Idea Transcript
Distinct roles for CREB-binding protein and p300 in hematopoietic stem cell self-renewal Vivienne I. Rebel*†, Andrew L. Kung*†, Elizabeth A. Tanner*, Hong Yang*, Roderick T. Bronson‡, and David M. Livingston*§ *Dana–Farber Cancer Institute and Harvard Medical School, 44 Binney Street, Boston, MA 02115; and ‡Department of Pathology, Harvard Medical School, Goldenson Building, 220 Longwood Avenue, Boston, MA 02115 Contributed by David M. Livingston, September 18, 2002
H
ematopoietic stem cells (HSC) are a rare population of multipotent cells responsible for life-long regeneration of blood cells. All HSC have the dual properties of self-renewal, i.e., the ability to produce new stem cells, and commitment to differentiation. Defects in either of these properties result in a lack of mature blood cell production. Little is known of the molecular mechanisms that determine whether an HSC selfrenews or commits to differentiation. There is circumstantial evidence suggesting that transcription regulating兾chromatin-modifying proteins, such as CREBbinding protein (CBP) and p300, play an instrumental role in stem cell fate decisions. In Caenorhabditis elegans, lack of CBP altered tissue specification by certain progenitor cells (1), and, in mice, CBP兾p300 expression was identified in fetal HSC but not in fetal, mature blood cells of various lineages (2). CBP and p300 are highly homologous proteins and participate in a myriad of cellular functions (3). Besides their function as chromatinremodeling proteins, CBP and p300 serve as coactivators of transcription in association with a diverse set of transcription factors (4, 5). Coactivation likely depends on their ability to acetylate nucleosomal proteins and兾or associated transcription factors (e.g., GATA-1, E2F, p53, and T cell factor). It also depends on their ability to recruit elements of the basal transcription machinery to promoters (6). Several studies describe redundancy in CBP and p300 function. However, there are also individuated p300 and CBP activities (7, 8). For example, in a comparison of p300 and CBP heterozygous knockout mice, p300⫹/⫺ animals showed no discernible phenotype, whereas CBP heterozygotes of similar genetic background (C57BL兾6) developed multiple phenotypic abnormalities, including hematopoietic failure and, eventually, hematologic malignancies (8). The hematologic signs observed in older CBP⫹/⫺ mice are reminiscent of human myelodysplastic syndrome, a disease that likely results from a defect in HSC (9, 10). We, thus, hypothesized that CBP contributes to normal hematopoiesis through the regulation of HSC behavior. Serial transplantation studies described herein demonstrate a critical role for CBP, but not p300, in maintaining an adequate pool of murine HSC through self-renewal. In contrast, p300, but not CBP, appears to contribute to normal hematopoiesis primarily through its support of hematopoietic differentiation. Methods Mice. CBP⫹/⫺ and p300⫹/⫺ mice (8) were backcrossed onto a C57BL兾6 background (N6). WT littermates served as controls. www.pnas.org兾cgi兾doi兾10.1073兾pnas.232568499
Congenic Ly5.1⫹ C57BL兾6 mice were purchased from the National Cancer Institute. Chimeric mice were created by injection of 8–12 embryonic stem (ES) cells into WT blastocysts, as previously described (11). Differences in strain backgrounds between blastocysts (BALB兾c or FVB) and ES cells (C57BL兾6) allowed for detection of chimerism based on coat color or allelic differences at the Ly-5 and H2-K loci [BALB兾c (Ly5.2⫹, H2Kd⫹), FVB (Ly5.1⫹, H2-Kq⫹), and C57BL兾6 (Ly5.2⫹, H2Kb⫹)]. All chimeric mice destined for fetal and neonatal analysis were created with FVB host blastocysts, because the expression of the Ly5 antigen on the cell surface of fetal liver (FL) hematopoietic cells was readily detectable whereas H2-K expression was not (data not shown). All experiments included the analysis of chimeric mice generated by injection of at least two, independent ES cell clones of each relevant genotype (WT vs. CBP⫺/⫺ vs. p300⫺/⫺). The mice were bred and maintained under microisolator conditions at the Redstone Animal Facility of the Dana–Farber Cancer Institute. All animal procedures were in accordance with the Institute’s policies regarding animal care and use. ES Cell Lines. The CBP⫺/⫺, p300⫺/⫺, and WT ES cell lines were
