|Home | About | Journals | Submit | Contact Us | Français|
Self-renewal is a defining characteristic of stem cells, however the molecular pathways underlying its regulation are poorly understood. Here we demonstrate that conditional inactivation of the Pbx1 proto-oncogene in the hematopoietic compartment results in a progressive loss of long-term hematopoietic stem cells (LT-HSCs) that is associated with concomitant reduction in their quiescence, leading to a defect in the maintenance of self-renewal as assessed by serial transplantation. Transcriptional profiling revealed that multiple stem cell maintenance factors are perturbed in Pbx1-deficient LT-HSCs, which prematurely express a large subset of genes, including cell cycle regulators, normally expressed in non-self-renewing multipotent progenitors. A significant proportion of Pbx1-dependent genes are associated with the Tgf-b pathway, which serves a major role in maintaining HSC quiescence. Prospectively isolated, Pbx1-deficient LT-HSCs display altered transcriptional responses to Tgf-b stimulation in vitro, suggesting a possible mechanism through which Pbx1 maintenance of stem cell quiescence may in part be achieved.
Hematopoietic stem cells (HSCs) are distinguished by their inherent capacity to self-renew and differentiate into multiple blood cell lineages throughout the life of the host. These crucial defining attributes of HSCs must be tightly regulated and intricately balanced in order to maintain homeostasis of the lympho-hematopoietic system and to prevent the emergence of blood cell malignancies.
A number of molecules have been implicated in the regulation of HSC self-renewal including cell-intrinsic transcription factors (Tel/Etv6, Gfi1, HoxB4, c-Myc, Zfx, FoxO1/3/4), signal transducers (Pten), and epigenetic regulators (Bmi-1, Ezh2), as well as extrinsic (niche-induced) factors, such as Tek/Angpt1 and Tgf-b signaling (Akala and Clarke, 2006; Galan-Caridad et al., 2007; Kamminga et al., 2006; Tothova et al., 2007). HSCs may also undergo various fates that prevent further self-renewal, including differentiation into more committed progenitor cells (inhibited by HoxB4, promoted by c-Myc), apoptosis (inhibited by Zfx and FoxO) or senescence (inhibited by Bmi1) (reviewed by Cellot and Sauvageau, 2007). To maintain a steady state pool of self-renewing cells and prevent exhaustion, HSCs must limit their cell divisions, and return to G0 until the next self-renewing cycle is required. At any one time only a small proportion of the LT-HSC population is actually in cycle (Venezia et al., 2004), and regulation of the ability to enter and exit quiescence is the most critical requirement for long-term maintenance of the stem cell pool.
Pbx1 is a TALE class homeodomain transcription factor that critically regulates numerous embryonic processes including morphologic patterning, organogenesis and hematopoiesis (DiMartino et al., 2001; Kim et al., 2002; Manley et al., 2004; Schnabel et al., 2003a; Schnabel et al., 2003b; Selleri et al., 2001). It is a component of hetero-oligomeric protein complexes that regulate developmental gene expression in part through modulation of Hox protein DNA-binding and context-dependent transcriptional effector functions (Saleh et al., 2000). As a complex with other TALE class proteins, Pbx1 may mark specific genes for activation by penetration of repressive chromatin and recruitment of coactivators (Berkes et al., 2004). As a major global developmental regulator, Pbx1 has been implicated in promoting progenitor cell proliferation in multiple tissues (Moens and Selleri, 2006). Its function in adult tissue stem cells, however, is unknown, although a potential role is suggested by the observation that Pbx1 is preferentially expressed in LT-HSCs compared to more mature ST-HSCs and multi-potent progenitor cells (MPPs) in the hematopoietic system (Forsberg et al., 2005; Kiel et al., 2005).
In this report, we employed a genetic approach to conditionally inactivate Pbx1 in the mouse hematopoietic system in vivo. In contrast to its role in promoting progenitor expansion, we report that Pbx1 positively regulates HSC quiescence. Our studies demonstrate that Pbx1 is a novel regulator of HSC self-renewal that contributes to maintenance of HSC quiescence through multiple pathways including the transcriptional program induced in response to Tgf-b signaling.
Pbx1-conditional ko mice were employed to study the role of Pbx1 in the mouse adult hematopoietic system (Fig. S1A). Homozygous Pbx1 floxed mice (Pbx1f/f) were crossed with Pbx1+/− mice (Selleri et al., 2001) expressing Cre under the control of the Tie2 promoter (Tie2Cre+.Pbx1+/− mice), which specifically drives Cre expression in endothelial cells and HSCs (Chang et al., 2004; Kisanuki et al., 2001). PCR analysis of the resulting Tie2Cre+.Pbx1f/− mice (referred to hereafter as “mutant” mice) showed specific and complete deletion of Pbx1 in different hematopoietic tissues, including peripheral blood (PB), bone marrow (BM), spleen, thymus, and lymph nodes (Fig. S1B), as expected from efficient excision of Pbx1 in HSCs.
