Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Threshold Levels of Gfi1 Maintain E2A Activity for B Cell Commitment via Repression of Id1

  • Jennifer Fraszczak,

    Affiliations Institut de recherches cliniques de Montréal (IRCM), Montréal, Québec, Canada, Département de microbiologie, infectiologie et immunologie, Université de Montréal, Montréal, Québec, Canada

  • Anne Helness,

    Affiliations Institut de recherches cliniques de Montréal (IRCM), Montréal, Québec, Canada, Division of Experimental Medicine, McGill University, Montreal, Canada

  • Riyan Chen,

    Affiliation Institut de recherches cliniques de Montréal (IRCM), Montréal, Québec, Canada

  • Charles Vadnais,

    Affiliations Institut de recherches cliniques de Montréal (IRCM), Montréal, Québec, Canada, Département de microbiologie, infectiologie et immunologie, Université de Montréal, Montréal, Québec, Canada

  • François Robert,

    Affiliations Institut de recherches cliniques de Montréal (IRCM), Montréal, Québec, Canada, Département de médecine, Université de Montréal, Montréal, Québec, Canada

  • Cyrus Khandanpour,

    Affiliation Department of Hematology, University Hospital, Essen, Germany

  • Tarik Möröy

    Tarik.Moroy@ircm.qc.ca

    Affiliations Institut de recherches cliniques de Montréal (IRCM), Montréal, Québec, Canada, Département de microbiologie, infectiologie et immunologie, Université de Montréal, Montréal, Québec, Canada, Division of Experimental Medicine, McGill University, Montreal, Canada

Threshold Levels of Gfi1 Maintain E2A Activity for B Cell Commitment via Repression of Id1

  • Jennifer Fraszczak, 
  • Anne Helness, 
  • Riyan Chen, 
  • Charles Vadnais, 
  • François Robert, 
  • Cyrus Khandanpour, 
  • Tarik Möröy
PLOS
x

Abstract

A regulatory circuit that controls myeloid versus B lymphoid cell fate in hematopoietic progenitors has been proposed, in which a network of the transcription factors Egr1/2, Nab, Gfi1 and PU.1 forms the core element. Here we show that a direct link between Gfi1, the transcription factor E2A and its inhibitor Id1 is a critical element of this regulatory circuit. Our data suggest that a certain threshold of Gfi1 is required to gauge E2A activity by adjusting levels of Id1 in multipotent progenitors, which are the first bipotential myeloid/lymphoid-restricted progeny of hematopoietic stem cells. If Gfi1 levels are high, Id1 is repressed enabling E2A to activate a specific set of B lineage genes by binding to regulatory elements for example the IL7 receptor gene. If Gfi1 levels fall below a threshold, Id1 expression increases and renders E2A unable to function, which prevents hematopoietic progenitors from engaging along the B lymphoid lineage.

Introduction

B-lymphocytes are the principal antibody producing cells and are indispensable for an efficient humoral immune response. B cell differentiation begins in the bone marrow (BM) by the generation of multipotent progenitors (MPPs) from hematopoietic stem cells (HSCs) [1, 2]. A subset of MPP cells that express Flt3 receptor, called lymphoid primed multipotent progenitors (LMPPs) [3, 4] become committed to the lymphoid lineage and generate the common lymphoid progenitors, CLPs [3, 5, 6]. Under specific conditions, both CLPs and LMPPs can generate T and B cells [79]. Commitment towards the B cell lineage starts after the CLP stage, when cells begin to express the marker B220 and to increase Rag1 expression. These cells, called pre-pro B cells, complete B-lineage commitment by differentiation into CD19 expressing pro B cells. They continue to differentiate along multiple stages until functional effector B cells [10].

A very complex network involving many transcription factors and cytokines such as Flt3L [11, 12] and IL-7 [13] [14] controls lymphoid commitment and B cell differentiation. B lineage specification is supported by the expression of three major transcription factors, E2A [1517], Ikaros [18, 19] and PU.1 [20, 21] in early progenitor cells including HSCs and MPPs, allowing the formation of LMPPs. The expression levels of these three transcription factors in progenitors are known to determine cell fate decisions. For instance, the transcription factor PU.1 regulates B lymphoid versus myeloid cell lineage choice in a dose dependent manner in hematopoietic progenitors, in particular in MPPs [2224]. Moreover, in CLPs, E2A controls the expression of EBF1, which is a major transcription factor in B cell differentiation [25, 26]. EBF1 acts in concert with E2A and Foxo1 to regulate genes essential for B cell development such as Pax5 [27]. These findings propose a transcriptional regulatory network that appear to function in a recurring manner to govern cell fate choices, especially in progenitors such as LMPPs that retain multilineage potential [5, 28].

The transcription repressor Gfi1 is a key element of this network orchestrating progenitor cell fate between myeloid and lymphoid lineages [29]. Indeed, Gfi1 deficient mice show impaired B cell differentiation, which can be partially rescued by reducing PU.1 expression levels [23], but its role in this complex transcriptional network is not fully understood.

To further understand how Gfi1 is functioning in early B lymphoid commitment and differentiation, we decided to study its participation in transcription factor regulatory circuits in early lymphoid and multipotential progenitors. We report here that a threshold level of Gfi1 protein expression is required to support B cell differentiation. We identify LMPPs as critical hematopoietic progenitors that require high levels of Gfi1 expression to repress Id1 to specifically sustain B-cell commitment. We propose a model in which Gfi1 is required to maintain Id1 at low levels, which leave E2A active to ensure the expression of those E2A target genes that are important for B cell differentiation.

Materials and Methods

Mice

The Institutional Review Board of the IRCM approved all animal protocols and experimental procedures were performed in compliance with the IRCM guidelines. The animals were euthanized by CO2 and all efforts were made to minimize suffering. GFI1 KO, GFI1 KI, GFI1-GFP and GFI1-P2A mice used in this study have been described previously [3034]. GFI1 KD mice were generated following a previously described strategy to generate GFI1 KI mice [32, 35]. GFI1 KI mice were obtained by inserting the human GFI1-encoding cDNA into the murine Gfi1 locus (32). In the KD mice, the targeted locus still retained the neo cassette in an antisense direction leading to a low expression of the human GFI1 knock-in transgene. MB1-cre mice were obtained from Jackson laboratories (Bar Harbor, Maine, USA) [36]. Mice have been bred on C57BL/6 background for at least 10 generations and were maintained in Specific-Pathogen-Free Plus environment.

OP9 and OP9-DL1 culture

For the B cell differentiation, sorted HSCs, MPPs, LMPPs, CLPs and pre-pro-B cells were cultured on plated OP9 cells in opti-alpha-modified Eagle Medium (OptiMEM) supplemented with IL-7 (10 ng/mL, Peprotech) and Flt3L (5 ng/mL, peprotech). After 10–12 days of co-culture, cells were analysed by flow cytometry (LSR I, BD Biosciences). For the T cell differentiation, CLPs or LMPPs were sorted and cultured on OP9 cells expressing DL1 in OptiMEM medium with SCF (10 ng/mL, R&D systems), Flt3L (5 ng/mL, Peprotech) and IL-7 (1 ng/mL, Peprotech). After 18–21 days of co culture, cells were analyzed by flow cytometry. For the myeloid differentiation, sorted LMPPs were cultured on plated OP9 cells in OptiMEM supplemented with myeloid cytokine mix described previously [37]. After 10 days, cells were analyzed by flow cytometry for Gr1 and Mac1 expression. Morpholino experiments were done with LSK cells sorted by flow cytometry were cultured in the presence of Flt3L and Il-7 for 1 hr under stroma-free conditions. Morpholinos were added to the culture in the presence of EndoPorter followed the manufacturer's instructions (Gene Tools, USA). After 4 hrs progenitor cells were transferred onto OP9 cells. After 7 days, cells were treated a second time with morpholinos and co-cultured on OP9 cells. B cell differentiation was assessed by flow cytometry for B220, CD43 and CD19. The OP9 and OP9-DL1 cells were kindly provided by Pr. Juan Carlos Zúñiga-Pflücker.

Chromatin immuno-precipitation (ChIP) assay

Histone acetylation and E2A ChIPs were performed on 0.1 x106 and 0.5 × 106 sorted LSK cells, respectively, cross-linked with 0.1% or 1% formaldehyde for 8 minutes and quenched with 125mM glycine. Samples were immuno-precipitated with 2 μg anti-E2A antibody (N-649 X; Santa Cruz Biotechnology), anti-H3K9ac antibody (Ab4441; Abcam), anti-H3K27ac antibody (Ab4729; Abcam) or control anti-IgG normal rabbit antibody (12–370; Millipore) coupled with Dynabeads beads coated with Protein G and A (Life Technologies). ChIP DNA was analyzed by quantitative PCR (qPCR) using SYBR green. The enrichment (relative to Input and a negative control intergenic region) was calculated using the ΔΔCt method. Primers are in S1 Table.

Western blot analysis

Cells were lysed in NP-40 buffer in the presence of protease inhibitor cocktail (Complete Mini, Roche Diagnostics). The following antibodies were used for immunoblotting and co-immuno-precipitation: Anti PU.1 (H-135, Santa Cruz), anti Id1 (C-20, Santa Cruz), anti Id2 (C-20, Santa Cruz), anti Id3 (C-20, Santa Cruz), anti E2A (N-649, Santa Cruz), anti Lamin B (C-20, Santa Cruz), anti Gfi1 (R&D systems) and anti β-Actin (Ac-15, Sigma-aldrich).

