Id1 restrains myeloid commitment, maintaining the self-renewal capacity of hematopoietic stem cells

Abstract

Appropriate hematopoietic stem cell (HSC) self-renewal reflects the tight regulation of cell cycle entry and lineage commitment. Here, we show that Id1, a dominant-negative regulator of E protein transcription factors, maintains HSC self-renewal by preserving the undifferentiated state. Id1-deficient HSCs show increased cell cycling, by BrdU incorporation in vivo, but fail to efficiently self-renew, leading to low steady-state HSC numbers and premature exhaustion in serial bone marrow transplant assays. The increased cycling reflects the perturbed differentiation process, because Id1 null HSCs more readily commit to myeloid differentiation, with inappropriate expression of myeloerythroid-specific genes. Thus, Id1 appears to regulate the fate of HSCs by acting as a true inhibitor of differentiation.

Hematopoietic stem cells (HSCs) have the ability to both maintain their own pool (self-renewal) and to constantly generate all mature peripheral blood cell lineages. The self-renewal capacity of HSCs is clearly sufficient to reconstitute the hematopoietic system upon stress, such as bone marrow transplantation or myeloablative therapy. The functional reserve of HSCs depends on the precise control of cell proliferation and cell fate determination, because either excessive HSC commitment to differentiate or impaired self-renewal divisions can result in HSC exhaustion in experimental models.

Decisions regarding self-renewal vs. commitment are based on microenvironmental cues, which predominantly use the Notch, Wnt, and Shh signaling pathways (1–7). However, the critical, cell-intrinsic regulators of HSC commitment are unknown.

Here, we demonstrate a unique function of the inhibitor of DNA binding 1 (Id1) dominant-negative basic helix–loop–helix transcription factor in restricting myeloid lineage commitment in HSCs and preserving their self-renewal capacity. Id1 is a member of the family of four proteins (Id1–4) known to inhibit the activity of the E protein bHLH transcription factors by restraining their ability to bind DNA. By modulating the function of transcription factors such as the E2A and Ets1, the Id proteins have been shown to regulate cell proliferation and cell fate determination both in vitro and in vivo (8–11).

We have examined the hematopoietic compartment of Id1^−/− mice and found a decreased frequency and diminished self-renewal capacity of HSCs in Id1-deficient bone marrow. These changes were associated with an increased turnover and accelerated myeloid differentiation of HSCs. Furthermore, Id1 null HSCs show increased expression of key myeloerythroid transcription factors, indicating a critical role of Id1 in the transcriptional repression of myeloid lineage commitment. Thus, Id1-dependent genetic programs are critical in preventing the premature differentiation and exhaustion of HSCs.

Results

Id1 Regulates the Frequency, Self-Renewal Capacity, and Myeloerythroid Commitment of HSCs.

We can readily detect Id1 mRNA and protein expression in the immature (Lin⁻) fraction but not in the differentiated (Lin⁺) fraction of murine bone marrow cells (Fig. 1A). Id1 mRNA expression is ≥4-fold higher in the most primitive Lin⁻ c-kit⁺ Sca-1⁺ (LKS) cells within the Lin⁻ compartment compared with various subsets of committed myeloid progenitor cells, represented by the Lin⁻ c-kit⁺ Sca-1⁻ (LKS⁻) phenotype (Fig. 1B). This finding suggested a regulatory role for Id1 in early hematopoiesis and led us to investigate the hematopoietic phenotype of Id1 null mice.

Fig. 1.

