499
Space-~active biomatemials in osteoblast cultu~s: an ul~astructural study J.M. Sautier, J.R. Nefussi and N, Forest
Laboratoire de Bi~logie-~~o~to~ogie, University Paris WI, Paris Cede& Frence
The tissua/biomaterial interface reactions of three biomaterials selected as candidates for hard tissue replacement were studied at the electron microscopical level after incubation with enzymatically isolated rat bone cells. An electron-dense layer was routinely observed between hydroxyapatite, coral, cytodex polymer and the neighbouring cells. This layer was visible before bone formation occurred, and was collagen free. The ultrastructural features revealed a needleshaped fiiamentous layer continuous with coral material, whereas hydroxyapatite or cytodex/ tissue interface was granular in appearance. These different structures may indicate reactive surfaces, depending on the composition of the substrate. Ke~wof~s:
Cell-~a~erjal
Received
June 1991; revised 22 July 1991; accepted 3 September 1991
IO
interface,
osfeobiasts,
During the last decade, a large number of biomaterials have been proposed as artificial bone fillers for repairing bone defects. Among these materials, a distinction can be made between bioinert and bioactive materials such as bioglasses and calcium-phosphate ceramics which have the capacity to achieve direct bond with bone. Reports following implantation of bioactive materials described the formation of an electron-dense layer at the interface bone/implant ‘*‘. However, little is known about the formation and the origin of this structure. The initial event at the interface between bioactive materials and neighbouring cells is of fundamental importance. However, thein vivu procedure did not allow examination of the response of specific cells to the substratum. Therefore, to characterize the interactions between bioactive materials and the adjacent tissue, cell cultures provided a good system. The purpose of this study was to use a bone cell culture model developed in our laboratory3 in the presence of coral granules, synthetic hydroxyapatite particles and polymer beads, and to study the so called ‘bone-bonding zone’ at the ultrastru~tural level, ~ATERIAL§
AND METHODS
Bone cells were isolated from fetal rat calvaria as previously described3. The parietal and frontal bones of 21-d Sprague Dawley rat fetuses were dissected, taking care to remove all the suture areas, Bone fragments were digested for 2 h at 37°C in phosphate buffered solution (PBS) with 2.5 mgiml collagenase (Sigma). After this Correspondence Biomaterials
to Dr J.M. Sautier
1992,
Vol. 13 No. 6
ultfastfuctufe
period, the cells were dissociated from the bone by pipetting, washed three times in PBS with 10% fetal calf serum (FCS, Boehringer Mannheim) and cultured in 100 mm diameter culture dishes in Dubelcco’s Modified Eagle Medium (DMEM), containing 10% FCS, 50 p g/ml ascorbic acid, 10 mM ~-glycerophosphate (Sigma) and antibiotics (50 units/ml penicillin, 50 ,ug/ml streptomycin, Gibco). 5 X lo5 cells were plated on 50 mm diameter culture dishes containing coral granules (95% calcium carbonate in the aragonite form) (Inoteb, France), commercial hy~oxyapatite (HA) particles (Pred, France) or cytodex-3 mic~ca~iers (Pharmacia, Upsalen, Sweden) which consisted of a denatured collagen layer covently linked to a dextran polymer matrix. On day 9 and 12 of culture the cells were prepared for transmission electron microscopy, following classical procedures. Briefly, cells were fixed in situ in Karnovsky solution (4% parafo~aldehyde, 1% glutaraldeh~de) for 1 h. After several rinses in 0.2 M sodium cacodylate buffer (pH 7.4) the cells were post-fixed in phosphate-buffe~d solution of osmium tetroxyde (170, pH 7.4). The cells were then dehydrated in graded alcohols and embedded in Epon, Ultrathin sections were then obtained with a diamond knife perpendicularly through nodules without decalcification and stained with uranyl acetate and lead citrate. RESULTS
Observations of g-day-old culture undecalcified sections, revealed at the electron microscopic level the presence of three different structures on the material surfaces. Q 1992 8u~e~orth-Heinemann 0142-9612/92/060400-03
Ltd
fn vitro surface-reactive
biomaterials:
