Clonogenic mast cell progenitors and their excess numbers in chimeric BALB/c mice with inactivated GATA-1
- The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville 3050, Victoria, Australia
-
Contributed by Donald Metcalf, October 9, 2007 (received for review September 3, 2007)
Abstract
In agar cultures of marrow cells from adult female BALB/c chimeric GATA-1Plt13/+ mice, a high frequency of unusual dispersed colonies was noted. Analysis showed that these were colonies of mast cells and that mast cell colony-forming cells (progenitors) could be detected in clonal cultures of adult marrow, neonatal marrow, or fetal liver if the combined stimulus of stem cell factor and interleukin-3 was used. Mast cell progenitors were in active cell cycle and showed an extensive capacity for self-generation. Mast cell colonies both from control GATA-1+/+ mice and GATA-1Plt13/+ mice could generate growth factor-dependent cloned cell lines that grew for >18 months. Surprisingly, the majority of the excessive numbers of mast cell progenitors in chimeric GATA-1Plt13/+ mice were transcribing the inactive Plt13 allele of GATA-1, suggesting that GATA-1 normally acts to restrict the emergence of committed mast cell progenitors. In sharp contrast, all eosinophil progenitors in these mice were transcribing the normal GATA-1 allele. No excess tissue mast cells were observed in GATA-1Plt13/+ mice, suggesting that the excess mast cell progenitors in these mice might be generating mast cells with a defective in vivo proliferative or tissue homing capacity.
Mast cells are important mediators of allergic responses after interaction of antigens with IgE bound to the mast cell surface (1). Mast cells are generated from progenitor cells in the bone marrow, and, after their migration to mucosal or connective tissues, they develop characteristic differences in their gross morphology and phenotype according to their location (2, 3).
To date, there has been some reluctance to attempt characterization of mast cell progenitors in clonal agar cultures because of early reports that colony macrophages can phagocytose metachromatic agar and then be confused with mast cells (4, 5).
GATA-1 is a highly conserved transcription factor expressed in erythroid, megakaryocytic, eosinophilic, and mast cells. GATA-1 is essential for the formation and maturation of erythroid cells and for survival in fetal life. GATA-1 is also necessary for the proper maturation of megakaryocytic cells, although mast cell development was possible (6). Studies on chimeric embryos, some of whose cells lack GATA-1, also documented that GATA-1 is essential for the formation of eosinophil progenitors in the fetal liver (7). Similarly, deletion of the high-affinity GATA-1 binding site in the GATA-1 promoter also led to selective loss of eosinophil lineage cells (8).
In mice lacking the distal promoter of the GATA-1 gene and having impaired production of GATA-1 (GATA-1low), mast cell differentiation was impaired (9); and where expression of GATA-1 in adult mice was suppressed by overexpression of dominant negative forms of GATA-1 or by iRNA, survival of mast cells was impaired, as was their response to IgE-induced degranulation (10).
Studies on mice with inactivation of the GATA-1 gene have been severely restricted because of the fetal death of such animals due to failure to form mature red cells (6). In experiments using BALB/c mice, we described the generation of Plt13 mice that carry an ENU-induced inactivating mutation in GATA-1 (11). The GATA-1 gene is located on the X-chromosome, and female mice heterozygous for the Plt13 mutation are chimeras of GATA-1+ and inactivated GATA-1Plt13 cells due to the lyonization of the X-chromosome. The GATA-1Plt13/+ mutation disrupts the initiation codon in exon-2 of the gene, and no GATA-1 protein of any type was detectable in GATA-1Plt13 megakaryocytes. This failure to detect protein in cells known to produce high levels of GATA-1 protein strongly suggests that the mutant GATA-1Plt13 allele does not produce detectable amounts of GATA-1 protein and that female cells exclusively expressing this mutated allele are essentially GATA-1 knockout cells. The Plt13 mouse is thrombocytopenic and has multiple abnormalities in hematopoietic populations (11). The chimeric state of Plt13 mice makes it possible to establish, by clonal analysis, which hematopoietic cells in Plt13 mice are derived from mutant GATA-1Plt13 cells.
Results
Mast Cell Colony-Forming Cells.
