Activation of Stem-Cell Specific Genes by HOXA9 and HOXA10 Homeodomain Proteins in CD34+ Human Cord Blood Cells
http://www.100md.com
《干细胞学杂志》
b Terry Fox Laboratory, British Columbia Cancer Agency, Vancouver, British Columbia, Canada;
c Department of Medicine and Comprehensive Cancer Center, University of California at San Francisco, California, USA
Key Words. Hematopoiesis ? Hematopoietic stem cells (HSCs) ? Homeobox genes ? Microarray ? Gene expression profiling
Correspondence: H. Jeffrey Lawrence, M.D., Department of Medicine, Hematology Research (151H), Veterans Affairs Medical Center, 4150 Clement St., San Francisco, CA 94121, USA. Telephone: 415-221-4810, ext. 3340; Fax: 415-750-6959; e-mail: jeffl@medicine.ucsf.edu
ABSTRACT
The 39 members of the HOX family of homeobox genes encode DNA-binding proteins, which play a key role in pattern formation along various body axes . The highly regulated deployment of HOX genes in complex overlapping domains over time and space during embryogenesis appears to be critical for the correct positioning of body appendages, strongly suggesting that a combinatorial HOX code may determine the identity of specific body segments. Several genes of the HOXA and HOXB clusters, including two adjacent genes HOXA9 and HOXA10, are expressed in primitive human and murine hematopoietic cells, and that expression is downregulated as cells differentiate, suggesting a role in early blood cell development .
The function of HOX genes in normal murine hematopoiesis has been explored both in knockout mice and in murine models of retrovirally driven overexpression. HOXA9 has an important role in normal hematopoiesis, as HOXA9-deficient mice have a variety of myeloid and lymphoid defects, as well as abnormalities in hematopoietic stem cell (HSC) function . By contrast, no consistent hematopoietic abnormalities have been described in mice lacking the HOXA10 gene. These findings suggest that, although both HOXA9 and HOXA10 are expressed in primitive hematopoietic cells, the two genes have distinct functions during normal blood cell differentiation. In accord with this notion is the fact that HOXA9 and HOXA10 proteins have little homology to one another outside of the 60-amino-acid DNA-binding homeodomain (HD).
However, overexpression of either HOXA9 or HOXA10 in murine marrow cells produces similar blood cell abnormalities, including a marked expansion of HSCs and committed progenitors with eventual transformation to acute myeloid leukemia (AML) . These similar biologic effects on blood cell development suggest that HOXA9 and HOXA10 modulate common genetic programs in blood cells when they are over-expressed. Moreover, in recent studies, we have documented striking overlap in the functional effects of NUP98-AbdB class HOX fusions, as tested with naturally occurring leukogenic fusion such as NUP98-HOXD13 and engineered fusions such as NUP98-HOXA10 . Such findings further suggest a high degree of target gene overlap, at least for HOX within the AbdB class of HOX (paralogs 9–13).
Since homeobox genes encode DNA-binding proteins, it has been presumed that HOX proteins function as transcription factors. Recent papers have described the transcriptional profiles of certain HD proteins using microarray and real-time reverse transcription polymerase chain reaction (RT-PCR) technologies in an effort to identify potential gene targets and their associated downstream molecular pathways. These studies have examined the genes modulated by the expression of HOXC13, HOXD3, the non-HOX HD proteins, PAX6, and the fly HOX1 homologue, labial, in nonhematopoietic tissues . Another study examined the expression profile of the leukemogenic fusion gene, NUP98-HOXA9, in an immortalized myeloid cell line , but no study published to date has attempted to identify genes modulated by a wild-type HOX protein in primary human hematopoietic cells.
In a recent paper, our laboratory published a description of the HOXA9 transcriptome in human leukemic cell lines, using a transient overexpression strategy in three cell lines, two myeloid and one lymphoid . In that study we observed modulation of a large number of genes within 24 hours of introduction of a HOXA9 expression vector in these cells. The modulated genes represented a wide variety of functional groups, including oncogenes, cell-cycle proteins, enzymes, membrane proteins, and other transcription factors. Interestingly, a number of these genes are known to be part of the transcriptome of normal HSCs and to be similarly modulated in primary samples of human AML, suggesting that these genes are authentic biologic targets of HOXA9 .
Remarkably, in that study the gene-expression profiles observed for the myeloid and lymphoid lines were dramatically different, indicating that the transcriptional effects of HOXA9 are highly dependent on cell context. This observation raises questions as to whether the HOXA9 targets identified in aneuploid immortalized myeloid cell lines would match the gene targets for HOXA9 in normal hematopoietic cells. To answer this question, we have studied the expression profile of the HOXA9 protein in human umbilical cord CD34+ cells. In addition, we have used this same system to examine the transcriptome of the related HOXA10 protein, permitting a comparison of the genes modulated by these two closely related transcription factors.
MATERIALS AND METHODS
Identification of Genes Modulated by Overexpression of HOXA9 and HOXA10
To identify and compare the genes modulated by HOXA9 and HOXA10 in normal human hematopoietic cells, we used a retroviral overexpression strategy in CD34+ cells obtained from human umbilical CB (Fig. 1 shows a schema of the experimental design). Real-time RT-PCR of RNA from day-3 samples revealed that a six- to sevenfold increase in mRNA levels of HOXA9 and HOXA10 (Fig. 2A) was achieved in the transduced cells. CB mRNA from HOXA9- and HOXA10-transduced samples were compared with MIG only–transduced controls through hybridization to cDNA microarrays to provide a differential gene expression profile. These microarray data were then subjected to stepwise statistical analysis.
Figure 1. Schematic diagram of transduction strategy for human CD34+ cells. Human umbilical cord cells were prestimulated with growth factors, aliquoted, and transduced with one of three retroviral constructs. Green fluorescent protein–positive (GFP+)/CD34+ cells were sorted from each transduction. Total RNA was isolated, and mRNA was amplified by two rounds of in vitro transcription. Gene-expression differences were analyzed by comparing either GFP control versus HOXA9-GFP or GFP control versus HOXA10-GFP samples, using cDNA microarrays. Abbreviations: IL, interleukin; MSCV, murine stem cell virus.
