Physiological Reviews

Role of Ion Channels and Transporters in Cell Migration

Albrecht Schwab, Anke Fabian, Peter J. Hanley, Christian Stock


Cell motility is central to tissue homeostasis in health and disease, and there is hardly any cell in the body that is not motile at a given point in its life cycle. Important physiological processes intimately related to the ability of the respective cells to migrate include embryogenesis, immune defense, angiogenesis, and wound healing. On the other side, migration is associated with life-threatening pathologies such as tumor metastases and atherosclerosis. Research from the last ∼15 years revealed that ion channels and transporters are indispensable components of the cellular migration apparatus. After presenting general principles by which transport proteins affect cell migration, we will discuss systematically the role of channels and transporters involved in cell migration.


Cell motility is central to tissue homeostasis in health and disease. Thus there is hardly any cell in the body that is not motile at a given point in its life cycle. After gastrulation cells have to reach the sites within the embryo that give rise to the different organs (11). Migration in the context of organ development and differentiation continues after birth. Neuroblasts still need to move postnatally to their final “place of work” within the brain (292, 293, 601). Migration of enterocytes along the crypt-villus axis occurs throughout the entire lifespan (651). Similarly, immune cells permanently patrol through our body and defend it against invading pathogens (162, 414, 541). Already more than 150 years ago, Julius Cohnheim described the concept of leukocyte diapedesis in great detail in his classical paper on inflammation and suppuration [“Über Entzündung und Eiterung” (98); commented on in Ref. 258; see also FIGURE 1, which is taken from a paper of one of Cohnheim's contemporaries (71)]. Wound healing of the skin requires the movement of keratinocytes and fibroblasts (362). Wounding also frequently affects the mucosa of the gastrointestinal tract, and migration of epithelial cells into the denuded area is a fast way to reestablish epithelial integrity (38, 127). Angiogenesis which is also triggered during wound healing is yet another example for a process depending on migration (560).

Figure 1.

Historical hand drawing of the recruitment of neutrophil granulocytes through the vascular wall. Although the figure is older than 140 years, the top small panels 1–6 clearly present the stages of rolling, diapedesis, and migration in the interstitial space. [From Caton (71).]

The downside of migration is represented by disease conditions in which cell migration constitutes a major pathophysiological element. The most prominent example is the formation of tumor metastases. Tumor cell migration is an essential step of the so-called metastatic cascade that leads to the spread of the disease within the body (341, 639). Rheumatoid arthritis and atherosclerosis are chronic inflammatory diseases involving the migration of leukocytes as well as that of synovial fibroblasts (319, 459) and smooth muscle cells, respectively (429, 576). These examples clearly illlustrate the diversity of the physiological and pathophysiological functions of migrating cells and highlight its medical importance.

The systematic study of cell migration goes back more than 150 years. First mechanistic models of amoeboid movement were developed in the second half of the 19th century (reviewed in Refs. 366, 367). The old models focussed on what was later to be defined as the cytoskeleton, which is one of the major intracellular motors of cell migration (103, 315, 441, 464). It goes of course without saying that there has been an enormous technical progress that can be illustrated by comparing the image acquisition of migration experiments. While Mast and co-workers projected the cell under observation with a camera lucida and made sketches of the cells' posterior ends in minute intervals (233), modern technologies allow frame rates of several thousand Herz, subpixel lateral resolution well below the refractive limit, or intravital observation of migrating cells (72, 486).

However, one component of the cellular migration machinery, namely, transport proteins in the plasma membrane, was largely neglected until quite recently. Studies from the last ∼15 years have unequivocally shown that the plasma membrane (277) and the transport proteins inserted therein also play important roles in cell migration. This will be the main focus of our review. We will provide an overview of mechanisms by which transport proteins contribute to the process of cell migration, we will systematically discuss relevant transport proteins, and finally, we will integrate their function in selected disease states.


A. Common Properties of Migrating Cells

One of the common features of all migrating cells is their polarization along their axis of movement, which can be seen in the presence or absence of directional stimuli (181, 402, 480). The polarization applies to morphology and function. This is particularly evident when cells are migrating on a two-dimensional substrate. Such conditions are typical for epithelial cells during wound healing (38) or for neutrophils adhering to the endothelium prior to transmigration (414, 449). The majority of cells, however, migrate in a three-dimensional environment. The front part of migrating cells is formed by a fanlike, 300-nm-thin, and organelle-free process, the lamellipodium (2, 3). Cell body and uropod form the rear part of the cell. Persistent migration is only possible when this polarization is maintained.

The polarization of migrating cells applies to many different aspects of the cellular migration apparatus. First of all, it concerns the opposing movements of the front and rear cell poles. Migration can be described as a repeated and highly coordinated cycle of protrusion of the lamellipodium and retraction of the rear part of the cell. Protrusion of the anterior and retraction of posterior parts of migrating cells requires that the respective components of the cellular migration machinery, e.g., polymerizing actin filaments at the cell front and contractile proteins at the rear end, are differentially localized and regulated within a migrating cell. We refer to recent reviews (103, 248, 315, 390, 464, 479) for a detailed overview of cytoskeletal mechanisms of migration. Force generated by the migration machinery must be transmitted onto the surrounding matrix so that the cell can eventually advance. The turnover of the respective cell adhesion receptors requires a constant supply of “new” receptors being delivered to the cell front resulting in an asymmetric distribution of integrins in migrating cells (52, 264, 314, 630). Finally, the polarization of migrating cells also comprises the subcellular distribution of other membrane proteins such as chemokine receptors (354) and transport proteins (see sects. III–V) as well as the composition of the plasma membrane itself (602).

B. Cell-Matrix Interactions During Cell Migration

Migration requires a coordinated formation and release of focal adhesion contacts to transmit force developed by the cellular migration motor to the underlying extracellular matrix (50, 54). In most cells integrins are central components of focal adhesion contacts that are comprised of multiprotein complexes also containing the Na+/H+ exchanger NHE1 (243, 453). Integrins are composed of different α and β subunits and serve as receptors for proteins of the extracellular matrix. The subunit composition of integrins determines their specificity for proteins of the extracellular matrix. Integrins are also important “hubs” for transmitting signals in the outside-in and inside-out direction, since they interact with a plethora of proteins on both sides of the plasma membrane. The abundance and complexity of protein-protein interactions at these contact sites is reflected by the term “adhesome” (655).

Importantly, cell adhesion turnover is also regulated by ionic mechanisms. The contribution of Ca2+ and protons has been studied in particular detail. The disassembly of focal adhesion components at the rear part of migrating cells involves the Ca2+-sensitive family of calpain proteases (68, 74). The turnover of LFA-1 adhesions in migrating T lymphocytes depends on calpain 2 activity. Inhibition or knockdown of calpain 2 impairs the disassembly of LFA-1 adhesions so that T lymphocytes cannot detach their rear part or become elongated, and LFA-1 is shed (567). The gradient of the intracellular Ca2+ concentration ([Ca2+]i) that has been found in many migrating cells including T lymphocytes contributes to the localized diassembly of focal adhesions, since [Ca2+]i is usually higher at their rear end than at their front (56, 95, 149, 202, 519, 625). Accordingly, calpain activity is highest in the rear part of T lymphocytes (567).

The pH dependence of cell adhesion was revealed by applying force measurements with atomic force microscopy, cell behavioral assays, and molecular dynamics simulations. RGD peptides that constitute a typical recognition motif for integrin binding of extracellular matrix proteins such as fibronectin bind most firmly to ανβ3 or α5β1 integrins in the plasma membrane of osteoclasts at pH 6.5 (321). Binding of collagen I to α2β1 integrins is pH dependent too. This was deduced from biochemical assays with isolated proteins (141) and from adhesion and migration experiments with human melanoma (MV3) cells in which the effects of variations of the extracellular pH (pHe) were explored. Adhesion and migration of α2β1 integrin-expressing melanoma cells have their optimum at pHe 6.8 and pHe 7.0, respectively (299, 554, 558, 561). Such a behavior is also found in ανβ3 integrin expressing CHO cells (436). Computational molecular dynamics simulations revealed a possible molecular mechanism for these effects. The conformational equilibrium of ανβ3 integrins and thereby their state of activation depends on pHe. An acidic extracellular environment (pHe <7.0) directly activates ανβ3 integrins and increases their ligand avidity by inducing a shift of their conformation towards the opening of the integrin headpiece (436).

Finally, there is a strong indirect or physical crosstalk between integrins and several K+ channels (e.g., KCa1.1, KV1.3, and KV11.1). KCa1.1 channels in endothelial and arteriolar smooth muscle cells are activated following αVβ3 or α5β1 integrin binding to vitronectin or fibronectin in a manner that involves a c-src-dependent tyrosine phosphorylation of the channel protein (274, 646). In lymphocytes, tumor or neuronal cells β1 integrin and KV1.3 or KV11.1 channel impact on each other so that adhesion is regulated by K+ channel activity (15, 20, 85, 325). Along the same lines, KV2.1 channels contribute to the activation of FAK in mesenchymal stem and corneal epithelial cells (239, 626) while TRPM8 channels inactivate FAK in PC-3 prostate carcinoma cells (647). In osteoblasts, the association of KCa1.1 with FAK could be promoted by exposing cells to a hyposmotic environment (477).

C. Chemotaxis

The directed migration towards (or away from) the source of an extracellular signal molecule (chemoattractant) is termed chemotaxis. The cells' ability to detect and to respond to such directional clues is not only a prerequisite for embryogenesis and the normal body homeostasis but also for the development of pathological conditions such as the formation of tumor metastases (504, 568, 639). The general view is that chemotaxis of eukaryotic cells occurs by spatial sensing. Cells detect concentration differences as small as 1% along their length axis, amplify them intracellularly, and thereby elicit the cellular responses underlying directed migration (504). The concentration differences of the chemoattractant will lead to a higher receptor occupancy at the part of the cell facing the source of the chemoattractant (245) resulting in asymmetrical stimulation of G protein-coupled chemoattractant receptors. This corresponds to the local excitation/global inhibition model of chemotaxis that produces a local membrane protrusion towards the chemoattractant source and retraction of the rear part of the cell (568).

