Formation of chondrous and osseous tissues in micromass cultures of rat frontonasal and mandibular ectomesenchyme

Differentiation (1990) 44: 197-206

Differentiation Ontogeny and Neoplasia Q Springer-Verlag 1990

Formation of chondrous and osseous tissues in micromass cultures of rat frontonasal and mandibular ectomesenchyme Robert M. Langille* and Michael Solursh Department of Biology, University of Iowa, Iowa City, IA 52242 USA Accepted in revised form May 11, 1990

Abstract. Rat frontonasal and mandibular mesenchyme was isolated from day-12 1/2 (stage-22) rat embryos and cultured at high density for up to 12 days. The stage chosen was based on the observation that mandibular mesenchyme at this stage became independent of its epithelium with respect to the production of both cartilage and bone. Frontonasal cultures developed aggregates of anastomosing columns of cells within 2 days. These grew as the cells enlarged, laying down an Alcian-blue-positive matrix by day 3 of culture. Significant mineral was detected by von Kossa staining by day 5 at which time the aggregates covered a large portion of the culture, eventually covering the entire micromass by day 10-12. Mandibular cultures developed centrally located nodular aggregates by 3 days of culture. These nodules increased in number, spreading outwards as the cells enlarged, laying down an Alcian-blue-positive matrix by day 4 and mineral by days 6 7 . At this time the nodules began to elongate and coalesce, but never covered the entire culture over the 12-day period. Antibody staining revealed that in both cultures the cells were initially positive for type I collagen. Subsequently, the aggregates began expressing type I1 collagen, followed by type X, which coincided with the onset of mineralization. At this time some cells were negative for these cartilage markers, but positive for osteoblast markers, bone sialoprotein 11, osteocalcin and type I collagen. In addition osteonectin and alkaline phosphatase were demonstrable in all of the aggregate cells late in the culture period. This provided clear evidence that chondroblast and osteoblast differentiation was proceeding within these cultures. The culture of rat facial mesenchyme should prove very useful, not only for the analysis of bone and cartilage induction and lineage relationships, but also in * Present address: Department of Anatomy, University of Ottawa, 451 Smyth Rd., Ottawa, ON, Canada K I H 8M5 * To whom offprint requests should be sent

furthering our knowledge of craniofacial differentiation, growth and pattern formation by extending our analysis to a mammalian system.

Introduction

The study of craniofacial development is important not only for the information on the organization and construction of this complex structure from many parts, but also for the information on the mechanisms and control of differentiation of ectomesenchyme, a cranial neural-crest-derived population of embryonic cells within the craniofacial region which gives rise to a wide variety of tissues, including a majority of the connective tissue and skeletal components of this region (for reviews see [23,26,15]). Much of the previous work has centered on experiments with chick embryos, primarily in vivo or organ culture, but more recently work has begun on the analysis of differentiation of chick embryo facial mesenchyme in culture (see review [43]) and the in vitro effects of modulators of differentiation such as retinoic acid [43, 221. Recently, differential staining [25] and labelling experiments [36, 37, 331 have given us direct evidence of neural crest contributions to a large area of mammalian facial mesenchyme equivalent to that seen in other vertebrates [15]. However, no reports have been published to date on the culture of mammalian facial mesenchyme with the goal of analyzing the differentiation of this tissue. The simultaneous induction of cartilage and bone observed to occur in the mandibular mesenchyme of the mouse by Hall [ 141, makes this the perfect tissue in which to study both the mechanisms of cartilage and bone induction and the lineage relationships between these two tissues, in addition to the other tissues derived from ectomesenchyme, thereby furthering the analysis of the con-

198

trol of differentiation and pattern formation in the mammalian face. Thus, the present study was undertaken to provide a careful analysis of the in vitro differentiation capacity of mammalian facial mesenchyme, particularily its chondrogenic capacity, and to compare the results with those obtained previously from avian embryos. In this paper we present the progress differentiation of rat mandibular and frontonasal mesenchyme within micromass cultures, documenting the expression of a variety of phenotypic aspects normally associated with skeletal tissue in vivo. Our analysis of the differentiation and phenotypic expression found to occur within these cultures clearly indicates that the differentiation of both chondroblasts and osteoblasts from primary mesenchyme is occurring in this system.

