Evidence of differential mass change rates between human breast cancer cell lines in culture

Evidence of differential mass change rates between human breast

cancer cell lines in culture

Elise A. Corbin

1,2,3

& Olaoluwa O. Adeniba

1,2

& Olivia V. Cangellaris

2,4

William P. King

1,2

& Rashid Bashir

2,4

Springer Science+Business Media New York 2017

Abstract Investigating the growth signatures of single

cells will determine how cell g rowth is regulated and cell

size is maintained. The ability to preci sely measure such

changes and alterations in cell size and cell mass could be

important fo r applications in cancer and drug screening.

Here, we measure the mass growth rate of individual be-

nign (MCF-10A), non-invasive (MCF-7), and highly-

invasive malignant (MDA-MB-231) bre ast cancer cells.

A micro-patterni ng technique was employed to allow for

the long-term growth of motile cells. Results show ma ss

growth rates at 4.8%, 1.2%, and 2.8% for MCF-10A,

MCF-7, and MDA-MB-231, demonstrating t hat normal

cells have a higher mass growth rate than cancerous cells.

All the cell lines show an increase in mass change rate

indicating that the mass accumulation rate is exponential

over a single cell cycle. The growth rates measured with

our MEMS sensor are compared with doubling times ob-

tained through conventional bulk analysis techniques, and

exhibit excellent agreement.

Keywords MEMS mass sensor

Breast cancer

Cell growth

rate

Cell micromechanics

Resonant frequency

1 Introduction

Cell growth is necessary for l ife, but the mechanism s that

control the regulation of cell mass, volume and growth rate

are still poorly understood (Lloyd 2013; Zangle and Te itell

2014). Cancer arises from a number of mutations in the

genetic makeup of a cell; through the cell cycle the mutated

genetic makeup is transferred to daughter cells and prolif-

eration causes these cells to mult iply (Lodish et al. 2003;

Alberts et al. 2002;Weinberg200 6 ).

Currently, a great deal is known about the conse-

quences of the mutations, but on a fundamental level there

is st ill much to learn. C ancerous mutations alter relevant

signaling pathways, which in-turn influence the interac-

tion and response of a cell to mechanical stimuli and

growth factors, cell fate, proliferation, transcription, mi-

gration, or differentiation. These pathways can change

how the cell cycle is regulated. In cancer, cells exhibit

unregulated growth and division as a consequence of

disrupted checkpoints within t he cell cycle, w hich is trig-

gered by the loss of proper signaling cues (Hanahan and

Weinberg 2000, 2011).

Cell growth consists of coordinated changes in both

mass and volume. A standard approach to measuring cell

growth is through accurate monitoring of cell size. A

Coulter counter is a well-established technique to measure

cell volume and has deepened our understanding as to how

cells regulate growth. However, cell mass is a more direct

and p recise indicator of ce ll size and it is more closely

indicative of biosynthetic processes in a cell. There are

many different approaches for measuring cellular growth

Electronic supplementary material The online version of this article

(doi:10.1007/s10544-017-0151-x) contains supplementary material,

which is available to authorized users.

* Rashid Bashir

rbashir@illinois.edu

Department of Mechanical Science and Engineering, University of

Illinois Urbana-Champaign, Urbana, IL 61801, USA

Micro and Nanotechnology Laboratory, University of Illinois

Urbana-Champaign, Urbana, IL 61801, USA

Perelman School of Medicine, University of Pennsylvania,

Philadelphia, P A 19104, USA

Department of Bioengineering, University of Illinois

Urbana-Champaign, Urbana, IL 61801, USA

Biomed Microdevices (2017) 19:10

DOI 10.1007/s10544-017-0151-x

and there is a debate as to whether looking at an aggregate

population is more correct and necessary (Cooper 2006;

