Stem Cells
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

First published online August 23, 2007
Stem Cells Vol. 25 No. 12 December 2007, pp. 3111 -3120
doi:10.1634/stemcells.2007-0250; www.StemCells.com
© 2007 AlphaMed Press

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
2007-0250v1
25/12/3111    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Reprints/Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Graham, S. M.
Right arrow Articles by Graham, G. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Graham, S. M.
Right arrow Articles by Graham, G. J.

TISSUE-SPECIFIC STEM CELLS

Transcriptional Analysis of Quiescent and Proliferating CD34+ Human Hemopoietic Cells from Normal and Chronic Myeloid Leukemia Sources

Susan M. Grahama, J. Keith Vassb, Tessa L. Holyoakea, Gerard J. Grahamc

aExperimental Haematology, Division of Cancer Sciences, University of Glasgow, Glasgow Royal Infirmary, Glasgow, United Kingdom;
bSection for Bioinformatics, Cancer Research U.K., Beatson Laboratories, Garscube Estate, Glasgow, United Kingdom;
cDivision of Immunology, Infection and Inflammation, University of Glasgow, Glasgow Biomedical Research Centre, Glasgow, United Kingdom

Key Words. Stem cells • Hemopoiesis • Leukemia • Chemokines

Correspondence: Gerard J. Graham, Ph.D., Division of Immunology, Infection and Inflammation, University of Glasgow, Glasgow Biomedical Research Centre, 120 University Place, Glasgow G12 8TA, United Kingdom. Telephone: 44-141-330-3982; Fax: 44-141-330-4297; e-mail: g.graham{at}clinmed.gla.ac.uk; Tessa L. Holyoake, Experimental Haematology, Division of Cancer Sciences, University of Glasgow, Glasgow Royal Infirmary, 10 Alexandra Parade, Glasgow G31 2ER, United Kingdom. Telephone: 44-141-211-4676; Fax: 44-141-211-0414; e-mail: Tlh1g{at}clinmed.gla.ac.uk

Received on April 4, 2007; accepted for publication on August 14, 2007.

First published online in STEM CELLS EXPRESS  August 23, 2007.

   ABSTRACT
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
Quiescent and dividing hemopoietic stem cells (HSC) display marked differences in their ability to move between the peripheral circulation and the bone marrow. Specifically, long-term engraftment potential predominantly resides in the quiescent HSC subfraction, and G-CSF mobilization results in the preferential accumulation of quiescent HSC in the periphery. In contrast, stem cells from chronic myeloid leukemia (CML) patients display a constitutive presence in the circulation. To understand the molecular basis for this, we have used microarray technology to analyze the transcriptional differences between dividing and quiescent, normal, and CML-derived CD34+ cells. Our data show a remarkable transcriptional similarity between normal and CML dividing cells, suggesting that the effects of BCR-ABL on the CD34+ cell transcriptome are more limited than previously thought. In addition, we show that quiescent CML cells are more similar to their dividing counterparts than quiescent normal cells are to theirs. We also show these transcriptional differences to be reflected in the altered proliferative activity of normal and CML CD34+ cells. Of the most interest is that the major class of genes that is more abundant in the quiescent cells compared with the dividing cells encodes members of the chemokine family. We propose a role for chemokines expressed by quiescent HSC in the orchestration of CD34+ cell mobilization.

Disclosure of potential conflicts of interest is found at the end of this article.


   INTRODUCTION
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
Under normal circumstances, the majority of hemopoietic stem cells (HSC) are quiescent (henceforth referred to as G0 HSC) [1, 2]. Although the reasons for this extensive quiescence are not clear, it has become apparent that major functional differences exist between dividing and G0 HSC [3, 4] that are reflected in their homing and mobilization ability [5–7]. G0 HSC have been shown to be preferentially mobilized by granulocyte colony-stimulating factor (G-CSF), and other agents and also show preferential homing to the bone marrow (BM) and more efficient engraftment compared with dividing HSC [7, 8]. Thus, quiescence is important for the migration of HSC into, and out of, the BM microenvironment [9].

In contrast, chronic myeloid leukemic (CML) stem cells are constitutively present in the circulation. We have also recently shown the presence of a significant quiescent (G0) population within the CD34+ population of hemopoietic cells from patients with CML. These G0, CD34+, CML cells are fully functional and are able to repopulate sublethally irradiated NOD/SCID mice [10]. This quiescent pool is likely to be therapeutically relevant as, in contrast to proliferating CD34+ CML cells, it appears to be resistant to the BCR-ABL inhibitor imatinib mesylate [11]. This suggests that these cells may be important mediators of post-therapy relapse in patients with CML [12–15].

To understand the molecular basis for the differences in migratory properties of normal and leukemic and G0, and dividing HSC, we have examined the transcriptome of human CD34+ peripheral blood (PB) cells sorted for cell cycle position. To do this, we have isolated G0 and dividing CD34+ hemopoietic cell populations from normal and patient samples using Hoechst (HST) and pyronin Y (Py) staining [10–12, 15]. This has allowed us to define distinct gene expression profiles for CD34+ cells at specific cell cycle positions and to contrast the transcriptional profile of normal and CML cells in both proliferative states. This study represents the first assessment of the transcriptome of primary quiescent and proliferating CD34+ hemopoietic cells and of its alteration in leukemia, and we believe that the data set we report will be of value in advancing our understanding of the transcriptional profile of primitive hemopoietic cells. Furthermore, our data highlight differences in the expression of genes reflecting the proliferative activity of normal and CML CD34+ cells and identify chemokines as the major class of genes that is differentially expressed between G0 and dividing CD34+ hemopoietic cells, both normal and leukemic.


