J. BIOMED. MATER. RES.
VOL. 10, PP. 161-174 (1976)
The Influence of Surface Chemistry on Implant Interface Histology: A Theoretical Basis for Implant Materials Selection* A. E. CLARK and L. L. HENCH, Department of Materials Science and Engineering, University of Florida, Gainesville, Florida, and H. A. PASCHALL, Department of Surgery, University of Florida, Chief of Orthopaedics, Gainesville V . A . Hospital, Gainesville, Florida
Summary A theory is proposed stating that an ideal implant material must have a dynamic surface chemistry that induces histological changes a t the implant interface which would normally occur if the implant were not present. Evidence for the validity of this theory is provided with a series of bone-implant studies which result in stable interfacial osteogenesis under specific implant surface chemistry conditions. Insufficient or excess surface ion concentrations produce negative osteogenesis and fixation results. Implantation of osteogenic implants in soft tissues also produces undesirable histological responses as proposed in the theory. A variety of surface chemical analyses of the implant are reviewed which provide a scientific basis for the implant surface theory.
THEORY The following theory is proposed for implant materials design and selection. An ideal implant material must have a dynamic surface chemistry that induces histological changes a t the implant interface which would normally occur if the implant were not present. This study examines the histological evidence for this theory in hard and soft tissues and describes the surface chemical basis of the theory. EVIDENCE FROM HARD TISSUE STUDIES Long experience with metallic implants in bone has led to the general observation of the development of a nonadherent fibrous cap*This paper was presented a t the Sixth Annual Symposium on Biomaterials, Clemson University, Clemson, South Carolina, April 1974.
161
162
CLARK, HENCH, AND PASCHALL
sule around the implant of varying thickness dependent upon the type of metal, implant location, geometry, stability of fixation, etc. The surface chemistry of the metallic implants is such that if reactions do occur, both the pH change and the ions leached into the physiological solutions are not normally associated with development of either immature or mature osteoid. As a consequence, interfacial osteogenesis does not occur. In contrast, a series of studies conducted on Ca, P, Na, and SiOz containing bioglasses and bioglass-ceramic implants show that these materials exhibit stable interfacial osteogenesis when appropriate combinations of implant surface chemistry and implant sites are These studies provide positive evidence for the theory advanced above. A summary of the evidence includes the following data. 1) The ions released from the implant surface are required for osteogenesis. (See Table I for implant compositions.) 2) Time dependent pH changes occurring at the implant interface are equivalent to the changes necessary for osteogenesis. 3) Ultrastructural changes of the implant surface provide a medium for collagen fibril and mucopolysaccharide bonding with an inorganic gel and mineralization to occur. 4) Hydroxyapatite precipitation takes place a t the implant-osteoid interface simultaneous with the precipitation of hydroxyapatite crystals away from the implant interface. 5 ) Mechanical loads applied to the bioglass-ceramic implantbone interface, either by chipping, microtoming, or torsional fracture of segmental replacements, always break away from the interface. Fracture lines stop at the interface. TABLE I Composition of Bioglass Implants Used for Developing Theory Weight Percentage Code No. 1. 2. 3. 4. 5. 6.
4.555 45S5F 45B5S5 45Sm 45S10 4592.5
SiO,
NazO
45 43 40 45 45 45
24.5 23 24.5 30.5 27.5 18.5
CaO 24.5 12 24.5 24.5 24.5 24.5
CaFz
16
PzO, 6.0 6.0 6.0 3 12
Bz03
5.0
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
163
If compositional alterations of the bone-bonding bioglass-ceramics are made, the surface chemistry dynamics change as will be discussed later. These changes lead to the following evidence in support of the proposed theory. 1) A slower reacting implant surface when placed midshaft in old femoral cortical bone in the rat (slow metabolic rate) produced a fibrous encapsulation; an implant of the same composition when placed in the metaphyseal region of young rat tibiae (high metabolic rate) induced interfacial osteogenesis and bonding. 2) Implants placed in the young rat metaphyseal region accompany the adjacent bone as growth occurs. 3) Altering the interfacial pH-time dependence by a factor of 100 does not significantly alter the osteogenesis as long as the interface becomes alkaline within five to six weeks. 4) A deficiency of phosphate ions released from the implant surface decreases osteogenic induction. 5 ) Implant surfaces that are too reactive and release a surplus of phosphate ions result in ectopic calcification of neighboring tissues and histological attack of the implant. Evidence for points 4) and 5 ) has been obtained from a new series of in. vivo experiments. Compositions 1, 4,5, and 6 (see Table I) were selected to study the influence of phosphorus additions on the behavior of bioglass implants. Samples of bioglass 5 x 5 X 1 mm were placed in defects created in the metaphysis of the tibia just distal to the epyphyseal plate of Sprague-Dawley male rats. The limbs were not immobilized and the animals were sacrificed at three and eight weeks. The tibiae were dissected clean of all soft tissues and the area of bone surrounding the bioglass was cut into 1 mm thick slices with bone on either side of the glass. These slices of bone and glass were then placed immediately in cold cacodylate buffered gluteraldehyde, fixed for 2 hr in the cold, and then washed with fresh cold buffer. The tissue sections were then placed in 2% osmium tetraoxide buffered at a pH of 7.4 and fixed for an additional hour. The tissue was then washed with an additional buffer and then rapidly dehydrated in graded alcohols and embedded in Epon 812. The blocks were then appropriately trimmed and sectioned on a Poster-Blum MT2 ultra microtome using glass knives.' Thick sections (1.0 pm) were prepared and stained with Richardson's methylene blue azure I1 or H & E stains with equivalent results.
