In any given system, there always exists a structural hierarchy- cells, tissues and organs; physically connected through a force-sensing network on the macroscale (organs, tissues) to the microscale (cells, protein complexes), to the nanoscale (individual proteins). These hierarchical structures constantly adapt to their micro-environment by mechanically stabilizing themselves through tensional forces transmitted along the continuum of the cytoskeleton filament systems. As a result, perpetual force-induced deformations of this architecture take place, which in certain circumstances lead to cell motility.
Such mechanotransduction is an integral part of cellular physiology and plays a critical role in organism homeostasis. Numerous models and design principles have emerged to describe how cells sense the geometry and physical forces of their environment, and how they convert these signals into biochemical pathways (reviewed in ,).
In general, mechanobiology can be described based on the individual cellular components involved. However, it is important to note that as in most biological systems, these components do not act in isolation, but instead contribute to a complex network of pathways. These pathways give rise to larger cellular processes, such as mechanosensing, intracellular signaling, nuclear mechanotransduction or even cell wide responses such as cell motility. These responses may once again encourage the cell to test its environment, thereby producing a feedback loop. Therefore, each seemingly discrete event or individual component exerts an influence on the others. This allows the system to continuously adjust to changes in its environment in order to maintain integrity by eliciting appropriate responses.
Cells test their environment through a diverse group of mechanosensory proteins and cellular structures. Mechanosensors use force-induced modification(s) and conformation-dependent biochemical reactions to transmit and convey information about the cells environment. Adhesion receptors and other membrane proteins that link the cell to its external environment act as mechanoreceptors, while the proteins/complexes that link these receptors to the internal cytoskeleton are nanoscale mechanosensory organelles (reviewed in ,). Cells may utilize motility as a means for their mechanosensors to reach out and sense their immediate environment.
2. Intracellular signaling
The initial mechanical signal occurs locally  and is transduced to other mechanosensors along the linked network  (activating functional modules). This leads to transient cytoskeletal rearrangements and in turn, generation of an equivalent opposing force within the extracellular matrix (ECM) resulting in remodeling of ECM components according to intracellular force changes. These events may lead to local force-induced deformation or regain homeostasis (for low-magnitude signal). Thus, the ECM and interconnected network of cytoskeletal elements play a central role as force conduits in tissues.
3. Signal integration to the nucleus
Signal integration happens by accumulation of low-magnitude signals over time (i.e. cyclic activation and deactivation of functional modules from different functions) until a ‘switch’ is turned on. This is important to make accurate decisions in order to produce a cohesive, high-fidelity response. Such repeat events also aid in acquired learning or ‘memory’ that prevents a precipitous response to a transient stimulus.
Repeated intracellular deformations are transmitted to the nucleus eventually altering its architecture (reviewed in ). Thus chromatin conformations and transcription patterns are modulated to produce a response (e.g. motility type).
The cellular response(s) to force occurs from seconds to minutes and the signals are integrated over space and time by mechanoresponsive pathways (reviewed in ). Mechanotransduction may result in a range of responses controlling cell shape, fate, motility, and, tissue growth and architecture.
Overall mechanobiology describes the relationship between a cell and its environment; how a cell can detect, measure and respond to the rigidity of its substrate and how these processes apply to larger biological systems. As the field of mechanobiology developed, several common themes applicable to various cell types and biological systems were described.
A key process in maintaining the balance of forces between a cell and its surroundings, is the ability of the cell to modulate its stiffness. This determines the elastic nature of the cytoskeleton and so in turn affects a diverse range of cellular processes . Currently three models have been widely proposed for the regulation of cytoskeletal stiffness by force:
i) Tensegrity: This model states that pre-existing tension within the architecture of both the cytoskeleton and the extracellular matrix (ECM) determines cytoskeletal rigidity upon the application of load (stress), such that stress is proportional to rigidity. This model assumes the cytoskeleton acts as a network of tensed cables interspersed by soft cellular material ,,. These cables tense and pull in response to an applied force in order to regain cellular stability.
ii) Semiflexible chain: On the assumption that actomyosin filaments are distributed uniformly throughout the cell, this model states that the filaments are non-linearly elastic (similar to cytosol) and stiffen under stress. Hence, actomyosin filaments can be defined as semi-flexible structures, which are suggested to respond to force isotropically i.e. uniformly throughout their structure irrespective of the directionality of the force applied ,,.
iii) Dipole polarization: On the assumption that actomyosin filaments are distributed uniformly throughout the cell, this model states that upon the application of force, the elastic filaments form dipoles. These dipoles propogate force through the cytoskeletal network through polarization and subsequent pulling on the filaments in direction-dependent manner. Hence, actomyosin filaments are suggested to respond to force anisotropically i.e. differentially throughout their structure dependent on the direction of the force applied ,,.
