Mechanical tugging force regulates the size of cell-cell junctions.
ABSTRACT Actomyosin contractility affects cellular organization within tissues in part through the generation of mechanical forces at sites of cell-matrix and cell-cell contact. While increased mechanical loading at cell-matrix adhesions results in focal adhesion growth, whether forces drive changes in the size of cell-cell adhesions remains an open question. To investigate the responsiveness of adherens junctions (AJ) to force, we adapted a system of microfabricated force sensors to quantitatively report cell-cell tugging force and AJ size. We observed that AJ size was modulated by endothelial cell-cell tugging forces: AJs and tugging force grew or decayed with myosin activation or inhibition, respectively. Myosin-dependent regulation of AJs operated in concert with a Rac1, and this coordinated regulation was illustrated by showing that the effects of vascular permeability agents (S1P, thrombin) on junctional stability were reversed by changing the extent to which these agents coupled to the Rac and myosin-dependent pathways. Furthermore, direct application of mechanical tugging force, rather than myosin activity per se, was sufficient to trigger AJ growth. These findings demonstrate that the dynamic coordination of mechanical forces and cell-cell adhesive interactions likely is critical to the maintenance of multicellular integrity and highlight the need for new approaches to study tugging forces.
- SourceAvailable from: Julie Gavard[show abstract] [hide abstract]
ABSTRACT: A major form of animal cell-cell adhesion results from the dynamic association of cadherin molecules, cytosolic catenins and actin microfilaments. Cadherins dynamically regulate the cytoskeleton. In turn, the actin cytoskeleton contributes to cadherin molecule oligomerization at cell contacts and to cell reshaping in response to environmental changes. Over the past two years, this evolutionarily conserved adhesion system has been intensively revisited in both its structural and functional aspects; this is illustrated by the remarkable progress in the determination of physical parameters of cadherin bonds (including force measurement) and the new insights into the role of alpha-catenin and the regulation of actin dynamics at cadherin contacts. Other recent studies uncover the important contribution of acto-myosin, microtubules and cell tension to adherens junction formation, cell differentiation and tissue reshaping/remodeling. An open challenge is now to integrate these new data with the diversity of cadherin adhesive complexes.Current Opinion in Cell Biology 11/2006; 18(5):541-8. · 11.41 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Apical constriction facilitates epithelial sheet bending and invagination during morphogenesis. Apical constriction is conventionally thought to be driven by the continuous purse-string-like contraction of a circumferential actin and non-muscle myosin-II (myosin) belt underlying adherens junctions. However, it is unclear whether other force-generating mechanisms can drive this process. Here we show, with the use of real-time imaging and quantitative image analysis of Drosophila gastrulation, that the apical constriction of ventral furrow cells is pulsed. Repeated constrictions, which are asynchronous between neighbouring cells, are interrupted by pauses in which the constricted state of the cell apex is maintained. In contrast to the purse-string model, constriction pulses are powered by actin-myosin network contractions that occur at the medial apical cortex and pull discrete adherens junction sites inwards. The transcription factors Twist and Snail differentially regulate pulsed constriction. Expression of snail initiates actin-myosin network contractions, whereas expression of twist stabilizes the constricted state of the cell apex. Our results suggest a new model for apical constriction in which a cortical actin-myosin cytoskeleton functions as a developmentally controlled subcellular ratchet to reduce apical area incrementally.Nature 12/2008; 457(7228):495-9. · 38.60 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Shaping a developing organ or embryo relies on the spatial regulation of cell division and shape. However, morphogenesis also occurs through changes in cell-neighbourhood relationships produced by intercalation. Intercalation poses a special problem in epithelia because of the adherens junctions, which maintain the integrity of the tissue. Here we address the mechanism by which an ordered process of cell intercalation directs polarized epithelial morphogenesis during germ-band elongation, the developmental elongation of the Drosophila embryo. Intercalation progresses because junctions are spatially reorganized in the plane of the epithelium following an ordered pattern of disassembly and reassembly. The planar remodelling of junctions is not driven by external forces at the tissue boundaries but depends on local forces at cell boundaries. Myosin II is specifically enriched in disassembling junctions, and its planar polarized localization and activity are required for planar junction remodelling and cell intercalation. This simple cellular mechanism provides a general model for polarized morphogenesis in epithelial organs.Nature 07/2004; 429(6992):667-71. · 38.60 Impact Factor
Mechanical tugging force regulates
the size of cell–cell junctions
Zhijun Liua,1, John L. Tanb,1, Daniel M. Cohena,1, Michael T. Yanga, Nathan J. Sniadeckia,
Sami Alom Ruiza, Celeste M. Nelsonb, and Christopher S. Chena,b,2
aDepartment of Bioengineering, University of Pennsylvania, Philadelphia, PA 19104; and
Johns Hopkins School of Medicine, Baltimore, MD 21205
bDepartment of Biomedical Engineering,
Edited by Shu Chien, University of California, San Diego, La Jolla, CA, and approved March 31, 2010 (received for review December 17, 2009)
Actomyosin contractilityaffects cellularorganizationwithintissues
in part through the generation of mechanical forces at sites of
cell–matrix and cell–cell contact. While increased mechanical load-
ing at cell–matrix adhesions results in focal adhesion growth,
whether forces drive changes in the size of cell–cell adhesions
remains an open question. To investigate the responsiveness of
adherens junctions (AJ) to force, we adapted a system of microfab-
ricated force sensors to quantitatively report cell–cell tugging force
and AJ size. We observed that AJ size was modulated by endothe-
lial cell–cell tugging forces: AJs and tugging force grew or decayed
with myosin activation or inhibition, respectively. Myosin-depen-
dent regulation of AJs operated in concert with a Rac1, and this
coordinated regulation was illustrated by showing that the effects
of vascular permeability agents (S1P, thrombin) on junctional sta-
bility were reversed by changing the extent to which these agents
coupled to the Rac and myosin-dependent pathways. Furthermore,
direct application of mechanical tugging force, rather than myosin
activity per se, was sufficient to trigger AJ growth. These findings
demonstrate that the dynamic coordination of mechanical forces
and cell–cell adhesive interactions likely is critical to the mainte-
nance of multicellular integrity and highlight the need for new
approaches to study tugging forces.