The hematopoietic organs appeared normal at birth, indicating proper embryonic development, but displayed major defects in young mutant mice. The thymus and spleen were dramatically reduced in size and showed pronounced histologic disorganization (Fig. 1A and not shown) compared to controls (Tie2Cre+.Pbx1+/f, Tie2Cre−.Pbx1+/f, or Tie2Cre−..Pbx1−/f mice, which did not differ in any of the phenotypes to be described and, hence, will not be further distinguished). Differences in the absolute numbers of CD45+ cells between mutant mice and controls (4-fold and 7-fold in spleen and thymus, respectively) indicated that the reduced organ sizes were mainly due to a lower content of hematopoietic cells (Fig. 1A). In the thymus, FACS analysis showed that all T-lineage cellular subpopulations were reduced, starting from the most immature DN1 stage (not shown), suggesting that the defect arose at a very early stage of T-lymphocyte development. The BM of mutant mice compared to controls was hypocellular (Fig. 1B), mainly accounted for by decreased B lineage cells as revealed by FACS analysis (Fig. 1C). The reduction in the B cell compartment was highly significant starting from the pro-B (B220+CD43+IgM−) cell stage with pre-B (B220+CD43−IgM−) cells constituting the population numerically most reduced. Thus, Pbx1 is necessary to maintain the postnatal hematopoietic compartment.
FACS analyses revealed marked reduction of common lymphoid progenitors (CLPs) (Fig. 2A and Fig. S2A), mild reductions in the total numbers of common myeloid progenitors (CMP, Fig. 2B) and to a lesser extent of granulocyte-monocyte progenitors (GMP), and reduced frequency of colony-forming cells in semi-solid culture (CFC, Fig. S2B). More importantly, there was a marked reduction of the HSC-enriched Lin−.c-Kit+Sca-1+ (KSL) cell population (Fig. 2C). Analysis of the LT-HSCs within this population (Kiel et al., 2005) revealed a considerable reduction in their numbers indicating a defect at the earliest stage of adult hematopoietic development (Fig. 2D), suggesting that the reduction of different progenitor cell populations predominantly originated from a reduced number of LT-HSCs. The described stem and progenitor abnormalities were progressive as the mice aged, being undetectable at birth (Fig. S2C), variable at 2-weeks of age (in 3 out of 5 mutants analyzed), and prominent from 3 weeks of age onwards.
To rule out the possibility that a vascular niche defect to which HSCs were exposed during development could be responsible for the observed hematopoietic abnormalities, Pbx1f/f mice were crossed with MxCre+.Pbx1+/− mice. Highly efficient Pbx1 deletion was induced with poly(I:C) in young MxCre+.Pbx1f/f (or −/f) mutant mice, and confirmed by qRT-PCR 3 to 4 weeks later (Fig. S3A). Defects in the stem cell compartment of MxCre+.Pbx1f/f (or −/f) mice were identical to those observed with Tie2Cre+.Pbx1−/f mice (Fig. 2E and S3B), and similar defects were also observed in CLPs, CMPs and B cells (Fig. S3C, D and not shown). To further exclude a primary BM microenvironmental cause underlying the HSC defect, BM cells from MxCre+.Pbx1f/f mice were transplanted into wt recipients, and deletion of Pbx1 in donor cells was induced with poly(I:C) after BM reconstitution. LT-HSCs were reduced as a result of Pbx1 excision exclusively in the hematopoietic compartment (Fig 2F). Furthermore, transplant of wt BM cells into MxCre+.Pbx1f/f (or −/f) mice (Fig S3E), showed that donor-derived HSCs were not reduced upon induction of Pbx1 deletion in mutant recipient mice, unequivocally demonstrating that the absence of Pbx1 from a non-hematopoietic compartment does not contribute to the HSC phenotype.
