

most CATG on the + strand in the known gene sequence on the genome, including introns, UTR, andhomology so on. The to the primate HIV-1 restriction factor Tri- 0 m5a.56,57 The expression of Trim5a has previously been described as ubiquitous,58 although the authors note the days post-natal. Scml4 Trim34 Tbx6 Dmrt2 Khlh6 Sp100 Zfp7 Unknown (MGC:62947) RIKEN 5830443L24 RIKEN 1810037B05 cDNA BC027373 RIKEN 9230105E10 Northern blot results were smeary. We, similarly, found ¼ expression of RIKEN cDNA 9230105E10 in all samples tested. However, its expression was elevated in tissues with higher levels of immune cells or in immune-cell-like cell lines. We thus tested purified immune cell populations for the expression of RIKEN cDNA 9230105E10. Far strand.

À higher levels (410-fold) of expression were seen in these populations than were observed in the whole tissues, suggesting that the majority of expression seen in the tissue samples may be due to immune cell contamina- Summary of the SAGE tag counts for transcripts validated by RT-qPCR tion. These data clearly show that expression of RIKEN

tags per hundred thousand; dpn cDNA 9230105E10 is elevated in immune cell popula- ¼ tions, consistent with the role it plays in preventing HIV most CATG on the GCTGCGTTCTTAGCTGTTAGAAGAGATGCTTACC 34328393 13507606 CCAATTTGTTCTCAGGATATCTTTTCTATCACAC 48928034 CCTGTCTTAGAGGAAGG 29789376 34419627 GACTTTGAAATCGAAGGGAGGCGGTGACCTTTGG 7305512 22095020 AAGACCCCCACTCTAGA 32450336 GAGCTCAACAAACCATT 21594498 GCGTATGTCCTATACCCTAAACCAGCAGCCATTC 28482155 20072084 TTAGAGAGGAGGGAAGG 24660348 0 Refers to the mean counts in 115 Non-spleen/Thymus LongSAGE Libraries available at www.mouseatlas.org. Tag position for transcript based resources (i.e. Refseq, MGC, and so on) is calculated by assigning the 3 For Ensembl genes, the position 1 tag is from the 3 3 Table 3 Sequence(A) Transcription factors ID Gene(B) Uncharacterized genes TPH a b infection. Spleen libraries Thymus libraries

Genes and Immunity SAGE analysis of spleen and thymus development BG Hoffman et al 107

Figure 1 RT-qPCR validation of the expression of various uncharacterized genes and transcription factors during spleen and thymus development. The expression of (a) seven transcription factors and (b) five uncharacterized genes was determined by RT-qPCR in manually dissected spleen and thymus tissues, from E15.5 and E18.5embryos, as well as adult (84 dpn) mice. For the RT-qPCR reactions, b-actin was used as the control and the thymus adult sample as the reference. The data shown is an average of the results obtained from three separate replicate RTs, each with duplicate reactions. Error bars indicate the standard deviation of the three averaged replicates.

Our results show that upregulation of RIKEN cDNA work raises the interesting possibility that HIV-1 infec- 9230105E10 expression can occur through both TLR4 and tion in primates may cause a TLR-induced upregulation TLR9 signaling pathways in macrophages and plasma- of Trim5a. Experimental evidence will be required to cytoid DC, respectively. The lack of LPS-induced validate this hypothesis and determine whether these upregulation in B cells suggests that this response may alterations in transcript levels are important in mediating occur primarily in cells involved in innate immunity. resistance to HIV-1 infection. Clearly, further studies are required to determine if these, In conclusion, the libraries created and analyzed here or other, pathways regulate RIKEN cDNA 9230105E10 are a useful resource for dissecting the signaling and expression in various immune cell populations. Our transcriptional networks regulating the development of

Genes and Immunity SAGE analysis of spleen and thymus development BG Hoffman et al 108