   MATERIALS AND METHODS
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
Patient Samples
Leukapheresis products were obtained from five chronic phase CML patients at diagnosis and before treatment. Normal samples were leukapheresis products from three G-CSF mobilized donors. Informed consent was obtained according to an ethical committee-approved protocol. Of the five CML patients, three were male and two were female, and of the three normal samples, two were obtained from male donors and one from a female donor.

CD34+ Cell Selection
CD34+ cells were enriched using the Isolex 50 system (NEXELL International, Brussels, Belgium, http://www.nexellinc.com) and cryopreserved in 10% dimethyl sulfoxide (Sigma-Aldrich, Poole, U.K., http://www.sigmaaldrich.com) in ALBA (4.5% human albumin solution; Scottish National Blood Transfusion Service). Flow cytometry analysis demonstrated that the enriched CML cells used for Hoechst/pyronin Y sorting were between 85% and 98% CD34+ and the normal cells between 76% and 85% CD34+. The CD34+ CML cells were also shown to be between 93% and 100% BCR-ABL-positive by fluorescence in situ hybridization.

Cell Culture
Cells were recovered from liquid nitrogen and washed once in phosphate-buffered saline (PBS)/2% fetal calf serum (FCS; Life Technologies, Paisley, U.K., http://www.lifetech.com) (PBS/2%). Cells were cultured overnight in Iscove's modified Dulbecco's medium (Sigma-Aldrich) supplemented with bovine serum albumin, insulin, and transferrin serum substitute (StemCell Technologies, Vancouver, BC, Canada, http://www.stemcell.com), 40 µg/ml low-density lipoprotein, and 10–4 M 2-mercaptoethanol. Serum-free medium (SFM) was further supplemented with a five-growth factor cocktail containing recombinant human, Flt3 ligand, and stem cell factor (100 ng/ml), interleukin (IL) 3, IL6 (StemCell Technologies), and G-CSF (Chugai Pharmaceutical, London, http://www.chugai-pharm.co.jp/hc/chugai_top_en.jsp), all at 20 ng/ml.

Flow Cytometry and Cell Sorting
Cells were recovered following overnight culture, washed in PBS/2%, and stained with 10 µM HST (Cambridge Bioscience, London, http://www.bioscience.co.uk) for 90 minutes at 37°C. Py (5 µg/ml) (Sigma-Aldrich) was added and incubated at 37°C for 45 minutes. Anti-CD34-fluorescein isothiocyanate (Becton, Dickinson and Company, Oxford, U.K., http://www.bd.com) was added for 15 minutes at room temperature before the cells were washed twice in PBS/2% containing HST and Py at equivalent concentrations. Prior to fluorescence-activated cell sorting (FACS) analysis, 1 µg/ml propidium iodide (PI) (Sigma-Aldrich) was added to the final wash. Cells were resuspended in PBS/2% containing HST and Py and kept on ice in the dark until they were sorted using the FACSVantage cell sorter (Becton Dickinson). Cell populations were sorted by gating first on the PI–/CD34+ cells and then on the HSTlo/Pylo (G0) and HST+/Py+ (G1/S/G2/M) cells. For further analysis of cell division, sorted cells were stained with 1 µM carboxyfluorescein diacetate succinimidyl diester (CFSE) (Cambridge Bioscience) in PBS/2% for 10 minutes at 37°C before being washed in ice-cold PBS/20% FCS and then in PBS/2%. Cells were then cultured for up to 3 days in supplemented SFM as described. The cells were recovered and stained with CD34-PE (Becton Dickinson) and PI before FACS analysis.

Isolation of Total RNA
Cells were placed in Trizol (Invitrogen, Paisley, U.K., http://www.invitrogen.com) and stored at –80°C. Total RNA was extracted from all samples according to the manufacturer's protocol. Contaminating genomic DNA was removed using the DNA-free kit (Ambion, Austin, TX, http://www.ambion.com).

Real-Time Polymerase Chain Reaction
TaqMan (Applied Biosystems, Warrington, U.K., http://www.appliedbiosystems.com) real-time polymerase chain reaction (PCR) was used for validation of microarray data. cDNA from samples not used for microarrays (processed using the same protocol) was applied to custom microfluidics cards containing probe/primer sets for the genes of interest. The cards were run on an ABI 7900 instrument (Applied Biosystems) using standard settings. GAPDH was included on the card as a housekeeping gene, and analysis was carried out using the relative quantitation {Delta}{Delta}–2 cycle threshhold method [16].

Microarrays
RNA for application to Affymetrix gene chips (Affymetrix UK, High Wycombe, U.K., http://www.affymetrix.com) was processed as described. The quality and quantity of RNA was assessed using the BioAnalyzer 2100 (Agilent Technologies, Cheshire, U.K., http://www.agilent.com) and the NanoDrop ND-1000 (NanoDrop Technologies, Wilmington, DE, http://www.nanodrop.com). The specific gene chips used were Affymetrix U133A human gene chips, which are capable of measuring the expression of 14,500 well-characterized human genes.

Thirty-five nanograms of each sample was subject to linear amplification using the Two Cycle Target Labeling Kit (Affymetrix). Processing was carried out at the Sir Henry Wellcome Functional Genomics Facility (SHWFGF) (University of Glasgow, Glasgow, U.K.).

