Griffith Research Online

https://research-repository.griffith.edu.au

Multiarray cell stretching platformfor high magnification real-timeimaging

AuthorHuang, Yuli, Nguyen, Nam-Trung, Lok, Khoi Seng, Foo Lee, Peter Peng, Su, Maohan, Wu, Min,Kocgozlu, Leyla, Ladoux, Benoit

Published2013

Journal TitleNanomedicine

DOI https://doi.org/10.2217/nnm.13.45

Copyright StatementCopyright 2013 Future Medicine Ltd. This is the author-manuscript version of this paper. Reproduced inaccordance with the copyright policy of the publisher. Please refer to the journal website for access tothe definitive, published version.

Downloaded fromhttp://hdl.handle.net/10072/52570

1

Multi-array cell stretching platform for high magnification

real-time imaging

Aims: This paper reports the development of a multi-array microchip for the application of mechanical strains onto

mono-layered cell cultures, with an emphasis on its real-time imaging capability. Materials & Methods: The

platform adopted a side-stretching mechanism. The cells were kept in the microscope focal plane throughout the

stretching process, allowing real-time microscopic imaging of cells in high magnifications (60×). A reconfigurable

array of 15 units was integrated into a single chip. Each unit was an assembly of three elastomeric layers and a

glass coverslip, with imprinted micro-pattern of 100~200 μm depth. An 8 μm thick membrane, in which the cells

were cultured, separated the top and bottom channels. A programmable pneumatic control system was developed

to actuate this platform. The deformation of this membrane was characterized in-situ using the particle tracking

technique. Multiple stretching experiments were conducted in this system with various cell-lines to verify its

desired characteristics. Results: The uniform strain has a maximum deformation of 69%. The observed acute and

long term cell morphological changes confirmed the stretching capability of this platform. The supreme imaging

capability has also been verified separately with the real-time imaging of transfected COS-7 stretching, and post-

stretching imaging of immunofluorescence stained PTK2. Conclusion: The platform reported here is a powerful

tool for studying mechanically-induced physiological changes in cells.

KEYWORDS: Mechanobiology, cyclic cell stretching, high magnification, pneumatic actuation,

BioMEMS, polymeric micromachining, microfluidics.

2

Introduction

Cellular mechanotransduction is a key process that regulates various cell functions such cell adhesion [1],

tissue formation [2], morphogenesis [3] and cell differentiation [4]. Biochemical networks and gradients seem to

be insufficient to fully explain cellular processes from single cell movements up to tissue patterning. Cell

interactions and shape changes involve mechanical properties of the environment such as its elasticity or the

application of external forces. As such, there is a number of emerging links between gene regulatory networks,

biochemical gradients, and physical forces, involving multiple feedback loops [5]. Such physical stimuli are

usually recreated in vitro. The development of various techniques to analyze the transmission of forces through

living cells has improved our knowledge on mechanotransduction at different scales from molecules to cells to

tissues. According to the different methods used to mechanically stimulate cells, the responses of cells are

heterogeneous in space and time. The techniques used to apply external forces or stresses on cells in a range of

stresses ranging from pN/μm2 up to tens of μN/μm2 are optical tweezers [6], magnetic tweezers [7], atomic force

microscopes (AFM) [8], pipette aspiration [9] and devices that probe cell mechanics over the whole cell size such

as fluid shear stress [10, 11] or cell stretching device [12, 13].

The influence of tensile strains has been extensively studied due to its simplicity and ubiquitous presences

in vivo. Both static and cyclic stimulations have been performed to analyze the cellular response to external forces.

For example, The study on smooth muscle cells and human umbilical vein endothelial cells suggested that such

stimulus led to the migration, better adherence and reorganization of stress fibers of these cells [14]. Similar

regulatory effects have also been found from other tests carried on the cardiovascular cells [15], as well as

bone/cartilage cells [16, 17]. Experimental observations showed that cells respond differently to static and

dynamically varying strains. For static or quasi-static strain, cells align parallel to the direction of applied strain [12,

18] whereas for cyclic strain, cells align away from the direction of the applied stretch [19, 20]. These alignments

also depend on whether the strain was applied in 2D or 3D and on cell types.

Another creative use of cyclic strain involves its combination with several other essential conditions to

recreate a holistic model of an organ in vitro, namely organ-on-chip. Huh et al. built a miniaturized lung-on-chip

platform, which consists of double cell layers co-cultured on two sides of a porous elastomeric membrane. The

alveolar cells adsorption of silica nanoparticles, as well as its resultant inflammatory responses was observed [21].