A major breakthrough in the field was the discovery that activation of phosphoinositide 3-kinase (PI3K) via G proteins causes the accumulation of PI(3,4,5)P3 at the cell pole facing the chemoattractant source (478) which is accentuated by the PI-3-phosphatase PTEN (phosphatase and tensin homolog deleted on chromosome 10) at the rear part of chemotacting cells (166, 246). However, chemotactic gradient sensing does not solely rely on PI(3,4,5)P3 signaling, since chemotaxis is possible even in the absence of polarized PI(3,4,5)P3 and type 1 PI3Ks (228). A variety of redundant chemotactic signaling pathways come into play when a cell has to prioritize between diverging chemotaxis gradients and needs to decide which path to take. This is reflected in the concept that neutrophils favor bacterial chemoattractants or complement components (“end-target chemoattractants”) over chemokines (“intermediary chemoattractants”) with p38 MAPK and PI3K being the respective intracellular signaling modules (154, 223, 224, 650).

An alternative model of chemotaxis suggests that it can occur without spatial gradient sensing. Instead, cells generate membrane protrusions continually, and chemoattractants only cause biased extension of existing extensions towards the higher concentrations. Thus chemotactic gradients rather modulate dynamics of already existing lamellipodia than causing the appearance of new ones (13).

Heterotrimeric G proteins usually belonging to the Gαi/o family and monomeric G proteins (e.g., Rho, Rac, Cdc42) are of central importance for chemotactic signal transduction. Pertussis toxin-sensitive members of the Gαi/o family stimulate PI3K via the Gβγ subunit. The resulting accumulation of PI(3,4,5)P3 accumulation and Rac activation contributes to the definition of the (new) front of the cell. Activation of Gα12/13 may provide “backness” through links to RhoGEFs (Rho guanine nucleotide exchange factors), which “turn on” Rho, required for retraction of the trailing end (629, 634). We could show that myosin IXb may act as RhoGAPs (GTPase-activating proteins) and selectively inhibit Rho at the leading edge of macrophages (210).

Activation of PLC-β isoforms and internal Ca2+ release are other common features of chemotactic signaling, mediated either through Gαq/11 or Gβγ subunits from activated Gαi/o (261). While it is established that chemoattractant stimulation frequently induces elevations of the intracellular Ca2+ concentration (112, 255), the molecular identity of the involved ion channels and their mechanisms of action are far less understood. Thus it is not yet known whether the crucial role of autocrine purinergic signaling mediated by chemoattractant-induced ATP release for neutrophil (80) and macrophage (301) chemotaxis is due to shaping intracellular Ca2+ signals. Some recent studies indicate that ion channels controlling the intracellular Ca2+ homeostasis are of critical importance for efficient chemotaxis (23, 39, 148, 530, 538) or for the related process of nerve growth cone steering (535, 536, 613, 628).

The Na+/H+ exchanger NHE1 is another transport protein that is crucial for the polarization of migrating cells (118, 126, 524, 556). In ameobae (Dictyostelium discoideum), a developmentally regulated NHE (DdNHE1) is needed for cell polarity and for efficient chemotaxis in response to cAMP. DdNHE1 is expressed only in chemotactically competent and not in vegetative cells, confirming its necessity for directed movement (443).


Many of the general principles by which ion transport proteins regulate cell migration are related to the housekeeping functions exerted by ion transport proteins such as setting the cell membrane potential as well as regulating cell volume, [Ca2+]i, and intra- and extracellular pH. Since the cell membrane potential and cell volume regulation involve many functionally diverse families of ion channels and transporters, their role in cell migration will be discussed in the following sections to provide a more unified view. The impact of [Ca2+]i and of pH on the cellular migration machinery will be discussed in detail as an introduction to the sections on Ca2+ and pH regulatory transport proteins, respectively (see sects. IVE and VA).

There is accumulating evidence that the role of some ion transport proteins in cell migration can be attributed to their nonconductive properties that do not depend on the transport of ions. They are either mediated by the ion-transporting protein itself or by associated ancillary subunits (e.g., β-subunits of voltage-gated Na+ channels or of the Na+-K+-ATPase). Transport proteins have a large repertoire of nonconductive properties including but not limited to the combination of channel and enzymatic activity in one protein (“chanzyme” TRPM2, -6, -7; Ref. 502), conformational coupling to other proteins (450) or the function as a signaling platform (330). In other cases, the nonconductive mechanisms have not yet been elucidated (627). Finally, ancillary units of transport proteins can even operate as adhesion receptors (14, 91, 444). We will refer to these effects in more detail when discussing the individual transport proteins.

A. Cell Membrane Potential

The membrane potential of migrating cells serves the same housekeeping and signaling functions like in any other eukaryotic cell. It indirectly affects cell migration by regulating cell volume as well as intra- and extracellular pH (see sects. IIIB and VA). In addition, the cell membrane potential plays a particularly important role in setting the electrical driving force for Ca2+ influx and controlling the gating behavior of voltage-dependent Ca2+ channels. Thus, in a cell expressing voltage-independent Ca2+ channels (e.g., TRP channels), [Ca2+]i rises in response to a hyperpolarization, while in excitable cells expressing voltage-gated Ca2+ channels a rise of [Ca2+]i can be induced by a depolarization. Functional coupling with Ca2+-sensitive K+ channels may then lead to a positive-feedback cycle promoting sustained Ca2+ influx (171) or to a negative feedback terminating Ca2+ influx (150), respectively. Along these lines, the depolarizing effect of TRPM4 channel-mediated cation influx has been linked to attenuating Ca2+ influx in nonexcitable cells (310). The functional importance of the cell membrane potential for cell migration is elegantly illustrated in neutrophils. In these cells, proton channels (VSOP/Hv1) prevent the depolarization of the cell membrane potential during respiratory burst and thereby sustain Ca2+ influx required for cell migration. Consequently, neutrophils from VSOP/Hv1−/− mice have a strong defect in chemokinetic migration (144).

In addition, a depolarization of the cell membrane potential exerts a modulatory effect on the cytoskeleton (88, 571, 612). Thus, in renal epithelial cells, a depolarization activates a signaling cascade consisting of ERK → GTP/GDP exchange factor GEF-H1 → Rho → Rho-kinase leading to myosin light chain phosphorylation (571, 612). In endothelial cells, the membrane potential is sensed by the cortical actin network and determines the actin polymerization/depolymerization ratio and thus cell stiffness. A depolarization of the cell membrane potential softens endothelial cells by depolymerizing the cortical actin cytoskeleton (62). Since the plasma membrane is only 5 nm thin, the seemingly small membrane potential of approximately −50 mV builds up an electric field of 107 V/m. Changes of this electrical field can, for example, control the conformation and thereby the activity of membrane proteins such as voltage-gated ion channels. The electric field across the plasma membrane also promotes the interaction between KV11.1 channels with β1-integrins and the recruitment of further signaling molecules such as FAK and Rac1 when the cell membrane potential hyperpolarizes upon KV11.1 activation (424, 450).

Moreover, the activity of voltage-dependent enzymes such as the phosphoinositide phosphatase Ci-VSP, a member of the PTEN family of phosphatidylinositol phosphatases, is controlled by the membrane potential (254, 400). Depolarization of the plasma membrane activates Ci-VSP and leads to cleavage of phosphatidylinositol-4,5-bisphosphate [PI(4,5)P2] and thereby regulates the activity of PI(4,5)P2-dependent ion channels. So far, it is not known whether Ci-VSP plays a role in cell migration by regulating the PI(4,5)P2 supply for the actin modulator cofilin (326) or for ion channels involved in cell migration. If this were the case, one would expect a temporal correlation between protrusive lamellipodial activity and membrane potential oscillations frequently observed in migrating cells (209, 526, 527). In such a scenario, a hyperpolarization of the cell membrane potential would shut down Ci-VSP, thereby increasing PI(4,5)P2 availability, which in turn would inhibit cofilin and protrusive activity at the leading edge. In this context, it is noteworthy that in Ciona intestinalis Ci-VSP is expressed in NADPH oxidase (gp91)-positive blood cells which include the highly motile phagocytes (419).

With the use of nanoparticles encapsulating voltage-sensitive fluorescent dyes, it was recently found that electric fields extend far beyond the plasma membrane into the cytosol (596). This raises the question whether there is a functional crosstalk between the electric field generated by the cell membrane and the electric fields originating from intracellular organelles (234) or macromolecular complexes (461, 595). At this stage, the relative contribution and tissue specificity of indirect housekeeping effects, of membrane-delimited processes such as Ci-VSP mediated PI(4,5)P2 cleavage, or of long-range electric field effects in controling cell migration is still an open issue.

B. Cell Volume

Cell migration can be described as a repetitive cycle of protrusion of the cell front that is followed by retraction of the rear part. In terms of cell volume, this cycle can be modeled as volume gain at the cell front and volume loss at the rear part of migrating cells (see FIGURE 2). Such a view is also supported by the frequent observation that the protrusion of the lamellipodium and the retraction of the rear part do not always occur simultaneously in migrating cells. Rather, one of these processes temporarily dominates so that the cell volume rises during the protrusion of the lamellipodium, whereas it decreases during the retraction of the rear part. Such (local) volume changes of up to 35% have been visualized and quantified in migrating cells using atomic force (515), scanning ion conductance (211), or multiphoton laser scanning fluorescence microscopy (620). Neutrophil granulocytes increase their volume upon chemotactic but not upon chemokinetic stimulation. When this gain of volume is prevented by exposing neutrophils to a hypertonic solution, chemotaxis is inhibited. Accordingly, neutrophils emigrated into inflamed lung tissue or into the abdominal wall have a larger cell volume than intravascular ones (490, 631).

Figure 2.

Cell volume changes during cell migration. Cell migration is a continuous cycle of protrusion of the cell front and retraction of the trailing end. This can be modeled as a cycle of isosmotic volume increase at the cell front and isomotic volume decrease at the rear end. This model is based on direct measurements of volume changes in migrating cells (211, 515, 620) and on the subcellular distribution of the relevant ion transport proteins and aquaporins. The molecular nature of the mechanosensitive Ca2+ channels implicated in this model is still elusive. The scheme illustrates how members of the “transportome” cooperate during cell migration.

