Springer-Verlag 1998

Cell Tissue Res (1998) 294:323±333

REGULAR ARTICLE

Andrea Pilloni ´ George W. Bernard

The effect of hyaluronan on mouse intramembranous osteogenesis in vitro

Received: 30 October 1997 / Accepted: 30 December 1997

Abstract Hyaluronan (HA) is an almost ubiquitous component of extracellular matrices. Early in embryogenesis mesenchymal cells migrate, proliferate and differentiate, in part, because of the influence of HA. Because many of the features of embryogenesis are revisited during wound repair, including bone fracture repair, this study was initiated to evaluate whether HA has an effect on calcification and bone formation in an in vitro system of osteogenesis. Enzyme-digested calvarial mesenchymal cells from 13-day-old mouse embryos were cultured in BGJb medium with rooster comb hyaluronan in seven different molecular weights (30, 40, 90, 160, 550, 660, and 1300 kDa). The dosages for each molecular weight were 0.5, 1.0 and 2.0 mg/ml. HA was added once to the medium at the plating of cells. After 10 days in culture, with low molecular weight hyaluronan (30 and 40 kDa) bone colonies were identifiable on a base of confluent fibroblasts. The number of colonies was larger than controls, particularly in the 1.0 and 2.0 mg/ml dosages of both 30 and 40 kDa of HA. Hyaluronan of high molecular weight, no matter what the dose, showed no significant bone colony formation, with apparent cell growth inhibition. Higher molecular weights were thereafter not included in this study. No statistically significant difference in the size of colonies was found when compared to controls in the 30 and 40 kDa bone colonies no matter what the dose. Key words Hyaluronan ´ Osteogenesis ´ Bone colonies ´ Calvaria ´ Mesenchyme ´ Mouse (Swiss Webster)

A. Pilloni University of Rome ªTor Vergataº, School of Dentistry, Section of Periodontics, Rome, Italy

)

G. W. Bernard ( ) University of California at Los Angeles, School of Dentistry, Section of Oral Biology, 10833, Le Conte Avenue, Los Angeles, CA 90024, USA

Introduction The extracellular matrix (ECM) plays an active and complex role in regulating the behavior of the cells that contact it, influencing their development, migration, proliferation, shape, and metabolic functions (Toole 1973). Polysaccharide glycosaminoglycans (GAGs), along with fibrous and adhesive proteins, such as collagen, elastin and fibronectin, respectively, make up the matrix (Yanagashita 1993). With the exception of hyaluronan (HA), GAGs contain sulfate groups and under physiological conditions, form a highly hydrated, gel-like ground substance in which the fibrous proteins are embedded (Hardingham and Fosang1992). HA, during the past years, has received increasing interest from biologists and clinicians because of its well known participation in the early events of tissue formation and repair. It is an almost ubiquitous component of the extracellular matrices discovered in bovine vitreous by Meyer and Palmer in 1934. It is a relatively simple molecule, built from a repeating sequence of nonsulfated disaccharide units consisting of d-glucuronic acid and N-acetyl-d-glucosamine linked by b(1®3) and b(1®4) glycosidic linkages, respectively. It contains up to several thousand sugar residues. HA has by far the highest molecular weight of the GAGs (it ranges from 104 to 107) and is thought to facilitate cell migration during tissue morphogenesis and repair. Since the early 1970s, with the works of Kvist and Finnegan (1970) and of Toole and Gross (1971), HA has been found in variable amounts in all tissues and fluids of adult animals and is especially abundant in early embryos. Because of its simplicity, HA is thought to represent the earliest evolutionary form of GAG. Toole et al. in 1972 showed the correlation between the production of HA and mesenchymal cell movement, on the one hand, and HA destruction and cell differentiation on the other. More recently HA has been found to play an important role in wound healing, by sharing certain common features with the early events during embryonic development of many organs (Hay 1980; Ruggiero et al. 1987; Bertolami 1984; Bertolami et al. 1988; Bertolami and Messadi

