1161
Journal of Cell Science 113, 1161-1166 (2000) Printed in Great Britain © The Company of Biologists Limited 2000 JCS1174
Clonal mesenchymal progenitors from human bone marrow differentiate in vitro according to a hierarchical model Anita Muraglia1, Ranieri Cancedda1,2,3 and Rodolfo Quarto2,* 1Centro di Biotecnologie Avanzate, Genova, Italy 2Laboratorio di Differenziamento Cellulare, Istituto Nazionale per la Ricerca sul 3Dipartimento di Oncologia, Biologia e Genetica, Universita’ di Genova, Italy
Cancro, 16132 Genova, Italy
*Author for correspondence (e-mail:
[email protected])
Accepted 27 January; published on WWW 7 March 2000
SUMMARY Bone marrow stromal cells can give rise to several mesenchymal lineages. The existence of a common stem/progenitor cell, the mesenchymal stem cell, has been proposed, but which developmental stages follow this mesenchymal multipotent progenitor is not known. Based on experimental evidence, a model of mesenchymal stem cell differentiation has been proposed in which individual lineages branch directly from the same progenitor. We have verified this model by using clonal cultures of bone marrow derived stromal fibroblasts. We have analyzed the ability of 185 non-immortalized human bone marrow stromal cell clones to differentiate into the three main lineages: osteo-, chondro- and adipogenic. All clones but one differentiated into the osteogenic lineage. About one third of the clones differentiated into all three lineages analyzed. Most clones (60-80%) displayed an osteo-chondrogenic potential. We have never observed clones with a differentiation potential
limited to the osteo-adipo- or to the chondro-adipogenic phenotype, nor pure chondrogenic and adipogenic clones. How long the differentiation potential of a number of clones was maintained was assessed throughout their life span. Clones progressively lost their adipogenic and chondrogenic differentiation potential at increasing cell doublings. Our data suggest a possible model of predetermined bone marrow stromal cells differentiation where the tripotent cells can be considered as early mesenchymal progenitors that display a sequential loss of lineage potentials, generating osteochondrogenic progenitors which, in turn, give rise to osteogenic precursors.
INTRODUCTION
questions. (1) Do clones derived from a single BMSC display a multilineage differentiation potential? (2) Which combination of lineages is revealed and what frequency do they have? (3) Do multilineage potent cells maintain their own multilineage differentiation potential through a number of mitotic divisions? In our study, we used an in vitro analytical approach which, although not necessarily predictive of the ‘in vivo’ cell behavior, would provide information on the differentiation ability of human mesenchymal progenitors and on their intrinsic commitment at the clonal level. For this purpose, we isolated, expanded and characterized 185 non-immortalized clones from human bone marrow cultures. Their osteo-chondro-adipogenic potential was assessed by in vitro assays. All BMSC clones analyzed were able to differentiate in vitro into the osteogenic lineage and a high percentage displayed osteo-chondro-adipogenic potential. The phenotypic osteo-adipo- or chondro-adipoassociation was not found. The differentiation potential, assessed at increasing cell doublings in a number of clones, was progressively lost. The adipogenic potential is apparently the first to be repressed, whereas the osteogenic and the
During its entire lifespan, the skeleton is subjected to constant remodeling in order to keep bone tissue biomechanical properties unaltered (Parfitt, 1984). Remodeling and fracture healing are representative examples of the regeneration capacity of bone tissue, for which a compartment of mesenchymal stem and/or progenitor cells is required. Bone marrow is a possible repository for such cells (Beresford et al., 1992; Caplan, 1991; Friedenstein, 1990; Johnstone et al., 1998; Owen, 1988). Mesenchymal stem cells (MSC) are defined as pluripotent cells that self-maintain throughout the organism’s life and whose progeny give rise to skeletal tissues: cartilage, bone, tendon, ligament and marrow stroma. The current model for mesenchymal differentiation proposes that lineage progenitors are directly derived from the MSC (Caplan, 1991; Friedenstein, 1990; Owen, 1988). According to this model one would expect to find differentiation into a random combination of phenotypes upon appropriate stimulation of clonal bone marrow stromal cells (BMSC). We therefore investigated the differentiation ability of human BMSC clones in order to answer the following
Key words: Differentiation, Clone, Hierarchy, Bone marrow stromal cell, Mesenchymal stem cell
1162 A. Muraglia and others chondrogenic potentials proceed together branching only very late, with the loss of the chondrogenic differentiation ability. This study provides evidence that BMSC can also undergo multiple differentiation pathways at a clonal level, and that tripotential clones occur with high frequency in the bone marrow. Clones enter senescence and progressively lose their typical mesenchymal multi-lineage potential, giving rise to biand monopotential cells. The existence of fixed combinations of lineages (i.e. osteoadipo-chondro-, osteo-chondro- and osteogenic clones) and the sequential loss of lineage potential in tripotent clones suggest a deterministic model of BMSC differentiation.
