(The FASEB Journal. 1999;13:S3-S15.)
© 1999 FASEB
How cells (might) sense microgravity
DONALD INGBER1
Departments of Pathology & Surgery, Children's Hospital and Harvard Medical School, Boston, Massachusetts 02115, USA
1Correspondence: Enders 1007-Surgical Research, Children's Hospital, 300 Longwood Ave., Boston, MA 02115. E-mail: ingber{at}a1.tch.harvard.edu
 |
ABSTRACT
|
|---|
This article is a summary of a lecture presented at an ESA/NASA Workshop
on Cell and Molecular Biology Research in Space that
convened in Leuven, Belgium, in June 1998. Recent studies are reviewed
which suggest that cells may sense mechanical stresses, including those
due to gravity, through changes in the balance of forces that are
transmitted across transmembrane adhesion receptors that link the
cytoskeleton to the extracellular matrix and to other cells (e.g.,
integrins, cadherins, selectins). The mechanism by which these
mechanical signals are transduced and converted into a biochemical
response appears to be based, in part, on the finding that living cells
use a tension-dependent form of architecture, known as tensegrity, to
organize and stabilize their cytoskeleton. Because of tensegrity, the
cellular response to stress differs depending on the level of
pre-stress (pre-existing tension) in the cytoskeleton and it involves
all three cytoskeletal filament systems as well as nuclear scaffolds.
Recent studies confirm that alterations in the cellular force balance
can influence intracellular biochemistry within focal adhesion
complexes that form at the site of integrin binding as well as gene
expression in the nucleus. These results suggest that gravity sensation
may not result from direct activation of any single gravioreceptor
molecule. Instead, gravitational forces may be experienced by
individual cells in the living organism as a result of stress-dependent
changes in cell, tissue, or organ structure that, in turn, alter
extracellular matrix mechanics, cell shape, cytoskeletal organization,
or internal pre-stress in the cell-tissue matrix.—Ingber, D. How cells
(might) sense microgravity.
Key Words: mechanotransduction • cytoskeleton • tensegrity • integrins • cell shape
|
INTRODUCTION
|
|---|
THIS PAPER IS BASED on an invited lecture I presented
at the Belgium National Academy of Sciences in the opening symposium of
the European Space Agency(ESA)2/National Aeronautics and Space Administration (NASA) Workshop on
Cell and Molecular Biology Research in Space, which convened
in Leuven, Belgium in June 1998. The organizers chose the title of my
lecture: How cells (MIGHT) sense microgravity. The word,
"might," was probably added as an afterthought by a wise meeting
organizer. This was a prudent choice because although it is clear that
physical forces, such as those due to gravity, are fundamental
regulators of tissue development, little is known about how living
cells sense these signals and convert them into a biochemical response.
This transduction process, which is at the core of gravity sensation,
is known as mechanotransduction; and this is what I will focus on
today.
Past work on mechanotransduction has revealed that certain cells have
evolved specialized crystal structures that respond directly to the
force of gravity. These dense crystals are called statoliths, literally
"standing stones," or otoliths, as in the case of the sensory cells
of the inner ear. When we move our heads, these dense crystals slide
over the receptor cells like tiny lead weights, and it is the resulting
localized distortion of the cell surface and interconnected
cytoskeleton (CSK) that is somehow sensed by the cell. The statolith
represents an elegant mechanism for mechanotransduction, however, it
does not explain how all of the cells in the body sense gravity.
One of the most common changes observed in astronauts who undergo
long-term spaceflight is bone resorption. In fact, it has been known
for over a century that bone matrix is deposited in distinct patterns
that precisely map out engineering lines of tension and compression for
any structure of that size and shape under similar loading conditions
(1)
. If the loading pattern is altered or an astronaut is
placed in microgravity, the bone immediately remodels. We now know that
living cells within bone (osteoblasts, osteoclasts) are responsible for
this remodeling. This means that individual cells must be able to sense
changes in physical forces in their local environment that are caused
by gravity and that they respond in the most efficient manner possible:
by putting new matrix where it is needed and removing it from where it
is not. However, no one has ever found statoliths in bone cells or in
cells within other tissues that are equally sensitive to perturbation
by mechanical stresses (e.g., muscle, skin, blood vessels, lung,
cartilage, nerve, etc.). In fact, sensitivity to mechanical forces now
seems to be a general property of all living tissues and all cultured
cells (2-6)
.
To understand how gravitational forces alter cell function, we must
place this form of developmental control in context of what we have
learned in recent years about other forms of cell regulation.
Understanding mechanotransduction in the context of cell and tissue
morphogenesis has been the major focus of my laboratory for the past 20
years. When I first entered the morphogenesis field, all of the
attention was on the identification and isolation of regulatory
factors. This led to the discovery of the importance of soluble growth
factors and insoluble extracellular matrix (ECM) molecules, in addition
to mechanical forces, as critical regulators of tissue growth and
development. However, in recent years, the focus has shifted from
discovery of mitogens and morphogens to analysis of the mechanism by
which extracellular stimuli elicit an intracellular response, or what
is known as "signal transduction."
Over the past few years I have been to many conferences and workshops
that focus on signaling. Unfortunately, each lecturer seems to have his
or her favorite signaling pathway: a few arrows leading to MAP kinase,
a particular tyrosine kinase, or a G protein is often offered as an
answer to how a complex function, such as cellular growth, is
controlled. It seems unlikely, however, that we will ever fully
understand cell regulation if we focus on any single signaling
mechanism in isolation. The reality in living tissues is that cells
normally sense multiple simultaneous inputs. For example, at the same
time a cell in a healing wound binds a growth factor, it may be forming
new ECM adhesions at its base, and through these adhesions it may feel
the pull of the surrounding tissue, due in large part to the force of
gravity. However, each individual cell produces only one concerted
response: it either grows or moves or differentiates or it dies
locally. Thus, the key is not which pathways are turned on, but how all
these signals are integrated inside the cell. Furthermore, we have to
understand how this integration occurs within the structural complexity
of the living cell.
