Obligate Role of Anti-Apoptotic MCL-1 in the Survival of Hematopoietic Stem Cells

See allHide authors and affiliations

Science  18 Feb 2005:
Vol. 307, Issue 5712, pp. 1101-1104
DOI: 10.1126/science.1106114


Apoptosis is important in controlling hematopoietic stem cell (HSC) numbers. However, the specific BCL-2 family member(s) that regulate HSC homeostasis are not precisely defined. We tested myeloid leukemia–1 (MCL-1) as an attractive candidate that is highly expressed in HSCs and regulated by growth factor signals. Inducible deletion of Mcl-1 in mice resulted in ablation of bone marrow. This resulted in the loss of early bone marrow progenitor populations, including HSCs. Moreover, growth factors including stem cell factor increased transcription of the Mcl-1 gene and required MCL-1 to augment survival of purified bone marrow progenitors. Deletion of Mcl-1 in other tissues, including liver, did not impair survival. Thus, MCL-1 is a critical and specific regulator essential for ensuring the homeostasis of early hematopoietic progenitors.

Hematopoietic stem cells (HSCs), which give rise to all cells of hematopoietic lineages, can undergo several developmental fates. These include self-renewal or differentiation into multipotent progenitors that give rise to mature hematopoietic cells (1). This process is tightly controlled and regulated by apoptosis (2). Constitutive expression of human BCL-2 in vivo under the control of the mouse major histocompatibility class I promoter causes accumulation of HSCs and enhances their ability to form colonies in vitro and to reconstitute the hematopoietic system of lethally irradiated recipients (3). Thus, apoptosis appears to be an important mechanism for regulating HSC numbers, but whether one or more anti-apoptotic BCL-2 family members is required for this homeostatic control has not been established (4).

Members of the BCL-2 family of proteins are critical regulators of cell death. Proapoptotic BH3 domain–only members of the BCL-2 family respond to death signals and activate the multidomain death effector proteins BAX and BAK, which constitute an obligate gateway to the intrinsic death signaling pathway that operates through the mitochondria and the endoplasmic reticulum (5, 6). Conversely, anti-apoptotic BCL-2 family members bind and sequester BH3-only molecules, thus preventing activation of BAX and BAK (7). Individual BCL-2 family members have specific roles during hematopoiesis. Bcl-2-deficient mice do not have overt problems during early hematopoiesis, but over time peripheral lymphocytes are vulnerable to apoptosis, especially after activating stimuli (8). Bcl-XL-deficient chimeras created by complementation of Rag2-deficient blastocysts generate lymphocyte progenitors; however, the survival of CD4+CD8+ thymocytes is compromised (9, 10). BCL-XL is also indispensable for the late maturation of both primitive and definitive erythroid progenitors (11). Mcl-1 deficiency results in peri-implantation lethality of mouse embryos (12). Conditional deletion of Mcl-1 during early lymphocyte differentiation increases apoptosis and arrests development at the cytokine-dependent pro-B cell and double-negative T cell stages (13). Furthermore, induced deletion of Mcl-1 results in loss of mature B and T lymphocytes, demonstrating a role for MCL-1 in the maintenance of the mature immune system (13).

MCL-1 is highly expressed in HSC, and its expression levels decreased in further differentiated progenitor populations such as the common myeloid progenitor (CMP) (14) and the common lymphoid progenitor (CLP) (15) as assessed by microarray analysis (16). To confirm this expression profile, we used real-time quantitative polymerase chain reaction (PCR) to detect the relative amounts of Mcl-1 mRNA in purified progenitor cells (13, 17). Amounts of Mcl-1 mRNA were higher in HSCs than in CLPs or CMPs. Megakaryocyte erythroid progenitors (MEPs) and granulocyte monocyte progenitors (GMPs) had still lower amounts (Fig. 1A).

Fig. 1.

Fatal failure of BM after deletion of Mcl-1. (A) Amount of Mcl-1 mRNA in wild-type BM progenitor populations purified by flow cytometry. The mean ± the standard error of the mean (SEM) is presented of triplicate assays in which Mcl-1 mRNA expression units were normalized to those of HPRT. (B) Kaplan-Meier survival graph for cohorts of 10 mice injected with three intraperitoneal doses of pI-pC (400 μg perdose). Log-rank test statistical analyses indicate that P = 0.0003. (C) Histological analysis of femurs from (left) MxCref/wt and (right) MxCref/null mice 11 days after three doses of pI-pC. Sections were stained with hematoxylin and eosin. (D) A total of 106 cell surface–marked (Ly5.2+) BM cells from MxCref/wt or MxCref/null mice were adoptively transferred along with 105 wild-type (Ly5.1+) BM cells into lethally irradiated (950 Rad) Ly5.1+ congenic recipients. After 5 weeks of engraftment, chimeric mice were treated with three doses of pI-pC, and the percentage of peripheral blood cells expressing Ly5.2 or Ly5.1 was determined by flow cytometry. The mean percentage of Ly5.2+ blood cells from three chimeric mice ± SEM is presented for each time point.