MATERIALS AND METHODS Cell culture Marrow aspirates were obtained under general anesthesia from iliac crest of healthy donors during BMT harvest procedures as part of a protocol approved by the competent ethical authority. Informed consent was obtained from all individuals included in this study. Donor age ranged between 5 months and 30 years. Bone marrow samples were washed twice with PBS (phosphate-buffered saline, pH 7.2) and the mononucleated cells (MNC) were counted with a nuclear stain (0.1% Methyl Violet in 0.1 M citric acid). Clones were obtained by limiting dilution: briefly, 2×103 nucleated marrow cells were plated in each well of a 96-multiwell plate in Coon’s modified Ham’s F12 medium supplemented with 10% FCS (Mascia Brunelli, Milano, Italy). 4-5 plates were prepared for each marrow sample. From the initial seeding, clones were cultured in standard medium both in the presence (FGF+) and in the absence (FGF−) of 1 ng/ml human recombinant Fibroblast Growth Factor-2 (FGF-2) (Austral Biologicals, San Ramon, CA, USA). Cultures were examined daily for the appearance of stromal colonies. Wells containing more than one colony were not considered. All the clones that reached confluence were detached with 0.05% trypsin-0.01% EDTA (Sigma, St Louis, MO, USA) and each of them plated in four wells (replicas) of a 24-multiwell plate. Out of the four replicas of each clone, three were stimulated for differentiation experiments when cells reached confluence and one was frozen for future use.
Adipogenic differentiation Clones were stimulated for 3 weeks in medium supplemented with 1% FCS, 1.0×10−7 M dexamethasone and 1.0×10−9 M insulin. Lipid droplets were revealed by staining with Sudan Black IV. Maintenance of clones multipotentiality through mitotic divisions After the characterization at the stage of replicas, four tripontential (osteo-chondro-adipogenic) and four bipotential (osteochondrogenic) clones were selected for the determination of maintenance of differentiation ability through mitotic divisions. Each frozen replica was thawed and plated in a 6 cm Petri dish with standard medium with or without FGF-2, depending on the conditions in which the clone was obtained. When cells reached confluence, they were trypsinized, counted for the determination of cell doublings, and divided into four equal parts. Three parts were plated each in a well of a 24-well plate and stimulated for differentiation towards the three lineages as described above, and the fourth part was replated in a 6 cm Petri dish to continue expansion. Statistical analysis Statistical analyses were performed by simple linear regression on the non-transformed data, and by the Mann-Whitney U-test.
