Gene Repression by Pax5 in B Cells Is Essential for Blood Cell Homeostasis and Is Reversed in Plasma Cells

Immunity 24, 269–281, March 2006 ª2006 Elsevier Inc.

DOI 10.1016/j.immuni.2006.01.012

Gene Repression by Pax5 in B Cells Is Essential for Blood Cell Homeostasis and Is Reversed in Plasma Cells Alessio Delogu,1 Alexandra Schebesta,1 Qiong Sun,1 Katharina Aschenbrenner,1 Thomas Perlot,1,2 and Meinrad Busslinger1,* 1 Research Institute of Molecular Pathology Vienna Biocenter Dr. Bohr-Gasse 7 A-1030 Vienna Austria

Summary The transcription factor Pax5 represses lineageinappropriate genes and activates B cell-specific genes in B lymphocytes. By identifying 110 Pax5repressed genes, we now demonstrate that Pax5 downregulates diverse biological activities including receptor signaling, cell adhesion, migration, transcriptional control, and cellular metabolism at B cell commitment. The T lymphoid or myeloid expression of these genes demonstrates that Pax52/2 pro-B cells and common lymphoid progenitors display lymphoid and myeloid promiscuity of gene expression. These lineage-inappropriate genes require continuous Pax5 activity for their repression, as they are reactivated in committed pro-B cells and mature B cells following conditional Pax5 deletion. Pax5-repressed genes are also reexpressed in plasma cells, which depend for normal function on Cd28 and Ccr2 reactivation. The loss of Pax5 during terminal differentiation thus contributes to the plasma cell transcription program. Finally, ectopic expression of the Pax5-repressed chemokine gene Ccl3 in B cells results in increased osteoclast formation and bone loss, demonstrating that Pax5-mediated gene repression is essential for normal homeostasis of hematopoietic development. Introduction Hematopoietic development is initiated in the bone marrow of adult mice by the decision of HSCs to differentiate to common myeloid progenitors (CMPs) or lymphoidprimed multipotent progenitors (LMPPs). Erythroid cell types arise from CMPs, while myeloid cells can develop from either CMPs or LMPPs (Adolfsson et al., 2005). Activation of the lymphoid Rag1 and Rag2 genes in LMPP cells characterizes the emergence of the earliest lymphocyte progenitors (ELPs), which differentiate to common lymphoid progenitors (CLPs) with their characteristic B, T, and NK cell potential (Busslinger, 2004). The entry of lymphoid progenitors into the B cell pathway depends on the transcription factors E2A, EBF, and Pax5 (BSAP). E2A and EBF coordinately activate the expression of B lymphoid genes in the earliest B cell progenitors (Busslinger, 2004). Activation of the B cellspecific transcription program is, however, not sufficient *Correspondence: [email protected] 2 Present address: Howard Hughes Medical Institute, The Children’s Hospital, Harvard Medical School, Boston, Massachusetts 02115.

to commit early progenitors to the B lymphoid lineage in the absence of Pax5. Homozygous Pax5 mutation arrests B cell development at an early pro-B cell stage (also known as pre-BI or fraction B cells) in the bone marrow (Nutt et al., 1997). Pax52/2 pro-B cells retain an extensive self renewal and broad lymphomyeloid potential characteristic of uncommitted progenitors (Nutt et al., 1999; Rolink et al., 1999). These Pax52/2 pro-B cells are in vitro clonable lymphoid progenitors with a latent myeloid potential, as they realize their lymphoid potential more efficiently than their myeloid options (Heavey et al., 2003). The restoration of Pax5 expression suppresses the multilineage potential of Pax52/2 pro-B cells while simultaneously promoting their development to mature B cells (Nutt et al., 1999). Conversely, the loss of Pax5 expression by conditional gene inactivation converts committed pro-B cells with a restricted B lymphoid potential into lymphoid progenitors with a broad developmental potential (Mikkola et al., 2002). Collectively, these data identified Pax5 as the critical B-lineage commitment factor. Uncommitted hematopoietic progenitors including Pax52/2 pro-B cells promiscuously express genes of different lineage programs by a process known as ‘‘lineage priming’’ (Hu et al., 1997; Nutt et al., 1999; Miyamoto et al., 2002; Akashi et al., 2003). At B-lineage commitment, Pax5 fulfills a dual role by activating B cell-specific genes and simultaneously repressing lineage-inappropriate genes (Nutt et al., 1999). Pax5 activates target genes coding for essential components of (pre)BCR signaling such as the signaling chain Iga (mb-1, CD79a), the stimulatory coreceptor CD19, and the central adaptor protein BLNK (Nutt et al., 1997; 1998; Schebesta et al., 2002). Hence, the transactivation function of Pax5 facilitates signal transduction from the pre-BCR and BCR, which constitute important checkpoints in B cell development. In contrast, the Pax5-dependent repression of the Csf1r (M-CSFR) and Notch1 genes renders committed B cells no longer responsive to the myeloid cytokine M-CSF or to T cell-inducing Notch1 ligands (Nutt et al., 1999; Souabni et al., 2002). Based on these two genes, we hypothesized that the repression function of Pax5 contributes to B-lineage commitment by shutting down inappropriate signaling systems. To gain further insight into the repression function of Pax5, we have now identified 110 Pax5-repressed genes coding for a plethora of proteins involved in cellcell communication, adhesion, migration, nuclear processes, and cellular metabolism. The characterization of these Pax5-repressed genes led to the following conclusions. The Pax52/2 pro-B cells and CLPs express not only lymphoid genes but also a multitude of myeloid genes, which are repressed by Pax5 at B-lineage commitment. Pax5 is continuously required to repress these genes, as they are reactivated in pro-B cells and mature B cells following conditional Pax5 deletion. Pax5-repressed genes are also reactivated in plasma cells, indicating that the physiological loss of Pax5 during terminal differentiation contributes to the plasma cell transcription program. Two of the reactivated genes, Cd28 and

Immunity 270

Ccr2, proved to be important for plasma cell differentiation or function. Finally, forced expression of the chemokine gene Ccl3 (MIP-1a) in B cells resulted in increased osteoclast formation and bone loss, indicating that the repression function of Pax5 is essential for normal hematopoiesis to occur. Results Gene Repression by Pax5 in Pro-B Cells To systematically search for Pax5-repressed genes, we created a specialized ‘‘pro-B cell’’ microarray by spotting 5000 potentially Pax5-repressed and 5000 potentially Pax5-activated transcripts from cDNA libraries that were generated by subtracting the transcripts between Pax5+/+ and Pax52/2 pro-B cells in both directions. This pro-B cell microarray was hybridized with Cy3- and Cy5-labeled probes prepared from in vitrocultured Pax5+/+ and Pax52/2 pro-B cells, respectively. Because the wild-type and mutant pro-B cells were grown under identical conditions as homogeneous cell populations free of other hematopoietic cell types, all gene expression changes could be attributed to the presence or absence of Pax5, which constitutes the only genetic difference between the two pro-B cell types. cDNA clones with a Cy5/Cy3 fluorescence ratio of > 2 (repressed) or < 0.5 (activated) were selected for further evaluation, demonstrating that Pax5 represses 44% and activates 56% of all differentially expressed genes (data not shown). Here we describe the characterization of the repressed Pax5 target genes. Sequencing of all cDNA clones with a >2-fold repression ratio revealed considerable redundancy and thus high reliability of the hybridization data, as each target gene was represented on average by w10 different cDNA clones on the pro-B cell microarray. This analysis identified 110 distinct Pax5-repressed genes, which code for six secreted proteins, 24 receptors or cell adhesion molecules, 19 intracellular signal transducers, six cytoskeleton-associated molecules, 17 metabolic proteins, nine proteases or protease inhibitors, 12 chromatin components or transcription factors, and 17 proteins of unknown function (see Table S1 in the Supplemental Data available with this article online). The expression of 83 genes was validated by RT-PCR analysis of wildtype and Pax52/2 pro-B cells (Figure 1A), and the mRNA level of selected genes was additionally quantitated by RNase protection assay (Figure S1), demonstrating that the vast majority of the identified genes was repressed >10-fold by Pax5. Five Pax5-repressed genes code for cell surface proteins, whose expression could be analyzed at the single-cell level by flow cytometry (Figure 1B). All Pax52/2 pro-B cells uniformly expressed the receptors Flt3 and CD28 as well as the cell adhesion molecules CD11a (Itgal), CD47, and Sca1 (Ly6a) on their surface in contrast to wild-type pro-B cells. In summary, we have identified a plethora of Pax5-repressed genes that define new roles for Pax5 in the downregulation of diverse cellular functions at B-lineage commitment. Promiscuous Expression of Lymphoid and Erythromyeloid Genes in Pax52/2 Pro-B Cells We next classified the different Pax5-repressed genes according to their lineage-specific expression pattern.