Expression profiling

Expression profiling was carried out using Affymetrix arrays according to published procedures [38]. Raw data files were analyzed using the R Software [R reference], including the affy [39] and Limma packages [40]. Data was normalized using the Robust Multi-array Average (RMA) approach.

Morpholinos

LSK cells were sorted and cultured in the presence of Flt3L and IL-7 for 1 hr under stroma-free conditions. Morpholinos were added to the culture in the presence of EndoPorter followed the manufacturer's instructions (Gene Tools, USA). After 4 hours, progenitor cells were transferred on OP9 cells. After 4 days, the cells were harvested from the co-culture and cultured in the presence of new cytokines and morpholinos with Endoporter for 4 hours. Then the cells were transferred on a new fresh OP9 cell layer. Every 4 days, cells were transferred on new OP9 cells and new cytokines were added.

The morpholino sequences: Anti-Id1 #1: CGGCACTGCCACTGGCGACCTTCAT; Anti-Id1 #2: AGAACAGAGTGTGGGAAGAGAACAA; Control: CCTCTTACCTCAGTTACAATTTATA were used as previously described in kosan et al. [37].

Statistical analysis

Two tailed student’s t-test was used to calculate p-values where indicated. A p-value ≤0.05 was considered as statistically significant.

GEO accession codes for publicly available data sets

E2A, Foxo1 and EBF1, GSE21978; Brg1 and Ikaros, GSE66978; PU.1, GSE21512.

Supplemental method about FACS stainings and RNA isolation are provided in S1 Appendix.

Results

The control of B cell development required a certain level of Gfi1

We have generated mutant mice by gene targeting, in which the endogenous Gfi1 gene is replaced by the human Gfi1 cDNA (termed “KI” for “knockin”) [32]. During the analysis of these mice, we noticed that when the Neo cassette was not deleted, the expression level of human Gfi1 was reduced to about 20–30% of the expression level of the control WT or KI mice without Neo cassette (Fig 1A and 1B) [35]. This provided us with several mouse models expressing different Gfi1 levels (Fig 1A and 1B): WT mice expressing a normal level of murine Gfi1, KI mice expressing a normal level of human Gfi1 instead of the murine Gfi1, heterozygote mice for Gfi1 (Het) expressing around 50–60% of WT Gfi1 levels, animals expressing a reduced level of human Gfi1 (termed “KD” for “knockdown”) and KO mice that were previously described [31] and completely lack Gfi1 expression (Fig 1A and 1B). We found decreased frequencies and absolute numbers of B220+CD19+, B220+IgM-, B220lowIgM+ and B220highIgM+ B cells (Fig 1C–1E) in the BM of KD mice, similar to what is seen and was previously described for Gfi1 KO mice [23, 4143] but no significant difference between Het, KI and WT mice. This suggests that a minimal threshold level of Gfi1 expression exists below which B cell differentiation is no longer maintained. Similar decreases in B cell frequencies were also seen in the BM from Mx-Cre Gfi1flox/flox mice where Gfi1 is deleted along with the neo marker gene after pIpC injections (S1A–S1C Fig), excluding an effect of the neo marker in KD mice. In Gfi1P2A/P2A mutant mice (P2A), in which Gfi1 is functionally disabled by a mutation in its N-terminal SNAG repressor impairing binding to LSD1 [30], we observed the same defect in B cell frequencies and numbers as in KD and KO mice (S1D–S1F Fig). A competitive transplantation experiment showed that Gfi1 KD and, as already described, Gfi1 KO BM cells [44] were unable to efficiently reconstitute hematopoiesis in recipient mice (Fig 1F). Among CD45.2+ cells in the blood of the transplanted recipient CD45.1 mice, the percentage of cells expressing B220 and CD19 was also lower in mice transplanted with KD BM cells compared to mice transplanted with WT and KI cells while the percentage of myeloid cells was not affected (Fig 1G) suggesting that the defect is likely intrinsic to the transferred cells. However, since the change in B cell reconstitution is similar to the change in overall reconstitution it cannot be excluded that the defect occurs already in the HSCs or MPPs.

thumbnail
Fig 1. The control of B cell development by Gfi1 is dose dependent.

(A) Western blot of Gfi1 levels in thymocytes isolated from Gfi1 WT, KI, Het, KD and KO mice. Lamin B was used as a loading control. Low exp. = Low exposure, high exp. = high exposure. (B) Semi quantification of relative Gfi1 expression in thymocytes from WT, KI, Het, KD and KO mice with ImageJ software by using chemiluminescent film system. (C) Flow cytometry analysis of BM B220+CD19+ B cells from WT, KI, Het, KD and KO mice. (D) Flow cytometry analysis of BM B cells from WT, KI, Het, KD and KO mice. Cells were stained for B220 and IgM surface markers. Each plot is representative of the experiment average. (E) Absolute number of BM B220+CD19+ B cells, BM B220+IgMneg, BM B220lowIgM+ and BM B220highIgM+. (F) and (G) CD45.1 mice were lethally irradiated (9,5 Gy) and transplanted with an equal number of CD45.1 BM cells and CD45.2 BM cells from WT, KI, KD or KO mice. After 4 and 8 weeks, blood cells from the transplanted mice were analyzed for the expression of CD45.1 and CD45.2 markers (F) and also for B220 and CD19 or Gr1 and Mac1 (G). All FACS plots in Fig 1 are representative of at least five independent experiments and a minimum of five mice were used to determine absolute numbers of B cell subsets in BM.

https://doi.org/10.1371/journal.pone.0160344.g001

Gfi1 is important at the stage where early B cell progenitors are formed

A significant reduction in the frequencies of B cell progenitors contained in both Hardy fractions A (pre-pro B cells, B220+CD43+BP-1-HSA-) and B (pro B cells, B220+CD43+BP-1-HSA+) were observed in KD and KO mice (Fig 2A) and also in P2A mice (S2A Fig). The cell surface levels of CD19 contained in fraction B were similar between all animals (Figs 2B and S2B) suggesting that the few fraction B cells remaining in KO or KD are committed B cells. Cell numbers of fractions A and B were reduced in KD, KO and P2A mice compared to WT, KI and Het (Figs 2C and S2C). Another gating strategy confirmed the reduction of the pre-pro B cell population, which is identical to Fraction A, in KD and KO mice (Figs 2D and 2E and S2D and S2E). The surface levels of CD93 and c-Kit (Fig 2F) in pre-pro B cells (lin-CD19-B220+CD43+) were similar between all the mice suggesting that the few cells present in fraction A of KO or KD mice underwent regular differentiation. Another gating strategy [45] confirmed that absence (KO) or low dose of Gfi1 (KD) lead to a decrease of B cell progenitors, whereas Het mice have a sufficient level of Gfi1 to produce normal numbers of B cell progenitors (Fig 2G and 2H).

thumbnail
Fig 2. Gfi1 is important at the stage where early B cell progenitors are formed.

(A) Flow cytometry analysis of B220, CD43, HSA, and BP-1 surface expression on BM cells. B220+CD43+ cells were electronically gated and analyzed for HSA and BP-1 expression to identify Fraction A (A, B220+CD43+HSA-BP1), Fraction B (B, B220+CD43+HSA+BP1), Fraction C (C, B220+CD43+HSA+BP1+) and Fraction C’ (C’, B220+CD43+HSAhighBP1+). (B) Flow cytometry analysis of CD19 expression on pro B cells. Fraction B cells were analyzed for surface CD19 expression. (C) Absolute numbers of Fraction A-C’ from WT, KI, Het, KD and KO mice. (D) Expression of Lin-, CD19, B220, and CD43 on BM cells. LinCD19 cells were analyzed for the expression of B220 and CD43. (E) Absolute numbers of pre-pro B cells (Lin-CD19-B220+CD43+) contained in the Fraction A from WT, KI, Het, KD and KO mice. (F) Expression of CD93 and c-Kit on pre-pro B cells from WT, KI, Het, KD and KO mice. Lin-CD19-B220+CD43+ cells were analyzed for CD93 expression or c-Kit on the surface. (G) Flow cytometry analysis of B220, CD19, c-Kit, IgM and IgD surface expression on BM cells. B220+CD19+c-Kit-/c-kit+ cells were gated for IgM and IgD expression to define Pro B cells (B220+CD19+c-Kit+IgM-IgD-), Pre B cells (B220+CD19+c-Kit-IgM-IgD-), Immature B cells (B220+CD19+c-Kit-IgM+IgD-) and mature B cells (B220+CD19+c-Kit-IgMlow/+IgD+). (H) Cell numbers of Pro, Pre, Immature (Imm.) and Mature (Mat.) B cells in the BM. All FACS plots are representative of a minimum of five (A, B and D) or at least two (F and G) independent experiments. At least five mice or 2 mice were respectively used to determine absolute number of B cell fractions and B cell progenitors (G and H).

