J O U R N A L O F M AT E R I A L S S C I E N C E : M AT E R I A L S I N M E D I C I N E 1 2 ( 2 0 0 1 ) 8 7 1 ± 8 7 7

Biomimetic coatings functionalized with adhesion

peptides for dental implants
S. ROESSLER, R. BORN, D. SCHARNWEBER, H. WORCH
Technische UniversitaÈt Dresden, Institut fuÈr Werkstoffwissenschaft, D-01062 Dresden,
Germany
A. SEWING, M. DARD
Merck Biomaterial GmbH, D-64271 Darmstadt, Germany

A complete biological integration into the surrounding tissues (bone, gingiva) is a critical
step for clinical success of a dental implant. In this work biomimetic coatings consisting
either of collagen type I (for the gingiva region) and hydroxyapatite (HAP) or mineralized
collagen (for the bone interface) have been developed as suitable surfaces regarding the
interfaces. Additionally, using these biomimetic coatings as a matrix, adhesion peptides
were bound to further increase the speci®city of titanium implant surfaces. To enhance cell
attachment in the gingiva region, a linear adhesion peptide developed from a laminin
sequence (TWYKIAFQRNRK) was bound to collagen, whereas for the bone interface, a cyclic
RGD peptide was bound to HAP and mineralized collagen using adequate anchor systems.
The biological potential of these coatings deduced from cell attachment experiments with
HaCaT human keratinocytes and MC3T3-E1 mouse osteoblasts showed the best results for
collagen and laminin sequence coating for the gingiva region and mineralized collagen and
RGD peptide coatings for regions with bone contact. Our concept opens promising
approaches to improve the biological integration of dental implants.
# 2001 Kluwer Academic Publishers

Introduction tissue next to the implant (e.g. bone: mineralized

Over the past years surface modi®cation of dental collagen) to create a natural microenvironment (extra-
implants focused mainly on micro and macro textures cellular matrix) for cells [21±23].
(e.g. titanium plasma spray coatings, etching) and to a We have developed a new strategy for dental implants
lesser extent on plasma spray coatings of calcium which is based on both concepts. The basis for the
phosphate phases (CPP) in the region of bone contact biomimetic coatings is provided by the composition at
[1±3]. For the gingiva area polished surfaces are favored the natural interface itself. Human gingival tissue
to minimize plaque adhesion, which may lead to comprises mainly of collagen type I. The main organic
periimplantitis [4±7]. However, metallic surfaces will component of bone is also collagen type I with
not be integrated through speci®c biological mechan- hydroxyapatite (HAP) as the main inorganic part.
isms. Therefore, biomimetic coatings consisting of collagen
Immediately after contact with the biological ¯uids the type I for the gingival region as well as mineralized
implant surface is conditioned by proteins, glycoproteins collagen and HAP for the bone interface have been
etc., whereas the ®nal composition is a result of a prepared on titanium surfaces for the latter both using a
dynamic process determined by adsorption and deso- biomimetic approach. Biomimetic coatings can be
rption (Vroman effect [8]). The original surface functionalized to enhance cell attachment. For the
characteristics in¯uence the adsorption and conformation gingival region a linear adhesion peptide using a laminin
of proteins [9, 10]. As a consequence host cells interact sequence (TWYKIAFQRNRK) [24] speci®c for a6 b4
with the conditioned metal surface. A practical applica- and a6 b1 integrins was selected, whereas a cyclic av -
tion of these naturally occurring events is seen in the selective RGD peptide [25] was taken for the bone
design of biological features on implant surfaces in order interface.
to induce a speci®c cellular response [10±12]. This
objective may be addressed by different approaches. One
concept is the immobilization of either adhesion proteins
(collagens, ®bronectin, laminin, vitronectin), bioactive Materials and methods
peptides (RGD, KRSR) or growth factors (BMP, TGF-b) Ti6Al4V samples (ASTM 136) (1 10 mm, height 2±
[12±20]. Another concept is seen in the formation of 3 mm) were prepared by polishing to roughness values
coatings by biomimetic processes that mimic the host below 25 nm (RMS) on a 100 mm length scale. Prior to
0957±4530 # 2001 Kluwer Academic Publishers 871
use ultrasonic cleaning was done in 1% triton X-100, subsequently mineralized by cathodic polarization in a
3 x†
acetone and ethanol for 20 min each. Ca2‡ =Hx PO4 -containing solution ( pH 6.4, 36  C) at
10 A m 2 for 15 min.

