Cancer cells are surrounded by a fluid–mosaic membrane that provides a highly dynamic structural barrier with the microenvironment, communication filter and transport, receptor and enzyme platform. This structure forms because of the physical properties of its constituents, which can move laterally and selectively within the membrane plane and associate with similar or different constituents, forming specific, functional domains. Over the years, data have accumulated on the amounts, structures, and mobilities of membrane constituents after transformation and during progression and metastasis. More recent information has shown the importance of specialized membrane domains, such as lipid rafts, protein–lipid complexes, receptor complexes, invadopodia, and other cellular structures in the malignant process. In describing the macrostructure and dynamics of plasma membranes, membrane-associated cytoskeletal structures and extracellular matrix are also important, constraining the motion of membrane components and acting as traction points for cell motility. These associations may be altered in malignant cells, and probably also in surrounding normal cells, promoting invasion and metastatic colonization. In addition, components can be released from cells as secretory molecules, enzymes, receptors, large macromolecular complexes, membrane vesicles, and exosomes that can modify the microenvironment, provide specific cross-talk, and facilitate invasion, survival, and growth of malignant cells. Cancer Res; 75(7); 1169–76. ©2015 AACR.

Cell membranes represent important cellular barriers and first-contact structures of normal and cancer cells. Extracellular signals from ions, hormones, cytokines, enzymes, growth and motility factors, receptors, extracellular matrix (ECM), other stromal elements, and subcellular membrane vesicles must first interact with the cell membrane to initiate signaling processes. Therefore, cell or plasma membranes are cellular filters that can selectively transmit signals and substances from outside cells and from adjacent cells into a cell's interior. Conversely, they can also release signals and molecules to other cells and the micro- and macro-environment in a complex process that has been termed “social cell biology” (1). In addition, cells are compartmentalized into organelles by various intracellular membrane structures that are responsible for biosynthesis, energy production, replication, transportation, recycling, destruction, secretion, and other cellular activities.

Cell membranes are intimately involved in the biochemical events that define cancers, and in particular, they are intensely involved in cancer metastasis (2). In addition, the establishment of metastases also requires a complex interplay between malignant cells, normal cells, stroma, and ECM in their new microenvironments, and these interactions are primarily mediated through cell membranes (3).

An important concept that maintains cell membrane structure is that amphipathic membrane components self-associate to exclude water interactions on their hydrophobic surfaces, whereas the hydrophilic portions of their structures interact with the aqueous environment (4). Thus, membrane glycerolphospholipids self-assemble to form lipid bilayers (5) due to the energy provided by the hydrophobic effect and van der Waals forces (6). Membrane integral globular proteins interact with membrane lipids through their acyl structures due to hydrophobic forces and much less to hydrophilic interactions between lipid head groups and protein hydrophilic amino acids (4, 6, 7).

Membrane proteins are operationally of three types: integral, peripheral, and membrane-associated (7). Integral proteins are globular and tightly bound to membranes by mainly hydrophobic forces and intercalated into the membrane lipid bilayer, whereas peripheral membrane proteins are bound to membranes by electrostatic or other forces. Peripheral membrane proteins can be removed from membranes without destroying basic membrane microstructure and are important in providing membrane attachment sites, scaffolding, tethering or membrane-supporting structures, membrane curvature-promoting components, and attachment points for soluble enzymes and signaling molecules (7, 8).

When membrane are distorted, deformed, compressed, or expanded, different forces and components react to the physical perturbations (9, 10). For example, certain peripheral membrane proteins can bind and cause deformation by forming crescent-shaped α-helical bundles that bind to membranes via electrostatic and some hydrophobic interactions (9), causing curvature by bending membranes to fit peripheral protein structure (9, 10).

The third class of membrane proteins is often isolated with cell membranes, but they are actually not membrane proteins (7). These are cytoskeletal and associated signaling structures at the inner cell membrane surface, and at the outer surface they include certain ECM components. These membrane-associated components are parts of dynamic structures involved in stabilizing membranes (and thus cells) and immobilizing membrane components. They are especially important in cellular activities, such as cell adhesion and motility, growth, endocytosis, exocytosis, signal transduction, and other important activities (11, 12).

First proposed in 1972, the fluid–mosaic membrane model (8) has proven its usefulness in describing basic nanoscale cell membrane structure for over 40 years (11). Although this model has been remarkably consistent with data collected on biologic membranes since 1925, it was inevitable that the original model could not explain aspects of membrane dynamics. For example, the concepts that membrane domains and cell membrane–associated structures are important in controlling the lateral mobilities and distributions of cell membrane proteins were not yet discovered (11–14).

Models of cell membrane structure produced a few years after the original model (8) were much less homogeneous (7). They contained additional information on protein and lipid aggregations and their segregation into membrane domains, cytoskeletal, and ECM interactions, among other features (7, 11). Nonetheless, in subsequent revisions of the fluid–mosaic model, all of the basic elements at the nanoscale level were retained (11–14). However, the arrangements of lipids and proteins into more compact structures and domains that maximized their mosaic nature along with the addition of dynamic hierarchical membrane organization produced a much more detailed description of its organization (Fig. 1A; refs. 11–14).

Figure 1.

A hypothetical cancer cell undergoing change to an invasive phenotype and beginning the process of invasion. A, a representation of the cell membrane that contains membrane domain structures and membrane-associated cytoskeletal and extracellular structures. The cell membrane has been peeled back at the right to reveal the bottom membrane surface and membrane-associated cytoskeletal elements that form barriers (corrals) that limit the lateral motions of some of the integral membrane proteins. In addition, membrane-associated cytoskeletal structures are indirectly interacting with integral membrane proteins at the inner membrane surface along with matrix or ECM components at the outer surface. Although this diagram presents possible mechanisms of integral membrane protein mobility restraint, it does not accurately represent the sizes and structures of integral membrane proteins, lipid domains, or membrane-associated cytoskeletal structures. B, exosomes or intracellular membrane vesicles are released from larger vesicles that fuse with the cell membrane. MF, microfilaments; MT, microtubule. Reprinted from BBA Biomembranes, 1838 (6), Nicolson GL, The Fluid–Mosaic Model of Membrane Structure: Still relevant to understanding the structure, function and dynamics of biological membranes after more than 40 years, 1451–1456, Copyright 2013, with permission from Elsevier.

