Introduction

The interaction of stem cells with their microenvironment, composed of both cellular and acellular components forms the notion of ‘stem cell niche’ that enables the stem cells to maintain their self-renewing ability, multipotency and undifferentiated state [1]. The same idea of niche is applicable to the cancer stem cells, a rare subpopulation of cancer cells thought to be accountable for tumor initiation, heterogeneity, drug resistance and reemergence following remission [2].

Cancer microenvironment often characterized by tissue hypoxia, reactive oxygen species, extracellular matrix remodeling and plethora of factors secreted by mesenchymal stem cells, immune cells and tumor cells actively modulates a multitude of signaling pathways like JAK/STAT, Hedgehog, Wnt, Notch, NF-κβ etc. to help CSCs in maintaining their stemness [3]. Besides the chemical players, physical properties like stiffness and tension of the microenvironment are also increasingly drawing attention to their roles in the development and progression of cancer [4, 5]. An increased risk of developing cancer in a variety of tissues is found to be associated with increased stiffness featured by high stromal collagen content and the presence of fibrotic lesions [6, 7]. The high stiffness of tumors facilitates the activation of mechanosensitive biochemical pathways enhancing the cell-cycle, EMT, cell motility and renders the tumor metastatic [8]. Now scientists are curious to understand how the mechanical properties of cancer microenvironment perturb the expression of CSC markers and associated traits.

Mainstream biologists find it difficult to assess these physicomechanical properties by conventional tools used in biology. In this context, the demand of a new discipline to fill the void of our knowledge is satiated by the advent of mechanobiology, an interdisciplinary field of science at the interface of biology, physics, and engineering. It employs both physical devices and biological techniques to look into how the changes in the mechanical properties of cells and tissues contribute to development, cell differentiation, cell physiology and diseases [9].

In order to unveil the biomechanical aspect of understanding the uniqueness of cancer stem cells and their niche, the basis of metastasis, neovasculogenesis and other important hallmarks of cancer, more reliable, as well as easily tractable model systems, are being designed [10,11,12]. This article critically reviews the participation of CSCs and the associated niche in cancer from a mechanobiological perspective.

Cancer Stem Cells

Since the discovery of cancer stem cells (CSCs) in the year 1997 by Bonnet and Dick in leukemia, they have been shown to exist in several other types of solid tumors including colon, breast, brain, and skin [13,14,15,16,17]. There are mainly two models best known to explain tumor initiation and progression, namely: the stochastic model and the hierarchical model. Whereas the stochastic model counts each tumor cell to have equal tumorigenicity, the hierarchical model suggests only CSCs have all the potential to proliferate and differentiate [18]. Figure 1 depicts the schema of these two models. The current review aims to examine the CSC hypothesis from a mechanobiological viewpoint not going into a comparison of the advantages and pitfalls of different models.

Fig. 1
figure 1

Schematic representation of the two familiar models of cancer initiation and progression. a. Stochastic model: According to this model, each cancer cell of a tumor is equally potent to continuously divide and form tumor. b. Hierarchical model: According to this model, only few cells called cancer stem cells of a tumor are capable of dividing, differentiating and reinitiating tumor. Other cells are not generally able to perpetuate a tumor, though they possess the ability to dedifferentiate into cancer stem cells

Full size image

CSCs present a very small percentage of the overall cancer cell population [19, 20]. No single isolation protocol is exclusive and sufficient for delineating CSCs from other cancer cells. Tumor cells are considered as CSCs if they are simultaneously: (a) positive for specific surface markers (like CD133/ prominin1) (b) part of side population (c) able to form floating spheres in serum-free medium and (d) able to form new tumors when implanted in mice [21]. CSCs exhibit a less adhesive strength towards the basement membrane and hence can be mechanistically enriched from a majority of other cancer cells before being definitively validated by surface marker analysis [22]. Biomechanical profiling of CSCs of different cancers (ovarian, breast, hepatic) using atomic force microscopy or rheological technique like micropipette aspiration has demonstrated that these cells are significantly softer and highly deformable than non-malignant, intermediates and even aggressive late-stage cancer cells [23,24,25]. The soft phenotype has also been shown to be sensitive to the anti-cancer agents [23]. These findings may lead to reaching the goal of efficiently identifying and targeting CSCs from a host of other tumor cells exploiting their mechanical uniqueness.