isolated de novo as previously described (11). Briefly, blastocysts were harvested 3.5 days post coitum from (C57BL兾6) CBP⫹/⫺ ⫻ CBP⫹/⫺ or p300⫹/⫺ ⫻ p300⫹/⫺ matings. Every ES cell line was genotyped by Southern blot analysis, and nullizygosity was confirmed by Western blot analysis. WT ES cells were rendered neomycin resistant by stable transfection with a neomycin phosphotransferase expression vector. p300⫺/⫺ and CBP⫺/⫺ ES cells were neomycin resistant due to the presence of the neomycin phosphotransferase expression cassette used in gene targeting (8, 12). In Vitro ES Cell Differentiation. Induction of in vitro hematopoietic differentiation by using a two-step differentiation protocol was performed as described (13, 14) with a few modifications. The primary differentiation step was initiated in liquid culture by using ultra low adherence plates (Costar) and Iscove’s modified Dulbecco’s medium (IMDM) supplemented with 15% FBS (StemCell Technologies, Vancouver), 2 mM Glutamine, 0.1 mM nonessential amino acids (Life Technologies, Grand Island, NY), and 1.5 ⫻ 10⫺4 M monothioglycerol (Sigma). Two days later, the embryoid bodies (EB) that were formed in liquid culture were transferred to methylcellulose-containing medium. After a total of 14 days of culture, EB were harvested, dispersed into single cell suspensions, and assayed in a secondary differentiation step for the presence of hematopoietic colony-forming cells (CFC). CFC were quantitated 12 days after plating.
Abbreviations: CBP, CREB-binding protein; HSC, hematopoietic stem cells; BM, bone marrow; FL, fetal liver; PB, peripheral blood; EB, embryoid bodies; H-EB, hematopoietic EB; CFC, colony-forming cells; ES, embryonic stem. †V.I.R.
and A.L.K. contributed equally to this work.
§To
whom correspondence should be addressed. E-mail: david㛭livingston@dfci. harvard.edu.
PNAS 兩 November 12, 2002 兩 vol. 99 兩 no. 23 兩 14789 –14794
CELL BIOLOGY
Hematopoietic stem cells (HSC) are tightly regulated through, as yet, undefined mechanisms that balance self-renewal and differentiation. We have identified a role for the transcriptional coactivators CREB-binding protein (CBP) and p300 in such HSC fate decisions. A full dose of CBP, but not p300, is crucial for HSC self-renewal. Conversely, p300, but not CBP, is essential for proper hematopoietic differentiation. Furthermore, in chimeric mice, hematologic malignancies emerged from both CBPⴚ/ⴚ and p300ⴚ/ⴚ cell populations. Thus, CBP and p300 play essential but distinct roles in maintaining normal hematopoiesis, and, in mice, both are required for preventing hematologic tumorigenesis.
Fig. 2. Contribution to hematopoiesis by CBP or p300 null cells in adult chimeric mice. Chimeric mice were created by injecting WT (n ⫽ 22 animals; gray symbols), p300⫺/⫺ (n ⫽ 20; open symbols), or CBP⫺/⫺ (n ⫽ 29; black symbols) ES cells into WT blastocysts of either BALB兾c (squares) or FVB (circles) origin. Each symbol represents a comparative analysis in a single animal of the contribution to PB cells and to coat color by the progeny of the above-noted ES cells, at ⬇8 wk of age. The relative contribution to coat color was estimated independently by two investigators, and the relative contribution of ES cells to the hematopoietic system was determined by flow cytometry.
