SEARCH

SEARCH BY CITATION

Keywords:

  • B cell development;
  • fetal liver;
  • microenvironment

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

The microenvironments, in which B lymphocytes develop in fetal liver, are largely still unknown. Among the nonhematopoietic cells, we have identified and FACS-separated two subpopulations, CD45TER119VCAM-1+ cells that are either CD105highLYVE-1high or CD105lowALCAMhigh. Immunohistochemical analyses find three of four c-Kit+IL-7Rα+B220lowCD19SLC B progenitors in contact with vascular endothelial-type LYVE-1high cells on embryonic day 13.5. One day later c-Kit+IL-7Rα+ cells develop to CD19− and +, SLC-expressing, DHJH-rearranged pre/pro and pro/preB-I cells. Less than 10% are still in contact with LYVE-1high cells, but half of them are now in contact with mesenchymally derived ALCAMhigh liver cells. All of these ALCAMhigh cells, but not the LYVE-1high cells produce IL-7 and CXCL12, while both produce CXCL10. Progenitors and pro/preB-I cells are chemoattracted in vitro toward CXCL10 and 12, suggesting that lymphoid progenitors with Ig gene loci in germline configuration enter the developing fetal liver at E13.5 from vascular endothelium, attracted by CXCL10, and then migrate within a day to an ALCAMhigh liver cell microenvironment, differentiating to DHJH-rearranging, surrogate light chain-expressing pre/proB and pro/preB-I cells, attracted by CXCL10 and 12. Between E15.5 and E16.5 preB-I cells expand 10-fold in continued contact with ALCAMhigh cells and begin VH- to DHJH-rearrangements in further differentiated c-KitIL-7Rα preBII cells. STEM Cells 2013;31:2800–2812


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Hematopoietic cell development appears ordered by different types of stromal cells that support different cell lineages through specialized cell contacts and differential cytokine and chemokine production and reception [1]. In bone marrow, entry into lymphoid and B-lymphocyte development is supported by the cytokine interleukin-7 (IL-7) and by the chemokine CXCL12, the latter recognized by its receptor, CXCR4 [[2, 3]] and produced by reticular cells and by preadipocytes, that appear to be of mesenchymal origin [[4, 5]]. IL-7 induces the development of common lymphoid progenitors (CLP) from multipotent progenitors (MPP) that express the corresponding receptor (IL-7R).

In bone marrow, B-cell development begins in c-Kit+IL-7Rα+CD19CD93+ FLT3+ progenitor pre/proB and pro/preB-I cells with the expression of surrogate light chain (SLC), composed of VpreB and λ5 [[6, 7]], and DH- to JH-rearrangements at the IgH-chain locus. The transcription factor Pax5 then induces a series of B-lymphocyte lineage-related genes, such as CD19, so that c-Kit+IL-7Rα+CD19+DH-JH-rearranged, SLC+ preBI cells develop [[8, 9]]. In preBI cells Pax5 also induces VH- to DHJH-rearrangements, and the productively rearranged IgH chain loci express IgμH chains, that are deposited on the surface of c-Kit-IL-7Rα large preBII cell [[6, 7]]. In these preB cell receptor (preBcR)+ preBII cells, VL- to JL-rearrangements are induced and repertoires of sIgM-expressing immature B cells are formed. Microenvironments interacting with this early development of B lymphocytes in bone marrow have been described [[10, 11]].

Much less is known of the hematopoiesis, including B lymphopoiesis that occurs in fetal liver [12]. On embryonic day (E) 10.5 of the mouse, transplantable hematopoietic stem cells (HSCs), of extraembryonic, yolk sac, and of intraembryonic aorta/gonad/mesonephros (AGM) origin, as well as B-lymphocyte progenitors are found to circulate in the intravascular spaces of embryonic blood [[13-16]]. Thereafter, from E12.5, fetal liver becomes a major B lymphopoietic and myelopoietic organ where progenitors and precursors develop progressively with time until mature B lymphocytes appear between E18.5 and E19.5 (birth) [17]. On E15, IgμH chain-expressing cells have been detected [18]. In general, B-cell development in fetal liver is expected to follow much of the molecular steps and cellular stages detected in bone marrow. As in bone marrow B-lymphocyte development in fetal liver is supported by IL-7 [[19, 20]] and by CXCL12 [2]. It generates a wave of B1a cells that prefer to populate the peritoneal cavities [21]. Their B cell progenitors and precursors appear to be distinct from the majority in bone marrow, for example, by B220 expression, or responses to thymic stromal lymphopoietin (TSLP) [22].

Earlier experiments done in vitro have suggested that support by mesenchymally derived, preadipocytic stromal cells is necessary for B-cell development of fetal liver cells [[17, 23]]. Among nonhematopoietic cells, vascular cell adhesion molecule (VCAM)-1-positive cells are good candidates for supporting hematopoiesis, since very late antigen-4, the ligand of VCAM-1, is expressed on early hematopoietic cells and plays important roles in hematopoiesis [[24, 25]]. Furthermore, CD105 is expressed on stromal cells involved in hematopoiesis [26]. Here, we characterize different nonhematopoietic cell population by the expression of characteristic surface markers identified in fluorescence-activated cell sorting (FACS) analyses. We find that VCAM-1+CD105high cells express high levels of the adhesion molecule lymphatic vessel endothelial receptor (LYVE-1). Another marker for this phenotypic characterization is the activated leukocyte cell adhesion molecule (ALCAM) [27]. ALCAM has been found to be expressed in the developing liver at E14.5 [28] as well as in bone marrow, where ALCAM+ nonhematopoietic cells were found to support hematopoiesis [29]. We find here that VCAM-1+CD105low fetal liver cells express high levels of ALCAM. We characterize by reverse transcriptase polymerase chain (RT-PCR) the functions of these cells to produce chemokines and cytokines and, in particular, IL-7 by the expression in cells of a mouse strain expressing a reporter-IL-7-GFP-gene [4]. We use LYVE-1- and ALCAM-specific mAbs to stain sections of fetal liver between E13.5 and E16.5 to detect nonhematopoietic microenvironments and localize early hematopoietic progenitors in contact with these microenvironments by an IL-7Rα-specific mAb, and precursor B cells by a huCD25-specific mAb that detects this reporter molecule on cells expressing the λ5-gene of SLC [30]. We further identify the phenotypes of different stages of B-cell development that attach to these nonhematopoietic cells by FACS and their status of Ig-gene rearrangements by PCR. Our experiments suggest that subsequent stages of progenitors and precursor B cells in fetal liver have preferential cooperating nonhematopoietic cell partners in their microenvironment.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Mouse Strains and Cell Preparations

