Peptide-immobilized nanoporous alumina membranes for enhanced osteoblast adhesion

Peptide-immobilized nanoporous alumina membranes for enhanced osteoblast adhesion

ARTICLE IN PRESS Biomaterials 26 (2005) 1969–1976 www.elsevier.com/locate/biomaterials Peptide-immobilized nanoporous alumina membranes for enhanced...

534KB Sizes 5 Downloads 45 Views

ARTICLE IN PRESS

Biomaterials 26 (2005) 1969–1976 www.elsevier.com/locate/biomaterials

Peptide-immobilized nanoporous alumina membranes for enhanced osteoblast adhesion Erin E. Leary Swan1, Ketul C. Popat1, Tejal A. Desai Department of Biomedical Engineering, Boston University, 44 Cummington Street, Boston, MA 02215, USA Received 10 May 2004; accepted 5 July 2004 Available online 11 August 2004

Abstract Bone tissue engineering requires the ability to regulate cell behavior through precise control over substrate topography and surface chemistry. Understanding of the cellular response to micro-environment is essential for biomaterials and tissue engineering research. This research employed alumina with porous features on the nanoscale. These nanoporous alumina surfaces were modified by physically adsorbing vitronectin and covalently immobilizing RGDC peptide to enhance adhesion of osteoblasts, bone-forming cells. X-ray photoelectron spectroscopy (XPS) and scanning electron microscopy (SEM) were used to characterize the modified nanoporous alumina surface. Survey and high-resolution C1s scans suggested the presence of RGDC and vitronectin on the surface and SEM images confirmed the pores were not clogged after modification. Cell adhesion on both unmodified and modified nanoporous alumina was compared using fluorescence microscopy and SEM. RGDC was found to enhance osteoblast adhesion after 1 day of culture and matrix production was visible after 2 days. Cell secreted matrix was absent on unmodified membranes for the same duration. Vitronectin-adsorbed surfaces did not show significant improvement in adhesion over unmodified membranes. r 2004 Elsevier Ltd. All rights reserved. Keywords: Alumina; Bone tissue engineering; Nanotopography; Osteoblast; RGD peptide; XPS

1. Introduction Alumina has been used extensively as a substrate for bone tissue engineering applications [1]. Current clinical applications include orthopedic prostheses and dental implants. The biocompatibility of alumina has been proven, and researchers continue to search for new and improved applications [2]. Efforts to use smooth surfaced alumina have been widely reported, but more recent studies have moved into the realm of textured alumina, which shows better bone in-growth [3–5]. A segment of current research in bone tissue engineering is focused on porous and nanophase alumina. In one study, aluminum was layered on titanium and then anodized to produce pores in the Corresponding author. Tel: +1-617-358-3054; fax: +1-617-3586766. E-mail address: [email protected] (T.A. Desai). 1 Both authors contributed equally to this work.

0142-9612/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2004.07.001

coating of the order of 200 nm [6,7]. These substrates were cultured with osteoblasts, and cell adhesion was investigated. Using this procedure, pores less than 100 nm where not achieved. One group of researchers used a process of indirect fused deposition to create pores of the order of 150–480 mm but found no influence on osteoblast proliferation from the varied pore sizes [8,9]. In another study using alumina nanoparticles of grain size ranging from 24 to 167 nm, osteoblast adhesion increased by 50% with decreasing grain size. [10–14]. The use of nanoscale alumina is advantageous because the feature size matches the size range of inorganic particles in bone [14]. The first two studies mentioned above were not able to investigate the effect of nanoscale topography on osteoblast due to fabrication limitations, while the third study supports the notion that nanoscale topography directly influences bone cell behavior. In this study, a two-step anodization process was used to create nanopores with a diameter in the range of 30–80 nm. Variation of the anodization