Methods Tissues. Virgin female rats (Wistar) were mated with bucks overnight and those females which had vaginal smears containing spermatozoa were individually segregated the next morning and that day designated as day 0. On the 11th or 12th day, rat embryos corresponding to stages 20 (11 1/2 days, 16-22 somites) or 22 (12 1/2 days, 29-35 somites [7]) were removed in a sterile manner from pregnant females which had previously been anesthetized with ether and killed by cervical dislocation. The embryos were removed in utero and placed immediately in cold Puck's saline G then freed from the uterus and all extraembryonic membranes. Mandibles and frontonasal masses were dissected off the head with fine forceps and prepared for grafting or cell culture as described below. Organ culrure. Mandibles were immersed in 2.5% trypsin-0.43% pancreatin in tryodes at 4" C [I41 for 1/2-3/4 h, until the epithelia was obviously well loosened. At this point the epithelia were easily removed with fine forceps and the remaining intact mesenchymal components of the mandibles were placed in cold tyrodes with 50% horse serum to stop any further enzymatic reactions. The mandibles were then either grafted onto chorioallantoic membranes (CAM) of day-8 host chicken embryos after the detailed methods of Hall [13], or cultured in vitro, both for 9 days. Hosts for CAM grafting were first candled to identify sites with good vascularization and then windows were cut in the outer shell after Hall [13]. Mandibles to be grafted were positioned onto squares of 0.45-pm Millipore filter paper (cut approximately 5 mm x 5 mm) in a drop of saline, then placed, inverted, onto the CAM at a junction of two or more large blood vessels. The windows were immediately sealed with the shell piece and cellophane tape, and the host eggs placed in an egg incubator (without rotation) for the 9-day culture period. Survival of the grafts was 20 out of 22 or 91%. Mandibles grown in vitro were placed on 0.45-pm Millipore filters and cultured on wire mesh grids at the air-media interface. Media consisted of BGJb (Gibco, NY, USA) plus 15% fetal calf serum and antibiotics. Controls consisted of both intact, untreated mandibles and mandibles which were treated in enzyme in an identical manner to that above, but did not have the epithelia dissected away. All of the controls were grafted onto the CAMS of host chicks for the same 9-day period. Survival of these cultured tissues was 8 out of 12 or 67%. Control and experimental tissues were then fixed in 10% buffered formalin and stained whole with Alcian blue to identify cartilage and Alizarin red to identify bone, as described by Kelly and Bryden [20] and Hanken and Wassersug [16].

Cell culrure. Mandibles and frontonasal masses were kept separate and treated identically. Upon removal they were washed in calci-

um- and magnesium-free saline G and a suspension of mesenchyma1 cells obtained by the method of Ahrens et al. [l]. Briefly, the tissues were exposed to 0.1% trypsin and 0.1 % collagenase in 10% chick serum for 10 rnin at 37" C, pipetted to loosen the epithelia, then filtered through two layers of #20 Nitex (Tetko Inc., NY, USA). After the total cell number was calculated for the resulting mesenchymal cell suspensions, they were centrifuged at 400 g for 6 min and resuspended in medium (modified essential medium, MEM; 10% fetal calf serum and antibiotics; all from Gibco, NY, USA) at a density of 2 x 10' cells/ml or greater in order to plate the cells at high density [l]. Ten-microliter drops of media and cells were placed in 35-mm tissue culture dishes and allowed to adhere at 37°C for 2 h prior to the addition of 2ml medium. This medium was changed daily over the 12-day culture period. Each day, some cultures were removed and fixed in either two changes of acet0ne:methanol (1 :1) for a total of 10 min prior to immunocytochemical analysis or Kahle's fixative [12] for 20 rnin prior to histological analysis. Microscopy. Each day live cultures were analyzed by phase microscopy for the extent of histodifferentiation. Photomicrographs were taken in some instances in order to obtain a series of photomicrographs of live cultures over the entire 12-day culture period. Those cultures prepared for histological analysis were washed subsequent to fixation in three changes of phosphate-buffered saline (PBS) and stained as follows. Most of the cultures were stained overnight in 0.5% Alcian blue (in 0.1 N HCI), destained with three changes of 0.1 N HCI, washed further in PBS (some counterstained with Carazzi's Hematoxylin [6]) and mounted in glycerol. The rest of the cultures were stained by the von Kossa technique [31] alone or in combination with Alcian blue. Those cultures fixed for immunostaining were subsequently washed in three changes of PBS, and incubated in a primary antibody solution for 45 min. The antibodies used include anti-collagen I (rabbit polyclonal [Pab], Pasteur Institute, DMI, Westbrook, ME diluted 1/40), anti-collagen I1 (rabbit Pab and mouse monoclonal [Mab], supplied by T.F. Linsenmeyer), anti-collagen X (Pab made to chick type X but recognizes some mammalian type X including rat [2, 21]), anti-alkaline phosphatase (Mab, made against bone alkaline phosphatase, provided by G. Rodan), anti-osteonectin (Pab, courtesy of Larry Fisher, NIH), anti-bone sialoprotein I1 (Mab; WVIDl(9C5), Developmental Studies Hybridoma Bank), anti-meromyosin (Mab; MF-20 [3]; Developmental Studies Hybridoma Bank), anti-osteocalcin (goat Pab, supplied by J. Lian). Some cultures were treated with testicular hyaluronidase (1 mg/ ml hyaluronidase in PBS) for 30 min at room temperature prior to exposure to the primary antibody. Additionally, any cultures which exhibited significant precipitate (generally those grown for 12 days), were pretreated with 0.25 M disodium EDTA for 30 rnin at room temperature prior to antibody labelling. After primary antibody incubation and an additional three changes of PBS, tetramethylrhodamine isothiocyanate or fluorescein isothiocyanate labelled anti-rabbit (for Pabs) or anti-mouse (for Mabs) IgGs (all 1/300 in PBS; Cappel laboratories Inc, Cochranville, PA, USA) were then added as the secondary antibodies and the cultures incubated for an additional 45 min after which they were again washed in PBS and mounted in glycerol containing p-phenylenediamine [ 191. In some instances the cultures were doublc-staincd with a combination of one of thc above Pab followed by a Mab. Control sections were incubated with tissue culture supernatant conditioned by mouse myeloma cells, rabbit serum or both, depending on antibody staining being compared.