Id1 regulates the maintenance of steady-state bone marrow cellularity, LKS cell frequency and myeloerythroid lineage determination. (A) Id1 protein is readily detectable in the nuclei of a subset of Lin⁻ but not Lin⁺ bone marrow cells. Arrows point to representative overlay images of Id1 (red fluorescence) and nuclear DNA DAPI staining (blue fluorescence) (B) Id1 mRNA level is highest in the most primitive hematopoietic cells and decreases after myeloid commitment. FACS-purified subsets of immature bone marrow hematopoietic cells were purified by multiparameter flow-cytometric sorting and analyzed for Id1 RNA expression by qPCR. Id1 expression was normalized to HPRT to obtain relative mRNA levels. The results represent an average (±SEM) of four measurements from two independent sorting experiments. (C) Decreased frequency of primitive LKS hematopoietic cells in Id1^−/− animals. Shown profiles are representative of flow-cytometric staining for Lin⁻ c-kit⁺ Sca-1⁺ (HSC and MPP) and Lin⁻ c-kit⁺ Sca-1⁻ (HPC) in Id1^+/+ and Id1^−/− bone marrow. The numbers indicate average (±SEM) percentage of nucleated bone marrow cells. (n = 15; ∗∗, P < 0.01). (D and E) Representative flow-cytometric profiles of LT-HSC, ST-HSC, MPP, and myeloid progenitor subsets (CMP, GMP, MEP) in Id1^+/+ and Id1^−/− bone marrow by using the published multiparameter surface staining criteria. The numbers indicate the average (±SEM) percentage of LKS⁺ and LKS⁻ populations, respectively (n = 4; ∗, P < 0.05).

The cellularity of the Id1 null bone marrow was ≈75% of normal [supporting information (SI) Fig. 6]. However, the distribution of the differentiated (Lin⁺) bone marrow subsets is unperturbed (SI Fig. 7), suggesting that the decreased cellularity could arise from a stem/progenitor cell homeostatic abnormality rather than from defects in the maturation of their progeny. Loss of Id1 does not significantly impair mature blood cell production, because we observe normal myeloid cell numbers and only marginally (10–15%) decreased red blood cell counts in the peripheral blood of Id1^−/− mice (SI Table 2). The significant (≈50%) decrease in circulating lymphocytes indicates an as yet undescribed impairment in lymphoid homeostasis in these mice.

A flow-cytometric examination of immature bone marrow hematopoietic cells shows a 2- to 4-fold decrease in the steady-state frequency of primitive, pluripotent LKS cells in the absence of Id1, with a normal frequency of the committed progenitors (LKS⁻) (Fig. 1C). There was no change in the frequency of long-term (CD34⁻flt3⁻) or short-term (CD34⁺flt3⁻) self-renewing HSCs within the Id1^−/− LKS cells (Fig. 1D), indicating a proportional reduction in both Id1^−/− HSC subsets (12–14). The number of side-population cells, identified by Hoechst 33342 dye efflux capacity, which possess a functional HSC phenotype (15), is also markedly reduced in Id1^−/− bone marrow (SI Fig. 8). Furthermore, wild-type transplant recipients of Id1^−/− bone marrow show reduced frequencies of LKS cells, indicating an intrinsic function of Id1 in HSCs (SI Fig. 9).

Although the frequency of committed progenitors (LKS⁻) is not significantly altered in Id1^−/− mice, the relative proportion of megakaryocyte–erythroid and granulocyte–monocyte progenitors (MEPs and GMPs, respectively) is inverted: the MEP/GMP ratio is 2.6 in Id1^−/− murine bone marrow vs. 0.5 in the wild-type bone marrow (Fig. 1E). This finding represents a true skewing in myeloerythroid commitment rather than perturbed cell surface marker expression, because both the Id1^+/+ and Id1^−/− immunophenotypically defined progenitor subsets show the appropriate colony-forming capacity in vitro (SI Table 2). Therefore, the absence of Id1 in vivo favors megakaryocyte–erythroid commitment of multipotent progenitors, consistent with the previously suggested role of Id1 in suppressing erythroid differentiation (16, 17).

To best quantify the number of functional HSCs in Id1^−/− bone marrow, we performed a competitive repopulation assay using varying numbers (2, 0.67, 0.22, and 0.074 × 10⁵) of Id1^+/+ or Id1^−/− CD45.2⁺ donor bone marrow cells together with a constant number (2 × 10⁵) of congenic CD45.1⁺ competitor bone marrow cells to transplant into lethally irradiated recipients. Consistent with the reduction in LKS and side-population bone marrow cells, the frequency of competitive repopulating units in Id1^−/− bone marrow is decreased ≈3-fold based on a limiting dilution analysis of peripheral blood repopulation 3 months after transplant (Table 1). Despite the decrease in HSC frequency, we do observe comparable levels of lymphomyeloid chimerism after a competitive bone marrow transplant with nonlimiting numbers (2 × 10⁵) of Id1^+/+ or Id1^−/− donor cells (Fig. 2A). This finding indicates the normal (or even enhanced) repopulating activity of the remaining Id1^−/− HSCs that can compensate for their diminished frequency.

