Cohesive cancer invasion of the biophysical barrier of smooth muscle

Abstract

Smooth muscle is found around organs in the digestive, respiratory, and reproductive tracts. Cancers arising in the bladder, prostate, stomach, colon, and other sites progress from low-risk disease to high-risk, lethal metastatic disease characterized by tumor invasion into, within, and through the biophysical barrier of smooth muscle. We consider here the unique biophysical properties of smooth muscle and how cohesive clusters of tumor use mechanosensing cell–cell and cell–ECM (extracellular matrix) adhesion receptors to move through a structured muscle and withstand the biophysical forces to reach distant sites. Understanding integrated mechanosensing features within tumor cluster and smooth muscle and potential triggers within adjacent adipose tissue, such as the unique damage-associated molecular pattern protein (DAMP), eNAMPT (extracellular nicotinamide phosphoribosyltransferase), or visfatin, offers an opportunity to prevent the first steps of invasion and metastasis through the structured muscle.

1. Introduction

Several cancers of organs surrounded by smooth muscle rank in the top 10 cancers for incidence and deaths for both men and women. These include colon, bladder, prostate, uterine, and gastric cancers [1]. A feature shared by these cancers is that, when organ confined, they are not lethal with treatment; they become lethal when they escape the primary organ and metastasize. A principal means of escape for tumors from the original site is invasion through the smooth muscle surrounding the exterior of the organ [2]. Another common means of escape from the primary tumor is through intravasation or extravasation through veins, which requires tumor clusters to migrate through a smooth muscle layer sheathing the vein wall, and move through an even thicker layer in escaping through arteries [3].

For this review, we will examine the features of smooth muscle invasive cancer which are physiologically relevant in six primary cancer types: prostate, bladder, colorectal, uterine, esophageal, and gastric tumors. We focus primarily on adeno-carcinomas in the prostate and bladder, and we will not discuss sarcomas as these tumors arise in muscle and do not invade through the muscle. The most researched and documented of these cancers is muscle invasive bladder cancer (MIBC); however, each cancer shares the basic feature that once a tumor invades the smooth muscle surrounding the gland of origin, the survival rate falls drastically.

The smooth muscle is a unique biophysical tumor microenvironment and tumor invasion into and through this barrier is a first step toward metastatic progression. We examine the biophysical characteristics and forces of the smooth muscle, the biosensing integrin adhesion receptors that bind the laminin ECM, and potential triggers for smooth muscle invasion. An ability to predict, reduce, or prevent the invasion of smooth muscle would decrease the mortality rates for these cancers through prevention of metastasis and provide additional considerations for improved treatment decisions.

1.1. Prostate cancer

Prostate cancer (PCa) is the most commonly diagnosed cancer among men in the USA (excluding non-melanoma skin cancer) with an annual age-standardized incidence rate of 112.6 per 100,000, as estimated by the Surveillance, Epidemiology, and End Results Program (SEER) between 2011 and 2015 [4]. PCa has a 98% overall 5-year survival rate, and a 100% 5-year survival rate if it remains organ confined, which drops to 30% 5-year survival if it escapes the gland [5]. The peripheral zone of the prostate gland (where most cancer arises) is composed of a smooth muscle stroma which forms a prostate capsule (Fig. 1, panel a), and a majority of epithelial tumors display traits of collective invasion into the surrounding smooth muscle layer, including cell–cell adhesions, the presence of E-cadherin, and occurrence of other cell–cell adhesion receptors [6, 7]. While the loss of E-cadherin signaling is a necessary but not sufficient marker of advanced disease [8, 9], metastatic disease expresses E-cadherin [10–12], providing an example of tumor phenotype plasticity.

Fig. 1.

Loss of organ-confined disease in prostate and bladder cancer is associated with an increased risk of biochemical recurrence, distant metastases, and lower cancer-specific survival. The prostate gland (panel a schematic) is under the bladder and surrounded by a smooth muscle layer (gray), referred to as a capsule. Organ-confined cancer is classified as T1 and T2, whereas tumor extending past the capsule into periprostatic adipose tissue (PPAT) (yellow) is classified as T3 and associates with high-risk disease. The bladder (panel b schematic), similar to the prostate gland, contains stages of organ confined disease (carcinoma in situ (CIS), Ta, T1, T2) and when the smooth muscle layer is breached, this is considered high-risk bladder cancer (T3, T4) resulting in the need for more aggressive therapy. Current research suggests that in both organ types, the PPAT or the perivesical fat provides potential chemoattractants for tumor progression. [Figure adapted from a diagram by © Cancer Research UK (2002), used with permission]

Glandular escape occurs by the invasion of the smooth muscle capsule [13, 14], which is especially rich in laminin 511 [13], an abundant extracellular matrix (ECM) protein trimer found predominantly in basement membranes [15] and which self-assembles into a mesh-like network or partition [16]. Beyond the smooth muscle surrounding the prostate is a layer of adipose tissue, the periprostatic adipose tissue (PPAT), that may provide chemoattractants to promote the escape of tumor clusters from the gland [17, 18]. Aggressive PCa is defined by extracapsular extension (ECE) and the clinical staging increases from T1, T2, to a T3 score. High-risk (aggressive) disease invades into and through the smooth muscle capsule, gaining access to nerves and vessels resulting in hematogenous spread to distant sites, primarily bone [19]. Although invasion through the smooth muscle is required for metastasis, it is rarely a metastatic destination [20].

