Gene expression analysis of major lineage-defining factors in human bone marrow cells: Effect of...

Gene Expression Analysis of Major Lineage-Defining Factors in HumanBone Marrow Cells: Effect of Aging, Gender, and Age-Related Disorders

Ying Jiang,1,2,3 Hajime Mishima,4 Shinsuke Sakai,4 Yin-kun Liu,2 Yoshimi Ohyabu,1 Toshimasa Uemura1,5

1Nanotechnology Research Institute (NRI), National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Higashi, Tsukuba Ibaraki,305-8566, Japan, 2Fudan University, Shanghai, People’s Rebublic of China, 3Medical School of Nantong University, Jiangsu, People’s Republic ofChina, 4Department of Orthopedic Surgery, Control of Musculoskeletal Systems, Advanced Biomedical Applications, Graduate School ofComprehensive Human Sciences, University of Tsukuba, 2-1-1 Tennodai, Tsukuba Ibaraki, 305-8572, Japan, 5Tokyo Medical and Dental University,1-5-45 Yushima Bunkyo-ku, Tokyo, 113-8519, Japan

Received 12 June 2007; accepted 30 November 2007

Published online 26 February 2008 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jor.20623

ABSTRACT: Adult bone marrow cells (BMCs) include two populations:;mesenchymal stem cells (MSCs), which can differentiate into bone,cartilage, and fat; and hematopoietic stem cells (HSCs), which produce all mature blood lineage. To study the effect of aging, gender, and age-related disorders on lineage differentiation, we performed quantitative RT-PCR to examine mRNA expression of the major factors definingBMC lineage, cbfa1 for osteoblasts, ppar-gamma for adipocytes, sox9 for chondrocytes, and rankl for osteoclasts, in bone marrow from80 healthy subjects and patients (14–79 years old) with two age-related disorders: osteoarthritis (OA) and rheumatoid arthritis (RA). Twoapoptosis-related genes, bcl-2 and drak1, were studied. RANKL and PPAR-Gamma levels exhibited a clear positive correlation with age infemale patients, but not in males, with a slight age-related decline in CBFa1 transcripts. DRAK1 expression showed an age-associatedascending trend with significantly greater transcripts of RANKL and DRAK1 in females ( p< 0.01). Compared with age-matchedcontrols, RApatients exhibited increased RANKL, PPAR-Gamma, and DRAK1 mRNA levels (p<0.05), and OA showed the higher RANKL and PPAR-Gamma transcripts (p< 0.05). Furthermore, SOX9 and DRAK1 expressions in the RA group were higher than in the OA group (p<0.05). Ourdata indicate that aging and age-related disorders affect gene expressions differently, suggesting that in aging, the lineage of bone marrowcells was modified with prominent changes in decreased bone marrow osteoblastogenesis, increased adipogenesis and osteoclastogenesis,while in age-related disorders, marrow adipogenesis and the activity or number of osteoclasts may play an important role in the pathogenesisof arthritic bone loss. � 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 26:910–917, 2008

Keywords: bone narrow; aging; lineage; transcription factor; apoptosis

Human BMCs contain two populations of bone marrowstem cells: mesenchymal stem cells (MSCs) and hema-topoietic stem cells (HSCs). The differentiation of bonemarrow multipotent stem cells toward a specific lineage isdependent on hormonal and local factors activating specifictranscription factors.1 Attempts have recently been madeto identify the mechanisms responsible for the lineage-specific differentiation of human BMCs. Cell–cell signal-ing, ECM (the extracellular matrix), and cytokines mightplay important roles in triggering a particular pathway;2,3

however, aging might be a major factor to regulate thelineage of BMCs. There is evidence to suggest that withage, the number of adipocytes in bone marrow increases,resulting in the appearance of fatty marrow.4 Aging is alsoassociated with a sustained loss of bone, due to reducedosteoblastic bone formation, improved osteoclastic boneresorption, and an increased volume of marrow adiposetissue.5,6 Also, a patient’s disease should be taken intoconsideration. For example, osteoporotic patients developmore adipocytes in their bones, which exhibit an inferiorosteoblast differentiating microenvironment, and therewas a significant reduction in in vitro chondrogenic andadipogenic activity in cultures of OA patient-derived MSCscompared with normal cultures, which may contribute toincreased bone density and loss of cartilage.7

To elucidate the effect of factors, such as age, disease,and gender in the ability of BMCs to differentiation,quantitative RT-PCRs were performed to detect the

mRNA expression of phenotype-specific transcriptionfactors that determine the lineage commitment ofBMCs. The first, ppar-gamma, a member of the ligand-activated nuclear receptor superfamily, is regarded as aadipogenic-specific transcription factor,8 inducing thetranscriptional activation of many different targetgenes involved in lipid metabolism.9 The second, Cbfa1(core-binding factor alpha 1), is an osteoblast-specifictranscription factor that is essential for the differ-entiation of mesenchymal stem cells into mature osteo-blasts and for maintaining the differentiated function inthese cells during bone formation and remodeling.10