Figure 2. Genes modulated by HOXA9 and HOXA10 as identified by significance analysis of microarrays (SAM) and Eisen Cluster analysis. (A): Bar chart representing the average of triplicate quantitative reverse transcription polymerase chain reaction (QRT-PCR) measurements of HOXA9 or HOXA10 mRNA in HOX gene–green fluorescent protein–positive (GFP+)–transduced versus GFP+ only–transduced cells. The bars represent standard error. (B): This SAM plot represents modulated genes that are shared between HOXA9 day-3 and HOXA10 day-3 datasets. SAM data for all genes that were deemed significant were ranked by the magnitude of their observed "d" scores, or difference from the comparator group. The MIG-control transfected group was the comparator group for changes in common in both HOXA9 and HOXA10. The differences from the comparator group were in either direction: upregulated genes (depicted in red) or downregulated (green). The delta was set to support a false discovery rate of 1% or less. Genes with expression levels that are statistically beyond delta in either direction are plotted either above (induced, red) or below (repressed, green) the comparator or control group. Genes whose expressions did not change more than the set delta in either direction were considered to be not statistically significantly different at the set false discovery rate. (C): Cluster analysis of cDNA microarray data. HOXA9 and HOXA10 microarray data were analyzed by the Eisen Hierarchical Cluster program and visualized with TreeView. The cluster shown represents 115 genes with a minimum of four data points that were at least twofold up- or downregulated. Red, induced; green, repressed; black, no change; gray, missing data. Group i includes induced genes common to HOXA9 and HOXA10; group ii includes genes that are induced by HOXA9 and repressed by HOXA10; group iii includes genes that are activated by HOXA10 and are downregulated by HOXA9; group iv includes repressed genes common to HOXA9 and HOXA10. Data shown for transduced cells harvested at day 3.
First, statistical analysis was performed using the SAM algorithm, which considers all genes with even small changes in expression. SAM identified a large number of genes of similar expression patterns in HOXA9 and HOXA10 day 3–transduced samples (Fig. 2A). Of the approximately 42,000 cDNA array clones, 1,998 genes were similarly modulated by both HOXA9 and HOXA10, as compared with the MIG control. By setting the value at 0.614, the FDR was 0.87%, and only 17 out of 1,998 genes were predicted to be false. Thus, SAMs show that these related HOX proteins are able to modulate many of the same genes. The small number of differentially modulated genes (i.e., upregulated by HOXA9 and downregulated by HOXA10, or the reverse) was represented almost entirely by expressed sequence tags of unknown function.
In a second step, Eisen Cluster analysis was performed, introducing threshold values to increase the likelihood of the identified genes being truly regulated and of biological significance. (Fig. 2B). Triplicate microarray data from day 3–harvested HOXA9- or HOXA10-transduced samples were filtered by the Eisen Cluster analysis program for genes, with four out of six observations showing twofold induction or repression. Using these more stringent criteria, 115 genes were modulated by one or both of the over-expressed genes. The gene expression profiles for HOXA9 and HOXA10 are shown in Figure 2C, broken down into four groups (i–iv). Group i consists of genes that are similarly regulated by HOXA9 and HOXA10. Groups ii and iii show genes that are differentially modulated by HOXA9 and HOXA10—for example, upregulated by HOXA9 yet downregulated by HOXA10. The largest group, group iv, represents genes that are repressed by both HOXA9 and HOXA10. This Eisen Cluster analysis suggests that there are many more genes regulated in a similar fashion by HOXA9 and HOXA10 than are differentially regulated.
Annotated genes that passed the Eisen Cluster filter for at least twofold differential expression, which were similarly regulated in more than one cord, and which were also determined significant by SAM, are presented in Tables 1 and 2. Table 1 is a list of 35 named genes that were either induced or repressed by HOXA9 and were also modulated by HOXA10. Table 2 lists genes that were modulated by HOXA10 after 3 days and remained modulated or were not modulated until 6 days of overexpression. While several genes were shown to be modulated at both time points, a larger number of genes was identified in the day-6 sample, which probably reflects the progressive modulation of secondary and tertiary genes over time. Several mRNAs that were previously detected in CD34+ cells are noted in both tables, showing that both HOX proteins modulate a number of genes that are active in HSCs .
Table 1. Genes modified by HOXA9 overexpression in human CD34+ cord blood cells
Table 2. HOXA10-modulated genes in human CD34+ cord blood cells
HOXA9 and HOXA10 Modulate Genes Involved in HSC Functions
Of the genes found to be significantly upregulated by overexpression of both HOXA9 and HOXA10, four genes were chosen for validation by real-time RT-PCR because of prior data showing these genes to be subject to modulation by HOXA9 in immortalized cells lines and/or because these genes were known to be expressed in HSCs. All of these genes were validated by triplicate analyses of RNA from HOXA10 day-3 samples (Fig. 3). The amplitude of the difference in gene expression as measured by real-time RT-PCR was often greater than that observed by microarray analysis, though results were always concordant in the direction of change.
Figure 3. QRT-PCR confirms the changes in gene expression that were seen in the microarray studies. The averages of triplicate gene-expression changes were measured by real-time RT-PCR, then calculated by the delta-delta Ct method, using the HOXA10 (day-3) sample, and compared with the fold changes seen on the microarrays. Gray, QRT-PCR; black, microarray analysis. Error bars represent standard error.
Three of the genes studied, ALDH1, ERG, and VLCS-H1, have all been reported to be expressed in CD34+ cells, compared with differentiated populations, suggesting an important role for these genes in normal HSC biology . VLCS-H1, or very long–chain acyl-CoA synthetase, had a 10-fold induction measuring with real-time RT-PCR, while ETS-related gene (ERG) showed an average 2.4-fold induction in HOXA10-transfected cells. While ERG expression was not found to be upregulated in our previous study with myeloid cell lines that overexpress HOXA9, a closely related gene v-ETS2 was. Aldehyde dehydrogenase 1 (ALDH1), which showed a 6.3-fold induction of expression by real-time RT-PCR, had been shown to be a target for HOXA9 in our previous microarray analysis, which included transient transcription assays with an ALDH1 reporter construct. IRX3, or Iroquois-class HD protein 3, which showed a 3.9-fold induction of gene expression by HOXA10 using real-time RT-PCR, was shown to be an early gene target in our previous study but has not been identified as part of the normal stem cell transcriptome. However, it is noteworthy that another member of the Iroquois family, IRX5, has been shown to be positively regulated by HOXB4 , another HOX protein known to be expressed in hematopoietic cells and to have the capacity to expand HSCs.
Activation of the Wnt Signaling Pathway by HOXA9 and HOXA10
HOXA9 and HOXA10 were found to positively regulate genes in the Wnt signaling pathway (Fig. 4). Wnt10B demonstrated a 4.5-fold increase in HOXA9-transduced cells, with a 3.1-fold increase in HOXA10-transduced cells, and this upregulation was validated by real-time PCR. In addition, two Wnt receptors, Frizzled 1 and Frizzled 5, were also upregulated, with Frizzled 1 showing approximately twofold upregulation by both HOXA9 and HOXA10, and Frizzled 5 showing two- to fourfold upregulation by HOXA10; these results were also validated by real-time PCR. The ability of these two HOX proteins to activate Wnt and Frizzled genes, given the key role of this pathway in stem cell self-renewal, may explain, at least in part, the biologic effects of HOX proteins on primitive hematopoietic cells.
Figure 4. HOXA9 and HOXA10 activate genes in the Wnt signaling pathway. Both (A) HOXA9 and (B) HOXA10 induced expression of Wnt pathway components Wnt10B, Frizzled 1 (FZD1), and Frizzled 5 (FZD5). The averages of triplicate gene-expression changes were measured by QRT-PCR using SYBR green I dye chemistry (gray) and compared with microarray analysis (black). Error bars represent standard error.