That volume changes remain restricted to one cell pole can be attributed at least partially to the poroelastic structure of the actin meshwork and cytoplasm within the lamellipodium. It is able to resist osmotic water flow, thereby preventing or retarding the immediate equalization of local volume changes (77, 251, 252, 386). Moreover, the stiffness of the lamellipodium gradually increases towards the leading edge so that the cell body is softer and more easily deformable than the lamellipodium (311, 463). In addition, the lamellipodium is attached more firmly to the substratum than the cell body (153, 515). These properties prevent the lamellipodium from decreasing its volume to the same extent as the cell body upon concomitant K+ and Cl channel activation and restrict the volume loss to the rear part of migrating cells. Local volume changes can thereby “mechanically” support protrusion and retraction of front and rear parts of migrating cells, respectively. When cells are migrating through a tortuous three-dimensional extracellular matrix, volume changes may also be required to overcome narrows (200, 470, 549, 581, 620).

A specialized form of localized volume change in migrating cells is represented by the phenomenon of “blebbing.” It is commonly explained on the basis of a contraction of the actomyosin network that causes an increase of the intracellular hydrostatic pressure. Local weakening of the cortical cytoskeleton allows the cytosol to be squeezed into the rapidly forming blebs of the plasma membrane. That is, blebbing is explained as a consequence of intracellular volume shifts. However, for the following reasons, we propose that transport of solutes and water across the plasma membrane contributes to the expansion of blebs as well. The poroelastic nature of the cytoskeleton impairs the immediate equalization of local volume changes (77, 386), and ion transport proteins including aquaporins are in the “correct” subcellular position to allow rapid solute and water uptake from the extracellular space (118, 119, 197, 200, 287, 492, 497, 556). Finally, aquaporins are found in the plasma membrane of blebs (242), and increasing the extracellular osmolarity inhibits bleb formation (653).

Cell volume regulation critically depends on the activity of ion channels and transporters (see Ref. 229 for a current comprehensive review). Practically all cells return to their normal volume following an osmotic perturbation by either activating K+ and Cl channels and releasing KCl and cell water (regulatory volume decrease, RVD) or by activation of Na+-K+-2Cl cotransport, Na+/H+ exchange, and nonselective cation channels leading to the net uptake of KCl and water (regulatory volume increase, RVI). Employing the terminology of volume regulation, cell migration can then be described as a cycle of isosmotic RVI at the cell front and RVD at the rear part. Accordingly, many of the ion channels and transporters participating in cell volume regulation are part of the cellular migration machinery. Examples include KCa3.1 (105, 508, 527, 530), VRAC (357, 513), KV1.3 (368), various TRP channels (617, 625), NHE1 (118, 217, 287, 481, 554), NKCC1 (200, 475, 527), or aquaporins (435) to name just a few. A quantitative evaluation of NHE1-mediated Na+ fluxes in MDCK-F cells revealed that it is of a magnitude sufficient to contribute to lamellipodial protrusion (287).

The above-mentioned volume model of cell migration requires the concerted, interdependent activity of a whole set of channels and transporters that belong to many different families of transport proteins. To account for this complexity, we would like to introduce the concept of a migration-asociated “transportome”. This concept implies that it is not the isloated function of a single ion channel or transporter but rather the activity of a whole network of transport proteins that is required for efficient migration. It appears that some transportome members such as aquaporins, KCa3.1, or NHE1 are found in most migrating cells, while others like proton channels (144) are expressed cell-type specifically so that the transportome meets the special needs of a given cell (see sects. IV and V for a detailed discussion of these transport proteins).

The link between cell volume regulation and cell migration becomes more evident when considering the rapid and pronounced polymerization and depolymerization of the actin cytoskeleton following osmotic shrinkage and osmotic swelling, respectively (204, 265, 447, 482, 525). Volume-dependent modulation of the actin cytoskeleton also concerns the activity and subcellular localization of proteins associated with actin filaments such as the myosin light chain (121), myosin II (447), cortactin (123), and the ERM protein ezrin (474). Hence, by setting the “correct” cell volume, ion channels and transporters have an important impact on the organization of the actin cytoskeleton. Intermediaries between cell volume perturbations and the actin cytoskeleton include Rho, Rac, Cdc42, and PI(4,5)P2, which are highly sensitive to cell volume changes and which constitute important signaling modules in the control of actin dynamics (122, 123, 286, 335, 407, 425, 474, 652). Importantly, PI(4,5)P2 also regulates the activity of many ion channels and transporters involved in cell migration (340) such as KV1.3 (49, 325), TRPC1 and -6 (39, 78, 109, 149, 205, 346), TRPV1 (617), TRPM4 (23, 408), TRPM7 (495), NHE1 (8, 558), and NCX1 (134, 648). Moreover, many studies have shown that ion channels and transporters involved in volume regulation and cell migration are regulated by the actin cytoskeleton and associated proteins (198, 230, 265, 286, 323, 370, 371, 373, 411). Such findings lend strong support to the concept of a mutual interdependence of ion channel and transporter function, cell volume regulation, and cytoskeletal dynamics during cell migration.


In the following section we will systematically summarize the current state-of-the-art concerning the contribution of molecularly defined ion channels to the cellular migration machinery. We will limit our discussion to the most relevant channels. This discussion is complemented by TABLE 1 listing the channels found to be involved in cell migration. When indicated we will also refer to studies investigating the role of transport proteins in the movement of nerve growth cones, which is a process closely related to cell migration.

View this table:
Table 1.

Ion and water channels involved in cell migration

A. K+ Channels

K+ channels form the largest family of ion channels with more than 70 members in mammals ( They were among the first transport proteins that were recognized to play a crucial role in cell migration (17, 18, 527). Their role in cell migration was the topic of a recent review (521) that will be updated in this section.

1. Voltage-gated K+ channels

KV1.3, KV10.1, and KV11.1 channels are the best studied voltage-gated K+ channels in the context of cell migration. They are of particular importance for cells of the immune system and for tumor cells where they constitute potential therapeutic, diagnostic, and/or prognostic targets (see also sect. VI, A and B).

A) KV1.

KV1.1 channels are reported to play a role in wound healing of intestinal and gastric epithelial cells (472, 540, 615). The effect of KV1.1 channels on migration of these epithelial cells is ascribed to setting the electrochemical driving force for Ca2+ influx and subsequently triggering Ca2+-dependent RhoA signaling. KV1.1 expression itself is regulated by intracellular polyamine synthesis.

However, most studies on the role of KV1 channels in cell migration are focused on KV1.3 channels (20, 175, 282, 325, 368, 416, 610). Blockade or reduced expression of KV1.3 channels is invariably followed by an inhibition of migration. This is of particular physiological relevance for the immune system since KV1.3 channels play a crucial role in orchestrating the immune response (59). Migration of effector memory T lymphocytes (368) or macrophages (175, 610) depends on KV1.3 function. Migration of alveolar macrophages is also inhibited by KV1.5 knockdown (440). The almost complete inhibition of migration of effector memory T lymphocytes in a model of delayed-type hypersensitivity by a specific KV1.3 blocker was explained on the basis of suppressed Ca2+ signaling leading to insufficient integrin activation (368). Possibly, this “housekeeping” effect of KV1.3, controlling [Ca2+]i (see sect. IIIA), is reinforced by its physical interaction and conformational coupling with β1 integrins, i.e., by a nonconductive property of KV1.3 channels (20, 325). Voltage-dependent gating of KV1.3 activates β1-integrins. Consequently, KV1.3 channel blockers prevent integrin activation.

B) KV10.1 (EAG1) AND KV11.1 (ERG1).

These voltage-gated K+ channels will be discussed together since they share a medically important property. They are regulators of tumor cell proliferation and migration, and upregulated expression of KV10.1 and KV11.1 channels in many tumor cells is negatively correlated with patient prognosis (5, 6, 182, 308, 309, 432, 451). KV11.1 channels have similar nonconductive properties as KV1.3 channel in that they also form complexes with β1-integrins in which both proteins reciprocally activate each other, presumably by conformational coupling (reviewed in Ref. 450). KV11.1 channels are activated upon integrin-mediated adhesion to fibronectin or laminin (17, 19), and integrin signaling depends on KV11.1 channel activity (85). Functionally, KV11.1 channels strongly affect neurite outgrowth in neuroblastoma cells (17, 19) and cell migration by controlling cell-matrix interactions and downstream signaling cascades. KV11.1 channels form a macromolecular signaling complex with the vascular endothelial growth factor receptor-1, FLT-1, and β1-integrins in acute myeloid leukemia (AML) cells. β1-Integrins mostly recruit the KV11.1B isoform. This interaction is required for FLT-1 signaling activation and AML cell migration (in vitro and in vivo) both of which are inhibited when KV11.1 channels are blocked. Importantly, the FLT-1/KV11.1/integrin complex is found in primary AML blasts from patients, too, and KV11.1 expression in leukemia patients correlates with a higher probability of relapse and shorter survival periods (451).

KV10.1 and KV11.1 channels also regulate the migration of nontumorous cells. Thus KV10.1 contributes to the migration of rat cerebellar granule cells (338). In the zebrafish model, KV11.1 antisense RNA morpholinos and pharmacological inhibitors attenuate neutrophil recruitment in response to an infectious stimulus (55).

