BONE TISSUE ENGINEERING is an emerging interdisci-plinary field involving principles of the life sciences
and engineering concerned with the formation of three-
dimensional bone substitutes by culturing osteogeniccells on natural or synthetic polymer scaffolds.1 The ul-timate goal is to create a living osteoinductive bone sub-stitute, which can overcome the current disadvantages ofautograft and allograft bone substitutes. Although vari-
Surface Modification by Complexes of Vitronectin and GrowthFactors for Serum-Free Culture of Human Osteoblasts
IRIS SCHLEICHER, M.D.,1 ANTHONY PARKER, B.S.,2 DAVID LEAVESLEY, Ph.D.,2ROSS CRAWFORD, Ph.D.,1,3 ZEE UPTON, Ph.D.,2 and YIN XIAO, Ph.D.2
ABSTRACT
Cell attachment, expansion, and migration in three-dimensional biomaterials are crucial steps foreffective delivery of osteogenic cells into bone defects. Complexes composed of vitronectin (VN), in-sulin-like growth factors (IGFs), and insulin growth factor-binding proteins (IGFBPs) have been re-ported to enhance cell attachment, proliferation, and migration in a variety of cell lines in vitro. Theaim of this study was to examine whether prebound complexes of VN and IGFs � IGFBPs couldfacilitate human osteoblast serum-free expansion in vitro and enhance cell attachment, prolifera-tion, and migration in three-dimensional biomaterial constructs. Human osteoblasts derived fromalveolar bone chips and the established human osteoblast cell line Saos-2 were used. These cells wereseeded on tissue culture plates and porous scaffolds of type I collagen sponges and polyglycolic acid(PGA), which had been coated with VN � IGFBP-5 � IGF-I. Cell attachment, proliferation, andmigration were evaluated by cell counting, confocal microscopy, and scanning electron microscopy.The number of attached human osteoblasts was significantly higher in VN-coated polystyrene cul-ture dishes. Furthermore, significant increases in cell proliferation were observed when growth fac-tors were bound to these surfaces in the presence of VN. In the two scaffold materials examined,greater cell attachment was found in type I collagen sponges compared with PGA scaffolds. How-ever, coating the scaffolds with complexes composed of VN � IGF-I or VN � IGFBP-5 � IGF-I en-hanced cell attachment on PGA. Moreover, the presence of VN � IGFBP-5 � IGF-I resulted in sig-nificantly greater osteoblast migration into deep pore areas as compared with untreated scaffoldsor scaffolds treated with fetal calf serum. These results demonstrated that complexes of VN �IGFBP-5 � IGF-I can be used to expand osteoblasts in vitro under serum-free conditions and en-hance the attachment and migration of human osteoblasts in three-dimensional culture. This in turnsuggests a potential application in surface modification of biomaterials for tissue reconstruction.
1Prince Charles Hospital, Brisbane, Queensland, Australia.2Tissue BioRegeneration and Integration, Science Research Centre, Institute of Health and Biomedical Innovation, Queens-
land University of Technology, Brisbane, Queensland, Australia.3Medical Engineering, Queensland University of Technology, Brisbane, Queensland, Australia.
ous materials have been tested for the potential con-struction of bone scaffolds, major challenges remain infacilitating osteogenic cell attachment, migration, andformation of natural bone matrix in three-dimensionalbiomaterials. In addition, currently cell expansion re-quires fetal calf serum (FCS). The restriction of animalproducts for human therapeutical goods suggest that ex-pansion of osteogenic cells in vitro under serum-free con-ditions is vital if the cell and biomaterial constructs areto be used therapeutically.
Vitronectin (VN) is an important component of the in-terstitial extracellular matrix (ECM) and has critical ad-hesive functions during development,2 angiogenesis, andwound healing.3 It is thought that binding of the heparin-binding domain (HBD) or an RGD (Arg-Gly-Asp) se-quence or other domains within VN mediates extracellu-lar matrix anchoring, cell spreading, and cell migration.4,5
VN, also known as S protein, “serum spreading factor,”or epibolin, has been detected as an essential mediator ofadhesion and spreading in many cells in vitro.6
Insulin-like growth factors (IGFs) are mitogenic formany types of cell and also have the ability to stimulatetissue-specific cellular responses such as bone matrixsynthesis.7,8 The abundant presence of IGFs in adultbone9 indicates that IGFs play important roles in the reg-ulation of bone cell activity.10,11 Furthermore, both sys-temic and local administration of IGF-I have shown ef-fects on bone cell proliferation and bone formation11
Investigations have demonstrated that VN can specifi-cally bind proteins of the IGF system and that the insulin-like growth factor-binding proteins (IGFBPs) play an im-portant role in modulating IGF-I binding to VN.12,13 Inaddition, it has been demonstrated by our laboratory thatMCF-7 breast cancer cells and HaCAT human skin ker-atinocytes exhibit enhanced functional responses to VN,IGFBP, and IGF-I complexes.14–16 Therefore, we testedosteoblast attachment, proliferation, and migration on tis-sue culture plastic and polymer scaffolds treated withcomplexes containing VN, IGFBP-5, and IGF-I to con-firm our hypotheses that IGFs bound to surface-adsorbedVN can stimulate and enhance cell attachment, prolifer-ation, and migration in the absence of serum in two-di-mensional monolayer culture and three-dimensional cul-ture on porous biomaterials in vitro.