HOXA9 and HOXA10 Repress as Well as Activate Gene Expression
The Eisen Cluster analysis in Figure 2 and Tables 1 and 2 show that many genes are downregulated by both HOX proteins, as was seen in our previous expression profile study in hematopoietic cell lines. Among the downregulated genes in this survey is CYBB, a respiratory burst oxidase component pg91(phox), which is known to be expressed in maturing myeloid cells and has previously been shown to be repressed by HOXA10 (Table 2) . Other significantly downregulated genes included Tip60 and vimentin. Vimentin is an intermediate filament protein that, in blood cells, increases with monocytic differentiation. Support for the notion that the vimentin gene could be a target of HOXA proteins comes from the observation that its expression is lost in ovarian tumors that overexpress HOXA7 . Furthermore, the vimentin promoter and enhancer region has been shown to have binding sites for HOX proteins . Thus, at least two genes associated with myeloid differentiation are inhibited by HOXA9 and/or HOXA10. Tip60 is a 60-kDa Tat– interactive protein that normally represses STAT3 and is involved in sensing DNA damage and inducing apoptosis . Thus down-regulation of Tip60 could be associated with improved survival and increased genomic instability, thereby predisposing to malignant transformation.
Inhibition of the Erythroid Differentiation Program by HOXA9 and HOXA10
To identify other molecular pathways affected by HOXA9 and HOXA10, these microarray data were also analyzed with the Gen-MAPP program designed by the Gladstone Institute in affiliation with the University of California at San Francisco (http://www.genmapp.org/download.asp). This analysis demonstrated that HOXA9 (1) and HOXA10 (3) inhibit enzymes of the heme bio-synthetic pathway. Likewise, HOXA9 (4) and HOXA10 (2) also inhibit globin genes (Fig. 5). Not included in the figure is inhibition of the pathway rate-limiting enzyme, ALAS-2 by HOXA10 (Table 2). SAM showed that HOXA9 and HOXA10 also down-regulate globin genes expressed from the ?-globin and -globin loci (Tables 1 and 2; Fig. 5). Thus, both HOXA9 and HOXA10 appear to effect a general repression of erythroid-specific genes, and this is consistent with previous observations .
Figure 5. HOXA9 and HOXA10 inhibit the expression of components of the heme biosynthetic pathway and globin. Analysis of microarray data by Gen MAPP (Gene MicroArray Pathway Profiler) revealed that one enzyme of the heme synthetic pathway was downregulated at least twofold by HOXA9 and threefold by HOXA10 (days 3 and 6). Two hemoglobin genes were identified by Eisen Cluster analysis as being repressed by both HOXA9 and HOXA10 overexpression. Two additional hemoglobins were repressed by HOXA9 overexpression alone. See Tables 1 and 2.
DISCUSSION
Special thanks are given to Dr. Neal Fischbach for his comments and thoughtful reading of the manuscript. Dr. Lawrence is a VA Career Development award recipient. This study was supported by NIH grant DK48642 (H.J.L., R.K.H.), a grant from the Veterans Affairs Administration (H.J.L.), and a grant from the National Cancer Institute of Canada with funds from the Terry Fox Foundation (R.K.H.).
REFERENCES
Krumlauf R. Hox genes in vertebrate development. Cell 1994;78:191–201.
Sauvageau G, Lansdorp PM, Eaves CJ et al. Differential expression of homeobox genes in functionally distinct CD34+ subpopulations of human bone marrow cells. Proc Natl Acad Sci U S A 1994;91:12223–12227.
Pineault N, Helgason CD, Lawrence HJ et al. Differential expression of Hox, Meis1, and Pbx1 genes in primitive cells throughout murine hematopoietic ontogeny. Exp Hematol 2002;30:49–57.
Lawrence HJ, Helgason CD, Sauvageau G et al. Mice bearing a targeted interruption of the homeobox gene HOXA9 have defects in myeloid, erythroid, and lymphoid hematopoiesis. Blood 1997;89:1922–1930.
Izon DJ, Rozenfeld S, Fong ST et al. Loss of function of the homeobox gene Hoxa-9 perturbs early T-cell development and induces apoptosis in primitive thymocytes. Blood 1998;92:383–393.
Thorsteinsdottir U, Sauvageau G, Hough MR et al. Overexpression of HOXA10 in murine hematopoietic cells perturbs both myeloid and lymphoid differentiation and leads to acute myeloid leukemia. Mol Cell Biol 1997;17:495–505.
Kroon E, Krosl J, Thorsteinsdottir U et al. Hoxa9 transforms primary bone marrow cells through specific collaboration with Meis1a but not Pbx1b. EMBO J 1998;17:3714–3725.
Thorsteinsdottir U, Mamo A, Kroon E et al. Over expression of the myeloid leukemia-associated Hoxa9 gene in bone marrow cells induces stem cell expansion. Blood 2002;99:121–129.
Pineault N, Abramovich C, Ohta H et al. Differential and common leukemogenic potentials of multiple NUP98-Hox fusion proteins alone or with Meis1. Mol Cell Biol 2004;24:1907–1917.
Tkatchenko AV, Visconti RP, Shang L et al. Overexpression of Hoxc13 in differentiating keratinocytes results in downregulation of a novel hair keratin gene cluster and alopecia. Development 2001;128:1547–1558.
Chauhan BK, Zhang W, Cveklova K et al. Identification of differentially expressed genes in mouse Pax6 heterozygous lenses. Invest Ophthalmol Vis Sci 2002;43:1884–1890.
Leemans R, Loop T, Egger B et al. Identification of candidate downstream genes for the homeodomain transcription factor Labial in Drosophila through oligonucleotide-array transcript imaging. Genome Biol 2001;2:1–15.
Miyazaki YJ, Hamada J, Tada M et al. HOXD3 enhances motility and invasiveness through the TGF-beta-dependent and -independent pathways in A549 cells. Oncogene 2002;21:798–808.
Ghannam G, Takeda A, Camarata T et al. The oncogene Nup98-HOXA9 induces gene transcription in myeloid cells. J Biol Chem 2004;279:866–875.
Dorsam ST, Ferrell CM, Dorsam GP et al. The transcriptome of the leukemogenic homeoprotein HOXA9 in human hematopoietic cells. Blood 2004;103:1676–1684.
Gomes I, Sharma TT, Mahmud N et al. Highly abundant genes in the transcriptosome of human and baboon CD34 antigen-positive bone marrow cells. Blood 2001;98:93–99.
Hawley RG, Lieu FH, Fong AZ, et al. Versatile retroviral vectors for potential use in gene therapy. Gene Ther 1994;1:136–138.
Baugh LR, Hill AA, Brown EL et al. Quantitative analysis of mRNA amplification by in vitro transcription. Nucleic Acids Res 2001;29:E29.
Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 2001;98:5116–5121.
Vandesompele J, De Paepe A, Speleman F. Elimination of primer-dimer artifacts and genomic coamplification using a two-step SYBR green I real-time RT-PCR. Anal Biochem 2002;303:95–98.
Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001;25:402–408.
Wagner W, Ansorge A, Wirkner U et al. Molecular evidence for stem cell function of the slow-dividing fraction among human hematopoietic progenitor cells by genome-wide analysis. Blood 2004;104:675–686.
Theokli C, Morsi El-Kadi AS, Morgan R. TALE class homeodomain gene Irx5 is an immediate downstream target for Hoxb4 transcriptional regulation. Dev Dyn 2003;227:48–55.
Lu Y, Goldenberg I, Bei L et al. HoxA10 represses gene transcription in undifferentiated myeloid cells by interaction with histone deacetylase 2. J Biol Chem 2003;278:47792–47802.
Naora H, Montz FJ, Chai CY et al. Aberrant expression of homeobox gene HOXA7 is associated with müllerian-like differentiation of epithelial ovarian tumors and the generation of a specific autologous antibody response. Proc Natl Acad Sci U S A 2001;98:15209–15214.
Benazzouz A, Duprey P. The vimentin promoter as a tool to analyze the early events of retinoic acid-induced differentiation of cultured embryonal carcinoma cells. Differentiation 1999;65:171–180.
Xiao H, Chung J, Kao HY et al. Tip 60 is a co-repressor for STAT3. J Biol Chem 2003;278:11197–11204.
Bernstein C, Bernstein H, Payne CM et al. DNA repair/pro-apoptotic dual-role proteins in five major DNA repair pathways: fail-safe protection against carcinogenesis. Mutat Res 2002;511:145–178.
Buske C, Feuring-Buske M, Antonchuk J et al. Overexpression of HOXA10 perturbs human lymphomyelopoiesis in vitro and in vivo. Blood 2001;97:2286–2292.
Gnatenko DV, Dunn JJ, McCorkle SR et al. Transcript profiling of human platelets using microarray and serial analysis of gene expression. Blood 2003;101:2285–2293.
Shim MH, Hoover A, Blake N et al. Gene expression profile of primary human CD34+CD38low cells differentiating along the megakaryocyte lineage. Exp Hematol 2004;32:638–648.
Thompson A, McGonigle G, Grier D et al. Defining Complete HOX Signatures in Murine Hematopoietic Cells by Q-PCR. Proc Amer Soc Hematol Blood 2003;202:2100a.
Fischbach NA, Rozenfeld S, Shen W et al. HOXB6 over-expression in murine bone marrow immortalizes a myelomonocytic precursor in vitro and causes hematopoietic stem cell expansion and acute myeloid leukemia in vivo. Blood 2005;105:1456–1466.
Crooks GM, Fuller J, Petersen D et al. Constitutive HOXA5 expression inhibits erythropoiesis and increases myelopoiesis from human hematopoietic progenitors. Blood 1999;94:519–528.
Shen W, Chrobak D, Krishnan K et al. HOXB6 protein is bound to CREB-binding protein and represses globin expression in a DNA binding-dependent, PBX interaction-independent process. J Biol Chem 2004;279:39895–39904.
Kappen C. Disruption of the homeobox gene Hoxb-6 in mice results in increased numbers of early erythrocyte progenitors. Am J Hematol 2000;65:111–118.
Sengupta PK, Lavelle DE, DeSimone J. The 87-kD A-globin enhancer-binding protein is a product of the HOXB2 (HOX2H) locus. Blood 1994;83:1420–1427.
Gomez-Skarmeta JL, Modolell J. Iroquois genes: genomic organization and function in vertebrate neural development. Curr Opin Genet Dev 2002;12:403–408.
Guidato S, Prin F, Guthrie S. Somatic motoneuron specification in the hindbrain: the influence of somite-derived signals, retinoic acid and Hoxa3. Development 2003;130:2981–2996.
Peters T, Ausmeier K, Dildrop R et al. The mouse Fused toes (Ft) mutation is the result of a 1.6-Mb deletion including the entire Iroquois B gene cluster. Mamm Genome 2002;13:186–188.
van der Hoeven F, Schimmang T, Volkmann A et al. Programmed cell death is affected in the novel mouse mutant Fused toes (Ft). Development 1994;120:2601–2607.
Volkmann A, Doffinger R, Ruther U et al. Insertional mutagenesis affecting programmed cell death leads to thymic hyperplasia and altered thymopoiesis. J Immunol 1996;156:136–145.
Anderson MK, Hernandez-Hoyos G, Diamond RA et al. Precise developmental regulation of Ets family transcription factors during specification and commitment to the T cell lineage. Development 1999;126:3131–3148.
Pereira DS, Dorrell C, Ito CY et al. Retroviral transduction of TLS-ERG initiates a leukemogenic program in normal human hematopoietic cells. Proc Natl Acad Sci U S A 1998;95:8239–8244.
Faergeman NJ, Knudsen J. Role of long-chain fatty acyl-CoA esters in the regulation of metabolism and in cell signalling. Biochem J 1997;323( Pt 1):1–12.
Storms RW, Trujillo AP, Springer JB et al. Isolation of primitive human hematopoietic progenitors on the basis of aldehyde dehydrogenase activity. Proc Natl Acad Sci U S A 1999;96:9118–9123.
Baum C, Fairbairn LJ, Hildinger M et al. New perspectives for cancer chemotherapy by genetic protection of haematopoietic cells. Expert Rev Mol Med 1999;1999:1–28.
Russo J, Barnes A, Berger K et al. 4-(N,N-dipropylamino) benzaldehy de inhibits the oxidation of all-trans retinal to all-trans retinoic acid by ALDH1A1, but not the differentiation of HL-60 promyelocytic leukemia cells exposed to all-trans retinal. BMC Pharmacol 2002;2:4.
Purton LE, Bernstein ID, Collins SJ. All-trans retinoic acid delays the differentiation of primitive hematopoietic precursors (lin–c-kit+Sca+) while enhancing the terminal maturation of committed granulocyte/monocyte progenitors. Blood 1999;94:483–495.
Purton LE, Bernstein ID, Collins SJ. All-trans retinoic acid enhances the long-term repopulating activity of cultured hematopoietic stem cells. Blood 2000;95:470–477.
Kastner P, Lawrence HJ, Waltzinger C et al. Positive and negative regulation of granulopoiesis by endogenous RARalpha. Blood 2001;97:1314–1320.
Lako M, Lindsay S, Lincoln J et al. Characterisation of Wnt gene expression during the differentiation of murine embryonic stem cells in vitro: role of Wnt3 in enhancing haematopoietic differentiation. Mech Dev 2001;103:49–59.
Reya T, Duncan AW, Ailles L et al. A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 2003;423:409–414.
Lisovsky M, Itoh K, Sokol SY. Frizzled receptors activate a novel JNK-dependent pathway that may lead to apoptosis. Curr Biol 2002;12:53–58.
Wang HY. WNT-frizzled signaling via cyclic GMP. Front Biosci 2004;9:1043–1047.