In Vitro Assay for Clonogenic Cells. CFC of erythroid and兾or
myeloid cell types were assayed in methylcellulose-containing medium, supplemented with the appropriate cytokines (StemCell Technologies). For each sample, two duplicate dishes were analyzed after 12–14 days of culture. For adult bone marrow (BM) samples, 1.5 ⫻ 104 cells were plated per dish. For FL samples, 2 ⫻ 104 cells were plated in each dish. Cells were cultured with and without 1.4 mg兾ml G418 (GIBCO) to determine the proportion of ES cell-derived CFC in the BM or FL from chimeric mice. In Vivo Competitive Repopulation Assay. Competitive repopulation
Fig. 1. Heterozygous CBP HSC decline with age. HSC measurements in WT mice are represented by the striped bars, in p300⫹/⫺ mice by the open bars, and in CBP⫹/⫺ animals by the black bars. (A) Depicted is the average number of HSC (⫾SD) per femur of mice younger than 3 mo (⬍3 mo) or between 9 and 12 mo of age (9 –12 mo). Older mice with signs of BM hypocellularity and兾or splenomegaly were not included in this analysis. The results of three independent experiments were combined. (B) Secondary recipients each received ⬇1 ⫻ 106 cells of the indicated genotype present in a total cell dose that did not differ significantly among the three groups: 5.2 (⫾4.4) ⫻ 106 cells for recipients of BM cells from primary WT recipients, 7.0 (⫾1.3) ⫻ 106 cells from primary p300⫹/⫺ recipients, and 6.4 (⫾2.1) ⫻ 106 cells from primary CBP⫹/⫺ recipients. The ability of HSC to reconstitute irradiated recipients after two rounds of transplantation is expressed relative to WT. Depicted are the combined results of two independent experiments. In total, cells of 11 WT, 18 CBP⫹/⫺, and 16 p300⫹/⫺ primary BM recipients were transplanted into secondary recipients. (C) BM cells collected as a pool from 2–3 secondary recipients, each of which experienced a similar level of Ly5.2⫹ PB and BM reconstitution (WT, 9.5% Ly5.2⫹ cells in the BM; p300⫹/⫺, 8.6%; and CBP⫹/⫺, 11.4%), were transplanted into tertiary recipients. Increasing doses of BM cells of the relevant genotypes (Ly5.2⫹) were transplanted into three groups of five recipients each. The proportion of recipient mice within each group that experienced multilineage reconstitution is indicated. 14790 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.232568499
assays were performed as described (15). Brief ly, Ly5.1⫹ C57BL兾6 mice were used as recipients. Whole body irradiation was administered in a split dose (500 and 450 cGy), by using a 137Cs source. Increasing doses of BM or FL cells (Ly 5.2⫹), together with a standard dose of 1.2 ⫻ 105 Ly-5.1⫹ marrow cells (to ensure short-term survival), were transplanted into groups of irradiated mice. At least 3 mo after transplantation, peripheral blood (PB) was obtained to determine whether Ly5.2⫹ HSC had repopulated the hematopoietic system. Recipients were viewed as multilineage reconstituted (‘‘positive’’) if ⬎1% of all nucleated blood cells were Ly5.2⫹, and this population included both myeloid and lymphoid cells, as determined by their side scatter properties. To obtain the frequency of HSC among the transplanted Ly5.2⫹ cells, Poisson statistics were applied to the analysis of the number of negative (nonreconstituted) mice. All BM transplantation experiments were performed in the same manner. For secondary transplantations, only those primary recipients with ⱖ95% probability of having been repopulated with a single HSC were used as donors (15). Histopathology. Full necropsy was performed on all adult chimeric mice that were killed. Gross tumors were microdissected and histologically analyzed. Where feasible, the contribution to the tumor by cells of a particular genotype was determined by Southern Rebel et al.
Table 1. Detailed hematologic analysis of chimeric mice % ES cell-derived cells Genotype of ES cell CBP⫺/⫺ 1 2 3 4 5 6 7 8 9 10 p300⫺/⫺ 11 12 13 14 15 16 17 WT兾WT* 18 19 20 21 22 23 24 25
Age at time of death, wk
Coat color
PB
BM (all cells)
BM (CFC)
In BM transplant recipients†
14 28 28 28 34 39 47 50 59 70
40 60 85 5 75 75 40 50 70 40
3 0 0 0 0 0 0 0 0 0.3
0 0 0 0 0 0 0 0 0 0.6
0 0 0 0 ND 0 0 1 0 0
— ND ND ND ND ND — — — —
18 18 20 60 92 92 92
10 5 20 60 80 75 60
0 0 0 6 3 10 4
0 0 2 4 6 32 20
0 0 3 32 19 52 42
ND ND ⫹⫹ ND ⫹⫹ ⫹⫹ ⫹⫹
81 81 81 81 81 71 71 71
7 7 10 75 75 95 85 90
0 12 37 68 52 22 28 37
0 10 24 32 24 11 25 31
ND ND ND ND ND ND ND ND
Hematologic malignancy
⫹
⫹ ⫹ ⫹
⫹
⫹⫹⫹ ⫹⫹⫹
⫹⫹⫹
blot analysis, as previously described (8, 12), or by semiquantitative PCR of the neomycin phosphotransferase gene. Results Monoallelic Loss of CBP Results in Inadequate HSC Self-Renewal. To explain the previously observed hematopoietic failure in CBP⫹/⫺ mice (8), we searched for a possible role of CBP in HSC function. First, the number of HSC in BM from young and old WT, CBP⫹/⫺, and p300⫹/⫺ mice were quantified in a competitive repopulation assay. Purified BM cells were transplanted into lethally irradiated recipients at limiting dilution, aiming to reconstitute them with as few HSC as possible. This procedure, therefore, allows one to quantitate HSC (15). In strain-matched WT and p300⫹/⫺ animals, the total number of HSC per femur increased with age (Fig. 1A), as expected for C57BL兾6 animals (16, 17). However, in older (C57BL兾6) CBP⫹/⫺ mice (9–12 mo), the abundance of functional HSC was significantly lower than in age-matched WT animals (*, P ⬍ 0.03). Thus, aging CBP⫹/⫺ mice develop both hematopoietic failure (8) and HSC depletion. One potential cause of age-related loss of HSC in CBP heterozygotes would be a defect in HSC self-renewal. To explore this possibility, BM cells from animals of various genotypes were subjected to serial transplantation. Primary transplant recipients were used as donors for secondary transplantation (Fig. 1B), and, to stress the system further, the combined BM from two to three secondary reconstituted recipients was subjected to a third round of standardized transplantation (Fig. 1C). Because the donor mice for the secondary transplantations had each been Rebel et al.