To address whether MCL-1 is required for the maintenance of hematopoietic cells we bred conditional Mcl-1flox(f)/f mice with mice expressing the Cre recombinase under the control of the endogenous Mx1 locus, which can be transiently activated in response to type-I interferon (IFN) (13, 18). We induced in vivo deletion by intraperitoneal administration of polyinocinic-polycytidylic acid (pI-pC) to MxCreMcl-1f/null mice (13, 18). Within 12 to 21 days after treatment, the majority of MxCreMcl-1f/null mice, but not MxCreMcl-1f/wt control mice, began to appear moribund and were killed (Fig. 1B). At sacrifice, MxCreMcl-1f/null mice were extremely anemic and had severely reduced numbers of bone marrow (BM) cells (Fig. 1C).

MxCre mice can induce gene deletion in various IFN-responsive tissues in response to pI-pC (18). To examine whether the effect of deletion was from effects on BM cells or from deletion in nonhematopoietic cells, we generated BM chimeric animals. Experimental BM expressing the Ly5.2 cell surface marker from MxCreMcl-1f/wt or MxCreMcl-1f/null BM was adoptively transferred with congenic wild-type carrier BM expressing Ly5.1 cell surface marker into lethally irradiated Ly5.1 congenic animals and allowed to engraft for 5 weeks. Animals were treated with pI-pC, and blood was monitored for Ly5.2+ cells in the periphery. Whereas the mice reconstituted with MxCreMcl-1f/wt BM maintained greater than 85% Ly5.2+ peripheral blood cells, the peripheral blood from MxCreMcl-1f/null reconstituted mice lost Ly5.2+ cells from the periphery and the number of Ly5.1+ carrier BM cells increased over 4 weeks (Fig. 1D). These data demonstrate that deletion of Mcl-1 from chimeric animals results in the loss of BM in a cell-autonomous manner.

To determine whether Mcl-1-deleted cells give rise to colonies when cultured in vitro, we harvested BM from MxCreMcl-1f/wt or MxCreMcl-1f/null mice 3 days after pI-pC treatment. After 7 days, colony numbers and morphology were scored, revealing that BM isolated from the treated MxCreMcl-1f/null mice generated fewer colonies than did BM from littermate controls (Fig. 2A) (19). Genomic DNA was isolated from the resulting colonies, and PCR assays were used to distinguish between the Mcl-1f, Mcl-1deleted, and Mcl-1null alleles (13). The deleted allele, Mcl-1deleted, was readily detected by PCR in the isolated BM 3 days after pI-pC treatment (day 0) (Fig. 2B). However, after 7 days of in vitro culture, only Mcl-1f and Mcl-1null alleles were detectable in the hematopoietic colonies (Fig. 2B). This indicates that BM progenitors lacking both copies of Mcl-1 are not viable and that the reduced numbers of resultant colonies are the result of BM progenitors that escaped the Cre-mediated deletion.

Fig. 2.

Loss of BM progenitor populations including hematopoietic stem cells after Mcl-1 deletion. (A) Recovered colony number from cultured BM from mice (Mcl-1f/wt or MxCref/null) harvested 3 days after one intraperitoneal dose of pI-pC. Data are presented as mean number of colonies from triplicate performed assays ± SEM counted 7 days after plating of 10,000 BM cells. (B) Genomic DNA from cultured BM [as in (A)] was analyzed by PCR to determine the representation of Mcl-1wt, Mcl-1flox, Mcl-1deleted, and Mcl-1null alleles. The day 0 BM samples are from mice injected with pI-pC (3 days before sampling), and the day 7 DNA are from cells harvested from the colonyforming assay. (C) The Sca-1 and c-Kit profile of mature hematopoietic lineage-negative BM (negative for B220, Ter119, CD19, CD3, CD4, CD8, Gr1, Mac-1, and CD127). Numbers indicate the percentage of lineage-negative BM for HSCs (c-Kit+Sca-1+, right box) and early hematopoietic progenitors (c-Kit+Sca-1, left box) subsets as determined by flow cytometry from MxCref/wt (left plot) and MxCref/null (right plot) mice 7 days after one dose of pI-pC. Data are representative of more than five mice analyzed per genotype. (D) Number of BM cells from subsets (mean ± SEM) for 5 mice from each Mcl-1f/wt, MxCref/wt, Mcl-1f/null, or MxCre f/null genotype at 8 weeks of age.

To examine the hematopoietic populations from MxCreMcl-1f/wt and MxCreMcl-1f/null animals, we analyzed BM 7 days after a single intraperitoneal dose of pI-pC. The total number of BM cells present in the MxCreMcl-1f/null mice was lower compared with that in pI-pC-treated littermate controls. When the HSC (mature hematopoietic lineagec-Kit+Sca-1+) and progenitor populations (lineagec-Kit+Sca-1)(14) were analyzed by flow cytometry, both populations were depleted in BM from treated MxCreMcl-1f/null animals (Fig. 2C) (20). Furthermore, the number of BM cells in mice lacking Mcl-1 was 10% of that in control animals 7 days after pI-pC-induced deletion (Fig. 2D). This included large decreases in the HSC and other progenitor populations; however, other more differentiated progenitor populations such as erythroid progenitors (Ter119+) were more subtly reduced (Fig. 2D).

