HOXA4 Provides Stronger Engraftment Potential to Short-Term Repopulating Cells Than HOXB4

Abstract

Genes of the HOX4 paralog group have been shown to expand hematopoietic stem cells (HSCs). Endogenous expression of HOXA4 is 10-fold higher than HOXB4 in embryonic primitive hematopoietic cells undergoing self-renewal suggesting a more potent capacity of HOXA4 to expand HSC. In this study, we provide evidence by direct competitive bone marrow cultures that HOXA4 and HOXB4 induce self-renewal of primitive hematopoietic cells with identical kinetics. Transplantation assays show that short-term repopulation by HOXA4-overexpressing multilineage progenitors was significantly greater than HOXB4-overexpressing progenitors in vivo, indicating differences in the sensitivity of the cells to external signals. Small array gene expression analysis showed an increase in multiple Notch and Wnt signaling -associated genes, including receptors and ligands, as well as pluripotency genes, for both HOXA4- and HOXB4-overexpressing cells, which was more pronounced for HOXA4, suggesting that both HOX proteins may assert their affects through intrinsic and extrinsic pathways to induce self-renewal of primitive hematopoietic cells. Thus, HOXA4 increases short-term repopulation to higher levels than HOXB4, which may involve Notch signaling.

Introduction

Homeobox (HOX) transcription factors are key regulators of hematopoiesis (reviewed in Alharbi et al. [1]). HOXB4 was the first HOX gene shown to play a role in the expansion of human and mouse hematopoietic stem cells (HSCs) by promoting self-renewal divisions [2,3]. Importantly, HOXB4-overexpressing HSC retained their full differentiation potential, but remained susceptible to external signals, as their numbers did not increase beyond those normally found in mice. The finding that exposure to HOXB4 recombinant protein also enhanced HSC numbers, made HOXB4 an attractive candidate for clinical ex vivo expansion of HSC [4,5].

Our group recently demonstrated that overexpression of HOXA4, which is highly homologous to HOXB4, also led to a net expansion of functional mouse HSC in vitro [6]. This was accompanied with expansion of total bone marrow (BM) cultures by HOXA4 that were up to 100-fold more important than control after 3 weeks of culture [6], and were comparable to earlier reported expansions with HOXB4-overexpressing progenitor cells [7]. The capacity to expand primitive hematopoietic cells appeared to be attributed to all paralog four members as evidenced in an embryonic stem cell overexpression model [8]. Moreover, coculture of human CD34⁺ cells on stromal cells producing either HOXB4 or HOXC4 protein was shown to augment primitive cell numbers with similar magnitude [9]. These observations suggest that functional redundancy within the HOX network not only occurs in developmental programs [10,11], but also in hematopoiesis.

BM transplantation is dependent on the presence of long-term (LT) HSC, which can sustain the production of diverse blood cell types for extended periods (>20 weeks) in myeloablated mice [12]. The repopulation capacity of immediate descendants of LT-HSC, such as short-term HSC and early committed progenitors is limited in time (<20 weeks) and becomes more restricted with differentiation due to decreasing self-renewal potential.

The expression of HOX genes, predominantly the A and B clusters, also decreases with maturation [13–16], suggesting a role for HOX genes in the self-renewal potential of hematopoietic cells. Interestingly, quantitative reverse transcription–polymerase chain reaction (qRT-PCR) analysis showed 10-fold higher expression of HOXA4 than HOXB4 in E14 fetal liver populations enriched for HSC [13]. The fact that fetal liver HSC are undergoing intensive self-renewal at that time of development to establish the HSC reservoir suggests that HOXA4 may also be an important determinant of HSC self-renewal under physiological conditions. These observations indicate that HOXA4 may be a more potent candidate for ex vivo expansion of HSC than HOXB4 in a clinical context.

In this study, using retroviral overexpression, we have directly compared the effect of HOXA4 and HOXB4 on the capacity of primitive hematopoietic cells to expand in vitro and to engraft in vivo.