J. M. Sautier et al.
However, in both cases this interface was visible before bone formation occurred and appeared like an electrondense collagen free layer. An electron-dense layer of variable thickness was present on the surface of the coral without interposition of soft tissue (Figure 2). This structure was composed of needle-like materials, densely packed and continuous with the aragonite crystals of coral. However, due to the dense homogeneous structure of this material the interface between the electron-dense band and coral was not clearly visible. In some places, the plasmic membrane of osteoblasts was in contact with the mineralized layer, whereas further outside a dense collagenous matrix was visible. Ultrastructural observation of HA revealed the presence of a loose grouping of individual crystal separated by electron-lucent spaces. An electmn-dense layer of z 300 nm wide was visible at the periphery of the HA mass (Figure 2). In addition, this structure was composed of a granular material, devoid of detectable matrix and an amorphous granular material extended between the synthetic HA crystals to the central part of the HA mass. Between the granular electron-dense layer and the plasmic membrane of an osteoblast, a mineralized matrix composed of needle-shaped crystals was deposited on the first formed electron-dense layer. A prominent structure characterized the cytodex surface (Figure 3). This structure appeared like an amorphous electron-dense layer, 200 nm wide and granular in appearance. Between this layer and the plasmic membrane of osteoblasts, a collagen matrix has developed.
401
Figure2 On day 9 of culture, around the HA crystal (open arrows) mass, a granular electron-dense layer (arrows) is visible. Between the plasmic membrane of an osteoblast (0) and the electron-dense layer, a bone mineralized matrix (MM) is noted (x31000)
Figure3 On day 9 of culture, a granular electron-dense layer (arrows) is visible at the surface of the cytodex (Cyt) polymer. Between this structure and the plasmic membrane of an osteoblast (0) a collagen matrix (asteriscs) has developed (X1000)
In both cases, once the initial mineralized layer was formed, bone formation occurred following a classical pattern of calcification with an intimate contact with the materials. Figure 4 showed the coral/bone interface on day 12 of culture and the relationships between osteoblasts to osteoid matrix and osteoid to bone mineralized matrix. Figure1 On day 9 of culture, an electron-dense layer (arrows) composed of needle-shaped filaments is visible on the coral (C) surface. A cell process of an osteoblast (0) is in contact with the electron-dense layer and further outside a collagenous matrix (CM) is noted (X31000)
-
DISCUSSION The bioactive properties of implant surfaces are of fundamental importance in a successful bone bonding.
_Biomaterials
1992, Vol. 13 No. 6
402
in vitro surface-reactive
Figure 4
On day 12 of culture, laying the plasmic membrane of an osteoblast (0), an osteoid matrix (OS) has developed. A mineralized front (arrow) is noted between the osteoid matrix (OS) and the mineralized matrix (MM) which is in contact with coral (C) biomaterial (white arrows) (X2000)
In the present study, three materials of different compositions showed the formation of a mineralized layer on their surfaces, without collagen matrix formation and before bone apposition. The fo~ation of an electrondense layer have been f~quently described at the interface between bone and HA’, ‘, 4. However, this structure was only observed on decalcified sections and recently, Daculsi et al.’ regarded this electron-dense layer as an artifact introduced during microscopy preparation processes (e.g., decalcification). In the present study, this layer was observed on undecalcified sections and for the first time, before or at the onset of bone formation. A partial dissolution of the calciumphosphate ceramics have been reported following implantation of calcium-phosphate ceramics in osseous sites”. It can be speculated that leached substances from the surface of the material increased calcium and phosphate ion concentration allowing a mineralization process and mediated the bonding of the implant to bone. Such biodegradation was also observed with bioglasses, resulting in the formation of a silica-Ca/P rich layer which mineralized at the surface of the implant7. A different process probably occurred at the coral/ tissue interface. The needle-shaped filaments observed on the coral surface has morphological appearance with biological apatite crystallites. Coral material could provide a bioactive surface and might act as a nucleating centre for mineralization. The third material was a polymer and mineralization of its surface arised from a different phenomenon and cannot be explained by the presence of calcium and phosphorus in the material. It is generally accepted that protein adsorption occurred on material surfaces’. However, when incubated without cells, mineralization
Biomaterials
1992,Vol. 13 No. 6
biomaterials:
J.M. Saufier ef a/.
of the cytodex surface did not occur, showing that the formation of the electron-dense layer was an active cellular biological processg. Specific bone non-collagenous proteins could be involved in the formation of this interface, since it has been shown that polyanionic proteins immobilized on a stable support induced mineral in vitro”‘. Due to their Cazf binding properties, polyanionic proteins such as proteoglycans, phosphoproteins or osteocalcin could be absorbed on the cytodex surface and provide nucleation sites which initiated mineral formation. The use of bone cell cultures appears of great interest in investigating the step-by-step surface-reactivity of biomaterials. In this study, it was shown that when various bioactive materials were incubated with rat bone cells, the resulting interface consisted of an electrondense layer before bone formation occurred. However, further investigations appeared necessary to identify the organic and physical components of these structures.