Douville et al. developed a similar microfluidic alveolar model, where surface tension caused by the air/liquid

interface is added on top of cyclic strain. It showed that a combination of fluid and solid mechanical stresses

contribute to cell death and detachment of alveolar epithelial cells [22]. Kim et al. used cyclic strain for mimicking

the peristalsis motions in gut tissue, where human intestinal epithelial (Caco-2) cells were co-cultured with

3

microbial flora above [23]. Such devices can be considered as an extension of a cell culture model to provide an

alternative to animal and clinical studies for drug screening and toxicology applications.

In the field of tissue engineering, the application of cyclic strain is essential for biomimetic stimulation of

certain stem cells [24]. For example, numerous studies have demonstrated that cyclic mechanical strain promoted

the cardiomyogenesis of stem cell-derived cardiomyocytes, which was critical for blood vessel tissue reparation

[25-28]. Likewise, the cyclic strain was also found helpful for the regeneration of other tissues including skin [29]

and skeletal muscles [30]

Various mechanisms have been used for applying the strain on adherent cells such as the flexible bottom

membrane distension [31], polyethylene terephthalate (PET) culture plate bending [32], 3D-hydrogel matrix

stretching [16] and etc. As macro-scale devices usually consume a large amount of sample and reagents. The large

footprint makes these devices incompatible with microscope for high-throughput real-time imaging. Thus, macro-

scale devices are rapidly phased out by micro-scale counterparts. In this new development, miniaturized pneumatic

balloons [33, 34] or membrane supported by static/movable posts [14, 35, 36] were two commonly adopted

stretching mechanisms. However, these devices limit the cell imaging capability by either moving the cells out of

focal plane or keeping them too far away from the lens due to its geometric constraints.

The cellular response to mechanical stimulations in real-time and over a long time period under high

optical magnification can lead to better understanding of cell mechanobiology at various scales from molecular

processes to tissue reshaping. Therefore, new devices are needed to provide high-resolution in-focal imaging

during the cell stretching process, which allows detailed analysis of mechanotransduction and other strain induced

cell behaviors.

Materials & Method

Device Fabrication

Each stretcher chip is an assembly of 4 layers, Figure 1. From the top to the bottom, the first three layers

are individually fabricated in polydimethylsiloxane (PDMS), while the last layer is a glass coverslip (Fisher

Scientific, #1 24×55 mm). Layers 1 and 3 were made by PDMS replica molding. In this process, a master-mold

was made from the standard photo-lithography on SU-8. To allow easy release, the master-mold is pre-coated with

Octafluorocyclobutane (C4F8) using a deep reactive-ion etching system (STS, 1 minute). To make layer 1, PDMS

was casted on the master-mold and cured at room temperature for 48 hours. Access-holes were subsequently

punched on layer 1 (Harris Uni-Core puncher, 0.75 mm). To make layer 3, a glass-slide was first pre-coated with

C4F8 in the aforementioned way. The coverslip was then attached to the coated side of the glass slide on its two

long edges using Kepton tape (3M), which provides structural support throughout the fabrication process. PDMS

4

was then poured on the master-mold, and the coverslip was slowly pressed down against the mold from one side to

allow air bubbles to be squeezed out. A 40 Pa static pressure was loaded on the coverslip, and the structure was

cured at 80ºC for 1 hour as a whole.

Layer 2 was a membrane made by spin-coating PDMS on a silicon\acrylic wafer at 4000 rpm for 3 minutes,

which gives a final thickness of 8±1 μm, as measured from microscopic cross-sectional view cut out from 5

different samples. This layer was bonded with the layer 3 by oxygen-plasma surface treatment (NT-2, BSET EQ,

120W 30s), and was thermally cured at 50ºC for 10 min. Upon completion, the portions of membrane that block

the access-holes were manually removed using tweezers with ultra-fine tip, (Cole-parmer), first to make the bottom

half cell channel accessible, and second to join the top and bottom halves of vacuum channels. In order to join the

vacuum micro-channels, precise alignment was necessary. The stack of layer 2~4 was plasma-treated concurrently

with layer 1 in the same way as mentioned above. The two pieces were then brought under a microscope of 4×

magnification for alignment and bonding. During this process, an in-house made alignment device was used for

increasing the reliability and speed.

From our numerical simulation (not presented here), the maximum attainable strain is highly dependent on

the channel height. However, increasing channel height is constrained by the need for a high optical magnification.

To address these different requirements, two configurations were separately made. Configurations A and B with

200 μm and 100 μm half channel height was used for high strain and high resolution applications, respectively.