Results Organ culture

Prior to analyzing the ability of facial mesenchyme to express skeletal phenotypes in cell culture, the first step was to establish a stage at which at least some of the

199 Table 1. Effects of culturing day-1 1 1/2 and -12 1/2 (stages-20 and -22) rat mandibular mesenchyme with or without mandibular epithelia in vitro or grafted onto the chorioallantoic membrane (CAM) of chick hosts

Conditions With epithelia’ Day 11 1/2 Day 12 1/2 Without epithelia Day 11 112 Day 12 1/2

n

No. with cartilage

6 6

6 5b

8

0

8

5

No. with bone

Includes three intact mandibles and three mandibles subjected to enzyme treatment but without epithelial removal One of the intact untreated day-1 1 1/2 (stage-20) mandibles did not stain well with Alcian blue and was scored as cartilage-deficient

a

Cell culture

Based on the observed competence of stage-22 rat mandibular mesenchyme to form both cartilage and bone without the presence of facial epithelia, this stage of facial mesenchyme was used to initiate cell cultures. By this stage there has been sufficient development of the facial region so that the maxillae, hyoid arch and frontonasal processes are large or prominent enough to facilitate easy mesenchyme isolation for cell culture. Thus, at the outset the study included the culture of stage-22 frontonasal, maxillary and hyoid arch mesenchymes as well as mandibular. Of these, only the frontonasal and mandibular mesenchymes displayed detectable skeletal histogenesis over a 12-day culture period as detailed below (observations for the maxillary and hyoid arch cultures not shown). Thus, the present study was restricted to cultures of mandibular and frontonasal mesenchyme.

mesenchyme of the face is competent to form cartilage and bone. Following the only previous work of this type [14], which was performed on mouse mandibles, the inductive state of the rat mandibular mesenchyme with respect to skeletal tissue differentiation was analyzed. Mandibles from day 11 1/2 (stage 20) and 12 1/2 (stage 22) rat embryos were grown in culture or on chick CAMS either with or without their epithelia. As shown in Table 1 stage-22 mandibular mesenchyme is capable of bone and cartilage differentiation when cultured in isolation while stage-20 mandibles require the presence of epithelia in order to produce cartilage and bone.

Within 24 h of attachment, cells in both frontonasal (FN) and mandibular (MD) cultures had formed a confluent ‘pavement’ pattern. Like facial mesenchyme of the chick [42], that of the rat displayed a varied morphology depending on the origin. FN cultures progressed more quickly than MD cultures and displayed anastomosing columns of enlarging cells by day 3 (Fig. 1 a), which were most easily identifiable as cell ag-

Fig. 1a-f. Progress of differentiation in live micromass cultures of frontonasal (FN; a-c) and mandibular (MD; d-9 mesenchyme. Aggregating cells (arrows) are evident in both FN and M D cultures at 3 days (a, d). In FN cultures by day 5 (b), the groups of aggregating cells enlarge and form anastomosing columns of phase-bright cells; a portion of one is shown together with two clusters of neuron-like cells (arrowheads). Expansion of these enlarged cell masses occurs to cover most of the culture; by day 8 (c) recruitment is

largely confined to the periphery where a secondary aggregate (*) is expanding to meet the main aggregate region. In MD culture, discrete nodules (arrow) of cells enlarge and become phase-bright by day 6 (e). Expansion is limited largely to elongation and fusion of nodules formed earlier in culture as demonstrated by the three nodules in this day-12 culture (9.Note the matrix material between the cells. Peripheral flattened cells line the nodules (arrowhead). Bar. 0.5 mm

Morphology of live cultures

200

Table 2. Summary of Alcian-blue and von Kossa staining and various extracellular matrix molecules and related enzymes localized in rat facial cultures at specific stages of differentiation Specific molecule

Undifferentiated a pavement of cells

Alcian-blue-positive matrix

Absent

von Kossa-positive matrix

Absent

Collagen type 1 Collagen type I1

Present Absent

Collagen type X BSP I1

Absent Absent

Osteonectin Alkaline phosphatase Osteocalcin

-d Absent

Aggregation and cell enlargement

Present (appears late) Present (appears late) Present Present (appears very early) Present Present (appears late) Present -

Aggregated cells fully enlarged and phase-bright' Aggregated region

Nonaggrega ted region

Present

Absent

Prescnt Prcscnt Prcscnt

Present (weaker than aggregate) Present Absent

Present Present

Absent Absent

Present Present Present

Present Absent Absent

Frontonasal cultures day 1 ; mandibular cultures days 1-2 Frontonasal cultures days 2-5; mandibular cultures days 3-6 Frontonasal cultures days 5-12; mandibular cultures days 6-12 Not performed BSP, bone sialoprotein

a

Fig. 2a-f. Alcian blue (a, b, d, e) and von Kossa (c, f) staining demonstrate cartilage matrix production and mineralization in FN (a-c) and MD (d-f) culture. Haematoxylin counterstain reveals wide areas of cell aggregation in a 3-day FN culture (a), while darker Alcian-blue-positive areas of cartilage matrix ( a r r o w ) are just appearing. Note the neuronal-like clusters, which stain intensely with haematoxylin (arrowhead). On day 10 (b), Alcian-blue-positive cartilage matrix covers a wide area of the cultures. Mineraliza-

tion of FN cultures is extensive, as demonstrated in c at 12 days. MD cultures also form cell aggregates as shown by haematoxylin staining at 4 days (d), but more-limited skeletogenic differentiation as demonstrated by Alcian-blue-positive nodules of cartilage matrix (arrows). At 12 days (e), the only aggregates are the centrally located nodules (N), most of which stain intensely with Alcian blue. At this time (9, the M D nodules are also positive for calcium mineralization. Bar, 100 pm

gregates when stained with hematoxylin. These cells grew phase-bright by days 5-6 (Fig. 1 b), and by days 7-8 the aggregates they formed began to display precipitated material in the extracellular matrix (ECM); (Fig. 1c), which became quite dense by day 12. Despite very thorough removal of all epithelia (both neural and

epidermal) from the FN processes, these cultures consistently displayed additional clusters of cells which were identified as neurons based on their polar morphology and axon-like interconnections. The presence of these cells had little apparent effect on the differentiation of skeletal/connective tissues in FN cultures when com-

Fig. 3a-d. Initial skeletal cytodifferentiation. Phase-contrast micrograph of a 2-day M D culture reveals a pavement of cells (a), which are positive for type I collagen (b). By day 3 in another MD culture (c), some of the aggregating cells (arrow) begin to exhibit type I1 collagen staining (a). Bar. 100 pm

Fig. 4a-f. Advanced cartilage differentiation. With the enlargement of the aggregating cells, as in this phase-contrast micrograph of a day-5 MD nodule (a), collagen type 11 staining is found throughout the entire area of enlarged cells, and type X collagen can now be visualized, as demonstrated in a 6-day M D culture doublestained for type I1 (b) and type X (c) collagen. As the cells become

fully phase-bright and the aggregate areas begin to coalesce, as in this day-9 FN culture viewed under phase contrast (d), type X collagen staining of the aggregate areas becomes identical to that for type I1 as demonstrated by a double-stained day-9 FN culture (e, type 11; f, type X; arrows indicate the same cells in both cultures as a reference). Bur. 100 pm

202

pared to M D cultures, which lacked these neuronal-like cells. When left 10-1 2 days the growth of anastomosing columns of cells covered the neural clusters and partially or completely obscured them. In M D cultures, centrally located whorls first appeared in the pavement of cells by day 2, and these cells had organized into distinct nodules by day 3 (Fig. 1 d). Enlargement of the nodular cells took about 3 more days (Fig. le), after which they too became phase-bright, on days 6-7 (Fig. I f ) . At this point, the nodules elongated and became surrounded by a single layer of flattened peripheral cells. Such peripheral cells were rarely observed around the anastomosing columns of cells within F N cultures except around isolated foci of enlarged cells, which are occasionally found in these latter cultures. As elongation of the nodules in the M D cultures continued, they began to coalesce and new nodules began to differentiate in a peripheral position relative to the centrally located ones. Precipitate was not seen in the ECM until days 9-12. Although no higher

Fig. 5a-i. Evidence for bone and muscle differentiation. Ten-day MD culture double-stained for type I (a) and type I1 (d) collagen. Cells containing each collagen type do not correspond. When similar MD cultures are double-stained for bone sialoprotein I1 (BSP 11) (b) and collagen type I1 (e). it is evident that again positive cells for each antigen do not correspond. A 12-day MD (c and 9 culture displays a large amount of extracellular precipitate, even after sodium EDTA treatment. The decalcified matrix material

incidence of cell death could be observed visually between the M D and FN cultures, the M D cultures needed to be plated at between 3-4 x lo5 cells/pl or 1 1/2-2 times the number usual for micromass cultures for consistent aggregation of the cells into nodules to occur. Dgferentiation of skeletal tissues

Both M D and F N cultures stained with Alcian blue (at pH l), which is indicative of a cartilaginous matrix, and with von Kossa, which is indicative of calcium phosphate mineralization. Alcian-blue-positive matrix was first identifiable, although weakly, in cultures with cell aggregates (FN day 3 of culture, M D day 4; Table2, Fig. 2a, d) and reached a maximum staining intensity at the stage when the aggregate cells were fully enlarged and phase-bright (day 5 F N ; day 6-7 MD; Table2, Fig. 2 b, e). The cell aggregates displayed calcified matrix only after significant cell enlargement had occurred (day