RESULTS BMSC cloning 256 BMSC clones (136 in FCS and 120 in FGF-2, respectively) were selected from about 2000 clones obtained from 1×107 marrow nucleated cells that were plated. No significant differences were observed in cloning efficiency and in phenotype distribution among donors of different age. Three different phenotypes were observed: (1) fibroblastic elongated cells, (2) large flattened cells and (3) thin starshaped cells (Fig. 1). Out of the original 256 clones, 80 FGF+
In vitro differentiation The potential of BMSC clones to differentiate into chondrogenic, osteogenic and adipogenic lineage was verified at different passages. Cells were enzymatically detached from culture dishes and plated at 5×104 cells/cm2 in 24-well culture plates. Each culture plate was kept in complete medium (with or without FGF-2) until confluence. Cultures were then stimulated with the appropriate differentiation medium according to the conditions described below. Stimulated and unstimulated multicolony BMSC cultures were used as controls. Chondrogenic differentiation Clones were stimulated for 1 week in standard medium supplemented with 2.0×10−4 M ascorbic acid, 1 ng/ml human recombinant TGFβ1 (Austral Biologicals, San Ramon, CA). Type II collagen expression was revealed by immunostaining with the monoclonal antibody CIICI (Developmental Studies Hybridoma Bank). Osteogenic differentiation Clones were stimulated for 2 weeks in standard medium supplemented with 2.0×10−4 M ascorbic acid, 7×10−3 M βglycerophosphate, 1.0×10−8 M dexamethasone. Osteocalcin expression was revealed by immunostaining with an antiserum against bovine osteocalcin kindly provided by Dr Simon Robbins.
Fig. 1. Phenotype of cultured human BMSC. Three cell phenotypes were observed in BMSC colonies: spindle shaped cells (A) were the most abundant, and large flattened cells (B) and starshaped cells (C) formed the other colonies. Bar, 40 µm.
Differentiation of human mesenchymal progenitors 1163 Table 1. Frequency and differentiation potential of clonal BMSC FGF− Phenotype O
C
A
+ + − + + − − −
+ + + − − + − −
+ − + + − − + −
18 84 0 0 3 0 0 0
17 80 0 0 3 0 0 0
27 48 0 0 4 0 0 1
34 60 0 0 5 0 0 1
105
100
80
100
Total
and 105 FGF− clones were able to reach confluence after passage in one well of a 24-well plate. All of them displayed the fibroblastic phenotype. At this stage, they were divided into four replicas. In vitro BMSC clone differentiation BMSC clones were grown in complete medium (+FGF-2 and –FGF-2) until they reached confluence. They were then stimulated to differentiate into the three lineages under the appropriate conditions. Fig. 2 displays the typical differentiation pattern of three different clones. All clones but one differentiated towards the osteogenic lineage, showing positivity for OC expression. About 95% of the clones also expressed type II collagen and a significant fraction (17% and 34%, –FGF-2 and +FGF-2, respectively) also assumed the adipogenic phenotype (Table 1). When comparing the clone replicas induced towards the different lineages, a hierarchy of differentiation patterns was evident. In fact, most of the clones (80% FGF− and 60% FGF+, respectively) were able to undergo both osteogenesis and chondrogenesis (OC clones). A very significant percentage (17% FGF− and 34% FGF+, respectively) were able to differentiate into all three induced lineages (OCA clones), and a small percentage (3% FGF− and 5% FGF+, respectively) were osteogenic only (O clones). Interestingly, associations of phenotypes such as chondro-adipogenic or osteo-adipogenic, and pure phenotypes such as chondrogenic or adipogenic, were never observed (Table 1).
% of total
Number of clones
% of total
The frequency and differentiation potential of clonal BMSC were assessed as indicated in Materials and Methods. All clones but one displayed the potential to differentiate towards the osteogenic lineage as revealed by osteocalcin expression. About 95% of the osteocalcin-positive clones became positive for type II collagen; 17% (FGF−) and 34% (FGF+) displayed the potential to originate all three phenotypes induced. O, Osteogenic phenotype; C, Chondrogenic phenotype; A, Adipogenic phenotype.
