Out-of-Equilibrium Microrheology inside Living Cells

Claire Wilhelm
*
Laboratoire Matie`re et Syste`mes Complexes (MSC), UMR 7057, CNRS and University Paris Diderot, Paris, France
(Received 12 December 2007; published 9 July 2008)
Both forced and spontaneous motions of magnetic microbeads engulfed by Dictyostelium cells have
served as experimental probes of intracellular dynamics. The complex shear modulus G

(w), determined
from active oscillatory measurements, has a power-law dynamics and increases with the probe size,
reecting intracellular structural complexity. The combined use of passive microrheology allows one to
derive the power spectrum of active forces acting on intracellular phagosomes and to test the validity of
the uctuation-dissipation theorem inside living cells.
DOI: 10.1103/PhysRevLett.101.028101 PACS numbers: 87.17.Rt, 83.85.Cg, 87.16.Uv, 87.80.Fe
During the diverse processes that govern the survival
and development of living organisms, each cell must react
continuously, in a perfectly orchestrated manner, to a
variety of internal and external constraints. During the
past two decades, micromanipulations have been devel-
oped to mechanically deform living cells in a controlled
manner. Most of these approaches target the whole cell [1]
or involve local constraints applied to the cell surface [2
4]. Biophysical methods capable of deforming the intra-
cellular environment (active intracellular microrheology)
are rare, however, because of the experimental difculties
inherent in introducing a probe into a living cell and
manipulating it in a noninvasive manner. Some recent
reports describe the use of engulfed magnetic microbeads
or endocytosed magnetic nanoparticles submitted to a
magnetic force [5] or torque [6] and embedded granules
submitted to optical forces [7]. However, most studies of
intracellular mechanics have focused on spontaneous dis-
placement of intracellular granules or internalized mi-
crobeads in order to deduce local dynamic properties
(passive intracellular microrheology) [810].
In an in-equilibrium system, the uctuation-dissipation
theorem (FDT) links thermal uctuations (passive mea-
surements) to the response to external perturbation (active
measurements). For biological systems, recent studies have
shown the violation of the FDT. Deviation from equilib-
rium has been observed with an in vitro model system
consisting of an actin network with embedded myosin
motors [11] and has been pioneered for living cells by
Bursac et al. [12]. In this case, violation of the FDT was
reported by comparing both forced and spontaneous mo-
tions of magnetic beads attached to the cell membrane and
bound to deeper structures of the cytoskeleton. Inside the
cell, in the cytoplasm, direct evidence of the deviation from
equilibrium is lacking. In this Letter, the combination of
active and passive intracellular microrheology is used to
explore internal cell mechanics, with respect to probe size
and to the activity of intracellular motors.
Magnetic beads of various sizes (0.312.8 m diame-
ter) were engulfed inside the cytoplasm of Dictyostelium
cells by the process of phagocytosis. When submitted to an
external magnetic eld, these so-obtained magnetic phag-
osomes acquire a magnetic moment m and self-organize
in chains containing N beads, as shown in Figs. 1 (electron
microscopy) and 2(a) (videomicroscopy).
The active microrheology experiments were based on
the recently developed rotational magnetic microrheology
approach [13], in which chains of magnetic microbeads are
manipulated by means of a sinusoidally oscillating mag-
netic eld [Fig. 2(b)]. Briey, the magnetic torque applied
to a chains of N beads
m

o

sinj2( 0)|,2 balances

the viscoelastic torque
ve
kVG0. The magnetic and
geometric factors
o
and k are given in Ref. [13]. G is
the material intrinsic complex shear modulus G

(w)
G
/
(w) +|G
//
(w), where G
/
and G
//
relate to the elastic
and viscous responses, respectively. Angles
o
e
|wi
and 0 0
o
e
|wi|
are described in Fig. 2(b). For small
( 0) values,
m

ve
gives G
/
and G
//
as a function of
the loss of amplitude 0
o
,
o
and the phase lag of the
chains: G
/
(
o
,kV)(
o
,0
o
)

sin() 1 and G
//

(
o
,kV)

