Tissue mechanics in tumor heterogeneity and aggression

Abstract

Tumorigenesis ensues within a heterogeneous tissue microenvironment that promotes malignant transformation, metastasis and treatment resistance. A major feature of the tumor microenvironment is the heterogeneous population of cancer-associated fibroblasts and myeloid cells that stiffen the extracellular matrix. The heterogeneously stiffened extracellular matrix in turn activates cellular mechanotransduction and creates a hypoxic and metabolically hostile microenvironment. The stiffened extracellular matrix and elevated mechanosignaling also drive tumor aggression by fostering tumor cell growth, survival, and invasion, compromising antitumor immunity, expanding cancer stem cell frequency, and increasing mutational burden, which promote intratumor heterogeneity. Delineating the molecular mechanisms whereby tissue mechanics regulate these phenotypes should help to clarify the basis for tumor heterogeneity and cancer aggression and identify novel therapeutic targets that could improve patient outcome. Here, we discuss the role of the extracellular matrix in driving cancer aggression through its impact on tumor heterogeneity.

The heterogeneous tissue ecosystem and tumor progression and aggression

Tumors are complex ecosystems where mutated tumor cells reside within a spatially heterogeneous cellular and acellular desmoplastic tumor microenvironment (TME). The desmoplastic TME is characterized by a mixture of inflammatory, mesenchymal, neural, vascular cells, and adipocytes, that interact within a fibrotic and stiffened extracellular matrix (ECM) where they encounter assorted soluble factors and secreted vesicles [1,2] (Figure 1).

Figure 1. Extracellular matrix (ECM) stiffness as a central hub in tumor heterogeneity and progression.

Cartoon depicting how a remodeled, crosslinked, and stiffened tumor stroma impacts diverse cancer hallmarks that drive tumor heterogeneity and promote cancer aggression.

The fibrosis that characterizes the desmoplastic TME develops prior to and concurrent with malignant transformation, and contributes to tumor progression and aggression [3,4]. The resultant stiffened, fibrotic ECM stroma, which is highly heterogeneous, engages tumor cell receptors including integrins (see Glossary), syndecans, and discoidins that potentiate the growth, survival, and invasion of the transformed cells to drive their malignant progression and metastasis. The fibrotic TME promotes the malignant phenotype of the cancer cells and the stromal cells by inducing their metabolic reprogramming and stress adaptation. This includes induction of an epithelial-to-mesenchymal transition (EMT) in the cancer cells that increases their motility and survival, as well as modifications in the metabolic phenotype of the tumor cells and the infiltrating immune cells that creates an immunosuppressive microenvironment favoring treatment resistance and metastasis [5–8]. The fibrotic TME also increases the frequency of cancer stem cells (CSCs) either by supporting the growth and survival of pre-existing genetically transformed CSCs, or by promoting the transdifferentiation of tumor cells into stem-like cancer cells. CSCs contribute to intratumoral heterogeneity and cancer plasticity, which promote therapy resistance, immune evasion, metastasis and tumor recurrence [9]. A stiffened tumor ECM can additionally directly increase genomic instability and enhance reactive oxygen species (ROS)-dependent mutations within the tumor cells to increase genetic heterogeneity within the tumor tissue to promote treatment resistance and metastatic potential [10–17]. By this means, the heterogeneous stiffened TME stroma contributes to tumor aggression and therapy resistance and drives metastasis and poor patient outcome. Importantly, the resident and recruited cellular components of the TME are also highly heterogeneous and additionally respond to the fibrotic, stiffened ECM stroma, so that their modified phenotypes further contribute to tumor progression and aggression. In this review, we summarize findings pertaining to the heterogeneity of tumor cells, as well as the cellular and noncellular stroma, emphasizing the role of tissue mechanics and how it can drive cancer progression, metastasis, and treatment resistance. We finish by highlighting considerations when designing therapeutic modalities to improve the treatment response of heterogeneous cancers.

Intrinsic cues drive intratumor heterogeneity

Cancer is initiated by genetic and epigenetic ‘hits’ in the cells that generate the neoplasia [18–20]. The advent of single cell-genomic sequencing studies has permitted a high-resolution roadmap of tumor evolution and identified the impressive intratumor heterogeneity of cancer cells across diverse cancer types [20–22]. Clinical studies have further emphasized the critical role played by tumor heterogeneity in driving metastasis and treatment resistance, and have underscored the importance of clarifying molecular mechanisms driving the heterogeneous cancer phenotype [23–26].

The origins of tumor cell heterogeneity could be because cancers can arise from different cells of origin. This includes distinct cell types or cells of the same type within the same tissue that have undergone different genetic, epigenetic, and/or transcriptomic alterations [19,27,28]. These genetic alterations give rise to varying tumor subtypes/phenotypes that increase tumor heterogeneity. For example, in pancreatic ductal adenocarcinoma (PDAC), acinar cell-derived PDAC is enriched in a molecular signature characteristic of a classical subtype, while duct-derived PDAC recapitulates features of a basal-like subtype [29]. In a glioblastoma, the location of the cancer cell within the tumor correlates with the mesenchymal subtype being located within the tumor core and proneural cells in the region of the tumor migrating into the normal parenchyma [30]. In breast cancer, tumor subtype is closely correlated with expression of oncogenic mutations; PIK3CA mutations in mammary basal cells lead to luminal tumors, whereas mutations in BRCA1 and TP53 in mammary basal cells promote formation of basal-like tumors [31,32].

Genomic instability within the transformed cellular fraction contributes to heterogeneity in gene expression profiles within cancer cells within the same tumor [33]. For example, single-cell RNA sequencing (scRNA-seq) studies have highlighted the significance of cancer cell intrinsic transcriptional states in driving various cancer subtypes, again underscoring the importance of dynamic phenotypic states in the development of intratumoral heterogeneity [34–37]. Intriguingly, analyses of transcriptomic and genetic variation in colorectal tumors suggest that phenotypic plasticity within tumors can occur stochastically, even without heritable underlying epigenetic changes [38].

CSCs, which may originate from a low abundance subpopulation of long-lived tissue-resident stem cells that have sustained oncogenic mutations, and that retain the potential for self-renewal and multidirectional differentiation, also explain tumor cell heterogeneity and contribute to cancer aggression [39]. Long-lived CSCs may be responsible for tumor initiation, maintenance of tumor propagation, and resistance to, as well as relapse after, therapy [40].

