- Email: [email protected]
Behaviour of human osteoblasts cultured on bioactive glass coatings
Biomaterials 19 (1998) 1019 — 1025
Behaviour of human osteoblasts cultured on bioactive glass coatings A. Oliva!,*, A. Salerno!, B. Locardi", V. Riccio#, F. Della Ragione!, P. Iardino!, V. Zappia$ ! Institute of Biochemistry of Macromolecules, Faculty of Medicine, Second University of Naples, Naples, Italy " Chair of Pediatric Orthopaedics, Faculty of Medicine, Second University of Naples, 80138 Naples, Italy # Stazione Sperimentale Del Vetro, Venice, Italy $ Institute of Food Science and Technology, CNR, Avellino, Received 2 January 1997; accepted 30 November 1997
Abstract Two new formulations of bioactive glasses were used as coatings on titanium alloy (TiAl6V4) implants for prosthetic applications in the orthopaedic field. The biocompatibility of these bioglasses, as well as their osteoconductive properties, were assessed by employing primary cultures of human osteoblasts. A nonbioactive glass, the titanium alloy and polystyrene surface were used as controls. The results obtained demonstrated that the two bioglasses elicited a rapid and strong proliferative response by osteoblasts, which spread, formed a close layer and then expressed the specific osteoblastic marker i.e. osteocalcin. In comparison, cells grew on the nonbioactive glass to a much minor extent, similar to that of polystyrene control, showing individual cellular elements not forming a compact sheet, but expressed levels of osteocalcin clearly higher than both the polystyrene control and the two bioglasses. Finally, a very low proliferative rate of osteoblasts and the synthesis of hardly detectable osteocalcin amounts were observed with the titanium alloy. In conclusion, our studies indicate that the new bioactive glasses are effective in stimulating osteoblast growth and differentiation. ( 1998 Published by Elsevier Science Ltd. All rights reserved Keywords: Human osteoblasts; Bioactive glass; Implant coatings; Bone substitutes
1. Introduction Biomaterials for hard tissue applications can be classified as bioinert or bioactive materials [1]. Bioinert materials have neither a positive nor a negative effect on bone growth, while the bioactive ones induce specific biological activity, namely they promote osteogenesis and establish an interfacial bond with the tissues to be implanted. However, this classification is controversial: indeed, several authors reject the existence of bioinert materials since there is always some kind of interactions between every type of material and the physiological environment. Since their discovery by Hench [2], new types of bioactive glasses have been developed and used in prosthetic applications and in repair of bone defects, due to their
* Corresponding author. Fax:#39 81 441688.
well-documented biocompatibility, as well as their osteoconductive and bone-bonding properties [2—7]. Among the various coating materials that are currently used in orthopaedic surgery, bioglasses (i) allow the prosthesis to adapt to the bone cavity; (ii) prevent the formation of fibrous tissue at the prosthesis—bone interface; (iii) favour a strong chemical bond between implant and bone tissue. In addition, bioactive glass has been shown to be a promising bone substitute material in experimental bone defects as well as in a number of clinical trials in the odontostomatology field [6, 7]. Bioglasses are silicate glasses containing sodium, calcium and phosphate as the main components. When exposed to body fluids, bioactive glasses undergo corrosion with a leaching of alkali ions resulting in the formation of a silica gel and a calcium phosphate layer on their surface. Successively, the calcium phosphate layer will recrystallize into hydroxycarbonate apatite [2, 3, 8]. Bone-bonding properties of bioglasses are based on the formation of this layer. Compared to synthetic hydroxyapatite, the surface layer of bioactive glasses is
0142-9612/98/$19.00 ( 1998 Published by Elsevier Science Ltd. All rights reserved. PII S 0 1 4 2 - 9 6 1 2 ( 9 7 ) 0 0 2 4 9 - 4
1020
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
more similar, in terms of crystallinity, to the apatite of bone tissue and consequently, a greater bone bonding has been reported for bioactive glasses than for hydroxyapatite [5]. The development of bioglass coatings for orthopaedic prostheses has been facilitated by the introduction of plasma spray techniques, in which granular materials are sprayed on the surface of a metal. The process takes place by injecting a gas flow (plasma spray) into a chamber in which the temperature is raised up to 10 000—30 000 K by means of an electric arc. Powders with a bead size less than 80 lm are pushed into the chamber, heated and forcefully sprayed through a nozzle onto the surface of the metal to be coated [9]. Among the advantages of this metodology there are the high deposition speed and the decreased modification of the metal substrate. With respect to the morphological characteristics, the disadvantages may be due to the presence of excessive macro- and microporosity in the coating layer, which can result either in the alteration of the mechanical properties or in the modification of the biological responses. Biocompatibility and performance of biomaterials can be usefully investigated by employing an in vitro model in a preliminary screening [10, 11] and successively testing in laboratory animals only the materials which are cytocompatible. Indeed, the use of cell culture systems is a valuable aid that can permit (i) a prior selection of biomaterials, avoiding unnecessary suffering of living beings; (ii) the direct observation of cell—material interaction, without the complexity of in vivo model; (iii) the rapid evaluation of biochemical markers of cellular phenotype. It has to be underlined, however, that the in vitro approach must be considered only the first step in the analysis of a new material and needs to be complemented by in vivo tests. This is particularly true for hard tissue application of biomaterials for which another crucial aspect has to be taken into account, namely, the mechanical compatibility. In an ideal in vitro approach, biomaterials should match the cell populations typical of the implant site, which, in the case of hard tissue prostheses, are mainly human osteoblasts [12, 13]. So far, conversely, the great majority of investigations have been performed on both established cell lines and primary cultures of rodent osteoblastic cells. However, the biocompatibility results obtained employing these models cannot be directly extrapolated to the normal human bone condition. Indeed, generally, cell lines as well as rodent cells are much more resistant to stressing conditions than the normal human counterpart. Therefore, we selected primary cultures of human osteoblasts as the most appropriate model to study in vitro the biocompatibility and performance of orthopaedic and odontostomatological materials [14]. The aim of the present study was to analyse the behaviour of human osteoblasts cultured on bioactive glass coatings. More precisely, we projected and obtained two
new kinds of bioactive glasses and compared them to control samples consisting in a nonbioactive glass, titanium alloy as such and the bottom of polystyrene plates. In particular, we assessed cell adhesion and morphology, proliferation rate, material colonization, as well as the expression of the peculiar biochemical parameter of osteoblastic phenotype, namely osteocalcin.
