Designing of a Si-MEMS device with an integrated skeletal muscle cell-based bioactuator

Designing of a Si-MEMS device with an integrated skeletal muscle cell-based bio-actuator

Hideaki Fujita &Van Thanh Dau &Kazunori Shimizu &Ranko Hatsuda &Susumu Sugiyama &Eiji Nagamori

Published online:19October 2010

#Springer Science+Business Media,LLC 2010

Abstract With the aim of designing a mechanical drug delivery system involving a bio-actuator,we fabricated a Micro Electro Mechanical Systems (MEMS)device that can be driven through contraction of skeletal muscle cells.The device is composed of a Si-MEMS with springs and ratchets,UV-crosslinked collagen film for cell attachment,and C2C12muscle cells.The Si-MEMS device is 600μm×1000μm in size and the width of the collagen film is 250~350μm,which may allow the device to go through small blood vessels.To position the collagen film on the MEMS device,a thermo-sensitive polymer was used as the sacrifice-layer which was selectively removed with O 2plasma at the positions where the collagen film was glued.The C2C12myoblasts were seeded on the collagen film,where they proliferated and formed myotubes after induction of differentiation.When C2C12myotubes were stimulated with electric pulses,contraction of the collagen film-C2C12myotube complex was observed.When the edge of the Si-MEMS device was observed,displacement of ~8μm was observed,demon-strating the possibility of locomotive movement when the

device is placed on a track of adequate width.Here,we propose that the C2C12-collagen film complex is a new generation actuator for MEMS devices that utilize glucose as fuel,which will be useful in environments in which glucose is abundant such as inside a blood vessel.

Keywords Collagen-sheet .C2C12.Muscle .MEMS

1Introduction

Recent development of micro-fabrication technology has enabled the design and fabrication of micro-machines that can go through small blood vessels,raising the possibility of the production of mechanical intelligent drug delivery system machines.Several MEMS devices with transporta-tion capability have been designed and fabricated (Desai et al.1999;Pham et al.2006),however,they were not intended for in vivo use and thus an actuator was not built into the devices,the transportation being achieved through periodical movement of a rail or surrounding electrical potential.Thus,the design of a new device with a built-in actuator that can work in wet and ionic conditions is required for in vivo use,but the design of such a device is challenging due to the difficulty in making a small but powerful actuator that can work in wet and ionic environments.Molecular motors are some interesting candidates for such actuators,and several attempts have been made to employ a molecular motor as an actuator (Suzuki et al.1997;Hiratsuka et al.2001),and the use of molecular motors in MEMS devices has been reported (Limberis et al.2000;Soong et al.2000).In spite of researchers ’efforts,production of sufficient force to drive devices using molecular motors has not been reported to date because of the extremely small force per molecule and fragility of proteins.

Electronic supplementary material The online version of this article (doi:10.1007/s10544-010-9477-3)contains supplementary material,which is available to authorized users.H.Fujita :K.Shimizu :E.Nagamori (*)Toyota Central R&D Labs.Inc.,41-1Yokomichi,

Nagakute,Aichi 480-1192,Japan e-mail:nagamori@mosk.tytlabs.co.jp

V .T.Dau :R.Hatsuda :S.Sugiyama

Research Institute for MicroSystem Technology,Ritsumeikan University,

Kusatsu,Shiga 525-8577,Japan

Biomed Microdevices (2011)13:123–129DOI 10.1007/s10544-010-9477-3

Recently,several attempts to use muscle cells as an actuator were made,mostly involving primary cultured cardiomyo-cytes(Yasuda et al.2001;Garcia-Webb et al.2007).The advantage of the use of muscle cells is that it can resolve the disadvantages of the use of a molecular motor as an actuator mentioned above.Numerous myosin motors are packed at high density into a well organized structure within a sarcomere,providing~2μN per cardiomyocyte(Shepherd et al.1990;Iribe et al.2007).Furthermore,cells provide a protein regeneration system,an energy supply and waste disposal system,and an excitation system which are all important for actuators.On the other hand,the use of cardiomyocytes has several drawbacks such as spontaneous contraction resulting in difficulty regarding controlled con-traction and the need of animal sacrifice because cardio-myocytes do not proliferate.Thus,use of a skeletal muscle myoblast cell line will be more ideal for bio-actuators because contraction can be controlled through electric pulses and myoblasts can undergo mitosis,enabling proliferation. Furthermore,skeletal muscle can generate a stronger force compared to cardiomyocytes,which is a preferable charac-teristic for an actuator.

