The Journal of Immunology

TGF-b1 Accelerates Dendritic Cell Differentiation from Common Dendritic Cell Progenitors and Directs Subset Specification toward Conventional Dendritic Cells

Piritta Felker,*,† Kristin Sere´,*,† Qiong Lin,*,† Christiane Becker,*,† Mihail Hristov,‡ Thomas Hieronymus,*,† and Martin Zenke*,†

Dendritic cells (DCs) in lymphoid tissue comprise conventional DCs (cDCs) and plasmacytoid DCs (pDCs) that develop from com- mon DC progenitors (CDPs). CDPs are Flt3+c-kitintM-CSFR+ and reside in bone marrow. In this study, we describe a two-step culture system that recapitulates DC development from c-kithiFlt32/lo multipotent progenitors (MPPs) into CDPs and further into cDC and pDC subsets. MPPs and CDPs are amplified in vitro with Flt3 ligand, stem cell factor, hyper–IL-6, and insulin-like growth factor-1. The four-factor mixture readily induces self-renewal of MPPs and their progression into CDPs and has no self- renewal activity on CDPs. The amplified CDPs respond to all known DC poietins and generate all lymphoid tissue DCs in vivo and in vitro. Additionally, in vitro CDPs recapitulate the cell surface marker and expression profile of in vivo CDPs and possess a DC-primed transcription profile. TGF-b1 impacts on CDPs and directs their differentiation toward cDCs. Genome-wide gene expression profiling of TGF-b1–induced identified instructive transcription factors for cDC subset specification, such as IFN regulatory factor-4 and RelB. TGF-b1 also induced the inhibitor of differentiation/DNA binding 2 that suppresses pDC development. Thus, TGF-b1 directs CDP differentiation into cDCs by inducing both cDC instructive factors and pDC inhibitory factors. The Journal of Immunology, 2010, 185: 5326–5335.

+ + + endritic cells (DCs) are professional APCs that are im- DC progenitors were described: a Flt3 c-kit CX3CR1 macrophage/ portant in inducing adaptive immune responses and main- DC progenitor (MDP) (7, 8) and an Flt3+c-kitintM-CSFR+ com- D taining T cell tolerance (1–3). DCs are divided into mon DC progenitor (CDP) (9, 10). MDPs give rise to macrophages several subsets according to their localization, phenotype, and and DCs, whereas CDPs are DC-restricted and do not generate other function. The main DC subsets are conventional DCs (cDCs) and cell types. Indeed, MDPs were shown to be more heterogeneous and plasmacytoid DCs (pDCs) found in lymphoid organs and migra- to contain CDPs (8, 11, 12). tory tissue DCs that are spread throughout peripheral organs (2, Frequently, DCs are generated from bone marrow or cord blood- by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. 3). DC development shows remarkable plasticity, and DCs have derived hematopoietic stem/progenitor cells or blood monocytes been shown to develop from both lymphoid and myeloid compart- in vitro. The major cytokines used in culturing DCs are Flt3 ligand ments via Flt3-expressing progenitors (4–6). Recently, clonogenic (Flt3L) and GM-CSF, and, more recently, M-CSF (13–17). Flt3L and M-CSF cultures give rise to cDC and pDC subsets, whereas hi *Department of Cell Biology, Institute for Biomedical Engineering, †Helmholtz In- GM-CSF induces the differentiation of CD11b DCs. According to stitute for Biomedical Engineering, and ‡Institute for Molecular Cardiovascular Re- the current view, Flt3L-generated DCs are considered as in vitro search, Medical Faculty, Rheinisch-Westfa¨lische Technische Hochschule, Aachen, Germany equivalents of steady state DCs, as found, for example, in spleen. Flt3L is the major cytokine for DC development in steady state Received for publication December 8, 2009. Accepted for publication August 20, 2010. in vivo; Flt3L is a potent inducer of both cDCs and pDCs in human https://www.jimmunol.org This work was supported by Grants SFB542 and SPP1356 from the German Research and mouse (15, 18, 19). Accordingly, Flt3L-deficient mice have Foundation (to M.Z.). a severe reduction in DC numbers in several lymphoid organs (15, 20, The microarray data presented in this article have been submitted to the Gene Ex- 21). Interestingly, administration of M-CSF induces cDC and pDC pression Omnibus database (http://www.ncbi.nlm.nih.gov/geo) under accession num- generation in vivo (17), and M-csf–deficient (CSF-1op/op)micehave ber GSE22432. reduced numbers of splenic DCs (22), indicating the in vivo rele- Address correspondence and reprint requests to Dr. Martin Zenke, Department of Cell Biology, Institute for Biomedical Engineering, Medical Faculty, Rheinisch- vance of M-CSF in DC development. In contrast, GM-CSF induces Downloaded from Westfa¨lische Technische Hochschule Aachen, Pauwelsstrasse 30, 52074 Aachen, differentiation of inflammatory DCs that have no steady-state Germany. E-mail address: [email protected] counterpart in vivo. GM-CSF appears to be involved in MDP and The online version of this article contains supplemental material. CDP homeostasis in bone marrow as well as in nonlymphoid tissue Abbreviations used in this paper: C, negative control; cDC, conventional dendritic cell; dermal DC development (21). However, Gm-csf– and Gm-csfr– CDP, common dendritic cell progenitor; CFU-G, CFU granulocyte; CFU-GEMM, deficient mice have only a very mild reduction in DC numbers in CFU granulocyte, erythrocyte, macrophage, megakaryocyte; CFU-GM, CFU granulo- cyte/macrophage; CFU-M, CFU macrophage; DC, dendritic cell; Flt3L, Flt3 ligand; lymphoid tissues where Flt3L is the more crucial factor (21, 23). GEMM, granulocyte, erythrocyte, macrophage, megakaryocyte; GO, ; TGF-b1 is a pleiotropic cytokine involved in a variety of bio- HLH, helix-loop-helix; HSC, hematopoietic stem cell; Id2, inhibitor of differentiation/ DNA binding 2; IGF-1, insulin-like growth factor-1; IRF, IFN regulatory factor; MDP, logical processes, such as development, differentiation, apoptosis, macrophage/DC progenitor; MHC-II, MHC class II; MPP, multipotent progenitor; and cell survival (24). TGF-b1 is crucial for development of PCA, principal component analysis; pDC, plasmacytoid dendritic cell; PI, propidium Langerhans cells, the cutaneous contingent of migratory DC, both iodide; SCF, stem cell factor. in vivo and in vitro (25–27). Yet, the influence of TGF-b1on Copyright Ó 2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00 development of other DC subsets is so far unknown.

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0903950 The Journal of Immunology 5327

