A nuclear basis for mechanointelligence in cells

Cells seem to respond to mechanical stimuli as if guided by human-like intelligence, or sometimes by pathological self-destructive impulses. At the core of this cellular mechanointelligence is mechanical memory, in which cells remember mechanical cues or forces that they have experienced in the past and respond by changing themselves or their environment over timescales that are long compared to those of other cellular processes. Mechanotransduction factors into development and into normal daily function of cells, and mechanical memory enables cells to adapt to their physical environment and respond appropriately to mechanical cues and stresses. However, nearly every cell type can remember and react to mechanical stimuli in a way that causes pathology, from blood clotting to fibrotic disease in solid tissues. The need to control mechanobiological responses is pressing, and a question at the heart of mechanobiology is how mechanical cues that make their way into the cell nucleus can trigger pathways that change cells permanently. Pathways for stress to reach the nucleus are well established (Fig. 1), but how stressing the nucleus causes permanent change to cells and gene expression is less clear. A key step in this sequence of events in the cell has now been revealed in a recent PNAS paper by Rashid et al., through the application of fluorescence correlation spectroscopy (FCS) to cells undergoing mechanical stimulus (1): Stressing of the cell’s cytoskeleton can lead to stressing of chromatin, leading to changes in mobility in the nucleus that last for tens of minutes. These new results are direct evidence of the existence of a time window of mechanosensation over which cells can lock in memory that drives phenotypic transitions (2–4).

From the mechanobiological perspective, key mysteries are how cells sense mechanical signals from their surrounding environment and convert them into biochemical responses and mechanical actions, and how this in turn regulates important biological processes such as gene expression, differentiation, and proliferation. That this happens is well established. A mechanical stimulus can lead to changes in the subsequent behavior of cells after cessation of mechanical signals for a time interval that depends on the magnitude, type, and duration of the mechanical signals (5–8).

These changes can be transient. An example is the fibroblast cells that maintain collagenous tissues. Applying static compressive forces for a 1-h interval will reduce the level of activity in the actomyosin motors that drive cell contraction, disrupt apical actin filaments, and lead to the shuttling of histone deacetylase 3 (HDAC3) into the nucleus (9). This, in turn, triggers a fundamental change in the packing of genetic material in the nucleus. The proportion of denser genetic material—heterochromatin—increases by the action of the HDAC3 that enters the nucleus. Histones, the proteins that form the core of the nucleosomes that are the basic units of chromatin structure, undergo posttranslational modifications at their N terminals when genes are activated. This causes the chromatin structure to become more relaxed and open, presumably allowing easier access to DNA for transcription factors and other proteins involved in gene expression. HDAC3 removes the acetyl group, resulting in the less-dense euchromatin transitioning into a more compact and tightly packed heterochromatin structure, which can prevent access to DNA by transcriptional machinery. In the case of compressed fibroblasts, removal of the acetylation marks on the histone tails results in condensation of chromatin and in cells entering a transcriptionally less active state demonstrated by transcriptome analysis (9). However, this is transient: All of these changes in the cytoskeleton and nucleus reverse 1 h after removing the compressive force, indicating that the application of short-term static forces induced short-term and reversible alterations in cell behavior.

Growing evidence suggests that changes in cell behavior induced by mechanical signals can remain for a much longer time if cell exposure time to mechanical signals is long enough. For example, culturing stem cells on two-dimensional substrates for a longer duration (e.g., several days) shows that both stiff and soft substrates can induce mechanical memory and long-term changes in cells, enabling them to retain the signature of their past substrate stiffness for many days after moving to a new substrate with different stiffness (10, 11).

How are mechanical signals transmitted from the extracellular environment to the nucleus and how do these signal transmissions induce short or long-term changes in chromatin structure and gene expression patterns (12)? Mechanical memory in cells is believed to be mediated by changes in the cytoskeleton. The cytoskeletal network of proteins that provides structural support to cells and helps them maintain their shape can be remodeled when cells are subjected to mechanical stress. This can alter the ways that the cell responds to subsequent stresses. This process is thought to involve many different signaling pathways and molecular mechanisms, including the activation of mechanosensitive ion channels and the redistribution of cytoskeletal proteins.