Recently,it was reported that when the C2C12murine myoblast cell line was seeded onto and allowed to differentiate on UV-crosslinked thin collagen film,the collagen film-myotube complex exhibited active tension with a few micro-meters of shortening(Fujita et al.2010). Because collagen film can be fabricated small enough to be integrated into a Si-MEMS device,the collagen film-myotube complex may be usable as a bio-actuator.To demonstrate the possibility of the collagen film-myotube complex being used as an actuator,we have designed a MEMS device that can be driven by the C2C12-collagen film complex,which moves on electric pulse stimulation. Here,we propose that the C2C12-collagen film complex is a new generation actuator which utilizes glucose as an energy source.The actuator developed in this study will be useful when a micro to millimeter sized actuator is needed, such as for a micro-machine for in vivo use.The actuator can also be utilized in Micro Total Analysis System (μTAS),a small device which holds multiple laboratory functions on single chip,where small actuator is needed to power pump or valve in the system.

2Materials,methods and experimental setup

2.1Fabrication of a Si-MEMS device with springs

and ratchets

A schematic illustration of the designed Si-MEMS device is presented in Fig.1(a)and an SEM image of the fabricated device is shown in Fig.1(b).The device was fabricated using a(100)silicon on insulator wafer consisting of a 10μm thick device layer and a1μm thick buried oxide layer.Positive photoresist(OFPR-800)was spin-coated onto a wafer and patterned with a photomask.The wafer was etched by deep reactive ion etching,followed by buffered HF etching to remove the buried oxide layer,a Si-MEMS device being yielded.The thickness of the Si-MEMS device is10μm.The Si-MEMS device is composed of two main components that are connected by two main springs.Several devices having different spring constants between k=0.15~2μN/μm were designed be-cause there was large diversity of force produced by myotubes previously reported,ranging from0.5to10kPa (McMahon et al.1994;Dennis et al.2001;Shimizu et al. 2010b).The spring constants were calculated using Eq.1 k?E?I=en?uTe1Twhere n is the number of peaks,E is Young’s modulus and I is internal moment of the springs,respectively.The internal moment of the springs were calculated with Eq.2

I?h?w3=12e2T

where w and h are the width and thickness of the spring, respectively.Shape factor u was defined by Eq.3

u?2=3?D3tp rD2t4p r2Dt1=2?p r3e3T

where D is half span of the main spring.Electrical stimulation experiments described below were performed with a device having a main spring of k=0.15μN/μm. Each component harbors two sub-springs with a spring constant of k=1μN/μm,which hold the ratchet mecha-nism.The Si-MEMS device was thus designed so that the collagen film-myotube bio-actuator bridges the two main components.The contracting force of myotubes is expected to retract the main springs and approximate the main components when myotubes are excited,whereas the restitution force of the main springs separates the main components during relaxation.Repeated excitation and relaxation is expected to result in locomotive movement in an inchworm manner when the device is placed on an appropriate track on which the ratchet system is effective. The Si-MEMS device is held to the base at five positions (indicated by grey arrowheads in Fig.1(a)),which are removed with a micro-manipulator before the experiment.