ARTICLE IN PRESS 1970

E.E. Leary Swan et al. / Biomaterials 26 (2005) 1969–1976

voltage from 30 to 60 V allowed for precise control over the surface topography with a uniform pore size distribution. This technique of producing nanoporous alumina using anodization provides ease of control over the size of porous structure. The cellular response of adhesive cells can be regulated by modification of the surface chemistry of the substrate. It is known that modification of substrates with proteins and peptides influences cell adhesion. Methods of surface modification include physical adsorption of adhesive proteins and covalent attachment of peptides RGD, –FHRRIKA–, –KLKSQLVKRK–, etc. [15–17]. In this study, both physical and covalent attachment methods were used and compared for their affects on osteoblast adhesion. Vitronectin, a cellular adhesive protein, was physically adsorbed and arginine-glycine-aspartic acidcysteine (RGDC), a cellular adhesive peptide, was covalently immobilized on the surface of nanoporous alumina. Vitronectin has been physically adsorbed and shown to enhance osteoblast adhesion on alumina [18]. Also, covalently immobilized RGDC has been shown to promote cell attachment on various substrates [15,19]. Fifty-five percent of osteoblast adhesion is mediated by the RGDC sequences in adhesive proteins [20]. X-ray photoelectron spectroscopy (XPS) and scanning electron microscopy (SEM) were used to characterize the modified surfaces in terms of their chemical composition and surface topography. Osteoblast response to unmodified and modified nanoporous alumina was investigated using fluorescence microscopy and SEM. Further, bone matrix production was characterized using energy-dispersive spectroscopy (EDS).

2. Materials and method 2.1. Fabrication of nanoporous alumina membranes Flat nanoporous alumina membranes were fabricated from an aluminum sheet using a two-step anodization process. The detailed procedure is described elsewhere [21]. In brief, 99.9% pure aluminum (Alfa Aesar) was anodized in two steps at 60 V. At the end of first anodization step the oxide was etched to form nucleation sites for later uniform pore formation. The second anodization and further etching of aluminum resulted in uniform pore size distribution with a diameter of 72 nm. These porous membranes were used in place of commonly used porous oxidized aluminum to eliminate the effects of aluminum from the study. 2.2. Cleaning/hydroxylating alumina membranes The alumina membranes were boiled in 30% hydrogen peroxide for 15 min to clean the surface. This step also introduced –OH groups on the surface, which

facilitated subsequent surface modification. Further, after treating with hydrogen peroxide, these films were boiled in deionized water for 15 min. The films were then air-dried and stored in argon until further surface modification. 2.3. Adsorption of vitronection on nanoporous alumina membranes Clean and dry alumina membranes were transferred into vials with vitronectin (Sigma) dissolved in phosphate-buffered saline (PBS) with a concentration of 50 mg/ml. Adsorption was allowed to proceed at room temperature for 24 h. Upon completion of adsorption, the membranes were thoroughly washed with deionized water for removal of non-adsorbed protein and salts from the buffer solution. The membranes were dried and stored in argon till further use. 2.4. Immobilization of RGDC peptide on nanoporous alumina membranes The RGDC peptide immobilization on nanoporous alumina membranes was achieved using an aminosilane linker [15,22]. Silanization was performed by incubating alumina membranes in 5% solution of (3-aminopropyl)triethoxysilane (APTES) (Aldrich) in toluene for 2 h. After the reaction, the membranes were washed twice with chloroform and acetone and several times with deionized water. In the next step, the membranes were incubated for 1 h at room temperature with a solution of 3.3 ml of N, N-dimethyl-formamide (DMF) (Alfa Aesar) containing 25 mg of N-succinimidyl-3-maleimidoproprionate (Aldrich) for a final concentration of 28 mM. The maleimide-grafted membranes were then washed twice with DMF and several times with deionized water. To prevent hydrolysis, the membranes were immediately subjected to the next step and were incubated for 2 h at room temperature with 300 ml of pure water containing 1.35 mg of RGDC (Bachem) for a final concentration of 10 mM. The RGDC-grafted membranes were thoroughly washed with deionized water, dried and stored in argon till further analysis. Fig. 1 shows the reaction scheme for the RGDC immobilization on the alumina surface. 2.5. Surface characterization 2.5.1. XPS To determine the surface composition of alumina membranes, XPS analysis was carried out. The alumina membranes were mounted on an XPS stage. Three spots per sample were analyzed. The analysis was conducted on a Kratos AXIS Ultra Imaging X-ray Photoelectron Spectrometer with a monochromatic Al-Ka-X-ray small spot source (1486.6 eV) and multichannel detector. A concentric hemispherical analyzer was operated in the

