SEARCH

SEARCH BY CITATION

Keywords:

  • Human stem cells;
  • CD133+ cells;
  • Mesenchymal stem cells;
  • Hypoxia;
  • Transcriptome;
  • Proliferation;
  • Differentiation Bone marrow;
  • Cord blood

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Umbilical cord blood (UCB) and bone marrow (BM)-derived stem and progenitor cells possess two characteristics required for successful tissue regeneration: extensive proliferative capacity and the ability to differentiate into multiple cell lineages. Within the normal BM and in pathological conditions, areas of hypoxia may have a role in maintaining stem cell fate or determining the fine equilibrium between their proliferation and differentiation. In this study, the transcriptional profiles and proliferation and differentiation potential of UCB CD133+ cells and BM mesenchymal cells (BMMC) exposed to normoxia and hypoxia were analyzed and compared. Both progenitor cell populations responded to hypoxic stimuli by stabilizing the hypoxia inducible factor (HIF)-1α protein. Short exposures to hypoxia increased the clonogenic myeloid capacity of UCB CD133+ cells and promoted a significant increase in BMMC number. The differentiation potential of UCB CD133+ clonogenic myeloid cells was unaltered by short exposures to hypoxia. In contrast, the chondrogenic differentiation potential of BMMCs was enhanced by hypoxia, whereas adipogenesis and osteogenesis were unaltered. When their transcriptional profiles were compared, 183 genes in UCB CD133+ cells and 45 genes in BMMC were differentially regulated by hypoxia. These genes included known hypoxia-responsive targets such as BNIP3, PGK1, ENO2, and VEGFA, and other genes not previously described to be regulated by hypoxia. Several of these genes, namely CDTSPL, CCL20, LSP1, NEDD9, TMEM45A, EDG-1, and EPHA3 were confirmed to be regulated by hypoxia using quantitative reverse transcriptase polymerase chain reaction. These results, therefore, provide a global view of the signaling and regulatory network that controls oxygen sensing in human adult stem/progenitor cells derived from hematopoietic tissues.

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


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Adult human bone marrow (BM) contains a diverse range of cells, including multipotent hematopoietic, vascular, and mesenchymal stem cells and their progeny. In this specialized organ, hematopoiesis is regulated by the complex interaction of CD133+ hematopoietic stem/progenitor cells (HSC/HPC) with mesenchymal stem cell-derived stromal niche cells [1]. Bone marrow mesenchymal stem cells also have the ability to form bone, cartilage, and fat, by differentiating into osteoblasts, chondrocytes, and adipocytes, respectively [2, [3], [4]5]. Thus, these mesenchymal cells have the potential to be used for tissue repair. They have also been reported to improve the bone marrow engraftment of transplanted umbilical cord blood (UCB)-derived HSC/HPC and to ameliorate graft-versus-host disease when infused during allogeneic HSC transplantation [6, 7], further extending their potential as cellular therapeutics.

UCB sourced at birth has become an accepted source of HSC/HPC for transplantation, particularly for children for whom a matched related or unrelated donor is unavailable (reviewed in [8]). Disadvantages of UCB include both limited cell numbers and delayed times to engraftment. These can be improved or overcome by cotransplantation of dual UCB units with BM mesenchymal cells (BMMC) or of UCB HSC with more mature progeny. The therapeutic potential of UCB CD133+ cells and BMMC may be improved, for instance, by expanding these cells ex vivo and modulating their homing and engraftment potential. However, the fate of these cells depends on both intrinsic and extrinsic signals [9, [10], [11]12] and is determined by the fine equilibrium between proliferative and differentiation potentials. Thus, understanding what controls the proliferation and differentiation of UCB CD133+ cells and BMMC is of great importance for their future clinical use. In this respect, it is of interest that the BM is one of the few organs of the body that is maintained in a hypoxic state [13], and hypoxia plays an important role in fetal development and cell differentiation [14]. Low oxygen levels are also associated with pathological conditions such as the formation and growth of tumors [15], including leukemias in BM [16], wounding [17], arthritic joints [18], and ischemic heart disease [19]. Although the effect of hypoxia in regulating global gene expression has been reported extensively in cancer cells, macrophages, and endothelial cells, there are significantly fewer studies that have evaluated changes in overall gene expression under hypoxic conditions in human stem and progenitor cells derived from hematopoietic organs [2]. It has been reported, for instance, that HSC/HPC increase their repopulating and in vitro clonogenic potentials when exposed to low oxygen levels (1.5% and 3% oxygen [20, 21]). Other studies show that bone marrow-derived mesenchymal cells exposed to hypoxia for more than 72 hours upregulate proangiogenic factors that, in a paracrine fashion, promote the proliferation and migration of endothelial and smooth muscle cells [22]. In addition, long-term or chronic exposure of bone marrow-derived mesenchymal stem cells to hypoxia has been reported to inhibit their adipogenic, chondrogenic, and osteogenic differentiation potentials [2, 23].