CSCs may be generated through genetic/epigenetic reprogramming of oncogenic allytransformed cancer cells into tumor cells that exhibit a CSC phenotype [41]. Some CSCs may derive from de-/transdifferentiated somatic cells [42] including through induction of an EMT [43]. Intrinsic factors, such as specific genetic mutations in tumor suppressor genes or protooncogenes such as MYC, can induce CSC-like characteristics by disrupting growth control mechanisms and activating specific oncogenic pathways [44]. Both mouse cortical astrocytes and neural stem cells can serve as cells of origin for high-grade glioblastomas upon activation of the epidermal growth factor (EGF) receptor pathway and inactivation of both Ink4a and Arf. Ink4a/Arf^−/− and EGF stimulated astrocytes can dedifferentiate to Nestin⁺A2B5 ⁺ progenitor cells that sustain and promote tumor progression [45]. Inactivation of Nf1 and p53 in neurons can induce dedifferentiation of neurons to a stem/progenitor state and foster the formation of gliomas in mice [46]. Importantly, CSCs can exhibit a range of phenotypes and genotype heterogeneity. For example, glioma stem cells that are double positive for CD133 and EGF receptor possess stronger tumorigenic potential than single positive cells; while in squamous cell carcinoma, both CD34 high and low cells cause the emergence of tumors with varying invasive and proliferative potential [47,48].

The cellular stroma and tumor heterogeneity

Emerging data indicate that variation among tumors (inter- and intratumoral heterogeneity) can be influenced not only by intrinsic cancer cell regulators but also by extrinsic factors from the TME [49,50]. The evolution of genetically modified cells into malignant lesions is promoted by dynamic interactions with a heterogeneous cellular and acellular TME [18,19]. Adding to this heterogeneity, single cell gene expression analysis emphasized the diversity of phenotypes in the cellular TME, and suggest that this variability is maintained by a combination of selective and stochastic processes [51]. Further contributing to this TME heterogeneity is variability in noncellular stromal ECM composition, organization, and mechanical features [52–55].

The tumor stroma refers to noncancerous cells and noncellular components of the TME that surround and interact with cancer cells within the tissue. Cellular components include mesenchymal stromal cells (MSCs), cancer-associated fibroblasts (CAFs), immune cells (neutrophils, macrophages, T cells), nerve cells, endothelial cells, pericytes, and adipocytes. Noncellular components include the ECM, extracellular vesicles, and secreted factors such as growth factors, cytokines, chemokines, and morphogens as well as metabolites. Depending on the biological context, as well as the fact that the TME itself is highly dynamic and heterogeneous, the tumor stroma can exert both tumor-suppressive or tumor-promoting effects [56] (Figure 2).

Figure 2. Mechanisms underlying fibrotic extracellular matrix (ECM)-dependent tumor progression.

Cartoon illustrating the way in which a fibrotic stroma and a stiffened ECM drive tumor progression and aggression. (A) Cancer-associated fibroblasts (CAFs), tumor-associated macrophages (TAMs), and activated fibroblasts deposit, remodel, and crosslink the tumor ECM to promote tumor cell growth, survival, and an epithelial-to-mesenchymal transition (EMT) that expands cancer stem cell frequency. (B) A fibrotic, stiffened tumor stroma induces hypoxia to promote cancer cell stress adaptation and metabolic rewiring that enhances their growth and survival and treatment resistance. (C) A stiffened ECM enhances the recruitment of protumoral myeloid and lymphoid cells and drives immune reprogramming to facilitate tumor escape from antitumoral immunity. (D) The cancer cells and the cells from the tumor microenvironment, including infiltrating immune cells, secrete soluble factors (cytokines, growth factors, exosomes, metabolites) that stimulate stromal fibroblasts and mesenchymal cells to synthesize, secrete, remodel, and crosslink ECM that drives tumor heterogeneity and fosters tumor progression and aggression. Abbreviations: CCL2, C-C motif chemokine ligand 2; FGF, fibroblast growth factor; IL-6, interleukin-6; TGF-β, transforming growth factor beta; VEGF, vascular endothelial growth factor.

Stromal fibroblasts and tumor heterogeneity.

Stromal fibroblasts, which are the major cellular TME population in the tumor, are specialized mesenchymal cells that synthesize, deposit, degrade, reorganize and crosslink ECM proteins to shape the biochemical and biomechanical properties of the ECM. Stromal fibroblasts secrete soluble factors that stimulate immune cell infiltration and drive the growth, survival, and invasion of the tumor cells and endothelial cells, nerves, adipocytes, immune cells, and MSCs. A secreted factor that is a major activator of stromal fibroblasts is transforming growth factor beta (TGF-β) [57]. TGF-β is secreted and stored in the ECM as a latent form that can be activated through either protease activity or mechanical force. Thus, a stiffened ECM, via ligation of integrins, promotes TGF-β activation to create a positive feedback loop that amplifies pro-tumorigenic/pro-fibrotic processes, including ECM remodeling and cytokine production.

Activated stromal fibroblasts can transform into CAFs, which are spindle-shaped highly contractile mesenchymal cells that remodel the ECM and secrete tumorigenic and immunomodulatory cytokines. CAFs are transcriptionally heterogeneous, as revealed by single-cell technology [58–61]. CAF classification across tumor types converges on three main subtypes that include: (i) myofibroblastic CAFs (myCAFs) that are highly contractile and express genes related to the production and secretion of ECM components, (ii) inflammatory CAFs (iCAFs) which are cells that are enriched for the expression of cytokine and chemokine genes, and (iii) antigen-presenting CAFs (apCAFs) which exhibit antigen-presenting properties (e.g., MHCII expression) [62]. Each of these CAF phenotypes can exert independent and overlapping effects on the TME and tumor progression. The TME also contains MSCs, which are multipotent non-hematopoietic stem cells that not only have immunomodulatory properties, but, like myCAFs, modify the ECM. MSCs contribute to tumor progression through their ability to secrete soluble factors that directly promote tumor cell growth and invasion, stimulate immune infiltration and modulate antitumor immunity by depositing, remodeling, and stiffening the ECM [63–65].