Mice were bred and maintained under specific pathogen free conditions at the Bundesinstitut für Risikobewertung, and experiments were performed according to institutional guidelines and German Federal laws on animal protection. Fetal liver was obtained from timed matings of C57BL/6 mice (Charles River, Wilmington, MA, http://www.criver.com/en-US/Pages/home.aspx). The embryonic stage was designated relative to E0.5, the day of vaginal plug formation. A single cell suspension of fetal liver was prepared by treatment with 1 mg/mL collagenase (Worthington Biochemical Corp., Vassar Ave., Lakewood, NJ, http://www.worthington-biochem.com/default.html) and with 0.1% DNase I (Worthington Biochemical Corp.) at 37°C for 30 minutes, thereafter passed through a 21-gauge needle with a 1 mL syringe. The transgenic mouse strain expressing the human CD25 gene under the control of the lambda5 promoter [30] and the IL-7-GFP-reporter mouse strain [4] have both been backcrossed for at least six generations to C57BL/6. Animal protocols were approved by Landesamt für Gesundheit und Soziales, Germany.

RT-PCR

Total RNA was purified using TRIZOL (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) and used as the template to synthesize cDNA using Superscript III reverse transcriptase (Invitrogen) primed with oligo-dT (Invitrogen) from 1 μg of total RNA. Gene expression was analyzed by PCR using primers described in supporting information Materials and Methods.

In the detection of mRNA of IL-7 and of TSLP, nested PCR was performed. For IL-7, the primers of IL-7 S2 and IL-7 AS1 were used followed by using primer of IL-7 S3 and IL-7 AS1. For TSLP, the primers of TSLP S2 and TSLP AS2 were used followed by using primer of TSLP S2 and TSLP AS. Quantitative PCR is described in supporting information Materials and Methods.

Flow Cytometry

Flow cytometry analyses on total single cell preparations were performed using the BD LSR II analyzer (BD Biosciences, San Diego, http://www.bdbiosciences.com). Antibodies for staining are described in supporting information Materials and Methods. To detect cytoplasmic proteins, cells were fixed with Fix Buffer I (BD Biosciences) for 10 minutes at 37°C and permeabilized in Phosflow Perm/Wash Buffer I (BD Biosciences) for 10 minutes at room temperature followed by 10 minutes blocking with Fcγ-R blocking Ab (2.4G2) and were then stained with the specific antibodies. Antibodies for staining are described in supporting information Materials and Methods.

Immunohistochemical Staining

For staining of sections, fetal livers were fixed in 4% paraformaldehyde and equilibrated in 30% sucrose/phosphate buffered saline. Cryostat sections were stained and mounted with Prolong Gold Antifade Reagent. All confocal microscopy was carried out on a LSM 710 (Carl Zeiss, Jena, Germany, http://www.zeiss.com). Antibodies for staining are described in supporting information Materials and Methods.

All sections were analyzed with a Zeiss LSM 710 confocal microscope. Images were obtained with 63× objective lens and digitally stitched using the LSM 710 software. Representative confocal images were processed and cropped using LSM Image Browser (Zeiss).

Migration Assay

c-Kit+IL-7Rα+ cells enriched by magnetic sorting were transferred to the upper layer of 5 μm pore polycarbonate membrane (Corning, Acton, MA, http://www.corning.com/lifesciences) while the lower chamber contained 100 ng/mL (or lower concentrations, where indicated) of CXCL10 or CXCL12. After 2 hours at 37°C, cells that had migrated to the lower chamber were counted and analyzed by flow cytometry.

Coculture of Nonhematopoietic Cells with Hematopoietic Cells

Details are described in supporting information Materials and Methods.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Nonhematopoietic Cell Populations Separable by Differential Expression of CD45, TER-119, VCAM-1, and CD105

In order to isolate single cells of different populations of nonhematopoietic cells, fetal liver at E14.5 was enzymatically digested. FACS distinguished CD45+TER-119+ hematopoietic cells from VCAM-1+ nonhematopoietic cells [11]. Two populations of these cells, VCAM-1+ and VCAM-1 cells were identified in the CD45 and TER-119 negative fraction. The VCAM-1+ population, which expresses intercellular adhesion molecule (ICAM)-1 that is reported to be expressed in stromal cells [[31, 32]], was further separated into CD105 high-expressing or negative cells (Fig. 1A).