The current understanding of DC development is mainly based Flow cytometry and cell sorting on studies of genetically modified mice and knockout models that Abs directly conjugated to biotin, FITC, PE, PE-Cy5, PerCP-Cy5.5, PE- lack transcription factors (reviewed in Refs. 3, 28, and 29). STAT3 Cy7, allophycocyanin, or Pacific Blue specific for the following Ags were is a central transcription factor in Flt3 signaling, and deletion used: AA4.1/CD93/Ly68 (AA4.1), B220 (RA3-6B2), CD3ε (145-2C11), of STAT3 severely impairs DC development (30). PU.1, an Ets CD4 (GK1.5), CD8a (53-6.7), CD11b (M1/70), CD11c (N418 or HL3), family transcription factor, is expressed highly in hematopoetic CD19 (1D3), CD24 (M1/69), CD45.1 (A20), CD45.2 (104), CD115 (AFS98), CD117 (ACK2), CD127 (A7R34), CD135 (A2F10.1), Gr-1 progenitors and DCs (31). Mice reconstituted with Pu.1-deficient (RB6-C5), MHC class II (MHC-II; M5/114.15.2), PDCA-1 (JF05- hematopoietic cells have severe hematopoietic defects and lack 1C2.4.1), Sca-1 (D7), SiglecH (440c), and Ter119/Ly-76 (TER-119) (all cDCs (32, 33). Ikaros, a zinc finger transcriptional regulator, is from BD Biosciences, San Jose, CA; eBioscience, San Diego, CA; or required for the normal development of early hematopoietic pro- Miltenyi Biotec, Bergisch Gladbach, Germany). Pacific Blue-labeled F4/80 (Cl:A3-1) Ab was from AbD Serotec (Du¨sseldorf, Germany). Purified genitors (34). Ikaros deficiency leads to either the ablation of all CD49d (R1-2) and CD49f (GoH3) Abs were from BD Biosciences. DC subsets or the absence of specific DC subsets, which depends Lineage-positive cells were stained with biotin-conjugated Lineage Cell on the type of mutation introduced (35, 36). Other transcription Detection Cocktail (Miltenyi Biotec). Streptavidin-FITC, streptavidin-PE, factors have a more selected impact on DC subset specification. and streptavidin-allophycocyanin were from Invitrogen or eBioscience. The development of cDCs is influenced by the IFN regulatory Streptavidin-PerCP-Cy5.5 was from BD Biosciences. Secondary goat anti- rat Ab conjugated to PE was from Invitrogen. Cells were treated with anti- factors (IRFs) IRF-2, IRF-4, and IRF-8 (also known as ICSBP) FcR–blocking reagent (Miltenyi Biotec) and 3% mouse serum prestaining. (37), the helix-loop-helix (HLH) transcription factor inhibitor of differentiation/DNA binding 2 (Id2) (38), and the NF-kB/Rel family member RelB (39, 40), among others. pDC development depends on the basic HLH factor E2-2 (41, 42), the Ets-domain transcription factor Spi-B (43), and IRF-4 and IRF-8 (37). Al- though critically adding to our understanding of DC development, it is worthwhile to note that these studies so far concentrated on the phenotype in mature DCs. It is yet unclear at which stage and how these factors influence DC development. In this study, we investigated the early steps of DC develop- ment from multipotent progenitors (MPPs) and CDPs, and the impact of TGF-b1 on DC subset specification from CDPs. CDPs are very scarce cells in vivo; thus, we used an in vitro culture sys- tem to obtain high numbers of CDPs. We show consecutive dif- ferentiation of MPPs to CDPs and further to DC subtypes and provide functional and molecular analyses of this differentiation pathway. We also demonstrate that TGF-b1 accelerates DC dif- ferentiation and directs differentiation toward cDCs by inducing a cDC-affiliated transcription profile in CDPs. by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. Materials and Methods Mice C57BL/6 mice, CD45.1 (B6.SJL-Ptprca Pepcb/BoyJ), GFP transgenic mice (44), and CX3CR1-GFP knockin mice (45) were maintained under specific pathogen-free conditions in the central animal facility of RWTH Aachen University Hospital, Aachen, Germany. All animal experiments were ap- proved by local authorities in compliance with the German animal pro- tection law. https://www.jimmunol.org Cells and cell culture Bone marrow suspensions were prepared, and cells were cultured in RPMI 1640 medium supplemented with 10% FCS, 100 U/ml penicillin/strepto- mycin, 2 mM L-glutamine (all from Invitrogen Life Technologies, Carls- bad, CA), and 50 mM 2-ME (Sigma-Aldrich, St. Louis, MO) with 25 ng/ml Flt3L (PeproTech, London, U.K.), 30 U/ml stem cell factor (SCF; Downloaded from PeproTech), 5 ng/ml IL-6/soluble IL-6R fusion (hyper–IL-6, kindly provided by S. Rose-John, Kiel, Germany), and 40 ng/ml long-range FIGURE 1. Amplification of MPP and CDP populations from bone insulin-like growth factor-1 (IGF-1; Sigma-Aldrich). After 3 d of cul- ture, cells were subjected to Ficoll-Hypaque density gradient centrifuga- marrow. Bone marrow cells were cultured with Flt3L, SCF, hyper–IL-6, and tion (density, 1.077 g/ml). Cells were kept at 2 3 106 cells/ml and re- IGF-1. A and B, In vitro-amplified bone marrow progenitors and in vivo gfp/+ freshed with medium and cytokines every second day. Progenitor cells lineage-negative bone marrow cells from CX3CR1 mice (A, B,re- were harvested on day 7 of culture. spectively) were stained for c-kit, Flt3, M-CSFR/CD115, and IL-7Ra/ To induce DC differentiation, progenitor cells were washed and cul- CD127 and analyzed by flow cytometry. In vitro cultures contained Flt32/lo tured further with 50 ng/ml Flt3L, 200 U/ml GM-CSF, or 20 ng/ml M-CSF c-kithi MPPs and Flt3+c-kitintM-CSFR+ CDPs, which showed the same (all from PeproTech) for 5–10 d. surface marker profile as MPPs and CDPs from bone marrow. Data shown are B cells were generated in cocultures with OP9 cells. B cell medium was representative of at least three independent experiments. C,PCAofinvivo alpha-MEM supplemented with 10% FCS, 100 U/ml penicillin/strepto- CDPs and in vitro-generated MPPs, CDPs, cDCs, and pDCs. Each symbol mycin, 2 mM L-glutamine, and 100 mM 2-ME with 1% conditioned su- pernatant from IL-7–producing J558 cells and 5 ng/ml Flt3L (PeproTech). represents a microarray data set: blue triangles, MPPs; orange squares, in vitro rTGF-b1 (R&D Systems, Minneapolis, MN) was used at 10 ng/ml CDPs; green diamonds, in vivo CDPs; red dots, cDCs; purple triangles, pDCs. concentration for short-term (#24 h) and at 3 ng/ml concentration for The two principal components PC1 and PC3 showed variances of 46.3 and long-term (.24 h) treatments. 15.9%, respectively. 5328 TGF-b1 IN DC DIFFERENTIATION FROM COMMON DC PROGENITORS

Routinely, staining with the respective isotype was used as control. Apo- Bone marrow transplantations and in vivo differentiation ptosis was measured with Annexin V-FITC apoptosis detection kit (BD potential of progenitor cells Biosciences). Cells were analyzed by FACSCanto II (BD Biosciences) using FlowJo software (Tree Star, Ashland, OR). MPPs and CDPs were obtained by FACS sorting from bone marrow cul- Amplified MPPs and CDPs were sorted as Flt32/loc-kithiCD11c2 cells tures of C57BL/6 mice at day 7. A total of 1 3 106 CD45.2+ MPPs or CDPs and Flt3+c-kitintM-CSFR+CD11c2IL-7Ra2 cells, respectively, at day 7 of was injected i.v. into CD45.1+ mice along with a radioprotective dose of culture. Cultures were deprived of cytokines for 1.5–2 h prestaining for 1 3 105 CD45.1+ bone marrow cells. Recipient mice were lethally irra- flow cytometry. In vivo CDPs were sorted from bone marrow as lineage- diated (twice at 5.5 Gy in 3-h intervals). At 10 d and 28 d after cell transfer, negative Flt3+c-kitintM-CSFR+IL-7Ra2 cells. Biotinylated Abs against recipient mice were sacrificed, and bone marrow and spleen were analyzed B220, CD11b, Gr-1, CD11c, CD3ε, Ter119, and NK1.1 and streptavidin- for donor-derived hematopoietic cells by flow cytometry. Three recipients FITC were used to stain lineage-positive cells. Cell sorting was performed per group were used. with an FACSAria device (BD Biosciences). Semiquantitative PCR Colony formation assay RNA was isolated using RNeasy Mini or Micro Kit with DNaseI diges- tion (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. In vitro-amplified MPPs and CDPs were FACS sorted and plated in Reverse transcription reaction was performed with 1 mg RNA using a High semisolid methylcellulose (MethoCult GF M3434, StemCell Technolo- Capacity cDNA RT Kit (Applied Biosystems, Foster City, CA). RT-PCR gies, Vancouver, British Columbia, Canada) containing insulin, transferrin, was performed with 50 ng cDNA (for primer sequences, see Supplemental SCF, IL-3, IL-6, and erythropoietin (5000 cells/ml). Colonies were eval- Table III). uated by microscopy at day 6, and colonies consisting of .40 cells were counted. Representative colonies were picked at day 8, and cytospins were DNA microarray analysis stained with Diff/Quick (Dade Behring, Du¨dingen, Switzerland). Erythroid differentiation was assessed by neutral benzidine staining. RNA was isolated using RNeasy Mini or Micro Kit with DNaseI diges- tion (Qiagen) as above and subjected to microarray analysis as before (38, Cell proliferation assay with CFSE labeling 46). Briefly, sample preparation was performed according to the Expres- sion Analysis Technical Manual (Affymetrix, Santa Clara, CA). GeneChip For monitoring cell proliferation, cells were labeled with CFSE (Molec- One-cycle Target Labeling Kit (Affymetrix) and 1 mg total RNA were used ular Probes, Eugene, OR). Briefly, cells were washed and resuspended in for in vitro-generated MPP, CDP, cDC, and pDC samples. GeneChip 3‘IVT PBS at 1 3 107 cells/ml. Labeling was performed with 3.5 mM CFSE for Express Kit (Affymetrix) and 50 ng or 250 ng total RNA were used for 10 min at 37˚C and stopped by addition of culture medium. Cells were in vivo CDPs and samples of TGF-b1 kinetics, respectively. Biotin-labeled washed twice with medium and analyzed by flow cytometry 24, 48, and cRNA was hybridized on Affymetrix Mouse Genome 430 2.0 Arrays 72 h postlabeling. Cell numbers were determined with an electronic cell (Affymetrix). GeneChip arrays were stained, washed, and scanned ac- counter device (CASY1, Scha¨rfe Systems, Reutlingen, Germany). cording to the manufacturer’s protocol. by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd.

