Characterization of a Transitional Preplasmablast Population in the Process of B Cell to Plasma Cell Differentiation

This information is current as Michel Jourdan, Anouk Caraux, Gersende Caron, Nicolas of September 29, 2021. Robert, Geneviève Fiol, Thierry Rème, Karine Bolloré, Jean-Pierre Vendrell, Simon Le Gallou, Frédéric Mourcin, John De Vos, Alboukadel Kassambara, Christophe Duperray, Dirk Hose, Thierry Fest, Karin Tarte and Bernard Klein Downloaded from J Immunol 2011; 187:3931-3941; Prepublished online 14 September 2011; doi: 10.4049/jimmunol.1101230 http://www.jimmunol.org/content/187/8/3931 http://www.jimmunol.org/

Supplementary http://www.jimmunol.org/content/suppl/2011/09/14/jimmunol.110123 Material 0.DC1 References This article cites 38 articles, 18 of which you can access for free at: http://www.jimmunol.org/content/187/8/3931.full#ref-list-1 by guest on September 29, 2021 Why The JI? Submit online.

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

• 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 © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Characterization of a Transitional Preplasmablast Population in the Process of Human B Cell to Plasma Cell Differentiation

Michel Jourdan,* Anouk Caraux,* Gersende Caron,†,‡ Nicolas Robert,x Genevie`ve Fiol,x Thierry Re`me,*,x Karine Bollore´,x Jean-Pierre Vendrell,x,{ Simon Le Gallou,† Fre´de´ric Mourcin,† John De Vos,*,x,{ Alboukadel Kassambara,* Christophe Duperray,*,x Dirk Hose,‖,# Thierry Fest,†,‡ Karin Tarte,†,‡ and Bernard Klein*,x,{

The early steps of differentiation of human B cells into plasma cells are poorly known. We report a transitional population of CD20low/2CD382 preplasmablasts along differentiation of human memory B cells into plasma cells in vitro. Preplasmablasts lack documented B cell or plasma cell (CD20, CD38, and CD138) markers, express CD30 and IL-6R, and secrete Igs at a weaker level 2 high 2

than do plasmablasts or plasma cells. These preplasmablasts further differentiate into CD20 CD38 CD138 plasmablasts and Downloaded from then CD202CD38highCD138+ plasma cells. Preplasmablasts were fully characterized in terms of whole genome transcriptome profiling and phenotype. Preplasmablasts coexpress B and plasma cell transcription factors, but at a reduced level compared with B cells, plasmablasts, or plasma cells. They express the unspliced form of XBP1 mRNA mainly, whereas plasmablasts and plasma cells express essentially the spliced form. An in vivo counterpart (CD19+CD20low/2CD382IL-6R+ cells) of in vitro-generated preplasmablasts could be detected in human lymph nodes (0.06% of CD19+ cells) and tonsils (0.05% of CD19+ cells). An open

access “B to Plasma Cell Atlas,” which makes it possible to interrogate expression in the process of B cell to plasma cell http://www.jimmunol.org/ differentiation, is provided. Taken together, our findings show the existence of a transitional preplasmablast population using an in vitro model of plasma cell generation and of its in vivo counterpart in various lymphoid tissues. The Journal of Immunology, 2011, 187: 3931–3941.

he production of high-affinity Ig-producing plasma cells in which they interact each other. Activated B cells then migrate to (PCs) is the end product of a complex network of cell follicles and initiate a germinal center (GC) reaction characterized interaction, gene rearrangements, and mutations. Ag en- by an extensive centroblastic proliferation associated with ran-

T by guest on September 29, 2021 counter induces both B and T cells to move to the outer T cell zone, dom somatic hypermutation in Ig (1) that are dependent on activation-induced deaminase (AID). CXCR4+ centroblasts are found within the dark zone of the GC, in close contact with poorly *INSERM, Unite´ 1040, 34000 Montpellier, France; †INSERM, Unite´ 917, 35043 Rennes, France; ‡Poˆle Cellules et Tissus, Centre Hospitalier Universitaire de Rennes, characterized dark zone follicular dendritic cells (FDCs) that do 35033 Rennes, France; xCentre Hospitalier Universitaire de Montpellier, Institut de not produce CXCL13 (CXCR5 ligand) but express high levels of Recherche en Biothe´rapie, 34295 Montpellier, France; {Unite´ de Formation et de ‖ CXCL12 (CXCR4 ligand) (2, 3). The internalization of CXCR4 Recherche de Me´decine, Universite´ Montpellier 1, 34000 Montpellier, France; Med- izinische Klinik und Poliklinik V, Universita¨tsklinikum Heidelberg, 69120 Heidel- after CXCL12 exposure allows the rapid CXCR5-dependent mi- berg, Germany; and #Nationales Centrum fu¨r Tumorerkrankungen, D-69120 Heidel- gration of centroblasts to the zone, acquiring a centrocyte berg, Germany phenotype, and these centrocytes compete with each other for the Received for publication April 27, 2011. Accepted for publication August 3, 2011. capture of Ag presented as immune complexes by CXCL13+ light This work was supported by grants from the Ligue Nationale Contre le Cancer zone FDCs. B cells with high-affinity Ig will pick up greater (e´quipe labellise´e 2009), Paris, France, and from L’Institut National du Cancer (Grant RPT09001FFA). amounts of Ag, thus receiving a BCR-mediated signal, together M.J. and A.C. designed research, performed the experiments, and wrote the paper; with a complex set of FDC-derived growth factors and adhesion K.T., T.F., G.C., F.M., and S.L.G. performed the experiments to characterize cells molecules (3–5). FDC/B cell interaction is short-lived but allows in vivo and corrected the paper; G.F. and N.R. provided technical assistance; K.B. the capture of Ag for further presentation to a specific subset of and J.-P.V. performed ELISPOT assays; D.H. provided normal bone marrow samples; A.K. provided assistance for bioinformatics analysis; J.D.V. designed the Amazonia GC-restricted follicular helper T cells (TFH). TFH provide them Web site and T.R. designed the RAGE Web site; C.D. provided assistance for cy- with essential survival, class switch recombination, and differen- tometry experiments; and B.K. is the senior investigator who designed research and tiation signals, in particular CD40L and IL-21, that will contribute wrote the paper. to the NF-kB/IRF4–dependent downregulation of BCL6 and the Address correspondence and reprint requests to Prof. Bernard Klein, INSERM, Unite´ 1040, Institut de Recherche en Biothe´rapie, Centre Hospitalier Universitaire de IRF4/STAT3-dependent upregulation of Blimp1. Owing to their Montpellier, Hoˆpital Saint Eloi, Avenue Augustin Fliche, 34295 Montpellier, France. lower affinity for Ag, the vast majority of centrocytes are not E-mail address: [email protected] rescued by this FDC/TFH system and are engulfed by tingible The online version of this article contains supplemental material. body macrophages (6). The GC is a dynamic structure, with GC Abbreviations used in this article: AID, activation-induced deaminase; BM, bone B cells moving cyclically between the dark and light zones to marrow; cy, cytoplasmic; D, day; FDC, follicular dendritic cell; GC, germinal center; MBC, memory B cell; ODN, oligodeoxynucleotide; PB, plasmablast; PC, plasma undergo repeated rounds of mutation and selection (2, 7). A cell; prePB, preplasmablast; s, surface; SAM, significance analysis of microarray; successful selection will eventually promote GC B cells to give sCD40L, soluble CD40L; SI, stain index; TFH, follicular helper T cell; XBP1s, XBP1 rise to early Ig secreting plasmablasts (PBs) or memory B cells spliced; XBP1u, XBP1 unspliced. (MBCs) under presently unknown signals. MBCs persist after im- Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 munization both as Ag-dependent memory within residual GC-like www.jimmunol.org/cgi/doi/10.4049/jimmunol.1101230 3932 CHARACTERIZATION OF HUMAN PREPLASMABLASTS structures, in close contact to FDCs and T cells, and Ag- a lower level than do PBs or PCs. The in vivo counterpart of these independent memory dispersed outside B cell follicles (8). In prePBs could be found in various lymphoid tissues. case of secondary immunization, MBCs can be restimulated to highly proliferate and differentiate into PBs and PCs. Materials and Methods The migration pattern of PBs has been recently documented in Reagents mice (9). PBs travel through the T cell zone toward the medul- Human recombinant IL-2 and IFN-a were purchased from R&D Systems lary cords. They are first found in the outer T cell zone close to (Minneapolis, MN), IL-6 and IL-15 were from AbCys (Paris, France), and 2 CD11c+CD8a dendritic cells, which highly produce IL-6. IL-10 was from PeproTech (Rocky Hill, NJ). Mouse (or rat when in- In vivo imaging has shown that PBs have a unique pattern of dicated) mAbs conjugated to Alexa Fluor 488, Alexa Fluor 647, allo- migration in T cell zone characterized by long linear paths that are phycocyanin, allophycocyanin-H7, allophycocyanin-Cy7, FITC, PerCP- Cy5.5, PE, and PE-Cy7 specific for human CD19 (clones HIB19 and randomly oriented (10). This migration pattern is dependent on an SJ25C1), CD24 (clone ML5), CD27 (clones L128 and M-T271), CD30 integrin ICAM1/2 axis and independent on a chemokine gradient. (clone BerH8), CD38 (clone HIT2), CD43 (clone 1G10), CD45 (clone Once arrived in the medullary zone, PBs further differentiate in HI30), CD138 (clone MI15), CXCR5 (clone RF8B2), IgG (clone G18- contact with Gr1+CD11b+F4/80+ monocytes/macrophages that 145), IgM (clone G20-127), and Ki67 (clone B56) were purchased from BD Biosciences (Le Pont De Claix, France); CD20 (clone B9E9), CD126 highly produce CXCL12, APRIL, and IL-6 (9). PBs are retained (IL-6R, clone M91), and CD138 (clone B-A38) from were Beckman in the medullary zone by the CXCL12 gradient but still keep Coulter (Fullerton, CA); CCR10 (rat, clone 314305) was from R&D moving. One putative reason of this continuous moving is to be Systems; and IgG (polyclonal goat Ab) was from SouthernBiotech (Bir- able to share the limiting PC niche. Most PBs will die in med- mingham, AL). ullary cords, and the surviving ones will exit the lymph node into Cell samples Downloaded from lymphatic vessels through a SP1-dependent gradient (11) and Peripheral blood cells from healthy volunteers were purchased from the enter the blood circulation in which they survive for a short period French Blood Center (Etablissement Franc¸ais du Sang Pyre´nne´es-Me´di- unless they can find a survival niche either in the bone marrow te´ranne´e, Toulouse, France). After removal of CD2+ cells using anti-CD2 (BM), spleen, or mucosa-associated lymphoid tissues, in which magnetic beads (Invitrogen, Cergy Pontoise, France), CD19+CD27+ MBCs they further differentiate into long-lived PCs (12). were sorted with a multicolor fluorescence FACSAria device with a purity