Let me provide you with the take-home message right now: we have found
that living cells are literally hard-wired so that they can filter the
same set of chemical inputs to produce different functional outputs and
this mechanism is largely controlled mechanically, through physical
distortion of transmembrane adhesion receptors on the cell surface that
preferentially transmit stresses to the internal CSK. Cells in all
tissues may sense changes in gravitational acceleration through
associated alterations in the balance of forces distributed between
their adhesions and the CSK, rather than through direct activation of
any single "gravioreceptor" molecule. The basis for these claims
will become clear from a review of recent work from our laboratory, as
will be described below. A more thorough discussion of the
mechanotransduction field can be found in recent reviews
(2-6)
.
|
CONTROL OF CELL GROWTH AND FUNCTION BY CELL DISTORTION
|
|---|
A simple example of how mechanical forces can impact on cell
behavior comes from studies in which we made cell distortion or cell
shape an independent variable. It has been long known that cells deform
from spherical to flat when they attach to an ECM substrate and that
this distortion is mediated in part by the action of tractional forces
(CSK tension) that cells exert on their own adhesions (7)
.
Our concept for controlling cell shape was to create small adhesive
islands that would prevent cell spreading but would promote local
clustering of cell surface ECM receptors (integrins) and associated
activation of signaling pathways inside the cell. Because of the
presence of surrounding nonadhesive Teflon-like regions, the cells
would be physically restricted in terms of how far they could extend
and flatten. By including saturating amounts of soluble mitogens in the
medium and a constant high concentration of immobilized ECM ligand,
such as fibronectin, the size of the island could be varied
independently while maintaining these other regulatory factors
constant.
We accomplished this through the use of a soft lithography-based
micropatterning technology (8
, 9
) that was
developed by George Whitesides' laboratory (Department of Chemistry,
Harvard University) as an alternative method for creating microchips
for the computer industry (10)
. When we plated capillary
endothelial (CE) cells on circular islands 20 µm in diameter that
were separated from neighboring islands by wider (40 µm) nonadhesive
(polyethylene glycol-coated) regions, we obtained spherical cells that
remained limited to the size, shape, and position of the engineered
island (Fig. 1
, A and B). When we plated the same cells on
square islands, we produced square cells with 90° corners (Fig. 1, C
and D).
View larger version (63K):
[in this window]
[in a new window]
|
Figure 1. Control of cell shape using micropatterned substrates. A)
Pattern with 20-µm-wide circular islands; B) CE cells
grown for 24 h on 20-µm circular islands; C) pattern
with different sized square islands of indicated size; D)
cells grown on corresponding square islands; E ) pattern
with multiple, small (5 µm) circular adhesive islands that are more
closely spaced (10 µm); F ) single cell spread over many
similar sized circular islands shown in E (see ref. 9
for
more details).
|
|
To explore the effects of cell distortion on cell function, Christopher
Chen, a student in my laboratory, cultured cells on individual square
adhesive islands of varying size. Using this approach, he was able to
demonstrate that CE cells can be geometrically switched between gene
programs for growth and apoptosis (programmed cell death)
(9)
. Cell growth increased in an exponential fashion as
the adhesive area available for cell spreading increased from
approximately 1000 to 3000 µm2. Conversely, as cell
spreading was restricted to smaller sizes and growth was turned off,
apoptosis was switched on. Cell distortion, rather than a particular
form (e.g., round vs. square), was the critical element here because
cells that spread to a similar degree (equal projected cell area)
exhibited similar behavior. Recently, Laura Dike and Chris Chen
extended these studies by showing that linear substrates that hold CE
cells in a moderate degree of extension, which induces neither growth
nor apoptosis, switch on differentiation (capillary tube formation) in
CE cells (Dike, L., Chen, C., and Ingber, D., unpublished
observations). Similar results were obtained earlier by plating CE
cells on dishes coated with varying ECM densities that differed in
their ability to resist cell tractional forces (11)
. Thus,
mechanical interactions between cells and their ECM appear to play a
critical role in cell regulation by switching cells between different
gene programs (Fig. 2
).
View larger version (18K):
[in this window]
[in a new window]
|
Figure 2. Schematic diagram showing mechanical switching between different gene
programs in CE cells. CE cells grow when spread, die when fully
retracted, and differentiate into capillary tubes if maintained at a
moderate degree of extension while also forming cell-cell contacts
(e.g., on thin micropatterned lines).
|
|
|
ADHESION COMPLEXES AS SITES FOR SIGNAL INTEGRATION
|
|---|
How can this work? How can cell adhesion and spreading shift cells
between different gene programs in the presence of a saturating amount
of soluble growth factors? Part of the explanation appears to be based
on how cells restructure their CSK when they adhere to ECM. Cell
attachment to ECM is mediated by binding of transmembrane integrin
receptors that cluster in spot weld-like anchoring sites, known as
focal adhesions (FAs), where they physically interconnect with the
actin CSK (12)
. This connection is mediated by binding
interactions between the cytoplasmic face of the integrin receptors and
actin-associated molecules in the cytoplasm, such as vinculin and
talin.
We (13
, 14
) and others (15
,
16
) discovered that many of the signaling molecules that
are turned on by integrins and mitogenic receptors are not floating
around in lipid bilayer; rather, they are immobilized on the CSK and
specifically concentrated within the FA at the site of integrin
binding. This includes tyrosine kinases, such as c-src and FAK kinase,
as well as the Na+/H+ antiporter, inositol
lipid kinases, and others. In fact, we even found that a subset of
high-affinity growth factor receptors (FGF receptors) are concentrated
within the FA (13)
. Thus, part of the mechanism of signal
integration appears to be based on the spatial organization of
signaling molecules within the cell and their physical association with
CSK elements within the FA. Similar integration likely occurs in
cell-cell adhesions because signaling molecules also concentrate in the
CSK adhesion complexes that connect to transmembrane adhesion receptors
(e.g., cadherins) in these regions (12)
.