myCAFs and MSCs synthesize and deposit abundant quantities of ECM into the TME, including fibrillar collagens, fibronectin, tenascin, multiple proteoglycans, and hyaluronic acid [66]. myCAFs and some MSCs also remodel these ECM proteins through YAP-Rho-ROCK-Myosin II actomyosin signaling and contractility, and by secreting and activating various metalloproteinases (MMPs) and cathepsins. Ultimately, myCAFs and MSCs reorganize the ECM into linearized bundles through scaffolding with facet collagens such as collagen XXII, and through interactions with other ECM proteins, including fibronectin, as well as via crosslinking through lysyl oxidases (LOX) and lysyl hydroxylases [67–72]. The net result is an extensively remodeled, stiffened, fibrotic stromal ECM that contains aligned collagenous/fibronectin highways upon which tumor cells can be seen migrating [3,73,74]. Through these linearized collagenous/fibronectin bundles, stromal mesenchymal cells exert long-range forces that influence the direction of tumor cell migration and enhance endothelial cell and immune cell invasion [75]. The stiffened, remodeled ECM also potentiates the activation of TGF-β to foster the induction of EMT in the cancer cells [76]. A stiffened TME in colorectal cancer potentiates TGF-β and other family ligand (Activin A) activation to promote EMT that drives metastasis [77]. The stiffened ECM also enhances the activity of G-protein-coupled receptors (GPCRs) and receptor tyrosine kinases such as epidermal growth factor receptor (EGFR) in cancer cells and stromal cells. ECM stiffness influences tumor cell growth, survival and migration through integrin and discoidin receptor mediated adhesion-linked mechanotransduction [6,68,78–82]. By stiffening the ECM, the mesenchymal stem cells and myCAFs also increase the solid stress in tumors that impedes lymph drainage and can increase capillary pressure [83]. The net result of myCAF and MSC activity is the recruitment and transdifferentiation of more CAFs and MSCs, and an acceleration of tumor progression through direct interactions with the genetically/epigenetically modified cancer cells [67,84,85]. Thus, MMP-dependent ECM degradation of the basement membrane surrounding premalignant lesions fosters the invasion of cancer cells into the surrounding interstitial matrix to drive malignant transformation. MMP activation also releases critical growth factors that promote cell growth and survival and drive the migration that is necessary for vascular and lymphatic extravasation that permits metastasis. Initially, the stiffened tumor ECM promotes angiogenesis by driving endothelial cell proliferation and migration [86]. However, excessive ECM deposition and fibrosis eventually compromise blood vascular integrity to alter endothelial/pericyte coverage and impede tumor vascularization and blood flow that leads to hypoxia [87]. Not surprisingly, hypoxia drives tumor aggression and metastasis, and has itself been shown to stimulate an EMT in cancer cells through activation of hypoxia inducible factor [88]. By this means, myCAFs and MSCs contribute to abnormal interstitial fluid flow and elevated solid stress, altered shear stress, and compromised tissue perfusion that contribute to tumor progression and heterogeneity through a feedforward cycle.

In summary, a reorganized and stiffened ECM by myCAFs and MSC contributes to tumor heterogeneity and malignancy by: (i) enhancing the activation of key oncogenic-linked pathways including Rho/ROCK, YAP/TAZ, PI3K/AKT, MAPK/ERK, and FAK signaling, mediated either through activation of cellular tension or through specific ligation of key cellular receptors such as fibronectin-ligation of ${\alpha }_{5{\beta }_1}$ and/or ${\alpha }_{\text{v}{\beta }_3}$ integrins or via engagement of discoidin receptors or syndecans [89–93], which promotes malignant transformation, growth, survival, invasion and metastasis; (ii) inducing angiogenesis through stimulating pro-angiogenic factors, such as vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF), or altering the behavior of endothelial cells, to modify oxygen and nutrient supply in the tumor [94]; (iii) creating migratory networks, or fostering the formation of invadopodia, ultimately facilitating the invasion of cancer cells and metastatic spread. Invadopodia also contribute to ECM degradation and the formation of pores/tunnels for cancer cell migration. By forging associations with focal adhesion complexes, invadopodia provide mechanical links that enable cancer cells to exert force during invasion and stimulate migration-related signaling pathways [55]. Other effects include (iv) inducing hypoxia and compromising drug delivery through vascular dysfunction [95]; (v) activating signaling pathways that promote drug and chemotherapy resistance [96,97]; (vi) inducing TGF-β activation that potentiates ECM deposition, remodeling and crosslinking, that represses antitumor immunity and induces an EMT [76].

ECM and tumor heterogeneity.

Given the heterogeneity of stromal fibroblasts and MSCs it is not surprising that the composition, organization, and stiffness of the tumor ECM is itself highly heterogeneous, and this heterogeneity also contributes to tumor heterogeneity [52,53]. Atomic force microscopy (AFM), shear rheology, and unconfined compression measurements revealed that the stiffness and viscoelasticity of the tumor stroma increase progressively during malignant transformation and correlate with tumor grade and aggression [53,68,70,81,98]. Thus, pancreatic intraepithelial neoplasia (PanINs) are surrounded by a stiffened, fibrotic stroma that further increases with the development of pancreatic adenocarcinoma [82]. Similarly, the stroma adjacent to ductal carcinoma lesions are stiffer and exhibit greater viscoelasticity than that measured in normal breast tissue, and these biophysical features increase further once the tissue transforms [99–102]. Breast tumor aggression represented by comparison between luminal A and B, Her2-positive, and triple-negative breast cancer (TNBC) correlates with increased ECM deposition, remodeling, and stiffening, and collagen crosslinking [53,69,70]. Importantly, scanning AFM measurements have revealed that the stiffness of the tumor stroma is quite variable [52,53]. For instance, the core of breast tumors is very compliant, whereas the invasive front of breast tumors is significantly stiffer [53]. A machine-learning convolution method entitled STIFMaps, which permits a survey of the stiffness of the tissue stroma reflecting collagen architecture, revealed a profound heterogeneity in the stiffness of the breast tumor stroma. STIFMaps analysis of human breast cancers identified heterogeneously stiff regions within the tumor tissue that correlated with elevated epithelial mechanotransduction (elevated activated β₁ integrin) and an EMT (positive SLUG transcription factor staining) [52]. Tumor heterogeneity, as revealed by STIFMap correlation with EMT markers, was able to predict poor patient outcomes in women with Her2⁺ breast cancers [52].

Tumor heterogeneity also promotes metastasis by enhancing the generation of the pre-metastatic niche [103,104]. Shedding of secreted factors and extracellular vesicles from the primary tumor and stromal cells triggers ECM remodeling, vascular permeability, angiogenesis, immunosuppression, and recruitment/activation of stromal cells [105]. A stiffened ECM also increases secretion of exosomes to foster liver metastasis [106]. Activation of resident normal fibroblasts in the metastatic tissues by extracellular vesicles similarly activates CAFs and induces ECM remodeling [107]. The remodeled ECM alters the composition, organization, and stiffness of the metastatic niche that enhances tumor cell growth and survival, and modulates the recruitment and activation of immune cells and endothelial cells [55].

The remodeled, stiffened, fibrotic tumor ECM also significantly influences antitumor immunity and – because the ECM is highly heterogeneous – this, too, will contribute to immune cell heterogeneity. To begin with, elevated tissue tension activates pSTAT3 to increase the level of multiple cytokines that stimulate myeloid cell infiltration and activation [82]. The infiltrating immune cells display an amoeboid migration and, in response to chemokine stimulation, use the oriented ECM bundled fibers as substrates for migration [108]. The infiltrating myeloid cells can, in turn, induce additional ECM remodeling to further stiffen the stroma by secreting abundant quantities of TGF-β, which activates CAFs and induces lysyl oxidases and lysyl hydroxylases [70]. Moreover, TGF-β has profound immunosuppressive activity [109]. In a mouse model of PDAC, TGF-β signaling in normal (universal) fibroblasts is essential for the development of LRRC15⁺ myCAFs, which are drivers of tumor progression and immunosuppressio n [110]. The TGF-β pathway has been reported to be rhythmically regulated by the circadian cloc k [111]. Likewise, circadian dysregulation of the TGF-β/SMAD4 and other oncogenic pathways have been associated with carcinogenesis and metastatic spread [112–114].