$ 95% (see Fig. 1A). Cells produced in the culture system were FACS- http://www.jimmunol.org/ Several studies have contributed to better understand the pro- sorted using FITC-conjugated anti-CD20 mAb and PE-conjugated anti- cess of differentiation of human centrocytes or MBCs into PBs CD38 mAb for day (D) 4 CD20highCD382 cells, D4 CD20lowCD382 2 2 2 and then mature PCs (reviewed in Ref. 13). The usual features of cells, and D4 CD20 CD38 cells, as well as D7 PBs (CD20 CD38+). 2 + PBs are cell cycling, Ig secretion, high CD38 expression, and loss D10 PCs (CD20 CD138 ) were FACS-sorted using FITC-conjugated anti- CD20 mAb and PE-conjugated anti-CD138 mAb. The purity of FACS- of B cell markers unlike CD19 (14–16). Those of mature PCs are sorted cell populations was $95% as assayed by cytometry. We looked for stop in cell cycle, Ig secretion, high CD38 expression, and CD138 prePBs in tonsil, lymph node, BM, and peripheral blood samples. Tonsils proteoglycan expression (17). We have shown that tonsil-purified were obtained from routine tonsillectomies performed at Chidren’s Clin- CD20+CXCR42CXCR5+ centrocytes already express PC tran- ique La Sagesse at Rennes (France). Reactive nonmalignant lymph node biopsies were collected at Centre Hospitalier Universitaire de Rennes. scription factors (IRF4, Blimp-1, XBP1) together with B cell After mincing, tonsillar and lymph node mononuclear cells were isolated by guest on September 29, 2021 transcription factors (BCL6, PAX5), although it is not clear by Ficoll-density gradient centrifugation. BM aspirates were collected whether centrocytes are a heterogeneous cell population contain- from adult patients undergoing thoracic surgery (Centre Hospitalier Uni- ing centrocytes and cells already committed to PC differentiation versitaire de Rennes) and BM mononuclear cells were isolated by Ficoll density gradient centrifugation. Tonsils, lymph node biopsies, and BM (18). In an elegant study using cellular affinity matrix technology aspirates were all collected after subject recruitment followed Institutional to capture secreted Igs, Arce et al. (19) have shown the existence Review Board approval and written informed consent process according to 2 of a minor population of CD38 /low Ig-secreting cells in tonsils the Declaration of Helsinki. high 2/low together with classical CD38 Ig-secreting cells. CD38 Cell cultures cells weakly expressed CD27 compared with CD38high cells. The CD382/low Ig-secreting cells were not detected in the peripheral In step one, B cell activation, all cultures were performed in IMDM high (Invitrogen) and 10% FCS (Invitrogen), supplemented with 50 mg/ml blood or BM. They were thought to be the precursors of CD38 human transferrin and 5 mg/ml human insulin (Sigma-Aldrich, St Louis, Ig-secreting cells in an in vitro model of B cell differentiation MO). Purified MBCs were cultured with IL-2 (20 U/ml), IL-10 (50 ng/ml), using PBMCs and tetanus toxoid activation. CD382/low Ig- and IL-15 (10 ng/ml) in six-well culture plates (1.5 3 105/ml in 5 ml/well). secreting cells were also reported in another model of B cell Phosphorothioate CpG ODN 2006 (10 mg/ml; Sigma-Aldrich) (22), histidine-tagged recombinant human soluble CD40L (sCD40L; 50 ng/ml) differentiation mimicking T cell help deploying CD40L, IL-2, and and anti-polyhistidine mAb (5 mg/ml) (R&D Systems) were added at IL-10 activation of MBCs (20). Given their poor ability to survive culture start. In step two, PB generation, at D4 culture, the cells were 2 upon CD40 activation removal, these CD38 /low Ig-secreting cells harvested, washed, and seeded at 2.5 3 105/ml with IL-2 (20 U/ml), IL-6 were initially assumed to be early precursors of short-term sur- (50 ng/ml), IL-10 (50 ng/ml), and IL-15 (10 ng/ml). In step three, PC generation, at D7 culture, cells were washed and cultured at 5 3 105/ml viving plasma cells (20). In a following study, this group has with IL-6 (50 ng/ml), IL-15 (10 ng/ml), and IFN-a (500 U/ml) for 3 d. shown that the ability to secrete Igs was mainly restricted to CD27high cells, independently of CD38 expression (21). Cell cycle analysis and immunophenotypic analysis Using a similar model of PC generation based on initial CD40L The percentage of cells in the S phase of the cell cycle was determined using and oligodeoxynucleotide (ODN) activation of MBCs, we show propidium iodide, and data were analyzed with ModFit LT software (Verity in this study that CD382CD202/lowCD27low Ig-secreting cells Software House, Topsham, ME) (23). Cells were stained with combina- tion of four to seven mAbs conjugated to different fluorochromes. Surface (termed preplasmablasts, prePBs) precedes the generation of high high staining was performed prior to cell fixation and permeabilization. The CD38 CD27 PBs and then PCs. PrePBs fully survive and Cytofix/Cytoperm (BD Biosciences) was used for intracellular staining differentiate into PBs and PCs upon removal of CD40 activation. of IgM, IgA, IgG, and Ki67 Ag, according to the manufacturer’s recom- A full molecular and phenotypic characterization of prePBs is mendations. Flow cytometry analysis was performed with a FACSAria provided, showing that these cells express specifically CD30, IL- cytometer using FACSDiva 6.1 (Becton Dickinson, San Jose, CA) and with a Cyan ADP cytometer driven by Summit software (Beckman Coulter). 6R, and cytoplasmic Igs, express weakly CD27, do not express For data analysis, CellQuest (Becton Dickinson), Summit, Kalusa (Beck- B cell or PC markers (CD38 and CD138), and secrete Igs at man Coulter), and Infinicit 1.3 (Cytognos, Salamanca, Spain) softwares The Journal of Immunology 3933 were used. The fluorescence intensity of the cell populations was quanti- fied using the stain index (SI) formula: (mean fluorescence intensity obtained from the given mAb minus mean fluorescence intensity obtained with a control mAb)/(2 3 SD mean fluorescence intensity obtained with the same control mAb) (24). The extensive phenotypic study of tonsil cells was carried out using the multicolor flow cytometry methodology and Abs we previously reported in detail (25). Analysis of Ig secretion ELISPOT. D4 CD202CD382, D7 CD202CD38+, and D10 CD202CD138+ cells were purified using cytometry cell sorting and cultured for 18 h in ELISPOT plates (Millipore, Bedford, MA), which were precoated with goat anti-human IgM, IgA, or IgG polyclonal Abs (Caltag Laboratories, Birlingame, CA). After nine PBS washings, alkaline phosphatase-conju- gated goat anti-human IgM, IgA, or IgG Abs (Tebu-Bio, Le Perray- en-Yvelines, France) were added for 6 h at 37˚C. After three PBS washings, a mixture of NBT/5-bromo-4-chloro-3-indolyl phosphate substrate (Sigma-Aldrich) was added and reaction was stopped with distilled water. IgM-, IgA-, and IgG-secreting-cells were enumerated and immunospot size was assessed using the Biosys Bioreader 5000 apparatus (Biosys, Miami, FL). ELISA. D4 CD202CD382, D7 CD202CD38+, and D10 CD202CD138+ cells were cytometry cell sorted, cultured for 24 h, and culture supernatants Downloaded from were harvested. IgM, IgA, and IgG secretions were assessed by ELISA using goat anti-human Igs for coating and secondary HRP-conjugated Abs specific for human g-, a-, or m-H chains, respectively (all from Jackson ImmunoResearch Laboratories, Newmarket, U.K.) as previously described (18). FIGURE 1. A 4-d activation of MBCs by sCD40L and CpG ODN yields two cell populations. MBCs were purified and cultured for 4 d using

Real-time RT-PCR analysis http://www.jimmunol.org/ sCD40L and CpG ODN activation and IL-2 plus IL-10 plus IL-15 cyto- Total RNA was extracted using the RNeasy kit (Qiagen, Valencia, CA) and kines. At D4, the cell phenotype was assayed with fluorochrome-conju- reverse transcribed with a reverse transcription kit (Qiagen). The Assays-on- gated anti-CD20 and anti-CD38 mAbs, or isotype-matched control mAbs. Demand primers and probes and the TaqMan Universal Master Mix were used according to the manufacturer’s instructions (Applied Biosystems, Fluorescence was determined with a FACScan device. Results are those of Courtaboeuf, France). The primers used for assessing specific expression 1 experiment representative of 17. A, Purification and phenotype of D0 of spliced or unspliced forms of XBP1 mRNA were from Applied Bio- starting MBCs. B, Phenotype of D4-activated cells. The vertical and hor- systems (Hs0329085 and Hs02856596 primers). Real-time RT-PCR was izontal lines in the second dot plot in A (right panel) and B correspond to performed using the ABI Prism 7000 sequence detection system and the positive threshold defined with fluorochrome-conjugated isotype con- normalized to b2-microglobulin for each sample, and compared with the trol mAbs. The second vertical half line in B defines the border between high low/2 values obtained for a known positive control using the following equation: CD20 and CD20 populations. by guest on September 29, 2021 100/2DDCt, where DDCt=DCt unknown 2 DCt positive control as de- scribed (26). with our bioinformatics platforms (RAGE, http://rage.montp.inserm.fr/) Microarray hybridization and bioinformatic analysis (27) and Amazonia (http://amazonia.transcriptome.eu/) (28). The cluster- ing was performed and visualized with the Cluster and TreeView softwares RNA was extracted and hybridized to U133 Plus 2.0 (29). Gene differentially expressed between cell populations was de- GeneChip microarrays, according to the manufacturer’s instructions termined with the significance analysis of microarray (SAM) statistical (Affymetrix, Santa Clara, CA). data from D4 CD20high 2 2 2 2 microarray analysis software (30). CD38 cells, D4 CD20lowCD38 cells, and DCD20 CD38 cells are deposited in the ArrayExpress public database (http://www.ebi.ac.uk/ Statistical analysis microarray-as/ae/, accession no. E-MEXP-3034). Gene expression data from D0 MBCs, D7 PBs, D10 PCs and purified BM PCs were from the Statistical comparisons were made with the nonparametric Mann–Whitney ArrayExpress public database (http://www.ebi.ac.uk/microarray-as/ae/, U test and the unpaired or paired Student t test using SPSS software. A p accession number E-MEXP-2360). Gene expression data were analyzed value of #0.05 was considered as significant.

Table I. Expression of surface and cytoplasmic IgH isotypes by MBCs and D4 activated cells

D0 D4

MBCs CD20highCD382 CD20lowCD382 CD202CD382 CD202CD38+

% SI % SI % SI % SI % SI sIgM 43 6 12 36 6 8656 3106 1656 596 3316 2a,b 7 6 2306 1a,b 6 6 1 cyIgM 45 6 10 9 6 2606 11 30 6 8446 14a 115 6 41a 18 6 13a,b 135 6 71a 21 6 14a,b 173 6 78a sIgA 27 6 6256 10 20 6 5116 1146 2176 2296 5b 17 6 1a 37 6 5b 16 6 1a cyIgA 25 6 6486 15 25 6 5376 8256 9 154 6 38a 33 6 10 172 6 44a 44 6 12a,b 209 6 61a,b sIgG 26 6 5336 396 276 3156 1a 12 6 4286 3a,b 12 6 5276 9186 9 cyIgG 27 6 5106 4206 6116 4296 6a 24 6 7a 41 6 4a,b 32 6 10a,b 33 6 7a 48 6 13a,b Ki67 2 6 1NA936 1246 2966 1546 6a 98 6 0a 60 6 8a,b 100 6 0a 77 6 11 S-phase 0.5 6 0NA256 8NA566 4a NA 56 6 3a NA 46 6 12a NA MBCs were cultured as described in Fig. 1. Starting MBCs and D4-activated cells were labeled with fluorochrome-conjugated anti-CD20 and anti-CD38 mAbs, and with fluorochrome-conjugated anti-human IgM, IgA, IgG, and Ki67 mAbs or isotype-controlled mAbs before or after cell permeabilization or with propidium iodide. For each cell population, data are the mean percentage 6 SD of positive cells and the mean staining indexes 6 SD from three to five separate experiments. aMean expression is significantly different from that in D4 CD20highCD382 cells. bMean expression is significantly different from that in D4 CD20lowCD382 cells. NA, not applicable. 3934 CHARACTERIZATION OF HUMAN PREPLASMABLASTS