The underlying cause for reduced numbers of Pbx1-null stem and progenitor cells was not a consequence of increased levels of in vivo apoptosis, since annexin V staining did not reveal increased retention in KSL cells (not shown) or LT-HSCs from mutant mice (Fig. S2D), or in myeloid or lymphoid progenitors (not shown). Conversely, assessment of cellular proliferation by BrdU pulse labeling experiments revealed significantly higher BrdU incorporation in all KSL cell subpopulations, starting from LT-HSCs (Fig. 3A). A more detailed analysis of the cell cycle by Hoechst/Pyronin Y staining revealed that far fewer mutant KSL and LT-HSCs were in the G0 phase, with a much higher proportion in SG2M (Fig. 3B). These data suggested that increased proliferation likely contributes to the reduced numbers of LT-HSCs by diminishing the pool of quiescent stem cells.
In competitive repopulation assays, escalating doses of total BM cells from mutant or control mice (Ly5.2) were infused into lethally-irradiated congenic recipients (Ly5.1), in competition with a fixed number of BM cells from Ly5.1/Ly5.2 mice. FACS analysis of blood leukocytes from transplanted animals revealed that BM cells from mutant mice were completely unable to compete with wt cells even when transplanted at a 10-fold higher dose of mutant cells than competitor cells (Fig. 4A). Donor Tie2Cre+.Pbx1−/f cells were undetectable also in the BM (not shown). The inability to compete with wt BM was not simply due to a lower content of mutant stem cells since the same number of total BM cells from mutant or control mice were transplanted, and the difference in the proportion of KSL and of LT-HSCs was much less than 10-fold (Fig. 2C and D, FACS plots). These data demonstrate a profound functional impairment of LT-HSCs in the absence of Pbx1.
To assess if Pbx1-deficient stem cells are intrinsically capable of engraftment and/or differentiation, different doses of BM cells from mutant or control mice were infused into lethally irradiated recipients in the absence of competitor cells. Analysis of PB for cells of donor origin after transplant showed that in the absence of competition efficient engraftment by Pbx1-null BM cells was achieved, but only at high transplanted cell doses (Fig. 4B; only myeloid cell data are shown for simplicity, since their high turnover provides a good measure of stem cell activity). At lower cell doses (2 × 105 or less total BM cells) engraftment was mostly short term (middle and right panels) and less efficient (right panel), in accordance with an insufficient number of functional HSCs transplanted from mutant mice. Once engrafted, multi-lineage differentiation was achieved (Fig. S4), confirming that differentiation and migration were not abrogated by lack of Pbx1. PCR performed on genomic DNA isolated from donor blood leukocytes several weeks post-transplant confirmed complete excision of the floxed Pbx1 allele (Fig. 4C), demonstrating that reconstitution was not mediated by rare un-deleted cells present in the BM of conditional ko mice prior to transplant. Thus, Pbx1-null HSCs can engraft, differentiate, and reconstitute hematopoiesis in lethally irradiated recipients, but only under non-competitive conditions.
Despite engraftment by Pbx1-deficient cells, progressive graft failure was observed long term after transplant (>20 weeks, Fig. 4B), in accordance with a defect at the level of LT-HSCs. Indeed, 20 weeks after transplant with 5×105 Pbx1-null BM cells phenotypically defined LT-HSCs were below the limit of detection in the BM of recipient mice (Fig. 4D). Accordingly, in methylcellulose in vitro colony assays performed at the same time point, CFU-GEMM colonies were not detected after plating of FACS-sorted (Ly5.1) donor BM cells from mice transplanted with Pbx1-null cells (Fig. 4E, top), nor were any colonies derived from high proliferative potential progenitors (Fig. 4E, bottom). Moreover, similar to the BM of mutant mice prior to transplant (Fig. S2), the number of total colonies obtained was significantly reduced, mainly due to a decreased number of BFU-E and CFU-GM, indicating a defect at the level of progenitors in addition to LT-HSCs, as predicted by the lower proportion and absolute number of CMP in the conditional ko mice (Fig. 2B). In the BM of recipients transplanted with cells from Pbx1 mutants, the proportion of B cells derived from the donor was also considerably lower (Fig. 4F). Due to their advanced age at the time of analysis, transplanted mice exhibited small thymi and a very low proportion of CLPs independent of the genotypes of the donor mice. However, a reduced proportion of B cells in the BM (but not in the blood) was in accordance with a progressive CLP or early B-cell progenitor defect. Hence, most of the features of the conditional ko mice (low numbers of stem, progenitor, and B cells in the BM) were reproduced in the transplanted mice, and thus constitute intrinsic stem and progenitor cell defects.