The third factor, Sox9, was the first transcription factorto be identified as essential for commitment of mesen-chymal cells to the chondrogenic lineage and to chon-drocyte differentiation.11 Last, RANKL, also calledosteoclast differentiation factor (ODF), was an absoluterequirement for osteoclast development and bone mod-eling both in vivo and in vitro, and also functions as asurvival factor for osteoclast precursors. At the sametime, we also examined the expression of two apoptosis-related gene, drak1 and Bcl-2, to make clear the effectof different factors on apoptosis of BMCs. Drak1 firstcloned by Sanjo et al. in 1998 from human placentacDNA.12 Kojima et al.13 reported that Drak1 was ex-pressed strongly in osteoclasts, but weakly or not at allin osteoblasts, suggesting that Drak1 is closely involvedin the regulation of osteoclastogenesis and osteoclastapoptosis. Bcl-2 is an oncogenic protein that acts byinhibiting programmed cell death.14,15 Enhanced ex-pression of Bcl-2 family members has been implicated inthe pathogenesis of RA.16

910 JOURNAL OF ORTHOPAEDIC RESEARCH JULY 2008

Correspondence to: Toshimasa Uemura (T: 81-29-861-2724; F: 81-29-861-2789; E-mail: [email protected])

� 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.

In this paper, we report that changes in the differ-entiation potential of BMCs are associated with advanc-ing age, gender, and pathological agents. Moreover, thesignificantly high expression level of DRAK1 in bonemarrow of patients with OA and RA suggest thatapoptosis is involved in the pathogenesis of these twodiseases.

MATERIALS AND METHODSPatientsA total of 80 patients, ranging in age from 14 to 79 years, wereenrolled in the study, including 27 men (aged 14–70, mean 48)and 53 women (aged 15–79, mean 53). We obtained fullconsent from all patients in accordance with the local ethicscommittees at both the National Institute of AdvancedIndustrial Science and Technology (AIST) and the Universityof Tsukuba. The primary diagnosis was traumatic femoralneck fracture in 29 of the cases, osteoarthritis (OA) in 37, andrheumatoid arthritis (RA) in 14. The distributions of diagnosesbetween the male and female subjects are as follows: trauma:11 females and 18 males; OA: 30 females and 7 males; RA: 12females and 2 males. Except in the analysis of disease-relateddifferences, RA patients were excluded from age-dependentand gender-dependent analyses because RA may affect ske-letal metabolism. Patients who met the entry criteria werestratified into younger group (<55 years old) and older group(�55 years old) according to the previous reports.17–19 Theresults of age dependence of pelvic fracture,17 human bone-derived cells,18 and cemented total knee arthroplasty forgonarthrosis18 showed reasonable evidences for dividing sub-jects into two groups under or over 55 years.

Human Bone MarrowBone marrow aspirates (5 mL) were obtained from theacetabulum during hip arthroplasty for OA and RA or duringhip replacement for traumatic femoral neck fracture withsigned informed consent. The bone marrow samples from thepatients with traumatic femoral neck fracture had no patho-logical changes, and were regarded as normal control. Allsamples were immediately mixed with TRIzol reagent and thenstored frozen at �708C prior to use.

Quantitative RT-PCRTotal RNA was isolated from human bone marrow cells usingTRIzol reagent. For reverse transcription, the reactionmixture contained 1 mg of RNA, 0.125 mM oligo(dT) primer,and 0.25 units of Avian Myeloblastosis Virus reverse tran-scriptase (TaKaRa) in a total volume of 20 mL. The reactionwas run for 10 min at 308C, and 30 min at 508C, and stopped byheating for 2 min at 958C. Aliquots (2 mL) of the products wereamplified in a reaction mixture (20 mL) containing Light-CycleTM-FastStart DNA Master SYBR Green I 2 mL, 0.5 mMof each primer, and 1.6 mM MgCl2 using a LightCyclerTM

(Roche Molecular Biochemicals, Indianapolis, IN). After pre-incubation at 958C for 10 min, a PCR was performed with40 cycles of denaturation at 958C for 10 s, annealing at 588C for10 s, and elongation at 728C for 16 s. A single fluorescencereading was taken in each cycle following the elongationstep. The primers used were as follows: ppar-gamma (humanppar-gamma mRNA, L40904.2), forward sequence: 50- TGCCT-TGCAGTGGGGATGT-30 and reverse sequence: 50- ATCGCC-CTCGCCTTTGCTT-30; DRAK1 (human DRAK1 mRNA,AB011420.1), forward sequence: 50-GCCAAGATTGTCGG-ATG-30 and reverse sequence: 50-GTCTGCAACACACTGAT-30;

sox9 (human sox9 mRNA, Z46629.1), forward sequence:50-CAAGTTCCCCGTGTGC-30 and reverse sequence: 50-GCAAGTGCGGGTACTG-30; adipsin(human adipsin mRNA,GenBank accession number M84526.1), forward sequence: 50-GGCAACCGCAAGAAGC-30 and reverse sequence: 50-CCAATGATCCTCCCACC-30, and for human BCL-2, CBFa1,RANKL and the internal standard glyceraldehyde-3-phos-phate dehydrogenase (GAPDH), the LightCyclerTM Primer Setfrom Roche Molecular Biochemicals. The relative mRNA levelsfor each condition were determined by performing quantitativeRT-PCR three times for each of two independent batches.

Statistical AnalysisAverage amounts of the transcripts of PPAR-gamma, RANKL,CBFa1, SOX9, DRAK1, BCL-2, and Adipsin were calculated,expressed as the arithmetic mean�SE, and plotted on graphs.Mann-Whitney and Kruskal-Wallis H nonparametric testswere used for statistical evaluation. The statistical signifi-cance of differences among the experimental groups wasestablished at the p< 0.05 (*) and p< 0.01 (**) levels.