High throughput parallelization

An array of 15 stretching units was integrated on one chip with the same size of a glass coverslip to

increase the experimental throughput (Figure 2). Every three identical units in a column share one common

vacuum port. The 5 different columns were actuated individually with different magnitudes and frequencies within

the same time-span.

An in-house built pneumatic control system applies programmable stretching magnitude and frequency to

the respective columns (Figure 3). The system consists of a vacuum pump (Dryfast 2034, Welch), a vacuum

controller (LabAid 1640, Welch), an 8×3 solenoid valve array (S10MM-31-12-3, pneumadyne), a current

amplifying circuit and a control program (Labview). The constant vacuum source output from vacuum controller

was supplied to each column through the valve array. As the motion of valves can be specified in the control

program, individual stretching frequencies can be supplied to each column. The upper limit of the frequency

is100 Hz due to the response time of the solenoid. However, as the system has a single vacuum source, the vacuum

strength is shared among all columns. In order to provide different stretching magnitude to each column, the cell

channel widths were adjusted according to the relationship: , where is the maximum strain, is the

5

maximum displacement at a given vacuum strength is the channel width. The example in Figure 2 demonstrates

a chip designed for simultaneous application of 5 different linearly incremental strains.

Strain Characterization

Quantitative analysis of the induced strain on the PDMS membrane is conducted in stretching units with

configuration A. During the fabrication process of the membrane, fluorescent micro-beads (Thermo Scientific,

1.0 μm, EX/EM 542/612 nm, diluted to 0.25% solids) were transferred onto the membrane via oxygen-plasma

assisted contact printing. Sequential recording were taken at discrete vacuum levels from the atmosphere pressure

to the minimum pressure reachable of the pump, with a constant increment of 88 kPa. The displacement of each

particle at the respective vacuum level, defined as , was tracked using ImageJ (NIH). As the

membrane chamber is shaped as a narrow rectangular stripe with a length to width ratio over 25, plain strain is

assumed in the vertical cross-sectional plane, and thus the corresponding strain in the width direction was derived

as .

Cell culturing protocol

The upper cell channel was cleaned and coated with collagen-gel as reported by Huh et al. [21]. After

two hours of coating, the excessive gel in the channel was gently flushed away by cell growth medium. The cells

were trypsinized (TrypLE™ Express, Life technologies) and resuspended to the concentration of 5~10 million/ml.

10 µl sample is enough for seeding of one unit. A pipette tip was loaded with 10 µl cells and inserted into the

channel inlet, while another empty pipette tip was inserted in the channel outlet to gently redraw 5 µl to let cells

flow into the channel. The stretcher was subsequently placed into an incubator (5% CO2, 37˚C) for 1 to 3 hours

until cells form attachment with the PDMS membrane, and then fresh medium is added into one pipette tip. Driven

by gravity, the medium slowly flows across the channel until the liquid levels of both sides are balanced. The chip

was further incubated overnight. When a confluent cell monolayer was formed on the membrane, the chip was

transferred onto the microscope stage housed inside a mini incubator, and connected to the pneumatic control

system for stretching experiments.

Cell-membrane strain dependence

A similar approach as in the PDMS membrane strain characterization was adopted for the analysis of the

relationship between strain on PDMS membrane and on cells. Devices with configuration B was used for better

visualization of cell features. The strain for cell was measured from the displacement of cell prominent features,

such as the focal adhesion, nucleus or boundary of the cells. Two cell-lines, monkey fibroblast (COS-7) and benign

human foreskin fibroblast (Fibro), were used separately in the tests. The test for each cell-line was triplicated, i.e.

each cell type was cultured in 3 units of a column, which were stretched under the identical condition. 5 readings

6

of either PDMS membrane or cell deformation were collected at each vacuum level. In total 15 readings for each

condition were averaged and their standard deviations were calculated for analysis.

In the experiment, both cells were cultured at a lower concentration of approximately 1 million/ml, such

that each cell has a larger spreading area and the cell boundaries can be easily distinguished. Differential

interference contrast (DIC) with 40× magnification was used for recording. As the cells were cultured in a lower

density, individual cells of either type were able to spread to a length between 20 to 200 µm along the stretching

direction. The features on larger cells, with longitudinal spreading more than 50 µm, were selected for

measurement to ensure the measurement accuracy.