adherens junction ∣ mechanotransduction ∣ myosin ∣ PDMS ∣ traction force
sion in the actin cytoskeleton (1, 2). Studies in model organisms
have revealed that this actomyosin activity is critical for tissue
morphogenesis and dynamic regulation of cell–cell contacts. Dur-
ing gastrulation, apical constriction and an inward displacement
of cell–cell junctions is driven by myosin II activity, allowing cells
of the ventral furrow to invaginate (3, 4). Additionally, myosin II
is essential for stereotypical cell shape changes and redistribution
of cell–cell contacts that accompany tissue extension/expansion
by cell intercalation in both Drosophila (5, 6) and Xenopus mod-
els (7, 8). Maintenance of cell–cell junctions in embryogenesis
also appears to depend on myosin as loss of myosin heavy chain
IIA disrupts E-cadherin and beta-catenin recruitment to cell–cell
junctions and impairs cell–cell adhesion in embryoid bodies (9).
Together, these studies demonstrate a requirement for myosin II
in the organization of cell–cell junctions and imply that myosin-
dependent forces may play a role in regulating cell–cell adhesion.
Among the different types of cell–cell junctions, the adherens
junctions (AJ), in particular, appear sensitive to changes in
myosin-dependent cytoskeletal tension. Inhibition of Rho kinase,
MLCK, myosin ATPase activity or expression interferes with the
growth and maintenance of adherens junctions in several systems
(10–15). E- and N-cadherin cluster into focal adhesion-like struc-
tures when cells are plated on cadherin-coated substrates, and
this clustering requires actomyosin contractility (14, 16, 17).
Yet, unlike the mechanosensitivity of focal adhesions (18), force-
dependent assembly of cadherins cell–cell junctions is not well
established. Nevertheless, focal adhesions and AJs depend on
ells generate contractile forces against their substratum and
on surrounding cells primarily through myosin-generated ten-
the actin cytoskeleton for anchoring to the matrix or neighboring
cells, respectively, and to withstand substantial forces (19).
To test the functional importance of forces at AJs, various
systems have been developed to load cell–cell adhesions using
atomic force microscopy (AFM), pipettes, or magnetically
trapped beads (20–27). Pulling beads bound to cadherins with
∼1–150 pN forces are sufficient to elicit cellular signaling and ac-
tin assembly (20, 24, 27). Force-induced breakage of E-cadherin
bonds, however, requires much stronger forces (∼200 nN) (23).
While these studies describe a wide range of forces that cells can
experience at cell–cell adhesions, these studies fail to capture the
role of endogenous, cell-generated forces experienced at the AJ.
Here we describe a method using microfabricated force
sensors that reports quantitatively on both the force applied to
the cell–cell contact and AJ size. We find that endothelial cells
generate substantial forces (up to ∼120 nN) that pull normal
to the face of the cell–cell contact, designated as the intercellular
tugging force. Furthermore, we demonstrate a robust correlation
between AJ size and tugging force and reveal that both junctional
parameters are directly responsive to myosin activity. This acto-
myosin contractility cooperates with a Rac-mediated pathway to
determine the ultimate size of AJs. Furthermore, acute stimula-
tion with RhoA or the application of exogenously pulling forces
initiates a rapid AJ growth. These findings demonstrate that AJs
undergo mechanically induced growth in response to loading and
suggest that cell-generated forces can influence the architecture
of tissues by regulating the strength of cell–cell adhesions.
A MicrofabricatedForce SensorFor Measuring Cell–Cell Tugging Force.
To measure tugging force, we retooled a previously described ap-
proach that uses an array of elastomeric microneedles to measure
traction forces (28) (Fig.1 A–C andEq. 1 in Fig. 1B).Cells exist in
a quasi-static equilibrium where net force sums to zero (28, 29)
(n.b., an imbalance of 1 nN would accelerate a 10 ng cell at
100 m∕s2). For a cell in contact with a neighbor, the net force
encompasses both traction forces and the intercellular tugging
force experienced at the cell–cell contact (Fig. 1 D–F). Because
the net force remains zero, the intercellular tugging force, Fc, is
equal in magnitude and opposite in direction to the measured net
traction force reported by the microneedle array (Fig 1F and
Eq. 2 in Fig. 1F). This approach relies on the formation of a single
contacting interface between two cells; calculating forces across
multiple cell–cell contacts is mathematically insoluble. Whereas
Author contributions: Z.L., J.L.T., D.M.C., M.T.Y., S.A.R., C.M.N., and C.S.C. designed
research; Z.L., J.L.T., D.M.C., M.T.Y., and S.A.R. performed research; Z.L., J.L.T., D.M.C.,
M.T.Y., and C.M.N. contributed new reagents/analytic tools; Z.L., J.L.T., D.M.C., M.T.Y.,
and S.A.R. analyzed data; and Z.L., D.M.C., N.J.S., and C.S.C. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1Z.L., J.L.T., and D.M.C. contributed equally to this paper.
2To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/lookup/suppl/
9944–9949 ∣ PNAS ∣ June 1, 2010 ∣ vol. 107 ∣ no. 22www.pnas.org/cgi/doi/10.1073/pnas.0914547107
random seeding yields few cells forming a single cell–cell contact,
efficient pair formation was obtained by bowtie micropatterns
(28) (Fig. 1 G and H).
To determine whether cells exert a measurable intercellular
Fc, we assayed forces in endothelial cells seeded as singlets or
in contacting pairs on the microneedle substrates (Fig. 1 I
and J). In contacting cells, but not singlets, a substantial Fc
(∼40 nN), oriented roughly normal to the face of the AJ, was
observed and was sustained over several hours (Fig. 1 J and K).
