Characterization of Early Stages of Human B Cell Development by Expression Profiling

This information is current as Marit E. Hystad, June H. Myklebust, Trond H. Bø, Einar A. of September 25, 2021. Sivertsen, Edith Rian, Lise Forfang, Else Munthe, Andreas Rosenwald, Michael Chiorazzi, Inge Jonassen, Louis M. Staudt and Erlend B. Smeland J Immunol 2007; 179:3662-3671; ;

doi: 10.4049/jimmunol.179.6.3662 Downloaded from http://www.jimmunol.org/content/179/6/3662

Supplementary http://www.jimmunol.org/content/suppl/2008/03/12/179.6.3662.DC1 Material http://www.jimmunol.org/ References This article cites 59 articles, 28 of which you can access for free at: http://www.jimmunol.org/content/179/6/3662.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 25, 2021

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Characterization of Early Stages of Human B Cell Development by Profiling1

Marit E. Hystad,* June H. Myklebust,*ʈ Trond H. Bø,† Einar A. Sivertsen,* Edith Rian,* Lise Forfang,* Else Munthe,‡ Andreas Rosenwald,§ Michael Chiorazzi,¶ Inge Jonassen,†# Louis M. Staudt,¶ and Erlend B. Smeland2*ʈ

We have characterized several stages of normal human B cell development in adult bone marrow by gene expression profiling of hemopoietic stem cells, early B (E-B), pro-B, pre-B, and immature B cells, using RNA amplification and Lymphochip cDNA microarrays Hierarchical clustering of 758 differentially expressed clearly separated the five populations. We used gene sets to .(6 ؍ n) investigate the functional assignment of the differentially expressed genes. Genes involved in VDJ recombination as well as B lineage- associated transcription factors (TCF3 (E2A), EBF, BCL11A, and PAX5) were turned on in E-B cells, before acquisition of CD19. Several transcription factors with unknown roles in B lymphoid cells demonstrated interesting expression patterns, including ZCCHC7 and ZHX2. Downloaded from Compared with hemopoietic stem cells and pro-B cells, E-B cells had increased expression of 18 genes, and these included IGJ, IL1RAP, BCL2, and CD62L. In addition, E-B cells expressed T/NK lineage and myeloid-associated genes including CD2, NOTCH1, CD99, PECAM1, TNFSF13B, and MPO. Expression of key genes was confirmed at the level by FACS analysis. Several of these Ags were heterogeneously expressed, providing a basis for further subdivision of E-B cells. Altogether, these results provide new information regarding expression of genes in early stages of human B cell development. The Journal of Immunology, 2007, 179: 3662–3671. http://www.jimmunol.org/

arly human B cell development occurs in the bone mar- cific recombination enzymes RAG1, RAG2, and TdT (6). The ex- row (BM)3 as hemopoietic stem cells (HSC) develop via pression of pre-BCR, composed of IgH chains and surrogate L E various B lineage-restricted precursors into immature B chains (VpreB and ␭14.1), is a hallmark for the pre-B cell popu- (IM-B) cells, which then leave the BM and enter the periphery. lation. Signaling through pre-BCR promotes L chain (VJL) rear- The different stages can be identified by the expression of CD Ags rangement and allelic exclusion at the Ig H chain locus. Once VJL and the rearrangement status of Ig H and L chains, and the current rearrangements are successful, L chains are expressed and combine consensus is that B lineage-committed cells pass through a with H chains as well as Ig␣/Ig␤ to form a functional BCR expressed ϩ ϩ Ϫ

CD34 CD10 CD19 common lymphoid progenitors (CLP) early on IM-B cells. by guest on September 25, 2021 B (E-B) stage before they mature via CD34ϩCD19ϩCD10ϩ Early B cell development is controlled by a hierarchical regu- pro-B, CD34ϪCD19ϩ large pre-B I and II, and CD34ϪCD19ϩ latory network of transcription factors including PU.1, E2A, EBF, small pre-B II into CD34ϪCD19ϩsIgMϩ IM-B cells (1, 2). Galy et and Pax5 (7). E2A and EBF are required for the initiation of E-B al. (3) were the first to identify human CLP by demonstrating that cell development by regulating the expression of B lineage-spe- CD34highLinϪCD10ϩ cells had B, T, NK, and dendritic cell po- cific genes, including PAX5, which is essential for B cell com- tential. Later studies have indicated that this population is biased mitment maintenance (8, 9). Furthermore, development of lym- toward B lineage development and termed the population E-B cells phoid cells is supported by cytokines and interactions with stromal (4, 5). Rearrangements of the VDJ H chain locus are characteristic cells, as demonstrated by in vitro studies of human B cell devel- of pro-B cells and require the expression of the lymphocyte-spe- opment (10). Genome-wide gene expression profiling has been performed in murine precursor B cells (11, 12). Recently, van Zelm et al. (13) *Department of Immunology, Institute for Cancer Research, Rikshospitalet-Radium- hospitalet Medical Centre, Oslo, Norway; †Department of Informatics, University of investigated the Ig gene rearrangement steps in correlation to the Bergen, Bergen, Norway; ‡Institute of Pathology, Faculty Division Rikshospitalet, expression of developing B cells from the BM University of Oslo, Oslo, Norway; §Institute of Pathology, University of Wurzburg, Wurzburg, Germany; ¶Metabolism Branch, National Cancer Institute, Bethesda, MD of healthy children. However, whereas the later stages of human B 20892; and ʈCentre for Biomedicine, University of Oslo, Oslo, Norway; and #Com- lineage development are well-characterized, the transitions from putational Biology Unit, University of Bergen, Bergen, Norway HSC into B lineage-committed cells have been explored in less Received for publication May 21, 2007. Accepted for publication June 20, 2007. detail. The aim of this study was to characterize the earliest steps The costs of publication of this article were defrayed in part by the payment of page of human B cell development in adult BM by gene expression charges. This article must therefore be hereby marked advertisement in accordance profiling. Total RNA was extracted from HSC, E-B, pro-B, pre-B, with 18 U.S.C. Section 1734 solely to indicate this fact. and IM-B cell populations from human BM, amplified, and hy- 1 This work was supported by the Norwegian Cancer Society, the Norwegian Re- search Council, and the Functional Genomics (FUGE) Program of the Norwegian bridized to Lymphochip cDNA microarrays. Altogether, this study Research Council. provides new insight into the molecular processes that take place 2 Address correspondence and reprint requests to Dr. Erlend B. Smeland, Department of in early human B cell differentiation. Immunology, Institute of Cancer Research, Rikshospitalet-Radiumhospitalet Medical Centre, 0310 Oslo, Norway. E-mail address: [email protected] 3 Abbreviations used in this paper: BM, bone marrow; HSC, hemopoietic stem cell; Materials and Methods E-B, early B; IM-B, immature B; CLP, common lymphoid progenitors; MPO, my- Cell separation and FACS analysis eloperoxidase; FDR, false discovery rate; GO, ; GC, germinal center; MEF, myocyte enhancer factor; MLP, multilineage progenitor cells; ALL, acute lym- BM aspirates from healthy adult donors were obtained after informed con- phoblastic leukemia. sent and approval by the Ethics Committee of the Norwegian Radium www.jimmunol.org The Journal of Immunology 3663

Table I. List of Abs and fixation/permeabilization methods used in FACS analysis and in cell-sorting experiments

Used at Concentration Ab Company (Prod. No.) (␮g/ml)