Mitchison 2003; Mitchison 2005). When dealing with

time-dependent measurements, it is known that studying

bulk properties will often overlook single cell events and

provide misleading analysis (Di Carlo et al. 2012;DiCarlo

and Lee 200 6 ). Measuring the growth of single cells has

many challenges to overcome (Popescu et al. 2014), includ-

ing the ability to handle and manipulate individual cells in

real time. H owever, the use of microelectromechanical sys-

tems (MEMS) has provided opportunities to make signifi-

cant advances (Bryan et al. 2010; So n et al. 2012). Several

precise MEMS measurement techniques have been recently

developed t o better understand how the biophysical prop-

erties of a cell affect its cycle progression and behavior in

disease. This includes magnetic twisting cytometry (Wang

and Thampatty 2006), optomechanical measures (Park

et al. 2015), micropipette aspiration (Evans and Yeung

1989; Hochmuth 2016;Satoetal.2016), quartz crystal

microbalance (QCM) (Li et al. 2005), atomic force m icros-

copy (AFM) (Cross et al. 2007; Kuz netsova et al. 2007;

Plodinec et al. 2012), suspended microchannel resonator

(SMR) (Son et al. 2012), and quantitative phase microsco-

py (QPM) (Popescu et al. 2008). Although QPM m ethods

are widely used to study adherent cells, they can only mea-

sure the ‘dry mass’ of live cel ls; thus, they rely on a refrac-

tive index difference to measure the non-aqueous (proteins,

nucleic acids, and lipid m olecules) mass density of the cell.

Meanwhile, allowable sample sizes in QCM are based on

electrode dimensions, which limit its adoption. SMR uses a

flow-through technique that senses the ‘buoyant mass’ of a

cell – the volumetric difference between the densities of a

cell and that of the media replaced by the cell – which is

similar to the optically measured ‘dry m ass’ as described in

QPM. This limits the types of cel ls able to be studied with

SMR. As an alternative to these methods, our pedestal

resonant sensor measures the whole cell apparent mass

and thus could potentially illustrate critical elements of

cell growth.

In this paper, we use our MEMS resonant mass sensors

to compare t he long-term growth of benign epithelial cells

(MCF-10A) with cancer cells of both high (MDA-MB-

231) a nd low invasive properties (MCF-7). In order to

measure long-term growth, we confine and trap single cells

on the pedestal of the MEMS sensor and repeatedly ob-

serve its resonant frequency to explore the way that cells

accumulate mass, grow, and divide. This approach is aided

by the use of cell patterning on the suspended resonant

devices in order to maintain capture of the cells over

time (Corbin et al. 2014a). From this approach, we find

that all the cell lines examined exhibit a grow th rate that

increases with cell mass, but that each cell line exhibits its

own unique mass change rate.

2 Experimental methods

2.1 MEMS resonant pedestal sensor

The MEMS resonant mass sensor comprises four beam-

springs suspending a square platform (60 × 60 μm

), as has

been described previously (Corbin et al. 2014b;Milletetal.

2012;Parketal.2010). The pedestal design both allows for a

large cell capture region and also minimizes the variation of

the displacement amplitude across the vibrating platform to

provide uniform mass sensitivity. These sensors are arrayed

on a chip in a 9 × 9 format of 81 sensors to maximize through-

put and capture efficiency. It operates in a first resonance

mode for mass sensing, which is a vertical displacement vi-

bration with resonant frequencies of approximately 160 kHz in

air and 60 kHz in liquid. This sensor is driven by passing an

actuation current through the sensor in a static magnetic field

to generate a Lorentz force that forces vibration.

2.2 Micro-patterned surface functionalization

In order to improve cell capture and retention, cells are

micro-patterned on the pedestal sensors through a selec-

tive functionalization and backfill passivation technique

(Corbin et al. 2014a, b). First, a hyd rophobic layer of

hexamethyldisilizane (HMDS) was applied to the sensor

surface through vapor deposition to promote efficient de-

position of Pluronic® F127 (Dorvel et al. 2010). Due to

the delicate nature of the structure, a photoresist transfer

technique (Yeom and Shannon 2010)wasusedto

provide a uniform layer of photoresist to be patterned.

The sample was then developed and exposed to oxygen

plasma to remove HMDS from the openings in the pho-

toresist for deposition of collagen type I. After rinsing

the surface with PBS, the chip was soaked in acetone to

lift-off the photoresist, leaving collagen selectively on

the pedestals of the sensors and surrounded by HMDS

everywhere else. At this point, pluronic was deposited on

the HMDS covering the beam springs and a thin edge of

the pedestal sensor through a backfill technique. This has

the effect of selectively promoting cell adhesion on the

collagen-functionalized pedestal while blocking unwant-

ed adhesion at sites with pluronic.