164
CLARK, HENCH, AND PASCHALL
(b) (a) Photomicrograph of a 45s m bioglass-bone interface at 8 weeks (b) Transmission electron micrograph of a 455 m bioglass-bone interface at 3 weeks (43,700X).
Fig. 1. (1,320X).
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
165
A diamond knife was used to cut the thin sections. Sections were poststained with saturated fresh alcoholic uranyl acetate and lead citrate. Figure l a is a light microscope section of a bone-45Sm bioglass interface at eight weeks. The thin layer of black material (G) lining the edge of the bone may be the silica rich gel layer which forms on the glass implant surface (I),as has been observed in the in vitro analyses. Small pieces of glass implant (I) are attached to the gel layer. Only a small density of viable osteocytes (0)are present in the newly formed bone (B). The osteoid front (OF) is characterized by a lack of active bone formation and very few active osteoblasts. Figure lb is a transmission electron micrograph of a three week bone-45S co bioglass interface. The material which exhibits the regular fracture pattern (GF) appears to be the silica-rich gel layer described above. The evidence in support of this conclusion follows. 1) The gel layer formed in vitro on the 45Sm glass is very soft and nonadherent. Thus, the glass implant particles (I) in Figure la, should break away during final microtomy of the section. None are seen in any TEM sections. Also, the relative softness of the gel should produce a uniform fracture pattern such as seen a t GF, with long nonbranching fracture lines protruding deep within the gel layer. 2) The thickness of the gel layer in Figure l b is 1-2 pm, measured to the end of the fracture lines and the region where the density of collagen fibers is low (GE). This is the same magnitude of the thickness seen in Figure l a . In vitro studies have shown that there is relative stability of the gel film thickness between the three to eight week period. Figure 2a is a photomicrograph of a 4585 glass-bone interface. I n contrast to Figure l a , large pieces of bioglass (I) are intimately attached t o bone and several normal osteocytes (0) are present in the mineralized area. There is a well-defined layer of osteoblasts actively engaged in laying down new bone (OF) and this front is separated from the mineralized area by a transition zone of partially mineralized bone. These features indicate that inducement of the normal ossification process has been achieved. An electron micrograph illustrating a 4585 glass-bone interface a t six weeks is shown in Figure 2b. A layer of amorphous material (2000-3000 A thick) separates the glass implant I from mineralized bone B. Hydroxyapatite crystals which extend across the zone connect the implant to the bone structure.
166
CLARK, HENCH, AND PASCHALL
(b) Fig. 2 . (a) Photomicrograph of a 4585 bioglass-bone interface a t 3 weeks (1300X). (b) Transmission electron micrograph of a 4585 bioglass-bone interface a t 6 weeks (22,lOOX).
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
167
The effect of an excessive amount of phosphorus in a bioglass is illustrated in Figure 3. Figure 3 is a photomicrograph of a 4552.5 glass-bone interface a t eight weeks. An important feature to note is that the implant I has been separated from the bone B by an interval containing a capillary (C) which on electron microscopy reveals intercellular crystallization (X) induced by the excess phosphorus. It was observed that the gel layer (G) remained attached to the bone when the interval containing the capillary separated the glass (1)-gel layer interface. Note also the unhealthy appearance of the osteocytes (0). They have withdrawn from their lacuner walls and the nuclei are pyknotic. There is also a n absence of new bone formation a t the osteoid front (OF). As added evidence for the importance of implant surface chemistry on osteogenesis, a series of uncrystallized, partially crystallized, and nearly fully crystallized 4585 implants were compared as rat femoral implants. Equivalent osteogenesis and interfacial bonding was observed for all three microstructures and crystallographic states. In vitro analyses showed nearly equivalent surface chemical behavior due t o the presence of a residual glassy phase left after crystallization.
EVIDENCE FROM SOFT TISSUE RESPONSES Whenever muscle is injured or foreign material is implanted into muscle, healing occurs by formation of a collagenous scar or fibrous encapsulation of the implant without direct attachment t o the implant. Bone induction in muscle has not been successful, although autogenous bone grafts implanted in muscle are revascularized and new bone is formed within the graft. Therefore, a n ideal implant for muscle or other soft tissues which behave in the same fashion, i.e., tendon, ligaments, or skin, would be one which induces a collagenous scar with direct attachment to the implant. With these facts in mind, several possibilities of the type of response to the implants in muscle would be: 1) induction of bone formation around the implant thus securely anchoring i t to the muscle. 2) fibrous tissue encapsulation of the implant without attachment ; 3) fibrous tissue encapsulation with direct collagen bonding to the implant; 4) encapsulation and bursa formation about the implant; or 5) tissue rejection with acute and chronic inflammatory response with sterile abcess formation and extrusion of the implant.