Figure 3 Analysis of the expression of RIKEN 9230105E10 in response to TLR agonists. RIKEN 9230105E10 expression was determined by RT-qPCR in three separate preparations of (a) FACS-sorted B220 þ plasmacytoid and B220À myeloid Flt-3L- derived DC treated or untreated with CpG (TLR9 agonist) and (b) macrophages and B cells treated or untreated with LPS (TLR4 agonist). For the RT-qPCR reactions, RIKEN 9230105E10 expression was normalized to the expression of 18S rRNA and the fold Figure 2 Analysis of the expression of RIKEN 9230105E10, which difference was compared to the bronchial lavage sample as a encodes a murine homolog of the primate HIV-1 restriction factor reference. The data shown is an average of the results obtained from Trim5a, in various tissues and cell lines. RIKEN 9230105E10 three separate replicate RTs, each with duplicate reactions. Error expression was determined in triplicate preparations of (a) bars indicate the standard deviation of the three averaged bronchial lavage (BAL), the spleen, testis, skeletal (skel) muscle, replicates. *Statistically significant difference between the untreated lung, liver, kidney, heart, gut, brain, thymus, peripheral blood and treated samples at Pp0.05 based on a Student’s t-test. mononuclear cells (PBMC) and embryonic stem (ES) cells; (b) the cell lines J774A.1, EL-4, AW 264.7, Wehi 3, 18.81, MEL, K46 and NIH3T3 and (c) primary hematopoietic cells including eosinophils, þ þ þ neutrophils, CD4 T cells, CD8 T cells, CD25 T cells, Treg cells, B mechanisms by which the body responds to HIV-1 cells, mast cells, Flt-3L-derived bone marrow dendritic cells (Flt-3L BMDCs) and macrophages. For the RT-qPCR reactions, RIKEN infection. 9230105E10 expression was normalized to the expression of 18S rRNA and the fold difference was compared to the BAL sample as a reference. The data shown is an average of the results obtained from three separate replicate reverse transcriptions, each with duplicate Materials and methods reactions. Error bars indicate the standard deviation of the three Generation of SAGE libraries averaged replicates. Embryos obtained from timed pregnant female C57Bl/6 mice were staged using Theiler criteria for the isolation of E15.5 and E16.5 spleen, as well as E15.5 and E18.5 lymphocytes, the spleen or the thymus. Our work further thymus, tissues. Adult tissues were collected from 84- shows the regulated and cell-type specificity of expres- day post-natal C57Bl/6 mice. All tissues were manually sion of one of our identified candidates, a mouse dissected and placed immediately into RNAlater (Invi- homolog of Trim5a, which not only demonstrates the trogen). For the E15.5 spleen library, 25 spleens were utility of our SAGE libraries but also suggests new pooled, whereas 10 were used for the E16.5 library.