Imaging capability

The real-time imaging capability of the chip was verified by stretching COS-7 cells and Rat Basophil

Leukemia (RBL) cells. In this experiment, COS-7 is transfected with lifeact-ruby, which directs the expression of

Red Fluorescent Protein (RFP) on the cell actin-filaments; whereas RBL is transfected with clathrin-Green

Fluorescent Protein (GFP). After culturing the cells on the device, a static strain of 8% was applied. Both the

relaxed and stretched positions were recorded by 2D confocal imaging with the magnification of 60× (UPLSAPO

60XW, numerical aperture (NA) of 1.2, working distance (WD) of 280 μm). The relax and stretch state of cells

were labeled with green and red pseudo-color respectively, and these two images were superimposed for better

comparison.

The imaging capability for post-analysis was verified on immunofluorescence stained kidney epithelial

cells from long-nosed potoroo (PTK2). In this process, the cells were fixed by formaldehyde (4%, 10 min), and

then permeablized by cold acetone (-20˚C, 5 min). Subsequently, the cytoskeleton components including the actin-

filament and the micro-tubule are co-stained by Alexa Fluor568 Phalloidin (5 U/ml, Gibco) and Mouse anti-α-

Tubulin-Alexa 488 (5 μg/ml, Gibco) for 20 minutes, followed by staining the nucleus with 4',6-diamidino-2-

phenylindole (DAPI) (1 μg/ml, Thermo Scientific) for another 5 minutes. As stretching is no longer needed, the

bottom coverslip was detached to reduce the distance between the cells and the objective lens. Consequently cells

can be imaged with a high magnification of 100× (UPLSAPO 100XO, N.A.: 1.4, W.D.: 130 μm).

Long term stretching effects

The tendency for cells to align perpendicular the direction of strain was observed in many different cell

types. Similar experiments were carried out on PTK2 cells to verify the capability of this device to induce long

term effects to cells. Two different cyclic strains with identical frequency (5% and 10%, 0.1 Hz) were applied to

the cells. The tests for both stretching regimens and the non-stretched control were triplicated. Cells were imaged

with 40× DIC via multi-stage acquisition. Time-lapse recordings were taken for more than 4 consecutive hours at

7

an interval of 5 minutes. Quantitative analysis on cell orientation over the time showed the transitional pattern of

cell alignment. To verify that the cell reorientation did occur due to cytoskeleton reorganization instead of merely

selective cell detachment along the unfavorable directions, post-stretching analysis was also carried out on the

same sample used in the above time-lapse recording. Fixation and immunofluorescence staining was immediately

applied to the set of cells stretched under 5% strain and the set serving as control. Confocal microscopy with

different magnifications was used for the visual comparison.

Result & Discussion

Strain Characterization

Figure 4 shows the distribution of displacement and strain on the membrane, as well as their relationship

with the applying vacuum. At all vacuum pressure levels, linear fittings of displacement yield a coefficient of

determination ( ) greater than 0.97. The accuracy of linear fittings confirms the linear proportionality between

the displacement and location, as also predicted in the simulation (data not shown). On the other hand, the line

fittings of strain versus location data showed a small gradient of less than ±0.6×10-4 µm-1. Both results are

evidences of the strain uniformity on the membrane, ensuring the identical experimental condition for all cells

loaded into one unit. Additionally, all fitted lines converge at one point, and all of their y-intercepts are within the

range of ±1.5 µm. This convergence suggests that the membrane is stretched along the geometric centerline of the

rectangular membrane. In other words, stretching is perfectly plane-symmetric. Comparing the y-intercepts of all

strain data, a constant increment is observed with the increasing vacuum strength, with a ratio approximately of 10%

per 88 kPa. This trend is also visually observed in the kymograph in Figure 5, as the vertical locations of each bead

form straight lines as the pressure increases. That means, the strain increases linearly with increasing vacuum

strength, which is another desirable feature in terms of device control. As the input of the device (vacuum strength)

is linearly proportional with the output of the device (strain), the control of the device is straightforward.

Furthermore, when the maximum vacuum strength is applied (98.7 KPa), the strain of a uniaxial stretcher reached

69%, which outperforms many existing commercial stretching devices, either in macro or micro scale [35-37].

In the study of cell-membrane strain dependence, the strain of both cell-lines has shown a clear correlation

with the strain of the membrane (Figure 6). From the graph, the difference between the two is well within the error

range. Plotting both strains against the vacuum strength shows a clear linear relationship. The small standard

deviation (<1%) also suggests that at each given vacuum level, the strain is highly uniform and reproducible

among different stretching units. However, as part of the compromise for greater image resolution, the maximum

strain was reduced from 69% in configuration A to 15% in configuration B.