2.2Integration of collagen film into the Si-MEMS device The procedure for integration of the collagen film into the Si-MEMS device is illustrated in Fig.2(a).To position the collagen film over the gap between the main components,

we used poly-N-isopropylacrylamide (PNIPAAm)as a sacrifice layer,as reported previously with some modifica-tions to remove Cr/Au film from a microdevice (Xi et al.2005).PNIPAAm dissolves in water when the temperature is below 32°C.Thus,PNIPAAm can be removed without any thermal damage to the cells.First,1μl of 5%PNIPAAm (Polysciences,Inc.,Warrington,PA)in ethanol was poured onto the Si-MEMS device,followed by drying for 1day.Consequently,the Si-MEMS device and the base were buried in PNIPAAm.Next,a stencil mask was placed on the Si-MEMS device so that only the position where the collagen film was to be glued were exposed using polyimide tape,followed by O 2plasma etching treatment at 100W for 3min to remove PNIPAAm in a rectangular shape using a high-frequency oxygen plasma generator (PiPi;Yamato Materials,Tokyo,Japan).The position where PNIPAAm was removed is illustrated in Fig.2(b)(blue),and removal of the PNIPAAm layer can be seen as a

whitish rectangle in a bright field micrograph (Fig.2(c)).Gel-type crazy glue (Toagosei Co.Ltd.,Tokyo,Japan)was placed at the position where the PNIPAAm had been removed.The position where the gel-type crazy glue was placed is shown in Fig.2(b)(pink).UV-crosslinked collagen film with a thickness of 20μm (Koken,Tokyo,Japan)was cut into rectangles of 250~350μm in width and ~1mm in length with a scalpel,and placed on the glue.The position where the collagen film was placed is shown in Fig.2(b)(green).The PNIPAAm layer surrounding the Si-MEMS device prevents adhesion of the device to the base and protects the device from damage by the glue.A bright field micrograph of the device after integration of the collagen film is shown in Fig.2(d).Because PNIPAAm becomes hydrophilic below 32°C,it can be removed by lowering the temperature below 32°C in the medium.2.3Cell and cell culture

Murine skeletal muscle cells,C2C12,were obtained from the RIKEN Bioresource Center (Ibaraki,Japan)and cultured as described previously (Fujita et al.2007).Briefly,the cells were cultured in growth medium (GM)consisting of Dulbecco ’s Modified Eagle Medium (DMEM;Invitrogen,Carlsbad,CA)containing 10%fetal bovine serum (FBS;ICN Biomedicals,Inc.,Aurora,OH),100U/ml potassium penicillin G,and 100μg/ml streptomycin sulfate (Invitrogen)for proliferation,and in differentiation medium (DM)consisting of DMEM containing 2%horse serum (Montarras et al.1989)(HS,JRH Biosciences,Inc.,Lenexa,KS),100U/ml potassium penicillin G,and 100μg/ml streptomycin sulfate for differentiation.A medium change was performed every day.To seed cells onto the Si-MEMS-collagen film-integrated device,the device was adhered to an ultralow attachment dish (Corning,New York,NY)using PNIPAAm.A cloning ring of a diameter of 7mm was placed around the device and C2C12cells were seeded onto the device at a density of 105cells/cm 2.Excess cells were washed off with GM at 4h after seeding.The cells reached confluence the following day and the medium was changed to DM.

2.4Immuno-fluorescence microscopy

To visualize the myotubes,cytochemical experiments were performed as previously described (Fujita et al.2009b ).Briefly,C2C12cells were fixed in PBS containing 4%paraformaldehyde for 10min,permeabilized with PBS containing 2%of Triton X-100,washed with PBS three times,and then incubated with the primary antibody for 1h.A monoclonal anti-αactinin (sarcomeric)antibody (EA-53;Sigma,Saint Louis,MO)was used at the dilution of 1:1000.Then,the cells were incubated with the secondary

antibody

Fig.1(a )Schematic drawing of the design of the Si-MEMS device.Three types of the device were fabricated with different spring constants for the main spring.(b )SEM image of the fabricated device (left ),after buried oxide layer was removed by vapor HF etching.Free-standing MEMS structure was successfully fabricated without sticking phenomenon (insets ).Scale bar,200μm

conjugated with Alexa488and Alexa546-phalloidin.All fluorescent probes used in the immunofluorescence study were from Invitrogen.The cells stained with the anti α-actinin antibody were regarded as myotubes.The specimens were washed three times with PBS and then studied under a fluorescence microscope (BX-51;Olympus,Tokyo,Japan).2.5Electrical stimulation of the device and measurement of movement