MATERIALS AND METHODS
Cell culture
Human osteoblast-like cells (Saos-2), originally iso-lated from a human osteosarcoma, were obtained fromthe American Type Culture Collection (HTB-85; ATCC,Rockville, MD). Human osteoblasts used for this studywere isolated from alveolar bone as described previ-
SURFACE MODIFICATION USING VITRONECTIN AND GROWTH FACTORS
ously.17 Briefly, normal human alveolar bone specimenswere obtained from young, healthy patients (12 to 16years old) at orthodontic clinics with ethics approval fromthe University of Queensland (Brisbane, Australia). Thetissue samples were first treated by collagenase digestionand then used as explants for establishment of cell cul-ture. The cells obtained were subcultured and character-ized by morphological and functional criteria, such asmineralization potential and mRNA expression of bonematrix proteins.17
Cells were maintained in culture in 75-cm2 flasks(Nalge Nunc International, Rochester, NY) containingminimal essential medium � formulation (á-MEM;GIBCO/Invitrogen, Auckland, New Zealand) (Saos-2cells) or Dulbecco’s modified Eagle’s medium (DMEM;GIBCO/Invitrogen) (human osteoblasts) supplementedwith 10% (v/v) FCS (Thermo Trace/Thermo Elec-tron, Melbourne, Australia), 1% penicillin–streptomycin(GIBCO/Invitrogen), and 1% nonessential amino acids(ICN Pharmaceuticals, Costa Mesa, CA).
Prebinding of vitronectin, IGFBP-5, and IGF-I
Vitronectin (Promega, Annandale, Australia), IGFBP-5 (S. Firth, University of Sydney, Sydney, Australia), andIGF-I (GroPep, Adelaide, Australia) were sequentiallybound to tissue culture plates, collagen sponges (ICNPharmaceuticals), and polyglycolic acid (PGA; AlbanyInternational Research, Mansfield, MA) according tomethods described previously.14–16 Briefly, for prelimi-nary studies using Saos-2 cells, VN (52 ng/well) inserum-free �-MEM (sf-�MEM) was added to 96-well tis-sue culture plates (Nalge Nunc International) and incu-bated for 2 h at 37°C. Wells were then blocked with 1%bovine serum albumin (BSA; Sigma-Aldrich, St. Louis,MO) in HEPES binding buffer (HBB) (0.01 M HEPES,0.12 M NaCl, 0.005 M KCl, 0.0012 M MgSO4 � 7H2O,0.008 M D-glucose) for 30 min. IGFBP-5 (70 ng/well) in0.5% BSA–HBB either alone or in combination withIGF-I (17.5 ng/well) was then added and incubated at 4°Covernight.
For subsequent studies using primary human os-teoblasts, plastic 24- and 96-well tissue culture plates(Nalge Nunc International), type I collagen sponges, andPGA (both precut to fit wells of 96-well plates) werecoated with VN at either 300 ng/well (24-well plates, col-lagen, and PGA scaffolds) or 100 ng/well (96-well plates)and allowed to incubate at 4°C overnight. The VN solu-tion was then removed and IGFBP-5 at 100 ng/well (24-well plates and collagen and PGA scaffolds) or 30 ng/well(96-well plates) was added and incubated at 4°C for 4 h.Finally, the IGFBP-5 solution was removed and IGF-I at100 ng/well (24-well plates, collagen and PGA scaffolds)or 30 ng/well (96-well plates) was added and incubatedat 4°C overnight.
1689
Attachment and proliferation assays
In the studies using Saos-2 cells, attachment and pro-liferation were measured by spectrophotometric detectionof the cleavage product (formazan) of the tetrazolium salt4-[3-(4-iodophenyl)-(2,4-dinitrophenyl)-2H-5-tetrazo-lio]-1,3-benzene disulfonate (WST-1; Roche Diagnos-tics, Brisbane, Australia). Briefly, subconfluent culturesof Saos-2 cells were serum starved for 4 h in sf-�MEM,trypsinized, and seeded onto precoated wells (5 � 103
cells per well in 100 �L of sf-�MEM) and incubated for3 or 48 h in a humidified 5% CO2–95% air atmosphereat 37°C (standard conditions). After incubation, themedium was removed and each well was washed oncewith sf-�MEM (attachment assay), or the medium wasnot removed (proliferation assay), before addition ofWST-1 reagent to each well according to the manufac-turer’s instructions, and incubated for a further 2 h un-der standard conditions. Plates were then placed in aBeckman microplate reader (Beckman Coulter, Fullerton,CA) and the absorbance of the formazan dye cleavageproduct was quantitated at 450 nm, using 650 nm as ref-erence. The results were corrected and converted to cellnumber using data from plate blanks (serum-free mediumand WST-1 alone).
Subsequent attachment and proliferation studies usedhuman osteoblast cells between passages 3 and 10 in culture. Human osteoblasts were prelabeled with a 2-�Ci/mL concentration of [3H]thymidine (5.0 Ci/mL; ICNPharmaceuticals) during their exponential growth phase.3H-labeled osteoblasts were then trypsinized and seededinto precoated wells of 96-well plates (7.5 � 103 cellsper well in 200 �L of serum-free DMEM [sf-DMEM]),in the presence or absence of scaffolds, and incubated for2 h at 37°C. Nonadherent cells were removed by gentlewashes with phosphate-buffered saline (PBS) and the at-tached cells were measured by counting radioactivitywith a scintillation counter (LS 6000SC; Beckman Coul-ter).
For proliferation assays, human osteoblasts wereserum starved in sf-DMEM for 24 h, trypsinized, andseeded onto precoated wells of 24-well plates (2 � 104
cells per well in 1 mL of sf-DMEM). Four hours later,50 �L of [3H]thymidine (10 �Ci/mL in PBS) was thenadded to the culture medium and cells were cultured fora further 72 h. Cells seeded into wells, without VN or innormal growth medium containing 10% FCS, were em-ployed as controls. DNA synthesis was measured by de-termination of [3H]thymidine uptake, using a scintillationcounter. Assays were performed in triplicate, where n �3 per assay.