Interestingly, a higher relative frequency of bipotent clones was found in younger donors (0-5, 11 years old) as compared to older donors (22, 30 years old) (85.3±6.0% versus 46.5±12.5%; values are means ± s.d., P=0.014; Table 2). Conversely, older donors had a relatively higher frequency of tripotent clones (44.9±19.1% versus 14.7±6.0%; means ± s.d., P=0.027) (Table 2). BMSC clones life span and maintenance of differentiation potential The ability of clones to maintain their differentiation potential through mitotic divisions in culture was determined for four 25
Doubling number
Fig. 2. Differentiation of BMSC clones in vitro. BMSC clones were grown in complete medium until confluence. They were then maintained in differentiation medium for different time intervals as described in Materials and Methods. Primary BMSC cultures were used as controls for the specificity of the differentiation conditions. Osteogenesis was revealed by immunostaining with antibodies against osteocalcin; chondrogenesis was revealed by immunostaining with antibodies against type II collagen and adipogenesis by Sudan Black staining. A number of clones were able to differentiate into the three lineages (OCA clones), a significant percentage were able to express only osteocalcin and type II collagen (OC clones) and a minor percentage was pure osteogenic (O clones). Ctrl+, stimulated and Ctrl−, unstimulated multicolony BMSC cultures.
FGF+
Number of clones
20
15
OC OCA
10
5
0 0
20
40
60
80
100
120
Time in days Fig. 3. Life span and differentiation potential of BMSC tripotential clones. The life span of three OCA +FGF clones was assessed and the maintenance of the original differentiation potential was verified at increasing cell doublings. OCA clones reached 22-23 doublings in about 80 days of culture. By then, they had lost any apparent proliferation potential. The 3 clones maintained the original differentiation potential up to 19 doublings, but they lost the adipogenic potential by the 22nd doubling still maintaining the OC potential. Symbols indicate the actual time points of each passage. O, Osteogenic phenotype; C, Chondrogenic phenotype; A, Adipogenic phenotype.
1164 A. Muraglia and others Table 2. Donor age related frequency and differentiation potential of clonal BMSC FGF− OCA Donor age (years)
FGF+
OC
O
OCA
OC
O
Number
%
Number
%
Number
%
Number
%
Number
%
Number
%
3 6 1 3 5
15.8 14.6 5.8 60 21.7
16 35 16 2 15
84.2 85.4 94.1 40 65.2
0 0 0 0 3
0 0 0 0 13.1
1 8 − 11 7
14.3 22.8 − 61.1 36.8
6 27 − 7 8
85.7 77.2 − 38.9 42.1
0 0 − 0 4
0 0 − 0 21.1
0.5 3.5 11 22 30
The frequency of tri-, bi- and monopotential clones was related to the age of the relative donor. A statistically significant increase in the frequency of bipotential (OC) clones was found in donors of pediatric age. Monopotential clones were found only in the oldest (30 year old) donor analysed. O, Osteogenic phenotype; C, Chondrogenic phenotype; A, Adipogenic phenotype.
OCA and four OC clones. OCA clones were able to reach 2223 doublings after about 80 days in culture (Fig. 3). The three OCA (FGF+) clones studied had lost their adipogenic potential when tested at passage 3 (equivalent to 22 doublings), but they maintained their osteo-chondrogenic potential. The only OCA (FGF−) clone tested lost both the chondrogenic and the adipogenic potential after 22 doublings. OC clones had a comparable life span reaching 20-22 doublings. Interestingly enough, the two OC (FGF−) clones lost chondrogenic potential, but maintained their osteogenic potential. Out of the two OC (FGF+) clones tested, one lost chondrogenic potential, and the other one maintained both osteo- and chondro-genic potential (Table 3). DISCUSSION The bone medullary cavity of the postnatal organism contains hemopoietic lineage cells as well as a heterogeneous cell population making up the marrow stroma (Beresford, 1989; Friedenstein, 1990; Owen, 1988). This unique cellular population includes cells of the fibroblastic, reticular, adipogenic and osteogenic lineages (Beresford, 1989). Marrow stroma also provides the support for hemopoiesis (Dexter et Table 3. Doubling related differentiation potential of triand bi-potential clones Clone 1 (FGF+) 2 (FGF+) 3 (FGF+) 4 (FGF−) 5 (FGF+) 6 (FGF+) 7 (FGF−) 8 (FGF−)
Doubling number
O
C
A
Doubling number
O
C
A
19 19 19 18 18 19 18 18
+ + + + + + + +
+ + + + + + + +
+ + + + − − − −
22 23 23 23 19 22 19 20
+ + + + + + + +
+ + + − + − − −
− − − − − − − −
Clone life span ranged between 19 and 23 cell doublings. OCA clones reached 22-23 doublings, whereas OC clones had a slightly shorter life span reaching 20-22 doublings. The three OCA (FGF+) clones lost their adipogenic potential by 22 doublings, but maintained their osteochondrogenic potential. The only OCA (FGF−) clone tested lost both chondrogenic and adipogenic potentials by 22 doublings. The two OC (FGF−) clones lost chondrogenic, but maintained osteogenic potential. One of the two OC (FGF+) clones tested lost chondrogenic potential, whereas the other one maintained both osteo- and chondrogenic potential. O, Osteogenic phenotype; C, Chondrogenic phenotype; A, Adipogenic phenotype.