FIGURE 2. MPPs show multilineage differentiation potential and are upstream from CDP. In vitro-amplified MPP and CDP were FACS sorted at day 8 (A), and 5000 cells were seeded in semisolid methylcellulose con- taining insulin, transferrin, SCF, IL-3, IL-6, and eryth- ropoietin (B). B, MPP-derived colonies consisting of .40 cells were scored for CFU-M, CFU-G, CFU-GM, and CFU-GEMM phenotype at day 6 (MPP). CDPs https://www.jimmunol.org generated colonies of 10–20 cells in size, and colonies containing .10 cells were considered (CDP). The CFU-GEMM colonies rarely turned red and were phe- notypically indistinguishable from mixed CFU-GM colonies (day 6), and therefore, the CFU-GM numbers shown comprise both CFU-GM and CFU-GEMM. C–E, Amplified MPPs and CDPs were FACS sorted as in A Downloaded from and cultured further under amplification conditions (Flt3L SCF, hyper–IL-6, and IGF-1). Cultures were analyzed at days 2, 5, and 8. C, MPP/CDP development was monitored by staining cells for c-kit and Flt3. D, Cumulative cell numbers were determined at regular time intervals and are depicted. E, Spontaneous DC differentiation was monitored by measuring CD11c expression. Representative data from two independent experiments are shown. The Journal of Immunology 5329

Gene expression levels were determined by GCRMA algorithm in R/ repertoire (Supplemental Fig. 1A–D). When Flt3L, SCF, hyper– Bioconductor (47). Principal component analysis (PCA) was done by IL-6, and IGF-1 were added individually or in different combi- prcomp in R package stats. Hierarchical clustering was performed using nations, the most pronounced response in proliferation was mea- Pearson correlation coefficient and the average linkage method and rep- resented by dendrogram and/or heatmap. Differential expression between sured for cytokine cocktails containing Flt3L (Supplemental Fig. two conditions was analyzed using Student t tests. The transcripts with 1E). The amplified progenitors gave rise to all hematopoietic a fold change .2 and p values ,0.05 were considered as being differ- lineages in vivo, including all DC subsets, when transferred into entially expressed. Raw p values were adjusted by Benjamini and Hoch- lethally irradiated recipients, which demonstrates their multi- berg’s method (48). For TGF-b1 data sets, the differentially expressed transcripts were lineage potential (Supplemental Fig. 2). + detected between any two time points. Fuzzy c-means algorithm in R/ Progenitor cells showed spontaneous differentiation into CD11c Bioconductor package Mfuzz (49) was applied to reveal the gene ex- DCs, which increased with culture time: at day 5, essentially no pression pattern in TGF-b1 kinetics data. Specifically, for each time point CD11c+ cells were measured, whereas at day 7, CD11c+ cells (0, 4, 8, 12, and 24 h), the gene expression values of the biological rep- constituted ∼10% of the culture (Supplemental Table I and data not licates were averaged. Then, the differentially expressed transcripts with averaged expression were clustered by Fuzzy c-means algorithm with shown). Thus, the culture contained cells that are primed for DC parameter k = 6 and m = 2 (49). The algorithm was performed on 100 differentiation. Therefore, we analyzed our progenitor culture for independent runs, and the one with highest silhouette values was selected the presence of the recently described CDP (9, 10). Indeed, we (49, 50). The R/Bioconductor package GOstats (51) was used for Gene found a Flt3+c-kitintM-CSFR+IL-7Ra2 population, which also Ontology (GO) (52) enrichment analysis. Data sets were submitted to Gene Expression Omnibus database (http:// expressed the chemokine CX3CR1 (Fig. 1A) (9–11). Thus, www.ncbi.nlm.nih.gov/geo) under accession number GSE22432. based on the surface marker expression and the differentiation potential of the cells (see below), we refer to this cell population Results as CDPs. In vitro-amplified bone marrow progenitors comprise MPPs In addition, the culture contained a population of Flt32/loc-kithi and CDPs cells (Fig. 1A). High c-kit expression correlates with immature Mouse bone marrow cells were cultured in the presence of Flt3L, stem/progenitor cell phenotype and Flt32/loc-kithi cells were found SCF, hyper–IL-6, and IGF-1. The culture induced propagation of to exhibit multilineage differentiation potential (see below), and synchronously proliferating Flt3+ cells with typical multipotent we will hereafter refer to this cell population as MPPs. Finally, we hematopoietic stem/progenitor cell morphology and surface Ag compared the surface marker repertoire of in vitro-generated CDPs

FIGURE 3. DC differentiation potential of sorted MPPs and CDPs. In vitro-amplified MPPs and CDPs were FACS sorted and cultured with Flt3L, M-CSF, or GM-CSF (5–8 d) to generate DC subsets and analyzed by FACS (A–C, respectively). A, CD11c+ cells were by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. subdivided into pDCs and cDCs (CD11b2B220+ and CD11b+B2202, respectively) at day 8 of differentiation culture with Flt3L. B, pDCs (CD11b2SiglecH+, pre- gated on CD11c+ cells) obtained with M-CSF at day 5 of differentiation culture. SiglecH+ cells were also B220+ (data not shown). C, CD11c+MHC-IIhi DCs obtained with GM-CSF at day 5 of differentiation culture. D, Differentiation of MPPs and CDPs into CD11c+ cells in Flt3L M-CSF, or GM-CSF cultures

https://www.jimmunol.org was followed in time as indicated. E, Cell proliferation of MPPs and CDPs in Flt3L, GM-CSF, or M-CSF cultures was followed in time and is presented as fold change relative to the number of cells seeded. F and G, CD45.2+ MPPs and CDPs were obtained by in vitro culture and FACS sorted as Flt32/loc-kithiCD11c2 and Flt3+c-kitintM-CSFR+CD11c2IL-7Ra2 cells, re-

Downloaded from spectively. Sorted MPPs (F) and CDPs (G) were injected i.v. into lethally irradiated CD45.1+ recipients along with a radioprotective dose of CD45.1+ bone marrow. Recipient mice were sacrificed 10 d post- transplantation, and spleen was analyzed for DC sub- sets. Donor-derived pDCs were identified as CD45.2+ CD11c+PDCA-1+CD11b2 cells and cDCs as CD45.2+ CD11c+PDCA-12CD11b2 (CD8a+ cDCs) and CD45.2+ CD11c+PDCA-12CD11b+ (CD11b+ cDCs); filled histo- gram, isotype control; open histogram, pDCs; open his- togram with dashed line, CD11c+PDCA-12 cDCs. Three recipients per group were used. 5330 TGF-b1 IN DC DIFFERENTIATION FROM COMMON DC PROGENITORS