In clonal cultures, adult GATA-1Plt13/+ marrow cells were unremarkable in producing the expected numbers and types of colonies when stimulated by granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), macrophage colony-stimulating factor (M-CSF), or IL-3 (Table 1). However, in cultures stimulated by the combination of stem cell factor (SCF)+IL-3+ erythropoietin (EPO), GATA-1Plt13/+ cells from the marrow or spleen generated elevated numbers of unusual loosely dispersed colonies of variable size (Fig. 1 A and Table 1).
Colony formation by cells from GATA-1Plt13/+ and GATA-1+/+ mice
Mast and blast cell colonies. (A) Diffuse mast cell colony grown from GATA-1Plt13/+ marrow cells by using SCF+IL-3+EPO. (B) Multicentric blast colony grown from GATA-1+/+ marrow cells, using SCF+G-CSF. (C) Mast colony cells with metachromatic granules grown in methylcellulose culture, stained with May–Grünwald Giemsa. (D) Macrophage colony cells grown in methylcellulose, stained with May–Grünwald Giemsa. (E) Colony mast cells strongly labeled by anti-kit sera. (F) Colony granulocytes and macrophages failing to be labeled by anti-kit sera.
In cultures stained with acetylcholinesterase plus Luxol Fast blue and hematoxylin, these dispersed colonies were quite distinctive, permitting the differential colony counts shown in Table 1. The cells in dispersed colonies were unusually uniform in shape and size. Their nuclei stained less intensely than the blast cells in multicentric blast cell colonies (Fig. 1 B). The cells had a round, slightly excentric nucleus, and the cytoplasm had a speckled appearance and was slightly brown in color. Characteristically, there were always mitotic cells in dispersed colonies. In contrast, blast colony cells had a relatively narrow basophilic cytoplasm or no visible cytoplasm. Dispersed colonies were sometimes similar in general shape and size to some macrophage colonies, but macrophage colony cells were of irregular size and had more bulky cytoplasm that was often irregular in shape.
It was noteworthy that, although the combination of stem cell factor plus IL-3 plus erythropoietin stimulated blast cell, granulocytic, macrophage, eosinophil, megakaryocyte, and some erythroid colony formation, the dispersed colonies developing in these cultures contained none of these cells. There were, however, infrequent blast colonies containing megakaryocytes in the cultures, and these may have been generated by multipotential progenitors. As shown in Table 1, IL-3 alone stimulated the formation of few or no dispersed colonies, and stem cell factor alone was similarly inactive (data not shown).
When the dispersed colonies were picked off, cytocentrifuged, then stained with May-Grünwald Giemsa, the cells were found to contain a uniform population of metachromatic granules and the dispersed colonies were recognized to be mast cell colonies. This identification might have been regarded as suspect because of early reports that macrophage colony cells can phagocytose metachromatic agar and then resemble mast cells (4, 5). However, a paired comparison of dispersed colony cells with macrophage colony cells (grown by using M-CSF) showed that macrophage metachromatic granules were consistently different. They were inconstant, irregular in size, and usually associated with a vacuole. Confirmation that the dispersed colony cells were indeed mast cells was obtained from colonies of GATA-1Plt13/+ cells grown in methylcellulose, using SCF+IL-3+EPO or M-CSF. As shown in Fig. 1 C and D, dispersed colony mast cells remained filled with metachromatic granules, but colony macrophages had a bulky cytoplasm filled with unstained vacuoles. Further confirmation of the mast cell nature of the dispersed colony cells was obtained by showing that cytocentrifuged preparations of dispersed colony cells stained strongly with antibodies to c-Kit, an identifying feature of mast cells (Fig. 1 E and F).
With practice, mast cell colonies could be identified in unstained bone marrow cultures with reasonable confidence. In reliability tests, sequential candidate mast cell colonies were identified in unstained cultures then cytocentrifuged and stained with May–Grünwald Giemsa. Of 36 such colonies presumptively identified as mast cell colonies, 32 were composed wholly of mast cells, 2 were blast cell colonies, and 2 were macrophage colonies. Estimates of mast cell colony numbers from unstained marrow cultures were therefore reasonably accurate, if a little imprecise. Identification of mast cell colonies responding to stimulation by SCF+IL-3+EPO was much easier in unstained cultures of GATA-1Plt13/+ neonatal marrow and fetal liver, because mast cell colonies were the most frequent colony type present in these cultures and frequencies were sharply higher than in control cultures of GATA-1+/+ cells (Table 1).