2. Materials and methods 2.1. Preparation of glasses We used the materials and the methods described below to obtain new formulations of bioactive glass (Biovetro) [9] and a bioinert glass for titanium plate coatings. Biovetro belongs to a group of compounds whose biocompatibility and osteoconductive properties have been previously assessed [2—7]. The glasses had the following composition: Na O (7—24%), K O (0.5—6%), CaO 2 2 (8—42%), MgO (1—3%), Al O (0.1—2%), SiO (46—72%) 2 3 2 and P O (0—7%). They were made at SEIPI laboratories 2 5 using an electric furnace and melting at 1350°C a mixture of raw materials of analytical grade purity. The glasses obtained were indicated as BVA, BVF and BVH. The first sample contained 6% P O and CaO/Na O 1 : 1. 2 5 2 BVF sample had 7% P O and a ratio CaO/Na O 9 : 1. 2 5 2 BVH, instead, was a sodium—calcium silicate glass absolutely lacking P O . 2 5 After the melting process, the glass was cast, crushed and transformed into powder of grain size suitable to the spraying step. The coating technique was achieved by using plasma spray equipment operating at a controlled atmosphere in the presence of an electric arc of power 25—30 kW and a modified torch with parameters that had been studied carefully for its use with the inorganic materials under investigation. Analysis of the glasses after the coating procedure did not demonstrate any remarkable modification ((0.8%) in the composition of the raw materials. The metal support was made up of a titanium alloy, TiAl6V4, similar to that used for a prosthesis, in the form of a disc 15 mm large and 3.3 mm thick. The samples were blasted by 5 atmospheres throw of 0.7 mm granules white corundum (Al O ). Some of the titanium disks 2 3 were directly assayed (titanium controls), while the others were coated as described before. The thickness of the coatings was kept constant at about 80 lm. After sterilization with c-rays, the samples were prepared for cell plating. 2.2. Preparation of the model system The isolation technique is based on the ability of osteoblasts to migrate from cancellous bone chips previously
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
treated with collagenase, [14, 15]. The bone specimen was obtained from tarsal scaphoid of a 22 yr old woman undergone reconstructive surgery. After a short treatment (60 min at 37°C) with collagenase at 1 mg ml~1 to remove the fibroblastic contaminants, the bone fragments were plated in 100 mm dishes containing Dulbecco’s Modified Eagle’s Medium (DMEM) with 10% fetal calf serum, sodium ascorbate (50 lg ml~1), penicillin (100 units ml~1) and streptomycin (100 lg ml~1). The cultures were incubated at 37°C in a 5% CO humidified 2 atmosphere. Within 2—3 weeks, cells migrated from bone chips and colonized the whole dish. They were then trypsinized and plated in the same medium at a density of 5]103 cm~2. The medium was changed every 3 days until the cells reached confluence. Cultures at fourth passage were used in our experiments. 2.3. Characterization of the model system The osteoblastic phenotype was assessed on the basis of (i) the high levels of alkaline phosphatase activity and its stimulation by the hormonal form of vitamin D (1.25 3 (OH) D ); (ii) the production of a specific noncol2 3 lagenous protein, namely osteocalcin; (iii) the mineralization of the extracellular matrix. As regards osteocalcin, it is the major noncollagenous bone protein as well as the only known osteoblast-specific polypeptide; the expression of its gene is under the control of the hormonal form of vitamin D . While the 3 production of this protein by the untreated cells was under the lower limit of the detection procedure (0.01 ng ml~1 of medium), the addition of 100 nM 1,25 (OH) D to confluent osteoblast cultures significantly 2 3 induced the osteocalcin gene expression [14, 15]. Further evidence of the osteoblastic phenotype was provided by the osteogenic capacity expressed by in vitro mineralization of the extracellular matrix. When 10 mM b-glycerophosphate, which acts as a phosphate donor, was added to the medium of confluent cultures, the extracellular matrix underwent calcification within three weeks, as confirmed by the intense staining with the von Kossa method [14]. This result supports the presence of high levels of alkaline phosphatase as well as the role of this enzyme in the initiation of the osteogenic process [16, 17]. 2.4. MTT assay The MTT assay is a very simple and useful tool for evaluating cell vitality and proliferation [18]. The key component is 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide or MTT. Mitochondrial dehydrogenases of living cells reduce the tetrazolium ring, yielding a blue formazan product which can be measured spectrophotometrically. The amount of for-
1021
mazan produced is proportional to the number of viable cells present. MTT (5 mg ml~1 in DMEM without phenol red) was added to the wells in an amount equivalent to 10% of the culture medium. After an incubation of 4 h at 37°C, the liquid was aspirated and the insoluble formazan produced was dissolved in isopropanol. The optical densities were measured at 570 nm, subtracting background absorbance determined at 690 nm. 2.5. Osteocalcin measurement In order to evaluate the production of osteocalcin in our model, confluent cultures of cells were incubated for 48 h in QBSF-51 medium (SIGMA) in the presence of 100 nM 1,25 (OH) D [14, 15]. The amount of the poly2 3 peptide in the cell medium was estimated by a commercial enzymeimmunoassay (EIA) employing highly specific monoclonal antibodies and peroxidase-labelled osteocalcin. 2.6. Scanning electron microscopy After the removal of the medium, the specimens were fixed with 2% glutaraldehyde in 0.01 M phosphate buffer, pH 7.4, for 1 h at room temperature. The cells, after being rinsed several times in the same buffer, were post-fixed in 1% osmium tetraoxide for 30 min at 4°C and dehydrated in an ethanol—water series of 30, 50, 70, 95% ethanol, each step taking 10—20 min, followed by 1 h in 100% ethanol. The samples were critical point dried and coated with a thin layer of gold (Edwards E306) and examined using a Cambridge electron microscope, set at 20 kV. 2.7. Statistical analysis Data are presented as the mean$standard deviation.
