Erratum to “Maintenance of differentiation potential of human bone marrow mesenchymal stem cells...

www.elsevier.com/locate/ybbrc

Biochemical and Biophysical Research Communications 326 (2005) 527–538

BBRC

Maintenance of differentiation potential of human bone marrowmesenchymal stem cells immortalized by human telomerasereverse transcriptase gene in despite of extensive proliferation

Basem M. Abdallaha, Mandana Haack-Sørensena, Jorge S. Burnsa, Birgitte Elsnaba,Franz Jakobb, Peter Hoklandc, Moustapha Kassema,*

a Laboratory for Molecular Endocrinology (KMEB), Department of Endocrinology and Metabolism, University Hospital of Odense, Odense, Denmarkb Experimental and Clinical Osteology, Orthopaedic Department, University of Wurzburg, Wuerzburg, Germany

c Department of Hematology, University Hospital of Aarhus, Aarhus, Denmark

Received 2 November 2004Available online 24 November 2004

Abstract

Human bone marrow mesenchymal stem cells (hMSC) represent a population of stem cells that are capable of differentiation intomultiple lineages. However, these cells exhibit senescence-associated growth arrest and phenotypic changes during long-term in vitroculture. We have recently demonstrated that overexpression of human telomerase reverse transcriptase (hTERT) in hMSC recon-stitutes telomerase activity and extends life span of the cells [Nat. Biotechnol. 20 (2002) 592]. In the present study, we have per-formed extensive characterization of three independent cell lines derived from the parental hMSC-TERT cell line based ondifferent plating densities during expansion in culture: 1:2 (hMSC-TERT2), 1:4 (hMSC-TERT4), and 1:20 (hMSC-TERT20).The 3 cell lines exhibited differences in morphology and growth rates but they all maintained the characteristics of self-renewingstem cells and the ability to differentiate into multiple mesoderm-type cell lineages: osteoblasts, adipocytes, chondrocytes, and endo-thelial-like cells over a 3-year period in culture. Also, surface marker studies using flow cytometry showed a pattern similar to thatknown from normal hMSC. Thus, telomerization of hMSC by hTERT overexpression maintains the stem cell phenotype of hMSCand it may be a useful tool for obtaining enough number of cells with a stable phenotype for mechanistic studies of cell differenti-ation and for tissue engineering protocols.� 2004 Elsevier Inc. All rights reserved.

Keywords: Telomerase; Stem cell; Differentiation; Proliferation; Osteoblast

Human mesenchymal stem cells or marrow stromalcells (hMSC) represent a non-hematopoietic cell popula-tion present in bone marrow with the ability for self-re-newal and multilineage differentiation [2,3]. Isolationand characterization of hMSC have been performedby several investigators using a variety of techniquesincluding plastic adherence [4], positive selection usingSTRO-1 antibody [5], selection based on adherence to

0006-291X/$ - see front matter � 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.bbrc.2004.11.059

* Corresponding author. Fax: +45 65912943.E-mail address: [email protected] (M.

Kassem).

fibronectin in low serum condition [6] or in vitro growthunder low oxygen tension in presence of hematopoieticcells [7]. In all these methods, populations of cells wereobtained with the ability to differentiate into differentmesoderm-type cells including osteoblasts, adipocytes,chondrocytes, myocytes, endothelial cells, and possiblyalso to endoderm and ectoderm lineage cells includingneural cells and heptaocytes [8].

Long-term in vitro culture of hMSC is associatedwith reduced proliferative potential due to cellularsenescence leading finally to growth arrest, a phenome-non known as replicative senescence [9,10]. Also, during

mailto:[email protected]

528 B.M. Abdallah et al. / Biochemical and Biophysical Research Communications 326 (2005) 527–538

long-term culture the senescent cells exhibit impaireddifferentiation capacity, thus limiting the ability to ex-pand hMSC to a large number of cells needed for ther-apeutic applications [10,11].

Replicative senescence is a general phenomenon expe-rienced by most somatic diploid cells due to the telomereloss occurring during cell replication. Telomere length ismaintained by telomerase, a cellular ribonucleoproteincomplex consisting of an integral RNA component(hTR), which serves as the telomeric template, a cata-lytic subunit (human telomerase reverse transcriptase,hTERT), which has reverse transcriptase activity, andassociated protein components [12]. We [1] and others[1,13,14] have previously demonstrated that culturedhMSC lack telomerase activity because of absence ofhTERT expression and that the hMSC undergo progres-sive telomere shortening during serial passaging [10,15].We have also demonstrated that ectopic overexpressionof hTERT in hMSC restored the telomerase activity andelongated telomere lengths as well as extended life spanof the cells [1]. Similar results have also been obtained inother cell types including fibroblasts [16], endothelialcells [17], and chondrocytes [18].

In addition to its role in maintenance of cell prolifer-ation potential, the role of telomerase in maintaining thebiological characteristic of stem cells in long-termcultures is not known. One recent study has suggestedimpaired differentiation potential of MSC in telome-rase-deficient mice (deficient in RNA component:mTR�/�) [19]. Thus, we examined the effect of mainte-nance of high levels of telomerase activity on self-re-newal and differentiation capacity of hMSC duringlong-term culture using the hMSC-TERT cell lines pre-viously established in our laboratory [1]. Our resultsdemonstrate that maintenance of telomerase activity issufficient for maintenance of stem cell characteristicsof hMSC during long-term culture in vitro.

Materials and methods

Cell culture

The establishment and characterization of hMSC-TERT cell linehave been described previously [1]. hMSC-TERT cells were grown in astandard growth medium consisting of phenol red-free minimalessential medium (MEM) supplemented with 10% fetal calf serum(FBS) (Gibco Invitrogen, Tastrup, Denmark), batch tested for growthof primary human bone marrow stromal cells and 1% penicillin/streptomycin (Gibco Invitrogen). All cells were maintained in ahumidified incubator at 37 �C and 5% CO2.

For the present studies, we have established three different celllines from the parental cell line (hMSC-TERT) based on differentsplit ratios used during subculturing. At 8th passage post-retroviraltransduction (population doubling level, PDL 23), cells were main-tained at a split ratio of 1:2 at each subsequent culture creatinghMSC-TERT2 cell line or at a split ratio of 1:4 to obtain hMSC-TERT4 cell line or at a split ratio of 1:20 creating hMSC-TERT20cell line.

Cell proliferation studies

Long-term cell growth in vitro was determined by calculatingcumulative population doubling level (PDL). At confluence the cellswere incubated with trypsin/EDTA (Gibco Invitrogen) and cells werecounted using a hemocytometer. At each passage, the initial cellnumber and number of cells at confluence were counted and popula-tion doubling PD was determined using the formula: log N/log 2 whereN is the number of cells at confluence divided by the initial cell num-ber. Cumulative PDL is thus the sum of population doublings.