To this end, we hybridized the pro-B cell microarray individually with Cy3-labeled cDNA probes (red) prepared from wild-type pro-B cells, thymocytes, macrophages, and erythroblasts in combination with a Cy5-labeled Pax52/2 pro-B cell probe (blue; Figure 2). The blue color in the first two vertical rows confirmed that all identified genes were transcribed in Pax52/2 pro-B cells but repressed in wild-type pro-B cells. Of these, 25 genes were preferentially expressed in thymocytes, 25 genes in macrophages, and four genes in erythroblasts (Figures 2A–2C). A smaller proportion of the genes showed a mixed lineage phenotype, with five, four, and ten genes belonging to the macrophage-erythroblast, macrophage-thymocyte, and mixed lineage clusters, respectively (Figures 2D–2F). A large percentage of Pax5repressed genes were either not expressed in any of the three hematopoietic cell types analyzed (13 genes, Figure 2G) or were expressed only at a very low level compared to Pax52/2 pro-B cells (23 genes, Figure 2H). Some genes in the two latter clusters may, however, be expressed in hematopoietic cell types other than those analyzed, which is exemplified by J-chain (Igj) expression in plasma cells (Rinkenberger et al., 1996). Interestingly, the ‘‘no/low expression’’ clusters also contain two neuronal genes coding for the glutamate receptor subunit Gria2 and the ectoderm-neural cortex-1 (Enc1) protein as well as two muscle-specific genes coding for the transcription factor Mef2c and fast skeletal muscle troponin I (Tnni2). In conclusion, the Pax52/2 pro-B cells coexpress lymphoid, myeloid, erythroid, and nonhematopoietic genes, which are repressed by Pax5 at B cell commitment. Myeloid and Lymphoid Promiscuity of Gene Expression in CLPs Recent expression analyses of hematopoietic progenitors led to the conclusion that the myeloid-restricted CMPs express genes of different myeloid lineages, while the lymphoid-restricted CLPs express only genes of lymphoid lineages (Miyamoto et al., 2002; Akashi et al., 2003). Since the Pax52/2 pro-B cells as lymphoid progenitors are closely related to CLPs (Heavey et al., 2003; Balciunaite et al., 2005), we were surprised to see that Pax52/2 pro-B cells express as many myeloid as lymphoid genes. To rule out any in vitro culture artifact, we analyzed the expression of selected myeloid and T lymphoid genes in MPPs (LSK), CMPs, CLPs, pro-B, pre-B, immature B cells, thymocytes, and myeloid Gr1+Mac1+ cells, which were FACS sorted as highly purified cell populations (Figure S2). RT-PCR analysis of the Pax5-repressed genes in these sorted cell types led to three major conclusions (Figure 3). First, these genes were repressed in vivo from the committed pro-B cell to the immature B cell stage consistent with their identification as Pax5-repressed genes. Second, all genes analyzed were expressed in CLPs as predicted by their expression in Pax52/2 pro-B cells. Third, the myeloid but not lymphoid genes were also expressed in the CMPs and MPPs. The myeloid Pax5-repressed genes can furthermore be grouped into two classes. Seven genes (Emb, Flt3, Ramp1, Lmo2, AI789751, Wbscr5, and Ccl3) were expressed at similar levels in CLPs and myeloid Gr1+Mac1+ cells, demonstrating that their expression in the CLPs cannot be caused by myeloid cell

Multilineage Gene Repression in B Cells 271

Figure 1. Pax5-Dependent Gene Repression in Pro-B Cells (A) RT-PCR analysis. Differential expression of the indicated genes was validated by RT-PCR analysis using 10-fold serial dilutions of cDNA prepared from in vitro cultured wild-type (+/+) and Pax5 mutant (2/2) pro-B cells. Hypoxanthine phosphoribosyltransferase (Hprt) expression was used to normalize the cDNA input and fibroblast-specific fibronectin 1 (Fn1) transcripts to control for the absence of stromal ST2 cell contamination (controls). Representative results of one of two pro-B cell pairs analyzed are shown. The annotation and alternative names of the Pax5repressed genes is provided in Table S1. (B) Flow cytometric analysis of the indicated cell surface proteins on cultured wild-type (dashed line) and Pax52/2 (black line) pro-B cells.

contamination. In contrast, four genes (Ccl9, Fc3r1g, Ccr2, and Lilrb4) were expressed at about a 50-fold higher level in myeloid Gr1+Mac1+ cells compared to CLPs. It is thus possible, although not very likely due to the double sorting procedure used, that the transcripts of these genes in the purified CLP population originate from a minor contamination of myeloid cells. Moreover, it is important to note that all four genes were also expressed in the contamination-free, in vitrocultured Pax52/2 pro-B cells (Figure 1). We conclude, therefore, that the CLPs, like the Pax52/2 pro-B cells, exhibit not only lymphoid but also myeloid promiscuity of gene expression.

Reversibility of Pax5-Mediated Gene Repression in Pro-B Cells We next investigated whether the Pax5-repressed genes are reactivated upon conditional Pax5 loss in

CreED-30 Pax5F/F pro-B cells expressing a Cre-ER fusion protein (Mikkola et al., 2002). In this induction system, the floxed (F) exon 2 of Pax5 is efficiently deleted upon Cre-ER activation, leading to a shift from functional Pax5 mRNA to the truncated transcript within 1 day of 4-hydroxytamoxifen (OHT) treatment (Figure 4A; controls). The Pax5 protein was reduced to low levels at day 4 (Mikkola et al., 2002), when the expression of the activated Pax5 target gene Cd19 started to be downregulated (Figure 4A). In parallel, the Pax5-repressed genes were reactivated in the same pro-B cells, thus providing unequivocal independent confirmation that the identified genes are negatively regulated by Pax5 (Figure 4A). Surprisingly, many genes were reactivated with similar kinetics as the J-chain (Igj) gene (Figure 4A), which is known to be a direct target of Pax5 repression (Rinkenberger et al., 1996). Together, these data indicate a continuous requirement of Pax5 for gene repression in committed pro-B cells.

Immunity 272

Figure 2. Expression of Pax5-Repressed Genes in Different Hematopoietic Lineages The mouse pro-B cell microarray was hybridized with a Cy5-labeled cDNA probe prepared from cultured Pax5 mutant (2/2) proB cells (blue color) in combination with a Cy3-labeled probe prepared from the following lineage-specific cells (Lin+; red color): two different wild-type (+/+) pro-B cell lines, bone marrow-derived macrophages (M), Thy1.2+ thymocytes (T), or bone marrow-derived erythroblasts (E). The hybridization data are depicted as expression ratios of the individual genes in the Lin+ cells relative to Pax52/2 pro-B cells according to the color scale shown.