(0strains intensely for BSP I1 (c). By 5 (FN cultures) or 6 (MD cultures) days the aggregates were also positive for alkaline phosphatase (g) and osteocalcin (h; both 10-day MD cultures). Nonskeletal differentiation also occurred in the facial cultures. Individual myocytes (i; 6-day MD culture) differentiated very early, outside the aggregate regions, here stained with an antibody to muscle meromyosin. Bur, 100 pm

203

5 FN; day 6-7 MD; Table 2) and calcification continued to the point where by day 8-9 (FN) or days 10-12 (MD) staining with von Kossa was intense, especially in the FN cultures (Fig. 2c, f), and distinct matrix precipitate could be easily observed in some cultures (Fig. 5 f). Occasionally, some of these older cultures displayed nodules which had very poor or no affinity for Alcian blue, but still displayed calcified matrix as revealed by the von Kossa technique. The presence of specific markers (collagens type I, 11, X, osteonectin, bone sialoprotein 11, osteocalcin and alkaline phosphatase) of skeletal/connective tissue differentiation was analyzed within the cultures over the 12-day period. The results are summarized in Table 2. Prior to any overt differentiation (excluding the neuronal cells of the FN cultures) cultures were positive for only type I collagen (Fig. 3a, b). Aggregating cells in cultures from both facial regions (FN day 2; MD day 3) began to demonstrate type I1 collagen synthesis (Figs. 3c, d ; 4a, b). By the time the cells were enlarged and becoming phase-bright (6-7 days, MD; 5 days, FN) the aggregates displayed intense staining with type I1 collagen antibody, and various cells which were positive for type X collagen could now be observed (Fig. 4c). Type X collagen staining increased rapidly thereafter until all of the cells positive for type I1 collagen were also positive for type X (Fig. 4d-f). The cell aggregates also became positive for alkaline phosphatase at this time (Fig. 5g). Bone sialoprotein I1 (BSP-11), a bone-specific molecule [lo], was first detected within aggregate cells only after they had become phase-bright (Table 2). BSP-I1 appeared to be associated with cells which were not positive for type I1 collagen by analysis of double-stained preparations (Fig. 5c, d). Interestingly when day-I2 cultures with significant precipitate were treated with EDTA and then exposed to anti-BSP-11, the antibody not only stained the cells, but also the remaining matrix material (Fig. 5 b, c). Double staining of cultures at this later stage with anti-collagens type I and I1 revealed that these two molecules were also restricted to different cells within the aggregates (Fig. 5a, b). The enlarged cells within the aggregates were also shown to be positive for osteocalcin (Fig. 5 h), which was absent from the interaggregate areas while osteonectin was observed throughout the cultures (not shown). Myocytes were also observed within both types of cultures as was confirmed by their staining for muscle myosin (Fig. 5i), as has been found in chick embryo facial cultures [43]. For the present study no effort was made either to quantify their numbers or detail their relationships with the other cell types differentiating within these cultures.

Discussion The analysis of the differentiation and phenotypic expression found to occur within cultures of rat facial mesenchyme clearly indicates that both chondrogenic and osteogenic differentiation occurred in these cultures.

Further, the immunological and histochemical evidence indicates that skeletogenic differentiation occurred in a definite progression. As with other types of mesenchyme in vivo [40] and in vitro [41] all of the cells were initially positive only for type I collagen. The very early expression of type I1 collagen by the aggregating (nonneural) cells within the cultures, and production of Alcian blue-positive matrix within the coalescing nodules of the mandibular cultures and the areas of cell aggregation of the frontonasal cultures, clearly indicates that cartilage differentiation occurred and was the first skeletal phenotype to differentiate in these cultures, as is the case in vivo [9]. Subsequently, the co-localization of type X, initially with only some cells but over time with all of the cells which expressed type I1 collagen, coupled with the presence of alkaline phosphatase and extensive mineralization, suggests that differentiation of cartilage within these cultures proceeded to a state of highly mineralized and possibly hypertrophic cartilage. A similar level of hypertrophy with the expression of type I1 and X collagen can be obtained from chick limb and facial mesenchyme and sternal cartilage mesenchyme, but to date, such a high level of type X collagen expression has been found only after these cells have been placed in gel culture [4, 351. Mineralization of a cartilaginous matrix has also been found to occur in high-density cultures of mouse limb bud mesenchyme grown for 14 days or more, but at present it is not known whether these cells produce type X collagen [1I]. The presence of osteonectin within the aggregates lends further support for the similarity of phenotypic expression found within the cultures as compared to differentiation in vivo. Osteonectin is a prominent feature of mineralized skeletal tissue [38] and has been found in osteoid formation in bone marrow cultures [24], although it is not restricted solely to bone as was once throught [34] and indeed has recently been localized in the mandibular condylar cartilage of the rat [8] and found extensively within skeletogenic and somitogenic areas of the mouse embryo [28]. Concurrent with the onset of type X collagen expression and mineralization was the appearance within the aggregates of cells which were negative for type I1 collagen but instead stained positive for bone sialoprotein 11, which to date has been found to be produced exclusively by bone cells or bone cell lines [ 10, 321. Furthermore, in contrast to the gradual transition from type I collagen to type I1 as observed in in vitro chondrogenesis of chick limb mesenchyme by von der Mark and von der Mark [41], double staining revealed cells within the centers of the aggregates which were negative for type 11, but intensely positive for type I collagen. Although type I collagen is found in a variety of connective tissues, it is the major collagen of bone [34] and is not found in cartilage matrix. This evidence, coupled with the observation that many cells within the aggregates were also positive for osteocalcin, a gamma-carboxyglutamate-rich protein produced exclusively by bone cells et al. 17, 181, clearly indicates that some of the cells within these mineralizing aggregates were forming bone