Microarray Analysis
Raw data were normalized using robust multichip average and used to produce lists of differentially expressed genes using rank products (RP) [17, 18]. RP involves the production of lists of up- and downregulated genes and assigning a statistical measure to these differences according to the consistency of the differential expression levels across replicates. Importantly, for single-channel arrays such as the Affymetrix gene chips used in the present study, all possible pairwise comparisons of gene expression changes are calculated, and thus the statistical analysis implicitly reflects the consistency of any reported changes across replicate chips. The gene lists prepared using RP are then processed using iterative group analysis (IGA) to identify statistically significantly changed groups of genes [19]. IGA assigns differentially expressed genes to gene groups defined according to Gene Ontology (http://www.geneontology.org) and measures the statistical strength of these changes in gene class according to both the numbers of differentially expressed genes assigned to each class and the relative position of the genes in the RP lists. Importantly, by focusing on groups of genes instead of single genes, one can use group members as internal replicates to determine statistical significance. An additional advantage of the IGA statistical process is that it does not require all members of a gene grouping to change simultaneously or even in the same direction and is thus less restrictive than other Gene Ontology-based methods. All of the above procedures were done in the SHWFGF at the University of Glasgow and comprise the routine array data analysis pipeline of that facility (http://www.gla.ac.uk/functionalgenomics/rp/affy_analysis.html).

GeneSpring software (Agilent Technologies) was used for the generation of the gene lists reported in supplemental online Tables 2–7. Lists of genes with least variance were produced using the R package (http://www.r-project.org). Other statistics were generated using Mann-Whitney tests or analysis of variance. An arbitrary cut off of threefold differential expression was selected for the presentation of the data generated from the GeneSpring-based analysis. The reason for the choice of threefold was that reducing the fold expression cutoff markedly increased the numbers of genes, in some cases to unmanageable levels. Importantly, the IGA analysis reported in Table 2 is independent of assignment of arbitrary cutoff levels and relates to the statistical robustness of the differences in gene expression as assessed using RP.


   RESULTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
Isolation of G0 and Dividing HSC
Following defrosting, CD34+-enriched cells from patient and control PB harvests were incubated overnight in SFM with growth factors to enhance recovery (Materials and Methods). The FACS plots in Figure 1 show representations of the gating strategy used to isolate patient and control CD34+ cells in two different phases of the cell cycle. Cells were gated as viable PI–, CD34+ populations (Fig. 1A) and further gated, as shown on the HST/Py plot, according to cell cycle position. Figure 1B shows a typical HST/Py plot for CD34+ CML cells, which is similar to that seen for normal cells. G0 cells exhibited low HST/Py staining and were identified as a distinct population in the HST/Py FACS plot [10–12, 15]. Cells in G1 demonstrated Py staining but were low for HST, and cells in S/G2/M were positive for both HST and Py. For the array analysis, two gates were used on the HST/Py plot, generating G0 and G1/S/G2/M fractions, with a gap between the sorting gates included to minimize cross-contamination (Fig. 1B). Ideally we would have used molecular or flow cytometric methods to confirm the quiescent nature of cells generated within the G0 gate from our five CML and three normal samples. However, the small number of cells obtained using this sorting approach precluded such subsequent analyses. Nevertheless, we have previously confirmed the quiescent nature of the normal and CML G0 populations by showing that they display high expression of p21 and low expression of Ki-67, cdc25, and cyclins D1–D3 [10, 15]. We are therefore confident that both normal and CML CD34+ cells within this sorting gate are indeed quiescent.


Figure 1
View larger version (25K):
[in this window]
[in a new window]

  		Figure 1. Strategy for the isolation of viable CD34+ cells in different phases of the cell cycle. This figure shows the gating strategy for the isolation of viable (R1), CD34+ (R2) stem cells (A) and the sorting of these cells into G0 and G1/S/G2/M phases of cell cycle using a combination of Hoechst and pyronin Y staining (B). The gates used to isolate G1 and G0 cells are shown, and the mean percentage of cells to be found within each of these gates is indicated. These percentages are means from three separate normal CD34+ cell sorts and five CML CD34+ cell sorts. Abbreviations: CML, chronic myeloid leukemia; FS, forward scatter; PI, propidium iodide; SS, side scatter.

 
Transcriptional Profiling of G0 and Dividing CD34+ Cells
Because of the low amounts of RNA obtained (supplemental online Table 1), the RNA from each of the samples had to be amplified prior to labeling and applying to Affymetrix GeneChip arrays. Note that the sample labeled normal 3 G0 did not yield sufficient cells or RNA and therefore was not included in the array analysis. To minimize any variability resulting from the RNA amplification, this step was performed at the same time and using the same linear amplification kit for each sample. Data analysis and bioinformatics were performed as described in Materials and Methods.

Comparison of the Affymetrix data identified 860 genes as being similarly and consistently expressed across all four groups (normal and CML, G0 and dividing). In addition to genes involved in general metabolic pathways, this cohort contained numerous genes associated with primitive hemopoietic cell function and self-renewal. For example, seven members of the Notch/Wnt pathways were present, as well as a number of other primitive cell-associated genes, including GATA2, RUNX1, HOXC4, ITAG6, ALDH, and E2F4 (data not shown). Many other regulators of transcription, signal transduction, and chromatin modification, all previously associated with a primitive hemopoietic cell phenotype, were also present. Thus, a common primitive hemopoietic cell transcriptional signature was evident in each of the four sorted populations.

Analysis of Differentially Expressed Genes
Table 1 shows the number of genes changed more than threefold between the sample groups; individual gene lists are presented in supplemental online Tables 2–7. The greatest numbers of genes changed were in the normal G0 versus normal dividing (188) and normal G0 versus CML dividing (332) groups, with only low numbers being changed between CML G0 and dividing cells (37), suggesting that CML G0 cells are closer to their cycling counterparts than normal G0 cells are to theirs. Interestingly, comparison of CML versus normal G0 cells revealed a high degree of transcriptional variability, again indicating a fundamental difference between quiescent normal and CML cells. In contrast, normal and CML dividing cells were transcriptionally very similar. Euclidean distance modeling of the relatedness of the sample groups (Fig. 2) confirmed that normal and CML dividing cells were closest in gene expression, with CML G0 cells transcriptionally closer to dividing cells than to normal G0 cells. These general conclusions were also borne out by IGA [19], which identified only limited numbers of gene groupings as being significantly changed between dividing normal and CML cells (Table 2).