324

1994). Numerous data also show correlation between HA and cell adhesion mediated by HA-binding domains (Goldstein et al. 1989, Stamenkovic et al. 1989). Moreover, HA seems to promote cell migration and proliferation while inhibiting cell differentiation (Knudson et al. 1989; Toole 1991) and there is evidence of a structural involvement with growth factors (Bertolami and Messadi 1994). Hyaluronan in calcified tissues Hjertquist and Vejlens in 1968 showed that HA is present in long bones and amounts to 3% of the total glycosaminoglycans. There is general agreement that chondroitin 4sulfate proteoglycan is the major proteoglycan in all human bones, cementum and dentin (Engfeldt and Hjerpe 1976; Prince et al. 1983; Franzen and Heinegard 1984; Sato et al. 1985; Bartold et al. 1988; Waddington et al. 1989). Studies on the synthesis, distribution, and degradation of HA have demonstrated that this component of the ECM plays a vital role in the initiation of cellular condensations preceding the differentiation of mesenchymal cells to cartilage (Knudson and Toole 1985). Precisely how HA exerts this action has not been determined, but that it has an action is now well illustrated in a number of skeletal systems. HA forms the backbone to which proteochondroitin sulfate is attached, thus forming large aggregates within the extracellular cartilage matrix (Wiebkin and Muir 1973; Motoky and Mulliken 1990; Wornom and Buchman 1992). Hyaluronan in calcification In 1904, Pfaundler coined the term Kalksalzfänger (lime salt catcher) for the calcium-concentrating properties of an uncertain molecule in bone formation and hypothesized that calcium binding was a crucial step in the calcification process. A role for GAGs and proteoglycans in biological mineralization has been postulated since the early 1960s (Campo and Dziewiat 1963). Yet Cuervo et al. in 1973 found that sulfated proteoglycans in the form of aggregates have a potent inhibitory effect on mineral growth when tested in vitro. They concluded that such data were consistent with the hypothesis that shielding or sequestration of ªearly mineral embryosº occurs within the radiating branches of the aggregate proteoglycan molecules. Mineral growth is thereby prevented when the aggregate form of proteoglycans is present in sufficient concentration to build up such shielding forces. In 1976 Howell and Pita found a strong relationship between the decrease of proteoglycan concentration and size and the transition from resting cartilage to the cartilage calcified zone of the growth plate. Chen and Boskey in 1985 emphasized the importance of proteoglycans in the mechanism of calcification, by reporting that charge density plays a significant role in determining their ability to inhibit hydroxyapatite formation both in vitro and in vivo,

leading Boskey and Dick (1991) to report that hyaluronan interactions with hydroxyapatite do not interfere with crystal proliferation and growth. In fact, Boskey et al. (1991) stated that undersulfated proteoglycans such as HA created an environment in which mineralization occurred too rapidly, suggesting the importance of sulfate content of proteoglycan for the regulation of mineral formation, and HA for its initiation and proliferation. These reports, along with the predominant role played by HA in morphogenesis, cell migration, differentiation, and adhesion, and more importantly as the only nonsulfated proteoglycan of the ECM, lend support to the objectives of the present study. It is the main purpose of this study to determine whether HA plays a role in osteogenesis, using a model of intramembranous bone formation in vitro. Because different molecular weights of HA could have different physiological effects, a second objective of this study was to determine whether diverse molecular weights and different dosages would affect osteogenesis. To this end, HA was added to known numbers of collagenase-dispersed calvarial mesenchymal cells, using different dosages and molecular weights (mw) that varied between 30 and 1300 kDa. These cells, after adhesion to Petri dishes, differentiated into osteoblasts and fibroblasts, which developed into bone colonies. These colonies were analyzed for number and size.

Materials and methods Cell culture system of isolated mesenchymal cells The osteogenic tissue culture system, first developed in 1977 by Marvaso and Bernard, began with enzyme-digested mouse calvarial mesenchymal cells. Swiss Webster mice, 12- to 13-day time-pregnant (Simonsen), were anesthetized with 0.4 ml of sodium Nembutal given intraperitoneally. Fetuses were aseptically removed one at a time, the epithelium removed from the head and the calvariae dissected away from the brain with iridectomy scissors. Portions of the coronal ends of temporal, frontal, parietal and occipital mesenchyme were removed and immediately placed into BGJb medium (Gibco Laboratories, Life Technologies, Grand Island, NY). Pooled tissue obtained from several mice was then transferred into a sterile 60-mm plastic Petri dish (Fisher Scientific) containing 12 ml sterile Hanks' balanced salt solution (HBSS) without calcium or magnesium. By means of an iridectomy scissors the mesenchymal tissue was cut in small pieces (approximately 1 mm”1 mm2) and then transferred into a test tube with 20 mg crude collagenase, and 50 mg dextrose. The Petri dish was then placed on a rotating platform for 90 min at 37C. The resulting enzymatic dispersion of cells was transferred to a 50-ml sterile, conical centrifuge tube. The cells were then centrifuged for 10 min, the supenatant removed and the resulting pellet resuspended in BGJb media supplemented with 0.1% (v/v) of 29.2 mg/ml solution of l-glutamine, 0.1% (v/v) of a 10 000 U/ml solution of penicillin-streptomycin, 0.1% (v/v) of a 250 mg/ml solution of Fungizone (Gibco Laboratories) and 10% (virus-screened) heat-inactivated fetal calf serum. This procedure was repeated three times to ensure complete removal of collagenase. After the final resuspension, an aliquot was removed and transferred to a haemocytometer for counting. The cell suspension was then appropriately diluted or concentrated to 4”105 cells / ml. Aliquots of 4 ml each containing 1.6”106 cells were plated on 60-mm sterile plastic Petri dishes and grown at 37C, 100% humidity in an atmosphere of 95% air and 5% CO2 in a Forma 3028 incubator (Forma Scientific, Marietta, Ohio) for 10 days. To ensure that undifferentiated, un-

