Hydroxy apatite microspheres enhance gap junctionalintercellular communication of human osteoblastscomposed of connexin 43 and 45

Ryusuke Nakaoka, Saifuddin Ahmed, Toshie TsuchiyaDivision of Medical Devices, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya-ku,Tokyo 158-8501, Japan

Received 26 September 2004; revised 14 December 2004; accepted 14 December 2004Published online 17 June 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.a.30328

Abstract: The aseptic loosening of artificial joints with as-sociated periprosthetic bone resorption may be partly due tothe suppression of osteoblast function to form new bone bywear debris from the joint. To assess the effect of wear debrison osteoblasts, effects of model wear debris on gap junc-tional intercellular communication (GJIC) of normal humanosteoblasts were estimated. The GJIC activity of the osteo-blasts after a 1-day incubation with the microspheres wassimilar to that of normal osteoblasts. However, hydroxyapatite particles, which have been reported to enhance thedifferentiation of osteoblasts in contact with them, enhancedthe GJIC function of the osteoblasts. From RT-PCR studies,not only connexin 43 but also connexin 45 is suggested toplay a role in the GJIC of the osteoblasts in an early stage of

coculture with the microspheres, although it is still unclearhow these connexins work and are regulated in the GJIC anddifferentiation. However, this study suggests that there is arelationship between the early levels of GJIC and the differ-entiation of the cells. Therefore, estimating the effect ofbiomaterials, even in the microsphere form, on the GJIC ofmodel cells, with which the biomaterials may be in contact invivo, can provide important information about their biocom-patibility. © 2005 Wiley Periodicals, Inc. J Biomed Mater Res74A: 181–186, 2005

Key words: gap junctional intercellular communication; hu-man osteoblasts; microspheres; hydroxy apatite; connexin

INTRODUCTION

Biomaterials implanted into the harsh environmentof the body cannot maintain their original shape, oreven their desired function, sometimes resulting inundesirable side effects. One well-known example isthe aseptic loosening of artificial joints observed inmany patients who underwent a total joint replace-ment 5 to 25 years ago. It has already been reportedthat aseptic loosening with associated periprostheticbone resorption is partly due to the activation of mac-rophages and osteoclasts by wear debris from theartificial joint.1–14 Macrophages stimulated by weardebris in vitro release significant amounts of inflam-matory mediators such as interleukin-1, interleukin-6,prostaglandin E2, collagenase, and tumor necrosis fac-tor.6–14 In addition, the biological effects of wear de-bris may depend on the type of material used as well

as the shape, size, and amount of the debris.4–11 There-fore, it is important to estimate the biocompatibility ofbiomaterials with not only their original shape butalso possible transformed shapes after their usage.

During the last decade, we have been researchingthe inhibitory potential of many kinds of biomaterialson gap junctional intercellular communication (GJIC)as an index for their biocompatibility.15–18 GJIC is afunction that plays an important role in maintainingcell and tissue homeostasis by exchanging low molec-ular weight molecules, which results in regulating cellgrowth, development, and differentiation of cells.19,20

Therefore, it is reasonable that disruption of this func-tion is the cause of many kinds of diseases. In a pre-vious report,18 we examined the inhibitory activity ofpolymer microspheres, which were used as modelwear debris from biomedical polymer in vivo, on theGJIC of rodent-derived fibroblasts. We concluded thatestimating the inhibitory activity of the microsphereson the GJIC might be useful for considering their sideeffects in the body. In other words, it may be possibleto predict whether wear debris causes aseptic loosen-ing of artificial joints by estimating their effect on GJICfunction.

No benefit of any kind will be received either directly orindirectly by the authors

Correspondence to: R. Nakaoka; e-mail: nakoaka@nihs.go.jp

© 2005 Wiley Periodicals, Inc.

However, it must be noted that the effects of themicrospheres may be different when the effects on theGJIC of human-derived cells are estimated. Osteo-blasts have been reported to communicate with oneanother via GJIC function, and the function is believedto be critical to the coordinated cell behavior necessaryin bone tissue development.21,22 Therefore, the ques-tion is raised whether wear debris has an inhibitoryeffect on the GJIC and the GJIC inhibition has a rela-tion with the aseptic loosening of artificial joints. Be-cause we have already observed some precoated poly-mer microspheres around 5 �m in diameter showedthe potential to inhibit GJIC of fibroblasts contactingwith them,23 we estimated effects of various micro-spheres around 5 �m in diameter on GJIC functionusing normal human osteoblasts to discuss the rela-tionship between the GJIC and the differentiation ofosteoblasts. In this study, we employed fluorescencerecovery after photobleaching (FRAP) analysis for es-timating the GJIC function,17 and assessed the poten-tial effect of many kinds of microspheres on the GJIC.

MATERIALS AND METHODS

Microspheres

Monodispersed polystyrene (PS) microspheres (5 �m indiameter) were purchased from Japan Synthetic Rubber Co.,Ltd. (Tokyo, Japan). Low-density polyethylene (PE) micro-spheres were generously supplied by Sumitomo Seikachemicals Co., Ltd. (Tokyo, Japan). Alumina (Al2O3) micro-spheres were obtained from the Association of Powder Pro-cess Industry and Engineering. Sintered hydroxy apatitemicrospheres (HA, 7.2 �m in diameter) were prepared andsupplied by Ube Material Industries, Ltd. A Multisizer II(Coulter Electronics Inc., Hialeah, FL) was used to determinethe average diameter of PE and alumina microspheres: 6.4and 5.1 �m, respectively. Microspheres were sterilized bydispersing them in a 70% ethanol solution, followed bycentrifugation in sterile conditions to remove the ethanolsolution. The microspheres were dispersed in sterile meth-anol for cell differentiation tests at specified concentrations.The suspension of microspheres in methanol was added to35-mm type I collagen-coated cell culture dishes (Asahitechno glass, Chiba, Japan), and the plates dried overnight atroom temperature. The obtained microsphere-coated dishes(100 �g/dish) were subjected to the assays.

Cell culture

Normal human osteoblasts (NHOst) were purchased fromBioWhitteker Inc. (Walkersville, MD). The standard cultureof NHOst was performed using alpha minimum essentialmedium (Gibco) containing 20% fetal calf serum (FCS)(Kokusai Shiyaku Co., Ltd., Tokyo, Japan). The cells were

maintained in incubators under standard conditions (37°C,5%-CO2-95%-air, saturated humidity). All assays were per-formed using alpha minimum essential medium containing20% FCS, supplemented with 10 mM beta-glycerophos-phate. NHOst (1 �105 cells/dish/2.5 mL medium) werecultured on microsphere-coated dishes for estimating theeffect of the microspheres interacted from the bottom of thecells. To estimate the effect of microspheres on cells adheredto the culture plates, the NHOst cells were cultured withmicrosphere-containing medium (100 �g/2.5 mL medium)after they had adhered to the collagen-coated dishes. Thetest cells were cultured while changing the medium threetimes when the measurement of GJIC was performed after a7-day incubation.

Measurement of GJIC activities

NHOst cultured with microspheres were subjected to flu-orescence recovery after photobleaching (FRAP) analysis toestimate the inhibitory activity of these microspheres towardthe GJIC. FRAP analysis was carried out according to anoriginal procedure by Wade et al.,24 with some modifica-tions.17 Briefly, NHOst were plated on microsphere-coateddishes and incubated for 1 or 7 days. After a wash withphosphate buffer saline (PBS) containing MgCl2 and CaCl2[PBS(�)], the cells were incubated for 5 min at room tem-perature in PBS(�) containing 5,6-carboxyfluorescein diac-etate (7 �g/mL, excitation 488 nm and emission 515 nm).After the washing off of excess extracellular dye withPBS(�), the cells in the test dishes in PBS(�) were subjectedto the FRAP analysis. In the control experiment, cells wereinoculated on an untreated glass bottom dish and treatedwith the same procedure as the tested cells. Cells in contactwith test microspheres and at least two other cells weresubjected to FRAP analysis under an Ultima-Z confocal mi-croscope (Meridian Instrument, Okemos, MI) with a 10�objective lens at room temperature. The cells were photo-bleached with a 488-nm beam and the recovery of fluores-cence intensity was subsequently monitored at 1-min inter-vals for a total period of 4 min. The data obtained from morethan seven independent cells were expressed as the averageof fluorescence recovery rate in comparison to the rate ob-tained from NHOst cultured without microspheres.

