Often the plastic response of a cell allows for beneficial adaptations that lead to maintaining physiological homeostasis. When there is a mechanical perturbation that disrupts this stability, the cell will adapt to this change in an attempt to return the system to a mechanically stable state, or mechanostasis (Chan et al., 2011). The heart is a highly plastic organ that demonstrates the transition between alternate stable states. In cases of increased hemodynamic load (such as pregnancy, repeated exercise, or postnatal growth) the heart muscle will undergo hypertrophic growth to accommodate the new output requirements (Tingare et al., 2013). The left ventricle of well-trained athletes can exceed the mass of non-athletes by as much as 60% (Hill and Olson, 2008). Despite this tremendous alteration, the cells maintain their plasticity; the physiological hypertrophy is reversible when the increased demand is removed. The end of an intensive training schedule leads to cardiac atrophy of 10%–22% in humans (Perhonen et al., 2001). This is also observed after pregnancy when a woman’s heart will decrease in size and return to the pre-pregnancy mass (Tingare et al., 2013).

These highly plastic responses are enabled by cellular proteins that sense the changes in mechanical cues and adapt accordingly. In the case of cardiac muscle size and function, the plastic response arises from remodeling of cardiac cells and the surrounding ECM. For example, increased mechanical loading due to pressure overload has been shown to cause sarcomere deposition in parallel, leading to the increase in myocyte cross-sectional area ultimately causing concentric hypertrophy. In contrast, volume overload often results in the elongation of myocytes due to the deposition of sarcomeres in series, which leads to dilation of the ventricle (Lammerding et al., 2004a; Lyon et al., 2015). Specifically, a change in the mechanical stretch is sensed by many mechanosensors in cardiac tissue: stretch activated ion channels (Yamazaki et al., 1998; Reed et al., 2014), integrins and integrin associated proteins (melusin, integrin-linked kinase) (White et al., 2006), cell surface receptors (G-protein-coupled receptors and angiotensin II type receptors), cytoskeletal and sarcomeric proteins (titin, myosin, or the small LIM domain protein MLP) (Lyon et al., 2015), and the nucleus (Lammerding et al., 2004a). Ultimately, the mechanical signal alters the cardiomyocyte phenotype via changes in gene expression resulting from changes in calcium concentration and activation of transcription pathways such as MAPK, JAK/STAT, PKC, PI3K, and Hippo-YAP/TAZ (Ruwhof and van der Laarse, 2000; Lammerding et al., 2004a; Saucerman et al., 2019). The focal adhesion-integrin complex is a main mechanosensor in cardiac fibroblasts and is known to activate many pathways leading to the production of ECM and fibrosis, including MAPK and p38 pathways (Davis and Molkentin, 2014; Saucerman et al., 2019). In addition to the activation of transcription factors, epigenetic mechanisms that regulate gene transcription could also provide another layer of gene transcription regulation by external mechanical cues (Saucerman et al., 2019).

Epigenetic regulation, or the processes involving chemical modifications of chromatin that influence gene transcription, allows for both phenotypic plasticity in response to environmental cues and the formation of stable memories. Epigenetic mechanisms, such as histone modifications or DNA methylation, are reversible (although DNA methylation is typically more stable than histone modifications) (Tzelepis et al., 2017). At the same time, epigenetic mechanisms can be inherited through cell division, thus maintaining a stable epigenetic memory across generations (Henikoff and Greally, 2016; Kim and Costello, 2017). On a single gene level, the epigenetic modifications along with the spatial organization of chromatin by architectural proteins, give rise to the overall 3-dimensional (3D) chromatin architecture. The spatial organization of chromosomes is dynamic and adaptable to environmental changes (Solovei et al., 2009; Uhler and Shivashankar, 2017) and has been shown to be altered in disease states (Seelbinder et al., 2021; Walker et al., 2021; Heo et al., 2022). However, 3D architectures are also known to be cell type specific and thus demonstrate developmental stability (Seelbinder et al., 2021; Winick-Ng et al., 2021). Therefore, the cellular epigenetic regulation and alterations in chromatin architecture that determine the epigenetic landscape can control both the stability and phenotypic plasticity of cells.