Materials and Methods

Retroviral construction and transduction

B6SJL mice were intravenously injected with 150 mg/kg of 5-fluorouracil (5-FU; Mayne Pharma (Canada), Inc.) to recruit HSC into cell cycle, as previously reported [6]. After 4 days, BM cells were isolated from these mice and cocultured for 2 days on confluent layers of the GP+E-86 packaging cell line stably producing MSCV-HOXA4-GFP, MSCV-HOXB4-GFP, MSCV-HOXA4-YFP, or MSCV-GFP retrovirus [17] in Dulbecco's modified Eagle's medium (DMEM; Wisent, Inc.) supplemented with 15% heat-inactivated fetal bovine serum (FBS; PAA Laboratories, Inc.), 6 ng/mL interleukin (IL)-3, 10 ng/mL IL-6, 100 ng/mL stem cell factor (SCF), 10⁻⁵ M 2-mercaptoethanol (Mallinckrodt Baker, Inc.), 50 ng/mL gentamicin (Wisent, Inc.), 10 ng/mL ciprofloxacin (Wisent, Inc.), and 6 ng/mL polybrene (TekniScience, Inc.).

Animals and transplantation assays

For transplantation assays, congenic C57BL/6 (CD45.2) and B6SJL (CD45.1) mice (The Jackson Laboratories) were used. Mice were bred and maintained in a specific pathogen-free animal facility of the HMR Research Center. Transduced (HOXA4, HOXB4, HOXA4/B4, or control GFP) BM cells from B6SJL mice treated or not with 5-FU were intravenously injected into sublethally irradiated (800 cGy) C57Bl/6 mice together with 2 × 10⁵ fresh BM cells. Peripheral blood (PB) repopulation was monitored every 2 weeks by flow cytometry for the presence of fluorescent cells. After 20 weeks posttransplantation, mice were sacrificed. All mouse experiment protocols were approved by the Animal Care Committee of the HMR Research Center.

In vitro cultures

Transduced BM cells were sorted on a fluorescence-activated cell sorting (FACS) Aria III with DiVa software (BD Bioscience) and seeded at different doses in duplicate in BM expansion medium (DMEM, 15% FBS, 6 ng/mL IL-3, 10 ng/mL IL-6, 100 ng/mL SCF, 10⁻⁵ M 2-mercaptoethanol, 50 ng/mL gentamicin, and 10 ng/mL ciprofloxacin). Doubling times were calculated using the Doubling Time website [18].

Flow cytometry analysis

BM cells from in vitro cultures were stained once a week for primitive hematopoietic population markers using the following conjugated antibodies: B220-biotin (bio), Gr1-bio, CD11b (MAC1)-bio, CD3-bio, TER119-bio, CD48-bio, c-Kit-APC, and Sca1-PE/Cy7. PB repopulation of the chimeras was monitored every 2 weeks by flow cytometry for the presence of GFP⁺ and/or YFP⁺ cells, and their contribution to the myeloid, B cell and T cell lineages (LINs), was determined using the following conjugated antibodies: B220-APC, CD11b (MAC1)-A700, and CD3-bio. More than 20 weeks posttransplantation, mice were sacrificed and analyzed by flow cytometry for the contribution of GFP⁺ and/or YFP⁺ cells to the myeloid, lymphoid, and erythroid LINs using the following conjugated antibodies: CD11b (MAC1)-Pacific Blue, Gr1-bio, B220-APC/Cy7, TER119-APC, CD4-APC/Cy7, and CD8α-APC. Biotinylated antibodies were stained with PerCP/Cy5.5-conjugated streptavidin. (All conjugated antibodies and streptavidin were purchased at BioLegend or BD Pharmingen.)

Clonogenic progenitor assays

Clonogenic progenitor assays for myeloid progenitors were performed by plating cells from HOXA4, HOXB4, or both (HOXA4/B4)-overexpressing BM cultures or from long-term repopulated HOXA4-, HOXB4-, HOXA4/B4-, or GFP-overexpressing BM chimeras in DMEM containing 1% Methocel MC (Sigma-Aldrich) supplemented with 10% FBS, 5.7% bovine serum albumin (TekniScience, Inc.), 10⁻⁵ M 2-mercaptoethanol, 5 U/mL erythropoietin, 10 ng/mL IL-3, 10 ng/mL IL-6, 50 ng/mL SCF, 2 mM glutamine (Life Technologies), and 200 mg/mL transferrin (Wisent). Colonies were scored as previously described [19].