Prostate cancer also intravasates and extravasates through a smooth muscle layer in vessels, and Batson’s venous plexus is involved—a system of valve-less veins that directly communicate with the vertebral column [21, 22], which is the major site of prostate cancer bone metastasis—so tumor clusters can move in any direction [23, 24]. In perineural invasion (PNI), to reach the nerves at the peripheral zone, the tumor also moves through a bed of muscle [14]. Most metastatic models of PCa have not addressed how the tumor escapes the gland through the smooth muscle layer (see, for example [25, 26]).

The presence of ECE defines the pT3a stage of prostatic adenocarcinoma. ECE is associated with an increased risk of biochemical recurrence, metastasis, and overall mortality [27, 28]—and is a required component in both the College of American Pathologists and the International Collaboration on Cancer Reporting guidelines [27]. ECE is difficult to detect in a prostate needle biopsy due to limitations of the sampling technique but is detectable by MRI [28] and has been observed by the histopathology examination of radical prostatectomy (RP) specimens for decades [29].

The normal prostate gland resides in a bed of muscle with a clear boundary whereas a tumor will invade and disrupt the muscle boundary (Fig. 2, panels a–c). Both mesenchymal and epithelial tumor cells can invade surrounding tissue either as single cells, chain-like files, clusters, or collective strands [2] and the angulated clusters of tumor cells moving through the muscle (Fig. 2, panel c, inset) are pathognomonic for reactive stroma and invasive disease [30]. The modes of escape are often adaptable and interchangeable, as cells can vary the strength of cell–cell adhesion and actomyosin contractility in response to determinants of the environment [2], including the ability of cell clusters to move single-file through capillaries and reassemble into clusters once that passage is traversed [31]. There is reason to suggest that tumor cell clusters make a similar adaptation to move through smooth muscle [32] in order to withstand the rigidity of the muscle barrier.

Fig. 2.

Normal prostrate glands respect the muscle barrier whereas cancer invades it. Serial sections of a needle biopsy were either stained with H and E (panel a), stained for HMWCK specific for normal glands (N) (panel b), or stained for muscle-specific desmin (panel c). Areas of muscle (Muscle), normal glands (N) and angulated tumor in disrupted muscle (Tumor) were observed. Black bar, 100 μm; inset × 5 zoom

The investigation of molecular expression changes in muscle would assist the need for ECE biomarkers as the assessment of extraprostatic extension is made challenging by the prostate gland not having a molecularly well-defined capsule. The simplest determinant for ECE is invasion into PPAT surrounding the prostate capsule (Fig. 1, panel a) since the occurrence of fat within the prostate gland is rare [27, 28]. ECE in the RP sample is well-studied and is a significant part of the tumor staging process—all RP specimens containing ECE are classified as pathological stage pT3 [28, 33]. ECE found at RP is correlated with a poor prognosis, which is why it is often included in postoperative nomograms and predicts negative consequences such as biochemical recurrence after RP [28, 34, 35]. Occasionally, ECE is also noted on the pathology report from the prostate biopsy, although its detection on the prostate biopsy is fairly uncommon [28]. As of 2016, there was only one study (conducted between 1997 and 2009 at Johns Hopkins University) that reported outcomes of patients with ECE on prostate biopsy, finding an ECE prevalence by histopathology of only 0.19% [28, 36]. This may be an underestimate since MRI/ultrasound fusion-guided targeted biopsy now can detect ECE not found on standard biopsy [37]. In contrast, while the perineural invasion is present in most RP specimens, it is generally not an independent predictor of post-surgical outcome and, therefore, has not been a required component for reporting guidelines—some studies suggest otherwise, but so far PNI has not been widely accepted on an outcome indicator [27].

1.2. Bladder cancer

Urinary bladder cancer is the eighth most common cause of cancer death in men [1], who experience this form of cancer roughly three times more often than women [1]. Even when the tumor remains confined within the gland, the 5-year survival rate is only 70%, dropping to 36% if the tumor escapes the bladder [38] in MIBC. Like the prostate, the bladder has an epithelial layer sitting above the basement membrane (Fig. 1, panel b), a single layer of cells separating the epithelial layer from the lamina propria, a sheet of extracellular material that acts as a filtration barrier and support structure for the mucosal layer [39]. The lamina propria is composed of areolar connective tissue (a rich semi-liquid matrix containing fibrocytes, plasma cells, macrophages, mast cells, and various white blood cells); this is intertwined with the muscularis propria [39], comprised of three muscle layers. A tumor that invades this layer can metastasize to other areas in the body via the lymphatic system and/or blood vessels [39]. Beyond the thick layer of smooth muscle, the bladder, like the prostate, is surrounded by a layer of adipose tissue, known as perivesical fat [39]. This outer layer of adipose tissue, again much like in the prostate, appears to coordinate with the tumor in escaping the gland [40, 41]. As is true in prostate cancer, loss of E-cadherin signaling in the primary tumor indicates aggressive disease and phenotype plasticity and suggests a poor prognosis [42, 43].