2.3 Cell mass measurement

The mass measurement procedure involves comparing the

resonant frequency of a loaded sensor with its original,

unloaded state to determine the mass of the loaded object.

This is a well-characterized method described fully elsewhere

(Corbin et al. 2013;Milletetal.2012), therefore it will only be

briefly introduced here. The measurement system combines

electromagnetic actuation, a lock-in amplifier, and a laser

10 Page 2 of 7 Biomed Microdevices (2017) 19:10

Doppler vibrometer (LDV) to measure the velocity of the

vibrating MEMS sensor platform to ultimately determine the

resonant frequency of the device. This is achieved by moni-

toring the difference in phase between applied actuation cur-

rent and sensor vibration. The excitation frequency is updated

based on this phase until converging upon the resonant fre-

quency. This procedure is used to estimate the resonant fre-

quency of the devices in a series of different states to extract

the mass of the adhered cell.

Calculating object mass requires three separate resonant

frequency measurements. First, the empty sensor resonant fre-

quency was measured in air to obtain the spring constant.

Next, the micro-pa tterning procedure was completed by

backfilling with pluronic to passivate the remaining chip sur-

faces. We then placed a PDMS chamber over the sensor area,

added fresh warm complete cell culture media, and sealed the

chamber with a sterile glass coverslip. The sensor was then

loaded into the incubation chamber of our system and allowed

to stabilize to 37 °C prior to the measurement. The second

empty sensor frequency was then measured in liquid to deter-

mine the reference frequency, which is reduced from the in-air

frequency due to hydrodynamic loading. At this point cells

were randomly seeded on the sensor area and allowed to at-

tach for 1 h before being gently rinsed to remove non-adhered

cells. The sensor is then restabilized at 37 °C in the incubation

chamber with the necessary conditions for each cell type (i.e.

humidity and CO

level). Finally, the in-liquid resonant fre-

quency of the mass loaded sensors were measured for com-

parison with the empty reference sensor frequency to calculate

the attached mass. This measurement of the loaded sensor

frequency was repeated over time to produce a growth profile.

Previous studies (Sato et al. 2016)showthatthrough250in-

media measurements, the sensor resonant frequency can be

determined within ±0.94 Hz (95% Confidence Internal (CI)),

yielding a mass resolution of 8.5 pg within a confidence in-

terval of 95% in liquid. With this sensor mass resolution per-

centage of ~1% of a typical mammalian cell mass (~1 ng), we

can guarantee that within a CI = 95%, an exponential and

linear growth model can be differentiated by our resonant

mass sensor (See Supplemental Information).

2.4 Breast epithelial cell culture

Normal human breast epithelial cells (MCF-10A) were cul-

tured in Dulbecco’s M odified Eagle Medium/Ham’sF-12

(Gibco) with 5% horse serum, 20 ng/mL EGF, 0.5 mg/mL

hydrocortisone, 100 ng/mL cholera toxin, 10 μg/mL insulin,

and 1% penicillin streptomycin. Human breast adenocarcino-

ma cells (MCF-7) were cultured in Dubecco’s Modified Eagle

Medium (Gibco) with 10% fetal bovine serum and 1% peni-

cillin streptomycin. Highly metastatic human breast adenocar-

cinoma cells (MDA-MB-231) were cultured in Leibovitz’s

Pluronic® F127

collagen

Standard Sensor

Micro-patterned Sensor

pedestal

spring

reference pad

Fig. 1 Overview of sensor

design with select

functionalization and passivation:

a Cartoon of the sensor layout;

and b cartoon of the desired

patterning with collagen

selectively patterned in the center

of the pedestal and with pluronic

backfilled everywhere else c

Bright field image of a single

released non-patterned pedestal

sensor that has been seeded with

human colon cancer cells (HT29).

It is shown that cells are able to

attach to the springs and

anywhere else that could

ultimately compromise the cell

measurement. d Bright Field

image of a single released

patterned pedestal sensor seeded

with human breast cells (MCF-

10A) and then rinsed to remove

non-attached cells, leaving only

the cell captured on the platform

Biomed Microdevices (2017) 19:10 Page 3 of 7 10

L-15 Medium (Sigma-Aldrich) with 10% fetal bovine serum

and 1% penicillin streptomycin. MCF-7 and MCF-10A were

cultured at 5% CO

, 100% humidity, while MDA-MB-231

were cultured at 100% humidity and no CO

Cells were introduced onto the sensors at a total of 9000

cells p er chip and allowe d t o a dhere. The sensors were

rinsed with fresh growth media and the culture chamber

was sealed with a sterilized glass cover slip for the

measurement. Mass measurements were taken approxi-

matelyevery20minforupto24h.