Western blot analysis

Cells were lysed in cell lysis buffer (50 mM Tris–HCl, pH 8.0, 100 mM NaCl, 1% Triton X-100, 5 mM EDTA (Mallinckrodt Baker, Inc.) with protease inhibitor cocktail (BD Bioscience). Proteins were subjected to SDS-PAGE and transferred to a nitrocellulose membrane. Membranes were incubated with a monoclonal anti-HOXB4 antibody (1:1,000 dilution; Developmental Studies Hybridoma Bank at the University of Iowa, Iowa City, IA) followed by a secondary polyclonal anti-rat antibody coupled with horseradish peroxidase (Santa Cruz Biotechnologies, Inc.) or a monoclonal anti-Flag antibody (1:1,000 dilution; Sigma-Aldrich) followed by a secondary polyclonal anti-rabbit antibody coupled with horseradish peroxidase (Jackson Immuno Research Laboratories, Inc.). For actin detection, membranes were incubated with a monoclonal anti-actin antibody (1:1,000 dillution; Millipore) followed by a secondary polyclonal anti-mouse antibody coupled with horseradish peroxidase (Jackson Immuno Research Laboratories, Inc.).

HSC purification and transduction

For HSC purification, BM cells from wild-type (C57BL/6 and/or B6SJL) mice were harvested and LIN-negative cells were purified using the Mouse Hematopoietic Progenitor Cell Enrichment Kit (STEMCELL Technologies). LIN⁻ cells were stained using conjugated antibody c-Kit-APC, Sca1-PerCPCy5.5, and CD150-PECy7 (BioLegend) and c-Kit⁺Sca1⁺CD150⁺ HSC were sorted on a FACS Aria III with DiVa software (BD Biosciences). After overnight culture in prestimulation media (DMEM, 15% FBS, 6 ng/mL IL-3, 40 ng/mL IL-6, 100 ng/mL SCF, 10⁻⁵ M 2-mercaptoethanol, 50 ng/mL gentamicin, and 10 ng/mL ciprofloxacin), HSC were subjected to two rounds of retroviral infection by spinoculation at 2,250 rpm for 90 min using MSCV-HOXA4-GFP, MSCV-HOXB4-GFP, or MSCV-GFP retrovirus. Two days after the last round of infection, GFP⁺ cells were sorted on a FACS AriaIII.

RNA isolation and amplification

Total RNA was isolated from HOXA4⁺, HOXB4⁺, or control-transduced HSC using TRIzol reagent (Life Technologies), treated with DNase-I (Life Technologies), and purified using the RNeasy MinElute Kit (Qiagen). RNA was amplified using the MessageAmp II aRNA Amplification Kit (Life Technologies). cDNA was prepared from 500 ng of amplified RNA or 1 μg of total RNA using MMLV-RT (Life Technologies) and random primers (Life Technologies) according to the manufacturer's protocols.

Quantitative reverse transcription–polymerase chain reaction

Primers for human and mouse HOX genes were used according to previously validated sequences [20,21]. Primers for candidate target genes were selected using the Primer Bank [22]. qRT-PCR was performed on an ABI 7500 Thermal Cycler (Applied Biosystems) using SYBR Green (Applied Biosystems). Triplicates were accepted in a 0.5 CT range. Relative quantification was achieved using the ΔΔCt method.

Results

HOXA4 and HOXB4 BM cells have equally strong proliferation potential

To compare the effect of HOXA4 or HOXB4 overexpression on the growth of primitive hematopoietic cells, cultures were initiated with primary BM cells overexpressing HOXA4-IRES-YFP, HOXB4-IRES-GFP, or both (Fig. 1A). qRT-PCR analysis confirmed the overexpression of HOXA4 and/or HOXB4, which were up to 1,000-fold higher than the endogenous levels (Fig. 1B). All BM cultures underwent massive proliferation resulting in average expansions of 5 × 10⁶-fold over the initial numbers after 4 weeks for all three conditions (Fig. 1C), which is 100-fold over the expansion of control cultures as reported before [6]. Co-overexpression of HOXA4 and HOXB4 did not further enhance proliferation.

FIG. 1.