1.3. Colorectal cancer

An estimated 53,200 people will die from colorectal cancer in the USA in 2020 [1]. As is true in the other epithelial cancers, muscle invasive disease has a poor outcome. Once the tumor invades and escapes the smooth muscle layer (AJCC stage 2), there is an 80% rate of recurrence and eventual death [44]. Colorectal cancer features a consistent proliferation of epithelial cells, some of which lose their normal structure and formation (compact and regular arrangement and shape, cellular polarity, intercellular junctions, and adherence to the basal lamina) and take on characteristics of mesenchymal cells (oval or fusiform shape, lack of polarization, and cytoskeleton reorganization) [45]. While this is often referred to as the epithelial-to-mesenchymal transition (EMT), we refer to the resulting cohesive clusters of tumor cells as a mixed phenotype, taking characteristics of both epithelial and mesenchymal cells [6]. Others have now suggested a mixed phenotype plasticity in cancer is a common part of the EMT spectrum [46].

There are three histological layers of the colorectum: the submucosa (SM), muscularis propria (MP), and subserosa (SS) [47]. The MP is exclusively composed of smooth muscle bundles and contains tight connective tissue, while the SM and SS are mainly composed of loose connective tissue [47, 48]. Takatsuna et al. found considerably greater numbers of myofibroblasts—cells with features of both a fibroblast and a smooth muscle cell in phenotype—located around these three invasive layers in the infiltrating type compared to the expanding type [47]. Furthermore, they found a higher density of myofibroblasts in the MP level, but not the SM and SS layers, correlated with a significantly reduced overall survival (slightly over 60% at 80 months) [47].

The increase in myofibroblasts would seem to indicate a disruption in the integrity of the smooth muscle as a result of the invasive tumor. In prostate cancer tissue, the angulated tumor clusters disrupting the smooth muscle are well-known [30, 49]. Smooth muscle disruption is a feature in other smooth muscle invasive tumors, as well. It remains unknown if other host biological factors, such as age, endocrine dysfunction, or inflammation that weaken the smooth muscle would be a confounding factor for tumor invasiveness.

1.4. Summary

A major gap in our knowledge is how and why epithelial tumors migrate into and through the laminin-rich muscle to escape the gland. There is considerable similarity between PCa, bladder cancer, colorectal cancer, and other epithelial tumors since these can escape their origin site through smooth muscle invasion [50–52] and exhibit phenotype plasticity. Identifying how these tumors move into and through the smooth muscle, withstand the biophysical forces of muscle, and what triggers this transition could be a major step toward reducing the metastatic spread and increasing survival. It is likely that both the tumor and the muscle responses are reciprocal and iterative events that drive toward metastasis to distant sites.

2. Smooth muscle as an organ defining element in bladder and prostate

2.1. Defining the muscle layer surrounding the bladder and prostate

Smooth muscle layers, together with connective tissue, form the anatomical layers of submucosa, mucosa, and serosa surrounding hollow organs or organ systems [2]. Both the prostate and the bladder are encased in a layer of smooth muscle [53–56], and both contain testosterone-sensitive cells [57, 58]. Smooth muscle is involuntary, non-striated muscle, functionally different from skeletal and cardiac muscles, and is found in the walls of the urinary bladder, uterus, stomach, intestines, and prostate, as well as in arteries and veins in the circulatory system, and in the tracts of the respiratory, urinary, and reproductive systems [59, 60]. Smooth muscle is differentiated from skeletal muscle in several ways, the most obvious being its ability to be contracted and controlled involuntarily. The nervous system uses smooth muscle to control subsystems throughout the body that operate outside of the control of the organism [60]. Fully differentiated smooth muscle contains the cytoplasmic molecular markers of desmin (an intermediate filament), calponin (an actin-binding protein), and laminin-binding adhesion receptors, including the biosensing α7β1 integrin [30, 49, 61, 62].

2.2. Biophysical characteristics and forces of the involuntary muscle

Individual smooth muscle cells, known as myocytes, are spindle-shaped (wide in the middle and tapered at both ends) and, like striated muscle, can tense and relax but not through the intention of the organism. However, smooth muscle tissue demonstrates greater elasticity and function within a larger length-tension curve than does striated muscle. This ability to stretch and still maintain contractility is important in organs like the intestines and urinary bladder. In the relaxed state, each cell is 20–500 μm in length (thousands of times shorter than skeletal muscle fibers), and these cells produce their own connective tissue, the endomysium [59].

Smooth muscle is organized both as single-unit smooth muscle (most common) and as multi-unit smooth muscle, each occurring in different areas of the body and exhibiting different characteristics [59]. Multi-unit smooth muscle cells seldom contain gap junctions, and so are not electrically coupled. Accordingly, contraction does not move from one cell to the next; rather, it is limited to the cell that was excited originally. Stimuli for multi-unit smooth muscles come from autonomic nerves or hormones but not from stretching. This muscle type occurs around large blood vessels, in the respiratory airways, and in the eyes [59].