based on CD20 and CD38 expressions at the end of a 4-d activation by sCD40L and CpG ODN (step one). These two populations were CD20highCD382 cells and CD20low/2 cells comprising, re- spectively, 18 6 7 and 82 6 7% of D4 cells (Fig. 1B). The CD20 and CD38 FACS dot plot data suggest that the CD20low/2 cell population may account for a continuous wave of differentiation of CD20lowCD382 cells (27 6 7% of D4 cells) into CD202CD38+ cells (17 6 7% of D4 cells) passing through a transitional CD202 CD382 stage (30 6 10% of D4 cells) (Fig. 1B). To demonstrate this, we further investigated the characteristics of D4 CD20low CD382, CD202CD382, and CD202CD38+ subsets (referred to below as CD20low/2 subsets). Surface and cytoplasmic Ig expression and cell cycle in D4 CD20low/2 cell subsets and in D4 CD20highCD382 cells The expression of surface (s) (sIgM, sIgA, sIgG) and cytoplasmic (cy) IgH isotypes (cyIgM, cyIgA, cyIgG) was looked for in the various D4 cell populations (Table I). Starting MBCs comprised + + the expected percentages of sIgM (43 6 12%), sIgA (27 6 6%), Downloaded from and sIgG+ (26 6 5%) cells (25, 31), and cell permeabilization did not increase the cytometry SI significantly, indicating that MBCs expressed mainly sIgs (Table I). In CD20lowCD382,CD202 CD382, and CD202CD38+ cells, the SI with anti-IgM or anti-IgA 2 2 FIGURE 2. Ig secretion by D4 CD20 CD38 cells, D7 PBs, and D10 Abs increased 10- to 20-fold after cell permeabilization (p , PCs. D4 CD202CD382 cells, D7 CD202CD38+ PBs, and D10 CD202 + 0.05), indicating that these cells were strongly involved in cyIg http://www.jimmunol.org/ CD138 PCs were flow cytometry cell sorted and cultured for 18 h in production (Table I, Supplemental Fig. 1). The increase in SI with ELISPOT plates or cultured for 24 h and culture supernatants were har- the anti-IgG mAb was significant (p , 0.05) but less pronounced vested. The number of IgM-, IgA-, and IgG-secreting cells was assessed by ELISPOT. Results are (A) the mean number of IgM-, IgA-, and IgG-se- (2- to 4-fold) due to the use of allophycocyanin- and PerCP- creting cells, and (B) the mean number of immunospots with a size $200 Cy5.5-fluorochome–conjugated Abs, resulting in a lower SI as 2 2 2 + mm determined in three separate experiments. C, IgM, IgA, and IgG reported (32). Of note, CD20 CD38 and CD20 CD38 cells secretions were assessed by ELISA and results are the mean Ig produc- comprised increased IgG+ and decreased IgM+ cell numbers 2 tion in micrograms per day and per 106 cells determined in three separate compared with CD20lowCD38 cells (p , 0.05; Table I). Con- 2 experiments. *The mean is significantly different from that in D4 CD20 versely, D4 CD20highCD382 cells expressed mainly sIgM+ (65 6 2 CD38 cells (p # 0.05). **The mean is significantly different from that in 3%), and cell permeabilization slightly increased the SI for IgM by guest on September 29, 2021 2 + D7 CD20 CD38 PBs (p # 0.05). and IgA (p , 0.01), unlike IgG. Additionally, CD20low/2CD382 cells had a lower cytoplasmic Ig content compared with CD202 Results CD38+ cells as indicated by fluorescence intensity (Table I, Sup- Activation of MBCs by sCD40L and CpG ODN yields two plemental Fig. 1). The level of Ig secretion by the D4 CD202 distinct cell populations CD382 cell subset was further compared with that of in vitro- In the first step of differentiation of MBCs into PCs, two distinct generated D7 PBs or D10 PCs using ELISPOT and ELISA. cell populations could be identified in 17 separate experiments There were 8% IgG-secreting cells in D4 CD202CD382 cells,

FIGURE 3. Plasmablast and plasma cell differentiation potential of D4 CD20low/2CD382 prePBs and of D4 CD20highCD382 cells. Purified MBCs were cultured for 4 d using sCD40L and CpG ODN activation and IL-2 plus IL- 10 plus IL-15 cytokines. At D4, CD20high CD382,CD20lowCD382,andCD202 CD382 cells were FACS sorted and cultured for 3 d with PB culture con- ditions (IL-2 plus IL-6 plus IL-10 plus IL-15). D7 cells were washed and cul- tured for an additional 3 d with plasma cell culture conditions (IL-6 plus IL-15 plus IFN-a). At each culture step, the cell phenotype was assayed with fluoro- chrome-conjugated anti-CD20, anti- CD38, and anti-CD138 mAbs, or with isotype-matched control mAbs. Fluores- cence was determined with a FACScan device. Results are those of one experi- ment representative of four, and the fold expansions are the mean values 6 SD of those in four experiments. The Journal of Immunology 3935 that is, 2.5- and 3.7-fold less than in D7 PBs or D10 PCs (p , 0.02; Fig. 2A), and with a 3- to 7-fold smaller immunospot size ($200 mm; Fig. 2B). The number of IgA-secreting cells in D4 CD202CD382 cells was not significantly different from those in D7 PBs or D10 PCs (Fig. 2A). There were 3–5% IgM-secreting cells in the D4 CD202CD382 cell subset as in D7 PBs and D10 PCs (Fig. 2A), with no difference in IgM immunospot size. The measurement of Ig production in culture supernatants confirmed ELISPOT data. D4 CD202CD382 cells secreted 5- to 7-fold less IgG, 1.5- to 2-fold less IgA, and similar IgM levels than did D7 PBs or D10 PCs (Fig. 2C). Cell cycling was looked for with anti-Ki67 and propidium io- dide staining (Table I). Starting MBCs were quiescent and all D4 cell populations were cycling with 93–100% Ki67+ cells. CD20low CD382, CD202CD382, and CD202CD38+ cells had a higher proportion of cells in the S phase (46–56%) than did CD20high CD382 cells (25%, p , 0.05). These results indicate that the D4 CD20low/2 cell population 2 2 comprised mostly CD20low/ CD38 cells that expressed cyIgs Downloaded from and secreted Igs, but at a weaker level than D7 PBs and D10 PCs. Given the Ig secretion, the lack of CD38 and the lack or weak expression of CD20 B cell marker, these CD20low/2CD382 cells are termed prePBs in the following. D4 CD20low/2CD382 prePBs generate CD202CD38high PBs http://www.jimmunol.org/ and then CD138+ plasma cells, unlike D4 CD20highCD382 cells The PC differentiation potential of D4 CD20low/2CD382 prePBs was assayed culturing these cells for three additional days in PB step two culture conditions (Fig. 3). CD20low/2CD382 cells were FIGURE 4. Unsupervised and supervised clustering of gene expression low 2 2 2 sorted into CD20lowCD382 and CD202CD382 cells to evaluate profiles of purified D4 CD20 CD38 and CD20 CD38 prePBs and of D4 CD20highCD382 B cells. MBCs from five donors (numbered 1–5) were whether these two populations had a similar differentiation po- cultured for 4 d with sCD40L, CpG ODN, and IL-2 plus IL-10 plus IL-15 tential and thus could actually correspond to a unique cell subset. cytokines. D4 CD20highCD382, CD20lowCD382, and CD202CD382 cells This was the case since a vast majority of both D4 CD20low by guest on September 29, 2021 2 2 2 2 were FACS sorted and their gene expression profiling was assayed with CD38 and D4 CD20 CD38 cells differentiated into D7 CD20 Affymetrix U133 Plus 2.0 microarrays. A, An unsupervised hierarchical high CD38 PBs in four separate experiments (73 6 15 and 87 6 clustering was run with the 5000 probe sets with the highest SD (log 7%, respectively) in association with a 4.2- or 3.3-fold cell ex- transform, center genes and arrays, uncentered correlation, and average 2 pansion, respectively (Fig. 3). When further put into step three linkage). The dendogram shows that CD20highCD38 samples and 2 2 plasma cell culture conditions for 3 d (D7 to D10), the sorted CD20low/ CD38 samples are grouped in two separate clades (r = 0.44). low/2 2 low 2 2 2 CD20lowCD382 or CD202CD382 cells generated CD202 In the CD20 CD38 clade, CD20 CD38 and CD20 CD38 cells CD38highCD138+ PCs (51 6 7 and 72 6 5%, respectively; n =4; originating from the same donor cluster together for three of five donors, high 2 indicating that these two populations are close. B, The probe sets differ- Fig. 3). Conversely, D4 CD20 CD38 cells gave rise to few high 2 2 2 + 2 6 entially expressed between D4 CD20 CD38 and CD20 CD38 cells CD38 CD20 cells (23 13%, n = 4) in step two culture con- were determined with a SAM supervised analysis for pairs (Wilcoxon ditions, and did not survive when further put in step three plasma statistic, 2-fold ratio, 0% false discovery rate), identifying 495 unique cell culture conditions (Fig. 3). Thus, these in vitro differentiation genes. When a gene was assayed by several probe sets, the probe set with 2 data reveal the presence of only two populations of CD38 cells at the highest variance was used. An unsupervised hierarchical clustering was 2 2 D4 of culture: first, actively cell cycling CD20low/ CD38 prePBs run on this 495 unique gene list. The normalized expression value for each that efficiently differentiated into CD202CD38high PBs and then gene is indicated by a color, with red representing high expression and CD202CD38highCD138+ PCs; and second, CD20highCD382 cells green representing low expression. that poorly differentiated and survived into PB and then PC cul- ture conditions. pervised analysis, since only 23 probe sets of 5000 were low 2 2 low/2 2 differentially expressed between CD20 CD38 and CD20 Molecular characterization of D4 CD20 CD38 prePBs CD382 cells (ratio $ 2, 1000 permutations, false discovery rate The gene expression of sorted D4 CD20highCD382, CD20low 0%; results not shown). Comparing CD202CD382 and CD20high CD382, and CD202CD382 cells obtained from MBCs from five CD382 cells, 695 probe sets fully discriminated the two cell healthy donors was profiled using Affymetrix U133 Plus 2.0 populations using paired SAM supervised analysis (ratio $ 2, microarrays. An unsupervised hierarchical clustering using the 1000 permutations, 0% false discovery rate) (Fig. 4B). They 5000 probe sets with the highest SD made it possible to cluster encoded for 495 unique genes, of which 267 were upregulated in CD20highCD382 cells together and CD20low/2CD382 prePBs in CD20highCD382 cells and 228 in CD202CD382 cells. Table II another cluster (Fig. 4A). Of note, inside the CD20low/2CD382 shows the top 50 genes that were upregulated in each of the two cluster, CD20lowCD382 and CD202CD382 cells originating from cell populations, and the 495 unique genes are shown in Supple- a same donor were grouped together for three of the five donors, mental Table I. CD202CD382 prePBs highly expressed Ig genes emphasizing that these two cell subsets constituted a homoge- in agreement with their increased cytoplasmic Ig expression, neous population (Fig. 4A). This was confirmed using paired su- IL-6R, BCMA (TNFRSF17), and CD59 that are PC markers. 3936 CHARACTERIZATION OF HUMAN PREPLASMABLASTS

Table II. CD202CD382 and CD20highCD382 top 50 overexpressed genes

CD202CD382 Genes CD20highCD382 Genes

Gene ID Gene Name Fold Change Score (d) Gene ID Gene Name Fold Change Score (d) 205945_at IL6R 18.41 4.44 232739_at SPIB 0.01 22.68 235275_at BMP8B 15.78 3.16 224402_s_at FCRL4 0.01 25.48 203914_x_at HPGD 14.66 2.65 226818_at MPEG1 0.01 22.91 204798_at MYB 10.80 7.09 235385_at MARCH-I 0.01 23.98 222392_x_at PERP 9.68 3.86 223343_at MS4A7 0.01 24.75 203373_at SOCS2 9.35 5.81 204959_at MNDA 0.01 24.11 221790_s_at LDLRAP1 8.87 3.35 266_s_at CD24 0.01 24.43 201324_at EMP1 8.67 3.79 204249_s_at LMO2 0.01 214.17 206641_at TNFRSF17/BCMA 6.86 5.37 219517_at ELL3 0.01 28.97 201243_s_at ATP1B1 6.67 2.67 228055_at NAPSB 0.02 29.02 233500_x_at CLEC2D 6.45 3.15 217418_x_at MS4A1/CD20 0.02 214.56 201678_s_at DC12 6.22 3.95 205987_at CD1C 0.03 27.57 39248_at AQP3 6.08 9.19 231093_at FCRH3 0.03 25.07 200983_x_at CD59 5.32 2.49 228153_at IBRDC2 0.03 22.97 206729_at TNFRSF8/CD30 5.25 2.61 207339_s_at LTB 0.04 24.80 204254_s_at VDR 5.20 3.22 208018_s_at HCK 0.05 25.33