In vitro BM cultures. (A) Overview of the in vitro experimental strategy, including a representative flow cytometry profile showing cells expressing HOXA4-YFP, HOXB4-GFP, or both. (B) qRT-PCR analysis for endogenous and overexpressed HOXA4 and HOXB4 genes in total BM cells transduced with HOXA4-YFP, HOXB4-GFP, or both vectors. (C) Representative growth curves of individual cultures of total BM cells transduced with HOXA4-YFP, HOXB4-GFP, or both (n = 3). Number of (D) total myeloid progenitors and (E) primitive GEMM progenitors in individual BM cultures transduced with HOXA4-YFP, HOXB4-GFP, or both vectors (n = 2). (F) Representative growth curve of HOXA4-YFP and HOXB4-GFP BM cells in cocultures (n = 3). Number of (G) total myeloid progenitors and (H) primitive GEMM progenitors in HOXA4-YFP and HOXB4-GFP in coculture (n = 2). *P ≤ 0.05 two-tailed Student's t-test. BM, bone marrow; CFC, colony-forming cell; GEMM, granulocytic–erythroid–megakaryocyte–monocyte; 5-FU, 5-fluorouracil; qRT-PCR, quantitative reverse transcription–polymerase chain reaction.

The numbers of myeloid progenitor cells, assessed by colony-forming cell (CFC) assays, were also comparable with few significant transitory differences (Fig. 1D, E), resulting in a net expansion up to 15 × 10⁶-fold. Importantly, the number of granulocytic–erythroid–megakaryocytic–monocytic (GEMM)-CFC was equally increased by HOXA4 and HOXB4 (Fig. 1E), indicating self-renewal of primitive hematopoietic cells in vitro. Consistently, flow cytometry analysis showed an increase of LIN⁻CD48⁻c-Kit⁺Sca1⁺ primitive cells (data not shown). In competition cultures, the proportions of HOXA4 and HOXB4 cells fluctuated around 50% and contained comparable numbers of progenitors, indicating equal proliferative potentials for HOXA4 and HOXB4 on hematopoietic cells in vitro (Fig. 1F-H and data not shown).

HOXA4 promotes better short-term hematopoietic repopulation than HOXB4

To assess the potential of HOXA4 and HOXB4 HSC to repopulate lethally irradiated recipients, HOXA4 or HOXB4-transduced BM cells were transplanted at low doses (10,000–30,000 BM cells, corresponding to 5–15 HSC) along with 200,000 total BM cells (Fig. 2A). All mice that received HOXA4 or HOXB4-transduced BM cells were repopulated at higher levels in the periphery than control chimeras (Fig. 2B). Interestingly, compared to HOXB4, the short-term repopulation was significantly superior for HOXA4 BM cells (Fig. 2B). Flow cytometry showed that this elevated short-term repopulation by HOXA4 was associated with an increased repopulation of B cells (8 weeks, Fig. 2C, left panel), but was changed to higher myeloid contributions compared to control at long-term repopulation (20 weeks, Fig. 2C, right panel). Only transient fluctuations in myeloid and T cell reconstitution were observed for HOXB4 (Fig. 2C, left panel).

FIG. 2.

Hematopoietic reconstitution in vivo. (A) Overview of the in vivo experimental strategy used in this study. (B) Engraftment (in %GFP) of HOXA4 (n = 15), HOXB4 (n = 5), and control chimeras (n = 13) in the periphery over time. (C) PB analysis of HOXA4, HOXB4, and control-transduced BM chimeras for myeloid and lymphoid LIN contribution at 8 weeks (left panel) and 20 weeks (right panel) reconstituted mice assessed by flow cytometry. B cells and T cells were detected using antibodies against B220 and CD3 surface markers, respectively. Antibody for MAC1 was used to detect myeloid cells. (D) Hematopoietic organ repopulation (in %GFP) of long-term (>20 weeks) HOXA4, HOXB4, and control GFP chimeras by flow cytometry analysis. For HOXA4 and control chimeras the results are pooled from three independent experiments.*P ≤ 0.05 (over control); **P ≤ 0.05 (HOXA4 vs HOXB4) one-tailed (B) and two-tailed (C, D) Student's t-test. LIN, lineage; PB, peripheral blood; Spl, spleen; Thy, thymus.

Higher levels of engraftment were also observed in HOXA4 and HOXB4 hematopoietic organs (Fig. 2D), which were sustained by 2-fold higher frequencies of myeloid progenitors (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/scd).