Invasive tumors are likely to encounter the single-unit smooth muscle rather than the multi-unit type based on muscle prevalence. Single-unit muscle fibers are joined by gap junctions, allowing the whole muscle to contract as a single unit (known as visceral muscle) and are found in the walls of all visceral organs other than the heart (which has unique cardiac muscle in its walls) [59]. Single-unit fibers are not constrained by the formation and limited elasticity of sarcomeres, therefore providing visceral smooth muscle with a stress-relaxation response. This is apparent in the muscles of a hollow organ, such as the bladder or stomach, that stretch and expand as they fill, causing mechanical stress that then triggers contraction and is immediately followed by relaxation to prevent the organ from an untimely release of its contents. Generally, visceral smooth muscle produces slow, steady contractions that allow substances, such as urine in the bladder, to accumulate and be released—for clarity, the external urethral sphincter of the bladder is comprised of striated muscle, which allows voluntary control for release urine [63].

Myocytes are covered by a basement membrane and bounded by loose collagen-rich fibrillar mesh-works that afford both tissue cohesion and plasticity of tissue movement during contraction, and provide a relatively low resistance for migrating cells, while the basement membranes provide abundant ligands, including laminin, for cell adhesion [2]. The biophysical measurements of muscle elasticity and the dynamic physical contraction/relaxation cycles demonstrate a unique microenvironment for invading tumor cells. The complex biophysical properties include stiffness, elasticity, and contraction forces with physical cues that likely guide tumor invasion [2]. Furthermore, these biophysical determinants will likely be changing as the tumor invades through the muscle since the elastic modulus of smooth muscle is approximately 100 times stiffer than that of fat, which surrounds the prostate and bladder [64].

Epithelial derived tumors interact directly with the smooth muscle ECM (such as laminin and collagen) environment using biophysical and biochemical sensing receptors called integrins on their cell surfaces. Integrins are ECM receptors that organize adhesions for cell mechanics and transmission forces in changing ECM environments during the normal processes of embryonic development and morphogenesis and in pathological processes such as epithelial tumor progression during invasion and metastasis [65–67]. Furthermore, in smooth muscle, α7β1 integrin, specific for laminin ECM binding, specifically the α7B variant, is highly expressed and inactivating mutations will result in loss of muscle integrity [61, 62, 68, 69]. The role of α7β1 integrin is to mediate muscle cell interactions with the laminin ECM, and the role of α6β1 integrin (also specific for laminin) is to mediate tumor cell–ECM interactions and smooth muscle invasion [32], which raises the interesting possibility of integrin coordination between the tumor and the normal muscle during laminin-based invasion.

As the tumor invades the smooth muscle surrounding the primary organ of origin, it will be important to identify the various factors that promote this invasion. It may be that the combination of biomechanical forces, chemoattractants, biochemical signals, and hormonal factors all can make the muscle less resistant to invasion and more malleable when confronted with invasive tumor. Understanding the ECM triggers unique to the smooth muscle remodeling by the tumor, collectively referred to as a licensing factor for metastasis [70], integrated with the tumor-intrinsic receptor activation mechanisms, will likely lead to new detection and therapeutic opportunities.

3. Muscle invasive epithelial cancers

3.1. Muscle destruction by the tumor and a decrease in survival

Smooth muscle invasion (SMI) originates from both epithelial tumors, after tumor cells have crossed a basement membrane, and mesenchymal tumors that originate from the collagen-rich interstitium without passing a basement membrane [2]. Invasion arises as a sequential process as epithelial tumor cells invade the connective tissue dividing the epithelial and muscle layers and then invades the muscle itself. The tumor cells travel along the denser outer smooth muscle that connects collagenous tissue or they invade and spread along and within the orientation of individual muscle cells [2]. SMI in bladder cancer can occur as cells form a linear strand, suggesting a need to overcome elastic forces in the muscle, and further suggesting collective migration along the linear endomysium [2, 71]. Previous research has shown that cohesive clusters of tumor cells can form linear strands in order to fit through a tight space, such as capillaries, and then reform as clusters once they have traversed the constraint [31]. During early invasion, colorectal tumor cells move in a budding formation, oriented linearly between smooth muscle cells, concurrent with considerable immune cell infiltration [2].

Aggressive tumor invasion causes muscle tissue to degrade, leading to weakened and dysfunctional muscle structure that has been replaced with poorly differentiated tumor tissue [72]. Consequently, pronounced invasion of smooth muscle by endometrial, bladder, and gastric cancers correlate with decreased survival—a reduction of 3- or 5-year survival of 15–30% versus minimal muscle involvement [2]. Table 1 lists six different cancer types associated with a decrease in survival due to invasive/metastatic disease. The loss of the normal muscle tissue homeostasis during tumor-induced remodeling is a prime candidate as a determining factor for sustained invasion within and through the muscle, an initial step of metastasis.

Table 1.