View this table:
[in this window]
[in a new window]

  		Table 1. Numbers of genes changed more than threefold

 


Figure 2
View larger version (11K):
[in this window]
[in a new window]

  		Figure 2. Euclidean distance analysis of the closeness of the array data sets. This diagram shows the results of Euclidean distance modeling of the relatedness of the transcriptional profiles of normal and CML, G0, and dividing CD34+ cells. Specifically this diagram shows the relative expression of all genes more than threefold changed between the sample groups. Abbreviation: CML, chronic myeloid leukemia.

 


View this table:
[in this window]
[in a new window]

  		Table 2. Gene groups identified by iterative group analysis

 
Differential Expression of Cell Cycle Genes
IGA [19] and GeneSpring analyses identified cell cycle genes as representing a significant component of the transcriptional differences between normal and CML dividing and G0 cells (Tables 2, 3). Notably, in dividing normal cells, there was significant upregulation of genes associated with early (G1/S-phase) cell cycle events (Table 3) confirming that compared with the G0 cells, most dividing cells had entered the early phases of the cell cycle following overnight growth factor stimulation (Materials and Methods). In addition, these data serve as a separate validation of our quiescent/dividing gating strategy.


View this table:
[in this window]
[in a new window]

  		Table 3. Table of cell cycle genes upregulated in dividing versus G0 cells

 
Intriguingly, for CML (Tables 2, 3; supplemental online Tables 2–7), there was only limited corresponding upregulation of early cell cycle genes in the dividing versus G0 cells. There was, however, an upregulation of genes associated with later cell cycle events of mitosis, spindle formation, and DNA replication (Table 3). Although these data do not provide insights into the relative cell cycle status of unstimulated CML and normal CD34+ cells, they do suggest that the CML cells were more advanced in cell cycle than their normal counterparts following overnight cytokine stimulation (a functional analysis of this difference is given below and in Fig. 5). In addition, comparison of normal and CML G0 cells revealed higher levels of transcription of early cell cycle genes in the G0 CML cells (Table 2D), further strengthening the notion that CML G0 cells are transcriptionally more akin to proliferating cells than to their normal quiescent counterparts. The functional implications of these data are described below.

Classification of Other Differentially Expressed Genes

G0 Versus Dividing Cells.   In addition to differences in cell cycle gene expression, normal dividing CD34+ cells displayed higher levels of transcripts for genes involved in DNA repair and oxygen transport than their quiescent counterparts (Table 2). In contrast, dividing CD34+ CML cells displayed very few significant transcriptional differences versus G0 cells other than cell cycle genes (Tables 2, 3; supplemental online Tables 2–7).

Much more similarity was seen in the reciprocal comparison of genes upregulated in G0 normal and CML cells compared with their dividing counterparts. Normal G0 cells displayed increased levels of expression of many histone genes, as well as genes involved in regulation of cell death and antigen presentation, whereas CML G0 cells expressed increased levels of erythroid genes and the aldo-keto reductase genes AKR1C1 and AKR1C2.

Strikingly, the family of genes exhibiting the most significant upregulation, in both normal and CML G0 cells compared with their dividing counterparts, was the chemokine family [20] (Fig. 3). The differentially expressed chemokines were predominantly inflammatory CXC chemokines (CXCR2 and CXCR3 ligands; Fig. 3). Interestingly, among the chemokines, the CXCR2 ligands (CXCLs 1, 2, 3, and 6) were most strongly differentially expressed. These results, which we obtained from GeneSpring analysis, were confirmed by IGA analysis, which identified additional group members and ranked chemokines statistically as the top changed group in the G0 versus dividing cells for both normal and CML samples (Table 2; supplemental online Tables 2–7).


Figure 3
View larger version (26K):
[in this window]
[in a new window]

  		Figure 3. Venn diagram analysis of GeneSpring data selected for threefold, fourfold, and fivefold differences. This figure shows Venn diagrams revealing the lists of genes that show common changes in expression in the CML G0 and DIV CD34+ cells as well as in the Norm G0 and DIV CD34+ cells. Consistent gene changes apparent following a cutoff of threefold, fourfold, and fivefold differences in transcription levels are shown. Abbreviations: CML, chronic myeloid leukemia; DIV, dividing; Norm, normal.

 

CML Versus Normal CD34+ Cells.   Aside from cell cycle genes, IGA analysis did not identify any other significantly altered classes of genes as being upregulated in dividing CML compared with normal cells. GeneSpring analysis highlighted only four genes as being more highly expressed by the dividing CML cells than their normal counterparts (Table 2). In examining the genes that were expressed at lower levels in CML, IGA analysis identified genes involved in antigen presentation and processing as being the most significantly altered. In addition, the dividing CML cells expressed lower levels of chemokines than normal cells, although the differences in expression levels were low. GeneSpring analysis identified 12 genes as being expressed at level more than threefold lower in dividing CML compared with normal cells. Again, many of these were genes associated with antigen presentation and processing.

In contrast to dividing cells, the transcriptional profiles of CML and normal G0 cells were more different (Fig. 2). IGA analysis (Table 2) identified significantly higher expression of genes associated with active biological processes in CML compared with normal G0 cells. The top changed groups were hemoglobin complex and oxygen transport. GeneSpring analysis identified 83 genes expressed at a level more than threefold higher in CML compared with normal G0 cells. These genes included many involved in erythroid cell function. IGA analyses again revealed genes involved in antigen processing and presentation, as well as chemokine genes, as being expressed at lower levels in CML G0 compared with normal G0 cells. GeneSpring analysis highlighted 55 genes as being expressed at lower levels by G0 CML compared with G0 normal cells. These again included genes involved in antigen processing and presentation but also included genes such as the HSC marker AC133 [21].