2.5 Bulk doubling time measurements

Since cell-collagen interaction can strongly change cellular

behavior, we performed doubling time experiments on colla-

gen coated plastic dishes and non-collagen coated dishes for

reference. We plated the all the cell lines at the same starting

density of 318 cells/mm

and culture conditions as with the

single cell analysis experiments and imaged them for 24 h.

Using ImageJ we determined the cell count and found that

MCF-7, MCF-10A, and MDA-MB-231 had doubling times

of 43.7 ± 7.4, 20.1 ± 1.9, and 26.7 ± 8.8, respectively.

3 Results and discussion

Direct, long-term growth profiles of breast epithelial cell lines

are measured using MEMS resonant pedestal sensors with

micro-patterned surfaces for selective functionalization and

passivation (Fig. 1). The cell lines st udied include MDA-

MB-231 and MCF-7, which are cancerous with high and

low invasive potential, respectively, and a benign cell line,

MCF-10A, for comparison. Repeated mass measurements of

single cells captured on the mass sensor reveal the increase of

cellular mass due to growth over the cell cycle. Figure 2 shows

examples of individual cell growth curves that continue until

the cell division where, interestingly, a temporary sharp de-

crease in apparent mass is detected (Movie S1-S3; Online

Resource 1–3). These sharp decreases in apparent mass can

be seen in the selected growth profiles at approximately 10,

13, and 9 h for MDA-MB-231, MCF-7, and MCF-10A, re-

spectively; however, each investigated cell divided at a differ-

ent time as our populations were not synchronized. During

mitosis the dividing cell will partially detach from the plat-

form, thus decreasing the contact area and altering the shape

of the cell (Park et al. 2010). This geometry change can lead to

a reduction of the inertial loading of the cell decreasing the

apparent mass as described by a 2-DOF model (Corbin et al.

2015; Corbin et al. 2013;Parketal.2010; Corbin et al. 2016).

The result of the 2-DOF model revealed that geometry and

viscoelasticity of the target, or cell, influences the mass mea-

surement. Cells are known to be soft and cancer cells are

known to be even softer ( Cross et al. 2007;Gucketal.

2016; Wirtz et al. 2011). The viscoelastic properties of cells

can lead to the cell operating out of phase with our sensor

causing a geometry change of the cell. However, it is impor-

tant to note that the increase in apparent mass during the cell

growth represents a true increase in cell mass and is not a

geometry or contact area artifact (Park et al. 2010).

To better investigate the growth dynamics of the benign

and cancerous epithelial cell populations, we analyzed the

MDA-MB-231

MCF-7

Cell Mass (ng)

MCF-10A

Time (hr)

Cell Mass (ng) Cell Mass (ng)

Time (hr)

Fig. 2 Mass measurement of adherent cells versus time for each cell line:

a MDA-MB-231; b MCF-7; and c MCF-10A. Each growth profile shows

an increase of a single adherent cell, then will go through a cell division

that is marked by a sudden decrease in cell mass, once division has

completed, the growth profile continues

10 Page 4 of 7 Biomed Microdevices (2017) 19:10

individual temporal mass profiles of all the cell lines. The

derivative of the cell mass profiles provides the instantaneous

mass change rate. For a given cell line, the data from each

individual measured cell is pooled together, and Fig. 3 (a -

c) show histograms of the instantaneous mass change rate data

divided into four groups based on their instantaneous mass. It

is clearly evident from these histograms that mass change rate

does depend on the instantaneous cell mass. The small oscil-

lations in mass data observed in Fig. 2 are expected and can

lead to negative mass change rates. These oscillations can

arise from a combination of short-term variations in cell vis-

coelasticity, density, or adhesion of the cell over the cell cycle.