2. Ca2+-sensitive K+ channels (KCa 3.1, KCa1.1, KCa2.3)

Ca2+-sensitive K+ channels are divided into three groups: KCa1.1 (formerly designated as BK), KCa2.1–2.3 (SK1–3), and KCa3.1 (IK, SK4, or Gardos channel). They are all involved in cell migration, with KCa3.1 channels being studied the most extensively. The important role of KCa3.1 channels in cell migration is underscored by their expression profile. They are expressed in virtually all migrating cells including most immune cells (49, 59, 105, 136, 137, 171, 209, 508, 533, 590), platelets (509), osteoclasts (146), vascular cells (51, 84, 289, 564, 578, 590), fibroblasts (106, 523), and several model cell lines used to study cell migration (70, 263, 279, 527, 528). Moreover, KCa3.1 channels are stage-dependently (over-)expressed in tumors pointing to a possible role in the metastatic cascade (216, 257, 306, 422, 431, 510, 523, 573, 616). In most cases, inhibition or genetic deletion of KCa channels results in reduced migratory activity and impaired chemoaxis (see FIGURE 3). Conversely, increased expression of KCa3.1 channels frequently correlates with increased migratory activity (528, 533, 564). The contribution to migration of different calcium-sensitive K+ channels is cell type specific. KCa1.1 channels, for example, are expressed in glioma and microglial cells. Yet, they only modulate migration of glioma cells but not that of microglia (298, 508, 553, 621).

Figure 3.

Dependence of MDCK-F cell migration on the activity of KCa3.1 channels. Blocking KCa3.1 channels with various K+ channel inhibitors leads to a proportionate decrease of KCa3.1 channel activity and migration. [Adapted from Schwab et al. (527).]

The Ca2+ sensitivity of Ca2+-activated K+ channels couples their activity to other Ca2+-dependent events during the process of cell migration (see sect. IVE). Their activity follows spatial gradients and temporal fluctuations of [Ca2+]i so that they are generally more active at the rear part than at the front of migrating cells (520, 527, 528) and exhibit an oscillating activity (146, 526). Consequently, topical application of the KCa3.1 channel blocker charybdotoxin inhibits migration only when it is directed towards the rear part of migrating MDCK-F cells (520). It is important for migrating cells that KCa channel activity is able to fluctuate because tonic pharmacological KCa channel activation impairs migration (298, 525). Transient elevations of [Ca2+]i preceding rear end retraction (317) trigger bursts of KCa channel activity and K+ efflux (110). The resulting local volume loss at the rear end (515) acts in concert with cytoskeletal Ca2+-dependent mechanisms underlying rear end retraction such as actomyosin contraction or calpain-mediated integrin release. An alternative view is that in glioma cells KCa channel-mediated volume loss predominantly affects invadopodia, thereby facilitating the invasion of the narrow extracellular space of brain tissue (376, 620). KCa channel-mediated volume changes are most likely accomplished by their parallel operation with volume-regulated anion channels (VRAC) (357, 358, 513) or ClC3 channels (376, 620) so that the combined application of KCa3.1 and Cl channel blocker leads to an almost complete inhibition of migration (70). In melanoma cells, additional KCa3.1 channel-dependent mechanisms support rear end retraction. By increasing [Ca2+]i (see below and sect. IIIA) KCa3.1 channels drive the secretion of the so-called melanoma inhibitory activity (MIA) at the rear part of the cells (510) leading to facilitated cell detachment (25).

So far, it is not yet known which function KCa3.1 channels exert at the cell front (528). They could serve a similar role as in a secretory epithelium and recycle K+ taken up by the Na+-K+-2Cl cotransporter NKCC1 which is concentrated at the front of migrating cells (200). Alternatively, maximal activation of KCa3.1 channels in the vicinity of a chemoattractant source could act synergistically with chemokine receptor internalization (489) to bring migration to a stop at a focus of inflammation. This view is supported by the observation that KCa3.1 channel activation at the cell front slows down migrating cells (522, 528).

In nonexcitable cells, KCa3.1 and KCa2 channels are involved in a positive feedback loop controlling the intracellular Ca2+ concentration (see sect. IIIA) (267, 306). This relation to [Ca2+]i was proposed to account for the promigratory effect of KCa2.3 channels in breast cancer, melanoma, and colon epithelial cells. [Ca2+]i and migratory activity are correlated with the level of KCa2.3 channel expression (76, 455, 457). KCa2.3 expresion itself is controlled by the adenomatous polyposis gene (Apc) in colonic epithelial cells (458). In monocytes, the feedback between KCa channels and [Ca2+i] was elucidated a step further by identifying TRPC6 and TRPV1 channels as being of equal importance for chemotaxis as KCa3.1 channels (506, 507). In line with the concept of the transportome, it is therefore conceivable that TRPC6 and TRPV1 channels locally cooperate with KCa3.1 channels by supplying the latter with Ca2+ needed for their activation.

In addition to the above-mentioned mechanisms related to their transport function, KCa channels also regulate migration by directly interacting with important molecular modules of the cellular migration apparatus such as FAK (477), cortactin (584), or integrins (α5β1 or αVβ3) (274, 633). Finally, there is a growing body of evidence that KCa3.1 channels are also part of growth factor/chemokine signaling cascades so that inhibition of the channels prevents the stimulation of migration by these agents (105, 279, 508, 530, 533).

3. Inwardly rectifying K+ channels

A) KIR4.2.

Kir4.2 channels are regulated by intracellular polyamines such as spermine or spermidine that block outward current through the channel. With the use of MEFs, CHO, or glioma cells, it was shown that the association of α9β1 integrins with spermidine/spermine acetyltransferase (SSAT) links migration to the activity of Kir4.2 channels (115, 604). Kir4.2 channels are colocalized with α9β1 integrins at the leading edge of the lamellipodium. This also places the polyamine catabolizing enzyme SSAT in the proximity of the channel and thereby weakens its inward rectification and allows (locally) more K+ efflux. A detailed analysis of the migratory behavior of α9β1 integrin and Kir4.2 expressing cells revealed a pronounced increase of migratory persistence. It was suggested that locally enhanced K+ efflux promotes the formation of a single dominant lamellipodium while at the same time preventing the formation of additional protrusions (115). These findings provide an instructive example for the importance of the local and spatially restricted regulation of ion channels in cell migration.

B. Voltage-Gated Na+ Channels

At first sight it seems counterintuitive that voltage-gated Na+ channels (NaV) can have any function in nonexcitable peripheral migrating cells. However, careful electrophysiological analysis of NaV channel currents in breast cancer cells revealed that there is indeed sustained NaV1.5 activity (“window current”) at membrane voltages between −60 and −20 mV. This is a voltage range typically encountered in migrating cells (176). Moreover, NaV1.7 expression in a subset of dendritic cells resulted in a depolarized cell membrane potential (285).

The first description of NaV channel expression in small-cell lung cancer cells dates back more than 30 years (434). It took a few years, however, to recognize that the expression of NaV channels correlates with metastatic potential, i.e., cellular motility/invasiveness (196, 546). Initially, a transient tetrodotoxin-sensitive Na+ current was found only in strongly metastatic Mat-Ly-Lu rat prostate cancer cells while it was absent in any of the corresponding weakly metastatic AT-2 cells (196). Transfection of a NaV (NaV1.4) into a weakly invasive human prostate cancer cell line significantly increased its invasiveness (31). In the meantime, the importance of NaV channels for the metastatic behavior has been confirmed for many different tumor cells. Inhibition of NaV (using TTX, siRNA, polyclonal antibodies) impairs motility and invasive behavior (see FIGURE 4) (47, 125, 158, 165, 170, 307, 426, 484, 485). In addition, NaV channels also contribute to cell motility and invasion of a number of different immune cells such as lymphocytes (159), macrophages (69), microglia (36), and dendritic cells (285).

Figure 4.

Involvement of NaV channels in motility/invasion of metastatic MDA-MB-231 breast cancer cells. The NaV channel blocker tetrodotoxin (10 μM, TTX) inhibits motility in a transwell assay (A), Matrigel invasion (B), and migration in an electric field of 3 V/cm (C). C depicts cell trajectories that are normalized to a common starting point. Cell migration was monitored i) under control conditions in the absence of an electric field, ii) in the presence of an electric field (3 V/cm), and iii) in the combined presence of an electric field (3 V/cm) and the NaV blocker tetrodotoxin. The ability of the cells to respond to the electric field is largely attenuated when NaV channels are blocked. [Modified from Fraser et al. (158), by permission of the American Association for Cancer Research.]

Importantly, NaV channels expressed in metastatic tumor cells were found to be fetal splice variants, as seen clearly in the case of NaV1.5 in human breast cancer cell (158). This issue is consistent with the epigenetic nature of cancer and oncofetal gene expression. To the best of our knowledge there is only one report showing association of a NaV mutation with glioblastoma (266). The clinical potential of neonatal NaV1.5 expression in human breast cancer was discussed earlier (see Ref. 427 for review).

Several mechanisms contribute to NaV's positive impact on cell migration and invasion (see Ref. 48 for a recent review). NaV1.5 promotes breast cancer cell invasion in a cysteine cathepsin-dependent manner by inducing a pericellular acidification (53, 176). In macrophages and melanoma cells, intracellularly localized NaV1.6 supports the formation of podosomes and invadopodia, respectively, by causing shifts of intracellular Na+ and Ca2+ and thereby affecting signaling pathways that link intracellular NaV activation to actin cytoskeleton dynamics (69). NaV channels also regulate migration in a nonconductive way by their associated β-subunits that act as cell adhesion molecules (46, 91, 113). In cerebellar granule neurons, α- and β1-subunits of NaV channels assemble to macromolecular complexes with contactin, and initiate a signaling cascade through fyn kinase leading to neurite outgrowth and migration (48). In oligodendrocyte precursor (NG2) cells NaV channels contribute to GABA-promoted migration by inducing a sustained elevation of [Na+]i which in turn is responsible for causing a Na+/Ca2+ exchanger (NCX)-mediated increase of [Ca2+]i (589). Finally, functional expression of NaVs is highly dynamic and under the control of a range of regulatory mechanisms, including growth factors (129, 599). On the other hand, NaV1.5 activity itself is a high level regulator controlling the transcription of gene networks involved in colon cancer invasion such as those of cell migration (236).


