Sebastien Chateauvieux, Jean-Laurent Ichant, Bruno Delorme, Vincent Frouin, Genevive Pitu, Alain Langonn, Nathalie Gallay, Luc

Senseb, Michle T. Martin, Kateri A. Moore and Pierre Charbord

Physiol Genomics 29:128-138, 2007. First published Dec 19, 2006; doi:10.1152/physiolgenomics.00197.2006 You might find this additional information useful... Supplemental material for this article can be found at: http://physiolgenomics.physiology.org/cgi/content/full/00197.2006/DC1 This article cites 48 articles, 27 of which you can access free at: http://physiolgenomics.physiology.org/cgi/content/full/29/2/128#BIBL This article has been cited by 3 other HighWire hosted articles: Hepatocyte differentiation of mesenchymal stem cells from human adipose tissue in vitro promotes hepatic integration in vivo H Aurich, M Sgodda, P Kaltwasser, M Vetter, A Weise, T Liehr, M Brulport, J G Hengstler, M M Dollinger, W E Fleig and B Christ Gut, April 1, 2009; 58 (4): 570-581. [Abstract] [Full Text] [PDF] Embryonic stromal clones reveal developmental regulators of definitive hematopoietic stem cells C. Durand, C. Robin, K. Bollerot, M. H. Baron, K. Ottersbach and E. Dzierzak PNAS, December 26, 2007; 104 (52): 20838-20843. [Abstract] [Full Text] [PDF] Bone homing of mesenchymal stem cells by ectopic {alpha}4 integrin expression S. Kumar and S. Ponnazhagan FASEB J, December 1, 2007; 21 (14): 3917-3927. [Abstract] [Full Text] [PDF] Updated information and services including high-resolution figures, can be found at: http://physiolgenomics.physiology.org/cgi/content/full/29/2/128 Additional material and information about Physiological Genomics can be found at: http://www.the-aps.org/publications/pg

Downloaded from physiolgenomics.physiology.org on June 6, 2010

This information is current as of June 6, 2010 .

Physiological Genomics publishes results of a wide variety of studies from human and from informative model systems with techniques linking genes and pathways to physiology, from prokaryotes to eukaryotes. It is published quarterly in January, April, July, and October by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright 2005 by the American Physiological Society. ISSN: 1094-8341, ESSN: 1531-2267. Visit our website at http://www.the-aps.org/.

Physiol Genomics 29: 128 138, 2007. First published December 19, 2006; doi:10.1152/physiolgenomics.00197.2006.

Molecular prole of mouse stromal mesenchymal stem cells

Sebastien Chateauvieux,1 Jean-Laurent Ichante,2 Bruno Delorme,1 Vincent Frouin,2 Genevieve Pietu,2 Alain Langonne,1,3 Nathalie Gallay,1 ` Luc Sensebe,1,3 Michele T. Martin,2 Kateri A. Moore,4 and Pierre Charbord1 `
Institut National de la Sante et de la Recherche Medicale, Equipe-ESPRI/EA-3855, Universite Francois Rabelais, Faculte de Medecine, Tours; 2Commissariat a lEnergie Atomique, Service de Genomique Fonctionnelle, ` Evry; and 3Etablissement Francais du Sang-Centre-Atlantique, Tours, France; and 4Department of Molecular Biology, Lewis Thomas Laboratory, Princeton University, Princeton, New Jersey
Submitted 6 September 2006; accepted in nal form 17 December 2006
1

Chateauvieux S, Ichante J-L, Delorme B, Frouin V, Pietu G, Langonne A, Gallay N, Sensebe L, Martin MT, Moore KA, Charbord P. Molecular prole of mouse stromal mesenchymal stem cells. Physiol Genomics 29: 128 138, 2007. First published December 19, 2006; doi:10.1152/physiolgenomics.00197.2006.We determined a transcriptional prole specic for clonal stromal mesenchymal stem cells from adult and fetal hematopoietic sites. To identify mesenchymal stem cell-like stromal cell lines, we evaluated the adipocytic, osteoblastic, chondrocytic, and vascular smooth muscle differentiation potential and also the hematopoietic supportive (stromal) capacity of six mouse stromal cell lines from adult bone marrow and day 14.5 fetal liver. We found that two lines were quadripotent and also supported hematopoiesis, BMC9 from bone marrow and AFT024 from fetal liver. We then ascertained the set of genes differentially expressed in the intersection set of AFT024 and BMC9 compared with those expressed in the union set of two negative control lines, 2018 and BFC012 (both from fetal liver); 346 genes were upregulated and 299 downregulated. Using Ingenuity software, we found two major gene networks with highly signicant scores. One network contained downregulated genes that are known to be implicated in osteoblastic differentiation, proliferation, or transformation. The other network contained upregulated genes that belonged to two categories, cytoskeletal genes and genes implicated in the transcriptional machinery. The data extend the concept of stromal mesenchymal stem cells to clonal cell populations derived not only from bone marrow but also from fetal liver. The gene networks described should discriminate this cell type from other types of stem cells and help dene the stem cell state. differentiation; hematopoiesis; cytoskeleton; transcription; gene network
DURING DEVELOPMENT, HEMATOPOIESIS takes place in sequential sites (reviewed in Ref. 16). It develops during embryonic life within the extraembryonic annex, the yolk sac, and within the intraembryonic region, the aorta-gonad-mesonephros (AGM) region. During fetal life, it proceeds within the liver. From birth onward, bone marrow is the major, if not exclusive, site of hematopoiesis in mammals. Hematopoietic stem cells (HSCs) migrate, via the bloodstream, from one site to the next. As early as the AGM stage, HSCs have acquired their essential properties of self-renewal and multipotentiality, giving rise to all blood lineages. It is in the fetal liver that HSCs show

maximal proliferation capacity: only at this stage is amplication of the HSC pool observed. Cells from the microenvironment provide different niches for hematopoietic cells. In the HSC niche (reviewed in Refs. 1, 24, 30, 48), the stemness, i.e., the balance between self-renewal and commitment, is insured by physical contact between the HSC and a stromal cell (i.e., hematopoietic supportive) and its associated extracellular matrix. In the bone marrow, many recent experiments strongly suggest that the stromal component would be an osteoblastic-type cell or a highly specialized microvascular cell (1). In the fetal liver, fewer studies have tried to dene the stromal component of the niche. We have suggested (7) that a cell in epithelial-to-mesenchymal transition would constitute such a stromal component; the nal adult fates of such a cell would be hepatocytes and sinusoid vascular smooth muscle cells that would no longer possess the stromal capacity. In all major hematopoietic sites (AGM, fetal liver, and bone marrow), a new type of stem cells has been described: the mesenchymal stem cells (MSCs) (6, 29, 36). MSCs are multipotential cells that differentiate into adipocytes (A), osteoblasts (O), chondrocytes (C), and vascular smooth muscle cells (V) (reviewed in Refs. 5, 13). It has been suggested that MSCs are identical to stromal cells used as feeder layers for long-term cultures of HSCs (37). In this work, we searched for mesenchymal clonal lines with A, O, C, and V differentiation potential and with similar stromal capacity. We found that two lines had such potential, one being from adult bone marrow, BMC9, and the other from the fetal liver, AFT024. We then tried to establish the transcriptomic prole specic for stromal MSCs independent of the hematopoietic site. To this end, we ascertained the set of genes differentially expressed in the intersection set of AFT024 and BMC9 compared with those expressed in the union set of two negative control non-MSC lines. This set would account for the quadripotentiality and the stromal potential, i.e., the specic capacity of this particular type of stem cells.
MATERIALS AND METHODS