The progressive graft failure and exhaustion of mutant LT-HSCs (KSL, CD150+CD48−CD41− or KSL, CD34−CD135−cells) (Mansson et al., 2007; Rossi et al., 2005) in the primary transplant recipients (Fig. 4D) suggested a defect in self-renewal. Therefore, secondary transplant experiments were performed to confirm the reduction of functional Pbx1-null LT-HSCs in the BM of primary transplant recipients. Twenty weeks after primary transplant, total BM cells from mice non-competitively reconstituted with 5 × 105 BM cells/mouse were harvested and sorted according to the donor Ly5.1 marker. Sorted donor cells were then transplanted into secondary recipients at escalating doses without competitor cells. Pbx1-deficient cells were unable to engraft secondary recipients (Fig. 5A). Analysis of the blood of secondary recipients before death revealed low but detectable chimerism (Fig. 5B), indicating that the few HSCs originally derived from mutant mice homed in the BM of secondary recipients, but failed to fully reconstitute. Therefore, phenotypic and functional LT-HSCs rapidly declined to undetectable levels following sequential transplants of BM cells from mutant mice to primary and secondary recipients (summarized in Fig. 5C).
Hematopoietic reconstitution was also assessed in situ in primary mutant mice by monitoring survival upon sequential injections of the cell cycle-dependent myelotoxic agent 5-fluorouracil (5-FU) (Cheng et al., 2000), an assay that does not depend on engraftment. The requirement for a relatively long follow-up prompted the use of the MxCre conditional ko model, since Tie2Cre+.Pbx1−/f mice have a short life-span. Five weeks after the last poly(I:C) injection, mutant and control mice received 5-FU at 7 day intervals to kill proliferating cells including hematopoietic progenitors, thus providing a stimulus for HSCs to proliferate in order to replenish the hematopoietic system. Pbx1-deficient mice succumbed after the second 5-FU injection, whereas exhaustion of the HSC pool occurred significantly later in control mice (Fig. 5D). Mutant mice died of hematopoietic failure, with pancytopenia in the blood and BM (not shown).
Taken together, our data demonstrate that Pbx1 maintains the hematopoietic system by regulating LT-HSC self-renewal.
Global gene expression profiling analyses were performed to identify Pbx1 dependent genes and pathways involved in the regulation of HSC self-renewal. For these studies, Pbx1 deletion was induced in MxCre+.Pbx1f/f mutant mice, which allowed the use of older animals to obtain a much higher yield of LT-HSCs and also the distinction between LT- and ST-HSCs in adult mice based on the expression of CD34 and CD135 markers in KSL cells. Cells were prospectively sorted from BM harvested 4 weeks after the last injection of poly(I:C).
Analysis of gene expression data from mutant and control LT-HSCs revealed 241 non-redundant differentially expressed genes, of which 61 were up-regulated and 180 down-regulated in Pbx1 mutants (Fig. 6A and Table S1). Microarray results were confirmed by qRT-PCR for a small subset of the genes (Table S1). Classification based on functional annotation revealed that a large proportion of up-regulated genes in Pbx1 mutant LT-HSCs belonged to cell cycle ontogeny groups (Fig 6B), consistent with the observed increase in cycling stem cells leading to loss of their self-renewal potential. Notably, approximately 8% of the down-regulated genes belonged to the Tgf-b signaling pathway (Fig 6B and Table S2), which has been previously implicated in maintenance of HSC quiescence. Other differentially regulated transcripts included genes known to be involved in self-renewal of HSCs (Hlf and Meis1), or to have roles in differentiation, the ubiquitin pathway, cell adhesion, chromatin-modification, and various signaling pathways including the MAPK pathway, which has been implicated in HSC maintenance (Ito et al., 2006). Thus, the transcriptional changes suggest that multiple stem cell-maintenance mechanisms are perturbed in the absence of Pbx1.
The lack of long-term self-renewal, but preserved reconstitution potential, as well as increased proliferation, suggested that phenotypically defined Pbx1-deficient LT-HSCs display functional features of ST-HSCs and/or MPPs. To test this hypothesis, the Pbx1 mutant LT-HSC gene set was interrogated for genes whose expression normally distinguishes LT- from ST-HSCs. To generate the latter gene set, LT- and ST-HSCs were prospectively sorted from the BM of wt mice and their global gene expression profiles were determined by microarray (Fig. 6C). Bioinformatic analysis revealed a gene set of 2,468 non-redundant transcripts that were differentially expressed between normal LT- and ST-HSCs sub-populations (Table S3), a higher number than previously reported using different sorting and analysis strategies (Forsberg et al., 2005; Zhong et al., 2005). This gene set (LT vs ST-HSCs) was then compared with the Pbx1-dependent LT-HSCs gene set. Among the transcripts down-regulated in mutant LT-HSCs, 17% were also down-regulated in normal ST-HSCs compared to LT-HSCs. Furthermore, among the transcripts up-regulated in mutant LT-HSCs, a substantial proportion (37%) was also up-regulated in normal ST-HSCs. Thus, many of the genes aberrantly expressed in phenotypically defined LT-HSCs from Pbx1 mutant mice are normally expressed in ST-HSCs, but not wt LT-HSCs. The mutant LT-HSC gene signature was also compared with a previously published gene set differentially expressed in MPPs versus LT-HSCs (Kiel et al., 2005). A considerable overlap was observed between Pbx1 mutant LT-HSC and normal MPP gene expression profiles (Fig. 6D), further confirming that in the absence of Pbx1, LT-HSCs prematurely express a transcriptional sub-program shared with downstream ST-HSCs and MPPs.