image

Figure 1. Characterization of nonhematopoietic cells in fetal liver. (A): Fluorescence-activated cell sorting analysis of nonhematopoietic cell populations from the fetal liver here shown as a representative example from E14.5. We first excluded hematopoietic CD45+ and/or TER-119+ erythrocytes. The arrows show the gating strategy for nonhematopoietic cells. VCAM-1+ but CD45TER-119 cells in the middle window were further separated into CD105high and CD105low cells in the right window. The data are representative of three independent experiments. (B): Total cell numbers, each dot representing determination of a separate fetal liver (n = 3–5/day). (C): Ratios of VCAM-1+CD105high and of VCAM-1+CD105low in fetal liver cells are shown (n = 3/day). (B, C): Three independent experiments have been done. Error bars show SDs. (D): Analysis of expressions of surface molecules. The open curves show CD45TER-119ICAM-1+VCAM-1+CD105low, the filled curves CD45TER-119ICAM-1+VCAM-1+CD105high cells. Data are representative of two independent experiments. (E): Analysis of ALCAM and LYVE-1 expression in the sorted fractions. The second row shows the fractions which were sorted. VCAM-1+CD45TER-119 cells in the left panel were further separated into fractions 1 and 2 in the right panel. Each number in the third row is correlated to the upper row. Data are representative of two independent experiments. (F): Photographs show a representative location of ALCAM+ cells (blue) and LYVE-1+ cells (green), detected by immunohistochemical staining at E14.5. The yellow and red box in the second picture from the left are magnified in the third, respectively, fourth picture from the left. Yellow arrows point to ALCAMhigh cells while the yellow arrowhead shows a LYVE-1high cell. White bars indicate 10 μm. The data are representative of three independent experiments. Abbreviations: ALCAM, activated leukocyte cell adhesion molecule; ICAM, intercellular adhesion molecule; LYVE, lymphatic vessel endothelial receptor; VCAM, vascular cell adhesion molecule.

Download figure to PowerPoint

These cell populations were already detectable at E12.5. Their relative presence between E13.5 to E15.5 did not change much, while the total number of cells increased continuously from E13.5 to E15.5 (Fig. 1B, 1C). FACS analyses showed that endothelial cell-related membrane proteins, VEGFR3, TIE2, CD31, and Sca-1 are expressed in the CD105high cells (Fig. 1D). On the other side, the CD105low population showed a higher surface expression of the previously reported liver-relevant genes DLK1, NGFR, CD133 [[33-35]], and of BP-1 [36]. Most importantly for our studies we found that LYVE-1 was highly expressed on the VCAM-1+CD105high cell population, while ALCAM was highly expressed on CD105low cells (Fig. 1E). These results suggest that CD105high cells have characters of endothelial cells, while CD105low cells include hepatoblasts or their progenitors.

The Distribution of LYVE-1High Cell and ALCAMHigh Cells in the Fetal Liver

To study the localization of these two types of nonhematopoietic cells in sections of fetal liver between E13.5 and 16.5, immunohistochemical stainings were performed with sections of fetal liver from E13.5-16.5 with LYVE-1- and ALCAM-specific antibodies. It can be seen that ALCAM was strongly detected all over the inside of the fetal liver as contiguous lines of connected cells, especially also on the outer rim of the fetal liver (shown in Fig. 1F for E14.5). By contrast, all LYVE-1 expressing cells were detected only inside fetal liver.

Thus, pictures in Figure 1F indicate that staining of sections detected only the highest LYVE-1- and ALCAM-expressing cells. A lower expression of ALCAM, detected in FACS analyses on LYVE-1 high cells, and a lower expression of LYVE-1, detected on ALCAMhigh cells (Fig. 1E) was apparently not detectable in the immunohistochemical analyses as LYVE-1/ALCAM-double positive cells.

In conclusion, two types of nonhematopoietic cells, that is, LYVE-1+CD105high VEGFR3+TIE2+CD31+Sca-1+ endothelial cells and ALCAM-1highCD105lowDLK1+NGFR+CD133+ liver cells can be separated by FACS and distinguished by their localizations in sections of fetal liver of E13.5 to E16.5.

Expression of IL-7 and CXCL12 Is Detected in ALCAMHigh Cells, but not in LYVE-1+ Cells

Next, we tested in the LYVE-1+ and the ALCAMhigh cells the expression of cytokines and chemokines involved in B lymphopoiesis and myelopoiesis by RT-PCR (Fig. 2A). LYVE-1+ cells expressed only Kit-ligand, but no other cytokine or chemokine tested. ALCAMhigh cells expressed Kit-ligand and, interestingly, IL-7, CXCL12 as well as Csf-1. Expression of Csf-1 was also detected in VCAM-1, CD45TER119 and CD45+TER119+ hematopoietic fetal liver cells. TSLP could not be detected in any of fractions, in fact, in RNA of total fetal liver cells on E14.5.

image

Figure 2. Analyses of chemokine and cytokine expressions in fetal liver cell populations. (A): Detection of gene expression by reverse transcriptase polymerase chain reaction (RT-PCR). (B): Fluorescence-activated cell sorting profiles of GFP expression of fetal liver cell fractions shown in Figure 1. Open lines show the cells from IL-7-GFP reporter mice, filled lines those from C57BL/6 mice. Cell populations are numbered as in Figure 1A. (A, B): The data are representative of two independent experiments. (C): The photographs show a representative staining of GFP (green) and ALCAM (red) in sections of E14.5 fetal liver from IL-7-GFP-reporter mice in the upper rows and GFP (green) and LYVE-1 (red) in the lower rows. Sections from WT E14.5 fetal liver were stained with same antibodies as a control. Right panels show the area of middle panels indicated by yellow rectangles. Yellow arrowheads indicate GFP positive cells. Yellow arrows point to GFP-negative LYVE-1-positive cells. White bars indicate 10 μm. The data are representative of three independent experiments. (D): The expression of CXCL10 was measured by quantitative RT-PCR of E13.5 fetal liver cell fractions shown in Figure 1. Values of expression are relative to Gapdh expression = 1. Abbreviations: ALCAM, activated leukocyte cell adhesion molecule; DAPI, 4′,6-diamide-2-phenylindole; GFP, green fluorescent protein; HPRT, hypoxanthine-guanine phosphoribosyl transferase; IL-7, interleukin 7; M-CSF, macrophage colony-stimulating factor; WT, wild type.

Download figure to PowerPoint

IL-7 expression was also monitored in fetal liver of a IL-7-GFP reporter mouse strain [4]. Consistent with the RT-PCR data, green fluorescent protein (GFP) expression was observed in ALCAMhigh cells, but not in LYVE-1+ cells (Fig. 2B). Expression of the reporter-IL-7-GFP seen in FACS analysis appeared to be homogeneous over all ALCAMhigh cells, indicating that all cells express IL-7 (Fig. 2B).