221760_at MAN1A1 5.04 12.00 208820_at PTK2 0.05 22.79 Downloaded from 220306_at FAM46C 5.01 3.56 210279_at GPR18 0.05 25.82 209457_at DUSP5 4.90 3.08 238009_at SOX5 0.06 25.17 205885_s_at ITGA4 4.73 2.58 213293_s_at TRIM22 0.06 27.82 203397_s_at GALNT3 4.61 4.02 219014_at PLAC8 0.06 24.11 206632_s_at APOBEC3B 4.61 4.79 228617_at BIRC4BP 0.06 22.80 217127_at CTH 4.59 5.13 206126_at BLR1 0.07 24.05 1554242_a_at COCH 4.59 3.11 204581_at CD22 /// MAG 0.07 24.39 202241_at TRIB1 4.58 9.99 216080_s_at FADS3 0.07 23.79 http://www.jimmunol.org/ 224802_at NDFIP2 4.52 2.81 221234_s_at BACH2 0.07 24.31 209921_at SLC7A11 4.52 3.40 213111_at PIP5K3 0.07 211.02 201397_at PHGDH 4.52 2.93 225123_at SESN3 0.08 24.27 203474_at IQGAP2 4.44 5.23 215933_s_at HHEX 0.08 25.42 200951_s_at CCND2 4.13 8.53 205128_x_at PTGS1 0.08 22.75 203066_at GALNAC4S-6ST 4.09 3.15 218032_at SNN 0.09 22.97 218018_at PDXK 4.08 2.66 205922_at VNN2 0.09 22.89 219118_at FKBP11 4.03 2.99 204440_at CD83 0.09 27.96 204900_x_at SAP30 3.97 4.52 203186_s_at S100A4 0.09 23.34 2

202558_s_at STCH 3.97 4.01 221011_s_at LBH 0.10 4.42 by guest on September 29, 2021 224851_at CDK6 3.96 3.14 204994_at MX2 0.10 23.19 211464_x_at CASP6 3.92 2.59 206983_at CCR6 0.10 24.88 200628_s_at WARS 3.87 4.70 200696_s_at GSN 0.10 22.79 217824_at UBE2J1 3.85 5.86 219836_at ZBED2 0.10 27.02 225512_at ZBTB38 3.78 7.76 1559263_s_at ZC3H12D 0.10 26.97 202468_s_at CTNNAL1 3.67 3.44 227458_at PDCD1LG1 0.10 25.18 218073_s_at TMEM48 3.60 3.70 38521_at MAG 0.10 27.58 228964_at PRDM1 3.57 4.17 230110_at MCOLN2 0.11 25.00 209695_at PTP4A3 3.55 3.79 223751_x_at TLR10 0.11 25.58 203971_at SLC31A1 3.53 3.51 229937_x_at LILRB1 0.11 24.08 216044_x_at FAM69A 3.50 3.58 226748_at LYSMD2 0.11 23.05 225520_at MTHFD1L 3.49 9.94 203233_at IL4R 0.11 23.57 222385_x_at SEC61A1 3.47 2.86 224499_s_at AICDA 0.12 25.11 201206_s_at RRBP1 3.44 3.22 215127_s_at RBMS1 0.12 23.48 226771_at ATP8B2 3.41 4.57 1552807_a_at SIGLEC10 0.12 22.99

Additionally, they expressed CD30 (TNFRSF8) and EMP1 (Table (CD30, IL-6R, and CD43), in agreement with gene expression II, Supplemental Table I). Conversely, CD20highCD382 cells data. highly expressed genes for B cell Ags, that is, CD19, CD20 B to plasma cell atlas (MS4A1), CD22, and CD24 (Table II, Supplemental Table I). low/2 2 An open access “B to Plasma Cell Atlas” was developed allowing Phenotype of D4 CD20 CD38 cells interrogating gene expression of D4 CD20low/2CD38- prePBs and FACS analysis confirmed Affymetrix data. D4 CD20low/2CD382 CD20highCD382 B cells together with those of MBCs, PBs, early cells expressed higher levels of IL-6R, CD30, CCR10, and CD43 PCs, and BMPCs (31). For convenient atlas use, D4 CD20lowCD382 than did D4 CD20highCD382 cells (Fig. 5, Table III). Conversely, cells, D4 CD202CD382 cells, and D4 CD20highCD382 cells are D4 CD20highCD382 cells expressed higher levels of CD24, CD19, termed CD20low prePBs, CD202 prePBs, and CD20high B cells. This CD45, CXCR5, and CD27 than did D4 CD20low/2CD382 cells open access atlas can be visualized on our Amazonia Web site (Fig. 5, Table III). Of note, CD27 dramatically increased again (http://amazonia.transcriptome.eu/index.php?zone=PlasmaCell). when D4 CD20low/2CD382 cells differentiated into D7 PBs and Fig. 6A displays the expression of genes coding for transcription D10 PCs as documented previously (31). D4 CD20lowCD382 cells factors controlling B cell or PC fates, and Fig. 6B displays had a phenotype close to that of CD202CD382 cells, but with the validation of these gene expressions by real-time RT-PCR. a significantly weaker expression of prePB and PB markers CD20low/2 prePBs weakly expressed PAX5 B cell transcription The Journal of Immunology 3937 Downloaded from http://www.jimmunol.org/

FIGURE 5. Phenotype and expression of homing molecules of D4 CD20lowCD382 and D4 CD202CD382 prePBs and of D4 CD20highCD382 cells. MBCs were cultured as described in Fig. 1. Cells were stained for CD20 and CD38 and the cell phenotype was analyzed by gating on D4 CD20highCD382 low 2 2 2

cells, D4 CD20 CD38 cells, and D4 CD20 CD38 cells. Black histograms show FACS labeling with anti-CD24, IL-6R, CD30, CD19, CD45, CXCR5, by guest on September 29, 2021 CD27, CCR10, and CD43. Gray histograms display the corresponding negative control mAbs. Data from one experiment representative of three to five are shown. factor and PAX5 target genes (IRF8, BACH2, EBF1, SPIB) (Fig. 5-fold in D7 PBs and D10 PCs that expressed XBP1s mRNA 6A,6B) as well as BCL6 gene (Fig. 6B). CD20low/2 prePBs mainly (Fig. 6C). The expression of genes coding for B cell and PC expressed PC genes PRDM1 and XBP1 at surface markers or homing molecules are displayed in Supple- a weaker level than PBs or early PCs (Fig. 6A). Because XBP1 mental Figs. 2 and 3. Affymetrix data are in agreement with the mRNA has to be spliced by IRE1 endonuclease to encode for an phenotype of prePBs reported above: lack of B cell Ags and of active XBP1 (33), spliced XBP1 (XBP1s) and unspliced CD38. Of interest, prePBs specifically expressed CD30 and EMP1 (XBP1u) mRNAs were quantified. D4 prePBs expressed high levels genes, unlike B cells, PBs, early PCs, or BMPCs. PrePB genes were of XBP1 mRNA, mainly XBP1u mRNA. The XBP1s/XBP1u significantly higher and B cell genes lower in CD202 prePBs than mRNA ratio was 0.3 in prePBs, and this ratio was increased 3.5- to in CD20low prePBs.

Table III. Expression of membrane markers by MBCs and D4 activated cells

D0 MBCs D4 CD20highCD382 D4 CD20lowCD382 D4 CD202CD382

Membrane Markers % SI % SI % SI % SI CD24 88 6 4826 15 68 6 21 25 6 41 31 6 36a 5 6 7a 5 6 4a,b 0.7 6 0.5a,b IL-6R 0.6 6 0.9 0 6 006 006 0266 8a 0.5 6 0.1a 60 6 9a,b 1.3 6 0.4a,b CD30 0 6 006 0306 5 0.9 6 0.4 70 6 7a 7 6 0a 87 6 4a 17 6 2a,b CD19 100 6 0456 13 100 6 0 225 6 73 92 6 3a 64 6 12a 89 6 4a,b 41 6 6a,b CD45 100 6 0 718 6 51 100 6 0 782 6 497 100 6 0 425 6 245a 100 6 0 353 6 193a,b CXCR5 97 6 276 1986 0206 4616 11a 6 6 2a 15 6 2a,b 1.5 6 0.4a CCR10 1.3 6 0.5 0 6 0216 9 1.7 6 1.1 42 6 6a 4.2 6 1.6 41 6 5 3.9 6 1.4 CD43 9 6 1 2.0 6 0.4 18 6 3 1.8 6 0.6 14 6 3 1.3 6 0.5 31 6 6b 3.6 6 1.5 CD27 100 6 096 1846 7316 9666 10a 20 6 7a 42 6 12a,b 11 6 4a,b MBCs were cultured as described in Fig. 1. Starting MBCs, D4-activated cells were labeled with fluorochrome-conjugated anti-CD20 and anti-CD38 mAbs and with indicated fluorochrome-conjugated mAbs or isotype-controlled mAbs. For each cell population, data are the mean percentage 6 SD of positive cells and the mean staining indexes 6 SD from three to five separate experiments. aMean expression is significantly different from that in D4 CD20highCD382 cells. bMean expression is significantly different from that in D4 CD20lowCD382 cells. 3938 CHARACTERIZATION OF HUMAN PREPLASMABLASTS

In vivo detection of CD20low/2CD382 prePBs The presence of prePBs, defined as CD19+CD20low/2CD382 CD1382 cells, was looked for in human tonsils and reactive lymph nodes. Fig. 7A shows that gating consecutively on CD19+ cells and then CD202 cells made it possible to detect a minor population of CD382 and CD1382 cells in seven of seven tonsil samples and in seven of nine lymph node samples. This CD19+ CD20low/2CD382CD1382 population accounted for 0.06 6 0.02 (range, 0.02–0.08%) and 0.05 6 0.03% (range, 0.02–0.09%) of CD19+ cells in tonsil and positive lymph node samples, re- spectively. CD19+CD20low/2CD382CD1382 cells could not be detected in BM (five samples) and peripheral blood (five samples, data not shown). Looking for cytoplasmic Ig expression, approx- imately two thirds of these CD19+CD20low/2CD382CD1382 cells expressed high levels of Ig L chains after cell permeabilization (Fig. 7B). The other third expressed one log weaker levels of Ig L chain after cell permeabilization and we previously showed these cells are B cells that express surface Ig L chains (25). Fig. 7C shows that a large part of in vivo CD19+CD20low/2CD382 Downloaded from CD1382 cells also expressed IL-6R, unlike CD19+CD20+ cells. Thus, 45 6 5% of CD19+CD20low/2CD382CD1382 cells are IL- 6R+ in tonsils (range, 36–51%) and 36 6 14% are IL-6R+ in lymph nodes (range, 18–57%). http://www.jimmunol.org/ Discussion Owing to the rarity of plasma cells and their anatomic location, the early stages of differentiation of B cells into plasma cells are not fully elucidated, in particular in . Using an in vitro model mimicking T cell help, we report a transitional prePB stage in the differentiation of human memory B lymphocytes into PBs, and then plasma cells. These prePBs are characterized by the loss of B cell markers, except CD19, a decrease in CD27, the lack of