(
o
,0
o
)

sin(). Videomicroscopy with an ultra-

fast camera (up to 1000 frames per second) was used to
detect the chain oscillations at frequencies varying from
0.2 to 20 Hz. The oscillation of at least 20 different chains,
in different cells, was analyzed for a given number N of
FIG. 1. Electron micrographs of the Dictyostelium cell interior
containing chains of magnetic beads within phagosomes ar-
ranged along the magnetic eld direction.
PRL 101, 028101 (2008)
P HYS I CAL RE VI E W L E T T E RS
week ending
11 JULY 2008
0031-9007,08,101(2),028101(4) 028101-1 2008 The American Physical Society
beads in the chain, at a given frequency, and for two
different angles
o
(15.5

and 28

). No deviation was
observed for different
o
values, ensuring the linearity of
the measurement. Both G
/
and G
//
increased with fre-
quency [Fig. 3(a)] and followed a power law w
n
, with
the same exponent n of between 0.5 and 0.6. The G
/
,G
//
ratio did not vary with frequency and was found close to 1,
and all of the data could be tted [Fig. 3(a)] with a power-
law constant phase model for the complex shear modulus:
G

(w) G
o
(|w,w
o
)
n
, (1)
with w
o
1 s
1
. Values for G
o
and n are given in the
Fig. 3 caption.
This power-law behavior is inconsistent with the dis-
crete number of time constants predicted by simple visco-
elastic models [5,14]. The power-law exponent n is
indicative of the degree of solid- or liquidlike mechanical
behavior of the probed material (solidlike for n 0 and
uidlike for n 1). The exponent found here for the
intracellular rheological behavior (n 0.50.6) is much
larger than that obtained when a cell is stretched between
microplates (n 0.20.3) [1], locally deformed by torque
(n 0.20.25) [2] or force (n 0.180.21) [4] exerted
on membrane-bound microbeads or poked with a microtip
(n 0.22) [3]. This is in keeping with the only other
active microrheology study inside the cell which yielded
power-law behavior with an exponent of 0.5 by measuring
the response of intracellular granules trapped and displaced
with optical tweezers [7]. Results of intracellular passive
one-point microrheology experiments measuring uctua-
tions of intracellular particles or granule positions submit-
ted only to thermal forces also yielded an exponent of 0.5
[8]. Interestingly, the exponent was 0.33 in lamellar re-
gions, which more closely resemble the cortical cytoske-
leton. Our active measurements therefore support an
intracellular medium that is much more uidlike than the
membrane and cortical cytoskeleton, with almost equiva-
lent elastic and dissipative contributions.
The major nding concerns the dependence of the rheo-
logical measurement with the probe size: Figure 3(b)
shows changes in the exponent n and the prefactor G
o
as
a function of the length of the probed chains. Although n
varies very little with the probe size, G
o
increases linearly,
by 1 order of magnitude (from 0.27 to 2.4 Pa), when the
chain length L increases from hundreds of nanometers to
the scale of the whole cell. This increase reects a stiffer
medium at larger scales, coherent with structural complex-
ity and the involvement of more elements probed while
increasing the probe size. Indeed, among thousands of
components contributing to the architecture of the cyto-
plasm, intracellular membranes have a length ranging from
nanometers to a few microns, and intracellular cytoskele-
ton laments organize into bers a few microns long or
remain isolated with a typical length of a few nanometers.
The disposal of an active magnetic probe microrheolog-
ical measurement and its comparison with spontaneous
motion of embedded probes opens a new window to ex-
plore the deviation from equilibrium inside living cells and
to evaluate the forces generated by molecular motors. The
FIG. 3 (color online). (a) Mean values of G
/
(w) and G
//
(w)
obtained with different probe sizes. Deviation (not shown for
clarity) was in the range of 40% for every measure. The solid
lines are the t of Eq. (1) G

(w) G
o
(|w,w
o
)
n
to the data. For
chains containing two beads (N 2), G
o
(0.27 0.04) Pa,
G
o
(0.92 0.06) Pa, and G
o
(2.47 0.1) Pa and n
(0.59 0.03), n (0.63 0.02), and n (0.55 0.03) for
beads of 0.3, 1, and 2.8 m diameters, respectively.
(b) Exponent n and prefactor G
o
as a function of the probe
length L NJ, with N the number of beads in the chain probe
and J the bead diameter.
FIG. 2. (a) Oscillation of chains of magnetic phagosomes
(1 Hz
o
28