al., 1977; Tavassoli and Friedenstein, 1983). Furthermore, stromal fibroblasts are able to differentiate into cartilage, bone, fat, muscle and other connective tissues (Beresford et al., 1992; Caplan, 1991; Friedenstein, 1990; Johnstone et al., 1998; Krebsbach et al., 1997; Martin et al., 1997; Owen, 1988). Their counterparts in vitro are colony forming units-fibroblasts (CFU-f), which in turn give rise to bone marrow fibroblasts (BMF), bone marrow stromal cells (BMSC) and mesenchymal stem cells (MSC), possibly corresponding to a single cell population. These cells reproduce bone and cartilage when implanted in vivo (Friedenstein, 1990; Gerasimov et al., 1986; Goshima et al., 1991; Gundle et al., 1995; Krebsbach et al., 1997; Martin et al., 1997; Ohgushi et al., 1989). In other studies, a relationship was found between adipogenesis and osteogenesis (Bennett et al., 1991; Kodama et al., 1983; Park et al., 1999), and it was finally shown that these cells possess osteo-chondrogenic potential at a clonal level (Gerasimov et al., 1986; Kuznetsov et al., 1997), although they may differ in osteogenic capacity (Kuznetsov et al., 1997). Similar studies have been performed on clonal cell lines of immortalized human and mouse marrow fibroblasts (Dennis et al., 1999; Fried et al., 1993; Mathieu et al., 1992). Very recently, Dennis and coworkers have shown that a clone derived from marrow cells of a transgenic mouse carrying a gene for conditional immortality possesses a quadripotential differentiation ability since it supports osteoclastogenesis and is able to differentiate to the osteogenic, chondrogenic and adipogenic lineages (Dennis et al., 1999). Furthermore, Pittenger and colleagues have shown the multilineage differentiation potential of clones isolated from adult human bone marrow cultures (Pittenger et al., 1999). Based on these experimental findings, a model for the MSC differentiation was proposed where all the different mesenchymal lineages originate from the same stem cell under the influence of different microenvironments (Caplan, 1991; Friedenstein, 1990; Owen, 1988). According to this stochastic model, one would expect to find all the possible phenotypes or associations of phenotypes in cell culture systems. It should be also recalled that cells of mesenchymal origin display the ability of interconversion from one cell type to another at a stage later than that of the multipotential stem cell (Bennett et al., 1991; Beresford et al., 1992; Park et al., 1999). The clonal analysis that we performed on a number of primary BMSC cultures led us to isolate and characterize a large number of clones. 17% of the clones isolated and cultured in standard medium (i.e. without FGF-2) were able to differentiate into the three lineages analyzed. This frequency
Differentiation of human mesenchymal progenitors 1165 was increased (34%) when the clones were isolated and expanded in the presence of FGF-2. This evidence suggests that a high proportion of primary BMSC behave as multipotent progenitors and that the addition of FGF-2 to the culture medium increases their percentage, possibly either by preventing or by delaying BMSC from undergoing a defined differentiation pathway. Our data show important differences in the relative frequency of the three classes of clones. The relative frequency of tripotential clones in younger donors is lower than in older donor marrows, but the absolute frequency is not significantly different. If we correct these values on the basis of the CFU-f frequency observed in the normal population, which is about twice as high in younger than in older donors (Galotto et al., 1999), younger donors have a frequency of tripotential clones that is significantly higher than in older donors. This is a novel observation, but is in agreement with evidence that CFU-f frequency is inversely proportional to the age (Egrise et al., 1992; Quarto et al., 1995; Tsuji et al., 1990) and with the apparent decrease in differentiation potential found in BMSC from senescent animals (Quarto et al., 1995). Moreover, our data also show an increased frequency (both relative and absolute) of bipotential clones (OC) in marrows from younger donors. It is quite possible that this reflects an expansion of a more mature cellular compartment in organisms undergoing active skeletal growth such as children; furthermore it may explain the relative reduction of the more immature compartment. Pittenger et al. (1999) have recently reported the osteoadipogenic phenotypic association in 2-3 out of the six