168
CLARK, HENCH, AND PASCHALL
Fig. 3. Photomicrograph of a 4582.5 bioglass-bone interface at 8 weeks (1320X ) .
Recognizing the above possible physiological reactions to an implant, a study was done t o determine the histological response of muscle to various compositions of bioglass-ceramic Small pieces of each material that exhibit variable surface reactivity (Table I) measuring 1 x 1 x 0.1 cm were implanted under aseptic techniques in the lateral thigh muscles of male Sprague-Dawley rats. The animals were sacrificed a t three, six, and 16 weeks. The bioglass and surrounding muscle were removed as a block and fixed in cold cacodylate buffered (pH 7.4) gluteraldehyde, dehydrated in graded alcohols and propylene oxide, and embedded in Epon 812. The blocks were sectioned for light and electron microscopy. All of the compositions evoked a similar tissue response which varied quantitatively according to the surface reactivity of the material. At three weeks, there was a resolving inflammatory reaction with early fibrous encapsulation of the implants. Figure 4a is a photomicrograph of a 4585 glass-muscle interface at three weeks. Several layers of elongated fibroblasts (F) are separating the glass implant (I) from the muscle tissue (MT). An electron micrograph of the glass cell interface (Fig. 4b) reveals a single layer of macrophagelike cells (MC) between the glass implant (I) and the fibroblasts (F).
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
169
(b) Fig. 4. (a) Photomicrograph of a 4555 bioglass-muscle interface at 3 weeks (1,670X). (b) Transmission electron micrograph of the area adjacent to the implant surface of (a) (47,500X).
170
CLARK, HENCH, AND PASCHALL
These cells have attacked and are altering the implant surface. There is a band three or four cell layers thick where cells have ingested bioglass (IB) within cisternae. At six weeks, there was a synovial membrane indentifiable about all implants. A t 16 weeks, there was evidence of continued attack of the implants by the synovial cells and phagocytosis of the ceramic particles. The less surface reactive materials showed the slowest development of attack. These findings demonstrate that although certain bioglass-ceramic compositions (4555, 45B &5, and 45S5F) fulfill the criteria of a n ideal implant material for bone, in that they induce normal bone formation with direct attachment to the implant, they do not fulfill the criteria of an ideal implant material in soft tissues. The reactive surfaces of the materials release into the microenvironment surrounding the implant Ca and P ions which ordinarily are not found in muscle in any significant quantities. Therefore, the materials evoke an inflammatory reaction which continues into a form of rejection. A similar form of rejection for the same reasons has been found in equivalent implants in porcine tendon.
DYNAMIC SURFACE CHEMISTRY EVIDENCE The histological reactions of bioglasses and bioglass-ceramics are significantly influenced by composition as indicated in both hard and soft tissue responses. Therefore, it is important to define the roles of the individual ionic species in the surface chemical reactions of these materials in order t o establish a surface chemical theory of response. Conventional silicate glasses have a continuous three-dimensional structure. The basic repeat unit is a tetrahedran with four oxygen atoms at the corners and a silicon atom a t the center. Oxygen atoms which are common t o two tetrahedra are referred to as bridging, since they provide the linkage necessary to extend the structure. Incorporation of calcium and sodium ions modifies the structure by tying up oxygen atoms which can no longer serve as a bridge between neighboring tetrahedra. Due to their low silica (45 wt yo)and high modifier content the normal structure of glass is drastically altered in bioglasses. The average structure consists of two-dimensional chains with 4 to 20
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
171
tetrahedra per chain.g The sodium and calcium act as ionic bridges between these chains. Hence, their role as covalent network modifier has been reversed to that of ionic network formers providing additional stability to the structure. When bioglasses are exposed to an aqueous environment, sodium and calcium are preferentially leached from the surface, leaving behind a silica-rich layer.6 Because of the low silica content of the glass, the structure of this layer for certain compositional ranges may be relatively open and adherence to the bulk glass is fragile. The addition of phosphorus significantly alters the corrosion process.1° Initially a silica-rich layer is produced. However, this layer is followed by an amorphous calcium phosphate layer which crystallizes with time.lOJ1 Increasing the phosphorus content reduces the time required for the calcium phsophate layer to form. Furthermore, the adherence of the calcium phosphate layer is superior to that of the silica-rich layer. Figure 5 is a composition profile of the surface of a 4535 bioglass exposed to an aqueous solution buffered at a pH of 7.4 and maintained at 37°C for 1 hr. The profile was obtained by ion milling the
1
Atomic Percsnt
'"t
2oD 0
0
1
2
3
4
5
Depth From L r f a c e ci2[04
Fig. 5 .