Van Den Berg DJ, Sharma AK, Bruno E et al. Role of members of the Wnt gene family in human hematopoiesis. Blood 1998;92:3189–3202.(Department of Medicine, V)
c Department of Medicine and Comprehensive Cancer Center, University of California at San Francisco, California, USA
Key Words. Hematopoiesis ? Hematopoietic stem cells (HSCs) ? Homeobox genes ? Microarray ? Gene expression profiling
Correspondence: H. Jeffrey Lawrence, M.D., Department of Medicine, Hematology Research (151H), Veterans Affairs Medical Center, 4150 Clement St., San Francisco, CA 94121, USA. Telephone: 415-221-4810, ext. 3340; Fax: 415-750-6959; e-mail: jeffl@medicine.ucsf.edu
ABSTRACT
The 39 members of the HOX family of homeobox genes encode DNA-binding proteins, which play a key role in pattern formation along various body axes . The highly regulated deployment of HOX genes in complex overlapping domains over time and space during embryogenesis appears to be critical for the correct positioning of body appendages, strongly suggesting that a combinatorial HOX code may determine the identity of specific body segments. Several genes of the HOXA and HOXB clusters, including two adjacent genes HOXA9 and HOXA10, are expressed in primitive human and murine hematopoietic cells, and that expression is downregulated as cells differentiate, suggesting a role in early blood cell development .
The function of HOX genes in normal murine hematopoiesis has been explored both in knockout mice and in murine models of retrovirally driven overexpression. HOXA9 has an important role in normal hematopoiesis, as HOXA9-deficient mice have a variety of myeloid and lymphoid defects, as well as abnormalities in hematopoietic stem cell (HSC) function . By contrast, no consistent hematopoietic abnormalities have been described in mice lacking the HOXA10 gene. These findings suggest that, although both HOXA9 and HOXA10 are expressed in primitive hematopoietic cells, the two genes have distinct functions during normal blood cell differentiation. In accord with this notion is the fact that HOXA9 and HOXA10 proteins have little homology to one another outside of the 60-amino-acid DNA-binding homeodomain (HD).
However, overexpression of either HOXA9 or HOXA10 in murine marrow cells produces similar blood cell abnormalities, including a marked expansion of HSCs and committed progenitors with eventual transformation to acute myeloid leukemia (AML) . These similar biologic effects on blood cell development suggest that HOXA9 and HOXA10 modulate common genetic programs in blood cells when they are over-expressed. Moreover, in recent studies, we have documented striking overlap in the functional effects of NUP98-AbdB class HOX fusions, as tested with naturally occurring leukogenic fusion such as NUP98-HOXD13 and engineered fusions such as NUP98-HOXA10 . Such findings further suggest a high degree of target gene overlap, at least for HOX within the AbdB class of HOX (paralogs 9–13).
Since homeobox genes encode DNA-binding proteins, it has been presumed that HOX proteins function as transcription factors. Recent papers have described the transcriptional profiles of certain HD proteins using microarray and real-time reverse transcription polymerase chain reaction (RT-PCR) technologies in an effort to identify potential gene targets and their associated downstream molecular pathways. These studies have examined the genes modulated by the expression of HOXC13, HOXD3, the non-HOX HD proteins, PAX6, and the fly HOX1 homologue, labial, in nonhematopoietic tissues . Another study examined the expression profile of the leukemogenic fusion gene, NUP98-HOXA9, in an immortalized myeloid cell line , but no study published to date has attempted to identify genes modulated by a wild-type HOX protein in primary human hematopoietic cells.
In a recent paper, our laboratory published a description of the HOXA9 transcriptome in human leukemic cell lines, using a transient overexpression strategy in three cell lines, two myeloid and one lymphoid . In that study we observed modulation of a large number of genes within 24 hours of introduction of a HOXA9 expression vector in these cells. The modulated genes represented a wide variety of functional groups, including oncogenes, cell-cycle proteins, enzymes, membrane proteins, and other transcription factors. Interestingly, a number of these genes are known to be part of the transcriptome of normal HSCs and to be similarly modulated in primary samples of human AML, suggesting that these genes are authentic biologic targets of HOXA9 .
Remarkably, in that study the gene-expression profiles observed for the myeloid and lymphoid lines were dramatically different, indicating that the transcriptional effects of HOXA9 are highly dependent on cell context. This observation raises questions as to whether the HOXA9 targets identified in aneuploid immortalized myeloid cell lines would match the gene targets for HOXA9 in normal hematopoietic cells. To answer this question, we have studied the expression profile of the HOXA9 protein in human umbilical cord CD34+ cells. In addition, we have used this same system to examine the transcriptome of the related HOXA10 protein, permitting a comparison of the genes modulated by these two closely related transcription factors.
MATERIALS AND METHODS
Identification of Genes Modulated by Overexpression of HOXA9 and HOXA10
To identify and compare the genes modulated by HOXA9 and HOXA10 in normal human hematopoietic cells, we used a retroviral overexpression strategy in CD34+ cells obtained from human umbilical CB (Fig. 1 shows a schema of the experimental design). Real-time RT-PCR of RNA from day-3 samples revealed that a six- to sevenfold increase in mRNA levels of HOXA9 and HOXA10 (Fig. 2A) was achieved in the transduced cells. CB mRNA from HOXA9- and HOXA10-transduced samples were compared with MIG only–transduced controls through hybridization to cDNA microarrays to provide a differential gene expression profile. These microarray data were then subjected to stepwise statistical analysis.
Figure 1. Schematic diagram of transduction strategy for human CD34+ cells. Human umbilical cord cells were prestimulated with growth factors, aliquoted, and transduced with one of three retroviral constructs. Green fluorescent protein–positive (GFP+)/CD34+ cells were sorted from each transduction. Total RNA was isolated, and mRNA was amplified by two rounds of in vitro transcription. Gene-expression differences were analyzed by comparing either GFP control versus HOXA9-GFP or GFP control versus HOXA10-GFP samples, using cDNA microarrays. Abbreviations: IL, interleukin; MSCV, murine stem cell virus.
Figure 2. Genes modulated by HOXA9 and HOXA10 as identified by significance analysis of microarrays (SAM) and Eisen Cluster analysis. (A): Bar chart representing the average of triplicate quantitative reverse transcription polymerase chain reaction (QRT-PCR) measurements of HOXA9 or HOXA10 mRNA in HOX gene–green fluorescent protein–positive (GFP+)–transduced versus GFP+ only–transduced cells. The bars represent standard error. (B): This SAM plot represents modulated genes that are shared between HOXA9 day-3 and HOXA10 day-3 datasets. SAM data for all genes that were deemed significant were ranked by the magnitude of their observed "d" scores, or difference from the comparator group. The MIG-control transfected group was the comparator group for changes in common in both HOXA9 and HOXA10. The differences from the comparator group were in either direction: upregulated genes (depicted in red) or downregulated (green). The delta was set to support a false discovery rate of 1% or less. Genes with expression levels that are statistically beyond delta in either direction are plotted either above (induced, red) or below (repressed, green) the comparator or control group. Genes whose expressions did not change more than the set delta in either direction were considered to be not statistically significantly different at the set false discovery rate. (C): Cluster analysis of cDNA microarray data. HOXA9 and HOXA10 microarray data were analyzed by the Eisen Hierarchical Cluster program and visualized with TreeView. The cluster shown represents 115 genes with a minimum of four data points that were at least twofold up- or downregulated. Red, induced; green, repressed; black, no change; gray, missing data. Group i includes induced genes common to HOXA9 and HOXA10; group ii includes genes that are induced by HOXA9 and repressed by HOXA10; group iii includes genes that are activated by HOXA10 and are downregulated by HOXA9; group iv includes repressed genes common to HOXA9 and HOXA10. Data shown for transduced cells harvested at day 3.