Effect of microspheres on calcium deposition byNHOst

The amount of calcium deposited during a 7-day incuba-tion of the cells were evaluated as follows: NHOst werecocultured with either precoated or added microspheres in24-well collagen-coated culture plates (Asahi techno glass,Chiba, Japan) for 1 week (2 �104 cells/20 �g microspheres/well/500 �L medium). After the cells were fixed in formal-dehyde, 0.5 mL of 0.1 M HCl was added to each well afterwashing the cells with PBS. The amounts of calcium dis-solved in HCl were estimated using a Calcium detecting kit(Calcium-C test Wako, Wako, Osaka, Japan) according to themanufacturer’s direction.

182 NAKOAKA, AHMED, AND TSUCHIYA

RT-PCR for estimating expression of connexins

According to the method reported by Ichikawa et al.,25

RT-PCR was performed to detect the expression of connexinmRNA in NHOst. After culturing NHOst with microspheresfor a scheduled time, total RNA was extracted from theNHOst using TRIZOL� reagent (Invitrogen Corp., Carlsbad,CA) according to the manufacturer’s instructions. After dis-solving the RNA in diethylpyrocarbonate-treated water, thetotal RNA concentration was measured spectrophotometri-cally using Genequant (Amarsham Biosciences Corp., Pisca-taway, NJ). RNA samples were adjusted to a minimumconcentration among collected samples in each experimentand reversibly transcribed to cDNA using Superscript™ II(Invitrogen Corp.). For PCR amplification of human con-nexin 45, Takara Ex-Taq™ (Takara Shuzo Co., Ltd., Shiga,Japan) was used with Ex-Taq™ buffer consisting of 20 pmoleach of two human connexin-45 specific primers (forward5�GTGGCAACTCCCTCTGTGAT3� and reverse 5�GGATC-CTCAAGTTCCCTCCT3�). For PCR amplification of humanconnexin 26, 32, and 43, Takara LA-Taq™ (Takara ShuzoCo., Ltd.) was used with Ex-Taq™ buffer consisting of 6pmol each of the human connexin-specific primers (for con-nexin 26, forward 5�ATGGATTGGGGCACGC3� and reverse5�TTAAACTGGCTTTTTTGACTTCCC3�. For connexin 32,forward 5�ATGAACTGGACAGGTTTGTACACCTTGCTC3�and reverse 5�TCAGCAGGCCGAGCAGCGG3�. For connexin43, forward 5�ATGGGTGACTGGAGCGCCTTAGGC3� andreverse 5�CTAGATCTCCAGGTCATCAGGCCG3�). ThePCR profile for connexin 45 involved pretreatment at 95°Cfor 2 min, followed by 35 cycles of denaturation at 95°C for45 s, annealing at 54°C for 45 s, and extension at 72°C for90 s. The PCR profile for connexin 26, 32, and 43 (35 times)was as follows: pretreatment at 95°C for 2 min, denaturationat 95°C for 30 s, annealing at 54°C for 30 s, and extension at72°C for 120 s. Reaction products were analyzed by electro-phoresis in 1.5% (w/v) agarose gel, followed by staining of

the products by SYBR� Green I (Takara Shuzo Co., Ltd.) anddetection of a 566-bp (connexin 45), 671-bp (connexin 26),852-bp (connexin 32), and 1149-bp (connexin 43) band, re-spectively. For the standardization of connexin cDNA, PCRamplification of glyceraldehyde-3-phosphate dehydroge-nase (GAPDH) mRNA in each sample was performed usingGAPDH-specific primers (forward 5�CCCATCACCATCT-TCCAGGAGCGAGA3� and reverse 5�TAAGTAGGACAA-CAAGGAGGTCGTGACGACGC3�; product size 578-bp).All reactions included negative controls without cDNA.

Statistical analysis

All data were expressed as the mean value � the standarderror of the means of the obtained data and treated statisti-cally with Student’s t test.