IgM FITC DakoCytomation (F0203) 1/40 dilution CD10 PE-Cy7 BD Biosciences (341102) 0.50 CD10 allophycocyanin BD Biosciences (332777) 0.60 CD34 allophycocyanin BD Biosciences (345818) 2.50 CD34 PE BD Biosciences (345802) 2.50 CD38 FITC Beckman Coulter (A07778) 0.30 CD38 PC5 Beckman Coulter (A07780) 0.80 CD38 PE DakoCytomation (R7144) 2.50 CD19 PC5 Beckman Coulter(A07771) 1.25 CXCR4 PE BD Biosciences (555974) 1/10 dilution CD62L BD Biosciences (341012) 1.00 CD2 FITC DakoCytomation (F0767) 2.50 CD127 PE Immunotech (PN IM1980) 1/5 dilution J chaina (rabbit) Gift from P. O. Brandtzaeg 1.25 Preimunized rabbit seruma Gift from P. O. Brandtzaeg 1.25 TdT FITCb DakoCytomation (F7139) Bcl-2 PEb BD Biosciences (340576) MPO PEc DakoCytomation (R7209) Downloaded from IgG1 FITC DakoCytomation (X0927) IgG2a FITC DakoCytomation (X0933) IgG1 PE DakoCytomation (X0928) IgG2b PE DakoCytomation (X0951) IgG2a PE BD Biosciences (349053) IgG1 allophycocyanin BD Biosciences (345818)

IgG1 PECy7 BD Biosciences (348808) http://www.jimmunol.org/ IgG1 PC5 Beckman Coulter (A07798)

a The cells were fixed in 4% paraformaldehyde for 10 minutes at room temperature, washed once in PBS, and then perme- abilized in 0.1% Triton-X 100 for 30 min at room temperature, washed once in PBS, and incubated with 50% methanol for 10 min at room temperature, washed twice in PBS, and then stained with the intracellular Ab. The anti-J chain Ab was a gift from P. O. Brandtzaeg, Oslo, Norway. b The cells were fixed and permeabilized by incubating with 100% methanol for 10 min at room temperature, washed once in PBS, and then stained with anti-J-chain Ab. Mouse anti-rabbit Ig PE was used as secondary layer. c The Intrastain kit from DakoCytomation was used as described by the manufacturer. by guest on September 25, 2021 Hospital. CD34ϩ or CD10ϩ cells were isolated from BM by the use of (20). Mean expression profiles from spots representing the same sequence immunomagnetic beads (CD34 Dynabeads, or sheep anti-mouse IgG- accession number were joined by calculating the mean log ratio of these coated Dynabeads labeled with anti-CD10 mAb; Invitrogen Life Technol- spots. All subsequent analysis was based on these mean expression pro- ogies) as previously described (14, 15). CD38ϪCD34ϩ HSC and files. Genes displaying differential expression between populations were CD38ϩCD34ϩCD10ϩCD19Ϫ E-B cells were obtained by FACS sorting on detected using a two sample t test with pooled variance estimate. Genes a FACSDiVa cell sorter (BD Biosciences) after labeling of CD34ϩ BM were sorted after significance, and the significance threshold was selected cells with anti-CD38 FITC, anti-CD34 allophycocyanin, anti-CD10 so that the false discovery rate (FDR) for the reported genes was 10%. PECy7, and anti-CD19 PECy5. CD34ϩCD10ϩCD19ϩIgMϪ pro-B, CD34Ϫ Finally, only genes with at least a 2-fold change between stages were kept. CD10ϩCD19ϩIgMϪ pre-B and CD34ϪCD10ϩCD19ϩIgMϩ IM-B cells Hierarchical cluster analysis was performed by cluster and the results were were obtained by FACS sorting after labeling of CD10ϩ BM cells with visualized by TreeView (21). The expression level of each gene was - anti-CD10 PECy7, anti-CD19 PECy5, anti-IgM FITC, and anti-CD34 al- ative to its median expression level across all samples. The proliferation lophycocyanin (details in Table I). All cell populations were collected in signature was developed by Staudt and colleagues (18). The other gene sets 5-ml tubes, washed once in PBS, and subjected to RNA isolation. Detec- are based on information from Gene Ontology (GO; http://source.stan- tion of intracellular was performed after labeling CD34ϩ BM cells ford.edu/cgi-bin/source/sourceSearch). The GO was used to select, within with Abs used to identify the various subpopulations, followed by fixation the significant probe sets, for probe sets annotated with GO terms repre- and permeabilization and subsequent staining with Abs (anti-Bcl-2, anti- senting genes associated with the variable lineages B, T, NK, dendritic cell, myeloperoxidase (MPO), anti-J chain, or anti-TdT; Table I). The labeled and myeloid, or genes associated with apoptosis, adhesion, regulation of cells were analyzed in LSR II (BD Biosciences) using FlowJo software transcription, and signaling. In the cell lineage marker gene sets, we also (www.flowjo.com). included selected genes (supplementary table IA).4 RNA extraction, amplification, labeling, and hybridization Gene pairs for discrimination of HSC, E-B, and pro-B stages, Each cell population was isolated from six donors and analyzed indepen- and E-B gene expression profile 4 4 dently. Total RNA was extracted from 1 ϫ 10 –8 ϫ 10 cells by the RNA Seventy-two genes were selected because they characterized differences Microprep kit (Stratagene), and subjected to two cycles of in vitro tran- between the three first stages (one stage vs rest or one stage vs next stage), scription according to a slightly modified protocol from Baugh et al. (16), had at least a 2-fold change, and were coding for membrane proteins (se- previously described by Nygaard et al. (17). The quality of the aRNA was lected by GO terms) or because they were of particular interest. All pairs evaluated by the Agilent 2100 Bioanalyzer. The gene expression profiling of these 72 genes were examined to determine whether the expression was performed using the Lymphochip cDNA microarrays, and the analysis levels of some of these gene pairs discriminated between the three stages. was performed as previously described (18). Data files for all experiments To test whether sample groups could be separated given the expression of can be found at www.lymphochip.nih.xx. two genes, we used the scoring function: Statistical analysis ͑ ៮ ͒2 ͸͸ d xij, xl. ͑ ៮ ͒2 Ͼ d xij, xi. Genes with 30% missing values were removed from the data set. The i j remaining missing values were estimated and imputed with LSimpute_ adaptive (19). Subsequently, each gene expression profile was mean cen- tered. These steps were performed using J-Express software (Molmine) 4 The online version of this article contains supplemental material. 3664 GENE EXPRESSION PROFILING OF E-B CELLS

FIGURE 1. Differential gene expres- sion during human B cell development. A, Immunomagnetic selection and subse- quent FACS were used to isolate the five populations from adult human BM. Shown are the FACS dot plots with sort- ing gates to obtain CD34ϩCD38Ϫ hemo- poietic stem cells (HSC), CD34ϩCD38ϩ CD10ϩCD19Ϫ E-B cells (E-B), CD34ϩ CD10ϩCD19ϩIgMϪ progenitor B cells (pro-B), CD34ϪCD10ϩCD19ϩIgMϪ precursor B cells (pre-B), and CD34Ϫ CD10ϩCD19ϩIgMϩ IM-B cells. B,Ex- pression of the genes encoding the selec- tion markers confirmed the validity of the approach. C, Hierarchical clustering of the 758 genes that differed significantly between two subsequent stages of differ- entiation (Ն2-fold change, FDR 10%). Total RNA was extracted from the puri- Downloaded from fied cell populations, amplified, and hy- bridized to Lymphochip cDNA microar- rays. Independent experiments from six different donors were performed for each cell population. Color changes within a row indicate expression levels relative to http://www.jimmunol.org/ the average of the sample population.

where Lineage-associated markers

min ͑ ៮ ͒ The expression of genes known to participate in B cell differenti- l : li d xij, xl. , ation and activation (B cell-associated markers), illustrated the