Downloaded from physiolgenomics.physiology.org on June 6, 2010

Cell Lines Lines used in this work were either 14.5 days postcoitum (dpc) fetal liver lines (AFT024, 2012, 2018, and BFC012) provided by K.A. Moore or bone marrow lines (BMC9 and BMC10) kindly provided by James E. Dennis (Case Western Reserve University, Cleveland, OH). All lines were immortalized using SV-40 large T-antigen. Large temperature-sensitive T-antigen (TSa-58) lines were made by retroviral transduction into the fetal liver stroma and clonally derived; the

Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org). Address for reprint requests and other correspondence: P. Charbord, Laboratoire dHematopoiese, Faculte de medecine, batiment Dutrochet, 10 Bvd ` Tonnelle, Tours 37032, France (e-mail: charbord@med.univ-tours.fr). 128

1094-8341/07 $8.00 Copyright 2007 the American Physiological Society

STROMAL MSC CHARACTERIZATION

129

clonal nature was veried by Southern blot analysis, which detected a single, unique proviral integration locus in the genomic DNA. The bone marrow lines had been obtained from transgenic immortomice generated by embryonic stem (ES) cells transfected with thermosensitive T under the major histocompatibility complex promoter. The lines were developed from isolated clones. Fetal liver cell lines were cultured at 33C in 5% CO2 in high-glucose DMEM with 10% (vol/vol) screened fetal calf serum (FCS, Hyclone), 0.1% -mercaptoethanol, 2 mM L-glutamine, and antibiotics. Bone marrow cell lines were cultured at 33C in 5% CO2 in -MEM with 10% FCS, 100 IU/ml (50 ng/ml) -interferon, 2 mM L-glutamine, antibiotics, and antimycotics. Differentiation Studies Osteoblastic differentiation was induced by culture in DMEM (high glucose) medium with 10% FCS, 0.1 M dexamethasone, 2 mM -glycerophosphate, and 150 M ascorbate-2-phosphate. Cells were seeded at 10,000 cells/cm2 and incubated for 21 days at 37C. Medium was changed every 4 days. Adipogenic differentiation was induced by culturing in DMEM (high glucose) medium with 20% FCS, 1 M dexamethasone, 0.35 M hydrocortisone, 0.5 mM isobutyl-methylxanthine (IBMX), 100 ng/ml insulin, and 60 M indomethacin. Cells were seeded at 20,000 cells/cm2 and incubated for 12 days at 37C. Medium was changed every 4 days. Chondrogenic differentiation was obtained in micropellets (3 105 cells/pellet) incubated at 37C for 21 days in 500 l of chondrogenic medium composed of DMEM (high glucose) with 0.1 M dexamethasone, 1 mM sodium pyruvate, 170 M ascorbic acid-2phosphate, 350 M proline, 1 insulin-transferrin-selenium, and 10 ng/ml transforming growth factor (TGF)- 1. Medium was changed every 4 days. RNA studies required the use of 20 30 micropellets. Vascular smooth muscle differentiation was obtained in long-term culture medium: McCoys 5A medium with 12.5% screened horse serum, 12.5% FCS, 20 M L-glutamine, 0.8 mM L-serine, 0.15 mM L-asparagine, 1 mM sodium pyruvate, 5 mM sodium bicarbonate, 1 M hydrocortisone, and antimycotics and antibiotics. Cells were passaged at the end of the rst week; during the two remaining weeks, medium was changed every 4 days. Cytochemical staining. Alkaline phosphatase was evaluated, after 4 and 15 days of culture in osteogenic medium, with the Alkaline Phosphatase Substrate Kit (Bio-Rad), added directly to the cell layer for 20 min at room temperature. For evaluation of the mineralized matrix, the cell layer was xed with 4% (vol/vol) formaldehyde, then stained by von Kossas stain using 5% (wt/vol) silver nitrate (Sigma) under ultraviolet light for 45 min, followed by 5% (wt/vol) sodium thiosulphate (Sigma) for 2 min: areas of mineralization were evaluated with LUCIA software (Nikon). For Nile O Red staining (NRO), cells were xed with 4% formaldehyde and stained with 1 g/ml Nile-red-O (Sigma) for 30 min. For safranin O staining, micropellets were xed with 4% formaldehyde for 24 h and then embedded in parafn. Slides were stained with 0.1% (wt/vol) safranin O (Sigma) for 5 min and counterstained with 0.02% (wt/vol) Fast Green for 3 min. Immunouorescence. Cells were cultured on Lab-Tek chamber slides (Nunc). Conuent layers were xed and permeabilized with methanol for 10 min. Slides were incubated for 1 h with mouse anti-human smooth muscle myosin heavy chains (hSMV, Sigma), followed by Alexa-488-conjugated goat anti-mouse antibody (Interchim), for 30 min. After staining, slides were mounted with Vectashield DAPI (Vector) and examined with a Leica DMR microscope (Leica Microsystems). RT-PCR. Total RNA was extracted using the RNeasy Tissue Kit (Qiagen) according to the manufacturers instructions. RNA concentration was determined by spectrophotometry at 260 nm, and RNA integrity was evaluated by electrophoresis of 2 g of total RNA in 1%
Physiol Genomics VOL
29