Of note, in view of the reports of multipotential precursors generating mast cells, erythroid cells, and megakaryocytes in cultures of GATA-1low cells (9), no dispersed mast cell colony contained acetylcholinesterase-positive megakaryocytes or erythroid cells either at day 7 or 14 of incubation, and the addition of thrombopoietin to the stimulating mixture did not alter this absence.
Further verification of the mast cell nature of these colonies was obtained by picking off individual candidate mast cell colonies and using them to initiate cloned cell lines, using SCF+IL-3+EPO as the proliferative stimulus. As shown in Table 2, the success rate in establishing these mast lines was very high for fetal liver, neonatal marrow, and adult marrow colonies. Mast cell lines were also able to grown from colonies in cultures of GATA-1+/+ cells, although it was more difficult to find candidate mast cell colonies in fetal liver cultures, and some macrophage colonies were chosen by mistake. Mast cell lines grown from GATA-1Plt13/+ and GATA-1+/+ colonies were identical in appearance (Fig. 2), contained no cells of other lineages, and uniformly lacked markers for lymphoid, erythroid or myeloid cells but were positive for c-Kit and CD41, a recently recognized marker of mast cells (12). The cultures were split at weekly intervals, and there were no consistent growth rate differences between GATA-1Plt13/+ and GATA-1+/+ cell lines. The cells from 12 of 13 GATA-1Plt13/+-derived cell lines tested were consistently more strongly positive for CD41 than were cells from 10 of 11 GATA-1+/+-derived colonies (Fig. 2).
Success rate in establishing continuous mast cell lines from individual mast cell colonies grown in agar
Continuous cloned mast cell lines. (A) Morphology of cells grown from a GATA-1+/+ marrow colony, stained with May–Grünwald Giemsa. (B) Morphology of cells grown from a GATA-1Plt13/+ marrow colony, stained with May–Grünwald Giemsa. (C) FACS profiles showing labeling with anti-CD41 of four mast cell lines grown from GATA-1+/+ and four grown from GATA-1Plt13/+ marrow colonies. Note stronger labeling of GATA-1Plt13/+ cell lines. The light lines indicates the profiles of cells labeled with the isotype control serum. (D) Stimulation of colony-formation in agar by cells from five mast cell lines grown from GATA-1Plt13/+ marrow colonies. Note that the cells were only stimulated to proliferate by SCF+IL-3+EPO, SCF+IL-3, or, less well, by SCF+GM-CSF. Each culture contained 200 cells, and colony counts were performed after 7 days of incubation. Data are mean colony counts from duplicate cultures.
The most remarkable feature of the cell lines derived from randomly chosen mast cell colonies from cultures of GATA-1Plt13/+ cells was that PCR analysis showed that 24 of 25 tested were the progeny of cells transcribing the nonfunctional Plt13 allele. This finding was reinforced by a further analysis of 12 of 12 mast cell colonies grown from GATA-1Plt13/+ cells, which showed that all were composed of cells using the Plt13 allele (Fig. 3). These combined data indicate that the large majority of the excess numbers of mast cell progenitors in GATA-1Plt13/+ mice were GATA-1Plt13 cells and that this allele permitted the excess generation of committed mast cell progenitors.
Genotype of mast cell and eosinophil progenitors. RNA was extracted from whole spleen (SPL), individual colonies grown from mice heterozygous for the Plt13 mutation or wild-type littermates. Mast cell colonies were grown by using SCF, IL-3, and EPO and were identified based on their morphology; eosinophil colonies were grown by using IL-5. cDNA was prepared, and a region including the second exon of GATA-1 was amplified by using two rounds of PCR. The resulting PCR products were treated (+) with the restriction endonuclease NcoI, and fragments were resolved on a 3% agarose gel. PCR products that carry the Plt13 mutation are resistant to cleavage and were identified by their larger size.
