JLC, DEW, and AJW conceived and designed the experiments. JLC, DEW, and AJW performed the experiments. JLC, DEW, and AJW analyzed the data. JLC wrote the paper. ILW assisted in the analysis of the data and reviewed the manuscript.
¤
¤
JLC is currently employed by and has stock options in Cellerant Therapeutics, a company seeking to transplant human HSCs. As a former advisory board member of Amgen, ILW owns significant Amgen stock. He also cofounded and consulted for Systemix; cofounded Cellerant Therapeutics, a spin-off from Systemix Novartis to transplant human HSCs; and is a cofounder and a director of Stem Cells, Inc., which is involved in the isolation and study of human central nervous system stem cells, liver-repopulating cells, and pancreatic islet stem/progenitor cells.
The major site of hematopoiesis transitions from the fetal liver to the spleen and bone marrow late in fetal development. To date, experiments have not been performed to evaluate functionally the migration and seeding of hematopoietic stem cells (HSCs) during this period in ontogeny. It has been proposed that developmentally timed waves of HSCs enter the bloodstream only during distinct windows to seed the newly forming hematopoietic organs. Using competitive reconstitution assays to measure HSC activity, we determined the localization of HSCs in the mid-to-late gestation fetus. We found that multilineage reconstituting HSCs are present at low numbers in the blood at all timepoints measured. Seeding of fetal bone marrow and spleen occurred over several days, possibly while stem cell niches formed. In addition, using dual-chamber migration assays, we determined that like bone marrow HSCs, fetal liver HSCs migrate in response to stromal cell-derived factor-1α (SDF-1α); however, unlike bone marrow HSCs, the migratory response of fetal liver HSCs to SDF-1α is greatly increased in the presence of Steel factor (SLF), suggesting an important role for SLF in HSC homing to and seeding of the fetal hematopoietic tissues. Together, these data demonstrate that seeding of fetal organs by fetal liver HSCs does not require large fluxes of HSCs entering the fetal bloodstream, and that HSCs constitutively circulate at low levels during the gestational period from 12 to 17 days postconception. Newly forming hematopoietic tissues are seeded gradually by HSCs, suggesting initial seeding is occurring as hematopoietic niches in the spleen and bone marrow form and become capable of supporting HSC self-renewal. We demonstrate that fetal and adult HSCs exhibit specific differences in chemotactic behavior. While both migrate in response to SDF-1α, fetal HSCs also respond significantly to the cytokine SLF. In addition, the combination of SDF-1α and SLF results in substantially enhanced migration of fetal HSCs, leading to migration of nearly all fetal HSCs in this assay. This finding indicates the importance of the combined effects of SLF and SDF-1α in the migration of fetal HSCs, and is, to our knowledge, the first demonstration of a synergistic effect of two chemoattractive agents on HSCs.
New results on the migratory behavior of blood cell precursors in the early embryo might be relevant to bone marrow transplants and other clinical therapies.
During fetal development, the primary anatomical concentration of hematopoietic stem cells (HSCs) changes location several times. The migration of blood-borne progenitors is essential for the establishment of hematopoiesis in subsequent hematopoietic tissues (
While the mechanisms that influence HSC homing and colonization are not completely understood, several experimental models suggest possible regulatory factors. The homing to and colonization of fetal hematopoietic organs by circulating HSCs likely require homing receptor/addressin interactions in the vascular lumen, followed by chemokine/chemokine receptor interactions, integrin/receptor binding, and growth/survival factors. Homing of lymphocytes and leukocytes has been well documented to involve first homing receptor/vascular addressin interactions, resulting in cell tethering and rolling on blood vessel endothelium. The rolling cells respond to a chemoattractant, produced by endothelial or stromal cells within the tissue, by firm adherence to the vessel wall mediated by integrin/receptor interactions. The adhered cells subsequently traverse the vessel wall, migrating toward the increasing gradient of chemoattractant (
Correct localization of HSCs throughout ontogeny may also involve other specific interactions with the hematopoietic microenvironment (
In order to test the hypothesis that fetal HSC migration is a timed developmental event, we collected blood from embryos ranging in age from 12.5 to 17.5 dpc to use in competitive reconstitution assays to measure long-term reconstituting hematopoietic stem cell (LT-HSC) activity. Our results indicate that mouse fetal HSCs are found constitutively rather than episodically in fetal circulation and are present at low numbers throughout mid-to-late fetal development. We also measured the seeding of the fetal spleen and fetal bone marrow during this period. The seeding of these organs is a gradual process occurring over several days and does not appear to involve a large influx of HSC. Finally, we found that FL HSCs migrate in response to the chemokine SDF-1α and that this response is substantially enhanced in the presence of SLF. The enhanced chemotactic response of HSCs to the combination of SLF and SDF-1α is a property of FL HSCs, but not adult BM HSCs.