Collagen coating
Collagen type I coating of Ti6Al4V samples was Adhesion peptides
performed according to Geiûler et al. [26]. Brie¯y, for The linear laminin sequence was synthesized with a thiol
coating with ®brillar collagen (acid soluble calfskin anchor (NMI, TuÈbingen, Germany) and bound to amino
collagen, Fluka, Deisenhofen, Germany) samples were groups of collagen using sulfosuccinimidyl 4-( p-
incubated in a suspension of 65 mM phosphate buffer maleimidophenyl)butyrat (SMPB). The peptide cyclo
( pH 7.0, 25  C) with a collagen concentration of about (-RGDfK) utilized with thiol or phosphonate anchor
1 mg ml 1 for 15 min and rinsed with distilled water. groups [28] was bound to collagen or CPP, respectively.
The binding was performed with concentrations of
0.1 mmol for the RGD peptide and 1 mmol for the
Hydroxyapatite coating laminin sequence in the coating solution, which are in the
The HAP coating on Ti6Al4V was achieved by a two step range of saturation coverage for titanium samples given
process. First HAP was deposited on titanium by an by ELISA measurements.
electrochemically controlled method. The process of Scanning electron microscopy (SEM, DSM 982
deposition is described in detail in Roessler et al. Gemini, Carl Zeiss Oberkochen, Germany) at low
[27]. Brie¯y, cathodic polarization was carried out acceleration voltage (1 kV) was utilized for morphology
using galvanostatic mode 100 A m 2 † in a Ca2 ‡ = characterization. Fourier Transform Infrared
3 x† Spectroscopy (FTS 2000, Perkin-Elmer) was used to
Hx PO4 -containing electrolyte (0.03 M CaCl2 ‡
0:02 M NH4 H2 PO4 ) at pH 6.4 and 37  C for 1 h. Second characterize the chemical compositions of the deposited
anodic polarization in phosphate buffer ( pH 12, 36  C) CPP and the mineralized collagen.
using galvanostatic mode 10 A m 2 ; 40 VSCE † was Cell attachment was measured in triplicate using
carried out to increase oxide thickness and hence partially hexosaminidase test 1 h after seeding [29]. ELISA
incorporate the HAP layer into the anodic oxide layer. The measurements were related to cell number and enzyme
coated samples were washed in distilled water and dried in activity by measuring absorbance in wells with known
air. number of cells. Plating ef®ciency was calculated
relative to the enzyme activity of the seeded cell
number (50.000).
Mineralized collagen coating
Mineralization of collagen was accomplished by an
electrochemically controlled process in a Ca2‡ = Results
3 x†
Hx PO4 -containing electrolyte at near physiological Morphology and chemical
conditions for pH (6.4) and temperature (36  C) under composition
cathodic polarization of the sample. Prior to mineral- Morphology (SEM) and chemical composition (FTIR) of
ization collagen ®brils were adsorbed onto a HAP layer the biomimetic coatings comprising either collagen,
previously deposited on the titanium surface by an HAP or mineralized collagen are described in the
electrochemically controlled procedure (see above) and following. A polished titanium sample which was used

Figure 1 SEM-image of a Ti6Al4V surface (reference).