Table 1.

Limiting dilution competitive repopulation assay

Donor genotype	Cell dose	Positive recipients, %	CRU (% BMNC)
ID1+/+	2 × 105	100	0.023 ± 0.012*
	0.67 × 105	100
	0.22 × 105	100
	0.074 × 105	80
ID1−/−	2 × 105	100	0.007 ± 0.003*
	0.67 × 105	100
	0.22 × 105	80
	0.074 × 105	40

BMNC, bone marrow nucleated cell; CRU, competitive repopulating units.

*95% confidence interval.

Fig. 2.

Decreased self-renewal capacity of multilineage repopulating HSCs in the absence of Id1. (A) Robust competitive long-term lymphomyeloid repopulating ability of Id1^−/− bone marrow cells. Flow-cytometric analysis of donor myeloid (Gr1⁺) and B-lymphoid (B220⁺) cell contribution in the peripheral blood 5 months after transplant (based on percentage of CD45.2⁺ cells). Bar graphs represent average (±SEM) values for three mice per group. The results are representative of two independent experiments. (B and C) Defective maintenance of the HSC self-renewal capacity in the absence of Id1. Bar graph represents average (±SEM) values of donor contribution to hematopoiesis in the peripheral blood and Lin⁻ bone marrow in groups of secondary recipients of Id1^+/+ (gray bars) vs. Id1^−/− (red bars) bone marrow cells (based on percentage of CD45.2⁺ cells) (n = 6).

We next performed a secondary competitive repopulation analysis to assess the self-renewal capacity of Id1^−/− HSCs under conditions of repeated replicative stress. Secondary transplantation was performed 5 months after the primary (noncompetitive) engraftment of Id1^+/+ or Id1^−/− CD45.2⁺ bone marrow cells in lethally irradiated CD45.1⁺ donor mice. We transplanted 2 × 10⁵ CD45.2⁺ donor cells together with equal numbers of CD45.1⁺ competitors into secondary recipients and monitored the secondary CD45.2⁺ donor cell engraftment over 5 months. At 1 month after transplant, the level of peripheral blood chimerism derived from wild-type and Id1 null secondary transplanted bone marrow cells is similar (Fig. 2B). However, at 2 months, we began to observe a competitive disadvantage of Id1-deficient cells that became highly significant at 5 months after transplant in both the blood and the bone marrow of the recipient mice (Fig. 2 C and D). Thus, Id1-deficient HSCs have diminished self-renewal capacity and exhaust more rapidly than Id1 expressing cells. In addition to the decrease in steady-state HSC frequency, these results demonstrate a critical role of Id1 in HSC maintenance under conditions of transplant-related stress.

Increased Turnover of LKS Cells in the Absence of Id1.

Numerous studies have shown a critical role of cell cycle regulators in the maintenance of HSC self-renewal capacity (18–23). Because Id1 is induced by mitogens and is thought to be essential for proper cell cycle progression in several cell types in vitro (11, 24), we measured the growth rate of single, sorted LKS cells in response to cytokine stimulation. Unexpectedly, the cell expansion was significantly faster in Id1^−/− than in Id1^+/+ cells during the first 48 h, whereas the growth rates were comparable between 48 and 96 h in culture (Fig. 3A). This finding implies the faster recruitment of Id1^−/− LKS cells into the cell cycle in response to hematopoietic growth factors, with no significant defects in cell cycle progression.

Fig. 3.

Increased cycling status of Id1^−/− LKS cells. (A) Enhanced in vitro proliferation of sorted Id1 LKS⁺ cells in response to cytokine stimulation. LKS cells were sorted into a 96-well plate (1 cell per well), and clonal expansion was followed for 4 days by manual cell counting under the microscope. The results show the average (±SEM) numbers of cells per well (50 wells were scored for each genotype) and are depicted on a log scale. These data are representative of two independent sorting experiments. (B) Higher rate of S-phase entry in Id1^−/− LKS compartment in vivo. After 3 days of BrdU administration in drinking water, bone marrow cells were simultaneously stained for surface markers and BrdU incorporation into newly synthesized DNA. Gray histograms indicate isotype Ab control; red histograms indicate anti-BrdU Ab. The profiles shown are representative of those seen in individual mice. Numbers indicate the percentage (±SEM) of BrdU⁺ LKS⁺ cells (n = 4; ∗, P < 0.05).

A significant fraction of HSCs are quiescent under steady-state conditions in vivo (19, 25, 26), with LKS cells entering the cell cycle more slowly than committed progenitor cells (27). To measure the cycling status of Id1^−/− HSCs in vivo, we tracked DNA synthesis by adding BrdU to the drinking water for 3 days and then immunostaining bone marrow cells for nuclear BrdU and for LKS cell surface markers. Significantly more LKS cells were BrdU⁺ in the Id1^−/− animals than the Id1^+/+ animals (79.0 ± 3.9% positive in Id1^−/− vs. 49.7 ± 7.4% positive in Id1^+/+), demonstrating a higher rate of S-phase entry in vivo (Fig. 3B).

Premature Myeloid Commitment of LKS Cells in the Absence of Id1.

HSC homeostasis reflects the balance between self-renewal and differentiation, which must be regulated to maintain an appropriate number of HSCs yet generate sufficient numbers of mature blood cells during steady-state and stress hematopoiesis. The more rapid exhaustion of Id1^−/− HSCs suggests a defect in this process, and to test the propensity of Id1^−/− LKS cells to commit to myeloid differentiation, we used the in vivo colony-forming unit spleen (CFU-S) assay. Colonies that form in the spleen of lethally irradiated recipient mice 12 days after transplantation (CFU-S_d12) arise from primitive progenitors and HSCs, whereas splenic colonies that appear on day 8 (CFU-S_d8) are generated from more committed myeloerythroid progenitor cells (28, 29). As expected, wild-type sorted LKS cells generate almost exclusively day-12 colonies; however, Id1 null LKS cells give rise to a similar number of day-12 colonies but also a significantly increased number of day-8 colonies (Fig. 4A). The CFU-S_d8 colony-forming ability of Id1 null LKS cells resembles the more differentiated common myeloid progenitor (CMP) and MEP cells (28), indicating the premature commitment to myeloid differentiation of these Lin⁻ c-kit⁺ Sca-1⁺ cells.

Fig. 4.

Premature myeloid commitment within the Id1^−/− LKS compartment. (A) Spleen colony-forming assay shows increased CFU-S_d8 activity by using sorted Id1^−/− LKS⁺ cells. Five hundred sorted LKS⁺ cells from Id1^+/+ or Id1^−/− bone marrow were transplanted into lethally irradiated recipients, and spleen colonies were scored on days 8 and 12 after transplant. Numbers indicate average colony numbers (±SEM) (n = 12) from three independent experiments. (B and C) In vitro stem/progenitor cell differentiation depends on the cell intrinsic expression of Id1. The results shown are representative of three independent experiments. The numbers indicate the percentage of cells remaining in the lineage-negative gate after 5 days in liquid culture. The accelerated expression of lineage markers in the absence of Id1 expression (red arrow) is reversed by retroviral expression of Id1 in primitive Id1^−/− LKS⁺ cells (blue arrow) (B) but not Id1^−/− LKS⁻ cells (C).

