Calcif Tissue Int (1991) 49:221-225 Calcified Tissue International �9 1991 Springer-Verlag New York Inc.

Comment

Clonal Osteogenic Cell Lines Express Myogenic and Adipocytic Developmental Potential

A k i r a Y a m a g u c h i 1 a n d A r n o l d J . K a h n z

~Department of Oral Pathology, School of Dentistry, Showa University, Tokyo Japan; and ZDepartment of Growth and Development, School of Dentistry, University of California, San Francisco, 521 Parnassus Avenue, San Francisco, California 94122, USA

Received July 18, 1990

S u m m a r y . Clonal osteoblastic cell lines were isolated from neonatal rat calvariae and characterized with regard to a number of features associated with authentic osteoblasts. These included elevated alkaline phosphatase activity (rela- tive to fibroblasts), PTH and PGE2-stimulated increases in cAMP, the predominant synthesis of type 1 collagen, and the production of a mineralized matrix in vitro. By these criteria, five clones with osteoblast-like phenotypes were identified (ROB-C8a, CI 1, C20, C23, and C26) which varied somewhat in shape, levels of alkaline phosphatase activity, and in re- sponsiveness to PTH and PGE2. C l l , C20, and C23 re- sponded to both effector substances, whereas C8a only re- sponded to PTH and C26 only responded strongly to PGE 2. Upon further examination, two of the clones (C23 and C26) were also found to exhibit significant muscle myotube for- mation after reaching confluence, and three of the clones (C8a, C l l , and C26) showed marked adipocyte differentia- tion after treatment with dexamethasone. Overall, these data add further supporting documentation to (1) the suspected ontogenetic relationships of osteoblasts to other connective tissue cells, and (2) the concept that osteoblastic cells asso- ciated with neonatal rat calvariae are in various stable stages of differentiation and developmental commitment.

Key words : Osteoblastic cell lines - c A M P - P T H - P G E 2 - A L P - Mitotubes - Adipocytes.

To date, several osteoblastic cell lines have been success- fully employed for the analysis of the phenotype of osteo- blasts [1-8]. Based on these studies, it has become increas- ingly clear that osteoblastic cells in vitro express a spectrum of phenotypic characteristics, perhaps reflecting different stages of osteoblastic differentiation as "captured" by these lines when they were first established [3, 5-8].

Osteoblasts are believed to arise from a relatively prim- itive, mesenchymal progenitor cell that is also capable of producing other cell types belonging to the connective tissue family including cartilage, muscle and fat (adipocytes) [%12]. However, with very few exceptions (e.g., ref [5]),

Offprint requests to: A. J. Kahn

little attention has been paid to the phenotype stability of cultured osteoblastic cells and, in particular, to their capac- ity to differentiate into the other kinds of connective tissue cells.

In the present study, we isolated a number of clonal cell lines by outgrowth from calvariae of 1-day-old rats. During the process of the characterization of these clones, we ob- served not only a diversity of osteoblastic phenotypes among them, but also a varying capacity to differentiate into myotubes and adipocytes. These findings reaffirm the het- erogeneous nature of the osteoblastic population in situ and the close ontogenetic relationship of various members of the skeletal connective tissue family.

M a t e r i a l s a n d M e t h o d s

Establishment o f Osteoblastic Cell Lines

Small bone fragments dissected from parietal bones of 1-day-old Sprague-Dawley rats were cultured in alpha-MEM containing 10% fetal bovine serum (GIBCO, NY or HyClone, UT), penicillin (100 U/ml), and streptomycin (100 mg/ml). After 10 days of culture, those cells migrating from the explants were harvested by trypsin/EDTA digestion (0.05% trypsin and 0.02% EDTA) and passaged, twice weekly, to establish the parent ROB osteoblastic cell line. Clones and subclones were then prepared from different passages of ROB. The initial cloning was performed with 20th passage ROB cells using cloning rings. One of these clones (C8) was later recloned by limiting dilution to yield the line designated C8a. Other subclones (C11, C20, C23, and C26) were generated from 53rd passage ROB cells by the same dilution technique.