RESULTS

Figure 1 shows effects of various microspheres onGJIC of NHOst in contact with the microspheres for 1and 7 days. The microspheres were precoated on 35-mmculture dishes before cell seeding. When the NHOstwere cultured with precoated PS, PE, and alumina mi-crospheres, their GJIC level was similar to that in NHOstcultured on a normal culture dish. On the other hand,the GJIC level was 1.5 times that of NHOst on the normaldish when they were cultured with precoated hydroxyapatite microspheres. After 7 days, the GJIC of NHOst incontact with microspheres became similar to that of nor-mal NHOst, irrespective of the type of microsphere. Thechange in GJIC of NHOst in contact with added micro-spheres is shown in Figure 2. As seen in Figure 1, hy-

Figure 1. Effect of precoated microspheres on gap junc-tional intercellular communication of NHOst estimatedfrom fluorescence recovery rates of target cells. The recoveryrates of the cells on untreated culture dishes on days 1 and7 were used as standards of all obtained data, respectively.(*p � 0.01 against culture dish).

Figure 2. Effect of added microspheres on gap junctionalintercellular communication of NHOst estimated from flu-orescence recovery rates of target cells. The recovery rates ofthe cells on untreated culture dishes on days 1 and 7 wereused as standards of all obtained data, respectively.

HYDROXY APATITE MICROSPHERES 183

droxy apatite microspheres enhanced their GJIC after a1-day culture compared to cells on a normal plate. Thedegree of enhancement of GJIC is, however, smaller thanthat seen in NHOst in Figure 1, and no significant dif-ference was observed between NHOst in contact withthe hydroxy apatite microspheres and those culturedwithout microspheres. In addition, Figure 2 indicatesthat addition of PS microspheres into a culture of NHOstinhibited GJIC.

To consider the effects of tested microspheres on notonly GJIC but also the differentiation of NHOst,changes in the amount of calcium deposited after a1-week coculture of NHOst with various micro-spheres were estimated. From Figure 3, it is suggestedthat there is the possible relation between the GJIC ofNHOst cocultured with microspheres for 1 day and

the amount of calcium deposited after a 1-week cocul-ture with the same microspheres.

To clarify which connexins exist in NHOst, we per-formed RT-PCR to detect mRNA of connexin 26, 32,43, and 45 in NHOst cultured on a normal culturedish. Figure 4 shows the result of RT-PCR to amplifythe mRNA from whole RNA collected from NHOstcultured for 1 and 7 days. As shown in the figure, onlyconnexin 43 and 45 were detected in NHOst. Whencells were cultured for 7 days, connexin 43 was de-tected at a lower level than that detected after the1-day culture, while connexin 45 was not detected.

Figures 5 and 6 show the results of RT-PCR toamplify mRNA of connexin 43 and 45 in NHOst cul-tured with various precoated microspheres. TheNHOst cultured with microspheres did not expressmRNA of connexin 43, except those with PE micro-spheres. After 7 days, the expression was suppressedin the normal NHOst while the expression was ob-served in NHOst cultured with microspheres, irre-spective of kind of the microsphere. On the otherhand, mRNA expression of connexin 45 was sup-pressed after a 1-day culture of NHOst only withalumina microspheres, followed by a decrease in ex-pression of the mRNA after their 7-day culture.

Figure 3. Relationship between GJIC on day 1 and calciumdeposition ratio after 7-day coculture of NHOst with variousmicrospheres (r2 � 0.74).

Figure 4. Expression of mRNA of various connexins (Cx)in NHOst cultured for 1 and 7 days. The number of NHOstcultured on 35-mm collagen-coated culture dishes was 2 �105. RT-PCR cycles of each lane are expressed at the bottomof the figure.

Figure 5. Expression of connexin 43 (Cx 43) mRNA inNHOst cultured with various precoated microspheres. Thenumber of PCR cycles for connexin 43 and GAPDH is 35 and20, respectively. Lane 1: without microspheres; lane 2: withPS microspheres; lane 3: with PE microspheres; lane 4: withalumina microspheres; lane 5: with HA microspheres.

Figure 6. Expression of connexin 45 (Cx 45) mRNA inNHOst cultured with various precoated microspheres. Thenumber of PCR cycles for connexin 45 and GAPDH is 40 and20, respectively. Lane 1: without microspheres; lane 2: withPS microspheres; lane 3: with PE microspheres; lane 4: withalumina microspheres; lane 5: with HA microspheres.