In contrast to the widely accepted view that NaV channels importantly contribute to tumor cell migration/invasion data on the involvement of the epithelial Na+ channel ENaC and on acid-sensing ion chanels (ACICs) in cell migration are still sparse. ENaC inhibition or knockdown impairs the migration of vascular smooth muscle (192), trophoblastic (BeWo) (116), corneal endothelial (88), glioma cells (268), and neurite growth in PC12 cells (135). Conversely, the mineralocorticoid hormone aldosterone, which promotes ENaC transcription, accelerates “wound” closure in scratch assays in an ENaC-dependent way. In BeWo cells, ENaC is more abundant at the wound edge (116). This is consistent with the observation that the cell membrane potential of corneal endothelial cells is more depolarized at the wound edge (88). This ENaC-mediated depolarization of the cell membrane potential was proposed to stimulate cell migration by triggering actin cable formation at the wound edge. Alternatively, ENaC could accelerate migration by producing a local volume gain at the cell front, since it is localized at the leading edge of the lamellipodium.

Acid-sensing ion channels (ASICs) belong to the same superfamily of cation channels as ENaC (275). Their proton sensitivity could make them attractive signaling molecules for adapting the behavior of tumor cells to their acidic environment (603). Indeed, enhanced expression of ASIC2a and -3 was found in adenoid cystic carcinoma tissue (649). However, so far a role of ASIC1, -2, and -3 in cell migration has only been demonstrated in astrocytes, glioblastoma (268, 608, 609), and smooth muscle cells (192, 193). The common result of these studies is that ASIC2 is a negative regulator of migration in astrocytes or A10 vascular smooth muscle cells by suppressing a voltage-independent, amiloride-inhibitable Na+ inward current while ASIC1 and ASIC3 are promigratory. ASIC2 expression in the plasma membrane is inversely regulated by heat shock protein 70 (Hsc70). Thus an increase in Hsc70, as seen in malignant glioma, leads to a decrease in the cell surface expression of ASIC2 (194, 609). In a chronic-hypoxia induced pulmonary artery hypertension model, ASIC1 and ASIC3 keep pulmonary artery smooth muscle cells in a synthetic migratory phenotype and prevent the activation of a complex bone morphogenic protein (BMP) signaling cascade that would drive the differentiation towards a procontractile phenotype (75).

D. Cl Channels

For reasons of electroneutrality, cations crossing the plasma membrane must be accompanied by anions. Similarly, osmotically relevant salt transport leading to (local) changes of cell volume requires the simultaneous movement of cations and anions. Thus the evidence for the involvement of Cl channels in cell migration is very strong. In fact, research on the role of Cl channels in glioma cell migration culminated in first clinical trials (352, 548). Nonetheless, the number of publications on Cl channels in cell migration is still relatively low. Part of the delay in Cl channel research in the field of migration is most likely due to the fact that the molecular identity of some Cl channels such as the volume-regulated anion channel (VRAC) is still elusive (139).

1. ClCs, VRAC

The strongest evidence for the involvement of Cl channls in cell migration has been obtained for ClC3 and VRAC. In glioma cells, Cl is accumulated far above electrochemical equilibrium (presumably by NKCC1) so that Cl channel activity (in conjunction with K+ channel activity) elicits massive Cl efflux which is accompanied by cell shrinkage (201). In some migrating tumor cells, Cl efflux can also be mediated by KCl cotransporter-4 (KCC4; Ref. 82). Numerous studies show that Cl channel activity is required for glioma cell invasion (107, 347, 376, 470, 549, 620) (see FIGURE 5). A model was put forward according to which hydrodynamic volume changes mediated by ion channels and transporters support invasion of glioma cells into and through the narrow and tortuous extracellular space of brain tissue (107, 201, 375, 376, 620). ClC3 channels also contribute to the migration of neutrophils (396, 611), HeLa cells (356), nasopharyngeal carcinoma cells (355), and wound closure in Xenopus laevis embryos (163). Possibly, the intracellular Ca2+ concentration plays a coordinating role since ClC3 channels are indirectly regulated by [Ca2+]i via their association with calmodulin-dependent protein kinase II (107).

Figure 5.

Involvement of volume-regulated Cl channels in glioma cell movement. A–D: sequential images of a migrating D54MG glioma cell. Times indicate when the images were acquired after obtaining a whole cell patch-clamp recording. The patch pipette is visible as a dark shadow at the lower left hand corner. E: time course of NPPB-sensitive whole cell current recorded at holding potentials of −40 mV (open symbols) and +40 mV (closed symbols). The current amplitude steadily increased after the cell began to extend a leading process (white arrow) and contracted its trailing edge. Letters correspond to the images in A-D. Data acquisition was begun 125 s after establishing a whole cell recording. F: ramp currents from the data in E. Letters correspond to the time points in E. [From Ransom et al. (470).]

Several studies revealed that the migration of fibroblasts (513), monocytes (280), microglia (659), or nasopharyngeal cancer cells (357, 358) requires VRAC activity. The expression of the H-ras oncogene in fibroblasts enhances the volume sensitivity of VRAC leading to faster VRAC-dependent migration (513). VRAC activity is highly sensitive to changes of the plasma membrane cholesterol concentration. Cholesterol depletion leads to VRAC activation (286, 324). In this context it is interesting to note that the plasma membrane cholesterol concentration is twice as high at the cell front than at the rear part of migrating endothelial cells (602) that also express VRAC (324). Such asymmetric distribution of cholesterol could sensitize VRAC at the rear part of migrating cells to changes of cell volume. Thereby, the distinct chemical composition of the plasma membrane at the front or rear part of migrating cells could lead to local differences of VRAC activity at the opposing cell poles and facilitate their cooperation with KCa3.1 channels that are also more active at the rear part (520).

2. Other Cl channels

There are only a few reports on the role of other Cl channels in cell migration (22, 250, 505, 594), and mechanisms of action have not yet been elucidated in detail. Members of the TMEM16/ANO family are particularly interesting candidates since they are Ca2+-sensitive Cl channels (139, 304) so that their activity could easily be coordinated with Ca2+-sensitive K+ channels and/or with other Ca2+-dependent mechanisms of cell migration (see sect. IVE). Yet, their role in cell migration still needs to be further investigated. So far, this has only been shown for TMEM16A/ANO1 whose overexpression is highly correlated with the occurrence of distant metastases in head and neck squamous cell carcinoma (HNSCC) (22).

3. GABA receptors

In the developing cerebral cortex, migration of interneurons and neuroblasts is strongly affected by the function of ionotropic GABAA and GABAC receptors (43, 120, 220). Depending on the expression level of the K+/Cl cotransporter KCC2 and the resulting shift of the Cl equilibrium potential, activation of GABAA receptors either leads to a depolarization (low KCC2 expression, high [Cl]i) or to a hyperpolarization (high KCC2 expression, low [Cl]i). A GABAA receptor-induced depolarization promotes migration of interneurons by activating L-type Ca2+ channels while a GABAA receptor-induced hyperpolarization constitutes a stop signal by abrogating Ca+ channel ativation (43). In oligodendrocyte progenitor cells (NG2 cells), GABAA receptors cooperate with voltage-activated Na+ channels and the Na+/Ca2+ exchanger NCX1 and promote migration by inducing a depolarization of the cell membrane potential, too (see sect. IVB) (589). These studies (43, 589) are remarkable for two reasons. On the one hand, they highlight the importance of the cell membrane potential of excitable neuronal cells in controlling cell migration. On the other hand, they illustrate the cell-specific composition of the GABA receptor-associated transportome that is needed for efficient cell migration, KCC2 and L-type Ca2+ channels in interneurons, and voltage-gated Na+ channels and NCX1 in oligodendrocyte progenitor cells (NG2 cells).

E. Calcium Regulatory Transport Proteins

Cell migration is a Ca2+-dependent process (161, 292, 448, 658) since the migration machinery comprises many Ca2+-sensitive effector molecules. Examples include myosin II (34, 37), calpain (155, 346, 600), calcineurin (101, 314), Ca2+/calmodulin-dependent protein kinase (140), gelsolin (377), integrins, and S100 proteins (575) as well as a number of ion channels and transporters. In addition, Ca2+ signaling during migration also induces gene transcription (591).

The effects of [Ca2+]i on the cellular migration machinery are spatially and temporally regulated with precision. Thus [Ca2+]i is not uniform throughout the cytosol of migrating cells. It has been known for almost 20 years that there is a gradient of [Ca2+]i along the length axis of migrating cells. In general, [Ca2+]i is higher at the rear part than at the front of migrating cells (56, 202, 519). This global gradient of [Ca2+]i is at least in part due to the subcellular distribution of Ca2+ storing organelles. The lamellipodium is virtually organelle-free (3, 102) so that Ca2+ release from intracellular Ca2+ stores occurs predominantly in the cell body (519). Recent studies provided a refinement of this view in that zones of locally elevated [Ca2+]i and short-lived “Ca2+ flickers” or localized Ca2+ transients were revealed at the leading edge of the lamellipodium of MDCK-F cells and fibroblasts, respectively (101, 149, 625). These local Ca2+ signals that are superimposed on the background of the global front-rear Ca2+ gradient are required for directional migration by leading to a calcineurin-dependent modulation of cell adhesion. Possibly, Ca2+ flickers also regulate calpain II that is enriched at the front of chemotaxing neutrophils constituting a “frontness” signal promoting polarization (417).

In addition to local and spatially restricted variations of [Ca2+]i and short-lived Ca2+ flickers, there is also another temporal component of [Ca2+]i signaling. Transient global elevations of cytosolic Ca2+ are required among others for migration of neutrophils (226, 353, 359), vascular smooth muscle cells (503), transformed epithelial cells (519), keratocytes (132), amoebae (345), cortical interneurons (43), and neuroblasts (292). Similarly, Ca2+ transients are involved in regulating nerve growth cone motility (183).

The complexity of Ca2+ signaling in cell migration is further increased by its crosstalk with other signaling pathways. Notably, there is a marked crosstalk between Ca2+ and cAMP/protein kinase A-dependent signaling (41, 237), the latter being also an important regulator of cell motility (378).