Reactivation of Repressed Genes in Mature B Cells upon Pax5 Loss To study whether Pax5 is still required for gene repression in mature B cells, we took advantage of the Cd19cre line, which efficiently deletes the floxed Pax5 allele only in mature B cells leading to upregulation of CD25 (Horcher et al., 2001). We therefore sorted mature Pax5-deficient (Pax5D/2) B cells as CD25+IgM+ cells from the lymph nodes of Cd19-cre Pax5F/2 mice after depletion of non-B cells, whereas control Pax5D/+ B cells were isolated as CD252IgM+ cells from Cd19-cre Pax5F/+ mice (Figure S3). RT-PCR analysis confirmed deletion of the floxed Pax5 allele in all sorted Pax5D/2 B cells, leading to decreased expression of the activated Pax5 target gene Cd19 (Figure 4B; Horcher et al., 2001). Importantly, the loss of Pax5 led to efficient reactivation of the repressed genes Ccl3, Ccr5, Flt3, Emb, Fc3r1g, Cd28, and J-chain (Igj) in mature Pax5D/2 B cells, while the same genes were repressed in control Pax5D/+ B cells (Figure 4B). Moreover, these seven genes were expressed in Pax5D/2 B cells at a comparable or even higher level compared to Pax52/2 pro-B cells (Figure 4B) similar to the Csf1r (M-CSFR) gene (Tagoh et al., 2004). Hence, the repression of these lineage-inappropriate genes requires the continuous presence of Pax5 from the pro-B to the mature B cell stage. Interestingly, the Pax5 loss in mature B cells strongly activated the Prdm1 gene coding for the plasma cell transcription factor Blimp1 (Shapiro-Shelef et al., 2003). In contrast, Pax5 inactivation in either mature B cells (Figure 4B) or pro-B

cells (Figure 1) did not affect expression of the transcription factor genes Bcl6 and Xbp1, which are essential for germinal center B cell and plasma cell development, respectively (Dent et al., 1997; Reimold et al., 2001). This genetic evidence thus suggests a role for Pax5 in Prdm1 repression, while the regulation of Bcl6 and Xbp1 does not depend on Pax5 function. Expression of Pax5-Repressed Genes in Plasma Cells Under physiological conditions, Pax5 expression is downregulated during antigen-driven development of mature B cells into immunoglobulin-secreting plasma cells (Busslinger, 2004). To investigate whether the Pax5-repressed genes are reactivated during terminal plasma cell differentiation, we stimulated B220+CD19+ B cells from the spleen of Vav-bcl2 transgenic mice with the polyclonal activator LPS for up to 9 days (Figure 5A). LPS stimulation led to the downregulation of Pax5 expression within 3 days. Starting at day 3, the plasma cell-specific genes Cd138 (Syndecan-1) and Xbp1 were activated together with a multitude of Pax5repressed genes including Prdm1 (Blimp1) (Figure 5A; Figure S4). Hence, the physiological loss of Pax5 during terminal differentiation leads to reactivation of lineagepromiscuous genes in plasma cells. We next isolated in vivo plasma cells from heterozygous IgHGFP/+ mice, which express the green fluorescent protein (GFP) gene under the control of the endogenous immunoglobulin heavy-chain (IgH) locus (Figure S5A).

Multilineage Gene Repression in B Cells 273

Figure 3. Coexpression of Pax5-Repressed Myeloid and T Lymphoid Genes in CLPs The expression of the indicated Pax5-repressed genes was analyzed by RT-PCR in the various hematopoietic cell types shown. The cDNA input of each sorted cell population was normalized for equal expression of the small ribosomal protein gene S16 prior to PCR analysis of 5-fold serial dilutions of the cDNA. The different cell types were isolated by FACS sorting (Figure S2) as described in Experimental Procedures.

IgHGFP/+ mice were immunized by intraperitoneal injection of sheep red blood cells, and plasma cells were sorted 14 days later as a homogeneous population of Lin2GFPhiCD138hi cells from the bone marrow and spleen (Figure 5C; Figures S5B and S5C). These plasma cells expressed Pax5 at an w50-fold lower level compared to splenic B cells (Figure 5B). Interestingly, the Pax5-repressed genes, which were activated during LPS-induced differentiation of B cells, were also expressed in the sorted plasma cells of both origins (Figure 5B). Transcripts of the Prdm1 (Blimp1), J-chain (Igj), and CD28 genes were even more abundant in plasma cells compared to Pax52/2 pro-B cells, whereas Emb, Itgal, Ly6a, Notch1, Fc3r1g, Serpinb6b, and Myl4 were expressed at similar levels in both cell types. Flow cytometric analysis furthermore revealed high expression of CD28, Sca1 (Ly6a), and CD11a (Itgal) on bone marrow plasma cells (GFPhiCD138hi, back line) compared to splenic B cells (dashed line) in IgHGFP/+ mice (Figure 5C). We conclude, therefore, that the downregulation of Pax5 expression during terminal differentiation contributes to the plasma cell transcription program by facilitating the reactivation of previously repressed genes. Reactivation of Cd28 and Ccr2 Is Essential for Normal Plasma Cell Function We next investigated whether reactivation of the Pax5repressed genes Cd28, Ccr2, and Ccl3 is important for

plasma cell function. The costimulatory receptor CD28 of T cells enhances signaling through the T cell receptor by binding to the ligands CD80 (B7.1) and CD86 (B7.2) on activated B cells, which is essential for germinal center formation in lymphoid organs (Shahinian et al., 1993; Ferguson et al., 1996). The chemokine receptor CCR2 (binding the ligand CCL2 [MCP-1]) and the chemokine CCL3 (MIP-1a; binding to the receptor CCR5) both mediate macrophage chemotaxis and regulate effector T cell function (Kuziel et al., 1997; Trifilo et al., 2003). As these three genes are expressed in multiple hematopoietic cell types including T cells, we used competitive bone marrow reconstitution to generate mice that specifically lack one of the three genes only within the B-lymphoid lineage. To this end, we transplanted lethally irradiated Rag22/2 mice with a 1:1 mixture of bone marrow from Vav-cre Pax5F/F mice and from Cd282/2, Ccr22/2, Ccl32/2or wild-type mice (Figure 6A). As Vav-cre Pax5F/F mice lack B lymphocytes due to hematopoietic-specific deletion of the B lymphoid gene Pax5 (Figure 6B; Figure S7C), all B cells and plasma cells in the reconstituted Rag22/2 mice can develop only from the transplanted Cd282/2, Ccr22/2, Ccl32/2 or wild-type HSCs. In contrast, all T cells in the Cd282/2 reconstituted mice expressed the costimulatory receptor CD28, which thus provided a selective advantage for the development of T cells derived from the Vav-cre Pax5F/F HSCs (Figure 6C). Six weeks after transplantation, the

Immunity 274

Figure 4. Reversibility of Pax5-Mediated Gene Repression in Pro-B and Mature B Cells (A) Reactivation of repressed genes upon Pax5 loss in committed pro-B cells. Transcripts of the indicated genes were analyzed by RT-PCR at different days after 4-hydroxytamoxifen (OHT) treatment of pro-B cells from CreED-30 Pax5F/F mice. PCR products corresponding to the fulllength (FL) Pax5 mRNA or truncated Pax5 transcript lacking exon 2 (DE2) are indicated. (B) Reversibility of Pax5-mediated gene repression in mature B cells. Mature control or Pax5-deficient B cells were isolated free of contaminating non-B cells from the lymph nodes of Cd19-cre Pax5F/+ mice (abbreviated as Pax5D/+) or Cd19-cre Pax5F/2 (Pax5D/2) mice, respectively (Figure S3). Both cell types were analyzed for expression of the indicated transcripts by RT-PCR of 5-fold serial cDNA dilutions.