Like adult stem cells, CSCs display ‘stemness’, a remarkable ability to self-renew, differentiate and to balance between quiescence, proliferation, and regeneration [26]. CSCs exhibit a substantially high capacity of self-renewal, a high degree of plasticity, continuous proliferation, ability to differentiate conforming to intra-tumor heterogeneity and a pronounced resistance to stressful factors and drugs [3, 26,27,28,29,30]. Because of their ability to initiate tumor formation, disseminate to other locations and to re-populate and grow into a new tumor mass, CSCs are interchangeably termed as tumor-initiating cells (TICs) or tumor-repopulating cells (TRCs) [31]. CSCs share many common pathways with adult stem cells like JAK/STAT, Hedgehog, Wnt, Notch, PTEN/AKT/P13K, NF-κB, MAPK/ERK, and TGFβ [26, 32]. Some of these pathways like Wnt and TGFβ cross-talk with mechanosensitive transcription factor YAP/TAZ [33, 34]. Growing evidence shows that mechanical properties of extracellular matrix (ECM) regulate the self-renewal and differentiation of adult stem cells and hence are important in tissue engineering [35, 36]. So, it is quite natural to be inquisitive about ‘if and how’ the tumor microenvironment modulates ECM composition and corresponding mechanical properties conferring to a hospitable niche for CSCs.

Tumor Microenvironment: A Mechanical Panorama

In 1889, Stephen Paget proposed that a favorable interaction between the metastatic cells (seed) and the tumor microenvironment (soil) lies in the root of an organ-preference pattern of breast cancer metastasis [37]. Before examining the applicability of this ‘seed-and-soil’ hypothesis with reference to CSCs and their niche, in this section, we will peep into the biomechanical aspects of the tumor microenvironment, the ‘home’-ground of CSCs.

It has long been practiced by medical doctors to diagnose a tumor on the basis of the differences in tissue rigidity sensed by palpation [38]. When the microenvironment is in a healthy state, it can help protect tumorigenesis while an unhealthy microenvironment becomes an accomplice [39]. Multiple studies, in recent days, have also confirmed that tumor tissue stiffness is much higher than that of its normal counterpart and is strongly correlated with disease progression and clinical outcome [40, 41]. This observation falls in with the fact that the TME experiences an increased deposition and dynamic remodeling of ECM proteins, the universal packing material of living tissues [42]. So, it is intriguing to look into the physicomechanical details of TME in conjunction with the key cellular components that actively sculpt the tumor-associated ECM.

Builders of the Mechanical Milieu

Tumor microenvironment (TME) is asymmetric aggression of a bunch of cellular and acellular components. Other than the cancer cells themselves, the cellular components include vascular cells (endothelial cells and pericytes), immune cells (mast cells, neutrophils, monocytes, macrophages, myeloid-derived suppressor cells) and most importantly the cancer-associated fibroblasts (CAFs). The acellular compartment consists of the extracellular matrix proteins (collagen, fibronectin, laminin etc.) and conditions like hypoxia [39, 43].

Fibroblasts are the cells specialized for secreting ECM proteins that provide the scaffold for the tissue morphogenesis and homeostasis [44]. During adult epidermal wound healing, otherwise quiescent fibroblast cells undergo differentiation and enormous expansion to smooth-muscle myosin (α-SMA)-positive myofibroblasts, expressing stress-fibers [45]. Cancers exceptionally allow the continuous recruitment and conversion of fibroblasts into active myofibroblasts, alternatively cited as CAFs to the tumor sites justifying the notion that “Cancers are wounds that do not heal”. Such kind of differentiation is often referred to as mesenchymal-to-mesenchymal transition (MMT). The CAFs not only originate from the fibroblast precursors but also from the other stromal cells by transdifferentiation, rendering them the most abundant cells in a tumor tissue stroma [46]. Also, there are reports showing that CAF activity is aggravated by the paracrine actions of cancer cells and other cancer-associated reactive stromal cells, especially mast cells, M2 macrophages and endothelial cells [43, 47, 48]. Other than their participation in ECM composition and remodeling, soluble factors form CAFs and other stromal cells are found to extensively cross-talk with the mechanically induced signaling pathways [49].