The application of these models to biological scenarios has proven most successful in the case of the tensegrity hypothesis (as reviewed in ). The models described above lead us to consider the following three key concepts, which characterize the mechanotransduction of forces in a cellular environment.
1) Mechanical signal propagation is rapid
Compared to soluble, ligand-induced signal transduction (by diffusion or motor-based translocation), mechanical forces applied to a cell are transmitted more than a 1000 times faster, along cytoskeletal filaments. For example, the activation of Src kinase via mechanical stimulation has been shown to occur in under a second, whilst activation via chemical stimulation requires tens of seconds or longer .
2) Prestressed cell structures promote long distance force propagation
The tensegrity model characterizes the cell as a hard-wired entity, composed of prestressed cytoskeletal filaments and an elastic cytosol. This is in contrast to other models representing the cell as a homogeneous elastic solid. In the latter case where all stress bearing components, namely the cytoskeleton and cytosol, have the same stiffness, force signals will rapidly decay (according to St Venants principle).
In contrast, in a prestressed environment, as defined by the tensegrity model, mechanical signals can be channeled through the cytoskeleton that is able to stiffen relative to the surrounding cytoplasm , . The decay therefore occurs at a much slower rate than when transmitted through softer components and so the force signal is able to travel further.
The tensegrity model also supports the fact that the higher the stiffness differentials between intracellular components, the longer the distance of force propagation. The elastic cytosol and prestressed cytoskeleton are therefore ideally suited to long distance force propagation. A similar process is suggested in context of the stiff nuclear and intranuclear structures relative to the softer cytosol ,.
3) Mechanochemical conversion can be induced from a distance
Mechanotransducers under tension are mechanically anisotropic i.e. elicit a response dependent on the direction of stress loading. Externally applied stress is distributed to points a few microns away from that of the applied force, according to the distribution at the time of pre-existing tensile forces (prestress) ,. A mechanochemical response can therefore be observed at local as well as distant sites, depending on the directionality of the response, as governed by the orientation of filaments interacting with the point of force application (reviewed in ).
Cellular systems function at a nanometer level and use a highly dynamic set of components. Both of these factors act as primary constraints that prevent the generation of momentum within the cell. Energy is therefore introduced into the system by high energy ligands, such as ATP and GTP (adenosine and guanosine triphosphate, respectively). The effect of energy transfer, across a cellular system, on molecular processes follows the two principles below:
1) Energy is harnessed by capturing conformational states
The assembly and activity of functional modules (i.e. protein complexes) requires ligand hydrolysis. The resultant energy introduced into these modules can be stored in particular conformations of proteins in a complex and can be later used to drive reactions and/or transfer energy onto other molecules by change of conformations.
2) Molecular dynamics are regulated by force sensing
When in a complex, a protein has bonds that mediate protein-protein interactions as well as those that maintain its own conformation. Upon force application, the dissociation of bonds between proteins in the complex competes with the dissociation of bonds within the protein, with the latter instance being favored as this allows the complex as a whole to maintain tension. This has been demonstrated using actin, filamin and a-actinin .
A change in conformation can be explained as a transition between two energy minima that are separated by a high energy state that slows down the transition . Applied force favors this transition by lowering the energy requirement and altering the energy minima i.e. stabilizing the new conformation  (see figure below). Thus conformational changes are therefore dependent on force sensing, in terms of both the magnitude and duration of the detected force ,. The conformational changes pass the signal onto neighboring molecules by exposing catalytic sites for initiating reactions and binding sites for signaling and/or cytoskeletal components.
With increase in load, cytoskeletal and cell adhesion structures initially display catch bond behavior, where bond lifetime increases until a threshold is reached when subsequent load then results in slip bond behavior i.e. reduction of lifetime leading to bond dissociation .