There are two important points here. The first is that proximity
between different transducing molecules provides a way to integrate
integrin and growth factor signaling pathways at the cell surface. For
example, Helen McNamee, Martin Schwartz, and I found that integrins can
regulate inositol signaling by controlling synthesis and local
availability of the phosphatidylinositol-bis-phosphate
substrate within the FA, whereas growth factors (e.g., platelet-derived
growth factors) control downstream signaling by regulating the
breakdown of this inositol lipid through activation of phospholipase C
(14
, 17
).
The second point is the finding that these signaling molecules
essentially function when in a "solid state," that is, when
immobilized on insoluble CSK scaffolds. This finding is recurring again
and again in cell biology: many of the enzymes and substrates that
mediate DNA synthesis, RNA processing, protein synthesis, and
glycolysis also appear to function when immobilized on insoluble
scaffolds in the cytoplasm and nucleus (18)
. This finding
provides a possible mechanism to explain the incredible efficiency of
biochemical reactions that are observed in living cells and cannot be
mimicked in a test tube. However, it also has important implications
for how stress-induced changes in cell shape and CSK structure could
alter biochemical reactions and change gene expression, given that the
FA represents a preferred site for mechanical signal transfer across
the cell surface (19-21)
, as I will describe below.
|
FUNCTIONAL CONTROL BY CELL SHAPE AND CYTOSKETAL TENSION
|
|---|
The integrin signaling elicited within FAs has been shown to be
critical for control of cell growth and function (12
,
22
). Thus, one could argue that cell shape and mechanical
distortion of the CSK are not important; instead, it might be the
increase in total area of cell-ECM contacts and associated integrin
binding that dictates whether cells will grow or die on large versus
small adhesive islands. To explore this further, we effectively broke
up a single small adhesive island (which would not support spreading or
growth) into many smaller FA-sized islands (3–5 µm in diameter) that
were spread out and separated by nonadhesive barrier regions (Fig. 1E
) (9)
. When cells were plated on these
substrates, their processes stretched from island to island and the
cells exhibited an overall extended form similar to cells on large
islands (Fig. 1F
). However, the total area of cell-ECM
contact was identical to that exhibited by non-growing cells on the
smaller islands. The question was: which is the most critical element
for cell growth: the ECM contact area available for integrin binding or
cell shape?
The answer is that cell shape is the most critical determinant of cell
function, at least in the presence of optimal growth factors and high
ECM binding. DNA synthesis was high in the cells that spread over
multiple small islands, whereas apoptosis was completely shut off
(9)
. Thus, cell shape per se appears to govern
how individual cells will respond to chemical signals (soluble mitogens
and insoluble ECM molecules) in their local microenvironment, as first
proposed by Folkman and Moscona (23)
. This mechanism for
establishing local growth differentials may play a critical role in
morphogenesis in all developing systems (24
,
25
).
What is the mechanism by which cell shape exerts its effects on cell
function? Clearly, the full answer is unknown. However, recent studies
have revealed that cell shape exerts its effects on cell cycle
progression many hours after early growth signals are turned on. When
synchronized CE cells are presented with an optimal ECM coating
concentration, saturating growth factors, and a substrate that promotes
cell spreading, the cells synchronously enter S phase approximately
16–20 h later. Yet, even with optimal growth factor and integrin
stimulation and associated early signaling pathways turned on, growth
can be inhibited if the cells are rounded up any time before 12 h
(e.g., using trypsin or cytochalasin D) (26
,
27
). This finding is consistent with work from other
laboratories which show that adhesion and CSK integrity are required
for cells to pass through the G1/S transition (28
,
29
).
Recent work by Sui Huang in my group has confirmed that even with early
signaling cascades (e.g., MAP kinase) fully activated, CE cells will
not enter S phase unless cell spreading is also promoted
(27)
. Furthermore, cell shape exerts these effects on
growth by harnessing the cell's molecular machinery that is normally
responsible for control of the late G1/S restriction point, including
the key cell cycle regulators: cyclin D1 and the cdk inhibitor,
p27kip1. The importance of cell shape appears to be that it
represents a visual manifestation of an underlying balance of
mechanical forces that in turn convey critical regulatory information
to the cell. For instance, we were able to show that pharmacological
inhibitors of actomyosin-based tension generation can produce the same
cell cycle arrest as cell rounding, without altering global cell shape
(27)
. Thus, changes in mechanical stresses, including
those imposed by gravity, could impact on cellular growth and
biochemistry by altering the net balance of forces within the cell and
its CSK.
|
CELLULAR TENSEGRITY AND ESTABLISHMENT OF THE CELLULAR FORCE BALANCE
|
|---|
These are interesting experimental results, but the question
remains: how could altering stresses applied to ECM or the cell surface
change CSK shape and alter cell function? This is a difficult mechanism
to understand if one accepts the conventional view of the cell being
composed of a viscous fluid cytosol surrounded by an elastic surface
membrane. Instead, we have found that living cells are hard-wired by a
continuous series of molecular struts (microtubules, cross-linked
microfilament bundles), cables (contractile microfilaments), and ropes
(intermediate filaments) in the form of a discrete CSK network that
mechanically connects specific transmembrane adhesion receptors on the
cell surface to discrete contacts on the nucleus in the center of the
cell (19
, 20
, 30
). Furthermore,
we have found that cells use a particular form of architecture, known
as tensegrity, to organize and mechanically stabilize this CSK network
(31)
.
Tensegrity comes out of the Buckminster Fuller school of geodesic
architecture. It emerged, in part, from Fuller trying to understand why
the geodesic dome is so efficient in terms of its ability to carry a
given load with a minimal amount of building materials. He realized
that it is not what it is made of: it could be constructed from wood
struts or aluminum poles. Rather, it is how these elements distribute
and balance mechanical stresses in three dimensions that determine the
dome's stability. Most man-made buildings are compression-dependent in
that the weight of one brick, due to gravity, compresses down on the
brick below and it is through this continuous compression that
stability results. However, if one were to drive a truck into the side
of a brick wall, the wall would shatter into pieces. In contrast,
Fuller realized that the dome gains its omnidirectional stability from
continuous tension that is resisted locally by a subset of its
structural elements (31
, 32
).