repopulated by only one or two Ly5.2⫹ HSC, the presence of Ly5.2⫹ blood cells in multiple secondary and tertiary transplant recipients would indicate extensive HSC self-renewal in the primary recipients. Compared with primary recipients of HSC from WT donors, BM HSC preparations from only half of the primary recipients of CBP⫹/⫺ HSC gave rise to reconstitution of secondary recipients (Fig. 1B). Moreover, none of the tertiary recipients of low doses of CBP⫹/⫺ cells from secondary recipients were reconstituted, unlike animals that received comparable numbers of WT cells (Fig. 1C). A much less prominent reduction in self-renewal capability was detected among p300⫹/⫺ HSC (Fig. 1 B and C). Taken together, these data demonstrate that loss of a single copy of CBP results in a reduction in HSC self-renewal capacity below a critical level, resulting in the depletion of HSC with age and associated hematopoietic failure. By contrast, the relatively small reduction in self-renewal of p300⫹/⫺ HSC was neither sufficient to cause HSC depletion (Fig. 1 A), nor was it associated with a clinically evident hematologic phenotype (8). Critical Role for CBP in Maintenance of HSC Pools. To further define
the role of CBP and p300 in hematopoiesis, we studied the effect of the complete absence of either protein in HSC self-renewal and differentiation. Others have reported defective primitive hematopoiesis (i.e., yolk sac) in CBP nullizygous mice (18) or mice homozygous for a truncated form of CBP (19). However, the study of HSC during definitive hematopoiesis (i.e., in FL and, ultimately, in adult animals), which can be detected in the PNAS 兩 November 12, 2002 兩 vol. 99 兩 no. 23 兩 14791
CELL BIOLOGY
ND, not done. *For transplantation purposes, the BM from mice nos. 18 –20, 21, and 22, and 23–25 were combined. †The — indicates no multi-lineage ES cell-derived reconstitution in the transplanted recipients; ⫹⫹ indicates between 25 and 50%; and ⫹⫹⫹ indicates ⬎50%.
Fig. 3. CBP⫺/⫺ and p300⫺/⫺ hematopoiesis in FL and in newborn chimeric animals. (A and B) FL cells isolated from E14.5 embryos were transplanted into lethally irradiated C57BL兾6 animals. The combined results of three independent transplantation experiments are shown in each panel. (A) Competitive repopulation of lethally irradiated recipients by unseparated FL single cell suspensions. The number of ES cell-derived FL cells (⫻105) injected per recipient mouse is indicated. (B) Secondary transplantations of marrow cells from primary donors reconstituted with 1–2 ES cell-derived HSC were performed to compare self-renewal ability of FL HSC of the indicated genotype. The ability to reconstitute secondary recipients is expressed relative to WT. (C and D) Newborn chimeric mice were killed 0 –1 day after birth. Liver (C), BM (D), PB, and spleen from 8 –9 chimeric neonates per genotype were harvested and analyzed for ES cell-derived (Ly5.2⫹) hematopoietic cells. Results of the analyses of PB and spleen (data not shown) were concordant with results of analyzing liver and BM hematopoietic cells within the same animal. Each symbol represents one neonate, and the horizontal bars represent the average of the data for each genotype.