Surprisingly GFP expression in the ALCAMhigh cell population was found to be positive for all cells, seen in the homogeneous distribution of fluorescence over all the cells, even though the expression itself appeared weak (Fig. 2B). It suggests that some of the ALCAMhigh cells express not only the chemokine CXCL12 but also more than one cytokine, that is, IL-7 and Csf-1.

Expression of the IL-7-GFP gene reporter was too low to be detectable in sections. Therefore, we used anti-GFP antibody to enhance the staining of the fetal liver sections and detected GFP+ cells that coexpressed ALCAM, but not LYVE-1 (Fig. 2C). The ALCAM+ cells were found to be located inside of the fetal livers (Fig. 1F). As shown in Figure 1F, ALCAM-positive cells also lined the outer ridge of fetal liver but no GFP signal was detectable in these cells. These data suggest that the ALCAM-positive cells lining the outer rim of fetal liver do not express IL-7 at levels detectable by staining of sections, and that IL-7 expression is confined to the ALCAMhigh cells inside. This also suggests that the ALCAM+ GFP-negative cells lining the outer rim might not be converted to single cells by the enzymatic digestion procedure that we used. In contrast to CXCL12, CXCL10 was found expressed both in LYVE-1high and in ALCAMhigh cells (Fig. 2D). This indicates that the two nonhematopoietic cell fractions of fetal liver could have distinguishing functions in the attractions and migrations of hematopoietic progenitors.

IL-7Rα+ Cell Precursors Reside Next to LYVE-1High Cells in E13.5 Fetal Liver

IL-7 controls lymphoid commitment of MPP to CLP, to early T-lymphocyte as well as B-lymphocyte development through CD19 pre/proB cells, CD19+ pro/preB-I cells, and preBII cells. To detect these progenitors, we stained sections of fetal liver with a monoclonal antibody specific for the IL-7Rα-chain.

IL-7Rα+ cells appeared scattered as single cells on E13.5 (Fig. 3A). Interestingly, in eight independent experiments, a total of 50 out of 66 of IL-7Rα+ cells were found in close proximity to LYVE-1+ cells (Tables 1, 2). In contrast, in the same eight independent experiments, only one of a total of 51 IL-7Rα+ cells was found near ALCAMhigh cells. These results suggest that LYVE-1high endothelial cells might interact with IL-7Rα+ pro-lymphoid and lymphoid progenitor cells.

image

Figure 3. Next neighbor analyses of IL-7Rα+ cells in immunohistochemical sections of E13.5 fetal liver, and of huCD25-SLC+ cells in sections from E14.5 to E16.5. (A): The photograph shows a representative localization of an IL-7Rα+ cell (red) at E13.5, attached to LYVE-1high cells (green), not to ALCAMhigh cells (blue). The yellow-boxed area in the middle is enlarged in the right panel. A yellow arrowhead indicates an IL-7Rα+ cell stained with rat anti-IL-7Rα antibody followed by biotin-conjugated anti-rat IgG and by streptavidin-Alexa Fluor 555. (B) huCD25+ cells (red) stained with rat anti-huCD25 followed by Cy3 conjugated anti-rat IgG antibody, ALCAM+ cells (blue) with goat anti-ALCAM antibody followed by Alexa Fluor 647-conjugated anti-goat IgG antibody and LYVE-1+ cells with Alexa Fluor 488-conjugated anti-LYVE-1 antibody. The photographs show huCD25+ cells (red) in close proximity to ALCAMhigh cells (blue), but not to LYVE-1high cells (green) on E14.5, 15.5 and 16.5. The yellow-boxed areas in the middle are shown enlarged in the right panels. White bars indicate 10 μm. For fluorescence-activated cell sorting analyses of the c-Kit+IL-7Rα+ cells of E13.5 and of the huCD25+ cells on E14.5–16.5 see Figure 4A. The data are representative of more than three independent experiments. Abbreviations: ALCAM, activated leukocyte cell adhesion molecule; DAPI, 4′,6-diamide-2-phenylindole; IL-7, interleukin 7; LYVE, lymphatic vessel endothelial receptor.

Download figure to PowerPoint

Table 1. Numbers (percentages) of IL-7Rα+ progenitors and of SLC+ pro and preB cells in close proximity to LYVE-1+ and to ALCAM+ cells in immunohistochemical sections of fetal liver between E13.5 and 16.5
 StainingContact with LYVE-1+Contact with ALCAM+
  1. The numbers shown are the sums of experiments with sections of fetal livers from eight E13.5, five E14.5, four E15.5 and four E16.5 fetuses. n.d. = not done.

  2. a

    Fluorescence-activated cell sorting analyses (Fig. 4A) show that at E13.5 around 95%, at E14.5 85%, at E15.5 80%, and at E16.5 25% of these cells are IL-7Rα+ huCD25(SLC)-.

  3. Abbreviations: ALCAM, activated leukocyte cell adhesion molecule; IL-7, interleukin 7; LYVE, lymphatic vessel endothelial receptor; SLC, surrogate light chain.

E13.5IL-7Rα+a50/66 (75.8%)1/51 (2.0%)
 SLC+n.d.n.d.
E14.5IL-7Rα+a4/45 (8.9%)n.d.
 SLC+2/53 (3.8%)32/53 (60.4%)
E15.5IL-7Rα+a0/72 (0%)n.d.
 SLC+3/96 (3.1%)48/96 (50.0%)
E16.5IL-7Rα+a6/91 (6.6%)n.d.
 SLC+5/121 (4.1%)90/121 (74.4%)
Table 2. Numbers (percentages) of IL-7Rα+ or of huCD25+(SLC)+ cells in close proximity to LYVE-1high and ALCAMhigh nonhematopoietic cells in independent experiments
  Contact with ALCAMhigh cellsContact with LYVE-1high cells
  1. The numbers of cells counted in these experiments are shown. The data of each experiment were obtained from a fetal liver of a different embryo.