In the present study, the proliferative capacity and differentiation potential of UCB CD133+ cells and cultured BMMC after brief exposures to normoxic and hypoxic conditions were assessed. In addition, we used high-density oligonucleotide micoarrays to establish and compare the transcriptional profiles of these cells in normoxia and hypoxia. Differential gene expression analysis identified common and unique genes regulated by hypoxia in both populations, revealing expression of two complementary receptor-ligand pairs, VLA-4/FN1 and CXCR4/CXCL12, known to play a key role in hematopoietic cell homing, adhesion, and migration. These results underscore the importance of carrying out a comparative study of these two cell populations. Furthermore, we identified novel genes differentially regulated by hypoxia in both stem/progenitor cells that may be used as therapeutic targets to improve stem cell transplantation in hypoxic tissue microenvironments.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Cell Isolation and Culture

UCB units were collected from normal births of consenting donors at the John Radcliffe Hospital, Women's Centre (Oxford, U.K.) with ethical approval and informed written consent. UBC CD133+ cells were isolated by positive immunoselection using UCB CD133 cell magnetic cell separation (MACS) system (Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com/en) [24] and were always found to be approximately 96.7 ± 0.9% (mean ± SEM) CD133+. Cells were seeded at 5 × 105 cells per milliliter in serum-free StemBioA medium (StemBio Research, Villejuif, France), with cytokines: 100 ng/ml Flt3-Ligand, stem cell factor and interleukin-6, and 10 ng/ml thyroid peroxidase (R&D Systems, Abingdon, U.K., http://www.rndsystems.com) and incubated overnight at 37°C in a humidified 5% (vol/vol) CO2 incubator under normoxic (21% [vol/vol] O2) conditions before use.

BMMC were purchased from Cambrex (L3F0551/L3F0664/L1F2155/L4F1560/L5F0138) at passage two (p2) and cultured in mesenchymal stem cell growth medium (MSCGM) according to manufacturer's protocols (Cambrex Bioscience, Cambridge, U.K., http://www.cambrex.com/). Cells were maintained under normoxic conditions until use. BMMC were used at passage four (p4) in all experiments. The phenotype of the p4 BMMC was analyzed by flow cytometry. They were positive for the following markers: CD105/endoglin, CD166/Alcam, CD29, and CD44; and negative for CD45, as described by the manufacturer.

Hypoxic Incubations and Cell Proliferation Assays

BMMC were plated at a density of 14,000 cells per cm2 in MSCGM and UCB CD133+ cells were seeded at a density of 5 × 105 cells per milliliter in StemBioA medium (StemBio Research) supplemented with cytokines. Cells were incubated overnight under normoxic (21% [vol/vol] O2) conditions, and media were then replaced and cells incubated for a further 24 hours under either normoxic or hypoxic (1.5% [vol/vol] O2) conditions. Postincubation, cells were either harvested into TRIzol (Invitrogen, Paisley, U.K., http://www.invitrogen.com) for RNA or protein extraction, or harvested and counted using a standard hemocytometer. Viable cell number was assessed by trypan blue exclusion. The mean and standard deviation were calculated for three independent experiments. Statistical analysis was carried out using a t test.

Colony-Forming Assay

Three independent UCB CD133+ cell preparations were cultured (5 × 105 cells per milliliter) in serum-free StemBioA medium [25] with cytokines for 24 hours under normoxic (20% oxygen) or hypoxic (1.5% oxygen) conditions as described herein. At the end of the incubation period, cells were plated in methocel medium optimized for growth of mixed colonies (colony-forming units [CFUs]) containing granulocytic, erythroid, monocytic, and megakaryocytic lineages (-GEMM); granulocytic and monocytic lineages (-GM); graunlocytic lineages (-G); or monocytic lineages (-M); and primitive erythroid burst-forming units (BFU-E; Stem Cell Technologies, Vancouver, Canada) at concentrations of 200 and 500 cells per milliliter essentially as described previously [26]. Colonies developing at days 14 and 21 were scored according to published criteria [27]. The mean and SEM of three independent experiments was calculated. Statistical analysis was carried out using Student's t test.

BMMC Differentiation Assays

Posthypoxic or -normoxic incubation, BMMC (n = 3 independent isolates) were trypsinated and incubated in osteoblast, adipocyte, or chondrocyte differentiation media or basal media according to the manufacturer's instructions (Cambrex). Adipocytes were detected 21 days postdifferentiation by incubation with oil red O solution. Osteoblasts were detected 14 days postdifferentiation by staining using the Von Kossa stain [28]. Chondrocytes were detected by type II collagen staining at day 21 postdifferentiation [29]. Briefly, micromass sections were blocked for 1 hour and incubated with goat anti-collagen II (AbCam, Cambridge, U.K., http://www.abcam.com) primary antibody, for 1 hour at room temperature. Alexa 488-conjugated secondary rabbit anti-goat antibody (Invitrogen) was used for detection. Negative staining controls used isotype-matched antibodies in place of the primary antibody. Cells were cultured for similar times in the absence of differentiation media and then stained as above. Chondrogenic micromass pellets were weighed before being frozen and sectioned. Statistical analysis was carried out using Student's t test.

Flow Cytometry

Analysis of cell surface markers was determined by incubating cells at 4°C for 30 minutes with the relevant test antibodies or isotype controls. BMMC were analyzed by flow cytometry using fluorescein isothiocyanate (FITC)-conjugated anti-CD105-, -CD29-, -CD166-, and -CD44- or -CD45-conjugated antibodies or controls (Serotec, Oxford, U.K., http://www.serotec.com). UBC CD133+ cells were analyzed for the CD133 marker using the AC133/2 antibody conjugated to phycoerythrin (PE) or the relevant isotype control (BD Biosciences, San Jose, CA, http://www.bdbiosciences.com). An LSRSII flow cytometer was used to assess the fluorescence intensities and analysis conducted using CELLQuest software (both from BD Biosciences).

Western Blotting

Cells were cultured under either normoxic or hypoxic conditions for periods of 4, 8, 12, or 24 hours before lysis radioimmune precipitation assay buffer containing proteinase inhibitor cocktail. Protein concentrations were determined using the Bio-Rad DC protein assay (BioRad, Amersham, U.K., http://www.bio-rad.com) [30]. Total cellular proteins (5–20 μg) were separated by electrophoresis on precast 10% (wt/vol) SDS-polyacrylamide gels (Invitrogen). Proteins were transferred to polyvinylidene fluoride membranes (Invitrogen) and immunoblotted with HIF1-α (1:250; BD Biosciences), centromeric protein F (CENPF) (AbCam) or α-tubulin (1:2,000; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) primary antibodies, and then horseradish peroxidase-conjugated secondary antibodies (1:2,500, Bio-Rad). Detection was performed using the ECL kit (Amersham-Pharmacia, Amersham, U.K., http://www.amershambiosciences.com) using standard methods [30].