Phenotypic Characterization o f the Cell Lines

Four criteria were used to establish the osteoblastic phenotype: el- evated expression of alkaline phosphatase (ALP), cAMP response to PTH and PGE2, synthesis of type I collagen, and production of a mineralized matrix. ALP activity was measured in cell lysates using p-nitrophenylphosphate as a substrate [2]. Responsiveness to PTH and PGE 2 was determined by radioimmunoassay in near confluent cultures containing 1% bovine serum albumin and 1 mM iso- butylmethylxanthine using a 3H-labeled cAMP assay kit (Amer- sham, IL). The types of collagen synthesized by each cell line were determined by polyacrylamide gel electrophoresis [13] using pro- teins labeled in vivo with L-3H proline 04]. Matrix mineralization was assessed in cells cultured on tissue culture plastic or in a type I collagen gel (Cell Matrix I, Nitta Gelatin Co., Osaka, Japan) in media supplemented with 5 mM beta-glycerophosphate and 50

222 A. Yamaguchi and A. J. Kahn: Myogenic and Adipocytic Developmental Potential

40 �9 ROB I OCSe

c23 1 % A cz6 / \

35 �9 c:,o / N

o_ o: ~ o I - w E ~ 30 e5 o c .

~_ t s / " ' , , \

U

5 /'~ " I I I

DAYS

Fig. 2. Changes in alkaline phosphatase activity as a function of time in culture. 5 • 104 cells of each cell line were inoculated into wells of 6-well plates, and the enzyme activity was measured as described in Materials and Methods. Values represent the mean and SD of triplicate wells.

Fig. 1. Morphology, at confluence, of the nonclonal maternal ROB cell line (a) and the five clonal cell lines: (b) C8a, (c) C1 l, (d) C20, (e) C23, (f) C26. Magnification •

ixg/ml ascorbic acid. The presence of mineral deposits was estab- lished by von Kossa staining.

Detection of Myotubes and Adipocytes

The presence of myotubes was confirmed by immunostaining for desmin, a muscle-specific intermediate filament, using the indirect immunoperoxidase technique, rabbit anti-desmin antibody (Medac Gesellschaft, Hamberg, FRG) and biotinylated anti-rabbit antibody (Vector Laboratory, CA). Adipocyte conversion was induced by treatment with 10-7 M dexamethasone after the first day in culture. After 9 days of treatment, the cultures were stained with Oil Red O to identify the adipocyte phenotype.

than that of the C8a and C26 clones. More importantly, al- kaline phosphatase activity for each clone is at least 6-fold greater than that of l-day-old rat skin fibroblasts at conflu- ence (1.3 --- 0.3 p.mol/hour/mg protein).

cAMP Production in Response to PTH and PGE 2

With the exception noted below, all the clones respond to PTH and/or PGE2 by markedly increasing intracellular cAMP levels. Three of the clonal l i n e s - - C l l , C20, and C23--are responsive to both agents showing 40- to 100-fold increases in cAMP level within 20 minutes (Table 1). In con- trast, C8a is sensitive only to PTH and C26 reacts weakly to PTH but strongly to PGE 2 (Table 1 ; Fig. 3). These responses are both dose dependent and rapid, being clearly evident within 4 minutes of exposure to each effector substance and near maximal by 8 minutes.

Results

Morphology of the Clonal Cell Lines

Figure 1 shows the morphology of ROB and each clone at confluence. The C8a and C26 subclones display a flattened, polygonal shape, as does clone C20. However, the cell size of the latter is smaller than that of C8a and C26. Two clones, C11 and C23, are composed of more oval-shaped cells which upon reaching confluency became layered.

Collagen Analysis

The types of collagen synthesized by each clonal cell line are shown in Figure 4. Each cell line synthesizes predominantly type I collagen (>90%) and small but variable amounts of type III and V collagens.