https://doi.org/10.1371/journal.pone.0160344.g002

Gfi1 is required for the generation of LMPPs and CLPs

To determine if the reduction of B cell progenitors is mostly due to an earlier defect during B cell development, we studied the LSK (Lin-c-Kit+Sca1+) compartment, which includes HSCs and the early lymphoid progenitors. Higher frequencies of LSK cells were found in KO and P2A mice confirming earlier studies [44, 46] and in KD mice compared to, Het, KI and WT animals (Figs 3A and 3B and S3A, S3D). Further, frequencies and absolute numbers of CLPs (Lin-IL7R+Flt3+Sca1medc-Kitmed) were severely reduced in KO, P2A and KD mice (Figs 3C and 3D and S3B). Although a similar effect of Gfi1 was also observed for LMPPs (Lin-Sca1+c-Kit+Flt3high), which were severely decreased in KD, KO and P2A mice (Figs 3E, 3G and S3C), both MPP (Lin-c-Kit+Sca1+Flt3low) and ST-HSC progenitors (Lin-c-Kit+Sca1+Flt3-CD48highCD150+) were equally increased in KD and KO mice compared to KI or WT mice (Fig 3F and 3H–3J) but the frequencies of these two subsets within the LSKFlt3low population was similar between the five different mouse mutants (KO, KD, KI, P2A and WT). The Gfi1 expression levels in heterozygous mice did not cause an increase in the LSK population or a decrease in the LMPP subset showing that a 50–60% expression level of Gfi1 is sufficient to support B cell differentiation. The decrease of LMPPs and CLPs in KD and KO mice is not due to an increase of the cell death compared to KI and WT mice (Fig 3K). These data suggest that differentiation blocks rather than cell death are causing the reduction in the LMPP and CLP populations (Fig 3L), when Gfi1 is severely down-regulated or absent suggesting that Gfi1 may play a role at different steps of the B cell differentiation.

thumbnail
Fig 3. Gfi1 is required for the generation of LMPPs.

(A) FACS analysis of LSK cells. BM cells from WT, KI, Het, KD and KO mice were analyzed for Lin-, c-Kit and Sca1 expression. (B) Absolute number of LSK cells in BM from WT, KI, Het, KD and KO mice. (C) FACS analysis of common lymphoid progenitors (CLP, Lin-IL7R+Flt3+Sca1lowc-Kit+) in the BM of WT, KI, Het, KD and KO mice. The percentages at the left and at the right on the dot plots respectively indicate the CLP relative frequency and the CLP absolute frequency. (D) Absolute numbers of CLP in the BM of WT, KI, Het, KD and KO mice. (E) LSK cells were analyzed for Flt3 expression in the BM of WT, KI, Het, KD and KO mice. MPP (Lin-Sca1+c-Kit-Flt3low) and LMPP (Lin-Sca1+c-Kit+Flt3+) are identified on the FACS plot. (F) LSKFlt3- cells were analyzed for CD48 and CD150 cell surface levels. (G) Absolute numbers of LMPP (Lin-Sca1+c-Kit+Flt3+) in the BM of WT, KI, Het, KD and KO mice. (H) Absolute numbers of MPP (Lin-Sca1+c-Kit+Flt3low) in the WT, KI, Het, KD and KO BM. (I) Absolute numbers of ST-HSC (Lin-Sca1+c-Kit+Flt3-CD150+CD48high) in the WT, KI, Het, KD and KO BM. (J) Absolute numbers of LT-HSC (Lin-Sca1+c-Kit+Flt3-CD150+CD48low) in the WT, KI, Het, KD and KO BM. (K) HSC (Lin-Sca1+c-Kit+Flt3-), MPP (Lin-Sca1+c-Kit+Flt3low), LMPP (Lin-Sca1+c-Kit+Flt3+) and CLP (Lin-IL7R+Sca1+c-Kit+) were gated and analyzed for cell death by Annexin V staining. (L) Schematic representation of lymphoid development from Hematopoietic Stem Cells (HSC) to Pro B cells with the different steps including Multipotent Primed Progenitors (MPP), Lymphoid Primed Multipotent Progenitors (LMPP), Common Lymphoid Progenitors (CLP) and Pre-pro B cells. ‘+++’ represents the population increased in the GFI1 KD and KO mice and ‘-’ represents the populations decreased in the GFI1 KD and KO mice. All FACS plots are representative of at least three independent experiments. A minimum of three mice were used to determine absolute numbers of different progenitor subsets in BM.

https://doi.org/10.1371/journal.pone.0160344.g003

Gfi1 is essential in LMPP and CLP cells for B cell lineage specification

HSCs, MPPs, LMPPs or CLPs from KO and KD mice were unable to direct B cell differentiation in vitro on OP9 cells (Figs 4A and S4A–S4C). However, sorted pre-pro B cells were able to generate B220+CD19+ cells after co-culture on OP9 cells (Fig 4B) suggesting that loss of Gfi1 affects B cell differentiation potential at the transition between MPPs and LMPPs. In addition, Gfi1 deletion after the pre-pro B cells stage (equivalent to Hardy Fraction A) using MB1-Cre Gfi1flox/flox mice has almost no effect on B-cell differentiation since the frequency of B220+CD19+ cells in the BM and B220+ cells in the spleen were close to WT levels (Figs 4C–4G and S4D), suggesting that Gfi1 is no longer required at this stage. The fact that expression of Gfi1 drops to very low levels at the pre-pro B cell stage is in agreement with this (Fig 4H). Although LMPPs from Gfi1 KD or KO mice were unable to sustain B cell differentiation, they maintained their ability to differentiate into T and myeloid cells (Fig 4I and 4J). Similarly, CLPs from KD or KO mice could not differentiate into B cells, but reduced levels (KD) of Gfi1 were sufficient to allow CLPs to differentiate into T cells (S4E Fig).

thumbnail
Fig 4. Gfi1 is essential in LMPP cells for B cell lineage specification.

(A) LMPPs from WT, KI, KD and KO were sorted and cultured on stromal OP9 cells in the presence of Flt3L (5 ng/mL) and IL-7 (10 ng/mL). After 10 or 12 days of co-culture cells were analysed for B220 and CD19 expression. The cell number for each mouse is indicated in bold. (B) Pre-pro B cells were sorted and cultured on OP9 cells in the presence of Flt3L (5 ng/mL) and IL-7 (10 ng/mL). After 10 or 12 days of co-culture, cells were analyzed for B220 and CD19 expression. (C) PCR analysis showing the deletion of Gfi1 flox allele in BM B220+CD19+ B cells and in splenic B220+ B cells. B220+CD19+ BM cells and B220+ splenic cells were sorted and lysed to extract DNA. (D) and (E) WT, Gfi1 KO (KO) and MB1-cre Gfi1 flox/flox (MB1/Gfi1fl) mice were analyzed for B cells in the BM (D) and spleen (E). (F) Absolute numbers of B220+CD19+ B cells in WT, Gfi1 KO (KO), Gfi1 flox/flox (Gfi1fl), MB1-cre (MB1/WT) and MB1-cre Gfi1 flox/flox (MB1/Gfi1fl) BM. (G) Absolute numbers of B220 B cells in WT, Gfi1 KO (KO), Gfi1 flox/flox (Gfi1fl), MB1-cre (MB1/WT) and MB1-cre Gfi1 flox/flox (MB1/Gfi1fl) spleens. (H) BM cells from Gfi1-GFP mice were gated for HSC, MPP, LMPP, pre-pro B cell (Pre-pro B) and GFP expression (blue line) was measured in each population. (I) Sorted LMPPs from WT, KI, KD and KO mice were cultured on OP9-DL1 stromal cells in the presence of SCF (10 ng/mL), Flt3L (5 ng/mL) and IL-7 (1 ng/mL). After 15 or 21 days of co-culture, cells were analyzed for CD44, CD25, CD4 and CD8 expression. (J) Sorted LMPPs from WT, KI, KD and KO were cultured on OP9 cells in the presence of myeloid cytokine mix as previously described [37]. After 10 days, cells were analysed for Gr1 and Mac1 expression by flow cytometry. FACS plots are representative of at least three independent experiments.

https://doi.org/10.1371/journal.pone.0160344.g004

The B cell program is impaired in MPPs and LMPPs in Gfi1 KO and KD mice

A comparison of mRNA expression profiles by microarray analysis of LSK cells sorted from WT and KO mice showed that the expression of genes implicated in B cell differentiation such as Il7r, Flt3, Ebf1 were reduced in the absence of Gfi1 (S5A Fig). The reduction of Il7r, Flt3, Ebf1 expression in KD and KO LSK cells was confirmed by real time PCR (S5B Fig). In addition, Pu1 and E2a [16, 47] levels did not change with Gfi1 levels in LSK, as observed by microarray analysis and real time PCR (S5A and S5B Fig). Since the LSK subsets are different between KO, KD and control mice, we verified the expression of these genes in the LSK Flt3low/- cells present in all mice, and found the same tendency, i.e. a severe or a low decrease in Il7r, Flt3, Ebf1 expression in KD and KO mice compared to WT and KI animals and no change for E2a and Pu.1 expression (Fig 5A). We also found no difference in PU.1 and E2A expression in Lin- cells between the different mice (Figs 5B and S5C). However, the expression of Id1, an E2A inhibitor [48] but not Id2 or Id3 was significantly higher at mRNA and protein levels than controls in KO and KD mice (Figs 5C and 5D and S5A and S5B, S5D). This was in agreement with the decrease of Il7r, Ebf1 [25] and Flt3, since they are E2A target genes. Moreover, other E2A targets that are essential for B cell commitment such as Ccr9, Rag1, Dntt or E2A target genes involved in lymphoid development like Notch1 [16] were also down-regulated in KD and KO progenitor cells compared to controls (Figs 5E and S5A, S5E). In addition, compared to control mice, KD and KO mice showed a significantly lower frequency of IL-7R+ cells in the LSK population and CCR9+ cells in the MPP subset (Fig 5F and 5G). This indicated that loss (KO) or a severe down-regulation (KD) of Gfi1 leads to a higher expression of Id1 in the progenitor cell subset, and lower expression levels of genes essential for B cell lineage.

thumbnail
Fig 5. The B cell program is impaired in MPPs and LMPPs in Gfi1 KO and KD mice.