First, statistical analysis was performed using the SAM algorithm, which considers all genes with even small changes in expression. SAM identified a large number of genes of similar expression patterns in HOXA9 and HOXA10 day 3–transduced samples (Fig. 2A). Of the approximately 42,000 cDNA array clones, 1,998 genes were similarly modulated by both HOXA9 and HOXA10, as compared with the MIG control. By setting the value at 0.614, the FDR was 0.87%, and only 17 out of 1,998 genes were predicted to be false. Thus, SAMs show that these related HOX proteins are able to modulate many of the same genes. The small number of differentially modulated genes (i.e., upregulated by HOXA9 and downregulated by HOXA10, or the reverse) was represented almost entirely by expressed sequence tags of unknown function.
In a second step, Eisen Cluster analysis was performed, introducing threshold values to increase the likelihood of the identified genes being truly regulated and of biological significance. (Fig. 2B). Triplicate microarray data from day 3–harvested HOXA9- or HOXA10-transduced samples were filtered by the Eisen Cluster analysis program for genes, with four out of six observations showing twofold induction or repression. Using these more stringent criteria, 115 genes were modulated by one or both of the over-expressed genes. The gene expression profiles for HOXA9 and HOXA10 are shown in Figure 2C, broken down into four groups (i–iv). Group i consists of genes that are similarly regulated by HOXA9 and HOXA10. Groups ii and iii show genes that are differentially modulated by HOXA9 and HOXA10—for example, upregulated by HOXA9 yet downregulated by HOXA10. The largest group, group iv, represents genes that are repressed by both HOXA9 and HOXA10. This Eisen Cluster analysis suggests that there are many more genes regulated in a similar fashion by HOXA9 and HOXA10 than are differentially regulated.
Annotated genes that passed the Eisen Cluster filter for at least twofold differential expression, which were similarly regulated in more than one cord, and which were also determined significant by SAM, are presented in Tables 1 and 2. Table 1 is a list of 35 named genes that were either induced or repressed by HOXA9 and were also modulated by HOXA10. Table 2 lists genes that were modulated by HOXA10 after 3 days and remained modulated or were not modulated until 6 days of overexpression. While several genes were shown to be modulated at both time points, a larger number of genes was identified in the day-6 sample, which probably reflects the progressive modulation of secondary and tertiary genes over time. Several mRNAs that were previously detected in CD34+ cells are noted in both tables, showing that both HOX proteins modulate a number of genes that are active in HSCs .
Table 1. Genes modified by HOXA9 overexpression in human CD34+ cord blood cells
Table 2. HOXA10-modulated genes in human CD34+ cord blood cells
HOXA9 and HOXA10 Modulate Genes Involved in HSC Functions
Of the genes found to be significantly upregulated by overexpression of both HOXA9 and HOXA10, four genes were chosen for validation by real-time RT-PCR because of prior data showing these genes to be subject to modulation by HOXA9 in immortalized cells lines and/or because these genes were known to be expressed in HSCs. All of these genes were validated by triplicate analyses of RNA from HOXA10 day-3 samples (Fig. 3). The amplitude of the difference in gene expression as measured by real-time RT-PCR was often greater than that observed by microarray analysis, though results were always concordant in the direction of change.
Figure 3. QRT-PCR confirms the changes in gene expression that were seen in the microarray studies. The averages of triplicate gene-expression changes were measured by real-time RT-PCR, then calculated by the delta-delta Ct method, using the HOXA10 (day-3) sample, and compared with the fold changes seen on the microarrays. Gray, QRT-PCR; black, microarray analysis. Error bars represent standard error.
Three of the genes studied, ALDH1, ERG, and VLCS-H1, have all been reported to be expressed in CD34+ cells, compared with differentiated populations, suggesting an important role for these genes in normal HSC biology . VLCS-H1, or very long–chain acyl-CoA synthetase, had a 10-fold induction measuring with real-time RT-PCR, while ETS-related gene (ERG) showed an average 2.4-fold induction in HOXA10-transfected cells. While ERG expression was not found to be upregulated in our previous study with myeloid cell lines that overexpress HOXA9, a closely related gene v-ETS2 was. Aldehyde dehydrogenase 1 (ALDH1), which showed a 6.3-fold induction of expression by real-time RT-PCR, had been shown to be a target for HOXA9 in our previous microarray analysis, which included transient transcription assays with an ALDH1 reporter construct. IRX3, or Iroquois-class HD protein 3, which showed a 3.9-fold induction of gene expression by HOXA10 using real-time RT-PCR, was shown to be an early gene target in our previous study but has not been identified as part of the normal stem cell transcriptome. However, it is noteworthy that another member of the Iroquois family, IRX5, has been shown to be positively regulated by HOXB4 , another HOX protein known to be expressed in hematopoietic cells and to have the capacity to expand HSCs.
Activation of the Wnt Signaling Pathway by HOXA9 and HOXA10
HOXA9 and HOXA10 were found to positively regulate genes in the Wnt signaling pathway (Fig. 4). Wnt10B demonstrated a 4.5-fold increase in HOXA9-transduced cells, with a 3.1-fold increase in HOXA10-transduced cells, and this upregulation was validated by real-time PCR. In addition, two Wnt receptors, Frizzled 1 and Frizzled 5, were also upregulated, with Frizzled 1 showing approximately twofold upregulation by both HOXA9 and HOXA10, and Frizzled 5 showing two- to fourfold upregulation by HOXA10; these results were also validated by real-time PCR. The ability of these two HOX proteins to activate Wnt and Frizzled genes, given the key role of this pathway in stem cell self-renewal, may explain, at least in part, the biologic effects of HOX proteins on primitive hematopoietic cells.
Figure 4. HOXA9 and HOXA10 activate genes in the Wnt signaling pathway. Both (A) HOXA9 and (B) HOXA10 induced expression of Wnt pathway components Wnt10B, Frizzled 1 (FZD1), and Frizzled 5 (FZD5). The averages of triplicate gene-expression changes were measured by QRT-PCR using SYBR green I dye chemistry (gray) and compared with microarray analysis (black). Error bars represent standard error.