CD38, and by the secretion of Igs at a lower level than CD38+ PBs by guest on September 29, 2021 or CD138+ plasma cells. Their prePB status is evidenced by their ability to fully differentiate into PBs that highly express CD38 within 3 d (CD19+CD202CD38high cells) and then plasma cells that express CD138 within additional 3 d (CD19+CD202CD38high CD138+ cells). Human CD382/low Ig-secreting cells were already reported but with challenging results regarding their behavior and phenotype. With a model of MBC activation by CD40L and cytokines close to the current one, Tangye et al. (20) reported that CD382 B cells poorly survived upon removal of CD40L activa- tion and were thought to be precursors of short-living plasma cells. In the current study, sorted CD20low/2CD382 prePBs effi- ciently differentiated into PBs and then plasma cells. An expla- nation could be the use of whole CD382 B cells in the study by Tangye et al. comprising activated B cells (here CD20highCD382 FIGURE 6. Expression of genes coding for transcription factors gov- cells) and prePBs, avoiding an efficient differentiation of prePBs. erning B cell to PC differentiation. The gene expression of the 54613 In a following study published by Avery et al. (21) using a similar Affymetrix probe sets in various B cell and plasma cell populations can be in vitro model, precursors of Ig-secreting cells were shown to be visualized using the Amazonia Web site (http://amazonia.transcriptome.eu/ high high 2 mainly CD27 cells, independently of CD38 expression, and index.php?zone=PlasmaCell): MBCs (clear blue), D4 CD20 CD38 CD27low cells had a reduced cell cycling. These data also do not fit cells termed CD20high B cells (dark blue), D4 CD20lowCD382 cells termed 2 2 2 with the current ones since prePBs have a reduced CD27 gene and CD20low prePBs (blue), D4 CD20 CD38 cells termed CD20 prePBs protein expression compared with MBCs, PBs, or plasma cells. (dashed pink), PBs (pink), early PCs (clear violet), and BMPCs (dark vi- low olet) gene expression profiling of prePBs, PBs, early PCs, and BMPCs Additionally, the CD27 prePBs were highly proliferating with were available from ArrayExpress public database (http://www.ebi.ac.uk/ microarray-as/ae/, accession nos. E-MEXP-3034 and E-MEXP-2360). A, Affymetrix signals of expression of genes coding for transcription factors level in the different cell populations was compared assigning the arbitrary controlling B cell and PC cell fate. Data are the mean value 6 SD of gene value 1 to the maximal expression. Data are the mean value 6 SD of mRNA expression determined in five separate experiments. B, Validation of gene level or ratio determined in three separate experiments. *The mean ex- expression of PAX5, BCL6, IRF4, PRDM1, and XBP1 by real-time RT- pression is significantly different from that in CD20high B cells (p # 0.05). PCR. C, Quantification of XBP1s and XBP1u mRNA by real-time RT-PCR **The mean expression is significantly different from that in CD20low and XBP1s/XBP1u ratio in CD202 prePBs, PBs, and PCs. The mRNA prePBs (p # 0.05). The Journal of Immunology 3939 Downloaded from http://www.jimmunol.org/

low/2 2 + FIGURE 7. Detection of CD20 CD38 IL-6R prePBs in human lymphoid tissues. Mononuclear cells isolated from tonsils (seven donors) and lymph by guest on September 29, 2021 nodes (nine donors) were stained for CD19, CD20, CD38, CD138, and IL-6R. A, A consecutive gating strategy on CD19+ and then CD202 cells made it possible to identify CD19+CD20low/2CD382CD1382 prePBs in lymphoid tissue samples. Numbers in the panels are the mean percentages 6 SD of prePBs in CD19+ cells in seven tonsils or nine lymph nodes. B, Cytoplasmic and membrane Ig L chain expression by CD19+CD20low/2CD382CD1382 prePBs. A high expression of k (F) and l (I) L chains corresponded to cytoplasmic L chains, and an intermediate expression to membrane L chains. Data are from one representative experiment of two separate experiments. C, CD19+CD20low/2CD382CD1382 prePBs IL-6R expression compared with CD19+CD20+ B cells. The bold histograms represent labeling with anti–IL-6R mAb and the light ones, with the control mAb. Data are from one representative experiment using tonsil cells.

50% of cells in the S phase. A possible explanation for these activation by histidine-bound CD30L and anti-histidine Abs. No differences is that we used CpG ODN and CD40L stimulation of modulation was found in our culture conditions (data not shown). memory B cells, whereas Tangye et al. used CD40L stimulation The loss of CD20 and most B cell markers in prePBs could be only. CpG ODN is known to promote B cell differentiation acti- likely explained by the decrease in PAX5 that controls expression vating TLRs. Thus, either CpG ODN can downregulate CD27 of B cell genes (34–36). PAX5 gene expression decrease in prePBs expression on prePBs that further differentiate into PBs, or the is associated with an upregulation of expression of IRF4 and differentiation of CD27low prePBs may require both CD40L and PRDM1 genes and downregulation of BCL6. Thus, the usual TLR activation. Arce et al. (19) identified CD38low/2 Ig-secreting crossregulation of B and plasma cell transcription factors should cells in tonsils. These CD38low/2 Ig-secreting cells were shown to occur in prePBs: NF-kB induces upregulation of IRF4 that asso- weakly express CD27, in agreement with our current data. These ciates with STAT3 to trigger PRDM1 gene activation and induces cells were suggested to be precursors of CD38high Ig-secreting downregulation of BCL6 gene expression, an inhibitor of PRDM1 cells because they were detected earlier in culture of tetanus gene activation. The resulting expression of PRDM1 gene product, toxoid-activated B cells. However, this was not proven formally BLIMP1, further represses BCL6 and PAX5 genes, leading to the using cell sorting and culturing these CD38low/2 cells. release of XBP1 gene suppression by PAX5. XPB1 mRNA has to The relevance of this transitional CD19+CD202CD382 prePB be spliced by the IRE1 endonuclease to encode for XBP1 that stage is highlighted by the peaked expression of CD30 gene and masters the induction of the unfold protein response (33). Pre-PBs protein. Memory B cells, PBs, early plasma cells, and BM plasma expressed mainly XBP1u mRNA, and PBs and PCs expressed cells do not express CD30. CD30 is a TNFR family member that mainly the XBP1s mRNA with a 3.5- to 5-fold increase of the is expressed on activated T cells, EBV-infected B cells, some large XBP1s/XBP1u mRNA ratio compared with prePBs. These data diffuse lymphoma, Reed–Sternberg cells, and anaplastic large cell again argue for the prePB status of D4 CD20low/2CD382 cells that lymphoma. As CD30 is an activation molecule, we looked for start to secrete Igs, but at a weaker level than PBs or PCs, and a modulation of prePB proliferation or differentiation using CD30 likely exhibit a less pronounced unfold protein response. An 3940 CHARACTERIZATION OF HUMAN PREPLASMABLASTS interesting point will be to identify the transcription factors con- References trolling the transient CD30 gene expression in prePBs, and then 1. Coffey, F., B. Alabyev, and T. Manser. 2009. Initial clonal expansion of germinal the expression of CD38 and of its ligand, CD31, in PBs. Cells with center B cells takes place at the perimeter of follicles. Immunity 30: 599–609. 2. Allen, C. D., T. Okada, and J. G. Cyster. 2007. Germinal-center organization and the main characteristics of these in vitro PBs could be identified cellular dynamics. Immunity 27: 190–202. in tonsils and lymph nodes at a low frequency gating on CD19+ 3. Allen, C. D., and J. G. Cyster. 2008. Follicular dendritic cell networks of primary CD20low/2CD382 cells. They accounted for 0.06% and 0.05% of follicles and germinal centers: phenotype and function. Semin. Immunol. 20: 14– + 25. CD19 cells in tonsils and positive lymph nodes, and they were 4. Park, C. S., S. O. Yoon, R. J. Armitage, and Y. S. Choi. 2004. Follicular dendritic not detected in peripheral blood and BM. We previously reported cells produce IL-15 that enhances germinal center B cell proliferation in membrane-bound form. J. Immunol. 173: 6676–6683. that PBs and B cells can be clearly identified by multicolor FACS 5. Yoon, S. O., X. Zhang, P. Berner, B. Blom, and Y. S. Choi. 2009. Notch ligands analysis using cell permeabilization and anti-k or l Ig L chain expressed by follicular dendritic cells protect germinal center B cells from ap- staining (25). PBs highly express k or l Ig L chains, and B cells optosis. J. Immunol. 183: 352–358. 6. Kranich, J., N. J. Krautler, E. Heinen, M. Polymenidou, C. Bridel, express intermediate levels. Using this technique, we found that A. Schildknecht, C. Huber, M. H. Kosco-Vilbois, R. Zinkernagel, G. Miele, and two-thirds of these cells expressed highly cytoplasmic Ig, and the A. Aguzzi. 2008. Follicular dendritic cells control engulfment of apoptotic other third expressed intermediate levels characteristic of surface bodies by secreting Mfge8. J. Exp. Med. 205: 1293–1302. low/2 7. Victora, G. D., T. A. Schwickert, D. R. Fooksman, A. O. Kamphorst, M. Meyer- Ig B cells. Arce et al. (19) also detected CD38 Ig-secreting Hermann, M. L. Dustin, and M. C. Nussenzweig. 2010. Germinal center dy- cells in tonsils, unlike peripheral blood and BM. These in vivo namics revealed by multiphoton microscopy with a photoactivatable fluorescent reporter. Cell 143: 592–605. cells expressed cytoplasmic Igs and IL-6R, as did in vitro prePBs, 8. Dogan, I., B. Bertocci, V. Vilmont, F. Delbos, J. Me´gret, S. Storck, unlike tonsil or lymph node B cells. CD30 could not be detected C. A. Reynaud, and J. C. Weill. 2009. Multiple layers of B cell memory with on these prePBs in vivo. This could be due to a rapid shedding of different effector functions. Nat. Immunol. 10: 1292–1299. Downloaded from 9. Mohr, E., K. Serre, R. A. Manz, A. F. Cunningham, M. Khan, D. L. Hardie, the protein in vivo following activation by CD30L (37). R. Bird, and I. C. MacLennan. 2009. Dendritic cells and monocyte/macrophages Recent immunochemistry and in vivo imaging studies have that create the IL-6/APRIL-rich lymph node microenvironments where plas- documented the traffic of maturing plasma cells from GC to mablasts mature. J. Immunol. 182: 2113–2123. 10. Fooksman, D. R., T. A. Schwickert, G. D. Victora, M. L. Dustin, medullary cords in mice (9, 10). In these studies, plasma cells were M. C. Nussenzweig, and D. Skokos. 2010. Development and migration of documented as IgG1-producing cells, which exit the GC to the plasma cells in the mouse lymph node. Immunity 33: 118–127. + 2 11. Kabashima, K., N. M. Haynes, Y. Xu, S. L. Nutt, M. L. Allende, R. L. Proia, and outer T cell zone in contact with CD11c CD8a dendritic cells http://www.jimmunol.org/ J. G. Cyster. 2006. Plasma cell S1P1 expression determines secondary lymphoid producing IL-6. They then migrate randomly though linear paths organ retention versus bone marrow tropism. J. Exp. Med. 203: 2683–2690. in the T cell zone and join the medullary cords in contact with 12. Radbruch, A., G. Muehlinghaus, E. O. Luger, A. Inamine, K. G. Smith, T. Do¨rner, and F. Hiepe. 2006. Competence and competition: the challenge of APRIL-producing monocytes/macrophages. These data fit well becoming a long-lived plasma cell. Nat. Rev. Immunol. 6: 741–750. with the molecular characterization and phenotype of prePBs 13. Jackson, S. M., P. C. Wilson, J. A. James, and J. D. Capra. 2008. Human B cell reported in this study. The need of T cell mimicking signals to get subsets. Adv. Immunol. 98: 151–224. 14. Jego, G., N. Robillard, D. Puthier, M. Amiot, F. Accard, D. Pineau, them in vitro suggests they are generated in the GC light zone J. L. Harousseau, R. Bataille, and C. Pellat-Deceunynck. 1999. Reactive plas- in vivo. The loss of CXCR5 expression in prePBs should abrogate macytoses are expansions of plasmablasts retaining the capacity to differentiate the attraction of these cells by CXCL13 producing FDC and T into plasma cells. Blood 94: 701–712. FH 15. Mei, H. E., T. Yoshida, W. Sime, F. Hiepe, K. Thiele, R. A. Manz, A. Radbruch, by guest on September 29, 2021 and CD40L activation and promote their migration to the outer and T. Do¨rner. 2009. Blood-borne human plasma cells in steady state are derived T cell zone. IL-6R expression by prePBs will prompt them to be from mucosal immune responses. Blood 113: 2461–2469. + 2 16. Caraux, A., M. Perez-Andres, M. Larroque, G. Requirand, Z. Y. Lu, T. Kanouni, stimulated by IL-6–producing CD11c CD8a dendritic cells. Of J. F. Rossi, A. Orfao, and B. Klein. 2011. Mobilization of plasma cells in healthy note, a removal of CD40L activation and a stimulation by IL-6 are individuals treated with granulocyte colony-stimulating factor for haemato- mandatory to induce prePBs to differentiate into PBs in the cur- poietic stem cell collection. Immunology 132: 266–272. 17. Medina, F., C. Segundo, A. Campos-Caro, I. Gonza´lez-Garcı´a, and J. A. Brieva. rent in vitro model (31), suggesting that this occurs in vivo when 2002. The heterogeneity shown by human plasma cells from tonsil, blood, and 2 prePBs migrate from GC light zone to CD11c+CD8a dendritic bone marrow reveals graded stages of increasing maturity, but local profiles of cells producing IL-6, which are found at the GC/T cell zone adhesion molecule expression. Blood 99: 2154–2161. 18. Caron, G., S. Le Gallou, T. Lamy, K. Tarte, and T. Fest. 2009. CXCR4 ex- border (9). The current identification of Ig-producing prePBs, pression functionally discriminates centroblasts versus centrocytes within human lacking CD20 and CD38, would allow for the precise location and germinal center B cells. J. Immunol. 182: 7595–7602. + 19. Arce, S., E. Luger, G. Muehlinghaus, G. Cassese, A. Hauser, A. Horst, traffic of these prePBs, compared with CD38 PBs and then K. Lehnert, M. Odendahl, D. Ho¨nemann, K. D. Heller, et al. 2004. CD38low IgG- + + CD38 CD138 plasma cells. PrePBs also express LFA-1 (ITGAL/ secreting cells are precursors of various CD38 high-expressing plasma cell ITGB2) and VLA-4 (ITGA4/ITGB1) integrins that could drive populations. J. Leukoc. Biol. 75: 1022–1028. 20. Tangye, S. G., D. T. Avery, and P. D. Hodgkin. 2003. A division-linked mech- plasma cell motility within lymph node through ICAM1/2 or anism for the rapid generation of Ig-secreting cells from human memory B cells. VCAM-1 gradient (10, 38). J. Immunol. 170: 261–269. In conclusion, we have fully characterized a human transitional 21. Avery, D. T., J. I. Ellyard, F. Mackay, L. M. Corcoran, P. D. Hodgkin, and S. G. Tangye. 2005. Increased expression of CD27 on activated human memory prePB stage. This opens the possibility to elucidate the mecha- B cells correlates with their commitment to the plasma cell lineage. J. Immunol. nisms controlling their generation, to track these prePBs in the 174: 4034–4042. 22. Hartmann, G., and A. M. Krieg. 2000. Mechanism and function of a newly process of normal plasma cell differentiation in vivo, after im- identified CpG DNA motif in human primary B cells. J. Immunol. 164: 944–953. munization to a given Ag, and to look for cancer disease char- 23. Jourdan, M., K. Mahtouk, J. L. Veyrune, G. Couderc, G. Fiol, N. Redal, acterized by the immortalization of the prePBs. C. Duperray, J. De Vos, and B. Klein. 2005. Delineation of the roles of paracrine and autocrine interleukin-6 (IL-6) in myeloma cell lines in survival versus cell cycle: a possible model for the cooperation of myeloma factors. Eur. Cytokine Netw. 16: 57–64. 24. Maecker, H. T., T. Frey, L. E. Nomura, and J. Trotter. 2004. Selecting fluoro- Acknowledgments chrome conjugates for maximum sensitivity. Cytometry A 62: 169–173. We thank staff of the Microarray Core Facility and the Cytometry Platform 25. Caraux, A., B. Klein, B. Paiva, C. Bret, A. Schmitz, G. M. Fuhler, N. A. Bos, of the Institute for Research in Biotherapy (Montpellier, France) for assis- H. E. Johnsen, A. Orfao, and M. Perez-Andres; Myeloma Stem Cell Network. tance. 2010. Circulating human B and plasma cells: age-associated changes in counts and detailed characterization of circulating normal CD1382 and CD138+ plasma cells. Haematologica 95: 1016–1020. 26. Jourdan, M., T. Reme, H. Goldschmidt, G. Fiol, V. Pantesco, J. De Vos, J. F. Rossi, D. Hose, and B. Klein. 2009. Gene expression of anti- and pro- Disclosures apoptotic in malignant and normal plasma cells. Br. J. Haematol. The authors have no financial conflicts of interest. 145: 45–58. The Journal of Immunology 3941