The functional and expression profiling studies suggested that Pbx1 regulates stem cell quiescence and exit from the LT self-renewing compartment. To further test this hypothesis, Gene Set Enrichment Analysis (GSEA) was employed to compare the Pbx1 regulated dataset with curated gene sets derived from diverse published experiments (Subramanian et al., 2005). To query the curated gene sets, the Pbx1 regulated dataset was reduced to a rank ordered list of genes expressed by mutant versus wt LT-HSCs (more than 9,684 non-redundant transcripts). In accordance with our hypothesis, GSEA revealed a strong correlation of genes down-regulated in the absence of Pbx1 with gene sets that identify HSCs (Table 1A and Fig. 6E). Conversely, the genes up-regulated in the absence of Pbx1 strongly correlated with gene sets typical of progenitors (Table 1B and Fig. 6E), as well as with cell cycle related gene sets. GSEA also revealed enrichment for Myc-activated genes and for genes involved in DNA damage control, possibly related to the perturbed cell cycle status of Pbx1 mutant LT-HSCs. Of note, genes down-regulated in the absence of Pbx1 were also enriched for B-cell gene sets, indicating that the CLPs and B-cell abnormalities observed in mutant mice most likely originated from a LT-HSC defect in lymphoid priming.
GSEA also confirmed a prominent perturbation of the Tgf-b pathway in Pbx1 mutant LT-HSCs as evidenced by significant enrichment of gene sets associated with Tgf-b pathway alterations in various cell types (Table 1A), including human cord blood CD34+ cells exposed to TGF-b in vitro (Fig. 6E). Therefore, the integrity of the Tgf-b signaling pathway was interrogated in prospectively purified mutant and wt LT-HSCs by assessing their transcriptional responses to Tgf-b stimulation in vitro. In contrast to wt LT-HSCs, mutant LT-HSCs did not up-regulate expression of several downstream transcripts in response to Tgf-b stimulation (Fig. 6F) including Smad7, Klf4 and Ski, as well as Cdkn1c (p57), recently implicated as an important regulator of HSC quiescence (Scandura et al., 2004; Umemoto et al., 2005; Yamazaki et al., 2006). Furthermore, unlike control LT-HSCs, Ccnd2 and Ccna2 were not down-regulated by Tgf-b simulation in mutant LT-HSCs, consistent with their in vivo cycling activity. These results together with the global transcriptional changes suggest that Pbx1 controls stem cell quiescence and self-renewal at least in part by affecting the response to Tgf-b.
In this study we demonstrate that the Pbx1 proto-oncogene and global developmental regulator serves a critical role within the hematopoietic system to support the self-renewal of adult HSCs as one of few known factors involved in maintaining their quiescence. Transcriptional profiling suggests that multiple stem cell maintenance programs are perturbed in the absence of Pbx1, including genes associated with the response to Tgf-b signaling. This suggests potential mechanisms for facilitating loss of quiescence, inappropriate cell cycle entry, and initiation of a transcriptional program resembling that of more mature multipotent progenitors. As a consequence, LT-HSCs lose their major defining characteristic, i.e. self-renewal, and progressively exhaust themselves in the absence of Pbx1.