Six different types of muscle invasive adenocarcinoma

Cancer type	Survival impact of invasive/metastatic disease	Reference
Colorectal cancer	80% recurrence rate in invasive disease	[44]
Bladder cancer	36% 5-year survival for regional disease	[38]
Gastric cancer	34% 5-year survival at Stage II, 8–20% at Stage III	[73]
Prostate cancer	30% 5-year survival for metastatic disease	[74]
Esophageal cancer	25% 5-year for regional disease	[1]
Uterine cancer	69% 5-year survival for regional disease; 17% in metastatic disease	[1]

3.2. The loss of differentiated muscle induced by the tumor and induction of smooth muscle cell migration

Smooth muscle cells (SMC) are the dominant cell type in prostate stroma, suggested as a key regulatory factor in maintaining homeostasis in the adult prostate via interaction with adjacent epithelial cells [53, 75]. Normal prostatic SMC can prevent or possibly reverse the tumorigenicity of established prostate tumors by provoking differentiation and repressing the growth of tumor cells [53]. The loss of SMC or a loss of smooth muscle differentiation has been postulated to promote the loss of stromal control over epithelial proliferation and differentiation [53]. Influencing the tumor by the prostatic SMC is similar in concept to the known process whereby prostate tumor-associated fibroblasts promote the transformation of nontumorigenic prostate epithelial cells [76]. Reactive stroma, as defined by the histologic features of muscle disruption (observed by loss of desmin, an intermediate filament of the cytoskeleton), is a predictor of biochemical recurrence in prostate cancer [30, 49]. The appearance of muscle-specific desmin in the serum is a marker for decreased recurrence-free survival in prostate cancer [49] and associated with a decreased survival and poor prognosis in colon cancer [77]. One key ECM protein called cysteine-rich angiogenic inducer 61 (CCN1/CYR61) is a key mediator of growth factor– induced migration of smooth muscle cells [78]. CCN1 protein expression is induced and secreted by a variety of normal cell types (fibroblasts, epithelial cells, endothelial cells, smooth muscle cells, and neurons) and tumors localizing on the cell surface and within the ECM. CCN1/CYR61 transcription is induced in response to extracellular stimuli, including growth factors, inflammation, shear stress, mechanical stretch, and oxygen deprivation (reviewed in [79]).

The CCN1 protein contains domains that interact with a variety of adhesion receptors, including laminin-binding integrins (α6β1) and heparin sulfate proteoglycans. Extracellular engagement of α6β1 integrin in metastatic prostate cancer will increase the expression of CCN1 [80]. CCN1 is secreted by tumor cells after α6β1 integrin engagement, induces smooth muscle de-differentiation, and responds to the biomechanical forces inherent in smooth muscle. Taken together, these observations point to this axis as a likely candidate trigger and coordinator of laminin-binding integrins (α6β1 and α7Bβ1) to accomplish adenocarcinoma invasion of the smooth muscle.

4. Epithelial cancers invade muscle as cohesive clusters

We have previously described the cohesive metastasis phenotype in prostate cancer [6]. Many epithelial tumors exhibit traits of collective invasion into surrounding tissues, including cell–cell adhesions, the presence of E-cadherin, intermediate filament proteins (cytokeratin 8, 18, and 19, vimentin), a cytoskeletal binding protein (plectin) [81], and occurrence of other cell–cell adhesion receptors [6, 7].

Most invasive solid tumors exhibit a predominantly collective migration phenotype, in which cohesive clusters of cells invade the peritumoral stroma [6, 82]. Tumor invasion, for example in the bladder, prostate, stomach, or colon, occurs as a multi-step process when epithelial tumor cells first invade the connective tissue that separates the epithelial and muscle layers, which is then followed by muscle invasion [2]. Bladder tumors exhibit cohesive migration as cells invade the smooth muscle surrounding the bladder [2, 81] in a chain-like fashion, suggesting either forced or collective migration orienting along the linear endomysium [39, 40]. Muscle invasive tumor cells generate invadopodia, filamentous actin–based (F-actin) membrane protrusions that are necessary for the breakdown of the ECM and migration through muscle [83].

Many of the basic principles also apply to other epithelial cancers. Cohesive clusters of human prostate cancer cells frequently are observed at multiple stages of dissemination—as a lymph node metastasis, within the vasculature as clustered circulating tumor cells (CTCs), or within bone. When these cohesive clusters are analyzed, they contain no mitotic figures and are Ki-67-negative, both in breast cancer CTCs [84] and in prostate cancer metastases [85, 86]. These results suggest that single CTCs may not be the primary origin of metastatic tumors but, rather, that cohesive CTC clusters, which have been identified as more efficient than individual CTCs in seeding distant metastases [87, 88] (reviewed in [31]), should be a primary therapeutic target.

Friedl and Gilmour identified three basic qualities of collective cell migration: (1) cell clusters remain connected and the cell–cell junctions are preserved; (2) multicellular polarity and actin cytoskeleton organization produce adhesive friction and protrusion of the leading edge of the cluster while preserving cell–cell junctions; and (3) clusters of moving cells often change the tissue structure of vessels as they travel, either by removing obstacles or by generating secondary modifications of the ECM, including the creation of a basement membrane [7]. An important aspect of CTC cluster migration is the establishment of “leader” cells and “follower” cells within the cluster that have been found in all migrating collectives described—including morphogenesis, wound repair, and cancer invasion [7]. The molecular mechanisms underlying single-cell polarization and migration are well-known, and the same basic mechanisms apply to collective movement (reviewed in [89]).