RNA Isolation

Total RNA was extracted from TRIzol-treated cell samples, following the manufacturer's instructions [31]. Potentially contaminating chromosomal DNA was removed from samples by treatment with DNase (Promega, Southampton, U.K., http://www.promega.com). To ensure RNA quality, further purification was carried out using RNAeasy mini kits (Qiagen, Crawley, U.K., http://www1.qiagen.com/) according to the manufacturer's instructions. RNA was eluted in diethylpyrocarbonate-treated water, and an aliquot of each sample was used to test the quantity of RNA by spectrophotometry and the quality by gel electrophoresis.

Microarray Experiments

All the experiments were designed in compliance with Minimum Information About a Microarray Experiment (MIAME) guidelines to ensure that the microarray data can be properly interpreted and independently verified [32]. RNA and cRNA (labeled probes described herein) quality was checked using RNA 6000 Nano Assay on Agilent Bioanalyzer 2100 (Agilent Technologies UK, Stockport, U.K., http://www.agilent.com). Throughout this study, we used commercially available high-density oligonucleotide Affymetrix HG-U133A arrays (Affymetrix, High Wycombe, U.K., http://www.affymetrix.com). All of the procedures and hybridizations were performed in triplicate for two biological samples and according to the GeneChip Expression Technical Manual (http://www.affymetrix.com/support/technical/manual/expression_manual.affx). For BMMC, these were independent isolates from single donors. For CD133+ cells, UCB units were pooled to provide sufficient cell numbers. In brief, 1 μg of RNA was converted to double-stranded cDNA with SuperScript II (Invitrogen) using T7-(dT)24 primer (Affymetrix). This was used in an in vitro transcription (IVT) assay to generate a biotin-labeled cRNA probe, using Enzo Bioarray High-yield RNA Transcript labeling kit (Affymetrix). The resulting cRNA was purified using the GeneChip sample clean-up module (Affymetrix) Fragmented cRNA (15 μg) was used in 300 μl of hybridization cocktail containing spiked controls (Affymetrix), 0.1 mg/ml Herring Sperm DNA (Promega), 0.5 mg/ml acetylated bovine serum albumin (Invitrogen). Two-hundred milliliters of this hybridization cocktail was used on each chip and incubated at 45°C for 16 hours in the hybridization oven, rotating at 60 rpm. After hybridization, the arrays were processed using a GeneChip Fluidics Station 400 according to recommended protocols (Affymetrix) of double-staining and posthybridization washes. Fluorescent images were captured using gene Array Scanner 2500 (Affymetrix).

Data Analysis and Quality Check

Gene transcript levels were determined from data image files using algorithms in Gene Chip Operating Software (GCOS; Affymetrix). Global scaling was performed to compare genes from chip to chip. Report files were generated according to the software and data quality check performed. Data analyses were performed using Data Mining Tool software (Affymetrix), GeneSpring 7.2 (Agilent Technologies UK) and Microsoft Excel (Microsoft Corp., Redmond, WA). Statistical analysis was carried out using analysis of variance to identify probes for which a significant difference (p < .05) in mean hybridization intensity was found between the two conditions. Data were filtered initially on the basis of the probes awarded a “P” (present status) in both batches of cells according to the GCOS algorithm. Fold changes were calculated by comparison of mean intensity hybridization values in normoxia compared to the matched sample exposed to hypoxia. A threshold change of 1.75-fold was selected on the basis that this cut-off captured many of the genes known to be regulated by hypoxia. Genes in these lists were classified according to Gene Ontology (GO; http://www.geneontology.org) with roles in biological process and/or molecular function defined.

Quantitative Reverse Transcriptase-Polymerase Chain Reaction

Total RNA (0.5 μg) obtained individually from three to six BMMC or CD133+ cord blood units was reverse transcribed to cDNA using the Invitrogen thermoscript reverse transcriptase polymerase chain reaction (RT-PCR) kit with random hexamers, according to manufacturer's instructions. The resulting cDNA was stored at −80°C until use. Quantification of the relative mRNA expression was performed using TaqMan gene expression assays (“Assay-on-demand,” Applied Biosystems [ABI], Foster City, CA, http://www.appliedbiosystems.com). The reactions were carried out in a total volume of 25 μl using 50 ng of cDNA, ABI TaqMan master mix, primers and probe. Probes were labeled at the 5′-end with the 6-carboxyfluorescein (FAM) fluorophore and at the 3′-end with the TAMRA quencher. Thermocycling conditions used were 2 minutes at 50°C and 10 minutes at 95°C followed by 40 cycles of 95°C (15 seconds) and 60°C (1 minute). Detection was performed using the ABI Prism 7000 Sequence Detector System. Analysis of data was carried out using the software provided, cycle threshold (Ct) values were related to the relative standard curve, and the relative mRNA quantity was compared to the control group. The relative mRNA level and then fold change in hypoxia were calculated using the 2−ΔΔCt method using β-2-microglobulin as the calibrator. The mean and SEM were calculated in three to five independent experiments. Statistical analysis was carried out using Student's t test.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Hypoxia Affects the Proliferation of Human CD133+ Clonogenic Cells and Cultured Bone Marrow Mesenchymal Stem Cells