Telomerase activity assay

Cells were cultured in standard growth medium to 80% confluence.Protein extracts from cell pellets were prepared by using 1· Chaps lysismethod [20]. Telomerase activity was performed using a modifiedtelomeric repeat amplification method (TRAP assay) described bySzatmari and Aradi [21] and using TRAPeze telomerase detection kit(Intergen, Oxford, UK).

Quantitative real-time TRAP assay

The SYBR Green real-time quantitative TRAP (RTQ-TRAP) as-say was performed using a modification of the method described [22].Briefly, RTQ-PCR amplification was performed with serial dilution ofcell extracts, including heat-inactivated samples (1000, 100, 10, and 1cell equivalents, to ensure the linearity of the assay). PCR was con-ducted in a final volume of 20 lL with IQTM Supermix (Bio-Rad,Herlev, Denmark) containing universal primers TS 5 0-AATCCGTCGAGCAGAGTT-3 0 and RP 5 0-GCGCGG(CTTACC)3CTAACC-3 0.For amplification, samples were incubated for 30 min at 25 �C, fol-lowed by 35 cycles of 30 s at 95 �C and 30 s at 60 �C. The threshold ofCt value was set above the negative controls (lysis buffer only and heat-inactivated samples) to avoid any background interference. All sam-ples were analyzed in triplicate. The HT-1080 cell line [23] was used asan independent positive control reference. Telomerase activity wasrepresented as a percentage of the activity in the positive control.

Flow cytometry (FACS) analysis

Cells were cultured in standard growth medium to 80% confluence.Cells were trypsinized, stained for viability with propidium iodide (PI)for 10 min, and assayed for size and granularity based on forward andside scattering by FACScan flow cytometer (Becton–Dickinson, SanJose, CA) linked with Cell-Quest 3.1 software (Becton–Dickinson).

For surface marker studies the cells were cultured in standardgrowth medium to 100% confluence, then detached with trypsin, andwashed and stained with specific non-conjugated primary antibodiesfor 30 min on ice followed by immunofluorescent secondary antibod-ies. Samples were washed and fixed with 1% of paraformaldehyde untilanalysis with FACScan (Becton–Dickinson).

The quantitation of adipocyte differentiation (see below) analysisbased on Nile Red lipophilic stain was performed as described [24].

Cell differentiation studies

Osteoblast differentiation. Cells were seeded at a density of 104 cells/cm2 in 60 cm2 Petri-dishes (for RNA isolation) or in 6-well plates (forhistochemical staining) and grown for 24 h in standard growth med-ium. At 60–70% cell confluence, the medium was replaced by a newmedium supplemented with 10�8 M calcitriol (1,25 dihydroxy vitaminD3, vit.D3) (Leo, Ballerup, Denmark) and this medium was replacedevery third day. For in vitro mineralization experiments, cells wereplated at 104 cells/cm2 in 6-well plates in media supplemented with50 lg/mL LL-ascorbic acid (Wako Chemicals GmbH, Neuss, Germany),10 mM b-glycerophosphate (Calbiochem, Albertslund, Denmark), and

B.M. Abdallah et al. / Biochemical and Biophysical Research Communications 326 (2005) 527–538 529

10�8 M calcitriol. The medium and its supplementation were changedevery 3 days and the experiments were terminated after 7 days ofculture.

Chondrocyte differentiation.Aliquots of 2 · 105 cells were pelleted in15 mL polypropylene tube by centrifugation and cultured as high-density pellet cultures for 3 weeks in 0.5 mL of serum-free chemicallydefined medium consisting of DMEM (BioWhittaker, Walkersville,MD) supplemented with 10 ng/mL TGF-b1 (R&D,Minneapolis, MN),100 nM dexamethasone, 50 mg/mL ascorbate 2-phosphate, 100 mg/mL sodium pyruvate, 40 mg/mL proline, and ITS-plus (CollaborativeBiomedical Products, Cambridge, MA; final concentrations: 6.25 mg/mL bovine insulin, 6.25 mg/mL transferrin, 6.25 mg/mL selenous acid,5.33 mg/mL linoleic acid, and 1.25 mg/mL bovine serum albumin)[25,26]. Cells were incubated for 3 weeks at 37 �C in 5% CO2 and themedium was changed every 3–4 days. Control cells were culturedwithout TFG-b1. The pellets were either paraffin embedded for stainingor directly stored at �80 �C for RNA extraction.

Adipocyte differentiation. Cells were seeded at 3 · 104 cells/cm2 in60 cm2 Petri-dishes (for RNA isolation) or in 6-well plates (for histo-chemical staining and flow cytometry studies) and cultured in a stan-dard growth medium. At 90–100% cell confluence, the medium wasreplaced by high-glucose Dulbecco�s modified MEM (DMEM) (GibcoInvitrogen) containing 10% FBS and supplemented with adipogenic-induction mixture (AIM) [containing 10�7 M dexamethasone (dex),0.45 mM isobutyl methyl xanthine (IBMX), 2 · 10�6 M insulin (allfrom, Sigma–Aldrich, Vallensbaek strand, Denmark), and 1 lMRosiglitazone [(BRL49653) (Novo Nordisk, Bagsvaerd, Denmark)].The adipogenic medium was replaced every 3 days.

Endothelial cell differentiation. Cells were cultured at a density of1 · 102 cells/cm2 in 12-well plate in the previously described endothe-lial differentiation medium (EDM) [27] which consists of 60% lowglucose DMEM (Gibco) and 40% MCDB-201 medium (Sigma) sup-

Table 1PCR primer sequences and their amplified product size

Gene Primer sequence

b-Actin 50-AGCCATGTACGT50-AGTCCGCCTAGA

Collagen type 1 (Col 1) 50-TGACGAGACCAA50-CCATCCAAACCAC

Osteocalcin (OC) 50-CATGAGAGCCCT50-AGAGCGACACCC

Osteopontin (OPN) 50-CCAAGTAAGTCC50-GGTGATGTCCTC

CBFA1/Runx2 50-TCTTCACAAATCC50-TGGATTAAAAGG

ALP 50-ACGTGGCTAAGA50-CTGGTAGGCGAT

PPAR-c 2 50-TTCTCCTAT TGA50-CTCCACTTTGATT

Adiponectin (APM1) 50-TGTTGCTGGGAG50-ATGTCTCCCTTAG

ADD1 50-GGAGCCATGGAT50-ATCTTCAATGGAG

aP2 50-GCCAGGAATTTG50-TGG TTG ATT TT

LPL 50-GAGATTTCTCTGT50-CTGCAAATGAGA

VEGF (A) 50-CTACCTCCACCAT50-TGATTCTGCCCTC

EPSA1 50-GCGCTAGACTCC50-TGGCCACTTACTA

ETB 50-CTGCTGCACATCG50-GCTCCAAATGGC

plemented with 1· linoleic acid BSA 100 mg/mL BSA (Sigma), 10�8 Mdexamethasone (Sigma), 10�4 M ascorbic acid 2-phosphate (WAKO),1 · ITS (1.0 mg/mL insulin from bovine pancreas, 0.55 mg/mL humantransferrin, and 0.5 lg/mL sodium selenite) (Sigma), and 10 ng/mLVEGF-B (R&D systems, MN, USA). Cells were allowed to differen-tiate in EDM for 7 days and medium was exchanged every 3–4 days.