  2. Abbreviations: ALCAM, activated leukocyte cell adhesion molecule; IL-7, interleukin 7; LYVE, lymphatic vessel endothelial receptor.

E13.5 IL-7Rα+Experiment 10/22/2
 Experiment 20/53/5
 Experiment 30/311/17
 Experiment 40/45/6
 Experiment 50/139/14
 Experiment 61/107/8
 Experiment 70/66/6
 Experiment 80/87/8
E14.5 huCD25+Experiment 12/51/5
 Experiment 23/101/10
 Experiment 310/100/10
 Experiment 47/130/13
 Experiment 510/150/15
E15.5 huCD25+Experiment 115/322/32
 Experiment 213/231/23
 Experiment 311/220/22
 Experiment 49/190/19
E16.5 huCD25+Experiment 116/250/25
 Experiment 233/374/37
 Experiment 322/300/30
 Experiment 419/291/29

FACS analyses were done with the fetal liver cells of E13.5 to characterize the stages of prolymphoid and lymphoid development that were detected in sections. They were first analyzed for c-Kit and IL-7Rα expression, and the c-Kit+IL-7Rα+ cells were then analyzed for CD19, CD93, and huCD25 (SLC) surface expression (Fig. 4A). At E13.5, approximately 0.34% of single fetal liver cells were c-Kit+IL-7Rα+, of which two subpopulations, CD19CD93+huCD25 CLP-like cells and CD19CD93SLC cells could be distinguished (Fig. 4A). This second, c-Kit+IL-7Rα+CD93 population has not been seen in B-cell development of bone marrow.

image

Figure 4. Development of hematopoietic progenitors and precursor B cells in fetal liver between E13.5 and E16.5. (A): Analysis of c-Kit- and IL-7Rα-expressing cells, further analyzed for CD19 and CD93, for CD19 and huCD25, for CD93 and huCD25 expression, and for huCD25 and SLC. Representative fluorescence-activated cell sorting (FACS) profiles are shown for fetal liver on E13.5 (top) and E16.5 (bottom). Less than 0.1% of all CD19+ cells could be detected, by either surface or intracytoplasmic staining, as IgμH-expressing cells (not shown). (B): Numbers of c-Kit+IL-7Rα+ cells, further analyzed for CD19 and CD93, for CD19 and huCD25, for CD93 and huCD25 expression by FACS analyses (strategy detailed in Fig. 4A) between E13.5 and E16.5 (n = 3–5/day). (A, B): The data are representative of more than three independent experiments. (C): Frequencies of Ig-locus-rearrangements in c-Kit+IL-7Rα+ cells. Analyses of IgH chain and L chain gene rearrangements were done by polymerase chain reaction (PCR) amplification with DNA of c-Kit+IL-7Rα+ cells of E13.5 to E16.5 fetal liver, diluted in threefold steps to pg DNA per assay was carried out in eight replicates as described [[37, 38]]. The first round was done over 28 cycles and contained all 5′ and 3′ primers as first round primers. For the second PCR round, 1 μL of the first PCR amplification was reamplified with one specific 5′ primer and a specific 3′ primer. The second round was done over 35 cycles (20 seconds at 95°C, 1 minute at 65°C, 2 minutes at 72°C). The sequences of all primers are given in previous publications [[37, 38]]. For the determination of the number of cells given in B it was assumed that one mouse cell contains 6.0 pg DNA. The points in the figure indicate the lowest number of cells (equivalent to multiples of 6.0 pg DNA) at which the last PCR product was detected in the eight replicas of the threefold dilutions of DNA. Note that the approximate frequencies can only be as precise as threefold dilutions allow them to be. VH 7183- and VH 558-DHJH-rearrangements were also determined in DNA of total fetal liver hematopoietic cells at E15.5 and E16.5 (open squares and open reversed triangles). Abbreviation: IL-7, interleukin 7.

Download figure to PowerPoint

On E13.5 of fetal liver, c-Kit+IL-7Rα+ population could include lymphoid tissue inducer (LTi). LTi expressing α4β7+ cell population consists of approximately 26% in IL-7Rα+ cells on E12.5 and about 18% on E15.5 [[39, 40]]. More than 75% of IL-7Rα+ cells reside next to LYVE-1+ cells (Tables 1, 2). That implies that both LTi cells as well as CLP-like cells could be next to LYVE-1+ cells.

We conclude, that at E13.5, around 65% of the hematopoietic c-Kit+IL-7Rα+cells are CLP-like progenitors, making it likely that these cells, in sections detected as IL-7Rα+ cells, are found attached to LYVE-1+ endothelial cells. It cannot be excluded that the other 35% c-Kit+IL-7Rα+CD93 cells are also attached to LYVE-1+ cells.

When induced for “in vitro” differentiation to T- or B-lymphocyte development, the pool of c-Kit+IL-7Rα+ cells developed within 6 days of culture to 30.1% of DN1-4, 0.5% of DPCD4+CD8+ and 4.8%/0.1% of CD4/CD8 single positive T-lineage cells, as well as 67.7% of CD19+ B-lineage cells (supporting information Fig. S1). We have not done the more stringent in vitro development from single CLP of E13.5 cells [41]. Hence, we tentatively call this fraction of cells to include CLP-like cells.

PCR analyses of the status of Ig gene rearrangements performed on DNA of E13.5 c-Kit+IL-7Rα+ cells indicate that less than 1 of 50 cells have any DH- to JH-rearrangements, and less than 1 in 1,800 VH7183- or VH 558- to DHJH-rearrangements on the IgH-chain locus (Fig. 4C). VL- to JL-rearrangements on either the κ- or the λ-L-chain loci cannot be detected in as many as 1,800 cells. These PCR analyses support the FACS analyses of the E13.5 c-Kit+IL-7Rα+ cells to show that the majority are CLP-like cells with Ig gene loci in their non-rearranged forms and, hence, that three of four of them are attached to LYVE-1+ cells.