27. Re`me, T., D. Hose, J. De Vos, A. Vassal, P. O. Poulain, V. Pantesco, 32. Roederer, M. 2001. Spectral compensation for flow cytometry: visualization H. Goldschmidt, and B. Klein. 2008. A new method for class prediction based on artifacts, limitations, and caveats. Cytometry 45: 194–205. signed- algorithms applied to Affymetrix microarray experiments. BMC 33. Iwakoshi, N. N., A. H. Lee, P. Vallabhajosyula, K. L. Otipoby, K. Rajewsky, and Bioinformatics 9: 16. L. H. Glimcher. 2003. Plasma cell differentiation and the unfolded protein re- 28. Le Carrour, T., S. Assou, S. Tondeur, L. Lhermitte, N. Lamb, T. Re`me, sponse intersect at the transcription factor XBP-1. Nat. Immunol. 4: 321–329. V. Pantesco, S. Hamamah, B. Klein, and J. De Vos. 2010. Amazonia!: an online 34. Cobaleda, C., A. Schebesta, A. Delogu, and M. Busslinger. 2007. Pax5: the resource to Google and visualize public human whole genome expression data. guardian of B cell identity and function. Nat. Immunol. 8: 463–470. Open Bioinformatics J. 4: 5–10. 35. Pridans, C., M. L. Holmes, M. Polli, J. M. Wettenhall, A. Dakic, L. M. Corcoran, 29. Eisen, M. B., P. T. Spellman, P. O. Brown, and D. Botstein. 1998. Cluster G. K. Smyth, and S. L. Nutt. 2008. Identification of Pax5 target genes in early analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. B cell differentiation. J. Immunol. 180: 1719–1728. USA 95: 14863–14868. 36. Calame, K. 2008. Activation-dependent induction of Blimp-1. Curr. Opin. 30. Tusher, V. G., R. Tibshirani, and G. Chu. 2001. Significance analysis of Immunol. 20: 259–264. microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. 37. Eichenauer, D. A., V. L. Simhadri, E. P. von Strandmann, A. Ludwig, USA 98: 5116–5121. V. Matthews, K. S. Reiners, B. von Tresckow, P. Saftig, S. Rose-John, A. Engert, 31. Jourdan, M., A. Caraux, J. De Vos, G. Fiol, M. Larroque, C. Cognot, C. Bret, and H. P. Hansen. 2007. ADAM10 inhibition of human CD30 shedding increases C. Duperray, D. Hose, and B. Klein. 2009. An in vitro model of differentiation of specificity of targeted immunotherapy in vitro. Cancer Res. 67: 332–338. memory B cells into plasmablasts and plasma cells including detailed phenotypic 38. Luther, S. A. 2010. Plasma cell precursors: long-distance travelers looking for and molecular characterization. Blood 114: 5173–5181. a home. Immunity 33: 9–11. Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021 Supplemental Figure Legends

Supplemental Figure 1. Expression of cytoplasmic immunoglobulin in D4 cell subsets. A. Phenotype of D4 cell subsets. B. Expression of cyIgM, cyIgA, and cyIgG in D4 cell subsets. Results are those of one experiment representative of three, SI means Staining Index. C. Expression of cyIgM, cyIgA, and cyIgG in D4 cell subsets.

Results are mean SI obtained in three separate experiments. * The mean expression is significantly different from that in CD20high B cells. ** The mean expression is significantly different from that in D4 CD20low prePBs.

Supplemental Figure 2. Expression of genes coding for B or PC markers.

Data are the Affymetrix signals of expression of genes coding for B or PC markers in

MBCs, CD20high B cells, CD20low prePBs, CD20- prePBs, PBs, early PCs, and

BMPCs using the same color code as in Figure 6A. Data are the mean value ± SD of gene expression determined in 5 separate experiments. * The mean expression is significantly different from that in CD20high B cells. ** The mean expression is significantly different from that in D4 CD20low prePBs.

Supplemental Figure 3. Expression of genes coding for homing molecules.

Data are the Affymetrix signals of expression of genes coding for remarkable chemokine receptors or integrins in MBCs, CD20high B cells, CD20low prePBs, CD20- prePBs, PBs, early PCs, and BMPCs using the same color code as in Figure 6A.

Data are the mean value ± SD of gene expression determined in 5 separate experiments. * The mean expression is significantly different from that in CD20high B cells. ** The mean expression is significantly different from that in D4 CD20low prePBs.

A B

cyIgM SI=31SI=83 SI=110 SI=215 CD20- CD38+

CD38 Anti-IgM-FITC

CD20- CD20low CD20high cyIgA SI=37SI=111 SI=147 SI=167 CD38- CD38- CD38-

CD20 Anti-IgA-FITC

cyIgG SI=15SI=30 SI=37 SI=59

Anti-IgG-APC CD20high CD20low CD20- CD20- CD38- CD38- CD38- CD38+ C

300 300 * 70 ** *** * *** 60 250 *** 250 *** * * 50 200 200 * * 40 ** 150 * 150 30 * 100 100 20 cyIgG SI cyIgG cyIgA SI cyIgA cyIgM SI cyIgM 50 50 10

0 0 0

CD20high CD20low CD20- CD20- CD38- CD38- CD38- CD38+

Supplemental Figure 1 2000 2000 900 900 TNFRSF8=CD30 7000 EMP1 CD19 CD20 800 CD22 800 CD24 700 * 6000 * ** 1500 700 700 600 1500 ** 5000 600 600 500 4000 500 500 * 1000 1000 400 * 400 400 3000 * 300 ** 300 * 300 500 2000 200 500 200 200 100 100 1000 * * 100 * * 0 0 0 0 0 ** ** 0 **

600 3000 500 3500 CD40 AICDA CIITA HLA-DMB 7000 1000 MPEG1 500 2500 3000 CD83 400 6000 800 2500 400 2000 5000 300 2000 600 300 1500 * 4000 200 1500 3000 400 200 1000 * 1000 2000 100 * 500 100 200 ** * 500 * * 1000 0 0 ** 0 * 0 * 0 * * * ** 0 *** **

800 IL4R 500 TNF LTB CD27 CD38 1500 CD31 4000 2000 600 400 3000 2500 3000 1500 1000 300 400 2000 2000 200 1000 1500 * 500 200 * 1000 * 100 * * 1000 500 * ** ** * * ** 500 0 0 0 ** 0 0 0

TNFRSF17 CD59 700 IL6R * 4000 =BCMA * 600 ** 1600 ** 500 3000 1400 1200 * 400 1000 2000 * 300 *** 800 * 600 200 1000 400 100 200 0 0 0

s s s s s s s l B B B C C C P P P B cel eP eP M B r r ly M gh p - p B i ow h l 20 ear 20 20 D D D C C C

Supplemental Figure 2 2000 800 CCR1 1000 CCR6 CCR10 S1PR1=EDG1 1500 800 600 750 * 600 1000 ** 400 * * 500 400 250 200 * * * * 500 200 ** 0 0 0 0

600 2000 CCR7 CXCR5 ITGA4 * CD69 500 500 ** 5000 1500 4000 400 400 3000 300 1000 300 * 2000 200 * 200 * 500 ** 1000 100 100 * * * * 0 ** 0 0 ** 0

150 1000 ITGAX ITGB2 ITGB1 * 4000 KLF2 800 800 ** 100 3000 600 600 2000 400 400 50 * * 1000 ** * 200 200 * 0 0 ** 0 0

s s s s s s s C l B B B C C B cel P P P P P re re y M M B p p l B igh - h low ear 20 20 20 D D D C C C

Supplemental Figure 3 Supplemental Table I. CD20-CD38- and CD20highCD38-. Gene list.