Hypervascularization and hypoxia, two more signature components of solid tumors, directly and indirectly, result in ECM realignment, elevation of interstitial fluid pressure and shear stress. These mechanical properties reciprocally act on their causative agents to cause a vicious cycle [50]. Concisely, the cellular components along with the acellular factors provide TME with a unique mechanical identity by setting a mutually interactive network in action.

Mechanical Properties of Tumor Microenvironment

Earlier reviews based on the published reports have distinguished four major mechanical perturbations of TME as follows: (a) ECM stiffening (b) elevated interstitial fluid pressure (IFP) (c) increased interstitial fluid flow and (d) compressive/solid stress imparted by confined growth [51].

The ‘core matrisome’ of mammalian ECM is characterized by having about 300 different proteins, of which collagen, proteoglycans, laminins, fibronectins, and elastin are worth mentioning [52]. ECM composition in cancerous tissue is quantitatively altered by both CAFs and the resident cancer cells leading to qualitative changes in terms of rigidity, density, porosity, solubility, and topography [53, 54]. The disruption of the equilibrium between ECM synthesis and secretion, and alterations in the amount and activity of matrix-remodeling enzymes namely MMPs and LOX are responsible for the desmoplastic appearance of solid tumors [42]. Collagen I and fibronectin are the most abundant ECM constituents in TME. Other ECM proteins namely tenascin, decorin, fibromodulin, SPARC, lumican, osteopontin, periostin, versican, and hyaluronan have shown to be implicated in biochemical and biomechanical alterations of TME. The metastatic transformation has been found to be closely associated with the remodeling of basement membrane (BM) proteins (collagen IV, laminin, entactin) and linearization of collagen fibers [54,55,56]. Recent evidence also accuses the tensile stress generated by cellular actomyosin contractility in response to high ECM stiffness for reciprocally regulating the ECM stiffness [57,58,59].

TME is the reservoir of pro-angiogenic agents including ECM components and fibronectin) and paracrine factors like vascular endothelial growth factor (VEGF) [60]. Excessive development of aberrant and leaky vasculature along with the deposition of a large amount of ECM proteins and retention and immobilization of liquid by negatively charged hyaluronan collectively give rise to an elevated IFP [50, 61]. Increased IFP and consequential increment of interstitial fluid flow are linked to heightened chemotherapeutic resistance, induction of epithelial-to-mesenchymal transition (EMT), myofibroblast, collagen alignment and tissue hypoxia [50, 62]. Hypoxia, in turn, promotes CSC self-renewal, excess angiogenesis and secretion along with collagen and collagen-remodeling enzymes by cancer cells [63, 64].

The solid stress in growing tumor develops as a result of: (a) an increased density of cancer cells, stromal cells and ECM components within a defined periphery of the host tissue (defined as residual stress), and (b) the reciprocal compression by the adjacent host tissue (defined as reciprocal stress). Such compressive stresses can regulate tumor morphology, growth, and metastasis [57].

So, from the above discussion, it is obvious that TME creates a preparative and supportive biomechanical atmosphere for the disease to perpetuate. Now, in accordance with the scope of our present review, we would like to dig out the studies regarding the effects of biomechanical anomalies of TME on the CSCs and the signaling pathways involved.

Tumor Microenvironment ‘Niche’s Cancer Stem Cells: Mechanomolecules in Action

Cells can sense the internal and external mechanical fluctuations in a similar manner, yet more quickly they can sense the chemical changes around or inside them, and decide to take a due course of action which encompasses maintenance of cell size and shape, cell migration, cell competition, cell division and what not! Cellular mechanotransduction is the fancy name to describe the transmission of mechanical signals in the form of chemical cascades by the cells [65]. With our previous understanding of the biomechanical characteristics of CSCs and TME, we hereby discuss how TME creates an advantageous ‘niche’ for CSCs to turn on different mechanosensory pathways that make them special. Extrinsic forces operated by ECM constituents can affect native conformation and related interactions of a great variety of molecules (mechanosensors) which, in response, trigger biologically important reactions leading to covalent modification of enzymes, protein-protein interaction, cytoskeletal rearrangement, changes in gene expressions, and beyond. Mechanosensors commonly consist of ion channels, cytoskeletal proteins, junctional proteins, receptors etc. [65, 66]. Figure 2 simplistically summarizes the common mechanosignaling pathways and their implications in CSCs.