Imaging capability

8

Real-time imaging of intra-cellular activities is a highly popular approach adopted for the study of

mechanobiology. Two essential requirements for this kind of imaging are (i) sample to be imaged with optical

magnification greater than 60× and (ii) sample to stay in the microscopic focal plane. The first requirement ensures

that the cellular components of interest are clearly identifiable. In this case, both material of the chip provides an

excellent optical property (refractive index: glass 1.47~1.92, PDMS 1.403. UV transparence: Glass > 300 nm,

PDMS > 240 nm). Furthermore, the small WD meets the requirement of high NA objective lens. The second

requirement is essential for imaging intra-cellular activities. As the phenomenon of interest usually occurs in a

short time-span, there is little time allowance for the refocus process. Moreover, due to the small depth of field in

high magnifications, the inaccuracy of refocus might easily present misleading information. For instance, some

components disappear while their surrounding seems unchanged, or the fluorescence intensity is reduced for the

same feature. Keeping the cells in the focal plane helps to solve these problems, and provides non-interrupted and

reliable observations over time. The second requirement is fulfilled by the intrinsic characteristics of side

stretching mechanism, which in theory does not generate any vertical motion on the membrane. However, in many

practical situations, minor vertical deviations might occur due to the thinning effect of the membrane. In this case,

the thickness of half of the membrane is kept as low as 4 µm, hence its thinning at 15% strain (Poisson’s ratio of

0.5, uniaxial strain) is only around 0.5 µm, which is smaller than the depth of field of the 100× lens. As shown in

Figure 7, each actin-filament is clearly identifiable, suggesting that the required subcellular resolution is attainable

in our device. In both relaxed and the stretched states, the same components remained visible, and no significant

change in intensity was noticed. However, there are also cases where a certain section of cell components move

out of focal plane. In these cases, the focus of the same component section can be restored within the focal

adjustment of ±3 μm range, which suggests that the practical degree of out-of-focus problem is kept within ±3 μm

range. In the stretching of RBL, clathrin is a cytosolic protein that usually assembles on the plasma membrane and

display a punctuate appearance, was chosen as another marker to demonstrate the both extent of stretching and the

time resolution. The lifetime of a single puncta of clathrin is usually 1 minute or longer. The same sets of clathrin

puncta before and during stretching (images taken at 5 sec apart of each other) can be observed and displayed up to

15% strain along the stretching direction. The above results are solid proofs of the in-focus stretching capability.

As the bottom coverslip is detached in post-analysis, the cell sample can be brought even closer to the

microscopic lens to meet the highest demand on WD. A 100× oil immersion objective was used in this application

to image the two cell components. As shown in Figure 8, the image is sharper and the component structures are

more visible due to three possible reasons: (i) the externally applied immunofluorescence intensity is usually

stronger than the auto-fluorescence, (ii) the removal of cytosol in the fixed cells helps to reduce the background

noise and (iii) increase in magnification helps to increase the resolution. However, the channel of actin filaments is

9

omitted here due to the presence of micro-beads of the save wave frequency on the PDMS membrane, which

introduces confusion to image.

Long term stretching effects

As directly visualized from the time-lapse recording (Figure 9), continuous stretching induced a gradual

change in the cell orientation in both stretched samples, while no significant change in cell shape and orientation

was observed from the non-stretched control samples. As the shortening of cell in the longitudinal direction is

always accompanied by the elongation of cell in the latitudinal direction as well as a certain extend of cell

migration, it is obvious that this orientation is not due to selective detachment of cells. Comparing the two

stretched samples, it is also apparent that the rate of cell orientation is proportional to the magnitude of strain,

which is consistent with previous studies [30, 38, 39].

The merged fluorescence channel of 488 nm and 405 nm showing micro tubule and nucleus respectively

was selected for the analysis, Figure 10. By comparing the stretched channel with the non-stretched control at both

low magnification (10×) and high magnification (40×), it is clear that both the orientation of cells and the

orientation of cytoskeleton has been significantly influenced by the cyclic strain. This experiment again confirmed

that mechanical strain was a direct clue to the cytoskeleton reorganization.

Conclusions

The device reported in this paper consists of a reconfigurable array of 3×5 micro-scale cell stretching units.

