Perhaps one of the most well documented solid TME characteristics is associated ECM stiffening. While this is downstream of initial tumorigenesis, owing to the recruitment and activation of cancer-associated fibroblasts, there is an established link between ECM stiffening and metastasis. 2D and 3D (encapsulating) hydrogel platforms (such as collagen, gelatin, or alginate-based gels) find that increased ECM stiffness drives invasion in metastatic breast cancer cells, while non-cancerous cells did not exhibit such invasive phenotypes (Levental et al., 2009; Chaudhuri et al., 2014; Peela et al., 2016; Ondeck et al., 2019). This may owe to oncogene-mediated changes in mechanosensitivity, which alters the transduction of ECM stiffening (Panciera et al., 2020). While this mechanoperception, at least in part, utilizes established mechanosensitive transcriptional regulators YAP/TAZ, 3D encapsulation reduces cross-sectional force exposure, suggesting metastatic mechanosensation may operate through parallel, YAP-independent pathways (Lee et al., 2019). This reprogramming of mechanosensation further enhances phenotypic plasticity, whereby the viscoelasticity of metastatic cells is dynamic and environmentally impressionable compared to their non-metastatic counterparts (Tian et al., 2020). Interestingly, stiffness-dependent chemoresistance is also observed in 2D and 3D hydrogel platforms (Rice et al., 2017; Joyce et al., 2018). Moreover, this stiffness-mediated resistance was only observed in metastatic cell lines, suggesting that phenotypic plasticity and prosurvival activation in metastatic cancer cells are both mechanically coupled.

In addition to stiffening, cancer-associated ECM is more dense and aligned, forming ECM highways for invading cells and altering ligand spacing within and around the tumor. Changes in ligand availability alter integrin subunit involvement, clustering, and focal adhesion complex assembly, thus perturbing intracellular signaling cascades that influence cell behaviors, including migration, proliferation, and survival (Figure 1A; Levental et al., 2009; Jang and Beningo, 2019). As such, when interacting with a sparse, non-cancerous ECM, the invasive and proliferative tendencies of readily metastatic cells are suppressed, suggestive of a ligand-dependent reprogramming that is maintained with a change in microenvironment (Kaukonen et al., 2016). While normal and cancerous fibroblast-generated matrices demonstrate the importance of ligand density, these techniques cannot be well controlled, nor their variables (such as compounded stiffness) isolated. The advent of 2D nano-spaced ligand platforms permits the investigation of ECM density with single-cell resolution. Studies find that cancer cell morphology, motility, plasticity, and adhesion are manipulated in a ligand density-dependent manner (Lee et al., 2011; Amschler et al., 2014, 2018; Horzum et al., 2015). Interestingly, varying ligand density demonstrates a proportional exchange between cell-cell and cell-ECM adhesion (Horzum et al., 2015). Moreover, cells on controlled ligand spacing platforms have also exhibited shifts in states of EMP, implicating ECM density in metastatic progression (Marlar et al., 2016). These nano-spaced ligand platforms have recently been combined with blocking peptidomimetics to delineate integrin subtype involvement in breast cancer drug resistance. Young et al. (2020) demonstrated that metastatic breast cancer drug sensitivity was highly dependent on ligand spacing and integrin subtype, thus affirming ECM architecture’s influence on metastatic protection and progression. Finally, regarding in vitro models of the TME, the importance of co-culture platforms, through which ECM, phenotypic, and chemoreceptive norms are modified by accessory cell types, such as cancer-associated fibroblasts and macrophages, must also be acknowledged (Kuen et al., 2017; Plaster et al., 2019; Vennin et al., 2019; Huang et al., 2020; Libring et al., 2020; Lugo-Cintrón et al., 2020).

Recent studies demonstrate that cancer invasion speed increases with the degree of constriction in long, representative microtracks (Holle et al., 2019; Mosier et al., 2019; Wang et al., 2019). Interestingly, invasion speed also increases with the number of brief, periodic constriction challenges (Mak et al., 2013; Ma et al., 2018). These studies observe confinement-dependent motility changes, with some reporting a mesenchymal to amoeboid-type shift in locomotion, owing to an adhesion-independent reprogramming, as is observed in immune cell invasion (Reversat et al., 2020). This reprogrammed adhesion has been recently highlighted and is accompanied by cell softening consistent with decreases in force exertion on the surrounding ECM (Guck et al., 2005; Kristal-Muscal et al., 2013; Khan et al., 2018; Holenstein et al., 2019; Beri et al., 2020; Han et al., 2020). Importantly, these mechanical traits of an amoeboid phenotype are evident in patient samples (Swaminathan et al., 2011; Plodinec et al., 2012). These data assert an amoeboid transition under confinement as a distinct paradigm during cancer invasion that traditional models of epithelial-mesenchymal transition are unable to characterize.

Nuclear membranes are coupled to the ECM through cytoskeletal fibers in established mechanotransduction pathways (Holle et al., 2018). In amoeboid-transformed cells, reduced ECM coupling obscures classical models of mechanotransduction. However, mechanically gated nuclear pores stretch during nuclear deformation, promoting the shuttling of transcriptional regulators, and activating signaling cascades that modify cell migration and behavior (Elosegui-Artola et al., 2017; Venturini et al., 2020). Under extreme constriction, the nuclear envelope ruptures, resulting in the mixing of the cytosolic, and nuclear contents; a phenomenon that does not impede cancer invasion (Denais et al., 2016; Raab et al., 2016). Interestingly, metastatic cells become more invasive following nuclear envelope rupture, whereas non-cancerous cells undergo accelerated senescence (Nader et al., 2020). Therefore, a deformed or repeatedly ruptured nucleus under confinement may become an independently mechanosensitive apparatus responsible for guiding cell behavior during cancer invasion (Figure 1B). However, the role of such nuclear mechanics in cancer invasion remains a current topic of interest (Fu et al., 2012; Kirby and Lammerding, 2018; Mierke, 2019; Heo et al., 2020; Herráez-Aguilar et al., 2020; Lomakin et al., 2020).