Spatial and temporal fine tuning of [Ca2+]i contributes to coordinating the concerted action of different effector proteins. This can be exemplified by considering mechanisms underlying the retraction of the rear part of a migrating cell (132, 133, 345). A rise of [Ca2+]i triggers contraction of the actomyosin network following the phosphorylation of the myosin light chain (644). It induces calcineurin-mediated release of cell-matrix contacts (314), activates the calcium-regulated focal adhesion protein proline-rich tyrosine kinase-2 (Pyk2), as well as the effector proteins paxillin and p130(Cas) (7). Activation of KCa3.1 and indirectly of ClC3 channels (107) causes a local shrinkage of the rear part of migrating cells (515). Accordingly, [Ca2+]i transients occur shortly before the retraction of the rear part of migrating neutrophils or keratinocytes (143, 317). Finally, a rise of [Ca2+]i also triggers the exoytosis of intracellular vesicles and lysosomes so that silencing members of the synaptotagmin family of calcium-sensing vesicle-fusion proteins such as SYT7 impair chemotaxis and rear end retraction (100).

Like in any other cell, the [Ca2+]i of migrating cells is determined by Ca2+ influx, release and uptake of Ca2+ ions from intracellular stores, and extrusion of Ca2+ into the extracellular space. This is accomplished by a wide variety of channels, transporters, and pumps located in the plasma membrane and in the membranes of intracellular Ca2+ stores (94). However, despite the importance of [Ca2+]i as a coordinator of different components of the migration machinery, the identification of the spatial and temporal contribution of molecularly defined Ca2+ transport proteins to the mechanisms of cell migration has only begun. This applies in particular to those Ca2+ channels, pumps, and exchangers that are responsible for releasing Ca2+ from intracellular stores and for removing Ca2+ from the cytosol into intracellular stores or across the plasma membrane into the extracellular space.

1. TRP channels

A rapidly increasing number of studies has firmly established that transient receptor potential (TRP) channels constitute an important element of the cellular migration machinery. They are especially relevant for migrating cells since TRP channels are considered as polymodal cell sensors (409). They are activated by many external stimuli, mediate receptor-operated Ca2+ entry, and thereby enable cells to dynamically adapt their migrational behavior. This also includes responses to mechanical cues, although the mechanosensitivity of TRP channels is still a matter of debate (190, 360). Mammalian TRP channels are ubiquitously expressed, and they are all Ca2+ permeable except for TRPM4 and TRPM5 channels (174).


TRPC channels are central constituents of receptor-activated signaling pathways and as such are responsible for the so-called receptor-operated calcium entry (ROCE) downstream of G protein-coupled receptors. This mode of activation explains why an increasing number of studies point to a strong link between TRPC channel function and (chemotactically) directed migration.

I) TRPC1. TRPC1 channels are required for polarization of migrating cells (149) and/or for their response to external directional cues in vitro and in vivo (39, 148, 471, 473, 654) (see FIGURE 6). TRPC1 channel proteins localize to the cell front of glioma cells chemotactically stimulated with EGF (39) and of neutrophil granulocytes (281). Consistent with this subcellular distribution, TRPC1 underlies the local elevation of [Ca2+]i at the very cell front of transformed renal epithelial (MDCK-F) cells (149). Thus the impact of TRPC1 channels on local Ca2+ dynamics at the cell front and directional migration appears to be similar to that of TRPM7 channels (see below; Ref. 625). Importantly, TRPC1 channels primarily control the “steering” mechanism of MDCK-F and glioma cells, since they have only a minor effect on basal migration. In glioma and intestinal epithelial cells as well as myoblasts (346), TRPC1-dependent directional migration is correlated with TRPC1-mediated store-operated Ca2+ entry which in the case of intestinal epithelial cells is enhanced by forming a complex with STIM1 (39, 473). So far, the mitogen-activated protein kinase/ERK1/2 cascade (148, 654), PI3K/Akt (657), and calpain (346) have been identified as intracellular signaling pathways and effector proteins involved in TRPC1-dependent migration.

Figure 6.

TRPC1 channel-dependent chemotaxis of MDCK-F cells in an FGF-2 gradient. A and B: paths of individual MDCK-F cells migrating in an FGF-2 gradient with the concentration rising towards the right. The paths are normalized to a common starting point. MDCK-F cells overexpressing TRPC1 channels (A) move up the FGF-2 gradient while cells with silenced TRPC1 channels (B) do not respond to this chemokine gradient. C: summary of the chemotaxis experiments shown in A and B. The displacement of the cells into the direction of the FGF-2 gradient is plotted as a function of time. The open symbols (control) illustrate the random movement in the absence of an FGF-2 gradient. [Modified from Fabian et al. (148), with kind permission of Springer Science + Business Media.]

TRPC1 channels are also required for the guidance of nerve growth cones in response to a number of different neurotropic factors such as netrin-1, brain-derived neurotrophic factor (BDNF), or myelin-associated glycoprotein (MAG) (535, 536, 613, 628). TRPC1-dependent axon guidance involves among other factors intracellular Ca2+ signaling via calcineurin and Slingshot phosphatase (628).

II) TRPC5. Depending on the cell type, TRPC5 channel activity can have a positive or a negative impact on cell migration. Possibly, this apparent discrepancy can be accounted for by the magnitude, duration, and localization of the intracellular Ca2+ signal elicited by TRPC5 activation. Moreover, the presence or absence of TRPC1-TRPC5 heteromers that have a lower Ca2+ permeability than monomers can also contribute to the variable effects of TRPC5 on cell migration (303, 534, 559). In vascular smooth muscle cells, TRPC5 activation and the resulting (transient) rise of [Ca2+]i triggered by sphingosine-1-phosphate (636) or by oxidized phospholipids (9) correlates with motility. Similarly, TRPC5 promotes migration of angiotensin II-stimulated podocytes by activating Rac1 in a Ca2+-dependent manner thereby disassembling stress fibers (583). An inverse relation of TRPC5 channels with respect to cell motility is observed in lysophosphatidylcholine (LPC)-stimulated aortic endothelial cells. TRPC5 channels inhibit migration because they trigger a sustained elevation of [Ca2+]i once they have been recruited to the cell membrane in response to an initial TRPC6-mediated increase of [Ca2+]i (78). In hippocampal neurons, TRPC5 channels fulfill a dual task. They promote axon formation by mediating localized Ca2+ influx that activates Ca2+/calmodulin kinase kinase (CaMKK) and its target Ca2+/calmodulin kinase I (111). At the same time, TRPC5-dependent Ca2+ transients limit the linear growth of nerve growth cones presumably by stabilizing their adhesive contacts (191).

III) TRPC6. TRPC6 channels play an important role in the migration of endothelial cells and granulocytes. They are required for VEGF-dependent migration and sprouting in an in vitro angiogenesis assay of endothelial cells (172, 205). Notably, TRPC6 channels exhibit specificity for VEGF since FGF-induced tube formation is still present (172). In human pulmonary arterial endothelial cells, a mechanism underlying TRPC6 cell surface expression was identified that has important implications for migration of other cell types, too. TRPC6 channels interact with the COOH-terminal tail domain of PTEN (284), which is a central player in chemotaxis (223). Functional expression of TRPC6 channels in the plasma membrane and their contribution to angiogenesis requires their interaction with PTEN (284).

Endothelial cells play an important role in recruiting granulocytes from the bloodstream to sites of infection or inflammation with selectin-mediated interactions between endothelial cells and leukocytes being one of the earliest steps in this cascade. TRPC6 channels that are expressed in granulocytes (109, 379, 531) are involved in this process. This was assessed in a model of an allergic airway response (531). Less eosinophil granulocytes, IL-5, and IL-13 are isolated from bronchoalveolar lavage fluid from TRPC6−/− compared with that from wild-type (wt) mice. These observations can be explained in two ways. Either lower cytokine secretion leads to a decreased maturation and weaker stimulation of eosinophils, or the infiltration/migration process itself is impaired. The following observations lend support to the second possibility. In neutrophils, TRPC6-dependent Ca2+ signaling induced by soluble E-selectin is augmented by the simultaneous application of the chemoattractant platelet activating factor (PAF) (379). Moreover, TRPC6−/− neutrophils cover shorter distances on a two-dimensional glass surface than wt cells when stimulated with macrophage inflammatory protein-2 (MIP-2) (109). Finally, unpublished findings from our laboratory revealed a severe chemotaxis defect of neutrophils from TRPC6−/− mice. Thus TRPC6 channels appear to function as central players in granulocyte recruitment by integrating selectin and chemoattractant receptor signaling.

Besides their role in cell proliferation and cell cycle progression (60, 579), TRPC6 channels also promote tumor cell invasiveness. In glioblastoma, hypoxia-induced upregulation of TRPC6 channel expression, depending on Notch signaling, importantly contributes to an aggressive phenotype with increased invasiveness by activating the the calcineurin/NFAT pathway in a Ca2+-dependent way (90).


TRPV5 and -6 are the TRP channels with the highest Ca2+ selectivity. However, to the best of our knowledge, there are no studies directly investigating their role in cell migration. Nonetheless, such a role is conceivable given the fact that TRPV6 channels are partners of KCa3.1 channels in prostate cancer cells (306).

I) TRPV1. A series of papers characterized the interplay between TRPV1 channels and microtubules in regulating the movement of nerve growth cones of dorsal root ganglial (DRG) cells or of TRPV1 transfected neuroblastoma × DRG neuron hybrid cells (184186, 188). The COOH-terminal part of TRPV1 interacts with microtubules and thereby stabilizes them while the activation of TRPV1 channels promotes the disassembly of dynamic microtubules that can account for the shortening of DRG neurites following TRPV1 activation. At least part of the effects of TRPV1 channels on the microtubule network can be classified as a “nonconductive property” that is mediated by intracellularly localized NH2- and COOH-terminal domains of the channel (186). In hepatocyte growth factor (HGF)-stimulated HepG2 hepatoblastoma cells, activation of TRPV1 channels enhances migratory activity. Due to a differential effect of TRPV1 on stable and dynamic microtubules (185), it was speculated that activation of TRPV1 with capsaicin could have modulated microtubule dynamics at the rear part of the cells in a way that allows an easier retraction of this cell pole, thereby stimulating migration (617). In corneal epithelial cells, TRPV1 stimulates migration by transactivating the EGF receptor (643).