872
Figure 2 SEM-image of adsorbed collagen type I ®brils on Ti6Al4V.

as reference surface is shown in Fig. 1. Fig. 2 shows a length and 5 60 nm width and height. The HAP needles
titanium surface covered by a network of native collagen form a porous structure with no preferential direction of
®brils after adsorptive immobilization. The characteristic crystallite orientation. The thickness of the coating is
banding pattern of collagen type I ®brils (67 nm) is seen about 5 mm. Fig. 4 shows the mineralized collagen
and a ®bril width between 80±100 nm was measured with coating on a titanium surface. The mineralized collagen
atomic force microscopy (AFM). Due to adsorption, coatings consist basically of a two layer structure
®brils deform and expose a ratio of width to height of comprising a pure layer of HAP on the titanium surface
about 10 which indicates strong interactions with the with a mineralized collagen layer on top. Until complete
titanium surface [30]. Places in between the network of mineralization was achieved the characteristic banding
collagen ®brils appear free of collagen in SEM-images pattern of collagen ®brils (63±67 nm) remained visible.
but are homogenously covered by collagen molecules However, no periodic correlation of HAP crystals c-axes
and small aggregates as found with AFM [31]. It is related to the banding pattern of collagen was found,
important to note, that the titanium surface is homo- instead collagen ®brils are covered by HAP. The
genously covered by collagen either collagen ®brils or electrochemical parameters used here result in partial
small aggregates. The coating consisting of electro- mineralization of the collagen ®brils. The thickness of
chemically prepared HAP is shown in Fig. 3. The HAP the coating is about 4 mm.
crystals have a needle like appearance with 5 500 nm FTIR-spectroscopy has been used to analyze the

Figure 3 SEM-image of hydroxyapatite on Ti6Al4V.

873
Figure 4 SEM-image of mineralized collagen type I coating on Ti6Al4V.

chemical composition of the coatings and characteristic characteristic phosphate bands between 1000 and
spectra are shown in Fig. 5. As intensities resulting from 1100 cm 1 as well as the amide-bands and can be
adsorbed collagen ®lms were too low, thick collagen interpreted as a superposition of the two components
®lms were prepared by drying high concentrated HAP and collagen I. For the amide-I-band a broadening
collagen suspension on titanium. The spectrum of and shift to lower wavenumbers was found. FTIR-spectra
collagen I ®brils shows maxima at 1660, 1555 and of bone (cranial bone) are almost similar to our
1280 cm 1 which can be assigned to protein amide-I-, - biomimetic coatings (mineralized collagen). The shift
II-, and -III-bands respectively. The FTIR-spectrum of of the amide-I-band of mineralized collagen corresponds
HAP clearly shows splitting into the antisymmetric v3 to the broad amide-I-band of bone.
and v4 vibration modes (1050, 1090 cm 1 and
570,600 cm 1 †. The previously inactive v1 vibration of Cell adhesion
the free phosphate ion can also be seen in the HAP Cell adhesion experiments were carried out with HaCaT
spectrum. Additional bands at 874, 1587 and 1419 cm 1 human keratinocytes and MC3T3-E1 mouse osteoblasts
can be associated to carbonate ions (CO23 ) resulting representing two different cell types in contact with a
from the reaction of OH ions with carbon dioxide from dental implant. Fig. 6 shows the plating ef®ciency for the
air. The characteristic OH libration at 630 cm 1 is also basic coatings of collagen, mineralized collagen and
evident. The OH stretching vibration at 3570 cm 1 is HAP and for the corresponding states with bound
either very weak or, due to the reaction of CO2 with the adhesion peptides. MC3T3-E1 cells were not tested on
hydroxyl groups of the electrochemically controlled surfaces coated with the laminin sequence. According to
deposited HAP, not detectable. It is further conceivable their differentiation stage there is a risk that they express
that the OH stretching vibration is overlapped by a broad the corresponding integrin receptors leading to non-
band resulting from adsorbed water. FTIR- and XRD- speci®c results. For HaCaT cells the strongest in¯uence
spectra of CPP using a electrochemically assisted on cell adhesion is coming from collagen, leading to a
procedure have been published in detail previously plating ef®ciency of 37% for the collagen coating in
[27]. The spectrum of mineralized collagen shows the contrast to 2% for the uncoated control. This is con®rmed
by the mineralized collagen surfaces where values in the
range of 40% are reached, while hydroxyapatite is less
favored with 11%. Binding of the laminin sequence to
the collagen layer leads to a further increase of cell
adhesion (57%). For the mineralized collagen no
in¯uence of the laminin sequence could be observed.
The RGD peptide shows no signi®cant in¯uence on
HaCaT cells except on the hydroxyapatite surface. Cell
adhesion of MC3T3-E1 mouse osteoblasts is improved to
values around 32% for mineralized collagen and HAP
coatings, while the control reaches 5% and the collagen
coating 10%. The binding of the RGD adhesion peptide
in this case also improves the plating ef®ciency leading
Figure 5 FTIR-spectra for hydroxyapatite, collagen type I, mineralized to values above 50% independent on anchor and coating
collagen type I and bone (cranial bone). type.
874
Figure 6 Cell adhesion of MC3T3-E1 and HaCaT cells on different biomimetic coatings and their functionalized surface states: 1 Ti6Al4V reference,
2±4 collagen, 5±8 mineralized collagen, 9±10 hydroxyapatite. Adhesion peptides were bound to collagen (3, 4, 6, 8) or to hydroxyapatite (7, 10).