No difference in average traction force per microneedle was ob-
served between singlet and contacting cells (Fig 1L). To confirm
that the observed Fcreflects contractile forces transmitted to
the AJ, we tested whether decoupling VE-cadherin from the actin
cytoskeleton perturbed Fc. Expression of a mutant VE-cadherin
deficient in β-catenin binding (deletion of residues 703–784)
sing control cells (Fig. 1 M–O), whereas traction forces remained
similar (Fig. 1P). Thus, tugging forces are applied to the AJ,
and the transmission of such forces requires that VE-cadherin be
anchored to the catenin complex.
AJ size (area of β-catenin staining) was heterogeneous in the
bowtie patterns (Fig. 1 Q and R). Interestingly, the size of AJs
and the magnitude of Fcwere linearly correlated, resulting in an
apparent constant stress of ∼1 nN∕μm2(Fig. 1S). Traction force
per microneedle, a measure of global cellular contractility, was
uncorrelated with AJ size, indicating that changes in AJ size
are specific to changes in local tugging force (c.f. Fig. 1S vs. T).
Myosin II-Dependent Tugging Forces Regulate AJ Assembly. These
data suggested the possibility that cell–cell junction assembly,
much like focal adhesion assembly, is a force-dependent process.
To test this hypothesis, we antagonized cytoskeletal tension by
treatment with Y-27632 (a Rho kinase inhibitor) or blebbistatin
(myosin II ATPase inhibitor). Both inhibitors substantially de-
creased average traction force and Fcand caused a reduction
in AJ size, preserving the linear relationship between Fcand
junctions (Fig. 2 A–D). Likewise, siRNA-mediated knockdown
of either myosin IIA or IIB decreased traction force, Fc, and
junction size (Fig. 2 A–C and E).
These data are consistent with a reported requirement for
myosin in maintaining adherens junctions (9, 10, 12–15). How-
ever, whether increasing myosin-mediated tension promotes AJ
assembly is largely untested. We therefore assayed the effects of
the tension-inducing drugs nocodazole and calyculin-A in our sys-
tem. Both agents significantly increased traction force, Fc, and AJ
assembly (Fig. 2 B, C, and F), supporting a role for tension-in-
duced AJ assembly. To achieve direct molecular targeting of myo-
sin, we infected cells with lentivirally encoded phosphomimetic
mutants (T18D, S19D) of myosin regulatory light chain (MLC).
Phosphomimetic MLC, but not wild-type MLC, dramatically en-
hanced average traction force and Fcand induced more expan-
sive AJs (Fig. 2 B,C, and G).
Remarkably, compiling all of our studies with tension manip-
ulations revealed a general relationship between AJ size and Fc
that spanned a wide range of AJ areas (∼7–87 μm2) and Fc
(∼7–120 nN) (Fig. 2H). Using moving average and bootstrap
methods to calculate the mean and 99% percent confidence
interval, one can appreciate a linear relationship between AJs
and Fcin the smaller force/AJ regime, whereas the response
of AJ begins to saturate at forces higher than ∼70 nN, suggesting
an upper limit for mechanosensitive growth of AJs.
To examine whether this force–AJ relationship can be general-
ized, we characterized the response of AJs in conventional
monolayers to tension manipulations. Consistent with the bowtie
system, treatment of monolayers with tension antagonists re-
duced AJ levels as compared to unstimulated cells (Fig. 2 I–K).
Conversely, upregulation of tension increased the size of AJs in
the monolayers (Fig. 2 I,J, and L). Thus, tugging forces appear to
contractile forces that strongly deflect underlying microneedles (phase
contrast image; scale bar, 50 μm). (B and C) The vector sum of individual
traction forces, Tn, (Red Arrows) exerted by a cell is zero (Eq. 1) (28, 29).
(D and E) For a pair of contacting cells, the net force encompasses both
traction forces (Red Arrows) and the intercellular tugging force, Fc, (Blue
Arrows). (F) The intercellular tugging force is equal in magnitude and
opposite in direction to the measured net traction force reported on the
microneedle array (Eq. 2). Vector Fcplotted over cell A is defined as the
net tugging force that cell A is exerting on cell B at the cell–cell contact.
Cell B is expected to pull on cell A with an equal and opposite force. (G and
H) Cells were constrained to a bowtie pattern using microcontact printing
of fibronectin (Cyan) (G) and formation of adherens junctions in green
(anti-β-catenin, H). Microneedles and cell nuclei were counterstained with
DiI (Blue) and DAPI (Red), respectively. Scale bar, 10 μm. (I and J) Represen-
tative force vector plots of single (I) and paired (J) cells. Force vectors are
plotted as arrows showing the direction of force and magnitude (arrow
length; reference arrow for 10 nN shown above scale bar). Individual trac-
tion forces are shown in red and tugging force is shown in white. Small,
force imbalances are observed due to uncertainties in microneedle deflec-
tion measurements. The error in the vector sum of individual traction forces
equals the error of the individual forces multiplied by the square root of
the number of tractions. The microneedles typically have an error of ±3 nN,
and cells span on average 15 microneedles, leading to an expected noise
level of 12 nN in the determination of Fc. (K) Average Fcexperienced in
bowtie pairs of HPAECs. * p < 0.05 indicates comparison against single cells.
(L) Average traction force per microneedle was similar in singlet or contact-
ing cells. (M–O) Tugging forces require coupling between VE-cadherin and
the actin cytoskeleton. Cells expressing adeno-VE-Δ, a truncated VE-cadher-
in mutant lacking the β-catenin binding domain (residues 703–784), form
adherens junctions (anti-β-catenin, Green, N) but show greatly diminished
tugging force compared to adeno-GFP infected cells (M and O). * p < 0.05
indicates comparison against GFP. (P) Average traction force per micronee-
dle was unaffected by VE-Δ expression. (Q and R) AJ (anti-β-catenin) in
HPAECs are heterogeneous in size. Tugging force is shown in yellow. (S)
AJ size is linearly correlated with Fc, with a correlation coefficient
R2≈ 0.9. Each data point represents one pair of cells. (T) AJ size is not
correlated with the average traction force per microneedle. Error bars
on all graphs denote standard error of the mean.