In vitro vascular tubular formation assay

The tubular formation assay was performed as described by Reyeset al. [6]. Briefly, cells were induced for endothelial differentiation for 7days as described above. After 7 days of induction, cells were tryp-sinized, mixed with serum-free medium containing 10 ng/mL VEGF-B(R&D 751-VE), and replated on Matrigel (ECM-gel Sigma E1270) in96-well plates. Cultures were maintained at 37 �C 5% CO2, where thevascular tube-like structure was seen after 6 h.

RNA isolation and RT-PCR analysis

Total RNA was isolated from cultured cells using single stepmethod with Trizol according to the manufacturer�s instructions. Theintegrity and purity of total RNA was verified spectrophotometricallyand by gel-electrophoresis on 0.8% SeaKem agarose (BMA, Hellerup,Denmark). For reverse transcriptase (RT), first strand complementaryDNA was synthesized from 5 lg of total RNA using a commercialrevertAid H minus first strand cDNA synthesis kit (Fermentas, Hel-singborg, Sweden) according to manual�s instructions. Polymerase-chain reaction (PCR) was performed by using 2 lL of 1:10 dilutedcDNA on an automated T gradient thermocycler (Biometra, Gottin-gen, Germany) in a 20 lL reaction mixture that contained 20 pmole offorward and reverse primers (sequences are given in Table 1), 1· PCRbuffer, dNTP mixture 0.2 mM, and 0.2 U hot start Taq polymerase

Product size (bp)

TGCTA-30 F 800AGCA-30 RGAACTG-30 F 599TGAAACC-30 R

CACA-30 F 310TAGAC-30 RAACGAAAG-30 F 347GTCTGTA-30 RTCCCC-30 F 230ACTTGGTG-3 0 RATGTCATC-30 F 475GTCCTTA-3 0 RCCCAGAAAGC-3 0 F 307GCACTTTGG-3 0 RCTGTTCTACTG-3 0 F 234GACCAATAAG-3 0 RTGCACTTTC-3 0 F 261TGGGTGCAG-3 0 R

ACGAAGTC-3 0 F 107C CAT CCC AT-3 0 RATGGCACC-3 0 F 275CACTTTCTC-3 0 RGCCAAGTG-3 0 F 62CTCCTTCT-30 RGAGAACAT-3 0 F 70CCTGACCCTT-30 RTCATTGAC-3 0 F 53

CAGTCCT-30 R


(Amplicon, Bie & Berntsen A-S, Copenhagen, Denmark). After initialdenaturation at 95 �C for 15 min, 35 cycles of denaturation for 30 s at95 �C, annealing for 30 s at 60 �C, and extension for 1 min at 72 �Cwere carried out and the PCR was terminated by a final extension at72 �C for 7 min. PCR products were analyzed by 1.5% agarose gelelectrophoresis, visualized by ethidium-bromide staining, and photo-graphed with a Kodak gel documentation system. The specificity ofPCR products was confirmed by DNA sequencing and restrictionmapping of the amplified product.

Real-time PCR

Quantitative PCR was performed in an iCycler IQ detection system(Bio-Rad, Herlev, Denmark) by using SYBR Green I as a double-strand DNA-specific binding dye. Thermocycling was performed in afinal volume of 20 lL containing 3 lL cDNA sample (diluted 1:30),20 pmol of each primer, and 2 mM MgCl2; 0.2 mM dNTP mixture, 1·Taq reaction buffer, 0.5 U HotStart Taq DNA polymerase (Qiagen,VWR, Denmark), 0.5lL of 1:3000 dilution of SYBR Green I (RochMolecular Biochemicals, Denmark), and 10 nM Fluorescein Calibra-tion Dye (Bio-Rad); (for collecting well factor directly from theexperimental plate). The quantification of target gene and b-actinmRNA was performed in separate tubes using the primers in Table 1.The cycle conditions for iCycler were started by using experimentalplate as a source of collecting well factors according to the operatinginstructions of Bio-Rad, followed by denaturing step at 95 �C for10 min and 40 cycles of 95 �C for 30 s, 60 �C for 30 s, and 72 �C for1 min. Each reaction was run in triplicate and the fluorescence data werespecified for collection at the end of the extension step in every cycle.

To ensure specific amplification, a melting curve was done for eachPCR by increasing the temperature from 60 to 95 �C with temperatureincrement rate of 0.5 �C/10 s. For verifying the melting curve results,all samples were run in 1.5% agarose gel electrophoresis and visualizedwith ethidium bromide staining. Fold induction and expression levelsfor each target gene were calculated using the comparative Ct method½1=ð2DCt Þ� formula, where DCt is the difference between Ct target andCt-reference after normalization to b-actin mRNA (Perkin–Elmer�sUser Bulletin No. 2). Data were analyzed using the optical systemsoftware version 3.0 (Bio-Rad) and Microsoft Excel 2000 to generaterelative expression values.

Immunocytochemistry

Immunostaining on coverslips with culture expanded hMSC-TERT20 (PDL 437) was performed using DAKO En Vision+. In brief,coverslips were incubated for 1 h at RT with primary antibodies di-luted in ChemMate Antibody diluent�s (S2022, DAKO) againstSTRO-1 (1:25), CD71 (1:25), and CD105 (1:200). Coverslips weresubsequently washed three times in Tris buffered saline (TBS; 0.05 M,pH 7.4), then incubated for 30 min with secondary anti-mouse Ig/HRP-conjugated polymers (K4001; En Vision+, DAKO) and visual-ized with 3,3 0-diaminobenzidine tetrahydrochloride (DAB, S3000;DAKO) according to manufacturer�s instruction.