In model systems, leader cells respond to the microenvironment and, therefore, control the direction and speed of migration of the cohesive cluster, as well as being exposed to more external signals (for example, chemoattractants or mechanical stressors) and being largely responsible for ECM remodeling during migration. In the remainder of the cluster, cell–cell adhesions reduce formation of a classical leading-edge in the individual cells, suggesting that the mechanisms driving the migration of follower cells are different from those impacting the leader cells [89]. Furthermore, leader cells often appear less ordered and more mesenchyme-like, while cells at the back of a cluster tend to demonstrate more adhesive assemblages, such as tubular networks, and contain tight cell–cell junctions generally missing in leader cells [7]. These variances in cellular processes may support the existence of a mixed phenotype in cohesive clusters—neither fully epithelial nor fully mesenchymal—and explain why the proposed EMT has been difficult to prove (reviewed in [13, 90, 91])—in fact, the model has now been modified to include the concept of an EMT spectrum [46]. We note with interest that the remarkable molecular heterogeneity of phenotypes within epithelial cancer clusters was directly observed by immunohistochemistry [92] and may provide a selective advantage for tumors invading into and through a highly structured muscle during the first stage of metastasis.

5. Future research of the integrated biophysical responses in the tumor-muscle microenvironment

5.1. Integrin-mediated adhesion to ECM and cell–cell interactions during tumor progression

The mechanical stiffness of the ECM significantly influences normal cell function, stem cell differentiation, and tissue homeostasis [93–95]. Abnormalities in ECM stiffness, in which the ECM becomes dysregulated and disorganized, contribute to the onset and progression of various diseases, such as fibrosis and cancer [93, 96]. For example, ECM irregularities can dysregulate the activity of stromal cells, facilitate tumor-associated angiogenesis and inflammation, and thereby create a tumorigenic microenvironment [96]. Dysregulation of ECM dynamics can enable cellular dedifferentiation and cancer stem cell growth, leading to disrupted tissue polarity that promotes tumor invasion [96]. As a result, epithelial cells are directly affected by dysregulated ECM dynamics, leading to cellular transformation and metastasis [96]. Tumors are often much stiffer than the surrounding normal tissue; however, solid cancer tissues can be up to 10-fold stiffer than healthy tissues, which is correlated with tumor cell survival and enhanced proliferation [93], and this factor may support the tumor’s ability to penetrate smooth muscle.

In early prostate cancer, known as high-grade prostatic intraepithelial neoplasia (HG-PIN), the immediate ECM is dramatically altered with the loss of laminin 332 and attenuation of the normal basal cell component of the gland. Both of these normal components are essential for glandular tissue integrity and are not present in prostate cancer [97, 98]. Using a 3D model system, loss of a laminin-binding integrin essential for normal glandular architecture will result in the budding of the cancer into the alternative surrounding ECM [99]. The loss of the normal tissue architecture and loss of normal contextual cues in HG-PIN [100] coincides with the new and direct exposure of the luminal cells to a new form of laminin in muscle (laminin 511) and the biophysical forces inherent in the muscle. Interestingly, HG-PIN contains genomic instability [101–103] and suggests that the loss of early contextual signals or increased biophysical stressors results in genomic instability early in the disease. We note with interest that in model systems, biophysical stresses on the nucleus of single migrating cells cause excess DNA damage leading to heritable genomic variation [104]. It remains to be determined if this occurs in tumor clusters undergoing the biophysical stresses inherent in the smooth muscle microenvironment.

Cells can sense and process mechanical information embedded in their extracellular environment that determines their growth, motility, and differentiation [105], a process known as mechanosensing [106]. Mechanotransduction is any of the various mechanisms by which cells convert “sensed” mechanical information into electrochemical activity [93]. The conversion of mechanical force into biochemical signals is critical in the development, physiology, and pathology of most tissues and has been extensively reviewed elsewhere [88]. In this process, integrins play important roles, as mechanosensors and as direct mechanotransducers—as transmitters of force to other elements—but also as intermediaries on pathways initiated by other receptors [93].

Integrins, and integrin-mediated adhesions, are known to act as the primary molecular link attaching cells to the ECM and to function as bidirectional hubs transmitting signals between cells and their environment [107]. Integrins employ their combined biochemical and mechanical properties to sense, respond to, and interact with ECM of differing properties with precision [107]. Focal adhesions (FAs) are assemblies of integrins and other proteins that form when the integrin interacts with the ECM and act as essential structures regulating mechanosensing [108, 109]. FAs link the extracellular matrix to the intracellular cytoskeleton and mediate the transmission of forces from the cytoskeleton to the ECM of the cell, and vice versa [108, 110, 111]. Furthermore, FAs act as a link between integrin receptors and the actin microfilaments within the cell; connecting the actin cytoskeleton—and actin-mediated tension—is essential for preserving the integrity of FAs [111]. In addition, integrin nano-clusters can bridge the ECM fibers to form cell-matrix adhesions that are exquisitely sensitive to change [112]. The cooperation of integrins with other cell adhesion molecules such as E-cadherin provides lateral associations for a highly integrative network [113] that would organize cancer clusters and forces [67] required for invasion into and through a tensile smooth muscle.