Among the multiple stem cell maintenance factors that are altered in Pbx1-deficient LT-HSCs, a significant proportion are linked with the Tgf-b pathway, which is increasingly implicated in regulation of HSC quiescence. Tgf-b has recently been suggested as a possible BM niche signal necessary to induce LT-HSC “hibernation” through inhibition of lipid raft clustering in response to cytokine stimulation (Yamazaki et al., 2007). The CDKI p57 appears to be responsible for regulating this quiescent HSC state (Yamazaki et al., 2007). Consistent with these observations, cell cycle arrest of human cord blood CD34+ cells in response to Tgf-b requires p57 (Scandura et al., 2004). Moreover, p57 is responsible for cell cycle arrest of mouse KSL side population (SP) cells (Umemoto et al., 2005), suggesting that it is one of the main contributors to maintenance of LT-HSC quiescence. We observed higher expression levels of Cdkn1c in LT-HSCs compared to ST-HSCs (although the difference did not meet our stringent statistical requirements for inclusion in Table S3). This is consistent with previous studies performed with different methods and mouse strains showing that Cdkn1c is preferentially expressed in LT-HSCs compared to ST-HSCs and MPPs (Kiel et al., 2005), and at higher baseline levels in mouse CD34− KSL cells (highly enriched for LT-HSCs) than other CDKIs such as p21 or p27 (Yamazaki et al., 2006). Although the latter are also implicated in HSC or progenitor self-renewal, we did not observe their expression levels to be altered in Pbx1-deficient LT-HSCs in contrast to the marked reduction in p57.
In support of a possible role for Pbx1 on a pathway that regulates the balance between stem cell quiescence/hibernation and cell cycle entry, the Cdkn1c gene was not induced by Tgf-b stimulation of prospectively isolated Pbx1 mutant LT-HSCs. Although these data do not rule out a secondary role for Cdkn1c, it is one of very few in our Pbx1-dependent gene set to contain a Pbx1 heterodimer binding site in proximity to its promoter. Furthermore, Pbx1/Prep transcriptional complexes have been shown to interact with Smads (Bailey et al., 2004), the nuclear effectors of Tgf-b signaling, and Pbx1 transcriptional complexes regulate expression of the p21 gene in a myelomonocytic cell line (Bromleigh and Freedman, 2000). In epithelial cells, a FoxO/Smad complex is implicated in Tgf-b up regulation of p15ink4a, another CDKI (Gomis et al., 2006). Thus, there are precedents for regulation of CDKI genes by Pbx and/or Smad proteins. However, perturbed expression of Smad7, a Tgf-b pathway regulator that has been implicated in HSC self-renewal (Blank et al., 2006), was also observed in Pbx1-deficient HSCs in vitro and in vivo raising the possibility of a more global compromise of the response to Tgf-b. Since our data were obtained 4 weeks after Pbx1 deletion, it is theoretically possible that they reflect in part compensatory changes in the proliferation activity or differentiation state of the stem cell compartment. Further investigations are necessary to establish a causal relationship between the hematopoietic defects caused by Pbx1 loss and perturbed regulation of Tgf-b pathway-associated genes implicated in HSC self-renewal.
HSC reductions associated with the absence of Pbx1 were only apparent from approximately 3 weeks post-birth despite the fact that expression of the Tie2Cre transgene initiates in definitive HSCs with their emergence in the aorta-gonad-mesonephros and yolk sac during embryonic development. The postnatal onset of the Pbx1 phenotype likely reflects differences in the proliferation states of embryonic/fetal and adult BM HSCs. During development, HSCs undergo massive expansion while preserving their self-renewal capacity, whereas HSCs in the post-natal BM are mostly quiescent starting from 3–4 weeks of age (Bowie et al., 2006), and their ability to maintain quiescence correlates with the maintenance of self-renewal and engraftment potential (Passegue et al., 2005). Thus, onset of the Pbx1-deficient HSC phenotype in Tie2Cre.Pbx1 mice correlates with the post-natal requirement for HSC quiescence for maintenance of self-renewal.
A role for Pbx1 in maintaining stem cell quiescence significantly broadens its contributions beyond previous studies, which have consistently implicated it in promoting progenitor cell expansion during development of multiple organs (Brendolan et al., 2005; Kim et al., 2002; Manley et al., 2004; Selleri et al., 2001), including the hematopoietic system (DiMartino et al., 2001). Thus, Pbx1 appears to have divergent but functionally complementary roles in the maintenance of HSC quiescence versus promotion of progenitor proliferation, and our current results raise the possibility that other Pbx1 deficient phenotypes may in part reflect its impact on tissue-specific stem cells. The potential complexity of its contributions in the hematopoietic progenitor compartment was suggested by earlier studies demonstrating that HoxB4-mediated HSC in vitro expansion is further enhanced by concurrent reduction of Pbx1 expression (Krosl et al., 2003), which appeared inconsistent with its role in promoting progenitor expansion. However, a major function for Pbx1 in maintaining LT-HSC quiescence provides a likely explanation for its antagonistic effects since HoxB4 over-expression enhances HSC expansion without affecting long-term self-renewal potential. The divergent roles for Pbx1 in HSCs versus progenitors may reflect possible context-dependent contributions to the expression of CDKI and other cell cycle regulatory genes.