To evaluate the presence of rare HSC activity in the fetal circulation, blood was collected from fetuses ranging in age from 12.5 to 17.5 dpc. Blood from a single age group was pooled and assayed by competitive reconstitution. A quantity of fetal cells, measured by fetus equivalents (FEs), was injected into an adult, lethally irradiated congenic recipient, along with a radioprotective dose of host-type bone marrow. A FE was defined as the amount of blood collected from a single fetus of each age group. The amount of fetal blood transplanted ranged from 4 FE to 0.1 FE. The recipient mice were periodically bled and assayed for donor cells of the B, T, and myeloid lineages.
Cell suspensions of fetal blood (A), spleen (B), and bone marrow (C) were used to competitively reconstitute lethally irradiated recipients. The percentage of donor-derived peripheral blood leukocytes is presented for each dose assayed at 20 wk or more following reconstitution. The bar represents the mean percentage of donor-derived peripheral blood leukocytes in all recipients transplanted with each dose of fetal tissue, blood, spleen, or bone marrow. Fetal tissue from each stage embryo (12.5–17.5 dpc) was transferred in two to three experiments at multiple doses. Positive engraftment was determined by comparison to staining of control mice, which in most cases was less than 0.1%.
To illustrate an increase in circulating HSCs at 14.5 dpc, the percentage of recipients with donor multilineage reconstitution from fetal blood are plotted for each timepoint (12.5–17.5 dpc), for 1.0 and 0.5 FE.
Cells can be identified in fetal circulation, spleen, and bone marrow with the FL HSC phenotype: Lineage– c-Kit
Cell suspensions of fetal blood were used to competitively reconstitute lethally irradiated recipients. The fraction of mice exhibiting LT-MLR (B plus T lymphoid and myeloid) for greater than 20 wk, for each dose of fetal blood, is presented
To measure the kinetics of seeding of the fetal spleen and bone marrow by HSCs, these tissues were similarly assayed by competitive reconstitution for LT-HSC activity. Spleens were collected from 14.5 to 17.5 dpc and fetal bone marrow from 15.5 to 17.5 dpc. For these experiments, a FE was defined as the number of cells collected from a single fetal spleen or bone marrow collected from two femurs plus two tibia.
Seeding of the fetal spleen (A) and bone marrow (B) by progenitors unable to provide sustained myelopoiesis precedes colonization of these tissues by HSCs. Reconstituted mice were analyzed for donor contribution in the peripheral blood of B, T, and myeloid lineages at 4 wk post-transplant. Contour plots show gating of donor (Ly-5.2
Cell suspensions of fetal spleen or bone marrow were used to competitively reconstitute lethally irradiated recipients. The fraction of mice exhibiting LT-MLR for greater than 20 wk is presented for each dose of fetal spleen or marrow
To begin to assess whether SDF-1α or SLF may play a direct role in fetal HSC migration, we assayed the ability of FL HSCs to migrate in response to SDF-1α and/or SLF in dual-chamber migration assays. Lineage-depleted fetal liver or adult bone marrow cells were placed in the upper well of a 5-μm transwell chamber, and SDF-1α, SLF, or SDF-1α plus SLF was added to the lower chamber. To evaluate the chemotactic versus chemokinetic effects of SDF-1α and SLF, equal concentrations of factors were added to both the top and bottom chambers. Following a 2 h incubation at 37°C, the cells that had migrated to the lower chamber were collected, stained for HSC markers, and analyzed by FACS to determine the actual number of migrated HSCs. Cell numbers between wells were normalized by the addition of predetermined numbers of fluorescent beads to each well before cell collection (