An approach to measure tugging forces (A) Endothelial cells exert
Liu et al. PNAS
June 1, 2010
regulate junction size, regardless of experimental culture condi-
tions (bowties versus monolayers).
Vasoactive Compounds Regulate AJ Size Through Rho/Force- and
Rac-Dependent Pathways. While the aforementioned tension ma-
nipulations reveal an important linkage between myosin activity,
tugging forces, and regulation of AJ size, these treatments lack a
physiological context. We hypothesized that physiological stimuli
may also control cell–cell junction homeostasis in endothelium
through tugging forces. Inflammatory agents such as thrombin
have been proposed to disrupt cell–cell contacts through in-
creased RhoA-mediated myosin-generated contraction (31),
which represents a counterexample to the stimulatory effects
of myosin contractility on AJs observed here (Fig. 2). Sphingo-
sine-1-phosphate (S1P), another vasoactive compound, appears
to strengthen cell–cell contact (32) at a concentration that does
not impact cellular contractility (33), hinting at the possibility of a
force-independent modulation of junction size.
Application of thrombin or S1P to monolayers induced the ex-
pected disruption or enhancement of AJ assembly, respectively
(Fig. 3 A and B). To clarify how these permeability agents impact
AJs, we assayed AJs and Fcin the bowtie-microneedle system.
Traction force and Fcincreased in response to thrombin without
a concomitant increase in junction size (Fig. 3 C–E), resulting in a
precipitous rise in mechanical stress at the junction (from ∼1 to
∼8 nN∕μm2). In fact, cell–cell contacts were either disrupted by
thrombin or resulted in residual AJs that were shorter, fainter,
and less distinct relative to controls (Fig. 3B). In contrast, S1P
triggered AJ expansion in a tortuous path across the cell–cell
contact, without affecting Fc(Fig. 3 D and E), and thus reduced
junctional stress (to ∼0.5 nN∕μm2).
The uncoupling of junction size and Fcin S1P and thrombin-
treated cells implies the existence of a mechanism of AJ growth
that is independent of Fcor one that modulates the coupling of
AJ growth to Fc. One such mechanism could be the regulation of
AJs by the Rac1 GTPase. Activated Rac1 is localized to sites of
nascent cell–cell contact (15) and is implicated in regulating the
assembly kinetics and strength of AJs (34). Importantly, Rac1 is
activated in response to S1P (33, 35), and this activation is re-
quired for S1P-mediated junctional assembly (35). We therefore
investigated the relationship of Rac1 activity to AJ size and tug-
ging forces in the presence of vasoactive compounds. Measure-
ments of Rac and RhoA activity (an upstream mediator for
myosin/tension driven responses) revealed that thrombin robustly
activated RhoA, while suppressing Rac activity (Fig. 3F). In con-
trast, S1P specifically upregulated Rac activity, while leaving
RhoA unperturbed (Fig. 3F). These data suggested a direct cor-
relation between junction size and Rac activity. Rac activity can
be antagonized using NSC23766, which prevents the interaction
of Rac1 and its upstream activators Tiam1 and Trio (36). To test
whether Rac activity was required for S1P-mediated junctional
assembly, HPAECs were challenged with S1P in the presence
of NSC23766 (10 μM). NSC23766 abolished S1P-induced AJ
assembly (Fig. 3 H and I), demonstrating that Rac activation
contributes to junctional growth.
We then tested whether changes in Rho and Rac activity
also contributed to thrombin-induced junctional disassembly.
Pretreatment with Y27632 inhibited the thrombin-induced in-
crease in traction forces and Fcand prevented the disruption
of junctions (Fig. 3 G–I). Remarkably, expression of a constitu-
tively active Rac mutant (RacV12) (37) prevented thrombin-
induced disassembly (Fig. 3 G–I). These finding suggest that
Rac activity and tugging force are both important determinants
in the response of AJs to vasoactive compounds.
Rac Activity and Tugging Force Cooperate to Regulate AJ Size. To ex-
amine whether Rac activity indeed cooperates with Fcto regulate
AJ size, we examined the effect of jointly manipulating tension
and Rac activity. RacV12 alone induced junction assembly with-
out altering Fc, suggesting AJ growth does not require an increase
in basal Fc. However, treatment of RacV12 cells with blebbistatin
partially inhibited AJ assembly, demonstrating that Rac-induced
junctions remained dependent on baseline tugging forces
(Fig. 3 J–L). Conversely, with the Rac antagonist, NSC23766,
attenuated the ability of phosphomimetic myosin to promote
sion antagonists and agonists modulate average traction force per micronee-
dle and Fc. The following antagonists were used: blebbistatin (30 μM, Blebbi),
Y27632 (25 μM, Y27), siRNA against myosin IIA (siMyoIIA), myosin IIB
(siMyoIIB). Tension agonists included nocodazole (1 μM, Noc) calyculin-A
(1nM, Calyc), lentivirally encoding phosphomimetic light chain (ppMLC).
Control conditions included vehicle control (Ctrl), control siRNA (siCyclophi-
lin), no treatment (N/T), or lentiviral wild-type MLC (wtMLC). * p < 0.05 indi-
cates comparison against vehicle control, # p < 0.05 indicates comparison
against control siRNA, and Δ p < 0.05 indicates comparison against wtMLC.
(C) AJ size is regulated by changes in actomyosin-mediated tension. Bowtie
pairs exposed to the conditions in (A and B) and stained for β-catenin. (D–G)
Relationship between AJ size and Fc. Inhibition of Fcby blebbistatin, Y27632
(D) or treatment with siRNA against myosin II A or B (E) leads to reduced
junction size, whereas elevated Fcfollowing stimulation with nocodazole,
calyculin-A (F), or infected with lentivirus encoding phosphomimetic mutants
(G) increased junction size. AJ size and Fcmeasurements (at least 10 bowtie
pairs per condition) were clustered and plotted as an elliptical fit (D–G).
Statistical significance between conditions is reported in Table S1. (H) A scat-
ter plot showing the correlation between AJ size and Fcacross all conditions.