Figure 1.

A hypothetical cancer cell undergoing change to an invasive phenotype and beginning the process of invasion. A, a representation of the cell membrane that contains membrane domain structures and membrane-associated cytoskeletal and extracellular structures. The cell membrane has been peeled back at the right to reveal the bottom membrane surface and membrane-associated cytoskeletal elements that form barriers (corrals) that limit the lateral motions of some of the integral membrane proteins. In addition, membrane-associated cytoskeletal structures are indirectly interacting with integral membrane proteins at the inner membrane surface along with matrix or ECM components at the outer surface. Although this diagram presents possible mechanisms of integral membrane protein mobility restraint, it does not accurately represent the sizes and structures of integral membrane proteins, lipid domains, or membrane-associated cytoskeletal structures. B, exosomes or intracellular membrane vesicles are released from larger vesicles that fuse with the cell membrane. MF, microfilaments; MT, microtubule. Reprinted from BBA Biomembranes, 1838 (6), Nicolson GL, The Fluid–Mosaic Model of Membrane Structure: Still relevant to understanding the structure, function and dynamics of biological membranes after more than 40 years, 1451–1456, Copyright 2013, with permission from Elsevier.

Close modal

By sorting different membrane lipids and integral membrane proteins into specialized membrane domains based on their physical properties, Mouritsen and Bloom (15) proposed that such sorting was based primarily on hydrophobic interactions and some hydrophilic interactions. This prevents hydrophobic mismatches between lipids and proteins, thus preventing membrane distortions (15).

The fluid–mosaic membrane model also accounted for cell membrane asymmetry (8). Cell membranes are asymmetric in the display of their components (11, 16). The finding of asymmetric distributions of various lipids, proteins, and glycoproteins between the inner and outer leaflets of cell membranes is likely universal (11, 16, 17). Moreover, the disruption of membrane asymmetry in cell membranes is associated with cell activation, adhesion, aggregation, apoptosis, recognition by phagocytic cells, among other events. Of note, it is also associated with pathologic processes (17, 18).

What the original fluid–mosaic membrane model lacked was the integration of this structure with other cellular elements (7, 11). Cytoskeletal and ECM interactions are known to alter cell membrane macrostructure by restrictions in the freedom of movement (lateral mobility) of membrane proteins and also causing global movements of these and other components by tethering them to cellular or extracellular structures (7, 11, 12). This process can result in endocytosis of some macromolecular complexes at the cell surface. Receptor clustering, domain formation, submembrane plaque assembly, internalization, acidification of the resulting endosomes, degradation, and membrane recycling are all part of normal membrane recycling (19, 20). The mobility of integral membrane components can also be controlled by cell–cell and cell–ECM interactions (21).

Cell adhesion and receptor complexes that are immobilized by ECM or various interactions are capable of communicating signals that are transmitted through a dynamically assembled cytoskeleton or generating mechanical forces that can move cells or resist exterior mechanical stresses (22). This serial assembly of specialized components (ECM, integral membrane proteins, peripheral membrane proteins, adaptor proteins, cytoskeletal elements, among others) may have evolved to convert biochemical signals into mechanical forces that are important in cellular behavior. Although many of the membrane peripheral proteins have been identified as components involved in cytoskeletal interactions with membranes (23), membrane lipids are also important in these interactions as specialized lipid domains or “lipid rafts” (20, 24).

Membrane domains are dynamic structures that can be generated by ligand or ion binding, hydrophobic interactions (or other events) and can assemble into complex transmembrane super-structures. These complexes recruit additional peripheral proteins at the inner cell membrane surface to form transmembrane plaques that are competent for initiating cellular signaling via enzymatic processes or undergoing further attachment to cytoskeletal elements (3, 22, 25, 26).

Cell membranes should be considered completely integrated mechanostructures within tissues. They continuously interact with and link various intracellular structures to components outside the cell while receiving signals and contacts from the microenvironment and passing these signals on to elicit appropriate cellular responses. They also send out messages, maintain cell polarity, and mechanical properties while undergoing constant turnover of their constituent components. Thus, the basic structure of cell membranes has evolved from the original homogeneous concept to one that contains specific “domains” of varying sizes that form specific membrane regulatory and mechanical structures that are linked to other intra- and extracellular structures (11, 12) that are involved in many cellular properties characteristic of normal and cancer cells (3, 11).

In addition to their asymmetric distributions across the membrane, membrane lipids are also unevenly distributed in the membrane plane (5, 27, 28). Cholesterol is particularly important in cell membrane organization and is often found in specific membrane domains (27–29). This is thought to be due, in part, to cholesterol's affinity for both the fluid and solid phases of membranes (29). Cholesterol partitions into liquid-ordered/-disordered phases to roughly the same extent and changes the properties of the lipid phases (30). Sphingolipids are also important in the formation of ordered membrane lipid domains (31). Sphingomyelins and phosphatidylcholines constitute more than one half of plasma membrane phospholipids and form the main partners for cholesterol (32). Indeed, sphingomyelins and cholesterol are critically important in formation of small, ordered lipid domains (“lipid rafts”) that are generally surrounded by liquid-phase lipids (32, 33).

The formation of lipid rafts is a dynamic and reversible process that confers functional signaling properties to cell membranes. As mentioned, lipid rafts are characterized by enrichments of cholesterol and sphingolipids, which are held together by hydrogen bonds, charge pairing, and hydrophobic and van der Waals forces (32–34). Their constituents can quickly exchange with bulk membrane lipids as well as with lipids in other rafts (35). Lipid rafts are generally nanosized (<300 nm diameter, most ∼10–200 nm) domains that can contain integral and peripheral membrane proteins. The proteins sequestered into membrane rafts can turn these domains into functional signal transduction structures important in endocytosis, cell death regulation, and other events relevant to cancer therapy (34, 36).