When a cell periodically tests its external environment, a mechanical signal is converted to a chemical signal across the cell membrane. Adhesion receptors located on the external cell membrane bind to surface molecules (extracellular membrane (ECM) components or other cells). The strength of this adhesion is converted to force-induced conformations within mechanosensory molecules. These molecules are linked to a network of cytoskeletal filaments, through which the force can be transmitted. Initial mechanosensing is rapid, within a sub-second to second timescale, leading to an early mechanoresponse that occurs within seconds to minutes.
The strength of the initial mechanosensing event (e.g. the amount of force sensed at the cell membrane) stimulates a feedback system within the linked cytoskeletal network. Following the initial response, a cell generates more forces between the contractile filaments, proteins that link the ECM to the cytoskeleton and the cell membrane. The mechanosensitive sites along these filaments can promote a second wave of responses. For example, force-induced membrane tension, compliance or curvature can influence mechanosensors at the cell surface, leading to further focal adhesion assembly , cytoskeleton redistribution , cell movement , and ECM remodeling ,.
There are two types of mechanosensing; active and passive. Passive mechanosensing is also known as ‘outside-in’ mechanosensing, as it is defined by external forces being detected and transduced into the cell (as reviewed in ). These forces include tension, compression, shear stress and hydrostatic pressure. Conversely, active mechanosensing is known as ‘inside-out’ mechanosensing and is defined by internal forces being generated by the cell to detect changes in the external environment. A common example is the generation of cellular traction as means to survey the external environment.
Passive mechanosensing can be exemplified by the detection of fluid shear stress by endothelial cells, via a mechanosensory membrane protein complex that bridges the extracellular and intracellular environments . Another example is that of intraluminal pressure detection by arterial myocytes, via receptor-ion channel complexes .
Active mechanosensing can be exemplified by force generation to promote cell movement from soft substrates to stiffer substrates, in a process termed durotaxis  and in the detection of surface topology . The generation of forces needed for cell migration, namely traction, involves coordinated activity between focal adhesions and retrograde F-actin flow .
Mechanosensing, whether active or passive, leads to intracellular responses that are transduced through the cell and ultimately result in a context-specific response.
Cells and subcellular structures experience forces from a variety of sources. In general, forces are developed from within the cell via the cytoskeleton (endogenous forces) or come from outside the cell (applied forces). Forces exerted on the cell are often dynamic in nature, requiring the cell to constantly re-evaluate its status and adjust its internal and external morphology accordingly. Although the mechanosensors and mechanotransduction events occur locally at the cell periphery, the forces and biochemical signals are transmitted throughout the cell ,,, and are integrated over time ,. In general, they promote stiffening, softening and reorientation of cytoskeletal filaments ,,. This results in strengthening of the entire contractile machinery without disturbing cellular connectivity (reviewed in ).
During cytoskeletal assembly, the filament subunits encounter intermolecular bonding forces that attract (pulls or tenses) its neighboring subunits. These forces are counterbalanced by intramolecular resistance against being compressed. The stiff and flexible regions present in most components and associated proteins endow filaments with elasticity and ‘pre-stresses’ the entire system to resist deforming forces such as extension, bending and compression (reviewed in ).
In non-muscle cells, contractile resistance to deformation (upon force/ stress application) is provided by the actomyosin machinery. This machinery generates tensional forces by remodeling and exerts traction forces on cell-cell and cell-matrix adhesions (strain), thereby creating a resting tension within the cell (reviewed in ) (See ‘Figure: Contractile machinery provides resistance to deformation’ below).
This contractile tension is representative of the feedback system that a cell uses to couple external and internal mechanotransduction events (reviewed in ). Variation in the stress vs. strain homeostasis can influence which of the different mechanosensors work together to orchestrate a concerted response and governs how the signal is integrated ,.
External forces experienced by cells and tissues include tension, compression, shear, swelling and membrane curvature. Resilient and pre-stressed, the contractile machinery immediately responds to micrometric and nanometric variations in the geometry, topography or spatial distribution of their environment ,,,,,.
Forces applied at the macroscale cause a change in the strength of cell-cell or cell-matrix associations. This activates mechanosensors and signals down through the cellular network via the cell membrane, adhesion receptors and focal adhesions. Specific pathways focus the force onto protein complexes that comprise functional modules . These mechanotransduction events allow the cell to distinguish the chemical nature  and stiffness of the underlying surface , as well as specific types of extracellular matrix (ECM) fibers . These factors influence the downstream signaling and cytoskeletal events that lead to altered cell morphology .