(I–L) Tension manipulations affect AJs in monolayers. Immunofluorescence
images showing changes in AJs (anti- β -catenin) in HPAECs monolayers
exposed to tension-manipulating conditions (I). Quantification of monolayer
junctional response (J–L). * p < 0.05 indicates comparison against control
monolayers: Ctrl (J), siCyclophilin (K), or wtMLC (L). All scale bars are
10 μm. Error bars denote standard error of the mean.
Tugging force regulates the size of adherens junctions. (A and B) Ten-
www.pnas.org/cgi/doi/10.1073/pnas.0914547107 Liu et al.
junction growth (Fig 3 J,K, and M) further supporting the hypoth-
esis that Rac activity is permissive for force-induced junction
growth. Again, similar effects were observed in monolayers as
in our bowtie system, suggesting that these relationships between
Rho, Rac, and junction assembly are conserved in more general
settings (Fig. 3 N and O). Together, these studies suggest a model
whereby Rho-mediated myosin activity generates tugging forces,
and Rac mediates the ability of AJs to assemble in response to
those forces (Fig. 3P).
Tugging Force Is Necessary and Sufficient for Rho-dependent AJ As-
sembly. While the aforementioned studies argue that tugging
forces cause AJ assembly, the reliance on myosin manipulations
cannot preclude indirect effects of these manipulations on corti-
cal actin assembly. Moreover, single end-point studies do not pro-
vide any insights into the temporal coupling of these phenomena.
To address these shortcomings, we characterized the dynamic
response of AJs to the acute stimulation of contractility in one
cell of a bowtie pair. Microinjection of the activated RhoA-
Q63L produced an immediate and robust increase in both trac-
tion forces (the noninjected cell shows a delayed response) and Fc
(Fig. 4 B and D), whereas no effect was observed upon microin-
jection of wild-type RhoA (Fig. 4 A and C). Importantly, RhoA-
Q63L triggered AJ growth, as revealed by GFP-VE-cadherin
localization, within minutes of rising Fc(Movies S1 and S2 and
Fig. 4 E and F), in a tension-dependent manner (Movies S1 and
S2 and Fig. 4F). To confirm whether mechanical force is truly
onto one of the cells within the bowtie pair and pulled to apply an
ical loading of the cell–cell junction also promoted VE-cadherin
assembly within several minutes oftugging (Movies S3 andS4 and
Fig. 4GandH).These data provide directevidence that adherens
junction assembly responds to cell–cell forces.
Vasoactive compounds S1P and thrombin coordinately upregulate and
downregulate adherens junctions. (A) Monolayers exposed to vehicle
only (Ctrl), thrombin (0.1 μM, Thrb), or S1P (1 μM, S1P) show S1P-induced
AJ growth and thrombin-induced AJ disassembly (anti-β-catenin stain-
ing). (B) Quantification of changes in AJs in monolayer cultures shown
in A. * p < 0.05 indicates comparison against control. (C) Thrombin,
but not S1P, induces cellular tension. Bar graph showing average traction
force per microneedle reported in bowtie pairs stimulated with vasoac-
tive compounds. (D and E) Effects of thrombin and S1P on junctional size-
tugging force relationship. S1P and thrombin alter AJ formation (anti-β-
catenin) in bowtie pairs (D). Relationship between AJ and Fcdata plotted
asanelliptical fit(E).Thetrendbandandcontrolcondition fromFig.2are
replotted here for reference. (F) RhoA and Rac1 activity assay for Throm-
bin and S1P using Rho G-LISA and Rac G-LISA kits, respectively. *, #
p < 0.05indicates comparisonagainst baseline RhoA (*)orRac(#)activity.
(G) Manipulations of Rac activity do not alter cellular tension, as assayed
by average traction force per microneedle. In contrast, inhibition of Rho
kinase (Y27632, 25 μM) blocks thrombin-induced tension. * p < 0.05 indi-
cates comparison against control. (H and I) Altering the coupling of S1P
of these vasoactive compounds on junctions. S1P-induced AJ growth is
inhibited by treatment with NSC23766 (10 μM), whereas thrombin-in-
duced AJ disassembly is blocked by expression of constitutively active
Rac (RacV12). AJ junctions visualized by β-catenin localization (H). Decou-
pling thrombin from cellular tension or forcibly linking it to constitutively
active Rac restores the linear relationship between AJ and tugging force.
Similarly, inhibition of Rac1 with NSC23766 restores the normal AJ-tug-
ging force balance in S1P-stimulated cells. (J) Myosin-manipulation reg-
ulates cellular tension independently of Rac activity levels. Blebbistatin
(30 μM) downregulates average traction force per microneedle even in
the presence of RacV12, whereas phosphomimetic myosin upregulates
average traction force, even in the context of reduced Rac activity
(NSC23766 treatment). * p < 0.05 indicates comparison against RacV12;
and # p < 0.05 indicates comparison against NSC23766. (K– M) Interde-
pendence of Rac-induced and tension-induced AJ assembly. Inhibition
of myosin activity (blebbistatin, 30 μM) reduced the growth of AJs stimu-
growth of AJs stimulated by phosphomimetic myosin. AJs visualized by β-
catenin localization. (L and M) Effects of double Rac/myosin manipula-
tions in J on AJ-Tcrelationship. (N and O) Rac and tension-dependent
pathways control the size of AJs in monolayer culture. Representative
immunofluorescent micrographs of monolayers of cells exposed to the
treatments in (G–M). AJs are visualized by β-catenin staining (N). Quanti-
fication of changes in junction size in monolayers (O). * p < 0.05 indicates
comparison against Ctrl; # p < 0.05 indicates comparison against RacV12;
and Δ p < 0.05 indicates comparison against NSC23766. (P) Schematic of
mine the ultimate size and stability of AJs. We suggest three general re-
gimes: a low stress regime (Shaded Green) wherein high Rac activity
support AJs assembly independently of tugging force, a moderate stress
regime (White Region) wherein Rac activity and myosin-mediated
tugging force coordinate to mediate mechanosensitive growth of AJs,
and a high stress regime (Shaded Red) wherein high tugging forces/myo-
sin activity initiate breaking/disassembly of junctions due to insufficient
Rac-mediated junction assembly. All scale bars are 10 μm. Error bars on all
graphs denote standard error of the mean. Statistical significance
between conditions E, I, L, and M is reported in Table S1.