and MPPs with that of CDPs and MPPs isolated from bone marrow. of MPPs. To investigate the population hierarchy, sorted MPPs In vitro-generated CDPs and MPPs were found to exhibit similar and CDPs were cultured under amplification conditions (Flt3L, expression pattern as their in vivo counterparts (Fig. 1A,1B). Next, SCF, hyper–IL-6, and IGF-1) and analyzed by flow cytometry for CDPs and MPPs were subjected to genome-wide gene expression the MPP/CDP markers at days 2, 5, and 8 postsorting. Inter- profiling. PCA demonstrated that in vivo CDPs and in vitro- estingly, only MPPs were able to extensively proliferate and re- amplified CDPs are very similar on the global gene expression produce both MPPs and CDPs (Fig. 2C,2D). CDPs showed a level (Fig. 1C). significantly lower proliferation capacity and even under growth conditions differentiated into CD11c+MHC-II+ DCs (Fig. 2C–E MPPs exhibit multilineage potential and are upstream of CDPs and data not shown). To investigate the differentiation potential of the in vitro-generated In summary, all of this suggests that CDPs are positioned down- MPPs and CDPs, cells were subjected to FACS sorting (Fig. 2A) stream of MPPs and have a restricted differentiation potential. and subsequently to colony formation assay. MPPs generated 80 6 4 colonies per 5000 cells seeded (.40 cells/colony) (Fig. In vitro-amplified CDPs readily differentiate into multiple DC 2B). CFU granulocytes (CFU-Gs), CFU macrophages (CFU-Ms), subsets CFU granulocyte/macrophage colonies (CFU-GMs), and a low Next, we determined in more detail how MPPs and CDPs differ number of mixed colonies with erythrocytes and megakaryocytes in their DC differentiation potential and kinetics. Amplified MPPs (CFU granulocyte, erythrocyte, macrophage, megakaryocyte [CFU- and CDPs were sorted and DC differentiation was induced by GEMM]) were found (Fig. 2B, Supplemental Fig. 3A). On the con- addition of Flt3L, M-CSF, or GM-CSF. Both MPPs and CDPs gave trary, the number of colonies from CDPs was 10 times lower (on rise to pDCs and cDCs when cultured with Flt3L or M-CSF (Fig. average, seven colonies comprising 10–20 cells per 5000 cells). All 3A,3B). In addition, under GM-CSF conditions, both progenitors colonies generated from CDPs were CFU-M (Fig. 2B, Supplemen- gave rise to CD11c+MHC-II+ DCs (Fig. 3C). tal Fig. 3B). CDPs differentiated into CD11c+ DCs with considerably faster The limited differentiation potential of CDPs in the colony kinetics, reaching 70–90% CD11c+ cells already after 3 d in formation assay indicated that CDPs are positioned downstream comparison with 10–30% in MPP cultures (Fig. 3D). Consistent

FIGURE 4. CDPs exhibit a DC-primed gene ex- pression profile. MPPs and CDPs were obtained by in vitro culture and FACS sorted as Flt32/loc-kithi 2 + int + 2 2

by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. CD11c and Flt3 c-kit M-CSFR CD11c IL-7Ra cells, respectively. cDCs and pDCs were generated from amplified progenitors with Flt3L and FACS sorted at differentiation day 6 as CD11c+CD11b+ and CD11c+ CD11b2PDCA-1+B220+ cells, respectively. Total RNA was extracted and subjected to DNA microarray anal- ysis (A) and semiquantitative RT-PCR analysis (B). A, Hierarchical cluster analysis of gene expression in MPPs, CDPs, cDCs, and pDCs. Gene expression data of 1019 probe sets (encoding for 775 genes) that were https://www.jimmunol.org .2-fold differentially regulated between MPPs and CDPs are depicted in heatmap format. Each gene is represented by a single row of colored boxes. Blue, expression levels below median; white expression lev- els equal to median; red, expression levels above me- dian. Cluster I: Pan-DC genes; cluster II: pDC genes; cluster IV: cDC genes; and cluster VI: MPP genes. B, Downloaded from RT-PCR analysis of representative genes found in cluster analysis (A) from sorted MPPs, CDPs, cDCs (upper panel) and pDCs (lower panel). C, Correlation matrix of MPPs, CDPs, pDCs, cDCs and in vivo CDPs is depicted in heatmap format. The Pearson correlation coefficients of sample pairs were calculated for the dif- ferentially expressed genes between MPPs and CDPs shown in A. Red, high correlation; blue, low correla- tion. C, negative control. The Journal of Immunology 5331

Table I. Genes differentially regulated between MPP and CDP

Cluster Gene Function Gene I Pan-DC Membrane protein Flt3, Il10ra, Ccr5 Transcription factor Irf5, Irf8 Other Ctsh, Ctss, Casp6 II pDC Membrane protein Tlr7, Cd28, Cd38 Transcription factor Tcf4, Notch1 III MPP and DC Membrane protein Igf1r IV cDC Membrane protein Cx3cr1, Tlr2, Fcgr1 Transcription factor Id2 Other Lgmn, Ctsc V CDP Membrane protein Csf1r Transcription factor Id3 VI MPP Membrane protein Ncam1, Trfr2, Prom1, Itga6, Cd34 Transcription factor Gata2, Cebpa, Nfe2, Tal1, Gfi1 VII MPP and cDC Membrane protein Kit Selected genes of the clusters shown in Fig. 4A are depicted.

with the fast differentiation kinetics, the proliferative capacity devoid of myeloid, T, and NK cell potential in vivo (data not of CDPs was lower. CDPs showed a limited initial phase of pro- shown). liferation within the first days of culture, and then cell numbers Taken together, both MPPs and CDPs have the full potential rapidly declined upon differentiation (Fig. 3E). In contrast, MPPs to generate all DC subsets in vitro and in vivo. CDPs are con- showed a higher proliferative potential in response to Flt3L and siderably faster in DC differentiation and have a limited prolif- GM-CSF than CDPs and required a longer time to fully differ- eration capacity, which is consistent with their position in differ- entiate into DCs (Fig. 3D,3E). entiation hierarchy downstream of MPPs. We then proceeded to analyze the DC differentiation potential of MPPs and CDPs in vivo by transfer into lethally irradiated CDPs show a gene expression pattern primed for DC recipients. MPPs showed multilineage potential, as expected, and development gave rise to CD11b+ cDCs, CD8a+ cDCs, and pDCs in spleen and To obtain further insight into the CDP differentiation potential, bone marrow (Fig. 3F, Supplemental Figs. 4A,5A, and data not we performed genome-wide gene expression analysis of MPPs, shown). Importantly, also CDPs were able to differentiate into all CDPs, cDCs, and pDCs. A total of 775 genes were found to be DC subsets (Fig. 3G, Supplemental Fig. 4B). In spleen, CDPs differentially expressed by .2-fold between MPPs and CDPs, and preferentially gave rise to DCs, whereas in bone marrow CDP these genes were subjected to hierarchical cluster analysis (Fig. progeny was predominantly CD11c negative, comprising mainly 4A). Clusters I, II, and IV revealed a DC-like expression profile in CD19+ B cells (Fig. 3G, Supplemental Figs. 4B,5B). This B cell CDPs, comprising genes that are already expressed in CDPs and by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. potential was also observed in OP9 cocultures with IL-7 and further upregulated in cDCs and/or pDCs. Conversely, cluster VI Flt3L, in which both MPPs and CDPs were able to produce B220+ contained MPP-associated genes, and these genes were not ex- CD19+ B cells (Supplemental Fig. 5C,5D). CDPs were essentially pressed (or expressed at low levels) in CDPs and DCs. Interestingly,

FIGURE 5. TGF-b1 accelerates DC differentiation and directs subset specification toward cDCs. A, PCA https://www.jimmunol.org of TGF-b1–treated (24 h) and untreated (0 h) CDPs and of cDCs and pDCs based on the expression of genes differentially regulated .2-fold by TGF-b1 (3758 genes). The two principal components PC1 and PC3 showed variances of 45.1 and 18.4%, respectively. B, Bone marrow progenitors were amplified with Flt3L, SCF, hyper–IL-6, and IGF-1 for 7 d. DC differentiation Downloaded from was induced with Flt3L in the presence or absence of TGF-b1 (3 ng/ml). Cultures were analyzed for pDC (CD11c+CD11b2B220+) and cDC (CD11c+CD11b+ B2202) subsets at day 10 of differentiation. Repre- sentative data from at least three independent experi- ments are shown. C and D, MPPs and CDPs were obtained by FACS sorting at day 7 as in Fig. 2A, and DC differentiation was induced with Flt3L. Sixteen hours later, cultures were treated with TGF-b1 (3 ng/ ml) for 48 h. Apoptosis was analyzed with Annexin V and propidium iodide (PI) staining and viable cells (PI-) and apoptotic cells (Annexin V+ PI-) are depicted (C, D, respectively). Error bars represent SD from mean (n = 4). Data shown are from two independent experiments. 5332 TGF-b1 IN DC DIFFERENTIATION FROM COMMON DC PROGENITORS