embryo beginning at ⬇embryonic day 11 (E11; ref. 20), has been hampered due to the early (ⱕE11.5) embryonic lethality associated with these genotypes (12, 18, 19). Therefore, to test the effect of loss of both copies of CBP or p300 on definitive hematopoiesis, we created chimeric mice by injecting strainmatched CBP⫺/⫺, p300⫺/⫺, or WT ES cell lines (C57BL兾6; Ly-5.2⫹, H2-Kb⫹, neomycin resistant) into WT blastocysts. The host blastocysts were either of BALB兾c (Ly-5.2⫹, H2-Kd⫹) or FVB (Ly-5.1⫹, H2-Kq⫹) origin. Gross chimerism was readily achieved in the progeny of these mixed blastocysts. Indeed, there was as much as ⬇90% contribution to coat color by the progeny of ES cells of all genotypes (Fig. 2), indicating that nullizygous CBP and p300 ES cells regularly and efficiently contributed to at least one tissue in these animals. To learn whether the progeny of p300 and CBP nullizygous ES cells contribute to the hematopoietic system, PB from ⬇8-wk-old chimeric mice was analyzed. In WT animals, the degree of coat color chimerism was highly predictive of the degree of PB cell chimerism (Fig. 2). By contrast, even when there were high levels of coat chimerism in CBP⫺/⫺;WT兾WT chimeric animals, no CBP⫺/⫺ PB cells were detected. Surprisingly, in p300⫺/⫺;WT兾WT chimeric mice, there was also significant underrepresentation of p300⫺/⫺ PB cells. Thus, it appears that CBP and p300 are both required for normal, definitive hematopoiesis. Further analysis of chimeric mice revealed that the lack of CBP⫺/⫺ PB cells in CBP⫺/⫺;WT兾WT chimeras was accompanied by a lack of CBP⫺/⫺ cells in the BM, including CFC (Table 14792 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.232568499
1). In keeping with these results, transplantation of BM cells from five of these mice (nos. 1 and 7–10) failed to result in CBP⫺/⫺ hematopoietic reconstitution of lethally irradiated recipients, thus revealing the absence of CBP⫺/⫺ HSC in these mice. These results demonstrate that CBP is either essential for formation of HSC, or maintenance of an adequate pool of HSC. Conversely, although there was a paucity of p300⫺/⫺ PB cells in p300⫺/⫺;WT兾WT chimeric animals, there were abundant numbers of p300⫺/⫺ hematopoietic cells in the BM (Table 1). Transplantation of BM cells from four p300⫺/⫺;WT兾WT chimeric mice (nos. 13 and 15–17) resulted in multilineage hematopoietic reconstitution with p300⫺/⫺ cells in lethally irradiated WT recipients (Table 1). Thus, functional p300⫺/⫺ HSC are present in chimeric BM, but, in this environment, they are defective in producing differentiated progeny that reach the PB. These findings suggest that p300⫺/⫺ HSC are defective in proper hematopoietic differentiation. CBP Is Not Required To Initiate HSC Formation but Is Essential for Maintaining the HSC Pool. Given the absence of CBP⫺/⫺ HSC in
adult chimeric mice, we wondered whether CBP⫺/⫺ HSC were present at all at an earlier developmental stage. Therefore, the state of hematopoiesis was analyzed in FL from E14.5 embryos arising after blastocyst injection and, subsequently, in various tissues of newborn chimeric mice. We found that FL progenitor cells, assayed as CFC in methylcellulose, were readily detectable in chimeras of all three genotypes, although CBP⫺/⫺ and p300⫺/⫺ CFCs were slightly less abundant compared with those of WT Rebel et al.
Fig. 4. p300, but not CBP, is required for hematopoietic differentiation in vitro. In each experiment, four WT, three p300⫺/⫺, and two CBP⫺/⫺ ES cell lines were tested in an in vitro differentiation assay. (A) WT and CBP⫺/⫺ ES cells, but not p300⫺/⫺ ES cells, formed EB characteristic of structures that had undergone hematopoietic differentiation (arrows). They were scored as H-EB. (B and C) After 14 days of culture in primary differentiation medium, EB were analyzed for their content of hematopoietic CFC (B) and cell number (C). Each point represents the average value of the results of three independent experiments for clones of the indicated genotype. *, All three p300⫺/⫺ clones proved to be incapable of forming H-EB. Therefore, the number of CFC per H-EB could not be formally determined. (D) ES cells of each genotype have the potential to generate adipocytes in vitro. Shown are cells stained with Oil Red O after in vitro differentiation (21).