Immunostaining for surface markers CD10, CD29, and CD44 wasperformed on paraffin-embedded sections of NBF (neutral bufferedformalin) fixed hMSC-TERT20 (PDL 437) as described above for cellson coverslips.

ForCollagen type II staining, sectionswere pre-digested with 300 U/mL hyaluronidase in 50 mM Tris (pH 8.0), 30 mM sodium acetatecontaining 0.5 mg/mL bovine serum albumin (BSA) (Sigma) for 15 minat 37 �C, and incubated with the monoclonal antibodies II-II6B3 to ColII (Obtained from Department of Biological Sciences, University ofIowa, Iowa City, IA) for 1 h at 37 �C. Immunostaining was detectedcalorimetrically using the streptavidin-peroxidase Histostain-SP kit forDAB (Zymed Laboratories, San Francisco, CA). Control groups wereperformed without primary antibodies under identical conditions.

Cytochemical staining

Staining for alkaline phosphatase. Cells induced to osteoblast dif-ferentiation for one week as described were washed in PBS twice, fixedin methanol/formalin (9:1) for 1 min, and incubated with alkalinephosphatase (ALP) substrate solution (5 mg naphthol AS-TR phos-phate) in 25 mL water plus 10 mg Fast red TR, in 24 mL of 0.1 M Trisbuffer (pH 9.5) for 1 h at room temperature. Mayer�s hematoxylin wasused as a counterstain. Cells were photographed using Olympus in-verted system microscope I 50· (Olympus, Albertslund, Denmark)equipped with an Olympus digital camera C3040 zoom (magnification200·).

Alizarin red S staining for in vitro formed mineralized matrix. Thecells were induced to form mineralized matrix as described above, afterone week, cells were rinsed with phosphate-buffered saline (PBS), air-dried, and fixed in ice-cold 95% ethanol for 30 min at �20 �C. Sub-sequently, the cells and the matrix were stained with 40 mM Alizarinred-S (AR-S), pH 4.2, for 1 h at RT, washed extensively five times withdeionized water, and rinsed with PBS (without Mg2+ or Ca2+) for15 min.

Alcian blue staining for chondrocytes. Sections of paraffin-embed-ded micropellets were stained with alcian blue (Sigma) solution, pH2.5.

Oil red-O staining for adipocytes. Cells differentiated into adipo-cytes were washed twice in PBS and lipid staining was performed usingOil red-O stain. Cells were fixed in 10% formaldehyde for 1 h at 4 �C,rinsed once in 3% isopropanol, and stained for 1 h at 4 �C with filteredOil red-O staining solution (prepared by dissolving 0.5 g Oil red-Opowder in 60% isopropanol). Cells were counterstained with Mayer�shematoxylin for 5 min and then washed with water. Mature adipocytescontaining lipid droplets stained red.

In vivo transplantation of hMSC-TERT in immuno-deficient mice

Cells (5 · 105) mixed with hydroxyl-apatite/tricalcium phosphateceramic powder (HA/TCP, 40 mg:, Zimmer Scandinavia, Denmark)were transplanted subcutaneously into the dorsal surface of 8-week-old female NOD/SCID mice (NOD/LtSz-Prkdcscid) as describedpreviously [28]. The transplants were recovered eight weeks aftertransplantation, fixed in 70% ethanol, and dehydrated andembedded undecalcified in methyl methacrylate (MMA). Tissuesections (7.5 lm thick) were cut and stained with Goldner�s tri-chrome stain.

Results

Growth characteristics of three different hMSC-TERT

cell lines

We monitored the growth characteristics of the threehMSC-TERT cell lines (hMSC-TERT2, hMSC-TERT4, and hMSC-TERT20). As shown in Fig. 1,differences existed between the three cell lines. Theslowest growth was observed in hMSC-TERT2 with apopulation doubling time (PDT) of 16 days. On theother hand, hMSC-TERT4 and hMSC-TERT20 hada shorter PDT (Fig. 1) that decreased during long-termculture. hMSC-TERT4 exhibited an acceleration ingrowth rate at PDL 25 from PDT of 10.2 days toPDT of 5 days and later at PDL 139 to PDT of 1.9days (Fig. 1; Table 2). Similar phenomenon was ob-served in hMSC-TERT20 which had an initial PDT

Fig. 1. Long-term growth curves of three different independentlygrowing cell populations derived from parental hMSC-TERT cellsover a period of more than 3 years (A). (n) hMSC-TERT2 passagedwith a split ratio of 1:2; ( ) hMSC-TERT4 passaged with a split ratioof 1:4; (s) hMSC-TERT20 passaged with a split ratio of 1:20. Cellswere grown in standard growth medium and cell number wascalculated every one week over the indicated period.

Table 2Growth characteristics of three independent hMSC-TERT sub-culturepopulations

Cell type PDL PDT Morphology

hMSC-TERT2 27 16 SquamoidhMSC-TERT4 25 10.2 Squamoid

61 5.02 Fusiform139 1.9 Polygonal

hMSC-TERT20 47 4.1 Squamoid109 1.4 Polygonal430 1.2 Cuboidal

PDL, cumulative population doubling level; PDT, population dou-bling time in days.


of 4.1 days and a decreased in PDT to 1.4 days at PDL109. The PDT remained constant afterwards for morethan 350 population doublings (Fig. 1; Table 2). Thus,acceleration of growth rate was inversely proportionalto seeding density.

Changes in cell morphology correlated with the growth

rate

Similar to changes in cell growth rates, morphologi-cal differences were observed between the three differentcell lines. hMSC-TERT cells in the initial parental cellpopulation were large and flat-shaped cells (squamoid)were similar to normal hMSC (Fig. 2A (a)). This mor-phology was maintained in hMSC-TERT2, hMSC-

TERT4, and hMSC-TERT20 at early PDL and lowgrowth rate (PDT 16, 10.2, and 4.1, respectively) (Fig.3A (b, c, f)). However, with time in culture the cellsstarted to acquire a distinct morphology associated withthe changing growth rate within each cell line (Table 2).The initial squamoid morphology was kept in hMSC-TERT2 cells (Fig. 2A (b)). hMSC-TERT4 cells changedtheir morphology from squamoid to rather fusiform andin late passages to polygonal cells (Fig. 3A (c–e)). Final-ly, hMSC-TERT20 changed its morphology fromsquamoid to polygonal and in late passage to smallcuboidal cells (Fig. 2A (f–h)). Consistent with micro-scopic evaluation of cell morphology, the analysis of cellsize and granularity by FACS revealed cells at high PDLwith large size and high granularity for hMSC-TERT2and 4 compared to hMSC-TERT20 cells which weresmall in size and less granular (Fig. 2B).

hTERT expression and telomerase activity

Since hMSC-TERT2, hMSC-TERT4, andhMSC-TERT20 exhibited differences in their growthrates in long-term culture, we examined whether levelsof telomerase activity varied among them. Expressionlevel of the ectopic hTERT mRNA was similar in allthree cell lines at different PDL as determined by real-time RT-PCR (data not shown). On the protein level,as shown by TRAP assay, telomerase activity wasequally detected in the three cell lines at differentPDL (Fig. 3A) and their cellular levels of telomerasewere similar as quantified by real-time TRAP method(Fig. 3B).