Genes and Immunity SAGE analysis of spleen and thymus development BG Hoffman et al 109 Twenty-five and 10 thymuses were used for the E15.5 Quantitative real-time PCR and E18.5 thymus libraries, respectively. For the adult RNA from each tissue or cell line was prepared using libraries, three spleens or three thymuses were pooled. Trizol (Invitrogen). Primers were designed using Primer3 Total RNA was extracted using Trizol (Invitrogen). RNA (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www. yields of 24, 72 and 18 mg were obtained from the E15.5, cgiT) and spanned introns where possible. Amplicons were E16.5 and adult spleens, respectively, whereas 47, 22 and between 80 and 120 bp for efficient amplification. Primer 50 mg were obtained from the E15.5, E18.5 and adult efficiencies were determined using a dilution series of thymuses. RNA quality was assessed using a Bioanalyser mixed embryonic spleen and thymus cDNA. Only (Agilent). Roughly, 10 mg of total RNA was used to primers pairs with an efficiency greater than 0.9 were construct each SAGE library. The I-SAGE long kit was used in subsequent analyses. The primers used are listed used to make all libraries (Invitrogen). Ligations were in Table 4. An ABI 7500 real-time PCR system (Applied transformed by electroporation and the resulting colo- Biosystems) and SYBRs Green supermix (Applied nies picked and arrayed into a 384-well format using a Q- Biosystems) were used for validations of the SAGE data, Pix platform. Plasmid DNA was purified and sequenced whereas an ABI 7900 real-time PCR system (Applied using BigDye terminators (version 3) with sequencing Biosystems) and SYBRs Green supermix (Marsh) were products resolved using an ABI 3700 automated sequen- used for the quantification of RIKEN cDNA 9230105E10 cer. Bioinformatic screening of the libraries for contam- expression. Triplicate cDNAs were obtained by RT of 1 mg ination, excess replicate ditags and sequence quality of total RNA from newly isolated tissues. All reactions ensured that they were of high quality. were carried out in duplicate using 10 ng of generated cDNA in each reaction. For validation of the SAGE data, samples were normalized to b-actin and the fold increase SAGE data analysis calculated using 2ÀDDCT .68 For quantification of RIKEN SAGE data was analyzed using DiscoverySpace (http:// cDNA 9230105E10 in tissues and cell lines, samples were www.bcgsc.ca/bioinfo/software/discoveryspace/). All normalized to 18S rRNA, the absolute number was SAGE libraries were generated by the Mouse Atlas of quantified and then expressed relative to the bronchial Gene Expression project (www.mouseatlas.com). They lavage sample. were filtered to remove ditags and also for sequence quality using a 95% quality cut off for all tags (these Cell and tissue isolation, culture and FACS sorting percentages refer to a quality measure based on the Unless otherwise indicated, all reagents for cell isolation PHRED score67 of the bases in the sequence). Tag to gene and culture were purchased from StemCell Technologies mapping was performed using the mouse Refseq, MGC Incorporated (STI). Total bone marrow was flushed from and Ensembl databases using the CMOST plugin in the femurs and tibias of mice with Dulbecco’s phos- DiscoverySpace. Tags were considered sense position phate-buffered saline (PBS) containing 2% fetal bovine matches if they mapped in the sense orientation to the serum (FBS). Bone marrow cells were seeded at a density gene, and antisense matches if they mapped in the of 1.5 Â 106 total nucleated cells/ml in Petri-style culture opposite orientation. Tag ‘position’ was determined by dishes in medium consisting of (RPMI) supplemented sequentially numbering tags from the 30 most tag with 10% FBS, 100 U/ml penicillin, 100 mg/ml strepto- 0 (position 1) onward (i.e. next 3 most tag would be mycin, 2 mML-glutamine, 115 mM monothioglycerol position 2, and so on). A tag was considered unambig- (MTG; Sigma) and 150 ng/ml Flt-3L. DC cultures were uous if it matched a single sense position gene in all of maintained at 371C in a humidified incubator with 5% the databases, and ambiguous if it mapped to multiple CO2 and then stimulated by the addition of 5 mg/ml of genes in a sense position. Genes were considered over- CpG for the last 18 h of culture before sorting the DC into represented at Pp0.05 using Audic-Claverie statistics.33 B220À and B220 þ populations using a FACS-Vantage GO analysis of genes was performed using Discovery- Flow Cytometer/Cell Sorter (Becton Dickinson). Six Space. independent differentiations were performed and