Rebel et al.
p300ⴚ/ⴚ but Not CBPⴚ/ⴚ ES Cells Fail to Differentiate into Hematopoietic Cells in Vitro. The reduced self-renewal capacity of
CBP⫺/⫺ HSC likely explains why their progeny cannot maintain normal hematopoiesis. Self-renewal of p300⫺/⫺ HSC, however, was not dramatically affected. Indeed, there was a net accumulation of p300⫺/⫺ HSC兾early progenitor cells in the BM of p300⫺/⫺;WT兾WT mice with aging (compare the results in Fig. 3D and Table 1). The relative inability of p300⫺/⫺ HSC to produce differentiated progeny suggests a possible defect in hematopoietic differentiation. To test this hypothesis, the hematopoietic differentiation potential of primitive p300⫺/⫺ cells was assayed by using an in vitro differentiation protocol. For each genotype, multiple ES cell clones, including those that successfully generated chimeric mice, were used in a two-step differentiation protocol. In the first step, EB were exposed to specific cytokines in an effort to generate hematopoietic cells. The existence of hematopoietic cells is reflected by the presence of relatively large cells migrating out of EB (ref. 14; Fig. 4A). In our assay, WT and CBP⫺/⫺ ES cell lines each gave rise to bountiful hematopoietic EB (H-EB). By contrast, none of the p300⫺/⫺ clones generated H-EB. Confirming the apparent absence of hematopoietic cells in EB formed during the first-step cultures, second-step cultures initiated with p300⫺/⫺ cells generated no hematopoietic CFC (Fig. 4B). By contrast, CBP⫺/⫺ H-EB yielded normal or even supernormal numbers of hematopoietic progeny by comparison with WT H-EB cultivated under similar conditions. These data suggest indeed a significant role for p300, but not CBP, in hematopoietic differentiation. p300⫺/⫺ ES cells produced EB containing ⬇4-fold fewer cells PNAS 兩 November 12, 2002 兩 vol. 99 兩 no. 23 兩 14793
CELL BIOLOGY
origin (data not shown). Competitive repopulation assays were then performed with pooled FL cells in search of CBP⫺/⫺, p300⫺/⫺, and WT HSC. The overall degree of chimerism of these pooled FL cell suspensions was determined before injection by immunostaining for the presence of Ly5.2⫹ (ES cell-derived) and Ly5.1⫹ (FVB blastocyst-derived) cells. The proportion of FL cells of ES cell origin was 6.9 and 3.2% for CBP⫺/⫺, 2.3 and 4.3% for p300⫺/⫺, and 5.7 and 4.5% for WT兾WT chimeras. Thus, the fraction of cells derived from the indicated ES cells in the injected cell suspensions was similar for all three genotypes tested. Although CBP⫺/⫺ and p300⫺/⫺ HSC were each detected in the FL of the relevant chimeric mice, more CBP⫺/⫺ FL cells were needed to reconstitute a lethally irradiated recipient than those of WT or p300⫺/⫺ origin (Fig. 3A). In addition, CBP⫺/⫺ HSC derived from primary transplant recipients were significantly less active in reconstituting secondary recipients than WT or p300⫺/⫺ HSC (Fig. 3B), again consistent with the presence of a defect in HSC self-renewal. Thus, early in definitive hematopoietic development (i.e., at the FL stage), neither CBP nor p300 is essential for HSC formation per se. However, CBP, but not p300, is essential for HSC self-renewal. In an effort to characterize further the loss over time of CBP⫺/⫺ and p300⫺/⫺ mediated hematopoiesis, we also analyzed hematopoiesis in newborn mice. The results revealed markedly reduced numbers of CBP⫺/⫺ or p300⫺/⫺ cells in those hematopoietic tissues tested by comparison with the case of ES-derived WT兾WT cells (Fig. 3 C and D). In keeping with these observations, there were also very few differentiated hematopoietic progeny of either knock-out ES cell in the PB of adult chimeric animals (Fig. 2).