HSCs were assayed for their ability to chemotax in a transwell assay in response to the chemokine SDF-1α. Migrating cells were labeled with stem cell markers and analyzed by FACS to determine the actual frequency of migrated HSCs. Like adult BM HSCs, FL HSCs migrate in response to SDF-1α (A), although at reduced levels. The optimal concentration of SDF-1α for both fetal liver and adult BM HSCs was determined to be 10 nM. The migratory effect of SLF was also assayed on FL HSCs and adult BM HSCs. FL HSCs migrate equally well to SLF as SDF-1α, while adult BM HSCs showed a lesser response to SLF. SLF and SDF-1α acted synergistically in their chemoattractive effects on FL HSCs (B). To determine whether migration was due to chemokinetic effects of SDF-1α, SLF, or both, equal concentrations of factors were added to both the top and bottom wells (T&B). Data are presented as the percentage of input HSCs that migrate to the bottom chamber for a representative migration assay, each point was performed in triplicate. These data are representative of three independent experiments. The single asterisk shows a significant increase in percent migration over basal migration (
As reported previously, HSCs in the fetal liver double in number daily from 12.5 to14.5 dpc, then decrease in number at 15.5 dpc (
As measured by reconstitution, a slight peak of circulating embryonic HSCs appears at 14.5 dpc. The slight decrease of circulating HSCs at 15.5 dpc could also be due to HSCs leaving the circulation to seed the spleen and later the bone marrow. The fetal spleen and bone marrow are initially seeded by progenitors unable to provide detectable or sustained myelopoiesis, indicating additional requirements not found in these immature hematopoietic tissues are needed for LT-HSC seeding, maintenance, or both. Our data indicate that fetal liver-derived HSCs (c-Kit+ Thy-1.1lo Lineage– Sca-1+ Mac-1lo) begin to seed the fetal spleen and bone marrow on 14.5–15 dpc and 17.5 dpc, respectively. We propose that fetal HSCs are continuously entering circulation and functionally engraft specialized stem cell niches as they develop (
HSCs are found constitutively at low numbers in fetal blood following the onset of circulation. Seeding of developing hematopoietic tissues by long-term HSCs is gradual and is not due to a large influx of cells. The large decline in HSC numbers seen in the fetal liver following 14 dpc is most likely the result of differentiation signaled by the developing hepatic environment rather than a timed migration to the fetal spleen and bone marrow.
Our results stand in contrast with those previously reported by
Studies indicate that HSC trafficking to and retention in the bone marrow relies on the chemokine SDF-1α and its receptor CXCR4 (
We also found that the migratory response of fetal HSCs to SLF was equal to the response to SDF-1α. The migratory response of fetal HSCs to SDF-1α in combination with SLF was synergistic. This finding indicates the importance of the synergistic effects of SLF and SDF-1α in the migration of fetal HSCs. In contrast, SLF alone or in combination with SDF-1α did not evoke a greatly enhanced migratory response from adult BM HSCs. At best, the combination of SLF and SDF-1α had additive affects on the chemoattractive response of adult BM HSCs. A synergistic effect of two chemoactive agents has not, to our knowledge, been directly demonstrated before for HSCs. The substantial migratory response of 14.5 dpc FL HSCs to the combination of SLF and SDF-1α could explain the increased severity of the Steel mutant phenotype for bone marrow versus fetal liver hematopoiesis (
Our lab has recently described a physiological process in which low numbers of BM HSCs rapidly but constitutively traverse the bloodstream of normal mice to seed unoccupied bone marrow niches (
The C57BL/Ka-Thy-1.1/Ly-5.2 (Thy-1.1, Ly-5.2) donor and C57BL/Ka-Thy-1.2 (Thy-1.2, Ly-5.1) recipient mouse strains were bred and maintained at the Stanford University Laboratory Animal Facility, Stanford, California, United States. All mice were routinely maintained on acidified water (pH 2.5). Irradiated recipient mice were more than 8 wk old at the time of irradiation. All protocols were approved by the Administrative Panel on Laboratory Animal Care at Stanford University School of Medicine.