Cell migration assay using Transwells
Saos-2 cell migration through the microporous mem-branes of 12-�m (pore size) Transwells was measured bya modified version of the Transwell migration assay de-
SCHLEICHER ET AL.
scribed previously.18 Briefly, 1 mL of a 1-�g/mL VN/sf-�MEM solution or sf-�MEM alone was added to thelower chamber of each well of 12-well Transwell tissueculture plates (Costar; Corning Life Sciences, New York,NY) and incubated for 4 h. Subsequently, duplicate 0.5-mL aliquots of IGFBP-5 (0.1 �g/mL) in sf-�MEM con-taining 0.05% BSA together with either 0.5 mL of IGF-I (0.025 �g/mL) or 0.5 mL of sf-�MEM alone with0.05% BSA were then added to the coated/uncoatedlower chambers and incubated at 4°C overnight. Growthfactor solutions were then removed and the lower chambers were washed twice with 0.05% BSA/HBB/sf-�MEM followed by the addition of 1 mL of0.05%BSA/HBB/sf-�MEM to the lower chamber of eachTranswell. The plates were then returned to the incuba-tor until required. Saos-2 cells were harvested and pre-pared as detailed above. The upper chamber of each Tran-swell received 2 � 105 cells followed by incubation for5 h under standard conditions. Cells seeded into wellswithout VN were employed as controls. Unmigrated cellswere removed from the upper surface of Transwell in-serts by cleaning with a dampened cotton bud. Migratedcells on the lower surface of the membrane were thenfixed with 10% formaldehyde and subsequently stainedby immersing the membranes in crystal violet. Insertswere then immersed in a beaker of circulating tap waterto remove excess stain and allowed to air dry before ex-traction of crystal violet in 1 mL of 10% acetic acid. Theabsorbance of duplicate 100-�L subsamples was read ina Beckman microplate reader (Beckman Coulter). Resultsare expressed as the absorbance at 595 nm � standard er-ror of the mean (SEM). Assays were performed in trip-licate, where n � 2 per assay.
Scanning electron microscopy and confocal laserscanning microscopy
Confocal laser scanning microscopy (CLSM) and scan-ning electronic microscopy (SEM) were used to study cellattachment and migration in collagen sponge and PGA scaf-folds. Osteoblasts (1 � 105) were seeded onto precoatedcollagen and PGA scaffolds and cultured under serum-freeconditions for 24 h, 72 h, and 2 weeks, respectively.
For SEM, osteoblasts grown on collagen sponges andPGA scaffolds were fixed in 3% glutaraldehyde (in 0.1M cacodylate buffer, pH 7.4) overnight. Samples werewashed in 0.1 M cacodylate buffer, postfixed in 1% aque-ous osmium tetroxide for 30 min, and dehydrated througha graded series of ethanol (50, 70, 90, and 100%). Aftercritical point drying with CO2, the samples were sputter-coated with gold and examined with a Quanta 200 SEM(FEI, Hillsboro, OR) operating at 10 kV.
For CLSM, osteoblasts grown on collagen sponges andPGA scaffolds were fixed in 4% paraformaldehyde for20 min. Samples were washed with PBS, permeabilizedwith 0.2% Tween for 20 min, and incubated in propid-
1690
3500
3000
2500
2000
1500
1000
500
0VN only IGF-1 IGFBP-5 IGF-VIGFBP-5 FCS
VN+VN-
Cel
l num
ber
*
** *#
B
ium iodide (0.1 mg/mL) for 5 min. To visualize F-actin,cells were then incubated with Alexa Fluor 488 phalloidin(Molecular Probes/Invitrogen Australia, Waverley, Aus-tralia) diluted 1:20 in PBS (200 U/mL). This was fol-lowed by thorough washing in PBS. Matrices weremounted in 1:1 glycerol–PBS mounting medium andsamples were examined with a Leica TCS 4D confocallaser scanning microscope (Leica Microsystems, Wet-zlar, Germany) equipped with an argon–krypton laser.
RESULTS
VN is the key component in modulating bone cellattachment on tissue culture plastic
Two hours after seeding human osteoblasts onto tissueculture plates, cell attachment was significantly increased
SURFACE MODIFICATION USING VITRONECTIN AND GROWTH FACTORS
in VN-coated wells compared with non-VN-coated wells(VN positive versus VN negative; p � 0.05) (Fig. 1). Nofurther enhancement of osteoblast attachment was foundwhen additional IGF-I and IGFBP-5 were coated ontotissue culture wells prebound with VN. However, somedifferences in cell attachment were noted between Saos-2 cells and human osteoblasts in response to differentcoating surfaces. In the absence of VN, a significant in-crease in attachment of bone-derived cells was found intissue culture wells coated with IGF-I � IGFBP-5 (p �0.05) (Fig. 1A), whereas no difference was detected insimilarly coated wells when Saos-2 cells were used (Fig.1B). In the presence of VN, no significant difference wasfound between the two cell types regardless of treatmentwith VN � IGF-I/IGFBP-5 (p � 0.05, Student t test),and all treatments resulted in attachment that was equiv-alent to that obtained in the positive controls (plate coatedwith serum) (Fig. 1).