RESULTSThe potential of bone marrow stem cells to differentiateinto osteoblasts, adipocytes, chondrocytes (from MSCs),or osteoclasts (from HSCs) was examined by measuringchanges in the gene expression of major lineage-definingfactors. We analyzed by real-time quantitative RT-PCRthe expression of the osteoblast-specific gene CBFa1,adipocyte-specific gene PPAR-gamma, chondrocyte-specific gene SOX9, osteoclast-specific gene RANKL, andapoptosis-related genes bcl-2 and DRAK1, using humanbone marrow samples from patients with OA, RA, andtraumatic femoral neck fracture, aged from 14 to 79. Thepositive expression rate of PPAR-gamma, RANKL, SOX9,CBFa1, DRAK1, and BCL-2 was 95, 75, 94, 89, 85, and95%, respectively, in a total of 80 patients. The relativeintensity of the expression of these genes was normaliz-ed to that of the housekeeping gene GAPDH. Figure 1illustrates the expression of mRNA transcripts from80 patients for both males and females, distributed byage. Figure 2 and Figure 3 illustrate age- and gender-related changes in mRNA expression for all subjects.Figure 4 presents the difference in expression among thegroups with OA, RA, and normal bone marrow. Based on anonsignificant interaction of age and gender, age anddisease, and disease and gender ( p> 0.05), the datasummarized in different categories are independent andshow no interruption by any other factors.

Quantitative RT-PCR for PPAR-gamma ExpressionExpression of PPAR-gamma mRNA in the group ofsubjects <55 years (N¼32) ranged from 0.0386 to 1.424,with a mean value of 0.396 (Fig. 1). In contrast, theexpression level was almost fourfold higher in the oldergroup �55 years (N¼30), varying from 0.055 to 13.146,with a mean value of 1.494 (Fig. 2). The difference inexpression of PPAR-gamma with age was statisticallysignificant ( p¼0.019). The intensity of the mRNAexpression of PPAR-gamma tended to be higher infemales (N¼ 39) than males (N¼23), although the

GENE EXPRESSION ANALYSIS OF MAJOR LINEAGE-DEFINING FACTORS IN HUMAN BONE MARROW CELLS 911

JOURNAL OF ORTHOPAEDIC RESEARCH JULY 2008

difference was not significant (Fig. 3). This is becausesome of the females exhibited very high levels ofexpression, whereas others had levels as low as the males.Bone marrow cells of RA patients (N¼14) had higherPPAR-gamma mRNA levels than those of normalcounterparts (N¼26), p¼ 0.018. No significant differ-ences were found between OA and normal marrow(Fig. 4). To exclude the possibility that the increase in

the expression of PPAR-gamma originated from bonemarrow fat cells, we also detected the mRNA of adipsin,which is a marker gene for mature adipocytes. No age-related correlation for adipsin was found, but PPAR-gamma exhibited a clear positive correlation with age(Fig. 5). It is suggested that the high PPAR-gammaexpression level comes from mesenchymal stem cells(MSCs) not from fat cells.

Figure 1. Effects of age, gender, and disease onlineage-specific gene expression in human bone marrowfrom 80 subjects. Relative mRNA levels were determinedby quantitative real-time PCR assays using the standardcurve method. Data were normalized relative to GAPDHas an internal control.

Figure 2. Age-related differences in mRNA expression of humanbone marrow cells. Relative mRNA levels were determined byquantitative real-time PCR assays. Data were normalized relativeto GAPDH as an internal control and expressed as the mean�SE,significantly different between the young group (<55 years) and oldgroup (�55 years) at the *p< 0.05 level.

Figure 3. Gender-related differences in mRNA expression inhuman bone marrow cells. Relative mRNA levels were determinedby quantitative real-time PCR assays. Data were normalizedrelative to GAPDH as an internal control and expressed as themean�SE, significantly different between females and males atthe *p<0.05 level.

912 JIANG ET AL.


Quantitative RT-PCR for RANKL ExpressionAn increase in the amount of RANKL mRNA with agehas been observed in human bone marrow cells.Compared with the expression of RANKL in 24 youngpatients (mean 0.0799, range from 0.00236 to 0.664), theexpression in senescent patients was much higher(N¼25, mean 0.3308, range from 0.00216 to 1.516),p< .05 (Fig. 2). There was significantly more transcriptof RANKL in females than in males, 0.286� 0.074versus 0.060� 0.021, p<0.05 (Fig. 3). RANKL expres-sion was shown to be significantly increased in OA(mean 0.272� 0.073) and RA (mean 0.225� 0.084), com-pared to normal bone marrows (mean 0.0875� 0.0384,p< 0.05) (Fig. 4).

Quantitative RT-PCR for CBFa1 ExpressionAn age-dependent increase in CBFa1 mRNA expressionwas detected in 71 bone marrow samples from a total of80 patients (Fig. 2). The expression level in the youngergroup <55 years varied from 0.170 to 1.296, with a mean

value of 0.571, and was higher than that in the oldergroup (N¼29, mean 0.464, range from 0.086 to 1.735),p<0.05, although the difference was not significant.No significant correlation was found between CBFa1expression and patients’ age, gender (Figs. 2 and 3).