Validation of Microarray Results
Given the difficulty in accessing normal samples and the very poor yields of RNA from such samples, these validation assays were performed on PB cells from three independent CML patients. Initial attempts at measuring chemokines released by these cells using enzyme-linked immunosorbent assay or Luminex approaches were unsuccessful because of the very small numbers of sorted cells. Thus, validation of the array analyses was confirmed by quantitative PCR using microfluidic cards. Expression of genes identified as being differentially expressed in CML G0 cells compared with CML dividing cells was examined using RNA from dividing and G0 sorted CD34+ CML cell populations. RNA from unsorted CD34+ CML cells was used as a calibrator and expression in G0 or dividing cells is reported relative to the expression levels found in these unsorted cells. Assessment of CD34 expression was included in this analysis, and as expected, similar expression levels were detected in the unsorted G0 and dividing cell populations (Fig. 4). Additional results (Fig. 4) confirmed the array analyses, showing differential expression of the chemokines (CCL19, CXCL13, CXCL12, CXCL3, and CXCL5) and the receptor CXCR4 in CML G0 compared with dividing cells. These quantitative PCR results therefore validate the differences in gene expression revealed by the microarray study.


Figure 4
View larger version (30K):
[in this window]
[in a new window]

  		Figure 4. Validation of the array data. Gene expression differences detected using the Affymetrix gene chips were validated using quantitative polymerase chain reaction (PCR) approaches. These data show results of TaqMan real-time PCR analysis for the genes indicated using unsorted CML CD34+ cells as the calibrator sample and comparing the expression levels with those seen in the same cells sorted for dividing and G0 subpopulations. Note the relatively unchanged expression of CD34 in the three groups but the marked differences in expression of the other candidate genes in the G0 compared with the dividing populations. GAPDH was included as a housekeeping gene, and the relative expression levels were quantified by the {Delta}{Delta}–2 CT method. These data are representative of those obtained from three separate experiments and were derived using RNA from cells from a male CML patient that were 88% CD34+ and 94% BCR-ABL-positive. Abbreviations: CML, chronic myeloid leukemia; DIV, dividing.

 
CML and Normal G0 Cells Exhibit Different Proliferation Characteristics
The differences in cell cycle gene expression by the normal and CML CD34+ cells (Tables 2, 3) would predict altered cell cycle kinetics for normal versus CML cells in response to growth factor stimulation. The question of whether CML and normal stem cells exhibit significantly different cell cycle kinetics remains controversial [22–24]. We therefore examined this aspect of normal and CML CD34+ cell biology in more detail. Results from FACS analyses (as shown in Fig. 1B) showed that the proportion of cells residing in G0 in CML was lower than that from normal CD34+ cells in G1 (1.9% compared with 3.8%; p = .04). Figure 1B also shows that there were fewer CD34+ CML cells in G1 (50.6%) compared with normal (70.9%; p = .04), indicating that a greater proportion of CD34+ CML cells were in S/G2/M. This was in keeping with the cell cycle gene expression data and in broad agreement with previous reports of a higher proliferative response of CML, compared with normal, CD34+ cells to growth factors [23, 24].

To further examine the differences in division kinetics of CML and normal CD34+ cells, they were sorted into G0 and dividing populations, stained with CFSE, and cultured in growth factor-supplemented medium. The cells were then assessed for cell division at 36 hours and again at 3 days. These time points were chosen because we have previously shown substantial entry of quiescent CML cells into cell cycle following 3 days of exposure to cytokines [10, 11]. Thirty-six hours therefore represents a time point intermediate between the initiation and the establishment of this entry into cell cycle. Figure 5 shows that for CD34+ CML cells at 36 hours, the G0 cells had undergone up to three divisions, with only 14% remaining undivided. By 3 days, almost all of the CML G0 cells had gone into division, with only 0.9% remaining in the undivided peak. In contrast, for normal cells, at 36 hours 62% of the G0 cells remained undivided, and 54% remained undivided by 3 days. The dividing CML cells also underwent up to three divisions by 36 hours, with only 1.8% remaining undivided. By 3 days, only 0.2% from the dividing sample remained undivided. In contrast, 46% of the dividing normal cells had not completed a further cell division by 36 hours, and after 3 days, 21% of the dividing cells had not divided further. These observations are consistent with the gene data, suggesting that CML cells demonstrate an increased tendency to enter cell division.


Figure 5
View larger version (27K):
[in this window]
[in a new window]

  		Figure 5. Assessment of the proliferative capacity of CD34+ normal and CML cells following cytokine stimulation. Histograms of Hoechst/pyronin Y sorted G0 and dividing cells stained with carboxyfluorescein diacetate succinimidyl diester (CFSE) and cultured for 36 hours and for 3 days. The markers (dark arrows) show the percentages of undivided cells, from either G0 or dividing cells, remaining after each time period. In addition the residual G0 and dividing populations at each time point are indicated as percentages within each panel. Note that CFSE measurements were made only within the CD34+ cell gate. Abbreviations: CML, chronic myeloid leukemia; Div, dividing; hrs, hours.