1691
FIG. 1. Cell attachment on tissue culture plastic treated with or without VN in combination with IGFBP-5 and IGF-I. Cellswere seeded onto tissue culture wells that had been coated with VN � IGFBP-5 � IGF-I and incubated for 2 h in serum-freeculture. The attached cells were then measured. Cell attachment was significantly improved in the presence of VN. Cells cul-tured in medium with 10% FCS were used as positive control (FCS). Values shown represent means � SD of triplicate wellsfrom three experiments. *p � 0.05 relative to the VN-negative control; #p � 0.05 relative to VN-positive only. (A) Data derivedwith osteoblasts from alveolar bone; (B) data derived with Saos-2 cells.
4800
4000
3200
2400
1600
800
0VN only IGF-1 IGFBP-5 IGF-VIGFBP-5 FCS
VN+
VN-
Cel
l num
ber
* *
*
*
#
A
Visual observation of the cultures supported the quan-titative data, as cells cultured in the presence of VN or10% FCS appeared to be attached and had well-spreadmorphologies. However, cells seeded into wells withoutVN or FCS were less spread and appeared to be moreweakly attached with rounded morphologies (results notshown).
Bone cell proliferation and migration arestimulated by VN and IGF complexes
In general, cell proliferation was significantly higherfor all treatments in the presence of VN compared withtreatments without VN (p � 0.01) (Fig. 2). Specifically,when cells were cultured in the presence of VN/IGFBP-
SCHLEICHER ET AL.
5/IGF-I cell proliferation was found to be significantlyhigher than with VN alone, VN/IGF-I, and VN/IGFBP-5 but was equivalent to that measured for cells culturedin 10% FBS (Fig. 2). Interestingly, for Saos-2 cells nosignificant difference was found between VN alone,VN/IGF-I, and VN/IGFBP-5 (Fig. 2B). However, forbone-derived cells an increase in proliferation was de-tected with VN/IGF-I treatment compared with VN aloneand VN/IGFBP-5 (Fig. 2A). No significant difference incell proliferation was found in the absence of VN in bothcell lines, even in the presence of IGF-I or IGFBP-5 orIGF-I/IGFBP-5. These results indicate that in cell cultureVN was the major factor influencing cellular responsesand that it modulated the reaction of cells to IGF-I andIGFBP-5. However, cellular responses obtained in the
1692
50000
40000
30000
20000
10000
0
VN only IGF-1 IGFBP-5 IGF-VIGFBP-5 FCS
VN+
VN-
Cel
l num
ber
*
** #
#
A
30000
25000
20000
15000
10000
5000
0VN only IGF-1 IGFBP-5 IGF-VIGFBP-5 FCS
VN+
VN-
Cel
l num
ber
*
** #
* # #B
FIG. 2. Cell proliferation on tissue culture plastic treated with or without VN in combination with IGFBP-5 and IGF-I. Cellswere seeded onto tissue culture wells that had been coated with VN � IGFBP-5 � IGF-I and incubated for 72 h in serum-freeculture. Cell number was then measured by counting radioactivity or by measuring WST-1. Cell proliferation was significantlyincreased in the presence of VN. Cells cultured in medium with 10% FCS were used as positive control (FCS). Values shownrepresent means � SD of triplicate wells from three experiments. *p � 0.05 relative to the VN-negative control. #p � 0.05 rela-tive to VN-positive only. (A) Data derived with osteoblasts from alveolar bone; (B) data derived with Saos-2 cells.
presence of the complexes of VN/IGFBP-5/IGF-I re-vealed differences between the two types of osteoblasts.Interestingly, both cell lines showed the greatest responsewhen exposed to the trimeric combination of VN withIGFBP-5 and IGF-I (Fig. 2).
Cell migration
Transwell migration assays were performed in orderto assess whether IGFBP-5 together with IGF-I in thepresence of VN enhances the migration response of os-teoblasts. We found that cell migration was significantlyenhanced in the presence of VN (p � 0.01) relative totreatments without VN (Fig. 3). Indeed, virtually no mi-gration could be detected in the absence of VN, irre-spective of subsequent IGFBP-5 and/or IGF-I treatment,suggesting that the presence of VN was vital for sup-porting the migration of cells through the membranes(Fig. 3). In addition, when either IGF-I or IGFBP-5 wasprebound to the VN-coated Transwells, cell migrationthrough the membranes was significantly increased com-pared with that of VN-coated Transwells alone (p �0.01). Although there was no significant difference foundin cell migration in VN-coated Transwells prebound witheither IGF-I alone, IGFBP-5 alone, or IGFBP-5 and IGF-I, responses to the IGFBP-5/IGF-I combinations wereconsistently slightly higher compared with the respectivecomponents alone. Interestingly, cell migration in Tran-swells treated with IGFBP-5 alone in the presence of VNwas significantly increased relative to migration in Tran-swells treated with VN alone (p � 0.01), suggesting IGF-I-independent effects of IGFBP-5 on cell migration(Fig. 3).
SURFACE MODIFICATION USING VITRONECTIN AND GROWTH FACTORS
VN and IGF-I complexes enhance bone cellattachment and migration in scaffold materials
To determine whether the enhancement of osteoblastattachment, proliferation, and migration by VN and IGFcomplexes in two-dimensional cultures would facilitateosteoblast activities in three-dimensional culture, studiesusing type I collagen sponges and PGA scaffolds wereundertaken. Overall, osteoblasts attached to a greater ex-tent to the type I collagen scaffold compared with thePGA scaffold. An increase in cell attachment was notedon PGA scaffolds coated with VN compared with thosewithout VN coating. However, there was no differencein cell attachment between VN-treated and untreatedgroups on the type I collagen scaffolds (Fig. 4A). Withrespect to the PGA scaffolds, 1 h after cell seeding theinitial cell attachment was poor when precoating of VNcomplexes did not occur. Conversely, PGA scaffoldswith prebound VN improved cell attachment, as did pre-coating the scaffolds with trimeric complexes composedof VN/IGFBP-5/IGF-I. Indeed, cell attachment was sig-nificant higher compared with the presence and absenceof VN alone (p � 0.05) (Fig. 4B).