reconstituted mice were subcutaneously injected with the T cell-dependent antigen NIP-Ovalbumin in complete Freund’s adjuvant, and the spleen and serum were analyzed 14 days later (Figure 6A). Germinal center B cells (PNA+B220+) were detected by flow cytometry (Figure 6D), and immunohistochemistry identified abundant germinal centers in the spleen of the reconstituted mice (Figure 6E). ELISA measurements revealed comparable levels of total IgG1 in the serum of the different reconstituted mice (Figure 7F). Moreover, the titers of NIPspecific IgG1 were also similar in the mice reconstituted with wild-type or Ccl32/2 bone marrow (Figure 7F), indicating that CCL3 is not essential for plasma cell function despite expression of its receptor CCR5 (Figure 5B). In contrast, mice reconstituted with Cd282/2 or Ccr22/2

bone marrow had on average a 10-fold lower level of NIP-specific IgG1. These data therefore identified a B-lineage-intrinsic role for CD28 and CCR2 in controlling the differentiation or function of plasma cells beyond germinal center formation. Ccl3 Expression in B Cells Causes Increased Osteoclast Formation and Bone Loss The Pax5-repressed chemokine gene Ccl3 (MIP-1a) and the Tnfsf11 gene coding for the osteoprotegerin ligand (OPGL, RANKL) are both known to promote osteoclast formation and thus bone destruction (Kong et al., 1999; Oyajobi et al., 2003). The Pax5-dependent repression of Ccl3 (MIP-1a) and/or Tnfsf11 (OPGL) transcription in B cells may therefore be essential to prevent

Multilineage Gene Repression in B Cells 275

Figure 5. Expression of Pax5-Repressed Genes in Plasma Cells (A) Reactivation of Pax5-repressed genes during LPS-induced plasma cell differentiation. The indicated transcripts were analyzed by RT-PCR at different days after in vitro LPS treatment of sorted Lin2B220+CD19+ splenocytes from Vav-bcl2 transgenic mice. (B) Expression of Pax5-repressed genes in plasma cells. IgHGFP/+ mice (Figure S5A) were immunized with sheep red blood cells. After 14 days, plasma cells were sorted from the bone marrow and spleen as Lin2GFPhi CD138hi cells (Figures S5B and S5C) followed by RT-PCR analysis of 5-fold serial cDNA dilutions. Splenic B cells were sorted as Lin2B220+CD19+ cells. (C) Cell surface expression. The expression of CD28, Sca1, and CD11a was analyzed by flow cytometry on GFPhiCD138hi plasma cells (black lines) from the bone marrow of immunized IgHGFP/+ mice and compared to the corresponding expression of splenic B220+ CD19+ B cells (dashed lines).

excessive osteoclast formation leading to imbalanced hematopoiesis. To test this hypothesis, we used a transgenic strategy to maintain Ccl3 expression in all B lymphocytes, where the endogenous Ccl3 gene is normally repressed by Pax5 (Figures 1, 3, and 4). To this end, we inserted a mouse Ccl3 minigene together with a human (h) Cd2 indicator gene into the Cd19 locus (Figure 7A), which is exclusively expressed in B lymphocytes under the stringent control of Pax5 (Nutt et al., 1998). The Cd19Ccl3 BAC as well as a control Cd19hCd2 BAC, containing only the hCd2 gene, were injected into oocyte pronuclei to generate transgenic mice. At an expected frequency, we obtained Cd19hCd2 transgenic mice that expressed the hCd2 gene in all B lymphocytes but not in myeloid, erythroid, or T lymphoid cells, thus validating our transgenic strategy (Figure S6). Despite repeated attempts, we obtained, however, only one weakly expressing Cd19Ccl3 transgenic line, suggesting that constitutive Ccl3 expression in B cells may interfere with the survival to adulthood (data not shown). Hence, we next analyzed transgenic founder (G0) embryos at day

18.5 of gestation. Two embryos expressed the Cd19Ccl3 transgene in all B lymphocytes of the fetal liver (Figure 7B) at an average mRNA level that was comparable to that of Pax52/2 bone marrow pro-B cells (Figure 7C). Histological analysis uncovered a loss of trabecular bone in the long bones of the Cd19Ccl3 transgenic embryos compared to control littermates, which was best visualized by van Kossa staining of the mineralized bone areas (Figure 7D). The observed bone loss was accompanied by a corresponding increase in tartrate-resistant acid phosphatase (TRAP)-positive osteoclasts in Cd19Ccl3 transgenic relative to wild-type embryos (Figure 7D). Increased osteoclast formation was further demonstrated by abundant expression of the osteoclast-specific cathepsin K (Ctsk) gene (Rantakokko et al., 1996) in Cd19Ccl3 transgenic bones in contrast to wild-type bones. Hence, normal osteoclast development and thus homeostasis of the hematopoietic system depends on the repression of Ccl3 in B cells, which identifies a critical role for the repression function of Pax5.

Immunity 276

Figure 6. The Reactivation of Pax5Repressed Genes Is Important for Plasma Cell Function (A) Schematic diagram describing the bone marrow reconstitution and immunization of Rag22/2 mice. The T cell-dependent antigen NIP-Ovalbumin (NIP-Ova) was subcutaneously injected in complete Freund’s adjuvant. The bone marrow (BM) of Vav-cre Pax5F/F mice is referred to as Pax5D/D BM. (B) Absence of B cells (CD19+B220+) in the bone marrow of Vav-cre Pax5F/F mice in contrast to Cd282/2 mice. (C) Uniform expression of CD28 on thymocytes of Rag22/2 mice after competitive reconstitution with Cd282/2 and Pax5D/D bone marrow (Cd282/2 rec; back surface). (D) Presence of germinal center B cells (PNA+B220+) in the spleen of immunized mice, which were reconstituted (rec) with bone marrow of the indicated genotypes and Pax5D/D bone marrow. (E) Formation of normal germinal centers in the spleen of immunized reconstituted mice. Cryosections of the spleen were stained with FITC-anti-IgD antibody (blue) and biotinylated peanut hemagglutinin (PNA, brown) followed by detection with an alkaline phosphatase-coupled anti-FITC antibody (visualized with Fast Blue) and horseradish peroxidaseconjugated streptavidin (visualized with DAB), respectively. (F) Impaired immune response in mice with Cd282/2 and Ccr22/2 B-lineage cells. The serum titers of total IgG1 and NIP-specific IgG1 were determined by ELISA 14 days after immunizing reconstituted mice of the indicated genotypes.

Pax5 mutant mice were recently shown to exhibit early-onset osteopenia due to increased osteoclast development (Horowitz et al., 2004). We could confirm

a massive loss of trabecular bone in the long bones of 12-day-old Pax52/2 mice (Figure S7A). Surprisingly, however, the formation of trabecular bone appeared to

Multilineage Gene Repression in B Cells 277

Figure 7. Ectopic Ccl3 Expression in B Cells Induces Osteoclast Formation and Bone Loss (A) Cd19Ccl3 transgene. The mouse Ccl3 cDNA, which was linked via IRES to a human (h) intracellularly truncated (t) Cd2 gene, was inserted into exon 1 of a Cd19 BAC. pA, polyadenylation site. (B) B cell-specific expression of the Cd19Ccl3 transgene. Fetal liver cells were isolated from a transgenic (tg) founder (G0) embryo (black line) and wild-type (wt) littermate (dashed line) at day 18.5 and stained with lineage-specific antibodies. The expression of hCD2 is displayed for B lymphoid (CD19+), myeloid (Mac1+Gr1+), and erythroid (Ter119+) cells. (C) Ccl3 mRNA levels. Ccl3 expression was compared by RT-PCR in sorted CD19+ fetal liver (FL) cells and cultured pro-B cells of the indicated genotypes. (D) Bone phenotype of Cd19Ccl3 transgenic embryos. Adjacent sections through the hindlimb (tibia) of transgenic and wild-type E18.5 embryos were analyzed by hematoxylin and eosin (HE), van Kossa, and TRAP staining and by in situ hybridization analysis of cathepsin K (Ctsk) expression. Van Kossa staining visualizes the mineralized trabecular and cortical bones in black color. Arrowheads point to TRAP-positive (red) and cathepsin K-expressing (blue) osteoclasts. Dashed lines indicate the contours of the bone. The same bone phenotype was observed with a second independent Cd19Ccl3 transgenic embryo.