Table 4 Primer sequences used for RT-qPCR

L-1810037B05 TGGGCATGAAGACAGACAAA L-Dmrt2 ATTTCGATCGGAAAGCAGTG R-1810037B05 GGTCCATTCATCTTCCCAGA R-Dmrt2 TAAGTCTCTGCCGTCATTGG L-5830443L24 GAGCAGCTCATCAAAGACCA L-Scml4 CTTCAGCCACCTGCCTTTT R-5830443L24 TCCGTCCATTTCTTCAGCTT R-Scml4 CGTTTGCCAGACACTCTTCA L-62947 TCCATCCACTCCAGAAGTCC L-Trim34 AGCAGGAGAAGGTGGAGACA R-62947 GGCCATGGGTAAAAGCTGTA R-Trim34 AAAGCCTGTTTGAATCCTTTG L-9230105E10 CCAGGGAACAAAGAAAGTTCC L-Zfp7 TGGTCTTTAAGCCGGAATTG R-9230105E10 GTTTTGAACCAGCGTCACCT R-Zfp7 CAAATCCTTGGTGCCTCTGT L-BC027373 GCAGGTCACACACTGCTCAT L-Neurod4 CCTGAATGATGCCTTGGATAA R-BC027373 CAGATCTGCCTTGGGGATAA R-Neurod4 GTTCCTTGCCAGTCGAAGAG L-Klhl6 GAAGTGGGTGGAGTTTGCAT L-Rslcan1 CTGGTCACATTTCTGGAGCA R-Klhl6 TCCATTTGTTGATGGAGGAA R-Rslcan1 GGCCTTTCCACACTCTTTACA L-Sp100 GAATCATCAGCTTGCCGAAT l-Zfp369 TTTCCAAGGCAAATGTGATG R-Sp100 GCTGCAAGACTCTGGGGTAG R-Zfp369 TTCATCAGGGGACTCTCAGC L-Tbx6 CTTCTACACCCTGCCGCTTT L-b-actin GCTCTTTTCCAGCCTTCCTT R-Tbx6 CAGAAATGCAGCTGAGTAGGG R-b-actin CGGATGTCAACGTCACACTT

RT-qPCR ¼ reverse transcription-quantitative polymerase chain reaction.

Genes and Immunity SAGE analysis of spleen and thymus development BG Hoffman et al 110 roughly 20 000 B220 þ cells and 500 000 B220À cells were mice. Funding for this project was provided by Genome obtained from each. Canada through Genome BC (CDH) and the National NIH3T3 and MEL cells were maintained in DMEM Institute of Health AI29564 (JT), with infrastructure (high glucose) supplemented with 10% fetal calf serum support provided by the BC Cancer Foundation. Kristi 5mML-glutamine and streptomycin–penicillin. Embryo- L Williams is supported in part by the 2004 Amgen/Focis nic stem cells (ESC) were maintained on a layer of Fellowship Award, whereas Cheryl D Helgason is a irradiated mouse embryo fibroblasts and fed daily with a scholar of the Canadian Institutes of Health Research and complete change of ESC maintenance medium consisting the Michael Smith Foundation for Health Research. of high glucose DMEM supplemented with 15% ESC- tested FBS, 0.1 mM nonessential amino acids, 2 mML- glutamine, 1000 U/ml LIF, 100 U/ml penicillin, 100 mg/ References ml streptomycin and 100 mM MTG (Sigma). All other cell lines were maintained in RPMI-1640 supplemented with 1 Lu J, Chang P, Richardson JA, Gan L, Weiler H, Olson EN. The 10% fetal calf serum, 2 mML-glutamine, 1 mM sodium basic helix–loop–helix transcription factor capsulin pyruvate, 0.1 mM nonessential amino acids and strepto- controls spleen organogenesis. Proc Natl Acad Sci USA 2000; mycin–penicillin. All cells were grown at 371C with 5% 97: 9525–9530.