(A) Analysis of Il7r, Flt3, Ebf1, Pu.1, E2a and Id2 gene expression by RT-qPCR in sorted LSKFlt3-/low cells from WT, KI, KD and KO mice. All values were normalized to the expression of the Gapdh gene and are presented relative to cDNA from WT mice. (B) E2A protein levels from lineage negative cells from WT, KI, KD and KO BM. β-Actin levels represent the loading control. (C) Analysis of Id1 expression by RT-qPCR in sorted LSKFlt3-/low from WT, KI, KD and KO mice. The values were normalized to the expression of Gapdh and are presented relative to cDNA from WT mice. (D) Id1 protein levels in lineage negative cells from WT, KI, KD and KO BM. β-Actin expression represents the loading control. (E) Analysis of E2A target gene (Ccr9, Dntt and Rag1) expression by RT-qPCR in sorted WT, KI, KD and KO LSKFl3-/low. The values were normalized to the expression of Gapdh and are presented relative to cDNA from WT mice. (F) IL-7R cell surface levels on WT, KI, KD and KO LSK cells. (G) CCR9 cell surface levels on WT, KI, KD and KO MPP cells. For all PCR data, at least three independent experiments were performed, each done in triplicates. FACS plots are representative of at least two (G) or three (F) independent experiments.

https://doi.org/10.1371/journal.pone.0160344.g005

Gfi1 regulates Id1 expression and by extension E2A targets for B cell differentiation

Gfi1 occupies the Id1 promoter in MLL-ENL transformed cells suggesting a direct regulation of Id1 expression by Gfi1 (Fig 6A) [49, 50]. A reporter gene assay using an Id1 promoter-driven luciferase gene showed that Gfi1, but not the non-functional Gfi1P2A mutant repressed Id1 expression in both 293T cells and 3T3 cells (Fig 6B). Id1 mRNA levels were not affected by loss or low doses of Gfi1 in pre-pro B cells where Gfi1 expression is absent (S6A Fig) and Flt3 and Il7r mRNA are readily detected. In U2OS cells, in which Gfi1 can be induced by doxycycline, Id1 mRNA expression was reduced upon induction of Gfi1 as well as its protein level (Fig 6C and 6D). This reduction of Id1 levels correlated with a decreased acetylation of histone H3 at lysine 9 (H3K9ac), a marker of transcriptionally active genes, at the Id1 promoter (Fig 6E), suggesting a transcriptional repression of the Id1 gene by Gfi1, a known recruiter of histone deacetylases [51]. Moreover, the higher expression of Id1 in Gfi1 KD and KO mice correlated with a decrease in H3K9ac at the E2A target genes Flt3 and Il7r, but not at the control Gapdh promoter (Fig 6F–6H). Interestingly, we observed that E2A binds to an Il7r evolutionary conserved putative enhancer region as well as to the IL7r promoter in LSK cells (Fig 6J and 6L), using previously published E2A ChIP-seq data from murine pro-B cells (Fig 6I; [27]). Brg1, Foxo1 and Ikaros were also found to occupy the same sites as E2A at the IL7r locus (Fig 6I) and often target the same regions as E2A at other B-cell specific genes that are affected by Gfi1 (S6B–S6H Fig). Upon the loss of Gfi1, however, binding of E2A was no longer detected at the Il7r enhancer site in sorted LSK cells by ChIP-qPCR (Fig 6J) and also coincided with a reduction in H3K27 acetylation (H3K27ac) enrichment, a marker for transcriptionally active promoters and particularly for active enhancers, at this region (Fig 6K) [52]. Furthermore, E2A binding to the Il7r promoter was also decreased in the absence of Gfi1 (Fig 6L) and a reduced enrichment of H3K27ac at Il7r promoter was also found (Fig 6M), which was not the case at the control promoter regions of Gapdh or Tnrc5 (Fig 6N and 6O). These data suggest that the reduction of Gfi1 expression beyond a certain threshold leads to increased levels of Id1 expression, which in turn inhibits E2A and prevents it from activating B-cell specific genes.

thumbnail
Fig 6. Gfi1 regulates Id1 expression and by extension E2A targets for B cell differentiation.

(A) ChIP-seq data for Gfi1 and control IgG binding over the Id1 locus in MLL-ENL transformed cells. (B) 293T cells or 3T3 cells were transfected with a luciferase reporter fused to the Id1promoter (Addgene, plasmid #16048) and pcDNA3.1 empty vector (EV), pcDNA3.1 with Gfi1 (Gfi1) or pcDNA3.1 with mutant of Gfi1 (P2A) (for 293 T cells only) in varying concentrations. Data were normalized for transfection using β-galactosidase and are presented as average relative luciferase units ± SD and are representative of two independent experiments. (C-D) U2OS and U2OS-Gfi1 (U2OS cell line in which Gfi1 can be induced with doxycycline) were treated with or without doxycycline (1 μg/mL) for 16 hours. Untreated U2OS (Ctrl), U2OS treated with doxycycline (Dox), untreated U2OS-Gfi1 (Ctrl) and U2OS-Gfi1 treated with doxycycline (Dox) cells were used for RNA (C) or protein extraction (D). Expression of Id1 was measured and normalized to the expression of Gapdh and is presented as the fold change relative to cDNA from untreated U2OS cells or untreated U2OS-Gfi1 (C). β-Actin was used as a loading control (D). (E) H3K9ac ChIP-qPCR at the Id1 promoter in U2OS and U2OS-Gfi1 cells treated with or without doxocycline (1 μg/mL) for 16 hours. (F-H) H3K9ac enrichment by ChIP-qPCR at the promoters of Flt3, Il7r and Gapdh from LSK cells sorted from WT, KI, KD and KO mice. (I) E2A, PU.1, Brg1, Foxo1 and Ikaros targeted regulatory regions from RAG1 null pro-B cells by ChIP-seq across the Il7r locus using published data [27]. The localization of primers used for ChIP-qPCR in Figs 6 G-M is represented in the Figure. (J) E2A ChIP enrichment at an Il7r 3’ putative enhancer in Gfi1 WT and KO sorted LSK cells. (K) H3K27ac enrichment by ChIP-qPCR at the Il7r 3’ putative enhancer in Gfi1 WT and KO sorted LSK cells. (L) E2A ChIP enrichment at an Il7r promoter in Gfi1 WT and KO sorted LSK cells. (M) H3K27ac ChIP enrichment at an Il7r promoter in Gfi1 WT and KO sorted LSK cells. (N) and (O) H3K27ac ChIP enrichment at Gapdh and Tnf regions in Gfi1 WT and KO sorted LSK cells. For the ChIP experiments, error bars represent standard deviation of three independent experiments for Figs 6 E-H and two independent experiments for Figs 6 J-N.

https://doi.org/10.1371/journal.pone.0160344.g006

Reduced expression of Id1 in LSK cells rescues B cell differentiation

To further test whether Gfi1 regulates the Id1/E2A axis in B-lineage development, we set out to inhibit Id1 in progenitors that lack Gfi1 to rescue B cell differentiation. We used morpholinos directed against Id1, which were able to enter 3T3 cells at high efficiency and led to a specific reduction of Id1 expression (Fig 7A–7C). We transfected LSK cells from WT or KO mice with the control morpholino and the two specific morpholinos (Id1 mor#1 and Id1 mor#2) (Fig 7D) and followed B cell differentiation on OP9 cells. WT cells transfected with morpholino Id1 #2 differentiated well into B220+ cells compared to the control cells or cells transfected with control morpholino or Id1 #1 (Fig 7E). After 15 days of co-culture, few LSK cells from Gfi1 KO mice transfected with Id1 #2 were able to differentiate into B220+ B cells (Fig 7E). The low numbers of B220+ cells obtained at the end of the differentiation for both WT and KO LSK cells, make the conclusion about Id1 silencing in LSK cells from KO mice difficult. We observed that most of the WT or KO LSKs did not support the transfection with morpholinos and died. To counteract this, we used LSK cells from Gfi1KO p53-/+ mice and from p53-/+ animals as controls. We found that Gfi1KO p53-/+ LSKs were able to generate cells expressing B220 after transfection with Id1 #2 compared to untreated cells or cells treated with the scrambled control morpholino (Fig 7F and 7G). LSKs from p53-/+ animals were used as controls and developed B220+ cells when untreated and also when treated with the scrambled control morpholino (Fig 7F and 7G). Even though these results were obtained with cells missing one allele of p53 to counteract cell death, they indicated that reduction of Id1 expression could overcome the block in early B-cell differentiation induced by loss of Gfi1. The data suggest a model, in which besides its effect on PU.1, Gfi1 by repressing Id1 positively regulates E2A activity and E2A target genes required for B cell commitment (Fig 7H).

thumbnail
Fig 7. Reduced expression of Id1 in LSK cells rescues B cell differentiation.