Quantitative RT-PCR for DRAK1 ExpressionDRAK1 mRNA expression was twice greater in femalepatients (N¼36, mean 0.193, range from 0.013 to 0.598)than in males (N¼20, mean 0.107, range from 0.011 to0.384), p<0.05 (Fig. 3). Patients with RA (mean0.377�0.121) and OA (mean 0.192� 0.023) had higherDRAK1 mRNA levels than those with normal bonemarrow (mean 0.116� 0.029), p< 0.01 (Fig. 4). No age-related correlation was found for DRAK1 mRNA ex-pression (Fig. 2).

Quantitative RT-PCR for SOX9 ExpressionA significant difference in SOX9 mRNA expression wasfound between RA and OA patients (Fig. 4). The ex-pression level in the RA group was 2.01� 0.544, muchhigher than that in the OA group (1.035� 0.323, p<0.05).There were no significant correlation between SOX9 ex-pression and age or gender (Figs. 2 and 3).

Quantitative RT-PCR for BCL-2 ExpressionThere was more than twice the level of BCL-2 expres-sion in RA (1.225�0.585) than in OA (0.573�0.106)and normal BMCs (0.515�0.116), although the differ-ence was not significant (Fig. 4). No significant cor-relation was found between BCL-2 mRNA expressionand age or gender (Figs. 2 and 3).

DISCUSSIONCbfa1 is an essential transcription factor for theosteogenic/chondrogenic and odontogenic lineages.20–26

As a means of bone regeneration, a technique for thetransplantation of cultured bone27–37 has been devel-oped to overcome the disadvantages of autologous bonegrafts and simple implantation of biomaterial into bonedefects. In cultured bone transplantation therapy, cbfa1is not only a marker of osteoblast differentiation, butalso an activator of osteogenesis. We have recentlyreported the strong effect of Cbfa1 overexpression inosteoblast differentiation in vitro and in vivo,37 confirm-ing that the Cbfa1 expression is an appropriate markerof the osteogenesis of bone marrow cells. Yoshikawaet al.38 reported the age dependence of human bonemarrow cells on osteogenesis using a cultured bonetransplantation model, suggesting that the bone-regen-erating ability of human bone marrow cells may notdepend on age: they showed no gender dependence.Their results are almost consistent with our data of aslight decline of CBFa1 expression with age withoutgender dependence. Despite our common-sense expect-ations about age-related osteoporosis having genderdependence, there exists neither significant age depend-ence nor gender dependence in the osteogenic potentialof bone marrow cells. Osteoprogenitor cells might haveno siginificant effect on osteogenesis by themselves.

Figure 4. Quantitative RT-PCR of mRNA expression in OA, RA,and normal bone marrow cells. Data were normalized relative toGAPDH as an internal control and expressed as the mean�SE,significantly different among three group at the *p<0.05 and**p<0.01 level.

Figure 5. Change in expression of PPAR-gamma and Adipsinwith advancing age. PPAR-gamma clearly exhibited a positivecorrelation with age, while no age-related correlation was detectedfor adipsin.



Ppar-gamma, a member of the ligand-activatednuclear receptor superfamily, is regarded as a adipo-genic-specific transcription factor,8 inducing the tran-scriptional activation of many different target genesinvolved in lipid metabolism.9 The PPAR-gamma ex-pression in the bone marrow has marked age, gender,and disease dependence, despite a small or negligibledependence on age, gender, and desease in CBFa1 ex-pression as shown in Figures 2, 3, and 4. The bonemarrow cells from older, female, and RA patientsexpressed higher PPAR-gamma mRNA. In particular,aging is known to cause a decrease in the number of bone-forming osteoblasts and an increase in the number ofmarrow adipocytes.39 Our results can consistently ex-plain the formation of adipocytes; however, then seemcontradictory to osteoblast formation. That can be re-asonably explained as follows. Lecka-Czernik et al.40 andAli et al.41 reported the effect of ppar-gamma expressionon adipocyte versus osteoblast differentiation. Theyshowed evidence that the activation of ppar-gamma2stimulates adipogenesis and inhibits osteoblastogenesisin mesenchymal progenitor cells. Thus, the upregulationof PPAR-gamma in bone marrow is thought to decreasethe number of osteoblasts for bone formation.

DRAK1 [death-associated protein (DAP) kinase-related apoptosis-inducing protein kinase 1] was provento play an important role in osteoclast apoptosis byour group.13 DRAK1 is expressed not in osteoblasts butin osteoclasts in the bone, and our recent study ofbisphosphonate treatment on osteoclasts suggests thatDRAK1 expression might trigger osteoclast apoptosis.42

In bone marrow cells, the DRAK1 expression mayoriginate from hematopoietic stem cells, which will dif-ferentiate into osteoclasts. Despite a negligible agedependence of DRAK1 expression, the upregulation infemale bone marrow cells is observed, and may be due tothe induction of osteoclasts by mechanisms including thegender dependence of RANKL expressions describedbelow. A high expression of DRAK1 in the bone marrowcells of RA patients can be explained by the appearance ofosteoclasts in synovioids of RA patients.