Competitive transplantation (Fig. 3A) assays also demonstrated an early transient higher reconstitution by HOXA4 compared to HOXB4 cells in the periphery, and similar long-term repopulation (Fig. 3B), but in contrast to single HOX chimeras, no significant differences between HOXA4 and HOXB4 were observed in LIN distribution (B220⁺, CD3⁺, and MAC1⁺) (Fig. 3C). However, in the hematopoietic organs a significant higher contribution was observed in the BM for HOXB4-transduced cells (Fig. 3D). Concurrently, more HOXB4 positive B, myeloid, and erythroid cells were observed in the BM, but not in other organs (Fig. 3E and data not shown). In agreement with this observation, more HOXB4 progenitors were present in the BM, although frequencies of myeloid progenitors within the pool of either HOXA4 or HOXB4 cells were the same (Fig. 3F).

FIG. 3.

HOXA4 and HOXB4 in vivo competition. (A) Overview of the in vivo experimental strategy used in this study. (B) Engraftment of HOXA4 (%YFP) and HOXB4 (%GFP) in competitive chimeras (n = 5) in the periphery over time. (C) PB analysis of HOXA4, HOXB4, and control chimeras for myeloid and lymphoid LIN contribution in short-term (8 weeks, top panel) and long-term (20 weeks, lower panel) reconstituted mice assessed by flow cytometry. B cells and T cells were detected using antibodies against B220 and CD3 surface markers, respectively. Antibody for MAC1 was used to detect myeloid cells. (D) Hematopoietic organ repopulation (in %GFP and %YFP) of long-term (>20 weeks) competitive chimeras by flow cytometry analysis. (E) Flow cytometry analysis of LIN repopulation in BM of long-term (<20 weeks) competitive chimeras using antibodies against B220, MAC1, Gr1, and TER119 surface marker. (F) Frequency of HOXA4 and HOXB4 myeloid progenitor in BM of long-term (>20 weeks) competitive chimeras. *P ≤ 0.05 two-tailed Student's t-test.

Increased expression of Notch-related genes by HOXA4 and HOXB4

To gain insight into the molecular mechanisms by which HOXA4 and HOXB4 mediate the expansion of primitive hematopoietic cells, the expression levels of a panel of candidate genes were measured in HOX and control transduced CD150⁺LIN⁻c-Kit⁺Sca1⁺ (LKS) cells by qRT-PCR (Fig. 4A). Potential candidate target genes were selected based on previous screens with HOXB4 and/or HOXC4 (Supplementary Table S1) and based on their reported role in HSC biology, self-renewal, or cancer biology [3,9,23,24]. Ectopic expression levels for HOXA4 and HOXB4 were similar and about 100-fold above the endogenous levels (Fig. 4B, C). HOXA4 and HOXB4 modulated the expression of 51 of the 87 candidates tested (Supplementary Table S2). Among them, 23 were ≥2-fold upregulated by both HOXA4 and HOXB4, whereas 17 and 11 were ≥2-fold upregulated by either HOXA4 or HOXB4 only, respectively (Fig. 4D and Table 1). Differentially expressed genes were involved in processes such as regulation of transcription, cell adhesion, stem cell maintenance/differentiation, pluripotency, cell cycle, and proliferation (Supplementary Table S3).

FIG. 4.

Candidate gene expression in primary HSC. (A) Overview of the experimental strategy used in this study and representation of HOXA4-GFP, HOXB4-GFP, and control GFP retroviral vectors. (B) Western blot analysis for retroviral HOXA4 or HOXB4 expression in NIH-3t3 cells transduced with HOXA4, HOXB4, or control vectors. (C) qRT-PCR analysis for HOXA4 and HOXB4 genes in sorted HSC (LIN⁻c-Kit⁺Sca1⁺CD150⁺) transduced with HOXA4, HOXB4, or control vector. (D) Venn diagram showing the total number of genes that are ≥ 2-fold upregulated by HOXA4 and/or HOXB4 in HSC. (E) Log2 Fold change of candidate genes from the Notch signaling pathway in HSC overexpressing HOXA4 or HOXB4 compared to control HSC. Dark region corresponds to 1.5-fold change value. (F) Venn diagram showing the total number of the Notch signaling pathway genes that are ≥ 1.5-fold upregulated by HOXA4 and/or HOXB4 in HSC. HSC, hematopoietic stem cells; RQ, relative quantification; LKS, LIN⁻c-Kit⁺Sca1⁺.

Table 1.

Relative Expression of ≥2-Fold Up- or Downregulated Gene Following HOXA4 or HOXB4 Overexpression in Primary HSC

		Log2 fold change
Gene	Name	HOXA4	HOXB4
≥2-Fold upregulated genes by both HOXA4 and HOXB4
Dll1	Delta-like 1 (Drosophila)	3.13	3.18
Zeb2	Zinc finger E-box binding homeobox 2	2.33	2.22
Pou5f1	POU class 5 homeobox 1	2.30	1.95
Egr1	Early growth response 1	2.20	1.93
Zic1	Zic family member 1	2.19	1.42
Sox2	SRY (sex determining region Y)-Box 2	2.12	1.67
Egr2	Early growth response 2	2.03	2.28
Dach1	Dachshund homolog 1 (Drosophila)	1.81	1.04
Col1a1	Collagen, type I, alpha 1	1.70	1.50
Zfpm1	Zinc finger protein FOG family member 1	1.69	1.02
Cxcr1	Chemokine (C-X-C motif) receptor 1	1.66	2.24
Dlx1	Distal-less homeobox 1	1.65	1.21
Nr2f2	Nuclear receptor subfamily 2 group F member 2	1.61	1.90
Cd38	CD38 molecule	1.54	2.24
Pax1	Paired box 1	1.53	2.19
Notch1	Notch 1	1.50	1.55
Lmx1b	LIM homeobox transcription factor 1 beta	1.48	3.57
Itga2	Integrin, alpha 2 (CD49B alpha 2 subunit of VLA-2 receptor)	1.44	1.78
Gpx8	Glutathioneperoxidase 8 (putative)	1.44	2.47
Lin28b	Lin-28 homolog B (Caenorhabditis elegans)	1.34	1.29
Dlx2	Distal-less homeobox 2	1.33	1.64
Thbs1	Thrombospondin 1	1.18	1.66
Fzd1	Frizzled family receptor 1	1.03	2.20
≥2-Fold upregulated genes by HOXB4 only
Cdx2	Caudal type homeobox 2	0.78	3.80
Ccl3	Chemokine (C-C motif) ligand 3	0.38	2.03
Klf4	Kruppel-like factor 4 (gut)	−0.17	1.97
Slamf1	Signaling lymphocytic activation molecule family member 1	0.96	1.91
Timp3	TIMP metallopeptidaseinhibitor 3	−0.23	1.53
Zbtb16	Zinc finger and BTB domain containing 16	−0.46	1.50
Ebf1	Early B-cell factor 1	−0.28	1.49
EP300	E1A binding protein p300	−0.75	1.46
Pbx1	Pre-B-cell leukemia homeobox 1	0.75	1.25
Dll3	Delta-like 3 (Drosophila)	0.49	1.15
Hnf4a	Hepatocyte nuclear factor 4 alpha	0.75	1.12
≥2-Fold upregulated genes by HOXA4 only
Fzd4	Frizzled family receptor 4	3.06	0.94
Jag1	Jagged 1	2.91	0.23
Nr4a2	Nuclear receptor subfamily 4, group A, member 2	1.93	0.52
Cdkn1a	Cyclin-dependent kinase inhibitor 1A (p21. Cip1)	1.87	0.25
Cxcr4	Chemokine (C-X-C motif) receptor 4	1.77	−0.33
Hey1	Hairy/enhancer-of-split related with YRPW motif 1	1.73	0.79
Emcn	Endomucin	1.69	0.96
Mef2c	Myocyte enhancer factor 2C	1.48	0.70
Tgm2	Transglutaminase 2	1.43	−0.23
Cited2	Cbp/p300-interacting transactivator with Glu/Asp-rich carboxy-terminal domain 2	1.39	0.22
Dtx1	Deltexhomolog 1 (Drosophila)	1.38	0.61
Nfyb	Nuclear transcription factor Y beta	1.38	−0.37
Ikaros	IKAROS family zinc finger 1 (Ikaros)	1.16	0.15
Trim28	Tripartite motif containing 28	1.15	0.76
Nsd1	Nuclear receptor binding SET domain protein 1	1.14	0.91
Dnmt1	DNA (cytosine-5)-methyltransferase 1	1.08	0.70
Notch2	Notch 2	1.00	0.52
Inverse regulation by HOXA4 and HOXB4
Sox4	SRY (sex determining region Y)-box 4	1.11	−1.54

HSC, hematopoietic stem cells.