Conrad Waddington proposed the metaphor of the “epigenetic landscape” to describe the changes in cell phenotypic plasticity and stability with changes in epigenetic regulation (Figure 1A). The action of balls rolling down a hill represents the accumulation of epigenetic changes during development that lead to stable differentiated cell states (Goldberg et al., 2007) (Figure 1A). At the same time, the metaphorical balls can roll into valleys adjacent to the current position. This transition represents cell phenotypic changes in response to external cues (e.g., mechanical cues), which can alter the chromatin architecture of the cell (Figures 1A, B). For example, cardiac fibroblasts will become activated into a myofibroblast state in response to a stiffening environment (Wang et al., 2003), which has been shown to change the chromatin architecture (Walker et al., 2021). The phenotypic plasticity or stable memory of the altered state of chromatin architecture is denoted by the depth of the valley in Waddington’s epigenetic landscape.

An increasing body of literature demonstrates the connections between the mechanical environment and the epigenetic landscape of cells (Spagnol et al., 2016; Miroshnikova et al., 2017; Wagh et al., 2021; Dupont and Wickström, 2022). Seminal work in the field demonstrated that substrate stiffness guides differentiation of mesenchymal stem cells (MSCs), highlighting the influence of the mechanical environment on how stem cells “roll” down the epigenetic landscape (Engler et al., 2006) (Figure 1A). In fact, chromatin condensation upon dynamic loading of MSCs was shown to be mediated by the cytoskeleton (Heo et al., 2015; Heo et al., 2016). Furthermore, the chromatin remodeling of MSCs in response to stiff culture environments occurs through the increase in histone acetyltransferase (HATs) and a decrease in histone deacetylases (HDACs) (Killaars et al., 2019; Killaars et al., 2020) (Figure 1B). Such chromatin remodeling through upregulation or downregulation of chromatin modifiers has also been observed in response to compressive forces (Damodaran et al., 2018) and hydrostatic pressure (Maki et al., 2021). In addition, chromosome positioning within the nucleus is also influenced by mechanical cues of cell geometry, which in turn affects the spatial organization of genes thereby regulating gene expression (Maharana et al., 2016; Wang et al., 2017; Alisafaei et al., 2019). Interestingly, recent studies have shown that mechanical cues can even induce cellular reprogramming via chromatin modifications. For instance, substrate topological cues, such as the presence of microgrooves, can influence the epigenetic state of fibroblasts and enhance their reprogramming to iPSCs through inducing changes in H3 acetylation and methylation (Downing et al., 2013). Also, nuclear deformation in fibroblasts through their confinement in microfluidic channels has shown to increase the efficiency of their reprogramming to neurons via decreased levels of H3K9me3 and DNA methylation (Song et al., 2022). Together, these studies demonstrate that the physical environment can alter the slope of the hill in the epigenetic landscape, either accelerating the “rolling down” of differentiating cells or allowing cells to “roll back up” the landscape in reprogramming processes (Figure 1A).

Although the mechanisms of how the epigenetic landscape is influenced by mechanical cues are not completely understood, recent studies focus on 1) indirect biochemical cues arising from the activation of mechanosensitive proteins and 2) the direct transmission of force to chromatin influencing a change in accessibility of transcription (Kirby and Lammerding, 2018; Dupont and Wickström, 2022; Kalukula et al., 2022). Deformation of stretch activated ion channels, such as Piezo1 found in the perinuclear endoplasmic reticulum, have been shown to decrease levels of H3K9me3 marked heterochromatin in response to cell stretching, while levels of H3K9me3 remained unchanged in response to stretch with Piezo1 depleted cells (Nava et al., 2020). Mechanosensitive nuclear membrane protein, Emerin, enrichment around the nuclear membrane can lead to the alteration of heterochromatin anchoring to the lamina associated domain through influencing nuclear G-actin levels (Le et al., 2016). Additionally, deformation of nuclear pores can allow nuclear shuttling of chromatin modifying enzymes, such as histone deacetylase 3 (HDAC3) from the cytoplasm (Zuela-Sopilniak and Lammerding, 2022). Alternatively, the physical links through the cytoskeleton and the LINC complex can allow for physical deformation of chromatin through nuclear strain transfer, influencing gene transcription (Lombardi et al., 2011; Lombardi and Lammerding, 2011; Neelam et al., 2015; Ghosh et al., 2017; Ghosh et al., 2019; Ghosh et al., 2021; Jones et al., 2022). For example, magnetic twisting of cell surface proteins induced transcription of a fluorescently tagged transgene depending on the direction and magnitude of the applied load (Tajik et al., 2016).