II) TRPV2. TRPV2 channels contribute to the migration and chemotaxis of macrophages (337, 403) and PC3 prostate cancer cells (393). In a murine macrophage cell line (TtT/M87) and PC3 cells, TRPV2 channels are translocated from the endoplasmic reticulum to the plasma membrane following stimlation with fMLP and LPC or lysophosphatidylinositol, respectively, which then underlies the sustained agonist-induced Ca2+ influx and chemotaxis. Importantly, TRPV2 expression is higher in samples from patients with metastatic disease than in solid primary tumors consistent with its role in tumor cell migration (394).

III) TRPV4. Mechanosensitive Ca2+ channels play an important role in coordinating the movement of the front and rear part of migrating cells and in the response to an altered mechanical load of cell-matrix contacts (317). However, so far the molecular nature of mechanosensitive Ca2+-permeable channels in migrating cells is still elusive. TRPV4 channels contribute to cell volume regulation (28, 30) and other mechanical responses such as ciliary beating under viscous strain (12) or mechanical hyperalgesia (10). Applying external forces to focal adhesion complexes via the movement of magnetic beads leads to an almost instantaneous activation of TRPV4 channels and Ca2+ influx into the stimulated endothelial cells (369). Such findings make TRPV4 channels an attractive candidate for the elusive mechanosensitive Ca2+ channel in cell migration. This view is further supported by observations that TRPV4 channels indeed affect cell migration. Stimulation of TRPV4 channels with their activator 4α-PDD increased lamellipodial dynamics of HepG2 hepatoblastoma cells (617). They are required for the realignment of endothelial cells under cyclic strain (580), and TRPV4 channels mediate arachidonic acid-induced migration of breast cancer-derived endothelial cells that express these channels at a much higher level than “normal” endothelial cells (152). On the other hand, a careful analysis of their role in migration of an immortalized neuroendocrine cell line (GN11) came to the opposite conclusion providing further evidence for the importance of the cellular context (transportome) of a given ion channel for cell migration and the temporal pattern of its activation.


TRPM channels were first dicovered as a tumor suppressor in melanoma (TRPM1), and expression of other TRPM channels has been linked with tumors (297). However, recent studies show that they are also involved in controlling immune cell function and motility of a number of other cell types. Some of its members, TRPM2, -6 and -7, are “chanzymes” combining the function of an ion channel and an enzyme in the same protein (174) so that their impact on cell migration is at least partially due to their non-conductive properties.

I) TRPM2. TRPM2 channels play an important role in neutrophil recruitment. They are involved in increasing the endothelial permeability in oxidant-stressed tissue (221), and they are required for chemotaxis of neutrophils towards fMLP (365, 442) and for chemokine production by macrophages (640). In neutrophil granulocytes, TRPM2 channels are activated by ADP-ribose that is formed intracellularly following stimulation with chemoattractants such as fMLP (222). Interestingly, TRPM2−/− neutrophils behave normally in a gradient of another chemokine, CXCL2 (640). Nonetheless, neutrophils do not infiltrate the colonic mucosa in a colitis model in TRPM2−/− mice, which involves elevated CXCL2 expression. In this model, TRPM2 channels support neutrophil chemotaxis rather indirectly since they are required for chemokine production by macrophages (640). The link between TRPM2 channels, inflammation, and chemokine production is given by reactive oxygen species (ROS) that are released by neutrophils and are potent activators of TRPM2 channels (215).

It is intriguing to note that distinct TRP channels in neutrophils appear to be coupled to distinct signaling pathways and thereby mediate the migratory response of neutrophils to different chemoattractants. TRPM2 channels are linked to the fMLP receptor, while TRPC6 channels appear to couple to the CXCR2 receptor (109). It is tempting to speculate that spatially restricted Ca2+ signals locally trigger the signaling cascades downstream of the respective receptors. Such local regulation could also involve intracellularly localized TRPM2 channels as evidenced in dendritic cells. In these cells, efficient chemotaxis requires TRPM2-mediated Ca2+ release from lysosomes (565).

II) TRPM4. TRPM4 is a Ca2+-activated channel permeable only for monovalent cations that controls the [Ca2+]i indirectly by setting the cell membrane potential (see sect. IIIA). So far, the role of TRPM4 channels in cell motility has been studied only in immune cells. In TRPM4−/− mice, migration of dendritic cells to lymphoid organs is impaired because chemotaxis is inhibited due to a Ca2+ overload-dependent downregulation of PLC-β2. When dendritic cells are matured with fixed Escherichia coli, the defect in chemotaxis overrides the increased migratory activity that has almost doubled in TRPM4−/− dendritic cells (23). A similar phenotype was observed in bone marrow-derived mast cells stimulated with stem cell factor (SCF) (538). These studies represent further examples for the dissociation of ion channels' impact on the “migratory motor” and the “steering wheel” of a migrating cell (see FIGURE 7). The impact of TRPM4 channels on cell motility critically depends on the expression level. Th2 cells express more TRPM4 channels than Th1 cells. Blocking TRPM4 channel function results in amplified intracellular Ca2+ signals in Th2 cells while they are reduced in Th1 cells. Conversely, migratory activity is impaired in Th2 cells, while it is increased in Th1 cells following TRPM4 inhibition (624).

Figure 7.

Migration and chemotaxis of dendritic cells depends on TRPM4 expression. A and B: trajectories of individual dendritic cells from wild-type (A) and TRPM4−/− (B) mice. The longer cell paths reveal that TRPM4 deficiency leads to increased migratory activity of dendritic cells that are stimulated with fixed E. coli. C: summary of chemotaxis experiments. Although wild-type dendritic cells migrate more slowly than TRPM4−/− cells, their chemotaxis towards CCL21 is more efficient. [From Barbet et al. (23), with permission from Nature Publishing Group.]

III) TRPM7. TRPM7 is a Ca2+- and Mg2+-permeable ion channel with a kinase domain (493, 494) that belongs to the family of α-kinases that is a relative of the Dictyostelium myosin heavy chain kinase that can induce myosin II filament disassembly (97). Thus TRPM7 is a prototype channel with pronounced nonconductive properties. Earlier studies focused on the impact of TRPM7 channels on cell adhesion (96, 562). Several other recent studies addressed the role of TRPM7 channels in cell migration (1, 61, 79, 169, 339, 496, 563). In migrating fibroblasts TRPM7 produces short-lived local increases of [Ca2+]i (“Ca2+ flickers”) at the leading edge (see FIGURE 8) (625). Their accentuation by mechanical stimulation is consistent with earlier studies on the cellular localization of mechanosensitive Ca2+ influx at the cell front (399) and with the reported mechanosensitivity of TRMP7 channels (415). Thus TRPM7 is another candidate channel for the elusive mechanosensitive Ca2+ channel of migrating cells. The frequency of TRPM7-mediated Ca2+ flickers also correlates with the cells' turning behavior in chemotactic gradients of PDGF (625). Moreover, in an in vitro wound healing assay with Swiss 3T3 fibroblasts (563) and during Xenopus laevis gastrulation (339), TRPM7 channels drive the polarization of migrating cells. Thus Ca2+ signals at the leading edge appear to be part of a local feedback system underlying cell polarity (99, 147). In line with the concept of a cell-specific transportome, the role of TRPM7 channels in directional persistence or polarization can be taken over by TRPC1 channels (149).

Figure 8.

TRPM7-mediated calcium flickers. A: overlay of calcium flicker ignition sites (red dots) and focal adhesions (green; integrin α5 staining). Enlarged views of calcium flickers, focal adhesions, and their overlay are shown to the right (top, middle, and bottom, respectively). N denotes the nucleus. B: calcium flickers in the lamellipodium before (left) and after (right) modifying traction force by applying an integrin ligand (RGDS peptide), blebbistatin, or cytochalasin. These treatments predominantly affect the flicker probability rather than its amplitude. [From Wei et al. (625), with permission from Nature Publishing Group.]

The story about TRPM7 and cell polarization took a new twist by revealing that Ca2+ can be substituted and/or supported by local Mg2+ influx (496). Thus the defect in cell polarization and in gastrulation following TRPM7 knockdown is rescued by expressing the Mg2+ transporter SLC41A2 (339, 563), and preventing Mg2+ influx by TRPM7 silencing impairs the migration of osteoblasts (1).


Orai and STIM proteins mediate Ca2+ release activated Ca2+ (CRAC) current with Orai proteins constituting the conductive pathway for Ca2+ and STIM being the sensor of the filling status of the ER Ca2+ stores (58). Since they are closely linked to TRP channel function, we will discuss these proteins here. Reflecting the eminent importance of CRAC current following receptor stimulation, several recent studies addressed the role of Orai/STIM in chemotactically or chemokinetically stimulated migration with a particular focus on cell adhesion. Current evidence indicates that integrin outside-in signaling is a common target of Orai1/STIM1-mediated Ca2+ signaling in cell migration. Orai1 plays a role in the recruitment of neutrophils from the blood flow to the endothelium (130, 501). Reduced Orai1 expression delays the onset of arrest and polarization of neutrophils on the endothelium under shear flow conditions because the accompanying intracellular Ca2+ signal is attenuated (501). Orai1-mediated Ca2+ influx is triggered by engagement of high-affinity LFA-1 whose clustering as well as actin polymerization at the cell front are promoted by colocalized Orai1 (130). The activation status of LFA-1 appears to play a decisive role in its interaction with Orai1, since the migration of T lymphocytes without shear stress is independent of Orai1 (567). STIM1-deficient T cells lack store-operated Ca2+ entry (SOCE) in response to CXCL11, CCL19, and CCL20 via stimulation of CXCR3, CCR7, and CCR6 receptors, respectively, leading to impaired chemotaxis. This defect contributes to the protection from experimental allergic encephalomyelitis (EAE) in T cell-specific STIM1-deficient mice (349). Orai1 and STIM1 are also involved in the migration of serum-stimulated breast (645) and EGF-stimulated cervical cancer cells (81). In both cancers, altered STIM1 expression indicates poor patient prognosis (81, 372). In intestinal epithelial cells, STIM1 acts in conjunction with TRPC1 to promote wound healing in a “scratch assay” of epithelial restitution (473). In vascular smooth muscle cells isolated from rat aorta, PDGF-induced Ca2+ entry and the resultant stimulation of migration depend on Orai1/STIM1. This is pathophysiologically relevant since the expression of Orai1/STIM1 is upregulated in balloon-injured carotid arteries (35). Mechanical injuries of the vascular wall lead to enhanced migration of the vascular smooth muscle cells. Orai1 expressed in endothelial cells contributes to VEGF-induced chemotaxis and angiogenesis (327).