204

matrix molecules and expressing osseous rather than chondrous phenotypes. These osseous phenotypic characters have also been identified either within mammalian osteoblasts or adjacent extracellular matrix when these cells have been cultured in vitro [34]. At present we can not state whether the in vitro progression of skeletogenesis observed within these cultures exactly mimics the process found in vivo. Clearly, formation of Meckel’s cartilage, the trabeculae and the cartilaginous nasal apparatus precedes the deposition of bone in the frontonasal and mandibular areas of rodents ([9], unpublished observations). However, bone deposition within the face proceeds by both intramembranous and endochondral bone formation. This is particularly the case with the dentary, which ossifies intramembranously along most of its length with the exception of the anterior portions, which ossify endochondrally [5], and the angular and condylar region, which possess secondary cartilage [9], the latter region displaying a significant center of endochondral ossification from which bone or cartilage form. In the case of the frontonasal process, although most of the bones such as the premaxilla and nasal are membranous [9], the posterior portion of the nasal septum is replaced endochondrally [9, 301. Thus, it will be important in future studies to discern whether the cells expressing osseous phenotypes in rat facial mesenchyme cultures arise independently of the cartilaginous cells, or in a transition from the later cells or whether indeed they arise by both pathways. The immunochemical evidence cited above suggesting that cells within the mandibular and frontonasal cultures are differentiating along an osteogenic pathway is further corroborated by our findings that, at least in the case of the mandibular process, the mesenchyme from stage-22 rat embryos cultured on host chick CAMS was capable of forming bone and cartilage without any further epithelial interaction, as is the case for the equivalent-stage mouse mandibular mesenchyme [141. Clearly, the cells which were placed in micromass culture in subsequent experiments had already been induced to form both cartilage and bone. In the present study we made no attempt to verify the independence for cartilage and bone formation of stage-22 frontonasal mesenchyme from its epithelia. At present these exist no other studies on the timing of skeletogenic induction in the various craniofacial regions of the mammalian skeleton with the exception of Hall’s work on the mouse mandible [14]. Previous work on chick frontonasal mesenchyme [29,39] has revealed that this tissue displays a more complex relationship between the mesenchyme and adjacent epithelia with respect to the induction to form cartilage and bone. Since the main purpose of the present study was to establish a culture system of mammalian facial mesenchyme in which to analyze the mechanisms of skeletogenesis, further testing of the timing of epithelial dependence/independence of the various regions of craniofacial mesenchyme beyond that obtained for the mandible will be the subject of future studies. Nevertheless, the expression of cartilage and bone markers by these cells in a manner equivalent to that observed in mandibular cultures clearly suggests

that frontonasal mesenchyme from stage-22 embryos does not require epithelial interaction while the failure of maxillary and hyoid arch mesenchyme of the same stage to initiate any skeletogenesis within the same time period in our initial experiments suggests that epithelial independence for skeletogenesis does not occur at the same time over the entire facial region. Although facial mesenchyme from stage-22 frontonasal and mandibular processes of the rat show similar patterns of skeletal phenotypic expression, they differ with respect to the morphologies exhibited by the cultures and to a lesser extent with respect to the timing of differentiation. Just as has been found for the chick [22, 42, 431 the patterns of chondrogencic differentiation (as revealed by Alcian-blue staining) in micromass cultures of rat frontonasal and mandibular mesenchyme differ with respect to their morphology. Whereas mandibular cultures from either chick and rat differentiate as nodules (although fewer in number in the rat than the chick), those of the rat frontonasal differentiate as anastomosing columns of cells, which if grown long enough (> 9 days) will cover the entire culture while those of the chick frontonasal differentiate as one largely continuous sheet [22,42,43]. In all likelihood the initial differentiation of skeletogenic cells as anastomosing columns within rat frontonasal mesenchyme cultures was due to the presence of the neuron-like cells. At present, we do not know the origin of these cells nor whether they influence the differentiation of cartilage and bone cells within these cultures. To a lesser extent the differentiation of the rat frontonasal and mandibular mesenchyme differed with respect to the onset and progression of differentiation, with the mesenchyme of the mandible lagging behind that of the frontonasal by 2-3 days in general. Additionally, consistent differentiation within the mandibular cultures required a 1.5- to 2-fold increase in the number of cells plated despite there being no apparent difference between the two types of cultures in the amount of cell death. At present we d o not have an explanation for this difference. The onset of chondrogenic differentiation, as revealed by Alcian-blue staining, of the rat facial mesenchyme lags greatly behind that observed for chick [22, 42, 431 and mouse [27] limb-bud mesenchyme and rat limb-bud mesenchyme (Langille and Solursh, unpublished), which in all three cases demonstrates Alcianblue-positive staining after 24 h in culture compared with 3-5 days for rat facial cultures. As detailed above it is known that the rat mandibular and frontonasal mesenchymes from stage-22 embryos is no longer dependent on the epithelium in order to differentiate, based either on direct observations of CAM-grafted mesenchyme in the case of the mandible or indirect evidence of the ability of isolated frontonasal mesenchyme to express cartilage and bone phenotypes in micromass culture. Yet, it is not known whether or not the epithelium might have an effect on the rate or extent of facial mesenchyme differentiation in vitro, a question that we are at present still addressing.