ARTICLE IN PRESS E.E. Leary Swan et al. / Biomaterials 26 (2005) 1969–1976 CH3CH2O

OH

+ H2O2 Alumina surface

OH

Hydrogen peroxide

1971

CH3CH2O-Si-CH2CH2CH2-NH2

+

OH CH3CH2O (3-aminopropyl) triethoxysilane

O O

O

C O-Si-CH2CH2CH2-NH2

H-C-CH2CH2

+

O

N C

H

O N,N-dimethyl-formamide + N-succinimidyl 3-maleimidopropionate O O

O

C

O-Si-CH2CH2CH2-N-C-CH2CH2 O

N

H

C-RGD

+

Cellular adhesion peptide

C O O

O

O

C

O-Si-CH2CH2CH2-N-C-CH2CH2 O

H

C-RGD

N C O

Fig. 1. Reaction scheme for immobilizing RGDC on alumina membrane surface.

constant analyzer transmission mode to measure the binding energies of emitted photoelectrons. The binding energy scale was calibrated by the Au4f7/2 peak at 83.9 eV, and the linearity was verified by the Cu3p1/2 and Cu2p3/2 peaks at 76.5 and 932.5 eV, respectively. Survey spectra were collected from 0 to 1100 eV with pass energy of 160 eV, and high-resolution spectra were collected for C1s, N1s and Al2p peaks detected with pass energy of 10 eV. All spectra were referenced by setting the hydrocarbon C1s peak to 285.0 eV to compensate for residual charging effects. Data for percent atomic composition, atomic ratios and peak fit analysis parameters were calculated using the manufacturer-supplied software. 2.5.2. SEM The porous structure of the alumina before and after modification was imaged using the JEOL JSM 6700F SEM. The alumina membranes, like other non-conductive surfaces, were sputter coated with gold at a

thickness of 10 nm to minimize the negative charge accumulation on the sample surface. The sputter coater was set at current of 20 mA and pressure of 0.05 mbar for 20 s to deposit a 10 nm layer of gold. Then, the membranes were imaged in the SEM at a voltage of 15 kV. Images were taken at various magnifications between 10,000  and 25,000  and were used to determine whether the pores were clogged after modification. 2.6. Osteoblast cell studies 2.6.1. Cell culture Human fetal osteoblasts, designated hFOB 1.19 (ATCC), were used for cell adhesion studies. This cell line was obtained from a spontaneous miscarriage and transfected with a temperature-sensitive mutant gene of SV40 large T antigen. The cells exhibit normal osteoblast phenotype after differentiation [23,24]. Cells below passage 10 were used in all experiments. The

ARTICLE IN PRESS E.E. Leary Swan et al. / Biomaterials 26 (2005) 1969–1976

medium used for growing osteoblasts consisted of 1:1 ratio of DMEM:F12 (Invitrogen) with 10% fetal bovine serum (Sigma) and 0.3 mg/ml of G418 sulfate powder (ATCC). The medium was changed every 2–3 days, and the subculture was done at a ratio of 1:4. 2.6.2. Cell adhesion and morphology Alumina membranes were adhered to the bottom of six-well plates with medical-grade silicone (Dow–Corning) and allowed to cure overnight. The wells were placed under ultraviolet lights in a biological hood for 24 h and then soaked in 70% ethanol for 30 min

Table 1 Surface elemental composition of various alumina membranes C (%)

O (%)

Al (%)

Si (%)

N (%)

Unmodified

9.98

58.73

31.29







APTES Maleimide RGDC

20.09 24.93 27.73

51.46 49.6 49.68

24.00 20.69 18.75

2.86 1.98 0.34

1.59 2.80 3.50

0.065 0.135 0.185

Vitronectin

18.39

52.1

25.24



4.27

0.170

N/Al

35000 30000

O1s

60000

15000

286.8

289.0

10000

Binding Energy (eV)

20000

12000

Al2p

C1s

10000

800

(a)

700 600 500 400 Binding Energy (eV)