To further investigate the regulatory role of Id1 in regulating HSC differentiation, we restored the expression of Id1 using retroviral gene transfer. Consistent with the altered in vivo CFU-S potential, empty vector-transduced Id1^−/− LKS cells show accelerated lineage marker expression over 5 days in response to early acting cytokines (IL-3, IL-6, and SCF) (Fig. 4B). Retroviral transduction of the Id1 cDNA rescued the normal dynamics of lineage marker expression in the Id1 null HSCs, with no effect in the wild-type HSCs (Fig. 4B). Id1 overexpression had no effect on committed progenitor cells, whether isolated from wild-type or Id1 null mice (Fig. 4C). This finding demonstrates the cell-autonomous function of Id1 in controlling the early stages of HSC commitment rather than acting as a general inhibitor of differentiation.

Gene Expression in Id1^−/− LKS Reflects Enhanced Myeloid Lineage Priming.

To determine how the biological differences between Id1 null HSCs and their normal counterpart are reflected in their gene expression profiles, we used oligonucleotide microarrays (Affymetrix, Santa Clara, CA) to study hematopoietic stem/progenitor cell subsets. As shown (30, 31), a large number of transcripts are differentially expressed during the commitment of wild-type HSC cells to the CMP phenotype. In our experiments, 2,038 genes were reproducibly decreased and 388 increased during the wild-type HSC-to-CMP transition, by using a >2-fold change in relative expression as a cutoff. These gene clusters represent the transcriptional signatures of immaturity and self-renewal (in the HSC-associated transcripts) and myeloerythroid commitment (in the CMP-associated transcripts), respectively. The obtained data are consistent with several similar published microarray studies (30–33) (see SI Data Sets 1 and 2). Within these experimentally obtained gene clusters we selected certain transcripts that have been associated with myeloid commitment or the HSC phenotype, based on a detailed literature search (shown in Fig. 5 A and B). The expression levels of critical erythroid (such as GATA1, EKLF, GFI1B, and EPOR) and myeloid (CEBPA, M-CSFR, and MPO) genes were increased 1.5- to 3-fold in Id1^−/− LKS cells (Fig. 5A), whereas HSC-enriched transcripts (such as ANG, EVI1, MCL1, HOXA9, EYA1, and EYA2) (31, 32) showed normal expression levels in Id1-deficient LKS cells (Fig. 5B). The increased relative expression of CMP-associated transcripts in Id1 null LKS cells, without a corresponding decrease in the expression of HSC-associated genes is also apparent within the unselected, experimental CMP and HSC-specific gene clusters (SI Fig 10).

Fig. 5.

Myeloid lineage priming in Id1^−/− LKS cells. (A and B) Sorted subsets of Id1^+/+ and Id1^−/− cells were analyzed by Affymetrix oligonucleotide array transcript profiling. Myeloid and HSC-associated gene lists were compiled from the experimentally obtained CMP and HSC-associated clusters (SI Data Sets 1 and 2) by using published microarray studies (30–32) and a detailed literature search. (A) In the CMP cluster, 19 of 26 probe sets show similar expression in Id1^−/− LKS and CMPs, that is, higher than in the wild-type LKS cells. (B) Most of the probe sets in the LKS cluster show similar expression in Id1^+/+ and Id1^−/− LKS cells, that is, higher than in the CMPs. (C) Quantitative mRNA expression (qPCR) in Id1^−/− LKS cells of genes relevant for myeloid commitment, self-renewal, and cell cycle regulation in HSCs. The results were normalized to HPRT expression and expressed as average fold change (±SEM) for Id1^−/− over Id1^+/+ LKS cells. Three independent RT-PCR measurements were performed for each gene (∗, P < 0.05; ∗∗, P < 0.01).

To confirm the expression levels of the known regulators of HSC self-renewal, myeloerythroid lineage commitment, and cell cycle regulation in the Id1^−/− LKS cells, we performed quantitative (q)PCR analyses. We confirmed the increased expression of both c/EBPα and GATA-1 mRNA and the normal levels of transcriptional regulators of HSC self-renewal, Gfi1 (20), Bmi1 (34), GATA-2 (22), and HoxB4 (35, 36) (Fig. 5C). No changes were found in expression of the CDK inhibitors p18, p27, and p16 that regulate the HSC cell cycle progression and replicative capacity (18, 19, 23, 34, 37) (data not shown). We also used qPCR to verify that the mRNA levels of the immunophenotypic markers used to characterize and isolate various hematopoietic cell populations closely correlated with their cell surface expression (data not shown).