184 NAKOAKA, AHMED, AND TSUCHIYA

DISCUSSION

As shown in Figure 3, normal human osteoblasts(NHOst) in contact with microspheres showed differ-ent levels of calcium deposition only after a 1-weekculture, suggesting composition of the microspheresaffects NHOst differentiation level. The differentiationwas suppressed by the contact with PS, PE, and alu-mina microspheres, while HA microspheres showedthe potential to enhance the differentiation. It has beenreported that GJIC plays an important role in not onlythe homeostasis of cells but also their differentia-tion.19–22 In addition, GJIC is affected by the micro-sphere’s composition, as has been reported using afibroblast cell line.18 Therefore, the results shown inFigures 1 and 2 suggest that the enhanced differenti-ation of NHOst relates to GJIC enhancement on a1-day culture in contact with HA microspheres, espe-cially the precoated microspheres. In addition, on co-culture with other microspheres, GJIC was slightlysuppressed at 1 day, although no significant differencecompared to control NHOst was observed. We havealready studied effects of the microspheres on NHOstdifferentiation, and enhancement of calcium deposi-tion by coculture with the hydroxy apatite micro-spheres was observed. Figure 3 suggests a relationshipbetween the calcium deposition and GJIC on day 1.This also indicates that GJIC of the NHOst, in contactwith materials in the microsphere form, in the earlystage may be one factor affecting their differentiation.

It has been reported that GJIC of cells derived fromhuman osteoblasts is mainly composed of connexin 43and 45.22,26,27 In this study, it is also indicated thatGJIC of NHOst is composed of connexin 43 and 45(Fig. 4). Therefore, it is possible that changes in thelevel of their GJIC is ascribed to the change in mRNAexpression level of connexin 43 and 45 and their ex-pression ratio. From Figures 5 and 6, mRNA of con-nexin 43 was expressed only in normal NHOst andthose cultured with PE microspheres, while it wasslightly expressed in NHOst cocultured with HA. Onthe other hand, mRNA of connexin 45 was expressedin NHOst in all conditions, except those coculturedwith alumina microspheres. Because HA was ob-served to enhance GJIC of NHOst, this suggests thatconnexin 45 may play a role in GJIC at an early stage.This also suggests that a higher level of connexin 45than that of connexin 43 may be important in theenhancement of GJIC. However, although the mRNAexpression of neither connexin 43 nor 45 was observedin NHOst cocultured with alumina microspheres,their GJIC was similar to that of normal NHOst. More-over, it has reported that gap junctions formed byconnexin 43 are more permeable to negatively chargeddyes such as lucifer yellow, calcein, and carboxyfluo-rescein used in this study, more than those formed by

connexin 45, and an increase of connexin 43 expres-sion and GJIC function parallel osteoblast differentia-tion.22,28 These are inconsistent with our findings andindicate that not high expression, but a rapid decreaseof connexin 45 mRNA is probably very important forGJIC change and differentiation of the osteoblasts.Therefore, even though connexin 45 may play an im-portant role in the early stage of GJIC in NHOst, it isprobable that another connexins or other mechanismsof GJIC play a role in the GJIC of NHOst.

Because many proteins are involved in GJIC forma-tion,28 other mechanisms or proteins may be impor-tant in the GJIC change induced by the contact withthe microspheres. It has reported that cadherins,which are important proteins for form tight junctionbetween cells, control connexin 43-mediated GJIC.29,30

In addition, a microtubule network inside a cell hasbeen reported to play an important role as guidancefor delivery of connexons, which are composed of sixconnexin molecules, to the cell membrane to make gapjunctions.31 Usually, surface characteristics of materi-als affect cell attachment as well as cell morphology,suggesting signal cascades of cell attachment and cy-toskelton rearrangement in the cell were influenced bythe characteristics. Therefore, it is probable that a sur-face characteristic of the microspheres affect thesemolecules in NHOst, resulting in changes of GJICactivities. Further studies on changes in not only con-nexin molecules but also other molecules such as cad-herin, actin, and microtubule in NHOst, is necessaryto clarify the mechanism of GJIC. In the future, we willstudy the above, and find another molecules partici-pating in the GJIC of NHOst and the mechanismsregulating the connexins in NHOst.