CO2. 2 Hecksher-Sorensen J, Watson RP, Lettice LA, Serup P, Eley L, Tissue RNA was harvested from PBS-perfused mice to De Angelis C et al. The splanchnic mesodermal plate directs reduce the level of blood cell contamination. RNA was spleen and pancreatic laterality, and is regulated by Bapx1/ isolated from eosinophils (95% purity) harvested by Nkx3.2. Development 2004; 131: 4665–4675. bronchial lavage of OVA-sensitized mice after day 5 of 3 Dear TN, Colledge WH, Carlton MB, Lavenir I, Larson T, Smith AJ et al. The Hox11 gene is essential for cell survival challenge. Peritoneal neutrophils were harvested 6 h during spleen development. Development 1995; 121: 2909–2915. þ þ after thioglycollate injection. Spleen CD4 and CD8 T 4 Tribioli C, Lufkin T. The murine Bapx1 homeobox gene plays a cells were selected using MiniMacs (Miltenyi Biotec). critical role in embryonic development of the axial skeleton Macrophages were harvested 3 days postperitoneal and spleen. Development 1999; 126: 5699–5711. injection of thioglycollate and purified by plate adher- 5 Kanzler B, Dear TN. Hox11 acts cell autonomously in spleen ence for 2 h. CD25À splenic T cells were isolated using development and its absence results in altered cell fate of Mouse T-cell Immunocolumns (Cedarlane Laboratories), mesenchymal spleen precursors. Dev Biol 2001; 234: 231–243. according to the manufacturer’s protocol, and stained 6 Cyster JG. Lymphoid organ development and cell migration. with anti-CD25–biotin (clone 7D4; BD Biosciences Phar- Immunol Rev 2003; 195: 5–14. mingen), followed by streptavidin–PE. Cells were 7 Mebius RE. Organogenesis of lymphoid tissues. Nat Rev Immunol 2003; 3: 292–303. labeled with anti-PE microbeads (Miltenyi Biotech), and 8 Boehm T, Bleul CC, Schorpp M. Genetic dissection of thymus þ CD25 cells depleted following methods from the development in mouse and zebrafish. Immunol Rev 2003; 195: manufacturer (Miltenyi Biotech). The depletion of 15–27. þ CD25 T cells was greater than 90%. Treg were purified 9 Gill J, Malin M, Sutherland J, Gray D, Hollander G, Boyd R. from the spleens of donor mice, to 90–95% purity, using a Thymic generation and regeneration. Immunol Rev 2003; 195: previously described column-purification protocol.69 In 28–50. vitro differentiated mast cells were obtained by co- 10 Bockman DE. Development of the thymus. Microsc Res Technol culturing bone marrow with IL-3 and SCF for 8 weeks. 1997; 38: 209–215. Spleen B cells were purified from C57Bl/6 mice using a 11 Coffer PJ, Burgering BM. Forkhead-box transcription factors 4 B-cell purification column (Accurate Chemical & Scien- and their role in the immune system. Nat Rev Immunol 2004; : 889–899. tific Corp.) according to the manufacturer’s instructions. 12 Gordon J, Bennett AR, Blackburn CC, Manley NR. Gcm2 and Foxn1 mark early parathyroid- and thymus-specific domains in the developing third pharyngeal pouch. Mech Dev 2001; 103: 141–143. Acknowledgements 13 Su D, Ellis S, Napier A, Lee K, Manley NR. Hoxa3 and pax1 regulate epithelial cell death and proliferation during thymus Eosinophil and mast cell RNA was a kind gift from BH and parathyroid organogenesis. Dev Biol 2001; 236: 316–329. À 14 Manley NR, Capecchi MR. The role of Hoxa-3 in mouse Koller, CD25 and Treg cell RNA were a kind gift from J Serody and purified B cells were a kind gift of SH Clarke. thymus and thyroid development. Development 1995; 121: 1989–2003. This work was carried out as part of the Mouse Atlas of 15 Su DM, Manley NR. Hoxa3 and pax1 transcription factors Gene Expression Project, a collaborative project invol- regulate the ability of fetal thymic epithelial cells to promote ving the following Principle Investigators: Cheryl D thymocyte development. J Immunol 2000; 164: 5753–5760. Helgason, Pamela Hoodless, Steven Jones, Marco Marra, 16 Wallin J, Eibel H, Neubuser A, Wilting J, Koseki H, Balling R. Elizabeth M Simpson and Greg Riggins. Daniella Pax1 is expressed during development of the thymus Gerhardt and Robert Strausberg were also important epithelium and is required for normal T-cell maturation. contributors to this project. We would like to acknowl- Development 1996; 122: 23–30. edge members of the Genome Sciences Center sequen- 17 Hetzer-Egger C, Schorpp M, Haas-Assenbaum A, Balling R, cing, bioinformatics and SAGE library construction Peters H, Boehm T. Thymopoiesis requires Pax9 function in teams, in particular Anita Charters, Jennifer Asano, thymic epithelial cells. Eur J Immunol 2002; 32: 1175–1181. 18 Dallman MJ, Smith E, Benson RA, Lamb JR. Notch: control of Allen Delaney, Jaswinder Khattra and Asim Siddiqui, lymphocyte differentiation in the periphery. Curr Opin for their valuable contributions. We would also like to Immunol 2005; 17: 259–266. thank Adrian Burke and Caroline Astell for their 19 Robey EA, Bluestone JA. Notch signaling in lymphocyte assistance in project management and the staff of the development and function. Curr Opin Immunol 2004; 16: Joint Animal Facility at the BBCRC for maintaining the 360–366.