(wt/vol) agarose and 1 Tris-acetate/EDTA. RT-PCR was performed on 10 ng of total RNA using the SuperScript One-Step RT-PCR System with Platinum (Invitrogen) and Applied Biosystems PCRsystem 2700. Primer sequences are as follows (F, forward; R, reverse): Gapdh (F: ggtgaaggtcggtgtgaacg; R: catcatacttggcaggtttctcc; 55C, 761 bp); Acta2 (F: gctgttttcccatccatcgtg; R: tggtgatgatgccgtgttctatc; 56.2C, 148 bp); Agc1 (F: tcctctccggtggcaaagaagttg; R: ccaagttccagggtcactgttaccg; 55C, 270 bp); Bglap1 (F: ctgagtctgacaaagccttc; R: gctgtgacatccatacttgc; 60C, 313 bp); Des (F: cagatgagggagctggaggatc; R: tgggtgtgatatccgagagtgg; 56.1C, 449 bp); Fabp4 (F: ctcctcctcgaaggtttacaaaatg; R: cctttcataacacattccaccacc; 56C, 387 bp); Lpl (F: tgtaactgttctgccctcccc; R: ttctctcttgacagggcggc; 55C, 456 bp); Pparg (F: tcggaatcagctctgtggacc; R: aacagcttctccttctcggcc; 56C, 538 bp). For real-time quantitative RT-PCR, reverse transcription was performed using the HighCapacity cDNA Archive Kit (Applied Biosystems) on 3 g of total RNA; PCR was then performed using the Taqman-Low-Density Arrays Microuidic Cards (Applied Biosystems), according to the manufacturers instructions. Western blotting. For Western blotting, described in detail in Ref. 9, primary antibodies were mouse anti-human monoclonal antibodies (mAbs) from Sigma-Aldrich that are cross-reactive with mouse proteins: anti-vinculin (Vin11-5), anti-caldesmon (hCald), and antismooth muscle myosin heavy chains (hSMV). Secondary antibody was goat anti-mouse IgG conjugated to horseradish peroxidase (BioRad). Revelation by chemiluminescence was performed by ECLplus Western Blotting Detection Kit (Amersham Biosciences) and acquisition with Chemi-Smart 2000 using Chemocapt software (Vilbert-Lourmat). Flow cytometry. Cells were detached using phosphate-buffered saline (PBS) with 1 mM EDTA. Primary antibodies were puried rat anti-mouse mAbs from BD Biosciences: CD34 (RAM34), CD44 (IM-7), CD45 (30F-11), CD49d (R1-2AKR), CD49e (5H10-27), CD105 (MJ7/18), CD106 (429), and Sca-1 (D7) or their negative controls (rat IgG1, IgG2a, and IgG2b). Secondary antibody was phycoerythrin (PE)-conjugated goat anti-rat antibody. Cells were passed through a FACSCalibur ow cytometer (Becton-Dickinson) using a 488-nm argon laser. Samples were analyzed by collecting 10,000 events using Cell-Quest software (Becton-Dickinson). Hematopoiesis Supportive (Stromal) Capacity Cobblestone area-forming cell. The hematopoietic population of precursors was puried from 4- to 6-wk-old C57Bl/6J mouse cells. Sca1 /Lin cells were isolated following, rst, depletion of lineagepositive cells using a lineage cell depletion kit (Miltenyi) and, second, isolation of Sca1 cells with a PE selection kit (EasySep) using anti-Sca1 antibody conjugated to PE (D7, BD Biosciences). The protocol for the study was submitted to and approved by the local Ethical Committeee. The purity was 95%, as evaluated by ow cytometry. Sca1 /Lin cells were plated on irradiated (15Gy) conuent layers of AFT024 or BMC9 grown on 96-well plates, using 6 24 replicas of 7 cell concentrations (1200 cells/well). Cobblestone area-forming cell (CAFC) frequency was determined by limit-dilution analysis using Poisson statistics. Flow cytometry. Analysis of hematopoietic precursors co-cultivated for 5 wk with the stromal layers was carried out using ow cytometry. Cells were labeled with rat anti-mouse mAbs from BD Biosciences: anti-CD45 (30-F11) conjugated to allophycocyanin (APC), anti-Sca-1 (D7) conjugated to PE, and antibodies conjugated to FITC, either CD4 (GK1.5) CD8a (53-6.7) CD45R (B220), or CD11b (M1/70), CD51 (RMV-7), CD14 (55-3), and Ter119 (Ter-119). Gene Screening by CEA Chip The mouse CEA chip consists of 13,060 cDNA clones from early mouse embryo libraries constructed by the National Institute on Aging (NIA; http://www.nia.nih.gov) representing 11,000 genes. It also
www.physiolgenomics.org

Downloaded from physiolgenomics.physiology.org on June 6, 2010

130

STROMAL MSC CHARACTERIZATION

includes positive [actin, tubulin, oligo(dT), cot-1 DNA] and negative (Tris-EDTA/DMSO) controls present many times in the array for quality control. Microarrays were prepared by spotting PCR products as previously described. The DNA chip used for this experiment is described and available at Gene Expression Omnibus (GEO; GPL3438 and GPL3439). RNA extraction. For each line, RNA was extracted from two independent cultures. Total RNA was isolated using the Trizol reagent (Invitrogen) and dissolved in RNase-free water. The quality of RNA was controlled using an Agilent 2100 Bioanalyzer (Agilent Technologies), and concentration was measured by absorbance at 260 nm. For each target preparation, 20 g of total RNA were reverse transcribed using Superscript II RT (Life Technologies) in the presence of an oligo(dT)-primed reaction and an amino-modied nucleotide (aminoallyl-dUTP). Amino-modied cDNAs were puried through a Microcon Centricon 30 microconcentrator (Amicon) and ethanol precipitated. In a second step, monofunctional forms of Cy3 and Cy5 dyes (Amersham Biosciences) were coupled with the puried aminomodied cDNAs. Unincorporated uorescent molecules and salts were removed through the Microcon Centricon 30. Labeled cDNA was mixed with 10 g of poly(A) RNA (Boehringer), 10 g of tRNA (Life Technologies), and 10 g of mouse Cot1 DNA (Life Technologies). Puried labeled targets were hybridized to the array overnight as previously described (3). The cell lines 2018, BFC012, and BMC9 were each compared with AFT024. Each comparison was performed four times in both labeling orientations (8 hybridizations per comparison, 4 dye swaps). Microarray data processing. Fluorescence intensities of Cy3 and Cy5 were measured separately at 532 and 635 nm, respectively, with a laser scanner (Axon GenePix 4000A). Image analysis was performed with GenePix Pro 4.0 software (Axon). Spots with obvious blemishes were agged. Genes with null intensity values and low mean intensity were also annotated with a specic ag and excluded from the list of validated differentially expressed genes. Statistical analysis of experimental data was performed after an intensity-dependent LOWESS normalization (47) with BioConductor/Limma (version 1.8.16) (39), using a standard ANOVA performed on the logarithm of gene expression data without background subtraction as described in Ref. 12. We did not apply a threshold on the ratio values of gene expression but instead used the t-test to determine signicance as such: results signicantly different from 1 (P value 0.05) were determined from the eight values obtained for each comparison. The microarray data set used for this analysis is described and available from GEO (GSE4491). Gene network analysis. The interactions between genes identied as differentially expressed and all other genes were investigated using the Ingenuity Pathways Analysis (IPA). A series of networks was generated; a network is dened by IPA as the reection of all interactions of a given protein dened in the literature. The interaction indicates physical association or consists of induction/activation or repression/inactivation/degradation of one gene product by the other, eventually through another intermediary molecule (indirect interaction). The program calculates a statistical score based on a right-tailed Fishers exact test that calculates the t of each network to the set of studied genes. A score 2 indicates that there is a less than 1 in 100 chance that the studied genes are assembled into the network because of random chance. We considered only the networks with the highest scores ( 50). We introduced in the analysis two sets: that of upregulated genes and that of downregulated genes. The two major networks for up- and downregulated genes were then compared to nd out the interactions among genes belonging to either network.
RESULTS