Formation of Mineralized Matrices

Alkaline Phosphatase Activity

Changes in ALP activity for each cell line with time in cul- ture are shown in Figure 2. The maximum enzyme activity of the ROB, CI1, C20, and C23 cell lines is 2-4 times higher

In ROB, CI1, and C23 cell lines grown on tissue culture plastic, opaque regions appear in multilayered nodules after 10-12 days of culture. The mineralization in these opaque regions is clearly evident in von Kossa stained preparations (Fig. 5a). Although no appreciable amount of mineralization is detected in C8a, C20, and C26 clones in standard culture,

A. Yamaguchi and A. J. Kahn: Myogenic and Adipocytic Developmental Potential

Table 1. Effects of PTH and PGE2 on the cAMP production

Amounts of cAMP produced (pmol/well) Cell line Basal PTH PGE 2

C8a 4.0 • 0.3 159.0 • 8.9 5.0 • 1.0 CI1 1.5 -+ 0.7 119.0 • 4.2 61.0 • 7.0 C20 2.5 • 0.7 129.0 • 2.8 50.0 • 7.5 C23 2.0 • 0.0 111.5 • 6.3 142.5 • 13.4 C26 5.0 • 0.7 16.5 • 2.9 302.5 • 10.6

Cells were cultured for 6 days in 6-well plates. Cells were treated for 8 minutes with 200 ng/ml of bPTH(1-34) or 3.5 ~xg/ml PGE2, and the amount of cAMP produced was determined. Data are means • SEM of 2 wells

223

T 400

r- 2~176 C8G oo I 0 2 6

-~ I00 200 Q.

s :~ 5O IO0

cJ

~,o~ ~ s ~ ~ , ~ ,~. ,~

PTH PGE 2 PTH PGE 2

Fig. 3. Dose response in cAMP accumulation in C8a and C26 cells treated for 8 minutes with PTH (200 ng/ml) or PGE 2 (3.5 ixg/ml). Values are the mean - SD of triplicate wells.

Fig. 4. Analysis of collagen types synthesized by each clonal cell line. 3H-proline-labeled proteins in each clone were separated by SDS-PAGE as described in Materials and Methods.

these cell lines produce a large amount of mineralized ma- trices when incubated in a type I collagen gel supplemented with beta-glycerophosphate and ascorbic acid (Fig. 5b).

Muscle and Adipocyte Differentiation

Elongated, multinucleated cells appear in cultures of C23 and C26, reaching maximal levels after confluence. These cells exhibited spontaneous contractility, and stain posi- tively for desmin (Fig. 6a and b), thereby confirming the muscle (myotube) phenotype of these cells. The C8 and C11 subclones also form desmin-positive muscle cells, although the number and size of myotubes is smaller than those as- sociated with C23 and C26. The C20 clone appears to lack the ability to express the muscle phenotype. The muscle- forming properties of each clone are summarized in Figure 7.

The C26 subclone also forms a large number of adipo- cytes in response to treatment with dexamethasone for 7 days. The adipocyte phenotype was confirmed by the for- mation of large, Oil Red O positive, cytoplasmic droplets (Fig. 6c). The C8a and C l l subclones also produce adipo- cytes in response to dexamethasone, although the number of fat cells appearing in these clones is less than that observed in C26 (Fig. 7). Subclones C23 and C20 express minimal adipocyte-forming capability (Fig. 7).

Discussion

Among the calvarial cell lines established in this study, three

Fig. 5. Formation of mineralized matrices in vitro. (a) 21-day mono- layer culture of C 11 cells stained by the von Kossa method. Note the extensive areas of mineralization. Magnification x 75. (b) Mineral- ization of C26 cells cultured in collagen gel for 21 days. After fixa- tion, the gel was stained in situ by the yon Kossa technique. Note the areas of dark staining indicative of the presence of mineralized matrix. Magnification x 100.

clones (Cl l , C20, and C23) exhibit a wide spectrum of os- teoblastic characteristics as described previously [15]; high ALP activity, PTH and PGE2-dependent cAMP production, type I collagen synthesis, and a capacity to support matrix mineralization in vitro. C26 and C8a also express most of these characteristics but differ in their responsiveness to PTH and PGE2, respectively: C26 exhibits minimal respon- siveness to PTH, and C8a does not respond to PGE2. How- ever, in this latter regard, another well-characterized osteo- blastic cell line, ROS 17/2.8, also lacks this response mech- anism [15] as indeed do other cell lines derived from rat calvariae [3]. Thus, it is not certain how essential the PGE2 response mechanism is in identifying osteoblastic cells.