3. Results and discussion 3.1. Preparation of samples The samples (14 for each type of glass) were placed in a 24-multiwell plates in the presence of 1.5 ml of phosphate-buffered saline (PBS), pH 7.4 containing penicillin (100 units ml~1) and streptomycin (100 mg ml~1), at 37°C in a 5% CO humidified atmosphere. After 24 h the 2 media were removed and pH was measured. While BVH did not modify at this parameter, the two bioactive glasses and particularly BVA, considerably raised the pH (between 8.5 and 8.7). This alkaline pH shift, due to the leaching of cationic ions from the bioglass surface, required other three wash treatments (as indicated above) before pH stabilization at 7.4 value made possible cell
1022
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
plating. Also BVH was obviously submitted to this prolonged soaking treatment. Either polystyrene bottoms of 24-well plates or titanium alloy (TiAl6V4) discs as such, that is without any coating, were used as control samples and treated in a similar manner. 3.2. Biocompatibility studies In order to evaluate the biological response to the materials under investigation, 34 000 osteoblasts were seeded per each well, which corresponded to a cellular density of about 2]104 cm2. All the experiments were carried out at least in triplicate. Twenty-four hours after cell plating, the medium was removed from each well and the unattached osteoblasts were counted in order to evaluate the adhesion efficiency (Fig. 1). In all the samples, except those of BVA, the percentage of nonadherent cells ranged from 2 to 4, whereas in the BVA case about 8% of the osteoblasts remained unattached. This higher percentage could be due to a further leaching of cations, that, even if not detectable in medium in toto (after 4 day washings, pH was stable at 7.4 value), could locally, that is in strict environs of material, ‘disturb’ the delicate phase of cell anchorage. Four days after the cell plating, MTT test was carried out to evaluate both cell vitality and proliferation in different samples. The highest values obtained were observed in BVH and in the other two control samples, especially titanium alloy. In contrast, cell vitality and/or proliferation were lower in both bioactive glasses and particularly in BVF samples (Fig. 2).
Fig. 2. MTT test on osteoblast cultures after four days from cell plating. Abbreviations are specified in the legend of Fig. 1.
Fig. 3. MTT test on osteoblasts after eight days of culture. For the abbreviations see the legend of Fig. 1.
Fig. 1. Percentage of unattached osteoblasts after plating onto different materials. PS"polystyrene; TI"titanium alloy uncoated; BVA"titanium alloy coated with ‘bioglass A’; BVF"titanium alloy coated with ‘bioglass F’; BVH"titanium alloy coated with ‘glass H’.
At eighth day, the scenario was significantly changed (Fig. 3) and then totally modified at sixteenth day of culture (Fig. 4), when a considerable proliferative push occurred in cells on BVA and BVF glasses. Indeed, at 16th day MTT test values related to BVA and BVF samples were about 3 times greater than the controls. Furthermore, these values were 5- and 10-fold higher, respectively, than those detected for the same bioglasses at fourth day. In contrast, only small increases in cell proliferation (about 1.5 times) were evidenced for BVH and for the other controls.
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
Fig. 4. MTT test on cell cultures after 16 days from osteoblast plating. Abbreviations are specified in the legend of Fig. 1.
The very high MTT values obtained for BVA and BVF might be due to their different chemical reactivity which could result in distinct surface topography. This was especially evident at later than early times when no remarkable differences were observed. In contrast, BVH glass did not seem to be subjected to such a phenomenon. A temporal sequence of events has been demonstrated during the development of a fully differentiated osteoblastic phenotype, namely: (i) proliferation, (ii) extracellular matrix (EM) maturation, (iii) mineralization [18]. The proliferative phase, characterized by both cell number increase and synthesis of EM basic components, is followed by the expression of high levels of alkaline phosphatase causing a series of modifications in the extracellular matrix, which finally culminate in the deposition of hydroxyapatite crystals. In turn, the onset of mineralization induces the synthesis of the bone-specific protein, osteocalcin, strictly dependent on the hormonal form of vitamin D . This calcium-binding polypeptide, 3 that tightly binds to hydroxyapatite, represents the peculiar marker of final osteoblast differentiation and plays a pivotal role in bone remodelling, either functioning as a matrix signal for osteoclast differentiation [19] or acting by itself as a negative regulator of bone matrix deposition by osteoblasts [20]. Therefore, to verify the termination of proliferative phase as well as to assess the expression of differentiated osteoblastic phenotype, 22 days after cell plating, cultures were incubated in QBSF-51 in the presence of 100 nM 1,25 (OH )D . Forty eight hours later, the media were 2 3 aspirated and assayed for osteocalcin content, while the last MTT test was carried out (Fig. 5). If we compare the results of MTT at 24th day (Fig. 5) to those at 4th day (Fig. 2), some conclusions can be
1023
Fig. 5. MTT test on osteoblast cultures after 24 days from plating. Abbreviations are explained in the legend under Fig. 1.
Fig. 6. Osteocalcin measurement on cell cultures after 24 days from osteoblast plating. Results are expressed an ng osteocalcin/MTT absorbance. Abbreviations are specified in the legend of Fig. 1.
drawn. As regards the control samples, the values related to polystyrene, as well as those corresponding to BVH, were equivalent and almost 2-fold higher than those of starting, whereas vital cell number on titanium clearly decreased. MTT values of BVA and BVF remained the highest being more than double in comparison to BVH and polystyrene, but both were rather lower than the MTT test carried out at sixteenth day. This decrease could be attributable to a rapid shift from the proliferative to differentiative phase that was induced by the presence of the hormonal form of vitamin D . 3 Fig. 6 reports osteocalcin levels synthesized by cells grown on the different materials after 22-day culture and
1024
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
2-day incubation with 100 nM 1,25-dihydroxycolecalciferol. Results were expressed as ng osteocalcin/absorbance MTT test. It appeared evident that the highest osteocalcin levels were synthesized by cells on BVH, while osteoblasts grown on BVA and BVF produced this polypeptide in amounts which were greater than polystyrene but about 50% those of BVH. Therefore, the two bioactive glasses BVA and BVF seemed to greatly promote osteoblast proliferation, whereas the nonbioactive glass BVH appeared to stimulate a differentiative pathway. In contrast, levels hardly detectable of osteocalcin were synthesized by cells on titanium alloy, that appeared to be toxic in long-term culture. Indeed, only in the first week of culture the situation of osteoblasts on uncoated titanium seemed similar, if not better, than those of the other controls, in terms of growth and viability (see Figs. 2 and 3) while, successively, a clear decrease of these parameters was observed. Although we have not evaluated the leaching of toxic ion components of titanium alloy, the results reported above are similar to the literature data demonstrating that sublethal amounts of the ionic components of TiAl6V4 impaired cell proliferation and alkaline phosphatase secretion and severely suppressed the peculiar expression of the osteoblastic phenotype, i.e. osteocalcin synthesis and deposition of a mineralized matrix [21—23]. The lack of toxicity by coated titanium clearly indicated that glass coatings were effective barriers to ion leaching. In particular, in the case of bioactive glasses, this effect might be the consequence of metal phosphate formation or incorporation of metal ions in the hydroxyapatite structure [23]. Finally, in order to directly evaluate the interaction between materials and cells, scanning electron micro-
scopy was carried out on the 24-day cultures. This analysis revealed that with both BVA (Fig. 7) and BVF (Fig. 8) osteoblasts attached and completely spread on the samples forming a very close layer in which individual cells were indistinguishable. On the contrary, single cells with a spindle-like shape were evident on BVH (Fig. 9). These SEM data strongly confirmed the MTT results and were also in accord with literature findings evidencing that one of the main regulators of proliferative rate in anchoragedependent cells is the shape [24]. Indeed, cells which adhere to some material, but spread little, as in BVH case, proliferate at a lower rate than those of the same population that are flattened and well spread on another substratum, as for BVA and BVF samples.