After 24 h of culture, the osteoblasts were attached andspread on the type I collagen scaffolds. SEM analysis re-vealed flat cell bodies with many cellular processes grow-ing into the pore areas within deeper layers of the colla-gen scaffold on VN-coated and uncoated type I collagenscaffolds. The number of cells present was slightly lessin the collagen scaffolds without any form of pre-treatment compared with VN-treated collagen sponges;however, no difference was noted in cell morphology(Fig. 5).
1693
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0VN only IGF-1 IGFBP-5 IGF BP-5/IGF-1
VN+
VN-Abs
*
** # * #
#
FIG. 3. Migration of Saos-2 cells seeded into the upper chamber of 12-�m (pore size) Transwells in response to prebindingwith or without VN in combination with IGFBP-5 and IGF-I under serum-free conditions. Cell migration was significantly in-creased in the groups prebound with VN compared with VN-negative groups. Values shown represent means � SD of triplicatewells from three experiments. *p � 0.05 relative to the VN-negative control. #p � 0.05 relative to VN-positive only.
Interestingly, fewer cells were visible on the PGA scaf-folds after 24 h in culture. These attached cells appearedonly on PGA fibers and no cells could be detected in theporous areas between the fibers. Nevertheless, there wasan increase in cell number and improved attachment inPGA scaffolds prebound with VN (Fig. 5).
Confocal microscopy revealed that after 3 days of cul-ture, many osteoblasts migrated into the deeper pore ar-eas in type I collagen scaffolds prebound withVN/IGFBP-5/IGF-I complexes compared with collagenscaffolds without prebound VN. There was no differencein cell attachment to surfaces of treated or untreated typeI collagen sponges. However, cell numbers were dra-matically decreased at 20 �m from the surface comparedwith cell numbers at the surface in all treated and un-treated groups. However, at 100 �m from the surface,cells were detectable only in VN/IGFBP-5/IGF-I-treatedcollagen sponges; no cells were found in collagen
SCHLEICHER ET AL.
sponges without VN treatment or in collagen scaffoldstreated with FCS (Fig. 6).
DISCUSSION
Effective treatment of bone defects is a great challengein orthopedic surgery. Bone grafting is remarkably suc-cessful and provides an effective solution to many clini-cal orthopedic challenges. Nevertheless, there is only alimited supply of suitable bone graft materials and use ofthe patient’s own bone often creates significant problemsat the donor site. Hence synthetic/artificial substituteshave been designed to satisfy this unmet need. Biodegrad-able synthetic polymers have been used in constructs fortissue-engineering applications.19–21 However, anchor-age-dependent cells such as osteoblasts do not attach,grow, and differentiate well on a hydrophobic surface.22
1694
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0VN only IGF-1 IGFBP-5 IGF-VIGFBP-5 FCS
VN+
VN-Cel
l num
ber
A
8000
*
**7000
6000
5000
4000
3000
2000
1000
0
VN only IGF-1 IGFBP-5 IGF-VIGFBP-5 FCS
VN+
VN-
Cel
l num
ber
B
FIG. 4. Osteoblast attachment in serum-free culture on three-dimensional biomaterials treated with or without VN in combi-nation with IGFBP-5 and IGF-I. Cells cultured in medium with 10% FCS were used as positive controls (FCS). Values shownrepresent means � SD of triplicate wells from three experiments. *p � 0.05 relative to the VN-negative control. (A) Cell at-tachment on type I collagen sponges. (B) Cell attachment on PGA scaffolds. No significant difference was found in cell attach-ment on type I collagen sponges regardless of VN coating status. A significant increase in cell attachment was noted in PGAprebound with VN versus scaffolds where VN was not present.
The hydrophobicity of synthetic polymers such as PGAhas been recognized as a difficulty for their use as scaf-folds in tissue engineering, despite their advantages foruse as biomaterials (such as easy design and modifica-tion; controllable rate of biodegradation; and easy andscalable material production, purification, and process-ing23–25). Therefore, surface modification of biomateri-als is essential for three-dimensional scaffolds in tissueengineering.
This study examined whether osteoblast attachment,proliferation, and migration on scaffolds, including typeI collagen and PGA, could be improved by coating the scaffolds with extracellular proteins and bioactivefactors. Experiments, predominantly with IGF-I, haveshown that IGFs stimulate bone cell proliferation and for-mation of bone.26 The role of the IGF system in bone ho-meostasis and healing, combined with the abundance ofthe IGFs in bone tissue, suggest that IGFs are excellentgrowth-promoting agents to deliver locally for targetedbone regeneration. Indeed, several in vivo studies havedemonstrated the potential of such approaches.27,28 In thestudies reported here, it was noted that osteoblast attach-
SURFACE MODIFICATION USING VITRONECTIN AND GROWTH FACTORS
ment was significantly enhanced on tissue culture plas-tic precoated with VN compared with wells without VNcoating. However, no further enhancement of cell at-tachment was noted when additional IGF-I and IGFBP-5 were present, indicating that VN is the critical com-ponent in regulation of osteoblast attachment in thissituation. Previously published binding studies have con-firmed that VN, IGFBPs, and IGF-I form complexes ontissue culture plastic, that IGFBP-5 modulates binding of125I-labeled IGF-I to VN, and that the regulation of cel-lular activity is due to the proteins bound to the sub-strate.12,15
VN is a multifunctional glycoprotein present in bloodand in the extracellular matrix. VN contains an RGD se-quence, through which it binds to the integrin receptor�v�3 (�v�3), and is involved in cell attachment, spread-ing, and migration.5 The enhancement of osteoblast at-tachment by prebound VN was also noticed in our stud-ies examining VN-coated PGA scaffolds, which alsorevealed a significant increase in osteoblast attachmentrelative to non-VN-coated scaffolds. This is most likelybecause scaffolds fabricated from synthetic polymers,
1695
FIG. 5. SEM analysis of cell attachment on type I collagen and PGA scaffolds. After 24 h of serum-free culture, osteoblastsrevealed flat cell bodies with many processes growing into the pore areas within deeper layers of the collagen scaffold on VN-coated (A) and uncoated (B) type I collagen scaffolds. Attached osteoblasts appeared only on PGA fibers and no cells could bedetected in the porous areas between the fibers. There was an increase in cell number and improved visual attachment in PGAscaffold prebound with VN (C) compared with uncoated PGA (D). Data are from duplicate studies in three separate experiments.