be normal in Vav-cre Pax5F/F mice, indicating that hematopoietic-specific deletion of Pax5 by Vav-cre does not lead to increased osteoclast development (Figure S7B). The difference in bone phenotype between the Vav-cre Pax5F/F and Cd19Ccl3 mice is likely explained by the low number of Pax5-deficient pro-B cells in the Vavcre Pax5F/F bone marrow, which produce little CCL3 compared to all the abundant B lymphocytes of the Cd19Ccl3 mouse (Figure S7C). Hence, the severe osteopenia observed in Pax52/2 mice is not the result of Pax5 deficiency in the hematopoietic system but instead must be caused by the lack of another, nonhematopoietic function of Pax5. Discussion B cell commitment critically depends on the transcription factor Pax5 (BSAP), which activates B cell-specific

genes and simultaneously represses lineage-inappropriate genes at the onset of B cell development (Busslinger, 2004). Here we have identified 110 Pax5-repressed genes by cDNA microarray analysis, which provided important insight into the repression function of Pax5 during B lymphopoiesis as discussed below. Changes in Cell Signaling, Adhesion, and Migration at B Cell Commitment The Pax5-repressed genes fulfill diverse functions in different hematopoietic cell types, indicating that Pax5 downregulates a broad range of biological activities including cell-cell communication, cell adhesion, migration, nuclear processes, and cellular metabolism at B cell commitment. A major class of repressed genes codes for secreted proteins, transmembrane receptors, and intracellular signal transducers. The cytokine OPGL (RANKL; encoded by Tnfsf11) controls osteoclast development

Immunity 278

and lymphopoiesis (Kong et al., 1999), while the insulinlike growth factor 2 (Igf2) regulates the proliferation of HSCs and erythromyeloid progenitors (Zhang and Lodish, 2004). Several Pax5-repressed genes code for cell surface receptors such as the tyrosine kinase receptor Flt3 controlling the proliferation and differentiation of multipotent progenitors (MPPs) and CLPs (Stirewalt and Radich, 2003); the GPI-anchored Ly6a (Sca1) protein regulating HSC self-renewal, myeloid progenitor development, and T cell proliferation (Ito et al., 2003); the costimulatory receptor CD28 of T cells (Shahinian et al., 1993); the signal-transducing common g chain (FcRg; encoded by Fc3r1g) of activating Fc receptors on myeloid cells (Takai, 2002); the subunit Ramp1 of the calcitonin gene-related peptide (CGRP) receptor implicated in myeloid progenitor development (Broome et al., 2000); and the inhibitory receptor Gp49b (encoded by Lilrb4) of NK and mast cells (Feldweg et al., 2003). Intracellular signaling molecules encoded by Pax5-repressed genes include the tyrosine kinase Lck (Molina et al., 1992), the catalytic subunit Aa (encoded by Ppp3ca) of the calcineurin phosphatase (Zhang et al., 1996), and the adaptor molecule Grap2 (Mona, Gads; Yoder et al., 2001), which are all involved in transducing signals from the (pre)TCR in T cells. The transmembrane adaptor protein NTAL/ LAB (encoded by Wbscr5) mediates signaling of the Fcg and Fc3 receptors in myeloid cells (Brdicka et al., 2002), while the b spectrin Spnb2 (ELF) functions as an adaptor in TGFb signaling (Tang et al., 2003). In summary, the Pax5-dependent repression of these signaling molecules unequivocally demonstrates that Pax5 downregulates multiple signal transduction pathways in addition to the previously identified M-CSFR (Csf1R) and Notch signaling systems (Nutt et al., 1999; Souabni et al., 2002) and thus renders committed B lymphocytes unresponsive to lineage-inappropriate signals. Lymphoid progenitors undergo a dramatic change in cell migration and adhesion at B-lineage commitment. The committed pro-B cells (Flt32B220+CD19+c-Kit+) adhere to a specialized bone marrow niche consisting of IL-7-expressing CXCL122 stromal cells (Tokoyoda et al., 2004). In contrast, uncommitted pre-pro-B cells, which have the same cell surface phenotype (Flt3+ B220+CD192) and developmental potential as Pax52/2 pro-B cells (Balciunaite et al., 2005), are located in a separate bone marrow environment composed of VCAM-1+ IL-72 stromal cells expressing the chemokine CXCL12 (SDF1; Tokoyoda et al., 2004). Interestingly, CXCL12 induces an efficient chemotactic response in uncommitted pre-pro-B cells (Tokoyoda et al., 2004) and Pax52/2 pro-B cells (Figure S8B), indicating that this chemokine retains uncommitted lymphoid progenitors in the CXCL12+IL-72 stromal cell niche. The CXCR4 receptor for CXCL12 is, however, equally expressed in lymphoid progenitors and committed pro-B cells (Tokoyoda et al., 2004; Figure S8A), suggesting that CXCL12-induced changes in integrin signaling may account for the altered migration and adhesion behavior of lymphocytes at B cell commitment. The chemokine CXCL12 activates, by inside-out signaling, the high-affinity binding state of the integrin adhesion receptors VLA-4 (a4b1 binding to VCAM-1) and LFA-1 (aLb2 binding to ICAM-1; Constantin et al., 2000), which are both coexpressed on uncommitted pre-pro-B and Pax52/2 pro-B cells (To-

koyoda et al., 2004; Figure S8A). In contrast, LFA-1 expression is selectively downregulated on committed pro-B cells due to Pax5-dependent repression of the subunit aL integrin (Itgal, CD11a; Figure 1). Pax5repressed genes also code for the transmembrane protein CD47 (Brown and Frazier, 2001) and tetraspanin Tm4sf2 (Zemni et al., 2000), which regulate adhesiondependent signaling by interacting with integrin receptors, while Embigin (Emb) promotes cell adhesion in a yet unknown manner (Huang et al., 1993). In addition to the chemokine receptors CCR2 and CCR5, other Pax5-repressed regulators of cell migration and adhesion include the signal-transducing molecule Lsp1 (Harrison et al., 2004), the Rho guanine nucleotide exchange factor Vav3 (Gakidis et al., 2004), and the Rho GTPaseactivating protein Daam1 (Habas et al., 2001), which are involved in controlling the actin cytoskeleton. In summary, Pax5 alters the adhesion and migration properties of lymphoid progenitors by downregulating multiple components of chemokine and intergrin signaling pathways. Promiscuous Expression of Myeloid Genes in Lymphoid Progenitors The lineage priming hypothesis predicts that some transcriptional activation of lineage-specific differentiation programs precedes commitment to any particular hematopoietic lineage (Hu et al., 1997). Two previous reports demonstrated that MPPs (LSKs) coexpress myeloid and lymphoid genes, whereas differentiation to CMPs correlates with the downregulation of lymphoid genes and differentiation to CLPs with the repression of erythromyeloid genes (Miyamoto et al., 2002; Akashi et al., 2003). In view of these data, we were surprised about the expression of a multitude of myeloid genes in Pax52/2 pro-B cells, which we previously characterized as lymphoid progenitors (Heavey et al., 2003). Moreover, our analysis detected the expression of both lymphoid and myeloid Pax5-repressed genes in CLPs, while MPPs and CMPs predominantly express myeloid but not lymphoid Pax5-regulated genes. These data agree with the conclusion of Miyamoto et al. (2002) that the promiscuous transcription of myeloid genes is initiated in HSCs and MPPs and that the expression of lymphoid genes is only activated following lymphoid commitment of MPPs. However, our results also demonstrate that CLPs exhibit not only lymphoid but also myeloid promiscuity that is repressed only at B cell commitment by Pax5. This finding is likely to account for the latent myeloid potential of CLPs (Kondo et al., 2000) and uncommitted pre-pro-B cells (Balciunaite et al., 2005). Pax52/2 progenitors express, however, very few erythroid genes, consistent with the finding that HSCs give rise to erythroid-megakaryocytic progenitors before commitment to the myeloid and lymphoid lineages (Adolfsson et al., 2005). Reactivation of Pax5-Repressed Genes in Plasma Cells Based on genetic studies of fly and mouse development, it is thought that transcription factors initiate cell fate decision by altering gene expression patterns, while the transcriptional state of committed cells is