Fig. 7. Scanning electron micrograph of 24-day osteoblast culture grown on BVA (300]).
Fig. 9. Scanning electron micrograph of 24-day osteoblast culture grown on BVH (300]).
Fig. 8. Scanning electron micrograph of 24-day osteoblast culture grown on BVF (300]).
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
In conclusion, our results demonstrate that our new formulations of bioactive glasses firstly stimulate to a great extent osteoblast proliferation and colonization, and allow in a successive differentiative step the expression of the peculiar osteoblast biochemical marker, namely osteocalcin. Therefore, these biomaterials are very promising not only for prosthetic applications, but also as bone substitute materials for bone defects repair in orthopaedic and maxillo-facial surgery, as well as in odontostomatological field.
Acknowledgements This research was partly supported by the grant ‘National Programme of Research on Advanced Innovative Materials’ by the ‘Italian Ministry of University and Scientific and Technological Research’ and bestowed to ‘Tecnobiomedica S.p.A’.
References [1] Fleck C, Eifler D, Ondracek G, Watts JF. In: Krawczynski J, Ondracek G, editors. Biomaterials scientific series of the international bureau, vol. 16. 1993:18—81. [2] Hench LL, Splinter RJ, Allen WC, Greenlee TK. Bonding mechanisms at the interface of ceramic prosthetic materials. J Biomed Mater Res Symp 1971;2:117—41. [3] Hench LL, Paschall HA. Direct chemical bond of bioactive glassceramic materials to bone and muscle. J Biomed Mater Res Symp 1973;4:25—42. [4] Ducheyne P. Bioglass coatings and bioglass composites as implant materials. J Biomed Mater Res 1985;19:273—91. [5] Pajamaki KJJ, Lindholm TS, Andersson OH et al. Bioactive glass and glass ceramic coated hip endoprosthesis: experimental study in rabbit. J Mater Sci Mater Med 1995;6:14—18. [6] Nakamura T, Yamamuro T, Higashi S, Kokubo T, Itoo S. A new glass-ceramic for bone replacement: evaluation of its bonding to bone tissue. J Biomed Mater Res 1985;19:685—98. [7] Turunen T, Peltola J, Happonen RP, Yli-Urpo A. Effect of bioactive glass granules and polytetrafluoroethylene membrane on repair of cortical bone defect. J Mater Sci Mater Med 1995;5:639—41. [8] Hench LL, Stanley HT, Clark AE, Hall M, Wilson J. Dental applications of bioglass implants. In: Bonfield W, Hastings GW, Tanner KE, editors. Bioceramics, vol. 4. (Proceedings of Fourth International Symposium on Ceramics in Medicine, London, September 1991:231—8.
1025
[9] Gabbi C, Cacchioli A, Locardi B, Guadagnino E. Bioactive glass coating: physicochemical aspects and biological findings. Biomaterials 1995;10:515—20. [10] Vrouwenvelder WCA, Groot CG, de Groot K. Behaviour of fetal rat osteoblasts cultured in vitro on bioactive glass and nonreactive glasses. Biomaterials 1992;13:382—92. [11] Vrouwenvelder WCA, Groot CG, de Groot K. Better histology and biochemistry for osteoblasts cultured on titanium-doped bioactive glass: bioglass 45S5 compared with iron-, titanium-, fluorine- and boron-containing bioactive glasses. Biomaterials 1994;15:97—106. [12] Naji A, Harmand M-F. Cytocompatibility of two coating materials, amorphous alumina and silicon carbide, using human differentiated cell cultures. Biomaterials 1991;12:690—4. [13] Serre CM, Papillard M, Chavassieux P, Boivin G. In vitro induction of a calcifying matrix by biomaterials constituted of collagen and/or hydroxyapatite: an ultrastructural comparison of three types of biomaterials. Biomaterials 1993;14:97—106. [14] Oliva A, Della Ragione F, Salerno A et al. Biocompatibility studies on glass ionomer cements by primary cultures of human osteoblasts. Biomaterials 1996;17:1351—6. [15] Oliva A, Della Ragione F, Fratta M, Marrone G, Palumbo R, Zappia V. Effect of retinoic acid on osteocalcin gene expression in human osteoblasts. Biochim Biophys Res Commun 1993; 191:908—14. [16] Wuthier RE, Register TC. Role of alkaline phosphatase, a polyfunctional enzyme, in mineraling tissues. In: Butler WT, editor. The chemistry and biology of mineraling tissues. EBSCO Media, Birmingham, 1985:113—24. [17] Beertsen W, van den Bos T. The role of alkaline phosphatase in the mineralization of collagenous substrates. In: Davidovitch Z, editor, The biological mechanisms of tooth movement and craniofacial adaptation. Columbus, 1992:181—5. [18] Slater TF, Sawyer B, Strauli U. Studies on succinate-tetrazolium reductase systems. III. Points of coupling of four different tetrazolium salts. Biochim Biophys Acta 1963;77:383—93. [19] Lian JB, Stein GS, Owen TA et al. Regulation of osteocalcin gene expression during growth and differentiation of the bone cell phenotype. In: Davidovitch Z, editor, The biological mechanisms of tooth movement and craniofacial adaptation. Columbus, 1992:159—79. [20] Ducy P, Desbois C, Boyce B et al. Increased bone formation in osteocalcin-deficient mice. Nature 1996;382:448—52. [21] Thompson JG, Puleo DA. Effects of sub-lethal metal ion concentrations on osteogenic cells derived from bone marrow stromal cells. J Appl Biomater 1995;6:249—58. [22] Thompson JG, Puleo DA. Ti-6Al-4V ion solution inhibition of osteogenic cell phenotype as a function of differentiation timecourse in vitro. Biomaterials 1996;17:1949—54. [23] Sousa SR, Barbosa MA. Effect of hydroxyapatite thickness on metal ion release from Ti6Al4V substrates. Biomaterials 1996;17:397—404. [24] Folkman J, Moscona A. Role of cell shape in growth control. Nature 1978;273:345—9.