HIF-1α is known to be induced in many cell types, including hemopoietic and mesenchymal cells, when they are exposed to hypoxia (for review see [33]). To confirm this response in purified UCB CD133 + cells and BMMC, we cultured these two cell populations under two different oxygen tensions (21% oxygen [normoxia] and 1.5% oxygen [hypoxia]) for 24 hours and examined the induction of HIF-1α by immunoblotting. A representative Western blot is shown in Figure 1A. Low levels of endogenous HIF-1α were observed in the BMMC, but not in UCB CD133+ cells at 24 hours of culture in normoxia. With 24 hours of exposure to hypoxia, our results clearly demonstrate that the HIF-1α protein was stabilized/upregulated in both cell populations.

thumbnail image

Figure Figure 1.. Comparison of the cellular response of umbilical cord blood (UCB) CD133+ cells and BMMC to hypoxic preconditioning. UCB CD133+ cells and BMMC were exposed to either normoxia or hypoxia for 24 hours and lysates immunoblotted for HIF1α (A). UCB CD133+ cells and BMMC were exposed to normoxia or hypoxia for 24 hours before counting viable cell number (B). UCB CD133+ cells were incubated for 24 hours in StemBioA (StemBio Research) medium with cytokines under either normoxia or hypoxia before plating a total of 500 cells from each treatment group into methylcellulose media. After 14 or 21 days incubation cultures were scored according to their morphological characteristics (C). BMMC were exposed to varying lengths of time in normoxia or hypoxia before lysates were immunoblotted for CENPF (D). BMMC were plated at 14,000 cells per cm2 in mesenchymal stem cell growth medium and incubated for 24 hours under either normoxic or hypoxic conditions before induction to differentiate along the adipogenic or osteogenic lineage. Adipogenesis was detected 21 days post differentiation by incubation with oil red O solution. Osteogenesis was detected 14 days post differentiation using Von Kossa staining (E). Chondrogenesis was detected 21 days post differentiation by type II collagen staining (F). Control cultures were maintained in basal media for the same length of time and stained with oil red O, Von Kossa or type II collagen staining (E, F). The chondrogenic micromass pellets were weighed prior to being frozen and sectioned to carry out the type II collagen staining. The fold-change in pellet weight of three independent experiments is shown (F). The expression of adipogenic (PPAR-α), osteogenic (RUNX2), and chondrogenic (SOX9) factors was assessed by quantitative real-time polymerase chain reaction after 24 hours of exposure to hypoxia. Fold-change in mRNA levels of hypoxic compared to normoxic samples was determined in three independent experiments (G). VEGFA was used as control. Statistical analysis was carried out using a t test. ∗, p < .005. Abbreviations: BFU-E, erythroid burst-forming unit; BMMC, bone marrow mesenchymal cells; CENPF, centromeric protein F; CFU, colony-forming unit; G, graunlocytic lineages; GEMM, granulocytic, erythroid, monocytic, and megakaryocytic lineages; GM, granulocytic and monocytic lineages; H, hypoxia; HIF-1α, hypoxia inducible factor-1α protein; M, monocytic lineages; N, normoxia; PPAR, peroxisome proliferator-activated receptor.

Download figure to PowerPoint

To establish whether short exposure to low oxygen tensions affected the fate of these cells, we assessed both their proliferative and differentiation potential. For the UCB CD133+ cells, there was no apparent loss of cell viability, and the total cell numbers remained unchanged over the 24-hour period and under the conditions used (Fig. 1B). We assessed the day-14 and -21 clonogenic myeloid progenitor content of these cells after this 24-hour exposure to normoxia or hypoxia. After the brief exposure to normoxia, 18.6% ± 6.0% and 17.6% ± 5.6% of the UCB CD133+ cells formed myeloid colonies in vitro at day 14 and day 21, respectively. The distribution of progenitors forming mixed colonies containing CFU-GEMM, or containing CFU-GM, -G, or -M, and primitive BFU-E is shown in Figure 1C. UCB CD133+ cells were exposed to hypoxia for 24 hours before myeloid clonogenic analyses demonstrated a moderate increase in total colony numbers, from 18.6% ± 6.0% to 29.6% ± 5.2% (p = .003; n = 3) at day 14 and from 17.6% ± 5.6% to 24.4% ± 6.9% (p = .04; n = 3) at day 21 (Fig. 1C). BMMC subjected to hypoxia for 24 hours showed a 1.61- ± 0.9-fold (p < .05; n = 3) increase in viable cell number when compared to normoxia (Fig. 1B). To further examine the reason for the increased cell number in hypoxia in BMMC, we measured the level of CENPF protein, a marker of cell proliferation [34], in cultures exposed to normoxia or hypoxia over a 24-hour time period. Strikingly, CENPF expression was upregulated in cells exposed to hypoxia in the initial 4–8-hour period, suggesting that increased BMMC number in hypoxia resulted from more rapid cell cycle progression (Fig. 1D). To determine whether hypoxia affected the differentiation potential of BMMCs, we subjected the BMMC to hypoxia for 24 hours and then examined their ability to differentiate into adipocytes, osteoblasts, or chondrocytes. BMMC retained adipogenic and osteogenic potential (n = 3; Fig. 1E), but showed increased levels of chondrogenesis after this short exposure to hypoxia. Notably, a significant increase in chondrogenic pellet size was found (p < .05, n = 3; Fig. 1F). This was confirmed by examining the change in level of expression of the SOX9 chondrogenic master gene in BMMC exposed to normoxia or hypoxia for 24 hours. We observed a 1.54- ± 0.13-fold increase (mean ± SEM; p < .05; n = 3) in SOX9 expression in hypoxia (Fig. 1G). In parallel, the expression of adipogenic factors such as peroxisome proliferator-activated receptor (PPAR)-α [35] and osteogenic factors such as RUNX2 and Osterix [36] was also examined (Fig. 1G). The expression of PPAR-α and RUNX2 was unchanged by short exposures to hypoxia, whereas the mRNA levels of Osterix were not detectable above background. These results are consistent with the differentiation pattern observed herein (Fig. 1E), and suggest an increased propensity of BMMC exposed briefly to hypoxic conditions for chondrogenic differentiation.