The assembled device was immersed in the medium under the culture conditions,and myotubes cultured on the device for 7days were actuated by electrical signals from carbon electrodes placed on either side of the device.Electric pulses were generated with a PC with specially designed LabView software (National Instruments,Austin,TX)and a hand-made amplifier.For twitch contraction,electric pulses of 1Hz were applied.For tetanic contraction,20Hz pulses were applied for 11s.The voltage was fixed at 1V/mm for both stimulations.The movement of the edge of the device driven by the collagen film-myotube complex was moni-tored under a microscope (BX-51,Olympus)equipped with a warm plate (MP10DM,Kitazato Supply Co.Ltd.,Shizuoka,Japan)and recorded on the PC as a digital movie.

To measure the displacement of the edge of the device,the acquired movie was evaluated with image processing software (VISION software,National Instruments).

3Results

3.1Assembly of the device with myotubes

Figure 3(a)shows a whole view of the device observed under a bright field microscope 1day after seeding of C2C12cells.A fluorescence micrograph revealed that the cells were only attached to the collagen film;i.e.,not onto the Si-MEMS device (Fig.3(b)),because the Si-MEMS device was covered with PNIPAAm,which is known to prevent cell attachment (Takezawa et al.1990;Fujita et al.2009a ).Sarcomeric α–actinin positive cells were not observed because differentiation was not induced at this point.We placed the collagen film so that some part of it extended out of the Si-MEMS device,so that cells on the collagen film could be observed with a phase contrast microscope.Figure 3(c)shows a phase contrast image of the collagen film extending out of the Si-MEMS device,showing that cells were attached to the collagen

film.

Fig.2(a )Procedure for inte-grating the UV-crosslinked col-lagen film into the Si-MEMS device.(1)The Si-MEMS de-vice was buried in PNIPAAm to protect it from excess glue in step (4).(2)A stencil mask was placed so as to expose the

positions where the glue was to be placed in step (4)and the O 2plasma was applied.(3)After O 2plasma treatment,the stencil mask was removed.(4)Glue was placed where the PNIPAAm had been removed.(5)The UV-crosslinked collagen film was placed on the glue.(6)PNI-PAAm can be removed by low-ering the temperature below to 32°C in medium.(b )Positions where the PNIPAAm was removed (blue),the glue was placed (pink),and the collagen film was placed (green).(c )Bright field micrograph showing the removal of PNIPAAm by O 2plasma treatment.(d )Bright field micrograph after integra-tion of the collagen film

Because the device is placed in an ultralow attachment dish,cells can only be found on the collagen film.Cells reached confluence 1day after seeding because they were seeded at a high density (105cells/cm 2).In this study,cells only attached to the area where they were intended to be;i.e.,on the collagen film.When a mechano-bio hybrid machine is considered,it is important that the cells are placed only where they are planned to be,which is generally achieved by applying molecules exhibiting high cell affinity (Kane et al.1999;Shen et al.2008)or by surface modification (Ozcan and Hasirci 2007;Alves et al.2008;Shimizu et al.2010a ).The use of UV-crosslinked collagen film provides another simple method for the patterning of cells.In this study,we used pre-crosslinked collagen film,which enables study without a strong UV light source and an optical setup.On-site cross linking of collagen could be one of the possible approaches for integrating the UV-crosslinked collagen film and Si-MEMS device.This will be examined in the future.When the medium was changed to DM,myoblasts fused and formed myotubes (Fig.4(a)),as previously reported (Yaffe and Saxel 1977).A fluorescence micrograph showed the presence of sarcomeric-α-actinin positive cells on the collagen film (Fig.4(b)).A magnified image showed the presence of huge multi-nucleated cells,indicating that myotubes had been successfully formed through cell fusion