(A) 3T3 cells were transfected with control scrambled-Fitc or two different morpholinos (Fitc) (Id1 mor #1 and Id1 mor #2) targeting Id1 following the company’s instructions. The efficiency of the transfection was measured by flow cytometry. (B) Id1 expression was measured by Western blot in 3T3 cells transfected with scrambled-Fitc, Id1 morpholino # 1 (Id1 mor #1) or Id1 morpholino # 2 (Id1 mor#2). β-Actin was used as a loading control. (C) Relative protein quantity of Id1 was measured with Image J software. (D) LSK cells from WT or KO mice were sorted and transfected or not with the control morpholino or with morpholinos directed against Id1 and cultured on OP9 cells in the presence of IL-7 and Flt3L. For each condition, the percentages of Fitc+ cells are indicated on the histogramme plots. Black is for the control cells, blue is for the cells transfected with the scramble, orange is for the cells transfected with the morpholino1 against Id1 (Id1 mor #1) and green is for the cells transfected with the morpholino 2 against Id1 (Id1 mor #2). (E) FACS analysis of LSK cells from WT and KO mice, transfected or not with scramble-Fitc, Id1 mor #1 or Id1 mor #2 and cultured on OP9 cells in the presence of IL-7 and Flt3L. B220, and Mac1 expression was measured after gating on the Fitc positive cells for the LSK cells transfected with morpholinos. The cell number for each condition is indicated in bold. (F) LSK cells from WT p53+/- or KO p53+/- mice were sorted and transfected or not with the control morpholino or with the morpholino directed against Id1 and cultured on OP9 cells in the presence of IL-7 and Flt3L. For each condition, the percentages of Fitc+ cells are indicated on the histogramme plots. Black is for the control cells, blue is for the cells transfected with the scramble and green is for the cells transfected with the morpholino 2 against Id1 (Id1 mor #2). (G) FACS analysis of LSK cells from WT, p53+/- and KO, p53+/- mice, transfected or not with scrambled-Fitc, or Id1 mor #2 and cultured on OP9 cells in the presence of IL-7 and Flt3L. B220, and Mac1 expression was measured after gating on the Fitc positive cells for the LSK cells transfected with morpholinos. The cell number for each condition is indicated in bold. (H) Schematic representation of the role of Gfi1 in B cell differentiation. It has been shown that Gfi1 inhibits PU.1 activity preventing its support of myeloid differentiation [23]. Our data suggest that Gfi1 exerts an additional function in B cell lineage commitment by inhibiting Id1 and thus allowing E2A to support the expression of targets genes such as Il7r, Flt3, Rag and others necessary for B cell differentiation. Green line: pathway positively regulated, Red line: pathway negatively regulated, Dashed green line: positively supports B-lineage differentiation, Dashed red line: suppresses B-lineage differentiation.

https://doi.org/10.1371/journal.pone.0160344.g007

Discussion

We present evidence that a threshold level of Gfi1 is required to direct B cell commitment in multipotent hematopoietic progenitors. We demonstrate that Gfi1 mediates this function by enabling the transcription factor E2A to activate a specific set of target genes that are critical for early B cell development. Loss of or reduced Gfi1 expression leads to de-repression of Id1, inactivation of E2A and ultimately the loss of expression of B cell differentiation genes. Our findings provide new insights that define Gfi1 as an essential control element of the E2A-Id1 axis.

LMPPs and CLPs require Gfi1 threshold levels for B-lineage priming

To maintain B cell development in mice, and more specifically to maintain LMPP or production with B lineage potential, a minimal expression level of Gfi1 is required. Whereas a reduction of Gfi1 expression to 50–60% is tolerable for directing B-lineage commitment and B-cell differentiation, a reduction to 20–30%, affects early B-cell differentiation severely. This is supported by our data from experiments using mice expressing different defined levels of Gfi1 and the determination of progenitor cell numbers by flow cytometry and in vitro B-cell differentiation on OP9 cells. Our data also indicate that Gfi1 expression levels are critical at the transition from MPPs to LMPPs, since LMPP but not MPP numbers drop significantly when Gfi1 expression is below 50–60%. Gfi1 is also essential for the transition between the CLP and pre-pro B cell stage because we observed that KD and KO CLP cells were still unable to produce B cells while the KD and KO pre-pro B cells produced B cells under OP9 supported conditions. Although this suggests that expression of Gfi1 is important at different stages of B cell development, we decided to focus on the MPP/LMPP transition. Also, LMPPs, which have both lymphoid and myeloid potential [3, 53], require Gfi1 expression levels to be at least at 50–60% to sustain B-cell differentiation in vivo or in vitro. Our data also show that LMPP do not require Gfi1 at all to direct myeloid differentiation and only require a relatively low Gfi1 expression level at around 20–30% of WT levels to initiate T-cell differentiation. Why LMPPs do not need Gfi1 for T cell differentiation remains unclear. It is possible that LMPPs retain the capacity to activate pathways that can compensate for Gfi1 deficiency in T cell development. The presence of the chemokine CCR9 on the surface of LMPPs could offer an explanation, since CCR9 expressing progenitors from the BM have been shown to go directly to the thymus, bypassing the CLP and ETP stages, and differentiate into T cells [53].

Gfi1 gauges E2A activity in LMPPs by regulating the expression of Id1

Although it is known that LMPPs have lymphoid and myeloid potential [3, 53], the mechanisms that engage these cells along a specific lineage remain incompletely understood. Transcription factors like PU.1 [21], Ikaros [18, 19], E2A [17, 47], EBF1 [25], Pax5 [54] and Foxo1 [45, 55, 56] are important in this process and several studies have linked these factors in regulatory circuits based on mathematical models [57]. The B lymphoid priming of LMPPs involves the activation of a specific set of genes including Rag1, Dntt, Notch1, Ccr9, Il7r and Flt3, all of which are targets or downstream effectors of E2A [16, 47]. Furthermore, B cells do not develop in E2A deficient mice, but T cells can be found in these animals [17, 58]. Interestingly, the regions bound by E2A at loci such as Flt3, Notch1 and Dntt are also frequently occupied at the same position by Brg1, EBF1, Foxo1 and/or Ikaros emphasizing these areas as key regulatory regions. Our findings that Gfi1 deficient LSKs show decrease of E2A target genes and that Gfi1 deficient LMPPs lose B cell potential but retain T cell and myeloid potential similar to E2A deficient cells are in agreement with the notion that Gfi1 regulates E2A activity in hematopoietic progenitors. Our data further suggest that Gfi1 exerts this regulatory role of E2A in LSL Flt3- cells by controlling the expression of Id1 [48], a protein already described to play a role in B cell fate determination [59, 60]. The first indication that Gfi1 can regulate the expression of Id genes came from studies with Gfi1 deficient thymocytes [43]. More recently, it has been shown that Gfi1 can directly repress Id2 and Id3 genes in hematopoietic progenitors, while activating the expression of Id1 [41]. These observations regarding Id proteins stand in contrast to the findings presented in this study, which shows an increase of Id1 expression when Gfi1 expression is reduced or absent in LSK Flt3- cells. However, the same team published a recent paper where they found a higher Id1 expression in the total BM from Gfi1 KO mice, comforting the data found in our study [61]. Nevertheless, we didn’t find the same observation for Id2 or Id3 expression in Gfi1 KO progenitor cells [41, 61], which could be explained by a different regulation of Id proteins during B cell development. Our data from experiments such as Ch-IP and reporter gene assays strongly suggest that Gfi1 occupies the Id1 promoter and represses Id1 expression in LSK Flt3- progenitors. Future experiments have to be done to confirm these data in primary cells especially in B cell progenitors with a ChIP-PCR for Gfi1 but technical limitations such as cell number, antibody quality have to be overcome to perform this experiment. Finally, the fact that we can measure reduced histone H3K9 lysine acetylation at the Id1 promoter upon induction of Gfi1 expression is in agreement with this, given that Gfi1 can associate with HDACs at promoter sites [51]. Further evidence for the validity of this model comes from expression analyses, which allowed us to identify several E2A targets involved in B cell fate determination [16] that are decreased in the LSK compartment of Gfi1 KD and KO mice such as Il7r, Ebf1, Flt3, Ccr9, Rag1, Dntt, and Notch1. Our Id1 knock-down experiments that rescue the ability to drive B cell development in Gfi1 deficient progenitors, which we discovered express high levels of Id1, lends further support to this model. In this experiment, only severe reduction of Id1 protein levels was efficient to restore partially B cell development in Gfi1 deficient progenitors, which offers an explanation to why a previous crossing between Gfi1 KO mice and Id1 heterozygous mice was not sufficient to rescue B cell development [61]. Since Id1 deficient mice are causing a number of other defects [60, 62], experiments with morpholinos were conducted rather than with mice in which Gfi1 and Id1 were ablated at the same time. Although, we obtained only a partial rescue by knocking down Id1 in a p53+/- background, which is likely due to technical limitations of the experiment owing to the fragility of Gfi1 deficient cells, the result are in agreement with our hypothesis. As previously shown, knocking down Pu.1 had a similar effect and partially rescued B-cell development in Gfi1 KO mice [23]. It is therefore conceivable that a knock down of both Pu.1 and Id1 at the same time may completely rescue B cell development in Gfi1 KO mice. It remains to be shown whether a similar increase of Id1 exists at the CLP stage but in accordance with the high expression of Id1 observed in lineage negative cells, we hypothesize that Id1 is also involved in the developmental block at this stage. However other targets could be also take part in this developmental block, but further experiments have to be performed to address this question.