clones analyzed. Other authors have reported similar associations of mesenchymal phenotypes (Hicok et al., 1998; Houghton et al., 1998). Most of these studies were performed either by virally immortalized cells or by inducing the differentiated phenotype with pharmacological stimuli or by interfering with the basic mechanisms of gene expression. In particular, Pittenger and coworkers used a stimulation protocol to induce adipogenesis where the association of strong pharmacological adipogenic stimuli such as methylisobutylxanthine and indomethacine, in addition to dexamethasone, may have also induced the adipogenic phenotype in cells non-responsive to physiological stimuli (Pittenger et al., 1999). Under our conditions for inducing cell differentiation, we obtained clones able to differentiate as OCA (45/185), OC (132/185) and O (7/185). The other potential associations of phenotypes as well as the other pure phenotypes were never observed. Our experiments performed on a number of clones to determine their actual life span indicate that these cells are able to undergo a significant, although limited, number of mitotic divisions. Growth kinetics progress faster at the primary culture level and they gradually slow down during the following passages to reach a complete stop after about 22 cell doublings. BMSC clones show a progressive loss of differentiation potential through the mitotic divisions. In fact, the four OCA clones analyzed displayed a loss of their adipogenic potential by doubling stage 22; all but one still maintained their OC potential. Three out of four OC clones lost their chondrogenic potential between 19 and 22 cell doublings. These data are in agreement with the phenotypic association observed in clonal differentiation experiments. Furthermore,
they indicate the existence of a hierarchy in the BMSC differentiation pathway, where the adipogenic lineage diverges and becomes independent earlier, whereas the osteochondrogenic lineages proceed together, possibly diverging later. The hierarchy we have observed ‘in vitro’ could also be reconciled with a model recently proposed about lineage hierarchy ‘in vivo’ (Bianco and Cossu, 1999; Bianco et al., 1999). This model proposes that subsequent waves of lineages are not rigidly separated in the developing bone (Bianco and Cossu, 1999; Bianco et al., 1999). We are proposing that, during the process of osteogenic differentiation, the adipogenic is the first lineage to diverge. Therefore the proposed phenotypic interconversion should occur at progenitor level. Adipogenic differentiation in the adulthood is thus dependent upon the presence of multipotent progenitors in the bone marrow. We have observed that tripotential clones can also be derived from adult bone marrows, although with a reduced frequency. Our data also indicate that the osteogenic pathway is the ‘default’ lineage that these cells can progress through, possibly because of either an intrinsic commitment or the in vitro culture conditions representing a microenvironment favoring osteogenesis. It has to be acknowledged that osteocalcin expression, although considered as a specific marker for the osteogenic lineage, is not by itself predictive of the in vivo behavior of cells expressing it (Yamamoto et al., 1991). Kuznetsov and coworkers have in fact demonstrated that only about the 60% of clonal human marrow stromal fibroblasts are able to form bone when implanted in vivo (Kuznetsov et al., 1997). Furthermore, only about 40% of the bone forming clones was able to yield an extensive bone formation accompanied by hematopoietic tissue (Kuznetsov et al., 1997). This experimental evidence implies that with an in vivo assay it is also possible to distinguish at least three different classes of clones: (1) clones able to form bone and support hematopoiesis; (2) clones frankly osteogenic and (3) clones non-osteogenic. In conclusion, at a clonal level BMSC do display a multilineage differentiation potential. These cells proceed through a spontaneous process leading to osteogenic differentiation and progressively lose the multipotentiality. A hierarchy is evident in their differentiation pathway, with the adipogenic lineage diverging earlier and the osteochondrogenic lineages proceeding together, possibly branching later on. All together, these data suggest a deterministic differentiation model for bone marrow stromal cells. Partially supported by grants from Associazione Italiana Ricerca sul Cancro, Agenzia Spaziale Italiana (A.S.I.) and European Space Agency (E.S.A.).