This changes dramatically at E14.5 and also, thereafter, on E15.5 and E16.5. Less than 10% of the IL-7Rα+ cells are still found attached to LYVE-1+ cells (Tables 1, 2). We conclude that the LYVE-1+ endothelial cell environment is transient, although direct contact site for early lymphoid progenitors at E13.5 on their way to B-lymphoid cell differentiation.

A Majority of huCD25-SLC+ B-Cell Progenitors at E14.5 Are Found Next to ALCAMHigh Cells

One day later, at E14.5, expression of SLC becomes detectable inside and on the surface of a small number of the c-Kit+IL-7Rα+ lymphoid progenitors and precursors (Fig. 4A). Hence, four c-Kit+IL-7Rα+ subpopulations could now be distinguished: CD19CD93+huCD25 CLP-like cells, CD19CD93+huCD25+ pre/proB, CD19+CD93+huCD25+ pro/preB-I cells, as well as the CD19CD93huCD25 cells (Fig. 4A).

Unfortunately, neither the CD19-specific nor the CD93-specific monoclonal antibody used in our FACS analyses could detect progenitors, pre/proB and pro/preB-I cells in sections. Therefore, in order to localize at least pre/proB and pro/preB-I cells expressing SLC we resorted to transgenic mice expressing the huCD25 gene under the control of the promoter for the SLC-component λ5 [30]. In sections at E14.5, huCD25+ cells were frequently detected as a single cell in close proximity to ALCAMhigh cells (Fig. 3B). In fact, a total of 32 of 53 huCD25+ cells in five independent experiments were found to reside next to ALCAMhigh cells. Only 2 out of 53 cells were found in close proximity to LYVE-1high cells. These data suggest that at least half of the c-Kit+IL-7Rα+ progenitors that differentiate to SLC+ pre/proB and pro/preB-I cells before VH- to DHJH-rearrangements are now found in a different nonhematopoietic environment, that is, in proximity of ALCAMhigh cells. This is all the more interesting, since the FACS analyses (Fig. 4A, 4B) indicate that the relative proportions of the c-Kit+IL-7Rα+SLC+ in the total c-Kit+IL-7Rα+ cells change (i.e., from 2 in 10 on E14.5, and 5 in 10 on E15.5, to 8–9 in 10 on E16.5) and suggests that only the SLC+ cells attach.

Since immunohistochemical stainings only detect the SLC+, but not any SLC precursor cells, it remains to be determined with additional markers whether the SLC CLP-like progenitors and CD19CD93 cells that are still abundant (at least three of five c-Kit+IL-7Rα+ cells at E14.5) do attach to ALCAMhigh cells or not.

PCR analyses of the Ig gene-rearrangement status in c-Kit+IL-7Rα+ cells detected a dramatic, over 30-fold increase in DHJH-rearranged IgH-chain loci between E13.4 and E14.5 (Fig. 4C). These semiquantitative PCR analyses indicate that half of the c-Kit+IL-7Rα+cell now had at least one DHJH-rearranged IgH allele. These analyses also showed that the two VH-families, VH 7183 and VH 558, remain non-rearranged between E13.5 and E15.5. (Fig. 4B).

Half of the Expanding huCD25-SLC+ B-Cell Progenitors at E15.5 and 16.5 Remain Attached to ALCAMhigh Cells

Between E14.5 and E16.5 c-Kit+IL-7Rα+ cells expand by proliferation at least 10-fold (Fig. 4A, 4B). FACS analyses showed that SLC+ pre/proB and pro/preB-I cells increase selectively in numbers over SLC cells (Fig. 4A, 4B). Interestingly, the CD19CD93huCD25 cell population increased between E14.5 and E15.5, but strongly decreased between E15.5 and E16.5 (Fig. 4A, 4B).

The lack of VHDHJH-rearrangements between E14.5 and E15.5 suggests that pre/proB and pro/preB-I, but not yet large preBII cells, increase (Fig. 4C). By contrast, between E15.5 and E16.5 VH- to DHJH-rearrangements became detectable, not in c-Kit+IL-7Rα+, but in total fetal liver cells (i.e., in preBcR+c-Kit preBII cells), and they involved both 3′ (e.g., VH-7183) as well as 5′ (e.g., VH-J558) VH-segments of the IgH-chain locus.

In sections of E15.5 and E16.5 fetal liver stained for huCD25(SLC)-expressing cells, again, more than half of huCD25+ cells were found near to ALCAMhigh cells, and only few huCD25+ cells remained close to LYVE-1high cells (Fig. 3B; Tables 1, 2). Furthermore, some of the huCD25+ cells were seen in clusters after E15.5. These findings, that even at E16.5 at least half of all pre/proB and pro/preB-I cells remain in contact with ALCAMhigh cells suggest that these nonhematopoietic cells continue to support SLC-expressing pro/preB-I cells in their proliferation and differentiation from the pre/proB to pro/preB-I stages. Whether ALCAM+ cells also continue to support preBcR+ large preBII cells that are in the process or have completed the downregulation of SLC expression [[6, 7]] remains to be investigated with antibodies that can detect these later stages of pro/preB-I cell development by other selective markers in sections of fetal liver.