Modulation of force-AJ relationship by soluble factors. (A and B)
Liu et al. PNAS
June 1, 2010
Tugging forces are critical in morphogenesis, mechanical integ-
rity,andin generating gradientsofmechanical stressesto regulate
patterns of cell function (38–40). Prior research has emphasized
the role of externally applied forces necessary in breaking or re-
modeling cell–cell adhesions (23, 24, 27). Herein we demonstrate
an approach to directly measure endogenous stresses generated
between cells and illustrated that such tugging forces induce junc-
tion growth. Tugging force measurements were facilitated by bow-
tie patterns to constrain cells to a single contacting interface.
Because cell shape and cell–cell contact can alter cytoskeletal or-
ganization (41,42),resultsusingthe bowtieconfiguration maynot
extend to other settings. Yet many of the observed responses of
tugging forces and AJ size in bowties appear to be similar in
monolayers. Nevertheless, further characterization of tugging
forces in different settings remains an important challenge.
Interestingly, AJs adjust dynamically in response to changes in
tugging forces. While a requirement for myosin II in maintenance
of basal AJ size is well established (10–15), here we show that
activated myosin as well as direct application of exogenous tug-
ging forces induces AJ assembly that occurs within minutes of
force application. Our data suggest that mechanosensitive growth
is a fundamental mechanism for controlling AJ size and is strik-
ingly similar to what occurs at focal adhesions (18). As postulated
for focal adhesions, the physiologic consequences of this force-
dependent AJ growth are likely twofold. First, force-induced as-
sembly dissipates mechanical stress and may protect the junction
against breakage (43). Second, AJs are molecular signaling hubs
(44), and force-induced junction assembly may trigger signaling
that contributes to the process of mechanotransduction.
Although the mechanism of tugging-force-induced AJ growth
remains unknown, several models can be considered. In the “zip-
per” model, localized myosin activation at the expanding edges of
cell–cell contacts increases the surface area of cell–cell contact,
thus increasing the likelihood of transcadherin interactions (15).
An alternative push-pull model suggests that nascent cadherin
interactions develop at filopodial-like protrusions along the
cell–cell contact (45), and these filopodial-like protrusions are
stabilized by the application of normal tugging forces (46). In
either model, junction growth is predicated on membrane protru-
sive activity, a Rac-dependent process (47).
In this context, it is especially interesting that we observe that
force-dependent growth of AJs depends, in part, on Rac1 activity.
Rac1 may play a permissive role in force-dependent junction
growth by supporting the membrane protrusion activity impli-
cated in the aforementioned zipper or push-pull models. Alter-
natively, Rac1 may play a direct role in force-mediated actin
polymerization (16, 22, 48), cytoskeletal coupling of VE-cadherin
(16, 25, 48), or regulation of catenin function (49). Attenuation of
these Rac1-dependent processes may abrogate intracellular sig-
naling required for force-induced junction assembly. Interest-
ingly, we have also implicated the downregulation of Rac1 in
force-dependent junctional disassembly induced by thrombin.
Coordinate upregulation of tugging force (ppMLC) and downre-
gulation of Rac (NSC23766) did not, however, trigger AJ disas-
sembly. While thrombin may suppress Rac more potently than
NSC23766, thrombin-mediated disassembly may require addi-
tional biochemical events such as thrombin-induced phosphoryla-
tion of catenins (50), in addition to reciprocal up/downregulation
of tugging force and Rac1 activity. Regardless, our findings
ment of traction force in response to microinjection of constitutively active Rho protein (RhoA-Q63L) (B) but not wild-type Rho protein (wt) (A) for each of the
two cells in bowtie pairs. Dotted arrow indicates the time point of microinjection, and only Cell I (Purple Line) was microinjected. Whiskers show the range of
traction force observed in two independent bowtie movies. (C and D) Microinjection of RhoA-Q63L (D) but not wild-type Rho protein (C), acutely boosts
tugging forces for each of the two cells in bowtie pairs. Dotted arrow indicates the time point of microinjection. Whiskers show the range of two microinjected
bowtie pairs across all time points. (E) Representative image frames showing assembly of GFP-VE-cadherin following microinjection of RhoA-Q63L, 10 minutes
before and 16 minutes after microinjection, respectively. Dotted white lines indicate the location of the pipette and the red arrowheads point to locations of
increased AJ size. Scale bar is 10 μm. (F) RhoA-Q63L induced AJ assembly in a tension-dependent manner. The RhoA-microinjected pairs (Orange Squares)
showed an increase in junction size while pretreatment with Y27632 (25 μM) blocked this effect (Blue Squares). Dotted arrow indicates the time point of
microinjection. Whiskers show the range of two replicated experiments across all time points. (G) Exogenous tugging force stimulated VE-cadherin assembly.
Fluorescence micrograph of GFP-VE-cadherin distribution prior to micropipet-mediated mechanically tugging of the cell–cell junction. Dotted white lines
indicate the location of the pipette and the white arrow points to the direction of pipette movement. Region of interest (Green Box) shows distribution
of junctional VE-cadherin before (0 min) and after tugging (4 min), pseudocolored green and red, respectively in the overlay image. Yellow boxes denote
the regions of the AJ showing the strongest cadherin recruitment. Scale bar is 10 μm. (H) Quantification of AJ growth in response to mechanical tugging. AJ size
of bowtie pairs over 10 min intervals plotting for baseline (Green Line), with pipette on top of the cell but no tugging (Blue Line), or before and after tugging
one cell of the bowtie (Red). Black arrow indicates the time point of tugging with the micropipet. Whiskers show the standard error of the mean across six
Microinjection of constitutively active RhoA and direct mechanical tugging leads to junction assembly. (A and B) Time-series plots showing enhance-
www.pnas.org/cgi/doi/10.1073/pnas.0914547107 Liu et al.
mechanical forces, and junctional structure.