REFERENCES Bennett, J. H., Joyner, C. J., Triffitt, J. T. and Owen, M. E. (1991). Adipocytic cells cultured from marrow have osteogenic potential. J. Cell Sci. 99, 131-139. Beresford, J. (1989). Osteogenic stem cells and the stromal system of bone and marrow. Clin. Orthop. Rel. Res. 240, 270-280. Beresford, J. N., Bennett, J. H., Devlin, C., Leboy, P. S. and Owen, M. E. (1992). Evidence for an inverse relationship between the differentiation of adipocytic and osteogenic cells in rat marrow stromal cell cultures. J. Cell Sci. 102, 341-351.
1166 A. Muraglia and others Bianco, P. and Cossu, G. (1999). Uno, nessuno e centomila: searching for the identity of mesodermal progenitors. Exp. Cell Res. 251, 257-263. Bianco, P., Riminucci, M., Kuznetsov, S. and Robey, P. G. (1999). Multipotential cells in the bone marrow stroma: regulation in the context of organ physiology. Crit. Rev. Eukary. Gene Expr. 9, 159-173. Caplan, A. I. (1991). Mesenchymal stem cell. J. Orthop. Res. 9, 641-650. Dennis, J. E., Merriam, A., Awadallah, A., Yoo, J. U., Johnstone, B. and Caplan, A. I. (1999). A quadripotential mesenchymal progenitor cell isolated from the marrow of an adult mouse. J. Bone Min. Res. 14, 700-709. Dexter, T. M., Allen, T. D. and Lajtha, L. G. (1977). Conditions controlling the proliferation of haemopoietic stem cells in vitro. J. Cell Physiol. 91, 335344. Egrise, D., Martin, D., Vienne, A., Neve, P. and Schoutens, A. (1992). The number of fibroblastic colonies formed from bone marrow is decreased and the in vitro proliferation rate of trabecular bone cells increased in aged rats. Bone 13, 355-361. Fried, A., Benayahu, D. and Wientroub, S. (1993). Marrow stroma derived osteogenic clonal cell lines: putative stages in osteoblastic differentiation. J. Cell Physiol. 155, 472-482. Friedenstein, A. J. (1990). Osteogenic stem cells in the bone marrow. Bone Min. Res. 7, 243-272. Galotto, M., Berisso, G., Delfino, L., Podesta, M., Ottaggio, L., Dallorso, S., Dufour, C., Ferrara, G. B., Abbondandolo, A., Dini, G. et al. (1999). Stromal damage as consequence of high-dose chemo/radiotherapy in bone marrow transplant recipients. Exp. Hematol. 27, 1460-1466. Gerasimov, U. V., Friedenstein, A. J., Chailakhyan, R. K. and Shishkova, V. V. (1986). [Differentiation potentials of clonal strains of bone marrow fibroblasts]. Biull. Eksp. Biol. Med. 101, 717-719. Goshima, J., Golberg, V. M. and Caplan, A. I. (1991). The osteogenic potential of cultured-expanded rat marrow mesenchymal cells assayed in vivo in calcium phosphate ceramic blocks. Clin. Orthop. 262, 298-311. Gundle, R., Joyner, C. J. and Triffitt, J. T. (1995). Human bone tissue formation in diffusion chamber culture in vivo by bone derived cells and marrow stromal fibroblastic cells. Bone 16, 597-601. Hicok, K. C., Thomas, T., Gori, F., Rickard, D. J., Spelsberg, T. C. and Riggs, B. L. (1998). Development and characterization of conditionally immortalized osteoblast precursor cell lines from human bone marrow stroma. J. Bone Min. Res. 13, 205-217. Houghton, A., Oyajobi, B. O., Foster, G. A., Russell, R. G. and Stringer, B. M. (1998). Immortalization of human marrow stromal cells by retroviral transduction with a temperature sensitive oncogene: identification of bipotential precursor cells capable of directed differentiation to either an osteoblast or adipocyte phenotype. Bone 22, 7-16. Johnstone, B., Hering, T. M., Caplan, A. I., Goldberg, V. M. and Yoo, J.