The myeloid cells within the stiffened, fibrotic tumor can also become metabolically altered and transition into a pro-tumor, ECM synthetic phenotype. The metabolically rewired, ECM synthetic tumor-associated macrophages (TAMs) eventually deplete the TME of arginine, which is critical for CD8⁺ T cell viability, and increase levels of the metabolite ornithine that compromises CD8⁺ T cell viability and function to reduce antitumor immunity and compromise checkpoint inhibitor therapy response [115]. In addition, the highly dense, compacted ECM itself forms a physical barrier that can actually hamper immune cell migration and compromise the distribution and concentration of chemoattractants, thereby altering the navigation or retention of immune cells within the TME [116]. In support of this contention, recent findings illustrated how a stiff ECM can impede T cell migration and that reducing collagen crosslinking in PDAC and mammary tumor models permits CD8⁺ T cell infiltration and reduced primary tumor progression [117]. Indeed, the physical characteristics of the highly heterogeneous TME ECM can both enhance and abrogate T cell migration and activation, as well as promote or block pro-tumoral macrophage recruitment and activation [118].

The stiffened, dense tumor stroma can also contribute to tumor cell heterogeneity by promoting genomic instability to induce genomic alterations. For example, tumor cells migrating through a dense, stiffened ECM exhibit nuclear rupture that favors genomic alterations, possibly by increasing replication stress [10,13,16,119]. In keeping with this paradigm, tumor cells forced to migrate through rigid small pores could accumulate genetic anomalies, including chromosomal losses and gains reminiscent of the acquisition of genomic instability [13,15,120]. Similarly, tumor cells experiencing high solid stress demonstrated compromised mitotic integrity that was accompanied by sustained chromosomal anomalies [13,16]. Chronic inflammation, which is potentiated by tissue fibrosis and ECM stiffening, can additionally increase the mutational load to drive malignant transformation and elevate tumor cell heterogeneity through myeloid-mediated ROS-stimulated DNA damage. Consistently, the mammary epithelial cells in the breast tissue from women with high mammographic density and a stiffer stroma, contain a high frequency of DNA double-strand breaks – possibly due to elevated levels of ROS secreted by the infiltrating macrophages [121–123]. For instance, experimentally elevating ROS levels secreted by myeloid cells through genetically engineered targeted knockdown of the GPX4 enzyme increased the mutational load in intestinal cells and accelerated malignant transformation in the colon of mice lacking adenomatous polyposis coli (APC) [124]. Through these mechanisms, the inflamed and mechanically challenged heterogeneous TME can promote malignant transformation, tumor progression and aggression, and enhance tumor cell heterogeneity.

myCAF and iCAF-secreted factors also actively contribute to tumor heterogeneity by influencing cancer cell proliferation and migration, as well as via recruitment and activity of immune cells, and angiogenesis. While myCAFs secrete factors (e.g., TGF-β, FGF, VEGF) involved in tumor cell survival, invasion and angiogenesis, iCAFs secrete factors (e.g., IL6, IL8, CCL2, CXCL12) that contribute to immune cell recruitment and function [62]. Thus, the combined signaling of iCAFs and myCAFs impacts localized activity and spatial regulation of cancer cells, immune cells, or other stromal cells by the release of cytokines and growth factors that have been trapped in the ECM and are released upon ECM remodeling and proteolytic breakdown. Moreover, many ECM proteins are composed of cytokine-/chemokine-like domains that are liberated during proteolytic cleavage and, as so-called matrikines, regulate many biological processes related to tumor progression [55].

The cellular TME, CSCs, and tumor heterogeneity

The cellular TME can sustain the development and orchestrate the heterogeneity of CSCs to further promote tumor heterogeneity [125]. The interplay between CSCs and the cellular TME may even create a reinforcing loop of tumor-promoting factors that enhance CSC fitness and tumor progression through the generation of tumor heterogeneity. For instance, CAFs have been reported to regulate tumor-initiating cells in the liver through paracrine secretion of hepatocyte growth factor (HGF) [126]. In gliomas, platelet-derived growth factor was shown to promote the massive expansion of non-neoplastic glial progenitor cells through both autocrine and paracrine signaling [127]. In these studies, the expanded progenitor cells were found to accumulate further genetic alterations characteristic of glioma cells and to form gliomas when transplanted into mice [128]. In squamous cell carcinoma, oncogenic Ras activation was shown to lead to aberrant communication between cancer cells and their TME (perivascular immune and stromal cells) that resulted in activation of pathways that promoted cancer progression [129]. Similarly, stem-like melanoma cells have been found to preferentially localize near blood vessels, where endothelial cells were found to support their dedifferentiation and stimulate melanoma growth through NOTCH signaling [130].

Immune cells, including macrophages, generate permissive niches that promote growth, survival, and ultimately favor the expansion of CSCs. Intratumoral, resident and monocyte-derived macrophages were shown to generate a mosaic of distinct TAMs that exerted diverse functional roles in driving tumor progression, including supporting CSCs and their niche through the production of cytokines (IL-6, IL-1β, TNF-α, TGF-β) and chemokines (CCL2, CCL5, CCL8) [131–140]. TAMs also contribute to biological processes such as EMT and stemness maintenance. For instance, TAMs participate in tumor stroma remodeling by producing matrix-degrading enzymes and ECM components. TAMs also produce heparan sulfate proteoglycans, collagenous fibers, fibronectin, osteopontin, and the cross-linker F13a1 [36,141,142]. The communication between TAMs and CSCs involves juxtacrine signaling, activating critical pathways such as SHH, NOTCH, STAT3, PI3K/AKT, WNT/β-catenin, NANOG, and NF-kB [143–151]. These pathways are vital for the generation and maintenance of CSCs. Ultimately, the interaction between TAMs and CSCs can support stemness features in cancer cells to facilitate their growth and metastasis.