Cell–cell adhesion complexes join cells to form mechanically coherent tissues that can resist detachment forces and, recently, there is an increased understanding that the mechanical force and cell–cell adhesion has greater variety range and subtlety [114]. One advance that has contributed to our knowledge in this area was the discovery that force patterns are actively driven by cells—for example, deformations of the ECM or by neighboring cells [115], especially contraction forces—and produce morphogenetically significant results when coupled to cell–cell adhesion (reviewed in [114, 116]). These are extracellular forces impacting cell–cell adhesion. There are also intracellular forces, such as apoptosis or cell division, that impact these adhesions [105, 114].

Cadherin-based junctions and contractility are integrated, forming one of the key regulators of mechanotransduction in epithelia, where E-cadherin forms homophilic adhesive complexes [115–117]. Cortical tension and cell–cell adhesion are intimately linked. Actomyosin tension and cadherin-based adhesion both contribute to the interfacial energy in epithelial tissues and, additionally, cadherin adhesion mechanically joins the cortices of adjoining cells at their contacts [118] and allows cell-level forces to produce supracellular, tissue-level forces [116]. Epithelial dynamics depend on how the mechanical properties of their E-cadherin-based cell–cell junctions are regulated during processes such as apoptosis or cell extrusion, for example [116].

Both E-cadherin and α6β1 integrin are found in a cell–cell location in human prostate cancer that has invaded into and through the smooth muscle. Tumors co-express α6β1 integrin and E-cadherin in a cell–cell location and α6β1 integrin in a cell–ECM distribution [32]. The distribution of α6β1 integrin and E-cadherin expression persists in higher grade aggressive prostate cancer that is angulated and invading between the desmin filaments within smooth muscle [32]. This observation is in line with the recent assertion that E-cadherin and integrins together form a strong adhesive network capable of withstanding increased biophysical forces [112]. Using a xenograft model system and a Crispr-Cas9 gene editing approach, eliminating the α6β1 integrin in tumor cells resulted in no invasive networks in vitro and the xenograft tumor did not invade into the smooth muscle of the mouse [32]. Unexpectedly, the tumor cells expressing α6β1 integrin with an ectodomain mutation had a gain of phenotype with increased cell–cell adhesion and a 30-fold increase in cell–cell biophysical properties at the apparent expense of the cell–ECM interactions [32], also resulting in no smooth muscle invasion. The increase in cell–cell proteins such as E-cadherin and claudins (CLDN 4,7) are part of an epithelial-specific adhesome [119] and clusters of integrin and cadherin organize adhesions for cell mechanics [65, 67, 113]. Integrins also regulate force transmission within and between cells, thereby playing an essential role in transmitting tension during morphogenesis [66]. Since the α6β1 integrin is the preferred adhesion receptor for laminin 511 [120] (an abundant ECM protein in muscle) and this integrin can alter cell–cell adhesion proteins such as E-cadherin, it will be of significant interest to determine the role of the integrin-cadherin coordination during tumor invasion into and through the dynamic and tensile smooth muscle. Previous studies suggest that a coordinate regulation of cell–ECM and cell–cell adhesion machinery would promote cohesive cluster invasion in muscle.

5.2. Gland-encasing adipose tissue as a chemoattractant source

While understanding the biophysical and cell biological requirements of tumor invasion into and through the muscle is important, one compelling question is how this transition is triggered. While many local and systemic chemoattractant possibilities exist, we note that abdominal fat (visceral fat, as opposed to subcutaneous fat between the musculature and the skin) is necessary for health, acting as a cushion to the internal organs (Fig. 1) but is also implicated in an increased risk factor in multiple forms of cancer [121]. As has been noted in uterine cancer among others, there is evidence that visceral fat and subcutaneous fat have opposing patterns of gene expression that may increase cancer risk for visceral compared to subcutaneous fat [122]. Visceral fat, which is composed of white adipose tissue (WAT), also serves as lipid energy storage (in the form of triglycerides), whereas brown adipose tissue dissipates energy as heat due to its plentiful mitochondria and greater degree of enervation and capillarization [123]. High levels of visceral WAT, as seen in obesity, have been linked to an increased risk and aggressiveness of many types of cancer [121, 124], including esophageal adenocarcinoma, colorectal, uterine, ovarian, and nine other cancers [125]. There is additional evidence for the role of PPAT in prostate cancer [17, 18] and for perivesical fat in bladder cancer [40, 41]. It is now estimated that up to 40% of cancer deaths can be attributed to obesity [126] through myriad pathways, including inflammation, dysbiosis (severe imbalance in the microbiome), and other factors that influence malignancy [127].