It is difficult to visualize continuous tension in the fully
triangulated geodesic dome; however, it is clearly observed in the
tensegrity sculptures of Kenneth Snelson who was one of Fuller's
students. Snelson's tensegrity masts are composed of multiple steel
girders that can rise more than 60 feet in the air, yet none of the
girders physically touch. Instead, they are held up against gravity and
suspended in mid-air through interconnection with a continuous series
of high tension cables that follow geodesic (minimum path) lines.
Spherical models can be constructed using similar building rules
(Fig. 3
). These sculptures may look strange, however, this is how our
bodies are constructed. If you were to visit an anatomy laboratory, you
would see a human skeleton composed of bones that must be wired
together and hung from a stand in order to take on an upright human
form. In reality, our bodies are composed of 206 compression-resistant
bones that are pulled up against the force of gravity and stabilized
through interconnection with a continuos series of tensile muscles,
tendons, and ligaments. Furthermore, the mechanical stiffness of our
bodies is determined by the level of tone or pre-stress in our muscles
and not by osmotic forces. This is true for all types of animals,
independent of size or species. In an identical way, the stability of a
stick-and-string tensegrity structure (Fig. 3)
depends directly on the
pre-stress (preexisting tension) in its filament system.
View larger version (84K):
[in this window]
[in a new window]
|
Figure 3. A computer-generated diagram of a 6-strut and 24-cable tensegrity model
in its equilibrium form. Note that the struts to do not physically
contact one another; the structure is stabilized by the presence of a
continuous series of tension cables that provide tensional continuity
and pre-stress the entire structure (kindly provided by Molecular
Geodesics, Inc.).
|
|
Many years ago, when I was a student at Yale, I constructed a simple
tensegrity model made out of sticks and elastic string (Fig. 4
) and was able to show that this model can mimic many of the
behaviors of complex living cells (24
, 31
,
33
). Because of the internal tone in the elastic strings
of this model, it takes on a spherical shape when unanchored, just as
living cells do in suspension. However, when the model is anchored to a
rigid resisting substrate, it spontaneously spreads out and takes on a
flattened form, like an adherent cell. Furthermore, when the same model
is similarly attached to a flexible substrate, the model pulls that
substrate into compression wrinkles and takes on a more rounded form.
Nearly identical behavior has been observed in living cells cultured on
flexible silicone rubber substrates (7)
. We also showed
that a model built with a nucleus that was connected to the surface by
additional elastic strings nicely mimicked the coordinated changes in
cell and nuclear shape, as well as the polarization of the nucleus to
the base, that are observed in cells when they spread on a rigid
substrate (24
, 33-35
) (Fig. 4)
. More
recently, we and others have confirmed that the mechanical stiffness
(ability to resist shape distortion) of the cell depends directly on
the level of pre-stress within the CSK (36-39)
as well as
on the maintenance of tensional continuity (20
,
30
), as predicted by the tensegrity model.
View larger version (22K):
[in this window]
[in a new window]
|
Figure 4. A nucleated tensegrity cell model shown in its unattached and attached
forms. The larger cell model was constructed from aluminum poles and
elastic cord in the arrangement shown in Figure 3
. The nucleus is a
geodesic sphere composed of applicator sticks and white elastic
strings; the nucleus is connected to the larger cell model by black
elastic threads that cannot be seen because of the black background.
Left) the cell and nucleus in the unattached model both
exhibit spherical forms. Right) coordinated spreading of the
cell and nucleus, along with basal reorientation of the nucleus, are
observed when the model is anchored to a rigid substrate.
|
|
It turns out that geodesic architecture is also utilized at the
molecular level in cells. Many people assume that the CSK transforms
from gel to sol when cells detach from a substrate and round, and that
the cell effectively behaves like a balloon filled with molasses. In
reality, cells can transform between fully spherical and fully
flattened forms over a period of minutes to hours without altering the
total amount of F-actin or microtubule polymer in the cell
(40
, 41
). In fact, high-resolution freeze
etch electron microscopy has shown that the cytoplasm of round cells is
filled by a continuous filamentous actin network that is initially
isotropic (non-oriented) but progressively realigns into linear bundles
(stress fibers) along its base when the cell adheres to a rigid
adhesive substrate (42)
. Meanwhile, in the apical region
of the cells, this same filamentous network can rearrange into
polygonal actin nets that sometimes can extend to form highly
triangulated structures, including molecular geodesic domes composed of
actomyosin filaments (Fig. 5
) (42-44)
. This finding is a clear example of how
cells use tensegrity at the molecular level. It is interesting to note
that these structural interconversions between an isotropic network and
either a linear bundle or geodesic dome can be predicted using a simple
kinematic model that can transform from an unstable network into fully
triangulated tensegrity forms without disruption of network continuity
(33)
. In fact, using this simple kinematic model, we
created structures that precisely mimicked forms observed in the actin
CSK of living cells, including strut for strut and vertex for vertex
correspondence at the molecular level, as defined by electron
microscopy (33)
.
View larger version (160K):
[in this window]
[in a new window]
|
Figure 5. The filamentous actin cytoskeleton of a cultured fibroblast visualized
by rhodaminated-phalloidin staining. A partial arc-like region of a
molecular geodesic dome is observed within the cytoskeleton of the
center cell; note that the filopodia appear as short linear spikes that
initiate only at the vertices of this triangulated lattice. (This image
was kindly provided by Alan Hall and Kate Nobe, University College
London).
|
|
More recent computer simulation studies carried out in collaboration
with Robert Ezzell, Keith Oslakovic, and Ryota Matsuura (Molecular
Geodesics, Inc.; Cambridge, MA) reveal how the pull of actomyosin
filaments can induce buckling or bending within interconnected CSK
struts that bear some of the compressive loads (Fig. 6
). In fact, local compressive buckling of microtubules has been
visualized directly within the CSK of living cells through the use of
microtubules labeled with green fluorescent protein (45)
.