Discussion studied the deposition of CPP on titanium in the presence

The initial biological reactions occurring at the host of albumin and ®bronectin [44, 45]. However, these
tissue implant interface are essential for the success of processes do not resemble the natural composition of
implantation [10]. These events may be tailored towards bone. Using collagen as the organic component of the
a more physiological process by surface modi®cations bone matrix our mineralization method leads to a
which minimize nonspeci®c adsorption of proteins structure, which mimics, to some extent, the main
[17, 32], promote adhesion of cells [11, 20, 33] and composition and microstructure of bone. Further, the
present a microenvironment closely related to the host FTIR-spectra of mineralized collagen display phosphate
tissue [10, 21]. In order to achieve such surface properties vibrations and amide-bands that match those of bone.
the main components of the extracellular matrix Also the shift of the amide-I-band of mineralized
(collagen, HAP, mineralized collagen) have been collagen as compared to nonmineralized collagen
immobilized on Ti6Al4V using adsorption or biomimetic corresponds well with the broad amide-I-band of bone.
processes. It was shown that collagen I coatings could be The following methods are used for surface modi®ca-
achieved by a simple adsorption process which proved to tion of biomaterials with proteins or molecules:
be stable against concurrence adsorption [26]. On the adsorption, covalent coupling and self-assembly techni-
other hand CPP are considered the material of choice for ques [17, 32, 33, 46]. The latter are rather dif®cult on
the bone interface. In contrast to ref. [34±37] who also metal oxides, whereas covalent coupling requires
used electrochemistry for calcium phosphate deposition, intermediate layers. Biomimetic coatings which addi-
we found CPP presenting a high similarity to natural tionally have a biological function could offer an
CPP. In analogy to the natural process an amorphous alternative to the covalent coupling process via classical
calcium phosphate precursor is ®rst formed [27] which chemistry. The main biological function is seen in the
then transforms to nanocrystalline HAP needles. The microenvironment in which peptide ligands are pre-
dimensions of the HAP needles compared to HAP sented, which is an important factor for cell adhesion
crystals in bone were higher by about one order of [33].
magnitude [38] which was also reported by There are numerous studies suggesting that cell
Schirkanzadeh [37]. attachment plays a dominant role for further cellular
Aiming to achieve a more precise resemblance to the proliferation, differentiation and biochemical activity
complex structure of bone a biomimetic process for [47, 48]. Cell attachment of human gingival epithelial
mineralization of collagen was developed. Although cells on coated titanium surfaces have led to different
there are a number of studies on the development of results. Dean et al. [13] report the greater af®nity to
bone-like substitutes containing collagen and CPP [39± laminin coated surfaces, while Park et al. [12] only found
42] only a few studies address the formation of bone-like minor differences between collagen I, collagen IV and
structures on metallic implant surfaces. Self-assembled laminin coatings in comparison to titanium. For HaCaT
bilayers of positively and negatively charged poly(amino human keratinocytes we have shown the positive
acids) on titanium are used by Hwang et al. [43] to in¯uence of collagen I on cell attachment which can
induce the formation of organoapatite. Serro et al. still be increased by functionalization with the laminin
875
sequence. The ®nding that more HaCaT cells adhere on 5. M . Q U I R Y N E N , H . C . VA N D E R M E I , C . M . B O L L E N , A .
S C H OT T E , M . M A R E C H A L and G . I . D O O R N B U S C H , J. Dent.
collagen I coated titanium surfaces is in agreement with a
Res. 72 (1993) 1304.
recent publication of Klein et al. [49], who measured cell 6. M . Q U I RY N E N , C . M . B O L L E N , W. P A PA I O A N N O U , J . VA N
adhesion on sand-blasted surfaces coated with collagen I E L D E R E and D . VA N S T E O N B E R G H E , J. Oral Maxillofac.
and IV. The results for the a6 b4 and a6 b41 selective Impl. 11 (1996) 169.
laminin sequence is supported by RaÈisaÈnen et al. [50] 7. C . D . W U - Y U A N , K . J . E G A N H O U S E , J . C . K E L L E R and K . S .
WA LT E R S , J. Oral Implantol. 21 (1995) 207.
who demonstrated the important role of the a6 b4 integrin
8. L . V R O M A N and A . L . A D A M S , J. Colloid Interface Sci. 111
in cell attachment of HaCaT cells. Our cyclic RGD (1986) 391.
adhesion peptide has no in¯uence on HaCaT cell 9. D . M . B R U N E T T E , Int. J. Oral Maxillofac. Impl. 3 (1988) 231.
attachment except for the hydroxyapatite surface. In 10. C . J . K I R K P AT R I C K , M . WA G N E R , H . KO È H L E R , F.