(Fig.4(c)).Large myotubes were often observed at the periphery of the collagen film,and a magnified image showed that these myotubes at the periphery were attached to the sides of the collagen film (Fig.4(c)).Because the thickness of the collagen film was 20μm,about the same as the diameter of myotubes,these myotubes at the sides of the collagen film were aligned in the direction of the device,which is ideal for an actuator.Thus,these myotubes at the sides of the collagen film are expected to substantially contribute to the movement of the device.On the other hand,a significant number of the myotubes on the surface of the collagen film were off-axis,which may be improved by applying micro-scale features to the collagen film.It has been reported that cells can be aligned using surface with linear microgrooves (Vernon et al.2005),continuous wavy micropatterns (Lam et al.2006),or by electrospun nanofibers (Huang et al.2006).Applying such features on UV-crosslinked collagen film is expected to enhance the alignment of the myotubes.3.2Electrical stimulation of the device

Five days after induction of differentiation,the device was placed between two carbon electrodes and subjected to continuous electric pulse stimulation at 1Hz,2ms pulse duration.Continuous electrical pulse stimulation at 1Hz

is

Fig.3(a )Bright field micrograph of the device at 1day after seeding of C2C12cells.Scale bar,1mm.(b )Immuno-fluorescence micro-graph of the device at 1day after seeding of C2C12cells.The rear edge is shown.(Red )Actin filaments,(green )sarcomeric-α-actinin,and (blue )nuclei.Scale bar,200μm.Because the cells had not differentiated,α-actinin positive cells were not found.(c )Phase contrast micrograph of the collagen film extending out from the Si-MEMS device.Scale bar,200μ

m

Fig.4(a )Phase contrast micro-graph of the collagen film extending out from the Si-MEMS device.Scale bar,100μm.(b )Immuno-fluorescence micrograph of the device at 5days after induction of differentiation.Scale bar,1mm.(c )Magnified image of the boxed are in (b ).(Red)Actin filaments,(green )sarcomeric-α-actinin,and (blue )nuclei.Scale bar,100μm

expected to increase the active tension produced by C2C12myotubes (Fujita et al.2007).When the edge of the collagen film was observed under a phase contrast microscope 3days after initiation of electric pulse stimulation,contraction of myotubes and movement of the collagen sheet at 1Hz were observed (Supple.movie 1).When the temperature was decreased to below 32°C to dissolve PNIPAAm,and the linkage between the device and the base (Fig.1(a),arrow-heads)was broken with a micro-manipulator,the device was separated from the base and released (Fig.5(a)).This indicates that the protection of the device with PNIPAAm was successful and attachment of the Si-MEMS device to the base did not occur due to excess glue when the collagen film was glued to the Si-MEMS device (Fig.2(a)).

When the edge of the device was observed,the device did not show any movement when it was under 1Hz twitch stimulation (Fig.5(b),thin blue dashed line).On the other hand,when the device was stimulated by tetanus stimula-tion of 20Hz,displacement of the edge of the device was observed (Fig.5(b),thick red line;Supple.movie 2).This finding showed that active tension generated by the collagen film-myotube complex on twitch stimulation is not strong enough to overcome the resistance of the spring and thereby produce observable movement of the device.Tetanus

stimulation resulted in movement of the device indicating that the tetanus force is sufficient to retract the main spring,which is in good agreement with the previous finding that tetanus stimulation results in greater active tension even with C2C12myotubes (Shimizu et al.2010b ).We have performed experiments with Si-MEMS device with different spring constants,but only those having smallest spring constant showed the movement.Although the device was designed to perform locomotive movement on a linear track,it actually showed torsional movement (Supple.movie 2),indicating that the force generated by the collagen film-myotube complex was askew.This is probably due to the non-parallel orientation of the myotubes as to the collagen film (Fig.4).Because it has also been reported that aligning the myotubes is important in the power output of the actuator (Fujita et al.2010),improving the alignment of myotubes may improve the performance of our bio-MEMS device.