'-Silica
Rich-Diffuse
6
7
8
Cwl
I n t e r f a c e d B u l k Glass
Composition profile of a corroded 45S5 bioglass surface.
9
172
CLARK, HENCH, BND PASCHALL
surface, atom layer by atom layer, while simultaneously analyzing the surface atoms with Auger spectroscopy.'l The results confirm the formation of a silica rich gel layer which is followed by a Ca-PO4 rich layer. At 1 hr the Ca-P04layer has just started to develop and is only 0.4 pm thick. The silica-rich gel layer beneath is 2.25 pm thick. Additional aqueous exposure in vitro increases the Ca-PO4 layer to as much as 10 pm. The calcium fluoride and boron oxide compositional modifications also fit this same analysis of surface reactions.1° During the initial corrosion reaction, calcium fluoride additions to a bioglass containing phosphorus increase the surface silica concentration, while boron oxide additions reduce it. Calcium fluoride and boron oxide also exert a rate controlling influence on the formation of the calcium phosphate layer. Calcium fluoride accelerates the process while boron retards it. In effect, calcium fluoride in the implant surface accelerates the formation of a resistant surface film which retards further reactions, while boron oxide delays the same process. The consequence of the sequence of surface chemical reactions described above is osteogenesis. The sodium ion release from the
Mineralized Bone
Osteoid
Surface Layer
Bulk Class
Fig. 6. A summary of the influence of surface chemistry on bioglass implant reactions with hard tissue.
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
173
surface eventually overrides the effect of acidic enzymes associated with wound healing and bone repair and permits a local alkaline p H t o be maintained. Osteoblasts differentiate in the implant vicinity. The silica-rich gel and amorphous calcium phosphate layer produced on the implant surface incorporates collagen fibrils and mucopolysaccharides generated by the osteoblasts. Crystallization of the calcium phosphate layer proceeds simultaneous with mineralization within the collagen fibrils resulting in a cojointly crystallized boneimplant junction. A summary of the structure of the interfacial bonding in hard tissue based on the histological and surface chemistry studies is shown in Figure 6. A thin (500 A) zone of co-crystallization within the inorganic gel and the collagen produces the bonding. If insufficient phosphorous is present, a thick (2000 8) zone of incomplete crystallization is developed which is weak and unstable. With excess phosphorus content, cellular rejection of the implant occurs by macrophage-like destruction of the surface accompanied by ectopic calcification of tissues in the vicinity of the implant. If insufficient surface reactivity is present osteogenesis does not proceed, fibroblastic cellular membranes envelop the surface hyrolysis product, and fibrous capsule develops. Application of the successful development of segmental bone grafts using this theory is presented in another study.12 Use of the concepts in orthopedic joint fixation13 and design of soft tissue studies appear in other reports.?
References 1. L. L. Hench, R. J. Splinter, W. C. Allen, and T. K. Greenlee, Jr., J . Biomed. Res. Symp., No. 2, 6 , 117 (1972). 2. L. L. Hench, T. K. Greenlee, Jr., and W. C. Allen, Reports # 1, #2, #3, August 1970, 1971, and 1972, U. S. Army Med. R & D Contract No. DADA17-70-C-0001. 3. T. K. Greenlee, J r , C A. Beckham, A. R. Crebo, and J. C. Malmborg, J . Biomed. Muter. Res., 6 , 244 (1972). 4. C. A. Beckham, T. K. Greenlee, Jr., and A. R. Crebo, J. Calcified Tissue Res., 8. 2 (1971). 5. L. L. Hench and H. A. Paschall, J. Biomed. Mater. Res. Symp. No. 5, 8, 49 (1974). 6. L. L. Hench and H. A. Paschall, J . Biomed. Res. Symp., No. 4 , 7, 25 (1973). 7. H. A. Paschall, M. M. Rodabush, and J. T. McVey, Report #3, August 1972, U. S. Army Med. R & D Contract No. DADA-17-70-C-0001.
174
CLARK, HENCH, AND PASCHALL
8. H. A. Paschall, J. McVey, and M. Rodabush, Orthopaedic Research Society, in preparation. 9. H. J. L. Trap and J. M. Stevels, Glastechn. Ber. 32K, VI/31 (1959). 10. A. E. Clark and L. L. Hench, J. A m r . Ceram. Soc., (1974), in press. 11. C. G. Pantano, A. E. Clark, and L. L. Hench, J . Amer. Ceram. Soc., 57(9), 412 (1974). 12. G. Piotrowski, L. L. Hench, W. C. Allen, and G. J. Miller, J. Biomed. Mater. Reo. Symp. No. 6, 9, 47 (1975). 13. L. L. Hench, H. A. Paschall, W. C. Allen, and G. Piotrowski, National Bureau of Standards Special Publication 415, 19 (1975).