than WT or CBP⫺/⫺ ES cells (Fig. 4C), raising the question of whether the inability to differentiate along a hematopoietic lineage is nonspecific, and merely the result of an impediment in EB formation. To address this question, ES cells were induced to differentiate into adipocytes (21), which, like hematopoietic cells, are mesoderm derivatives. Neither the lack of p300 nor CBP prevented adipocyte differentiation (Fig. 4D), suggesting that mesodermal induction is present in p300⫺/⫺ EB. Thus, their inability to produce hematopoietic cells in vitro is not a general characteristic of mesoderm-derived lineages. p300 and CBP Are both Important for Preventing Hematologic Tumor Formation. Prior experimental evidence revealed that CBP pos-
sesses hematologic tumor suppressor activity (8). Consistent with this finding, one adult CBP⫺/⫺;WT兾WT chimeric animal was found to harbor a hematologic malignancy (thymic lymphoma). Southern blot analysis showed that the tumor contained only CBP⫺/⫺ cells (data not shown) (8). It is possible that the absence of additional hematologic tumors in CBP⫺/⫺;WT兾WT chimeric mice is linked to the lack of a contribution by CBP⫺/⫺ cells to the hematopoietic system. Despite the absence of tumors in p300 heterozygotes (8), hematologic tumors did appear in multiple p300⫺/⫺;WT兾WT chimeric animals (Table 1). Histiocytic sarcomas were observed in most p300⫺/⫺;WT兾WT mice that were analyzed, and these masses appeared to be composed largely, if not exclusively, of p300⫺/⫺ cells, as reflected by Southern blot analysis. Thus, both CBP and p300 appear to play a role in suppressing hematologic tumor development. Discussion Self-renewal and differentiation are the key processes through which the HSC pool and, thus, normal blood cell production and function are maintained. Here, we show that two paralogous proteins, CBP and p300, are required for optimal support of these processes. Interestingly, despite high protein homology and the performance of many overlapping functions, their roles are, at least in part, distinct during hematopoietic development. CBP is essential for HSC self-renewal, whereas p300 is not. The presence of both CBP and p300 in WT, highly purified Sca1⫹Linlo/⫺c-Kit⫹⫹CD34⫺ HSC was confirmed by Western blotting (data not shown). Thus, the dependence of HSC for self-renewal on CBP, as opposed to p300, is not a product of the presence of CBP and not p300 in these cells. Although p300 is not essential for HSC self-renewal, it appears to play an important role in hematopoietic differentiation. Sup1. Shi, Y. & Mello, C. (1998) Genes Dev. 12, 943–955. 2. Phillips, R. L., Ernst, R. E., Brunk, B., Ivanova, N., Mahan, M. A., Deanehan, J. K., Moore, K. A., Overton, G. C. & Lemischka, I. R. (2000) Science 288, 1635–1640. 3. Goodman, R. H. & Smolik, S. (2000) Genes Dev. 14, 1553–1577. 4. Blobel, G. A. (2000) Blood 95, 745–755. 5. Eckner, R. (1996) Biol. Chem. 377, 685–688. 6. Nakajima, T., Uchida, C., Anderson, S. F., Lee, C. G., Hurwitz, J., Parvin, J. D. & Montminy, M. (1997) Cell 90, 1107–1112. 7. Eid, J. E., Kung, A. L., Scully, R. & Livingston, D. M. (2000) Cell 102, 839–848. 8. Kung, A. L., Rebel, V. I., Bronson, R. T., Ch’ng, L. E., Sieff, C. A., Livingston, D. M. & Yao, T. P. (2000) Genes Dev. 14, 272–277. 9. Cooper, L. J., Shannon, K. M., Loken, M. R., Weaver, M., Stephens, K. & Sievers, E. L. (2000) Blood 96, 2310–2313. 10. Nilsson, L., Astrand-Grundstrom, I., Arvidsson, I., Jacobsson, B., HellstromLindberg, E., Hast, R. & Jacobsen, S. E. (2000) Blood 96, 2012–2021. 11. Hogan, B., Beddington, R., Costantini, F. & Lacy, E. (1994) Manipulating the Mouse Embryo: A Laboratory Manual (Cold Spring Harbor Lab. Press, Plainview, NY). 12. Yao, T. P., Oh, S. P., Fuchs, M., Zhou, N. D., Ch’ng, L. E., Newsome, D., Bronson, R. T., Li, E., Livingston, D. M. & Eckner, R. (1998) Cell 93, 361–372. 13. Helgason, C. D., Sauvageau, G., Lawrence, H. J., Largman, C. & Humphries, R. K. (1996) Blood 87, 2740–2749. 14. Wiles, M. V. & Keller, G. (1991) Development (Cambridge, U.K.) 111, 259–267.