Taken together, these data indicate that short hypoxic incubations can increase the myeloid clonogenic ability of UCB CD133+ cells. In contrast, hypoxia promotes both the rate of cell division and proliferation of the BMMC, while enhancing cells' ability to enter the chondrogenic lineage.

UCB CD133+ Cells and BMMC Share a Large Pool of Genes

To identify the gene fingerprint or signature of the UCB CD133 + cells and BMMC used here, we analyzed the transcriptional profiles of genes expressed in these cell populations using microarray analyses 24 hours after the cells had been exposed to normoxia or hypoxia. In these experiments, the UCB CD133+ cells were cultured in serum-free medium with cytokines, whereas the BMMC were incubated as semiconfluent cultures in MSCGM medium containing 10% serum during the 24-hour normoxic or hypoxic exposures. Total RNA isolated from these cells was labeled and hybridized to high-density Affymetrix oligo arrays containing approximately 22,283 gene probes (or 14,500 well-characterized human genes) as described in Materials and Methods. For each gene, the analysis software determined a mean fluorescent intensity for each probe set and genes were awarded a P status according to the Affymetrix GCOS algorithm. Data were filtered according to the genes being awarded a P in two different samples of UCB CD133+ cells or BMMC. Annotations were performed using the UniGene database, and nonredundant sets of genes were compiled for the two conditions, normoxia (Fig. 2A) and hypoxia (Fig. 2B). Taken together, these data indicate that a large percentage (approximately 80%) of the genes expressed in each cell population are shared between them, and that short exposures to hypoxia do not alter this proportion significantly.

thumbnail image

Figure Figure 2.. Overlapping gene expression in umbilical cord blood (UCB) CD133+ cells and bone marrow mesenchymal cells (BMMC). Venn diagrams showing the number of genes that are shared between or are unique to UCB CD133+ cells and BMMC in normoxia (A) and hypoxia (B). Abbreviation: BMSC, bone marrow-derived stem cells.

Download figure to PowerPoint

Gene Fingerprints of the UCB CD133+ Cells and BMMC Reveal Complementary Receptor-Ligand Pairs

To identify genes enriched in each cell population, we compared the transcriptional profiles, or fingerprints, of UCB CD133+ cells and BMMC obtained from the microarray experiments (see Fig. 2A). The average mean fluorescent intensity (MFI) of genes present in both batches of UCB CD133+ cells was compared to the average MFI of genes expressed in both batches of BMMC. Using a 10-fold change in gene expression as the cut-off, 175 genes were identified as enriched in UCB CD133+ cells compared to BMMC. Seventy-nine of these 175 genes were common to both cell types, whereas 96 genes were expressed in UCB CD133+ cells and not in BMMC. Using the same cut-off, we identified a total of 448 genes enriched at least 10-fold in BMMC compared to UCB CD133+ cells. Interestingly, only 8 of these genes were expressed in both cell types, and 440 genes were present in BMMC but not in UCB CD133+ cells.

Genes enriched in UCB CD133+ cells included known HSC/HPC markers such as CD133 (110-fold), KIT (50-fold), Tie1 (40-fold), and GATA2 (13-fold). The high expression level of transcription factors that support self-renewal, such as GATA2, confirmed the results of others [37, 38]. Enriched genes encoding chemokines and chemokine receptors in UCB CD133+ cells accounted for 8 of the 175 enriched genes. These included CXCL2 (15.4-fold), CXCL5 (45.8-fold), CCL1 (115.5-fold), CCL4 (158-fold), CCL3/CCL3L1/CCL3L3 (233-fold), CCL20 (56-fold), CCRL2 (16.3-fold), and CXCR4 (32.5-fold). In addition, it was noted that the integrin alpha four gene, ITGA4, was enriched 6.5-fold in UCB CD133+ cells. In BMMC, the corresponding ligands for CXCR4 and ITGA4 (VLA-4) were found to be enriched. These included CXCL12 (31.6-fold), the major ligand for CXCR4, and FN1 (approximately 139-fold) a known ligand for VLA-4. The high-level of ITGA4 and CXCR4 expression in UCB CD133+ cells and the corresponding ligands in BMMC highlights the known crosstalk between these cell types and their expressed receptor-ligand pairs in the bone marrow niche [39, [40], [41], [42]43]. The VLA-4/FN adhesion pathway plays a key role in the CXCL12-induced migration and subsequent adhesion of CD133+ HSC/HPC in vitro and in promoting the lodgement in and interaction of UCB hematopoietic progenitor cells with the bone marrow niche in vivo [39, [40], [41], [42]43]. In addition, the egression of these cells from the bone marrow also involves signaling via CXCL12/CXCR4 and VLA-4/FN axis (for reviews, see [40, 41, 43]). These data support the concept that progenitor cells overexpress molecules that play crucial roles in directing them to and holding them in their specific niche.