Table II. TGF-b1–regulated genes

TGF-b1 Effect Gene Function Gene

Upregulation Membrane protein Cd5, Cd11c/Itgax, Cd40, Cd74/MHC-II invariant chain, Cd80, Ccr1, Ccr7, Csf1r, Flt3, Il10ra, Il13ra1, MHC-II/H2-DMa, MHC-II/H2-DMb2 Transcription factor Bhlhe40/Stra13/Dec1, Ciita, Id2, Irf1, Irf4, Irf5, Irf9, , Klf6, Klf7, Nfkb1, Nfkb2, Nfkbie, Notch2, Rel, relB, Runx3, Ski, Skil, Vdr Other Card6, Card11, Ctss, Lyst Downregulation Membrane protein Cd28, Cd38, Tlr7 Transcription factor Cebpb, Gfi1, Id1, Id3, Klf2, Klf16, Myb, , Notch1 Selected genes regulated .2-fold by TGF-b1 are depicted.

CDPs clustered with cDCs and pDCs, indicating the close similarity expression analysis provides further support for the DC-primed char- between CDPs and DCs. acter of CDPs. Cluster I contains Pan-DC genes, such as Flt3 and Irf8 (Fig. TGF-b1 accelerates DC development and impacts on DC 4A,4B, Table I), that are important for cDC and pDC development subset specification (6, 37). Cluster I also includes the transcription factor Irf5 and the cytokine and chemokine receptors Il10ra and Ccr5. Further genes TGF-b1 signaling is of pivotal importance for DC development involved in DC function, including cathepsin and (25, 38, 61–63), yet the impact of TGF-b1 signaling on DC sub- caspase genes (Ctsh, Ctss,andCasp6), with cathepsin S (encoded by set specification from CDPs has not been analyzed. The in vitro Ctss) being particularly important for MHC-II Ag presentation (53). system described in this paper provides an excellent means to Cluster II identifies pDC affiliated genes, such as Tlr7 and the decipher molecular events in response to TGF-b1 in CDPs and transcription factor E2-2/Tcf4 (Fig. 4A,4B, Table I) (41, 42), and its consequences. Therefore, CDPs were obtained by in vitro cul- both genes were upregulated from MPPs to CDPs. The pDC clus- ture, FACS sorted, and treated with TGF-b1 for 4, 8, 12, and 24 h ter also comprised lymphoid-related genes, including the surface for genome-wide gene expression analysis. A total of 3758 genes markers Cd28 and Cd38 and the transcription factor Notch1, which were regulated .2-fold by TGF-b1 and further subjected to is expressed in pDCs (54). clustering. The analysis revealed clusters of up- and downregu- Cluster IV comprises cDC-associated genes, including myeloid lated genes that contained known TGF-b1 target genes, such as genes, like Id2, Tlr2, Fcgr1 (FcgR), and the chemokine receptor Smad7, Id2, and Irf4 (Supplemental Fig. 6) (24, 38, 62). Cx3cr1 (Fig. 4A,4B, Table I). Cluster IV also includes genes en- PCA analysis positioned TGF-b1–treated CDPs between un- coding proteins involved in Ag processing and presentation, such treated CDPs and DCs, indicating that TGF-b1 promotes DC as Lgmn and Ctsc (55). differentiation (Fig. 5A). To further extend this observation, we Cluster V identifies CDP-specific genes, which include Csf1r generated DCs in the presence and absence of TGF-b1 and ana- by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. (M-CSFR) and Id3 (Fig. 4A,4B, Table I). M-CSFR is one surface lyzed DC development by flow cytometry. Strikingly, TGF-b1com- marker used to characterize CDPs, and M-CSFR expression in- deed peaks in CDPs. Id3, similar to Id2, contributes to the re- stricted developmental potential of CDPs (38, 56). Cluster VI comprises hematopoietic stem/progenitor cell genes highly expressed in MPPs and downregulated in CDPs and DCs. These include several transcription factors related to cell prolife- ration and self-renewal, like Gata2, Cebpa, Nfe2, Tal1, and Gfi1 (Fig. 4A,4B, Table I) (57–59). Additionally, a number of stem/ https://www.jimmunol.org progenitor cell-related surface proteins, including Ncam1, Trfr2, Prom1, Itga6 (CD49f), and Cd34, are found in this cluster (57–59). Interestingly, the SCF receptor c-kit, used to characterize MPPs by flow cytometry (Figs. 1, 2A), is found in a separate cluster of genes with prominent expression in MPPs and cDCs (Cluster VII, Fig. 4A,4B, Table I). High c-kit expression on cDCs is consistent Downloaded from with FACS data and previous reports (K. Sere´ and M. Zenke, unpublished observations) (60). Then, MPPs, CDPs, cDCs, and pDCs, including in vivo CDPs from bone marrow, were subjected to correlation matrix analysis using the differentially expressed genes as in Fig. 4A. We dem- onstrate that in vitro-amplified CDPs and in vivo CDPs cluster together and exhibit a high correlation in their gene expression profile (Fig. 4C). FIGURE 6. Transcription factors involved in TGF-b1 signaling and DC Taken together, genes important for DC differentiation are upreg- development. A, Hierarchical clustering and heatmap representation of ulated from MPPs to CDPs, and their expression is further enhanced transcription factors involved in TGF-b1 signaling and DC development in DCs. Concomitantly, expression of proliferation/multipotency- obtained from gene array data (GEO accession number GSE22432). Genes associated genes declines upon CDP and DC differentiation. Impor- that are highly induced upon TGF-b1 treatment after 24 h and abundantly tantly, CDPs express both cDC- and pDC-specific genes, consistent expressed in cDCs are indicated (box). B, RT-PCR analysis of selected withtheircapacitytodifferentiateintobothDCsubsets.Thus,thisgene genes shown in A. The Journal of Immunology 5333