Microtubule curvature (buckling) also has been shown to decrease when
cells are treated with inhibitors of actomyosin-based tension
generation (46)
. In addition, mechanical measurements in
individual cultured cardiac myocytes confirm that increasing the total
mass of polymerized microtubules compromises the contractility of these
cells by physically resisting compressive (inward-directed) forces in
the CSK lattice, even though Z-band movement (actomyosin-based tension
generation) proceeds normally under these conditions (47)
.
Taken together, these data provide direct support for the conjecture
that microtubules bear some of the compressive loads in a cellular
tensegrity system (24
, 48
). However,
microtubules are not the only compression-bearing elements in the CSK.
For example, highly stiffened, cross-linked bundles of actin that form
the cores of filopodia (49)
also bear compression caused
by the centripetal pull of the surrounding cell cortex and surface
membrane in migrating cells (33)
. Based on computer
simulations, it is possible that rapidly polymerizing (elongating)
actin filaments in the leading edge may help to mechanically stabilize
local regions within the lamellipodium by pushing out against the
surrounding microfilament lattice and thereby pre-stressing the network
without active contraction.
|
CELLULAR TENSEGRITY AND MECHANOTRANSDUCTION
|
|---|
Tensegrity does more than predict pattern formation. It also
helps to explain how cells sense and respond to external mechanical
signals. For example, if one were to pull on a single wood beam in the
frame of a house, local bending and breakage would result. In contrast,
when a single element in a pre-stressed tensegrity structure is
stressed, all of the interconnected elements undergo global structural
rearrangements, even at a great distance (Fig. 7
). This results in a linear stiffening response: the mechanical
stiffness of these structures increase in direct proportion as the
level of applied stress is raised (Fig. 8A
). To our amazement, in 1993 we found that living cells
exhibit this precise behavior (19)
.
View larger version (110K):
[in this window]
[in a new window]
|
Figure 7. Tensegrity provides action at a distance. Application of a local stress
to the bottom element of a stick and elastic string tensegrity
structure (left) results in long-distance stress transfer and global
structural rearrangements throughout the entire model (right). This
response is accompanied by linear stiffening behavior (Fig. 8A
).
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Figure 8. Linear stiffening behavior in a tensegrity model (A) and
living CE cells (B). The stiffness of both structures
increases in direct proportion as the level of the applied stress is
raised (methods by which these data were obtained can be found in ref. 19
). Note that an isolated tension element from the tensegrity model
does not exhibit this behavior (A).
|
|
Ning Wang in my group developed a method to measure CSK stiffness in
living cells by allowing cells to bind to small (5 µm diameter)
ferromagnetic spheres coated with ECM ligands, applying controlled
twisting forces (shear stresses) to the bound surface receptors through
the beads, and simultaneously quantitating the degree of bead rotation
(angular strain; Fig. 9
) (19
, 36-39
, 47
,
50-52
). First, he observed that twisting on bound
metabolic receptors (acetylated low-density lipoprotein receptor)
(19)
or histocompatibility antigens (52)
was
met by very little resistance, even though they spanned the plasma
membrane. In contrast, twisting bound integrin receptors (which form
FAs that connect to the actin CSK) resulted in a proportional increase
in cellular stiffness (Fig. 8B
). In other words, living
cells behaved as if they were tensegrity structures. The importance of
this observation may be even more significant than seen at first
glance. For example, bioengineers have known for many years that most
normal tissues also exhibit linear stiffening behavior, however, this
response could never be explained starting from first mathematical
principles. Working in collaboration with Dimitrije Stamenovic, we now
can explain this fundamental property of living cells and tissues
mathematically, using tensegrity as a mechanistic basis
(53
, 54
).
View larger version (13K):
[in this window]
[in a new window]
|
Figure 9. A schematic diagram demonstrating how cytoskeletal mechanics can
be measured non-invasively in living cells through the use of
ligand-coated magnetic beads in combination with a magnetic twisting
device. Ferromagnetic microbeads (5-µm diameter) were allowed to bind
to cell surfaces for 10 min and unbound beads were removed before magnetic manipulation was initiated. Brief application of a strong
external magnetic field (1000 Gauss for 10 µs) resulted in
magnetization and alignment of the magnetic moments of all
surface-bound beads. Defined mechanical stresses (0–68
dynes/cm2) were then applied without remagnetizing the
beads using a weaker "twisting" magnetic field (0–25 Gauss)
applied perpendicular to the original field. The average bead rotation
(angular strain) induced by the twisting field was quantitated
using a magnetometer. In the absence of force transmission across the
cell surface, the spherical beads would twist in place by 90°
into complete alignment with the twisting field. In contrast,
transmission of force to the CSK would result in increased
resistance to deformation. Thus, we can measure changes in cytoskeletal
mechanics in living cells by measuring changes in the rate and degree
of bead rotation (See refs. 19
and 50 for more details).
|
|
The major point here is that if cells behave as tensegrity structures,
then changes in CSK and nuclear structure may be driven by altering the
balance of forces transmitted across the cell surface. However, the
model predicts and our data confirm that mechanical stresses are not
transmitted equally across all points on the surface membrane. Rather,
mechanical signals appear to be transferred into the cells specifically
at sites where the cell mechanically connects its internal scaffolds to
its external supporting structures, that is, within specialized
adhesion complexes, including FAs and cell-cell junctions. Experimental
studies have confirmed that integrins and other adhesion molecules
(e.g., cadherins, selectins) provide preferred paths for mechanical
signal transfer across the cell surface (19
,
36-39
, 47
, 50-52
), as
suggested by the tensegrity model (2
, 24
).
The finding that the FA mediates mechanical signal transfer is
especially interesting given that the FA represents a major site for
signal processing inside the cell. By focusing diverse stimuli on a
common site, mechanical signals may be simultaneously integrated with
signals from ECM and growth factors to produce a concerted response
(Fig. 10
). However, the overall response to stress may still be modulated
by altering the pre-stress in the CSK-ECM lattice.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 10. A schematic diagram depicting signal integration in cellular adhesion
sites. Signals from growth factors, ECM, and mechanical stresses all
converge on common sites where cells physically anchor to extracellular
matrix (focal adhesion) and to other cells (not shown). Because living
cells always generate active tension in their cytoskeleton, an
exogenous mechanical stress is imposed on a preexisting force balance.