In conclusion, the GJIC level of NHOst changes oncontact with microspheres, and is affected by the com-position of the microspheres. The GJIC level in theearly stage might be important in the differentiationcontrol of NHOst and the level may be controlledpartly by expression of connexin 43, connexin 45, andunclarified connexins in addition to other mechanismsregulating GJIC function. Detecting a biomaterial’seffect on the GJIC of human cells may be one usefulmethod for estimating its biocompatibility.

The authors appreciate the support of Health and LaborSciences Research Grants for Research on Advanced MedicalTechnology, Research on Health Sciences focusing on DrugInnovation, and Risk Analysis Research on Food and Phar-maceuticals, Ministry of Health, Labour and Welfare.

References

1. Willert HG, Semlitsch M. Reactions of the articular capsule towear products of artificial joint prosthesis. J Biomed Mater Res1977;11:157–164.

HYDROXY APATITE MICROSPHERES 185

2. Savio JA III, Overcamp LM, Black J. Size and shape of bioma-terial wear debris. Clin Mater 1994;15:101–147.

3. Wang W, Ferguson DJP, Quinn JMW, Simpson AHRW, Atha-nasou NA. Osteoclasts are capable of particle phagocytosis andbone resorption. J Pathol 1997;182:92–98.

4. Goodman SB, Fornasier VL, Lee J, Kei J. The histological effectson the implantation of different sizes of polyethylene particlesin the rabbit tibia. J Biomed Mater Res 1990;24:517–524.

5. Kubo T, Sawada K, Hirakawa K, Shimizu C, Takamatsu T,Hirasawa Y. Histiocyte reaction in rabbit femurs to UHMWPE,metal, and ceramic particles in different sizes. J Biomed MaterRes 1999;45:363–369.

6. Shanbhag AS, Jacobs JJ, Black J, Galante JO, Glant TT. Macro-phages/particle interactions: Effect of size, composition andsurface area. J Biomed Mater Res 1994;28:81–90.

7. Kim KJ, Itoh T, Tanahashi M, Kumegawa M. Activation ofosteoclasts-mediated bone resorption by the supernatant froma rabbit synovial cell line in response to polyethylene particles.J Biomed Mater Res 1996;32:3–9.

8. Voronov I, Santerre JP, Hinek A, Callahan JW, Sandhu J, Boyn-ton EL. Macrophage phagocytosis of polyethylene particles invitro. J Biomed Mater Res 1998;39:40–51.

9. Catelas I, Huk OL, Petit A, Zukor DJ, Marchand R, Yahia L.Flow cytometric analysis of macrophage response to ceramicand polyethylene particles: Effects of size, concentration, andcomposition. J Biomed Mater Res 1998;41:600–607.

10. Green TR, Fisher J, Stone MH, Wroblewski BM, Ingham E.Polyethylene particles of a “critical size” are necessary for theinduction of cytokines by macrophages in vitro. Biomaterials1998;19:2297–2302.

11. Green TR, Fisher J, Matthews JB, Stone MH, Ingham E. Effectof size and dose on bone resorption activity of macrophages byin vitro clinically relevant ultra high molecular weight poly-ethylene particles. J Biomed Mater Res (Appl Biomater) 2000;53:490–497.

12. Takei I, Takagi M, Santavirta S, Ida H, Hamasaki M, Ishii M,Fukushima S, Ogino T, Konttinen YT. Matrix metalloprotein-ases and tissue inhibitors of metalloproteinases in joint fluid ofthe patients with loose artificial hip joints. J Biomed Mater Res1999;45:175–183.

13. Trindade MCD, Schurman DJ, Maloney WJ, Goodman SB,Smith RL. G-protein activity requirement for polymethyl-methacrylate and titanium particle-induced fibroblast interleu-kin-6 and monocyte chemoattractant protein-1 release in vitro.J Biomed Mater Res 2000;51:360–368.

14. Sacomen D, Smith RL, Song Y, Fornasier V, Goodman SB.Effects of polyethylene particles on tissue surrounding kneearthroplasties in rabbits. J Biomed Mater Res (Appl Biomater)1998;43:123–130.

15. Tsuchiya T, Hata H, Nakamura A. Studies on the tumor-promoting activity of biomaterials: Inhibition of metabolic co-operation by polyetherurethane and silicone. J Biomed MaterRes 1995;29:113–119.