As recently described, we found that pluripotent HSCs do express detectable levels of certain myeloid lineage-specific genes. This characteristic is thought to reflect the priming of HSCs for myeloid differentiation, which occurs before true lineage commitment (30, 38, 39). We show that Id1 loss results in a premature induction of the myeloid differentiation program in LKS cells, without a significant change in HSC enriched transcripts, including the known regulators of HSC self-renewal. This finding is consistent with the increased level of transcriptional myeloid lineage priming in Id1^−/− HSCs cells, before their full commitment to the progenitor cell phenotype.

Despite the increased cycling status of Id1^−/− LKS cells, the expression of the CDK inhibitor p21, a critical effector of HSC quiescence (19), is significantly increased in these cells (Fig. 5E). This finding is consistent with the increased E2A activity in HSCs that would occur in the absence of Id1, because p21 has been shown to be induced by E2A and repressed by Id1 (40).

Discussion

We have identified quantitative and qualitative defects in the HSC pool of Id1-deficient mice using a variety of in vivo and in vitro approaches. Id1^−/− LKS cells demonstrate an increased tendency to differentiate and enter the cell cycle, leading to increased HSC turnover, a diminished steady-state HSC frequency, and the premature loss of HSC self-renewal capacity under conditions of replicative stress.

Most known transcriptional regulators of HSC homeostasis, including Bmi-1 (34), Gfi-1 (20), GATA-2 (22, 41), and HOXB4 (36), appear to have their primary effect on HSC cell cycle events, thereby regulating the rate and capacity of HSC self-renewal divisions. In contrast, we propose that Id1 expression maintains the normal balance of self-renewal and differentiation primarily by actively restraining the HSC commitment to myeloid differentiation.

At the molecular level, Id1^−/− LKS cells show a relative induction of the myeloerythroid differentiation program, including elevated levels of c/EBPα and GATA-1, master regulators of myeloid and erythroid lineage differentiation, respectively, suggesting that the recently described myeloid lineage priming in HSCs is actively regulated by Id1 to maintain a balanced rate of HSC commitment.

The relative increase in MEP frequency in Id1^−/− bone marrow, indicating a shift in lineage determination, is consistent with studies, including our own, that demonstrated that Id1 can block erythroid differentiation (16, 17). Therefore, in addition to maintaining the self-renewal capacity of hematopoietic stem cells, Id1 plays a regulatory role in myeloerythroid lineage choice during HSC differentiation. A recently revised model of HSC lineage determination has proposed that at least a partial restriction of the erythroid program could take place at the very early stage of HSC differentiation, before the lymphomyeloid switch (14). Thus, it is tempting to speculate that the increased frequency of MEPs in the absence of Id1 could result from a premature erythroid commitment occurring at the LKS stage.

Based on our data, the increased rate of cell cycle entry of Id1^−/− LKS cells could reflect the recruitment of relatively quiescent HSCs into the actively proliferating progenitor pool (42) rather than a primary effect of Id1 loss on the HSC cell cycle. Id1 was shown to prevent cell cycle arrest by inhibiting expression of the p16 and p21 CDK inhibitors (8, 40). Indeed, we found an increase in p21 mRNA in Id1^−/− HSCs, but this was not sufficient to prevent increased HSC cycling in the absence of Id1. As hypothesized (11), any proposed role of Id1 in cell cycle regulation must be context-dependent, because Id1 loss can result in either increased or decreased cell proliferation. Our findings suggest that these results may be explained by broader effects of Id1 on cell fate determination.

Our results suggest that the proper balance between myeloerythroid commitment and HSC self-renewal depends on stem cell-specific active silencing of bHLH transcription factors, mediated by Id1. Thus far, our preliminary data indicate that the premature in vitro differentiation of Id1 null HSCs can be delayed by lowering the levels of E2A (data not shown). Future studies, examining hematopoiesis in mice deficient for both Id1 and its putative molecular targets may be required to determine the relative contribution of E protein silencing vs. other possible mechanisms of Id1 function in regulating HSC behavior.