The importance of adipose tissue in tumor initiation, growth, and aggressiveness is founded on two observations: (1) increased research has demonstrated a role for obesity in at least 13 different cancers [125], and (2) adipocytes comprise a significant portion of the tumor microenvironment for several cancers, especially abdominally metastasizing cancers (e.g., gastric, colon, and uterine), promoting tumor growth [121, 128]. The presence of higher levels of visceral adipose tissue in obesity reflects a shift from type 2 anti-inflammatory cytokines to a type 1 pro-inflammatory state, which is associated with increased production of pro-inflammatory cytokines, such as TNF-α, IL-1β, and IL-6 [127].

Each of these pro-inflammatory cytokines is linked to aggressive, metastatic tumors and may act as chemoattractants that promote muscle invasion in epithelial cancers. For example, tumor necrosis factor-α (TNF-α), a protein made by monocytes/macrophages in response to an antigen or infection, is associated with increased migration and invasion of colon cancer cells [129]. Interleukin 1β (IL-1β) generates a cascade of inflammation regulators, resulting in increased tumor angiogenesis, enhanced invasiveness, and tumor-mediated immuno-suppression [130]. Interleukin-6 (IL-6) is produced by cancer cells and inflammatory and stromal cells and is a well-known promoter and marker of aggressive prostate cancer [131], as well as promoting chemoresistance in gastric cancer [132]. Furthermore, in both prostate cancer and bladder cancer, an adipose-secreted chemokine, CXCL1, chemoattracts adipose stromal cells (ASCs), mobilized from WAT, which can then be recruited by the tumor to stimulate cancer progression [41, 133]. Mesenchymal stromal cells in WAT, or ASCs, are perivascular cells that act as adipocyte progenitors [134]. In addition to the pro-inflammatory cytokines mentioned above, WAT also contains essential and adaptive immune cells, including macrophages, dendritic cells, mast cells, eosinophils, neutrophils, and T and B lymphocytes [121].

Human visceral WAT from prostate cancer patients contains macrophages that produce and secrete high levels of extracellular nicotinamide phosphoribosyltransferase, a pro-inflammatory cytokine called eNAMPT and visfatin [135], which functions as a damage-associated molecular pattern protein (DAMP) and master regulator of inflammation [136]. The eNAMPT-secreting macrophages are CD14+ and classified as myeloid-derived suppressor cells [137] and are increased in metastatic prostate cancer patients [138]. eNAMPT or visfatin is a secreted unique cytokine that was originally discovered as a pre-B cell colony-stimulating factor [139] and is present in muscle and adipose tissue. Although eNAMPT is associated with glucose homeostasis in obesity, peripheral actions on muscle are known [140]. Interestingly, eNAMPT will mediate enhancement of prostate cancer invasiveness into muscle tissues. Furthermore, prostate cancer invasion is dramatically attenuated by weekly treatment with an eNAMPT-neutralizing polyclonal antibody [141]. Therefore, eNAMPT has emerged as a novel biomarker of muscle invasive aggressive prostate cancer and a viable therapeutic target [141].

The considerable research demonstrating that cancers surrounded by WAT are linked to obesity suggests that tumor growth is driven by local exposure to adipose cells [121]. Although adipocytes are characteristically nonexistent within epithelial glands, in invasive tumors, the adipocytes come in direct contact with tumors, especially in the reproductive (prostate, uterus) and digestive organs [121]. Other epithelial cancers (gastric, colon, serous endometrial) seed or directly invade the omentum, a WAT pad that extends from the stomach over the gastrointestinal tract [142]. Tumor-associated WAT has been found to play a role in aggressive prostate cancer, and extracapsular extension into WAT beyond the boundaries of the gland is a primary indicator of aggressive disease [17, 143, 144]. Taken together these observations suggest that the surrounding WAT could act as a chemoattractant for the tumor to invade the weakened muscle that encapsulates the organs, driving metastasis.

6. Summary

The smooth muscle barrier offers a dynamic tensile and organized barrier to tumor clusters attempting to escape the originating gland. When the tumor successfully escapes, survival rates drop dramatically for nearly all epithelial cancers. A variety of physiologically relevant factors in the tumor microenvironment contribute to the ability of cohesive clusters of tumor cells to penetrate and move through the smooth muscle barrier, including cell–cell and cell–ECM mechanosensing receptors and the response of the muscle to the tumor. Understanding the phenotype plasticity of the tumor as a moving cluster of cells in the changing microenvironment of the muscle will provide a new perspective and likely uncover new mechanisms of mechanosensing coordination and tumor cell survival. The use of model systems to test the significance of the variety of pro-inflammatory cytokines secreted by the adipose tissue or their associated tumor-infiltrating macrophages will likely yield new relevant factors required for smooth muscle invasion. With the advent of positional phenotyping to understand the significance of tumor phenotype plasticity and the response of the smooth muscle to tumor invasion will likely reveal new diagnostic tools to aid conventional treatment decisions and new therapeutic targets to block metastasis. With an increased understanding of the specific adhesion receptors in the tumor clusters that are required for invasion, the response of the muscle to tumor clusters during invasion, and the potential for WAT exposure to act as a sustained chemoattractant for tumor to escape from the primary organ, we believe that a multi-omics and multi-investigator approach to invasive tumors may offer a pathway toward mitigating tumor invasion into smooth muscle and a reduction in metastatic disease.

Data availability

N/A