Strengthening of the contractile machinery in response to an external mechanical force, ultimately creates an opposite force within the ECM. This leads to ECM remodeling to regain homeostasis and reinforcement of adhesions to resist higher forces . This process therefore creates a feedback loop within the mechano-sensing and -transduction system.
Certain components of the cytoskeleton, namely the microtubules and actin filaments, bear compressive forces to counterbalance the tensile forces in the pre-stressed elements. Other elements, namely intermediate filaments, are necessary for long force transfer through the cytoplasm and for structural integration of the cytoplasm and nucleus (reviewed in ,,).
It is clear that mechanical forces and structural cues influence cell behavior by regulating signal processes. These signals go on to alter gene expression and reorganize the cytoskeleton and extracellular matrix (ECM), resulting in an orchestrated cellular response. The process by which mechanical stimuli are transmitted into the cell, to exert the aforementioned effects, is termed mechanotransduction.
Cytoskeletal filaments are suggested to transmit local stresses over long distances , and as such the cytoskeleton and linked extracellular matrix (ECM) are proposed as a means to focus force upon molecules that can transmit mechanotransduction (reviewed in ). The manner in which mechanical stimuli are transmitted into the cell is dependent on the coordinated activity of mechanosensors. The activity of these mechanosensors and subsequent signal integration is influenced by resting tension levels within the cell, which are set by the cytoskeletal system , (reviewed in ).
The cytoskeleton is therefore a primary factor in facilitating mechanotransduction. The adaptive cytoskeleton deforms through assembly and disassemby of its filaments in response to an applied force. This is akin to other dense sol-gel networks, which have monomers in solution and polymers coexisting in rapid equilibrium. This characteristic allows the cytoskeleton to shift through different physical forms (i.e. phase transitions) in the absence of a biochemical mechanism .
In general the application of force on an object can result in two different outcomes dependent on whether the object is attached to another structure or not (as reviewed in ). If unattached to anything else, an applied force will result in acceleration i.e. the force is translated. This is rarely true in a cellular context. If an object, such as the cell, is attached to another structure or surface, an applied force results in stress. Stress can lead to a mechanical distortion, presented as a morphological and/or structural change.
All bioactive molecules have extended and contracted forms, structural rearrangements, and nanoscale variations in their motion in living systems ,. Forced-induced conformational changes and/or altered assembly occurs in the cytoskeleton, focal adhesions and ECM components ,,,. Mechanosensory molecules (mechanosensors) have a broad set of structural regions or motifs that can be altered over a range of mechanical forces (as reviewed in ,,).
Mechanical forces influence mechanosensors to promote mechanotransduction in a variety of ways, listed below:
Focal adhesions are known to sense both chemical and physical properties of their matrix environment. Chemical sensing is mediated by the different types of receptors that may function additively, synergistically or antagonistically ,,,. However, even with the molecular repertoire of focal adhesions (FAs) having been unveiled substantially (reviewed in , ), the details of functional interplay between these proteins on sensing mechanical forces remain fairly elusive. From available data, it can only be suggested that FAs undergo regulated turnover at any given time governed by diverse environmental signals (reviewed in).
Matrix density, spacing ,, rigidity , orientation and geometry  are some of the physical parameters the focal adhesions and in turn, cells are known to be sensitive to. However, over years, the mechanisms that integrate such complex information have not been understood. Only recently, the key determinants of rigidity sensing by integrins have been described (reviewed in ,):
A remarkable feature of cell-matrix adhesions, which makes them more versatile than any individual surface receptor, is the large repertoire of mechanosensitive and signaling components in their cytoplasmic scaffold. Mechanotransduction most likely occurs through protein conformational changes in multi-modular proteins, that act as molecular switches  (reviewed in ) leading to subsequent phosphorylation signaling pathways or addition of components  (reviewed in ) under internal/external mechanical perturbation . These ultimately affect prominent protein-protein interactions allowing self-assembly and/or remodeling of the adhesion unit . These allow a multitude of possibilities that trigger force-induced signal transduction pathways involving the adhesion components and diverse targets, resulting in cascading events at a distance (reviewed in ).
Based on the hypothetical protein switches, two models (1 and 2 below) have been put forth to describe the physical mechanism of focal adhesion mechanosensitivity (reviewed in ). These are formulated based on the principle that applied force can reduce the energy requirement of conformational transition or compensates for the energy spent during transition. Recent studies that tracked the mechanosensing ability of FAs suggest that protein tyrosine kinases (PTKs) could regulate force threshold for sensing by FAs ,and hence supports the protein switch theory. A third model which is independent of the hypothetical protein switch concept also exists.