16. Nakaoka R, Tsuchiya T, Kato K, Ikada Y, Nakamura A. Studieson tumor-promoting activity of polyethylene: Inhibitory activ-ity of metabolic cooperation on polyethylene surfaces is mark-edly decreased by surface modification with collagen but notwith RGDS peptide. J Biomed Mater Res 1997;35:391–397.

17. Nakaoka R, Tsuchiya T, Nakamura A. The inhibitory mecha-nism of gap junctional intercellular communication induced by

polyethylene and the restorative effects by surface modifica-tion with various proteins. J Biomed Mater Res 2001;57:567–574.

18. Nakaoka R, Sakaguchi K, Tsuchiya T, Nakamura A. Studies onin vitro evaluation for the biocompatibility of various bioma-terials: Inhibitory activity of various kinds of polymer micro-spheres on metabolic cooperation. J Biomed Mater Res 2001;57:279–284.

19. Mensil M, Krutovskikh V, Omori Y, Yamasaki H. Role ofblocked gap junctional communication in non-genotoxic carci-nogenesis. Toxicol Lett 1995;82/83:701–706.

20. Yamasaki H. Role of disrupted gap junctional intercellularcommunication in detection and characterization of carcino-gens. Mutat Res 1996;365:91–105.

21. Sawada MS, Mano H, Hanada K, Kakudo S, Kameda T,Miyazawa K, Nakamaru Y, Yuasa S, Mori Y, Kumegawa M,Hakeda Y. Down-regulation of gap junctional intercellularcommunication between osteoblastic MC3T3–E1 cells by basicfibroblast growth factor and a phorbol ester (12-O-tetradeca-noylphorbol-13-acetate). J Bone Miner Res 1997;12:1165–1173.

22. Lecanda F, Towler DA, Ziambaras K, Cheng SL, Koval M,Steinberg TH, Civitelli R. Gap junctional communication mod-ulates gene expression in osteoblastic cells. Mol Biol Cell 1998;9:2249–2258.

23. Nakaoka R, Tsuchiya T. Biocompatibility of various kinds ofpolymer microspheres estimated from their effect on gap junc-tional intercellular communication of fibroblasts. Mater Trans2002;43:3122–3127.

24. Wade MH, Trosko JE, Steindler M. A fluorescence photo-bleaching assay of gap junctional-mediated communicationbetween human cells. Science 1986;232:525–528.

25. Ichikawa R, Tsuchiya T. A strategy for the suppression oftumorigenesis induced by biomaterials: Restoration of trans-formed phenotype of polyetherurethane-induced tumor cellsby Cx43 transfection. Cytotechnology 2002;39:1–8.

26. Donahue HJ, Li Z, Zhou Z, Yellowley CE. Differentiation ofhuman fetal osteoblastic cells and gap junctional intercellularcommunication. Am J Physiol Cell Physiol 2000;278:C315–C322.

27. Laing JG, Manley-Markowski RN, Koval M, Civitelli R, Stein-berg TH. Connexin45 interacts with zonula occludens-1 andconnexin43 in osteoblastic cells. J Biol Chem 2001;276:23051–23055.

28. Duffy SH, Delmar M, Spray DC. Formation of the gap junctionnexus: Binding partners for connexins. J Pathol Paris 2002;96:243–249.

29. Jongen WMF, Fitzgerald DJ, Asamoto M, Piccoli C, Slaga TJ,Gros D, Takeichi M, Yamasaki H. Regulation of connexin43-mediated gap junctional intercellular communication byCa2� in mouse epidermal cells is controlled by E-cadherin.J Cell Biol 1991;114:545–555.

30. Meyer RA, Laird DW, Revel JP, Johnson RG. Inhibition of gapjunction and adherens junction assembly by connexin andA-CAM antibodies. J Cell Biol 1992;119:179–189.

31. Lauf U, Giepmans BNG, Lopez P, Braconnot S, Chen SC, FalkMM. Dynamic trafficking and delivery of connexins to theplasma membrane and accretion to gap junctions in livingcells. Proc Natl Acad Sci USA 2002;99:10446–10451.

186 NAKOAKA, AHMED, AND TSUCHIYA