Received April 16, 1975
VOL. 10, PP. 161-174 (1976)
The Influence of Surface Chemistry on Implant Interface Histology: A Theoretical Basis for Implant Materials Selection* A. E. CLARK and L. L. HENCH, Department of Materials Science and Engineering, University of Florida, Gainesville, Florida, and H. A. PASCHALL, Department of Surgery, University of Florida, Chief of Orthopaedics, Gainesville V . A . Hospital, Gainesville, Florida
Summary A theory is proposed stating that an ideal implant material must have a dynamic surface chemistry that induces histological changes a t the implant interface which would normally occur if the implant were not present. Evidence for the validity of this theory is provided with a series of bone-implant studies which result in stable interfacial osteogenesis under specific implant surface chemistry conditions. Insufficient or excess surface ion concentrations produce negative osteogenesis and fixation results. Implantation of osteogenic implants in soft tissues also produces undesirable histological responses as proposed in the theory. A variety of surface chemical analyses of the implant are reviewed which provide a scientific basis for the implant surface theory.
THEORY The following theory is proposed for implant materials design and selection. An ideal implant material must have a dynamic surface chemistry that induces histological changes a t the implant interface which would normally occur if the implant were not present. This study examines the histological evidence for this theory in hard and soft tissues and describes the surface chemical basis of the theory. EVIDENCE FROM HARD TISSUE STUDIES Long experience with metallic implants in bone has led to the general observation of the development of a nonadherent fibrous cap*This paper was presented a t the Sixth Annual Symposium on Biomaterials, Clemson University, Clemson, South Carolina, April 1974.
161
162
CLARK, HENCH, AND PASCHALL
sule around the implant of varying thickness dependent upon the type of metal, implant location, geometry, stability of fixation, etc. The surface chemistry of the metallic implants is such that if reactions do occur, both the pH change and the ions leached into the physiological solutions are not normally associated with development of either immature or mature osteoid. As a consequence, interfacial osteogenesis does not occur. In contrast, a series of studies conducted on Ca, P, Na, and SiOz containing bioglasses and bioglass-ceramic implants show that these materials exhibit stable interfacial osteogenesis when appropriate combinations of implant surface chemistry and implant sites are These studies provide positive evidence for the theory advanced above. A summary of the evidence includes the following data. 1) The ions released from the implant surface are required for osteogenesis. (See Table I for implant compositions.) 2) Time dependent pH changes occurring at the implant interface are equivalent to the changes necessary for osteogenesis. 3) Ultrastructural changes of the implant surface provide a medium for collagen fibril and mucopolysaccharide bonding with an inorganic gel and mineralization to occur. 4) Hydroxyapatite precipitation takes place a t the implant-osteoid interface simultaneous with the precipitation of hydroxyapatite crystals away from the implant interface. 5 ) Mechanical loads applied to the bioglass-ceramic implantbone interface, either by chipping, microtoming, or torsional fracture of segmental replacements, always break away from the interface. Fracture lines stop at the interface. TABLE I Composition of Bioglass Implants Used for Developing Theory Weight Percentage Code No. 1. 2. 3. 4. 5. 6.
4.555 45S5F 45B5S5 45Sm 45S10 4592.5
SiO,
NazO
45 43 40 45 45 45
24.5 23 24.5 30.5 27.5 18.5
CaO 24.5 12 24.5 24.5 24.5 24.5
CaFz
16
PzO, 6.0 6.0 6.0 3 12
Bz03
5.0
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
163
If compositional alterations of the bone-bonding bioglass-ceramics are made, the surface chemistry dynamics change as will be discussed later. These changes lead to the following evidence in support of the proposed theory. 1) A slower reacting implant surface when placed midshaft in old femoral cortical bone in the rat (slow metabolic rate) produced a fibrous encapsulation; an implant of the same composition when placed in the metaphyseal region of young rat tibiae (high metabolic rate) induced interfacial osteogenesis and bonding. 2) Implants placed in the young rat metaphyseal region accompany the adjacent bone as growth occurs. 3) Altering the interfacial pH-time dependence by a factor of 100 does not significantly alter the osteogenesis as long as the interface becomes alkaline within five to six weeks. 4) A deficiency of phosphate ions released from the implant surface decreases osteogenic induction. 5 ) Implant surfaces that are too reactive and release a surplus of phosphate ions result in ectopic calcification of neighboring tissues and histological attack of the implant. Evidence for points 4) and 5 ) has been obtained from a new series of in. vivo experiments. Compositions 1, 4,5, and 6 (see Table I) were selected to study the influence of phosphorus additions on the behavior of bioglass implants. Samples of bioglass 5 x 5 X 1 mm were placed in defects created in the metaphysis of the tibia just distal to the epyphyseal plate of Sprague-Dawley male rats. The limbs were not immobilized and the animals were sacrificed at three and eight weeks. The tibiae were dissected clean of all soft tissues and the area of bone surrounding the bioglass was cut into 1 mm thick slices with bone on either side of the glass. These slices of bone and glass were then placed immediately in cold cacodylate buffered gluteraldehyde, fixed for 2 hr in the cold, and then washed with fresh cold buffer. The tissue sections were then placed in 2% osmium tetraoxide buffered at a pH of 7.4 and fixed for an additional hour. The tissue was then washed with an additional buffer and then rapidly dehydrated in graded alcohols and embedded in Epon 812. The blocks were then appropriately trimmed and sectioned on a Poster-Blum MT2 ultra microtome using glass knives.' Thick sections (1.0 pm) were prepared and stained with Richardson's methylene blue azure I1 or H & E stains with equivalent results.