Evidence that E2A acts on the promoter of the Il7r gene via a distal enhancer element

Il7r is the target gene that is most affected by the absence of Gfi1 in progenitors and Gfi1 has previously been implicated in the regulation of Il7r in thymocytes [63]. Our findings reported here suggest that Gfi1 regulates Il7r expression by controlling the activity of E2A via Id1. Two E2A binding sites at the Il7r locus were identified in murine pro B cells [27]. The conservation of the sequences suggested the existence of similar E2A dependent regulatory regions at the promoter of the Il7r gene and at a putative 3’ enhancer region downstream of the Il7r locus in other species. Both elements are also bound by other regulatory transcription factors such as Brg1, Foxo1 and Ikaros, further highlighting their importance as regulatory regions. E2A and some of these factors also occupy other genes driving B cell development such as Rag1/2, Dntt, Notch1, Ccr9, Il7r and Flt3, all of which are targets or downstream effectors of E2A which are down-regulated in Gfi1 deficient cells, supporting the notion that Gfi1 regulates this particular gene set via the Id1/E2A axis. Using ChIP-qPCR, E2A, H3K9ac and H3K27ac were detected at both sites in Gfi1 WT progenitors, which is consistent with transcriptional activation by E2A. Strikingly, the enrichment of both E2A and H3K27ac, a histone mark that can distinguish active enhancers from inactive/poised enhancers [52], is lost or reduced in Gfi1 KO cells at the promoter and at the putative distal enhancer region. This also supports our hypothesis that threshold levels of Gfi1 are required to permit E2A to bind to and transcriptionally activate its target genes. Gfi1 exerts this function by blocking Id1 transcription, enabling hematopoietic progenitors to maintain their potential for the B lineage.

Supporting Information

S1 Fig. The control of B cell development by Gfi1 is dose dependent.

(A) Western blot Gfi1 from thymocytes isolated from Mx-cre Wt, Gfi1 flox/flox mouse (WT) treated with pIpC and Mx-cre Tg, Gfi1 flox/flox mice (MxcreTg) treated with pIpC to delete Gfi1. Lamin B was used as a loading control. (B) and (C) FACS analysis of bone marrow B cells from Mx-cre Wt, Gfi1 flox/flox mouse (WT) treated with pIpC and Mx-cre Tg, Gfi1 flox/flox mice (MxcreTg) treated with pIpC to delete Gfi1. Bone marrow cells were stained for B220 and CD19 (B) or B220 and IgM (C). (D) FACS analysis of B220 and CD19 in the bone marrow of WT and P2A mice. (E) FACS analysis of B220 and IgM in the bone marrow of WT and P2A mice. (F) Cell numbers of B220+ CD19+ cells in the WT and P2A bone marrow. At least three mice were used to determine absolute numbers of B cell subsets in bone marrow.

https://doi.org/10.1371/journal.pone.0160344.s002

(PDF)

S2 Fig. Gfi1 is important at the stage where early B cell progenitors are formed.

(A) Flow cytometry analysis of B220, CD43, HSA, and BP-1 surface expression on bone marrow cells from WT or P2A mice. B220+CD43+ cells were electronically gated and analyzed for HSA and BP-1 expression to identify Fraction A (Fr.A, B220+CD43+HSA-BP1), Fraction B (Fr.B, B220+CD43+HSA+BP1), Fraction C (Fr.C, B220+CD43+HSA+BP1+) and Fraction C’ (Fr.C’, B220+CD43+HSAhighBP1+). (B) CD19 expression on pro B cells from WT and P2A mice. (C) Absolute numbers of Fraction A-C’ from WT and P2A mice. (D) Expression of Lin-, CD19, B220, and CD43 on bone marrow cells from WT and P2A mice. LinCD19 cells were analyzed for the expression of B220 and CD43. (E) Absolute numbers of pre pro B cells in WT and P2A bone marrow. At least three mice were used to determine absolute numbers of B cell subsets in bone marrow.

https://doi.org/10.1371/journal.pone.0160344.s003

(PDF)

S3 Fig. Gfi1 is required for the generation of LMPPs.

(A) LSK cell numbers in WT and P2A bone marrow. (B) CLP numbers in WT and P2A bone marrow. (C) Cell numbers of HSC, MPP and LMPP cells in the WT and P2A bone marrow. (D) BM cells from WT, KI, Het, KD and KO mice were stained with control isotypes to determine the CLP population gate. At least three mice were used to determine absolute numbers of B cell subsets in bone marrow.

https://doi.org/10.1371/journal.pone.0160344.s004

(PDF)

S4 Fig. Gfi1 is essential in LMPP cells for B cell lineage specification.

(A) HSC (Lin-Sca1+c-Kit+Flt3-), (B) MPP (Lin-Sca1+c-Kit+Flt3low) or (C) CLP (Lin-Il7R+Sca1+c-Kit+) were sorted and cultured on OP9 stromal cells in the presence of IL-7 (10 ng/mL) and Flt33L (5 ng/mL) for 10 or 12 days. Then cells were analyzed for B220 and CD19 expression by flow cytometry. The cell number on the FACS plots for each mouse is indicated in bold. (D) Gfi1 flox/flox (Gfi1fl), MB1-cre (MB1/WT) mice were analyzed for B cells in the bone marrow and spleen. (E) Sorted CLPs were cultured on OP9-DL1 cells in the presence of SCF (10 ng/mL), FLT3L (5 ng/mL) and IL-7 (1 ng/mL) for at least 15 days. The cells were analyzed for CD44, CD25, CD4 and CD8 expression. All FACS plots are representative of at least two independent experiments and at least three mice were used to determine absolute numbers of B cell subsets in bone marrow and spleen.

https://doi.org/10.1371/journal.pone.0160344.s005

(PDF)

S5 Fig. The B cell program is impaired in MPPs and LMPPs in Gfi1 KO and KD mice.

(A) Table showing the expression log2 fold change of selected up- and down-regulated genes between sorted WT and KO LSKs. (B) Il7r, Flt3, Ebf1, Pu.1, E2a, Id1, Id2 and Id3 expression in sorted LSK cells from WT, KI, KD and KO mice was measured by real time qPCR. Expression of the genes was normalized to Gapdh and is presented relative to cDNA from WT cells. (C) PU.1 expression in Lin- cells from WT, KI, KD and KO BM was measured by western blot. β-actin was used as a loading control. (D) Id2 and Id3 expression in lineage negative cells from WT, KI, KD and KO mice measured by western blot. β-actin was used as a loading control. (E) Ccr9, Notch1, Rag1 and Dntt expression in sorted LSK cells from WT, KI, KD and KO mice was measured by real time qPCR. Expression of the genes was normalized to Gapdh and is presented relative to cDNA from WT cells. Real time PCR are representative of at least 3 independent experiments and the western blot scans are representative of two different experiments.

https://doi.org/10.1371/journal.pone.0160344.s006

(PDF)

S6 Fig. Gfi1 regulates Id1 expression and by extension E2A targets for B cell differentiation.

(A) Il7r, Flt3 and Id1 expression in sorted pre-pro B cells from WT, KI, KD and KO was measured by real time qPCR. Expression of Il7r, Flt3 and Id1 was normalized to Gapdh and presented as the fold increase relative to cDNA from WT cells. (B-H) E2A, PU.1, Brg1, EBF1, Foxo1 and Ikaros targeted regulatory regions from RAG1 null pro-B cells by ChIP-seq across the Il7r, (C) Flt3, (D) Ccr9, (E) Ebf1, (F) Notch1, (G) Dntt, (H) Rag1 loci.

https://doi.org/10.1371/journal.pone.0160344.s007

(PDF)

S1 Table. Primers for mouse and human ChIP-qPCR.

https://doi.org/10.1371/journal.pone.0160344.s008

(PDF)

Acknowledgments

We thank Mathieu Lapointe for technical assistance; Marie-Claude Lavallée, Jo-Anny Bisson, and Jade Dussureault for excellent animal care; Eric Massicotte and Julie Lord for FACS and cell sorting; and Odile Neyret for assistance with molecular biology. J.F. was supported by a fellowship from the Fonds recherche Québec-Santé. C.V. was supported by a fellowship from the Canadian Institute for Health Research (CIHR). C.K is supported by funds of the Deutsche Krebshilfe and the Jakstädt-foundation. T.M. holds a Canada Research Chair (Tier 1). This work was supported by CIHR grants to T.M. (MOP– 111011) and F.R. / T.M. (MOP-106427).

Author Contributions

Conceptualization: JF TM. Formal analysis: JF AH CV. Funding acquisition: TM FR. Investigation: TM JF AH. Methodology: JF AH CV RC CK. Project administration: TM. Resources: TM FR. Software: JF CV. Supervision: TM. Writing - original draft: JF AH TM. Writing - review & editing: JF TM.