៮ by guest on September 25, 2021 i is group number, j is group element number, xij and xi. are the two- shift in gene expression during B lymphopoiesis (Fig. 2, B cell dimensional coordinates (expression level of the two genes) of each markers). As expected, several B lineage-associated genes, includ- samples and group centroid, respectively, and d() denotes the euclidean ing CD19, CD24, CD79B, and VPREB3, showed highest expres- distance. sion levels from the pro-B stage and through the subsequent stages (cluster IV). Interestingly, ADA, DNTT (TdT), and the Ig recom- Results bination genes RAG1 and RAG2 showed increased expression lev- Differential gene expression in HSC, E-B, pro-B, pre-B, and els already in the CD34ϩCD38ϩCD10ϩCD19Ϫ E-B population IM-B cell populations (4.5-, 26.1-, 7.3-, and 5.6-fold increase, respectively, compared The objective of this study was to perform gene expression pro- with the expression levels in HSC), together with HMGB1, filing of the earliest developmental stages of human B cell differ- HMGB2, and EZH2, suggesting that the rearrangement machinery entiation. RNA was extracted from 1 ϫ 104–8 ϫ 104 cells of each is activated in CD19-negative cells (cluster II). Additionally, TCF3 of the five populations from normal adult human BM: CD34ϩ (E2A), EBF, and PAX5, which are of crucial importance for B cell CD38Ϫ HSCcells;CD34ϩCD38ϩCD10ϩCD19Ϫ E-Bcells;CD34ϩ development and function, were expressed at higher levels in E-B CD10ϩCD19ϩIgMϪ pro-B cells; CD34ϪCD10ϩCD19ϩIgMϪ vs HSC (1.7-, 3.5-, and 8.0-fold increase, respectively), and were pre-B cells, and CD34ϪCD10ϩCD19ϩIgMϩ IM-B cells, followed further increased in the E-B to pro-B transition (2.9-, 1.9-, and by linear RNA amplification and hybridization to Lymphochip 4.4-fold, respectively; Figs. 2 (cluster IV) and 3B, and supplemen- cDNA microarrays (Fig. 1A). The mRNA expression of the genes tary table IB). As expected, genes known to be involved in im- encoding the selection markers corresponded closely to the expres- mune responses and Ag binding, including HLAs, TAP2, IGHM, sion of the respective proteins, confirming the validity of the ap- CD83, CD86, BLR1, and BTG1, showed the highest expression proach (Fig. 1B). Furthermore, expression of 758 genes differed levels in the pre-B and IM-B populations (cluster III). significantly between two subsequent stages of differentiation (Ն2- The E-B population also expressed several genes encoding T or fold change, FDR 10%). Hierarchical clustering of these genes NK lineage-associated markers, including CD2, SELL (CD62L), revealed a pattern that clearly separated the five consecutive cell CD99, CD244, and NOTCH1 (Fig. 2, T/NK cell markers). More- populations (Fig. 1C). Indeed, each population displayed a distinct over, NOTCH1, which is crucial for T lineage commitment and gene expression profile that was remarkably consistent among the development, was expressed in HSC, E-B and pro-B cells, but was individual donors (n ϭ 6). Furthermore, there was a stepwise down-regulated in later stages. In contrast, GATA3 which also is change in the gene expression pattern from HSC throughout the important for T lineage commitment was down-regulated in the other developmental stages. A total of 191, 189, 234, and 363 HSC to E-B transition. Interestingly, E-B cells had higher expres- genes changed in the HSC to E-B, E-B to pro-B, pro-B to pre-B, sion levels of myeloid-associated markers MPO, PECAM1, and and pre-B to IM-B cell transitions, respectively (total gene list in TNFSF13B, than HSC and pro-B cells (Fig. 2, myeloid cell supplementary table IB). markers). The Journal of Immunology 3665

FIGURE 2. Gene expression of lineage-associated markers. The 758 genes that differed significantly be- tween two subsequent stages of dif- ferentiation (Ն2-fold change, FDR 10%) were analyzed to visualize the gene expression of B, T, NK, or my- eloid lineage-associated markers. The Downloaded from lineage-associated gene sets were made from information based on Gene Ontology (GO) by using SOURCE (http://source.stanford.edu/ cgi-bin/source/sourceSearch) and by including selected genes (details in

supplementary table IA). The abbre- http://www.jimmunol.org/ viations of the populations are as in Fig. 1. by guest on September 25, 2021

Gene expression patterns related to function cent cells in the HSC population is in accordance with the finding To further investigate the functional assignment of the differen- of others (22, 23), and a progressive reduction in the percentage of tially expressed genes, we used the proliferation signature (18), or G0 cells during differentiation has also been detected in mice (24). created gene sets based on information from GO (details in Ma- The apoptosis gene set included 38 differentially expressed terials and Methods and supplementary table IA). The prolifera- genes which could be separated in three main clusters (Fig. 3A, tion signature included different cell cycle control genes, cell cycle apoptosis gene set). Of note, genes with highest mRNA expression checkpoint genes, and genes involved in DNA synthesis and rep- levels in HSC and particularly in E-B, included the antiapoptotic lication. Interestingly, this signature divided the populations gene BCL2 in addition to TNFSF13B (encoding the cytokine largely in two groups: the HSC and IM population showed low BAFF), but also the proapoptotic gene DAPK1 (cluster II). Several expression of virtually all genes in this signature (Fig. 3A, prolif- antiapoptotic genes such as BIRC5 (survivine), MYBL2 (B-), eration signature). In contrast, the pro-B and pre-B populations and FAIM, as well as suppressor of cytokine signaling 2 showed comprised high expression levels of nearly all the proliferation- highest expression levels in pro-B and pre-B cells (cluster III). associated genes, in agreement with the clonal expansion and rapid Finally, genes with highest expression levels in IM-B cells and proliferation which follow successful rearrangement of the H and partially also in pre-B cells, included several genes involved in L chains. Interestingly, the proliferation signature suggested that induction of apoptosis like BTG1, HRK, TRAF5, STK17B, TN- the E-B population was not as quiescent as HSC cells. Thirty pro- FRSF21, and BIK (cluster I). Taken together, the populations with liferation-associated genes were increased Ͼ2-fold in the HSC to highest similarity regarding expression of the genes in the apopto- E-B transition, including several genes involved in S phase, such sis gene set turned out to be the HSC and E-B populations. as CKEK1, TOP2A, UHRF1, PCNA, BZW1, and ILF3, and genes The adhesion gene set generated four main clusters which involved in the G2/M phase transition (KIF11, KIF14, and BUB1) clearly illustrated the shift in expression of cell-cell and cell-matrix (Fig. 3B). These results corresponded well with the findings that molecules as the cells mature (Fig. 3A, adhesion gene set). Genes Ϯ 13 2% of E-B cells were in S/G2/M phase of cell cycle, com- with high expression in HSC and E-B and low expression in the pared with 0.4 Ϯ 0.2% and 30 Ϯ 2% of HSC and pro-B cells, next three populations included ITGB2, CD44, and SELL (CD62L) respectively (n ϭ 4, data not shown). The high fraction of quies- (cluster I), while CD34, CDH2 (N-cadherin), CD99, and ITGA6 3666 GENE EXPRESSION PROFILING OF E-B CELLS Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 3. Gene expression changes according to function. The 758 genes that differed significantly between two subse- quent stages of differentiation (Ն2-fold change, FDR 10%) were merged to visu- alize their expression during the E-B cell development and shown according to their functional assignment to the different gene sets (apoptosis, adhesion, regulation of transcription or signaling). The prolifera- tion signature was developed previously (18). The other gene sets were obtained based on information from GO (details in supplementary table IA). A, Analyses of changes in gene expression pattern be- tween the five consecutive cell popula- tions. B, Changes in gene expression pat- tern in the HSC to E-B transition and the E-B to pro-B transition. The Journal of Immunology 3667

Table II. The stage-specific expressed genes in the HSC, E-B cell, and pro-B cell populationa

HSC E-B Cell Pro-B

Gene symbol Fold change t score Gene symbol Fold change t score Gene symbol Fold change t score