Small fluctuations of these properties occur naturally and have

a small impact on the apparent mass detected by our sensors.

y = 0.028x + 0.047

y = 0.012x + 0.0008

y = 0.048x + 0.22

0.0

0.2

0.4

0.6

0.8

1.0

1.4

1.6

1.2

0.50.0-0.5

1.0

0.5

-0.5

-1.5

-2.5

y = 0.76x - 1.227

y = 0.757x - 0.9137

y = 0.793x - 1.846

MCF7MCF10A MDA-MB-231

Doubling Time (hr)

Fig. 4 Analysis of mass change rate per unit mass of individual cells. A

five point moving average of changes in mass from all culture data points

of individual breast cells. a Apparent values: Average cells acquire 1.2%,

2.8%, 4.8% additional mass every hour for MCF-7, MDA-MB- 231, and

MCF-10A, respectively. (Inset) Log-log plot shows a power law of the

different cell lines where the slopes are less than unity verifying

consistency with scaling rules of energy consumption versus size of an

organism. b Average doubling times found through mass measurement

for MCF-7, MDA-MB- 231, and MCF- 10Awith error bars representing

standard error

Apparent

abc

Fig. 3 Analysis of cell growth rate and mass change rate versus mass. Four histograms accounting for the mass accumulation at specific mass values: a

MDA-MB-231; b MCF-7; and c MCF10-A

Biomed Microdevices (2017) 19:10 Page 5 of 7 10

Despite these mass fluctuations, there is a clearly identifiable

increase in cell mass over long periods of time, as confirmed

by simultaneous optical imaging.

Fig. 4a translates the distributions found in Fig. 3 to find the

cell mass growth rate indicated by the linear trends of the bin

max peaks for both the apparent and corrected masses. These

trends suggest that cells with greater mass also have an in-

creased rate of mass accumulation, regardless of cell line.

However, the cell lines do have different exponential growth

rates, which is the slope of the mass change rate against mass.

Figure 4a shows that MCF-10A, MCF-7, and MDA-MB-231

on average accumulate 4.8%, 1.2%, and 2.8% of their mass

every hour for the apparent mass values. Figure 4bpresents

how the growth rates translate to mass doubling times of 14.8,

58.2, and 25.1 h (= log (2)/ log (1 + rate)). To determine if

doubling times measured with the sensor differed between cell

lines, we used analysis of covariance (ANCOVA) to compare

the fitted mass change rates. Through this analysis we found

that the mass change rate, and thus the doubling time, signif-

icantly depended on cell line (p = 0.012). Post hoc tests be-

tween each cell line, with Bonferroni correction for multiple

comparisons, found that MCF-7 and MCF-10A differed sig-

nificantly (p = 0.013), but neither differed from MDA-MB-

231 significantly (p = 0.434 for MCF-7 and p = 0.182 for

MCF-10A). The inset of Fig. 4a presents log-log plots com-

paring mass change rate with cell mass. The growth follows

the rules of scaling energy consumption by having a slope less

than unity(Hou et al. 2008).

The American Type Culture Collection biological resource

center (ATCC) provides doubling times, or the times for cul-

ture to double in size, for each cell line. For comparison, a

population doubling time was considered for each cell line. In

these bulk population experiments we recreated the microen-

vironment of our single cell experiments by coating collagen

on plastic dishes. Table 1 collects the doubling time data of

each cell line determined by ATCC, our population experi-

ment, and our mass measurements. The MCF-10A cells ap-

pear to agree well with the bulk doubling time from ATCC,

while they tend to grow slightly faster compared to our bulk

doubling time measurements. The low invasive MCF-7 cells

grow more slowly than the ATCC doubling times and our bulk

doubling times. MDA-MB-231 (high invasive) cancer cell

line agrees well between our measured and our bulk doubling

times, however, it exhibits fairly substantial differences be-

tween measured and ATCC values. The differences in growth

values are based on the direct measurement of mass of indi-

vidual cells while both population and ATCC values are based

on bulk cell counting. Although ATCC is a world reference for

cell lines, the growing rate for each cell line varies with culture

media and methodology. Heterogeneity exists within a homo-

geneous population implying that cells are not created equal,

and yielding differences between cells such as size or mass.

Expression of subtle signature differences in growth among

individual cells can help explain why one cell may grow dif-

ferently or lead to metastasis. Our measurements show good

agreement with our bulk measures; however, a more in depth

statistical analysis is necessary to verify differences between

cell lines.