325

A

B Fig. 1 A Fibroblasts growing in vitro for 2±3 days, control, phase contrast. ”1000. B Fibroblasts growing to confluence in vitro for 2±3 days with low molecular weight hyaluronan, phase contrast. ”1000

calcified mesenchyme was obtained from the system, tissue taken from litter mates of each group of experimental fetuses was fixed at the time of surgery and later processed for light and electron microscopy. If any one of the semi-serial sections had revealed any differentiation or organization, the complete series of experiments derived from that tissue would have been discarded. HA was added to the tissue cultures, at different dosages and different molecular weights. Rooster comb purified powdered HA was used (FIDIA Advanced Biotechnologies, Abano Terme, Italy) and the following mo-

326 Fig. 2 Mineralized bone after 10 days in situ, von Kossa stain, phase contrast. ”400

lecular weights were tested: 30, 40, 90, 160, 550, 660 and 1300 kDa. Three different dosages were then selected for this study: 0.5, 1.0 and 2.0 mg/ml. Each trial consisted of a particular molecular weight, tested in the three different dosages and one control. A duplicate was obtained per trial, having a total of eight Petri dishes per MW tested. To guarantee statistical significance, based upon proper cellular density (Hayflick and Moorhead 1961), each dish was plated with 1.6”106 cells at the time of plating. To provide enough cells for each of the eight plates, six pregnant mice were necessary per trial. A final number of 42 Swiss Webster time-pregnant mice was used for the study and 56 Petri dishes were analyzed. The HA was added once only during the first plating of cells in the first medium. Medium was changed the first time after 48 h but HA was not added again. Thereafter, cells were fed with HA-free media every 24 h until day 10, when the resulting tissues were fixed. Microscopy Bone colony formation was monitored daily by phase-contrast microscopy (Zeiss Axiovert 35). Once embedded in Epon, bone colonies, number and size, were evaluated with a computerized digital analyzer (Summagraphics Bitpad). As in the method previously de-

scribed in 1977 by Marvaso and Bernard, bone cultures were fixed in situ in a solution of 2.5% glutaraldehyde in 0.1 M sodium cacodylate, buffered at pH 7.4 for 2 h, then washed with 0.1 M sodium cacodylate twice, postfixed for 1 h in 1% OsO4 in 0.1 M sodium cacodylate buffer, dehydrated in graded alcohols and embedded in Epon 812. Each duplicate was treated with von Kossa reagents for the determination of bound calcium. The Petri dish plastic was then removed from the embedded material and blocks were cut with glass knives on a Sorvall Porter-BlumH MT-l or MT-2 ultramicrotome (Sorvall, Newton, Conn.). Semi-thin sections (1 mm) were taken throughout the entire length of each embedded bone colony, collected on pre-cleaned glass slides (Becton Dickinson Labware, Oxnard, Calif.). Sections were stained with 1% aqueous toluidine blue in 1% borax. Photomicrographs were taken of the most representative fields with an Olympus BH-2 light microscope. Thin sections stained with uranyl acetate and lead citrate, were used for ultrastructural examinations. Transmission electron microscopy (TEM) was done with a Siemens lA Elmiskop operating at 80 kV. Within the cell culture tissue only the dense, osmiophilic clumps that represent cellular aggregates of bone colonies were examined. At least three randomly selected blocks from each Petri dish, embedded at each time point, were viewed by TEM.

327 Statistical analysis The study looked at mean number and mean size of each colony. The raw data was analyzed, using a two-way analysis of variance (ANOVA), a procedure where dosage (including control=zero dose) and molecular weight are the two factors. It also considered comparisons across dosages, with a hypothesis of forming a dose response curve for each molecular weight.

Results Phase-contrast microscopy Following 2±3 days as single cell suspensions, mesenchymal cells attached to the bottom of the Petri dish and differentiated into fibroblastic and osteoblastic lines. Fibroblastic growth showed typical sparse cellular density in the middle of the Petri dish in control dishes (Fig. 1B). In the presence of HA, cellular density was increased significantly all over the Petri dish (Fig. 1A). In general, with low molecular weight HA, particularly in the 1.0 and 2.0 mg/ml doses of 30 kDa and 40 kDa, cells became confluent in 72 h and bone colonies in the experimental dishes after 10 days showed a distinct von Kossa stained calcification pattern. (Fig. 2). The number of colonies was enhanced with the addition of 30 kDa and 40 kDa HA compared to the controls and it was visually evident immediately after von Kossa staining (Figs. 3, 8A,B). However, colony formation was almost insignificant at the higher molecular weights (550, 660 and 1300 kDa) when compared to controls, no matter what the dose. Since there were very few colonies, their size could not be measured with any significance. With the 30 kDa and 40 kDa HA there was no statistically significant difference in the size of colonies when compared to control dishes (Fig. 8C,D). Microscopy At day 10, osteogenic elements amongst a well-defined collagen matrix were easily identifiable (Fig. 4). Collagen fibrils and calcification of the matrix were noticed in close approximation to osteoblasts, and osteocytes were seen within the calcified matrix (Fig. 5). All the events that are characteristic of the initiation of intramembranous bone calcification were identified (Figs. 4±7) exactly as seen in vivo: (1) differentiation of mesenchymal cells into osteoblasts (Figs. 4, 5); (2) subsequent appearance of matrix vesicles (the initial calcification loci) in the extracellular space (Figs. 6, 7); (3) crystallization of hydroxyapatite within and about these vesicles (Figs. 6, 7); (4) growth of hydroxyapatite crystals into spheroidal or cylindrical nodules of bone (Fig. 7); and (5) fusion of these nodules into seams of woven or embryonic bone. (Figs. 4, 5, 7).

A

B

C Fig. 3 A Bone colonies after 10 days in vitro, in situ. Von Kossa stain, control. ”1. B Bone colonies, after 10 days in vitro, in situ. Von Kossa stain, with 1 mg/1 ml 40 kDa hyaluronan. ”1. C Bone colonies, in vitro. Von Kossa stain with 1 mg/1 ml 30 kDa hyaluronan. ”1

Statistical analysis Analysis of variance (ANOVA) methods were used to compare mean colony number and mean colony size among the four dose levels (control, 0.5 mg/ml, 10 mg/ ml, 2.0 mg/ml) and the two molecular weights that showed statistical significance (30 and 40 kDa). Mean and median values were similar, implying that an ANOVA was appropriate. For colony number (Fig. 8A,B), there is a roughly linear increasing trend with dose. While the effect of dose is significant (F 3,24=19,68, P<0.0001) this trend is not significantly different in 30 kDa versus 40 kDa (F for comparing 30 kDa

328 Fig. 4 Bone areas after 10 days in vitro, developing within a larger bone colony, as seen in Fig. 3A and 3B (cM calcified matrix, c collagen, oB osteoblast, M mesenchymal cell). Toluidine blue. ”200

to 40 kDa is F 1,24=0.03, P=0.85; for molecular weight main effect; F 3,24=1.43, P=0.26, for possible molecular weight and dose interaction. However, an analysis of size by plate showed that there is substantial variability between plates. Assuming that the 32 combinations of 30 and 40 kDa represent 32 independent plates, an analysis based on these 32 plates fails to show any statistical significance. That is, the difference in mean size at the various doses could be due to plate-to-plate variability alone (F for any molecular weight or dose effect, F 7,24=1.03, P=0.4346; Fig. 8C,D). In summary, the analysis of the data shows that when low molecular weight HA was added (namely, 30 and 40 kDa), no matter what the dose, there was an increase in number of colonies, but not in size, compared to controls.

Discussion During the past decade an increasing amount of studies have emphasized the concept that growth, form, and function of cells are modulated by highly specific interactions between the cells and their extracellular matrices and by polypeptide growth factors. Most of such research suggests that specific cellular interactions are important not only for cell growth but also for its morphology and function (Toole 1973). These studies show that extracellular matrices are cell- and tissue-specific, and progress has been made in characterizing these components and their specific functions (Toole 1991). This is also the case of calcified tissues, because: ªMineralized tissues are not simply the end result of precipitations of inorganic crystals from supersaturated solution in microenvironments that happen to be composed of organic molecules. They are, in fact, organic matrix-induced processes under the

329

Fig. 5 Woven bone seams after 10 days in vitro. Toluidine blue. ”200

constant control of living cells¼º (Fisher and Termine 1985). Hyaluronan possesses biochemical and physical properties suitable to perform an important role in the early events of osteogenesis as well as in many other tissues. Ever since the early 1950s it has been well established that considerable HA is synthesized in the early stages of callus formation during the repair of fractured long bones (Maurer and Hudack 1952). HA is a prominent extracellular matrix component during bone morphogenesis (Toole and Gross 1971) and large amounts of HA are present during the transition from mesenchymal cells to cartilage (Wiebkin et al. 1973; Handley and Lowther 1976; Kresse et al. 1994). Iwata and Urist in 1973 showed that when allogenic demineralized cortical bone matrix is implanted intramuscularly into rats, peak levels of hyaluronan are found within the matrix 2 days after implantation and, as cytodifferentiation ensues, the level of hyaluronan within the matrix declines and in 12 days bone marrow formation occurs. This is essentially what is expected to be found also in the intramembranous model of osteogenesis. During an early stage of intramembranous osteogenesis (as in endochondral) when only undifferentiated mesenchymal cells