Genes and Immunity SAGE analysis of spleen and thymus development BG Hoffman et al 111 20 Radtke F, Wilson A, Mancini SJ, MacDonald HR. Notch 41 You LR, Lin FJ, Lee CT, DeMayo FJ, Tsai MJ, Tsai SY. regulation of lymphocyte development and function. Nat Suppression of Notch signalling by the COUP-TFII transcrip- Immunol 2004; 5: 247–253. tion factor regulates vein identity. Nature 2005; 435: 98–104. 21 Guidos CJ. Notch signaling in lymphocyte development. 42 Allen III RD, Bender TP, Siu G. c-Myb is essential for early T Semin Immunol 2002; 14: 395–404. cell development. Genes Dev 1999; 13: 1073–1078. 22 Allman D, Aster JC, Pear WS. Notch signaling in hematopoi- 43 Bender TP, Kremer CS, Kraus M, Buch T, Rajewsky K. Critical esis and early lymphocyte development. Immunol Rev 2002; functions for c-Myb at three checkpoints during thymocyte 187: 75–86. development. Nat Immunol 2004; 5: 721–729. 23 Anderson AC, Robey EA, Huang YH. Notch signaling 44 Lieu YK, Kumar A, Pajerowski AG, Rogers TJ, Reddy EP. in lymphocyte development. Curr Opin Genet Dev 2001; 11: Requirement of c-myb in T cell development and in mature T 554–560. cell function. Proc Natl Acad Sci USA 2004; 101: 14853–14858. 24 Wahl MB, Heinzmann U, Imai K. LongSAGE analysis 45 Cai S, Han HJ, Kohwi-Shigematsu T. Tissue-specific nuclear significantly improves genome annotation: identifications architecture and gene expression regulated by SATB1. Nat of novel genes and alternative transcripts in the mouse. Genet 2003; 34: 42–51. Bioinformatics 2005; 21: 1393–1400. 46 Nie H, Maika SD, Tucker PW, Gottlieb PD. A role for SATB1, a 25 Saha S, Sparks AB, Rago C, Akmaev V, Wang CJ, Vogelstein B nuclear matrix association region-binding protein, in the et al. Using the transcriptome to annotate the genome. Nat development of CD8SP thymocytes and peripheral T lym- Biotechnol 2002; 20: 508–512. phocytes. J Immunol 2005; 174: 4745–4752. 26 Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry 47 Leiden JM. Transcriptional regulation of T cell receptor genes. JM et al. Gene ontology: tool for the unification of biology. The Annu Rev Immunol 1993; 11: 539–570. Gene Ontology Consortium. Nat Genet 2000; 25: 25–29. 48 Rothenberg EV, Taghon T. Molecular genetics of T cell 27 Wang VE, Tantin D, Chen J, Sharp PA. B cell development and development. Annu Rev Immunol 2005; 23: 601–649. immunoglobulin transcription in Oct-1-deficient mice. Proc 49 Su DM, Navarre S, Oh WJ, Condie BG, Manley NR. A domain Natl Acad Sci USA 2004; 101: 2005–2010. of Foxn1 required for crosstalk-dependent thymic epithelial 28 Messier H, Brickner H, Gaikwad J, Fotedar A. A novel POU cell differentiation. Nat Immunol 2003; 4: 1128–1135. domain protein which binds to the T-cell receptor beta 50 Meng A, Moore B, Tang H, Yuan B, Lin S. A Drosophila enhancer. Mol Cell Biol 1993; 13: 5450–5460. + + doublesex-related gene, terra, is involved in somitogenesis in 29 Sakaguchi S. Naturally arising Foxp3-expressing CD25 CD4 vertebrates. Development 1999; 126: 1259–1268. regulatory T cells in immunological tolerance to self and non- 51 Kim S, Kettlewell JR, Anderson RC, Bardwell VJ, Zarkower D. self. Nat Immunol 2005; 6: 345–352. Sexually dimorphic expression of multiple doublesex-related 30 Fontenot JD, Rudensky AY. A well adapted regulatory genes in the embryonic mouse gonad. Gene Expr Patterns 2003; contrivance: regulatory T cell development and the forkhead 3: 77–82. family transcription factor Foxp3. Nat Immunol 2005; 6: 52 Azakie A, Lamont L, Fineman JR, He Y. Divergent transcrip- 331–337. tional Enhancer Factor 1 (DTEF-1) regulates the cardiac 31 Fontenot JD, Rasmussen JP, Williams LM, Dooley JL, Farr AG, Troponin T promoter. Am J Physiol Cell Physiol 2005; 289: Rudensky AY. Regulatory T cell lineage specification by the C1522–C1534. forkhead transcription factor foxp3. Immunity 2005; 22: 53 Brendolan A, Ferretti E, Salsi V, Moses K, Quaggin S, Blasi F 329–341. et al. A Pbx1-dependent genetic and transcriptional network 32 Ochs HD, Ziegler SF, Torgerson TR. FOXP3 acts as a rheostat regulates spleen ontogeny. Development 2005; 132: 3113–3126. of the immune response. Immunol Rev 2005; 203: 156–164. 