REFERENCES 1
2
3
4
5
6
7 8
9
10
Ganeles, J., Listga~en, M.A. and Evian, C., Ult~st~ctura of durapatite-periodonial tissue interface in human intrabony defects, 1. Periodontal. 1986, 57 (3), 133-141 Van Blitterswijk, CA., Grote, J.J., Kruypers, W., Blok van Hoek, C.J.G. and Deams, W.T., Bioreactions at the tissue/ hydroxyapatite interface, EGomaterials 1985,6,243-251 Nefussi, J.R., Boy-Lefevre, M.L., Boulekbache, H. and Forest, N., Mineralization in vitro of matrix formed by collagenase digestion, osteoblasts isolated by Differentiation 1985, 29, 160-168 Van Blitterswijk, C.A., Grote, J.J.. Krujipers, W., Ileams, W.T. and de Groot, K., Macropore tissue ingrowth: a quantitative and qualitative study on hydroxyapatite ceramic, ~io~aterial 1986, 7, 137-143 Daculsi, G., Le Geros, R. and Deudon, C., Scanning and transmission electron microscopy and electron probe analysis of the interface between implants and host bone, Scanning Microsc. 1990, 4 (2), 309-314 Daculsi, G., Le Geros, R., Heughebaert, M. and Barbieux, of carbonate-apatite crystals after I ., Formation implantation of calcium phosphate ceramics, Calcif. Tissue Znt. 1990, 46, 20-27 Hench, L.L., Ceramic glasses and composites in medicine, Med. Znstr. 1973, 7 (21, 136-144 Klein, C.P.A.T., de Groot, K., Vermeiden, J.P.W. and van Kamp, G., Interaction of some serum proteins with hydroxyapatite and other materials, J. Biomed. Mater. Res. 1980,X4, 705-712 Sautier, J.M., Nefussi, J.R. and Forest, N., Mineralization and bone formation on microca~ier beads with isolated rat calvaria cell population, Calcif. Tissue Znt. [in press) Linde, A. and Lussi, A., Mineral induction by polyanionic dentin and bone proteins at physiolqgical ionic conditions, Connect. Tissue Res. 1989, 21, 197-203
Space-~active biomatemials in osteoblast cultu~s: an ul~astructural study J.M. Sautier, J.R. Nefussi and N, Forest
Laboratoire de Bi~logie-~~o~to~ogie, University Paris WI, Paris Cede& Frence
The tissua/biomaterial interface reactions of three biomaterials selected as candidates for hard tissue replacement were studied at the electron microscopical level after incubation with enzymatically isolated rat bone cells. An electron-dense layer was routinely observed between hydroxyapatite, coral, cytodex polymer and the neighbouring cells. This layer was visible before bone formation occurred, and was collagen free. The ultrastructural features revealed a needleshaped fiiamentous layer continuous with coral material, whereas hydroxyapatite or cytodex/ tissue interface was granular in appearance. These different structures may indicate reactive surfaces, depending on the composition of the substrate. Ke~wof~s:
Cell-~a~erjal
Received
June 1991; revised 22 July 1991; accepted 3 September 1991
IO
interface,
osfeobiasts,
During the last decade, a large number of biomaterials have been proposed as artificial bone fillers for repairing bone defects. Among these materials, a distinction can be made between bioinert and bioactive materials such as bioglasses and calcium-phosphate ceramics which have the capacity to achieve direct bond with bone. Reports following implantation of bioactive materials described the formation of an electron-dense layer at the interface bone/implant ‘*‘. However, little is known about the formation and the origin of this structure. The initial event at the interface between bioactive materials and neighbouring cells is of fundamental importance. However, thein vivu procedure did not allow examination of the response of specific cells to the substratum. Therefore, to characterize the interactions between bioactive materials and the adjacent tissue, cell cultures provided a good system. The purpose of this study was to use a bone cell culture model developed in our laboratory3 in the presence of coral granules, synthetic hydroxyapatite particles and polymer beads, and to study the so called ‘bone-bonding zone’ at the ultrastru~tural level, ~ATERIAL§
AND METHODS
Bone cells were isolated from fetal rat calvaria as previously described3. The parietal and frontal bones of 21-d Sprague Dawley rat fetuses were dissected, taking care to remove all the suture areas, Bone fragments were digested for 2 h at 37°C in phosphate buffered solution (PBS) with 2.5 mgiml collagenase (Sigma). After this Correspondence Biomaterials
to Dr J.M. Sautier
1992,
Vol. 13 No. 6
ultfastfuctufe