As a control to these mRNA studies, eosinophil colonies developing in agar cultures when stimulated by IL-5 were analyzed. Of 37 such colonies analyzed from cultures of GATA-1Plt13/+ cells, all colonies were GATA-1+ in allelic-type (Fig. 3), confirming earlier studies on fetal liver cells that no eosinophil progenitor formation occurs in the absence of GATA-1 (7). Less definitive were results from an analysis of blast and granulocytic colonies grown by using SCF+G-CSF, because many lacked detectable GATA-1 transcripts (data not shown).
No differences were noted in the regulators required to stimulate the proliferation of mast cell colonies grown from either GATA-1+/+ or GATA-1Plt13/+ cells or in the proliferation of mast cells lines derived from mice of either type. As shown by the typical results in Fig. 2 with colony formation by mast cell lines, unexpectedly, mast cell proliferation was not stimulated in primary cultures or cell line cultures by either IL-3 or SCF alone. Proliferative stimulation required the combinations of SCF+IL-3, SCF+IL-3+EPO or, less efficiently, SCF+GM-CSF.
Cycling Status of Mast Cell Progenitors and Tissue Mast Cell Numbers.
In three separate experiments, bone marrow cells from GATA-1Plt13/+ mice were incubated in vitro for 20 min with high concentrations of tritiated thymidine or with control medium alone. After thorough washing, subsequent quadruplicate clonal cultures revealed that there had been an average reduction of 46 ± 5% in the numbers of mast colonies developing after exposure to tritiated thymidine compared with numbers in control cultures. This was similar to the 35 ± 8% reduction in other colony types seen in these cultures stimulated by SCF+IL-3+EPO. The data indicated that mast cell colony-forming cells, like other committed progenitor cells, are in active cell cycle in young adult mice.
The high frequency of mast cell progenitors in BALB/c GATA-1Plt13/+ mice and their active cycling status led to the expectation that mast cell numbers in various tissues might be higher than in GATA-1+/+ mice. Data from mast cell counts in groups of 2-month-old BALB/c GATA-1Plt13/+ and GATA-1+/+ mice are shown in Table 3. Organs surveyed included those containing small mucosal-type mast cells, such as the spleen, and organs containing large connective tissue-type mast cells, such as the skin, organ capsule cells, or peritoneal cavity cells. As shown in Table 3, no consistent differences were noted in mast cell numbers between the two strains. Of particular note was the failure to find any metachromatic mast cells in any of the GATA-1Plt13/+ sternal bone marrows.
Tissue mast cell content
Recloning of GATA-1+/+ and GATA-1Plt13/+ Blast and Mast Cell Colonies.
GATA-1+/+ blast colonies grown by using SCF+G-CSF are multicentric and exceptionally large. Recloning of these multicentric BALB/c colonies, using GM-CSF in the secondary cultures, produced larger and more numerous secondary colonies than had been observed in studies using C57BL/6 cells (13) (Table 4), however, only a few BALB/c blast-type multicentric colonies were grown by using IL-3 or SCF+IL-3+EPO. This establishes only a minor capacity of these blast colony-forming cells for self-generation. The behavior of GATA-1Plt13/+ multicentric blast colonies grown with SCF+G-CSF was essentially identical to that of GATA-1+/+ cells. Of special note was the virtual absence in secondary cultures of any convincing mast cell colonies (only three possible mast cell colonies were observed in the colonies grown from 10 blast colonies, and these might have been macrophage colonies). The negative nature of these results suggests strongly that mast colony-forming cells are not the progeny of blast colony-forming cells that form multicentric colonies when stimulated by SCF+G-CSF.
Progenitor cell content of blast or mast cell colonies
In sharp contrast, recloning of GATA-1Plt13/+ presumptive mast colonies grown with SCF+IL-3+EPO generated an average of >400 daughter mast cell colonies per colony. Many of these were equal in size to, or larger than, the primary mast cell colonies, and the cells generating these colonies were clearly capable of extensive self-generation. The cells in mast cell colonies were, however, structured in proliferative capacity. Some cells formed only small colonies or clusters, and by day 14, such clusters and colonies had disappeared.