4Discussion and conclusion

The development of MEMS devices and μTAS revealed the need for a sub-millimeter sized motor/actuator that can be integrated into such a device.Thermal engines are assembled from numerous parts,the fuel supply system,combustion chamber,and gas intake/exhaust systems having to be combined,which makes it difficult to fabricate small engines.An electrostatic actuator has a simple structure and thus can be small,but the energy efficiency,power output,and working distance are low.Although other forms of actuators such as polymer-based ones have been investigated,we propose the use of a skeletal muscle cell as a sub-millimeter sized linear actuator.Myosin molecules provide power in muscle cells,but the use of the myosin molecules themselves is difficult due to the need of molecular alignment,difficulty in switching,and fragility of the molecules.The use of muscle cells solves some of the problems mentioned above,because numerous myosin motors can be integrated into muscle cells,the molecules being aligned so as to increase the power efficiency,an integrated switching system enabling electri-cal control of the actuator,and a regeneration system for the damaged molecules.In addition,unlike molecular motors that utilize ATP as an energy source,muscle cells utilize glucose,a stable and inexpensive chemical reachable even in an in vivo environment.A MEMS device that has an integrated actuator that utilizes glucose as an energy source may be operated inside a blood vessel and could be a model for the design of the future intelligent drug delivery systems.Such a device will enable to transport and release the drug at the site where the drug is needed.In C2C12cells,small population of cells with mitotic activity remains even after differentiation.It is important that cells with

mitotic

Fig.5(a )Phase contrast image of the device released from the base at 5days after induction of differentiation.Scale bar,400μm.(b )Movement of the edge of the Si-MEMS device during electric pulse stimulation.(Red )Tetanus stimulation and (blue )twitch stimulation at 1Hz.The time point at which stimulation was applied is shown above the graph

activity are eliminated when in vivo use is considered.This could be achieved by using reagents to kill cells with mitotic activity such as arabinofuranosylcytosine.

In this study,we did not attempt to place the integrated device on a track to induce locomotive movement because we observed several shortfalls with the design of the device;1) the movement was not linear,more bending being observed, and2)shortening was not observed on twitch stimulation and the distance was small even on tetanus contraction,only reaching~8μm.The bending is probably due to the design of the device,i.e.,the collagen film was placed on the device and myotubes were cultured on the collagen film,so that the myotubes were positioned~20μm above the device because the thickness of the collagen film was20μm.The use of thinner collagen film and seeding of the cells on the Si-MEMS device side of the film may abolish the bending movement. The small shortening distance could be due to the spring constant present in the collagen film,which we did not take into account when the Si-MEMS device was designed.Thus, by using thinner collagen film and decreasing the main spring constant,a greater shortening distance could be obtained.We have tried to use thinner UV-crosslinked collagen film but because of the fragility of the film,it was not possible to place the collagen film on the Si-MEMS device.In addition,UV-crosslinked collagen film thinner than20μm was not transparent,making it difficult to observe cells during cultivation.We have also tried on-site fabrication of UV-crosslinked collagen film,but because fabricated collagen film did not attach to the Si-MEMS device strong enough,the film was separated from the Si-MEMS device during cell cultivation(data not shown).Thus,it is important to have an architecture which enhances collagen film attachment to the Si-MEMS device in the design introduced in the future.

Here,we have demonstrated the possibility of using the collagen film-myotube complex as an actuator for MEMS devices.We have designed and fabricated a device composed of a Si-MEMS comprising springs and ratchets, UV-crosslinked collagen film and C2C12myotubes.Unlike cardiomyocytes,because C2C12cells can proliferate infinitely before differentiation,the use of the C2C12cell line enabled the fabrication of a bio-actuator without animal sacrifice.The device with an integrated bio-actuator was activated by electric pulse stimulation,allowing external control of the device.The collagen film-myotube complex could be usable as an actuator in MEMS devices where switching is necessary.Our study is to present a concept of skeletal muscle driven micro-machine for locomotion in the wet but glucose sufficient condition such as in blood vessel. Study presented herein highlighted the factors necessary for the locomotion of the device such as less spring constant, more collagen film area to harbor more myotubes,pattern to align myotubes,and architecture for holding collagen film to enable on-site collagen film production.Acknowledgements The authors would like to thank Yuki Morioka for the excellent technical assistance.