period, the cells were dissociated from the bone by pipetting, washed three times in PBS with 10% fetal calf serum (FCS, Boehringer Mannheim) and cultured in 100 mm diameter culture dishes in Dubelcco’s Modified Eagle Medium (DMEM), containing 10% FCS, 50 p g/ml ascorbic acid, 10 mM ~-glycerophosphate (Sigma) and antibiotics (50 units/ml penicillin, 50 ,ug/ml streptomycin, Gibco). 5 X lo5 cells were plated on 50 mm diameter culture dishes containing coral granules (95% calcium carbonate in the aragonite form) (Inoteb, France), commercial hy~oxyapatite (HA) particles (Pred, France) or cytodex-3 mic~ca~iers (Pharmacia, Upsalen, Sweden) which consisted of a denatured collagen layer covently linked to a dextran polymer matrix. On day 9 and 12 of culture the cells were prepared for transmission electron microscopy, following classical procedures. Briefly, cells were fixed in situ in Karnovsky solution (4% parafo~aldehyde, 1% glutaraldeh~de) for 1 h. After several rinses in 0.2 M sodium cacodylate buffer (pH 7.4) the cells were post-fixed in phosphate-buffe~d solution of osmium tetroxyde (170, pH 7.4). The cells were then dehydrated in graded alcohols and embedded in Epon, Ultrathin sections were then obtained with a diamond knife perpendicularly through nodules without decalcification and stained with uranyl acetate and lead citrate. RESULTS
Observations of g-day-old culture undecalcified sections, revealed at the electron microscopic level the presence of three different structures on the material surfaces. Q 1992 8u~e~orth-Heinemann 0142-9612/92/060400-03
Ltd
fn vitro surface-reactive
biomaterials:
J. M. Sautier et al.
However, in both cases this interface was visible before bone formation occurred and appeared like an electrondense collagen free layer. An electron-dense layer of variable thickness was present on the surface of the coral without interposition of soft tissue (Figure 2). This structure was composed of needle-like materials, densely packed and continuous with the aragonite crystals of coral. However, due to the dense homogeneous structure of this material the interface between the electron-dense band and coral was not clearly visible. In some places, the plasmic membrane of osteoblasts was in contact with the mineralized layer, whereas further outside a dense collagenous matrix was visible. Ultrastructural observation of HA revealed the presence of a loose grouping of individual crystal separated by electron-lucent spaces. An electmn-dense layer of z 300 nm wide was visible at the periphery of the HA mass (Figure 2). In addition, this structure was composed of a granular material, devoid of detectable matrix and an amorphous granular material extended between the synthetic HA crystals to the central part of the HA mass. Between the granular electron-dense layer and the plasmic membrane of an osteoblast, a mineralized matrix composed of needle-shaped crystals was deposited on the first formed electron-dense layer. A prominent structure characterized the cytodex surface (Figure 3). This structure appeared like an amorphous electron-dense layer, 200 nm wide and granular in appearance. Between this layer and the plasmic membrane of osteoblasts, a collagen matrix has developed.
401
Figure2 On day 9 of culture, around the HA crystal (open arrows) mass, a granular electron-dense layer (arrows) is visible. Between the plasmic membrane of an osteoblast (0) and the electron-dense layer, a bone mineralized matrix (MM) is noted (x31000)
Figure3 On day 9 of culture, a granular electron-dense layer (arrows) is visible at the surface of the cytodex (Cyt) polymer. Between this structure and the plasmic membrane of an osteoblast (0) a collagen matrix (asteriscs) has developed (X1000)
In both cases, once the initial mineralized layer was formed, bone formation occurred following a classical pattern of calcification with an intimate contact with the materials. Figure 4 showed the coral/bone interface on day 12 of culture and the relationships between osteoblasts to osteoid matrix and osteoid to bone mineralized matrix. Figure1 On day 9 of culture, an electron-dense layer (arrows) composed of needle-shaped filaments is visible on the coral (C) surface. A cell process of an osteoblast (0) is in contact with the electron-dense layer and further outside a collagenous matrix (CM) is noted (X31000)
-
DISCUSSION The bioactive properties of implant surfaces are of fundamental importance in a successful bone bonding.