Integral membrane proteins can interact within different membrane domains, but they must also interact with membrane lipids to produce an intact plasma membrane (8, 11, 15, 37). Specifically, portions of their structures must directly pair with the acyl chains of membrane phospholipids or the hydrophobic portions of other membrane lipids. This is accomplished by hydrophobic matching (15, 29, 37). The concept of hydrophobic matching between the hydrophobic core of the lipid bilayer and hydrophobic stretches of amino acids in integral membrane proteins is essential for understanding how cell membranes form a stable structure (11, 29, 38). If the hydrophobic portions of this structure are mismatched, an elastic distortion of the lipid matrix around the integral membrane protein occurs (15, 29, 37). This can produce protein conformational changes, potentially effecting protein function and protein–protein interactions, such as protein aggregation into membrane super-structures (37, 39). In addition, there are other physical forces, such as lateral pressure forces, lateral phase changes, membrane curvature, ionic interactions, among other forces, that must be considered to produce an overall tensionless membrane structure (11, 39).

Restrictions on the lateral movements of integral membrane proteins have been attributed to extracellular restrictions, such as ECM, the formation of specialized membrane domains (lipid rafts and larger lipid domains), large protein complexes, and peripheral membrane barriers at the inner membrane surface (Fig. 1A; refs. 11, 12). Jacobson and colleagues (13) have summarized the lateral movements of membrane proteins into distinct categories: (i) transient confinement by obstacle protein clusters (also called protein fenceposts or pickets); (ii) transient confinement into defined domains or corrals by a cytoskeletal meshwork; (iii) directed motion due to attachment to the cytoskeleton; and (iv) random diffusion in the fluid membrane. Therefore, the original description of integral membrane proteins freely diffusing in the membrane plane pertains to only one of these categories (8).

It is now believed that a substantial portion of integral membrane proteins is not capable of free lateral diffusion in the cell membrane; they are confined, at least transiently, to small membrane domains by barriers at the inner membrane surface (11–14). However, integral proteins can escape from one of these domains to an adjacent domain. They can even escape the domains altogether, unless they undergo aggregation and their size prevents extradomain movements. Consequently, the abilities of membrane proteins to move between adjacent domains may be related to their sizes, the sizes of the cytoplasmic barriers, and the complex interactions of these barriers with the cytoskeleton (14) and ECM (11).

The approximate areas of cell membrane receptor domains have been estimated to vary from 0.04 to 0.24 μm2, and the approximate transit times of membrane receptors in these membrane domains can vary from 3 to 30 seconds (14, 40). Thus, cell membrane domains can range in diameter from 2 to 300 nm. For example, actin-containing cytoskeletal-fenced domains have been found in approximate diameters of 40 to 300 nm, lipid raft domains in the range of 2 to 20 nm, and dynamic integral membrane protein complexes in domains of 3 to 10 nm in diameter (14, 40). The presence of different types of cell membrane domains and the selective presence of membrane proteins in these domains suggest another level of membrane compositional complexity beyond the original fluid–mosaic membrane structure (11–14).

Cell membranes must quickly respond to intracellular and extracellular signals and other microenvironmental events. To do this rapidly, it may be more efficient to have receptors prepositioned in the cell membrane within signaling domains so that they can undergo rapid aggregation into supramolecular signaling structures (14). The partitioning of plasma membranes to limit the dynamics of their integral membrane protein components (at least part of the time) to cytoskeletal-fenced corrals, or tethering them directly or indirectly to membrane-associated cytoskeletal elements, can create relatively stable membrane domains of increased receptor densities (40).

Signal transduction, cell activation, identification, differentiation, and other complex membrane-initiated events may require the presence of enhanced receptor densities within specific membrane domains to selectively promote cellular signaling. Kusumi and colleagues (14, 40) have proposed that cell membranes possess hierarchical architectures that consist of a basic fluid–mosaic membrane plus various membrane micro- and nanosized domains or compartments defined by cytoskeletal fencing and protein fenceposts, lipid rafts and other lipid domains, oligomeric integral membrane lipoprotein domains, and other structures. This complex structure is depicted simplistically in Fig. 1A. Within the hierarchical structure (including membrane domains) protein components are limited in their diffusion rates to those that are five to 50 times slower than when the same components are free to diffuse laterally in the membrane without restraint (41).

Cells package newly synthesized materials for transport to various organelles and to the cells' exterior by incorporating them into small membrane vesicles that are delivered to specific target membrane domains (42). This process is also used to rapidly repair damage to the plasma membrane and other cellular membranes (43). Invasive tumor cells use directed exocytosis and other pathways to release degradative enzymes, bioregulators, and other molecules into the microenvironment to facilitate invasion of ECM and tissue stroma and evade host defenses (Fig. 1; refs. 44–46). It is also a mechanism that is used to display newly synthesized receptors, enzymes, and other molecules on the exterior cell membrane surface and facilitate their turnover.

A critical event in the delivery of materials via small vesicles is membrane fusion (47). Membrane fusion events are also apparent at the cell level, for example, when adjacent cancer cells undergo fusion with other cancer cells or normal cells to produce aneuploid cells, complementation, and other characteristics found in progressive neoplastic cells (38, 48). Membrane fusion is obviously not a cancer-associated event; it occurs during many normal cellular processes, such as fertilization, myoblast formation, and bone homeostasis (38).

Directed vesicle transport and fusion inside cells depends to some degree on lipid composition, distribution, and acylation. For example, certain lipids, such as the sphingolipids and sterols found in lipid rafts, are concentrated in vesicles destined to fuse with the plasma membrane (42, 49, 50). Specific polyphosphoinositides with their tethered proteins may also help direct vesicles to particular membrane sites (51, 52). Membrane fusion is dependent on specialized membrane-binding fusion machinery composed of specific proteins (SNARE, SNAP, and SM proteins, among others) that pull adjacent membranes together to promote lipid bilayer fusion (53).