RANKL(ODF, TRANCE, OPGL) has been identifiedas essential for the differentiation and maintenance ofosteoclasts. RANKL is highly expressed on osteoblast/stromal cells, which are essentially involved in osteoclastformation from hemopoietic progenitor cells through amechanism involving cell-to-cell contact.43 As recentlyreviewed,44,45 a number of osteoclastogenic agents actby inducing the surface expression of RANKL. Theinduced expression of RANKL by osteoblast/stromal cellsenhanced osteoclastogenesis, leading to an increasednumber of osteoclast precursors in the bone marrowand to increased number of osteoclasts on trabecularbone.44–47 Deletion of the RANKL gene in mice resultedin severe osteopetrosis and a complete lack of osteoclastsas a result of an inability of osteoblasts to supportosteoclastogenesis.48 Although osteoprotegerin (OPG), adecoy receptor for RANKL, is also involved in osteoclas-togenesis by inhibiting osteoclast differentiation,49 the

increase in RANKL mRNA expression that results in anincreased RANKL/OPG ratio was much more prominentfor the induction of osteoclastogenesis. The evidenceindicated that RANKL mRNA was constitutively ex-pressed in osteoblasts and upregulated in the treatmentof osteotropic factors, while OPG expression in osteo-blasts was unchanged or slightly increased in thepresence or absence of bone resorbing factors.50,51 whichillustrated that the RANKL/RANK pathway is animportant mediator of increased osteoclastogenesis. Sojust as transcription factors Cbfa1 and PPAR-gamma arealready known to be key regulators for osteoblastic andadipogenic differentiation, RANKL is now regarded as acritical regulators for osteoclastic formation. Our resultsindicate that: (1) the expression level of RANKL mRNAin bone marrow cells increases with age; (2) comparedwith male subjects, female subjects have much more ofthe RANKL transcript in bone marrow cells; (3) inOA and RA, the level of RANKL expression in bonemarrow cells is much higher than normal. The enhance-ment of RANKL expression with age suggests a way toexplain the age-related increase in the capacity ofosteoblast/stromal cells to support osteoclastogenesis.52

In line with our results, Cao et al.53 suggested that thechanges in RANKL expression associated with agingmay contribute, at least in part, to the gradual loss ofbone that occurs with age and the development of age-related osteoporosis. However, some studies found noage-dependent change in the RANKL serum level inhealthy individuals54 or decreased in the serum RANKLconcentration with age.55 One explanation for thisdiscrepancy may be that serum RANKL levels do notconsistently reflect the local microenvironment in thebone. Besides bone, a variety of other extraskeletaltissues such as brain, heart, kidney, skeletal muscle, andskin can express RANKL mRNA, which highlights theimportance of assessing RANKL levels directly in thebone’s microenvironment.

Although some research demonstrated that a defi-ciency of estrogen may induce an upregulation of RANKLexpression in bone marrow cells,56,57 we have no directevidence to support such a conclusion. Our results showthat the RANKL expression level is lower in malepatients than in female counterparts. A possible explan-ation for this may be that androgen has an effect onRANKL expression.

Focal or diffuse bone loss represents a major unsolvedproblem in RA. The activation of osteoclastogenesismediated by the enhanced expression of RANKL playsa critical role in the bone destruction process in RA. In allexperimental results of arthritis studied thus far,inhibition of RANKL completely prevented bone lossand partially protected cartilage. As for RA patients,high serum levels of soluble RANKL58 and increasedexpression of RANKL mRNA in synovial tissues59 wereobserved, suggesting that RANKL is the principalmediator of bone loss in human RA. As far as we know,our study is the first to clarify the high level of RANKLmRNA in human bone marrow cells from RA patients.

914 JIANG ET AL.


Another intriguing aspect of our findings is that not onlyin RA, but also in OA, a condition which, in large joints, isnot usually associated with periarticular bone resorp-tion, enhanced expression of the RANKL transcript wasdetected, underlining that higher levels of RANKLmRNA seem to be linked to the bone remodeling thatoccurrs in this disease.

Sox9 is a transcription factor with a high-mobilitygroup (HMG-box) DNA-binding domain exhibiting ahigh degree of homology with that of the mammaliantestis-determining factor, SRY.60 It is a master regulatorof cartilage formation and may cooperate with SOX5 andSOX6 to regulate chondrogenesis.16 During chondro-genesis, SOX9 is expressed in all chondroprogenitors andall differentiated chondrocytes, and is regarded as achondrogenic-specific transcription factor.61 Until now,there has been no report about age-related and gender-related changes in the chondrogenesis of MSCs in bonemarrow, although O’Driscoll et al.62 confirmed an age-dependent decline in the chondrogenic potential ofundifferentiated MSCs in the periosteum. Our studyshowed no significant correlation between Sox9 expres-sion and aging, suggesting that the chondrogenicpotential of MSCs from human bone marrow tissues isnot affected by age. The nonsignificant improvement ofSOX9 expression in male patients may result from amale-specific activation of SOX9 by unknown factors,because SOX9 is a relative of the mammalian Y-linkedsex-determiningfactor, SRY. Osteoarthritis and rheu-matoid arthritis are characterized by a loss of articularcartilage, leading to an irreversible impairment of jointmotion. Unlike bone remodeling, in which nonresidentcells of hematopoietic and mesenchymal origin arerecruited to the bone surface to remodel the bone matrix,the cartilage remodeling process is conducted entirely bya single cell type, the chondrocyte. Adult bone marrowcontains a reservoir of mesenchymal stem cells (MSC)with the potential to become chondrocytes in vitro andin vivo. So determining the chondrogenic potential ofMSCs is helpful in clarifying the pathogenesis ofcartilage destruction in RA and OA. The new findingof our study is that the level of expression of SOX9mRNA in bone marrow cells was significantly higher inRA than in OA, implying that the initiation of cartilageloss differs between RA and OA, thus providing a clue forclinical intervention to promote cartilage remodeling.