Behaviour of human osteoblasts cultured on bioactive glass coatings A. Oliva!,*, A. Salerno!, B. Locardi", V. Riccio#, F. Della Ragione!, P. Iardino!, V. Zappia$ ! Institute of Biochemistry of Macromolecules, Faculty of Medicine, Second University of Naples, Naples, Italy " Chair of Pediatric Orthopaedics, Faculty of Medicine, Second University of Naples, 80138 Naples, Italy # Stazione Sperimentale Del Vetro, Venice, Italy $ Institute of Food Science and Technology, CNR, Avellino, Received 2 January 1997; accepted 30 November 1997
Abstract Two new formulations of bioactive glasses were used as coatings on titanium alloy (TiAl6V4) implants for prosthetic applications in the orthopaedic field. The biocompatibility of these bioglasses, as well as their osteoconductive properties, were assessed by employing primary cultures of human osteoblasts. A nonbioactive glass, the titanium alloy and polystyrene surface were used as controls. The results obtained demonstrated that the two bioglasses elicited a rapid and strong proliferative response by osteoblasts, which spread, formed a close layer and then expressed the specific osteoblastic marker i.e. osteocalcin. In comparison, cells grew on the nonbioactive glass to a much minor extent, similar to that of polystyrene control, showing individual cellular elements not forming a compact sheet, but expressed levels of osteocalcin clearly higher than both the polystyrene control and the two bioglasses. Finally, a very low proliferative rate of osteoblasts and the synthesis of hardly detectable osteocalcin amounts were observed with the titanium alloy. In conclusion, our studies indicate that the new bioactive glasses are effective in stimulating osteoblast growth and differentiation. ( 1998 Published by Elsevier Science Ltd. All rights reserved Keywords: Human osteoblasts; Bioactive glass; Implant coatings; Bone substitutes
1. Introduction Biomaterials for hard tissue applications can be classified as bioinert or bioactive materials [1]. Bioinert materials have neither a positive nor a negative effect on bone growth, while the bioactive ones induce specific biological activity, namely they promote osteogenesis and establish an interfacial bond with the tissues to be implanted. However, this classification is controversial: indeed, several authors reject the existence of bioinert materials since there is always some kind of interactions between every type of material and the physiological environment. Since their discovery by Hench [2], new types of bioactive glasses have been developed and used in prosthetic applications and in repair of bone defects, due to their
* Corresponding author. Fax:#39 81 441688.
well-documented biocompatibility, as well as their osteoconductive and bone-bonding properties [2—7]. Among the various coating materials that are currently used in orthopaedic surgery, bioglasses (i) allow the prosthesis to adapt to the bone cavity; (ii) prevent the formation of fibrous tissue at the prosthesis—bone interface; (iii) favour a strong chemical bond between implant and bone tissue. In addition, bioactive glass has been shown to be a promising bone substitute material in experimental bone defects as well as in a number of clinical trials in the odontostomatology field [6, 7]. Bioglasses are silicate glasses containing sodium, calcium and phosphate as the main components. When exposed to body fluids, bioactive glasses undergo corrosion with a leaching of alkali ions resulting in the formation of a silica gel and a calcium phosphate layer on their surface. Successively, the calcium phosphate layer will recrystallize into hydroxycarbonate apatite [2, 3, 8]. Bone-bonding properties of bioglasses are based on the formation of this layer. Compared to synthetic hydroxyapatite, the surface layer of bioactive glasses is
0142-9612/98/$19.00 ( 1998 Published by Elsevier Science Ltd. All rights reserved. PII S 0 1 4 2 - 9 6 1 2 ( 9 7 ) 0 0 2 4 9 - 4
1020
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
more similar, in terms of crystallinity, to the apatite of bone tissue and consequently, a greater bone bonding has been reported for bioactive glasses than for hydroxyapatite [5]. The development of bioglass coatings for orthopaedic prostheses has been facilitated by the introduction of plasma spray techniques, in which granular materials are sprayed on the surface of a metal. The process takes place by injecting a gas flow (plasma spray) into a chamber in which the temperature is raised up to 10 000—30 000 K by means of an electric arc. Powders with a bead size less than 80 lm are pushed into the chamber, heated and forcefully sprayed through a nozzle onto the surface of the metal to be coated [9]. Among the advantages of this metodology there are the high deposition speed and the decreased modification of the metal substrate. With respect to the morphological characteristics, the disadvantages may be due to the presence of excessive macro- and microporosity in the coating layer, which can result either in the alteration of the mechanical properties or in the modification of the biological responses. Biocompatibility and performance of biomaterials can be usefully investigated by employing an in vitro model in a preliminary screening [10, 11] and successively testing in laboratory animals only the materials which are cytocompatible. Indeed, the use of cell culture systems is a valuable aid that can permit (i) a prior selection of biomaterials, avoiding unnecessary suffering of living beings; (ii) the direct observation of cell—material interaction, without the complexity of in vivo model; (iii) the rapid evaluation of biochemical markers of cellular phenotype. It has to be underlined, however, that the in vitro approach must be considered only the first step in the analysis of a new material and needs to be complemented by in vivo tests. This is particularly true for hard tissue application of biomaterials for which another crucial aspect has to be taken into account, namely, the mechanical compatibility. In an ideal in vitro approach, biomaterials should match the cell populations typical of the implant site, which, in the case of hard tissue prostheses, are mainly human osteoblasts [12, 13]. So far, conversely, the great majority of investigations have been performed on both established cell lines and primary cultures of rodent osteoblastic cells. However, the biocompatibility results obtained employing these models cannot be directly extrapolated to the normal human bone condition. Indeed, generally, cell lines as well as rodent cells are much more resistant to stressing conditions than the normal human counterpart. Therefore, we selected primary cultures of human osteoblasts as the most appropriate model to study in vitro the biocompatibility and performance of orthopaedic and odontostomatological materials [14]. The aim of the present study was to analyse the behaviour of human osteoblasts cultured on bioactive glass coatings. More precisely, we projected and obtained two
new kinds of bioactive glasses and compared them to control samples consisting in a nonbioactive glass, titanium alloy as such and the bottom of polystyrene plates. In particular, we assessed cell adhesion and morphology, proliferation rate, material colonization, as well as the expression of the peculiar biochemical parameter of osteoblastic phenotype, namely osteocalcin.