this case as well as for the laminin sequence on B I T T I N G E R , M . OT T O and C . L . K L E I N , J. Mater. Sci. 8

mineralized collagen, which shows lower results then (1997) 131.

11. K . A N S E L M E , Biomaterials 21 (2000) 667.
on collagen itself, the integrin receptors may be 12. J . C . PA R K S , H . M . K I M and J . S . KO , J. Oral. Maxillofac. Impl.
in¯uenced by presence of divalent Ca2‡ ions [51]. 13 (1998) 826.
The hydroxyapatite and mineralized collagen coatings 13. J . W. D E A N , K . C . C U L B E R T S O N and A . M . D A N G E LO ,
act positively in cell attachment of MC3T3-E1 mouse J. Oral. Maxillofac. Impl. 10 (1995) 721.
14. K . C . D E E , T. T. A N D E R S E N and R . B I Z I O S , J. Biomed. Mater.
osteoblasts in contrast to collagen and titanium surfaces.
Res. 40 (1998) 371.
The binding of the RGD adhesion peptide greatly 15. D . M . F E R R I S , G . D . M O O D I E , P. M . D I O M O N D , C . W. D .
increases cell adhesion independently of coating type. G I O R A N N I , M . G . E H R L I C H and R . F. VA L E N T I N I ,
For collagen, values as high as for the calcium phosphate Biomaterials 20 (1999) 2323.
coatings are reached demonstrating the high ef®ciency of 16. B . F. LO W E N B E R G , R . M . P I L L A R , J . E . A U B I N , J . S O D E K