Forcing cancer cell stemness

Evidence suggest that the stiff, fibrotic tumor ECM can directly and indirectly promote cancer cell stemness and can serve as a viability niche for CSCs to favor tumor heterogeneity and cancer aggression [152,153]. To begin with, cell and tissue tension promote CSC growth and survival, and can foster a stem-like phenotype in tumor cells. Furthermore, ECM–CSC interactions can promote the CSC properties of cancer cells at the primary tumor by engaging specific transmembrane cellular receptors [154]. Similarly, a modified tissue ECM at the pre-metastatic niche can sustain the growth and survival of disseminated CSCs [155]. ECM stiffness, induced by excess ECM deposition/degradation, remodeling and crosslinking, can also influence CSC self-renewal and differentiation. For instance, soft fibrin matrices were found to promote self-renewal in melanoma CSCs [156–158], while hepatoma cells and breast cancer cells displayed enhanced stemness in a stiff matrix [159–161], and a stiff matrix enhanced CSC marker expression in breast CSCs [162,163]. Indeed, hydrogel networks with increasing tension can reprogram differentiated cancer cells from various cancer types into CSCs [164]. More recently, studies revealed how a stiff tumor ECM was able to increase CSC frequency in patient tissues, as well as in experimental murine models, by enhancing EGF receptor-dependent extracellular signaling regulated kinase (ERK) activation, which promoted progesterone-induced RANK signaling [121]. Additional mechanisms whereby the ECM can drive and sustain CSCs include: (i) inducing EMT to drive CSC phenotypes through signaling pathway activation, as was shown for collagen I, collagen XVII, and laminin-5 ligation-mediated activation of signaling that promoted EMT and stemness in cancer cells, possibly by activating Wnt [165–167]; (ii) promoting CSC self-renewal by engaging integrins to activate specific stemness-promoting transcriptional programs, as was shown by engagement of collagen I, laminin and fibronectin receptors to maintain tumor cell stemness [168–170]; (iii) functioning as a reservoir for growth factors that aid in CSC survival, as was demonstrated for matrix metalloprotease-released factors such as ciliary neurotropic factor and stem cell factor that critically promoted CSC self-renewal and chemoresistance [171,172]; (iv) modulating metabolic reprogramming in CSCs to facilitate CSC self-renewal and drug resistance through tension-enhanced engagement of integrins and induction of focal adhesions, as well as downstream signaling that triggers autophagy [71,173,174]. The findings demonstrate how CSC phenotype and diversity can be significantly impacted by the biochemical and biophysical properties of ECM of the tumor.

Tissue mechanics interact with CSCs

The interplay between genetic, epigenetic, and TME heterogeneity coordinates with tissue mechanics to drive cancer progression. For example, the genotype of tumor cells can increase epithelial cell contractility to induce ECM remodeling that contributes to tumor fibrosis [82,175]. Moreover, tissue-level forces can drive epigenetic modifications in tumor cell chromatin to alter gene expression through ECM stiffness and elevated tissue solid stress [176,177]. The distinctive mechanical attributes of tumor tissue stroma can also affect gene expression of tumor cells, including CSCs. The decision for self-renewal, differentiation, dormancy, or quiescence are intimately and reciprocally affected by mechanical force. For example, fluid shear stress application to cultured breast cancer MCF7 cells was shown to upregulate a number of CSC markers [178]. Curiously, ovarian CSCs have a softer elastic modulus than their more differentiated neoplastic precursors and aggressive ovarian cancer cell counterparts, suggesting that they may be protected from high solid stress. Indeed, ovarian CSCs undergo dramatic cytoskeletal remodeling that increases their elastic modulus upon antitumor agent treatment with sphingosine [179]. Regardless, the biomechanical properties of CSCs are unique, plastic, and may serve as a promising target for future therapeutic manipulation.

The CSC niche maintains and promotes CSC function, which is sustained by dynamic interactions with components of the TME (ECM, stroma, extracellular vesicle secretion and signaling), including the mechanical features of the tumor tissue. For example, the physical features of the TME, including ECM stiffness, alter CSCs by activating mechanotransduction. Consistently, blocking the IL-8 receptor CXCR1 in human breast CSCs inhibited mechanosignaling through FAK/AKT/FOXO3A to reduce the viability of CSCs [180]. The well-known mechano-activated pathway, YAP/TAZ, can regulate breast cancer aggression and chemotherapy resistance, likely by promoting the survival of CSCs [181]. Genetic mutations commonly found in various tumor types also regulate CSC function, as was shown in studies suggesting that TGF-β, TP53, and PTEN regulate CSC maintenance and differentiation [182–185]. In summary, the complexity of CSC viability, phenotype, and function is tightly coordinated by several factors, including genetic, cellular microenvironmental and biomechanical properties of the TME.

Therapeutic resistance and targeting of the TME

The heterogeneous fibrotic TME promotes malignant transformation, tumor progression and aggression, and compromises treatment. Substantial effort has been exerted towards developing therapies to address the heterogeneity including targeting CAFs and ameliorating tumor fibrosis to improve therapy (Figure 3). Nevertheless, while some studies have reported protective effects of CAF deletion, other studies have documented deleterious consequences following CAF targeting, suggesting that CAFs exert both tumor-promoting and tumor-suppressive functions depending on subtype, biological context, and timing of therapeutic intervention [62]. For instance, ablating alpha smooth muscle actin-positive fibroblasts in an experimental model in which KRas/TGF-β receptor 2 was targeted to the pancreatic epithelium to create aggressive PDACs, was found to potentiate the malignant phenotype [186]. These contradictory therapeutic results are likely due to the inherent heterogeneity of CAFs. For instance, while experimental studies in murine models demonstrated the ability of the CAF hedgehog inhibitor to significantly reduce pancreatic adenocarcinoma fibrosis and potentiate chemotherapy response clinical translation failed [187–190], subsequent studies using genetic models to ablate hedgehog signaling in stromal fibroblasts demonstrated rebound vascularity that promoted tumor cell growth and invasion [191,192]. Studies further demonstrated that the hedgehog receptors are confined to the alpha smooth muscle positive subset of CAFs while retaining the activity of other CAF subpopulations, emphasizing the importance of understanding the role played by distinct CAFs in cancer [62,193]. However, selective targeting of CAF subtypes is difficult largely due to their immense plasticity, overlapping marker expression, and lack of functional classification. Ultimately, a better understanding of functional subtypes and dynamic interaction with cellular and noncellular TME components, as well as identification of predictive CAF biomarkers, will be necessary before specific and effective targeting of CAFs, CAF-ECM units, or iCAFs or even MSCs can be implemented to treat cancer patients.

Figure 3. Extracellular matrix (ECM) and microenvironment-focused cancer therapies: targets and strategies.

(A) Schematic depicting therapeutic opportunities targeting the ECM in cancer, including inhibition of ECM deposition, disruption of ECM crosslinking, and modulation of mechanosignaling induced by tumor forces. The inserted table summarizes known inhibitors, their targets, and mechanisms of action. (B) Schematic illustrating therapeutic strategies targeting the tumor microenvironment, including immunotherapies and anti-angiogenic therapies. The inserted table lists known tumor microenvironment therapies, their targets, and mechanisms of action. Abbreviations: HA, hyaluronic acid; LOX, lysyl oxidase; MMP, metalloproteinase; PD1, programmed cell death; TAMS, tumor-associated macrophages; VEGF, vascular endothelial growth factor.