best supporter (18). The mesenchymal differentiation potential of the fetal liver lines was not evaluated in previous works. On the contrary, in another report, the adult bone marrow line BMC9 was shown to be tripotent (with A, O, and C differentiation potential) and supportive for osteoclasts, while BMC10 was decient in both A differentiation and osteoclast support (14). AFT024 and BMC9 Are Quadripotent Mesenchymal Lines After 12 days in adipogenic medium, many AFT024 and BMC9 cells contained large ( 5 m in diameter) NRO vesicles (Fig. 1A), and adipocytic markers (Fig. 2A) were induced (lipoprotein lipase; Lpl) or increased [fatty acid-binding protein (Fabp4) and peroxisome proliferator-activated receptor- (Pparg)]. After 21 days in osteogenic medium, many von Kossa mineralized areas were apparent (Fig. 1A), and for AFT024, the osteocalcin (Bglap) osteoblastic marker was increased (Fig. 2A). Alkaline phosphatase was signicantly increased in both lines at day 4 (P 0.05) and day 15 (P 0.001) (Fig. 2B). After 21 days in chondrogenic culture conditions, safranin O well-formed micropellets were apparent (Fig. 1A), and the aggrecan-1 (Agc1) chondrocytic marker was clearly induced in both lines (Fig. 2A). After 21 days of cells being in long-term culture medium inducing V differentiation, immunouorescence studies indicated the presence of microlaments containing smooth muscle myosin in most AFT024 cells and about one-half of BMC9 cells (Fig. 1B). -Smooth muscle actin (Acta2) transcripts in AFT024 and desmin (Des) transcripts in both lines were slightly increased (Fig. 2A). Western blotting studies indicated that the smooth muscle myosin heavy chain SM1 was enhanced in AFT024, h-caldesmon was clearly induced in both lines, and metavinculin was detected in both (Fig. 2C). Finally, all cells from both lines expressed, at high a level, CD106, Sca-1, and CD49e; on the other hand, there was no expression of CD45, CD34, CD49d, and CD90 (not shown). In 2012 and BMC10 cells, after A differentiation, most NRO vesicles (Fig. 1A) were of small size, and Fabp4 and Pparg were enhanced, while Lpl was not detected in BMC10 (Fig. 1B). After O differentiation, many von Kossa mineralized areas were apparent (Fig. 1A), and Bglap was increased (Fig. 2A). Alkaline phosphatase was signicantly increased at day 4 (P 0.001) in 2012 and day 15 (P 0.001) in both lines (Fig. 2B). After C differentiation, micropellets contained few safranin O areas (Fig. 1A). After V differentiation, Acta2 and Des transcripts remained unchanged (but the band was faint in 2012) (Fig. 2A); h-caldesmon, SM1, and metavinculin were increased in 2012, but h-caldesmon and SM1 remained barely detectable in BMC10 (Fig. 2C). The immune phenotype was as described for AFT024 and BMC9 (not shown). In 2018 and BFC012, after A differentiation, few cells with small-sized NRO vesicles were detected (Fig. 1A); in 2018, but not BFC012, Lpl, Fabp4, and Pparg were induced (Fig. 2A). After O differentiation, von Kossa mineralized areas were not detected in either lines (Fig. 1A), but in 2018, Bglap was detected before differentiation and unchanged after (Fig. 2A). Alkaline phosphatase was signicantly increased at day 4 (P 0.001) in 2018 and day 15 (P 0.001) in both lines (Fig. 2B), which suggested that both lines were able to differentiate, more or less readily, into O but were defective
www.physiolgenomics.org

Downloaded from physiolgenomics.physiology.org on June 6, 2010

Previous studies have shown that the four 14.5 dpc fetal liver lines (AFT024, 2012, 2018, and BFC012) varied greatly in their ability to support hematopoiesis, with AFT024 being the
Physiol Genomics VOL
29

STROMAL MSC CHARACTERIZATION

131

Fig. 1. Differentiation potential of the cell lines, cellular markers. A: histochemistry (all lines, 1 representative experiment of 3). The following is presented for each line: morphology (phase contrast) of cells cultured in proliferation medium, staining by von Kossa stain after 21 days in osteogenic medium, staining by Nile Red O after 12 days in adipogenic medium, and staining by safranin O on micropellets maintained for 21 days in chondrogenic medium. Bar 100 m. B: immunouorescence for AFT024 and BMC9 cells stained with anti-smooth muscle (SM) myosin; for vascular smooth muscle cell (V) differentiation, cells were cultured for 21 days in the long-term culture medium, before xation and staining.

Downloaded from physiolgenomics.physiology.org on June 6, 2010

for mineralization. After C differentiation, micropellets contained no safranin O areas (Fig. 1A), and Agc1 induction was weak in BFC012 (Fig. 2A). After V differentiation, Des transcripts were induced in 2018, but barely in BFC012, and Acta2 was unchanged (Fig. 2A); a faint h-caldesmon band was apparent, but no metavinculin, while SM1 was unchanged in BFC012 (Fig. 2C). CD49e expression was low in both lines; Sca-1 was not detected in BFC012, and CD34 was expressed in both lines (not shown). In conclusion, AFT024 and BMC9 were clearly quadripotent and showed an immune phenotype of murine MSCs. All other lines showed impaired differentiation at one or several
Physiol Genomics VOL
29

levels. The lines with the most impaired differentiation potential were 2018 and BFC012; these lines also showed unusual expression of some of the membrane markers. AFT024 and BMC9 Are Stromal Lines The hematopoietic supportive ability was studied in both lines by seeding irradiated cells with bone marrow mouse Sca1 /Lin cells (Fig. 3). The CAFC frequency (Fig. 3, C and D) from two HSC enrichments after 5 wk was 1/40 1/50 and 1/110 1/160 for AFT024 and BCM9, respectively. Flow cytometry studies (Fig. 3B) indicated the persistence of immature
www.physiolgenomics.org

132

STROMAL MSC CHARACTERIZATION

Fig. 2. Differentiation stromal potential of the cell lines, molecular markers. A: RT-PCR. Expression of differentiationspecic transcripts for adipocytes (Lpl, Fabp4, and Pprag), osteoblasts (Bglap), chondrocytes (Agc1), and vascular smooth muscle cells (Des and Acta2). Housekeeping gene Gapdh. Before extraction, cells were cultured in respective media as indicated in Fig. 1. C, control medium; I, induction medium. B: alkaline phosphatase concentration as expressed by the ratio of day 4 to day 0 and day 15 to day 0. Statistical signicance (8 experiments per time point for each line) was evaluated using ANOVA: *P 0.05 and **P 0.01. C: Western blotting. Expression of vascular smooth muscle-specic proteins showing the induction or not of the 200-kDa smooth muscle myosin heavy chain (SM1), of h-caldesmon (h-Cald), and of metavinculin (MetaVcl). -Act, -actin.