In contrast to UCB CD133+ cells, BMMC expressed a high proportion of extracellular matrix components, as well as cell adhesion molecules. One example is the gene coding for FN1, mentioned in the preceding paragraph. From the BMMC fingerprint in normoxia, we observed that genes encoding collagen (12 genes) and their modifying enzymes (7 genes) were highly enriched in this cell population, illustrating the importance of collagen synthesis in these cells. These included COL1A1 (60-fold), COL1A2 (2232-fold), COL3A1 (221-fold), COL4A1 (114-fold), COL4A2 (150-fold), COL5A1 (174-fold), COL5A2 (326-fold), COL6A3 (593-fold), COL8A1 (59-fold), COL11A1 (99-fold), COL14A1 (186-fold), and COL16A1 (24-fold). Collagen-modifying enzymes included procollagen hydroxylases such as P4H42 (21-fold), LOX (411-fold), LOXL1 (70.4-fold), and LOXL2 (105-fold), procollagen oxidase PLOD2 (10.7-fold), and procollagen endopeptidases PCOLCE (119-fold) and PCOLCE2 (28.2-fold). Among other genes enriched in BMMCs compared to UCB CD133+ cells were a number of growth and survival factors (seven genes) that were also highly expressed. These included ADM (11.7-fold), FGF5 (16.7-fold), FGF7 (43.8-fold), NGFβ (15.5-fold), STC2 (34.7-fold), VEGFA (from 4- to 70-fold), and VEGFC (16-fold).

Taken together these data show that UCB CD133+ cells and BMMC express many molecules reflecting their individual functions and also two complementary receptor-ligand pairs, VLA-4/FN1 and CXCR4/CXCL12, that are known to play an essential role in cell migration and in cell-cell interactions with microenvironmental niches [39].

The Molecular Response of UCB CD133+ Cells and BMMC to Hypoxia Is Cell Type-Specific

To gain a better understanding of the molecular mechanisms that control cellular responses of progenitor cells to hypoxia, we compared changes in global gene expression in UCB CD133 + cells and BMMC cultured in normoxic or hypoxic conditions for 24 hours. Differential gene expression regulated by hypoxia was determined by comparing pairwise normoxic and hypoxic samples for each cell population as described in Materials and Methods. The cut-off limits used by others for a significant fold change vary from threefold [38] to 1.85-fold [40] to 1.5-fold [41]. When we used a threshold of 1.5-fold, a total of 214 and 92 genes appeared to be differentially regulated by hypoxia at 24 hours in UCB CD1133+ cells and BMMCs, respectively (data not shown). We decided to use a threshold of 1.75-fold change because this cut-off included many of the genes known to be regulated by hypoxia (e.g., VEGFA, PGK1, and ENO2.). On the basis of this criterion, we identified a total of 183 genes differentially regulated by hypoxia at 24 hours in UCB CD133+ cells. Approximately 161 genes of the total 183 were upregulated (> 1.75-fold), whereas 22 genes were downregulated. Annotations were performed using GO, and the candidate genes were classified into categories according to biological processes. Figure 3A shows that 26.8% of the genes differentially regulated by hypoxia in UCB CD133+ progenitor cells have not been assigned to any biological process (unknown), and almost half of the genes (48%) were involved in one of the following categories: metabolism (13.1%), cell proliferation/survival (10.4%), transcription (9.3%), signal transduction (7.7%), and transport (6.6%). In BMMC, 45 genes were differentially regulated by hypoxia at 24 hours. Thirty-three [33] of these genes were upregulated by more than 1.75-fold, and 12 genes were downregulated. After we categorized genes using GO (Fig. 3B), approximately 26.7% of the genes differentially regulated by hypoxia in BMMC had unknown associated biological processes, and more than half (51.1%) could be assigned to one of the following categories: signal transduction (15.5%), cell proliferation/survival (8.9%), metabolism (8.9%), transcription (8.9%), and nucleic acid metabolism (8.9%). When hypoxia-responsive genes found in UCB CD133+ cells and BMMC were compared, only nine such genes were shared between the two progenitor cell populations. These included ADM, AK3L1/L2, BNIP3, ENO2, PGK1, SLC16A3, SOS2, TPI, and VEGFA (Table 1). Seven of the nine genes are known to be regulated by hypoxia in other cell types, such as endothelial or cancer cells [42, [43], [44], [45], [46], [47], [48], [49], [50]51]. The exceptions are AK3L1/L2, which is involved in metabolism, and SOS2, which is associated with signal transduction in Drosophila [52]. Three of the eight genes shared between UCB CD133+ cells and BMMC are involved in metabolism (ENO2, TPI, and PGK1), two are associated with cell proliferation and survival (VEGFA and BNIP3), and the rest are involved in transport (SLC16A3) and signal transduction (ADM). Taken together, these results show that only a small proportion of genes regulated by hypoxia are common to UCB CD133+ cells and BMMC, suggesting that their molecular response to low oxygen levels is largely cell type-specific.

thumbnail image

Figure Figure 3.. Major categories of genes regulated by hypoxia in umbilical cord blood (UCB) CD133+ cells and bone marrow mesenchymal cells (BMMC) profiles. Distribution according to biological processes of genes differentially regulated by hypoxia more than 1.75-fold in UCB CD133+ cells (A) or BMMC (B).

Download figure to PowerPoint

Table Table 1.. Genes differentially regulated by hypoxia that are shared between BMMC and umbilical cord blood CD133+ cells
Thumbnail image of

Interestingly, genes involved in cell proliferation and survival represented one of the largest categories of genes regulated by hypoxia in both UCB CD133+ cells and BMMC. In BMMC, these included BNIP3, DUSP1, FGF2, and VEGFA. BNIP3 is a proapoptotic factor known to be regulated by hypoxia in cancer cells [43]. DUSP1 encodes a dual specific phosphatase that protects against overactivation of HIF-1α [53], and possibly against HIF-1α-mediated cell cycle arrest [54, 55]. FGF2 and VEGFA are growth factors with potent proangiogenic and mitogenic characteristics that may exert paracrine and autocrine effects. Recently, it has been suggested that VEGFA has an autocrine mitogenic role in human bone marrow-derived mesenchymal stem cells [56]. In contrast to our results with BMMC, the overall proliferation of UCB CD133+ cells remained unaltered after 24-hour exposure to hypoxia, although we noted an increase in clonogenic potential. Genes differentially regulated by hypoxia associated with cell proliferation/survival in UCB CD133+ cells included, among others CCNG2, a negative regulator of cell cycle progression [57] and NEDD9, which is expressed in G1 phase arrest [58], suggesting that there might be a transient block in cell cycle progression during the exposure to hypoxia.