300

200

100

0

60000 50000 O1s

40000

Unmodified

Virtonectin

285.2

286.8

8000 6000

289.0

288.0

4000 2000 0

30000 20000

Binding Energy (eV)

C1s N1s

10000 0 1100 1000 900

20000

0

40000

0 1100 1000 900

25000

285.2

RGDC Maleimide Silane Unmodified

5000

Intensity (CPS)

Intensity (CPS)

80000

Intensity (CPS)

for sterilization. Osteoblasts were seeded at a density of 5000 cells/cm2. Adhesion of the cells was quantified 24 h after seeding. Cells were fluorescently dyed with 5-(and 6-) carboxyfluorescein diacetate (CFDA) (Molecular Probes) to visualize the cytoskeleton, and the images were recorded using an Olympus BX60 light microscope. Osteoblast morphology after adhesion was further examined using SEM. Prior to imaging, the osteoblasts were fixed in place on the alumina membranes and dehydrated. Using a procedure modified from Corning Life Sciences [25], the alumina membranes were rinsed twice in PBS and then soaked in the primary fixative of 3% glutaraldehyde (Sigma) in a solution of 0.1 M of sodium cacodylate (Polysciences) and 0.1 M of sucrose (Sigma-Aldrich) for 45 min. The above solution was replaced by a buffer containing 0.1 M of sodium

Intensity (CPS)

1972

800

(b)

700 600 500 400 Binding Energy (eV)

Al2p

300

200

100

0

Fig. 3. High-resolution C1s scans for unmodified and modified alumina membrane surfaces.

70000

Intensity (CPS)

60000 50000

O1s

Table 2 Results of deconvolution of high-resolution C1s spectra

40000 30000 20000 10000 0 1100 1000 900

(c)

800

N1s

C1s

700 600 500 400 Binding Energy (eV)

300

Al2p 200

100

0

Fig. 2. XPS survey scans for: (a) unmodified, (b) RGDC-immobilized and (c) vitronectin-adsorbed alumina membranes.

Unmodified APTES Maleimide RGDC Vitronectin

285.2 eV

286.8 eV

288.0 eV

289.0 eV

73.2 68.1 59.8 48.3 49.9

14.2 17.8 20.7 27.2 22.5

— — — — 12.5

12.6 14.1 19.5 24.5 15.1

ARTICLE IN PRESS E.E. Leary Swan et al. / Biomaterials 26 (2005) 1969–1976

cacodylate and 0.1 M of sucrose for 5 min and was again repeated. Further, the cells were dehydrated by replacing the buffer with increasing concentrations of ethanol (35%, 50%, 70%, 95% and 100%) for 10 min each. The cells were then dried by replacing ethanol by hexamethyldisilazane (HMDS) (Polysciences) for 10 min. The HMDS was removed, and the membranes were air-dried for 30 min. SEM imaging was conducted on the JEOL JSM 6100 SEM at voltages ranging from 10 to 20 kV after the membranes were sputter coated in gold in the Cressington 108 Sputter Coater. The sputter coater was set with identical settings as mentioned above. 2.6.3. Phenotype expression EDS was used to determine the elemental composition of cellular secretions during SEM analysis. The nodules produced by the cells were analyzed using the Oxford Instruments EDS Unit attached to the JSM 6100 SEM. Working distance of the sample was adjusted to 15 mm. The electron beam was set at 20 kV, and the specific markers of phosphorus and calcium were used to determine the presence of extracellular matrix produced by the osteoblasts.

1973

3. Results and discussion 3.1. Surface characterization 3.1.1. XPS analysis XPS analysis was performed on modified surfaces to ensure the presence of silane, maleimide, RGDC and vitronectin on the alumina membranes and was compared to unmodified surfaces. Survey scans were taken after each step of modification. No nitrogen was detected on the unmodified surfaces, and carbon was present due to impurities from processing. There was a distinct increase in C1s (285 eV) and N1s (400 eV) peaks with modification of silane, maleimide and RGDC. This was also followed by consequent decrease in Al2p (72 eV) peaks for the above surfaces. Si2p (100 eV) peak appeared on the APTES-modified surface; however, the intensity of this peak decreased with further maleimide and RGDC modification. Table 1 shows the surface elemental compositions for the various alumina membranes. The analysis on APTES-modified surfaces revealed 1.59% of nitrogen on the surface, suggesting desirable APTES binding. Furthermore, the amount of carbon was doubled after APTES modification. This

Fig. 4. (a) 10,000  SEM images for unmodified and RGDC-immobilized alumina membranes and (b) 25,000  SEM image for RGDCimmobilized alumina membranes.