By altering the balance of forces transmitted across the adhesion site,
the signaling machinery that is oriented on the cytoskeletal backbone
of the adhesion complex can be altered, thereby producing different
functional outputs, even though growth factor and ECM signals remain
constant. GF, growth factor signals; ECM, signals from extracellular
matrix; FA, focal adhesion complex; T, cytoskeletal tension.
|
|
|
HARD-WIRING IN THE CYTOSKELETON
|
|---|
Even with all of these results, many people still have a problem
with the tensegrity model. They prefer to view the cell as a
viscoelastic fluid surrounded by an elastic membrane. If cells are
hard-wired as suggested by the tensegrity model, then pulling rapidly
on integrin receptors on the cell surface would be expected to induce
structural changes in the nucleus in the center of the cell. In
contrast, if the cytoplasm was filled with a viscous fluid, these
stresses would rapidly dissipate in the cytosol. Indeed, Andrew
Maniotis in my lab demonstrated hard-wiring in living cells by pulling
on cell surface integrins with ECM-coated micropipettes and inducing
nucleolar realignment within the center of the nucleus, many
micrometers away from the site of stress application (20)
.
He and Krzysztof Bojanowski also demonstrated mechanical continuity
within the nucleus, including physical connections between all of the
chromosomes in the human genome (55)
. It is interesting to
note that DNA appeared to be responsible for this mechanical
connectedness.
Finally, other groups have shown that the mitotic spindle is itself a
tensegrity structure (56)
. By rapidly disrupting one
microtubule within the mitotic spindle using laser ablation, they were
able to demonstrate immediate buckling in the remaining spindle
microtubules. These results suggest that the microtubules act as
compression struts that push out and pre-stress the surrounding
chromatin network and associated nuclear matrix, thereby creating a
tensegrity force balance that stabilizes the spindle. This is a
beautiful example of the use of hierarchical organization in cellular
tensegrity (31
, 33
). Both the spindle and the
whole cell independently exhibit properties of tensegrity. Yet, at the
same time, the spindle remains tensionally coupled to integrins on the
surface of the mitotic cell (20)
. Hence, both the smaller
and larger system also function as one single, integrated tensegrity
network.
It is important to note that the term, hard-wiring, is used here
in the context of cell and nuclear structure to indicate mechanical
connectedness at any point in time when a force is applied to the cell
surface. The cell and CSK are also highly dynamic; individual molecules
and connectors may come and go. However, tensegrity defines how and
where these forces are transmitted and thus it helps to establish and
maintain characteristic molecular patterns. The hierarchical nature of
cellular tensegrity systems also may permit entire subunits to be
removed and replaced (e.g., integrins in FAs, CSK filaments in the
cytoplasm, cells in whole tissues) without compromising structure or
function at a larger scale.
|
CELLULAR CONTROL LIES IN THE BALANCE OF FORCES
|
|---|
In summary, the importance of these results and of the tensegrity
model is that they suggest that cellular control lies in the balance of
forces that are transmitted across cell surface adhesion receptors,
through the CSK, and into the nucleus (63)
. In this sense,
we can think of integrins and other transmembrane adhesion receptors as
mechanoreceptors because they are among the first cell surface
molecules to sense externally applied mechanical stresses and because
they transmit these signals across the surface over a specific
molecular pathway. Integrins within FAs, and both cadherins and
selectins within their own respective adhesion complexes
(12
, 52
, 57
), may be viewed as
part of a larger mechanochemical transduction machinery or signaling
organelle because they provide a way to interconvert mechanical signals
into a chemical response at the site of force application to the cell
surface (2
, 3
, 12
). For example,
our studies with CE cells revealed that while occupancy and clustering
of integrin receptors is alone sufficient to activate intracellular
signaling cascades and gene expression (13-16
,
57
, 58
), it is not sufficient for CE cell
survival. These cells require integrin signaling and a slight degree of
mechanical distortion (change in forces balanced across the FA) to
survive. If the spreading is not extensive, these cells enter what is
essentially a default differentiation cascade. To further shift cells
from differentiation to growth, the balance of forces transmitted
across integrins, through the FA, and to the CSK once again must be
altered. In other words, the same inputs produce different functional
outputs depending on changes in the level of forces transmitted across
the FA (Fig. 10)
.
Transmission of force across the cell surface and throughout a
tensionally integrated lattice would simultaneously result in CSK
rearrangements at multiple locations inside the cell and nucleus as
well as subsequent changes in molecular mechanics. In fact, chemically
or genetically deleting the FA protein, vinculin, disrupts mechanical
force transfer through the FA, interferes with these CSK
rearrangements, and prevents complex cell behaviors, such as movement,
even though local integrin binding and cell adhesion are unchanged
(21
, 59
). All of these normal behaviors
return when normal vinculin levels are restored through transfection.
If biochemistry does function in a solid state, then changing molecular
architecture and mechanics within the FA, CSK, or interconnected
nuclear scaffolds could have major effects of cellular biochemistry due
to local alterations in thermodynamic or kinetic parameters
(3)
. For example, if one were to rapidly pull on a
molecular filament or scaffold without breaking the structure, then
some or all of the molecular constituents that constitute that
structure must undergo some degree of local distortion. Changing the
shape of a molecule alters its kinetic and thermodynamic behavior and
hence, changes biochemistry (3
, 60
).
Is this true? Can mechanics control biochemistry in living cells? The
answer is clearly yes. A simple example is the observation that the
actin bundles that form the cores of the filopodia that lead cell
migration are not formed randomly. Rather, they preferentially
polymerize specifically from the vertices of actin geodesic domes or
triangulated nets that form in more central regions of the cell
(43
, 44
). In other words, while increasing
levels of free actin monomer or cleaving actin filaments may chemically
promote new actin polymerization, it is the tension-dependent
restructuring of the CSK that dictates where these new bundles form and
how they integrate with the rest of the cell. This is a clear example
of how cell mechanics and structure can locally impinge on a
biochemical reaction as well as on cell polarity.