In addition to a stem cell defect, a marked reduction was observed in the BM pro-and pre-B cell compartments, as well as a striking reduction in CLPs, suggesting a role for Pbx1 at a critical stage of lymphoid development where acute leukemia likely originates, and confirming previous results from our laboratory obtained with a different experimental approach (Sanyal et al., 2007). The extremely low number of CLPs in mutant mice prevented studies of their cell cycle status, although the cycling rates of pro- and pre-B cells were only mildly decreased (not shown), consistent with previous conclusions that Pbx1 may be dispensable from the pro-B stage onward (Sanyal et al., 2007). The striking CLP hypoplasia may reflect a cell intrinsic requirement for Pbx1 in CLPs, or alternatively be the consequence of insufficient CLP generation due to an aberrant differentiation program initiated by Pbx1-null HSCs. In support of the latter, GSEA revealed a defect in lymphoid priming present at the LT-HSC stage. The expression of lymphoid-specific genes in LT-HSCs and the concept of lineage priming in stem cells has been previously reported (Mansson et al., 2007; Rossi et al., 2005). In Pbx1-deficient mice, the lymphoid priming defect does not completely compromise further differentiation capacity, since mature B cells are present in BM, spleen and PB of primary ko mice and transplant recipients. In future studies it will be of interest to further characterize the roles of Pbx1 in HSC lymphoid priming and generation of CLPs, and the potential relevance of these contributions for Pbx1 oncoproteins in B-cell precursor leukemias.
Transgenic Tie2Cre, Mx1Cre, and Pbx1+/− mice have been described (Chang et al., 2004; Kisanuki et al., 2001; Kuhn et al., 1995; Selleri et al., 2001). These strains were inter-crossed to generate Tie2Cre+.Pbx1+/− or MxCre+.Pbx1+/− mice, which were subsequently bred with Pbx1f/f mice (to be described elsewhere) to obtain Tie2Cre+.Pbx1−/f, MxCre+.Pbx1−/f or MxCre+.Pbx1f/f mutants and their littermate controls, fully or partially backcrossed onto a C57BL/6 background. Mutant and control mice were genotyped by PCR. Tie2Cre+.Pbx1−/f were viable, but displayed a short life span and died of unknown causes between 3–7 weeks of age. Congenic Ly5.1 mice, purchased from Jackson Laboratories (Bar Harbor, ME), were maintained in the Stanford animal facility and used as recipients of BM transplantation experiments. Congenic C57BL/6 Ly5.2 mice were crossed with Ly5.1 mice to obtain Ly5.1/Ly5.2 competitor BM cells. All experiments were performed with the approval of and in accordance with Stanford’s Administrative Panel on Laboratory Animal Care.
Tissues were fixed in 10% (weight/volume) formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin.
BM cell suspensions were obtained by flushing of femurs and tibiae in sterile staining buffer (see below). PB was obtained by tail bleeding. Staining of cells for FACS analysis was performed in deficient RPMI (Irvine Scientific) containing 3% fetal calf serum, 1 mM EDTA and 0.01 M HEPES (staining buffer) using the following conjugated monoclonal antibodies (mAbs) obtained from either BD Pharmingen (BD) or eBioscience (San Diego, CA): Mac1/CD11b (M1/70), B220 (RA3–6B2), CD43 (S7), IgMb (AF6–78), TCRβ (H57–597), NK1.1 (PK136), TER119 (TER-119), cKit (2B8), Sca1 (D7 or E13–161.7), IL-7Rα (SB14), CD16/32 (2.4G3), CD34 (RAM34), CD48 (HM48–1), CD41 (MWReg30), Ly5.1 (A20), Ly5.2 (104), CD135 (AF2 10.1). The lineage cocktail included the following Cy7PE-conjugated mAbs: Mac1, Gr1 (RB6–8C5), B220, TER119, CD3 (145–2C12), CD4 (GK1.5), and CD8 (53–6.7). CD150 (TC15–12F12.2) was purchased from BioLegend. Stained cells were analyzed with an LSR1A flow cytometer or FACSAria (BD Biosciences, San Jose, CA). Cell sorts were performed using a FACS Vantage (BD Biosciences). Cell Quest Pro or Diva software (BD) and FlowJo (Tree Star) were used for data acquisition and analysis, respectively. For Annexin V staining, freshly isolated BM cells were first stained with the appropriate mAbs, then washed in binding buffer and incubated with Annexin V FITC (BD).