ARTICLE IN PRESS E.E. Leary Swan et al. / Biomaterials 26 (2005) 1969–1976

was expected due to the longer hydrocarbon chains present in APTES. The addition of maleimide resulted in a slight increase in nitrogen composition to 2.8% and a significant increase in carbon to about 25%. Grafting with the RGDC peptide resulted in significant increase in nitrogen to 3.5% as well as carbon to 28%. These results are also supported by the N/Al ratios (Table 1). Theoretically, the surface with RGDC should show sulfur, but since the cysteines are not present on the outer surface, the amount of sulfur is very negligible and not detectable by XPS [15]. For vitronectin adsorption on alumina membranes, there was a sharp increase in N1s (400 eV) and C1s (285 eV) peaks and subsequent decrease in Al2p (72 eV) peaks. The amount of nitrogen present on the surfaces was about 5%, suggesting significant protein on the surface. Fig. 2 shows the survey scans for unmodified, RGDC-immobilized and vitronectin-adsorbed alumina membranes. To further support the presence of APTES, maleimide, RGDC and vitronectin on alumina membranes, high-resolution C1s scans were taken (Fig. 3). The major hydrocarbon peak was at 285.2 eV. The binding energy at 286.8 eV is assigned to amines (CH2N) and alkoxy groups (CH2–O). From silane-modified to maleimidegrafted surfaces, there was addition of a peak at higher binding energy of 289.0 eV which is due to the carbon atoms in imide groups [C(QO)–N–C(QO)]. This confirmed the covalent attachment of maleimide groups on the surface. The increase at this position from maleimide-grafted to RGDC-immobilized surfaces was due to the carboxyl groups (COOH) and guanidine [NH–C(QNH)–NH2] carbons. C1s energies for amide functions (OQC–N) were expected to be around 288.0 eV but were difficult to deconvolute. However, for vitronectin all the four peaks, i.e. 285.2, 286.8, 288.0 and 289.0 eV were deconvoluted. Table 2 shows the deconvolution of high-resolution C1s spectra into different peaks. The decrease in the percentage of the peak at 285.2 eV subsequently with increase in peaks at other binding energies reflects the various modification steps. 3.1.2. SEM SEM was used to determine whether the porous structure of alumina membranes remained present after modification. Fig. 4(a) shows the 10,000  SEM images of unmodified and RGDC-immobilized membranes taken on the JEOL JSM 6700F SEM. The images showed comparable porous structure. To further visualize the nanoporous structure at higher magnification, 25,000  images were taken for RGDC-immobilized membranes (Fig. 4(b)). The high magnification images showed very little clogging of nanopores. Thus, it was not necessary to alter the concentration of RGDC to retain the pore size after modification.

3.2. Osteoblast adhesion studies Osteoblasts seeded on RGDC-immobilized membranes were counted after 1 day of culture and were compared to cell counts on unmodified membranes for the same duration. The cells were visualized using CFDA stain and were counted from fluorescence microscope images (Fig. 5). More cells were present on

Fig. 5. Fluorescence images of osteoblasts adhered on unmodified and RGDC-immobilized membranes.

700 600 Cell count

1974

500 400 300 200 100 0

Unmodified

Vitronectin adsorbed

RGDC immobilized

Fig. 6. Osteoblasts adhered on the membranes after 1 day of culture.