The osteogenic and adipogenic differentiation potential of BMMC exposed to hypoxia for 24 hours was unaltered, whereas the chondrogenic potential was increased. Consistent with these results, the mRNA levels of the adipogenic factor PPAR-α and osteogenic factor RUNX2 were unchanged in BMMC exposed to hypoxia. By contrast, the expression of SOX9, a key transcription factor responsible for chondrocyte differentiation, was increased by 1.54-fold (Fig. 1G). A similar increase in SOX9 expression was observed in the microarray analysis (1.66-fold). Recently, it has been reported that hypoxia induces Sox9 gene expression in murine mesenchymal cells [59].

Novel Genes Regulated by Hypoxia in UCB CD133+ Cells and BMMC

To validate the results obtained in the microarray analysis, we carried out real-time quantitative RT-PCR (qRT-PCR) on a number of selected genes to confirm changes in gene expression. We chose to test four of the genes that were shared between UCB CD133+ cells and BMMC: BNIP3 and VEGFA, both of which are known to be regulated by hypoxia, and AK3L1/L2 and SOS2. In addition, we also tested a panel of hypoxia-regulated genes that are unique to either UCB CD133+ cells or BMMC as representatives of the major categories described in Tables 1 and 2. These genes included BNIP3L, CDTSPL, CCL20, LSP1, NEDD9, PDK1, S100A10, SLC6A8, and TMEM45A for UCB CD133+ cells and EDG-1, EPHA3, and STC1 for BMMC. To our knowledge, CDTSPL, CCL20, LSP1, NEDD9, TMEM45A, EDG-1, and EPHA3 have not been described previously to be regulated by hypoxia, whereas BNIP3L, PDK1, S100A10, SLC6A8, and STC1 are known to be regulated by hypoxia [41, 60, 61]. The qRT-PCR analysis confirmed that the expression of BNIP3, VEGFA, and AK3L1/L2 was significantly (p ≤ .05) upregulated by hypoxia by more than 1.75-fold in both cell populations. However, the change in expression of SOS2 in either UCB CD133+ cells or BMMC exposed to hypoxia was not statistically significant using the primer set described. We additionally confirmed that BNIP3L, CCL20, LSP1, NEDD9, PDK1, S100A10, SLC6A8, and TMEM45A were significantly (p ≤ .05) upregulated and that CTDSPL was significantly (p = .05) downregulated by hypoxia in UCB CD133+ cells. In BMMC, STC1 and EPHA3 were found to be significantly (p = .03) upregulated by hypoxia, whereas a significant decrease in gene expression was observed for EDG-1 (p = .04). In summary, these results confirmed the differential expression of seven genes known to be regulated by hypoxia in other cell types and eight additional genes that had not been reported previously to be regulated by hypoxia.

Table Table 2.. Real-time quantitative polymerase chain reaction analysis of gene expression regulated by hypoxia in umbilical cord blood CD133+ cells and bone marrow-derived mesenchymal cells
Thumbnail image of

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

The ability of cells to sense and respond to low oxygen levels plays an important role in both pathological and physiological conditions. In this study, we characterized and compared, for the first time, the response of UCB CD133+ cells and BMMC to short incubations in hypoxia. Our results show that exposure to 1.5% oxygen for 24 hours stabilizes HIF-1α protein in both cell populations. We observed no overall effect of 24-hour exposure to hypoxia on the proliferation of the total UCB CD133+ cell population. However, we did find a significant increase in clonogenic myeloid cell numbers after 24 hours of hypoxic preconditioning, a finding consistent with previous studies in which hypoxia was shown to enhance both CFU and severe combined immunodeficiency repopulating ability cell numbers [20, 21]. It is of note that different oxygen concentrations appear to affect the fate of the more immature HSC/HPC in different ways. Whereas lower oxygen levels (approximately 1%) maintain the early progenitors in culture [62], higher concentrations are reported to induce their differentiation into hematopoietic lineages [63, 64].

We also found that short-term exposure to hypoxia clearly promoted proliferation and cell cycle progression of BMMC. To our knowledge, this has not been described before. This effect contrasts with results obtained by us and others on endothelial cells where hypoxia induces their cell-cycle arrest [54, 65]. Interestingly, Grayson et al. [66] have recently reported that in three-dimensional (3D) cultures, bone marrow-derived mesenchymal stem cells show increased proliferation when exposed to long-term chronic hypoxia (2% oxygen). When induced to differentiate into osteogenic, adipogenic and chondrogenic lineages after 24 hours of preconditioning in hypoxia, the BMMC used in our studies maintained their potential to form osteoblasts and adipocytes, while enhancing the expression of genes required for a chondrogenic differentiation switch. This correlated with an increase in SOX9 gene expression after exposure to hypoxia for 24 hours. In murine mesenchymal cells, the key chondrogenic transcription factor Sox9 has been recently found to be a direct target for HIF-1α and, therefore, regulated by hypoxia [59]. Others have analyzed the effects of hypoxia on mesenchymal stem cell differentiation in different ways [2, 23, 66]. In those studies, the mesenchymal stem cells were induced to differentiate in cultures subjected to chronic hypoxia (over 2–3 weeks). Grayson's [66] demonstration of increased adipogenic and osteogenic differentiation in long-term 3D cultures of bone marrow derived mesenchymal stem cells under these conditions contrasts with the results of Malladi et al. [23], who showed that long-term hypoxia reduces both the osteogenic and chondrogenic potential of adipose tissue-derived mesenchymal cells. In addition, other studies have reported that adipogenic differentiation of bone marrow-derived mesenchymal stem cells is inhibited by hypoxia through activation of the transforming growth factor-β/Smad signaling pathway [2]. Although the differences from our experiments can be explained by the fact that acute hypoxia preconditioning before BMMC differentiation was induced rather than chronic exposure to hypoxia of differentiating BMMC, it is more difficult to explain the differences observed among the experiments conducted by the groups of Grayson [66], Malladi [23], and Zhou [2]. In each case, 1%–2% oxygen was used, and all hypoxic incubations were long-term incubations (chronic hypoxia) while the cells were induced to differentiate. However, whereas Zhou et al. [2] used monolayer cultures, Grayson et al. [66] and Maladi et al. [23] cultured their progenitor cells in 3D scaffolds. It is, therefore, possible that different tissue architectures may account for the different responses in differentiation potential of BMMCs chronically exposed to hypoxia.