54 Miyamoto M, Fujita T, Kimura Y, Maruyama M, Harada H, 33 Audic S, Claverie JM. The significance of digital gene Sudo Y et al. Regulated expression of a gene encoding a expression profiles. Genome Res 1997; 7: 986–995. nuclear factor, IRF-1, that specifically binds to IFN-beta gene 34 Smith E, Sigvardsson M. The roles of transcription factors in B regulatory elements. Cell 1988; 54: 903–913. lymphocyte commitment, development, and transformation. J Leukocyte Biol 2004; 75: 973–981. 55 Matsuyama T, Kimura T, Kitagawa M, Pfeffer K, Kawakami T, et al. 35 Barrans SL, Fenton JA, Banham A, Owen RG, Jack AS. Strong Watanabe N Targeted disruption of IRF-1 or IRF-2 results expression of FOXP1 identifies a distinct subset of diffuse in abnormal type I IFN gene induction and aberrant Cell 75 large B-cell lymphoma (DLBCL) patients with poor outcome. lymphocyte development. 1993; : 83–97. Blood 2004; 104: 2933–2935. 56 Stremlau M, Owens CM, Perron MJ, Kiessling M, Autissier P, 36 Wlodarska I, Veyt E, De Paepe P, Vandenberghe P, Nooijen P, Sodroski J. The cytoplasmic body component TRIM5alpha Theate I et al. FOXP1, a gene highly expressed in a subset of restricts HIV-1 infection in Old World monkeys. Nature 2004; diffuse large B-cell lymphoma, is recurrently targeted by 427: 848–853. genomic aberrations. Leukemia 2005; 19: 1299–1305. 57 Yap MW, Nisole S, Lynch C, Stoye JP. Trim5alpha protein 37 Tarlinton D, Light A, Metcalf D, Harvey RP, Robb L. restricts both HIV-1 and murine leukemia virus. Proc Natl Acad Architectural defects in the spleens of Nkx2-3-deficient mice Sci USA 2004; 101: 10786–10791. are intrinsic and associated with defects in both B cell 58 Reymond A, Meroni G, Fantozzi A, Merla G, Cairo S, Luzi L maturation and T cell-dependent immune responses. et al. The tripartite motif family identifies cell compartments. J Immunol 2003; 170: 4002–4010. EMBO J 2001; 20: 2140–2151. 38 Hahm K, Sum EY, Fujiwara Y, Lindeman GJ, Visvader JE, 59 Iwasaki A, Medzhitov R. Toll-like receptor control of the Orkin SH. Defective neural tube closure and anteroposterior adaptive immune responses. Nat Immunol 2004; 5: 987–995. patterning in mice lacking the LIM protein LMO4 or its 60 Verthelyi D, Zeuner RA. Differential signaling by CpG DNA interacting partner Deaf-1. Mol Cell Biol 2004; 24: 2074–2082. in DCs and B cells: not just TLR9. Trends Immunol 2003; 24: 39 Wang Y, Wimmer U, Lichtlen P, Inderbitzin D, Stieger B, Meier 519–522. PJ et al. Metal-responsive transcription factor-1 (MTF-1) is 61 Vogel SN, Fitzgerald KA, Fenton MJ. TLRs: differential essential for embryonic liver development and heavy metal adapter utilization by toll-like receptors mediates TLR-specific detoxification in the adult liver. FASEB J 2004; 18: 1071–1079. patterns of gene expression. Mol Interven 2003; 3: 466–477. 40 Bardoux P, Zhang P, Flamez D, Perilhou A, Lavin TA, Tanti JF 62 Beutler B. Inferences, questions and possibilities in Toll-like et al. Essential role of chicken ovalbumin upstream promoter- receptor signalling. Nature 2004; 430: 257–263. transcription factor II in insulin secretion and insulin 63 Beutler B, Du X, Poltorak A. Identification of Toll-like receptor sensitivity revealed by conditional gene knockout. Diabetes 4 (Tlr4) as the sole conduit for LPS signal transduction: genetic 2005; 54: 1357–1363. and evolutionary studies. J Endotoxin Res 2001; 7: 277–280.