In the case of cardiac cells also, there is recent literature that highlights such mechanical regulation of the epigenetic landscape and its importance in cardiomyocyte development and function (Tingare et al., 2013; Stratton and McKinsey, 2016; Jarrell et al., 2019; Lityagina and Dobreva, 2021; Ross and Stroud, 2021; Powers and McCulloch, 2022). Connections from the cardiomyocyte nucleus to the external environment through microtubules and desmin intermediate filaments is required for the maintenance of the nuclear morphology and the loss of these connections results in aberrant gene expression and DNA damage (Heffler et al., 2020). Similarly, previous work has associated disrupted nuclear morphology with cardiac pathology (dilated cardiomyopathy) (Zhou et al., 2020) and the effects of transverse aortic constriction surgery (Karbassi et al., 2019). Cardiomyocyte nuclear morphology can be altered through both external cues and changes of internal cellular proteins, such as lamin proteins and proteins in the LINC complex, such as nesprins (Lammerding et al., 2004b; Ghosh et al., 2019) (Figure 1C). Work from our lab shows that when cardiomyocyte nuclear morphology is disrupted through culture on stiff substrates or through disruption of the LINC complex, the architecture of trimethylated H3K9 marked chromatin remodels, which we linked to dysregulation of the cardiomyocyte development and function (Seelbinder et al., 2021) (Figure 1C). Our work also supports a connection between nuclear strain and heterochromatin organization since regions of higher tensile strain in cardiomyocytes were more likely to overlap with regions of trimethylated H3K9 marked chromatin (Ghosh et al., 2019; Seelbinder et al., 2021). Additionally, stiff environments have been shown to induce changes in the chromatin architecture of cardiac interstitial cells that is linked to a persisting fibrotic phenotype (Walker et al., 2021). Key regulators of the chromatin changes of these cardiac interstitial cells included the CREB Binding Protein (Walker et al., 2022) and HDACs (Walker et al., 2021), while detachment of the nucleus from the cytoskeleton prevented the chromatin architecture changes. Taken together, these results support how the mechanical environment regulates the chromatin architecture, which in turn influences phenotypic changes of cardiac cells (Figures 1B, C).

Recent studies have found that mechanical environments can cause phenotypic transitions that are maintained, even after the mechanical stimulus is removed demonstrating a stable mechanical memory (Figure 1D; Table 1) (Tingare et al., 2013). A critical factor observed to be an important determinant of mechanical memory is the prolonged exposure to the mechanical stimulus. For example, Balestrini et al. have shown that when lung fibroblasts are primed in a pathologically stiff culture environment for 3 weeks, they display a myofibroblast phenotype for 2 weeks after transfer to healthy soft substrates (Balestrini et al., 2012). In the reverse experiment, lung fibroblasts cultured on healthy soft substrates for 3 weeks were partially protected from myofibroblast activation after shifting to the pathologically stiff substrates (Balestrini et al., 2012). Similarly, Yang et al. showed that hMSC exposure to a relatively high stiffness of 10 kPa for 10 days before lowering the stiffness to 2 kPa, led to a permanent (>10 days) nuclear localization of RUNX2, a pre-osteogenic transcription factor, and YAP/TAZ, the transcriptional regulator of differentiation induced by ECM stiffness. Interestingly, experiments probing the mechanism of mechanical memory showed that YAP/TAZ acts as a “rheostat” that regulates the reversibility or irreversibility of the osteogenic response depending on whether the mechanical dosing time exceeded a critical memory threshold (Yang et al., 2014) (Figure 1D). Another molecule identified to play a role in preserving the fibrotic mechanical memory in rat MSCs is microRNA-21. In fact, knocking down microRNA-21 was sufficient to erase accumulated mechanical memory from stiff priming (Li et al., 2017).