Lymphocyte Progenitors Migrate Toward CXCL12 and CXCL10 Differentially Expressed in Nonhematopoietic Cells In Vitro

Our findings that: (a) the receptor for CXCL12, CXCR4 was found to be expressed on the surface of c-Kit+IL-7Rα+ cells from E13.5 as well as on CD19+ pro/preB-I cells at 15.5 and E16.5 (Fig. 5A), (b) that the majority of the CD19CD93+SLC CLP-like cells at E13.5 were localized near LYVE-1+ endothelial cells, and the SLC+ pre/proB and pro/preB-I cells at E14.5 near ALCAMhigh liver cells, and that (c) CXCL12 appeared selectively expressed in ALCAMhigh, but not in LYVE-1+ cells prompted us to hypothesize that CXCL12 could attract c-Kit+IL-7Rα+ cells from LYVE-1high cells to migrate to ALCAMhigh cells, as well as, once arrived at the new site, to keep CD19+ precursor B cells in the ALCAMhigh environment.

image

Figure 5. Induction of the migration of c-Kit+IL-7Rα+ cells and CD19+ preBI cells by CXCL10 and CXCL12 in vitro. CXCR3- and CXCR4-expression detected by fluorescence-activated cell sorting (FACS) (A) and migratory response in a transwell chemotaxis assay to CXCL10 or CXCL12 of c-Kit+IL-7Rα+ cells from E13.5 (B) (n = 3/group) and CD19+ preB-I cells from E15.5 (C) (n = 2/group). (D): CXCR3-expression on c-Kt+IL-7R+Lin cells on E13.5 detected FACS. Migratory response in a transwell chemotaxis assay to CXCL10 of c-Kit+IL-7Rα+Lin cells from E13.5 (E) (n = 3/group). (F): c-Kit+IL-7Rα+Lin hematopoietic cells were cultured with or without sorted nonhematopoietic cells. After 2 days, B-cell development was measured as numbers of c-Kit+CD19+ cells. Two independent experiments are shown. (G): B220 expression on fetal liver (left panel) and on spleen cells (right panel). (A–E, G): Data are representative of two independent experiments. Error bars represent SD. p values were determined by t test. Abbreviations: ALCAM, activated leukocyte cell adhesion molecule; IL-7, interleukin 7; LYVE, lymphatic vessel endothelial receptor.

Download figure to PowerPoint

To test this hypothesis, we resorted to “in vitro” migration assays, in which c-Kit+IL-7Rα+ cells from fetal liver at E13.5 and with CD19+ pro/preB-I cells from E15.5 were exposed to a gradient of CXCL12. CXCL12 induced the migration of the E13.5 c-Kit+IL-7Rα+ cells in vitro (Fig. 5B). In addition, CD19+ pro/preB-I cells from E15.5 also migrated in response to CXCL12 (Fig. 5C). The migrated cells could be induced to develop in vitro to DN1-4, DPCD4+CD8+, and CD4 or CD8 single positive T-lineage cells, as well as CD19+ B-lineage cells (supporting information Fig. S1). These results show that c-Kit+IL-7Rα+ CLP-like cells at E13.5, and thereafter at least through E15.5, express CXCR4 and, hence, are chemoattracted by CXCL12 to ALCAMhigh cells until at least E16.5, while they proliferate and differentiate from pro/preB and pro/preB-I cells to preBII cells. It suggests but, does not prove, that CLP-like cells of E13.5 migrate to the chemokine-expressing ALCAMhigh cells, where they continue to be attracted.

Since we found CXCL10 expressed in LYVE-1+ cells (Fig. 2D), we reasoned that this chemokine might also function to attract E13.5 c-Kit+IL-7Rα+ cells. Indeed, FACS detected CXCR3 on the E13.5 c-Kit+IL-7Rα+ cells (Fig. 5D). Furthermore, migration assays show that E13.5 c-Kit+IL-7Rα+ cells are attracted to CXCL10 (Fig. 5E). These results suggest CXCL10 produced by LYVE-1+ cells attracts c-Kit+IL-7Rα+ progenitors at E13.5 and continues to attract c-Kit+IL-7Rα+ cells to ALCAM+ cells at E14.5 and thereafter.

ALCAM+ Cells Support the Differentiation of c-Kit+CD19+ Cells In Vitro

To investigate whether ALCAM+ cells support B-cell development, cocultivation was performed. Sorted c-Kit+IL-7Rα+Lin cells from E13.5 fetal liver were cultivated with or without sorted ALCAM+ or LYVE-1+ cells without exogenous addition of cytokines. In 2 days, CD19+ B cells were detected in threefold higher numbers in coculture with ALCAM+ cells than with or without LYVE-1+ cells (Fig. 5F). These results suggest that ALCAM+ cells can support B-cell development in vitro.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

In bone marrow, cells of the endosteum lining the inner bone surfaces, as well as vascular endothelial cells, basal lamina and perivascular cells, organized from the internal to the external side of the vessels [[42-45]] of blood sinusoids have been identified as major sites in bone marrow that provide cell contacts, chemokines and cytokines, and that secure interactions by which hematopoietic cells and nonhematopoietic cells differentiate [[46-48]]. In contrast, sinusoids in fetal liver contain endothelial (sinusoidal) cells without the basal membrane. These interactive sites of hematopoiesis and B-lymphopoiesis are largely still unknown in fetal liver, and are expected to differ to some degree from bone marrow, not the least because bone and its endosteum is missing. Furthermore, as the nonhematopoietic cell compartments in fetal liver keep growing with the hematopoietic cell compartments from E13.5 onward, the composition of the LYVE-1+ cell compartment and its assembly with other cells (such as pericytes and other perivascular cells) might change with the progression of development in fetal liver.