205

In summary, we have reported the ability of stage-22 (12 1/2-day) rat mandibular and frontonasal mesenchyme to express both cartilage and bone phenotypes in micromass culture. Further study of differentiation within these cultures will allow analysis of the simultaneous induction and differentiation of both bone and cartilage within a population of cells derived from a single source (cranial neural crest [151). The cultures are equally valuable for the analysis of the mechanisms involved in mammalian craniofacial differentiation, patterning and growth, which can be compared with the relatively large amount of information available for craniofacial development of the chick. Acknowledgements. The authors wish to thank Karen Jensen and Dwight Moulton for their excellent technical assistance and we wish to thank all of our colleagues, Drs. Fisher, Franzen, Grant, Heinegard, Lian, Linsenmeyer and Rodan, who very kindly donated their antibodies, without which this work would not have been possible. This work was supported by grants HD05505 and DE05837 to MS and an NSERC of Canada Postdoctoral Fellowship and University Research Fellowship to RML. The MF-20 and BSP-I1 hybridoma supernatants were obtained from the Developmental Studies Hybridoma Bank maintained by the Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, M D and the Department of Biology, University of Iowa, Iowa City, IA, under contract N01HD-6-2915 from the NICHD.

References 1 . Ahrens PB, Solursh M, Reiter, RS (1977) Stage related capacity for limb chondrogenesis in cell culture. Dev Biol 60:69-82 2. Ayad S, Kwan APL, Grant ME (1987) Partial characterization of type X collagen from bovine growth-plate cartilage. FEB Lett 220:181-186 3. Bader D, Masaki T, Fischman DA (1982) Immunochemical analysis of myosin heavy chain during avian myogenesis in vivo and in vitro. J Cell Biol95 :763-770 4. Bates GP, Schor SL, Grant ME (1987) A comparison of the effects of different substrata on chondrocyte morphology and the synthesis of collagen types IX and X. In Vitro Cell Dev Biol23 :374-380 5. Bhaskar SN, Weinmann JP, Schour I (1953) Role of Meckel’s cartilage in the development and growth of the rat mandible. J Dent Res 32:398-410 6. Carazzi D (191 1) Eine neue Hamatoxylinlosung. Z Wiss Mikrosk (Leipzig) 28 :273 7. Christie GA (1964) Developmental stages in somite and postsomite rat embryos, based on external appearance, and including some features of the macroscopic development of the oral cavity. J Morphol 114:26>286 8. Copray JCVM, Johnson PM, Decker JD, Hall SH (1989) Presence of osteonectin/SPARC in mandibular condylar cartilage of the rat. J Anat 162:43-51 9. deBeer GR (1985) The development of the vertebrate skull. Reprint of the 1971 version. Oxford Univ Press, Oxford 10. Franzen A, Heinegard D (1985) Isolation and characterization of two sialoproteins present only in bone calcified matrix. Biochem J 232:715-724 1 1 . Franzen A, Heinegard D, Solursh M (1987) Evidence for sequential appearance of cartilage matrix proteins in developing mouse limbs and in cultures of mouse mesenchymal cells. Differentiation 36: 199-210 12. Guyer M F (1953) Animal Micrology. Univ Chicago Press, Chicago, p 236 13. Hall BK (1978) Grafting of organs and tissues to the chorioal-