promised pDC development, whereas cDC differentiation remained Gene expression profiling revealed that DC-specific transcrip- unaffected (Fig. 5B). TGF-b1 inhibits B cell development by in- tion profiles are activated from MPPs to CDPs and further to the DC ducing apoptosis in B cell progenitors (56). Similarly, TGF-b1in- subsets. Consistent with their capacity to generate both cDCs and duced apoptosis in CDPs cultured with Flt3L (Fig. 5C,5D). pDCs, CDPs express Pan-DC genes as well as genes specific for To gain further insight into the molecular mechanism of TGF-b1 either DC subset. Interestingly, genes important for DC function, impact on DC differentiation, TGF-b1–regulated genes were sub- such as pattern recognition (Tlr2 and Tlr7), Ag processing (ca- jected to GO (http://www.geneontology.org) overrepresentation anal- thepsin S) (53), and MHC-II presentation (legumain) (55), were ysis. Interestingly, TGF-b1–induced genes were found in GO cate- already upregulated from MPPs to CDPs. Thus, CDPs harbor the gories related to immune cell activation and differentiation, Ag pre- molecular machinery that upon proper stimuli enables full DC sentation, and signal transduction (Supplemental Table II). On the differentiation and DC subset specification. contrary, TGF-b1–downregulated genes were involved in nucleic Concomitantly with lineage specification, hematopoietic cells acid and amino acid metabolism and cell cycle control (Supplemental cease proliferating and undergo cell cycle arrest. In line with the loss Table II), which relates to the growth inhibition and DC differentia- of self-renewal capacity and multilineage potential, CDP differen- tion promoting activity of TGF-b1. tiation from MPPs is accompanied by downregulation of stem/ Given the instructive role of transcription factors and cytokine- progenitor cell-associated gene expression. Thus, in parallel to ac- driven signaling in DC differentiation, we further focused on quiring a DC-primed receptor and transcription factor repertoire, transcription factors and surface molecules regulated by TGF-b1. CDPs downregulate receptors and transcription factors important TGF-b1 induced Pan-DC genes, such as Flt3, CD11c/Itgax, and for stem cell proliferation and self-renewal, such as c-kit and Gata2. Il10ra (Tables I, II). Moreover, TGF-b1 induced a cDC-like tran- The crucial role of TGF-b1 in Langerhans cell development scription profile in CDPs by upregulating the expression of Id2, is well established (25–27), whereas the effect of TGF-b1 on the members of the NF-kB signaling pathway (Nfkb1, Nfkb2, Nfkbie, development of other DC subsets remained elusive. Interestingly, relB, and Rel), MHC-II molecules, such as H2-DMa, H2-DMb2 TGF-b1 has most recently been implicated in regulating hema- and CD74 (MHC-II invariant chain), and costimulatory molecules, topoietic stem cells (HSCs), having a selective impact on myeloid- including Cd40 and Cd80 (Table II). The induction of a cDC-like biased HSCs at the expense of lymphoid-biased HSCs (64). Thus, transcription profile in TGF-b1–treated CDPs was also observed we employed our in vitro system to study the impact of TGF-b1 by hierarchical cluster analysis (Fig. 6A). Interestingly, expression on CDP differentiation and subset specification. We found that: 1) of pDC-affiliated genes, such as Cd28, Cd38, Tlr7,andNotch1,was TGF-b1 pushed the CDP transcription profile toward DCs within downregulated by TGF-b1 similar to the MPP-related genes Myb, 24 h, indicating that TGF-b1 accelerates DC differentiation from Myc,andGfi1 (Tables I, II). These results were further supported CDPs; 2) TGF-b1–treated CDPs acquired a cDC-affiliated tran- by RT-PCR analysis (Fig. 6B). scription factor repertoire, with the induction of cDC differentia- Taken together, our results suggest that TGF-b1 accelerates DC tion instructing transcriptional regulators, including IRF-4 and differentiation and subset specification of CDPs through upregu- several members of the NF-kB family (37, 39, 40); and 3) TGF-b1 lation of a specific set of transcription factors. As a consequence, induced transcription factors that inhibit pDC differentiation, such CDPs acquire a cDC-like molecular signature that instructs cDC as Id2 and IRF-1 (38, 65). differentiation and inhibits pDC development. The results reported in this study are very much in line with by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. the observation that ectopic Id2 expression compromises pDC Discussion development (66) and that Id2-deficient mice have increased pDC An increasingly precise picture of DC development is emerging, frequencies (38). Id2 is an HLH factor that lacks the adjacent basic including the identification of DC-restricted Flt3+c-kitintM-CSFR+ region required for DNA binding and thereby antagonizes activating CDPs (9, 10). In this study, we investigated CDPs amplified to HLH factors, like E2-2/Tcf4 and E2A, in a DNA-independent high cell numbers from bone marrow and show that they re- manner (67). E2-2/Tcf4 is an HLH factor essential for pDC deve- capitulate the cell surface marker profile, function, and gene ex- lopment, as demonstrated by gene knockout and knockdown stud- pression of the in vivo CDPs. A detailed molecular analysis dem- ies (41, 42). Conversely, overexpression of HLH factors, such as onstrates that CDPs exhibit a DC-primed gene transcription pro- HEB and E2A, were reported to promote pDC development (43). https://www.jimmunol.org file. Additionally, the in vitro system allowed investigation of 1) Bhlhe40/Stra13/Dec1 represents another HLH factor that is induced the consecutive development of MPPs into CDPs and their further in CDPs by TGF-b1. Bhlhe40/Stra13 acts as transcriptional re- differentiation into DC subsets and 2) the impact of TGF-b1on pressor while binding to E-box DNA elements and thus antagonizes CDP differentiation. We show that TGF-b1 directs differentiation other HLH factors in a DNA-dependent manner (68). Thus, it is toward cDCs at the expense of pDC development by inducing tempting to speculate that Bhlhe40/Stra13 might be yet another a cDC-affiliated transcription profile. Thus, the in vitro expansion transcription factor that represses pDC development. Interestingly, Downloaded from of the DC committed CDP and its manipulation by TGF-b1 might 24 h TGF-b1 treatment was sufficient to induce also genes related allow the development of new vaccine strategies. to DC function, such as costimulatory molecules, MHC-II genes, Cell commitment and differentiation are regulated by extrinsic and genes involved in Ag processing and presentation (Cd40, Cd80, clues, cytokines and growth factors, and intrinsic transcriptional H2-DMa , CD74, and Ctss). regulators. During development of specific hematopoietic lineages, In summary, our studies uncover the DC-primed expression the gene expression repertoire and the options of stem/progenitor repertoire of CDPs, which is very much in line with CDP func- cells become increasingly restricted, leading to the establishment tion as a lineage-committed progenitor cell. CDP is still bipotential, of a specific lineage from the choice of several. This involves the and we identify TGF-b1 as a determining factor that acts on the activation of lineage-affiliated genes and repression of unrelated bifurcation point of cDC versus pDC development and directs genes. Accordingly, during development from MPPs, our in vitro- cDC subset specification. generated CDPs upregulate receptors for all known DC-poietins Flt3L, GM-CSF, and M-CSF and readily differentiate into all DC Acknowledgments subsets in response to cognate ligands in vitro. In vivo, these CDPs We thank M. Okabe, S. Jung, F. Tacke, S. Rose-John, M. Busslinger, and U. generate all lymphoid tissue DC subsets posttransplantation. Just for providing mice and reagents. We also thank F. Tacke and F. Heymann 5334 TGF-b1 IN DC DIFFERENTIATION FROM COMMON DC PROGENITORS