Fig. 2
figure 2

Mechanosensory molecular pathways in cancer stem cell. Mechanical stress induced by stiffened ECM, composed of collagen, laminin, fibronectin, hyaluronan etc. can be sensed by mechanosensors (integrins, laminin, CD44 receptors, syndecans, DDRs, GPCRs etc.) and transduced to induce pathways specific for CSC survival, self renewal, drug resistance and progression. DDR: discoidin domain receptor; GPCR: G protein coupled rector; NF-κB: nuclear factor kappa-light-chain-enhancer of activated B cells; FAK: focal adhesion kinase; RhoGDI: Rho GDP-dissociation inhibitor; MLCK: Myosin light-chain kinase; MLC: Myosin light-chain; PKA, PKC: Protein kinase A, C; ERK: extracellular signal-regulated kinase

Full size image

Transmembrane glycoprotein CD44, a known receptor of hyaluronic acid (HA) is considered as an important CSC marker. On binding with growth factors of TME, CD44 and its isoforms form complexes with ezrin resulting in cytoskeletal remodeling and signaling to the nucleus. Tumor necrotic factor (TNF-α) mediated up-regulation of HA leads to the generation of CD44 variants by alternative splicing. The HA-CD44 interaction has been implicated in (a) the sustenance of stemness (Nanog expression), (b) tumor metastasis to liver, bone marrow and lungs and (c) drug resistance [67, 68]. CD133 or prominin is another transmembrane pentaspan glycoprotein and a known biomarker of CSCs. Type I collagen, the causal factor for increased ECM stiffness of tumor stroma, has been found to conditionally stimulate the CD133 expression in glioblastoma cells [69, 70]. CD133 expression is linked with CSC stemness, plasticity, and drug resistance [71]. Other such integral membrane proteins like syndecan-1 (CD138), discoidin domain receptor 1 (DDR1) act as receptors for ECM components like fibronectin and collagen respectively to induce mechanotransduction pathways in CSCs [72, 73].

Laminin (Lam)-α2, a non-collagenous ECM protein acts as a niche for glioblastoma stem cells by supporting their growth and self-renewal [74]. Breast CSCs produce Lam511 matrix which interacts with α6Bβ1 integrin to activate Hippo transducer TAZ which, in turn, induces the transcription of Lam511. This signaling also promotes CSC self-renewal and tumor initiation [75].

Versican, a large chondroitin sulfate proteoglycan is responsible for the emergence of various cancer hallmarks by its interaction with multiple membrane proteins including HA, integrins, CD44, microfibrillar fibulins and epidermal growth factor receptor (EGFR). CSC marker CD44 binds with versican to promote tumor progression and migration via expressing HA-mediated motility receptor (RHAMM) and MMP9 through the activation of JNK and NF-κB pathways [76].

Fibronectin, an essential component of ECM interacts with membrane integrins. Investigation on glioma stem-like cells revealed that fibronectin (Fn) favored cell survival via Erk activation; differentiation, proliferation and motility via the activation of Focal adhesion kinase/Paxillin/AKT signaling; and increased chemoresistance via upregulating P-glycoprotein expression [77].

Formation of macromolecular focal adhesion (FA) complexes is marked by engagement and clustering of integrins and associated proteins classified into (a) ‘integrin signaling layer’ consisting of focal adhesion kinase (FAK) and paxillin (b) ‘force-transduction layer’ made of talin and vinculin and (c) zyxin, VASP, α-actinin constituting the ‘actin regulatory layer’. The level of tyrosine phosphorylation of signaling molecules activates either Rac to protrusion and migration or Rho leading to adhesion growth and stabilization [78]. Binding of type I collagen of stiff ECM with Integrin β1 of CSC membrane is followed by the induction of FAK and subsequent autophosphorylation that recruits Src family kinases. These Src kinases activate the catalytic domain of FAK essential for the formation of the whole FA complex. FAK promotes CSC survival and metastasis in a kinase-dependent manner [79]. ILK, a serine-threonine kinase by nature has also been implicated in the assembly of FA and interaction of FA with actin cytoskeleton [80]. ECM stiffening and tissue hypoxia cooperatively generate the breast CSC pool via the activation of ILK and CD44 [31]. Activated ILK/PI3K/Akt pathway leads to up-regulation of self-renewal capacity in CSCs [81].

Caveolins (Cav) are integral membrane proteins densely populated over the lipid rafts and are involved in receptor-independent endocytosis [82]. Cav1 has been reported to mediate chemoresistance via the activation of Wnt-independent β-catenin/ABCG2 signaling pathway in breast CSCs [83]. Recently, Cav1 has been suggested to regulate a unique mechanotransduction response to substrate stiffness through an actin-dependent control of Yes-associated protein (YAP) [84]. This particular pathway needs further investigation to uncover its contribution to CSC hallmarks.

In several types of cancers, YAP/TAZ helps to sustain CSC features via its increased activity specifically within tissue regions exhibiting higher collagen cross-linking [85]. TAZ induces the self-renewal of non-CSCs and expansion of the pool of CSCs [86]. YAP expression marks CSCs and maintains CSC phenotype through Sox2-Hippo signaling pathway [87]. YAP/TAZ is also important for CSCs to display other hallmarks like EMT and chemoresistance. Shear stress-induced migration and invasion of cancer cells also require YAP onstage [88]. Fluid shear stress has also been shown to induce CSC-like phenotype in epithelial cell adhesion molecule (EpCAM) expressing MCF-7 breast cancer cells without induction of EMT, though the pathways involved are poorly understood [89].

Nuclear architecture and chromatin remodeling are closely related to gene expression and cell differentiation. A specialized multimolecular assembly called linker of nucleoskeleton and cytoskeleton (LINC) containing nesprin, SUN proteins and lamins, constitutes a functional connection of membrane adhesion molecules with nucleoplasm via actin and intermediate filament network [90]. LINC proteins have a close association with the differentiation status of embryonic stem cells, and their loss-of-function is proven to have roles in cancer metastasis [91, 92]. So, their roles in shaping the plasticity and metastatic ability of CSCs in response to external biophysical cues should be thoroughly probed.

The discovery by Tan et al. has pointed at epigenetic changes for melanoma CSCs’ self-renewal capacity and tumorigenic potential that soft matrices bring in through a mechanism involving reduced H3K9 methylation and increased Sox2 expression [93]. Conversely, glioblastoma CSCs show little change in proliferation, migration and spreading as a function of ECM stiffness [94]. Hence, the exact routes operational within the CSCs in response to extrinsic biomechanical cues of tumor ECM to modulate their intrinsic properties and showcase specified hallmarks require in-depth analysis across different cancer types by employing both in vivo and in vitro state-of-the-art approaches.

Tools for Novel Therapeutic Discovery

The tumor microenvironment provides cancer cells with a diverse set of extracellular cues to influence tumor cell behavior and function. Recent developments in the arena of tumor biomechanics and CSC biology, in particular, have chalked out an alternative explanation and potential therapeutic targets of tumor progression and metastasis. Hence, a detailed understanding of the mechanisms of mechanical cues in this context demands some realistic model systems, suitable for experimental set-ups. The model systems currently being used in this field of research are summarized in Fig. 3.

Fig. 3
figure 3

Model systems and platforms commonly used in mechanobiological experiments

Full size image

Stromal remodeling, tensional redistribution between tumor cells and the surrounding stroma and angiogenesis-driven fluid flux and shear stress in tumor-stroma: all these changes result in a significantly stiffer ECM than that of a normal tissue. Despite the classical 2D and 3D cell culture platforms, in order to genuinely mimic the mechanical nature of normal and tumor stroma, researchers have been dealing with a spectrum of biocompatible as well as widely tunable biomaterials, broadly grouped into naturally derived, synthetic and hybrid [95].