During embryonic development of mice, pluripotent HSCs (pHSCs) first develop independently extraembryonically in the yolk sac where most of these pHSCs do not express CD93 [49] and intraembryonically in the AGM region where all of them already express CD93 [50]. These progenitors migrate to fetal liver, as soon as blood circulation is established in the embryo at E8.5 [[13-16]]. Therefore, pHSCs, MPPs, or even CLPs might be able to form stable sites of hematopoiesis near embryonic vascular endothelium, pericytes, and other perivascular cells of blood vessel—hence, also at sites where embryonic blood circulates through fetal liver [24]. If so, then it might be expected that these (CD93 as well as CD93+) hematopoietic progenitors could be attracted to enter the fetal liver from embryonic blood by transmigration through LYVE-1+ vascular endothelium and perivascular cells, a process that we may have caught in our immunohistochemical sections at E13.5, showing LYVE-1+ cells in close contact with IL-7Rα+ cells. We identify here CXCL10 as one candidate of such a chemokine that could function to attract hematopoietic progenitors at E13.5 to transmigrate from embryonic blood into the developing fetal liver (Figs. 2D, 5E). Our model of the early B-cell development in contact with the microenvironments of fetal liver is shown in supporting information Figure S2. Inside fetal liver, at least the CD93+ cells are then expected to differentiate to B-lineage cells [51], and our observations point to ALCAM+ cells, which provide close contacts and CXCL12 and IL-7 for this B-cell differentiation, and to where these progenitors could migrate. Our immunohistochemical localization at E13.5 of 75% of the CLP-like c-Kit+IL-7Rα+ cells near LYVE-1+ nonhematopoietic cells might, in fact, have caught early hematopoietic progenitors during their transmigration through vascular endothelium from embryonic blood into fetal liver. Our finding that three of four c-Kit+IL-7Rα+ cells attach, and that around 40% of them are CD93 makes it likely that both CD93+ and CD93 cells attach to LYVE-1+ cells, hence, enter the inside of fetal liver via the same route. The c-Kit+IL-7Rα+CD93+ cells attached to LYVE-1+ cells at E13.5 are likely to derive from AGM-pHSC, while the CD93 cells might originate from yolk sac [50]. CD93+ cells should include multilineage lymphocyte progenitor cells while the CD93 cells may be typical yolk sac progenitors, with limited differentiation potential for primitive hematopoiesis, in fetal liver of our experiments already declining at E15.5. This interesting possible new aspect of hematopoiesis in fetal liver needs to be investigated in greater detail.

B-cell development in fetal liver generates mainly B-1a B cells from B220lo-neg progenitors (Fig. 5G) [[12, 18, 21, 52, 53]]. It has been argued that this B-cell development could occur in layered waves: one of it involving B220lo-neg cells responsive to the cytokine TSLP [54], another stimulated by IL-7. Unfortunately, we could not detect TSLP-expressing cells by FACS nor did the LYVE-1+ or the ALCAMhigh nonhematopoietic cell populations, nor total fetal liver express TSLP at levels detectable by RT-PCR analyses. Hence, our studies do not describe the possible contact sites that a TSLP-dependent B-cell development may take in fetal liver.

Within 1 day, between E13.5 and E14.5, CLP-like cells disappear from LYVE-1+ nonhematopoietic cells, and further differentiated huCD25 (SLC)+ B-cell precursors appear, at least half of them in close proximity to the IL-7-producing ALCAMhigh cells. For the next 3 days, between E14.5 and E16.5, B-cell development appears to follow ways that have been recognized in bone marrow to be IL-7-dependent [[6, 7, 17]]. Since the B-lineage progenitors all have not yet rearranged IgL-chain loci, it appears that contacts to ALCAMhigh cells support pre/proB and pro/preB-I cell proliferation and differentiation up to the preBcR-expressing stage of large preB-II cells.

We found that ALCAM+ nonhematopoietic cells, from E14.5 through E16.5, found inside fetal liver but not on the rim, produce both IL-7 and CXCL12 and continue to provide close contacts for at least half of all the developing progenitors, pre/proB cells and pro/preB-I cells, up to preBcR-expressing preBII cells. Such close contacts are also required when pro/preB-I cells proliferate on mesenchymally derived, preadipocytic stromal cells in the presence of IL-7 in vitro [[17, 23, 55]], mimicking the interactions of ALCAMhigh nonhematopoietic cells with CD19+SLC+ preB-I cells. While exogenous addition of recombinant IL-7 is needed in vitro to induce proliferation in cultures, comparably high concentrations of IL-7 might be generated in the interphase of ALCAMhigh cells and preB cells in the local environment of, but not everywhere else in fetal liver. IL-7 is not membrane-bound on stromal cells but secreted from the cells.

Removal of IL-7 from preB-I cells “in vitro” induces V(D)J-rearrangements in the Ig loci, leading to the development of sIgM+ B cells [23]. By comparison “in vivo,” the removal of preB-I cells from contacts with the IL-7-producing ALCAM+ environment could induce the progression of precursor B cells to the stage of sIgM+ immature B cells. In agreement with such an “in vivo” scenario after E16.5 in fetal liver, Johnson et al. [56] have found that, in bone marrow, some IL-7-expressing cells are separate from CXCL12-expressing cells, suggesting a mode of B-cell development that allows the downregulation of signaling by IL-7 by CXCL12-mediated attraction of precursor B cells away from an IL-7-producing- and into an IL-7-negative environment. The apparently discordant findings of stromal cells producing both IL-7 and CXCL12 (as in our results of the ALCAM+ cells on E14.5 and E15.5), or producing only IL-7 or only CXCL12, might be reconciled if stromal cells change their capacities to produce either both or only one cytokine/chemokine with development.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

We have separated nonhematopoietic cells of fetal liver into at least two subpopulations that are found in contact with cells of two subsequent stages of early B-cell development. During this development the earlier, c-Kit+IL-7Rα+CD19SLC B progenitors reside first next to LYVE-1+ cells, are attracted by CXCL10, then continue to migrate, and finally attach as more differentiated SLC-expressing pre/proB and pro/preB-I cells to ALCAMhigh cells that express IL-7, CXCL10, and CXCL12. These findings suggest that two early, subsequent B-lineage progenitors could control their development by two different nonhematopoietic cell environments in fetal liver.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

We thank J. Winkler, N. Dittberner, and T. Keiser for technical help; T. Nakayama and Shin-ichi Hayashi for encouragement; Andreas Radbruch and Simon Fillatreau for critical reading of our manuscript. M.T. was a Research Fellow of the Alexander von Humboldt Foundation and is currently a foreign research fellow of the Max Planck Society. This work was partially supported by a Kosellek grant (ME 2764/1-1) of the Deutsche Forschungsgemeinschaft to F.M.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
stem1421-sup-0001-suppfig1.tif19765KSupporting Information Figure 1.
stem1421-sup-0002-suppfig2.tif83102KSupporting Information Figure 2.
stem1421-sup-0003-suppInfo.doc76KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.