Discussion
The formation of slowly growing mast cell colonies in methylcellulose cultures is described in ref. 14. The present study describes a method for identifying and enumerating mast cell-committed progenitors in mouse hematopoietic tissues by their ability to form more rapidly growing distinctive dispersed colonies in agar cultures. The mast cell colonies in the present study were first recognized because of their presence in 10-fold higher numbers in cultures of chimeric heterozygous female Plt13 BALB/c mice with an inactivating mutation of the X-linked GATA-1 gene.
Individual mast cell colonies grown from BALB/c and Plt13 mice did not differ consistently in shape or size and both required stimulation either by SCF+IL-3 or by SCF+IL-3+EPO or less effectively by SCF+GM-CSF. The absence of any cells with metachromatic granules in the bone marrow of normal or Plt13 mice implies that committed mast cell progenitors are themselves nongranulated.
Plt13 mast cell progenitors were highest in frequency in the fetal liver with a progressive decline in frequency until adult life, where they remained 10-fold more frequent than in control BALB/c mice. This frequency was maintained unaltered at least until 18 months of age (data not shown).
The remarkable feature of the excess mast colonies grown from GATA-1Plt13/+ chimeric mice was that the vast majority of mast cell colony-forming cells were transcribing the mutated functionless GATA-1 allele. This implies that GATA-1 normally acts to restrict the emergence of committed mast cell progenitors. This was in sharp contrast to the absolute inability of cells using the mutated allele to generate eosinophil progenitor cells, in agreement with earlier studies (7, 8). The eosinophil data also complement previous inverse experiments in which overexpression of GATA-1 in granulocyte-macrophage lineage cells forced cells into the eosinophil lineage (7, 15).
The GATA-1Plt13 mast cell progenitors were in active cell cycle and could readily generate sustained clonal mast cell lines containing no other cell types. These progenitor cells and their progeny may well have had intrinsic abnormalities, because the cells exhibited abnormally high levels of CD41 and no excess numbers of mast cells accumulated in vivo either in the bone marrow or peripheral tissues. In this regard, previous studies on chimeric GATA-1-depleted fetal liver cells documented a normal ability of GATA-1 mast cells to proliferate in vitro (6), but in GATA-1low mice, mast cell differentiation was impaired (9), and, where expression of GATA-1 was suppressed by dominant negative forms of GATA-1, survival of mast cells was impaired, as was their response to IgE-induced degranulation (10). The present study on chimeric mice, where the mast cell progenitors used the inactivated GATA-1 allele, reinforces the possibility that, in the absence of GATA-1, mast cells may be unable to proliferate or migrate normally in vivo. This contrast between in vitro and in vivo behavior of GATA-1− cells may be a further example of the paradoxical behavior already documented for erythroid and megakaryocytic cells (6, 11).
A somewhat similar situation has recently been documented in mice with a conditional knockout of the SCL gene. Excess numbers of mast cell progenitors were reported in these mice, but no increases in mast cell numbers were noted in peripheral tissues until the mice were injected with IL-3 (16). In view of the dependency of GATA-1Plt13/+ mast cells on stimulation by the combination of SCF and IL-3 and the absence of IL-3 in unmanipulated mice (17), the injection of IL-3 might similarly increase mast cell numbers in the Plt13 spleen—the spleen being the most responsive to IL-3 stimulation of mast cell numbers (18). There would also be merit in examining the allelic type of tissue mast cells developing in Plt13 mice in response to worm infestation (19, 20) to establish whether GATA-1Plt13 mast cells were able to take part in such responses.
The situation in Plt13 mice contrasts intriguingly with that described in GATA-1low mice, where mast cell progenitors were absent from the marrow but novel progenitors were present that had a heightened proliferative activity. These cells again depended on a combination of growth factor stimuli, but where the cells were multilineage in potential, forming not only mast cells but megakaryocytes and erythroid cells (9, 21). Furthermore, the mast cells in peripheral tissues in these mice were abnormal and apoptotic. The two models both document the vital role played by GATA-1 in the biology of mast cell formation and the sharp differences in detail may be related to strain differences between the mice but more likely represent subtle consequences of low versus no expression of GATA-1 on the lineage commitment of early hematopoietic precursors.