The assembly of fusogenic proteins at the cell membrane constitutes a specialized dynamic membrane microdomain called a porosome (54). In some normal cells, porosomes appear ultrastructurally as “pits” measuring 0.5 to 2 μm in diameter containing depressions of 100 to 180 nm (55). Porosomes are responsible for directing exocytosis to particular sites at the cell surface. This is important for directed cell migration and invasion of ECM and stroma as well as for the normal function of cells secreting necessary proteins, glycoproteins, enzymes, bioregulators, and other important molecules.

To facilitate cell invasion, invading tumor cells can extend specialized actin-rich membrane protrusions called “invadopodia” that penetrate into surrounding ECM, stroma, and basement membranes (Fig. 1; refs. 56, 57). These specialized cell structures display and are associated with extracellular degradative enzymes and contain intracellular actin polymers and their regulators, such as cortactin, cofilin, N-WASP, Arp2/3, and facin (57–59). The comparable normal cell counterparts of invadopodia are called podosomes, and these structures are apparent in many normal cells under conditions of ECM and basement membrane invasion during embryogenesis, wound healing, inflammatory responses, and organ regeneration (58, 59). Differences found between invadopodia and podosomes are that they often differ in size, shape, density, and stability—notably, invadopodia are typically stable for much longer periods of time compared with podosomes (59).

Invadopodia are stimulated to form by various factors. These include growth factors, oncogenic transformation, epithelial–mesenchymal transition (EMT), hypoxia, adhesion receptors, chemokines, and degradative enzyme activity. They typically require initial attachment to ECM and linkage to cytoskeletal components to initiate formation (57–59). Beaty and Condeelis (60) have proposed that there are four stages (presumably after ECM attachment) involved in invadopodia formation: (i) assembly of a core actin (and accessory protein) structure, (ii) activation of associated kinases, assembly of actin polymers, recruitment of adhesion components, and their transmembrane linkage, (iii) actin polymerization and invadopodial elongation and stabilization, and (iv) microtuble and intermediate filament recruitment, elongation of the mature structure along with ECM degradation. Although there is some uncertainty as to the exact stage where adhesion receptors are important in invadopodia formation (initially or in the maturation, elongation stages; ref. 60), there is no doubt that adhesion is an important early element of the invadopodia process (56, 59, 61).

Invadopodia may require specialized cell membrane lipid domains to initiate their organization. Indeed, specialized lipid rafts have been found at the sites of invadopodia formation, and similar rafts are known to be involved in membrane vesicle trafficking, exocytosis, and actin polymerization at the cell membrane inner surface (62). Thus, membrane microdomain formation is an important aspect of the formation of invadopodia and podosomes and probably other processes that require directed membrane distortions, adhesion, vesicle fusion, matrix degradation, and other processes.

For cell invasion to occur, invading cells must have the ability to degrade matrix barriers and migrate along invasion pathways generated by ECM restructuring and destruction. Thus, ECM degradation is an important step in the metastatic process (44, 61–64). Various ECM-degrading enzymes are associated with isolated invadopodia (56, 59), and degradative enzymes appear to be released by exocytosis near sites of invadopodia (61, 65). In addition, cell membranes at invadopodia sites appear to bind to and mechanically orient loosened matrix components parallel to cell surfaces to assist in mechanical force generation into ECM tunnels (61, 65). Therefore, cell membranes, degradative enzymes, and invadopodia can mechanically and enzymatically restructure ECM to facilitate cell movement and invasion by reshaping ECM into tube-like structures in which cancer cells can invade. Eventually collective cell movements can form into a massive multicell invasive structure that penetrates along the tubes generated by invadopodia and then single cells (65).

Tumor cells naturally release small (0.1–2 μm diameter) extracellular membrane vesicles (EV) derived from budding plasma membranes and separately even smaller (<100-nm diameter) microvesicles called exosomes from exocytosed intracellular vesicles (Fig. 1; refs. 66, 67). The released membrane vesicles can contain various molecules, such as small fragments of DNAs, microRNAs, proteins (enzymes, biomodulators, and receptors), and carbohydrates (67–69). Along with their enriched plasma membrane components and various cell receptors, they can mediate a form of communication by transfer between tumor cells, resulting in exchange of cellular materials. They can also mediate communication between tumor and normal cells in the microenvironment (67–69). This is not a unique property of tumor cells—vesicles released from normal cells are found in virtually every extracellular fluid where they appear to play a role in normal cell communication and regulation of inflammation, coagulation, development, and other normal physiologic processes (69, 70). In various cancers, they can also affect tumor interactions with the microenvironment and promote progression, angiogenesis, invasion, and metastasis (69, 70–72).

Various factors can influence the release of EV and exosomes. Among these are cellular energy, intracellular levels of calcium, changes in membrane phospholipids, and other regulators of cytoskeleton-membrane interactions, membrane-acting enzymes, and other effectors of exocytosis, hypoxia, and oxidative and shear stresses (67, 71). Release of EV and exosomes are also affected by invadopodia, which can enhance their release and drive tumor invasive behavior (72). Released EV and exosomes are quite heterogeneous in their composition and reflect cell-to-cell variations in the cargos, membranes, and cells from which they are derived (70–73). Of significance is that most normal and benign cells do not release large quantities of EV and exosomes until they progress to invasive, metastatic phenotypes (73).

Tumor cell–released EV and exosomes can have profound effects on tumor microenvironment (70–72). In addition to their delivery of nucleic acids, bioactive phospholipids, degradative enzymes, receptors, growth and motility factors, and other cargo that can affect invasion and metastatic colonization, these vesicles can also stimulate angiogenesis, stromal reactions, and release of paracrine and other bioactive molecules from normal surrounding cells that condition the tumor microenvironment (67, 70, 71, 73). For example, the observation that signaling microRNAs in microvesicles can enhance endothelial cell migration and promote angiogenesis indicates that these vesicles have important molecular communication properties (74). Shed tumor membrane vesicles can break down, releasing their contents to the extracellular environment, and this has important consequences for invasion and cell motility. An example is the EV-release of a tumor-derived membrane glycoprotein (EMMPRIN) that stimulates fibroblast release of matrix metalloproteinases that facilitate tumor invasion and metastasis (75). EV and exosomes can also be involved in recruiting lymphoid cells that stimulate tumor growth. For example, blocking exosome secretion by inhibiting small GTPases results in decreased tumor growth and lung metastasis in a breast cancer model by decreasing neutrophil-stimulated tumor cell proliferation (76).