Although much remains to be learned about the pathway of osteoblastic differentiation, it is generally believed that it begins with the same progenitor cell that also gives rise to fibroblasts, adipocytes, skeletal muscle, and chondrocytes

224

Fig. 6. Myotubes and adipocytes formation by the C26 cell line. Desmin-positive, elongated myotubes in C26 cultures after 7 days of incubation. (a) Magnification • 100. (b) Magnification • 200. (e) Adipocytes in clone C26 after 8 days of treatment with dexametha- sone (1 • 10 - 7 M). Small vacuoles in the cytoplasm are stained with Oil Red O. Magnification • 100.

0 ,015

g �9 ~ 0 .010

o 0 ,005

0

40

30

c: 2 0 o o

~ ~ 0 -

< 0

-r

T

I i ~ - -

C S a C l l C 2 0 C 2 3 C26

f

II CSa C l l C20 C23 C26

Fig. 7. The degree of muscle and adipocyte conversion in each clone. To detect myotubes, cells were inoculated in 24-well plates and cultured for 7 days. Muscle conversion was expressed as a percentage of desmin-positive cells per estimated total cell number. Values are the mean and SE of triplicate wells. To stimulate adipo- cyte formation, the cells were treated with 10 7 M dexamethasone for 9 days. The number of the Oil Red O-positive cells as counted in a defined region (2.16 mm 2) of each culture using a microscope. The total number of adipocytes in each well was calculated from the number of fat cells determined per unit area. Adipose conversion was expressed as the percentage of adipocytes per total cell number in each culture. Values are given as the mean and SE of triplicated wells.

[9]. This view is supported by work from a number of labo- ratories [5, 10-12] including the now classical studies of Friedenstein on induced and determined osteoprogenitor cells [16]. The results of the present investigation are also consistent with this concept. All of the cell lines generated from newborn rat calvariae display some of the major char-

A. Yamaguchi and A. J. Kahn: Myogenic and Adipocytic Developmental Potential

acteristics associated with authentic osteoblasts, but vary in both the degree of expression of these characteristics and, more importantly, in their capacity to differentiate into al- ternative cell types. C8a and C11 show an ability to form adipocytes in the presence of dexamethasone but were poor muscle formers. C23 could express the muscle phenotype but not fat cell formation. C20 is very limited, under the conditions tested, in its ability to express alternative pheno- types. In contrast, C26 displays extensive muscle and adi- pocyte differentiation under appropriate circumstances. This clonal cell line also exhibits the lowest level of alkaline phos- phatase activity and the poorest response to PTH, suggest- ing that it is the least differentiated (developmentally com- mitted) of the various cell lines produced in this study. In this latter regard, C26 cells might be compared to the cell lines isolated by Grigoriadis et al. [5] from normal rat calva- ria (e.g. RCJ 3.1) and by Kellermann et al. [8] from mouse teratocarcinoma, and C3H 10T1/2 and Swiss 3T3 cells treated with 5-azacytidine [17]. In these instances, clonal "mesenchymal" cell lines also express multiple develop- mental potential including the ability to differentiate into chondrocytes, muscle cells, and adipocytes. We found no chondrocytes in C26 after 3 weeks in culture with dexameth- asone treatment (10-7 M, unpublished data) which was the most effective inducer of this cell type in RCJ 3.1 [5].

Though it is not yet possible to draw any definitive con- clusions about osteoblast lineage and differentiation, the data presently available are consistent with the following view: Osteoblasts are derived from a multipotential precur- sor cell located in the bone lining or within bone marrow. This developmental potential becomes more restricted as the cells mature within the context of an epigenetic mechanism that confers phenotypic stability even when presumptive os- teoblasts are isolated and cultured in vitro. In this sense, at least, bone tissue must be considered as consisting of a het- erogeneous cell population that cannot necessarily be ren- dered homogeneous by cloning and subculture. In fact, to achieve such homogeneity, it will probably be necessary not only to produce relatively immature cell lines (e.g., C26, RCJ 3.1) but also to identify those exogenous factors that control osteoblast differentiation in vivo so that they can be applied to such lines in vitro. Work to identify such factors, including those that restrict as well as promote osteoblast differentiation, is now in progress.