References

C.M.Alves,Y.Yang,

D.Marton,D.L.Carnes,J.L.Ong,V.L.Sylvia,

D.D.Dean,R.L.Reis,C.M.Agrawal,J Biomed Mater Res B

Appl Biomater87(1),59–66(2008)

R.G.Dennis,P.E.Kosnik2nd,M.E.Gilbert,J.A.Faulkner,Am J Physiol Cell Physiol280(2),C288–295(2001)

A.Desai,S.W.Lee,Y.C.Tai,Sens Actuators,A,Phys73(1–2),37–44

(1999)

H.Fujita,T.Nedachi,M.Kanzaki,Exp Cell Res313(9),1853–1865

(2007)

H.Fujita,K.Shimizu,E.Nagamori,Biotechnol Bioeng103(2),370–

377(2009a)

H.Fujita,K.Shimizu,E.Nagamori,Biotechnol Bioeng103(5),1034–

1041(2009b)

H.Fujita,K.Shimizu,E.Nagamori,Biotechnol Bioeng106(3),482–

489(2010)

M.G.Garcia-Webb,A.J.Taberner,N.C.Hogan,I.W.Hunter,Am J Physiol Heart Circ Physiol293(1),H866–874(2007)

Y.Hiratsuka,T.Tada,K.Oiwa,T.Kanayama,T.Q.Uyeda,Biophys J 81(3),1555–1561(2001)

N.F.Huang,S.Patel,R.G.Thakar,J.Wu,B.S.Hsiao,B.Chu,R.J.

Lee,S.Li,Nano Lett6(3),537–542(2006)

G.Iribe,M.Helmes,P.Kohl,Am J Physiol Heart Circ Physiol292(3),

H1487–1497(2007)

R.S.Kane,S.Takayama,E.Ostuni,D.E.Ingber,G.M.Whitesides, Biomaterials20(23–24),2363–2376(1999)

https://www.sodocs.net/doc/108164014.html,m,S.Sim,X.Zhu,S.Takayama,Biomaterials27(24),4340–4347(2006)

L.Limberis,R.J,Stewart.Nanotechnology(2),47(2000)

D.K.McMahon,P.A.Anderson,R.Nassar,J.B.Bunting,Z.Saba,A.

E.Oakeley,N.N.Malouf,Am J Physiol266(6Pt1),C1795–

1802(1994)

D.Montarras,C.Pinset,J.Chelly,A.Kahn,F.Gros,EMBO J8(8),

2203–2207(1989)

C.Ozcan,N.Hasirci,J Biomater Sci Polym Ed18(6),759–773(2007) P.H.Pham,

D.V.Dao,S.Amaya,R.Kitada,S.Sugiyama,J Micromech Microeng16(12),2532–2538(2006)

K.Shen,J.Qi,L.C.Kam,J.Vis.Exp.(22),(2008)

N.Shepherd,M.V ornanen,G.Isenberg,Am J Physiol258(2Pt2), H452–459(1990)

K.Shimizu,H.Fujita,E.Nagamori,Biotechnol Bioeng106(2),303–310(2010a)

K.Shimizu,H.Sasaki,H.Hida,H.Fujita,K.Obinata,M.Shikida,E.

Nagamori,Biomed Microdevices12(2),247–252(2010b)

R.K.Soong,G.D.Bachand,H.P.Neves, A.G.Olkhovets,H.G.

Craighead,C.D.Montemagno,Science290(5496),1555–1558 (2000)

H.Suzuki,A.Yamada,K.Oiwa,H.Nakayama,S.Mashiko,Biophys

J72(5),1997–2001(1997)

T.Takezawa,Y.Mori,K.Yoshizato,Biotechnology(N Y)8(9),854–856(1990)

R.B.Vernon,M.D.Gooden,https://www.sodocs.net/doc/108164014.html,ra,T.N.Wight,Biomaterials26

(16),3131–3140(2005)

J.Xi,J.J.Schmidt,C.D.Montemagno,Nat Mater4(2),180–184 (2005)

D.Yaffe,O.Saxel,Nature270(5639),725–727(1977)

S.I.Yasuda,S.Sugiura,N.Kobayakawa,H.Fujita,H.Yamashita,K.

Katoh,Y.Saeki,H.Kaneko,Y.Suda,R.Nagai,H.Sugi,Am J Physiol Heart Circ Physiol281(3),H1442–1446(2001)