ARTICLE IN PRESS E.E. Leary Swan et al. / Biomaterials 26 (2005) 1969–1976

1975

Fig. 7. SEM images showing osteoblast morphology after 1 and 2 days in culture on unmodified and RGDC-immobilized membranes.

the RGDC-immobilized membranes compared to unmodified membranes. Further, vitronectin-adsorbed membranes also showed increased cell adhesion when compared to unmodified membrane. Fig. 6 shows the number of cells adhered on unmodified, RGDCimmobilized and vitronectin-adsorbed membranes after 1 day of culture. As can be seen, RGDC-immobilized membranes had approximately four times more cells than unmodified membranes. Vitronectin-adsorbed membranes showed improved cell attachment compared to unmodified membranes; however, they were lower than RGDC-immobilized membranes. This may be due to the fact that RGDC was covalently attached and therefore more stable compared to vitronectin, which was physically adsorbed. Therefore, further analysis on vitronectin-adsorbed membranes was not performed. The morphology of the osteoblasts adhered on unmodified and RGDC-immobilized membranes was investigated using SEM analysis. The cells after 1 day were spherically shaped on both unmodified and RGDC-immobilized membranes. After 2 days, the osteoblasts on the RGDC-immobilized membranes had begun secreting bone matrix in the form of clumps. However, no matrix was apparent on unmodified membranes. This phenomenon may be due to the adhesive nature of RGDC allowing the cells to attach and increase their rate of matrix production compared to the unmodified membranes. Future work is now focused on investigating this occurrence. Fig. 7 shows SEM images of cells adhered on unmodified and RGDC-immobilized membranes after 1 and 2 days of incubation.

Fig. 8. Close up of EDS scan for osteoblast matix after 2 days on RGDC-immobilized membrane.

In order to confirm the presence of matrix on the RGDC-immobilized membranes, EDS scans were conducted on the substrate. The EDS scan for the modified membrane showed peaks at 3.69 and 2.013 keV representing calcium and phosphorous, respectively. These elements are considered markers of bone matrix. Hence, the presence of these elements confirmed that the material seen in the SEM images was indeed matrix. Fig. 8 shows the close up of the EDS scan of the matrix produced by the osteoblasts on the membrane. The aluminum peak was present due to the alumina

ARTICLE IN PRESS 1976

E.E. Leary Swan et al. / Biomaterials 26 (2005) 1969–1976

substrate on which the cells were cultured, and the gold peak was from the gold sputter coating used in preparation of the sample for SEM.

4. Conclusions Covalent immobilization of the cellular adhesive peptide, RGDC, was accomplished on nanoporous alumina, which has not been previously reported. The surfaces were characterized using XPS. High-resolution C1s scans suggest the presence of covalently immobilized RGDC moieties on the surface of nanoporous alumina. Further, as shown by SEM images, the covalent immobilization of RGDC did not clog the nanopores, allowing the osteoblasts to respond to both the substrate topography and surface chemistry. The peptide immobilization improved initial osteoblast adhesion to the alumina membranes as shown by fluorescence microscopy cell count. Matrix production was evident on the peptide-immobilized substrates after 2 days in culture, which was absent in case of osteoblasts on unmodified alumina. Analysis of the matrix by EDS showed calcium and phosphorous, markers of bone matrix.

Acknowledgements The authors would like to acknowledge Maggie Paulose and Prof. Craig A. Grimes of the Department of Electrical Engineering at The Pennsylvania State University, University Park for their assistance with SEM analysis. Funding support was provided by NSF (2354-BU-NSF-0033) and NIH (5-R21-EB00570-02). References [1] Shackelford JF. Bioceramics—current status and future trends. Bioceram Mater Sci Forum 1999;293:99–106. [2] Langer R, Vacanti JP. Tissue Eng. Science 1993;260(5100):920–6. [3] Chang YS, Oka M, Kobayashi M, Gu HO, Li ZL, Nakamura T, Ikada Y. Significance of interstitial bone ingrowth under loadbearing conditions: a comparison between solid and porous implant materials. Biomaterials 1996;17(11):1141–8. [4] Josset Y, Oum’Hamed Z, Zarrinpour A, Lorenzato M, Adnet JJ, Laurent-Maquin D. In vitro reactions of human osteoblasts in culture with zirconia and alumina ceramics. J Biomed Mater Res 1999;47(4):481–93. [5] Josset Y, Oum’Hamed Z, Dupont C, Trentesaux C, LaurentMaquin D. Examination of zirconia, alumina ceramics and titanium interactions on human osteoblasts in culture. Bioceram Key Eng Mater 2000;192-1:329–32.