The origin of the excess numbers of mast cell progenitors in Plt13 mice remains to be established. They may be more actively produced by mutated more ancestral cells, such as CFU-S, but, given their capacity for extensive self-renewal, their presence in excess numbers may merely represent self-renewal, possibly in the absence of an effective ability to expend themselves by generating maturing progeny in vivo.
Materials and Methods
Mice.
The Plt13 line of BALB/c strain mice was established during an ENU-mutagenesis screen designed to identify mutations that impair platelet number. The mutated gene was characterized as an inactivation of the GATA-1 gene on the X-chromosome, and only heterozygous female GATA-1Plt13/+ mice were viable. These showed thrombocytopenia but had excess numbers of GATA-1Plt13 megakaryocytes with aberrant cytoplasmic differentiation (11).
Cultures.
Clonal cultures were prepared in 35-mm Petri dishes containing 1 ml of DMEM with 20% FCS and 0.3% agar (22). Cultures contained 25,000 6- to 8-week-old adult marrow cells, 50,000 adult spleen cells, 10,000 1- to 3-day-old neonatal marrow cells, or 2,500 day-13 fetal liver cells and were stimulated by final concentrations of one or more of the following purified recombinant mouse proteins produced in this laboratory or purchased from PeproTech: 10 ng/ml GM-CSF, 10 ng/ml M-CSF, 10 ng/ml IL-3, 100 ng/ml SCF, 2 international units (IU)/ml erythropoietin, 10 ng/ml human G-CSF, or a combination of SCF+IL-3+EPO.
After 7 days of incubation at 37°C in a fully humidified atmosphere of 10% C02 in air, cultures were scored then fixed with 1 ml of 2.5% glutaraldehyde. After 4 h, cultures were floated intact onto glass slides, dried, then stained in sequence for acetylcholinesterase, Luxol Fast Blue and hematoxylin, then all colonies were counted and classified as described in ref. 22. Methylcellulose cultures were performed and analyzed as described in ref. 23.
To initiate cloned cell lines, individual 7-day colonies were removed from cultures, using a fine pipette, and transferred to 1-ml cultures of DMEM with 10% FCS, containing 100 ng of SCF, 10 ng of IL-3, and 2 IU EPO. Cultures were repassaged at weekly intervals.
Recloning of Colonies.
Individual 7-day colonies were transferred to 8 ml of agar medium, resuspended, then cultured in duplicate cultures containing 10 ng of GM-CSF, 10 ng of IL-3, or 100 ng of SCF with 10 ng of IL-3 and 2 IU EPO.
FACS Analysis.
FACS analysis of continuous cell lines was performed as described in ref. 24.
RT-PCR.
mRNA analysis of individual colonies was performed by using the GATA-1 primers described in ref. 11, with minor modifications. To amplify the GATA-1 message from single colonies, a second round of PCR was performed with nested primers (forward, 5′-CCACTAAGGTGGCTGAATCC; reverse, 5′-GTTGAGGCAGGGTAGAGTGC). PCR products were treated with the restriction enzyme NcoI and separated on agarose gels (3–5%).
Thymidine Suiciding.
Bone marrow cells from 8-week-old mice were subjected to thymidine suiciding as described in ref. 25.
Histochemistry.
Individual 7-day colonies were removed and resuspended in saline then cytocentrifuged onto slides, and the cells were fixed with paraformaldehyde then processed with anti-kit antibodies as described in ref. 26.
Acknowledgments
This work was supported by grants from the Cancer Council Victoria; National Health and Medical Research Council, Canberra Program Grant 257500; and United States National Institutes of Health Grant CA22556.
Footnotes
- *To whom correspondence should be addressed. E-mail: metcalf{at}wehi.edu.au
-
Author contributions: D.M., I.M., and W.S.A. designed research; D.M., I.M., S.M., and L.D.R. performed research; D.M., I.M., S.M., L.D.R., and W.S.A. analyzed data; and D.M. wrote the paper.
-
The authors declare no conflict of interest.
- © 2007 by The National Academy of Sciences of the USA