The overall footprint of the chip is 24 mm ×55 mm, which is compatible with the majority of microscope stage-

holders. Hence, high experimental throughput is made possible in this device by applying multi-stage imaging to

all stretching units in time-lapse recording. An automated pneumatic control system was also constructed, capable

of simultaneously providing 24 lines of actuation with individually adjustable frequencies and adjustable

magnitudes as a whole. Characterization of the device was conducted by tracking the membrane deformation with

florescent micro-beads. The results verified the strain uniformity and input/output linear proportionality. The

analysis of cell-membrane strain dependence confirmed the bonding quality between them. Moreover, the innate

in-focus stretching capability and its ultra-thin bottom thickness enabled the chip to be used for imaging

subcellular phenomena as verified by real-time imaging of transfected COS-7 cells. The strain-induced long term

effects to cells were also studied. Cell alignment was proven with PTK2 cells based on the visual examination of

the stretching process and post-stretching micro-tubule distribution under high magnification.

Financial & competing interest disclosure

10

Funding for this work was received from the Mechanobiological Institute and Singapore National Research

Foundation. The authors have no other relevant affiliations or financial involvement with any organization or

entity with a financial interest in or financial conflict with the subject matter or materials discussed in the

manuscript apart from those disclosed. No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research

The authors state that they have obtained appropriate institutional review board approval or have followed

the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In

addition, for investigations involving human subjects, informed consent has been obtained from the participants

involved.

Future perspective

Driven by the vast demand in the field of mechanobiology, it won’t be long before novel stretching

mechanisms emerge, which would be completely free from the geometric constraints for the gap between the lens

and the sample. Such mechanisms may inherit the characteristics of side-stretching, while the actuation is replaced

by more precise and compact method such as piezoelectricity, thermal expansion or electromagnet. Such

breakthrough would release the full potential of optical microscopy for the analysis of intra-cellular activities.

Executive summary

In-focus real-time imaging

The side stretching mechanism provides innate in-focus stretching capability, which permits continuous

imaging of a particular subcellular phenomenon during the stretching process.

The in-focus capability was further enhanced by adopting the ultra-thin membrane eliminating the problem

of membrane thinning.

High resolution imaging

The gap between the sample and the objective lens was reduced by a novel PDMS-coverslip assembly.

Both 2D visualization and 3D z-stacking are successfully carried out using 60× confocal microscopy.

The detachable coverslip bottom design enables successful post-stretching confocal imaging of 100×.

Programmable cell-stretching array

11

Parallelization of 3×5 stretching units reduces overall time consumption by 4/5 in long term stretching

experiments that require variation in condition, providing at the same time quantitatively meaningful triplication to

each condition.

Pneumatic control system provides precise cyclic actuation to each column of 3 units, with the frequency

individually adjustable and the magnitude individually reconfigurable.

12

Figure 1 Configuration for high optical resolution with cover-slip bottom to meet the WD requirement

13

Figure 2 (a) The cross section of a stretching unit at relaxed and stretched state. (b) The assembled multi-array stretching

device with 5×3 units. The bottom is a glass cover-slips used for high resolution imaging.

14

Figure 3 Schematic view of the pneumatic control system. The actuation power was supplied by a vacuum system, which

was marked as black lines in the sketch. The electric signals for vacuum strength and frequency control were marked as

green lines.

Liquid Trap

Vacuum Pump

Pressure Setpoint

+-

PID

Algorithm

Solenoid Valve

Stretching Unit

Solenoid Valve

Stretching Unit

Max 24 Units In Parallel

...

Signal

Breakdown

Current

Amplification

Solenoid Valve

Stretching Unit

Labview

Control

Program

15

Figure 4 Relationship between (a) displacement and (b)strain and the location in a representative strain unit. Only data for

three vacuum levels are displayed. Coefficient of determination is shown in (a) to verify the quality of linear fitting. The

fitting equations are shown in (b) to demonstrate the uniformity of strain and its linear increment with the vacuum strength.

16

Figure 5 (a) The membrane with the corresponding regions of interest at the relaxed state. (b-c) Kymograph of stretched

particles in two region of interests A and B at different applied pressure. Scale bar: 100 μm.

17

Figure 6 The strain provided by the membrane and strain experienced by the cells for (a) COS-7 and (b) Fibro. Scale bar:

10 μm.

(b)

18

Figure 7 Images from 60× 2D confocal microscopic view of (a) the actin-filament of a single COS-7 cell (b) the clathrin of

two neighboring RBL cells.The recordings at the relax and the stretch state of cells are colored with green and red

respectively, while both images are merged to show the comparison of cell morphology during stretching. Scale bar: 10 μm.

.

Figure 8 100× confocal immunofluorescence microscopic view of (a) none-stretched PTK2 control cells (b) PTK2 cell

stretched with 5% cyclic strain through 3 colour channels (from left to right), 1. The 488 nm channel showing the micro

tubules stained with Mouse anti-α-Tubulin-Alexa 488; 2. The 405nm channel showing the nucleus stained with 4',6-

19

diamidino-2-phenylindole (DAPI); 3. Image of merged multiple channels showing the overall appearance of cell components.