Acknowledgments. This work was supported in part by Grants-in- Aid (01870076 and 01870076) from the Ministry of Science, Educa- tion and Culture of Japan (A.Y.) and NIH Grant AR32788-05 (A.J.K.).

References

1. Majeska RJ, Rodan SB, Rodan GA (1980) Parathyroid hormone responsive clonal cell lines from rat osteosarcoma. Endocrinol- ogy 107:1494-1503

2. Partridge NC, Alcon D, Michelangeli VP, Ryan G, Martin TJ (1983) Morphological and biochemical characterization of four clonal osteogenic sarcoma cell lines of rat origin. Cancer Res 43:4308--4314

3. Aubin JE, Heersche JNM, Merrilees MJ, Sodek J (1982) Isola- tion of bone cell clones with differences in growth, hormone responses, and extracellular matrix production. J Cell Biol 92:452--461

4. Sudo J, Kodama H, Amagai Y, Yamamoto S, Kasai S (1983) In vitro differentiation and calcification in a new clonal osteogenic

A. Yamaguchi and A. J. Kahn: Myogenic and Adipocytic Developmental Potential 225

cell line derived from newborn mouse calvaria. J Cell Biol 96:191-198

5. Grigoriadis AE, Heersche JNH, Aubin JE (1988) Differentiation of muscle, fat, cartilage, and bone from progenitor cells present in a bone-derived clonal cell population: effect of dexametha- sone. J Cell Biol 106:2139-2151

6. Heath JK, Rodan SB, Yoon K, Rodan GA (1989) Rat calvarial cell lines immortalized with SV-40 large T antigen: constitutive and retinoic acid-inducible expression of osteoblastic features. Endocrinology 124:3060-3068

7. Guenther HL, Hofstetter W, Stutzer A, Muhlbauer R, Fleish H (1989) Evidence for heterogeneity of the osteoblastic phenotype determined with clonal rat bone cells established from trans- forming growth factor-B-induced cell colonies grown anchorage independently in semisolid medium. Endocrinology 125:2092- 2102

8. Kellermann O, Buc-Caron MH, Marie PJ, Lamblin D, Jacob F (1990) An immortalized osteogenic cell line derived from mouse teratocarcinoma is able to mineralize in vivo and in vitro. J Cell Biol 110:123-132

9. Owen M (1988) Marrow stromal stem cells. J Cell Sci (suppl) 10:63-76

10. Caplan AJ, Stoolmilaer AC (1973) Control of chondrogenic ex-

pression in mesodermal cells of embryonic limb. Proc Natl Acad Sci USA 70:1713-1717

11. Osdoby P, Caplan AJ (1976) The possible differentiation of os- teogenic elements in vitro from chick limb mesodermal cells. I. Morphological evidence. Dev Biol 52:283-299

12. Reddi AH, Huggins CB (1972) Biochemical sequences in the treatment of normal fibroblasts in adolescent rats. Proc Natl Acad Sci USA 69:1601-1605

13. Sykes B, Puddle B, Francis M, Smith R (1976) The estimation of two collagens from human dermis by interrupted gel electropho- resis. Biochem Biophys Res Commun 72:1472-1480

14. Kream BE, Rowe D, Smith MD, Maher V, Majeska R (1986) Hormonal regulation of collagen synthesis in a clonal rat os- teosarcoma cell line. Endocrinology 119:1922-1928

15. Rodan GA, Rodan SB (1984) Expression of the osteoblastic phenotype. In: Pech WA (ed) Bone and mineral research, vol. 2. Elsevier Science Publishing, Amsterdam, pp 244--285

16. Fridenstein AJ, Piatetzky-Shapiro II, Petrakova KV (1966) Os- teogenesis in transplants of bone marrow cells. J Embryol Exp Morph 16:381-390

17. Taylor SM, Jones PA (1979) Multiple new phenotypes induced in 10T1/2 and 3T3 cells treated with 5-azacytidine. Cell 17:771- 779