Similarly, despite encouraging results in murine models, success in clinically translating anti-fibrotic and/or ECM inhibitory agents has yielded mixed results (Table 1). Contributing to the failures is that the acellular TME is also highly heterogeneous and consequently can provide a spectrum of tumor-suppressive or -supportive micro niches that modulate tumor cell/CSC growth and survival, as well as immune evasion. The TME also evolves over time, fostering tumor cell adaptation that can promote cancer progression, aggression, and treatment resistance.

Table 1.

Clinical trials targeting ECM molecules and mechanosignaling pathways for solid tumors

Target	Agent	Agent type	Solid tumor type	Trial status	Refs
ECM
MMP-9	Andecaliximab (GS-5745)	Monoclonal antibody targeting MMP-9	Breast, colorectal, pancreatic, esophagogastric, and non-small cell lung cancer Gastric or gastroesophageal junction adenocarcinoma	Phase 1 – Completed Phase 2 – Completed	NCT01803282 NCT02864381
MMP-9, MMP-1, MMP-2, MMP-14 and MMP-7	Marimastat	Drug, broad-spectrum MMP inhibitor	Non-small cell lung cancer	Phase 3 – Completed	NCT00002911
MMP-2, MMP-3, MMP-9, MMP-13, and MMP-14	Prinomastat (AG3340)	Drug, broad-spectrum MMP inhibitor	Glioblastoma Multiforme	Phase 2 – Completed	NCT00004200
MT1-MMP	BT1718	Drug, inhibitor of MT1-MMP	Advanced solid tumors	Phase 1/2 – Completed	NCT03486730
ADAM10 and ADAM17	INCB7839	Drug, inhibitor of ADAM10 and ADAM17	Pediatric high-grade gliomas	Phase 1 – Ongoing	NCT04295759
LOXL2	Simtuzumab (GS-6624)	Monoclonal antibody targeting LOXL2	Pancreatic cancer Colorectal cancer	Phase 2 – Completed Phase 2 -Terminated	NCT01472198 NCT01479465
TGF-β	HCW9218	Bifunctional TGF-β Antagonist/IL-15 protein complex	Advanced pancreatic carcinoma	Phase 1b/2 – Ongoing	NCT05304936
TGFβ receptor I	Galunisertib (LY2157299)	Drug, inhibitor of TGFβ receptor I	Prostate cancer Hepatocellular carcinoma	Phase 2 – Ongoing Phase 2 – Completed	NCT02452008 NCT01246986
Mechanosignaling pathways
αv-integrin	Intetumumab (CNTO 95)	Anti-αv-integrin monoclonal antibody	Metastatic hormone refractory prostate cancer	Phase 2 – Completed	NCT00537381
α5β1 integrin	Volociximab (M200) ATN-161	Anti-α5β1 integrin monoclonal antibody small peptide antagonist of integrin α5β1	Pancreatic cancer Ovarian cancer, primary peritoneal cancer Brain and CNS tumors	Phase 2 – Completed Phase 1/2 – Completed Phase 1/2 – Completed	NCT00401570 NCT00635193 NCT00352313
α2 integrin, α3 integrin, α5 integrin, and β1 integrin	E7820	Drug, inhibitor of α2/α3/α5/β1 integrin	Locally advanced or metastatic colon or rectal cancer	Phase 1/2 – Completed	NCT01347645
TEAD–YAP	Verteporfin	Drug, inhibits TEAD–YAP association	Glioblastoma	Phase 1/2 – Recruiting	NCT04590664
TEAD	VT3989	Small-molecule inhibitor of the auto-palmitoylation of TEAD	Solid tumors, mesothelioma	Phase 1/2 – Recruiting	NCT04665206
FAK	Defactinib	Small-molecule FAK inhibitor	Gastric and stomach cancer Pancreatic ductal adenocarcinoma	Phase 2 – Recruiting Phase 2 – Recruiting	NCT06487221 NCT03727880
ROCK/AKT	AT13148	Small-molecule AGC kinase inhibitor	Advanced solid tumors	Phase 1 – Completed	NCT01585701
ERK1/ERK2	Ulixertinib (BVD-523)	Drug, inhibits ERK1/2	Pancreatic cancer	Phase 1 – Terminated	NCT02608229

Cancer stemness and therapy resistance

As malignancy ensues, the TME evolves and becomes increasingly hostile (acidic pH, hypoxia, ECM stiffness, metabolite limitations) to normal cell viability and differentiation. However, genetically mutated cancer cells that stress-adapt to this aberrant TME, and CSCs that are uniquely poised to withstand noxious TMEs, are not only able to grow and survive in this microenvironment but can also evade antitumor immunity and resist therapy.

Chemotherapy and radiation treatment induce a high level of tissue stress and can expand tissue CSC levels and/or the transdifferentiation of cancer cells into a stem-like phenotype. For example, prior studies showed that chemotherapy-induced extracellular vesicles can promote a cancer stemness phenotype [194], while CSC-derived exosomes could drive drug resistance in adjacent cancer cells [195]. Other studies have identified CAFs and distinct ECM components essential for creating an environmental niche that favors CSC survival and resistance to radiotherapy and chemotherapy [196]. A stiff ECM induces the EMT transdifferentiation of cancer cells into stem-like cells or CSCs and promotes their growth and survival [152]. Generally, cell-surface receptors/molecules (e.g., integrins, CD44, ion channels) detect mechanical cues and transmit these signals to activate downstream signaling (e.g., FAK, Rho-ROCK, YAP-TAZ) that ultimately drives cell growth, survival, and invasion. Mechanotransduction can expand the frequency of pre-existing tissue stem cells, or it can promote an EMT in cancer cells to drive their transdifferentiation into CSCs [52,121]. Thus, tissue-specific differences in sensitivity and adaptability to mechanosignaling may dictate impact on CSCs and are highly relevant for therapeutic targeting.

Regardless of which factors induce CSCs, such adaptive diversification of tumor cells is reflected by intratumoral phenotypic and genotypic heterogeneity and poses a significant therapeutic challenge. Indeed, a CSC gene signature has been associated with acute myeloid leukemia prognosis and basal-like human triple-negative breast cancers with a stem-like phenotype have been reported to show the worst outcomes [197,198]. The mechanistic basis for CSC resistance to chemotherapy and radiotherapy includes: (i) expression of multidrug resistance transporters; (ii) improved DNA repair and pro-survival capacity; (iii) induction of a prometastatic phenotype (EMT, evasion of anoikis, adaptation to hypoxia); (iv) increased autophagic activity; (v) increased glycoproteins that enhance the glycocalyx; and (vi) immune escape. For example, treatment-resistant gliomas with a CSC phenotype showed high expression of BCRP1, the ABC drug transporter [199], had elevated notch signalin g [200], and expressed higher galectin-1, a major glycoprotein [201]. Moreover, the ovarian CSC population was maintained by downregulation of the DNA damage binding protein 2 through increased expression of the microRNA miR-328 [202].