In addition to the regulation of microRNAs and transcription factors, recent studies focus on chromatin remodeling induced by histone modifications as a mechanism for retaining mechanical memory (Turner, 2002; Hathaway et al., 2012; Heo et al., 2015; Fan et al., 2017) (Figure 1D). In fact, a theoretical framework to model mechanical memory required the modulation of epigenetic plasticity due to long term mechanical priming that locks in mechanical memory (Peng et al., 2017; Mathur et al., 2020). Dynamic loading applied for 150 s to 3 h resulted in increased levels of chromatin condensation in MSCs. The increased levels of chromatin condensation only persisted in cells dynamically loaded for 3 h, while chromatin condensation of cells that were loaded for less time (less than 1 h) was reversible (Heo et al., 2015). The persistence of condensed chromatin could be prevented by using pharmacological treatments to inhibit EZH2. Similarly, work from our lab shows that remodeling of H3K9me3 marked chromatin occurs when chondrocytes are exposed to stiff substrates and is later remembered when chondrocytes are transferred to a soft 3D environment. However, the remodeling of H3K9me3 can be partially disrupted with inhibitors of epigenetic modifiers (KDM4, EZH2) (Scott et al., 2023; Scott et al., 2022). Many of these examples of mechanical memory and others (listed in Table 1) highlight how changes in chromatin architecture can persist, maintaining stable memories. Therefore, not only are the mechanically induced changes in chromatin architecture involved in the plastic response of the cell, but such changes can also persist after removal of the stimulus, contributing to the stable mechanical memory of the cell.

The effort to maintain mechanostasis in the presence of certain mechanical stimuli may lead to maladaptive responses that decrease the cell’s ability to function in the new environment. For example, a continuous excessive stress on the heart muscle, as caused by high blood pressure, often leads to thickening of the left ventricle walls of the heart, known as maladaptive cardiac hypertrophy (Dorn and Force, 2005; Shyu, 2009). Unlike the highly-reversible physiological hypertrophy, the hypertrophy induced by prolonged stress is usually followed by apoptosis or necrosis and has not been shown to be easily reversed (Jaalouk and Lammerding, 2009; Tingare et al., 2013). Regression of left ventricular remodeling due to pressure overload has been demonstrated with the removal of aortic band constrictions (Gao et al., 2005; Stansfield et al., 2007; Zhao et al., 2013; Yoshida et al., 2020) in animals as well as clinically with treatment of aortic stenosis after aortic valve replacements (Krayenbuehl et al., 1989; Matsumura et al., 2008; Hahn et al., 2013), demonstrating some degree of left ventricular remodeling plasticity. Additionally, regression of left ventricular remodeling has also been seen after weight loss (Kardassis et al., 2012). However, left ventricle dysfunction sometimes persists even after 3 years following weight loss (Alexander and Peterson, 1972). Aortic banding and de-banding studies suggest that exposure time to pressure overload may influence the regression and recovery of left ventricle remodeling (Zhao et al., 2013). Further exploration is needed to identify the conditions or mechanical threshold in which cardiac remodeling leads to heart failure (irreversible memory) and when there is potential for reversal of cardiac remodeling (reversible memory) (Figure 1D).

We propose a depiction of the cardiomyocyte response to loading in Figure 1E to demonstrate how the plastic response of cardiomyocytes changes with increased loading. Each valley in Figure 1E represents an alternate mechanically stable state, such as a physiological normal cardiomyocyte or a pathological hypertrophic cardiomyocyte. The depth of the valley is proportional to the inability to change phenotype. By exploring short and long-term changes in chromatin architecture of cardiac cells in response to various exposure times and degrees of mechanical loading, we may gain a more in-depth understanding of how plasticity or long-term disease states arise in the heart.