Bcl-2 is the first identified suppressors of programmedcell death.63 Overexpression of Bcl-2 in hematopoieticstem cells (HSC) has been shown to provide protectionagainst a wide variety of apoptosis-inducing challengesin vitro64 and in vivo.45 BCL-2–transgenic mice show anincrease in both the number and repopulation potentialof HSC in bone marrow.65 Thus, apoptosis appears to bean important mechanism regulating the homeostasisof HSC. To define the apoptotic death in human bonemarrow cells, BCL-2 mRNA expression was detected byquantitative RT-PCR. Our results showed that humanbone marrow cells express BCL-2, but no significantcorrelation was found between BCL-2 mRNA expression

and the patient age or gender. In RA, BCL-2 expression istwice the level found in OA or normal bone marrow,suggesting that a defective apoptosis in bone marrowcells is involved in the pathogenesis of RA. Consistentwith our results, Perlman et al.66 reported that BCL-2 ismore highly expressed in synovial tissue in RA than inOA. Also, in an antigen-induced model of RA, BCL-2 waspresent at sites of early erosion and correlated withscores for erosion and inflammation,67 thereby support-ing an important role for BCL-2 in the developmentof RA.

In conclsion, our data indicate that aging and age-related disorders affect gene expressions differently,suggesting that, in aging, the lineage of bone marrowcells is modified, with prominent changes in decreasedbone marrow osteoblastogenesis, increased adipogene-sis, and osteoclastogenesis, whereas in age-related dis-orders, marrow adipogenesis and the activity or numberof osteoclast may play an important role in the patho-genesis of arthritic bone loss.

ACKNOWLEDGMENTSThis study is carried out as a part of ‘‘Ground-based ResearchAnnouncement for Space Utilization’’ promoted by JapanSpace Forum and is partly supported by JST (Japan Scienceand Technology Agency).

REFERENCES1. Tuan RS, Boland G, Tuli R. 2003. Adult mesenchymal stem

cells and cell-based tissue engineering. Arthritis Res Ther5:32–45.

2. Baksh D, Song L, Tuan RS. 2004. Adult mesenchymal stemcells: characterization, differentiation, and application in celland gene therapy. J Cell Mol Med 8:301–316.

3. Dennis JE, Charbord P. 2002. Origin and differentiation ofhuman and murine stroma. Stem Cells 20:205–214.

4. Ricci C, Cova M, Kang YS, et al. 1990. Normal age-relatedpatterns of cellular and fatty bone marrow distribution in theaxial skeleton: MR imaging study. Radiology 177:83–88.

5. Chan GK, Duque G. 2002. Age-related bone loss: old bone, newfacts. Gerontology 48:62–71.

6. Justesen J, Stenderup K, Ebbesen EN, et al. 2001. Adipocytetissue volume in bone marrow is increased with aging and inpatients with osteoporosis. Biogerontology 2:165–171.

7. Murphy JM, Dixon K, Beck S, et al. 2002. Reducedchondrogenic and adipogenic activity of mesenchymal stemcells from patients with advanced osteoarthritis. ArthritisRheum 46:704–713.

8. Tontonoz P, Hu E, Spiegelman BM. 1994. Stimulation ofadipogenesis in fibroblasts by PPAR gamma 2, a lipid-activated transcription factor. Cell 79:1147–1156.

9. Wahli W, Braissant O, Desvergne B. 1995. Peroxisomeproliferator activated receptors: transcriptional regulators ofadipogenesis, lipid metabolism and more. Chem Biol 2:261–266.

10. Otto F, Thornell AP, Crompton T, et al. 1997. Cbfa1, acandidate gene for cleidocranial dysplasia syndrome, isessential for osteoblast differentiation and bone development.Cell 89:765–771.

11. Bi W, Deng JM, Zhang Z, et al. 1999. Sox9 is required forcartilage formation. Nat Genet 22:85–89.

12. Sanjo H, Kawai T, Akira S. 1998. DRAKs, novel serine/threonine kinases related to death-associated protein kinasethat trigger apoptosis. J Biol Chem 273:29066–29071.



13. Kojima H, Nemoto A, Uemura T, et al. 2001. rDrak1, a novelkinase related to apoptosis, is strongly expressed in active osteo-clasts and induces apoptosis. J Biol Chem 276:19238–19243.

14. Tsujimoto Y, Croce CM. 1986. Analysis of the structure,transcripts, and protein products of bcl-2, the gene involved inhuman follicular lymphoma. Proc Natl Acad Sci USA 83:5214–5218.

15. Reed JC. 1997. Double identity for proteins of the Bcl-2 family.Nature 387:773–776.

16. Liu H, Pope RM. 2003. The role of apoptosis in rheumatoidarthritis. Curr Opin Pharmacol 317–322.

17. Henry SM, Pollak AN, Jones AL, et al. 2002. Pelvic fracture ingeriatric patients: a distinct clinical entity. J Trauma 53:15–20.

18. Katzburg S, Lieberherr M, Ornoy A, et al. 1999. Isolation andhormonal responsiveness of primary cultures of human bone-derived cells: gender and age differences. Bone 25:667–673.

19. Stern SH, Bowen MK, Insall JN, et al. 1990. Cemented totalknee arthroplasty for gonarthrosis in patients 55 years old oryounger. Clin Orthop Relat Res 260:124–129.