Downloaded from physiolgenomics.physiology.org on June 6, 2010

Sca1 /CD45 hematopoietic cells along with more mature CD11b /CD45 granulocytes, CD14 /CD45 monocytes, and CD51 /CD45 megakaryocytes, without signicant difference between the two cell lines. Very few TER119 /CD45 erythroid cells were detected on BMC9. CD4 /CD45 , CD8 / CD45 , or CD45R /CD45 lymphoid cells were not detected. The presence of megakaryocytes, polymorphs, and monocytes was conrmed by cytology after May Grunwald Giemsa staining (Fig. 3A). In conclusion, both lines displayed hematopoietic support, with an advantage for AFT024 in terms of CAFC frequency. Molecular Prole of Stromal MSCs Results from the above cellular studies indicated that AFT024 and BMC9 were adequate models for stromal multipotential MSCs, whereas 2018 and BFC012 with severely defective stromal and differentiation potential were negative control lines. To determine the molecular prole specic for
Physiol Genomics VOL
29

stromal multipotential MSCs independent of the hematopoietic site, we ascertained the set of genes differentially expressed in the intersection set of AFT024 and BMC9 compared with those expressed in the union set of the two negative control lines 2018 and BFC012. The lines BMC10 and 2012, with partially defective differentiation and stromal potential, were not considered in the analysis. Our strategy was, rst, to determine the set of genes included in the union of genes expressed by 2018 and BFC012 (U1 2018 BFC012); second, to subtract this union set from genes expressed by AFT024 (S1 AFT024 U1) and by BMC9 (S2 BMC9 U1); and, third, to determine the intersection of S1 and S2 (S3 S1 S2). There were 346 upregulated and 299 downregulated genes in S3. Functional classication is shown in Fig. 4. Hierarchical clustering validated our strategy, the genes up- or downregulated in BMC9 (2 distinct extractions, 8 dye swaps) showing the greatest similarity to those in AFT024, whereas those in BFC012 showed the greatest dissimilarity (Fig. 5).
www.physiolgenomics.org

STROMAL MSC CHARACTERIZATION

133

Downloaded from physiolgenomics.physiology.org on June 6, 2010

Fig. 3. Stromal potential of AFT024 and BMC9. A: cytology of hematopoietic cells (Sca1 /Lin ) cultured for 5 wk on either line, showing the presence of megakaryocytes, polymorphs, and monocytes. B: ow cytometry studies performed on hematopoietic cells cultured for 5 wk on either line. C: cobblestone area formation from hematopoietic cells after 5 wk in culture. D: cobblestone area-forming cell (CAFC) frequency as measured after 5 wk for hematopoietic cells cultured at limit dilution.

Supplemental Table 1 (supplemental data are available at the online version of this article) gives the list of downregulated genes belonging to the following categories: cytoskeleton, transcription and chromatin remodeling, signaling, adhesion, trafcking, metabolism, cell cycle and apoptosis, protein synthesis and degradation, and other. Study of gene networks using Ingenuity software revealed a major network of 35 genes, with a highly signicant score of 64 (Network 1, Fig. 6A). Ten of the downregulated genes in this network coded for proteins implicated positively in osteogenic differentiation, proliferation, or transformation: protooncogene protein c-fos (Fos), parathyroid hormone-related protein (Pthlh), adrenomedullin (Adm), receptor activity-modifying protein 2 (Ramp2), apoptosis regulator Bcl-2 (Bcl2), ezrin (Vil2), estrogen receptor 1 (Esr1), proline-rich nuclear receptor co-activator 1 (Pnrc1), guanine nucleotide-binding protein G, -subunit (Gnas), and regulator of G protein signaling 2 (Rgs2). Supplemental Table 2 lists the upregulated genes grouped into the same categories as in Supplemental Table 1. Study of gene networks using the Ingenuity software revealed a major network of 35 genes, with a highly signicant score of 56 (Network 2, Fig. 6B). Most interactions in this network were
Physiol Genomics VOL
29

direct and consisted of physical association or induction. Transcripts belonged to two categories, transcripts for cytoskeletal molecules and transcripts implicated in the regulation of the transcriptional machinery (including chromatin remodeling). The genes of this network encoding proteins implicated in the cytoskeleton were as follows: the structural proteins tropomyosins (Tpm3, Tpm1), moesin (Msn), cytoplasmic 1 actin (Actb), cytoplasmic 2 actin (Actg1), the molecular chaperones for actin and tubulin, T-complex protein 1 subunit- (Cct3), - (Cct5), and - (Cct7), and molecules involved in the regulation of microlament or microtubule assembly/dynamics, rhophilin (Rhpn2), transforming protein RhoA (Rhoa), phospholipase D1 (Pld1), myosin regulatory light chain MRCL2 (Mrlc2), protein diaphanous homolog 3 (Diaph3), and ADPribosylation factor 1 (Arf1). Gene Ontology analysis conrmed the signicance (P 0.02 by Fishers exact test) of the overrepresentation of biological processes, cellular components, and molecular functions involved in cytoskeletal organization. The complete list of upregulated genes (Supplemental Table 2) conrmed the presence of many cytoskeletal genes not included in the network described above. The molecules of the network involved in the regulation of transcription were as
www.physiolgenomics.org

134

STROMAL MSC CHARACTERIZATION

Downloaded from physiolgenomics.physiology.org on June 6, 2010

Fig. 4. Functional classication of genes up- and downregulated in mesenchymal stem cells (MSCs). This classication was performed after analysis of the data in the literature for each of the genes.

follows: SWI/SNF-related matrix-associated actin-dependent regulator of chromatin (Smarcc1), the splicing factors heterogeneous nuclear ribonucleoprotein-A1 (Hnrpa1), -D0 (Hnrpd), -U (Hnrpu) and -K (Hnrpk), the arginine/serine-rich proteins 1 and 3 (Sfrs1, Sfrs3), ruvB-like 1 (Ruvbl1), actin-like protein 6A (Actl6a), putative RNA-binding protein 3 (Rbm3), GATAbinding protein 2 (Gata2), four and a half LIM domains protein 2 (Fhl2), zinc nger protein 638 (Znf638), high-mobility group nucleosome-binding domain-containing protein 2 (Hmgn2), protein ightless-1 homolog (Flii), and Cullin-1 (Cul1). Some of the cytoskeletal molecules are known to be associated to

molecules involved in the transcription machinery: Rhoa is functionally associated to Fhl2, as activated Rhoa induces the translocation of Fhl2 to the nucleus and its subsequent activation (31); Actl6a, Ruvbl1, and Smarcc1 are part of molecular complexes associating transcriptional regulators to molecules of the cytoskeleton such as Actb (15, 34). Two transcripts, Fhl2 and Ruvbl1, are implicated in the Wnt signaling pathway, known to modulate the transcriptional transactivation of -catenin in association with LEF1/TEF factors (45). Certain molecules of the network of downregulated genes (Network 1) are known to be physically associated to those of

Fig. 5. Hierarchical clustering of genes upand downregulated in MSCs. Each column indicates the level of gene expression in BMC9, 2018, and BFC012, as related to that in AFT024. Color scale indicates the degree of similarity. Gene clusters are shown at left and right sides (for upregulated and downregulated genes, respectively). The Sr-2 red box includes 11 genes present in the network of upregulated genes (shown in Fig. 6B).