) containing (from left to right) 0.3-, 1-, and

2.8-m-diameter magnetic beads, studied by means of video-
microscopy. Videos are available in supplementary auxiliary
les [18]. (b) The magnetic device used for microrheological
measurements inside the cell (left): Two permanent magnets
create a homogeneous eld of 79 mT, and a perpendicular pair
of coils is supplied with the same current intensity (2 and 4 A),
providing a magnetic eld of 21 and 42 mT, respectively. Angle
notations (middle) and typical angular response curve (right).
PRL 101, 028101 (2008)
P HYS I CAL RE VI E W L E T T E RS
week ending
11 JULY 2008
028101-2
spontaneous motion of a single phagosome within the
cytoplasm is described by a generalized Langevin equation
m
J
Ji
+
R
i
0
[(i r)(r)Jr E(i), in which (i) is the
phagosome velocity, [(i) a delayed friction function that
takes into account the viscoelastic properties, and E(i) the
force acting on the phagosome, including contributions of
both thermal Brownian forces and driving forces generated
by molecular motors.
Assuming that inertia is negligible for micron-sized
phagosomes of diameter o and that generalization of
Stokes law to viscoelastic materials states that G(w)
(|w)[(w)
6ro
, the Fourier transforms of E(i) and (i) are linked
as E(w)
6roG(w)
(|w)
(w). The power spectrum of intra-
cellular forces P
EE
(w), a Fourier transform of the correla-
tion function (E(i)E(i +r))
i
, then writes P
EE
(w)
}6roG

(w)}
2
w
2
P

(w), with P

(w) related to Fourier transform

(r
2
(w)) of the mean square displacement (MSD) (r
2
(i)):
w
2
(r
2
(w)) 4P

(w). P
EE
(w) can therefore be calculated
from the correlation function and shear modulus as
P
EE
(w) (3ro)
2
}G

(w)}
2
(r
2
(w)). (2)
In the absence of driving forces (at thermal equilibrium),
the FDT writes
(r
2
(w)) 4k
B
T
n
//
(w)
w
, (3)
where n

(w) n
/
(w) n
//
(w)
1
6roG

(w)
, giving
P
EE
(w) (6ro)k
B
T
G
//
(w)
w
. (4)
The uctuation-dissipation theorem [Eqs. (3) and (4)] is no
longer valid for systems out of equilibrium, as it is pre-
sumed for living cells, owing, in particular, to the activity
of molecular motors. Evaluation of P
EE
(w) for intracellu-
lar forces was rst proposed in Ref. [10], where it was
demonstrated that the sub- versus superdiffusive behaviors
of intracellular granules versus phagosomes can be related
to the activity of associated microtubule motors. More
recently, Lau et al. [15] calculated the power spectrum of
stress uctuations by comparing passive measurements
made with intracellular granules to active measure-
ments described by Fabry et al. [2] using magnetic beads
coupled through the membrane to the cortical cytoskele-
ton. However, both measurements are indirect, as the active
and passive probes differ, with different cellular locations,
hindering the quantication of motor activities. Here cor-
relation [passive microrheology, (r
2
(w))] and response
functions [active microrheology, G

(w)] were obtained

independently with the same probes, with the same intra-
cellular location. For the passive microrheology measure-
ments, the magnetic phagosomes 0.3, 1.0, and 2.8 m in
diameter were tracked every 0.01 s and the mean square
displacement (r
2
(i)) computed [see Fig. 4(a)]. The MSDs
of all beads were well described by power-law behavior
over the range 0.01 <i <4 s: (r
2
(i)) D

(i,i
o
)

, with
between 1.2 and 1.6 for all three types of beads, demon-
strating superdiffusive behavior. i
o
is taken to be 1 s, and
D

is expressed in units of m
2
. For each MSD, the power
spectral density was computed by Fourier transform, and
its mean value could be adjusted by a single power law
[Fig. 4(a)]
(r
2
(w)) D

w
w
o

(+1)
. (5)
w
o
is taken to be 1 s
1
, and Dis expressed in units of m
2
s.
Values for D and are given in the Fig. 4 caption. For the
active microrheology measurements, G

(w) is taken as the

measure made in the same cells with the same probes,
using doublets of beads (N 2), to approach the one-
bead probe conditions used for uctuations. P
EE
(w) then
follows a power-law behavior P
EE
(w) w