U. (1998). In vitro chondrogenesis of bone marrow-derived mesenchymal progenitor cells. Exp. Cell Res. 238, 265-272. Kodama, H., Koyama, H., Sudo, H. and Kasai, S. (1983). Adipose conversion of mouse bone marrow fibroblasts in vitro: their alkaline phosphatase activity. Cell Struct. Funct. 8, 19-27. Krebsbach, P. H., Kuznetsov, S. A., Satomura, K., Emmons, R. V., Rowe, D. W. and Robey, P. G. (1997). Bone formation in vivo: comparison of osteogenesis by transplanted mouse and human marrow stromal fibroblasts. Transplantation 63, 1059-1069. Kuznetsov, S. A., Krebsbach, P. H., Satomura, K., Kerr, J., Riminucci, M., Benayahu, D. and Robey, P. G. (1997). Single-colony derived strains of human marrow stromal fibroblasts form bone after transplantation in vivo. J. Bone Min. Res. 12, 1335-1347. Martin, I., Muraglia, A., Campanile, G., Cancedda, R. and Quarto, R. (1997). Fibroblast growth factor-2 supports ex vivo expansion and maintenance of osteogenic precursors from human bone marrow. Endocrinology 138, 4456-4462. Mathieu, E., Schoeters, G., vander Plaetse, F. and Merregaert, J. (1992). Establishment of an osteogenic cell line derived from adult mouse bone marrow stroma by use of a recombinant retrovirus. Calcif. Tissue Int. 50, 362-371. Ohgushi, H., Goldberg, V. M. and Caplan, A. I. (1989). Repair of bone defects with marrow cells and porous ceramic. Experiments in rat. Acta Orthop. Scand. 60, 334-339. Owen, M. (1988). Marrow stromal stem cells. J. Cell Sci. 10, 63-76. Parfitt, A. M. (1984). The cellular basis of bone remodeling: the quantum concept reexamined in light of recent advances in the cell biology of bone. Calcif. Tissue Int. 36, S37-45. Park, S. R., Oreffo, R. O. C. and Triffitt, J. T. (1999). Interconversion potential of cloned human marrow adipocytes in vitro. Bone 24, 549-554. Pittenger, M. F., Mackay, A. M., Beck, S. C., Jaiswal, R. K., Douglas, R., Mosca, J. D., Moorman, M. A., Simonetti, D. W., Craig, S. and Marshak, D. R. (1999). Multilineage potential of adult human mesenchymal stem cells. Science 284, 143-147. Quarto, R., Thomas, D. and Liang, C. T. (1995). Bone progenitor cell deficits and the age-associated decline in bone repair capacity. Calcif. Tissue Int. 56, 123-129. Tavassoli, M. and Friedenstein, A. (1983). Hemopoietic stromal microenvironment. Am. J. Hematol. 15, 195-203. Tsuji, T., Hughes, F. J., McCulloch, C. A. and Melcher, A. H. (1990). Effects of donor age on osteogenic cells of rat bone marrow in vitro. Mech. Ageing Dev. 51, 121-132. Yamamoto, T., Ecarot, B. and Glorieux, F. H. (1991). In vivo osteogenic activity of isolated human bone cells. J. Bone Min. Res. 6, 45-51.