References

  1. 1. Christensen JL, Weissman IL. Flk-2 is a marker in hematopoietic stem cell differentiation: a simple method to isolate long-term stem cells. Proceedings of the National Academy of Sciences of the United States of America. 2001;98(25):14541–6. Epub 2001/11/29. 261562798 [pii]. pmid:11724967; PubMed Central PMCID: PMC64718.
  2. 2. Morrison SJ, Weissman IL. The long-term repopulating subset of hematopoietic stem cells is deterministic and isolatable by phenotype. Immunity. 1994;1(8):661–73. Epub 1994/11/01. [pii]. pmid:7541305.
  3. 3. Adolfsson J, Mansson R, Buza-Vidas N, Hultquist A, Liuba K, Jensen CT, et al. Identification of Flt3+ lympho-myeloid stem cells lacking erythro-megakaryocytic potential a revised road map for adult blood lineage commitment. Cell. 2005;121(2):295–306. Epub 2005/04/27. S0092-8674(05)00158-3 [pii] pmid:15851035.
  4. 4. Mansson R, Hultquist A, Luc S, Yang L, Anderson K, Kharazi S, et al. Molecular evidence for hierarchical transcriptional lineage priming in fetal and adult stem cells and multipotent progenitors. Immunity. 2007;26(4):407–19. Epub 2007/04/17. S1074-7613(07)00213-0 [pii] pmid:17433729.
  5. 5. Laiosa CV, Stadtfeld M, Graf T. Determinants of lymphoid-myeloid lineage diversification. Annu Rev Immunol. 2006;24:705–38. Epub 2006/03/23. pmid:16551264.
  6. 6. Kondo M, Weissman IL, Akashi K. Identification of clonogenic common lymphoid progenitors in mouse bone marrow. Cell. 1997;91(5):661–72. Epub 1997/12/11. S0092-8674(00)80453-5 [pii]. pmid:9393859.
  7. 7. Kondo M, Scherer DC, Miyamoto T, King AG, Akashi K, Sugamura K, et al. Cell-fate conversion of lymphoid-committed progenitors by instructive actions of cytokines. Nature. 2000;407(6802):383–6. pmid:11014194.
  8. 8. Martin CH, Aifantis I, Scimone ML, von Andrian UH, Reizis B, von Boehmer H, et al. Efficient thymic immigration of B220+ lymphoid-restricted bone marrow cells with T precursor potential. Nature immunology. 2003;4(9):866–73. pmid:12925850.
  9. 9. Allman D, Sambandam A, Kim S, Miller JP, Pagan A, Well D, et al. Thymopoiesis independent of common lymphoid progenitors. Nature immunology. 2003;4(2):168–74. Epub 2003/01/07. ni878 [pii]. pmid:12514733.
  10. 10. Nagasawa T. Microenvironmental niches in the bone marrow required for B-cell development. Nature reviews Immunology. 2006;6(2):107–16. pmid:16491135.
  11. 11. Dolence JJ, Gwin K, Frank E, Medina KL. Threshold levels of Flt3-ligand are required for the generation and survival of lymphoid progenitors and B-cell precursors. Eur J Immunol. 2011;41(2):324–34. Epub 2011/01/27. pmid:21268003; PubMed Central PMCID: PMC3050551.
  12. 12. Miller JP, Izon D, DeMuth W, Gerstein R, Bhandoola A, Allman D. The earliest step in B lineage differentiation from common lymphoid progenitors is critically dependent upon interleukin 7. The Journal of experimental medicine. 2002;196(5):705–11. Epub 2002/09/05. pmid:12208884; PubMed Central PMCID: PMC2193997.
  13. 13. Corfe SA, Paige CJ. The many roles of IL-7 in B cell development; mediator of survival, proliferation and differentiation. Semin Immunol. 2012;24(3):198–208. Epub 2012/03/17. S1044-5323(12)00016-4 [pii] pmid:22421572.
  14. 14. Mackarehtschian K, Hardin JD, Moore KA, Boast S, Goff SP, Lemischka IR. Targeted disruption of the flk2/flt3 gene leads to deficiencies in primitive hematopoietic progenitors. Immunity. 1995;3(1):147–61. Epub 1995/07/01. pmid:7621074.
  15. 15. de Pooter RF, Kee BL. E proteins and the regulation of early lymphocyte development. Immunol Rev. 2010;238(1):93–109. Epub 2010/10/26. pmid:20969587; PubMed Central PMCID: PMC2992984.
  16. 16. Dias S, Mansson R, Gurbuxani S, Sigvardsson M, Kee BL. E2A proteins promote development of lymphoid-primed multipotent progenitors. Immunity. 2008;29(2):217–27. Epub 2008/08/05. S1074-7613(08)00314-2 [pii] pmid:18674933; PubMed Central PMCID: PMC2600583.
  17. 17. Bain G, Maandag EC, Izon DJ, Amsen D, Kruisbeek AM, Weintraub BC, et al. E2A proteins are required for proper B cell development and initiation of immunoglobulin gene rearrangements. Cell. 1994;79(5):885–92. pmid:8001125.
  18. 18. Schwickert TA, Tagoh H, Gultekin S, Dakic A, Axelsson E, Minnich M, et al. Stage-specific control of early B cell development by the transcription factor Ikaros. Nature immunology. 2014;15(3):283–93. Epub 2014/02/11. ni.2828 [pii] pmid:24509509.
  19. 19. Wang JH, Nichogiannopoulou A, Wu L, Sun L, Sharpe AH, Bigby M, et al. Selective defects in the development of the fetal and adult lymphoid system in mice with an Ikaros null mutation. Immunity. 1996;5(6):537–49. pmid:8986714.
  20. 20. Dakic A, Metcalf D, Di Rago L, Mifsud S, Wu L, Nutt SL. PU.1 regulates the commitment of adult hematopoietic progenitors and restricts granulopoiesis. The Journal of experimental medicine. 2005;201(9):1487–502. Epub 2005/05/04. jem.20050075 [pii] pmid:15867096; PubMed Central PMCID: PMC2213186.
  21. 21. Scott EW, Simon MC, Anastasi J, Singh H. Requirement of transcription factor PU.1 in the development of multiple hematopoietic lineages. Science. 1994;265(5178):1573–7. pmid:8079170.
  22. 22. DeKoter RP, Singh H. Regulation of B lymphocyte and macrophage development by graded expression of PU.1. Science. 2000;288(5470):1439–41. Epub 2000/05/29. 8531 [pii]. pmid:10827957.
  23. 23. Spooner CJ, Cheng JX, Pujadas E, Laslo P, Singh H. A recurrent network involving the transcription factors PU.1 and Gfi1 orchestrates innate and adaptive immune cell fates. Immunity. 2009;31(4):576–86. Epub 2009/10/13. S1074-7613(09)00415-4 [pii] pmid:19818654; PubMed Central PMCID: PMC4373467.
  24. 24. Zou GM, Chen JJ, Yoder MC, Wu W, Rowley JD. Knockdown of Pu.1 by small interfering RNA in CD34+ embryoid body cells derived from mouse ES cells turns cell fate determination to pro-B cells. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(37):13236–41. Epub 2005/09/06. 0506218102 [pii] pmid:16141339; PubMed Central PMCID: PMC1201612.
  25. 25. Kee BL, Murre C. Induction of early B cell factor (EBF) and multiple B lineage genes by the basic helix-loop-helix transcription factor E12. The Journal of experimental medicine. 1998;188(4):699–713. Epub 1998/08/26. pmid:9705952; PubMed Central PMCID: PMC2213347.
  26. 26. Lin H, Grosschedl R. Failure of B-cell differentiation in mice lacking the transcription factor EBF. Nature. 1995;376(6537):263–7. pmid:7542362.
  27. 27. Lin YC, Jhunjhunwala S, Benner C, Heinz S, Welinder E, Mansson R, et al. A global network of transcription factors, involving E2A, EBF1 and Foxo1, that orchestrates B cell fate. Nature immunology. 2010;11(7):635–43. Epub 2010/06/15. ni.1891 [pii] pmid:20543837; PubMed Central PMCID: PMC2896911.
  28. 28. Singh H, Medina KL, Pongubala JM. Contingent gene regulatory networks and B cell fate specification. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(14):4949–53. Epub 2005/03/25. 0500480102 [pii] pmid:15788530; PubMed Central PMCID: PMC555998.
  29. 29. Batliner J, Mancarelli MM, Jenal M, Reddy VA, Fey MF, Torbett BE, et al. CLEC5A (MDL-1) is a novel PU.1 transcriptional target during myeloid differentiation. Molecular immunology. 2011;48(4):714–9. pmid:21094529; PubMed Central PMCID: PMC3026634.
  30. 30. Fiolka K, Hertzano R, Vassen L, Zeng H, Hermesh O, Avraham KB, et al. Gfi1 and Gfi1b act equivalently in haematopoiesis, but have distinct, non-overlapping functions in inner ear development. EMBO Rep. 2006;7(3):326–33. Epub 2006/01/07. 7400618 [pii] pmid:16397623; PubMed Central PMCID: PMC1456886.
  31. 31. Karsunky H, Zeng H, Schmidt T, Zevnik B, Kluge R, Schmid KW, et al. Inflammatory reactions and severe neutropenia in mice lacking the transcriptional repressor Gfi1. Nat Genet. 2002;30(3):295–300. Epub 2002/01/26. ng831 [pii]. pmid:11810106.
  32. 32. Khandanpour C, Krongold J, Schutte J, Bouwman F, Vassen L, Gaudreau MC, et al. The human GFI136N variant induces epigenetic changes at the Hoxa9 locus and accelerates K-RAS driven myeloproliferative disorder in mice. Blood. 2012;120(19):4006–17. Epub 2012/08/31. blood-2011-02-334722 [pii] pmid:22932805.