, with
+ 1 2n. With thermal equilibrium, 1 n.
Violation of this relationship was rst observed with a
probe strongly coupled to an actin cytoskeleton via the
cell membrane, placing around 2 [12], as rst suggested
by Lau et al. [15]. Here, with the use of intracellular
probes, was found around 1.2. The value 2 corre-
sponds to a series of force steps, while 0 corresponds
to a series of instantaneous innite force pulses (uncorre-
lated spectrum). Avalue in between ( 1) should corre-
spond to smoothing of discontinuities in instantaneous
force pulses. This implies that motor activities are more
correlated and continuous for the actin cytoskeleton probed
through the cell membrane than intracellular motors re-
FIG. 4 (color online). (a) (r
2
(w)) as a function of frequency for
the three phagosome probes. Black lines correspond to the best
t with Eq. (3) (r
2
(w)) D(w,w
o
)
(+1)
: 1.38, 1.43,
and 1.39 and D 5.4 10
13
m
2
s, D 2.2
10
13
m
2
s, and D 2.4 10
14
m
2
s for beads of 0.3, 1, and
2.8 m diameters, respectively. Inset: MSD (r
2
(i)) computed
for each independent cell as (r
2
(i)) (jx(i + i) x(i)|
2
+
jy(i + i) y(i)|
2
), where (x, y)(i) is the bead position at time i,
i is the time lag, and angle brackets denote averaging over i.
(b) Effective temperature over the bath temperature as a function
of the probe diameter J
probe
. The set of equations (1)(3) then
gives the effective temperature T
eff
for the nonequilibrium
situation: k
B
T
eff
(w)
3roDG
o
2 sin(rn,2)
(w,w
o
)
n
, with w
o
1 s
1
.
The lines are visual guides.
PRL 101, 028101 (2008)
P HYS I CAL RE VI E W L E T T E RS
week ending
11 JULY 2008
028101-3
sponsible for phagosome transport. Besides, thanks to
independent measurements of intracellular stiffness and
mean square displacement, we derive quantitatively the
force uctuation spectrum. As it follows power-law behav-
ior, the Fourier transform P
EE
(w) is easily inverted, lead-
ing to the force correlation function:
(E(i)E(i +r))
i
E
o

r
r
o

1
, (6)
with r
o
set to 1 s and E
o
equal to 0.4, 2.7, and 5.7 pN for
0.3-, 1-, and 2.8-m beads, respectively. The force corre-
lation increases slightly with time in the probed time range
(0.011 s) with an exponent of about 0.2, a signature of
collective behavior with reinforcement of the force with
time, and an average value in the range of the stall force of
a single kinesin motor, which is about 5 pN. If one keeps in
mind the picture of molecular motors generating forces in
discontinuous force pulses, these statistical mean values of
the active forces are not unreasonable and support the
vision of cooperative motors acting together for active
transport of phagosome cargos [16].
The measured P
EE
(w) exhibit a strong mismatch with
the prediction at thermal equilibrium [Eq. (3)], demonstrat-
ing that the uctuation-dissipation theorem is no longer
valid inside living cells. To quantify the deviation from the
in-equilibrium situation, it has been proposed to replace the
bath temperature T in Eqs. (2) and (3) by a frequency-
dependent effective temperature T
eff
[17]. T
eff
(w) then
runs roughly as w
0.8
, for the accessible frequency range
(0.110 Hz): The lower the frequency, the higher the
effective temperature. Deviation from equilibrium is due
to the action of phagosome driving motors, the activities of
which should not interfere with high-frequency thermal
modes. Figure 4(b) shows the ratio of the effective tem-
perature to the bath temperature (37