High-throughput microfluidics-based platforms have come up with three remarkable advantages: (a) the ability of co-culturing cells in a spatially controlled manner (b) generation of and control over (signaling) gradients (c) the integration of perfusion/flow. Recently, a couple of experiments on microfluidic platforms have shown that migration of the tumor cells through microtracks laid down by the migrating CAFs and macrophages, is very much dependent on the overall topography of ECM, including its porosity, stiffness etc. [96]. As mentioned earlier in this review, CSCs are softer and deformable than non-CSCs [23]. So, it is quite possible to distinguish CSCs on the basis of their unique cellular mechanics. The use of high-throughput microfluidic devices instead of high-precision but low-throughput laboratory techniques like micropipette aspiration, atomic force microscopy and optical tweezers have created a new hope of dealing regular diagnostic affairs with a mechanistic approach [97, 98]. Such screening applications must be coupled with transcriptomic analyses in order to characterize the molecular mechanisms that regulate mechanical features of the CSCs in response to surrounding tissue stiffness.

Experiments require mimicking the kind of solid stress that tumor cells experience in order to replicate the situation in vitro. For this purpose 3D model systems with adjustable mechanical properties by varying the ECM elements are commonly employed. There are few specialized instruments to artificially generate mechanical force, like the commerciallyavailable FlexCell FX-5000 compression system [99]. Scientists have invested in biomaterials and parameters like differential drug-sensitivity to enrich and perpetually culture CSCs in vitro [100].

Particle-tracking microrheology (PTMR) uses ballistic fluorescent polystyrene tracer beads and statistically analyzes their brownian motion to provide a quantitative measurement of fluctuating intracellular stresses. Employing this novel technique and a prostate cancer cell line, a group of researchers from the University of Texas has shown and quantified how ECM stiffness regulates effective intracellular stiffness of cancer cells in a 3D matrix environment [101]. The possibility of using such techniques to pinpoint CSCs and related ECM based on their unique biomechanical identities needs fervent scientific enquiry.

Recently, a research group from University of California, Santa Barbara (UCSB) has created a biocompatible magnetic microdroplet of ferrofluid oil to investigate mechanical forces in cellular microenvironments and their spatiotemporal variations in vivo. Using this technique, they have found that tissue stiffness in live developing zebrafish embryos varies along the tail bud of the animals [102]. They aim to use this platform to study the mechanisms of tumor formation in multicellular spheroids and hope to understand how abnormal biomechanics can cause or promote cancer and other diseases. These kinds of technologies can be exploited to investigate mechanical responses of CSCs in vivo.

Recent in vivo experiments, though few, have shown how ECM stiffness influences CAF activation, tumor cell invasion etc. via the mechanobiologically important molecules like LOXL2, FAK, YAP, ROCK, Cav1and actomyosin [103]. By chemotherapeutically targeting important mechanosignaling pathways or by instructing the stromal cells to cause changes in ECM composition and thereby stiffness, one can manage to get rid of the CSCs and chances of relapse [104,105,106,107]. Nowadays, scientists are also attempting to develop integrative systems biology-models in order to analyze complex mechanobiological interactions across all levels of biological organization i.e. from atomistic to systemic scales.

Conclusion

The current state-of-the-art technologies hold great promises for better understanding and prospective application of mechanobiological modulations during tumorigenesis and tumor progression both in fundamental and translational research arenas. Since the discovery of cancer stem cells, they attracted much scientific attention because of their utmost biochemical and biomechanical uniqueness amongst the whole bunch of other cells in a malignant tumor. These cells are not solitary entities. Though representing a small subset of tumor cells, cancer stem cells are supposed to constantly interact with other cells and tumor ECM. Biophysical properties like rigidity, porosity, density etc. of tumor microenvironment are actively architected by its cellular and acellular constituents. The review has critically discussed how the tumor microenvironment provides a hospitable niche for the cancer stem cells by inducing several cross-connected mechanotransduction pathways to support exhibiting their distinctive phenotype. Simply targeting the intrinsic pathways (Wnt, Notch, Hedgehog) implicated in the self-renewal and survival of CSCs can lead to their differentiation and proliferation, and may also impact normal stem cell functions. As normal tissue stroma is mechanically very different from the CSC niche, researchers are endeavoring to come up with more targeted intervention against CSCs by perturbing the pathways activated by extrinsic mechanical cues. Innovations in terms of tools, platforms and model systems for the study of mechanobiology of CSCs and their niche are also gathering pace. Present trends are encouraging and it is well expected that there will be many breakthroughs in the coming years.