for cell sorting, J. Ober-Bloebaum for help with bone marrow transplanta- 23. Vremec, D., G. J. Lieschke, A. R. Dunn, L. Robb, D. Metcalf, and K. Shortman. tion experiments, S. Mitzka for excellent technical assistance, and A. Offer- 1997. The influence of granulocyte/macrophage colony-stimulating factor on dendritic cell levels in mouse lymphoid organs. Eur. J. Immunol. 27: 40–44. geld for expert secretary assistance. 24. Larsson, J., and S. Karlsson. 2005. The role of Smad signaling in hematopoiesis. Oncogene 24: 5676–5692. 25. Borkowski, T. A., J. J. Letterio, A. G. Farr, and M. C. Udey. 1996. A role for Disclosures endogenous transforming growth factor beta 1 in Langerhans cell biology: the The authors have no financial conflicts of interest. skin of transforming growth factor beta 1 null mice is devoid of epidermal Langerhans cells. J. Exp. Med. 184: 2417–2422. 26. Strobl, H., E. Riedl, C. Scheinecker, C. Bello-Fernandez, W. F. Pickl, References K. Rappersberger, O. Majdic, and W. Knapp. 1996. TGF-beta 1 promotes in vitro 1. Steinman, R. M., and J. Banchereau. 2007. Taking dendritic cells into medicine. development of dendritic cells from CD34+ hemopoietic progenitors. J. Immu- Nature 449: 419–426. nol. 157: 1499–1507. 2. Shortman, K., and S. H. Naik. 2007. Steady-state and inflammatory dendritic- 27. Strobl, H., C. Bello-Fernandez, E. Riedl, W. F. Pickl, O. Majdic, S. D. Lyman, cell development. Nat. Rev. Immunol. 7: 19–30. and W. Knapp. 1997. flt3 ligand in cooperation with transforming growth factor- 3. Merad, M., and M. G. Manz. 2009. Dendritic cell homeostasis. Blood 113: 3418– beta1 potentiates in vitro development of Langerhans-type dendritic cells and 3427. allows single-cell dendritic cell cluster formation under serum-free conditions. 4. Manz, M. G., D. Traver, T. Miyamoto, I. L. Weissman, and K. Akashi. 2001. Blood 90: 1425–1434. Dendritic cell potentials of early lymphoid and myeloid progenitors. Blood 97: 28. Zenke, M., and T. Hieronymus. 2006. Towards an understanding of the tran- 3333–3341. scription factor network of dendritic cell development. Trends Immunol. 27: 5. D’Amico, A., and L. Wu. 2003. The early progenitors of mouse dendritic cells 140–145. and plasmacytoid predendritic cells are within the bone marrow hemopoietic 29. Wu, L., and Y. J. Liu. 2007. Development of dendritic-cell lineages. Immunity precursors expressing Flt3. J. Exp. Med. 198: 293–303. 26: 741–750. 6. Onai, N., A. Obata-Onai, R. Tussiwand, A. Lanzavecchia, and M. G. Manz. 30. Laouar, Y., T. Welte, X. Y. Fu, and R. A. Flavell. 2003. STAT3 is required for 2006. Activation of the Flt3 signal transduction cascade rescues and enhances Flt3L-dependent dendritic cell differentiation. Immunity 19: 903–912. type I interferon-producing and dendritic cell development. J. Exp. Med. 203: 31. Carotta, S., A. Dakic, A. D’Amico, S. H. M. Pang, K. T. Greig, S. L. Nutt, and 227–238. L. Wu. 2010. The transcription factor PU.1 controls dendritic cell development 7. Fogg, D. K., C. Sibon, C. Miled, S. Jung, P. Aucouturier, D. R. Littman, A. Cumano, and Flt3 cytokine receptor expression in a dose-dependent manner. Immunity 32: and F. Geissmann. 2006. A clonogenic bone marrow progenitor specific for mac- 628–641. rophages and dendritic cells. Science 311: 83–87. 32. Guerriero, A., P. B. Langmuir, L. M. Spain, and E. W. Scott. 2000. PU.1 is 8. Auffray, C., D. K. Fogg, E. Narni-Mancinelli, B. Senechal, C. Trouillet, N. Saederup, required for myeloid-derived but not lymphoid-derived dendritic cells. Blood 95: J. Leemput, K. Bigot, L. Campisi, M. Abitbol, et al. 2009. CX3CR1+ CD115+ 879–885. CD135+ common macrophage/DC precursors and the role of CX3CR1 in their re- 33. Anderson, K. L., H. Perkin, C. D. Surh, S. Venturini, R. A. Maki, and B. E. Torbett. sponse to inflammation. J. Exp. Med. 206: 595–606. 2000. Transcription factor PU.1 is necessary for development of thymic and myeloid 9. Onai, N., A. Obata-Onai, M. A. Schmid, T. Ohteki, D. Jarrossay, and M. G. Manz. progenitor-derived dendritic cells. J. Immunol. 164: 1855–1861. 2007. Identification of clonogenic common Flt3+M-CSFR+ plasmacytoid and 34. Georgopoulos, K., M. Bigby, J. H. Wang, A. Molnar, P. Wu, S. Winandy, and conventional dendritic cell progenitors in mouse bone marrow. Nat. Immunol. 8: A. Sharpe. 1994. The Ikaros gene is required for the development of all lym- 1207–1216. phoid lineages. Cell 79: 143–156. 10. Naik, S. H., P. Sathe, H. Y. Park, D. Metcalf, A. I. Proietto, A. Dakic, S. Carotta, 35. Allman, D., M. Dalod, C. Asselin-Paturel, T. Delale, S. H. Robbins, M. O’Keeffe, M. Bahlo, A. Papenfuss, et al. 2007. Development of plasmacytoid G. Trinchieri, C. A. Biron, P. Kastner, and S. Chan. 2006. Ikaros is required for and conventional dendritic cell subtypes from single precursor cells derived plasmacytoid dendritic cell differentiation. Blood 108: 4025–4034. in vitro and in vivo. Nat. Immunol. 8: 1217–1226. 36. Wu, L., A. Nichogiannopoulou, K. Shortman, and K. Georgopoulos. 1997. Cell- 11. Waskow, C., K. Liu, G. Darrasse-Je`ze, P. Guermonprez, F. Ginhoux, M. Merad, autonomous defects in dendritic cell populations of Ikaros mutant mice point to T. Shengelia, K. Yao, and M. Nussenzweig. 2008. The receptor tyrosine kinase a developmental relationship with the lymphoid lineage. Immunity 7: 483–492. Flt3 is required for dendritic cell development in peripheral lymphoid tissues. 37. Gabriele, L., and K. Ozato. 2007. The role of the interferon regulatory factor Nat. Immunol. 9: 676–683. (IRF) family in dendritic cell development and function. Cytokine Growth Factor 12. Liu, K., G. D. Victora, T. A. Schwickert, P. Guermonprez, M. M. Meredith, Rev. 18: 503–510. K. Yao, F. F. Chu, G. J. Randolph, A. Y. Rudensky, and M. Nussenzweig. 2009. 38. Hacker, C., R. D. Kirsch, X. S. Ju, T. Hieronymus, T. C. Gust, C. Kuhl, T. Jorgas, by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. In vivo analysis of dendritic cell development and homeostasis. Science 324: S. M. Kurz, S. Rose-John, Y.Yokota,and M. Zenke. 2003. Transcriptional profiling 392–397. identifies Id2 function in dendritic cell development. Nat. Immunol. 4: 380–386. 13. Inaba, K., M. Inaba, N. Romani, H. Aya, M. Deguchi, S. Ikehara, S. Muramatsu, 39. Burkly, L., C. Hession, L. Ogata, C. Reilly, L. A. Marconi, D. Olson, R. Tizard, and R. M. Steinman. 1992. Generation of large numbers of dendritic cells from R. Cate, and D. Lo. 1995. Expression of relB is required for the development of mouse bone marrow cultures supplemented with granulocyte/macrophage thymic medulla and dendritic cells. Nature 373: 531–536. colony-stimulating factor. J. Exp. Med. 176: 1693–1702. 40. Wu, L., A. D’Amico, K. D. Winkel, M. Suter, D. Lo, and K. Shortman. 1998. 14. Brasel, K., T. De Smedt, J. L. Smith, and C. R. Maliszewski. 2000. Generation of RelB is essential for the development of myeloid-related CD8alpha- murine dendritic cells from flt3-ligand-supplemented bone marrow cultures. dendritic cells but not of lymphoid-related CD8alpha+ dendritic cells. Immunity Blood 96: 3029–3039. 9: 839–847. 15. Brawand, P., D. R. Fitzpatrick, B. W. Greenfield, K. Brasel, C. R. Maliszewski, 41. Cisse, B., M. L. Caton, M. Lehner, T. Maeda, S. Scheu, R. Locksley, and T. De Smedt. 2002. Murine plasmacytoid pre-dendritic cells generated from D. Holmberg, C. Zweier, N. S. den Hollander, S. G. Kant, et al. 2008. Tran- Flt3 ligand-supplemented bone marrow cultures are immature APCs. J. Immu- scription factor E2-2 is an essential and specific regulator of plasmacytoid https://www.jimmunol.org nol. 169: 6711–6719. dendritic cell development. Cell 135: 37–48. 16. Naik, S. H., A. I. Proietto, N. S. Wilson, A. Dakic, P. Schnorrer, M. Fuchsberger, 42. Nagasawa, M., H. Schmidlin, M. G. Hazekamp, R. Schotte, and B. Blom. 2008. M. H. Lahoud, M. O’Keeffe, Q. X. Shao, W. F. Chen, et al. 2005. Cutting edge: Development of human plasmacytoid dendritic cells depends on the combined generation of splenic CD8+ and CD8- dendritic cell equivalents in Fms-like action of the basic helix-loop-helix factor E2-2 and the Ets factor Spi-B. Eur. J. tyrosine kinase 3 ligand bone marrow cultures. J. Immunol. 174: 6592–6597. Immunol. 38: 2389–2400. 17. Fancke, B., M. Suter, H. Hochrein, and M. O’Keeffe. 2008. M-CSF: a novel 43. Schotte, R., M. Nagasawa, K. Weijer, H. Spits, and B. Blom. 2004. The ETS plasmacytoid and conventional dendritic cell poietin. Blood 111: 150–159. transcription factor Spi-B is required for human plasmacytoid dendritic cell 18. Maraskovsky, E., K. Brasel, M. Teepe, E. R. Roux, S. D. Lyman, K. Shortman, development. J. Exp. Med. 200: 1503–1509.