Mechanical signals from the extracellular matrix can be directly transmitted to the nucleus through physical cytoskeletal links (e.g., actin filaments, microtubules, and intermediate filaments) or can indirectly impact the nuclear structure by triggering nuclear translocation of epigenetic and transcriptional factors (13). Growing evidence shows that the direct cytoskeleton-mediated transmission of mechanical signals to the nucleus can induce rapid changes in cell behavior. For example, cyclic forces on the apical surface of cells are directly transmitted to the nucleus within less than 10 ms; these can stretch chromatin. This stretching can, in turn, induce rapid upregulation of genes, both endogenous genes and transgenes that have been artificially introduced into the genome through genetic engineering (14, 15). Furthermore, direct cytoskeleton-mediated transmission of mechanical signals to the nucleus can stretch the nuclear envelope and subsequently open nuclear pores to facilitate nuclear translocation of proteins that promote transcription of specific genes. A much-studied example is Yes-associated protein (YAP), a transcriptional co-activator that plays an important role in the Hippo signaling pathway that is involved in cell proliferation, differentiation, and tissue homeostasis. When YAP enters the nucleus and binds to DNA and other transcription factors, it regulates gene expression and subsequent activity of the Hippo pathway (16). Thus, mechanical stimuli can affect gene expression and drive both physiological and pathological processes (17–19).

But how does it all start, and what sets the timescales? A recent study in PNAS reveals that stretching of both chromatin and nuclear pores, induced by the direct cytoskeleton-mediated transmission of mechanical signals, plays an important role in the short-term after-effects of mechanical signals (1). In this study, application of local cyclic forces on the surface of cells for 10 min leads to chromatin decondensation, which in turn increases protein mobility inside the nucleus. This leads to elevated diffusivity of single proteins in the nucleoplasm that persists several minutes after the cessation of force. However, the elevated levels of chromatin decondensation and the nucleoplasm protein mobility reverse and return to baseline levels 1 h after force cessation, indicating that the application of short-term cyclic forces induces short-term and reversible alterations in cell behavior. Interestingly, it was found that the increased levels of chromatin decondensation and increased nucleoplasm protein mobility remain for a much longer time if transport through nuclear pores is blocked. Blocking nuclear transport either by treating cells with a nuclear pore blocker or by reducing tension in the nuclear envelope retains the elevated levels of chromatin decondensation and nucleoplasm protein mobility even 1 h after force cessation, indicating that nuclear transport through nuclear pore complexes regulates the duration of short-term after-effects of mechanical signals. These results offer a tantalizing glimpse into the mechanisms behind the initial induction of mechanical memory. Findings reveal a transient but potent window of activity in the nucleus, whose duration is modulated by diffusion through nuclear pore channels.

The way that this discovery was made is fascinating in its own right. From the perspective of FCS, the study achieves a long-standing aim of the field, namely to quantify fluctuations from equilibrium in the nucleus, and to thereby determine how the intranuclear mechanical environment changes with mechanical loading. FCS has long been used to study the dynamics of biomolecules in solution, and the efforts to study chromatin using FCS date back at least to the 1980 paper of Sorscher et al. (20), who quantified binding of ethidium bromide to DNA in isolated mammalian nuclei and concluded that nuclear viscosity changes with factors such as hydration. More recent applications of FCS have shown that nuclear confinement and associated nuclear elongation can also result in changes to effective viscosity (21). Although more work is needed to account for the confined nature of the nucleus and thereby measure true diffusion rates (22), the work of Rashid et al. (1) provides a foundation for evaluating the effects of confinement by nuclear pore channels on intranuclear diffusion. Results motivate further studies to test whether nuclear pore complexes can also erase or reinforce the long-term mechanical memory that remains in cells for days and even weeks after cessation of mechanical signals and provide new tools for doing so.