HOXA9 and HOXA10 Repress as Well as Activate Gene Expression
The Eisen Cluster analysis in Figure 2 and Tables 1 and 2 show that many genes are downregulated by both HOX proteins, as was seen in our previous expression profile study in hematopoietic cell lines. Among the downregulated genes in this survey is CYBB, a respiratory burst oxidase component pg91(phox), which is known to be expressed in maturing myeloid cells and has previously been shown to be repressed by HOXA10 (Table 2) . Other significantly downregulated genes included Tip60 and vimentin. Vimentin is an intermediate filament protein that, in blood cells, increases with monocytic differentiation. Support for the notion that the vimentin gene could be a target of HOXA proteins comes from the observation that its expression is lost in ovarian tumors that overexpress HOXA7 . Furthermore, the vimentin promoter and enhancer region has been shown to have binding sites for HOX proteins . Thus, at least two genes associated with myeloid differentiation are inhibited by HOXA9 and/or HOXA10. Tip60 is a 60-kDa Tat– interactive protein that normally represses STAT3 and is involved in sensing DNA damage and inducing apoptosis . Thus down-regulation of Tip60 could be associated with improved survival and increased genomic instability, thereby predisposing to malignant transformation.
Inhibition of the Erythroid Differentiation Program by HOXA9 and HOXA10
To identify other molecular pathways affected by HOXA9 and HOXA10, these microarray data were also analyzed with the Gen-MAPP program designed by the Gladstone Institute in affiliation with the University of California at San Francisco (http://www.genmapp.org/download.asp). This analysis demonstrated that HOXA9 (1) and HOXA10 (3) inhibit enzymes of the heme bio-synthetic pathway. Likewise, HOXA9 (4) and HOXA10 (2) also inhibit globin genes (Fig. 5). Not included in the figure is inhibition of the pathway rate-limiting enzyme, ALAS-2 by HOXA10 (Table 2). SAM showed that HOXA9 and HOXA10 also down-regulate globin genes expressed from the ?-globin and -globin loci (Tables 1 and 2; Fig. 5). Thus, both HOXA9 and HOXA10 appear to effect a general repression of erythroid-specific genes, and this is consistent with previous observations .
Figure 5. HOXA9 and HOXA10 inhibit the expression of components of the heme biosynthetic pathway and globin. Analysis of microarray data by Gen MAPP (Gene MicroArray Pathway Profiler) revealed that one enzyme of the heme synthetic pathway was downregulated at least twofold by HOXA9 and threefold by HOXA10 (days 3 and 6). Two hemoglobin genes were identified by Eisen Cluster analysis as being repressed by both HOXA9 and HOXA10 overexpression. Two additional hemoglobins were repressed by HOXA9 overexpression alone. See Tables 1 and 2.
DISCUSSION
Special thanks are given to Dr. Neal Fischbach for his comments and thoughtful reading of the manuscript. Dr. Lawrence is a VA Career Development award recipient. This study was supported by NIH grant DK48642 (H.J.L., R.K.H.), a grant from the Veterans Affairs Administration (H.J.L.), and a grant from the National Cancer Institute of Canada with funds from the Terry Fox Foundation (R.K.H.).
REFERENCES
Krumlauf R. Hox genes in vertebrate development. Cell 1994;78:191–201.
Sauvageau G, Lansdorp PM, Eaves CJ et al. Differential expression of homeobox genes in functionally distinct CD34+ subpopulations of human bone marrow cells. Proc Natl Acad Sci U S A 1994;91:12223–12227.
Pineault N, Helgason CD, Lawrence HJ et al. Differential expression of Hox, Meis1, and Pbx1 genes in primitive cells throughout murine hematopoietic ontogeny. Exp Hematol 2002;30:49–57.
Lawrence HJ, Helgason CD, Sauvageau G et al. Mice bearing a targeted interruption of the homeobox gene HOXA9 have defects in myeloid, erythroid, and lymphoid hematopoiesis. Blood 1997;89:1922–1930.
Izon DJ, Rozenfeld S, Fong ST et al. Loss of function of the homeobox gene Hoxa-9 perturbs early T-cell development and induces apoptosis in primitive thymocytes. Blood 1998;92:383–393.
Thorsteinsdottir U, Sauvageau G, Hough MR et al. Overexpression of HOXA10 in murine hematopoietic cells perturbs both myeloid and lymphoid differentiation and leads to acute myeloid leukemia. Mol Cell Biol 1997;17:495–505.
Kroon E, Krosl J, Thorsteinsdottir U et al. Hoxa9 transforms primary bone marrow cells through specific collaboration with Meis1a but not Pbx1b. EMBO J 1998;17:3714–3725.
Thorsteinsdottir U, Mamo A, Kroon E et al. Over expression of the myeloid leukemia-associated Hoxa9 gene in bone marrow cells induces stem cell expansion. Blood 2002;99:121–129.
Pineault N, Abramovich C, Ohta H et al. Differential and common leukemogenic potentials of multiple NUP98-Hox fusion proteins alone or with Meis1. Mol Cell Biol 2004;24:1907–1917.
Tkatchenko AV, Visconti RP, Shang L et al. Overexpression of Hoxc13 in differentiating keratinocytes results in downregulation of a novel hair keratin gene cluster and alopecia. Development 2001;128:1547–1558.
Chauhan BK, Zhang W, Cveklova K et al. Identification of differentially expressed genes in mouse Pax6 heterozygous lenses. Invest Ophthalmol Vis Sci 2002;43:1884–1890.
Leemans R, Loop T, Egger B et al. Identification of candidate downstream genes for the homeodomain transcription factor Labial in Drosophila through oligonucleotide-array transcript imaging. Genome Biol 2001;2:1–15.
Miyazaki YJ, Hamada J, Tada M et al. HOXD3 enhances motility and invasiveness through the TGF-beta-dependent and -independent pathways in A549 cells. Oncogene 2002;21:798–808.
Ghannam G, Takeda A, Camarata T et al. The oncogene Nup98-HOXA9 induces gene transcription in myeloid cells. J Biol Chem 2004;279:866–875.
Dorsam ST, Ferrell CM, Dorsam GP et al. The transcriptome of the leukemogenic homeoprotein HOXA9 in human hematopoietic cells. Blood 2004;103:1676–1684.
Gomes I, Sharma TT, Mahmud N et al. Highly abundant genes in the transcriptosome of human and baboon CD34 antigen-positive bone marrow cells. Blood 2001;98:93–99.
Hawley RG, Lieu FH, Fong AZ, et al. Versatile retroviral vectors for potential use in gene therapy. Gene Ther 1994;1:136–138.
Baugh LR, Hill AA, Brown EL et al. Quantitative analysis of mRNA amplification by in vitro transcription. Nucleic Acids Res 2001;29:E29.
Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 2001;98:5116–5121.
Vandesompele J, De Paepe A, Speleman F. Elimination of primer-dimer artifacts and genomic coamplification using a two-step SYBR green I real-time RT-PCR. Anal Biochem 2002;303:95–98.
Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001;25:402–408.