such as PGA and poly(lactic-co-glycolic acid) (PLGA),lack intrinsic cell adhesion signals. Cells adhere to thesepolymer scaffolds via extracellular molecules that are ad-sorbed on the polymer surface from the serum in culturemedium in vitro or in surrounding fluid in vivo.29
It has been found that vitronectin is the major proteinadsorbed from the serum onto synthetic polymers,30 andthat the adsorption of VN onto tissue culture polystyreneis correlated with the concentration of VN in the coatingsolution.31,32 It has been proposed that specificity of pro-tein-binding sites results from particular proteins beingable to attach to preferred chemical features on the poly-mer surface through nonrandom domains on the sur-face.31 The adsorption of VN onto PGA results in changesin hydrophobicity of PGA, which in turn affect the abil-ity of cell ligands to adhere.29 These alterations mightarise from a change in the conformation of adhesion li-gands on the solid surface or in their affinity for the sur-face.33–35 In addition, the VN receptor, �v�3, has beenfound to be responsible for mediating initial attachmentof osteoblasts and longer term adhesion to biomaterials.36
SCHLEICHER ET AL.
Precoating the hydrophobic surface has also been foundto activate several intracellular signals such as mitogen-activated protein kinase (MAPK), which improve osteo-blast adhesion.35 In experiments to extend the presentstudy, we intend to study cellular response using block-ing antibodies against VN and �v�3.
No significant difference was noticed between un-coated and coated type I collagen sponges. This is be-cause the natural polymer of type I collagen has manybinding sites for integrins. In particular, collagen has mo-tifs to facilitate binding of the �1 integrin subunit of os-teoblasts.37 Nevertheless, when IGF-I was prebound tothe VN-coated collagen sponges or membrane of Tran-swells, cell migration was significantly enhanced. Thisstimulation of cell migration was further strengthened bythe additional binding of IGF-I and IGFBP-5 to VN. IGF-I plays important roles in the regulation of bone cell ac-tivity including cell migration by stimulating an increasein �v�3 ligand binding.38 Indeed, our results indicate thatthe stimulation of osteoblast migration by IGF-I wasfound to require the presence of VN. In smooth muscle
1696
FIG. 6. Confocal microscopy shows that after 3 days of serum-free culture, many osteoblasts have migrated into the deeperpore areas in type I collagen scaffold prebound with VN/IGFBP-5/IGF-I complexes. At 100 �m from the surface cells were de-tectable only in collagen sponges treated with VN/IGFBP-5/IGF-I; no cells were found in collagen sponges without VN treat-ment or in collagen scaffolds treated with FCS. Data shown are representative of images from duplicate samples in three sepa-rate experiments. Original magnification: first row, �50; second and third rows, �200.
cells, IGF-I stimulation of cell migration also requires thepresence of VN, because of a requirement for ligand bind-ing by the �v�3 integrin (VN receptor).39
An IGF-I-independent effect of IGFBP-5 on osteoblastmigration was also noted in the presence of VN. In vas-cular smooth muscle cells (VSMCs) a similar effect hasalso been reported. Thus IGFBP-5 stimulates VSMC mi-gration through an IGF-independent mechanism. Themolecular basis underlying this ligand-independent ac-tion is explained by the concept that the IGFBP-5 aminodomain contains a putative trans-activation domain,which in turn suggests that IGFBP-5 is localized in theVSMC nucleus and possesses transcription-regulatoryactivity that is IGF independent.40 Whether the same mo-lecular mechanism underpins the responses we found inosteoblasts requires further study.
In addition to its apparent effectiveness as an IGF- andIGFBP-binding protein, VN provides additional functionsthat can stimulate cellular activities under serum-free con-ditions. Our data suggest that under serum-free conditionsosteoblasts attached and proliferated in tissue culturedishes prebound with VN/IGFBP5/IGF-I. This creates thepotential to expand osteogenic cells in serum-free culture.VN is an essential cell survival factor and is an importantmediator of cell anchorage to surfaces. On binding VN,VN receptors organize the intracellular cytoskeleton5 andthereby facilitate the function of most somatic cell types,including osteoblasts. VN also adsorbs readily to a widerange of surfaces5,41 and when adsorbed assumes a bio-logically active, multimeric conformation.41 In addition tomultiple anchorage sites, this may also reveal the bindingdomain(s) for proteins of the IGF system.16 Thus plastic-adsorbed VN plus IGF and IGFBPs may be responsiblefor the observed enhancement of osteoblast attachment,proliferation, and migration.16 It is also possible that bind-ing to VN enables IGFs to be presented to cell surface re-ceptors in a favorable orientation, or that the coactivationof the VN and IGF receptors results in synergy. The lat-ter has been suggested, for example, with respect to the�v�3 integrin and the insulin receptor.5 The applicationof VN to facilitate binding of IGF and IGFBPs thereforeoffers a simple and convenient method of bonding IGFfamily members to biomaterials, presenting and main-taining them locally at the required concentration. A log-ical extension of this therefore is that coactivation maysynergistically stimulate de novo bone regeneration at thetissue–implant interface.