164
CLARK, HENCH, AND PASCHALL
(b) (a) Photomicrograph of a 45s m bioglass-bone interface at 8 weeks (b) Transmission electron micrograph of a 455 m bioglass-bone interface at 3 weeks (43,700X).
Fig. 1. (1,320X).
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
165
A diamond knife was used to cut the thin sections. Sections were poststained with saturated fresh alcoholic uranyl acetate and lead citrate. Figure l a is a light microscope section of a bone-45Sm bioglass interface at eight weeks. The thin layer of black material (G) lining the edge of the bone may be the silica rich gel layer which forms on the glass implant surface (I),as has been observed in the in vitro analyses. Small pieces of glass implant (I) are attached to the gel layer. Only a small density of viable osteocytes (0)are present in the newly formed bone (B). The osteoid front (OF) is characterized by a lack of active bone formation and very few active osteoblasts. Figure lb is a transmission electron micrograph of a three week bone-45S co bioglass interface. The material which exhibits the regular fracture pattern (GF) appears to be the silica-rich gel layer described above. The evidence in support of this conclusion follows. 1) The gel layer formed in vitro on the 45Sm glass is very soft and nonadherent. Thus, the glass implant particles (I) in Figure la, should break away during final microtomy of the section. None are seen in any TEM sections. Also, the relative softness of the gel should produce a uniform fracture pattern such as seen a t GF, with long nonbranching fracture lines protruding deep within the gel layer. 2) The thickness of the gel layer in Figure l b is 1-2 pm, measured to the end of the fracture lines and the region where the density of collagen fibers is low (GE). This is the same magnitude of the thickness seen in Figure l a . In vitro studies have shown that there is relative stability of the gel film thickness between the three to eight week period. Figure 2a is a photomicrograph of a 4585 glass-bone interface. I n contrast to Figure l a , large pieces of bioglass (I) are intimately attached t o bone and several normal osteocytes (0) are present in the mineralized area. There is a well-defined layer of osteoblasts actively engaged in laying down new bone (OF) and this front is separated from the mineralized area by a transition zone of partially mineralized bone. These features indicate that inducement of the normal ossification process has been achieved. An electron micrograph illustrating a 4585 glass-bone interface a t six weeks is shown in Figure 2b. A layer of amorphous material (2000-3000 A thick) separates the glass implant I from mineralized bone B. Hydroxyapatite crystals which extend across the zone connect the implant to the bone structure.
166
CLARK, HENCH, AND PASCHALL
(b) Fig. 2 . (a) Photomicrograph of a 4585 bioglass-bone interface a t 3 weeks (1300X). (b) Transmission electron micrograph of a 4585 bioglass-bone interface a t 6 weeks (22,lOOX).
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
167
The effect of an excessive amount of phosphorus in a bioglass is illustrated in Figure 3. Figure 3 is a photomicrograph of a 4552.5 glass-bone interface a t eight weeks. An important feature to note is that the implant I has been separated from the bone B by an interval containing a capillary (C) which on electron microscopy reveals intercellular crystallization (X) induced by the excess phosphorus. It was observed that the gel layer (G) remained attached to the bone when the interval containing the capillary separated the glass (1)-gel layer interface. Note also the unhealthy appearance of the osteocytes (0). They have withdrawn from their lacuner walls and the nuclei are pyknotic. There is also a n absence of new bone formation a t the osteoid front (OF). As added evidence for the importance of implant surface chemistry on osteogenesis, a series of uncrystallized, partially crystallized, and nearly fully crystallized 4585 implants were compared as rat femoral implants. Equivalent osteogenesis and interfacial bonding was observed for all three microstructures and crystallographic states. In vitro analyses showed nearly equivalent surface chemical behavior due t o the presence of a residual glassy phase left after crystallization.
EVIDENCE FROM SOFT TISSUE RESPONSES Whenever muscle is injured or foreign material is implanted into muscle, healing occurs by formation of a collagenous scar or fibrous encapsulation of the implant without direct attachment t o the implant. Bone induction in muscle has not been successful, although autogenous bone grafts implanted in muscle are revascularized and new bone is formed within the graft. Therefore, a n ideal implant for muscle or other soft tissues which behave in the same fashion, i.e., tendon, ligaments, or skin, would be one which induces a collagenous scar with direct attachment to the implant. With these facts in mind, several possibilities of the type of response to the implants in muscle would be: 1) induction of bone formation around the implant thus securely anchoring i t to the muscle. 2) fibrous tissue encapsulation of the implant without attachment ; 3) fibrous tissue encapsulation with direct collagen bonding to the implant; 4) encapsulation and bursa formation about the implant; or 5) tissue rejection with acute and chronic inflammatory response with sterile abcess formation and extrusion of the implant.