Perfused microfluidic devices recreate one of the most formidable mechanical stresses of the metastatic cascade, fluidic shear stress. These platforms facilitate the physiological culture of endothelial cells, the permeability, and morphology of which are mechanically regulated by flow, thus, improving research translatability of endothelial traversal or junction modification (Wang et al., 2013; Sfriso et al., 2018). Perfused platforms demonstrate that both flow rate and pulsatility influence cancer-endothelial adhesion and subsequent traversal (Kühlbach et al., 2018). Furthermore, trans-endothelial migration is cooperated by disruptions in endothelial permeability owing to external mechanical and chemical perturbation; as is observed in cancer-associated macrophage activation or exposure to tissue-specific factors (Figure 1C; Jeon et al., 2013, 2015; Lee et al., 2014; Peng et al., 2019; Zavyalova et al., 2019). Not all cancer cells possess the ability to traverse the endothelium, which may reflect the phenotypic heterogeneity of metastatic cells (Jeon et al., 2013; Bertulli et al., 2018). Interestingly, mechanically resilient amoeboid phenotypes have been observed during intravasation, evidenced by macrophage-mediated RhoA activity (Kosla et al., 2013; Roh-Johnson et al., 2014). An important study by Chen et al. (2013) demonstrates the endothelium’s pliability as a mechanical barrier, visualizing increases in endothelial apertures throughout a single extravasation. They also report clusters of extravasating cells, which may increase local endothelial exposure to permeabilizing, proinflammatory cancer secretions, further destabilizing the endothelium and enhancing metastatic progression (Chen et al., 2013). In platforms that lack perfusion, metastatic cells still perturb the endothelium’s structural integrity to facilitate intravasation, findings that are supported in vivo (Nguyen et al., 2019).

While co-culture studies more accurately recapitulate cancer-endothelial interactions, they make it challenging to isolate the submicron constriction mechanics that may constitute intra- and extravasation. Recently, glass microfluidic devices with submicron constriction challenges were fabricated with femtosecond laser-assisted etching. While this platform does not recreate other essential variables, such as ECM stiffness, it demonstrates that metastatic cells are capable of submicron invasion and that, as previously reported, invasion speeds increase with constriction. Crucially, this mechanical challenge did not impair post-constriction proliferation or migration (Sima et al., 2020).

Circulating tumor cells form multicellular clusters in vitro and in vivo (Chen et al., 2013; Yu et al., 2013; Aceto et al., 2014; Au et al., 2016). This cell-cell adhesion, and subsequent engagement of adherens junction proteins, such as cadherin, initiates mechanically coupled antiapoptotic signaling cascades and thus, afford CTCs time to disseminate in suspension (Guadamillas et al., 2011; Li et al., 2019; Tang et al., 2020). Studies demonstrate that the mechanotransduction of shear stress may facilitate this phenotypic shift (Zhao et al., 2014; Yang et al., 2016; Follain et al., 2020). Moreover, cancer cells exposed to physiological shear stress are more invasive, proliferative, and chemoresistant than non-cancerous cells and CTCs in static conditions (Lee et al., 2017, 2018; Novak et al., 2019). Therefore, not only are CTCs resistant to anoikis and physiological shear stresses, but such stimuli potentiate metastasis (Figure 1D; Barnes et al., 2012; Zhang et al., 2018). CTC clusters remain highly deformable while maintaining cell-cell adhesions, permitting the navigation of capillary-sized constrictions (Au et al., 2016). Shear stress also enhances extravasation and migration in CTCs, owing to increases in cellular oxidative stress (Ma et al., 2017). Interestingly, while initially softer, metastatic cells stiffen following shear stress exposure, while their non-cancerous counterparts were unresponsive. This reinforces an oncogene-mediated reprogramming of cellular mechanosensitivity and cytoskeletal mechanoadaptation (Chivukula et al., 2015).

Throughout the metastatic cascade, one mechanical exposure seemingly prepares the invading cell for the next. Logically, this progression should be studied in an integrated fashion, rather than in isolation, as is traditional of reductionist research. As such, some metastasis-on-a-chip platforms allow researchers to study each stage of the metastatic cascade in a single microfluidic device (Sleeboom et al., 2018; Sontheimer-Phelps et al., 2019). These facilitate investigations into metastatic enigmata, like organotropism, whereby metastasizing cells have a secondary tissue preference (Figure 1E; Hoshino et al., 2015). While in its infancy, organotropic studies of metastasis do elude to disseminative preference and demonstrate stiffness-dependent TME escape, reflecting in vivo observations (Skardal et al., 2016; Aleman and Skardal, 2019). While cellular mechanics become difficult to resolve with increasing system complexity, important physical cues may be reproduced, and their effects on cancer progression, examined; such as the cyclic tension of respiration in a model of lung metastasis (Hassell et al., 2017). Such biomimetic systems also lend themselves to pharmacological and biochemical screening, granting researchers insight into how such conditioning modifies and influences the biophysics of metastatic microenvironments (Kim J. et al., 2020; Nashimoto et al., 2020; Palacio-Castaneda et al., 2020; Rajan et al., 2020a, b; Sharifi et al., 2020).