Physiol Genomics VOL

29

www.physiolgenomics.org

STROMAL MSC CHARACTERIZATION

135

Downloaded from physiolgenomics.physiology.org on June 6, 2010

Fig. 6. Studies of gene networks up- and downregulated in MSCs. A: major network (Network 1) of genes downregulated in MSCs (as given by the Ingenuity software). The no. under each gene name indicates the mean log signal ratio for AFT024 (2018 BCF012); similar values were obtained for BMC9 (2018 BCF012). B: major network (Network 2) of genes upregulated in MSCs. A and B: regular line for physical association, arrow line for induction/activation, blunt-ended line for repression/inactivation/degradation, and discontinuous line for indirect interaction. Details about the interactions are given in Supplemental Table 4.

the network of upregulated genes (Network 2) (Fig. 6): villin 2 (Network 1) to moesin and cytoplasmic 1 actin (Network 2); and NF- B inhibitor- and Fos (Network 1) to different heterogeneous nuclear ribonucleoproteins and to cytoplasmic 2 actin, ADP-ribosylation factor 1, and RhoA (Network 2) (2, 10, 11, 17, 19, 22, 32, 44). Eleven (of 35) molecules of the network of upregulated genes were localized in a specic cluster, Sr-2, that contained 46 genes (Fig. 5). The molecules included in the two networks and their interacting partners are listed in Supplemental Table 4.
DISCUSSION

Our study indicates that two mouse mesenchymal clonal lines, one from bone marrow (BMC9) and the other from fetal liver (AFT024), possess the characteristics of stromal MSCs. Both lines were able to differentiate into adipocytes, mineralizing osteoblasts, chondrocytes, and vascular smooth muscle cells and supported Sca1 /Lin mouse bone marrow hematopoietic precursors, allowing the generation of cobblestone areas and the differentiation of myeloid precursors. On the contrary, in the other lines, the mesenchymal differentiation potential was impaired at one or several levels. The differentiation defect was most severe in 2018 and BFC012. These lines gave rise to few adipocytes, did not generate mineralizing osteoblasts, gave, under chondrogenic condition, loose aggregates that were safranin O negative, and expressed, after V differentiation, low amounts of h-caldesmon. In addition, it has
Physiol Genomics VOL
29

been shown in a previous study that none of these two lines allowed the maintenance of HSCs (18). It has been suggested that the two populations, that of MSCs (isolated by their adherence to plastic and containing stem cells/ progenitors for the A, O, C lineages) and that of stromal cells (used as feeder layers for long-term cultures of HSCs), are identical (37). This hypothesis has been sustained for cultureexpanded bone marrow primary human MSCs (28). Our study extends this concept of stromal MSCs to clonal cell populations derived not only from bone marrow but also from fetal liver. To determine the transcriptomic prole specic for stromal multipotential MSCs independent of the hematopoietic site, we ascertained the set of genes differentially expressed in the intersection set of AFT024 and BMC9 compared with those expressed in the union set of the two negative control lines, 2018 and BFC012. We found that 645 transcripts were specifically regulated in the two stromal multipotential cell lines, with a similar number of up- and downregulated genes. Some of the upregulated transcripts appeared to be of special interest and already have been described in the literature on stromal cells: the serine/threonine kinase NLK (Nlk) and the Jumonji protein 1a and 2 (Jarid1a, Jarid2), because of their demonstrated role as stromal mediator and differentiating factor for stromal cells (23, 25), and the lymphocyte-specic adapter protein Lnk (Lnk), because of its implication in the interaction between HSCs and cells from the microenvironment (41).
www.physiolgenomics.org

136 Table 1. Stemness gene candidates

Trafcking Cytoskeleton Metabolism

STROMAL MSC CHARACTERIZATION

Signaling

Cops4 Kpna2 Laptm5 Snx2 Syt11 Tapbp

Actb Cct3 Cct5 Tpm3

Acadm Fasn Nme2 Slc25a5

Catnal Hdgf Ly6e M6pr Mlc1 Ppp2r5c Ywhaq

Transcription and Chromatin Remodeling

Cell Cycle and Apoptosis

Protein Synthesis and Degradation

Ccnb1 Ccng1 Diablo Hipk3 Nasp Ccnb1

Eif4g2 Hspa8 Pigf Psmb4 Psmc5 Sdf4

Cpsf3 Gata2 Hmgn2 Hnrpa1 Hnrpu Psip1 Ptma Sfrs3

Transcripts expressed in stromal mesenchymal stem cells (MSCs) and in other stem cell types. For details, see Supplemental Table 3 (supplemented data are available at the online version of this article).

As shown in Table 1, 42 transcripts in our set of upregulated genes were also found upregulated in other embryonic or adult hematopoietic or neural stem cells (4, 21, 27, 33, 38, 42, 43, 49). Such transcripts may be characteristic of the stem cell state (50). Studies using short interfering RNA (20, 40) should conrm this hypothesis. Because our gene chip encompasses about only one-third of the known mammalian genome, the number of up- and downregulated genes might be larger than what is described here. In particular, some categories found in other genetic screens as implicated in stromal support (8) are insufciently represented in the chip used here (especially proteoglycans and chemokines). We chose to study specic clones that were immortalized using a temperature-sensitive SV-40 T-antigen for the following reasons. 1) These clones have been studied in detail previously (14, 18), either for their stromal capacity (fetal liver lines) or for their differentiation potential (bone marrow lines). 2) The introduction of the viral sequence would insure sufcient proliferation, in particular when studying chondrogenic differentiation using the micropellet technique, requiring large cell numbers. 3) Mouse cells immortalized by T show a phenotype comparable to nontransduced cells (9). To assess the potential inuence of viral genes on the expression of cell genes, we have recently compared, using microuidic cards from Applied Biosystems, the expression of 380 transcripts in clones derived from primary cultures of mouse bone marrow MSCs vs. BMC9 and BMC10. We have found that 7% of the transcripts were up- or downregulated in cells with large T. The studied transcripts belonged to several categories. The least affected category was that of cytoskeletal components (only 3% of the genes in this category were signicantly modulated). On the other hand, the most affected category was that of molecules involved in cell cycle and apoptosis: 39% of the studied cyclins, cyclin inhibitors, and apoptosis inhibitors were upregulated in BMC9 and BMC10. Molecules implicated in the transcriptional machinery were only marginally affected (6%). These data indicate that the
Physiol Genomics VOL
29