 

   DISCUSSION
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
To date, a number of gene expression studies have been performed aimed at trying to examine the transcriptional profile of the HSC [25–34]. Indeed, a number of studies have attempted to define a general molecular fingerprint for stem cells [35, 36], although this approach has been of limited success [37, 38]. In this study, we have taken the analysis of the primitive hemopoietic cell transcriptome a step further by analyzing the gene expression profile of defined, sorted cell populations within the normal and leukemic human CD34+ cell compartments, namely G0 and dividing cells. It is important to point out that the differences in our approach mean that the data we have generated cannot be easily compared with those from most other array analyses of the HSC transcriptome, although others studies have also reported enhanced expression of cell cycle-related genes in CD34+ chronic phase CML cells compared with normal CD34+ cells [31, 32]. However, there have been previous studies with the similar aim of defining the transcriptional differences between quiescent and proliferating primitive hemopoietic cells. Venezia et al. [39] attempted to address this issue by comparing gene expression patterns in resting (predominantly quiescent) and 5-fluorouracil-treated (predominantly proliferating) primitive murine BM cells. Our data show limited overlap with that of Venezia et al. [39], possibly because the source of cells used in our study was human PB and also because of the variety of experimental differences in our approaches to the preparation of the cells for the array analysis. Interestingly, however, there is significant overlap between our data set and that of Oswald et al. [40], particularly with respect to the expression of chemokine genes. In that study, Oswald et al. [40] examined gene expression in primitive hemopoietic cells grown on a fibrillar collagen matrix. In keeping with our data, they show that expression of many of the chemokines that we report here is associated with the delayed proliferation seen with primitive hemopoietic cells grown in this manner.

Examination of the transcriptional data that we generated provides a number of insights into the functional differences between normal and CML quiescent and proliferating cells. For example, analysis of the expression of genes involved in cell cycle regulation indicates differences in proliferative responses of normal and CML cells to overnight cytokine stimulation. Specifically, normal CD34+ cells express a cohort of cell cycle genes indicative of early positioning within cell cycle. In contrast, CML cells express later cell cycle genes, suggestive of a more rapid transit through cell cycle following growth factor stimulation. We examined the functional relevance of this, and we show that both G0 and dividing CML cells proliferate more rapidly following cytokine stimulation than their normal counterparts. These transcriptional and biological data are therefore highly suggestive of an intrinsic, BCR-ABL-dependent enhancement of cell cycle transit of CML compared with normal CD34+ cells. Although our findings in this regard are similar to those of other groups [23, 24], they differ significantly from the results from Buckle et al. [22]. An important difference between our experimental approach and that of Buckle et al. [22] is that we have examined the proliferative response of the normal and CML-derived CD34+ cells to growth factor stimulation and not simply in unstimulated cells. Thus, we believe that our data provide a more accurate assessment of the proliferative potential of CML versus normal CD34+ cells. Together with the other previous studies [23, 24], it therefore appears that CD34+ cells from CML sources have a higher intrinsic proliferative activity than their normal counterparts.

In addition to cell cycle genes, other gene differences were highlighted by both IGA and GeneSpring analyses. The implications of the enhanced erythroid gene expression levels in G0 CML cells are not clear at present, although they may relate to the predisposition of HSC to erythroid differentiation (demonstrated in previous studies on the hemopoietic differentiation potential of murine embryonic stem cells [41]), which appears, therefore, to be exaggerated in G0 CML CD34+ cells. G0 CML cells also expressed higher levels of AKR1C1 and AKR1C2 than their proliferating counterparts. It has been reported that the AKR1C enzyme is present in myeloid leukemic cell lines and that its activity inhibits their differentiation [42]. We postulate that expression of these genes may indicate one mechanism that maintains a primitive phenotype in CML G0 CD34+ cells and prevents exhaustion of the malignant clone.

Of the most note were the marked differences in expression of chemokine genes, which were much more strongly expressed by G0 cells than by proliferating cells. The overexpressed chemokines, in general, were ligands for CXCR2 and CXCR3. The CXCR2 ligands were most significantly overexpressed by G0 CD34+ cells and with CXCR2 being the predominant neutrophil-expressed chemokine receptor, this suggests a relationship between G0 CD34+ cells and neutrophils. Notably, all the CD34+ cells used in this study were from PB and in the case of normal cells were the product of mobilization regimens. Recent data on HSC mobilization clearly implicate neutrophils as a key cell type in the mobilization process [43]. The likelihood is that neutrophil-derived proteases are required to disrupt local molecular adhesions anchoring HSC to stromal cells within BM niches [44–46]. One outstanding question relates to the mechanism of targeting neutrophils to the HSC niche during mobilization. As CXCR2 is the major neutrophil chemokine receptor, we postulate that chemokines expressed by quiescent HSC may be involved in neutrophil recruitment to the niche during mobilization. This would account for the preferential mobilization of G0 HSC and the requirement for CXCR2 in mobilization [47] and would ensure localization of neutrophil-protease release to the BM niche. This hypothesis is currently being tested in our laboratories.


   CONCLUSION
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
Disclosure of Potential...
 Acknowledgments
 References
 
In summary, we describe here the transcriptional profiling of normal and CML-derived CD34+ cells separated on the basis of their position in the cell cycle. The most striking conclusions from our study are as follows: (a) dividing CML and normal CD34+ cells are remarkably similar in their transcriptional profiles, suggesting that the impact of BCR-ABL on the CD34+ cell transcriptome is limited; (b) quiescent CD34+ CML cells are transcriptionally and proliferatively more similar to dividing than normal CD34+ cells; (c) the major class of genes that are overexpressed in quiescent cells of either normal or CML origin compared with proliferating cells are those encoding members of the chemokine family of proinflammatory mediators. We believe that these chemokines are likely to be involved in the regulation of release of CD34+ cells from the bone marrow niches and are currently testing this hypothesis.


   DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
Acknowledgments
 References
 
The authors indicate no potential conflicts of interest.


   ACKNOWLEDGMENTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
References
 
This work was funded by grants from the Scottish Chief Scientist's Office and the Leukemia Research Fund. G.J.G. and J.K.V. are supported by grants from Cancer Research UK.