2. Materials and methods 2.1. Preparation of glasses We used the materials and the methods described below to obtain new formulations of bioactive glass (Biovetro) [9] and a bioinert glass for titanium plate coatings. Biovetro belongs to a group of compounds whose biocompatibility and osteoconductive properties have been previously assessed [2—7]. The glasses had the following composition: Na O (7—24%), K O (0.5—6%), CaO 2 2 (8—42%), MgO (1—3%), Al O (0.1—2%), SiO (46—72%) 2 3 2 and P O (0—7%). They were made at SEIPI laboratories 2 5 using an electric furnace and melting at 1350°C a mixture of raw materials of analytical grade purity. The glasses obtained were indicated as BVA, BVF and BVH. The first sample contained 6% P O and CaO/Na O 1 : 1. 2 5 2 BVF sample had 7% P O and a ratio CaO/Na O 9 : 1. 2 5 2 BVH, instead, was a sodium—calcium silicate glass absolutely lacking P O . 2 5 After the melting process, the glass was cast, crushed and transformed into powder of grain size suitable to the spraying step. The coating technique was achieved by using plasma spray equipment operating at a controlled atmosphere in the presence of an electric arc of power 25—30 kW and a modified torch with parameters that had been studied carefully for its use with the inorganic materials under investigation. Analysis of the glasses after the coating procedure did not demonstrate any remarkable modification ((0.8%) in the composition of the raw materials. The metal support was made up of a titanium alloy, TiAl6V4, similar to that used for a prosthesis, in the form of a disc 15 mm large and 3.3 mm thick. The samples were blasted by 5 atmospheres throw of 0.7 mm granules white corundum (Al O ). Some of the titanium disks 2 3 were directly assayed (titanium controls), while the others were coated as described before. The thickness of the coatings was kept constant at about 80 lm. After sterilization with c-rays, the samples were prepared for cell plating. 2.2. Preparation of the model system The isolation technique is based on the ability of osteoblasts to migrate from cancellous bone chips previously
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
treated with collagenase, [14, 15]. The bone specimen was obtained from tarsal scaphoid of a 22 yr old woman undergone reconstructive surgery. After a short treatment (60 min at 37°C) with collagenase at 1 mg ml~1 to remove the fibroblastic contaminants, the bone fragments were plated in 100 mm dishes containing Dulbecco’s Modified Eagle’s Medium (DMEM) with 10% fetal calf serum, sodium ascorbate (50 lg ml~1), penicillin (100 units ml~1) and streptomycin (100 lg ml~1). The cultures were incubated at 37°C in a 5% CO humidified 2 atmosphere. Within 2—3 weeks, cells migrated from bone chips and colonized the whole dish. They were then trypsinized and plated in the same medium at a density of 5]103 cm~2. The medium was changed every 3 days until the cells reached confluence. Cultures at fourth passage were used in our experiments. 2.3. Characterization of the model system The osteoblastic phenotype was assessed on the basis of (i) the high levels of alkaline phosphatase activity and its stimulation by the hormonal form of vitamin D (1.25 3 (OH) D ); (ii) the production of a specific noncol2 3 lagenous protein, namely osteocalcin; (iii) the mineralization of the extracellular matrix. As regards osteocalcin, it is the major noncollagenous bone protein as well as the only known osteoblast-specific polypeptide; the expression of its gene is under the control of the hormonal form of vitamin D . While the 3 production of this protein by the untreated cells was under the lower limit of the detection procedure (0.01 ng ml~1 of medium), the addition of 100 nM 1,25 (OH) D to confluent osteoblast cultures significantly 2 3 induced the osteocalcin gene expression [14, 15]. Further evidence of the osteoblastic phenotype was provided by the osteogenic capacity expressed by in vitro mineralization of the extracellular matrix. When 10 mM b-glycerophosphate, which acts as a phosphate donor, was added to the medium of confluent cultures, the extracellular matrix underwent calcification within three weeks, as confirmed by the intense staining with the von Kossa method [14]. This result supports the presence of high levels of alkaline phosphatase as well as the role of this enzyme in the initiation of the osteogenic process [16, 17]. 2.4. MTT assay The MTT assay is a very simple and useful tool for evaluating cell vitality and proliferation [18]. The key component is 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide or MTT. Mitochondrial dehydrogenases of living cells reduce the tetrazolium ring, yielding a blue formazan product which can be measured spectrophotometrically. The amount of for-
1021
mazan produced is proportional to the number of viable cells present. MTT (5 mg ml~1 in DMEM without phenol red) was added to the wells in an amount equivalent to 10% of the culture medium. After an incubation of 4 h at 37°C, the liquid was aspirated and the insoluble formazan produced was dissolved in isopropanol. The optical densities were measured at 570 nm, subtracting background absorbance determined at 690 nm. 2.5. Osteocalcin measurement In order to evaluate the production of osteocalcin in our model, confluent cultures of cells were incubated for 48 h in QBSF-51 medium (SIGMA) in the presence of 100 nM 1,25 (OH) D [14, 15]. The amount of the poly2 3 peptide in the cell medium was estimated by a commercial enzymeimmunoassay (EIA) employing highly specific monoclonal antibodies and peroxidase-labelled osteocalcin. 2.6. Scanning electron microscopy After the removal of the medium, the specimens were fixed with 2% glutaraldehyde in 0.01 M phosphate buffer, pH 7.4, for 1 h at room temperature. The cells, after being rinsed several times in the same buffer, were post-fixed in 1% osmium tetraoxide for 30 min at 4°C and dehydrated in an ethanol—water series of 30, 50, 70, 95% ethanol, each step taking 10—20 min, followed by 1 h in 100% ethanol. The samples were critical point dried and coated with a thin layer of gold (Edwards E306) and examined using a Cambridge electron microscope, set at 20 kV. 2.7. Statistical analysis Data are presented as the mean$standard deviation.