Cell surface characteristics of hMSC-TERT cells

Using a combination of FACS analysis and immuno-cytochemistry, we characterized the surface markersknown to be associated with hMSC. Using FACS analy-sis, no significant quantitative or qualitative differenceswere observed between the cells during long-termculture. Both hMSC-TERT4 (PDL 193) and hMSC-TERT20 (PDL 360) were positive for: CD10 (75–99%),CD13, CD33 (60–80%), HLA-DR, CD49c, CD49f,CD29, and negative for CD3, CD14, CD19, CD20,CD34, CD45, CD54, CD56, and CD66 (Table 3). Usingimmunohistochemistry, hMSC-TERT20 (PDL 437)exhibited positive staining for CD29, CD10, CD44,CD71, CD105, and STRO-1 (Fig. 4) and did not stainfor markers of hematopoietic lineages including CD14,CD34, CD45 or CD117 (c-kit) (not shown).

hMSC-TERT maintained in vitro differentiation capacity

We have performed extensive studies of hMSC-TERTcells for differentiation into the three mesoderm-type lin-eages (osteoblasts, adipocytes, and chondrocytes).

Fig. 2. Cell morphology of hMSC and long-term passaged hMSC-TERT subpopulations. (A) (a) Primary hMSC, (PDL 10). (b) hMSC-TERT2 cells(PDL 27). (c) hMSC-TERT4 (PDL 25). (d) hMSC-TERT4 (PDL 106). (e) hMSC-TERT4 (PDL 214). (f) hMSC-TERT20 (PDL 60). (g) hMSC-TERT20 (PDL 155). hMSC-TERT20 (PDL 417). Cells were photographed after 1 week of cultures. Bar, 100 lm. (B) FACS analysis for size andgranularity of primary hMSC and three different hMSC-TERT at high PDL corresponding to that shown in figures a, b, e, and h. Cells were culturedin standard growth medium to 70% confluence and treated as described in Materials and methods for FACS analysis, cell size and granularity shownby forward light and side light scattering for 50,000 events.

Fig. 3. Telomerase activity in three established hMSC-TERT cell lines: (a) hMSC-TERT2 (PDL 64); (b) -TERT4 (PDL 74); (c) -TERT4 (PDL 189);(d) -TERT20 (PDL 115); and (e) -TERT20 (PDL 475) as determined by using (A) conventional TRAP assay and (B) real-time quantitative TRAPassay (RTQ-TRAP). Cells were cultured in standard growth medium to 80% confluence and lysed in Chaps buffer. TRAP and RTQ-TRAP methodswere performed for 500 cells as described in Materials and methods. Quantitative telomerase activity is expressed as a percentage of the activity in thepositive control (HT-1080). Values are expressed as means ± SD of three different measurements.


Table 3FACS analysis of hMSC-TERT4 and hMSC-TERT20

Antigen hMSC-TERT4 p62(PDL 193)

hMSC-TERT20 p64(PDL 360)

CD3 � �CD10 ++ +++CD13 +++ +++CD14 � �CD19 � �CD20 � �CD33 +++ +++CD34 � �CD45 � �CD54/ICAM � �CD56 � �CD66 � �HLA-DR + +++CD49c/(integrin a3) +++ +++CD49f/(integrin a6b4) +++ +++CD29/(integrin b1) +++ +++

Surface marker characterization of two independent sub-populationsof hMSC-TERT at high PDL. Cells were cultured in 10% FCS toconfluence, harvested, labelled with specific antibodies and analyzed byFACS. Percentage of positive cells is represented as mean of threemeasurements.(�, negative; +, 10–25%; ++, 25–75%; +++, 75–100%).


Osteoblast differentiation

All the three hMSC-TERT cell lines (hMSC-TERT2,hMSC-TERT4, and hMSC-TERT20) maintained duringtheir long-term culture the ability to respond to osteo-blast-inducing medium. As shown in Fig. 5A, osteo-blast-inducing medium increased ALP production andled to the formation of mineralized matrix visualized

Fig. 4. Immunocytochemistry characterization of hMSC-TERT. Immunostaembedded sections of formaldehyde-fixed cell pellets for CD10, CD29, andconditions for both CD71, CD105, and STRO-1 as described in Materialsantibody were performed by omitting the primary antibody for both stainin

by positive AR-S staining. Also, as shown by RT-PCRthe cells were able to upregulate the expression of earlyand late osteoblastic marker genes (Runx2/CBFA1,ALP, and osteocalcin). Similar results were obtained inhMSC-TERT4 and hMSC-TERT2 cultures (not shown).

Chondrocyte differentiation

Both hMSC-TERT4 and hMSC-TERT20 weretested for their ability to form chondrocytes in micro-mass cultures. As shown in Fig. 6A, hMSC-TERT20was able to form chondrocytes evidenced by positivestaining for alcian blue and type II collagen. Also,the induced hMSC-TERT4 and hMSC-TERT20 wereable to upregulate chondrocyte-specific genes (collagentype II and collagen type X) as demonstrated byRT-PCR (Fig. 6B).

Adipocyte differentiation

hMSC-TERT2, hMSC-TERT4, and hMSC-TERT20 can form adipocytes after induction byBRL, Dex, and IBMX. As shown in Figs. 7A and B,more than 70% of the hMSC-TERT20 cells formedmature adipocytes as assessed by Oil red-O stainingof accumulated lipid droplets and quantified by FACSanalysis based on Nile red staining. Also, the cells wereable to upregulate gene expression of both early(PPARc2) and late (aP2 and adiponectin) adipocyte-specific genes upon adipoctye induction (Fig. 7C). Sim-ilar results were obtained by hMSC-TERT2 andhMSC-TERT4 (not shown).

ining for hMSC-TERT20 (PDL 437) was performed either on paraffin-CD44, or on coverslips of expanded culture under standard growthand methods using DAKO En Vision+ system. Negative controls forg procedures. Bar, 50 lm.