Tumor cell–released EV and exosomes may also protect neoplastic cells from harmful chemicals such as chemotherapeutic drugs, oxidized phospholipids, irradiation, immune responses, and cell death signals (67, 77). Treating human and animal lung cancer cell lines with irradiation and hypoxia that do not stimulate apoptosis results in enhancing the release of EV that activate and chemoattract stromal and endothelial cells. Once activated with tumor-derived EV, the stromal cells then release several pro-angiogenic factors. When this stromal cell–conditioned media were used to stimulate tumor cells, the result was enhanced metastatic potential in vivo (78).

Recently, the use of EV and exosomes for modulating the microenvironment or producing new therapeutics for targeting of specific bioactive molecules to specific sites, such as brain, has been proposed (79). Whether such approaches prove useful clinically remain to be seen, but this novel approach could be an interesting way to target drugs to secondary sites.

The transition of an epithelial to mesenchymal cell phenotype or EMT in carcinoma cells has been proposed as one of the striking changes that accompany invasion and metastasis (80, 81). Unlike the multiple genetic changes that are typical of in vivo transformation and tumorigenesis, EMT appears to be primarily epigenetic and driven for the most part by microenvironmental signals (82). Changes that occur, such as loss of cell adhesion and cell junctions, modifications in cell shape, acquisition of cell motility, release of growth and motility factors and degradative enzymes, among other changes, are characteristic of invasive, malignant cells at the primary site (Fig. 1; refs. 80, 81, 83). At least three different regulatory pathways control cell polarity, adhesion, cell junctions, and other related properties (82). These pathways are driven by tumor microenvironment cross-talk and signaling events starting at the cell membrane surface and then along the relevant signaling circuits inside cells (83). Moreover, once malignant carcinoma cells have metastasized to secondary sites, they can apparently revert back to epithelial-like morphologies along with reexpression of epithelial markers, indicating that EMT has transient and potentially reversible characteristics (80, 81). Although EMT changes have been associated with invasion and metastasis in animal tumors, there remains an ongoing debate on whether this phenomenon is truly representative of malignant pathways in human cancers (84). Indeed, Tarin (85) has argued that invasive cell phenotypes occur after normal tissues are damaged and during embryogenesis and that there is a lack of convincing pathologic evidence in humans that EMT occurs when carcinomas metastasize.

A new approach to developing therapies for highly malignant carcinoma cells has been to target metastasis-related changes, such as cell surface components involved in cell junctions, adhesion, motility, growth, signaling, and other signatures of an invasive phenotype. One place that this could be important is in the identification and typing of circulating tumor cells (CTC; ref. 86). Enumeration of CTC and typing CTC biomarkers are being developed to assess risk of metastatic disease and hopefully predict the effects of therapy to prevent metastases (87). CTC are thought to be directly related to stem cells, the presumptive source of metastases, so Zhang and colleagues (88) isolated CTC from patients with breast cancer without brain metastases, grew them in culture, and subjected them to metastasis assays in immunosuppressed mice. They found that CTC from patients with breast cancer could be selected for a unique “brain-metastasis signature” (EpCAM/HER2+/EGFR+/HPSE+/Notch1+) that may explain, in part, the ability of CTC to form brain metastases. Using the unselected parental EpCAM CTC and the brain-metastasis marker-selected CTC, they found that only the latter were highly invasive and capable of forming brain metastases when xenographed into nude mice (88). Although this report involved only a few patients, future studies may be able to unlock the CTC-targeting signatures for metastasis to specific secondary sites. Such information could prove to be more useful than biopsies in predicting metastatic disease, especially site-specific metastases, and eventually preventing metastases from forming and treating the metastases that have formed (86, 87).

Although tremendous progress has been attained over the last decade in ascertaining the structural and functional roles of cell membranes in malignant processes, much remains to be discovered at the cellular and molecular level. In particular, we know very little about how microenvironmental signals and cross-talk can rapidly change the phenotypes of cancer and normal cells within the tumor microenvironment and how various signaling pathways, from the cell membrane to various cellular organelles, and vice versa, control these complex interactions. Uncoding the plasticity of this process may be essential in explaining metastatic behavior (89).

We are just beginning to understand the cell surface and surrounding properties of malignant cells (and normal cells) that are important in explaining metastasis to secondary sites as well as the properties of target organs for metastatic colonization. This information will not be easily forthcoming, but it will be essential in the eventual development of new therapeutic approaches to limit or destroy metastases. It will also be important in reducing the symptoms of cancer and eliminating the adverse effects of cancer therapy (90).

No potential conflicts of interest were disclosed.