Downloaded from and H. J. McKenna. 1996. Dramatic increase in the numbers of functionally 44. Okabe, M., M. Ikawa, K. Kominami, T. Nakanishi, and Y. Nishimune. 1997. mature dendritic cells in Flt3 ligand-treated mice: multiple dendritic cell sub- ‘Green mice’ as a source of ubiquitous green cells. FEBS Lett. 407: 313–319. populations identified. J. Exp. Med. 184: 1953–1962. 45. Jung, S., J. Aliberti, P. Graemmel, M. J. Sunshine, G. W. Kreutzberg, A. Sher, 19. Pulendran, B., J. Banchereau, S. Burkeholder, E. Kraus, E. Guinet, C. Chalouni, and D. R. Littman. 2000. Analysis of fractalkine receptor CX(3)CR1 function by D. Caron, C. Maliszewski, J. Davoust, J. Fay, and K. Palucka. 2000. Flt3-ligand targeted deletion and green fluorescent protein reporter gene insertion. Mol. Cell. and granulocyte colony-stimulating factor mobilize distinct human dendritic cell Biol. 20: 4106–4114. subsets in vivo. J. Immunol. 165: 566–572. 46. Hieronymus, T., T. C. Gust, R. D. Kirsch, T. Jorgas, G. Blendinger, M. Goncharenko, 20. McKenna, H. J., K. L. Stocking, R. E. Miller, K. Brasel, T. De Smedt, K. Supplitt, S. Rose-John, A. M. Mu¨ller, and M. Zenke. 2005. Progressive and E. Maraskovsky, C. R. Maliszewski, D. H. Lynch, J. Smith, B. Pulendran, et al. controlled development of mouse dendritic cells from Flt3+CD11b+ progenitors 2000. Mice lacking flt3 ligand have deficient hematopoiesis affecting hemato- in vitro. J. Immunol. 174: 2552–2562. poietic progenitor cells, dendritic cells, and natural killer cells. Blood 95: 3489– 47. Gentleman, R. C., V. J. Carey, D. M. Bates, B. Bolstad, M. Dettling, S. Dudoit, 3497. B. Ellis, L. Gautier, Y. Ge, J. Gentry, et al. 2004. Bioconductor: open software 21. Kingston, D., M. A. Schmid, N. Onai, A. Obata-Onai, D. Baumjohann, and development for computational biology and bioinformatics. Genome Biol. 5: R80. M. G. Manz. 2009. The concerted action of GM-CSF and Flt3-ligand on in vivo 48. Benjamini, Y., and Y. Hochberg. 1995. Controlling the false discovery rate: A dendritic cell homeostasis. Blood 114: 835–843. practical and powerful approach to multiple testing. J. R. Stat. Soc. B. 57: 289–300. 22. MacDonald, K. P., V. Rowe, H. M. Bofinger, R. Thomas, T. Sasmono, D. A. Hume, 49. Kumar, L., and E. Futschik M. 2007. Mfuzz: a software package for soft clus- and G. R. Hill. 2005. The colony-stimulating factor 1 receptor is expressed on tering of microarray data. Bioinformation 2: 5–7. dendritic cells during differentiation and regulates their expansion. J. Immunol. 50. Rousseeuw, P. 1987. Silhouettes: a graphical aid to the interpretation and vali- 175: 1399–1405. dation of cluster analysis. J. Comput. Appl. Math. 20: 53–65. The Journal of Immunology 5335

51. Alexa, A., J. Rahnenfu¨hrer, and T. Lengauer. 2006. Improved scoring of func- cells regulates T helper cell differentiation and allergic asthma. Nat. Med. 14: tional groups from gene expression data by decorrelating GO graph structure. 565–573. Bioinformatics 22: 1600–1607. 61. Fainaru,O.,E.Woolf,J.Lotem,M.Yarmus,O.Brenner,D.Goldenberg,V.Negreanu 52. Ashburner, M., C. A. Ball, J. A. Blake, D. Botstein, H. Butler, J. M. Cherry, Y. Bernstein, D. Levanon, S. Jung, and Y. Groner. 2004. Runx3 regulates mouse A. P. Davis, K. Dolinski, S. S. Dwight, J. T. Eppig, et al. 2000. Gene ontology: TGF-beta-mediated dendritic cell function and its absence results in airway in- tool for the unification of biology. Nat. Genet. 25: 25–29. flammation. EMBO J. 23: 969–979. 53. Nakagawa, T. Y., W. H. Brissette, P. D. Lira, R. J. Griffiths, N. Petrushova, 62. Ju, X. S., D. Ruau, P. Ja¨ntti, K. Sere´, C. Becker, E. Wiercinska, C. Bartz, J. Stock, J. D. McNeish, S. E. Eastman, E. D. Howard, S. R. Clarke, et al. 1999. B. Erdmann, S. Dooley, and M. Zenke. 2007. Transforming growth factor beta1 Impaired invariant chain degradation and antigen presentation and diminished up-regulates interferon regulatory factor 8 during dendritic cell development. collagen-induced arthritis in cathepsin S null mice. Immunity 10: 207–217. Eur. J. Immunol. 37: 1174–1183. 54. Pelayo, R., J. Hirose, J. Huang, K. P. Garrett, A. Delogu, M. Busslinger, and 63. Heinz, L. X., B. Platzer, P. M. Reisner, A. Jo¨rgl, S. Taschner, F. Go¨bel, and P. W. Kincade. 2005. Derivation of 2 categories of plasmacytoid dendritic cells H. Strobl. 2006. Differential involvement of PU.1 and Id2 downstream of TGF- in murine bone marrow. Blood 105: 4407–4415. beta1 during Langerhans-cell commitment. Blood 107: 1445–1453. 55. Manoury, B., E. W. Hewitt, N. Morrice, P. M. Dando, A. J. Barrett, and C. Watts. 64. Challen, G. A., N. C. Boles, S. M. Chambers, and M. A. Goodell. 2010. Distinct 1998. An asparaginyl endopeptidase processes a microbial antigen for class II hematopoietic stem cell subtypes are differentially regulated by TGF-beta1. Cell MHC presentation. Nature 396: 695–699. 56. Kee, B. L., R. R. Rivera, and C. Murre. 2001. Id3 inhibits B lymphocyte pro- Stem Cell 6: 265–278. genitor growth and survival in response to TGF-beta. Nat. Immunol. 2: 242–247. 65. Gabriele, L., A. Fragale, P. Borghi, P. Sestili, E. Stellacci, M. Venditti, G. Schiavoni, 57. Ivanova, N. B., J. T. Dimos, C. Schaniel, J. A. Hackney, K. A. Moore, and M. Sanchez, F. Belardelli, and A. Battistini. 2006. IRF-1 deficiency skews the dif- I. R. Lemischka. 2002. A stem cell molecular signature. Science 298: 601–604. ferentiation of dendritic cells toward plasmacytoid and tolerogenic features. 58. Ramalho-Santos, M., S. Yoon, Y. Matsuzaki, R. C. Mulligan, and D. A. Melton. J. Leukoc. Biol. 80: 1500–1511. 2002. “Stemness”: transcriptional profiling of embryonic and adult stem cells. 66. Spits, H., F. Couwenberg, A. Q. Bakker, K. Weijer, and C. H. Uittenbogaart. 2000. Science 298: 597–600. Id2 and Id3 inhibit development of CD34(+) stem cells into predendritic cell (pre- 59. Ma˚nsson, R., A. Hultquist, S. Luc, L. Yang, K. Anderson, S. Kharazi, S. Al- DC)2 but not into pre-DC1. Evidence for a lymphoid origin of pre-DC2. J. Exp. Hashmi, K. Liuba, L. Thore´n, J. Adolfsson, et al. 2007. Molecular evidence for Med. 192: 1775–1784. hierarchical transcriptional lineage priming in fetal and adult stem cells and 67. Kee, B. L. 2009. E and ID proteins branch out. Nat. Rev. Immunol. 9: 175–184. multipotent progenitors. Immunity 26: 407–419. 68. St-Pierre, B., G. Flock, E. Zacksenhaus, and S. E. Egan. 2002. Stra13 homo- 60. Krishnamoorthy, N., T. B. Oriss, M. Paglia, M. Fei, M. Yarlagadda, dimers repress transcription through class B E-box elements. J. Biol. Chem. 277: B. Vanhaesebroeck, A. Ray, and P. Ray. 2008. Activation of c-Kit in dendritic 46544–46551. by guest on September 30, 2021. Copyright 2010 Pageant Media Ltd. https://www.jimmunol.org Downloaded from