are found, HA reaches peak levels. In terms of its correlations with wound healing mechanisms and hard tissue development (Hay 1980) HA can be thought as a ªprimerº in cell regeneration. A well-defined sequence of events, already postulated by Weigel et al. (1986) clarified this and seems to match the outcomes of our study. When calvaria-mesenchymal cells were separated and well dispersed at the time of the first plating, HA was added to the first medium, which was maintained for 48 h. By virtue of its molecular characteristics, such as the ability to occupy large volumes of the ECM with little mass, hyaluronan functions as a space maintainer. Brian Toole, in 1991, showed that when excess HA surrounds cells from all sides it prevents them from aggregating. Under these conditions, the highly replicating cellular elements, such as calvarial mesenchymal cells during early stages of tissue formation were given the necessary extracellular space for the production of the daughter elements. Moreover, as previously mentioned, many studies have suggested a role for HA in cell migration. They all postulate different ways by which HA might exert an effect and they all relate to the need for a translocating cell to detach from its substratum, at least partially, as well as to re-attach. Whereas the ªfootpadsº deposited by attaching cells are enriched in heparan sulfate, those deposited by migrating cells contain more HA and undersulfated chondroitin sulfate (Rollins and Culp 1979). In the absence of HA, mesenchymal cells, once plated, attach to the bottom of the Petri dish and initiate the process of condensation, which is a critical step in cellular differentiation. What we call a colony is, in fact an aggregation, collection or condensation of mesenchymal cells and their progeny that are distinguished from the fibroblast cells that surround the colony. Therefore, such cells once condensed, are already limited in their expansive potential by the cell contact inhibition phenomena. No further space can be occupied by daughter cells, not only because of the early attachment of the progenitor cells, but also because the latter are reciprocally inhibited by contact. As soon as condensation occurs colony formation begins. On the other hand, in the presence of HA, primordial elements are given more time before they can attach to the bottom of the dish. In such case they can fully express their replicating potentials. If this is mostly due to excess HA in the medium itself or to an increased time of available growth factors within the extracellular matrix was not assessed by our study. The HA is then removed (48 h later) and so it is presumed that it is present during the attachment of cells to the surface of the dish. However, at this point in time, cellular elements are well distributed over the entire area of the plate. An increase in number of bone colonies is now expected to occur, leading to the final increased colony number when compared to controls. The size of each colony is the same for experimental and control dishes. However, the quality of bone colonies does not change with or without HA. After 48 h, the first medium is changed and replaced with a HA-free medium. This removal of HA is what presumably would have naturally occurred, in vivo, when hy-

330 Fig. 6 Mineralizing extracellular matrix (MV matrix vesicles, OCF oriented collagen fibers). Uranyl and lead. ”20 000

aluronidase would have been secreted by the first fibroblasts formed and by the still present mesenchymal cells in the process of differentiation. Cellular elements have now the necessary amount of HA for the formation of the HA-dependent cross-bridging. This concept is derived from recent findings on HA in cell regulation. In fact, Toole (1991) postulated the presence of large HA-dependent, pericellular coats during the initial formation of cartilage in early development of the limb. We also propose the above-mentioned multivalent cross-bridging by HA of HA receptor sites on adjacent cells as a mechanism for cell aggregation. The two mechanisms together, pericellular coats and cross-bridging, lead to the formulation of the following hypothesis. Soon after the removal of high levels of exogenous HA from the medium, cell coats are created that maintain intercellular space and inhibit cell adhesion, so that the cells can keep migrating and dividing. Once this is accomplished, the degradation of HA by hyaluronidase allows the aggregation phenomenon of multivalent cross-bridging by HA to HA receptor sites to occur. Fig. 7 Mineralizing stain of woven bone (BN bone nodules, WB woven bone, OCF oriented collagen fibers). Uranyl and lead. ”20 000

331

30 25 20 15 10 5 0

Control

A

0.5

1

Mean size of bone colonies (mm2)

0.7 0.6 0.5 0.4 0.3 0.2 0.1

C

15 10 5

1

Control

Fig. 8 Mean number (A, B) and size (C, D) of bone colonies at increasing doses of 30 kDa (A, C) and 40 kDa (B, D) hyaluronan

Weigel's hypothesis that matrix rich in HA is laid down in a cell-poor space, which in turn stimulates mesenchymal cell migration differentiation and enzyme secretion such as hyaluronidase, seems to fit with our outcomes. Cells in the process of replicating themselves need space. They have to avoid cell-to-cell contact inhibition. As far as cell-mediated events are concerned, a number of cell surface binding sites for HA have been reported on endothelial cells and fibroblasts (Eriksson et al. 1983; Raja and Weigel 1985; Banerjee and Toole 1992). In this regard, there is increasing evidence for the role of HA in angiogenesis (West and Kumar 1989), another crucial event in bone formation in vivo. Moreover, proteoglycans and growth factor activity have been correlated by several authors (Rouslahti 1989; Rouslahti and Yamaguchi 1991). Binding of growth factors to proteoglycans engenders formation of localized tissuebound reservoirs of growth factors. Bertolami and Messadi (1994) state that apart from their possible role

1

2

40 kDa Hyaluronan

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

2

Dose in mg/ml HA (p < 0.4346)

0.5

Dose in mg/ml HA (p < 0.0001)

1.0

0.8

0.5

20

B

30 kDa Hyaluronan

Control

25

0

0.9

0

40 kDa Hyaluronan

30

2

Dose in mg/ml HA (p < 0.0001)

1.0

Mean size of bone colonies (mm2)

35

30 kDa Hyaluronan Mean number of bone colonies

Mean number of bone colonies

35

D

Control

0.5

1

2

Dose in mg/ml HA (p < 0.4346)

as storehouses for growth factors, proteoglycans, such as HA, may interact with such factors to generate direct and specific cellular effects. More importantly, among the growth factor-HA interactions reported to be important in bone repair, platelet-derived growth factor (PDGF) and especially the TGF-b superfamily (with all the different bone morphogenetic proteins or BMPs) have been found. In fact, decorin, a small proteoglycan in bone that binds reversibly with TGF-b, may act as a reservoir for the growth factor within the tissue (Yamaguchi et al. 1990; Couchman and Woods 1993). Recent data, show that TGFb stimulates HA-dependent coat formation and HA synthesis (Munaim et al. 1991) and that basic fibroblast growth factor (FGF) is highest in precondensation stage limb buds, when HA synthesis and cell proliferation are maximal (Munaim et al. 1988; Seed et al. 1988). One explanation is that polysaccharide chains are too inflexible to fold up into the compact globular structures that polypeptide chains typically form. Moreover, they are strongly hydrophilic. Thus, GAGs (such as HA) tend to adopt highly extended, so-called random-coil conformations, which occupy a huge volume relative to their mass and form gels even at very low concentrations. Because of

332

this porous hydrated gel form there is rapid diffusion of water-soluble molecules such as growth factors and the migration of cells. This applies also to calcifying matrices, where calcium and phosphate ions have to be able to diffuse from site to site before reaching the suitable site for crystallization into hydroxyapatite. HA seems to do so through the mechanism of so-called ionic exclusion. Finally, as previously mentioned, Boskey et al. (1991) report on the possible role of HA in the mineralization process. Their study on brachymorphic mice (bm/bm) with an inherited undersulfated chondroitin sulfate molecule shows that HA affects the organization of the matrix and creates an environment in which mineralization occurs too rapidly and in a disorganized, random fashion. Moreover, crystals appear larger or are more perfect. The same mechanism might occur with HA in normal mineralization because it is the only non-sulfated proteoglycan, with chondroitum sulfate acting as a modulator of calcification. In conclusion, it is important to keep in mind that HA has a time-specific mode of action. As previously mentioned, once its task is accomplished, HA normally leaves the matrix, allowing cells to congregate. It is probably for this reason that the higher molecular weight HAs have shown cell growth inhibition. Even after removal from the first medium, the high molecular weight HA, which assumes in solution a denser gel-like structure, can still be present as ªexcessº both as matrix component and cell coating. It is therefore of paramount significance to remember that HA, under specific conditions, can either be a promoter molecule as well as a potent inhibitor. When Toole in 1971 reported on the inhibitory effects of HA in cartilage formation in vitro, there was no mention about either the molecular weight nor about the amount of time the HA was kept in the culture medium. The growing knowledge about the potential of hyaluronan clearly demonstrates that such a molecule is essential during the early stages of any newly forming or regenerating tissue. However, it is also evident that in such tissues, once their formation is complete, the amount of HA present becomes insignificant (e.g., cementum and dentin). The data from this research has indicated that using proper conditions and when properly applied to specific cellular elements low molecular weight HA fully expresses the osteogenic potential of mesenchymal cells through the subsequent differentiation and proliferation of osteoprogenitor cells as a function of its open spatial organization. This leads to an enhancement of bone colony formation in vitro.

References Banerjee SD, Toole BP (1992) Hyaluronan-binding protein in endothelial cell morphogenesis. J Cell Biol 119:643±652 Bartold PM, Miki I, Mcallister B, Narayanan AS, Page RC (1988) Glycosaminoglycans of human cementum. J Periodont Res 23:13±17

Bertolami CN (1984) Glycosaminoglycan interactions in early wound repair. In: Hunt TK, Heppenstall RB, Pimes E, et al (eds) Soft and hard tissue repair: biological and clinical aspects. Praeger, London, p 67 Bertolami CN, Messadi DV (1994) The role of proteoglycans in hard and soft tissue repair. Crit Rev Oral Biol 5:311±337 Bertolami CN, Argenta JG, Siebert EP, Bronson RE (1988) Distinctive fibroblastic subpopulations in skin and oral mucosa demonstrated by differences in glycosaminoglycan content. In Vitro Cell Dev Biol 24:1121±1126 Boskey AL, Dick P (1991) Hyaluronan interactions with hydroxyapatite do not alter in vivo hydroxyapatite cristal proliferation and growth. Matrix 11:442±446 Boskey AL, Maresca M, Wikstrom B, Hjerpe A (1991) Hydroxyapatite formation in the presence of proteoglycans of reduced sulfate content: studies in the brachymorphic mouse. Calcif Tissue Int 49:389±393 Brunette DM (1988) The effect of surface topography on cell migration and adhesion. In: Ratner BD (ed) Surface characterization of biomaterials. (Progress in biomedical engineering). Elsevier, New York Campo RD, Dziewiat DD (1963) Turnover of the organic matrix of cartilage and bone as visualized by autoradiography. J Cell Biol 18:19 Chen CC, Boskey AL (1985) Mechanisms of proteoglycan inhibition of hydroxyapatite growth. Calcif Tissue Int 37:395±400 Couchman JR, Woods A (1993) Structure and biology of pericellular proteoglycans. In: Roberts DD, Mecham RP (eds) Cell surface and extracellular glycoconjugates. Academic Press, Orlando, Fla, pp 33±82 Cuervo LA, Pita JC, Howell DS (1973) Inhibition of calcium phosphate mineral growth by proteoglycan aggregate fractions in a synthetic lymph. Calcif Tissue Res 13:1±10 Engfeldt B, Hjerpe A (1976) Glycosaminoglycans and proteoglycans of human bone tissue at different stages of mineralization. Acta Pathol Micro Biol Scand 84:95±106 Eriksson S, Fraser JRE, Laurent TZ, Pertoft H, Smedsrod B (1983) Endothelial cells are a site of uptake and degradation of hyaluronic acid in the liver. Exp Cell Res 144:223±231 Fisher LW, Termine JD (1985) Noncollagenous proteins influencing the local mechanisms of calcification. Clin Orthop Rel Res 200:362±385 Franzen A, Heinegard D (1984) Characterization of proteoglycans from the calcified matrix of bovine bone. Biochem J 224:59±66 Goldstein LA, Zhou DFH, Picker LJ, Minty CN, Bargatze RF, Ding JF, Butcher EL (1989) A human lymphocyte homing receptor, the hermes antigen, is related to cartilage proteoglycan core and link proteins. Cell 56:1063±1072 Handley CJ, Lowther DA (1976) Inhibition of proteoglycan biosynthesis by hyaluronic acid in chondrocytes in culture. Biochim Biophys Acta 444:69±74 Hardingham TE, Fosang AJ (1992) Proteoglycans: many forms and many functions. FASEB J 6:861±870 Hardingham TE, Muir H (1972) The specific interaction of hyaluronic acid with cartilage proteoglycans. Biochim Biophys Acta 279:401±405 Hay ED (1980) Development of the vertebrate cornea. Int Rev Cytol 63:263±291 Hayflick L, Moorhead PS (1961) Serial cultivation of human diploid cell strains. Exp Cell Res 25:585±621 Hjertquist SO, Vejlens L (1968) The glycosaminoglycans of dog compact bone and epiphyseal cartilage in the normal state and in hyperparathyroidism. Calcif Tissue Res 2:314±333 Howell DS, Pita JC (1976) Calcification of growth plate cartilage with special reference to studies on micropuncture fluids. Clin Orthop 118:208±229 Huggins CB, Reddi AH (1975) The formation of bone marrow in fibroblast-transformation ossicles. Proc Natl Acad Sci USA 72:2212±2220 Iwata H, Urist MR (1973) Hyaluronic acid production and removal during bone morphogenesis in implants of bone matrix in rats. Clin Orthop Rel Res 90:236±245

333 Ko JS, Bernard GW (1981) Osteoclast formation in vitro from bone marrow mononuclear cells in osteoclast-free bone. Am J Anat 161:415±425 Knudson CB, Toole BP (1985) Changes in the pericellular matrix during differentiation of limb bud mesoderm. Dev Biol 112: 308±318 Knudson WC, Biswas C, Li XQ, Nemec RE, Toole BP (1989) The role and regulation of tumor-associated hyaluronan. In: Evered D, Whelan J (eds) The biology of hyaluronan. Wiley, Chichester, pp 138±149 Ko JS, Bernard GW (1981) Osteoclast formation in vitro from bone marrow mononuclear cells in osteoclast-free bone. Am J Anat 161:415±425 Kresse H, Hausser H, Shonherr E (1994) Small proteoglycans. In: Jolles P (ed) Proteoglycans. Birkhäuser, Basel, pp 73±100 Kvist TN, Finnegan CU (1970) Distribution of glycosaminoglycans in axil region of developing chick embryo. I. Histochemical analysis. J Exp Zool 175:221±240 Laurent TC (1969) . In: Gerritsen T (ed) Modern separation methods of macromolecules and particles. Wiley, New York, p 199 Marvaso V, Bernard GW (1977) Initial intramembranous osteogenesis in vitro. Am J Anat 149:453±468 Maurer PH, Hudack SS (1952) The isolation of hyaluronic acid from callus tissue of early healing. Arch Biochem Biophys 38:49±53 Meyer K, Palmer JW (1934) The polysaccharide of the vitreous humor. J Biol Chem 107:629±634 Motoki DS, Mulliken JB (1990) The healing of bone and cartilage. Clin Plast Surg 17:527±544 Munaim SI, Klagsbrun M, Toole BP (1988) Developmental changes in fibroblast growth factor in the chicken embryo limb bud. Proc Natl Acad Sci USA 85:8091±8093 Munaim SI, Klagsbrun M, Toole BP (1991) Hyaluronan-dependent pericellular coats of chick embryo limb mesodermal cells: induction by basic fibroblast growth factor. Dev Biol 143:297±302 Paschalakis P, Vynios DH, Tsiganos CP, Dalas E, Maniatis C, Koutsoukos PG (1993) Effect of proteglycans on hydroxyapatite growth in vitro: the role of hyaluronan. Biochim Biophys Acta 1158:129±136 Pfaundler M (1904) Über die Elemente der Gewebsverkalkung und ihre Beziehung zur Rachitisfrage. Jahrb Kinderheilk 60:123±133 Prince CWF, Rahemtulla F, Butler WT (1983) Metabolism of rat bone proteoglycans in vivo. Bio Chem J 216:589±596 Raja RH, Weigel PH (1985) Binding and internalization of hyaluronate by isolated rat liver cells. J Cell Biol 101:426±430 Rollins BJ, Culp LA (1979) Glycosaminoglycans in the substrate adhesion sites of normal and virus-transformed murine cells. Biochemistry 18:141±148 Ross R, Greenlee TK Jr (1966) Electron microscopic attachment sites between connective tissue cells. Science 153:997±999 Rouslahti E (1989) Proteoglycans in cell regulation. J Biol Chem 264:13369±13372 Rouslahti E, Engvall E (1980) Complexing of fibronectin glycosaminoglycans and collagen. Biochim Biophys Acta 361:350 Rouslahti E, Yamaguchi Y (1991) Proteoglycans as modulators of growth factor activities. Cell 64:867±869 Ruggiero SL, Bertolami CN, Bronson RE, Damiani PJ (1987) Hyaluronidase activity of rabbit skin wound granulation tissue fibroblasts. J Dent Res 66:1283±1287

Sato SF, Rahetmulla F, Prince CW, Tomana M, Butler WT (1985) Proteoglycans of adult bovine compact bone. Connect Tissue Res 14:65±75 Seed J, Okwin BB, Hauschka SD (1988) Fibroblast growth factor levels in the whole embryo and limb bud during chick development. Dev Biol 128:50±57 Stamenkovic I, Amiot M, Pesando JM, Seed B (1989) A lymphocyte molecule implicated in lymph node homing is a member of the cartilage link protein family. Cell 56:1057±1062 Toole BP (1973) Hyaluronate and hyaluronidase in morphogenesis and differentiation. Am Zool 13:1061±1065 Toole BP (1991) Proteoglycans and hyaluronan in morphogenesis and diffferentiation. In: Hay ED (ed) Cell biology of extracellular matrix. Plenum Press, New York Toole BP, Gross J (1971) The extracellular matrix of the regenerating newt limb: synthesis and removal of hyaluronate prior to differentiation. Dev Biol 26:28±35 Toole BP, Jackson G, Gross J (1972) Hyaluronate in morphogenesis in vitro. Proc Natl Acad Sci USA 69:1384 Underhill CB, Toole BP (1979) Binding of hyaluronate to the surface of cultured cells. Cell Biol 82:475±482 Urist MR (1965) Bone; formation by autoinduction. Science 159:893±899 Urist MR (1976) Biochemistry of calcification. In: Bourne GH (ed) The biochemistry and physiology of bone. Academic Press, New York, pp 1±59 Vlodavsky I, Folkman J, Sullivan R, Fridman R, Ishaimic R, Sasse J, Klagsbru M (1987) Endothyelial cell-derived basic fibroblast growth factor: synthesis and deposition into subendothelial extracellular matrix. Proc Natl Acad Sci USA 84:2292 Waddington RJ, Embery G, Last KS (1989) Glycosaminoglycans of human alveolar bone. Arch Oral Biol 34:587±589 Weigel PH, Fuller GM, Le Boeuf RD (1986) A model for the role of the hyaluronic acid and fibrin in the early events during the inflammatory response and wound healing. J Theor Biol 119:219± 234 West DC, Kumar S (1989) Hyaluronan and angiogenesis. The biology of hyaluronan. (Ciba Foundation Symposium 143) Wiley, Chichester, pp 187±207 Wiebkin OW, Muir H (1973) Factors affecting the biosynthesis of sulphated glycosaminoglycans by chondrocytes in short term maintenance culture isolated from adult tissue. In: Kulonin E, Pikkarainen J (eds) Biology of fibroblast. Academic Press, New York, pp 231±236 Wiebkin O, Hardingham TE, Muir H (1973) Inhibition of sulfate incorporation in isolated adult chondrocytes by hyaluronic acid. FEBS Lett 37:42±46 Wornom IL, Buchman SR (1992) Bone and cartilaginous tissue. In: Cohen K, Diegelmann RF, Lindblat WJ (eds) Wound healing: biochemical and clinical aspects. Saunders, Philadelphia, pp 356±383 Yamaguchi Y, Mann DM, Rouslahti E (1990) Negative regulation of transforming growth factor-b by the proteoglycan decorin. Nature 346:281±284 Yanagashita M (1993) Function of proteoglycans in extracellular matrix. Acta Pathol Jpn 43:283±293