168
CLARK, HENCH, AND PASCHALL
Fig. 3. Photomicrograph of a 4582.5 bioglass-bone interface at 8 weeks (1320X ) .
Recognizing the above possible physiological reactions to an implant, a study was done t o determine the histological response of muscle to various compositions of bioglass-ceramic Small pieces of each material that exhibit variable surface reactivity (Table I) measuring 1 x 1 x 0.1 cm were implanted under aseptic techniques in the lateral thigh muscles of male Sprague-Dawley rats. The animals were sacrificed a t three, six, and 16 weeks. The bioglass and surrounding muscle were removed as a block and fixed in cold cacodylate buffered (pH 7.4) gluteraldehyde, dehydrated in graded alcohols and propylene oxide, and embedded in Epon 812. The blocks were sectioned for light and electron microscopy. All of the compositions evoked a similar tissue response which varied quantitatively according to the surface reactivity of the material. At three weeks, there was a resolving inflammatory reaction with early fibrous encapsulation of the implants. Figure 4a is a photomicrograph of a 4585 glass-muscle interface at three weeks. Several layers of elongated fibroblasts (F) are separating the glass implant (I) from the muscle tissue (MT). An electron micrograph of the glass cell interface (Fig. 4b) reveals a single layer of macrophagelike cells (MC) between the glass implant (I) and the fibroblasts (F).
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
169
(b) Fig. 4. (a) Photomicrograph of a 4555 bioglass-muscle interface at 3 weeks (1,670X). (b) Transmission electron micrograph of the area adjacent to the implant surface of (a) (47,500X).
170
CLARK, HENCH, AND PASCHALL
These cells have attacked and are altering the implant surface. There is a band three or four cell layers thick where cells have ingested bioglass (IB) within cisternae. At six weeks, there was a synovial membrane indentifiable about all implants. A t 16 weeks, there was evidence of continued attack of the implants by the synovial cells and phagocytosis of the ceramic particles. The less surface reactive materials showed the slowest development of attack. These findings demonstrate that although certain bioglass-ceramic compositions (4555, 45B &5, and 45S5F) fulfill the criteria of a n ideal implant material for bone, in that they induce normal bone formation with direct attachment to the implant, they do not fulfill the criteria of an ideal implant material in soft tissues. The reactive surfaces of the materials release into the microenvironment surrounding the implant Ca and P ions which ordinarily are not found in muscle in any significant quantities. Therefore, the materials evoke an inflammatory reaction which continues into a form of rejection. A similar form of rejection for the same reasons has been found in equivalent implants in porcine tendon.
DYNAMIC SURFACE CHEMISTRY EVIDENCE The histological reactions of bioglasses and bioglass-ceramics are significantly influenced by composition as indicated in both hard and soft tissue responses. Therefore, it is important to define the roles of the individual ionic species in the surface chemical reactions of these materials in order t o establish a surface chemical theory of response. Conventional silicate glasses have a continuous three-dimensional structure. The basic repeat unit is a tetrahedran with four oxygen atoms at the corners and a silicon atom a t the center. Oxygen atoms which are common t o two tetrahedra are referred to as bridging, since they provide the linkage necessary to extend the structure. Incorporation of calcium and sodium ions modifies the structure by tying up oxygen atoms which can no longer serve as a bridge between neighboring tetrahedra. Due to their low silica (45 wt yo)and high modifier content the normal structure of glass is drastically altered in bioglasses. The average structure consists of two-dimensional chains with 4 to 20
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
171
tetrahedra per chain.g The sodium and calcium act as ionic bridges between these chains. Hence, their role as covalent network modifier has been reversed to that of ionic network formers providing additional stability to the structure. When bioglasses are exposed to an aqueous environment, sodium and calcium are preferentially leached from the surface, leaving behind a silica-rich layer.6 Because of the low silica content of the glass, the structure of this layer for certain compositional ranges may be relatively open and adherence to the bulk glass is fragile. The addition of phosphorus significantly alters the corrosion process.1° Initially a silica-rich layer is produced. However, this layer is followed by an amorphous calcium phosphate layer which crystallizes with time.lOJ1 Increasing the phosphorus content reduces the time required for the calcium phsophate layer to form. Furthermore, the adherence of the calcium phosphate layer is superior to that of the silica-rich layer. Figure 5 is a composition profile of the surface of a 4535 bioglass exposed to an aqueous solution buffered at a pH of 7.4 and maintained at 37°C for 1 hr. The profile was obtained by ion milling the
1
Atomic Percsnt
'"t
2oD 0
0
1
2
3
4
5
Depth From L r f a c e ci2[04
Fig. 5 .