The liver is sensitive to MCL-1 deletion by the MxCre system; therefore we examined Mcl-1 deletion in this tissue (18). In liver, MCL-1 was undetectable by protein immunoblotting after pI-pC-mediated deletion (Fig. 3A) (21). However, histological analysis of liver sections from MxCreMcl-1f/wt and MxCreMcl-1f/null mice 14 days after pI-pC treatment did not reveal any abnormalities, suggesting that loss of MCL-1 expression does not lead to apoptosis of this tissue (22). To test this possibility, we adoptively transferred Ly5.1+ wild-type BM into sublethally irradiated Ly5.2+ MxCreMcl-1f/wt or MxCreMcl-1f/null mice. After 4 weeks of engraftment, the mice were treated with pI-pC, and cells were analyzed by flow cytometry. The Ly5.2+ peripheral lymphocytes from MxCreMcl-1f/null mice were lost within 2 weeks (Fig. 3B); however, the Ly5.1+ wild-type BM promoted the survival of the chimeric mice for more than 14 weeks after MxCre-mediated deletion. Liver lysates from pI-pC-treated MxCreMcl-1f/null chimeric mice contained no detectable MCL-1 expression (Fig. 3C). Thus, deletion of Mcl-1 in the liver is efficient, but nonhematopoietic effects of deletion do not appear to be responsible for the failure of the animals to survive.

Fig. 3.

Lethality of Mcl-1-deletion is BM-specific. (A) MCL-1 protein expression as determined by immunoblot of liver lysates from Mcl-1f/null, or MxCre f/null, Mcl-1f/wt, MxCref/wt mice 11 days after three doses of pI-pC. (B) Ly5.2+ Mcl-1f/wt or MxCref/null mice were sublethally irradiated (400 Rad) and adoptively transferred with 106 Ly5.1+ congenic wild-type BM cells. After 5 weeks of engraftment, three doses of pI-pC were administered to the chimeric mice, and weekly blood analysis was performed by flow cytometry for Ly5.2 or Ly5.1 expression. The average percentage of peripheral blood cells that were host-derived(expressed Ly5.2+) ± SEM was determined by flow cytometry for each recipient genotype. (C) Immunoblot analysis of liver MCL-1 protein expression from BM chimeric animals [Ly5.1+ wild-type BM into Ly5.2+ Mcl-1f/wt or MxCref/null mice, as in (B) 14 weeks after 3 doses of pI-pC].

Expression of MCL-1 is controlled by growth factor signaling pathways. Both mature lymphocytes and lymphoid progenitors increase expression of MCL-1 in response to interleukin (IL)-7 signaling (13). Stem cell factor (SCF) induces expression of MCL-1 in a human BM-derived cell line (23). To determine whether Mcl-1 is expressed in response to growth factors in HSCs, we used real-time PCR. The amount of Mcl-1 mRNA was greater 30 min after exposure of purified HSCs to SCF. IL-6 had a smaller effect, whereas culture with IL-11 did not induce expression (Fig. 4A) (24).

Fig. 4.

Hematopoietic growth factors induce Mcl-1 expression and require Mcl-1 to mediate progenitor survival. (A) Mean ± SEM of triplicate assays for Mcl-1 mRNA normalized to amount of HPRT mRNA for purified, wild-type HSCs after 30, 120, or 360 min of exposure to cytokine growth factors (20 ng/ml). (B) Percentage of cells undergoing apoptosis in cultured FACS-purified BM progenitor populations (HSC, CMP, CLP, and GMP) from Mcl-1f/f or wild-type mice 48 hours after transduction with MSCV-Cre-IRES-EGFP to induce deletion of Mcl-1. Each progenitor population was cultured in growth factors (Materials and Methods), and the percentage of cells undergoing apoptosis is presented for Cre+ (EGFP-expressing) or Cre (EGFP-negative) populations stained with annexin-V to detect phosphatidylserine exposed on the surface of dying cells. Data are representative of two independently performed experiments.

To assess whether Mcl-1 is required for the survival of cultured BM progenitor populations exposed to growth factors, we used retroviral transduction of Cre into purified BM progenitor populations from Mcl-1f/f or wild-type mice (25). The purified BM progenitor populations (HSC, CMP, CLP, and GMP) were cultured in appropriate growth factors (26). By 48 hours after retroviral transduction, more than 90% of Mcl-1f/f Creexpressing [enhanced green fluorescent protein positive (EGFP+)] progenitor cells (HSCs, CMP, and CLPs) were apoptotic (Fig. 4B). Expression of Cre in wild-type BM progenitor populations did not induce apoptosis (Fig. 4B). Therefore, survival of BM progenitors in vitro requires the expression of Mcl-1 induced by early-acting cytokine signals.

Although previous studies had implicated anti-apoptotic BCL-2 family members in regulating the homeostasis of hematopoietic progenitors (3), our studies indicate that a single anti-apoptotic BCL-2 family member, MCL-1, is essential for promoting the survival of HSC and other hematopoietic progenitors.

References and Notes

Stay Connected to Science

Navigate This Article