Wagner W, Ansorge A, Wirkner U et al. Molecular evidence for stem cell function of the slow-dividing fraction among human hematopoietic progenitor cells by genome-wide analysis. Blood 2004;104:675–686.
Theokli C, Morsi El-Kadi AS, Morgan R. TALE class homeodomain gene Irx5 is an immediate downstream target for Hoxb4 transcriptional regulation. Dev Dyn 2003;227:48–55.
Lu Y, Goldenberg I, Bei L et al. HoxA10 represses gene transcription in undifferentiated myeloid cells by interaction with histone deacetylase 2. J Biol Chem 2003;278:47792–47802.
Naora H, Montz FJ, Chai CY et al. Aberrant expression of homeobox gene HOXA7 is associated with müllerian-like differentiation of epithelial ovarian tumors and the generation of a specific autologous antibody response. Proc Natl Acad Sci U S A 2001;98:15209–15214.
Benazzouz A, Duprey P. The vimentin promoter as a tool to analyze the early events of retinoic acid-induced differentiation of cultured embryonal carcinoma cells. Differentiation 1999;65:171–180.
Xiao H, Chung J, Kao HY et al. Tip 60 is a co-repressor for STAT3. J Biol Chem 2003;278:11197–11204.
Bernstein C, Bernstein H, Payne CM et al. DNA repair/pro-apoptotic dual-role proteins in five major DNA repair pathways: fail-safe protection against carcinogenesis. Mutat Res 2002;511:145–178.
Buske C, Feuring-Buske M, Antonchuk J et al. Overexpression of HOXA10 perturbs human lymphomyelopoiesis in vitro and in vivo. Blood 2001;97:2286–2292.
Gnatenko DV, Dunn JJ, McCorkle SR et al. Transcript profiling of human platelets using microarray and serial analysis of gene expression. Blood 2003;101:2285–2293.
Shim MH, Hoover A, Blake N et al. Gene expression profile of primary human CD34+CD38low cells differentiating along the megakaryocyte lineage. Exp Hematol 2004;32:638–648.
Thompson A, McGonigle G, Grier D et al. Defining Complete HOX Signatures in Murine Hematopoietic Cells by Q-PCR. Proc Amer Soc Hematol Blood 2003;202:2100a.
Fischbach NA, Rozenfeld S, Shen W et al. HOXB6 over-expression in murine bone marrow immortalizes a myelomonocytic precursor in vitro and causes hematopoietic stem cell expansion and acute myeloid leukemia in vivo. Blood 2005;105:1456–1466.
Crooks GM, Fuller J, Petersen D et al. Constitutive HOXA5 expression inhibits erythropoiesis and increases myelopoiesis from human hematopoietic progenitors. Blood 1999;94:519–528.
Shen W, Chrobak D, Krishnan K et al. HOXB6 protein is bound to CREB-binding protein and represses globin expression in a DNA binding-dependent, PBX interaction-independent process. J Biol Chem 2004;279:39895–39904.
Kappen C. Disruption of the homeobox gene Hoxb-6 in mice results in increased numbers of early erythrocyte progenitors. Am J Hematol 2000;65:111–118.
Sengupta PK, Lavelle DE, DeSimone J. The 87-kD A-globin enhancer-binding protein is a product of the HOXB2 (HOX2H) locus. Blood 1994;83:1420–1427.
Gomez-Skarmeta JL, Modolell J. Iroquois genes: genomic organization and function in vertebrate neural development. Curr Opin Genet Dev 2002;12:403–408.
Guidato S, Prin F, Guthrie S. Somatic motoneuron specification in the hindbrain: the influence of somite-derived signals, retinoic acid and Hoxa3. Development 2003;130:2981–2996.
Peters T, Ausmeier K, Dildrop R et al. The mouse Fused toes (Ft) mutation is the result of a 1.6-Mb deletion including the entire Iroquois B gene cluster. Mamm Genome 2002;13:186–188.
van der Hoeven F, Schimmang T, Volkmann A et al. Programmed cell death is affected in the novel mouse mutant Fused toes (Ft). Development 1994;120:2601–2607.
Volkmann A, Doffinger R, Ruther U et al. Insertional mutagenesis affecting programmed cell death leads to thymic hyperplasia and altered thymopoiesis. J Immunol 1996;156:136–145.
Anderson MK, Hernandez-Hoyos G, Diamond RA et al. Precise developmental regulation of Ets family transcription factors during specification and commitment to the T cell lineage. Development 1999;126:3131–3148.
Pereira DS, Dorrell C, Ito CY et al. Retroviral transduction of TLS-ERG initiates a leukemogenic program in normal human hematopoietic cells. Proc Natl Acad Sci U S A 1998;95:8239–8244.
Faergeman NJ, Knudsen J. Role of long-chain fatty acyl-CoA esters in the regulation of metabolism and in cell signalling. Biochem J 1997;323( Pt 1):1–12.
Storms RW, Trujillo AP, Springer JB et al. Isolation of primitive human hematopoietic progenitors on the basis of aldehyde dehydrogenase activity. Proc Natl Acad Sci U S A 1999;96:9118–9123.
Baum C, Fairbairn LJ, Hildinger M et al. New perspectives for cancer chemotherapy by genetic protection of haematopoietic cells. Expert Rev Mol Med 1999;1999:1–28.
Russo J, Barnes A, Berger K et al. 4-(N,N-dipropylamino) benzaldehy de inhibits the oxidation of all-trans retinal to all-trans retinoic acid by ALDH1A1, but not the differentiation of HL-60 promyelocytic leukemia cells exposed to all-trans retinal. BMC Pharmacol 2002;2:4.
Purton LE, Bernstein ID, Collins SJ. All-trans retinoic acid delays the differentiation of primitive hematopoietic precursors (lin–c-kit+Sca+) while enhancing the terminal maturation of committed granulocyte/monocyte progenitors. Blood 1999;94:483–495.
Purton LE, Bernstein ID, Collins SJ. All-trans retinoic acid enhances the long-term repopulating activity of cultured hematopoietic stem cells. Blood 2000;95:470–477.
Kastner P, Lawrence HJ, Waltzinger C et al. Positive and negative regulation of granulopoiesis by endogenous RARalpha. Blood 2001;97:1314–1320.
Lako M, Lindsay S, Lincoln J et al. Characterisation of Wnt gene expression during the differentiation of murine embryonic stem cells in vitro: role of Wnt3 in enhancing haematopoietic differentiation. Mech Dev 2001;103:49–59.
Reya T, Duncan AW, Ailles L et al. A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 2003;423:409–414.
Lisovsky M, Itoh K, Sokol SY. Frizzled receptors activate a novel JNK-dependent pathway that may lead to apoptosis. Curr Biol 2002;12:53–58.
Wang HY. WNT-frizzled signaling via cyclic GMP. Front Biosci 2004;9:1043–1047.
Van Den Berg DJ, Sharma AK, Bruno E et al. Role of members of the Wnt gene family in human hematopoiesis. Blood 1998;92:3189–3202.(Department of Medicine, V)