'-Silica
Rich-Diffuse
6
7
8
Cwl
I n t e r f a c e d B u l k Glass
Composition profile of a corroded 45S5 bioglass surface.
9
172
CLARK, HENCH, BND PASCHALL
surface, atom layer by atom layer, while simultaneously analyzing the surface atoms with Auger spectroscopy.'l The results confirm the formation of a silica rich gel layer which is followed by a Ca-PO4 rich layer. At 1 hr the Ca-P04layer has just started to develop and is only 0.4 pm thick. The silica-rich gel layer beneath is 2.25 pm thick. Additional aqueous exposure in vitro increases the Ca-PO4 layer to as much as 10 pm. The calcium fluoride and boron oxide compositional modifications also fit this same analysis of surface reactions.1° During the initial corrosion reaction, calcium fluoride additions to a bioglass containing phosphorus increase the surface silica concentration, while boron oxide additions reduce it. Calcium fluoride and boron oxide also exert a rate controlling influence on the formation of the calcium phosphate layer. Calcium fluoride accelerates the process while boron retards it. In effect, calcium fluoride in the implant surface accelerates the formation of a resistant surface film which retards further reactions, while boron oxide delays the same process. The consequence of the sequence of surface chemical reactions described above is osteogenesis. The sodium ion release from the
Mineralized Bone
Osteoid
Surface Layer
Bulk Class
Fig. 6. A summary of the influence of surface chemistry on bioglass implant reactions with hard tissue.
SURFACE CHEMISTRY INFLUENCE ON IMPLANTS
173
surface eventually overrides the effect of acidic enzymes associated with wound healing and bone repair and permits a local alkaline p H t o be maintained. Osteoblasts differentiate in the implant vicinity. The silica-rich gel and amorphous calcium phosphate layer produced on the implant surface incorporates collagen fibrils and mucopolysaccharides generated by the osteoblasts. Crystallization of the calcium phosphate layer proceeds simultaneous with mineralization within the collagen fibrils resulting in a cojointly crystallized boneimplant junction. A summary of the structure of the interfacial bonding in hard tissue based on the histological and surface chemistry studies is shown in Figure 6. A thin (500 A) zone of co-crystallization within the inorganic gel and the collagen produces the bonding. If insufficient phosphorous is present, a thick (2000 8) zone of incomplete crystallization is developed which is weak and unstable. With excess phosphorus content, cellular rejection of the implant occurs by macrophage-like destruction of the surface accompanied by ectopic calcification of tissues in the vicinity of the implant. If insufficient surface reactivity is present osteogenesis does not proceed, fibroblastic cellular membranes envelop the surface hyrolysis product, and fibrous capsule develops. Application of the successful development of segmental bone grafts using this theory is presented in another study.12 Use of the concepts in orthopedic joint fixation13 and design of soft tissue studies appear in other reports.?
References 1. L. L. Hench, R. J. Splinter, W. C. Allen, and T. K. Greenlee, Jr., J . Biomed. Res. Symp., No. 2, 6 , 117 (1972). 2. L. L. Hench, T. K. Greenlee, Jr., and W. C. Allen, Reports # 1, #2, #3, August 1970, 1971, and 1972, U. S. Army Med. R & D Contract No. DADA17-70-C-0001. 3. T. K. Greenlee, J r , C A. Beckham, A. R. Crebo, and J. C. Malmborg, J . Biomed. Muter. Res., 6 , 244 (1972). 4. C. A. Beckham, T. K. Greenlee, Jr., and A. R. Crebo, J. Calcified Tissue Res., 8. 2 (1971). 5. L. L. Hench and H. A. Paschall, J. Biomed. Mater. Res. Symp. No. 5, 8, 49 (1974). 6. L. L. Hench and H. A. Paschall, J . Biomed. Res. Symp., No. 4 , 7, 25 (1973). 7. H. A. Paschall, M. M. Rodabush, and J. T. McVey, Report #3, August 1972, U. S. Army Med. R & D Contract No. DADA-17-70-C-0001.
174
CLARK, HENCH, AND PASCHALL
8. H. A. Paschall, J. McVey, and M. Rodabush, Orthopaedic Research Society, in preparation. 9. H. J. L. Trap and J. M. Stevels, Glastechn. Ber. 32K, VI/31 (1959). 10. A. E. Clark and L. L. Hench, J. A m r . Ceram. Soc., (1974), in press. 11. C. G. Pantano, A. E. Clark, and L. L. Hench, J . Amer. Ceram. Soc., 57(9), 412 (1974). 12. G. Piotrowski, L. L. Hench, W. C. Allen, and G. J. Miller, J. Biomed. Mater. Reo. Symp. No. 6, 9, 47 (1975). 13. L. L. Hench, H. A. Paschall, W. C. Allen, and G. Piotrowski, National Bureau of Standards Special Publication 415, 19 (1975).
Received April 16, 1975