CD20-CD38- genes CD20highCD38- genes Fold Fold Gene ID Gene Name Change Score(d) Gene ID Gene Name Change Score(d) 205945_at IL6R 18,41 4,44 232739_at SPIB 0,01 -2,68 235275_at BMP8B 15,78 3,16 224402_s_at FCRL4 0,01 -5,48 203914_x_at HPGD 14,66 2,65 226818_at MPEG1 0,01 -2,91 204798_at MYB 10,80 7,09 235385_at MARCH-I 0,01 -3,98 222392_x_at PERP 9,68 3,86 223343_at MS4A7 0,01 -4,75 203373_at SOCS2 9,35 5,81 204959_at MNDA 0,01 -4,11 221790_s_at LDLRAP1 8,87 3,35 266_s_at CD24 0,01 -4,43 201324_at EMP1 8,67 3,79 204249_s_at LMO2 0,01 -14,17 206641_at TNFRSF17 6,86 5,37 219517_at ELL3 0,01 -8,97 201243_s_at ATP1B1 6,67 2,67 228055_at NAPSB 0,02 -9,02 233500_x_at CLEC2D 6,45 3,15 217418_x_at MS4A1 0,02 -14,56 201678_s_at DC12 6,22 3,95 205987_at CD1C 0,03 -7,57 39248_at AQP3 6,08 9,19 231093_at FCRH3 0,03 -5,07 200983_x_at CD59 5,32 2,49 228153_at IBRDC2 0,03 -2,97 206729_at TNFRSF8 5,25 2,61 207339_s_at LTB 0,04 -4,80 204254_s_at VDR 5,20 3,22 208018_s_at HCK 0,05 -5,33 221760_at MAN1A1 5,04 12,00 208820_at PTK2 0,05 -2,79 220306_at FAM46C 5,01 3,56 210279_at GPR18 0,05 -5,82 209457_at DUSP5 4,90 3,08 238009_at SOX5 0,06 -5,17 205885_s_at ITGA4 4,73 2,58 213293_s_at TRIM22 0,06 -7,82 203397_s_at GALNT3 4,61 4,02 219014_at PLAC8 0,06 -4,11 206632_s_at APOBEC3B 4,61 4,79 228617_at BIRC4BP 0,06 -2,80 217127_at CTH 4,59 5,13 206126_at BLR1 0,07 -4,05 CD22 /// 1554242_a_at COCH 4,59 3,11 204581_at MAG 0,07 -4,39 202241_at TRIB1 4,58 9,99 216080_s_at FADS3 0,07 -3,79 224802_at NDFIP2 4,52 2,81 221234_s_at BACH2 0,07 -4,31 209921_at SLC7A11 4,52 3,40 213111_at PIP5K3 0,07 -11,02 201397_at PHGDH 4,52 2,93 225123_at SESN3 0,08 -4,27 203474_at IQGAP2 4,44 5,23 215933_s_at HHEX 0,08 -5,42 200951_s_at CCND2 4,13 8,53 205128_x_at PTGS1 0,08 -2,75 GALNAC4S- 203066_at 6ST 4,09 3,15 218032_at SNN 0,09 -2,97 218018_at PDXK 4,08 2,66 205922_at VNN2 0,09 -2,89 219118_at FKBP11 4,03 2,99 204440_at CD83 0,09 -7,96 204900_x_at SAP30 3,97 4,52 203186_s_at S100A4 0,09 -3,34 202558_s_at STCH 3,97 4,01 221011_s_at LBH 0,10 -4,42 224851_at CDK6 3,96 3,14 204994_at MX2 0,10 -3,19 211464_x_at CASP6 3,92 2,59 206983_at CCR6 0,10 -4,88 200628_s_at WARS 3,87 4,70 200696_s_at GSN 0,10 -2,79 217824_at UBE2J1 3,85 5,86 219836_at ZBED2 0,10 -7,02 225512_at ZBTB38 3,78 7,76 1559263_s_at ZC3H12D 0,10 -6,97 202468_s_at CTNNAL1 3,67 3,44 227458_at PDCD1LG1 0,10 -5,18 218073_s_at TMEM48 3,60 3,70 38521_at MAG 0,10 -7,58 228964_at PRDM1 3,57 4,17 230110_at MCOLN2 0,11 -5,00 209695_at PTP4A3 3,55 3,79 223751_x_at TLR10 0,11 -5,58 203971_at SLC31A1 3,53 3,51 229937_x_at LILRB1 0,11 -4,08 216044_x_at FAM69A 3,50 3,58 226748_at LYSMD2 0,11 -3,05 225520_at MTHFD1L 3,49 9,94 203233_at IL4R 0,11 -3,57 222385_x_at SEC61A1 3,47 2,86 224499_s_at AICDA 0,12 -5,11 201206_s_at RRBP1 3,44 3,22 215127_s_at RBMS1 0,12 -3,48 226771_at ATP8B2 3,41 4,57 1552807_a_at SIGLEC10 0,12 -2,99 SLC16A6 /// 230748_at LOC440459 3,41 3,29 215346_at CD40 0,12 -2,70 211430_s_at IGH 3,41 10,17 201315_x_at IFITM2 0,13 -4,05 205047_s_at ASNS 3,35 7,15 203932_at HLA-DMB 0,13 -7,09 242939_at TFDP1 3,35 3,64 210164_at GZMB 0,13 -2,91 201349_at SLC9A3R1 3,30 4,84 205306_x_at KMO 0,13 -4,92 201200_at CREG1 3,21 3,02 212203_x_at IFITM3 0,13 -2,78 222412_s_at SSR3 3,20 3,57 222891_s_at BCL11A 0,14 -5,39 222532_at SRPRB 3,18 2,42 224406_s_at FCRL5 0,14 -5,22 225368_at HIPK2 3,18 9,16 235401_s_at FCRLM1 0,14 -7,68 200924_s_at SLC3A2 3,18 2,81 221044_s_at TRIM34 0,14 -3,07 201565_s_at ID2 3,17 2,71 227646_at EBF 0,14 -6,06 237625_s_at IGKC 3,16 2,42 202464_s_at PFKFB3 0,14 -7,13 SUB1 /// 221727_at SUB1P1 3,15 5,95 209828_s_at IL16 0,15 -2,71 201160_s_at CSDA 3,12 3,52 218076_s_at ARHGAP17 0,15 -2,68 204142_at ENOSF1 3,11 5,55 205671_s_at HLA-DOB 0,16 -3,29 216449_x_at TRA1 3,10 6,60 223287_s_at FOXP1 0,16 -3,63 201021_s_at DSTN 3,09 3,50 201850_at CAPG 0,16 -3,71 1553530_a_at ITGB1 3,07 2,53 227697_at SOCS3 0,16 -4,16 230352_at PRPS2 3,06 4,92 215565_at DTNB 0,16 -5,78 222231_s_at PRO1855 3,06 11,04 205098_at CCR1 0,17 -4,47 203967_at CDC6 3,05 2,49 224374_s_at EMILIN2 0,17 -2,97 223062_s_at PSAT1 3,03 7,55 204352_at TRAF5 0,17 -5,83 214512_s_at SUB1 2,97 3,20 204446_s_at ALOX5 0,17 -7,65 221253_s_at TXNDC5 2,95 7,06 211962_s_at ZFP36L1 0,17 -7,14 218681_s_at SDF2L1 2,92 2,94 210538_s_at BIRC3 0,17 -6,09 201923_at PRDX4 2,91 4,79 228343_at POU2F2 0,18 -4,02 212110_at SLC39A14 2,91 4,18 222450_at TMEPAI 0,18 -3,68 204695_at CDC25A 2,90 3,40 230252_at GPR92 0,19 -3,66 226982_at ELL2 2,89 5,43 205859_at LY86 0,19 -3,46 224679_at MESDC2 2,88 3,22 232543_x_at ARHGAP9 0,19 -8,11 212501_at CEBPB 2,88 4,02 203281_s_at UBE1L 0,19 -3,53 200895_s_at FKBP4 2,85 3,49 202748_at GBP2 0,20 -3,48 203968_s_at CDC6 2,85 3,93 214022_s_at IFITM1 0,20 -3,21 226750_at LARP2 2,83 3,64 203271_s_at UNC119 0,20 -2,70 201468_s_at NQO1 2,83 2,60 204057_at IRF8 0,20 -6,78 219869_s_at SLC39A8 2,83 2,40 212827_at IGHM 0,20 -11,66 209340_at UAP1 2,83 4,22 212774_at ZNF238 0,21 -7,99 1553954_at ALG14 2,80 3,11 225570_at SLC41A1 0,21 -4,53 202375_at SEC24D 2,77 2,55 223162_s_at LCHN 0,22 -3,31 200617_at KIAA0152 2,76 3,56 238429_at TMEM71 0,22 -4,36 200889_s_at SSR1 2,75 5,39 213572_s_at SERPINB1 0,22 -3,58 204502_at SAMHD1 2,75 3,02 225775_at TSPAN33 0,22 -4,41 204639_at ADA 2,74 2,56 202732_at PKIG 0,23 -3,63 202721_s_at GFPT1 2,73 2,57 228234_at TICAM2 0,23 -5,28 213440_at RAB1A 2,73 2,48 218312_s_at ZNF447 0,23 -2,99 204033_at TRIP13 2,70 6,36 209558_s_at HIP1R 0,23 -4,00 200755_s_at CALU 2,69 2,63 200644_at MARCKSL1 0,23 -2,94 213836_s_at WIPI49 2,69 4,74 207777_s_at SP140 0,23 -9,22 223209_s_at SELS 2,67 4,59 204882_at ARHGAP25 0,23 -4,91 219110_at NOLA1 2,65 4,14 229597_s_at WDFY4 0,23 -3,91 223513_at CENPJ 2,65 2,61 242239_at NSUN6 0,23 -2,80 203200_s_at MTRR 2,65 3,26 209795_at CD69 0,23 -5,02 219806_s_at FN5 2,64 2,47 217763_s_at RAB31 0,23 -3,51 226150_at HTPAP 2,64 4,05 226878_at HLA-DOA 0,24 -4,25 223325_at TXNDC11 2,61 3,56 207992_s_at AMPD3 0,24 -4,95 202195_s_at TMED5 2,59 3,72 214177_s_at PBXIP1 0,24 -4,06 201399_s_at TRAM1 2,59 4,99 221969_at PAX5 0,24 -8,80 226935_s_at CRR9 2,57 6,90 34210_at CD52 0,25 -3,55 211275_s_at GYG 2,57 6,17 219183_s_at PSCD4 0,25 -3,15 209806_at HIST1H2BK 2,56 4,41 1563674_at FCRL2 0,25 -6,04 202763_at CASP3 2,56 9,41 203118_at PCSK7 0,25 -3,17 209017_s_at PRSS15 2,55 3,99 204912_at IL10RA 0,25 -5,00 209100_at IFRD2 2,53 3,12 203006_at INPP5A 0,25 -3,15 225143_at SFXN4 2,52 3,61 201260_s_at SYPL1 0,25 -5,64 203224_at RFK 2,52 2,68 219546_at BMP2K 0,25 -5,43 223413_s_at LYAR 2,51 3,65 218700_s_at RAB7L1 0,25 -6,43 201195_s_at SLC7A5 2,51 8,06 212813_at JAM3 0,25 -2,95 HSPA1A /// 200800_s_at HSPA1B 2,50 2,77 201362_at IVNS1ABP 0,25 -3,85 222642_s_at TMEM33 2,50 2,46 226530_at BMF 0,26 -3,67 213129_s_at GCSH 2,49 4,10 211676_s_at IFNGR1 0,26 -7,17 202655_at ARMET 2,48 8,01 244023_at SYK 0,26 -2,88 203857_s_at PDIA5 2,48 2,69 204866_at PHF16 0,26 -6,83 200929_at TMED10 2,48 3,05 223218_s_at NFKBIZ 0,26 -5,66 218350_s_at GMNN 2,47 6,50 236293_at RHOH 0,26 -2,69 218257_s_at UGCGL1 2,47 2,73 217783_s_at YPEL5 0,26 -4,00 222432_s_at GK001 2,47 6,59 218223_s_at CKIP-1 0,27 -3,03 222646_s_at ERO1L 2,45 6,16 213618_at CENTD1 0,27 -4,11 213113_s_at SLC43A3 2,45 5,49 202086_at MX1 0,27 -2,89 218817_at SPCS3 2,45 3,89 216060_s_at DAAM1 0,27 -4,80 207121_s_at MAPK6 2,45 3,62 56256_at SIDT2 0,27 -4,21 201096_s_at ARF4 2,45 6,41 204881_s_at UGCG 0,27 -2,99 208852_s_at CANX 2,44 4,39 202499_s_at SLC2A3 0,27 -4,99 225440_at AGPAT3 2,44 2,51 228258_at TBC1D10C 0,27 -4,66 203925_at GCLM 2,43 4,05 210201_x_at BIN1 0,27 -3,26 SLC2A3 /// 222977_at SURF4 2,43 2,65 222088_s_at SLC2A14 0,27 -2,86 FAM45B /// 221804_s_at FAM45A 2,42 3,13 201331_s_at STAT6 0,28 -5,09 201625_s_at INSIG1 2,42 2,60 227224_at RALGPS2 0,29 -6,41 211744_s_at CD58 2,40 2,94 204588_s_at SLC7A7 0,29 -2,72 223308_s_at WDR5 2,40 3,14 200788_s_at PEA15 0,29 -6,62 222404_x_at HSPC121 2,39 3,55 221601_s_at FAIM3 0,30 -3,39 208658_at PDIA4 2,39 2,89 206687_s_at PTPN6 0,30 -6,73 203405_at DSCR2 2,38 6,67 225415_at DTX3L 0,30 -7,53 200806_s_at HSPD1 2,38 3,95 222915_s_at BANK1 0,30 -2,79 203560_at GGH 2,37 6,25 224701_at PARP14 0,30 -3,10 202842_s_at DNAJB9 2,35 3,10 200678_x_at GRN 0,31 -2,86 209825_s_at UCK2 2,35 3,70 218543_s_at PARP12 0,31 -2,70 209273_s_at HBLD2 2,34 3,02 1555349_a_at ITGB2 0,31 -3,58 213523_at CCNE1 2,33 4,00 206398_s_at CD19 0,31 -9,40 202613_at CTPS 2,33 3,32 208012_x_at SP110 0,31 -4,70 208290_s_at EIF5 2,33 7,18 74694_s_at RABEP2 0,31 -2,79 204836_at GLDC 2,33 3,11 221565_s_at FAM26B 0,32 -3,61 218883_s_at MLF1IP 2,33 3,48 202656_s_at SERTAD2 0,32 -4,00 224759_s_at MGC17943 2,33 3,27 226459_at PIK3AP1 0,32 -7,29 203328_x_at IDE 2,32 2,39 206181_at SLAMF1 0,32 -3,77 201323_at EBNA1BP2 2,32 3,71 209568_s_at RGL1 0,32 -3,07 220565_at CCR10 2,32 2,95 225629_s_at ZBTB4 0,33 -3,17 200968_s_at PPIB 2,31 5,92 227606_s_at AMSH-LP 0,33 -3,07 219551_at EAF2 2,31 2,47 224735_at CYBASC3 0,33 -3,46 205632_s_at PIP5K1B 2,30 2,94 231747_at CYSLTR1 0,33 -2,71 224954_at SHMT1 2,30 3,18 227677_at JAK3 0,33 -4,40 202223_at ITM1 2,30 3,43 208944_at TGFBR2 0,33 -6,03 209674_at CRY1 2,30 3,40 223514_at CARD11 0,34 -3,46 222620_s_at DNAJC1 2,29 4,93 203837_at MAP3K5 0,34 -3,58 200799_at HSPA1A 2,29 3,43 202395_at NSF 0,34 -6,12 209314_s_at HBS1L 2,28 3,34 226474_at NOD27 0,34 -4,68 220651_s_at MCM10 2,28 3,82 202716_at PTPN1 0,34 -6,53 218488_at EIF2B3 2,28 2,58 218640_s_at PLEKHF2 0,35 -3,45 202433_at SLC35B1 2,28 4,40 213772_s_at GGA2 0,35 -5,41 221521_s_at Pfs2 2,27 4,53 203923_s_at CYBB 0,36 -2,66 NOMO1 /// NOMO2 /// 221853_s_at NOMO3 2,26 4,12 202269_x_at GBP1 0,36 -2,74 217800_s_at NDFIP1 2,25 2,87 207522_s_at ATP2A3 0,36 -3,61 1554572_a_at SUV39H2 2,24 2,44 50221_at TFEB 0,36 -3,05 201391_at TRAP1 2,24 4,10 203332_s_at INPP5D 0,37 -8,03 205393_s_at CHEK1 2,23 3,30 212631_at STX7 0,37 -5,57 201912_s_at GSPT1 2,22 5,65 223220_s_at PARP9 0,37 -3,37 217855_x_at SDF4 2,22 3,65 217984_at RNASET2 0,37 -3,20 218254_s_at SAR1B 2,21 3,03 200904_at HLA-E 0,37 -3,19 222680_s_at DTL 2,21 3,71 213521_at PTPN18 0,38 -5,33 202788_at MAPKAPK3 2,20 2,95 228869_at SLIC1 0,38 -6,30 202206_at ARL7 2,20 2,55 225230_at TMEM77 0,38 -3,67 213302_at PFAS 2,19 2,67 228771_at ADRBK2 0,38 -2,87 207707_s_at SEC13L1 2,19 3,96 221718_s_at AKAP13 0,39 -4,29 202857_at TMEM4 2,18 4,69 203528_at SEMA4D 0,39 -4,51 229595_at CHCHD4 2,16 4,67 202769_at CCNG2 0,39 -4,03 241937_s_at WDR4 2,16 2,76 203185_at RASSF2 0,39 -11,16 208864_s_at TXN 2,16 7,55 201432_at CAT 0,39 -5,87 201014_s_at PAICS 2,15 7,20 201170_s_at BHLHB2 0,39 -4,86 222843_at FIGNL1 2,15 3,91 200743_s_at TPP1 0,39 -8,58 210156_s_at PCMT1 2,15 3,38 221816_s_at PHF11 0,40 -3,99 212048_s_at YARS 2,14 5,94 222512_at NYREN18 0,40 -4,12 235113_at PPIL5 2,13 3,75 202902_s_at CTSS 0,40 -3,01 201516_at SRM 2,13 3,19 207677_s_at NCF4 0,40 -2,74 210691_s_at CACYBP 2,13 4,38 224925_at PREX1 0,40 -2,70 201011_at RPN1 2,13 6,16 218501_at ARHGEF3 0,40 -2,70 217850_at GNL3 2,13 2,99 201137_s_at HLA-DPB1 0,40 -8,06 221677_s_at DONSON 2,12 4,17 223049_at GRB2 0,41 -5,31 204127_at RFC3 2,12 3,48 203765_at GCA 0,41 -5,01 209434_s_at PPAT 2,12 2,41 202114_at SNX2 0,41 -2,65 208689_s_at RPN2 2,11 4,54 200965_s_at ABLIM1 0,41 -2,91 201459_at RUVBL2 2,11 3,03 229971_at GPR114 0,41 -3,25 213671_s_at MARS 2,10 5,29 212998_x_at HLA-DQB1 0,41 -3,75 203432_at TMPO 2,10 2,47 226148_at BTBD15 0,41 -2,90 201000_at AARS 2,10 4,68 210754_s_at LYN 0,41 -6,56 226422_at PTX1 2,09 2,53 225701_at AKNA 0,42 -4,96 200670_at XBP1 2,09 2,90 202446_s_at PLSCR1 0,42 -4,29 213017_at ABHD3 2,09 2,65 222108_at AMIGO2 0,42 -2,73 221597_s_at HSPC171 2,08 2,50 204197_s_at RUNX3 0,42 -5,76 210296_s_at PXMP3 2,08 5,67 230925_at APBB1IP 0,43 -5,66 201662_s_at ACSL3 2,08 5,92 209984_at JMJD2C 0,43 -2,90 1564520_s_at SKB1 2,08 2,55 226354_at LACTB 0,43 -3,92 223035_s_at FARSLB 2,07 7,41 201301_s_at ANXA4 0,43 -2,98 208813_at GOT1 2,07 2,63 202501_at MAPRE2 0,43 -5,60 220865_s_at TPRT 2,07 2,53 203241_at UVRAG 0,43 -2,80 219126_at PHF10 2,06 3,18 210046_s_at IDH2 0,43 -2,70 224866_at MLSTD2 2,06 3,46 200934_at DEK 0,43 -3,80 218264_at BCCIP 2,06 2,56 31845_at ELF4 0,44 -2,89 218696_at EIF2AK3 2,05 4,32 208894_at HLA-DRA 0,44 -5,30 203343_at UGDH 2,05 2,50 216250_s_at LPXN 0,44 -8,65 244519_at ASXL1 2,04 3,01 203567_s_at TRIM38 0,44 -4,45 212685_s_at TBL2 2,04 3,05 217478_s_at HLA-DMA 0,44 -4,22 204781_s_at FAS 2,04 2,47 212899_at CDC2L6 0,44 -3,08 227212_s_at PHF19 2,03 4,21 223358_s_at PDE7A 0,44 -3,67 202771_at FAM38A 2,03 3,40 213241_at PLXNC1 0,44 -3,51 217977_at SEPX1 2,03 2,61 204489_s_at CD44 0,44 -3,95 212295_s_at SLC7A1 2,03 5,99 201234_at ILK 0,44 -2,98 229097_at DIAPH3 2,02 3,47 201312_s_at SH3BGRL 0,45 -8,74 202369_s_at TRAM2 2,00 2,68 223199_at MKNK2 0,45 -4,03 227087_at INPP4A 2,00 4,32 205213_at CENTB1 0,45 -3,21 35254_at TRAFD1 0,45 -2,65 203879_at PIK3CD 0,45 -5,15 203710_at ITPR1 0,45 -3,64 208991_at STAT3 0,45 -6,79 201814_at TBC1D5 0,45 -4,47 209723_at SERPINB9 0,45 -3,48 204362_at SCAP2 0,45 -4,71 220059_at BRDG1 0,45 -3,32 201613_s_at AP1G2 0,45 -3,23 223391_at SGPP1 0,45 -3,42 218242_s_at SUV420H1 0,46 -2,99 209619_at CD74 0,46 -3,85 218396_at VPS13C 0,46 -3,78 203037_s_at MTSS1 0,46 -3,18 203685_at BCL2 0,46 -2,68 205126_at VRK2 0,47 -5,02 208998_at UCP2 0,47 -5,62 236295_s_at NOD3 0,47 -3,75 1553906_s_at FGD2 0,47 -5,20 223092_at ANKH 0,47 -3,97 200972_at TSPAN3 0,47 -3,07 1555797_a_at ARPC5 0,47 -2,76 219279_at DOCK10 0,47 -5,41 210592_s_at SAT 0,47 -5,66 202963_at RFX5 0,47 -7,49 222243_s_at TOB2 0,48 -2,71 221773_at ELK3 0,48 -3,26 223303_at URP2 0,49 -3,36 225010_at CCDC6 0,49 -4,16 226016_at CD47 0,49 -3,60 HLA-DQA1 /// HLA- 212671_s_at DQA2 0,49 -3,18 204061_at PRKX 0,49 -2,73 204028_s_at RABGAP1 0,49 -3,24 204949_at ICAM3 0,49 -3,65 227146_at QSCN6L1 0,49 -3,42 202080_s_at OIP106 0,49 -3,61 223553_s_at DOK3 0,49 -3,87 222146_s_at TCF4 0,50 -3,31 212144_at UNC84B 0,50 -3,90