Not only do tugging forces regulate cell–cell adhesions, but
cell–cell adhesions can, in turn, trigger many signals including
the Rho GTPases to regulate cellular mechanics (41). This inter-
play between force and adhesion highlights how channeling of
actomyosin-generated traction and tugging forces at the cell–
matrix and cell–cell interface provide a means to dynamically re-
organize cell–matrix and cell–cell adhesions locally, as well as the
architecture of tissues globally, and further support coordinated
crosstalk between the two mechanical interfaces of the cell with
its external environment. Understanding the balance between
forces across each of these two interfaces ultimately may help
to explain multicellular reorganizations such as occur during
tissue morphogenesis and demonstrates the need for better
approaches to describe these forces.
Materials and Methods
The experimental methods and protocols are described briefly; a detailed
description is included in SI Text. PDMS microneedle array substrates were
fabricated as in Tan et al. (28). Bowtie patterns had a total area of
1600 μm2. Microneedle deflections were calculated using Matlab to deter-
mine the displacement of the centroid of the microneedle at the tip (see
SI Text). Forces were then computed via a known spring constant of the
microneedle, (32 nN∕μm). AJ size was quantified based on thresholded
anti-β-catenin fluorescence images and validated against a more rigorous
volumetric analysis (Fig. S1). Mean pixel intensity of staining was measured,
to confirm that increases in AJ area did not occur at the expense of intensity
of staining (Fig. S1). Ellipse-fitting was performed with a least-squares ellipse
algorithm in Matlab (51). The trend band of junction size at different tugging
force levels is plotted as a moving average (Darker Gray Line) ?99% Confi-
dence Interval (CI, Lighter Gray Region) using Matlab.
ACKNOWLEDGMENTS. GFP-VE-cadherin and RacV12 adenovirus were gener-
ous gifts of S. Shaw/F. Luscinskas and A. Ridley respectively; K. G. Birukov,
R. A. Desai, J. G. N. Garcia, D. M. Pirone, A. Popel, D. Reich, and L. H. Romer
provided helpful discussions. This work was supported in part by grants from
the National Institutes of Health (EB00262, EB08396, GM74048, HL73305,
HL90747), the Material Research Science and Engineering Center, and the
Center for Engineering Cells and Regeneration of the University of Pennsyl-
vania. J.L.T. and C.M.N received support from the Whitaker Foundation. Ruth
L. Kirschstein National Research Service Awards supported D.M.C. and N.J.S.
1. Burridge K, Chrzanowska-Wodnicka M (1996) Focal adhesions, contractility, and
signaling. Annu Rev Cell Dev Biol 12:463–518.
2. Mege RM, Gavard J, Lambert M (2006) Regulation of cell–cell junctions by the
cytoskeleton. Curr Opin Cell Biol 18(5):541–548.
3. Dawes-Hoang RE, et al. (2005) folded gastrulation, cell shape change and the control
of myosin localization. Development 132(18):4165–4178.
4. Martin AC, Kaschube M, Wieschaus EF (2009) Pulsed contractions of an actin-myosin
network drive apical constriction. Nature 457(7228):495–499.
5. Bertet C, Sulak L, Lecuit T (2004) Myosin-dependent junction remodelling controls
planar cell intercalation and axis elongation. Nature 429(6992):667–671.
6. Rauzi M, Verant P, Lecuit T, Lenne PF (2008) Nature and anisotropy of cortical forces
orienting Drosophila tissue morphogenesis. Nat Cell Biol 10(12):1401–1410.
7. Rolo A, Skoglund P, Keller R (2009) Morphogenetic movements driving neural tube
closure in Xenopus require myosin IIB. Dev Biol 327(2):327–338.
8. Skoglund P, Rolo A, Chen X, Gumbiner BM, Keller R (2008) Convergence and extension
at gastrulation require a myosin IIB-dependent cortical actin network. Development
9. Conti MA, Even-Ram S, Liu C, Yamada KM, Adelstein RS (2004) Defects in cell
adhesion and the visceral endoderm following ablation of nonmuscle myosin heavy
chain II-A in mice. J Biol Chem 279(40):41263–41266.
10. Abraham S, et al. (2009) VE-Cadherin-mediated cell–cell interaction suppresses
sprouting via signaling to MLC2 phosphorylation. Curr Biol 19(8):668–674.
11. de Rooij J, Kerstens A, Danuser G, Schwartz MA, Waterman-Storer CM (2005) Integrin-
dependent actomyosin contraction regulates epithelial cell scattering. J Cell Biol
12. Ivanov AI, et al. (2007) A unique role for nonmuscle myosin heavy chain IIA in regula-
tion of epithelial apical junctions. PLoS One 2(7):e658.
13. Miyake Y, et al. (2006) Actomyosin tension is required for correct recruitment of
adherens junction components and zonula occludens formation. Exp Cell Res
14. Shewan AM, et al. (2005) Myosin 2 is a key Rho kinase target necessary for the local
concentration of E-cadherin at cell–cell contacts. Mol Biol Cell 16(10):4531–4542.
15. Yamada S, Nelson WJ (2007) Localized zones of Rho and Rac activities drive initiation
and expansion of epithelial cell–cell adhesion. J Cell Biol 178(3):517–527.
16. Gavard J, et al. (2004) Lamellipodium extension and cadherin adhesion: Two cell
responses to cadherin activation relying on distinct signalling pathways. J Cell Sci
17. Lambert M, et al. (2007) Nucleation and growth of cadherin adhesions. Exp Cell Res
18. Bershadsky AD, Balaban NQ, Geiger B (2003) Adhesion-dependent cell mechanosen-
sitivity. Annu Rev Cell Dev Biol 19:677–695.
19. Chen CS, Tan J, Tien J (2004) Mechanotransduction at cell–matrix and cell–cell contacts.
Annu Rev Biomed Eng 6:275–302.