C) as a function of
the probe diameter and at different frequencies. With
0.3-m-diameter probes, we found an effective tempera-
ture that is 20 times the bath temperature at 1 s
1
. By
contrast, the effective temperature is almost a hundredfold
the bath temperature with 1- and 2.8-m-diameter beads at
1 s
1
. The phagosomes of 0.3 m diameter have fewer
interactions with the networks, whereas phagosomes 1 and
2.8 m in diameter are more likely to come across cyto-
skeleton laments and undergo active transport.
In conclusion, using active magnetic bead rotational
microrheology, it is demonstrated that the intracellular
shear modulus has a power-law dynamics of exponent
between 0.50.6: The inside of the cell has a more liquid-
like behavior than the membrane and associated cyto-
skeleton. In addition, the absolute value of the shear modu-
lus increases with the probe size, reecting structural
complexity. By combining this active intracellular micro-
rheology with classical passive microrheology, this Letter
exhibits a strong violation of the uctuation-dissipation
theorem inside the cells. This deviation from the in-
equilibrium situation was estimated by measuring a
frequency-dependent effective temperature different from
the bath temperature or, alternatively, the power spectrum
of active intracellular forces, both of which are directly
coupled to the activity of biological motors.
The author acknowledges Jean-Claude Bacri, Damien
Robert, Florence Gazeau, and Francois Gallet for fruitful
discussions and Jacques Servais for technical help.
*[email protected]
[1] N. Desprat, A. Richert, J. Simeon, and A. Asnacios,
Biophys. J. 88, 2224 (2005).
[2] B. Fabry, G. N. Maksym, J. P. Butler, M. Glogauer, D.
Navajas, and J. J. Fredberg, Phys. Rev. Lett. 87, 148102
(2001).
[3] J. Alcaraz, L. Buscemi, M. Grabulosa, X. Trepat, B. Fabry,
R. Farre, and D. Navajas, Biophys. J. 84, 2071 (2003).
[4] M. Balland, N. Desprat, D. Icard, S. Fereol, A. Asnacios,
J. Browaeys, S. Henon, and F. Gallet, Phys. Rev. E 74,
021911 (2006).
[5] W. Feneberg, M. Westphal, and E. Sackmann, Eur.
Biophys. J. 30, 284 (2001).
[6] S. Marion, N. Guillen, J.-C. Bacri, and C. Wilhelm, Eur.
Biophys. J. 34, 262 (2005).
[7] M. Yanai, J. P. Butler, T. Suzuki, H. Sasaki, and H.
Higuchi, Am. J. Physiol. 287, C603 (2004).
[8] S. Yamada, D. Wirtz, and S. C. Kuo, Biophys. J. 78, 1736
(2000).
[9] Y. Tseng, T. P. Kole, and D. Wirtz, Biophys. J. 83, 3162
(2002).
[10] A. Caspi, R. Granek, and M. Elbaum, Phys. Rev. Lett. 85,
5655 (2000).
[11] D. Mizuno, C. Tardin, C. F. Schmidt, and F. C.
Mackintosh, Science 315, 370 (2007).
[12] P. Bursac, G. Lenormand, B. Fabry, M. Oliver, D. A.
Weitz, V. Viasnoff, J. Butler, and J. J. Fredberg, Nat.
Mater. 4, 557 (2005); P. Bursac, B. Fabry, X. Trepat, G.
Lenormand, J. P. Butler, N. Wang, J. J. Fredberg, and S. S.
An, Biochem. Biophys. Res. Commun. 355, 324 (2007).
[13] C. Wilhelm, J. Browaeys, A. Ponton, and J. C. Bacri, Phys.
Rev. E 67, 011504 (2003).
[14] C. Wilhelm, F. Gazeau, and J. C. Bacri, Phys. Rev. E 67,
061908 (2003).
[15] A. W. C. Lau, B. D. Hoffmann, A. Davies, J. C. Crocker,
and T. C. Lubensky, Phys. Rev. Lett. 91, 198101 (2003).
[16] K. B. Zeldovich, J. F. Joanny, and J. Prost, Eur. Phys. J. E
17, 155 (2005); S. Klumpp and R. Lipowsky, Proc. Natl.
Acad. Sci. U.S.A. 102, 17 284 (2005).
[17] N. Pottier, Physica (Amsterdam) 345A, 472 (2005); B.
Abou and F. Gallet, Phys. Rev. Lett. 93, 160603 (2004); S.
Jabbari-Farouji et al., Phys. Rev. Lett. 98, 108302 (2007).
[18] See EPAPS Document No. E-PRLTAO-101-068827 for
two videos showing the oscillations of chains of magnetic
phagosomes at 0.2 and 2 Hz frequencies. For more infor-
mation on EPAPS, see http://www.aip.org/pubservs/
epaps.html.
PRL 101, 028101 (2008)
P HYS I CAL RE VI E W L E T T E RS
week ending
11 JULY 2008
028101-4