MLLT3 10.14 ؊12.24 IL1RAP 3.85 ؊11.90 LIG4 6.69 ؊11.12 MEIS1 7.37 ؊10.01 IGJ 9.43 ؊7.17 CD79B 5.20 ؊10.91 CRHBP 11.05 ؊9.72 CYBB 2.60 ؊7.08 ZHX2 4.91 ؊10.85 HLF 10.37 ؊9.30 BCL2 2.23 ؊6.64 BACH2 2.68 ؊10.76 MYCN 6.14 ؊9.13 LOC57228 2.36 ؊5.97 CD19 3.89 ؊10.64 SCA1 2.87 ؊8.98 RFX1 2.22 ؊5.56 CD72 4.78 ؊10.50 ADRA2A 3.84 ؊8.82 MADH7 3.13 ؊5.33 EFNB2 2.42 ؊10.33 PTGS1 3.21 ؊8.69 S100A8 2.96 ؊5.14 TCL6 2.04 ؊9.40 ABCB1 3.14 ؊8.58 MPO 4.09 ؊4.86 QRSL1 3.62 ؊9.36 DKFZP564K1964 2.49 ؊8.07 C5orf13 3.19 ؊4.54 TCL6 4.44 ؊9.27 DNTT 33.52 13.01 FLNB 2.26 ؊4.41 MAPKAPK3 2.39 11.19 BLNK 9.15 12.36 SLC7A5 2.40 ؊4.23 PTPN22 2.60 9.86 UHRF1 19.43 12.11 P2RY5 2.53 ؊3.88 SELL 3.22 9.50 HMGB1 5.38 9.18 CCND2 3.08 ؊3.75 CD44 3.52 9.46 MEF2A 3.54 8.95 LDHB 3.85 ؊3.73 HRB2 2.61 8.41 EBF 4.85 8.66 HNRPK 2.08 ؊3.65 SLA 2.63 8.39 GLRX 4.20 8.60 ARL7 2.16 ؊3.63 DAPK1 5.30 8.00 Downloaded from RAD51 5.29 8.55 BASP1 2.36 ؊3.55 SAMSN1 2.65 7.26 RASGRP2 2.10 8.39 LRP10 2.11 6.75 SH3BP5 2.40 6.75 BCNP1 6.49 8.33 CTNNA1 2.11 5.45 LILRB1 4.32 6.52 HLA-E 2.08 5.12 HLA-DQB1 2.00 4.01

a Genes were ranked according to their t score (two-sample Student t test), maximum 10% FDR for HSC and pro-B, and maximum 30% FDR for E-B. Up-regulated genes in bold; down-regulated genes in italic. http://www.jimmunol.org/ showed high expression in the first three populations and were IM-B cells included the transcription factors IRF4, ICSBP1 down-regulated in the pre-B and IM-B cell populations (cluster II). (IRF8), NFATC4 and VDR, (cluster IV). Interestingly, several of CD44 and CDH2 have recently been implicated to play a role in the transcription factors that were expressed during the earliest interaction with the stem cell niche (25, 26). Interestingly, stages of B cell development were also found to be highly ex- CTNNA1 (␣-catenin), ICAM3, and SELP (P-selectin) are the pressed in germinal center (GC) cells (GC B cells or CD19ϩ pe- only genes in the gene set that were significantly down-regu- ripheral blood B cells, data not shown), including XBP1, BACH2, lated in the HSC to E-B transition, whereas all others were HMGB2, EZH2, MCM2, MCM7, TCF19, FOXM1, NR3C1, RB1, by guest on September 25, 2021 abundantly expressed also in E-B cells (Fig. 3B). As the cells UHRF1, , and SMARCA4. have rearranged the Ig H (pre-B) and L chains (IM-B), and the Finally, 127 genes coupled to intracellular signaling generated stroma dependency is declining (27), a new set of adhesion six clusters (Fig. 3A, signaling gene set). Genes with low expres- genes were up-regulated, including CD72, CD9, and VCAM1 sion in HSC, E-B, and pro-B, but that were highly up-regulated in (cluster IV). pre-B and IM-B cells, included the chemokine and cytokine re- Totally, 111 genes associated with transcription regulation were ceptors CXCR4 and IL4R, and several kinases including FGR, differentially expressed in the five populations (Fig. 3A, transcrip- TRAF5, and LYN (cluster II). Interestingly, the TNF family mem- tion gene set). Transcription factors with high mRNA expression bers, TNFSR13B (BLyS/BAFF) and TNFSF4 (OX40L), were in HSC and partly also in E-B, which were down-regulated in the down-regulated in pre-B and IM-B cells (cluster IV). Additionally, next three consecutive populations, included genes adaptor proteins and kinases were represented in this cluster, in- known to play a crucial role in normal development like HOXB6, cluding LCP2 (SLP76), FYN, MAP4K3 and MAPKAPK3. Thus, the HOXB2, and MEIS1, and genes important for the self-renewal po- mRNA levels of LYN and FYN, which are important in (pre-) BCR tential such as STAT5A (cluster I). Genes with high expression signaling, were inversely expressed in the five populations. Finally, levels in the first three populations (HSC, E-B, and pro-B cells), genes included in cluster VI were up-regulated in pro-B and pre-B and then down-regulated in the pre-B and IM-B populations in- cells, and partly in E-B cells. In addition to genes involved in cell cluded TIEG, XBP1, ERG, and TCF7L2 (cluster III). As earlier cycle regulation, this cluster also included BLNK (SLP65), an impor- described, the essential transcription factors EBF, TCF3, PAX5, tant downstream mediator of (pre-) BCR signaling. and LEF1 were up-regulated in HSC to E-B and in E-B to pro-B transitions and continued to be high during the next differentiation E-B gene expression profile steps (cluster VI). Interestingly, several other transcription factors To characterize the earliest steps in B cell development in further showed similar expression patterns including MEF2A, ZHX2, detail, we next compared genes differentially expressed in E-B ZCCHC7, DMTF1, NR3C1, FOXO1A, BACH2, and RB1. Cluster V compared with the neighbor populations HSC and pro-B cells (2- consisted of genes with high expression levels in 1) E-B and pro-B fold change, 30% FDR), giving a list of 22 genes which constituted or 2) pro-B and pre-B. The genes up-regulated in E-B and pro-B the E-B gene expression profile (Table II). Among these genes, 18 cells included BCL11A, which is essential for normal lymphopoi- showed higher and 4 showed lower expression levels in E-B com- esis (28) (cluster Va). Among the genes up-regulated in pro-B and pared with the HSC and pro-B populations. Interestingly, IGJ (J pre-B cells were transcription factors known to be involved in chain) was the gene showing the highest differential expression in the regulation of proliferation or differentiation including MYBL2, E-B population with a 9.4-fold increase compared with the two neigh- FOXM1, GFI1 and TCF19 (Cluster Vb). Finally, several of the boring populations. Additionally, the other up-regulated genes transcription factors with high expression levels in the pre-B and included the antiapoptotic gene BCL2, IL1RAP (essential signal 3668 GENE EXPRESSION PROFILING OF E-B CELLS

criminators for the three populations with score 12.1, 10.9, and 26.8, respectively (Fig. 4 and Table III).

Confirmation of the differential expression of selected genes by FACS analysis Among the genes identified in the E-B gene expression profile, we FIGURE 4. Characterization of HSC, E-B, and pro-B by gene pairs. A selected a set of genes to be examined at the protein level based on gene pair class separation test was performed to identify combinations of feasibility in FACS analyses. In addition, we included receptors markers that could better discriminate the E-B population. The ability of known to be involved during B lymphopoiesis, like CD127 (IL- the gene pairs to discriminate between the stages was ranked by a score, ϩ and a score above 10 gave excellent separation. Shown are dot plots for 7R␣) and CXCR4. Thus, CD34 cells were stained with Abs to three selected gene pairs, where each dot represents the median mRNA identify the HSC, E-B, and pro-B populations, in addition to Abs from individual donors: green, HSC; red, E-B; and blue, pro-B. to the selected proteins. Overall, there were good correlations be- tween mRNA and protein levels in the three populations (Fig. 5). E-B cells demonstrated a strong and relatively homogenous ex- transducing component of the functional IL-1R), CYBB (a subunit of pression of TdT protein and most of these cells also expressed cytochrome b), the MHC class II regulatory factor RFX1, the inhib- IL-1RAcP (IL1RAP), Bcl-2, CD127, and CXCR4. Moreover, we itory MADH7, S100A8 (a calcium-binding S100 protein in- detected a weak expression of J chain and CD2 in E-B, while HSC

volved in the regulation of cell cycle progression and differentiation) and pro B cells did not express detectable levels. The expression level Downloaded from and MPO (a heme protein synthesized during myeloid differentiation). of CD62L (SELL) was relatively homogeneous and strong in E-B cells, in contrast to a weaker and more heterogeneous expression in Characterization of HSC, E-B, and pro-B cells by gene pairs HSC and pro-B cells, respectively. Despite higher MPO mRNA levels A gene pair class separation test was performed to identify com- in E-B cells than in HSC and pro-B cells, we did not observe detect- binations of markers that could better discriminate the E-B popu- able levels of MPO protein in these cells (Fig. 5, B and C).