3. Results and discussion 3.1. Preparation of samples The samples (14 for each type of glass) were placed in a 24-multiwell plates in the presence of 1.5 ml of phosphate-buffered saline (PBS), pH 7.4 containing penicillin (100 units ml~1) and streptomycin (100 mg ml~1), at 37°C in a 5% CO humidified atmosphere. After 24 h the 2 media were removed and pH was measured. While BVH did not modify at this parameter, the two bioactive glasses and particularly BVA, considerably raised the pH (between 8.5 and 8.7). This alkaline pH shift, due to the leaching of cationic ions from the bioglass surface, required other three wash treatments (as indicated above) before pH stabilization at 7.4 value made possible cell
1022
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
plating. Also BVH was obviously submitted to this prolonged soaking treatment. Either polystyrene bottoms of 24-well plates or titanium alloy (TiAl6V4) discs as such, that is without any coating, were used as control samples and treated in a similar manner. 3.2. Biocompatibility studies In order to evaluate the biological response to the materials under investigation, 34 000 osteoblasts were seeded per each well, which corresponded to a cellular density of about 2]104 cm2. All the experiments were carried out at least in triplicate. Twenty-four hours after cell plating, the medium was removed from each well and the unattached osteoblasts were counted in order to evaluate the adhesion efficiency (Fig. 1). In all the samples, except those of BVA, the percentage of nonadherent cells ranged from 2 to 4, whereas in the BVA case about 8% of the osteoblasts remained unattached. This higher percentage could be due to a further leaching of cations, that, even if not detectable in medium in toto (after 4 day washings, pH was stable at 7.4 value), could locally, that is in strict environs of material, ‘disturb’ the delicate phase of cell anchorage. Four days after the cell plating, MTT test was carried out to evaluate both cell vitality and proliferation in different samples. The highest values obtained were observed in BVH and in the other two control samples, especially titanium alloy. In contrast, cell vitality and/or proliferation were lower in both bioactive glasses and particularly in BVF samples (Fig. 2).
Fig. 2. MTT test on osteoblast cultures after four days from cell plating. Abbreviations are specified in the legend of Fig. 1.
Fig. 3. MTT test on osteoblasts after eight days of culture. For the abbreviations see the legend of Fig. 1.
Fig. 1. Percentage of unattached osteoblasts after plating onto different materials. PS"polystyrene; TI"titanium alloy uncoated; BVA"titanium alloy coated with ‘bioglass A’; BVF"titanium alloy coated with ‘bioglass F’; BVH"titanium alloy coated with ‘glass H’.
At eighth day, the scenario was significantly changed (Fig. 3) and then totally modified at sixteenth day of culture (Fig. 4), when a considerable proliferative push occurred in cells on BVA and BVF glasses. Indeed, at 16th day MTT test values related to BVA and BVF samples were about 3 times greater than the controls. Furthermore, these values were 5- and 10-fold higher, respectively, than those detected for the same bioglasses at fourth day. In contrast, only small increases in cell proliferation (about 1.5 times) were evidenced for BVH and for the other controls.
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
Fig. 4. MTT test on cell cultures after 16 days from osteoblast plating. Abbreviations are specified in the legend of Fig. 1.
The very high MTT values obtained for BVA and BVF might be due to their different chemical reactivity which could result in distinct surface topography. This was especially evident at later than early times when no remarkable differences were observed. In contrast, BVH glass did not seem to be subjected to such a phenomenon. A temporal sequence of events has been demonstrated during the development of a fully differentiated osteoblastic phenotype, namely: (i) proliferation, (ii) extracellular matrix (EM) maturation, (iii) mineralization [18]. The proliferative phase, characterized by both cell number increase and synthesis of EM basic components, is followed by the expression of high levels of alkaline phosphatase causing a series of modifications in the extracellular matrix, which finally culminate in the deposition of hydroxyapatite crystals. In turn, the onset of mineralization induces the synthesis of the bone-specific protein, osteocalcin, strictly dependent on the hormonal form of vitamin D . This calcium-binding polypeptide, 3 that tightly binds to hydroxyapatite, represents the peculiar marker of final osteoblast differentiation and plays a pivotal role in bone remodelling, either functioning as a matrix signal for osteoclast differentiation [19] or acting by itself as a negative regulator of bone matrix deposition by osteoblasts [20]. Therefore, to verify the termination of proliferative phase as well as to assess the expression of differentiated osteoblastic phenotype, 22 days after cell plating, cultures were incubated in QBSF-51 in the presence of 100 nM 1,25 (OH )D . Forty eight hours later, the media were 2 3 aspirated and assayed for osteocalcin content, while the last MTT test was carried out (Fig. 5). If we compare the results of MTT at 24th day (Fig. 5) to those at 4th day (Fig. 2), some conclusions can be
1023
Fig. 5. MTT test on osteoblast cultures after 24 days from plating. Abbreviations are explained in the legend under Fig. 1.
Fig. 6. Osteocalcin measurement on cell cultures after 24 days from osteoblast plating. Results are expressed an ng osteocalcin/MTT absorbance. Abbreviations are specified in the legend of Fig. 1.
drawn. As regards the control samples, the values related to polystyrene, as well as those corresponding to BVH, were equivalent and almost 2-fold higher than those of starting, whereas vital cell number on titanium clearly decreased. MTT values of BVA and BVF remained the highest being more than double in comparison to BVH and polystyrene, but both were rather lower than the MTT test carried out at sixteenth day. This decrease could be attributable to a rapid shift from the proliferative to differentiative phase that was induced by the presence of the hormonal form of vitamin D . 3 Fig. 6 reports osteocalcin levels synthesized by cells grown on the different materials after 22-day culture and
1024
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
2-day incubation with 100 nM 1,25-dihydroxycolecalciferol. Results were expressed as ng osteocalcin/absorbance MTT test. It appeared evident that the highest osteocalcin levels were synthesized by cells on BVH, while osteoblasts grown on BVA and BVF produced this polypeptide in amounts which were greater than polystyrene but about 50% those of BVH. Therefore, the two bioactive glasses BVA and BVF seemed to greatly promote osteoblast proliferation, whereas the nonbioactive glass BVH appeared to stimulate a differentiative pathway. In contrast, levels hardly detectable of osteocalcin were synthesized by cells on titanium alloy, that appeared to be toxic in long-term culture. Indeed, only in the first week of culture the situation of osteoblasts on uncoated titanium seemed similar, if not better, than those of the other controls, in terms of growth and viability (see Figs. 2 and 3) while, successively, a clear decrease of these parameters was observed. Although we have not evaluated the leaching of toxic ion components of titanium alloy, the results reported above are similar to the literature data demonstrating that sublethal amounts of the ionic components of TiAl6V4 impaired cell proliferation and alkaline phosphatase secretion and severely suppressed the peculiar expression of the osteoblastic phenotype, i.e. osteocalcin synthesis and deposition of a mineralized matrix [21—23]. The lack of toxicity by coated titanium clearly indicated that glass coatings were effective barriers to ion leaching. In particular, in the case of bioactive glasses, this effect might be the consequence of metal phosphate formation or incorporation of metal ions in the hydroxyapatite structure [23]. Finally, in order to directly evaluate the interaction between materials and cells, scanning electron micro-
scopy was carried out on the 24-day cultures. This analysis revealed that with both BVA (Fig. 7) and BVF (Fig. 8) osteoblasts attached and completely spread on the samples forming a very close layer in which individual cells were indistinguishable. On the contrary, single cells with a spindle-like shape were evident on BVH (Fig. 9). These SEM data strongly confirmed the MTT results and were also in accord with literature findings evidencing that one of the main regulators of proliferative rate in anchoragedependent cells is the shape [24]. Indeed, cells which adhere to some material, but spread little, as in BVH case, proliferate at a lower rate than those of the same population that are flattened and well spread on another substratum, as for BVA and BVF samples.