expression of genes belonging to the cell cycle and apoptosis category (see Supplemental Tables 1 and 2) may have been inuenced by the viral sequence insertion. However, most genes present in the networks (vide below) are probably minimally inuenced by the presence of T. Using the Ingenuity software, we found two major networks with highly signicant scores 50. The network of 35 downregulated genes contained 10 transcripts positively implicated in osteoblastic differentiation, proliferation, or transformation. These data suggest that the intersection set of AFT024 and BMC9 is representative of bona de undifferentiated cells, in contrast to the set of 2018 and BFC012, which appears to represent cells already committed to the osteogenic pathway albeit unable to complete it by mineralization when placed in an osteogenic condition. The network of 35 upregulated genes belonged to two categories, transcripts for cytoskeletal molecules and transcripts implicated in the regulation of the transcriptional machinery (including chromatin remodeling). Cytoskeletal molecules included structural components of actin-based cytoskeleton, chaperones for actins and tubulins, and proteins involved in the signaling pathways leading to the assembly/disassembly of microtubules and microlaments. Some of the molecules implicated in the transcriptional machinery have been detected in protein complexes associated with cytoskeletal components, such as the NuA4 histone acetylase complex (15) or the BAF53 complex (34). Networks have not been emphasized in previous transcriptomic studies on stem cells (4, 21, 27, 33, 35, 38, 40, 42, 43, 46, 49). The networks described in the present study may serve as a basis for future experiments. First, one may assess whether the same networks are found in MSCs from primary cultures (polyclonal and clonal cultures). Second, one may investigate whether the hyperexpression of some of the network cytoskeletal components (either structural, such as moesin or cytoplasmic actin 1, or regulatory, such as RhoA) induces the hyperexpression or downregulation of some of the molecules implicated in the transcription machinery and in the differentiation pathways. As a nal point, it will be essential to verify whether the hyperexpression of a cytoskeletal component belonging to the gene network of upregulated molecules improves or hampers the differentiation and/or stromal potentials of the transduced cells. Genetic regulatory networks provide essential information on the generation of the different cell types (26). Future experiments should indicate whether the gene networks presently described are genetic regulatory networks (i.e., include molecules involved in positive or negative regulatory loops) that determine the maintenance in a nondifferentiated state of this particular stem cell type. In conclusion, this study describes gene networks for stromal MSCs from the two major hematopoietic sites that should help discriminate this cell type from other stem cells (ES, neural, hematopoietic, etc.) and help dene the stem cell state.
ACKNOWLEDGMENTS We thank Jorge Domenech, Olivier Herault, and Cyrille Petat for helpful discussion on different methodological aspects of this work. GRANTS This work was supported by grants from the European Union (Integrated Project Genostem no. 503161), from Fondation de France (no. 2002002384), www.physiolgenomics.org

Downloaded from physiolgenomics.physiology.org on June 6, 2010

STROMAL MSC CHARACTERIZATION from Association de Recherche sur le Cancer (no. 5827), from Electricite de France, and from the National Heart, Lung, and Blood Institute (NIH-HL58739). S. Chateauvieux was supported in part by the Societe Francaise dHematologie. REFERENCES 1. Adams GB, Scadden DT. The hematopoietic stem cell in its place. Nat Immun 7: 333337, 2006. 2. Adilakshmi T, Laine RO. Ribosomal protein S25 mRNA partners with MTF-1 and La to provide a p53-mediated mechanism for survival or death. J Biol Chem 277: 4147 4151, 2002. 3. Benchaouir R, Rameau P, Decraene C, Dreyfus P, Israeli D, Pietu G, Danos O, Garcia L. Evidence for a resident subset of cells with SP phenotype in the C2C12 myogenic line: a tool to explore muscle stem cell biology. Exp Cell Res 294: 254 268, 2004. 4. Bhattacharya B, Miura T, Brandenberger R, Mejido J, Luo Y, Yang AX, Joshi BH, Ginis I, Thies RS, Amit M, Lyons I, Condie BG, Itskovitz-Eldor J, Rao MS, Puri RK. Gene expression in human embryonic stem cell lines: unique molecular signature. Blood 103: 2956 2964, 2004. 5. Bianco P, Gehron Robey P. Marrow stromal stem cells. J Clin Invest 105: 16631668, 2000. 6. Campagnoli C, Roberts IA, Kumar S, Bennett PR, Bellantuono I, Fisk NM. Identication of mesenchymal stem/progenitor cells in human rsttrimester fetal blood, liver, and bone marrow. Blood 98: 2396 2402, 2001. 7. Chagraoui J, Lepage-Noll A, Anjo A, Uzan G, Charbord P. Fetal liver stroma consists of cells in epithelial-to-mesenchymal transition. Blood 101: 29732982, 2003. 8. Charbord P, Moore K. Gene expression in stem cell-supporting stromal cell lines. Ann NY Acad Sci 1044: 159 167, 2005. 9. Charbord P, Oostendorp R, Pang W, Herault O, Noel F, Tsuji T, Dzierzak E, Peault B. Comparative study of stromal cell lines derived from embryonic, fetal, and postnatal mouse blood-forming tissues. Exp Hematol 30: 12021210, 2002. 10. Chen CY, Xu N, Shyu AB. Highly selective actions of HuR in antagonizing AU-rich element-mediated mRNA destabilization. Mol Cell Biol 22: 7268 7278, 2002. 11. Davis M, Hatzubai A, Andersen JS, Ben-Shushan E, Fisher GZ, Yaron A, Bauskin A, Mercurio F, Mann M, Ben-Neriah Y. Pseudosubstrate regulation of the SCF(beta-TrCP) ubiquitin ligase by hnRNP-U. Genes Dev 16: 439 451, 2002. 12. Decraene C, Benchaouir R, Dillies MA, Israeli D, Bortoli S, Rochon C, Rameau P, Pitaval A, Tronik-Le Roux D, Danos O, Gidrol X, Garcia L, Pietu G. Global transcriptional characterization of SP and MP cells from the myogenic C2C12 cell line: effect of FGF6. Physiol Genomics 23: 132149, 2005. 13. Dennis JE, Charbord P. Origin and differentiation of human and murine stroma. Stem Cells 20: 205214, 2002. 14. Dennis JE, Merriam A, Awadallah A, Yoo JU, Johnstone B, Caplan AI. A quadripotential mesenchymal progenitor cell isolated from the marrow of an adult mouse. J Bone Miner Res 14: 700 709, 1999. 15. Doyon Y, Cote J. The highly conserved and multifunctional NuA4 HAT complex. Curr Opin Genet Dev 14: 147154, 2004. 16. Dzierzak E, Oostendorp R. Hematopoietic stem cell development in mammals. In: Hematopoiesis. A Developmental Approach. Oxford, UK: Oxford Univ. Press, 2001, p. 209 217. 17. Gutkind JS. Cell growth control by G protein-coupled receptors: from signal transduction to signal integration. Oncogene 17: 13311342, 1998. 18. Hackney JA, Charbord P, Brunk BP, Stoeckert CJ, Lemischka IR, Moore KA. A molecular prole of a hematopoietic stem cell niche. Proc Natl Acad Sci USA 99: 1306113066, 2002. 19. Hay DC, Kemp GD, Dargemont C, Hay RT. Interaction between hnRNPA1 and IkappaBalpha is required for maximal activation of NFkappaB-dependent transcription. Mol Cell Biol 21: 34823490, 2001. 20. Ivanova N, Dobrin R, Lu R, Kotenko I, Levorse J, DeCoste C, Schafer X, Lun Y, Lemischka IR. Dissecting self-renewal in stem cells with RNA interference. Nature 442: 533538, 2006. 21. Ivanova NB, Dimos JT, Schaniel C, Hackney JA, Moore KA, Lemischka IR. A stem cell molecular signature. Science 298: 601 604, 2002. 22. Johnston IM, Spence HJ, Winnie JN, McGarry L, Vass JK, Meagher L, Stapleton G, Ozanne BW. Regulation of a multigenic invasion programme by the transcription factor, AP-1: re-expression of a downregulated gene, TSC-36, inhibits invasion. Oncogene 19: 5348 5358, 2000. Physiol Genomics VOL
29