Another example comes from the work of Steve Heidemann and Robert
Buxbaum. They found that tensegrity-based transfer of mechanical loads
between contractile actomyosin filaments, microtubule struts, and ECM
tethers can explain how neurites extend their cell processes
(48)
. Transferring a compressive load off the end of the
microtubule and onto a newly formed ECM adhesion results in a change in
the critical concentration of tubulin in the cell. If the levels of
free tubulin remain constant, this would lead to the addition of
tubulin monomers onto the end of the microtubule and hence, extension
of the neurite process. Experimental studies confirm that neurites are
pre-stressed structures and that neurite extension can be controlled by
altering this type of cellular tensegrity force balance
(61)
. A similar tensegrity mechanism appears to act in
epithelial cells, however, alterations in this force balance manifest
themselves through changes in tubulin protein half-life, rather than
changes in total microtubule mass (62)
.
A third example comes from recent work by Marina Chicurel in my
laboratory. Working in collaboration with Robert Singer (Albert
Einstein Medical Center), we showed that altering the balance of
mechanical forces transmitted across cell surface integrins can induce
formation of a microcompartment specialized for protein synthesis in
the region of the FA (63)
. We had previously shown that CE
cells recruit many signaling molecules involved in growth control
(e.g., Na+/H+ antiporter, FGF receptor) to the
FA that forms at the site of integrin binding when they are allowed to
bind to small ECM-coated microbeads (13)
. Because many of
these molecules are also involved in control of protein synthesis, we
carried out high-resolution in situ hybridization with
oligonucleotide probes for poly A+ mRNA and ribosomes. We
observed integrin-dependent recruitment of both probes to the
bead-associated FA within minutes after bead binding. This recruitment
could be inhibited by chemically interfering with actomyosin-based
tension generation and enhanced in a stress-dependent manner by
magnetically twisting the magnetic beads bound to surface integrin
receptors but not to other transmembrane molecules (e.g., HLA antigen)
(63)
. In other words, formation of this microcompartment
was controlled specifically by transmitting stresses across the FA,
thereby altering the CSK force balance. More recent studies show that
cellular genes in the cAMP-dependent signaling pathway can be regulated
in a similar stress- and integrin-dependent manner (C. Meyer and D.
Ingber; unpublished results). Other studies similarly demonstrate that
application of mechanical forces to integrins preferentially activates
intracellular signaling mechanisms (reviewed in ref. 4
). In the case of
the neuromuscular synapse, transfer of stress across integrins results
in an almost immediate (< 10 ms) release of calcium inside the cell;
the effect is so rapid that it likely results from direct distortion of
integrins and associated calcium channels at or very near the site of
integrin binding (64)
.
|
IMPLICATIONS FOR GRAVITY SENSATION
|
|---|
What does this mean for how gravity influences cell and tissue
development? First of all, local distortion in the CSK appears to be
common to all mechanisms of cellular mechanotransduction. Certain
specialized mechanosensory cells utilize highly dense organelles or
microstructures (e.g., otoliths, stereocilia) to induce a localized
stress differential and associated mechanical strain in the CSK in
order to experience gravitational acceleration. Our experiments suggest
that other nonspecialized cells may similarly feel the pull of gravity
as a result of CSK distortion. However, this distortion may result
indirectly through changes in the mechanics or flexibility of the
cell's adhesions to ECM or to neighboring cells. These changes would
be caused by large-scale forces acting at the whole tissue or organ
level (e.g., bending of the femur, altered tone in muscles,
posture-dependent effects on the vasculature and lung) that, in turn,
alter the structure and form of their internal cellular and molecular
components.
Taken together, these results suggest that we should discard the
concept of a single gravity-specific receptor molecule and the idea
that mechanical signals are transmitted equally at all points on the
cell surface. Instead, we must try to understand gravity sensation in
the context of the structural complexity of living cells and tissues.
From this perspective, it is likely that in some cases the whole cell
or even the whole tissue must be viewed as "the" gravity sensor.
The importance of tensegrity is that it provides a way to focus
mechanical energy on specific molecular components and to integrate
part and whole, thereby tuning the entire cellular (and tissue)
response (3
, 31
, 32
). This
tuning function may involve local changes in cellular rearrangements,
cell shape, molecular architecture, chemical linkages, or modulation of
internal pre-stress. Gravity acting on the whole organism is in fact a
major contributor to pre-stress within individual tissues. When
organisms are placed in microgravity, they experience an acute decrease
in pre-stress on the macroscale which, due to the hierarchical
organization of living tissues (31)
, should trickle down
to produce corresponding changes in structure and mechanics at the
cellular and molecular level (Fig. 11
). Clearly, other factors, including cosmic radiation, changes in
fluid convection, vibration, and cabin conditions (e.g., constant
temperature, artificial light, etc.) also could contribute to the
effects of microgravity during spaceflight. However, future studies on
gravity sensation should incorporate tensegrity-based considerations
and strive to understand the minimal size of cellular aggregates
necessary to sense gravity as well as the importance of tissue
mechanics and CSK tone for the gravitropic response.
|
ACKNOWLEDGMENTS
|
|---|
This work was supported by grants from NASA and NIH. I would like
to thank all of my past and present students, fellows, technicians, and
collaborators for their critical contribution to the success of these
studies; Dr. Alan Hall and Kate Nobes of the University College London
for kindly providing their beautiful micrographs of actin geodesic
domes; and the staff of Molecular Geodesics, Inc. for their
collaborative efforts and for kindly providing computer-generated
images (the author is a consultant to and equity holder in Molecular
Geodesics, Inc.).
|
FOOTNOTES
|
|---|
2 Abbreviations: ESA, European Space
Agency; NASA, National Aeronautics and Space Administration; CSK,
cytoskeleton; ECM, extracellular matrix; CE, capillary endothelial;
FAs, focal adhesions.