Scale bar: 10 μm.

Figure 9 Time-lapse recording of PTK2 cell morphologic changes, imaged by 40× DIC. Scale bar: 50 μm.

Figure 10 10× confocal images of PTK2 cells with stained micro-tubule(green) and nucleus(blue): (a) Control sample

without stretching; (b) 5% cyclic strain loaded sample; (c) 40× zoom-in view of micro-tubule for the strain loaded sample.

Scale bar: 50 μm.

20

References

1. Discher DE, Janmey P, Wang YL: Tissue cells feel and respond to the stiffness of their substrate. Science 310(5751), 1139-1143 (2005).

2. Ingber DE: Mechanobiology and diseases of mechanotransduction. Annals of Medicine 35(8), 564-577 (2003).

3. Desprat N, Supatto W, Pouille PA, Beaurepaire E, Farge E: Tissue Deformation Modulates Twist Expression to Determine Anterior Midgut Differentiation in Drosophila Embryos. Developmental Cell 15(3), 470-477 (2008).

4. Engler AJ, Sen S, Sweeney HL, Discher DE: Matrix Elasticity Directs Stem Cell Lineage Specification. Cell 126(4), 677-689 (2006).

5. Chiquet M, Gelman L, Lutz R, Maier S: From mechanotransduction to extracellular matrix gene expression in fibroblasts. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 1793(5), 911-920 (2009).

6. Choquet D, Felsenfeld DP, Sheetz MP: Extracellular matrix rigidity causes strengthening of integrin- cytoskeleton linkages. Cell 88(1), 39-48 (1997).

7. Hu XC, Wang Y, Shi DR, Loo TY, Chow LWC: Immunomagnetic tumor cell enrichment is promising in detecting circulating breast cancer cells. Oncology 64(2), 160-165 (2003).

8. Chaudhuri O, Parekh SH, Lam WA, Fletcher DA: Combined atomic force microscopy and side-view optical imaging for mechanical studies of cells. Nature Methods 6(5), 383-387 (2009).

9. Zhou EH, Quek ST, Lim CT: Power-law rheology analysis of cells undergoing micropipette aspiration. Biomechanics and Modeling in Mechanobiology 9(5), 563-572 (2010).

10. Helmke BP, Rosen AB, Davies PF: Mapping mechanical strain of an endogenous cytoskeletal network in living endothelial cells. Biophysical Journal 84(4), 2691-2699 (2003).

11. Su FC, Wu CC, Chien S: Review: Roles of microenvironment and mechanical forces in cell and tissue remodeling. Journal of Medical and Biological Engineering 31(4), 233-244 (2011).

12. Eastwood M, Mudera VC, Mcgrouther DA, Brown RA: Effect of precise mechanical loading on fibroblast populated collagen lattices: Morphological changes. Cell Motility and the Cytoskeleton 40(1), 13-21 (1998).

13. Jungbauer S, Gao H, Spatz JP, Kemkemer R: Two characteristic regimes in frequency-dependent dynamic reorientation of fibroblasts on cyclically stretched substrates. Biophysical Journal 95(7), 3470-3478 (2008).

14. Zheng W, Jiang B, Wang D, Zhang W, Wang Z, Jiang X: A microfluidic flow-stretch chip for investigating blood vessel biomechanics. Lab on a Chip, (2012).

15. He Y, Shroff SG: Cyclic stretch downregulates arterial vascular connexin43 protein expression: An ex vivo study. Biomechanics and Modeling in Mechanobiology 9(2), 203-211 (2010).

16. Vanderploeg EJ, Wilson CG, Levenston ME: Articular chondrocytes derived from distinct tissue zones differentially respond to in vitro oscillatory tensile loading. Osteoarthritis and Cartilage 16(10), 1228-1236 (2008).

17. Honda K, Ohno S, Tanimoto K et al.: The effects of high magnitude cyclic tensile load on cartilage matrix metabolism in cultured chondrocytes. European Journal of Cell Biology 79(9), 601-609 (2000).

18. Collinsworth AM, Torgan CE, Nagda SN, Rajalingam RJ, Kraus WE, Truskey GA: Orientation and length of mammalian skeletal myocytes in response to a unidirectional stretch. Cell and Tissue Research 302(2), 243-251 (2000).

19. Smith PG, Garcia R, Kogerman L: Strain reorganizes focal adhesions and cytoskeleton in cultured airway smooth muscle cells. Experimental Cell Research 232(1), 127-136 (1997).