lation from HSC and pro-B. Totally, 2556 gene pairs were tested Protein expression of selected gene pairs, including CD38 and http://www.jimmunol.org/ and examined, resulting in 68 gene pairs with a score above 10 that CD44, were also analyzed at the single-cell level by FACS analysis. were denoted very good discriminators. The gene pairs DNTT/ As the protein expression of several of these gene pairs was - SELL, CD38/CD44, and IL1RAP/CD44 were all very good dis- tively heterogeneous, particularly in pro-B cells, we obtained a less

Table III. Gene pairs that separate HSC, E-B, and pro-B cell populations

Rank Score Gene 1 Gene 2 Rank Score Gene 1 Gene 2 by guest on September 25, 2021 1 26.8 IL1RAP CD44 25 12.7 DNTT IGJ 2 20.7 RASGRP2 28 12.1 SELL 3 19.7 MEIS1 41 11.2 CD72 4 18.4 CD86 44 11.0 CD79B 5 17.5 CD79B 29 12.1 SELL MEIS1 6 17.5 CTGF 38 11.5 ADRA2A 7 16.9 MERTK 43 11.1 PAX5 8 16.9 DNTT 64 10.1 CTGF 9 16.8 CD38 30 12.1 MEIS1 CD79B 10 16.7 MKI67 37 11.5 CD72 11 16.1 PAX5 42 11.2 CTGF 12 15.6 DDR1 47 11.0 CYBB 13 15.5 EFNB2 36 11.5 PAX5 CD72 14 14.3 CD81 61 10.2 EFNB2 16 14.1 PTPRE 18 13.3 CTGF CYBB 17 13.4 SELL 26 12.4 BCL2 19 13.1 CXCR4 33 11.8 ADRA2A 21 12.9 ADRA2A 46 11.0 CD38 22 12.8 CD50 54 10.8 LAIR1 24 12.7 CD72 55 10.8 IGJ 27 12.4 DLG5 56 10.7 DLG5 32 11.9 ADAM28 57 10.7 CD45 45 11.0 CD24 65 10.1 FLNB 51 10.8 CD10 66 10.1 DDR1 53 10.8 IL18R1 68 10.0 CD50 63 10.1 CD45 31 11.9 ADRA2A CD79B 15 14.2 RASGRP2 CYBB 58 10.4 CYBB 20 13.0 CD44 59 10.4 CYBB CD81 35 11.6 EFNB2 60 10.3 CD45 40 11.2 CTGF 62 10.2 DDR1 23 12.8 CD44 MEIS1 67 10.0 DLG5 34 11.7 CTGF 39 11.3 ADRA2A 48 10.9 CYBB 49 10.9 CD38 50 10.8 DLG5 52 10.8 PAX5 The Journal of Immunology 3669

that this population encompasses several subpopulations (1). In humans, most investigators have identified CLP as CD34ϩCD10ϩCD19Ϫ (3, 5, 29). Of interest, CD34ϩCD10ϩ CD19Ϫ cells expressing CD127 (IL-7R␣), were found to be biased toward the B lineage (30). We found that the majority of CD34ϩCD10ϩCD19Ϫ E-B cells expressed CD127 (77 Ϯ 1%), while only a small fraction coexpressed CD7 (3 Ϯ 1%, data not shown). This is in line with other data, which also indicated that CD34ϩCD7ϩ cells are biased to T lineage development (31). Interestingly, we found that the T/NK cell-associated gene CD2 was specifically expressed in E-B cells both at the mRNA and the protein level. CD2 expression has previously been detected on normal and neoplastic B progenitors (32), but the striking expression of CD2 on CD34ϩCD38ϩCD10ϩCD19Ϫ cells has not been previously recognized. In addition, other T/NK-asso- ciated genes were also expressed in E-B cells, including IL1RAP, SELL, CD244, CD99, and NOTCH1. Furthermore, E-B cells expressed mRNA for the myeloid-associated genes MPO, PECAM1, and TNFSF13B, although MPO protein Downloaded from expression could not be detected. Taken together, these data suggest that the E-B population might contain cells with mul- tilineage potential. Of note, Rumfelt et al. (33) recently dem- onstrated that murine CD19Ϫ B cell precursor populations re- vealed considerable lineage plasticity and could be divided into three subpopulations, including multilineage progenitor cells, http://www.jimmunol.org/ CLP cells which possessed lymphoid and myeloid potential (MLP), and pre-pro-B cells which possessed B/T potential. In- vestigations to further subdivide human CD34ϩ cells on the basis of CD2, CD10, CD22, CD62L, CD244, CXCR4, and CD127 expression, in combination with functional studies to explore the lineage plasticity, might be of value for better char- acterization of different E-B lineage populations in humans.

How early the initiation of VDJ rearrangement in B lineage by guest on September 25, 2021 cells starts is still debated and has been a focus in recent studies. We found that the E-B population had turned on transcription of several genes encoding proteins involved in gene rearrangement, FIGURE 5. Verification of the gene expression at the protein level. including RAG1, RAG2, DNTT (TdT), and ADA, indicating that the Five-color FACS analyses were used to examine the protein level of a recombinase machinery is initiated before expression of CD19. selected set of genes. The protein analysis was performed on freshly iso- This is in line with other studies, which also found high expression lated cells from human BM. Shown are FACS histograms of the selected ϩ Ϫ ϩ Ϫ protein (black line) or corresponding isotype control Ab (gray line) in the of TdT in human CD34 CD19 CD10 lin cells (4), and RAG1 ϩ Ϫ ϩ Ϫ various subpopulations (CD34ϩCD38Ϫ HSC, CD34ϩCD38ϩCD10ϩ and RAG2 in CD34 Lin CD10 CD7 cells (5). Moreover, it has ϩ ϩ ϩ Ϫ CD19Ϫ E-B, and CD34ϩCD38ϩCD10ϩCD19ϩ pro-B), extracellular pro- been shown that CD34 CD10 CD127 CD19 cells have initi- teins in A and intracellular proteins in B. Each protein was verified at least ated rearrangement of DJH (34). The initiation of rearrangement of twice. C, The mRNA expression of the selected genes. All the selected DJ in CD34ϩCD22ϩCD19Ϫ cells was shown by van Zelm et al. Ն H genes differed significantly between the different populations ( 2-fold (13), although they found that RAG1 or RAG2 expression levels change, FDR 10%), except for CD127 which did not change significantly were hardly up-regulated in this population, suggesting that rear- at the mRNA level. rangement can be initiated even with low levels of RAG1/RAG2. It should be noted that van Zelm et al. (13) defined pro-B cells as ϩ Ϫ ϩ good separation of these three populations by FACS analysis com- CD34 lin CD22 . Analysis of CD22 expression revealed that ϩ Ϫ ϩ Ϫ pared with use of (average) mRNA levels (data not shown). among CD34 CD19 cells, 10.6% were CD22 CD10 , 0.5% were CD10ϩCD22Ϫ, and 2.4% were CD22ϩCD10ϩ (data not Discussion shown), indicating that our E-B subpopulation (composed of CD22ϩ ϩ ϩ Ϫ In this study, we have characterized early human B cell develop- CD10 and CD10 CD22 cells) was only partly overlapping ϩ Ϫ ϩ ment by gene expression profiling of five consecutive populations with the CD34 lin CD22 population. In line with our findings, in adult BM: HSC, E-B, pro-B, pre-B, and IM-B cells using RNA RAG2 and TdT expression was already detectable in murine MLP amplification and Lymphochip cDNA microarrays. Gene sets were cells, while CLP cells were found to express TdT, RAG1, and used to characterize the different populations in further detail. As RAG2 at significant levels (33). We observed a homogenous and the earliest stages in human B lymphopoiesis are not clearly de- strong TdT protein staining of E-B cells, whereas HSC had no fined, we particularly focused on the characterization of the E-B detectable levels. This could indicate that there exist intermediate population and the changes in gene expression during the transi- precursors between CD34ϩCD38Ϫ cells and the CD34ϩCD38ϩ tion from HSC to E-B and from E-B to pro-B cells. CD10ϩCD19Ϫ E-B population with low or intermediate levels of In the present study, we have defined the population TdT and/or RAG genes, which were not included in our study. CD34ϩCD38ϩCD10ϩCD19Ϫ as E-B cells. Recent data suggest Indeed, recent data suggest that CD79A and CD127 expression 3670 GENE EXPRESSION PROFILING OF E-B CELLS