[6] Labat B, Chamson A, Frey J. Effects of gamma-alumina and hydroxyapatite coatings on the growth and metabolism of human osteoblasts. J Biomed Mater Res 1995;29(11):1397–401. [7] Karlsson M, Palsgard E, Wilshaw PR, Di Silvio L. Initial in vitro interaction of osteoblasts with nano-porous alumina. Biomaterials 2003;24(18):3039–46. [8] Bose S, Darsell J, Hosick HL, Yang L, Sarkar DK, Bandyopadhyay A. Processing and characterization of porous alumina scaffolds. J Mater Sci 2002;13(1):23–8. [9] Bose S, Darsell J, Kintner M, Hosick H, Bandyopadhyay A. Pore size and pore volume effects on alumina and TCP ceramic scaffolds. Mater Sci Eng C 2003;23(4):479–86. [10] Webster TJ, Siegel RW, Bizios R. Osteoblast adhesion on nanophase ceramics. Biomaterials 1999;20(13):1221–7. [11] Webster TJ, Siegel RW, Bizios R. Design and evaluation of nanophase alumina for orthopaedic/dental applications. Nanostruct Mater 1999;12(5–8):983–6. [12] Webster TJ, Ergun C, Doremus RH, Siegel RW, Bizios R. Enhanced functions of osteoblasts on nanophase ceramics. Biomaterials 2000;21(17):1803–10. [13] Gutwein LG, Webster TJ. Osteoblast and chondrocyte proliferation in the presence of alumina and titania nanoparticles. J Nanoparticle Res 2002;4(3):231–8. [14] Price RL, Haberstroh KM, Webster TJ. Enhanced functions of osteoblasts on nanostructured surfaces of carbon and alumina. Med Biol Eng Comput 2003;41(3):372–5. [15] Xiao SJ, Textor M, Spencer ND. Immobilization of the cell-adhesive peptide Arg-Gly Asp-Cys (RGDC) on titanium surfaces by covalent chemical attachment. J Mater Sci 1997;8: 867–72. [16] Stile RA, Shull KR, Healy KE. Axisymmetric adhesion test to examine the interfacial interactions between biologically modified networks and models of the extracellular matrix. Langmuir 2003;19:1853–60. [17] Stile RA, Healy KE. Thermo-responsive peptide-modified hydrogels for tissue regeneration. Biomacromolecules 2001;2:185–94. [18] Webster TJ, Schadler LS, Siegel RW, Bizios R. Mechanisms of enhanced osteoblast adhesion on nanophase alumina involve vitronectin. Tissue Eng 2001;7(3):291–301. [19] Wang D, Ji J, Sun Y, Shen J, Feng L, Elisseeff J. In situ immobilization of proteins and RGD peptide on polyurethane surfaces via poly(ethylene oxide) coupling polymers for human endothelial cell growth. Biomacromolecules 2002;3:1286–95. [20] Dee KC, Bizios R. Mini-review: proactive biomaterials and bone tissue engineering. Biotechnol Bioeng 1996;50(4):438–42. [21] Gong DW, Yadavalli V, Paulose M, Pishko M, Grimes CA. Controlled molecular release using nanoporous alumina capsules. Biomed Microdevices 2003;5(1):75–80. [22] Davis DH, Giannoulis CS, Johnson RW, Desai TA. Immobilization of RGD to /1 1 1S silicon surfaces for enhanced cell adhesion and proliferation. Biomaterials 2002;23:4019–27. [23] Harris SA, Enger RJ, Riggs BL, Spelsberg TC. Development and characterization of a conditionally immortalized human fetal osteoblastic cell line. J Bone Miner Res 1995;10(2):178–86. [24] Harris SA, Spelsberg TC. Immortalized human fetal osteoblastic cells. US Patent No. 5681701, 1997. [25] Corning Incorporated Life Sciences. Preparation of Costar Transwell inserts for scanning electron microscopy. Mar 9, 2004. /http://www.corning.com/Lifesciences/technical_information/ techDocs/sem_transwell_protocol.pdfS.