Also, the expression of several genes of the Wnt and Notch signaling pathways were affected, more notably by HOXA4. Among 30 Notch-related genes tested, the levels of 24 were at least 1.5-fold changed by the overexpression of either HOXA4 and/or HOXB4 (Fig. 4E, F and Table 2). These included Notch receptors (Notch1 and Notch2) and Notch ligands (Dll1, Dll3, and Jag1). The expression level of Notch2 was higher than Notch1 in primitive hematopoietic cells (data not shown), which corresponds with previous reports using reporter mice [25]. An upregulation of Notch target genes such as Hey1 and Runx1 indicates potential activation of the canonical Notch signaling pathway. Furthermore, in agreement with the larger B cell population the expression of Pax5 was increased in HOXA4-transduced CD150⁺LKS cells. Together these data suggest that the Notch signaling pathway plays a role in HOX4-mediated self-renewal of HSC.

Table 2.

Relative Expression of Genes Implicated in the Notch Signaling Pathway in HSC Overexpressing HOXA4 of HOXB4

		Log2 fold change
Gene	Name	HOXA4	HOXB4
Notch receptors
Notch1	Notch 1	1.50	1.55
Notch2	Notch 2	1.00	0.52
Notch ligands
Dll1	Delta-like 1 (Drosophila)	3.13	3.18
Dll3	Delta-like 3 (Drosophila)	0.49	1.15
Jag1	Jagged 1	2.91	0.23
Regulators
Dtx1	Deltex homolog 1 (Drosophila)	1.38	0.61
Dtx2	Deltex homolog 2 (Drosophila)	0.88	0.09
Dtx4	Deltex homolog 4 (Drosophila)	0.39	0.58
Dlx1	Distal-less homeobox 1	1.65	1.21
Dlx2	Distal-less homeobox 2	1.33	1.64
Numb	Numbhomolog (Drosophila)	−0.03	−0.09
Cofactors
Hdac1	Histone deacetylase 1	0.43	−0.59
Hdac2	Histone deacetylase 2	0.46	−0.32
EP300	E1A binding protein p300	−0.75	1.46
Targets
Hes1	Hairy and enhancer of split 1 (Drosophila)	0.39	−0.23
Hey1	Hairy/enhancer-of-split related with YRPW motif 1	1.73	0.79
Runx1	Runt-related transcription factor 1	0.83	0.22
Zfpm1	Zinc finger protein FOG family member 1	1.69	1.02
Kat2a	K(lysine) acetyltransferase 2A	0.21	−0.43
Ccnd1	Cyclin D1	0.75	−0.39
Cdkn1a	Cyclin-dependent kinase inhibitor 1A (p21, Cip1)	1.87	0.25
Cd44	CD44 molecule (Indian blood group)	0.95	0.59
Nr4a2	Nuclear receptor subfamily 4 group A member 2	1.93	0.52
Pax5	Paired box 5	0.59	−0.06
Genes related to the Wnt pathway that crosstalk with Notch
Fzd1	Frizzled family receptor 1	1.03	2.20
Fzd4	Frizzled family receptor 4	3.06	0.94
Egr1	Early growth response 1	2.20	1.93
Sox2	SRY (sex determining region Y)-box 2	2.12	1.67
Sox4	SRY (sex determining region Y)-box 4	1.11	−1.54
Dvl1	Dishevelled segment polarity protein 1	0.50	−0.35

Bold indicates ≥2-fold change in expression.

Discussion

In this study, we demonstrate that HOXA4 and HOXB4 promote significant expansion of BM cells in vitro through maintenance and expansion of the progenitor pool. The magnitude of expansion induced by HOXA4 and HOXB4 was similar in vitro; however, when transplanted into mice, HOXA4-transduced cells clearly provided better short-term repopulation of BM than HOXB4. This was associated with a higher contribution of lymphoid cells compared to HOXB4-transduced cells, indicating that HOXA4 is particularly important in restoring the B cell compartment, congruent with our previous observations [6].

In agreement with our data, it has been reported that other HOX4 paralog genes confer similar potential expansion advantages to mouse or human primitive hematopoietic cells in vitro, when overexpressed [8] or cultured on engineered stromal cells (HOXB4 or HOXC4) [9]. Thus, in the absence of cues from the hematopoietic niche and its possible physical restraints, the proliferative potential of hematopoietic stem and progenitor cells is equally enhanced by HOXA4 and HOXB4.