This report shows that an Id protein is required for the proper self-renewal of adult tissue stem cells and is a key transcriptional regulator of HSC lineage commitment. Our findings could provide a target for attacking leukemic transformation, because Id1-mediated maintenance of immaturity might be aberrantly used by leukemic stem cells. We recently identified Id1 as a gene up-regulated by the presence of in AML1-ETO in primary human hematopoietic cells that show extensive in vitro self-renewal (43). Future studies will be needed to address the requirement for Id1 expression in the maintenance of malignant cell self-renewal in experimental models of leukemic transformation.

Materials and Methods

Mice.

The generation of Id1-deficient mice was described (44). Mice used in all experiments were of the matching sex, age (8–10 weeks old), and genetic background (C57BL/6 or B6/Sv129, as indicated below). C57BL/6 (CD45.2⁺), B6.SJL (CD45.1⁺), and F1B6/SV129 (CD45.2⁺) mice were purchased from The Jackson Laboratory, Bar Harbor, ME. All animals were maintained in germ-free air Thorensten units in the Memorial Sloan–Kettering Cancer Center Animal Facility, according to IACUC-approved protocols.

Flow Cytometry.

Murine bone marrow hematopoietic cells were analyzed and isolated based on the expression pattern of surface markers by using flow cytometry. For the analysis of HSC and myeloid progenitor subset frequencies, we used 8- to 10-week-old, sex matched wild-type and Id1^−/− mice of B6/Sv129 background. Bone marrow nucleated cells were obtained from hindlimbs by flushing the bone with an insulin syringe and IMDM plus 10% FBS. Lineage markers were labeled by using a mixture of biotin-conjugated antibodies (CD3, CD4, CD5, CD8, B220, Gr1, CD11b, and Ter119) from BD Pharmingen (San Jose, CA). To eliminate the early lymphoid progenitors from the primitive LKS subset, we added the biotin-conjugated CD127 (IL-7RA) antibody (Pharmingen) to the lineage mixture. The expression of lineage markers was subsequently detected by using streptavidin-APC-Cy7 or a streptavidin-PerCP conjugates from Pharmingen. The FACS analysis of LT-HSCs, ST-HSCs, and myeloid progenitor subsets (CMP, GMP, and MEP) was performed as described (12–14, 45). The directly labeled anti-Sca-1 FITC or PE, c-kit APC, CD34 FITC, and Flt3 PE antibodies were also obtained from Pharmingen, whereas the Sca-1 PE-Cy7 and CD16/32 (FcRgII) PE-Cy7 reagents were from eBioscience (San Diego, CA). The four-color analysis (FITC, PE, PerCP, and APC) was performed by using the FACSCalibur cytometer from Becton Dickinson (San Jose, CA). The five-color analysis (FITC, PE, PE-Cy7, APC, and APC-Cy7) as well as all cell sorting was performed by using the MoFlo cell sorter from Cytomation (Fort Collins, CO).

Bone Marrow Transplantation.

Recipient mice were lethally irradiated with 10 Gy of whole-body irradiation. For the noncompetitive, primary, and secondary competitive bone marrow transplantation experiments, donors were C57BL6 wild-type or Id1^−/− (CD45.2⁺) mice, whereas recipients and competitor bone marrow cells were of B6.SJL (CD45.1⁺) background. Analysis of peripheral blood chimerism was performed by immunostaining for the CD45 congenic marker isoforms on B220⁺ (B-lymphoid) and Gr1⁺ (myeloid) cells 5 months after transplant. For the limiting-dilution competitive transplantation assay (LD–CRU assay) and the spleen colony-forming unit assay (CFU-S), we used B6/Sv129 wild-type or Id1^−/− donor (CD45.2⁺) and B6.SJL (CD45.1⁺) competitor bone marrow cells, transplanted into F1B6/SV129 (CD45.2⁺) lethally irradiated recipients. The CRU numbers were calculated by using limiting dilution analysis (L-Calc software; Stem Cell Technologies, Vancouver, BC, Canada) of lymphomyeloid donor (CD45.2⁺) engraftment 3 months after transplant. The percentage of recipient-derived CD45.2⁺ cells in mice receiving the B6.SJL (competitors) alone, was <2.5%, which was considered the threshold for positive lymphomyeloid engraftment. For the spleen colony-forming assay, lethally irradiated recipients were transplanted with 500 sorted wild-type or Id1^−/− LKS cells per mouse. Transplant recipients were killed, and macroscopic spleen colonies were scored 8 or 12 days after transplant.