Recently, a deeper understandin g of cell growth dy-

namics has revealed how cells grow individually and as

a population over time (Park et al. 20 1 0 ). This work has

expanded upon that initial finding and validates it through

the study of multiple cell lines. Here, we showed that the

average growth rates of the investigated adherent cell

lines increase with cell mass. It has been hypothesized

that size homeostasis can b e maintained in one of two

ways: through linear growth without regulation (Conlon

and R aff 2003), or ex ponential growth that requires

check-point regulations (Godin et al. 201 0;Tzuretal.

2009). While the specific regulatory factors defining

growth are still intangible, it is li kely that size is main-

tained by some signaling, possibly from environmental

cues or genetics; however, we observ e common trends

between different cell lines that are consistent with previ-

ous findings (Park et al. 2010).

4 Conclusion

This paper presents the use of MEMS resonant sensors to

investigate the differences in growth between benign and ma-

lignant adherent cancer cells through long-term mass mea-

surements. Cells from each investigated cell line show a n

increase in mass change rate with respect to mass; therefore,

the heavier cells accumulate more mass more quickly.

Through the measurements with the MEMS sensor, we were

able to determine that doubling time significantly depends on

cell type, suggestive of inherent differences in cell growth

depending on cancer phenotype. These doubling times also

agreed well with bulk measurements and standard reference

values from ATCC. The agreement in doubling time measures

supports the accuracy of mass measurements, and future stud-

ies can take advantage of these devices to explore instanta-

neous growth of individual cells.

Table 1 A comparison of the doubling times of the values obtained

from the MEMS resonant sensor growth measurements and the American

Type Culture Collection (ATCC) database values

Cell line Doubling time

Measured (h) Counted (h) ATCC (h)

MCF-10A 14.8 20.1 16

MCF-7 58.2 43.7 38

MDA-MB-231 25.1 26.7 38

10 Page 6 of 7 Biomed Microdevices (2017) 19:10

These measures have the ability to expand our understand-

ing of adherent cell growth using a non-destructive technique

capable of long-term observations. There are many highly

regulated processes during cell growth including the replica-

tion of intracellular material that lead to mass accumulation,

an increase in the physical size of the cell, and progression

through the cell cycle. Through direct measurement of indi-

vidual cell mass, we can better understand the mechanisms

that form the basis of uncontrolled proliferation in cancer.

The potential to expand the current system to include the

monitoring of cell cycle status through optical verification

with fluorescence promises a next generation of mass mea-

surement for identification of growth rates durin g specific

cycle phases.

References

B. Alberts, A. John son, J. Lewis, M. R aff, K. Roberts, P. Walter,

Molecular Biology of the Cell, 4th edn. (Garland Science, New

York, 2 002)

A.K. Bryan, A. Goranov, A. Amon, S.R. Manalis, Proc. Natl. Acad. Sci.

107, 999 (2010)

I. Conlon, M. Raff, J. Biol. 2 (2003)

S. Cooper, Theor. Biol. Med. Model. 3, 10 (2006)

E.A. Corbin, L.J. Millet, J.H. Pikul, C.L. Johnson, J.G. Georgiadis, W.P.

King, R. Bashir, Biomed. Microdevices. 15, 311 (2013)

E.A. Corbin, B.R. Dorvel, L.J. Millet, W.P. King, R. Bashir, Lab Chip 14,

1401 (2014a)

E.A. Corbin, L.J. Millet, K.R. Keller, W.P. King, R. Bashir, Anal. Chem.

86, 4864 (2014b)

E.A. Corbin, F. Kong, C.T. Lim, W.P. King, R. Bashir, Lab Chip 15,839

(2015)

E.A. Corbin, O.O. Adeniba, R.H. Ewoldt, R. Bashir, Appl. Phys. Lett.

108, 93701 (2016)

S.E. Cross, Y.-S. Jin, J. Rao, J.K. Gimzewski, Nat Nano 2, 780 (2007)

D. Di Carlo, L.P. Lee, Anal. Chem. 78, 7918 (2006)

D. Di Carlo, H. T. K. Tse, and D. R. Gossett, in Single-cell anal. Methods

Protoc., ed. by S. Lindström and H. Andersson-Svahn (Humana

Press, T otowa, 2012), pp. 1–10

B.Dorvel,B.Reddy,I.Block,P.Mathias,S.E.Clare,B.Cunningham,

D.E. Bergstrom, R. Bashir, Adv. Funct. Mater. 20,87(2010)

E. Evans, A. Yeung, Biophys. J. 56, 151 (1989)

M. Godin, F.F. Delgado, S. Son, W.H. Grover, A.K. Bryan, A. Tzur, P.

Jorgensen, K. Payer, A.D. Grossman, M.W. Kirschner, S.R.

Manalis, Nat Meth 7, 387 (2010)

J. Guck, S. Schinkinger, B. Lincoln, F. Wottawah, S. Ebert, M.

Romeyke, D. Len z, H.M. Erickson , R. Anan thakrishn an, D.

Mitchell, J. Käs, S. Ulvick, C. Bilby, Biophys. J. 88,3689

(2016)

D. Hanahan, R.A. Weinberg, Cell 100, 57 (2000)

D. Hanahan, R.A. Weinberg, Cell 144, 646 (2011)

R.M. Hochmuth, J. Biomech. 33, 15 (2016)

C. Hou, W. Zuo, M.E. Moses, W.H. Woodruff, J.H. Brown, G.B. West,

Science 322, 736 (2008)

T.G. Kuznetsova, M.N. Starodubtseva, N.I. Yegorenkov, S.A. Chizhik,

Micron 38, 824 (2007)

J. Li, C. Thielemann, U. Reuning, D. Johannsmann, Biosens.

Bioelectron. 20, 1333 (2005)

A.C. Lloyd, Cell 154, 1194 (2013)

H. Lodish, A. Berk, P. Matsudaira, C.A. Kaiser, M. Krieger, M.P. Scott,

S.L. Zipursky, J. Darnell, Molecular Cell Biology, 5th edn. (WH

Freeman, New York, 2003)

L.J. Millet, E.A. Corbin, R. Free, K. Park, H. Kong, W.P. King, R. Bashir,

Small 8, 2555 (2012)

J. Mitchison, Int. Rev. Cytol. 165 (2003)

J.M. Mitchison, Theor. Biol. Med. Model. 2, 4 (2005)

K.Park,L.J.Millet,N.Kim,H.Li,X.Jin,G.Popescu,N.R.Aluru,

K.J. Hsia, R. Bashir, Proc. Natl. Acad. Sci.

107, 20691 (2010)

K. Park, A. Mehrnezhad, E.A. Corbin, R. Bashir, Lab Chip 15,3460

(2015)

M. Plodinec, M. Loparic, C.A. Monnier, E.C. Obermann, R. Zanetti-

Dallenbach, P. Oertle, J.T. Hyotyla, U. Aebi, M. Bentires-Alj,

L.Y.H., C.-A. Schoenenberger, Nat Nano 7, 757 (2012)

G. Popescu, Y. Park, N. Lue, C. Best-Popescu, L. Deflores, R.R. Dasari,

M.S. Feld, K. Badizadegan, Am. J. Physiol.-Cell Physiol 295, C538

(2008)

G. Popescu, K. Park, M. Mir, R. Bashir, Lab Chip 14, 646 (2014)

M. Sato, N. Ohshima, R.M. Nerem, J. Biomech. 29, 461 (2016)

S. Son, A. Tzur, Y. Weng, P. Jorgensen, J. Kim, M.W. Kirschner, S.R.

Manalis, Nat. Methods 9, 910 (2012)

A. Tzur, R. Kafri, V.S. LeBleu, G. Lahav, M.W. Kirschner, Science 325,

167 LP (2009)

J.H.C. Wang, B.P. Thampatty, An Introductory Review of Cell

Mechanobiology, Biomechanics and Modeling in Mechanobiology

(2006), pp. 1–16

R.A. Weinberg, The biology of cancer, 1st edn. (Garland Science, New

York, 2 006)

D. Wirtz, K. Konstantopoulos, P.C. Searson, Nat. Rev. Cancer 11,512

(2011)

J. Yeom, M.A. Shannon, Adv. Funct. Mater. 20, 289 (2010)

T.A. Zangle, M.A. Teitell, Nat Meth 11, 1221 (2014)

Biomed Microdevices (2017) 19:10 Page 7 of 7 10