Multilineage Gene Repression in B Cells 279

subsequently maintained by epigenetic factors encoded by the Polycomb and Trithorax group genes (Ringrose and Paro, 2004). In particular, the chromatin-modifying Polycomb group protein complexes are responsible for the epigenetic maintenance of the repressed state (Ringrose and Paro, 2004). It was therefore surprising to see that the B-lineage commitment factor Pax5 is continually required to repress lineage-inappropriate genes throughout B cell development, as the mere loss of Pax5 leads to reactivation of Pax5-repressed genes not only in pro-B cells but also in mature B cells. The observed transcriptional plasticity of B cells could also explain why the suppression of Pax5 activity by C/EBP transcription factors allows B cells to transdifferentiate into macrophages under myeloid cytokine conditions (Xie et al., 2004). Physiological stimulation by foreign antigen induces, however, mature B cells to undergo terminal differentiation to plasma cells, which results in a radical change of gene expression, as most B cell-specific genes including Pax5 are repressed, while a distinct plasma cell-specific program is activated by Blimp1 and Xbp1 (Shaffer et al., 2002; 2004). Here we have shown that several Pax5-repressed genes are reexpressed in plasma cells. Hence, the downregulation of Pax5 expression during terminal differentiation contributes to the plasma cell transcription program by facilitating the reactivation of lineage-promiscuous genes. J-chain (Igj), Cd28, and Ccr2 are three of the Pax5-repressed genes, which are reactivated in plasma cells. While J-chain is known to be essential for the assembly and secretion of pentameric IgM in plasma cells (Niles et al., 1995), we have now discovered a B-lineage-intrinsic function of the CD28 and CCR2 receptors in plasma cell differentiation by using a competitive bone marrow reconstitution assay. Mice, which specifically lack Cd28 or Ccr2 in the B lymphoid lineage, are able to form normal germinal centers upon immunization with NIP-Ova. However, the titers of NIP-specific IgG1 were 10-fold reduced in the serum of these mice compared to controls, indicating that CD28 and CCR2 are important for post-GC differentiation or function of plasma cells. Future experiments will address whether CD28 and CCR2 signaling controls terminal differentiation, cell survival, or efficient Ig secretion of plasma cells. Whereas CD28 expression in T cells is required to promote the B-T cell collaboration leading to germinal center formation (Shahinian et al., 1993; Ferguson et al., 1996), we have now identified a B-lineage-intrinsic role for CD28 in plasma cells. Using conditional Pax5 inactivation in mature B cells, we reinvestigated the role of Pax5 in controlling the expression of the GC-specific transcription factor Bcl6 and the plasma cell regulators Xbp1 and Prdm1 (Blimp1). This genetic approach did not reveal a role for Pax5 in the regulation of Bcl6 and Xbp1 and thus contradicts previous transient transfection data, which showed a 2-fold repression of a minimal Xbp1 promoter construct by Pax5 overexpression in HeLa and B cells (Reimold et al., 1996). Importantly however, the loss of Pax5 resulted in activation of Blimp1 in mature B cells. Hence, Pax5 appears to be involved in the repression of Blimp1 and the corresponding plasma cell transcription program, similar to the role of Bcl6 in GC B cells (Shaffer et al., 2000).

Pax5-Dependent Gene Repression in B Cells Is Essential for Blood Cell Homeostasis Pax5 also represses the chemokine gene Ccl3 throughout B cell development. CCL3 (MIP-1a) is not only involved in the recruitment of leukocytes to sites of inflammation but is also expressed in multiple myeloma cells, where it is responsible, as a potent osteoclastogenic factor, for bone destruction in tumor patients (Oyajobi et al., 2003). To test whether the Pax5 repression function is essential for normal hematopoiesis, we maintained Ccl3 expression from the Cd19 locus throughout B cell development, which resulted in an osteopenic bone phenotype due to increased osteoclast formation. The loss of trabecular bone in Cd19Ccl3 transgenic mice is likely to affect the function of HSCs in the bone marrow, as specialized osteoblasts lining the trabecular bone surface act as key components of the HSC niche (Suda et al., 2005). We conclude, therefore, that normal hematopoiesis and bone formation depend on the B cell-specific repression of Ccl3 by Pax5. Experimental Procedures Detailed methods can be found in Supplemental Data. Mice Wild-type and mutant mice were maintained on the C57BL/6 background. Pro-B Cell Culture Pro-B cells were cultured on g-irradiated ST2 cells in IL-7 containing IMDM medium (Nutt et al., 1997). cDNA Microarray Analysis A pro-B cell microarray was created by spotting 10,000 cDNA inserts from libraries that were generated by bidirectional subtraction of the transcripts between Pax5+/+ and Pax52/2 pro-B cells using the PCRSelect cDNA Subtraction kit (Clontech). This microarray was hybridized with Cy3- and Cy5-labeled probes, scanned, and evaluated as described (Schebesta et al., 2002). Semiquantitative RT-PCR Analysis cDNA was prepared from different hematopoietic cell types, and semiquantitative PCR (Mikkola et al., 2002) was performed using the primers listed in Table S2. Conditional Pax5 Inactivation in Pro-B Cells Cultured pro-B cells from CreED-30 Pax5F/F mice were treated with 1 mM 4-hydroxytamoxifen (OHT) and withdrawn at daily intervals prior to semiquantitative RT-PCR analysis (Mikkola et al., 2002). FACS Sorting and Analysis Antibodies were obtained from PharMingen. The different cell types were isolated by two consecutive rounds of FACS sorting (Figure S2) as follows: LSK MPPs (Lin2IL-7Ra2Sca1hic-Kithi), CMP (Lin2Ly6C2 Sca12c-KithiCD31+), CLP (Lin2IL-7Ra+Sca1loc-Kitlo), pro-B (Lin2 CD19+c-Kit+), pre-B (Lin2CD19+CD25+IgM2), immature B cells (Lin2CD19+IgM+IgD2), granulocytes (Mac1+Gr1+), and thymocytes (Thy1.2+CD192). FACS reanalysis revealed a >99% purity of the double-sorted cell populations (Figure S2). Purification of Mature Pax5-Deficient B Cells Lymph node cells were depleted of non-B cells with anti-CD33, CD4, CD8a, CD11c, CD43, DX5, Gr1, Ly6C, Mac1, and Ter119 antibodies. The mature B cells of the Cd19-cre Pax5F/+ genotype were sorted as Lin2CD252IgM+ B cells and Pax5-deficient mature B cells of the Cd19-cre Pax5F/2 genotype as Lin2CD25+IgM+ B cells (Figure S3).