the chosen cyclic RGD peptide. A high selectivity can and A . H . M E L C H E R , ibid. 9 (1988) 302.
17. D . A . P U L E O and A . N A N C I , ibid. 20 (1999) 2311.
also be deduced from the negligible in¯uence of this 18. S . R O E S S L E R , C . W O L F, D . S C H A R N W E B E R and H . W O R C H ,
peptide on the HaCaT cells. Similar property has also J. Adhes. Sci. Technol. 14 (2000) 453.
been demonstrated by Kantlehner et al. [52] for other cell 19. H . W O R C H in ``Metals as Biomaterials'', edited by J. A. Helsen
types. and H. J.Breme (John Wiley & Sons Ltd, 1998) p. 177.
20. P. S C H A F F N E R , J . M E Y E R , M . D A R D , B . N I E S , S . V E R R I E R ,
H . K E S S L E R and M . K A N T L E H N E R , J. Mater. Sci.: Mater.
Med. 10 (1999) 837.
Conclusion 21. J . P. M A Z Z U CO T E L L I , C . K L E I N - S OY E R , A . B E R E T Z , C .
Biomimetic coatings consisting of the main components B R I S S O N , G . A R C H I P O F F and J . P. CA Z E N AV E , Int. J. Artif.
of bone can be produced on titanium implant surfaces Organs 14 (1991) 482.
22. S . RO E S S L E R , D . S C H A R N W E B E R , R . B O R N , H . WO R C H , A .
using electrochemical controlled methods. SEM and
S E W I N G and M . D A R D , in ``Proceedings of the 6th World
FTIR-spectroscopy of biomimetic coatings show struc- Biomaterials Congress'' (Kamuela, Hawaii, May 2000,
tural and chemical properties comparable to bone. Transactions) p. 1040.
Further, adhesion promoting peptides could be immobi- 23. H . B . W E N , J . R . D E W I J N , F. Z . C U I and K . D E G RO OT,
lized on collagen, HAP or mineralized collagen without J. Biomed. Mater. Res. 41 (1998) 227.
24. M . N O M I Z U , W. H . K I M , K . YA M A M U R A , A . U TA N I and
losing their functionality. Biomimetic coatings can be
S A N G - Y O , J. Biol. Chem. 270 (1995) 20583.
regarded as a good alternative to the conventional 25. M . K A N T L E H N E R , D . F I N S I N G E R , J . M E Y E R , P.
covalent coupling to metallic implant surfaces. Cell S C H A F F N E R , A . J O N C Z Y K , B . D I E F E N B AC H E R , B . N I E S
adhesion experiments provide evidence to our concepts and H . K E S S L E R , Angew. Chem. 111 (1999) 587.
to coat metal implants with the main components of the 26. U . G E I S S L E R , U . H E M P E L , C . W O L F, D . S C H A R N W E B E R ,
H . W O R C H and K . W. W E N Z E L , J. Biomed. Mater. Res. 51
extracellular matrix presenting additionally an adhesion
(2000) 752.
promoting peptide. Coating and adhesion peptide have to 27. S. RO E S S L E R , M. S TO E L Z E L , R. BORN, D.
be chosen regarding to the host tissue at the interface. S C H A R N W E B E R , H . W O R C H , A . S E W I N G and M . D A R D ,
ibid. Submitted May, 2001.
28. A . J O N C Z Y K , S . G O O D M A N , B . D I E F E N B AC H , A . S U T T E R
and H . K E S S L E R , WO-A97/14716.
Acknowledgment 29. U . L A N D E G R E N , J. Immunol. Meth. 67 (1984) 379.
The ®nancial support by BMBF (03N4021) is gratefully 30. S . R O E S S L E R , D . S C H A R N W E B E R and H . W O R C H , J. Mater.
acknowledged. The authors like to thank B. Blessing and Sci. Lett. 18 (1999) 577.
S. Perschbacher for performing the cell experiments and 31. S . R O E S S L E R , C . W O L F, D . S C H A R N W E B E R , H . W O R C H , A .
S E W I N G and M . D A R D , in ``Proceedings of the 6th World
BoneMaster GmbH, Dresden, for providing the hy-
Biomaterials Congress'' (Kamuela, Hawaii, May 2000,
droxyapatite coatings. Transactions) p. 1045.
32. A . N A N C I , J . D . W U E S T, L . P E R U , P. B R U N E T, V. S H A R M A ,
S . Z A L Z A L and M . D . M C K E E , J. Biomed. Mater. Res. 40 (1998)
References 342.
1. M. BRIEN, H. S C H L I E P H A K E , J . B E C K E R and P. A . 33. B . T. H O U S E M A N and M . M R K S I C H , Biomaterials 22 (2001)
R E I C H A R T, Z. ZahnaÈrtzl. Implantol. 14 (1998) 208. 943.
2. D . B U S E R , R . K . S C H E N K , S . S T E I N E M A N N , J . P. 34. S . B A N and S . M A R U N O , J. Biomed. Mater. Res. 42 (1998) 387.
F I O R E L L I N I , C . H . F O X and H . S T I C H , J. Biomed. Mater. 35. H . I S H I Z AWA and M . O G I N O , ibid. 31 (1996) 6279.
Res. 25 (1991) 889. 36. J . R E D E P E N N I N G , T. S C H L E S S I N G E R , S . B U R N H A M , L .
3. G . L . D E L A N G E and K . D O N AT H , Biomaterials 9 (1988) 302. L I P P I E L LO and J . M I Y A N O , ibid. 30 (1996) 287.
4. L . W. L I N D Q U I S T, B . R O C K L E R and G . E . CA R L S S O N , 37. M . S H I R K H A N Z A D E H , J. Mater. Sci.: Mater. Med. 9 (1998)
J. Prosthet. Dent. 59 (1988) 59. 67.