20. Bard L, et al. (2008) A molecular clutch between the actin flow and N-cadherin
adhesions drives growth cone migration. J Neurosci 28(23):5879–5890.
21. Baumgartner W, et al. (2000) Cadherin interaction probed by atomic force microscopy.
Proc Natl Acad Sci USA 97(8):4005–4010.
22. Baumgartner W, Schutz GJ, Wiegand J, Golenhofen N, Drenckhahn D (2003)
Cadherin function probed by laser tweezer and single molecule fluorescence in
vascular endothelial cells. J Cell Sci 116(Pt 6):1001–1011.
23. Chu YS, et al. (2004) Force measurements in E-cadherin-mediated cell doublets
reveal rapid adhesion strengthened by actin cytoskeleton remodeling through Rac
and Cdc42. J Cell Biol 167(6):1183–1194.
24. Ko KS, Arora PD, McCulloch CA (2001) Cadherins mediate intercellular mechanical
signaling in fibroblasts by activation of stretch-sensitive calcium-permeable channels.
J Biol Chem 276(38):35967–35977.
25. Kris AS, Kamm RD, Sieminski AL (2008) VASP involvement in force-mediated adherens
junction strengthening. Biochem Biophys Res Commun 375(1):134–138.
26. Panorchan P, et al. (2006) Single-molecule analysis of cadherin-mediated cell–cell
adhesion. J Cell Sci 119(Pt 1):66–74.
27. PotardUS, ButlerJP, WangN (1997) Cytoskeletal mechanics in confluent epithelial cells
probed through integrins and E-cadherins. Am J Physiol 272(5 Pt 1):C1654–1663.
28. Tan JL, et al. (2003) Cells lying on a bed of microneedles: An approach to isolate
mechanical force. Proc Natl Acad Sci USA 100(4):1484–1489.
29. Balaban NQ, et al. (2001) Force and focal adhesion assembly: A close relationship
studied using elastic micropatterned substrates. Nat Cell Biol 3(5):466–472.
30. Navarro P, et al. (1995) Catenin-dependent and -independent functions of vascular
endothelial cadherin. J Biol Chem 270(52):30965–30972.
31. Dudek SM, Garcia JG (2001) Cytoskeletal regulation of pulmonary vascular perme-
ability. J Appl Physiol 91(4):1487–1500.
32. McVerry BJ, Garcia JG (2004) Endothelial cell barrier regulation by sphingosine 1-phos-
phate. J Cell Biochem 92(6):1075–1085.
33. Shikata Y, Birukov KG, Garcia JG (2003) S1P induces FA remodeling in human
pulmonary endothelial cells: Role of Rac, GIT1, FAK, and paxillin. J Appl Physiol
34. Ehrlich JS, Hansen MD, Nelson WJ (2002) Spatio-temporal regulation of Rac1
localization and lamellipodia dynamics during epithelial cell–cell adhesion. Dev Cell
35. Mehta D, Konstantoulaki M, Ahmmed GU, Malik AB (2005) Sphingosine 1-phosphate-
induced mobilization of intracellular Ca2þ mediates rac activation and adherens
junction assembly in endothelial cells. J Biol Chem 280(17):17320–17328.
36. Gao Y, Dickerson JB, Guo F, Zheng J, Zheng Y (2004) Rational design and
characterization of a Rac GTPase-specific small molecule inhibitor. Proc Natl Acad
Sci USA 101(20):7618–7623.
37. Ridley AJ, Paterson HF, Johnston CL, Diekmann D, Hall A (1992) The small GTP-binding
protein rac regulates growth factor-induced membrane ruffling. Cell 70(3):401–410.
38. Nelson CM, et al. (2005) Emergent patterns of growth controlled by multicellular form
and mechanics. Proc Natl Acad Sci USA 102(33):11594–11599.
39. Scott IC, Stainier DY (2003) Developmental biology: Twisting the body into shape.
40. Ingber DE (2003) Tensegrity I. Cell structure and hierarchical systems biology. J Cell Sci
41. Nelson CM, Pirone DM, Tan JL, Chen CS (2004) Vascular endothelial-cadherin regulates
cytoskeletal tension, cell spreading, and focal adhesions by stimulating RhoA. Mol Biol
42. Parker KK, et al. (2002) Directional control of lamellipodia extension by constraining
cell shape and orienting cell tractional forces. FASEB J 16(10):1195–1204.
43. Vestweber D (2000) Molecular mechanisms that control endothelial cell contacts.
J Pathol 190(3):281–291.
44. Erez N, Bershadsky A, Geiger B (2005) Signaling from adherens-type junctions. Eur J
Cell Biol 84(2–3):235–244.
45. Vasioukhin V, Bauer C, Yin M, Fuchs E (2000) Directed actin polymerization is the
driving force for epithelial cell–cell adhesion. Cell 100(2):209–219.
46. Brevier J, Montero D, Svitkina T, Riveline D (2008) The asymmetric self-assembly
mechanism of adherens junctions: A cellular push-pull unit. Phys Biol 5(1):016005.
47. Hall A (2005) Rho GTPases and the control of cell behaviour. Biochem Soc Trans
48. Waschke J, et al. (2004) Requirement of Rac activity for maintenance of capillary
endothelial barrier properties. Am J Physiol-Heart C 286(1):H394–401.
49. KaibuchiK, KurodaS,Fukata M,NakagawaM (1999)Regulationof cadherin-mediated
cell–cell adhesion by the Rho family GTPases. Curr Opin Cell Biol 11(5):591–596.
50. Ukropec JA, Hollinger MK, Salva SM, Woolkalis MJ (2000) SHP2 association with
VE-cadherin complexes in human endothelial cells is regulated by thrombin. J Biol
51. Halir R, Flusser J (1998) Numerically stable direct least squares fitting of ellipses.
Proceedings of the 6th International Conference in Central Europe on Computer
Graphics and Visualization (WSCG, Plzen, Czech Republic), pp 125–132.
Liu et al.PNAS
June 1, 2010