1. Stress-driven mechanism
Stress-driven model proposes that stress-sensing happens at the FA-actin bundle interface and most of the proteins that act as molecular switch, change to an active conformation when the stress they experience exceeds a critical value . This model provides an explanation for matrix-rigidity dependent growth and maturation of FAs as well as the consequent ECM remodeling. Some mechanosensitive components that exhibit such force-dependent unfolding include integrins , talin , fibronectin , p130CAS , paxillin , FAK  and Src ,.
A recently proposed model argues that the initiation of adhesion growth in early stages as well as growth and decay of mature adhesions are ECM-stiffness-dependent, resulting from the inherent molecular properties of mechanosensitive adhesion components . The two molecular characteristics that are suggested to provide a force balance mediating these events are force-dependent state transition and strain-dependent surface binding. Hence, the stiffer the substrate higher is the probability of adhesion maturation. However, the role of mechanical force in structural transitions for most of these proteins are based on assumptions.
2. Strain-driven mechanism
In this model, FAs are envisioned as two-layered structures, where the lower integrin-containing layer interacts with the substrate and is mechanosensitive. When the upper layer is subjected to force due actin dynamics, a local strain is experienced on the top of mechanosensitive layer at the front edge of the cell . This compression activates the elastic plaque proteins in the lower layer, increases their affinity and triggers anisotropic growth of focal adhesions in the direction of pulling force ,.
3. Thermodynamic model
This model is based on the thermodynamic principles that stretching the focal adhesions decreases the chemical potential of any given protein within the plaque . According to this model, FA is elastic in nature and stretching in the direction parallel to plasma membrane happens due to contractile forces of the attached stress fibers. Thus reduction in chemical potential acts as a trigger to promote addition of new plaque components from the cytoplasm in the direction of pulling and dispersal occurs on force reduction (reviewed in , see figure below). Work done in later years also support this theory of differential dynamics at the proximal and distal ends  and implicate actin bundling proportional and parallel in direction to force application as a means of substrate elasticity sensing . They propose that the high tension exerted by attached stress fibers can cause active exchange of plaque proteins at the stress fiber-proximal portion of the FA, termed “heel”. Whereas at the distal “toe” region, minimal tension (threshold) prevails and thus the exchange rate is just sufficient to retain the components in place (reviewed in ).
Mechanotransduction is an iterative process, involving multiple rounds of mechano-sensing, transduction, and -response. It is a means by which force-induced conformational and/or biochemical changes result in the activation and amplification of intracellular signaling cascades. These cascades lead to a cellular response i.e. mechanoresponse, which feeds back into the loop of mechanotransduction. Consequently, a highly dynamic steady state can be reached through repeated cycles of force-sensing and concomitant cellular remodeling (reviewed in ). The cellular mechanoresponse often takes the form of a change in morphology and/or motility, which is facilitated by the generation of force by the cell itself.
Cell shape remodeling and cell locomotion specifically require the cell to produce two types of force; a protruding force to extend the leading edge forward and traction forces to move the cell body (reviewed in ,). Although this seems relatively simple, a number of factors contribute to generating the forces needed for modifying cell shape and for cell motility, including:
Forces produced by nucleotide binding, controlled hydrolysis, and release of inorganic phosphate
Energy from nucleotide hydrolysis is used to promote actin filament elongation , to promote the translocation of motor proteins along cytoskeletal filaments (reviewed in ) and is stored within cytoskeletal lattices to generate a resting tension within the cell .
Forces produced by filament self-assembly
Cytoskeletal filament self-assembly generates thermodynamic driving forces ,,, (reviewed in ,) for mechanical restructuring of the cell shape  and for driving cell motility ,,. Both polymer dynamics and their concentration can be directly modified by force, which serves as an inherent feedback system (reviewed by ).
Forces produced by higher-order structuring of the filaments
Cytoskeletal filaments generate tensional forces and exert traction forces on their adhesions to the extracellular matrix (ECM) or to other cells, thereby generating a resting tension within the cell (reviewed in ,).
The generation of force, whether at the level of single molecules or macromolecular filament structures, allows the cell to exert changes in response to specific mechanical stimuli. These changes are both varied and highly specific and include: