Biomechanics: a driving force behind metastatic progression

Metastatic progression, which begins with the invasion and migration of tumor cells from a primary tumor, marks a major turning point in the evolution of cancer. Indeed, it eventually leads to the formation of secondary tumors, the metastases, which are very often responsible for the patient’s death. Understanding the mechanisms controlling the different steps of this process, as well as those explaining the fundamental phenomenon of organotropism (i.e. the distribution of metastases in distant organs by a non-random and tumor-specific process), is essential to define new innovative therapeutic solutions. In this review paper, we will present how biomechanics is an essential element to this understanding, and will emphasize the importance of this orthogonal and promising angle of study as well as our laboratory’s focus on the late stages of dissemination, arrest and extravasation of circulating cancer cells and factors secreted by the primary tumor such as extracellular vesicles. Résumé. La progression métastatique, qui débute par l’invasion et la migration de cellules tumorales depuis une tumeur primaire, marque un tournant majeur dans l’évolution du cancer. En effet, elle conduit à terme à la formation des tumeurs secondaires, les métastases, responsables très souvent de la mort du patient. Comprendre les mécanismes contrôlant les différentes étapes de ce processus, ainsi que ceux expliquant le phénomène fondamental d’organotropisme (c’est-à-dire la distribution ∗Corresponding authors. # Contributed equally. ISSN (electronic) : 1768-3238 https://comptes-rendus.academie-sciences.fr/biologies/ 250 Florent Colin et al. des métastases dans les organes distants par un processus non aléatoire et tumeur-spécifique), est essentiel pour définir de nouvelles solutions thérapeutiques innovantes. Dans ce document de synthèse, nous présenterons comment la biomécanique est un élément essentiel à cette compréhension, et insisterons sur l’importance de cet angle d’étude orthogonal et prometteur ainsi que la focalisation de notre laboratoire sur les étapes tardives de dissémination, d’arrêt et d’extravasation des cellules cancéreuses circulantes et des facteurs sécrétés par la tumeur primaire comme les vésicules extracellu-


Introduction
Metastatic progression-the last step of the multistage carcinogenesis process-is ultimately leading to the development of metastases that are usually responsible for the death of cancer patients [1,2].It begins when malignant cancer cells escape from the primary tumor and its microenvironment, invade the patient's body through the lymphatic and/or the vascular systems, and finally reach distant organs.The colonization of these organs, an inefficient process requiring the conjunction of multiple helping factors, leads to the formation of secondary tumors called metastases.In support of Steven Paget's "seed and soil" concept [3] there are considerable evidences to suggest that the development of metastasis is induced by intrinsic characteristics of tumor cellsnotably their genetic and epigenetic alterations-as well as organ-specific environmental factors [4].This so-called "organotropism" has been extensively studied from a medical, biological and chemical perspective in order to understand and anticipate tumor and metastatic progression in patients.Insofar as the metastases are largely incurable and currently very difficult to detect early enough during the natural history of the pathology, the exploration of the molecular and cellular pathways leading to the establishment of site-specific metastasis are thus in the forefront of cancer research.Other factors promote metastasis, among which the formation of permissive microenvironments, the so-called premetastatic niche priming.This mechanism occurs when factors secreted and released by the primary tumor create, in organs distant to the primary tumor, an environment where extravasation and growth of disseminating cancer cells are facilitated [5].The discovery of the role of extracellular vesicles (EVs) in intercellular communication, and in our case of their role (in particular that of exosomes) in the promotion of the formation of the pre-metastatic niche (PMN) is the subject of intense research, including in our laboratory.
In recent years, we have seen the emergence of biomechanical studies-in addition to our ownaimed at understanding the contribution of fundamental driving forces and their involvement in each stage of the metastatic progression (Figure 1highlighted in yellow).During growth of solid tumors for instance, the increased interstitial pressure has been shown to promote invasion and concomitantly limit access to chemotherapeutic molecules [6].Later during the metastatic cascade-after invasion through complex extracellular matrix environments and intravasation through blood vessels of cancer cells evading from the primary tumor-the vascular tree will mechanically disseminate and guide circulating tumor cells to a first capillary bed, often a highly metastatic organ (i.e. the lungs in breast cancer, the liver in colorectal cancer).At the level of blood capillaries, hemodynamic parameters and the topography of the vessels themselves influence the sites of tumor cell arrest [7].This is also probably true for soluble pro-metastatic factors-including EVs-secreted by the primary tumor, suggesting that the locations of the pre-metastatic niches must themselves respond to this set of physical constraints.Although these aspects are well described, their real biological impacts remain only poorly described in vivo, particularly in the later steps that occur in and off the bloodstream.Our laboratory primary objective is to continue to further elucidate the central role of biomechanics on the entire metastatic cascade, from the tumor progression to the metastasis itself (Figure 1-red star references).To this end, we combine state-of-the-art sophisticated tools (notably microfluidics-based), in vitro and in vivo models (mouse and zebrafish) with high-resolution imaging-such as correlated light and electron microscopy [8]-and biophysical approaches to mimic pathological situations highly relevant to cancer research.In doing so, we aim to ultimately identify innovative therapeutic alternatives.

Forces that shape tumors and its microenvironment (TME)
A primary tumor and its highly-plastic microenvironment (TME) form an intricate non-homogeneous structure composed of different elements: the cancer cells themselves, a broad spectrum of stromal cells that can be activated including notably cancerassociated fibroblasts (CAFs), immune cells, and the lymphatic and blood vessels.It also includes the extracellular matrix (ECM) which contains, in addition to fiber-like molecules such as collagen and fibronectin, cell-secreted products (i.e.cytokine, chemokines, extracellular vesicles), various metabolites, and present hypoxic and acidic properties [4].The growth of the primary tumor (Figure 1-1.Growth) results from uncontrolled proliferation in a healthy tissue and creates solid stress, which in turn induces various forces and tensile stress around and within the tumor itself [9].This, in conjunction with the activation of neighboring stromal cells with high contractibility potential, results in the extracellular matrix becoming significantly stiffer and anisotropic, a feature correlated with a malignant tumor phenotype.We, and others, have previously demonstrated that the contractility of these stromal cells is a key factor in metastatic progression.In particular, we observed that Caveolin-1-a multifunctional scaffolding protein associated with cell-surface caveolae [10]-is a major regulator of tumor cell invasion and subsequent metastasis (Figure 1-2.Invasion) through the control of focal adhesion dynamics within tumor cells [11] and the biomechanical remodeling of the tumor-associated microenvironment by stromal cells [12].The microenvironment of the tumor, influenced by the increased forces and abnormal fluids flows, will gradually transition from an anti-malignancy to a pro-malignancy behavior [13], promote the migration of cancer cells through the extracellular matrix (Figure 1 On a side note, while angiogenesis and lymphangiogenesis of the primary tumor play a potentiating role [14], tumor dissemination is an event that can also be observed in early lesions [7].In conclusion, identifying and characterizing the flow-based mechanical constraints inherent to the tumor progression and metastasis cascade is of the utmost importance [15] and at the heart of our laboratory's efforts [7]. To understand the mechanical forces at work, it is necessary to consider the primary tumor and TME irrigating fluids, namely the interstitial fluid, the lymph and the blood circulation.Each has its own biochemical characteristics, cellular content and flow patterns.The interstitial fluid, originating from blood capillaries' leakiness and surrounding the cells, has a laminar flow heading towards the lymphatic draining sites.In a primary tumor, such flow is driven by the decreasing interstitial pressure gradient directed outward from its center (schematically its highest-pressure point).Lymphatic vessels exhibit a weakly pulsatile laminar flow, draining material from the primary tumor outward to the lymph nodes.Finally, the blood system, which presents different types of vessels (capillaries, veins and arteries), irrigates the primary tumor and delivers oxygen but also connects the growing tumor to distant parts of the organism.While capillaries and veins present a laminar flow of low to medium intensity, and a rather reduced pulsativity, they are supply and dissemination routes of the primary tumor (as illustrated in Figure 1), a phenomenon reinforced by their neo-formations following the release of pro-angiogenic factors from the tumor and its TME.The arteries present a laminar or turbulent flow of high intensity as well as a strong pulsativity and also participate to this process.Arterial circulation is a particularly hostile environment for the propagation of CTCs, since they exhibit the highest shear stresses [7].Altogether, lymph and blood circulation offer optimal platforms to tumor cells for colonizing distant lands.Remarkably, CTCs are not the only ones to disseminate through these fluids and vasculatures: other tumorsecreted factors such as extracellular vesicles, some of which have pro-metastatic properties, use similar strategies to the benefit of metastatic fitness of CTCs.

The pre-metastatic niche (PMN) priming
Lymphatic and blood vessels allow the dissemination of various tumor-secreted factors such as cytokines, chemokines, growth factors, matrix metalloproteinases, circulating tumor DNA, antigens and extracellular vesicles (Figure 1-EVs, in blue) [16,17].Given their abundance in the body fluids of patients and their short half-life, it is likely that the primary tumor and its TME release them consistently and regularly.Yet, this remains poorly understood and should provide fertile grounds for future research.These tumor-secreted factors are subjected to hemodynamic constraints [18]-similarly to CTCs-and may arrest in various metastasis-free organs where they can locally modify the microenvironment and create so-called "pre-metastatic niches" (PMNs), which are ideal sites for CTC targeting and metastases formation (Figure 1-7.Premetastatic niche priming).Extracellular vesicles (EVs) in particular-small lipid bilayer particles that serve as signaling cargoes and which can be quickly cleared from the blood by intravascular macrophages/monocytes [18] and endothelial cells of the PMNs-have been demonstrated to exhibit specific organotropism depending on their primary tumor of origin, a feature made possible by the exposure of a variable repertoire of surface adhesion molecules [19].Because they are transported by the blood flow, these non-inert circulating objects may "stop" in specific vascular regions as a function of their adhesive potential and the hemodynamic forces at play [18].In relation to this mechanism, we have recently demonstrated that the Ral GTPases control the biogenesis, the secretion and the content of tumor extracellular vesicles [20] that favor tumor metastasis in mouse syngeneic models of breast cancer [21].It appears thus likely that the formation of the PMNs depends not only on physical factors such as the architecture, permeability, leakiness, local flux and hemodynamic forces existing in the vasculatory system, but also on biological factors such as cell-adhesion molecules or intravesicular cargoes (i.e.proteins or nucleic acids) that make these extracellular vesicles important actors of the metastatic cascade.It is interesting to note at this point that EVs and CTCs might share similar targeting strategies for colonizing distant organs.Subsequently to the PMN priming process, cancer cells that have completed their migration and intravasation steps will be able to begin their intravascular journey as "circulating tumor cells" (CTC) until reaching their final destination.

Intravascular journey of a CTCs: flow forces, adhesion and extravasation
Steven Paget's concept of "seed and soil" [3]-while remaining a central principle-was somewhat challenged by James Ewing (as early as 1928) and Dale Coman [22] by the introduction of the concept that the mechanics of the bloodstream were responsible for the dissemination of tumors to secondary organs followed by the demonstration that the frequency of occurrence of metastases and the vascular architecture as well as the capillary nature of the vessels in the heavily affected target organs were correlated [23].Hematogenous circulation mechanics appeared thus responsible for the distribution of tumor cells in target organs.These two models, which may appear to be independent or even contradictory, actually represent two aspects of a single combinatorial "bio-mechanical" phenomenon.Indeed, although the non-specific and mechanical arrest of CTCs is an existing mechanism, the fate of these depends also on the organ in which they are located [24].Numerous works, including our own, have since demonstrated that the metastatic process results from multifactorial physical and biochemical selective events where fluid biomechanics is an essential element [4,7].As soon as a cancer cell escapes from the primary tumor, intravasates in the blood circulation and becomes a CTC, an essential parameter for its survival and intravascular arrest is the shear stress, which depends on the fluid viscosity and the velocity gradient.Concretely, the more a CTC moves away from the center of the vessel-which it does naturally thanks to its margination properties-the more it will be deformed by the shearing phenomenon (this is especially true in blood due to its high viscosity) (Figure 1-5.Dissemination, shear stressinduced cell death).When entering the lymphatic or venous system, the shear stress remains low and the friction with the membrane of the vascular endothelium presumably harmless.Some CTCs will succeed in reaching the arteries, where the shear stress is high and where the risks of collision are statistically important: there, cells might undergo deformation, fragmentation and ultimately death [25].This bottleneck results in the selection of CTCs with the most suitable properties for their survival-both physicochemically and physiologically with reported modification of genes expression [26]-, and promotes their patterning into more resistant cell clusters, either among themselves as CTC clusters [27] or in association with other cell types (fibroblasts, neutrophils or blood platelets) [7] (Figure 1-5.Dissemination, illustrated).CTCs that have survived collision events, shear stress, and escape the immune system during their migration will exploit their adhesive potential and the hemodynamic properties of the vessels that carry them to colonize specific and permissive regions [28,29], presumably PMNs in vivo.However, it is not so much when the CTCs are free within the circulation that they undergo the greatest stress, but when they attach and try to extravasate: one indication is that upon injection of tumor cells into the afferent vessel of metastatic organs, although 80% of the cells are thought to stably arrest one day after injection, less than 4% of injected cells efficiently form metastatic foci [30].Although the intravascular arrest process of CTCs can occur passively by physical occlusion of small capillaries [31] (Figure 1-6.Arrest, physical trapping), which is currently a subject of re-search of our laboratory, the active adhesion to the vessel wall has been also reported to participate-at least in part-to further extravasation [28].Adhesion between CTCs and the endothelium of the blood vasculature is only possible if the binding forces (i.e.ligand-receptor interaction) between them is greater than the forces exerted by the flow and shear stress [29] (Figure 1-8.Extravasation, active adhesion).This essential step can be influenced by the physicochemical characteristics of the vascular wall, the viscoelastic properties of the CTCs, but also by the homogenous or polarized distribution of the ligands and receptors.Different mechanisms/modelsthough not necessarily exclusive-have been proposed to describe the arrest and extravasation steps of CTCs.They could follow strategies similar to those of leukocytes, namely a rolling along the endothelial wall and then a (paracellular or transcellular) diapedesis step (Figure 1-8.Extravasation, diapedesis).The formation of "low adhesive strength-rapid activation" followed by "higher adhesive strength-slow activation" ligand-receptor bonds through a tug-ofwar with blood flow has been documented [28,29], and could explain both the rolling and arrest.However, our own observations tend to suggest that CTCs arrest intravascularly rather abruptly, with a risk of detachment: more work is thus needed to clarify this behavior.Furthermore, we demonstrated that CTCs do not only use diapedesis to leave the bloodstream and form secondary tumor foci: they can also use a flow-dependent endothelial remodeling mechanism [28] (Figure 1-8.Extravasation, endothelial remodeling), which is characterized by the formation of protrusions leading to the creation of new vascular lumens (pocketing), and the passive exclusion of CTCs from the vascular environment.Indeed, the vascular endothelium displays considerable mechano-sensing abilities that make it likely to sense, react and adapt to any conditions [32].
In the context of intravascular CTC dissemination, we recently observed that the VEGFR2-dependent pathway as a signaling node sensitive to flow forces and driver of tumor metastasis through endothelial remodeling [33].Doing so, we identified a potential therapeutical target in the fight against tumor metastases.Finally,-although dormancy phenomena (ranging from a few hours to years) have been reported [34]-the CTC will eventually initiate extravascular proliferation to form a secondary tu-mor called metastases (Figure 1-9.Metastatic outgrowth).The proliferation efficiency will arguably depend on the metastatic microenvironment characteristics, and presumably on its further remodeling by pro-tumor factors such as EVs.Then, the newformed metastases will, in turn, potentially initiate their own process of tumor progression and metastasis in the manner of the primary tumor, and dramatically accelerate the worsening of the cancer patient's health condition.

Conclusion and perspectives
Our ability to effectively treat cancer depends in large part on our ability to decipher, correctly diagnose and prevent cancer progression and metastasis.In this regard, research on the role played by mechanical forces in these processes is fundamental and complementary to other approaches.The different body fluids, due to the types of flows that animate them and their resulting constraints (notably shear stress in blood circulation), are responsible for the dissemination of secreted tumor factors and CTCs.They play a central role in the selection and preparation of pre-metastatic niche for the formation of metastases, thus preparing an ideal soil.They are also key elements in selecting CTCs capable of surviving during the metastatic process, whether during their migration, adhesion, arrest or extravasation-each of them requiring biomechanical and physiological adaptations-thus carrying the metastatic seed.We recently demonstrated that fluid forces are also key in activating flow sensing receptors (in our models, but presumably in cancer patients PMN as well) and promote extravasation processes and thus metastasis.The pervasiveness of biomechanics in solid cancers (some elements of which also apply to blood cancers) makes research efforts in this area relevant.Although many studies have focused in the past decade on understanding forces within the primary tumor, more work is needed to understand, for example, how these can tune the secretion and nature of tumorpromoting factors such as extracellular vesicles.To understand how shear stress and flow forces impact the secretion, dissemination and uptake of factors secreted by the primary tumor and its TME, and to what extent they are involved in their cancer-dependent organotropism, are major challenges.Furthermore, how mechanical properties of CTCs themselves impact their metastatic fitness is conceptually appealing and thus studied in our laboratory.A detailed understanding of the forces at play could allow the development of innovative therapeutics taking advantage of biomechanical constraints to specifically target pre-metastatic niches-in the manner of factors secreted by tumors, i.e. engineered drug delivery EVs-and block or delay their establishment and subsequent endothelial remodeling and metastasis.
The fragility of CTCs to shear stress must also be considered: if the use of drugs targeting blood pressure does not appear to be in the best interest of a cancer patient, tools allowing their continuous removal from the blood are already being developed [35] and more work is needed to understand whether their mechanical profiles are involved in such behavior.
Understanding the biomechanics at work in a patient will also require considering the parameters acting on them.The physical/physiological characteristics of a patient-in particular age [36]-the potential effects of the treatments followed, and finally their evolution during the natural history of the pathology are all elements that could enable more efficient and personalized therapies.Another axis is the study of the effect of factors secreted by tumor and non-tumor cells, of their possible pro-vs.anti-tumor competition during the metastatic cascade, as well as their potential involvement in a biomechanical modification of the PMN.

Figure 1 .
Figure 1.Biomechanics of the metastatic cascade: a schematic representation of the path of cancer cells and extracellular vesicles during the cascade of events leading to the formation of metastases.The nine major steps are indicated in black on a white background, large font, and are referenced in the text.The biomechanical forces involved are highlighted in yellow.There are two legends on the right: the first one on top gives the references of the laboratory studies (red stars, numbered).A second one on the bottom references the types of cells and factors.