20. Komori T, Yagi H, Nomura S, et al. 1997. Targeted disruptionof Cbfa1 results in a complete lack of bone formation owing tomaturational arrest of osteoblasts. Cell 89:755–764.

21. Mundlos S, Otto F, Mundlos C, et al. 1997. Mutationsinvolving the transcription factor CBFA1 cause cleidocranialdysplasia. Cell 89:773–779.

22. Otto F, Thornell AP, Crompton T, et al. 1997. Cbfa1, acandidate gene for cleidocranial dysplasia syndrome, isessential for osteoblast differentiation and bone development.Cell 89:765–771.

23. Inada M, Yasui T, Nomura S, et al. 1999. Maturationaldisturbance of chondrocytes in Cbfa1-deficient mice. Dev Dyn214:279–290.

24. Kim IS, Otto F, Zabel B, et al. 1999. Regulation of chondrocytedifferentiation by Cbfa1. Mech Dev 80:159–170.

25. Banerjee C, McCabe LR, Choi JY, et al. 1997. Runt homologydomain proteins in osteoblast differentiation: AML3/CBFA1 isa major component of a bone-specific complex. J Cell Biochem66:1–8.

26. Ducy P, Zhang R, Geoffroy V, et al. 1997. Osf2/Cbfa1: atranscriptional activator of osteoblast differentiation. Cell 89:747–754.

27. Caplan AI. 1991. Mesenchymal stem cells. J Orthop Res9:641–650.

28. Caplan AI, Bruder SP. 1997. Cell and molecular engineeringin bone. Principles of tissue engineering. Georgtown, TX:Landes p 603–619.

29. Petite H, Viateau V, Bensaid W, et al. 2000. Tissue-engineeredbone regeneration. Nat. Biotech 18:959–963.

30. Ohgushi H, Caplan AI. 1999. AI.Stem cell technology andbioceramics: from cell to gene engineering. J Biomed MaterRes 48:913–927.

31. Tanaka J, Uemura T. 2000. Inorganic materials: poroushydroxyapatite for tissue engineering Tanpakushitsu Kaku-san Koso. 45:2150–2155.

32. Dong J, Kojima H, Uemura T, et al. 2001. In vivo evaluation ofa novel porous hydroxyapatite to sustain osteogenesis of trans-planted bone marrow-derived osteoblastic cells. J BiomedMater Res 57:208–216.

33. Dong J, Uemura T, Kikuchi M, et al. 2002. Long-termdurability of porous hydroxyapatite with low-pressure systemto support osteogenesis of mesenchymal stem cells. BiomedMater Eng 12:203–209.

34. Dong J, Uemura T, Shirasaki Y, et al. 2002. Promotion of boneformation using highly pure porous beta-TCP combined withbone marrow-derived osteoprogenitor cells. Biomaterials 23:4493–4502.

35. Uemura T, Dong J, Wang Y, et al. 2003. Transplantation ofcultured bone cells using combinations of scaffolds and culturetechniques. Biomaterials 24:2277–2286.

36. Wang YC, Uemura T, Dong J, et al. 2003. Application ofperfusion culture system improves in vitro and in vivo osteo-genesis of bone marrow-derived osteoblastic cells in porousceramic materials. Tissue Eng 9:1205–1214.

37. Kojima H, Uemura T. 2005. Strong and rapid induction ofosteoblast differentiation by Cbfa1/Til-1 overexpression forbone regeneration. J Biol Chem 280:2944–2953.

38. Yoshikawa T, Ohgushi H, Ichijima K, et al. 2004. Boneregeneration by grafting of cultured human bone. Tissue Eng10:688–698.

39. Moerman EJ, Teng K, Lipschitz DA, et al. 2004. Agingactivates adipogenic and suppresses osteogenic programs inmesenchymal marrow stroma/stem cells; the role of PPAR-gamma transcription factor and TGF-beta/BMP signalingpathways. Aging Cell 3:377–389.

40. Lecka-Gernik B, Moerman EJ, Grant DF, et al. 2002.Divergent effects of selective peroxisome proliferator-activated receptor-gamma2 ligands on adipocyte versusosteoblast differentiation. Endocrinology 143:2376–2384.

41. Ali AA, Weinstein RS, Stewart SA, et al. 2005. Rosiglitazonecauses bone loss in mice by suppressing osteoblast differ-entiation and bone formation. Endocrinology 146: 1226–1235.

42. Liu B, Kojima H, Nemoto A, et al. 2002. Expression of novelkinase rDRAK1 in osteoclasts and its role in osteooclastapoptosis. 42nd Annual Meeting of the American Society forCell Biology, San Francisco.

43. Yasuda H, Shima N, Nakagawa N, et al. 1998. Osteoclastdifferentiation factor is a ligand for osteoprotegerin/osteoclastinhibitory factor and is identical to TRANCE/RANKL. ProcNatl Acad Sci USA 95:3597–3602.

44. Hofbauer LC, Khosla S, Dunstan CR, et al. 2000. The roles ofosteoprotegerin and osteoprotegerin ligand in the paracrineregulation of bone resorption. J Bone Miner Res 15:2–12.

45. Theill LE, Boyle WJ, Penninger JM. 2002. RANK-L andRANK: T cells, bone loss, and mammalian evolution. AnnuRev Immunol 20:795–823.