Mice received a single intraperitoneal injection of 5-bromodeoxyuridine (BrdU) 2 hrs prior to sacrifice. In order to have enough cells for the analysis of LT-HSC, BM cells were obtained by crushing of multiple bones (femurs, tibia, humerous, ulna, shoulder blades, sternum, spine). Progenitors were purified based on the expression of cKit using anti-CD117 magnetic microbeads and an automated cell separator (AutoMACS, Miltenyi Biotec, Auburn, CA), and then subjected to cell surface staining for detection of stem cells. Analysis of BrdU incorporation was performed using the FITC BrdU Flow Kit (BD Pharmingen) according to the manufacturer’s protocol. For H/PY staining, purified cKit+ cells were incubated with cell surface markers to detect HSC and then incubated with HOECHST 33342 (Invitrogen) and Pyronin Y (Sigma-Aldrich) as previously described (Passegue et al., 2005).
Transplantation of BM cells from mutant or control mice (with or without 2 × 105 competitor BM cells), or secondary transplantation of donor-derived BM cells from primary recipients, were performed by retro-orbital injection of lethally irradiated Ly5.1 mice (900 cGy). For secondary transplants, 3 doses of sorted Ly5.2 cells were injected: 1 × 106, 5 × 105 and 2 × 105 (3 mice/group). Data from the 3 different doses were pooled since the outcome did not correlate with the number of injected cells.
Two week old MxCre+.Pbx1f/f (or −/f) or control mice received poly(I:C) injections (InvivoGen, San Diego, CA) every other day for a total of 7 injections. Five weeks after the last injection, mice received 150 mg/kg 5-FU (ICN Pharmaceuticals, Inc., Costa Mesa, CA) ip at 7 day intervals.
BM cells (2 × 104) were seeded into methylcellulose-containing medium (methoCult 3234; Stem Cell Technologies) in the presence of SCF (20 ng/ml), IL-6, IL-3 (10 ng/ml each), and Epo (3 U/ml). Colonies were counted after 10 d of culture.
The significance of differences was determined by two-tailed Student’s t-test.
Two week old MxCre+.Pbx1f/f or MxCre−.Pbx1f/f mice were given 7 poly(I:C) injections every other day and sacrificed 4 weeks after the last injection. BM cells were obtained from multiple bones of individual mice, immunomagnetically selected for c-kit expression, then stained with conjugated mAbs (CD34, CD135, Sca-1, lineage cocktail, cKit) prior to sorting. Cells were maintained on ice when possible through all procedures. RNA from sorted cell populations (3–5,000 LT-HSC, 4–12,000 ST-HSC) was purified using Trizol and subjected to two rounds of amplification with a Two-Cycle Target Labeling and Control Reagent (Affymetrix, Santa Clara). Microarray experiments were performed in the Stanford PAN Facility using Affymetrix 430–2.0 arrays. Arrays were scanned with a Gene Chip Scanner 3000 (Affymetrix) running GCOS 1.1.1 software. Scanned data were exported to dChip software for normalization, statistical analysis and heat mapping. A perfect-match/mismatch (PM/MM) model was used for the calculation of expression values (Li and Wong, 2001). The threshold for considering genes differentially expressed was 1.3 fold-difference with a 90% confidence. Raw data will be available for download from Gene Expression Omnibus (http://ncbi.nlm.nih.gov/geo). The categorization of genes into functional groupings was based on the Gene Ontology (GO) classification system as well as on comprehensive evaluation of the relevant literature for all of the differentially regulated genes. Gene set enrichment analysis (Subramanian et al., 2005) was performed using GSEA v2.0 software available from the Broad Institute (http://www.broad.mit.edu/gsea).
For studies of Tgf-b-dependent transcriptional response, LT-HSC were sorted from poly(I:C)-treated MxCre+Pbx1f/f or control mice as described for the microarray experiments, then incubated for 4 h in StemPro-34SFM medium (GIBCO) with 2 ng/ml TGF-β1 (PeproTech, Rocky Hill, NJ) prior to RNA extraction with a PicoPure RNA isolation Kit (Arcturus, Mountain View, CA) and amplification with a RiboAmp RNA amplification kit (Arcturus). cDNA was prepared using Superscript First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA) and then subjected to real-time PCR using Taqman probes (Applied Biosystems, Foster City, CA).
We thank Maria Ambrus and Cita Nicolas for technical assistance, and Tim C. Somervaille for critically reading the manuscript. F.F. was supported by a fellowship from the American-Italian Cancer Foundation and an ASH Fellow Scholar Award in basic research. We acknowledge support from PHS grants CA42971 and CA90735.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.