Fig. 5. Osteoblast differentiation of hMSC-TERT20. Cells were induced in vitro to form osteoblasts by culturing them in osteogenic medium (MEMcontaining 10% FCS, ascorbic acid, b-glycerophosphate, and vitamin D3) for 7 days. (A) Alizarin red staining was used for matrix mineralizationwhile ALP staining was used for alkaline phosphatase. (B) Expression analysis of osteogenic markers by RT-PCR.

Fig. 6. Chondrocyte differentiation of hMSC-TERT. Cells were cultured as a pelleted micromass for 3 weeks in serum-free DMEM supplementedwith dexamethasone, ascorbic acid, LL-proline, and ITS-plus as described in Materials and methods. (A) Histochemical for Alcian blue andimmunohistochemical staining for Col II. (B) RT-PCR expression analysis of Col type II and X after 3 week induction in chondrogenic medium.


Differentiation into other cell lineages

The ability of hMSC-TERT to differentiate into othercell lineages was also examined. Most of these experi-ments were performed using hMSC-TERT20 since it isthe most extensively grown populations.

Endothelial cell differentiation

hMSC-TERT20 were able to differentiate into endo-thelial-like as demonstrated by RT-PCR. hMSC-TERT20 cells increased the mRNA expression levels of

VEGF A, EPAS-1 (HIF-2a, hypoxia-inducible tran-scriptional factor), and ETB (endothelin receptor typeB) after 3 and 7 days of induction (Fig. 8C). Quantifica-tion by real-time PCR (Fig. 8D) showed an increase ofVEGF A, ETB, and EPSA-1 transcripts of 7.3-, 8-, and4.4-folds, respectively, after 7 days of induction. Theexpression levels of VEGF receptors Flt-1 and KDRwere detectable but did not change during the differenti-ation process (data not shown). Immunocytochemicalstaining (Fig. 8A) showed that the induced hMSC-TERT20 cells stained positive for endothelial markervon Wilbrand Factor (vWF). Also, the characteristic

Fig. 7. Adipocyte differentiation of hMSC-TERT20. Cells were grown to become confluent and induced to form adipocytes by culturing in MEMcontaining 10% FCS, dexamethasone, IBMX, insulin, and BRL (Rosiglitazone). After 10 days of induction, the adipocyte formation was analyzedby: (A) Oil red-O staining. (B) Quantification of lipid accumulation with FACScan Flow cytometry based on Nile Red staining. (C) Expressionanalysis of some adipogenic markers by RT-PCR.

Fig. 8. Endothelial differentiation of hMSC-TERT. (A) Immunocytochemical staining of vWF. hMSC-TERT20 (PDL 450) cells were mixed withMatrigel, cultured for 14 days in standard growth medium without (a) or with 50 ng/mL VEGF-B (b), then parafin-embedded and stained for vWF.Bar, 50 lm. (B) Tubular formation by hMSC-TERT20. Cells were induced to endothelial differentiation for 7 days in EDM as described in M&M.The vascular tube architecture was photographed after 4 h of replating the differentiated cells on Matrigel in the presence of 10 ng/mL VEGF-B. Bar,100 lm. (C) Expression analysis of endothelial markers in hMSC-TERT20 after 3 and 6 days of induction in EDM as analyzed by RT-PCR and (D)quantified by real-time PCR. Expression of target genes was normalized to b-actin and represented as a fold induction over control non-treated cells.Data are shown as means of at least two independent experiments.


vascular tube formation was observed when the cellswere cultured on the matrigel with VEGF A (Fig. 8B).

In vivo transplantation of hMSC-TERT cells

The MSC phenotype of hMSC-TERT was also stud-ied using in vivo transplantation subcutaneously in im-mune-deficient mice. hMSC-TERT2 as well as earlypassage hMSC-TERT4 and hMSC-TERT20 formed ec-topic bone, marrow stroma supporting hematopoiesis

and adipocytes (Figs. 9A and B). However, at PDL256, hMSC-TERT20 formed tumors composed mostlyof mesoderm type cells [29].

Discussion

In the present study, we have demonstrated that telo-merized hMSC lines maintain long-term self-renewaland differentiation capacity. Also, using ‘‘gold stan-

Fig. 9. In vivo transplantation of hMSC-TERT. hMSC-TERT2 (PDL 60) cells mixed with HA/TCP and implanted subcutaneously in immuno-deficient mice for 8 weeks as described in Material and methods. Histological sections of two different plastic-embedded (A) hMSC-TERT2(PDL = 47) and (B) early TERT20 (PDL = 148) implants with HA/TC (H) (stained by Goldener�s Trichrome) demonstrating the regenerated bone(B), osteoid (O), marrow stroma supporting hematopoiesis (M), and adipocytes (A) (magnification 100·).


dard’’ assay for measuring stem cell characteristics ofMSC, hMSC-TERT cell lines were capable of formingbone, bone-marrow supporting stroma, and adipocyteswhen transplanted subcutaneously in immune-deficientmice. These findings suggest a new biological role for tel-omerase in maintenance of the biological integrity ofadult stem cells.

We observed differences in growth rates between thethree telomerized cell lines studied (hMSC-TERT2,hMSC-TERT4, and hMSC-TERT20) that were createdbased on differences in plating densities during long-term in vitro cultures, These differences are not mediatedby changes in hTERT transduction among the cell linessince the three cell lines exhibited similar levels of ecto-pic hTERT mRNA as well as telomerase activity. Simi-lar phenomenon has been observed and extensivelystudied by Prockop and co-workers. By initially platinghMSC at low densities they were able to obtain threemorphologically distinct cell types ranging from smallround cells with the highest growth rate to large and flatmature cells with decreased growth rate [30,31]. It is pos-sible that culturing MSC at lower densities imposesselective pressure that favors the growth of cells withgrowth advantage. We have recently reported that theincreased growth rate of hMSC-TERT4 at PDL 95and hMSC-TERT20 at PDL 256 was associated withdeletion of the Ink4a/ARF locus [29]. Since Ink4a/ARF

locus encodes two proteins (INK4a and ARF) that con-tribute to senescence-associated growth arrest [32,33], itis plausible that the Ink4a/ARF deletion contributed tothe increased growth rate in the presence of telomeraseactivity but the proof of this hypothesis remains to bedetermined in our model.