   REFERENCES
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 

  1. Bradford GB, Williams B, Rossi R et al. Quiescence, cycling, and turnover in the primitive hematopoietic stem cell compartment. Exp Hematol 1997;25:445–453.[Medline]

  2. Ponchio L, Conneally E, Eaves C. Quantitation of the quiescent fraction of longterm culture-initiating cells (LTC-IC) in normal human blood and marrow and the kinetics of their growth factor-stimulated entry into S-phase in vitro. Blood 1995;86:3314–3321.[Abstract/Free Full Text]

  3. Uchida N, He D, Friera AM et al. The unexpected G0/G1 cell cycle status of mobilized hematopoietic stem cells from peripheral blood. Blood 1997;89:465–472.[Abstract/Free Full Text]

  4. Gothot A, Pyatt R, McMahel J et al. Functional heterogeneity of human CD34+ cells isolated in subcompartments of the G0/G1 phase of the cell cycle. Blood 1997;90:4384–4393.[Abstract/Free Full Text]

  5. Glimm H, Oh IH, Eaves CJ. Human hematopoietic stem cells stimulated to proliferate in vitro lose engraftment potential during their S/G(2)/M transit and do not reenter G(0). Blood 2000;96:4185–4193.[Abstract/Free Full Text]

  6. Cashman J, Dykstra B, Clark-Lewis I et al. Changes in the proliferative activity of human hematopoietic stem cells in NOD/SCID mice and enhancement of their transplantability after in vivo treatment with cell cycle inhibitors. J Exp Med 2002;196:1141–1149.[Abstract/Free Full Text]

  7. Passegue E, Wagers AJ, Giuriato S et al. Global analysis of proliferation and cell cycle gene expression in the regulation of hematopoietic stem and progenitor cell fates. J Exp Med 2005;202:1599–1611.[Abstract/Free Full Text]

  8. Jetmore A, Plett PA, Tong X et al. Homing efficiency, cell cycle kinetics, and survival of quiescent and cycling human CD34+ cells transplanted into conditioned NOD/SCID recipients. Blood 2002;99:1585–1593.[Abstract/Free Full Text]

  9. Adams GB, Scadden DT. The hematopoietic stem cell in its place. Nat Immunol 2006;7:333–337.[CrossRef][Medline]

  10. Holyoake T, Jiang X, Eaves C et al. Isolation of a highly quiescent subpopulation of primitive leukemic cells in chronic myeloid leukemia. Blood 1999;94:2056–2064.[Abstract/Free Full Text]

  11. Graham SM, Jorgensen HG, Allan E et al. Primitive, quiescent, Philadelphia-positive stem cells from patients with chronic myeloid leukemia are insensitive to STI571 in vitro. Blood 2002;99:319–325.[Abstract/Free Full Text]

  12. Jorgensen HG, Allan EK, Graham SM et al. Lonafarnib reduces the resistance of primitive quiescent CML cells to imatinib mesylate in vitro. Leukemia 2005;19:1184–1191.[CrossRef][Medline]

  13. Drummond MW, Lush CJ, Vickers MA et al. Imatinib mesylate-induced molecular remission of Philadelphia chromosome-positive myelodysplastic syndrome. Leukemia 2003;17:463–465.[CrossRef][Medline]

  14. Jiang X, Lopez A, Holyoake T et al. Autocrine production and action of IL-3 and granulocyte colony-stimulating factor in chronic myeloid leukaemia. Proc Natl Acad Sci U S A 1999;96:12804–12809.[Abstract/Free Full Text]

  15. Holyoake TL, Jiang X, Jorgensen HG et al. Primitive quiescent leukemic cells from patients with chronic myeloid leukemia spontaneously initiate factor-independent growth in vitro in association with up-regulation of expression of interleukin-3. Blood 2001;97:720–728.[Abstract/Free Full Text]

  16. 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.[CrossRef][Medline]

  17. Irizarry RA, Hobbs B, Collin F et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 2003;4:249–264.[Abstract]

  18. Breitling R, Armengaud P, Amtmann A et al. H. Rank products: A simple, yet powerful, new method to detect differentially regulated genes in replicated microarray experiments. FEBS Lett 2004;573:83–92.[CrossRef][Medline]

  19. Breitling R, Amtmann A, Herzyk P. Iterative Group Analysis (iGA): A simple tool to enhance sensitivity and facilitate interpretation of microarray experiments. BMC Bioinformatics 2004;5:34.[CrossRef][Medline]

  20. Rot A, von Andrian UH. Chemokines in innate and adaptive host defense: Basic chemokinese grammar for immune cells. Annu Rev Immunol 2004;22:891–928.[CrossRef][Medline]

  21. Bhatia M. AC133 expression in human stem cells. Leukemia 2001;15:1685–1688.[Medline]

  22. Buckle AM, Mottram R, Pierce A et al. The effect of Bcr-Abl protein tyrosine kinase on maturation and proliferation of primitive haematopoietic cells. Mol Med 2000;6:892–902.[Medline]

  23. Traycoff CM, Halstead B, Rice S et al. Chronic myelogenous leukaemia CD34+ cells exit G0/G1 phases of cell cycle more rapidly than normal marrow CD34+ cells. Br J Haematol 1998;102:759–767.[CrossRef][Medline]

  24. Primo D, Flores J, Quijano S et al. Impact of BCR/ABL gene expression on the proliferative rate of different subpopulations of haematopoietic cells in chronic myeloid leukaemia. Br J Haematol 2006;135:43–51.[CrossRef][Medline]

  25. Akashi K, He X, Chen J et al. Transcriptional accessibility for genes of multiple tissues and hematopoietic lineages is hierarchically controlled during early hematopoiesis. Blood 2003;101:383–389.[Abstract/Free Full Text]

  26. Eckfeldt CE, Mendenhall EM, Flynn CM et al. Functional analysis of human hematopoietic stem cell gene expression using zebrafish. PLoS Biol 2005;3:e254.[CrossRef][Medline]