A20.1 (anti-Ly-5.1, CD45.2, FITC-conjugated; BD Biosciences [Pharmingen], Palo Alto, California, United States) and AL1-4A2 (anti-Ly-5.2, CD45.1, Texas Red conjugate) were used to analyze donor and host cells following reconstitution. Blood analysis included 6B2 (anti-B220), KT31.1 (anti-CD3), GK1.5 (anti-CD4), 53-6.7 (anti-CD8), 8C5 (anti-Gr-1), and M1/70 (anti-Mac-1).
The monoclonal antibodies used in immunofluorescence staining for HSC analysis included 2B8 (anti-c-Kit, APC conjugate), 19XE5 (anti-Thy-1.1, FITC conjugate), E13 (anti-Sca-1, Ly6A/E, Texas Red conjugate), and M1/70 (anti-Mac-1, PE conjugate). Lineage marker antibodies included 6B2 (anti-B220), KT31.1 (anti-CD3), GK1.5 (anti-CD4), 53-7.3 (anti-CD5), 53-6.7 (anti-CD8), Ter119 (anti-erythrocyte-specific antigen), 8C5 (anti-Gr-1), and M1/70 (anti-Mac-1). The antibodies were purified and conjugated within our lab. Each antibody was titrated and used at predetermined optimal concentrations (highest signal with lowest background following staining of control spleen or bone marrow cells).
Timed pregnancies of C57BL/Ka-Thy-1.1/Ly-5.2 mice were used to obtain embryos. The day the vaginal plug was observed was designated as 0.5 dpc. The uterus was removed and washed to remove maternal blood. Fetuses were carefully removed to prevent contamination with maternal blood. Fetuses were then decapitated in Hanks' balanced salt solution containing 5 mM EDTA and allowed to bleed out. Fetuses were passaged through several dishes of media until completely pale. Blood was combined and centrifuged. Blood was either prepared for injection by sedimenting RBCs in dextran followed by lysis of erythrocytes in 0.15 M ammonium chloride, 0.01 M potassium bicarbonate solution on ice, or remained unmanipulated and was injected directly into recipients. Results from both preparations were comparable. Spleens were obtained by first removing the spleen and stomach to a dish of Hanks' balanced salt solution containing 2% FBS. The spleens were then peeled from the surface of the stomach and placed in a clean dish of media. Femurs and tibia were removed and cleaned of muscle tissue. Spleens and bone were dissociated using the rubber end of a 1 ml-syringe plunger and filtered through nylon mesh. Blood and tissues were collected from fetuses obtained from at least three pregnant females for each timepoint measured. Fetuses that appeared developmentally advanced or delayed in any age group were discarded.
Adult recipient mice were lethally irradiated with a split dose of 950 rad as previously described (
FL HSCs and adult BM HSCs were prepared as for antibody staining. Cells were stained with a lineage cocktail of the same purified rat IgG monoclonal antibodies used for FACS sorting. The cells were then depleted by magnetic selection using anti-rat IgG beads as per manufacturer's instructions (Dynal Biotech, Oslo, Norway), followed by a 1h incubation in RPMI media (GIBCO-BRL, San Diego, California, United States) containing 10% FBS in a tissue culture flask at 37°C to remove adherent cells. Dual-chamber chemotaxis assays were performed using 24-well plates with 5-μm pore size inserts (Costar/Corning, Corning, New York, United States), as previously described (
Results shown in
The LocusLink (
We thank L. Jerabek for laboratory management, S. Smith for antibody preparation, and L. Hidalgo and D. Escoto for animal care. This work was supported by National Institutes of Health (NIH) grant 5R01 HL-58770 to ILW, NIH Training Grant in Molecular and Cellular Immunobiology 5T32AI07290-16 to JLC, National Institute of Allergy and Infectious Diseases training grant 5T32 AI-07290 to DEW, and American Cancer Society Grant PF-00-017-LBC to AJW.
aorta–gonad–mesenepheros region
bone marrow hematopoietic stem cell
colony-forming unit–spleen
days postconception
fetus equivalent
fetal liver hematopoietic stem cell
hematopoietic stem cell
long-term reconstituting hematopoietic stem cell
long-term multilineage reconstitution
stromal cell-derived factor-1α
Steel factor