Whether the enhanced short-term repopulation in HOXA4 BM recipients derives from a higher sensitivity of committed early progenitors, in particular those of the B cell LIN, to HOXA4 or from a stronger response of HSC is not clear, but a HSC subset with a greater proliferative short-term phenotype and better B cell reconstitution potential has been defined [26].

Competitive transplantation assays confirmed the advantage of HOXA4 in short-term repopulation, but the dominant lymphoid reconstitution of HOXA4 was absent in the presence of HOXB4-transduced cells. These data suggest that a paracrine effect between HOXA4- and HOXB4-transduced cells exist that could be direct or indirect through the hematopoietic niche as has been reported for HOXB4 [27].

qRT-PCR data on a panel of candidate genes demonstrated modulation of the Notch pathway, predominantly by HOXA4, in primitive hematopoietic cells. The Notch signaling pathway is known to play a critical role in HSC self-renewal [28–31] and some synergism between HOXB4 and Dll1 in the expansion of primate CD34⁺ cord blood cells has been reported [32,33]. Although crosstalk between the Notch and HOX pathway have been reported [34,35], we show for the first time that Notch genes are common downstream targets of HOXA4 and HOXB4 in primitive hematopoietic cells, suggesting that HOX4-induced self-renewal may involve Notch signaling. The Notch-related target genes included those coding for receptors, ligands, and cofactors, indicating that HOXA4 and HOXB4 may regulate the responsiveness of hematopoietic cells to Notch ligands, and amplify the signal response through enhanced ligand production. However, the enhanced lymphoid repopulation by HOXA4 BM cells could not be explained solely by higher Notch1 expression, which is important for lymphoid differentiation [25,36,37], as equal levels were found in HOXB4 cells.

In addition to the Notch pathway, the expression levels of Wnt-related genes were also increased by HOXA4 and HOXB4. Notch and Wnt are major pathways in HSC self-renewal and embryonic studies demonstrated that HOX genes integrate their signals to establish segment identity [34,38]. Crosstalk between these signaling cascades has also been demonstrated in HSC [39] and modulation of genes in both pathways maybe essential for the prolonged self-renewal of HSC induced by HOXA4 or HOXB4. Furthermore, the observation that activation of the Wnt signaling pathway by Wnt3a promotes short-term multilineage reconstitution of LIN⁻c-Kit⁻Sca1⁺ BM cells in vivo [40] indicate that Wnt signaling might contribute to the enhanced HOXA4-induced short-term repopulation, as HOXA4 overexpression induced higher expression of Wnt-associated genes.

qRT-PCR data also showed the increase in expression of several self-renewal and pluripotency genes, which indicate that HOXA4 and HOXB4 may activate intrinsic self-renewal pathways. Among these genes are Lin28b, a micro RNA-binding protein that, when overexpressed, increases self-renewal activity of adult HSC [41], and the transcription factors Sox2 and Pou5f1 (Oct4), also known for their critical role in pluripotency. The co-overexpression of the latter has also been shown to enhance the proliferation and differentiation of human mesenchymal stem cells [42]. Thus, upregulation of those genes by HOXA4 and HOXB4 likely contributes to the induction of HSC self-renewal.

In conclusion, our results show that overexpression of HOXA4 and HOXB4 results in the activation of genes involved in both intrinsic and extrinsic pathways, and suggest a potential role for the Notch pathway in association with Wnt signaling downstream HOXA4 and HOXB4 in primitive hematopoietic cells. Moreover, in the absence of niche-derived signals delivered, HSC-overexpressing HOX4 genes have the same potency in culture, but show paralog-specific differences in vivo. Together, based on our results, manipulation of the Notch pathway in conjunction with HOX4 merits further exploration for the expansion of primitive hematopoietic cells for therapeutic strategies.

Acknowledgments

The authors thank Martine Dupuis from the HMR flow cytometry platform for cell sorting and flow cytometry analysis. The authors also thank the staff of the animal care facility for taking care of the animals. The authors are also grateful for the excellent comments and suggestions on the article by Drs. Alex Thompson, Nicolas Pineault, and Heather Melichar. This study was supported by the Canadian Cancer Society Research Institute, grant no. 20399, and a Discovery grant of NSERC; M.F. is a fellow of the Cole Foundation; C.E.L.G. and M.F. are recipients of departmental scholarships.

Author Disclosure Statement

No competing financial interests exist.