The cell surface marker profiles of hMSC-TERT4and hMSC-TERT20 were very similar to that reportedfor normal hMSC [34–36]. It is interesting thathMSC-TERT cells maintained expression of STRO-1antigen which is considered a marker for the hMSCin bone marrow [5]. Similar to our findings, Shi et

al. [37] reported the maintenance of high levels ofSTRO-1 expression in hTERT-transduced hMSC.The association between high level of telomerase activ-ity and the maintenance of STRO-1 immunoreactivityis difficult to explain since the epitope recognized bythe STRO-1 antibody is not known. Our results maysuggest that the maintenance of STRO-1 immunoreac-tivity is secondary to the maintenance of stem cellcharacteristics conferred by the persistence of telome-rase activity. This hypothesis is supported by the pres-ervation of other surface markers typical for hMSC inhMSC-TERT, e.g., CD105/endoglin, CD29/b1 inte-grin, and CD44.

The effect of ectopic expression of hTERT on abol-ishing the senescence-associated growth arrest andextending the proliferative life-span has been demon-strated in several somatic cell types including fibroblasts,osteoblasts, endothelial cells, epithelial cells, and kerati-nocytes (for review, [38]). However, in addition to theseeffects on cell proliferation, there is an increasing recog-nition that telomerase activity may also contribute tothe biological functions of the cells. In the present studywe have found that hMSC-TERT cell lines respondedadequately to in vitro differentiation signals in contrastto the impaired responsiveness observed in senescentcells [39]. We [1] and others [37] have also reported en-hanced bone formation capacity of telomerized hMSCafter in vivo transplantation. Similar to our findings,the maintenance of the differentiation functions of cellshas also been observed in other cell types; hTERT-over-expression in endothelial cells led to an improved neo-vascularization after in vivo implantation compared tocontrol telomerase-negative cells [40,41]. Also, telomeri-zed human fetal hepatocytes exhibited an extended lifespan and maintenance of their liver-specific characteris-tics in vitro and in vivo [42]. The molecular mechanismsunderlying the maintenance (or even improvement) offunctional activity of the cells due to increased telome-rase activity is under investigation. For example in


hMSC, ectopic hTERT expression was associated withincreased gene expression of osteoblast-specific genes(e.g., Runx2/CBFA1 and osterix) [43]. It has also beenreported that telomerase activity contributes to a largernumber of cellular maintenance processes in addition totelomere maintenance. These functions include anti-ap-optotic effects, genomic stability, and DNA repair, andhave been termed telomerase ‘‘extracurricular activities’’and may positively contribute to the improved func-tional activity of the cells [44]).

Stem cells are defined as cells with both self-renewalcapacity and an ability to differentiate into various celltypes [45]. Our results suggest a new role for telome-rase activity in maintaining both functions in adultmesenchymal stem cells and provide a general ap-proach for long-term maintenance of adult stem cells.This notion is supported by the biological findings re-ported in telomerase-deficient mouse (mTR�/�) of im-paired hematopoietic cell proliferation, impairedwound healing, and hepatic regeneration after hepaticinjury [46]. Moreover, hMSC cultured from telome-rase-deficient mice (mTR�/�) exhibited impaired invitro chondrocyte and adipocyte differentiation com-pared with wild type littermates [47].

Our studies suggest that telomerization of hMSC is apotentially useful approach to obtain large number ofcells for mechanistic studies of self-renewal and differen-tiation of stem cells. For example, based on using thistelomerized hMSC model, we have recently identifieddlk1/Pref-1 (delta like1, transmembrane protein belongsto EGF-like protein family) as a new negative regulatorof osteoblast differentiation of hMSC [26]. Also, thesetelomerized cells may be useful for clinical protocols ofcell transplantation and tissue engineering [48]. Whileour data show that in late passage cells, some pheno-typic drift may occur, it is important to note that thisis not an inevitable outcome of hTERT-overexpression.For example, hMSC-TERT2 has been growing in thelaboratory for more than three years and still maintainsfull differentiation potential and normal genetic and epi-genetic profiles [29]. In this regard, isolation of hMSCwith naturally regulated telomerase activity may providea safer cell type for tissue engineering applications[7,8,34]. However, the possibility of obtaining a largenumber of these cells with maintained genomic stabilityduring long-term culture still remains to be identified.

Acknowledgments

We thank Tina K.L. Nielsen and Lone Christiansenfor excellent technical assistance, Ditte C. Andersen forCD markers staining, and Graham Leslie for help withFACS analysis. The helpful discussions and interactionswith the members of the Wuerzburg Osteology Researchgroup (R. Ebert andN. Schuetze) are gratefully acknowl-

edged. This work was supported by grants from DanishMedical Research Council, Danish Center for Stem CellResearch, Karen Elise Jensen�s foundation and NovoNordisk foundation, Dr. BMA is a recipient of a fellow-ship from Alfred Benzon�s foundation in Denmark.

References

[1] J.L. Simonsen, C. Rosada, N. Serakinci, J. Justesen, K. Stende-rup, S.I. Rattan, T.G. Jensen, M. Kassem, Nat. Biotechnol. 20(2002) 592–596.

[2] P. Bianco, P.G. Robey, Nature 414 (2001) 118–121.[3] D.J. Prockop, Science 276 (1997) 71–74.[4] M. Kassem, L. Mosekilde, E.F. Eriksen, J. Bone Miner. Res. 8

(1993) 1453–1458.[5] S. Gronthos, S.E. Graves, S. Ohta, P.J. Simmons, Blood 84

(1994) 4164–4173.[6] M. Reyes, T. Lund, T. Lenvik, D. Aguiar, L. Koodie, C.M.

Verfaillie, Blood 98 (2001) 2615–2625.[7] G. D�ippolito, S. Diabira, G.A. Howard, P. Menei, B.A. Roos,

P.C. Schiller, J. Cell Sci. 117 (2004) 2971–2981.[8] Y. Jiang, B.N. Jahagirdar, R.L. Reinhardt, R.E. Schwartz, C.D.

Keene, X.R. Ortiz-Gonzalez, M. Reyes, T. Lenvik, T. Lund, M.Blackstad, J. Du, S. Aldrich, A. Lisberg, W.C. Low, D.A.Largaespada, C.M. Verfaillie, Nature (2002).

[9] S.I.S. Rattan, in: S.C. Kaul, R. Wadhwa (Eds.), Aging of Cells inand outside the Body, Kuwer Academic Publishers, London,2003, pp. 1–8.

[10] K. Stenderup, J. Justesen, C. Clausen, M. Kassem, Bone 33(2003) 919–926.

[11] C.M. DiGirolamo, D. Stokes, D. Colter, D.G. Phinney, R. Class,D.J. Prockop, Br. J. Haematol. 107 (1999) 275–281.

[12] J.W. Shay, Y. Zou, E. Hiyama, W.E. Wright, Hum. Mol. Genet.10 (2001) 677–685.

[13] D. Parsch, J. Fellenberg, T.H. Brummendorf, A.M. Eschlbeck,W. Richter, J. Mol. Med. 82 (2004) 49–55.