precedes CD10 and CD19 expression and that these cells have DJH II and Fig. 5). The J chain determines the correct assembly of rearrangement and faintly express PAX5 (35, 36). IgA dimers and IgM pentamers (48), and accordingly shows a Crucial transcription factors involved in E-B lymphopoiesis, strong increase during the last stages of B cell activation (49). including E2A, EBF, PAX5, BCL11A, and LEF1, were also In line with our data, J-chain mRNA expression has previously turned on in the HSC to E-B transition, and showed a further been detected in B cell lines representing progenitors and also increase in the pro-B population. This is in agreement with in primary cells including E-B progenitors (CD34ϩCD19Ϫ), recent findings at the single-cell level in mice where E2A was pro-B cells, and in early fetal thymocytes (50–52). The tran- found to be expressed in MLP, whereas EBF and PAX5 were scriptional regulation of J-chain expression has been well-stud- expressed in the progression from CLP to pre-pro-B cells (33). ied in mice (53, 54), and myocyte enhancer factor-2 () is Some genes known to be activated by PAX5 were found to be one of the known positive regulators of IGJ gene expression. Of increased in the E-B population including BLNK (8.4-fold) and note, we found that MEF2A was highly up-regulated in E-B CD79A (1.9-fold), whereas another target gene, CD19, was cells, suggesting that the increased MEF2A levels could con- turned on in the pro-B cells. IGJ and XBP1, genes known to be tribute to J-chain expression in E-B cells. Pax5, a negative reg- repressed by PAX5, were down-regulated in pro-B and pre-B ulator of IGJ, was slightly up-regulated in E-B cells, and cells, respectively. The differential effects of PAX5 on various showed a further increase in the E-B/pro-B transition. It has target genes is in line with previous data (37) and is likely due been hypothesized that high levels of Pax5 prevent IGJ expres- to variable expression and/or activity of transcriptional coregu- sion during B cell development until Pax5 is down-regulated at lators interacting with PAX5 at the different stages in B cell the plasma cell stage (53, 54). Thus, the transient IGJ expres- development. Of note, we also identified a number of genes sion observed at the E-B stage could be due to high expression Downloaded from encoding transcription factors with expression pattern similar to of MEF2A, preceding high levels of Pax5 in pro B cells. The the described essential factors. EWSR1, ZCCHC7, MEF2A, and transient expression of J chain in early lymphoid progenitors, NR3C1 showed exactly the same expression pattern as EBF, before Ig expression, is intriguing, but so far there is no evi- whereas the expression of ARID5B, ZHX2, DMTF1, FOXO1A, dence for a functional role for J chain in early progenitors. BACH2, ZEP36L1, and RB1 was similar to the expression of Identification and characterization of the gene expression in

E2A and PAX5. Many of these genes have a well-characterized normal precursor B cell subsets are valuable for the study of http://www.jimmunol.org/ role in cell cycle regulation, whereas little is known regarding malignant B lineage cells, and can give a more precise reference other genes, in particular the zinc finger transcription factors for the cellular origin of precursor B cell acute lymphoblastic ZHX2 and ZCCHC7. Thus, these two transcription factors rep- leukemia (ALL). ALL is a heterogeneous disease with a number resent interesting candidates to be further explored regarding a of genetically distinct leukemia subtypes, and can arise from putative role in E-B cell development. any stage of development, including primitive cells with mul- Of note, several of the transcription factors showed a biphasic tilineage potential. Gene expression profiling (55–57), and sin- expression pattern during B cell development. We observed that gle nucleotide polymorphism array analysis have provided new IRF4 and ICSBP1 (IRF8) showed expression peaks in the pre-B insights into the underlying biology of ALL subtypes (58). In- and IM-B cells, respectively. In line with this, double knockout terestingly, Mullighan et al. (59) recently used single nucleotide by guest on September 25, 2021 mice (IRF4Ϫ/ϪIRF8Ϫ/Ϫ) were shown to have a complete block polymorphism arrays to detect genetic alterations in leukemic at the pre-B cell stage, demonstrating a crucial role for these cells from 242 ALL patients. PAX5 turned out to be the most factors in the transition from pre-B to IM-B cells (38). Both frequently mutated gene, giving reduced levels of PAX5 protein or genes are down-regulated at later stages until re-expressed in a generated hypomorphic alleles. Furthermore, deletions were also de- subpopulation of centrocytes and in plasma cells (39), where tected in TCF3, LEF1, IKZF1, and IKZF3, suggesting that direct dis- they control class-switch recombination and plasma cell differ- ruption of pathways controlling B cell development contributes to entiation, respectively (40, 41). BCL11A also demonstrated a B-ALL pathogenesis. Hence, comparison of gene expression profiles biphasic expression pattern with up-regulation in the E-B and of normal and neoplastic B cell progenitors may identify genes in- pro-B populations. BCL11A has previously been demonstrated volved in the oncogenic transformation and gene products of potential to be essential for lineage commitment of both B and T cells in prognostic and therapeutic relevance. mice (28), and is expressed also in GC cells, where it functions as a transcriptional repressor (42, 43). Interestingly, we identi- Acknowledgment fied several other transcription factors with biphasic expression We thank Prof. P. O. Brandtzaeg for kindly providing us with the patterns. Most of these factors are known to be important for the anti-J-chain Ab. cell cycle, RNA transcription, or cell proliferation, but also included Disclosures factors involved in V(D)J recombination such as HMGB2 and EZH2 The authors have no financial conflict of interest. (44, 45). In addition, we observed high expression levels in early stages of B cell development of other transcription factors with known References essential roles in GC B cells or plasma cells, including BACH2 and 1. Blom, B., and H. Spits. 2006. Development of human lymphoid cells. Annu. Rev. XBP1 (46, 47). However, there is no evidence for a functional role for Immunol. 24: 287–320. 2. LeBien, T. W. 2000. Fates of human B-cell precursors. Blood 96: 9–23. these factors in E-B cell progenitors. Taken together, the biphasic 3. Galy, A., M. Travis, D. Cen, and B. Chen. 1995. Human T, B, natural killer, and expression of several transcription factors during B cell differentiation dendritic cells arise from a common bone marrow progenitor cell subset. Immu- suggest the use of similar transcriptional networks in E-B cell stages nity 3: 459–473. 4. Imamura, R., T. Miyamoto, G. Yoshimoto, K. Kamezaki, F. Ishikawa, and in GC and plasma cells. H. Henzan, K. Kato, K. Takase, A. Numata, K. Nagafuji, et al. 2005. Mobiliza- One of the surprising findings in our study was that IGJ gene tion of human lymphoid progenitors after treatment with granulocyte colony- (J chain) expression was significantly up-regulated in E-B cells stimulating factor. J. Immunol. 175: 2647–2654. 5. Rossi, M. I., T. Yokota, K. L. Medina, K. P. Garrett, P. C. Comp, A. H. Schipul, compared with HSC and pro-B cells. The mRNA expression in Jr., and P. W. Kincade. 2003. B lymphopoiesis is active throughout human life, E-B cells was, however, much lower than in IM-B cells. In line but there are developmental age-related changes. Blood 101: 576–584. 6. Jung, D., C. Giallourakis, R. Mostoslavsky, and F. W. Alt. 2006. Mechanism and with this, we could detect weak, but distinct, protein expression control of V(D)J recombination at the immunoglobulin heavy chain locus. Annu. of J chain in E-B cells, but not in HSC or in pro-B cells (Table Rev. Immunol. 24: 541–570. The Journal of Immunology 3671