_Biomaterials
1992, Vol. 13 No. 6
402
in vitro surface-reactive
Figure 4
On day 12 of culture, laying the plasmic membrane of an osteoblast (0), an osteoid matrix (OS) has developed. A mineralized front (arrow) is noted between the osteoid matrix (OS) and the mineralized matrix (MM) which is in contact with coral (C) biomaterial (white arrows) (X2000)
In the present study, three materials of different compositions showed the formation of a mineralized layer on their surfaces, without collagen matrix formation and before bone apposition. The fo~ation of an electrondense layer have been f~quently described at the interface between bone and HA’, ‘, 4. However, this structure was only observed on decalcified sections and recently, Daculsi et al.’ regarded this electron-dense layer as an artifact introduced during microscopy preparation processes (e.g., decalcification). In the present study, this layer was observed on undecalcified sections and for the first time, before or at the onset of bone formation. A partial dissolution of the calciumphosphate ceramics have been reported following implantation of calcium-phosphate ceramics in osseous sites”. It can be speculated that leached substances from the surface of the material increased calcium and phosphate ion concentration allowing a mineralization process and mediated the bonding of the implant to bone. Such biodegradation was also observed with bioglasses, resulting in the formation of a silica-Ca/P rich layer which mineralized at the surface of the implant7. A different process probably occurred at the coral/ tissue interface. The needle-shaped filaments observed on the coral surface has morphological appearance with biological apatite crystallites. Coral material could provide a bioactive surface and might act as a nucleating centre for mineralization. The third material was a polymer and mineralization of its surface arised from a different phenomenon and cannot be explained by the presence of calcium and phosphorus in the material. It is generally accepted that protein adsorption occurred on material surfaces’. However, when incubated without cells, mineralization
Biomaterials
1992,Vol. 13 No. 6
biomaterials:
J.M. Saufier ef a/.
of the cytodex surface did not occur, showing that the formation of the electron-dense layer was an active cellular biological processg. Specific bone non-collagenous proteins could be involved in the formation of this interface, since it has been shown that polyanionic proteins immobilized on a stable support induced mineral in vitro”‘. Due to their Cazf binding properties, polyanionic proteins such as proteoglycans, phosphoproteins or osteocalcin could be absorbed on the cytodex surface and provide nucleation sites which initiated mineral formation. The use of bone cell cultures appears of great interest in investigating the step-by-step surface-reactivity of biomaterials. In this study, it was shown that when various bioactive materials were incubated with rat bone cells, the resulting interface consisted of an electrondense layer before bone formation occurred. However, further investigations appeared necessary to identify the organic and physical components of these structures.
REFERENCES 1
2
3
4
5
6
7 8
9
10
Ganeles, J., Listga~en, M.A. and Evian, C., Ult~st~ctura of durapatite-periodonial tissue interface in human intrabony defects, 1. Periodontal. 1986, 57 (3), 133-141 Van Blitterswijk, CA., Grote, J.J., Kruypers, W., Blok van Hoek, C.J.G. and Deams, W.T., Bioreactions at the tissue/ hydroxyapatite interface, EGomaterials 1985,6,243-251 Nefussi, J.R., Boy-Lefevre, M.L., Boulekbache, H. and Forest, N., Mineralization in vitro of matrix formed by collagenase digestion, osteoblasts isolated by Differentiation 1985, 29, 160-168 Van Blitterswijk, C.A., Grote, J.J.. Krujipers, W., Ileams, W.T. and de Groot, K., Macropore tissue ingrowth: a quantitative and qualitative study on hydroxyapatite ceramic, ~io~aterial 1986, 7, 137-143 Daculsi, G., Le Geros, R. and Deudon, C., Scanning and transmission electron microscopy and electron probe analysis of the interface between implants and host bone, Scanning Microsc. 1990, 4 (2), 309-314 Daculsi, G., Le Geros, R., Heughebaert, M. and Barbieux, of carbonate-apatite crystals after I ., Formation implantation of calcium phosphate ceramics, Calcif. Tissue Znt. 1990, 46, 20-27 Hench, L.L., Ceramic glasses and composites in medicine, Med. Znstr. 1973, 7 (21, 136-144 Klein, C.P.A.T., de Groot, K., Vermeiden, J.P.W. and van Kamp, G., Interaction of some serum proteins with hydroxyapatite and other materials, J. Biomed. Mater. Res. 1980,X4, 705-712 Sautier, J.M., Nefussi, J.R. and Forest, N., Mineralization and bone formation on microca~ier beads with isolated rat calvaria cell population, Calcif. Tissue Znt. [in press) Linde, A. and Lussi, A., Mineral induction by polyanionic dentin and bone proteins at physiolqgical ionic conditions, Connect. Tissue Res. 1989, 21, 197-203