In summary, this study found that complexes com-posed of VN, IGFBP-5, and IGF-I enhanced cell attach-ment and proliferation of human osteoblasts in serum-free in vitro culture. These findings suggest that thesecomplexes may provide a possible serum-free method forex vivo expansion of human osteoblasts. Because coatingof the complexes on synthetic biomaterials resulted in en-hanced cell attachment and enhanced cell migration intothe three-dimensional construct, this also suggests that
SURFACE MODIFICATION USING VITRONECTIN AND GROWTH FACTORS
these complexes can be used for surface modification ofbiomaterial scaffolds for tissue engineering. Further stud-ies have been planned to investigate whether this serum-free culture condition can support osteoblast differentia-tion and long-term culture.
ACKNOWLEDGMENTS
The authors thank Ms. C. Theodoropoulos and Ms. S.Singh for technical support in SEM and CLSM. The au-thors are also grateful for funds from the QUT StrategicCollaborative Program Scheme and for a Peter DohertyFellowship award from the NHMRC, Australia to Dr. YinXiao.
REFERENCES
1. Jadlowiec, J.A., Celil, A.B., and Hollinger, J.O. Bone tis-sue engineering: Recent advances and promising thera-peutic agents. Expert Opin. Biol. Ther. 3, 409, 2003.
2. Preissner, K.T. Structure and biological role of vitronectin.Annu. Rev. Cell Biol. 7, 275, 1991.
3. Preissner, K.T., and Seiffert, D. Role of vitronectin and itsreceptors in haemostasis and vascular remodeling. Thromb.Res. 89, 1, 1998.
4. Edens, R.E., LeBrun, L.A., Linhardt, R.J., Kaul, P.R., andWeiler, J.M. Certain high molecular weight heparin chainshave high affinity for vitronectin. Arch. Biochem. Biophys.391, 278, 2001.
5. Schvartz, I., Seger, D., and Shaltiel, S. Vitronectin. Int. J.Biochem. Cell Biol. 31, 539, 1999.
6. Weller, M., Wiedemann, P., Bresgen, M., and Heimann, K.[Vitronectin: Mediator of cell adhesion in proliferative vit-reoretinopathy?]. Fortschr. Ophthalmol. 87, 221, 1990.
7. McQueeney, K., and Dealy, C.N. Roles of insulin-like growth factor-I (IGF-I) and IGF-I binding protein-2(IGFBP2) and -5 (IGFBP5) in developing chick limbs.Growth Horm. IGF Res. 11, 346, 2001.
8. Jia, D., and Heersche, J.N. Insulin-like growth factor-1 and-2 stimulate osteoprogenitor proliferation and differentia-tion and adipocyte formation in cell populations derivedfrom adult rat bone. Bone 27, 785, 2000.
9. Mohan, S., and Baylink, D. IGF system components andtheir role in bone metabolism. In: Rosenfeld, R.G., andRoberts, C.T., Jr., eds. The IGF System. Totowa, NJ: Hu-mana Press, 1999, p. 457.
10. Linkhart, T.A., Mohan, S., and Baylink, D.J. Growth fac-tors for bone growth and repair: IGF, TGF-� and BMP.Bone 19, 1S, 1996.
11. Wergedal, J.E., Mohan, S., Lundy, M., and Baylink, D.J.Skeletal growth factor and other growth factors known tobe present in bone matrix stimulate proliferation and pro-tein synthesis in human bone cells. J. Bone Miner. Res. 5,179, 1990.
12. Upton, Z., Webb, H., Hale, K., Yandell, C.A., McMurtry,J.P., Francis, G.L., and Ballard, F.J. Identification of vit-ronectin as a novel insulin-like growth factor-II bindingprotein. Endocrinology 140, 2928, 1999.
13. Nam, T., Moralez, A., and Clemmons, D. Vitronectin bind-ing to IGF binding protein-5 (IGFBP-5) alters IGFBP-5modulation of IGF-I actions. Endocrinology 143, 30, 2002.
14. Hyde, C., Hollier, B., Anderson, A., Harkin, D., and Upton,Z. Insulin-like growth factors (IGF) and IGF-binding pro-teins bound to vitronectin enhance keratinocyte protein syn-thesis and migration. J. Invest. Dermatol. 122, 1198, 2004.
15. Kricker, J.A., Towne, C.L., Firth, S.M., Herington, A.C., andUpton, Z. Structural and functional evidence for the interac-tion of insulin-like growth factors (IGFs) and IGF bindingproteins with vitronectin. Endocrinology 144, 2807, 2003.
16. Noble, A., Towne, C., Chopin, L., Leavesley, D., and Up-ton, Z. Insulin-like growth factor-II bound to vitronectinenhances MCF-7 breast cancer cell migration. Endocrinol-ogy 144, 2417, 2003.
17. Xiao, Y., Qian, H., Young, W.G., and Bartold, P.M. Tis-sue engineering for bone regeneration using differentiatedalveolar bone cells in collagen scaffolds. Tissue Eng. 9,1167, 2003.