  33. 33. Yucel R, Kosan C, Heyd F, Moroy T. Gfi1:green fluorescent protein knock-in mutant reveals differential expression and autoregulation of the growth factor independence 1 (Gfi1) gene during lymphocyte development. The Journal of biological chemistry. 2004;279(39):40906–17. pmid:15252036.
  34. 34. Zarebski A, Velu CS, Baktula AM, Bourdeau T, Horman SR, Basu S, et al. Mutations in growth factor independent-1 associated with human neutropenia block murine granulopoiesis through colony stimulating factor-1. Immunity. 2008;28(3):370–80. Epub 2008/03/11. S1074-7613(08)00068-X [pii] pmid:18328744; PubMed Central PMCID: PMC2835494.
  35. 35. Hones JM, Botezatu L, Helness A, Vadnais C, Vassen L, Robert F, et al. GFI1 as a novel prognostic and therapeutic factor for AML/MDS. Leukemia. 2016. pmid:26847026.
  36. 36. Hobeika E, Thiemann S, Storch B, Jumaa H, Nielsen PJ, Pelanda R, et al. Testing gene function early in the B cell lineage in mb1-cre mice. Proceedings of the National Academy of Sciences of the United States of America. 2006;103(37):13789–94. Epub 2006/08/31. 0605944103 [pii] pmid:16940357; PubMed Central PMCID: PMC1564216.
  37. 37. Kosan C, Saba I, Godmann M, Herold S, Herkert B, Eilers M, et al. Transcription factor miz-1 is required to regulate interleukin-7 receptor signaling at early commitment stages of B cell differentiation. Immunity. 2010;33(6):917–28. Epub 2010/12/21. S1074-7613(10)00458-9 [pii] pmid:21167753.
  38. 38. Valk PJ, Verhaak RG, Beijen MA, Erpelinck CA, Barjesteh van Waalwijk van Doorn-Khosrovani S, Boer JMet al. Prognostically useful gene-expression profiles in acute myeloid leukemia. N Engl J Med. 2004;350(16):1617–28. Epub 2004/04/16. 350/16/1617 [pii]. pmid:15084694.
  39. 39. Gautier L, Cope L, Bolstad BM, Irizarry RA. affy—analysis of Affymetrix GeneChip data at the probe level. Bioinformatics. 2004;20(3):307–15. Epub 2004/02/13. 20/3/307 [pii]. pmid:14960456.
  40. 40. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47. Epub 2015/01/22. gkv007 [pii] pmid:25605792; PubMed Central PMCID: PMC4402510.
  41. 41. Li H, Ji M, Klarmann KD, Keller JR. Repression of Id2 expression by Gfi-1 is required for B-cell and myeloid development. Blood. 2010;116(7):1060–9. Epub 2010/05/11. blood-2009-11-255075 [pii] pmid:20453161; PubMed Central PMCID: PMC2938128.
  42. 42. Hock H, Hamblen MJ, Rooke HM, Traver D, Bronson RT, Cameron S, et al. Intrinsic requirement for zinc finger transcription factor Gfi-1 in neutrophil differentiation. Immunity. 2003;18(1):109–20. Epub 2003/01/18. S1074761302005010 [pii]. pmid:12530980.
  43. 43. Yucel R, Karsunky H, Klein-Hitpass L, Moroy T. The transcriptional repressor Gfi1 affects development of early, uncommitted c-Kit+ T cell progenitors and CD4/CD8 lineage decision in the thymus. The Journal of experimental medicine. 2003;197(7):831–44. Epub 2003/04/12. jem.20021417 [pii]. pmid:12682108; PubMed Central PMCID: PMC2193890.
  44. 44. Hock H, Hamblen MJ, Rooke HM, Schindler JW, Saleque S, Fujiwara Y, et al. Gfi-1 restricts proliferation and preserves functional integrity of haematopoietic stem cells. Nature. 2004;431(7011):1002–7. pmid:15457180.
  45. 45. Welinder E, Mansson R, Mercer EM, Bryder D, Sigvardsson M, Murre C. The transcription factors E2A and HEB act in concert to induce the expression of FOXO1 in the common lymphoid progenitor. Proceedings of the National Academy of Sciences of the United States of America. 2011;108(42):17402–7. Epub 2011/10/06. 1111766108 [pii] pmid:21972416; PubMed Central PMCID: PMC3198373.
  46. 46. Khandanpour C, Kosan C, Gaudreau MC, Duhrsen U, Hebert J, Zeng H, et al. Growth factor independence 1 protects hematopoietic stem cells against apoptosis but also prevents the development of a myeloproliferative-like disease. Stem Cells. 2011;29(2):376–85. Epub 2011/07/07. pmid:21732494.
  47. 47. Kee BL, Quong MW, Murre C. E2A proteins: essential regulators at multiple stages of B-cell development. Immunol Rev. 2000;175:138–49. Epub 2000/08/10. pmid:10933599.
  48. 48. Sun XH, Copeland NG, Jenkins NA, Baltimore D. Id proteins Id1 and Id2 selectively inhibit DNA binding by one class of helix-loop-helix proteins. Molecular and cellular biology. 1991;11(11):5603–11. Epub 1991/11/01. pmid:1922066; PubMed Central PMCID: PMC361931.
  49. 49. Bonadies N, Foster SD, Chan WI, Kvinlaug BT, Spensberger D, Dawson MA, et al. Genome-wide analysis of transcriptional reprogramming in mouse models of acute myeloid leukaemia. PloS one. 2011;6(1):e16330. pmid:21297973; PubMed Central PMCID: PMC3030562.
  50. 50. Khandanpour C, Phelan JD, Vassen L, Schutte J, Chen R, Horman SR, et al. Growth factor independence 1 antagonizes a p53-induced DNA damage response pathway in lymphoblastic leukemia. Cancer Cell. 2013;23(2):200–14. pmid:23410974; PubMed Central PMCID: PMCPMC3597385.
  51. 51. Duan Z, Zarebski A, Montoya-Durango D, Grimes HL, Horwitz M. Gfi1 coordinates epigenetic repression of p21Cip/WAF1 by recruitment of histone lysine methyltransferase G9a and histone deacetylase 1. Molecular and cellular biology. 2005;25(23):10338–51. pmid:16287849; PubMed Central PMCID: PMC1291230.
  52. 52. Creyghton MP, Cheng AW, Welstead GG, Kooistra T, Carey BW, Steine EJ, et al. Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proceedings of the National Academy of Sciences of the United States of America. 2010;107(50):21931–6. pmid:21106759; PubMed Central PMCID: PMC3003124.
  53. 53. Lai AY, Kondo M. Identification of a bone marrow precursor of the earliest thymocytes in adult mouse. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(15):6311–6. Epub 2007/04/04. 0609608104 [pii] pmid:17404232; PubMed Central PMCID: PMC1851047.
  54. 54. Urbanek P, Wang ZQ, Fetka I, Wagner EF, Busslinger M. Complete block of early B cell differentiation and altered patterning of the posterior midbrain in mice lacking Pax5/BSAP. Cell. 1994;79(5):901–12. pmid:8001127.
  55. 55. Dengler HS, Baracho GV, Omori SA, Bruckner S, Arden KC, Castrillon DH, et al. Distinct functions for the transcription factor Foxo1 at various stages of B cell differentiation. Nature immunology. 2008;9(12):1388–98. Epub 2008/11/04. ni.1667 [pii] pmid:18978794; PubMed Central PMCID: PMC2679692.
  56. 56. Amin RH, Schlissel MS. Foxo1 directly regulates the transcription of recombination-activating genes during B cell development. Nature immunology. 2008;9(6):613–22. pmid:18469817; PubMed Central PMCID: PMC2612116.
  57. 57. Laslo P, Pongubala JM, Lancki DW, Singh H. Gene regulatory networks directing myeloid and lymphoid cell fates within the immune system. Semin Immunol. 2008;20(4):228–35. Epub 2008/09/06. S1044-5323(08)00066-3 [pii] pmid:18771937.
  58. 58. Zhuang Y, Soriano P, Weintraub H. The helix-loop-helix gene E2A is required for B cell formation. Cell. 1994;79(5):875–84. pmid:8001124.
  59. 59. Cochrane SW, Zhao Y, Perry SS, Urbaniak T, Sun XH. Id1 has a physiological role in regulating early B lymphopoiesis. Cell Mol Immunol. 2011;8(1):41–9. Epub 2011/01/05. cmi201058 [pii] pmid:21200383; PubMed Central PMCID: PMC3058379.
  60. 60. Cochrane SW, Zhao Y, Welner RS, Sun XH. Balance between Id and E proteins regulates myeloid-versus-lymphoid lineage decisions. Blood. 2009;113(5):1016–26. Epub 2008/10/18. blood-2008-06-164996 [pii] pmid:18927439; PubMed Central PMCID: PMC2635070.
  61. 61. Kim W, Klarmann KD, Keller JR. Gfi-1 regulates the erythroid transcription factor network through Id2 repression in murine hematopoietic progenitor cells. Blood. 2014;124(10):1586–96. Epub 2014/07/24. blood-2014-02-556522 [pii] pmid:25051963; PubMed Central PMCID: PMC4155270.
  62. 62. Perry SS, Zhao Y, Nie L, Cochrane SW, Huang Z, Sun XH. Id1, but not Id3, directs long-term repopulating hematopoietic stem-cell maintenance. Blood. 2007;110(7):2351–60. pmid:17622570; PubMed Central PMCID: PMCPMC1988946.
  63. 63. Ligons DL, Tuncer C, Linowes BA, Akcay IM, Kurtulus S, Deniz E, et al. CD8 lineage-specific regulation of interleukin-7 receptor expression by the transcriptional repressor Gfi1. The Journal of biological chemistry. 2012;287(41):34386–99. pmid:22865857; PubMed Central PMCID: PMC3464544.