[14] S. Zimmermann, M. Voss, S. Kaiser, U. Kapp, C.F. Waller,U.M. Martens, Leukemia 17 (2003) 1146–1149.

[15] M. Kveiborg, M. Kassem, B. Langdahl, E.F. Eriksen, B.F. Clark,S.I. Rattan, Mech. Ageing Dev. 106 (1999) 261–271.

[16] A.G. Bodnar, M. Ouellette, M. Frolkis, S.E. Holt, C.P. Chiu,G.B. Morin, C.B. Harley, J.W. Shay, S. Lichtsteiner, W.E.Wright, Science 279 (1998) 349–352.

[17] R.E. Nisato, J.A. Harrison, R. Buser, L. Orci, C. Rinsch, R.Montesano, P. Dupraz, M.S. Pepper, Am. J. Pathol. 165 (2004)11–24.

[18] S. Piera-Velazquez, S.A. Jimenez, D. Stokes, Arthritis Rheum. 46(2002) 683–693.

[19] L. Liu, C.M. DiGirolamo, P.A. Navarro, M.A. Blasco, D.L.Keefe, Exp. Cell Res. 294 (2004) 1–8.

[20] N.W. Kim, M.A. Piatyszek, K.R. Prowse, C.B. Harley, M.D.West, P.L. Ho, G.M. Coviello, W.E. Wright, S.L. Weinrich, J.W.Shay, Science 266 (1994) 2011–2015.

[21] I. Szatmari, J. Aradi, Nucleic Acids Res. 29 (2001) E3.[22] H. Wege, M.S. Chui, H.T. Le, J.M. Tran, M.A. Zern, Nucleic

Acids Res. 31 (2003) E3.[23] S. Rasheed, W.A. Nelson-Rees, E.M. Toth, P. Arnstein, M.B.

Gardner, Cancer 33 (1974) 1027–1033.[24] B.M. Abdallah, C.H. Jensen, G. Gutierrez, R.G. Leslie,

T.G. Jensen, M. Kassem, J. Bone Miner. Res. 19 (2004)841–852.

[25] J.U. Yoo, T.S. Barthel, K. Nishimura, L. Solchaga, A.I. Caplan,V.M. Goldberg, B. Johnstone, J. Bone Joint Surg. Am. 80 (1998)1745–1757.


[26] A.M. Mackay, S.C. Beck, J.M. Murphy, F.P. Barry, C.O.Chichester, M.F. Pittenger, Tissue Eng. 4 (1998) 415–428.

[27] M. Reyes, A. Dudek, B. Jahagirdar, L. Koodie, P.H. Marker,C.M. Verfaillie, J. Clin. Invest. 109 (2002) 337–346.

[28] K. Stenderup, C. Rosada, J. Justesen, T. Al Soubky, F. Dagnaes-Hansen, M. Kassem, Biogerontology 5 (2004) 107–118.

[29] N. Serakinci, P. Guldberg, J.S. Burns, B. Abdallah, H. Schrodder,T. Jensen, M. Kassem, Oncogene 23 (2004) 5095–5098.

[30] D.C. Colter, R. Class, C.M. DiGirolamo, D.J. Prockop, Proc.Natl. Acad. Sci. USA 97 (2000) 3213–3218.

[31] D.C. Colter, I. Sekiya, D.J. Prockop, Proc. Natl. Acad. Sci. USA98 (2001) 7841–7845.

[32] M. Serrano, H. Lee, L. Chin, C. Cordon-Cardo, D. Beach, R.A.DePinho, Cell 85 (1996) 27–37.

[33] C.J. Collins, J.M. Sedivy, Aging Cell 2 (2003) 145–150.[34] S. Gronthos, A.C. Zannettino, S.J. Hay, S. Shi, S.E.

Graves, A. Kortesidis, P.J. Simmons, J. Cell Sci. 116 (2003)1827–1835.

[35] M.K. Majumdar, M.A. Thiede, J.D. Mosca, M. Moorman, S.L.Gerson, J. Cell Physiol. 176 (1998) 57–66.

[36] M.F. Pittenger, A.M. Mackay, S.C. Beck, R.K. Jaiswal, R.Douglas, J.D. Mosca, M.A. Moorman, D.W. Simonetti, S. Craig,D.R. Marshak, Science 284 (1999) 143–147.

[37] S. Shi, S. Gronthos, S. Chen, A. Reddi, C.M. Counter, P.G.Robey, C.Y. Wang, Nat. Biotechnol. 20 (2002) 587–591.

[38] M.R. Wada, P. Polly, A. Nagpal, S.C. Kaul, in: S.I.S. Rattan(Ed.), Modulating Aging and Longevity, Kluwer AcademicPublisher, London, 2003, pp. 85–99.

[39] M. Kveiborg, S.I. Rattan, B.F. Clark, E.F. Eriksen, M. Kassem,J. Cell Physiol. 186 (2001) 298–306.

[40] S. Murasawa, J. Llevadot, M. Silver, J.M. Isner, D.W. Losordo,T. Asahara, Circulation 106 (2002) 1133–1139.

[41] J. Yang, U. Nagavarapu, K. Relloma, M.D. Sjaastad, W.C. Moss,A. Passaniti, G.S. Herron, Nat. Biotechnol. 19 (2001) 219–224.

[42] H. Wege, H.T. Le, M.S. Chui, L. Liu, J. Wu, R. Giri, H. Malhi,B.S. Sappal, V. Kumaran, S. Gupta, M.A. Zern, Gastroenterol-ogy 124 (2003) 432–444.

[43] S. Gronthos, S.Q. Chen, C.Y. Wang, P.G. Robey, S.T. Shi, J.Bone Miner. Res. 18 (2003) 716–722.

[44] S. Chang, R.A. DePinho, Proc. Natl. Acad. Sci. USA 99 (2002)12520–12522.

[45] F.M. Watt, B.L. Hogan, Science 287 (2000) 1427–1430.[46] K.L. Rudolph, S. Chang, H.W. Lee, M. Blasco, G.J. Gottlieb, C.

Greider, R.A. DePinho, Cell 96 (1999) 701–712.[47] L. Liu, C.M. DiGirolamo, P.A. Navarro, M.A. Blasco, D.L.

Keefe, Exp. Cell Res. 294 (2004) 3116–3121.[48] J.W. Shay, W.E. Wright, Nat. Biotechnol. 18 (2000) 22–23.

Erratum to “Maintenance of differentiation potential of human bone marrow mesenchymal stem cells...

Documents

Transcript of Erratum to “Maintenance of differentiation potential of human bone marrow mesenchymal stem cells...