20. Wang JHC, Grood ES: The strain magnitude and contact guidance determine orientation response of fibroblasts to cyclic substrate strains. Connective Tissue Research 41(1), 29-36 (2000).

21

21. Huh D, Matthews BD, Mammoto A, Montoya-Zavala M, Yuan Hsin H, Ingber DE: Reconstituting organ-level lung functions on a chip. Science 328(5986), 1662-1668 (2010).

22. Douville NJ, Zamankhan P, Tung YC et al.: Combination of fluid and solid mechanical stresses contribute to cell death and detachment in a microfluidic alveolar model. Lab on a Chip - Miniaturisation for Chemistry and Biology 11(4), 609-619 (2011).

23. Kim HJ, Huh D, Hamilton G, Ingber DE: Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow. Lab on a Chip, (2012).

24. Simmons CS, Petzold BC, Pruitt BL: Microsystems for biomimetic stimulation of cardiac cells. Lab on a Chip 12(18), 3235-3248 (2012).

25. Gwak SJ, Bhang SH, Kim IK et al.: The effect of cyclic strain on embryonic stem cell-derived cardiomyocytes. Biomaterials 29(7), 844-856 (2008).

26. Clause KC, Tinney JP, Liu LJ, Keller BB, Tobita K: Engineered early embryonic cardiac tissue increases cardiomyocyte proliferation by cyclic mechanical stretch via p38-MAP kinase phosphorylation. Tissue Engineering - Part A 15(6), 1373-1380 (2009).

27. Park JS, Huang NF, Kurpinski KT, Patel S, Hsu S, Li S: Mechanobiology of mesenchymal stem cells and their use in cardiovascular repair. Frontiers in Bioscience 12(13), 5098-5116 (2007).

28. Gong Z, Niklason LE: Use of human mesenchymal stem cells as alternative source of smooth muscle cells in vessel engineering. Methods in molecular biology (Clifton, N.J.) 698, 279-294 (2011).

29. Kobayashi M, Spector M: In vitro response of the bone marrow-derived mesenchymal stem cells seeded in a type-I collagen-glycosaminoglycan scaffold for skin wound repair under the mechanical loading condition. Molecular & cellular biomechanics : MCB 6(4), 217-227 (2009).

30. Bayati V, Sadeghi Y, Shokrgozar MA et al.: The evaluation of cyclic uniaxial strain on myogenic differentiation of adipose-derived stem cells. Tissue and Cell 43(6), 359-366 (2011).

31. Ueki M, Tanaka N, Tanimoto K et al.: The effect of mechanical loading on the metabolism of growth plate chondrocytes. Annals of Biomedical Engineering 36(5), 793-800 (2008).

32. Mawatari T, Lindsey DP, Harris AHS, Goodman SB, Maloney WJ, Smith RL: Effects of tensile strain and fluid flow on osteoarthritic human chondrocyte metabolism in vitro. Journal of Orthopaedic Research 28(7), 907-913 (2010).

33. Shimizu K, Shunori A, Morimoto K, Hashida M, Konishi S: Development of a biochip with serially connected pneumatic balloons for cell-stretching culture. Sensors and Actuators, B: Chemical 156(1), 486-493 (2011).

34. Wu MH, Wang HY, Liu HL et al.: Development of high-throughput perfusion-based microbioreactor platform capable of providing tunable dynamic tensile loading to cells and its application for the study of bovine articular chondrocytes. Biomedical Microdevices, 1-10 (2011).

35. Moraes C, Chen JH, Sun Y, Simmons CA: Microfabricated arrays for high-throughput screening of cellular response to cyclic substrate deformation. Lab on a Chip - Miniaturisation for Chemistry and Biology 10(2), 227-234 (2010).

36. Simmons CS, Sim JY, Baechtold P et al.: Integrated strain array for cellular mechanobiology studies. Journal of Micromechanics and Microengineering 21(5), (2011).

37. International F, Corporation: FX-4000™ Flexcell® Tension Plus™ & Tissue Train® System Specifications. (2007).

38. Wang JHC, Goldschmidt-Clermont P, Wille J, Yin FCP: Specificity of endothelial cell reorientation in response to cyclic mechanical stretching. Journal of Biomechanics 34(12), 1563-1572 (2001).

39. Morioka M, Parameswaran H, Naruse K et al.: Microtubule dynamics regulate cyclic stretch-induced cell alignment in human airway smooth muscle cells. PLoS ONE 6(10), (2011).