The transcriptional profiles of UCB CD133+ cells and BMMC revealed that our results are in agreement with the work of others. For instance, 77% (4,188) of the genes present in BMMC in our studies are also present in the gene profile of bone marrow-derived mesenchymal stem cells carried out previously by Serial Analysis of Gene Expression [67]. In addition, 93% (38 of 41 genes) of the HSC markers found to be enriched in UCB CD133+ cells identified previously by Jaatinen et al. [38] are also present in our study. Hematopoietic progenitor cell markers including CD133, KIT, Tie 1, and GATA2 were found in UCB CD133+ cells in this study, as well as genes encoding chemokines, chemokine receptors and genes involved in immune response. In contrast, genes that encode cell adhesion molecules, extracellular matrix components, growth factors and survival factors were enriched in BMMC. Interestingly, UCB CD133+ cells expressed two hypoxia-regulated genes, CXCR4 and ITGA4, which are important molecules involved in stem cell homing, whereas BMMC expressed the corresponding CXCR4 and ITGA4 ligands, CXCL12 and FN1. This stresses the significance of comparing these cell populations and sheds light on receptor/ligand interactions that are important in regulating stem cell homing and stem cell fate.

We found 183 genes differentially expressed in UCB CD133+ cells in hypoxia and 45 genes in BMMC when a 1.75-fold change in expression threshold was chosen on the basis of the upregulation of other known hypoxia-inducible genes (e.g., VEGFA, PGK1, ENO2). Among these hypoxia-induced genes, only nine of them were shared between the two cell populations, confirming that the response to low oxygen levels was cell type-specific. A significant proportion of the genes identified in our microarray study have been described before to be regulated by hypoxia, some of them as examples of potent mitogenic and proangiogenic factors with established roles in vascular biology (e.g., VEGFA, VEGFC [41]). Notably, we have identified genes not known to be regulated by hypoxia previously. Among those, we have confirmed a selected number (seven in total) by qRT-PCR, namely CDTSPL, CCL20, LSP1, NEDD9, and TMEM45A in UCB CD133+ cells and EDG1 and EPHA3 in BMMC. Interestingly, genes involved in cell proliferation or survival represented one of the largest categories of genes regulated by hypoxia in both UCB CD133+ cells and BMMC, emphasizing the importance of these genes during the initial responses to hypoxia [2, 23]. Hypoxia-responsive genes associated with cell proliferation or survival shared between BMMC included BNIP3 and VEGFA. BNIP3 is a proapoptotic factor known to be regulated by hypoxia in cancer cells [43]. VEGFA is a growth factor with very potent mitogenic and proangiogenic characteristics. It has been reported that VEGFA and its receptors play a major role in the maintenance and self-renewal of human hematopoietic stem and progenitor cells [72]. More recently, Mayer et al. [56] have suggested an autocrine or paracrine mitogenic role of VEGFA in human BM-derived mesenchymal stem cells. Although more detailed studies would be required to understand the mechanisms by which hypoxia enhances proliferation in BMMC and clonogenic progenitors in UCB CD133+ cells, it is plausible that an autocrine role of factors such as VEGFA, CXCR4, and CXCL12 may contribute to the maintenance of HSC/HPC and BMMC exposed to hypoxia.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

We have found that hypoxia can alter the proliferation and clonogenic capacity of BMMC and UCB CD133+ cell subsets without affecting their differentiation potential. Gene profiling experiments have identified several genes not previously described to be regulated by hypoxia that will provide a platform to elucidate the molecular pathways that control stem/progenitor cell responses to hypoxia.

Disclosure of Potential Conflicts of Interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

The authors indicate no potential conflicts of interest.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

We thank Prof. Marcela Contreras for her continuous support and Emma Frith for her help with the CFU assays. This work is supported by grants from National Health Service (NHS) Blood and Transplant (S.M.W., A.L.H.), the Leukaemia Research Fund UK (S.M.W., E.M.R., M.R.), the Medical Research Council (MRC) UK (D.B., K.D., S.M.W.), the Wellcome Trust (S.M.W.), the E.U. Framework 6 Strategic Targeted Research Project Programme (S.M.W.), the British Heart Foundation (S.M.W., E.M.R.), and the National Translational Cancer Research Network (A.L.H., S.M.W.). This research was carried out at the National Blood Service Oxford and the MRC Unit, Department of Human Anatomy and Genetics, University of Oxford, and benefits from research and development (R&D) funding received from the NHS R&D Directorate (S.M.W.). E.M.R. and S.J.M.H. contributed equally to this paper and are joint first authors. M.R. is currently affiliated with Genetic and Gene Therapy Laboratory, Centre for Basic Research, Biomedical Research Foundation of the Academy of Athens, Athens, Greece.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References