876
38. R . Z . L E G E RO S , in ``Calcium Phosphates in Oral Biology and 47. E . D . H AY, in ``Cell biology of extracellular matrix'' (Plenum
Medicine'' (Basel, KraÈger, 1991) p. 110. Press, New York, 1991) p. 343.
39. J . H . B R A D T, M . M E R T I G , A . T E R E S I A K and W. PO M P E , 48. A. BRUCE, B. DENNIS, L. JULIAN, R. MARTIN, R. KEITH
Chem. Mater. 11 (1999) 2694. and D . W. J A M E S , in ``Molecular biology of the cell'' (Garland,
40. Y. D O I , T. H O R I G U C H I , Y. M O R I WA K I , H . K I I TA G O , T. New York, 1994) p. 963.
K A J I M O T O and Y. J . I . I WAY A M A , J. Biomed. Mater. Res. 31 49. M . K L E I N , T. H O H L F E L D , P. M O O R M A N N and H .
(1996) 43. M E N N E K I N G , Int. J. Oral. Maxillofac. Impl. 15 (2000) 247.
41. C . D U , Z . F. C U I , W. Z H A N G , Q . L . F E N G and K . D E G RO O T , 50. L. RA È ISAÈ NEN, M. KO È NOÈ NEN, J. JUHANOJA, P.
ibid. 50 (1999) 518. VA R P AVA A R A , J . H A U TA N I E M I and M . H O R M I A ,
42. K . S . T E N H I U S E N , R . M A R T I N , M . K L I M K I E W I C Z and P. W. J. Biomed. Mater. Res. 49 (2000) 79.
B RO W N , ibid. 29 (1995) 803. 51. B . D I E F E N B AC H , Ph.D. Thesis, TH Darmstadt, Germany (1995).
43. J . J . H WA N G , K . J A E G E R , J . H A N CO C K and S . I . S T U P P, ibid. 52. M . K A N T L E H N E R , P. S C H A F F N E R , D . F I N S I N G E R ,
47 (1999) 504. J . M E Y E R , A . J O N C Z Y K , B . D I E F E N B AC H E R , B . N I E S , G .
44. A . P. V. A . D E S E R R O , A . C . F E R N A N D E S , B . J . V. HO È L Z E M A N N , S . L . G O O D M A N and H . K E S S L E R ,
S A R A M A G O , J . L I M A and M . A . B A R B O S A , Biomaterials 18 Chembiochem. 1 (2000) 107.
(1997) 963.
45. A . P. V. A . D E S E R R O , A . C . F E R N A N D E S and B . J . V.
S A R A M A G O , J. Biomed. Mater. Res. 49 (2000) 345.
46. M . M R K S I C H and G . M . W H I T E S I D E S , Ann. Rev. Biophys. Received 14 May
Biomol. Struct. 25 (1996) 55. and accepted 21 May 2001

877