1.
Pastor-Pareja
JC
,
Xu
T
. 
Dissecting social cell biology and tumors using Drosophila genetics.
Annu Rev Genet
2013
;
47
:
51
74
.
2.
Nicolson
GL
. 
Cell surfaces and cancer metastasis.
Hospital Pract
1982
;
17
:
75
86
.
3.
Xu
R
,
Bondreau
A
,
Bissell
MJ
. 
Tissue architecture and function: dynamic reciprocity via extra- and intra-cellular matrices.
Cancer Metastasis Rev
2009
;
28
:
167
76
.
4.
Singer
SJ
. 
The molecular organization of membranes.
Annu Rev Biochem
1974
;
43
:
805
33
.
5.
Edidin
M
. 
Lipids on the frontier: a quarter century of cell-membrane bilayers.
Nat Rev Mol Cell Biol
. 
2003
;
4
:
414
8
.
6.
Cramer
WA
,
Engelman
DM
,
von Heijne
G
,
Rees
DC
. 
Forces involved in the assembly and stabilization of membrane proteins.
FASEB J
1992
;
6
:
3397
402
.
7.
Nicolson
GL
. 
Transmembrane control of the receptors on normal and tumor cells. I. Cytoplasmic influence over cell surface components.
Biochim Biophys Acta
1976
;
457
:
57
108
.
8.
Singer
SJ
,
Nicolson
GL
. 
The Fluid Mosaic Model of the structure of cell membranes.
Science
1972
;
175
:
720
31
.
9.
Zimmerberg
J
,
Kozlov
MM
. 
How proteins produce cellular membrane curvature.
Nat Rev Mol Cell Biol
2006
;
7
:
9
19
.
10.
Baumgart
T
,
Capraro
BR
,
Zhu
C
,
Das
SL
. 
Theromodynamics and mechanics of membrane curvature generation and sensing by proteins and lipids.
Annu Rev Phys Chem
2011
;
62
:
483
506
.
11.
Nicolson
GL
. 
The Fluid-Mosaic Model of Membrane Structure: still relevant to understanding the structure, function and dynamics of biological membranes after more than 40 years.
Biochim Biophys Acta
2014
;
1838
:
1451
66
.
12.
Nicolson
GL
. 
Update of the 1972 Singer–Nicolson Fluid-Mosaic Model of membrane structure.
Discoveries
2013
;
1
:
e3.
13.
Jacobson
K
,
Sheets
ED
,
Simson
R
. 
Revisiting the fluid mosaic model of membranes.
Science
1995
;
268
:
1441
2
.
14.
Kusumi
A
,
Fujiwara
TK
,
Chadda
R
,
Xie
M
,
Tsunoyama
TA
,
Kalay
Z
, et al
Dynamic organizing principals of the plasma membrane that regulate signal transduction: commemorating the fortieth anniversary of Singer and Nicolson's fluid-mosaic model.
Annu Rev Cell Dev Biol
2012
;
28
:
215
50
.
15.
Mouritsen
OG
,
Bloom
M
. 
Mattress model of lipid–protein interactions in membranes.
Biophys J
1984
;
46
:
141
53
.
16.
Rothman
JE
,
Lenard
J
. 
Membrane asymmetry.
Science
1977
;
195
:
743
53
.
17.
Daleke
DL
. 
Regulation of transbilayer plasma membrane phospholipid asymmetry.
J Lipid Res
2003
;
44
:
233
42
.
18.
Pomorski
TS
,
Hrafnsdottir
S
,
Devaux
PF
,
van Meer
G
. 
Lipid distribution and transport across cellular membranes.
Semin Cell Dev Biol
2001
;
12
:
139
48
.
19.
Geiger
B
,
Yehuda-Levenberg
S
,
Bershadsky
AD
. 
Molecular interactions in the submembrane plaque of cell-cell and cell–matrix adhesions.
Acta Anat
1995
;
154
:
42
62
.
20.
Chichili
CR
,
Rogers
W
. 
Cytoskeleton–membrane interactions in membrane raft structure.
Cell Mol Life Sci
2009
;
66
:
2319
28
.
21.
Salas
PJ
,
Vega-Salas
DE
,
Hochman
J
,
Rodriguez-Boulan
E
,
Edidin
M
. 
Selective anchoring in the specific plasma membrane domain: a role in epithelial cell polarity.
J Cell Biol
1988
;
107
:
2363
76
22.
Geiger
B
,
Bershadsky
A
. 
Assembly and mechanosensory function of focal contacts.
Curr Opin Cell Biol
2001
;
13
:
584
92
.
23.
Schwarz
US
,
Gardel
ML
. 
United we stand: integrating the actin cytoskeleton and cell-matrix adhesions in cellular mechanotransduction.
J Cell Sci
2012
;
125
:
3051
60
.
24.
Janmey
PA
,
Lindberg
U
. 
Cytoskeletal regulation: rich in lipids.
Nat Rev Mol Cell Biol
2004
;
5
:
658
66
.
25.
Cho
W
. 
Building signaling complexes at the membrane.
Sci. STKE
2006
;
2006
:
pe7
.
26.
Geiger
B
,
Bershadsky
A
,
Pankov
R
,
Yamada
KM
. 
Transmembrane extracellular matrix-cytoskeleton crosstalk.
Nat Rev Mol Cell Biol
2001
;
2
:
793
805
.
27.
Somerharju
P
,
Virtanen
JA
,
Cheng
KH
. 
Lateral organization of membrane lipids. The superlattice view.
Biochim Biophys Acta
1999
;
1440
:
32
48
.
28.
Simons
K
,
Sampaio
JL
. 
Membrane organization and lipid rafts.
Cold Spring Harb Perspect Biol
2010
;
3
:
a004697
.
29.
Mouritsen
OG
. 
Model answers to lipid membrane questions.
Cold Spring Harb Perspect Biol
2011
;
3
:
a004622
.
30.
Lindblom
G
,
Orädd
G
. 
Lipid lateral diffusion and membrane heterogeneity.
Biochim Biophys Acta
2009
;
1788
:
234
44
.
31.
Lingwood
D
,
Simons
K
. 
Lipid rafts as a membrane-organizing principle.
Science
2010
;
327
:
46
50
.
32.
Ramstedt
B
,
Slotte
JP
. 
Sphingolipids and the formation of sterol-enriched ordered membrane domains.
Biochim Biophys Acta
2006
;
1758
:
1945
56
.
33.
van Meer
G
,
Voelker
DR
,
Feigenson
GW
. 
Membrane lipids. Where they are and how they behave.
Nat Rev Mol Cell Biol