Fig. 7. Scanning electron micrograph of 24-day osteoblast culture grown on BVA (300]).
Fig. 9. Scanning electron micrograph of 24-day osteoblast culture grown on BVH (300]).
Fig. 8. Scanning electron micrograph of 24-day osteoblast culture grown on BVF (300]).
A. Oliva et al. / Biomaterials 19 (1998) 1019—1025
In conclusion, our results demonstrate that our new formulations of bioactive glasses firstly stimulate to a great extent osteoblast proliferation and colonization, and allow in a successive differentiative step the expression of the peculiar osteoblast biochemical marker, namely osteocalcin. Therefore, these biomaterials are very promising not only for prosthetic applications, but also as bone substitute materials for bone defects repair in orthopaedic and maxillo-facial surgery, as well as in odontostomatological field.
Acknowledgements This research was partly supported by the grant ‘National Programme of Research on Advanced Innovative Materials’ by the ‘Italian Ministry of University and Scientific and Technological Research’ and bestowed to ‘Tecnobiomedica S.p.A’.
References [1] Fleck C, Eifler D, Ondracek G, Watts JF. In: Krawczynski J, Ondracek G, editors. Biomaterials scientific series of the international bureau, vol. 16. 1993:18—81. [2] Hench LL, Splinter RJ, Allen WC, Greenlee TK. Bonding mechanisms at the interface of ceramic prosthetic materials. J Biomed Mater Res Symp 1971;2:117—41. [3] Hench LL, Paschall HA. Direct chemical bond of bioactive glassceramic materials to bone and muscle. J Biomed Mater Res Symp 1973;4:25—42. [4] Ducheyne P. Bioglass coatings and bioglass composites as implant materials. J Biomed Mater Res 1985;19:273—91. [5] Pajamaki KJJ, Lindholm TS, Andersson OH et al. Bioactive glass and glass ceramic coated hip endoprosthesis: experimental study in rabbit. J Mater Sci Mater Med 1995;6:14—18. [6] Nakamura T, Yamamuro T, Higashi S, Kokubo T, Itoo S. A new glass-ceramic for bone replacement: evaluation of its bonding to bone tissue. J Biomed Mater Res 1985;19:685—98. [7] Turunen T, Peltola J, Happonen RP, Yli-Urpo A. Effect of bioactive glass granules and polytetrafluoroethylene membrane on repair of cortical bone defect. J Mater Sci Mater Med 1995;5:639—41. [8] Hench LL, Stanley HT, Clark AE, Hall M, Wilson J. Dental applications of bioglass implants. In: Bonfield W, Hastings GW, Tanner KE, editors. Bioceramics, vol. 4. (Proceedings of Fourth International Symposium on Ceramics in Medicine, London, September 1991:231—8.
1025
[9] Gabbi C, Cacchioli A, Locardi B, Guadagnino E. Bioactive glass coating: physicochemical aspects and biological findings. Biomaterials 1995;10:515—20. [10] Vrouwenvelder WCA, Groot CG, de Groot K. Behaviour of fetal rat osteoblasts cultured in vitro on bioactive glass and nonreactive glasses. Biomaterials 1992;13:382—92. [11] Vrouwenvelder WCA, Groot CG, de Groot K. Better histology and biochemistry for osteoblasts cultured on titanium-doped bioactive glass: bioglass 45S5 compared with iron-, titanium-, fluorine- and boron-containing bioactive glasses. Biomaterials 1994;15:97—106. [12] Naji A, Harmand M-F. Cytocompatibility of two coating materials, amorphous alumina and silicon carbide, using human differentiated cell cultures. Biomaterials 1991;12:690—4. [13] Serre CM, Papillard M, Chavassieux P, Boivin G. In vitro induction of a calcifying matrix by biomaterials constituted of collagen and/or hydroxyapatite: an ultrastructural comparison of three types of biomaterials. Biomaterials 1993;14:97—106. [14] Oliva A, Della Ragione F, Salerno A et al. Biocompatibility studies on glass ionomer cements by primary cultures of human osteoblasts. Biomaterials 1996;17:1351—6. [15] Oliva A, Della Ragione F, Fratta M, Marrone G, Palumbo R, Zappia V. Effect of retinoic acid on osteocalcin gene expression in human osteoblasts. Biochim Biophys Res Commun 1993; 191:908—14. [16] Wuthier RE, Register TC. Role of alkaline phosphatase, a polyfunctional enzyme, in mineraling tissues. In: Butler WT, editor. The chemistry and biology of mineraling tissues. EBSCO Media, Birmingham, 1985:113—24. [17] Beertsen W, van den Bos T. The role of alkaline phosphatase in the mineralization of collagenous substrates. In: Davidovitch Z, editor, The biological mechanisms of tooth movement and craniofacial adaptation. Columbus, 1992:181—5. [18] Slater TF, Sawyer B, Strauli U. Studies on succinate-tetrazolium reductase systems. III. Points of coupling of four different tetrazolium salts. Biochim Biophys Acta 1963;77:383—93. [19] Lian JB, Stein GS, Owen TA et al. Regulation of osteocalcin gene expression during growth and differentiation of the bone cell phenotype. In: Davidovitch Z, editor, The biological mechanisms of tooth movement and craniofacial adaptation. Columbus, 1992:159—79. [20] Ducy P, Desbois C, Boyce B et al. Increased bone formation in osteocalcin-deficient mice. Nature 1996;382:448—52. [21] Thompson JG, Puleo DA. Effects of sub-lethal metal ion concentrations on osteogenic cells derived from bone marrow stromal cells. J Appl Biomater 1995;6:249—58. [22] Thompson JG, Puleo DA. Ti-6Al-4V ion solution inhibition of osteogenic cell phenotype as a function of differentiation timecourse in vitro. Biomaterials 1996;17:1949—54. [23] Sousa SR, Barbosa MA. Effect of hydroxyapatite thickness on metal ion release from Ti6Al4V substrates. Biomaterials 1996;17:397—404. [24] Folkman J, Moscona A. Role of cell shape in growth control. Nature 1978;273:345—9.