46. Tsukii K, Shima N, Mochizuki S, et al. 1998. Osteoclastdifferentiation factor mediates an essential signal for boneresorption induced by 1 alpha, 25-dihydroxyvitamin D3,prostaglandin E2, or parathyroid hormone in the micro-environment of bone. Biochem Biophys Res Commun 246:337–341.

47. Lacey DL, Timms E, Tan HL, et al. 1998. Osteoprotegerinligand is a cytokine that regulates osteoclast differentiationand activation. Cell 932:165–176.

48. Kong YY, Yoshida H, Sarosi I, et al. 1999. OPGL is a keyregulator of osteoclastogenesis, lymphocyte development andlymph-node organogenesis. Nature 397:315–323.

49. Simonet WS, Lacey DL, Dunstan CR, et al. 1997. Osteopro-tegerin: a novel secreted protein involved in the regulation ofbone density. Cell 89:309–319.

50. Takeshita S, Arai S, Kudo A. 2001. Identification andcharacterization of mouse bone marrow stromal cell linesimmortalized by temperature-sensitive SV40 T antigen:supportive activity for osteoclast differentiation. Bone 29:236–241.

51. Kiviranta R, Morko J, Alatalo SL, et al. 2005. Impaired boneresorption in cathepsin K-deficient mice is partially compen-sated for by enhanced osteoclastogenesis and increasedexpression of other proteases via an increased RANKL/OPGratio. Bone 36:159–172.

52. Glowacki J. 1995. Influence of age on human marrow. CalcifTissue Int 56:S50–S51.

916 JIANG ET AL.


53. Cao J, Venton L, Sakata T, et al. 2003. Expression of RANKLand OPG correlates with age-related bone loss in male C57BL/6 mice. J Bone Miner Res 18:270–277.

54. Ziolkowska M, Kurowska M, Radzikowska A, et al. 2002. Highlevels of osteoprotegerin and soluble receptor activator ofnuclear factor kappa B ligand in serum of rheumatoidarthritis patients and their normalization after anti-tumornecrosis factor alpha treatment. Arthritis Rheum 46:1744–1753.

55. Liu JM, Zhao HY, Ning G, et al. 2005 Relationships betweenthe changes of serum levels of OPG and RANKL with age,menopause, bone biochemical markers and bone mineraldensity in Chinese women aged 20–75. Calcif Tissue Int 76:1–6.

56. Eghbali-Fatourechi G, Khosla S, Sanyal A, et al. 2003. Role ofRANK ligand in mediating increased bone resorption in earlypostmenopausal women. J Clin Invest 111:1221–1230.

57. Ikeda T, Utsuyama M, Hirokawa K. 2001. Expression profilesof receptor activator of nuclear factor kappaB ligand, receptoractivator of nuclear factor kappaB, and osteoprotegerinmessenger RNA in aged and ovariectomized rat bones. J BoneMiner Res 16:1416–1625.

58. Ziolkowska M, Kurowska M, Radzikowska A, et al. 2002. Highlevels of osteoprotegerin and soluble receptor activator ofnuclear factor kappa B ligand in serum of rheumatoidarthritis patients and their normalization after anti-tumornecrosis factor alpha treatment. Arthritis Rheum 46:1744–1753.

59. Takayanagi H, Iizuka H, Juji T, et al. 2000. Involvement ofreceptor activator of nuclear factor kappaB ligand/osteoclast

differentiation factor in osteoclastogenesis from synoviocytesin rheumatoid arthritis. Arthritis Rheum 43:259–269.

60. Ng LJ, Wheatley S, Muscat GE, et al. 1997. SOX9 binds DNA,activates transcription, and coexpresses with type II collagenduring chondrogenesis in the mouse. Dev Biol 183:108–121.

61. Zhao Q, Eberspaecher H, Lefebvre V, et al. 1997. Parallelexpression of Sox9 and Col2a1 in cells undergoing chondro-genesis. Dev Dyn 209:377–386.

62. O’Driscoll SW, Saris DB, Ito Y, et al. 2001. The chondrogenicpotential of periosteum decreases with age. J Orthop Res19:95–103.

63. Vaux DL, Cory S, Adams JM. 1988. Bcl-2 gene promoteshaemopoietic cell survival and cooperates with c-myc toimmortalize pre-B cells. Nature 335:440–442.

64. Fairbairn LJ, Cowling GJ, Reipert BM, et al. 1993. Suppres-sion of apoptosis allows differentiation and development of amultipotent hemopoietic cell line in the absence of addedgrowth factors. Cell 74:823–832.

65. Domen J, Cheshier SH, Weissman IL. 2000. The role of apop-tosis in the regulation of hematopoietic stem cells: overexpres-sion of Bcl-2 increases both their number and repopulationpotential. J Exp Med 191:253–264.

66. Perlman H, Georganas C, Pagliari LJ, et al. 2000. Bcl-2expression in synovial fibroblasts is essential for maintainingmitochondrial homeostasis and cell viability. J Immunol 164:5227–5235.

67. Perlman H, Liu H, Georganas C, et al. 2001. Differentialexpression pattern of the antiapoptotic proteins, Bcl-2 andFLIP, in experimental arthritis. Arthritis Rheum 44: 2899–2908.



Gene expression analysis of major lineage-defining factors in human bone marrow cells: Effect of...

Documents

Transcript of Gene expression analysis of major lineage-defining factors in human bone marrow cells: Effect of...