2008
;
9
:
112
24
.
34.
Simons
K
,
Gerl
MJ
. 
Revitalizing membrane rafts: new tools and insights.
Nat Rev Mol Cell Biol
2010
;
11
:
688
99
.
35.
Quinn
PJ
,
Wolf
C
. 
The liquid-ordered phase in membranes.
Biochim Biophys Acta
2009
;
1788
:
33
46
.
36.
Neumann
AK
,
Itano
MS
,
Jacobson
K
. 
Understanding lipid rafts and other related membrane domains.
F1000 Biol Rep
2010
;
2
:
31
6
.
37.
Bagatolli
LA
,
Ipsen
JH
,
Simonsen
AC
,
Mouritsen
OG
. 
An outlook on the organization of lipids in membranes: searching for a realistic connection with the organization of biological membranes.
Prog Lipid Res
2010
;
49
:
378
89
.
38.
Lu
X
,
Kang
Y
. 
Cell fusion as a hidden force in tumor progression.
Cancer Res
2009
;
69
:
8536
9
.
39.
Mouritsen
OG
. 
Lipids, curvature and nano-medicine.
Eur J Lipid Sci Technol
2011
;
113
:
1174
87
.
40.
Kusumi
A
,
Suzuki
KG
,
Kasai
RS
,
Ritchie
K
,
Fujiwara
TK
. 
Hierarchical mesoscale domain organization of the plasma membrane.
Trends Biochem Sci
2011
;
36
:
604
15
.
41.
Poste
G
,
Papahadjopoulos
D
,
Nicolson
GL
. 
Local anesthetics affect transmembrane cytoskeletal control of mobility and distribution of cell surface receptors.
Proc Natl Acad Sci USA
1975
;
72
:
4430
4
.
42.
Diaz-Rohrer
B
,
Levental
KR
,
Levental
I
. 
Rafting through traffic: membrane domains in cellular logistics.
Biochim Biophys Acta
2014
;
1838
:
3003
13
.
43.
Andrews
NW
,
Almeida
PE
,
Corrotte
M
. 
Damage control: cellular mechanisms of plasma membrane repair.
Trends Cell Biol
2014
;
24
:
734
42
.
44.
Deryugina
EI
,
Quigley
JP
. 
Matrix metalloproteinases and tumor metastasis.
Cancer Metastasis Rev
2006
;
25
:
9
34
.
45.
Hammond
E
,
Khurana
A
,
Shridhar
V
,
Dredge
K
. 
The role of heparanases and sulfatases in the modification of heparan sulfate proteoglycans within the tumor microenvironment and opportunities for novel cancer therapeutics.
Front Oncol
2014
;
4
:
195
.
46.
Franco
SJ
,
Huttenlocher
A
. 
Regulating cell migration: calpains make the cut.
J Cell Sci
2005
;
118
:
3829
38
.
47.
Leabu
M
. 
Membrane fusion in cells: molecular machinery and mechanisms.
J Cell Mol Med
2006
;
10
:
423
7
.
48.
Duelli
D
,
Lazebnik
Y
. 
Cell fusion: a hidden enemy?
Cancer Cell
2003
;
3
:
445
8
.
49.
McMaster
CR
. 
Lipid metabolism and vesicle trafficking: more than just greasing the transport machinery.
Biochem Cell Biol
2001
;
79
:
681
92
.
50.
Holthuis
JC
,
van Meer
G
,
Huitema
K
. 
Lipid microdomains, lipid translocation and the organization of intracellular membrane transport.
Mol Membr Biol
2003
;
20
:
231
41
.
51.
van Meer
G
,
Sprong
H
. 
Membrane lipids and vesicular traffic.
Curr Opin Cell Biol
2004
;
16
:
373
8
.
52.
Martin
TF
. 
PI(4,5)P2-binding effector proteins for vesicle exocytosis.
Biochim Biophys Acta
2014
:
pii
:
S1388-1981(14)00195-4
.
53.
Sudhof
TC
,
Rothman
JE
. 
Membrane fusion: grappling with SNARE and SM proteins.
Science
2009
;
323
:
474
7
.
54.
Leabu
M
,
Niculite
CM
. 
Porosome: a membrane microdomain acting as the universal secretory portal in exocytosis.
Discoveries
2014
;
2
:
e29
.
55.
Jena
BP
,
Cho
SJ
,
Jeremic
A
,
Stromer
MH
,
Abu-Hamadah
R
. 
Structure and composition of the fusion pore.
Biophys J
2003
;
84
:
1337
43
.
56.
Buccione
R
,
Caldieri
G
,
Ayala
I
. 
Invadopodia: specialized tumor cell structures for degradation of the extracellular matrix.
Cancer Metastasis Rev
2009
;
28
:
137
49
.
57.
Yamaguchi
H
. 
Pathological roles of invadopodia in cancer invasion and metastasis.
Eur J Cell Biol
2012
;
91
:
902
7
.
58.
Gimona
M
,
Buccione
R
,
Courtneidge
SA
,
Linder
S
. 
Assembly and biological role of podosomes and invadopodia.
Curr Opin Cell Biol
2008
;
20
:
235
41
.
59.
Revach
O-Y
,
Geiger
B
. 
The interplay between the proteolytic, invasive, and adhesive domains of invadopodia and the roles in cancer invasion.
Cell Adhes Migration
2014
;
8
:
1
11
.
60.
Beaty
BT
,
Condeelis
J
. 
Digging a little deeper: the stages of invadopodia formation and maturation.
Eur J Cell Biol
2014
;
1172
:
115
23
.
61.
Friedl
P
,
Wolf
K
. 
Proteolytic interstitial cell migration: a five step process.
Cancer Metastasis Rev
2009
;
28
:
129
35
.
62.
Yamaguchi
H
,
Oikawa
T
. 
Membrane lipids in invadopodia and podosomes: key structures for cancer invasion and metastasis.
Oncotarget
2010
;
1
:
320
8
.
63.
Marchetti
D
,
McQuillan
D
,
Spohn
WC
,
Carson
DD
,
Nicolson
GL
. 
Neurotrophin stimulation of human melamona cell invasion: selected NT enhancement of heparanase activity and heparanase degradation of specific heparan sulfate subpopulations.
Cancer Res
1996
;
56
:
2856
63
.
64.
Hotary
K
,
Li
XY
,
Allen
E
,
Stevens
SL
,
Weiss
SJ
. 
A cancer cell metalloprotease triad regulates the basement membrane transmigration program.