Current therapeutic approaches designed to target CSCs have relied upon targeting stemness-associated factors/signaling pathways (e.g., Shh, Wnt/β-catenin, Notch, NF-κB, Rho/ROCK). Unfortunately, these factors/signaling pathways are not confined to CSCs, which severely complicates treatment dosing and toxicity. Moreover, efficient therapeutic windows for targeting CSCs remain unclear. Hence, systematic studies and a comprehensive understanding of CSC-specific markers and targets, and development of effective combination therapies, or optimization of drug dosing and timing will be crucial to develop successful therapies that improve patient outcomes.

Targeting of the ECM

Dynamic interactions between the mutated cancer cells and the fibrotic, stiffened TME ECM are a conserved feature of all solid cancers and pose a serious impediment to effective cancer treatment. Accordingly, a comprehensive understanding of the features of the tumor ECM and tumor cell mechanotransduction has potential to lead to the discovery of therapeutic targets and development of innovative treatment strategies [203]. What is already clear from prior studies is that a fibrotic, stiff tumor stroma can contribute to treatment resistance through diverse mechanisms including: (i) directly modulating cancer cell growth, survival, invasion and transdifferentiation into CSCs; (ii) compromising drug delivery by impeding the vasculature and creating hypoxia; and (iii) inhibiting antitumor immunity by preventing T cell infiltration and driving a pro-tumor myeloid cell phenotype. Current therapeutic strategies aimed at targeting biomechanical features of the TME include factors that stimulate ECM synthesis, deposition, remodeling and crosslinking, for example, TGF-β inhibition [85]. Pharmacological approaches to inhibit TGF-β or TGFBR inhibition have shown beneficial responses in animal models [85] and may improve radiation response in patients [204,205] (NCT02581787, NCT01401062). However, many of these inhibitors have deleterious systemic effects, including compromising cardiac function and/or eliciting rebound tumor cell growth and dissemination, preventing their clinical adoption [206]. Another ECM targeting strategy, that was exploited early on, was to inhibit ECM degradation through repression of the activity of the MMPs implicated in tumor cell metastasis. Once again, however, this approach had catastrophic consequences when applied clinically to treat patients – likely in part due to the lack of specificity of many of the MMP inhibitors used and the diverse pro- and antitumor activity of diverse tissue MMPs [1]. Other approaches have included targeting many of the collagen crosslinking enzymes, including LOX and lysyl oxidase like-2 (LOXL-2) which showed impressive antitumor and antimetastatic phenotypes in murine models of breast, pancreas, and lung cancer [68,207,208]. Similarly, however, early clinical trial results were either nonconclusive or deleterious, and progress has been hampered. More recently, a pan LOX inhibitor was developed, and thus far, Phase 1 and 2 clinical trials have shown promising results in treating PDAC patients with high optimism pending for results obtained in the Phase 3 clinical trials [209]. Additionally, anti-ECM treatments have included targeting the high hyaluronic acid levels in PDACs, which increases tumor tissue solid stress and severely impedes vascular integrity and drug delivery. Hyaluronidase treatment showed spectacular results in a murine model of PDAC and promising effects in subsets of PDAC patients with high hyaluronic acid tumor levels but was ultimately discontinued [210–212]. A major impediment to all of these approaches is that many carry non-negligible risks such as interfering with physiological/tumor-suppressive functions of TGF-β, uncontrolled release of signaling factors from the ECM, and overstimulation of inflammation or pro-tumorigenic pathways, as well as accelerated cancer cell migration and invasion [85]. Optimizing treatment windows (across tumor stage or circadian day) and personalizing treatment strategies (stiff versus soft matrix environments) may mitigate some of these risks but will need to be further investigated. Other strategies aimed at therapeutically targeting the ECM include focusing on ECM proteins such as specific fibronectins (fibronectin extra domain A and B) [85,213,214], and receptors such as adhesion proteins (e.g., integrins) [215,216]. For example, a humanized ${\alpha }_{\text{v}{\beta }_3}$ integrin function blocking antibody showed exciting efficacy for the treatment of aggressive glioblastoma first in murine models and thereafter in clinical trials to treat glioblastoma (GBM) patients [217]. However, the therapy was effective only in a subset of GBM patients and was therefore not advanced further. In the future, systematic comparison of ECM organization and biomechanical features across cancers, as well as development of materials and model systems that better recapitulate ECMs in vitro, should help to identify treatments that target either the ECM or its receptors to improve anticancer therapy.

Targeting of CAFs

The therapeutic resistance in tumors induced by subsets of CAFs is primarily due to their ability to deposit and reorganize the ECM into fibrotic, physical barriers surrounding the tumor that both inhibit immune cell infiltration and impede the tumor vasculature and lymphatic drainage. The net result is a hypoxic, physically constrained tumor with high solid stress that creates stressadapted resistant CSCs and prevents drug delivery and tumor regression. These CAFs also secrete abundant cytokines and growth factors and activate latent TGF-β to stimulate cancer cell growth, survival, and invasion, drive cancer cell transdifferentiation into CSCs, and reprogram myeloid cells to become immune suppressive [218–220]. Preclinical studies have identified a plethora of anti-CAF treatments including most notably hedgehog inhibitors that effectively reduced tumor fibrosis and potentiated chemotherapy responsiveness in murine models of PDAC [187]. However, early clinical adoption proved unsuccessful, emphasizing the importance for a better understanding of pro- and anti-tumorigenic functions of CAFs as well as their evident heterogeneity [221]. Indeed, CAF heterogeneity is a major challenge for designing appropriate targeted anti-CAF treatments due largely to: (i) the lack of specific markers and signaling pathways; (ii) CAF plasticity and dependance on the biological context; (iii) the dual functional role of CAFs within the TME. These findings suggest that fine-tuning the balance of CAF composition within the TME might improve therapeutic success. For example, targeted deletion of LRRC15⁺ myCAFs in KPR tumors recalibrated the CAF populations toward a more normal (universal) nonmalignant fibroblast state that improved outcome in an experimental model [110]. Other studies claim that pharmacological approaches can reprogram/shift CAFs towards myofibroblastic or inflammatory subtypes that can restrain tumor growth or increase response to immunotherapy, although further studies will be required to confirm and expand these findings [222]. CAFs can also modulate antitumor immunity by secreting factors that inhibit immune cell activity. To overcome this phenotype, combination therapies have been developed that target CAF-derived immunosuppressive signals such as CXCL12 and TGF-β in combination with immune checkpoint blockers and have potential to improve the clinical efficacy of immunotherapy [223]. For example, combination therapy targeting the interaction of FAP⁺ CAF-derived CXCL12 with CXCR4 and PD1 blockage expanded the benefit of chemotherapy in patients with PDAC in a Phase 2 clinical trial [224]. Accordingly, targeting the interactions between dendritic cells, TAMs, and other myeloid cells has potential to offer exciting new avenues to design CAF-specific immunotherapies to improve the efficacy of cancer patient treatment (Table 1).