|
REFERENCES
|
|---|
-
Koch, J. C. (1917) The laws of bone architecture. Am. J. Anat. 21,177-298
-
Ingber, D. E. (1991) Integrins as mechanochemical transducers. Curr. Opin. Cell Biol. 3,841-848[Medline]
-
Ingber, D. E. (1997) Tensegrity: the architectural basis of cellular mechanotransduction. Annu. Rev. Physiol. 59,575-599[Medline]
-
Chicurel, M., Chen, C. S., Ingber, D. E. (1998) Cellular control lies in the balance of forces. Curr. Opin. Cell Biol. 10,232-239[Medline]
-
Davies, P. F. (1995) Flow-mediated endothelial mechanotransduction. Physiol. Rev. 75,519-560[Abstract/Free Full Text]
-
Vandenburgh, H. H. (1992) Mechanical forces and their second messengers in stimulating cell growth in vitro. Am. J. Physiol. 262,R350-R355[Abstract/Free Full Text]
-
Harris, A. K., Wild, P., Stopak, D. (1980) Silicone rubber substrata: a new wrinkle in the study of cell locomotion. Science 208,177-180[Abstract/Free Full Text]
-
Singhvi, R., Kumar, A., Lopez, G., Stephanopoulos, G. N., Wang, D. I. C., Whitesides, G. M., Ingber, D. E. (1994) Engineering cell shape and function. Science 264,696-698[Abstract/Free Full Text]
-
Chen, C. S., Mrksich, M., Huang, S., Whitesides, G., Ingber, D. E. (1997) Geometric control of cell life and death. Science 276,1425-1428[Abstract/Free Full Text]
-
Prime, K. L., Whitesides, G. M. (1991) Self-assembled organic monolayers: model systems for studying adsorption of proteins at surfaces. Science 252,1164-1167[Medline]
-
Ingber, D. E., Folkman, J. (1989) Mechanochemical switching between growth and differentiation during fibroblast growth factor-stimulated angiogenesis in vitro: role of extracellular matrix. J. Cell Biol. 109,317-330[Abstract/Free Full Text]
-
Yamada, K. M., Geiger, B. (1997) Molecular interactions in cell adhesions complexes. Curr. Opin. Cell Biol. 9,76-85[Medline]
-
Plopper, G., McNamee, H., Dike, L., Bojanowski, K., Ingber, D. E. (1995) Convergence of integrin and growth factor receptor signaling pathways within the focal adhesion complex. Mol. Biol. Cell 6,1349-1365[Abstract]
-
McNamee, H. P., Liley, H. G., Ingber, D. E. (1996) Integrin-dependent control of inositol lipid synthesis in vascular endothelial cells and smooth muscle cells. Exp. Cell Res. 224,116-122[Medline]
-
Miyamoto, S., Akiyama, S., Yamada, K. M. (1995) Synergistic roles for receptor occupancy and aggregation in integrin transmembrane function. Science 267,883-885[Abstract/Free Full Text]
-
Miyamoto, S., Teramoto, H., Coso, O. A., Gutkind, J. S., Burbelo, P. D., Akiyama, S. K., Yamada, K. M. (1995) Integrin function: molecular hierarchies of cytoskeletal and signaling molecules. J. Cell Biol. 131,791-805[Abstract/Free Full Text]
-
McNamee, H., Ingber, D., Schwartz, M. (1993) Adhesion to fibronectin stimulates inositol lipid synthesis and enhances PDGF-induced inositol lipid breakdown. J. Cell Biol. 121,673-678[Abstract/Free Full Text]
-
Ingber, D. E. (1993) The riddle of morphogenesis: a question of solution chemistry or molecular cell engineering?. Cell 75,1249-1252[Medline]
-
Wang, N., Butler, J. P., Ingber, D. E. (1993) Mechanotransduction across the cell surface and through the cytoskeleton. Science 260,1124-1127[Abstract/Free Full Text]
-
Maniotis, A., Chen, C., Ingber, D. E. (1997) Demonstration of mechanical connections between integrins, cytoskeletal filaments, and nucleoplasm that stabilize nuclear structure. Proc. Natl. Acad. Sci. USA 94,849-854[Abstract/Free Full Text]
-
Ezzell, R. M., Goldmann, W. H., Wang, N., Parasharama, N., Ingber, D. E. (1997) Vinculin promotes cell spreading by mechanically coupling integrins to the cytoskeleton. Exp. Cell Res. 231,14-26[Medline]
-
Clark, E. A., Brugge, J. S. (1995) Integrins and signal transduction pathways, the road taken. Science 268,233-239[Abstract/Free Full Text]
-
Folkman, J., Moscona, A. (1978) Role of cell shape in growth control. Nature 273,345-349[Medline]
-
Ingber, D. E., Jamieson, J. D. (1985) Cells as tensegrity structures: architectural regulation of histodifferentiation by physical forces tranduced over basement membrane. Andersson, L. C. Gahmberg, C. G. Ekblom, P. eds. Gene Expression During Normal and Malignant Differentiation ,13-32 Academic Press Orlando, FL.
-
Ingber, D. E., Folkman, J. (1989) How does extracellular matrix control capillary morphogenesis?. Cell 58,803-805[Medline]
-
Ingber, D. E., Prusty, D., Sun, Z., Betensky, H., Wang, N. (1995) Cell shape, cytoskeletal mechanics, and cell cycle control in angiogenesis. J. BioMechanics 28,1471-1484[Medline]
-
Huang, S., Chen, C. S., Ingber, D. E. (1998) Control of cyclin D1, p27Kip1 and cell cycle progression in human capillary endothelial cells by cell shape and cytoskeletal tension. Mol. Biol. Cell 9,3179-3193[Abstract/Free Full Text]
-
Hansen, L., Mooney, D., Vacanti, J. P., Ingber, D. E. (1994) Integrin binding and cell spreading on extracellular matrix act at different points in the cell cycle to promote hepatocyte growth. Mol. Biol. Cell 5,967-975[Abstract]
-
Bohmer, R-M., Scharf, E., Assoian, R. K. (1996) Cytoskeletal integrity is required throughout the mitogen stimulation phase of the cell cycle and mediates the anchorage-dependent expression of cyclin D1. Mol. Biol. Cell 7,101-111[Abstract]
TOP
ABSTRACT
INTRODUCTION