Genes Dev
2006
;
20
:
2673
86
.
65.
Friedl
P
,
Wolf
K
. 
Tube travel: the role of proteases in individual and collective cancer cell invasion.
Cancer Res
2008
;
68
:
7247
9
.
66.
Beaudoin
AR
,
Grondin
G
. 
Shedding of vesicular material from the cell surface of eukaryotic cells: different cellular phenomen.
Biochim Biophys Acta
1991
;
1071
:
203
19
.
67.
Turturici
G
,
Tinnirello
R
,
Sconzo
G
,
Gerael
F
. 
Extracellular membrane vesicles as a mechanism of cell-to-cell communication: advantages and disadvantages.
Am J Physiol Cell Physiol
2014
;
306
:
C621
C633
.
68.
Cocucci
E
,
Racchetti
G
,
Meldolesi
J
. 
Shedding mircrovesicles: artifacts no more.
Trends Cell Biol
2009
;
19
:
43
51
.
69.
Ratajczak
J
,
Wysocznski
M
,
Hayek
F
,
Janowska-Wieczorek
A
,
Ratajczak
MZ
. 
Membrane-derived microvessels: important and under-appreciated mediators of cell-to-cell communication.
Leukemia
2006
;
20
:
847
56
.
70.
Martins
VR
,
Dias
MS
,
Hainaut
P
. 
Tumor-cell–derived microvesicles as carriers of molecular information in cancer.
Curr Opin Oncol
2013
;
25
:
66
79
.
71.
Maheshwari
S
,
Singh
AK
,
Arya
RK
,
Pandey
D
,
Singh
A
,
Datta
D
. 
Exosomes: emerging players of intercellular communication in tumor microenvironment.
Discoveries
2014
;
2
:
e26
.
72.
Hoshino
D
,
Kirkbride
KC
,
Costello
K
,
Clark
ES
,
Sinha
S
,
Grega-Larson
N
, et al
Exosome secretion is enhanced by invadopodia and drives invasive behavior.
Cell Rep
2013
;
12
:
1159
68
.
73.
D'Souza-Schorey
C
,
Clancy
JW
. 
Tumor-derived microvesicles: shedding light on novel microenvironment modulators and prospective cancer biomarkers.
Genes Dev
2012
;
26
:
1289
99
.
74.
Li
J
,
Zhang
Y
,
Liu
Y
,
Dai
X
,
Li
W
,
Cai
X
, et al
Microvesicle-mediated transfer of microRNA-150 from monocytes to endothelial cells promotes angiogenesis.
J Biol Chem
2013
;
288
:
23586
96
.
75.
Sidhu
SS
,
Meingistab
AT
,
Tauscher
AN
,
LaVail
J
,
Basbaum
C
. 
The microvesicle as a vehicle for EMMPRIN in tumor–stromal interactions.
Oncogene
2004
;
23
:
956
63
.
76.
Bobrie
A
,
Krumeich
S
,
Reyal
F
,
Recchi
C
,
Molta
LF
,
Seabra
MC
, et al
Rab27a supports exosome-dependent and -independent mechanisms that modify the tumor microenvironment and can promote tumor progression.
Cancer Res
2012
;
72
:
4920
30
.
77.
Kahlert
C
,
Kalluri
R
. 
Exosomes in tumor microenvironment influence cancer progression and metastasis.
J Mol Med
2013
;
91
:
431
7
.
78.
Wysoczynski
M
,
Ratajczak
MZ
. 
Lung cancer secreted microvesicles: underappreciated modulators of microenvironment in expanding tumors.
Int J Cancer
2009
;
125
:
1595
603
.
79.
Alvarez-Erviti
L
,
Seow
Y
,
Yin
H
,
Betts
C
,
Lakhal
S
,
Wood
MJ
. 
Delivery of siRNA to the mouse brain by systemic infection of targeted exosomes.
Nat Biotechnol
2011
;
29
:
341
5
.
80.
Thiery
JP
,
Acloque
H
,
Huang
RYJ
,
Nieto
MA
. 
Epithelial–mesenchymal transitions in development and disease.
Cell
2009
;
139
:
871
90
.
81.
Wells
A
,
Chao
VL
,
Grahovac
J
,
Wu
G
,
Lauffenburger
DA
. 
Cell motility in carcinoma metastasis as modulated by switching between epithelial and mesenchymal phenotypes.
Front Biosci
2014
;
16
:
815
37
.
82.
Elsum
IA
,
Martin
C
,
Humbert
PO
. 
Scribble regulates an EMT polarity pathway through modulation of MAPK-ERK signaling to mediate junctional formation.
J Cell Sci
2013
;
126
:
3990
9
.
83.
Sheel
C
,
Eaton
EN
,
Li
SH
,
Chaffer
CL
,
Reinhardt
F
,
Kah
KJ
. 
Paracrine and autocrine signals induce and maintain mesenchymal and stem cell states in the breast.
Cell
2011
;
145
:
926
40
.
84.
Bastid
J
. 
EMT in carcinoma progression and dissemination: facts, unanswered questions, and clinical observations.
Cancer Metastasis Rev
2012
;
31
:
277
83
.
85.
Tarin
D
. 
The fallacy of epithelial mesenchymal transition in neoplasia.
Cancer Res
2005
;
65
:
5996
6001
.
86.
Castle
J
,
Shaker
H
,
Morris
K
,
Tugwood
JD
,
Kirwan
CC
. 
The significance of circulating tumour cells in breast cancer: a review.
Breast
2014
;
23
:
552
60
.
87.
Turner
N
,
Pestrin
M
,
Galardi
F
,
De Luca
F
,
Malorni
L
,
Di Leo
A
. 
Can biomarker assessment on circulating tumor cells help direct therapy in metastatic breast cancer?
Cancers
2014
;
6
:
684
707
.
88.
Zhang
L
,
Ridgway
LD
,
Wetzel
MD
,
Ngo
J
,
Yin
W
,
Kumar
D
, et al
The identification and characterization of breast cancer CTCs competent for brain metastasis.
Sci Transl Med
2013
;
5
:
180ra48
.
89.
Valastyan
S
,
Weinberg
RA
. 
Tumor metastasis: molecular insights and evolving paradigms.
Cell
2011
;
147
:
275
92
.
90.
Nicolson
GL
. 
Lipid replacement therapy: a nutraceutical approach for reducing cancer-associated fatigue and the adverse effects of cancer therapy while restoring mitochondrial function.
Cancer Metastasis Rev
2010
;
29
:
543
52
.