Immunity 280

LPS Stimulation of Splenic B Cells Sorted Lin2B220+CD19+ splenocytes of Vav-bcl2 mice were treated for up to 9 days with 10 mg/ml LPS in IMDM medium containing 10% fetal calf serum. Isolation of Plasma Cells IgHGFP/+ mice were immunized by intraperitoneal injection of 108 sheep red blood cells, and, 14 days later, plasma cells were sorted from the bone marrow and spleen as Lin2GFPhiCD138hi cells (Figure S5). Immunization and Determination of Ig Titers Lethally irradiated Rag22/2 mice were transplanted with a 1:1 mixture of 106 bone marrow cells from Vav-cre Pax5F/F mice and Cd282/2, Ccr22/2, Ccl32/2, or wild-type mice. Six weeks later, the reconstituted mice were immunized by subcutaneous injection of 50 mg NIP-Ovalbumin in complete Freund’s adjuvant. The titers of total and NIP-specific serum IgG1 were determined by ELISA 14 days after immunization. Generation of the Cd19Ccl3 Transgene Mouse Ccl3 cDNA was inserted together with an IRES-hCd2t gene into the mouse Cd19 BAC RP23-407K24 by standard BAC modification (Schebesta et al., 2002). Transgenic mice were generated by injection of supercoiled BAC DNA at w1 ng/ml into pronuclei of C57BL/ 6xCBA F1 zygotes. Histological Analysis and In Situ Hybridization Long bones of E18.5 embryos were fixed, decalcified, and sectioned prior to hematoxylin and eosin or TRAP staining. Decalcification was omitted for van Kossa staining. Bone sections were in situ hybridized with a digoxygenin-labeled cathepsin K RNA probe. Cryosections of the spleen were stained with a FITC-anti-IgD antibody and biotinylated PNA followed by detection with an alkaline phosphatase-coupled anti-FITC antibody (visualized with Fast Blue) and with horseradish peroxidase-conjugated streptavidin (visualized with DAB). Supplemental Data Supplemental Data include Supplemental Experimental Procedures, Supplemental References, eight figures, and two tables and can be found with this article online at http://www.immunity.com/cgi/ content/full/24/3/269/DC1/. Acknowledgments We thank P. Steinlein for assistance with cDNA microarray analysis, M. Schebesta for help with cDNA library construction, C. Hartmann for advice on osteoclast experiments, D. Tarlinton for expert help with ELISA measurements, and W. Kuziel for providing Ccr22/2 mice. This research was supported by Boehringer Ingelheim, by the Austrian Industrial Research Promotion Fund, by the EU FP6 funding for the Epigenome Network of Excellence, and by the Austrian GEN-AU initiative (financed by the Bundesministerium fu¨r Bildung, Wissenchaft and Kultur). Received: March 10, 2005 Revised: January 17, 2006 Accepted: January 20, 2006 Published: March 21, 2006 References Adolfsson, J., Ma˚nsson, R., Buza-Vidas, N., Hultquist, A., Liuba, K., Jensen, C.T., Bryder, D., Yang, L., Borge, O.-J., Thoren, L.A.M., et al. (2005). Identification of Flt3+ lympho-myeloid stem cells lacking erythro-megakaryocytic potential: a revised road map for adult blood lineage commitment. Cell 121, 295–306. Akashi, K., He, X., Chen, J., Iwasaki, H., Niu, C., Steenhard, B., Zhang, J., Haug, J., and Li, L. (2003). Transcriptional accessibility for genes of multiple tissues and hematopoietic lineages is hierarchically controlled during early hematopoiesis. Blood 101, 383–389.

Balciunaite, G., Ceredig, R., Massa, S., and Rolink, A.G. (2005). A B220+ CD117+ CD192 hematopoietic progenitor with potent lymphoid and myeloid developmental potential. Eur. J. Immunol. 35, 2019–2030. Brdicka, T., Imrich, M., Angelisova, P., Brdickova, N., Horvath, O., Spicka, J., Hilgert, I., Luskova, P., Draber, P., Novak, P., et al. (2002). Non-T cell activation linker (NTAL): a transmembrane adaptor protein involved in immunoreceptor signaling. J. Exp. Med. 196, 1617–1626. Broome, C.S., Whetton, A.D., and Miyan, J.A. (2000). Neuropeptide control of bone marrow neutrophil production is mediated by both direct and indirect effects on CFU-GM. Br. J. Haematol. 108, 140– 150. Brown, E.J., and Frazier, W.A. (2001). Integrin-associated protein (CD47) and its ligands. Trends Cell Biol. 11, 130–135. Busslinger, M. (2004). Transcriptional control of early B cell development. Annu. Rev. Immunol. 22, 55–79. Constantin, G., Majeed, M., Giagulli, C., Piccio, L., Kim, J.Y., Butcher, E.C., and Laudanna, C. (2000). Chemokines trigger immediate b2 integrin affinity and mobility changes: differential regulation and roles in lymphocyte arrest under flow. Immunity 13, 759–769. Dent, A.L., Shaffer, A.L., Yu, X., Allman, D., and Staudt, L.M. (1997). Control of inflammation, cytokine expression, and germinal center formation by BCL-6. Science 276, 589–592. Feldweg, A.M., Friend, D.S., Zhou, J.S., Kanaoka, Y., Daheshia, M., Li, L., Austen, K.F., and Katz, H.R. (2003). gp49B1 suppresses stem cell factor-induced mast cell activation-secretion and attendant inflammation in vivo. Eur. J. Immunol. 33, 2262–2268. Ferguson, S.E., Han, S., Kelsoe, G., and Thompson, C.B. (1996). CD28 is required for germinal center formation. J. Immunol. 156, 4576–4581. Gakidis, M.A., Cullere, X., Olson, T., Wilsbacher, J.L., Zhang, B., Moores, S.L., Ley, K., Swat, W., Mayadas, T., and Brugge, J.S. (2004). Vav GEFs are required for b2 integrin-dependent functions of neutrophils. J. Cell Biol. 166, 273–282. Habas, R., Kato, Y., and He, X. (2001). Wnt/Frizzled activation of Rho regulates vertebrate gastrulation and requires a novel Formin homology protein Daam1. Cell 107, 843–854. Harrison, R.E., Sikorski, B.A., and Jongstra, J. (2004). Leukocytespecific protein 1 targets the ERK/MAP kinase scaffold protein KSR and MEK1 and ERK2 to the actin cytoskeleton. J. Cell Sci. 117, 2151–2157. Heavey, B., Charalambous, C., Cobaleda, C., and Busslinger, M. (2003). Myeloid lineage switch of Pax5 mutant but not wild-type B cell progenitors by C/EBPa and GATA factors. EMBO J. 22, 3887– 3897. Horcher, M., Souabni, A., and Busslinger, M. (2001). Pax5/BSAP maintains the identity of B cells in late B lymphopoiesis. Immunity 14, 779–790. Horowitz, M.C., Xi, Y., Pflugh, D.L., Hesslein, D.G., Schatz, D.G., Lorenzo, J.A., and Bothwell, A.L. (2004). Pax5-deficient mice exhibit early onset osteopenia with increased osteoclast progenitors. J. Immunol. 173, 6583–6591. Hu, M., Krause, D., Greaves, M., Sharkis, S., Dexter, M., Heyworth, C., and Enver, T. (1997). Multilineage gene expression precedes commitment in the hematopoietic system. Genes Dev. 11, 774–785. Huang, R.-P., Ozawa, M., Kadomatsu, K., and Muramatsu, T. (1993). Embigin, a member of the immunoglobulin superfamily expressed in embryonic cells, enhances cell-substratum adhesion. Dev. Biol. 155, 307–314. Ito, C.Y., Li, C.Y., Bernstein, A., Dick, J.E., and Stanford, W.L. (2003). Hematopoietic stem cell and progenitor defects in Sca-1/Ly-6A-null mice. Blood 101, 517–523. Kondo, M., Scherer, D.C., Miyamoto, T., King, A.G., Akashi, K., Sugamura, K., and Weissman, I.L. (2000). Cell-fate conversion of lymphoid-committed progenitors by instructive actions of cytokines. Nature 407, 383–386. Kong, Y.-Y., Yoshida, H., Sarosi, I., Tan, H.-L., Timms, E., Capparelli, C., Morony, S., Oliveira-dos-Santos, A.J., Van, G., Itie, A., et al.