lantoic membrane of the embryonic chick. TCA Manual 4: 881884 14. Hall BK (1980) Tissue interactions and the initiation of osteogenesis and chondrogenesis in the neural crest-derived mandibular skeleton of the embryonic mouse as seen in isolated murine tissues and in recombinations of murine and avian tissues. J Embryol Exp Morphol 58:251-264 15. Hall BK (1988) The Neural Crest, including a facsimilie reprint of the neural crest by Sven Horstadius. Oxford Science Publ, Oxford 16. Hanken J, Wassersug R (1981) The visible skeleton. Funct Phot 16:22-26,# 17. Hauschka PV (1984) Osteocalcin and its functional domains. In: Butler WT (ed) The Chemistry and Biology of Mineralized Tissues. EBSCO Media, Birmingham, AL 18. Hauschka PV, Reid ML (1978) Timed appearance of a calcium binding protein containing gamma-carboxyglutamic acid in developing chick bone. Dev Biol 65:426-434 19. Johnson GD, Araujo NC (1981) A simple method of reducing the fading of immunofluorescence during microscopy. J Immunol Meth 43 :349-350 20. Kelly WL, Bryden MM (1983) A modified differential stain for cartilage and bone in whole mount preparations of mammalian fetuses and small vertebrates. Stain Techno1 58: 131-134 21. Kwan APL, Freemont AJ, Grant ME (19x6) Immunoperoxidase localization of type X collagen in chick tibiae. Biosci Rep 6: 155-1 62 22. Langille RM, Paulsen DF, Solursh M (1989) Differential effects of physiological concentrations of retinoic acid in vitro on chondrogenesis and myogenesis in chick craniofacial mesenchyme. Differentiation 4 0 : 8 4 9 2 23. LeDouarin N (1982) The Neural Crest. Cambridge Univ Press, Cambridge 24. Maniatopoulos C, Sodek J, Melcher AH (1988) Bone formation in vitro by stromal cells obtained from bone marrow of young adult rats. Cell Tissue Res 254:317-330 25. Nichols DH (1986) Formation and distribution of neural crest mesenchyme to the first pharyngeal arch region of the mouse embryo. Am J Anat 176:221-231 26. Noden DR (1987) Interactions between cephalic neural crest and mesoderm populations. In: Maderson PFA (ed) Development and evolution of the neural crest. J Wiley and Sons, London, p 89-1 19 27. Owens EM, Solursh M (1981) In vitro histogenic capacities of limb mesenchyme from various stage mouse embryos. Dev Biol88:297-311 28. Sage H, Vernon RB, Decker J, Funk S, Iruela-Arispe ML (1989) Distribution of the calcium-binding protein SPARC in tissues of embryonic and adult mice. J Histochem Cytochem 37:819829 29. Schowing J (1968) Mise en evidence du r6le inducteur de I’enckphale dans I’osteogenesc du crine embryorinaire du poulet. J Embryol Exp Morphol 19:83-94 30. Scott JH (1953) The cartilage of the nasal septum. Br Dent J 95:3743 31. Sheehan DC, Hrapchak BB (1980) Theory and Practice of Histotechnology. The CV Mosby Co, Toronto 32. Silberman M, Reddi AH, Hand AR, Leaprnan R. von der Mark K, Franzen A (1987) Chondroid bone arises from mcscnchymal stem cells in organ culture of mandibular condyles. J Craniofac Genet Dev Biol7: 59-79 33. Smits-van Prooije AE, Vermeij-Keers C, Dubbledam JA, Mentink MMT, Poelmann RE (1987) The formation of mesoderm and mesectoderm in presomite rat embryos cultured in vitro using WGA-Au as a marker. Anat Embryol 176:71-77 34. Sodek J, Berkman FA (1987) Bone cell cultures. Meth Enzymol 1451303-324 35. Solursh M, Jensen KL, Reiter RS, Schmid TM, Linsenmeyer TF (1986) Environmental regulation of type X collagen production by cultures of limb mesenchyme, mesectoderm and sternal chondrocytes. Dev Biol 117 :9&101

206 36. Tan SS, Morriss-Kay GM (1985) The development and distri-

bution of the cranial neural crest in the rat embryo. Cell Tissue Res 240 :4 0 3 4 16 37. Tan SS, Morriss-Kay GM (1986) Analysis of cranial neural crest cell migration and early fates in postimplantation rat chimaeras. J Embryo1 Exp Morphol98:21-58 38. Termine JD, Kleinman HD, Whitson SW, Conn KM, McGarvey ML, Martin GR (1981) Osteonectin, a bone-specific protein binding mineral to collagen. Cell 26:9%105 39. Tyler MS (1983) Development of the frontal bone and cranial meninges in the embryonic chick: an experimental study of tissue interactions. Anat Rec 206:61-70 40. von der Mark K, von der Mark S, Gay S (1976) Study of

differential collagen synthesis during development of the chick embryo by immunofluorescence. I. Preparation of collagen type I and type I1 specific antibodies and their application to early stages of the chick embryo. Dev Biol48 :237-249 41. von der Mark K, von der Mark S (1977) Immunological and biochemical studies of collagen type transition during in vitro chondrogenesis of chick limb mesoderm cells. J Cell Biol 13:136141 42. Wedden SE, Lewin-Smith MR, Tickle C (1986) The pattern

of chondrogenesis of cells from facial primordia of chick embryos in micromass culture. Dev Biol 117:71-82 43. Wedden SE, Ralphs JR, Tickle C (1988) Pattern formation in the facial primordia. Development Suppl 103: 3 1 4