Concluding remarks and future perspectives

The fibrotic, stiff tumor ECM plays a pivotal role in regulating cancer aggression by activating mechanotransduction to promote tumor cell growth, survival, and invasion, and by enhancing tumor heterogeneity and cancer cell stemness, which are key drivers of metastasis and therapy resistance. The stiff tumor ECM enhances tumor heterogeneity by promoting the growth and survival of select genetically modified tumor cells, by fostering genomic instability and/or increasing mutagenic load, by creating supportive niches for CSCs, by promoting tumor cell EMT into CSCs, and contributing to CSC maintenance and expansion. The ECM also serves as a reservoir for growth factors and cytokines, that support CSC survival and expansion and foster tumor cell growth and invasion and dissemination to drive tumor progression. Stromal fibroblasts are important modulators of the mechanical properties of the TME through their ability to remodel the ECM and secrete pro-fibrotic, pro-tumorigenic cytokines and exosomes. Indeed, CAF-induced ECM stiffening has significant implications for tumor progression. For example, a stiff ECM activates mechanotransduction in cancer cells to potentiate oncogenic signaling and inhibit tumor suppressor expression that promotes tumor cell growth, survival, and invasion and malignant progression. The stiffened tumor stroma also stimulates angiogenesis and eventually contributes to vascular collapse to induce hypoxia and tumor aggression and impede drug delivery that drives treatment resistance. The stiff stroma additionally creates a physical barrier that hinders T cell invasion and migration and promotes a pro-tumor myeloid phenotype that compromises CD8 T cell activity to interfere with antitumor immunity and immune checkpoint therapy. Clearly, understanding the molecular mechanisms underlying the ECM’s influence on cancer stemness has profound implications for the development of new therapeutic strategies (see Outstanding questions). Current treatments often fail to eliminate CSCs, leading to tumor relapse and metastasis. By targeting the ECM, stromal fibroblasts, and/or mechanotransduction, it may be possible to disrupt the supportive CSC niche, to reduce their frequency, and ultimately improve long-term patient outcome. Approaches such as inhibiting ECM stiffening, blocking specific integrin interactions, or targeting mechanotransduction represent promising avenues for future research and clinical application.

Outstanding questions.

What is the impact of specific CAF subtypes on tumor cell heterogeneity and CSCs?

What is the impact of ECM tension on tumor cell heterogeneity, and is this mediated through CSCs?

How does the TME drive tumor cell genomic heterogeneity, and is this through genomic instability or targeted mutations?

What is the impact of ECM tension on antitumor immunity and tumor heterogeneity?

What is the relationship between the oncogenic/tumor suppressor tumor genotype and TME heterogeneity?

What ECM phenotype permits CSC growth and survival to promote metastatic progression in the pre-metastatic niche?

What is the optimal strategy and timing for therapeutic targeting of tissue fibrosis and/or mechanotransduction to reduce tumor heterogeneity, CSC expansion, and tumor aggression?

Can combination therapies targeting the fibrotic ECM enhance tumor sensitivity to treatment or synergize with TME-targeting therapies (e.g., immunotherapies)?

Highlights.

The desmoplastic tumor microenvironment (TME) is defined by a heterogeneous extracellular matrix (ECM) and cellular landscape that drive tumor progression, aggression, and therapy resistance.

The stiffened ECM stroma fuels tumor heterogeneity, malignant transformation, and therapy resistance.

Heterogeneous cancer-associated fibroblasts (CAFs) and tumor-associated macrophages (TAMs) in the tumor stroma are key drivers of fibrosis, which, through activated mechanosignaling, promotes tumor growth, survival, invasion, and immunosuppression.

Cancer stem cell (CSC) viability, phenotype, and function are tightly controlled by genetic factors, the cellular microenvironment, and the TME’s biomechanical properties, and drive tumor heterogeneity.

Glossary

Amoeboid migration

cell migration strategies with one or more of the following characteristics: rapid single cell crawling, high actomyosin contractility, hydrostatically driven blebs, low adhesion to the substrate or frequent shape changes.

Anoikis

a form of programmed cell death that occurs in anchorage-dependent cells when they detach from the surrounding extracellular matrix (ECM).

Fibrillar collagens

fibrillar collagens (types I, II, III, V, XI, XXIV and XXVII) constitute a subgroup within the collagen family that contains rigid, rod-like molecules with three subunits, α-chains, folded into a right-handed collagen triple helix. Their biosynthesis is a critical pathway controlling development, growth, homeostasis and aging.

Fibrin

a non-globular protein formed from fibrinogen, a soluble protein found in blood plasma. When tissue damage results in bleeding, fibrinogen is converted at the wound into fibrin by the action of thrombin, a clotting enzyme.

Fibronectin

a high-molecular-weight glycoprotein of the ECM that binds to integrins and other ECM proteins such as collagen, fibrin, and heparan sulfate proteoglycans and mediates cell-ECM interactions.

Focal adhesions

integrin-containing, multiprotein structures that form mechanical links between intracellular actin bundles and the extracellular substrate in many cell types.

Glycocalyx

a thin, negatively charged layer of membrane-bound glycoproteins, proteoglycans, glycosaminoglycans, and plasma proteins that coats the apical surface of cells.

Hyaluronic acid

a linear glycosaminoglycan, an anionic, gel-like, polymer, found in the extracellular matrix.

Hydrogel network

a porous, water insoluble, three-dimensional network of polymers that swell in aqueous solvents.

Integrins

transmembrane receptors that help cell–cell and cell–extracellular matrix adhesion. Upon ligand binding, integrins activate signal transduction pathways that mediate cellular signals.

Invadopodia

specialized actin-rich protrusions that extend from the cell membrane to the surrounding ECM.

Juxtacrine signaling

ligands from the inducing cell interact with receptors of adjacent responding cells.

Lysyl hydroxylase

a glycosyltransferase known to post-translationally modify collagens through the addition of galactosyl or glucosyl sugars.

Lysyl oxidases (LOX)

extracellular enzymes responsible for initiating covalent crosslink formation in collagen fibrils by oxidatively deaminating specific lysine and hydroxylysine residues.

Mechanotransduction

the process by which a cell senses and responds to mechanical stimuli.

Stiffness

elastic modulus (E) expressed in units of force per area (Pa) or pounds per square inch (psi).

Syndecans

transmembrane proteoglycans thought to act as coreceptors, especially for G-proteincoupled receptors.

Tenascin

extracellular matrix glycoproteins that modulate cell adhesion, migration, and growth.

Transforming growth factor beta (TGF-β)

a multifunctional cytokine belonging to the transforming growth factor superfamily that plays a central role in wound healing and in tissue repair.

Viscoelasticity

a material property where the material exhibits both elastic and viscous characteristics when deformed. This means the material can store and dissipate energy during deformation, resulting in a time-dependent response to stress.