Multilineage Gene Repression in B Cells 281

(1999). OPGL is a key regulator of osteoclastogenesis, lymphocyte development and lymph-node organogenesis. Nature 397, 315–323. Kuziel, W.A., Morgan, S.J., Dawson, T.C., Griffin, S., Smithies, O., Ley, K., and Maeda, N. (1997). Severe reduction in leukocyte adhesion and monocyte extravasation in mice deficient in CC chemokine receptor 2. Proc. Natl. Acad. Sci. USA 94, 12053–12058. Mikkola, I., Heavey, B., Horcher, M., and Busslinger, M. (2002). Reversion of B cell commitment upon loss of Pax5 expression. Science 297, 110–113. Miyamoto, T., Iwasaki, H., Reizis, B., Ye, M., Graf, T., Weissman, I.L., and Akashi, K. (2002). Myeloid or lymphoid promiscuity as a critical step in hematopoietic lineage commitment. Dev. Cell 3, 137–147. Molina, T.J., Kishihara, K., Siderovski, D.P., van Ewijk, W., Narendran, A., Timms, E., Wakeham, A., Paige, C.J., Hartmann, K.U., Veillette, A., and Mak, T.W. (1992). Profound block in thymocyte development in mice lacking p56lck. Nature 357, 161–164. Niles, M.J., Matsuuchi, L., and Koshland, M.E. (1995). Polymer IgM assembly and secretion in lymphoid and nonlymphoid cell lines: evidence that J chain is required for pentamer IgM synthesis. Proc. Natl. Acad. Sci. USA 92, 2884–2888. Nutt, S.L., Urba´nek, P., Rolink, A., and Busslinger, M. (1997). Essential functions of Pax5 (BSAP) in pro-B cell development: difference between fetal and adult B lymphopoiesis and reduced V-to-DJ recombination at the IgH locus. Genes Dev. 11, 476–491. Nutt, S.L., Morrison, A.M., Do¨rfler, P., Rolink, A., and Busslinger, M. (1998). Identification of BSAP (Pax-5) target genes in early B-cell development by loss- and gain-of-function experiments. EMBO J. 17, 2319–2333. Nutt, S.L., Heavey, B., Rolink, A.G., and Busslinger, M. (1999). Commitment to the B-lymphoid lineage depends on the transcription factor Pax5. Nature 401, 556–562. Oyajobi, B.O., Franchin, G., Williams, P.J., Pulkrabek, D., Gupta, A., Munoz, S., Grubbs, B., Zhao, M., Chen, D., Sherry, B., and Mundy, G.R. (2003). Dual effects of macrophage inflammatory protein-1a on osteolysis and tumor burden in the murine 5TGM1 model of myeloma bone disease. Blood 102, 311–319. Rantakokko, J., Aro, H.T., Savontaus, M., and Vuorio, E. (1996). Mouse cathepsin K: cDNA cloning and predominant expression of the gene in osteoclasts, and in some hypertrophying chondrocytes during mouse development. FEBS Lett. 393, 307–313. Reimold, A.M., Ronath, P.D., Li, Y.-S., Hardy, R.R., David, C.S., Strominger, J.L., and Glimcher, L.H. (1996). Transcription factor B cell lineage-specific activator protein regulates the gene for human X-box binding protein 1. J. Exp. Med. 183, 393–401. Reimold, A.M., Iwakoshi, N.N., Manis, J., Vallabhajosyula, P., Szomolanyl-Tsuda, E., Gravallese, E.M., Friend, D., Grusby, M.J., Alt, F., and Glimcher, L.H. (2001). Plasma cell differentiation requires the transcription factor XBP-1. Nature 412, 300–307. Ringrose, L., and Paro, R. (2004). Epigenetic regulation of cellular memory by the Polycomb and Trithorax group proteins. Annu. Rev. Genet. 38, 413–443. Rinkenberger, J.L., Wallin, J.J., Johnson, K.W., and Koshland, M.E. (1996). An interleukin-2 signal relieves BSAP (Pax5)-mediated repression of the immunoglobulin J chain gene. Immunity 5, 377–386. Rolink, A.G., Nutt, S.L., Melchers, F., and Busslinger, M. (1999). Long-term in vivo reconstitution of T-cell development by Pax5deficient B-cell progenitors. Nature 401, 603–606. Schebesta, M., Pfeffer, P.L., and Busslinger, M. (2002). Control of pre-BCR signaling by Pax5-dependent activation of the BLNK gene. Immunity 17, 473–485. Shaffer, A.L., Yu, X., He, Y., Boldrick, J., Chan, E.P., and Staudt, L.M. (2000). BCL-6 represses genes that function in lymphocyte differentiation, inflammation, and cell cycle control. Immunity 13, 199–212. Shaffer, A.L., Lin, K.-I., Kuo, T.C., Yu, X., Hurt, E.M., Rosenwald, A., Giltnane, J.M., Yang, L., Zhao, H., Calame, K., and Staudt, L.M. (2002). Blimp-1 orchestrates plasma cell differentiation by extinguishing the mature B cell gene expression program. Immunity 17, 51–62. Shaffer, A.L., Shapiro-Shelef, M., Iwakoshi, N.N., Lee, A.-H., Qian, S.-B., Zhao, H., Yu, X., Yang, L., Tan, B.K., Rosenwald, A., et al.

(2004). XBP1, downstream of Blimp-1, expands the secretory apparatus and other organelles and increases protein synthesis in plasma cell differentiation. Immunity 21, 81–93. Shahinian, A., Pfeffer, K., Lee, K.P., Ku¨ndig, T.M., Kishihara, K., Wakeham, A., Kawai, K., Ohashi, P.S., Thompson, C.B., and Mak, T.W. (1993). Differential T cell costimulatory requirements in CD28deficient mice. Science 261, 609–612. Shapiro-Shelef, M., Lin, K.-I., McHeyzer-Williams, L.J., Liao, J., McHeyzer-Williams, M.G., and Calame, K. (2003). Blimp-1 is required for the formation of immunoglobulin secreting plasma cells and pre-plasma memory B cells. Immunity 19, 607–620. Souabni, A., Cobaleda, C., Schebesta, M., and Busslinger, M. (2002). Pax5 promotes B lymphopoiesis and blocks T cell development by repressing Notch1. Immunity 17, 781–793. Stirewalt, D.L., and Radich, J.P. (2003). The role of FLT3 in haematopoietic malignancies. Nat. Rev. Cancer 3, 650–665. Suda, T., Arai, F., and Hirao, A. (2005). Hematopoietic stem cells and their niche. Trends Immunol. 26, 426–433. Tagoh, H., Schebesta, A., Lefevre, P., Wilson, N., Hume, D., Busslinger, M., and Bonifer, C. (2004). Epigenetic silencing of the c-fms locus during B-lymphopoiesis occurs in discrete steps and is reversible. EMBO J. 23, 4275–4285. Takai, T. (2002). Roles of Fc receptors in autoimmunity. Nat. Rev. Immunol. 2, 580–592. Tang, Y., Katuri, V., Dillner, A., Mishra, B., Deng, C.-X., and Mishra, L. (2003). Disruption of transforming growth factor-b signaling in ELF bspectrin-deficient mice. Science 299, 574–577. Tokoyoda, K., Egawa, T., Sugiyama, T., Choi, B.-I., and Nagasawa, T. (2004). Cellular niches controlling B lymphocyte behavior within bone marrow during development. Immunity 20, 707–718. Trifilo, M.J., Bergmann, C.C., Kuziel, W.A., and Lane, T.E. (2003). CC chemokine ligand 3 (CCL3) regulates CD8+-T-cell effector function and migration following viral infection. J. Virol. 77, 4004–4014. Xie, H., Ye, M., Feng, R., and Graf, T. (2004). Stepwise reprogramming of B cells into macrophages. Cell 117, 663–676. Yoder, J., Pham, C., Iizuka, Y.-M., Kanagawa, O., Liu, S.K., McGlade, J., and Cheng, A.M. (2001). Requirement for the SLP-76 adaptor GADS in T cell development. Science 291, 1987–1991. Zemni, R., Bienvenu, T., Vinet, M.C., Sefiani, A., Carrie, A., Billuart, P., McDonell, N., Couvert, P., Francis, F., Chafey, P., et al. (2000). A new gene involved in X-linked mental retardation identified by analysis of an X;2 balanced translocation. Nat. Genet. 24, 167–170. Zhang, C.C., and Lodish, H.F. (2004). Insulin-like growth factor 2 expressed in a novel fetal liver cell population is a growth factor for hematopoietic stem cells. Blood 103, 2513–2521. Zhang, B.W., Zimmer, G., Chen, J., Ladd, D., Li, E., Alt, F.W., Wiederrecht, G., Cryan, J., O’Neill, E.A., Seidman, C.E., et al. (1996). T cell responses in calcineurin Aa-deficient mice. J. Exp. Med. 183, 413– 420.