7. Matthias, P., and A. G. Rolink. 2005. Transcriptional networks in developing and 34. Bertrand, F. E., L. G. Billips, P. D. Burrows, G. L. Gartland, H. Kubagawa, and mature B cells. Nat. Rev. Immunol. 5: 497–508. H. W. Schroeder, Jr. 1997. Ig DH gene segment transcription and rearrangement 8. Urbanek, P., Z. Q. Wang, I. Fetka, E. F. Wagner, and M. Busslinger. 1994. before surface expression of the pan-B-cell marker CD19 in normal human bone Complete block of early B cell differentiation and altered patterning of the pos- marrow. Blood 90: 736–744. terior midbrain in mice lacking Pax5/BSAP. Cell 79: 901–912. 35. Reynaud, D., N. Lefort, E. Manie, L. Coulombel, and Y. Levy. 2003. In vitro 9. Horcher, M., A. Souabni, and M. Busslinger. 2001. Pax5/BSAP maintains the identification of human pro-B cells that give rise to macrophages, natural killer identity of B cells in late B lymphopoiesis. Immunity 14: 779–790. cells, and T cells. Blood 101: 4313–4321. 10. Kurosaka, D., T. W. LeBien, and J. A. R. Pribyl. 1999. Comparative studies of 36. Sanz, E., M. varez-Mon, A. Martinez, and A. de la Hera. 2003. Human cord blood different stromal cell microenvironments in support of human B-cell develop- CD34ϩPax-5ϩ B-cell progenitors: single-cell analyses of their gene expression ment. Exp. Hematol. 27: 1271–1281. profiles. Blood 101: 3424–3430. 11. Hoffmann, R., T. Seidl, M. Neeb, A. Rolink, and F. Melchers. 2002. Changes in 37. Johnson, K., M. Shapiro-Shelef, C. Tunyaplin, and K. Calame. 2005. Regulatory gene expression profiles in developing B cells of murine bone marrow. Genome events in early and late B-cell differentiation. Mol. Immunol. 42: 749–761. Res. 12: 98–111. 38. Lu, R., K. L. Medina, D. W. Lancki, and H. Singh. 2003. IRF-4,8 orchestrate the 12. Hoffmann, R., T. Seidl, L. Bruno, and M. Dugas. 2003. Developmental markers pre-B-to-B transition in lymphocyte development. Genes Dev. 17: 1703–1708. of B cells are superior to those of T cells for identification of stages with distinct 39. Falini, B., M. Fizzotti, A. Pucciarini, B. Bigerna, T. Marafioti, M. Gambacorta, gene expression profiles. J. Leukocyte Biol. 74: 602–610. R. Pacini, C. Alunni, L. Natali-Tanci, B. Ugolini, et al. 2000. A monoclonal antibody 13. van Zelm, M. C., M. van der Burg, D. de Ridder, B. H. Barendregt, (MUM1p) detects expression of the MUM1/IRF4 protein in a subset of germinal E. F. E. de Haas, M. J. T. Reinders, A. C. Lankester, T. Revesz, F. J. T. Staal, and center B cells, plasma cells, and activated T cells. Blood 95: 2084–2092. J. J. M. van Dongen. 2005. Ig gene rearrangement steps are initiated in early 40. Klein, U., S. Casola, G. Cattoretti, Q. Shen, M. Lia, T. Mo, T. Ludwig, human precursor B cell subsets and correlate with specific transcription factor K. Rajewsky, and R. la-Favera. 2006. Transcription factor IRF4 controls plasma expression. J. Immunol. 175: 5912–5922. cell differentiation and class-switch recombination. Nat. Immunol. 7: 773–782. 14. Myklebust, J. H., E. B. Smeland, D. Josefsen, and M. Sioud. 2000. Protein kinase 41. Lee, C. H., M. Melchers, H. Wang, T. A. Torrey, R. Slota, C. F. Qi, J. Y. Kim, C-␣ isoform is involved in erythropoietin-induced erythroid differentiation of P. Lugar, H. J. Kong, L. Farrington, et al. 2006. Regulation of the germinal center CD34ϩ progenitor cells from human bone marrow. Blood 95: 510–518. gene program by interferon (IFN) regulatory factor 8/IFN consensus sequence- 15. Smeland, E. B., S. Funderud, G. Kvalheim, G. Gaudernack, A. M. Rasmussen, binding protein. J. Exp. Med. 203: 63–72. L. Rusten, M. Y. Wang, R. W. Tindle, H. K. Blomhoff, and T. Egeland. 1992. 42. Liu, H., G. Ippolito, J. Wall, T. Niu, L. Probst, B. S. Lee, K. Pulford, A. Banham, Downloaded from Isolation and characterization of human hematopoietic progenitor cells: an effec- L. Stockwin, A. Shaffer, et al. 2006. Functional studies of BCL11A: character- tive method for positive selection of CD34ϩ cells. Leukemia 6: 845–852. ization of the conserved BCL11A-XL splice variant and its interaction with 16. Baugh, L. R., A. A. Hill, E. L. Braun, and C. P. Hunter.2001. Quantitative anal- BCL6 in nuclear paraspeckles of germinal center B cells. Mol. Cancer 5: 18. ysis of mRNA amplification of in vitro transcription. Nucleic Acids Res. 29: e29. 43. Pulford, K., A. H. Banham, L. Lyne, M. Jones, G. C. Ippolito, H. Liu, 17. Nygaard, V., A. Løland, M. Holden, M. Langaas, H. Rue, F. Liu, O. Myklebost, P. W. Tucker, G. Roncador, E. Lucas, S. Ashe, et al. 2006. The BCL11AXL Ø. Fodstad, E. Hovig, and B. Smith-Sørensen. 2003. Effects of mRNA amplifi- transcription factor: its distribution in normal and malignant tissues and use as a cation on gene expression ratios in cDNA experiments estimated by analysis of marker for plasmacytoid dendritic cells. Leukemia 20: 1439–1441.

variance. BMC Genomics 4: 11. 44. Dai, Y., B. Wong, Y. M. Yen, M. A. Oettinger, J. Kwon, and R. C. Johnson. http://www.jimmunol.org/ 18. Alizadeh, A. A., M. B. Eisen, R. E. Davis, C. Ma, I. S. Lossos, A. Rosenwald, 2005. Determinants of HMGB proteins required to promote RAG1/2-recombi- J. C. Boldrick, H. Sabet, T. Tran, X. Yu, et al. 2000. Distinct types of diffuse large nation signal sequence complex assembly and catalysis during V(D)J recombi- B-cell lymphoma identified by gene expression profiling. Nature 403: 503–511. nation. Mol. Cell. Biol. 25: 4413–4425. 19. Bo, T. H., B. Dysvik, and I. Jonassen. 2004. LSimpute: accurate estimation of miss- 45. Su, I. H., A. Basavaraj, A. N. Krutchinsky, O. Hobert, A. Ullrich, B. T. Chait, and ing values in microarray data with least squares methods. Nucleic Acids Res. 32: e34. A. Tarakhovsky. 2003. Ezh2 controls B cell development through histone H3 20. Dysvik, B., and I. Jonassen. 2001. J-Express: exploring gene expression data methylation and Igh rearrangement. Nat. Immunol. 4: 124–131. using Java. Bioinformatics 17: 369–370. 46. Muto, A., S. Tashiro, O. Nakajima, H. Hoshino, S. Takahashi, E. Sakoda, 21. Alizadeh, A., M. Eisen, R. E. Davis, C. Ma, H. Sabet, T. Tran, J. I. Powell, D. Ikebe, M. Yamamoto, and K. Igarashi. 2004. The transcriptional programme L. Yang, G. E. Marti, D. T. Moore, et al. 1999. The lymphochip: a specialized of antibody class switching involves the repressor Bach2. Nature 429: 566–571. cDNA microarray for the genomic-scale analysis of gene expression in normal 47. Reimold, A. M., N. N. Iwakoshi, J. Manis, P. Vallabhajosyula, E. Szomolanyi-Tsuda, and malignant lymphocytes. Cold Spring Harb. Symp. Quant. Biol. 64: 71–78. E. M. Gravallese, D. Friend, M. J. Grusby, F. Alt, and L. H. Glimcher. 2001. Plasma