Genes and Immunity SAGE analysis of spleen and thymus development BG Hoffman et al 112 64 Dunne A, O’Neill LA. Adaptor usage and Toll-like receptor 67 Ewing B, Green P. Base-calling of automated sequencer signaling specificity. FEBS Lett 2005; 579: 3330–3335. traces using phred. II. Error probabilities. Genome Res 1998; 65 Hirschfeld M, Ma Y, Weis JH, Vogel SN, Weis JJ. Cutting edge: 8: 186–194. repurification of lipopolysaccharide eliminates signaling 68 Pfaffl MW. A new mathematical model for relative quantifica- through both human and murine toll-like receptor 2. tion in real-time RT-PCR. Nucleic Acids Res 2001; 29: e45. J Immunol 2000; 165: 618–622. 69 Taylor PA, Panoskaltsis-Mortari A, Swedin JM, Lucas PJ, Gress 66 Dinel S, Bolduc C, Belleau P, Boivin A, Yoshioka M, Calvo E et al. Reproducibility, bioinformatic analysis and power of the RE, Levine BL et al. L-Selectin(hi) but not the L-selectin(lo) SAGE method to evaluate changes in transcriptome. Nucleic CD4+25+ T-regulatory cells are potent inhibitors of GVHD Acids Res 2005; 33: e26. and BM graft rejection. Blood 2004; 104: 3804–3812.

Supplementary information accompanies the paper on Gene and Immunity website (http://www.natutre.com/gene)

Genes and Immunity