18. Leavesley, D.I., Ferguson, G.D., Wayner, E.A., andCheresh, D.A. Requirement of the integrin �3 subunit forcarcinoma cell spreading or migration on vitronectin andfibrinogen. J. Cell Biol. 117, 1101, 1992.
20. Gunatillake, P.A., and Adhikari, R. Biodegradable syntheticpolymers for tissue engineering. Eur. Cell Mater. 5, 1, 2003.
21. Piskin, E. Biodegradable polymeric matrices for bioartifi-cial implants. Int. J. Artif. Organs 25, 434, 2002.
22. Anselme, K. Osteoblast adhesion on biomaterials. Bioma-terials 21, 667, 2000.
23. McFarland, C.D., Mayer, S., Scotchford, C., Dalton, B.A.,Steele, J.G., and Downes, S. Attachment of cultured hu-man bone cells to novel polymers. J. Biomed. Mater. Res.44, 1, 1999.
24. Altankov, G., and Groth, T. Fibronectin matrix formationby human fibroblasts on surfaces varying in wettability. J.Biomater. Sci. Polym. Ed. 8, 299, 1996.
25. Altankov, G., Grinnell, F., and Groth, T. Studies on thebiocompatibility of materials: Fibroblast reorganization ofsubstratum-bound fibronectin on surfaces varying in wet-tability. J. Biomed. Mater. Res. 30, 385, 1996.
26. Bautista, C.M., Baylink, D.J., and Mohan, S. Isolation ofa novel insulin-like growth factor (IGF) binding proteinfrom human bone: A potential candidate for fixing IGF-IIin human bone. Biochem. Biophys. Res. Commun. 176,756, 1991.
27. Laffargue, P., Fialdes, P., Frayssinet, P., Rtaimate, M.,Hildebrand, H.F., and Marchandise, X. Adsorption and re-lease of insulin-like growth factor-I on porous tricalciumphosphate implant. J. Biomed. Mater. Res. 49, 415, 2000.
28. Raschke, M., Wildemann, B., Inden, P., Bail, H., Flyvb-jerg, A., Hoffmann, J., Haas, N.P., and Schmidmaier, G.Insulin-like growth factor-1 and transforming growth fac-tor-�1 accelerates osteotomy healing using polylactide-coated implants as a delivery system: A biomechanical andhistological study in minipigs. Bone 30, 144, 2002.
Frontiers in Tissue Engineering. New York: Elsevier Sci-ence, 1998, pp. 121–137.
30. Kim, B.S., Nikolovski, J., Bonadio, J., Smiley, E., andMooney, D.J. Engineered smooth muscle tissues: Regulat-ing cell phenotype with the scaffold. Exp. Cell Res. 251,318, 1999.
31. Steele, J.G., Dalton, B.A., Johnson, G., and Underwood,P.A. Adsorption of fibronectin and vitronectin onto Pri-maria and tissue culture polystyrene and relationship to themechanism of initial attachment of human vein endothelialcells and BHK-21 fibroblasts. Biomaterials 16, 1057, 1995.
32. Steele, J.G., Dalton, B.A., Johnson, G., and Underwood,P.A. Polystyrene chemistry affects vitronectin activity: Anexplanation for cell attachment to tissue culture polystyrenebut not to unmodified polystyrene. J. Biomed. Mater. Res.27, 927, 1993.
33. Vermeer, A.W., Giacomelli, C.E., and Norde, W. Adsorp-tion of IgG onto hydrophobic teflon: Differences betweenthe F(ab) and F(c) domains. Biochim. Biophys. Acta 1526,61, 2001.
34. Zoungrana, T., Findenegg, G.H., and Norde, W. Structure,stability, and activity of adsorbed enzymes. J. Colloid. In-terface Sci. 190, 437, 1997.
35. Lee, Y.J., Park, S.J., Lee, W.K., Ko, J.S., and Kim, H.M.MG63 osteoblastic cell adhesion to the hydrophobic sur-face precoated with recombinant osteopontin fragments.Biomaterials 24, 1059, 2003.
36. Rezania, A., and Healy, K.E. Integrin subunits responsiblefor adhesion of human osteoblast-like cells to biomimeticpeptide surfaces. J. Orthop. Res. 17, 615, 1999.
37. Basle, M.F., Lesourd, M., Grizon, F., Pascaretti, C., andChappard, D. [Type I collagen in xenogenic bone materialregulates attachment and spreading of osteoblasts over the�1 integrin subunit]. Orthopade 27, 136, 1998.
38. Maile, L.A., Imai, Y., Clarke, J.B., and Clemmons, D.R.Insulin-like growth factor I increases �v�3 affinity by in-creasing the amount of integrin-associated protein that isassociated with non-raft domains of the cellular membrane.J. Biol. Chem. 277, 1800, 2002.
39. Jones, J.I., Doerr, M.E., and Clemmons, D.R. Cell migra-tion: Interactions among integrins, IGFs and IGFBPs. Prog.Growth Factor Res. 6, 319, 1995.
40. Xu, Q., Li, S., Zhao, Y., Maures, T.J., Yin, P., and Duan,C. Evidence that IGF binding protein-5 functions as a li-gand-independent transcriptional regulator in vascularsmooth muscle cells. Circ. Res. 94, E46, 2004.
41. Horbett, T.A., and Lew, K.R. Residence time effects onmonoclonal antibody binding to adsorbed fibrinogen. J.Biomater. Sci. Polym. Ed. 6, 15, 1994.
Address reprint requests to:Yin Xiao, Ph.D.
Level 4, Q-blockLife Sciences
Queensland University of Technology2 George Street