Retroviral Transduction and in Vitro Differentiation Assay.

A transient production system for generating ecotropic retroviruses was used to transduce murine bone marrow cells enriched for HSC. Briefly, the Phoenix-eco retroviral producer cells were used to produce retroviruses by transfection of the retroviral vectors MIGR1-Id1 (Id1-IRES-GFP) or MIGR1 empty vector used as a control (46). Retroviral supernatant was collected 48 and 72 h after transfection. Sorted hematopoietic stem/progenitor cells were cultured for 24 h in the presence of early acting cytokines (100 ng/ml SCF/10 ng/ml IL-3/10 ng/ml IL-6) (PeproTech, Rocky Hill, NJ) in serum-free media (X-vivo 15; Cambrex, CA), to induce proliferation. The cells were transduced by one round of spinoculation in Retronectin (TaKaRa, Shiga, Japan)-coated plates in the presence of 8 μg/ml of Polybrene (Sigma, St. Louis, MO), which resulted in 15–30% GFP⁺ cells. The next day, transduced cells were washed, cultured in the presence of cytokines for an additional 72 h (i.e., a total of 5 days in culture), and analyzed for lineage marker expression within the GFP⁺ gate by FACS.

In Vivo BrdU Incorporation.

Analysis of the in vivo BrdU incorporation into LKS cells was performed by using the FITC BrdU Flow Kit (Pharmingen) after a single i.p. injection of BrdU (Sigma Aldrich), 1 mg per 6 g of mouse weight and admixture of 1 mg/ml of BrdU to drinking water for 3 days. Mice were B6/Sv129 females, 8–10 weeks old. Cell surface markers were identified by using the lineage, c-kit and Sca-1 antibodies, as described above.

RNA Isolation and Gene Expression Analysis.

For gene expression profiling of primary hematopoietic stem cell and progenitor cell subsets, we pooled hindlimb bone marrow cells from three mice. Freshly isolated cells were sorted by surface marker expression, and total RNA was extracted by using the RNAEasy kit (Qiagen, Valencia, CA). To generate sufficient sample quantities for oligonucleotide gene chip hybridization and qPCR experiments, we used the GeneChip Two-Cycle cDNA Synthesis kit (Affymetrix) for cRNA amplification and labeling. The amplified cRNA was labeled and hybridized to the MOE430 Plus 2 oligonucleotide arrays (Affymetrix) or reverse-transcribed by using the First Strand cDNA Synthesis kit (Invitrogen) for qPCR experiments.

The Affymetrix gene expression profiling data were normalized by using the published Robust Multiarray Average (RMA) algorithm (47) using the GeneSpring 7 software (Agilent Technologies, Palo Alto, CA).

Messenger RNA expression levels for critical regulators of HSC differentiation and self-renewal were verified by qPCR using the SYBR green PCR Mastermix (Applied Biosystems, Foster City, CA), optimized sequence specific oligonucleotide primer pairs (provided upon request), and the 7900 Real Time PCR system (Applied Biosystems). The relative expression of target genes was calculated by using the standard curve method and normalized to HPRT mRNA content.

Statistical Analysis.

All P values (statistical significance) were determined by using the unpaired, two-tailed t test.

Abbreviations

CMP

common myeloid progenitor

GMP

granulocyte–monocyte progenitor

HSC

hematopoietic stem cell

LKS

Lin⁻ c-kit⁺ Sca-1⁺

LKS⁻

Lin⁻ c-kit⁺ Sca-1⁺

MEP

megakaryocyte–erythroid progenitor