22. Bradford, G. B., B. Williams, R. Rossi, and I. Bertoncello. 1997. Quiescence, cell differentiation requires the transcription factor XBP-1. Nature 412: 300–307. by guest on September 25, 2021 cycling, and turnover in the primitive hematopoietic stem cell compartment. Exp. 48. Brandtzaeg, P., and H. Prytz. 1984. Direct evidence for an integrated function of Hematol. 25: 445–453. J chain and secretory component in epithelial transport of immunoglobulins. Na- 23. Cheshier, S. H., S. J. Morrison, X. Liao, and I. L. Weissman. 1999. In vivo ture 311: 71–73. proliferation and cell cycle kinetics of long-term self-renewing hematopoietic 49. Brandtzaeg, P., and F. E. Johansen. 2005. Mucosal B cells: phenotypic characteris- stem cells. Proc. Natl. Acad. Sci. USA 96: 3120–3125. tics, transcriptional regulation, and homing properties. Immunol. Rev. 206: 32–63. 24. Passegue, E., A. J. Wagers, S. Giuriato, W. C. Anderson, and I. L. Weissman. 2005. 50. Erlandsson, L., P. Akerblad, C. Vingsbo-Lundberg, E. Kallberg, N. Lycke, and Global analysis of proliferation and cell cycle gene expression in the regulation of T. Leanderson. 2001. Joining chain-expressing and -nonexpressing B cell popu- hematopoietic stem and progenitor cell fates. J. Exp. Med. 202: 1599–1611. lations in the mouse. J. Exp. Med. 194: 557–570. 25. Jin, L., K. J. Hope, Q. Zhai, F. Smadja-Joffe, and J. E. Dick. 2006. Targeting of 51. McCune, J. M., S. M. Fu, and H. G. Kunkel. 1981. J chain biosynthesis in pre-B CD44 eradicates human acute myeloid leukemic stem cells. Nat. Med. 12: cells and other possible precursor B cells. J. Exp. Med. 154: 138–145. 1167–1174. 52. Bertrand, F. E., L. G. Billips, G. L. Gartland, H. Kubagawa, and H. W. Schroeder, 26. Zhang, J., C. Niu, L. Ye, H. Huang, X. He, W. G. Tong, J. Ross, J. Haug, Jr. 1996. The J chain gene is transcribed during B and T lymphopoiesis in hu- T. Johnson, J. Q. Feng, et al. 2003. Identification of the haematopoietic stem cell mans. J. Immunol. 156: 4240–4244. niche and control of the niche size. Nature 425: 836–841. 53. Rao, S., K. Saoussen, E. R. Gackstetter, and M. E. Koshland. 1998. Myocyte 27. Glodek, A. M., M. Honczarenko, Y. Le, J. J. Campbell, and L. E. Silberstein. enhancer factor-related B-MEF2 is developmentally expressed in B cells and 2003. Sustained activation of cell adhesion is a differentially regulated process in regulates the immunoglobulin J chain promoter. J. Biol. Chem. 273: B lymphopoiesis. J. Exp. Med. 197: 461–473. 26123–26129. 28. Liu, P., J. R. Keller, M. Ortiz, L. Tessarollo, R. A. Rachel, T. Nakamura, 54. Wallin, J. J., J. L. Rinkenberger, S. Rao, E. R. Gackstetter, M. E. Koshland, and N. A. Jenkins, and N. G. Copeland. 2003. Bcl11a is essential for normal lym- P. Zwollo. 1999. B cell-specific activator protein prevents two activator factors phoid development. Nat. Immunol. 4: 525–532. from binding to the immunoglobulin J chain promoter until the antigen-driven 29. Miller, J. S., V. McCullar, M. Punzel, I. R. Lemischka, and K. A. Moore. 1999. stages of B cell development. J. Biol. Chem. 274: 15959–15965. Single adult human CD34ϩ/LinϪ/CD38Ϫ progenitors give rise to natural killer 55. Yeoh, E. J., M. E. Ross, S. A. Shurtleff, W. K. Williams, D. Patel, R. Mahfouz, cells, B-lineage cells, dendritic cells, and myeloid cells. Blood 93: 96–106. F. G. Behm, S. C. Raimondi, M. V. Relling, A. Patel, et al. 2002. Classification, 30. Ryan, D. H., B. L. Nuccie, I. Ritterman, J. L. Liesveld, C. N. Abboud, and subtype discovery, and prediction of outcome in pediatric acute lymphoblastic R. A. Insel. 1997. Expression of interleukin-7 by lineage-negative hu- leukemia by gene expression profiling. Cancer Cell 1: 133–143. man bone marrow progenitors with enhanced lymphoid proliferative potential 56. Ross, M. E., X. Zhou, G. Song, S. A. Shurtleff, K. Girtman, W. K. Williams, and B-lineage differentiation capacity. Blood 89: 929–940. H. C. Liu, R. Mahfouz, S. C. Raimondi, N. Lenny, et al. 2003. Classification of 31. Haddad, R., P. Guardiola, B. Izac, C. Thibault, J. Radich, A. L. Delezoide, pediatric acute lymphoblastic leukemia by gene expression profiling. Blood 102: C. Baillou, F. M. Lemoine, J. C. Gluckman, F. Pflumio, and B. Canque. 2004. 2951–2959. Molecular characterization of early human T/NK and B-lymphoid progenitor 57. Chiaretti, S., X. Li, R. Gentleman, A. Vitale, K. S. Wang, F. Mandelli, R. Foa, cells in umbilical cord blood. Blood 104: 3918–3926. and J. Ritz. 2005. Gene expression profiles of B-lineage adult acute lymphocytic 32. Uckun, F. M., A. Muraguchi, J. A. Ledbetter, T. Kishimoto, R. T. O’Brien, leukemia reveal genetic patterns that identify lineage derivation and distinct J. S. Roloff, K. Gajl-Peczalska, A. Provisor, and B. Koller. 1989. Biphenotypic mechanisms of transformation. Clin. Cancer Res. 11: 7209–7219. leukemic lymphocyte precursors in CD2ϩCD19ϩ acute lymphoblastic leukemia 58. Downing, J. R., and C. G. Mullighan. 2006. Tumor-specific genetic lesions and and their putative normal counterparts in human fetal hematopoietic tissues. their influence on therapy in pediatric acute lymphoblastic leukemia. Hematology Blood 73: 1000–1015. 2006: 118–122. 33. Rumfelt, L. L., Y. Zhou, B. M. Rowley, S. A. Shinton, and R. R. Hardy. 2006. 59. Mullighan, C. G., S. Goorha, I. Radtke, C. B. Miller, E. Coustan-Smith, J. D. Dalton, Lineage specification and plasticity in CD19Ϫ early B cell precursors. J. Exp. K. Girtman, S. Mathew, J. Ma, S. B. Pounds, et al. 2007. Genome-wide analysis of Med. 203: 675–687. genetic alterations in acute lymphoblastic leukaemia. Nature 446: 758–764.