137

23. Kitajima K, Kojima M, Nakajima K, Kondo S, Hara T, Miyajima A, Takeuchi T. Denitive but not primitive hematopoiesis is impaired in jumonji mutant mice. Blood 93: 8795, 1999. 24. Kopp HG, Avecilla ST, Hooper AT, Rai S. The bone marrow vascular niche: home of HSC differentiation and mobilization. Physiology Bethesda 20: 349 356, 2005. 25. Kortenjann M, Nehls M, Smith AJ, Carsetti R, Schuler J, Kohler G, Boehm T. Abnormal bone marrow stroma in mice decient for nemo-like kinase, Nlk. Eur J Immunol 31: 3580 3587, 2001. 26. Loose M, Patient R. Global genetic regulatory networks controlling hematopoietic cell fates. Curr Opin Hematol 13: 229 236, 2006. 27. Ma X, Husain T, Peng H, Lin S, Mironenko O, Maun N, Johnson S, Tuck D, Berliner N, Krause DS, Perkins AS. Development of a murine hematopoietic progenitor complementary DNA microarray using a subtracted complementary DNA library. Blood 100: 833 844, 2002. 28. Majumdar MK, Thiede MA, Haynesworth SE, Bruder SP, Gerson SL. Human marrow-derived mesenchymal stem cells (MSCs) express hematopoietic cytokines and support long-term hematopoiesis when differentiated toward stromal and osteogenic lineages. J Hematother Stem Cell Res 9: 841 848, 2000. 29. Mendes SC, Robin C, Dzierzak E. Mesenchymal progenitor cells localize within hematopoietic sites throughout ontogeny. Development 132: 11271136, 2005. 30. Moore KA, Lemischka IR. Stem cells and their niches. Science 311: 1880 1885, 2006. 31. Muller JM, Metzger E, Greschik H, Bosserhoff AK, Mercep L, Buettner R, Schule R. The transcriptional coactivator FHL2 transmits Rho signals from the cell membrane into the nucleus. EMBO J 21: 736 748, 2002. 32. Ostrowski J, Kawata Y, Schullery DS, Denisenko ON, Higaki Y, Abrass CK, Bomsztyk K. Insulin alters heterogeneous nuclear ribonucleoprotein K protein binding to DNA and RNA. Proc Natl Acad Sci USA 98: 9044 9049, 2001. 33. Park IK, He Y, Lin F, Laerum OD, Tian Q, Bumgarner R, Klug CA, Li K, Kuhr C, Doyle MJ, Xie T, Schummer M, Sun Y, Goldsmith A, Clarke MF, Weissman IL, Hood L, Li L. Differential gene expression proling of adult murine hematopoietic stem cells. Blood 99: 488 498, 2002. 34. Park J, Wood MA, Cole MD. BAF53 forms distinct nuclear complexes and functions as a critical c-Myc-interacting nuclear cofactor for oncogenic transformation. Mol Cell Biol 22: 13071316, 2002. 35. Phinney DG, Hill K, Michelson C, DuTreil M, Hughes C, Humphries S, Wilkinson R, Baddoo M, Bayly E. Biological activities encoded by the murine mesenchymal stem cell transcriptome provide a basis for their developmental potential and broad therapeutic efcacy. Stem Cells 24: 186 198, 2006. 36. Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, Mosca JD, Moorman MA, Simonetti DW, Craig S, Marshak DR. Multilineage potential of adult human mesenchymal stem cells. Science 284: 143147, 1999. 37. Prockop DJ. Marrow stromal cells as stem cells for nonhematopoietic tissues. Science 276: 7174, 1997. 38. Ramalho-Santos M, Yoon S, Matsuzaki Y, Mulligan RC, Melton DA. Stemness: transcriptional proling of embryonic and adult stem cells. Science 298: 597 600, 2002. 39. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol 3: Article 3, 2004. 40. Song L, Webb NE, Song Y, Tuan RS. Identication and functional analysis of candidate genes regulating mesenchymal stem cell self-renewal and multipotency. Stem Cells 24: 17071718, 2006. 41. Takizawa H, Kubo-Akashi C, Nobuhisa I, Kwon SM, Iseki M, Taga T, Takatsu K, Takaki S. Enhanced engraftment of hematopoietic stem/ progenitor cells by the transient inhibition of an adaptor protein, Lnk. Blood 107: 2968 2975, 2006. 42. Terskikh AV, Easterday MC, Li L, Hood L, Kornblum HI, Geschwind DH, Weissman IL. From hematopoiesis to neuropoiesis: evidence of overlapping genetic programs. Proc Natl Acad Sci USA 98: 7934 7939, 2001. 43. Terskikh AV, Miyamoto T, Chang C, Diatchenko L, Weissman IL. Gene expression analysis of puried hematopoietic stem cells and committed progenitors. Blood 102: 94 101, 2003. 44. Tsukita S, Yonemura S. Cortical actin organization: lessons from ERM (ezrin/radixin/moesin) proteins. J Biol Chem 274: 3450734510, 1999. www.physiolgenomics.org

Downloaded from physiolgenomics.physiology.org on June 6, 2010

138

STROMAL MSC CHARACTERIZATION addressing single and multiple slide systematic variation. Nucleic Acids Res 30: e15, 2002. 48. Yin T, Li L. The stem cell niches in bone. J Clin Invest 116: 11951201, 2006. 49. Zhou G, Chen J, Lee S, Clark T, Rowley JD, Wang SM. The pattern of gene expression in human CD34( ) stem/progenitor cells. Proc Natl Acad Sci USA 98: 13966 13971, 2001. 50. Zipori D. The stem state: plasticity is essential, whereas self-renewal and hierarchy are optional. Stem Cells 23: 719 726, 2005.

45. Wei Y, Renard CA, Labalette C, Wu Y, Levy L, Neuveut C, Prieur X, Flajolet M, Prigent S, Buendia MA. Identication of the LIM protein FHL2 as a coactivator of beta-catenin. J Biol Chem 278: 5188 5194, 2003. 46. Wieczorek G, Steinhoff C, Schulz R, Scheller M, Vingron M, Ropers HH, Nuber UA. Gene expression prole of mouse bone marrow stromal cells determined by cDNA microarray analysis. Cell Tissue Res 311: 227237, 2003. 47. Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J, Speed TP. Normalization for cDNA microarray data: a robust composite method

Downloaded from physiolgenomics.physiology.org on June 6, 2010

Physiol Genomics VOL

29

www.physiolgenomics.org