  27. Zhong JF, Zhao Y, Sutton S et al. Gene expression profile of murine long-term reconstituting vs. short-term reconstituting hematopoietic stem cells. Proc Natl Acad Sci U S A 2005;102:2448–2453.[Abstract/Free Full Text]

  28. Phillips RL, Ernst RE, Brunk B et al. The genetic program of hematopoietic stem cells. Science 2000;288:1635–1640.[Abstract/Free Full Text]

  29. Georgantas RW 3rd, Tanadve V et al. Microarray and serial analysis of gene expression analyses identify known and novel transcripts overexpressed in hematopoietic stem cells. Cancer Res 2004;64:4434–4441.[Abstract/Free Full Text]

  30. Ng YY, van Kessel B, Lokhorst HM et al. Gene-expression profiling of CD34+ cells from various hematopoietic stem-cell sources reveals functional differences in stem-cell activity. J Leukoc Biol 2004;75:314–323.[Abstract/Free Full Text]

  31. Kronenwett R, Butterweck U, Steidl U et al. Distinct molecular phenotype of malignant CD34(+) hematopoietic stem and progenitor cells in chronic myelogenous leukemia. Oncogene 2005;24:5313–5324.[CrossRef][Medline]

  32. Diaz-Blanco E, Bruns I, Neumann F et al. Molecular signature of CD34(+) hematopoietic stem and progenitor cells of patients with CML in chronic phase. Leukemia 2007;21:494–504.[CrossRef][Medline]

  33. Zheng C, Li L, Haak M et al. Gene expression profiling of CD34+ cells identifies a molecular signature of chronic myeloid leukemia blast crisis. Leukemia 2006;20:1028–1034.[CrossRef][Medline]

  34. Shojaei F, Trowbridge J, Gallacher L et al. Hierarchical and ontogenic positions serve to define the molecular basis of human hematopoietic stem cell behavior. Dev Cell 2005;8:651–663.[CrossRef][Medline]

  35. Ramalho-Santos M, Yoon S, Matsuzaki Y et al. "Stemness": Transcriptional profiling of embryonic and adult stem cells. Science 2002;298:597–600.[Abstract/Free Full Text]

  36. Ivanova NB, Dimos JT, Schaniel C et al. A stem cell molecular signature. Science 2002;298:601–604.[Abstract/Free Full Text]

  37. Evsikov AV, Solter D. Comment on "‘Stemness’: Transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature.". Science 2003;302–393 author reply 393.

  38. Fortunel NO, Otu HH, Ng HH et al. Comment on "‘Stemness’: Transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature.". Science 2003;302:393 author reply 393.

  39. Venezia TA, Merchant AA, Ramos CA et al. Molecular signatures of proliferation and quiescence in hematopoietic stem cells. PLoS Biol 2004;2:e301.[CrossRef][Medline]

  40. Oswald J, Steudel C, Salchert K et al. Gene-expression profiling of CD34+ hematopoietic cells expanded in a collagen I matrix. STEM CELLS 2006;24:494–500.[Abstract/Free Full Text]

  41. Baird JW, Ryan KM, Hayes I et al. Differentiating embryonal stem cells are a rich source of haemopoietic gene products and suggest erythroid preconditioning of primitive haemopoietic stem cells. J Biol Chem 2001;276:9189–9198.[Abstract/Free Full Text]

  42. Desmond JC, Mountford JC, Drayson MT et al. The aldo-keto reductase AKR1C3 is a novel suppressor of cell differentiation that provides a plausible target for the non-cyclooxygenase-dependent antineoplastic actions of nonsteroidal anti-inflammatory drugs. Cancer Res 2003;63:505–512.[Abstract/Free Full Text]

  43. Pruijt JF, Verzaal P, van Os R et al. Neutrophils are indispensable for hematopoietic stem cell mobilization induced by interleukin-8 in mice. Proc Natl Acad Sci U S A 2002;99:6228–6233.[Abstract/Free Full Text]

  44. Lapidot T, Dar A, Kollet O. How do stem cells find their way home? Blood 2005;106:1901–1910.[Abstract/Free Full Text]

  45. Papayannopoulou T. Current mechanistic scenarios in hematopoietic stem/progenitor cell mobilization. Blood 2004;103:1580–1585.[Abstract/Free Full Text]

  46. Lapidot T, Petit I. Current understanding of stem cell mobilization: The roles of chemokines, proteolytic enzymes, adhesion molecules, cytokines, and stromal cells. Exp Hematol 2002;30:973–981.[CrossRef][Medline]

  47. Pelus LM, Horowitz D, Cooper SC et al. Peripheral blood stem cell mobilization. A role for CXC chemokines. Crit Rev Oncol Hematol 2002;43:257–275.[Medline]




This article has been cited by other articles:


Home page
 BloodHome page
A. S. M. Yong, K. Keyvanfar, N. Hensel, R. Eniafe, B. N. Savani, M. Berg, A. Lundqvist, S. Adams, E. M. Sloand, J. M. Goldman, et al.
Primitive quiescent CD34+ cells in chronic myeloid leukemia are targeted by in vitro expanded natural killer cells, which are functionally enhanced by bortezomib
Blood, January 22, 2009; 113(4): 875 - 882.
[Abstract] [Full Text] [PDF]

Home page
 ASH Education BookHome page
C. H. Jamieson
Chronic Myeloid Leukemia Stem Cells
Hematology, January 1, 2008; 2008(1): 436 - 442.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
2007-0250v1
25/12/3111    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Reprints/Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Graham, S. M.
Right arrow Articles by Graham, G. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Graham, S. M.
Right arrow Articles by Graham, G. J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
STEM CELLS THE ONCOLOGIST CME ALPHAMED PRESS JOURNALS