14794 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.232568499
porting this view are the results of in vitro ES cell differentiation assays (Fig. 4) and those of the analyses of hematopoietic cells in adult p300⫺/⫺;WT兾WT chimeric animals, in which there was an accumulation of p300⫺/⫺ HSC兾progenitor cells in the BM, accompanied by a relative absence of mature cells in the PB (Table 1 and Fig. 2). However, p300⫺/⫺ HSC transplanted into previously irradiated, WT animals appeared to have overcome this differentiation defect. Specifically, abundant, well-differentiated p300⫺/⫺ progeny were detected in the PB of these animals. These seemingly contrasting observations lead to the speculation that the presence of p300⫺/⫺ stromal cells in the BM microenvironment may affect the differentiation of p300⫺/⫺ HSC兾progenitor cells. The notion that the microenvironment in which HSC reside is crucial to their fate is well established (22), and the possibility that the role of p300 in hematopoietic differentiation is, in part, through non-cell autonomous mechanisms is currently under investigation. It remains to be seen whether CBP and兾or p300 exert their influence in HSC self-renewal and differentiation through the execution of certain chromatin remodeling functions, the recruitment and兾or acetylation of highly specific transcription factors allowing the regulation of possibly stage- and lineagespecific target genes, or a combination of such effects. In this regard, it is worthwhile noting that Ikaros, another chromatin remodeling protein, appears to exert its effect on HSC and the development of both lymphoid and myeloid lineages through a combination of both of the above-noted mechanisms (23). Although overexpression (24) or, in other instances, depletion of certain gene products (25–27) has, thus far, been shown to affect self-renewal of nonleukemogenic, definitive HSC, CBP is the first gene in which monoallelic loss perturbs normal, adult HSC function. In this regard, understanding the specific biochemical contributions of CBP and p300 to hematopoiesis should begin to shed light on molecular mechanisms governing the most basic HSC fate decisions. Such knowledge could also shed light on the mechanism(s) by which p300 and CBP suppress hematologic malignancies. We thank Julie Kujawa and Jill Gerber for their expert assistance with animal experiments, Cheryl Helgason for sharing her knowledge of in vitro ES cell differentiation with us, and Evan Rosen for assistance with adipocyte differentiation. We also thank M. Walhout, R. Shivdasani, and J. Dekker for their critical reviews of the manuscript. This work was supported by grants from the National Cancer Institute (to D.M.L. and A.L.K.) and the Leukemia and Lymphoma Society (to V.I.R.). V.I.R. and A.L.K. were both supported by the Claudia Adams Barr Program in Cancer Research. H.Y. was a fellow of the Sharf-Green Research Fund. 15. Rebel, V. I., Miller, C. L., Thornbury, G. R., Dragowska, W. H., Eaves, C. J. & Lansdorp, P. M. (1996) Exp. Hematol. 24, 638–648. 16. Morrison, S. J., Wandycz, A. M., Akashi, K., Globerson, A. & Weissman, I. L. (1996) Nat. Med. 2, 1011–1016. 17. Sudo, K., Ema, H., Morita, Y. & Nakauchi, H. (2000) J. Exp. Med. 192, 1273–1280. 18. Tanaka, Y., Naruse, I., Hongo, T., Xu, M., Nakahata, T., Maekawa, T. & Ishii, S. (2000) Mech. Dev. 95, 133–145. 19. Oike, Y., Takakura, N., Hata, A., Kaname, T., Akizuki, M., Yamaguchi, Y., Yasue, H., Araki, K., Yamamura, K.-I. & Suda, T. (1999) Blood 93, 2771–2779. 20. Ling, K. W. & Dzierzak, E. (2002) Curr. Opin. Immunol. 14, 186–191. 21. Rosen, E. D., Sarraf, P., Troy, A. E., Bradwin, G., Moore, K., Milstone, D. S., Spiegelman, B. M. & Mortensen, R. M. (1999) Mol. Cell 4, 611–617. 22. Lemischka, I. R. (1997) Stem Cells 15, 63–68. 23. Georgopoulos, K. (2002) Nat. Rev. Immunol. 2, 162–174. 24. Sauvageau, G., Thorsteinsdottir, U., Eaves, C. J., Lawrence, H. J., Largman, C., Lansdorp, P. M. & Humphries, R. K. (1995) Genes Dev. 9, 1753–1765. 25. Cheng, T., Rodrigues, N., Shen, H., Yang, Y., Dombkowski, D., Sykes, M. & Scadden, D. T. (2000) Science 287, 1804–1808. 26. Ohta, H., Sawada, A., Kim, J. Y., Tokimasa, S., Nishiguchi, S., Humphries, R. K., Hara, J. & Takihara, Y. (2002) J. Exp. Med. 195, 759–770. 27. Rebel, V. I., Hartnett, S., Hill, G. R., Lazo-Kallanian, S. B., Ferrara, J. L. & Sieff, C. A. (1999) J. Exp. Med. 190, 1493–1504.
Rebel et al.