2. P2X receptors

In addition to the purinergic P2Y receptors (301), some members of the P2X family of ATP-gated ion channels (P2X1, P2X4, and P2X7) have been implicated in cell migration and chemotaxis, especially in immune cells (316, 421). ATP itself, however, is not a chemoattractant but rather acts as an autocrine amplifier for other chemoattractants such as C5a (249). P2X1, P2X4, and P2X7 are commonly found in cells of the immune system (117, 209, 619). At first sight, P2X7 receptors are least likely to be involved in cell migration, since channel activation requires greater than ∼300 μM ATP, well beyond saturating concentrations (∼100 μM ATP) for the other P2X family members (412). However, P2X7 receptors are highly expressed in some cancers including breast and thyroid cancer (545, 547). In breast cancer cells, P2X7 together with KCa2.3 contributes to cystein cathepsin-dependent tumor cell invasiveness (259). The proinvasive effect of P2X7 receptors depends on the activity of KCa2.3 channels, which are probably activated by the P2X7-mediated rise of the intracellular Ca2+ concentration.

Pharmacological inhibition or siRNA knockdown of P2X4 receptors decreases ATP-induced chemotaxis (421) and morphine-induced migration of microglial cells (235) by as yet undefined mechanisms. P2X4 or other P2X receptor subtypes could contribute to Ca2+ flickers at the leading edge (625). Alternatively, P2X4-mediated cation influx could lead to osmotic water uptake and local swelling, if it occurred predominantly at the cell front. P2X1 receptor-deficient neutrophils migrate more slowly than wild-type cells, and the authors deduced that P2X1 activation leads to increased Rho activity and trailing end retraction by an unknown mechanism (316).

3. Na+/Ca2+ exchanger

The temporal and spatial “shape” of intracellular Ca2+ signals critically depends on transport proteins mediating the export of Ca2+ across the plasma membrane or into intracellular stores. Their role in cell migration is presently underestimated. One of the relevant transport proteins in this context is the NCX (348). Depending on the electrochemical gradients, however, the Na+/Ca2+ exchanger can export (forward mode) or import Ca2+ (reverse mode). NCX is inhibited by the isothiourea derivative KB-R7943, which is >10-fold more potent in inhibiting the reverse than the forward mode (253). We showed that in migrating MDCK-F cells the Na+/Ca2+ exchanger is one of the major components of the Ca2+ efflux machinery required for cell migration (134). Moreover, we observed that (forward) NCX activity is also necessary for PAF-stimulated migration of human neutrophil granulocytes (unpublished observation from our laboratory). NCX also participates in the migration of oligodendrocyte progenitor (NG2) cells (589), primary cardiac myofibroblasts (466), tendon fibroblasts (499), microglia (244), pancreatic cancer (131), and corneal endothelial cells (89). In these cell types, Na+/Ca2+ exchange was reported to be operating in the reverse mode. However, reverse mode of operation was not shown in all cells. In microglia or in cardiac myofibroblasts, it was inferred from the use of KB-R7943 (244, 466). Some caution is justified at this point since the KB-R7943 sensitivity strongly depends on the NCX1 splice variant. The cardiac splice variant, NCX1.1, is relatively insensitive to KB-R7943 when operating in the forward mode (IC50 >30 μM) (253), while splice variants NCX1.3 and NCX1.7 are inhibited in their forward mode by KB-R7943 with IC50 values of 2–3 μM (206). To the best of our knowledge, it is not yet known which NCX isoforms are expressed in primary cardiac myofibroblasts or microglial cells.

F. Aquaporins

The discovery that aquaporins constitute important components of the cellular migration machinery was a major breakthrough in the field (see FIGURE 9; reviewed in Refs. 128, 344, 435). It provided strong support for the model that ion channels and transporters involved in cell migration can do so by inducing local changes of cell volume. Regardless of the cell type, genetic deletion of aquaporins in mice or knockdown of aquaporins in cultured cells is accompanied by reduced migratory activity. The involvement of aquaporins 1, 3, 4, 5, 7, and 9 in cell migration was shown in in vivo and in vitro in endothelial (242, 395, 497), melanoma (238), glial (374, 498), renal proximal tubular (213), corneal epithelial (322), gastric epithelial (218) and gastric carcinoma cells (241), chondrocytes (331), keratocytes (214, 492), neutrophils (270, 342), dendritic cells (212), fibroblasts (63), lung adenocarcinoma (83), and neuronal stem cells (294). The widespread “use” of multiple aquaporins in cell migration clearly points to the universal nature of this mechanism. Several studies have shown that aquaporins are concentrated at the leading edge of the lamellipodium (213, 342, 497, 498) and that water flux is increased in motile cell regions of neutrophils (342). It was proposed that a local osmotic disequilibrium at the cell front induced by local ion fluxes [e.g., by Na+/H+ and Cl/HCO3 exchange (287), Na+-K+-2Cl (200, 404, 527) or K+-Cl cotransport (659)] or actin polymerization (428) leads to local water flow supporting the formation of protrusions (435). Accordingly, lamellipodial activity is enhanced upon expression of aquaporins (497). A theoretical analysis comes to a similar result, namely, that an osmotic gradient elicits water fluxes that result in the movement of cells (256). Water flow into the lamellipodium would also be in line with the so-called Brownian ratchet model. This model predicts that bending of actin filaments due to thermal fluctuations makes room for actin monomers to polymerize with their barbed ends pointing towards the plasma membrane at the leading edge. When the elongated actin filaments recoil, they exert an elastic pushing force on the membrane and thereby protrude the leading edge of the lamellipodium (390). As a consequence of local water influx, intracellular constituents such as actin monomers are diluted. The resulting concentration gradient may drive actin polymerization in growing filopodia by steepening the gradient from the cell body into the filopodium and thereby promote forward flux of actin monomers (343). Water flow across the plasma membrane of the leading edge also provides an additional protrusive force because the poroelastic nature of the cytoplasm (386) prevents the immediate dissipation of a localized volume increase at the front. In addition to their transport-related effects on cell migration, aquaporins also elicit other effects based on their impact on the [Ca2+]i (294) or nonconductively by interacting with other intacellular proteins such as Lin-7/β-catenin (395) or PKC-ξ and Cdc42 (343).

Figure 9.

Involvement of aquaporins in cell migration. A: AQP1−/− mice have an extended survival with subcutaneous melanoma (left). This beneficial effect of AQP1 knockout is due to impaired angiogenesis. The number of vessels in subcutaneous melanoma is lower in AQP1−/− than in wild-type mice (right). In vitro analysis shows that (B) wound healing of cultured endothelial cells occurs more slowly in cells from AQP1−/− than from wild-type mice (left; blue, initial wound edge; red, after 24 h) which is quantified as the wound edge speed (right). C: transfecting CHO cells with AQP1 enhances their migratory activity which is illustrated by longer paths of individually tracked cells (left). Cells were tracked over 4 h, and arrows indicate initial positions. Middle and right: AQP1 expression increases the number of ruffles at the leading edge of the lamellipodium, which correlates well with the increased protrusive activity. [From Saadoun et al. (497), with permission from Nature Publishing Group.]

The contribution of aquaporins to the cellular migration machinery has important pathophysiological implications and therapeutic potential in oncology (435). AQP1 contributes to tumor angiogenesis (497) and drives metastases by promoting tumor cell extravasation (238). Its (over)expression in breast cancer cells (430) like that of AQP5 in chronic myeloid leukemia cells (73) and non-small-cell lung cancer (NSCLC) (350) correlates with poor prognosis. Moreover, aquaporins are important modulators of the wound healing process. AQP1 and AQP3 are required for efficient restitution of the wounded corneal epithelium (322, 492). Similarly, AQP3 is involved in wound healing of the skin (63, 214). The beneficial effects of EGF on wound closure are closely correlated with EGF-induced AQP3 expression (63). On the other hand, genetic deletion of AQP4 greatly reduces glial scar formation (21, 498). Finally, the role of aquaporins in the migration of immune cells qualifies them as potential targets in the treatment of inflammatory diseases (212, 270, 328, 342).


TABLE 2 gives a synopsis of this section, listing all transporters and their function in cell migration that will be discussed in detail in the following part.

View this table:
Table 2.

Transporters involved in cell migration

A. pH Regulatory Transport Proteins

Intra- and extracellular pH homeostasis reflects the integrated function of intra- and extracellular buffers, pH-regulatory transport proteins, and the production/consumption of protons by metabolism (42). Precise pH regulation by transport proteins that move acids or bases across the plasma membrane is mandatory for all cells (488), since the protonation of proteins is a crucial determinant of their function. Accordingly, the function of proteins comprising the cellular migration apparatus is also pH dependent. That cell migration is a pH-dependent process was discovered already in the 1920s (233). Since then, many studies have shown that intra- and extracellular pH (pHi and pHe) are critical determinants for efficient cell migration (57, 118, 217, 287, 299, 351, 436, 476, 481, 552, 554, 556). The interrelation of pH and cell migration is of particular pathophysiological importance in tumor metastasis and innate immune cell function.

In solid tumors, hypoxia insufficient vascularization and tumor anemia with a reduced oxygen transport capacity of the blood lead to a switch to glycolytic metabolism resulting in an extensive production of lactate and protons (225, 300, 465). pHi is defended and may even become more alkaline than in normal cells (177) by increasing acid export so that the extracellular compartment is acidified (87, 569). Thus, in solid tumor tissues, pHe reaches values of pHe 6.5 and lower (603). This characteristic metabolic tumor microenvironment in turn increases the metastatic potential (67, 179, 622). The acidic and lactate-enriched tumor microenvironement can also provide an explanation for the inability of the immune system to eliminate tumor cells. Lactate stimulates the migration of tumor cells, while it inhibits migration of monocytes (180) and impairs the differentiation of tumor-associated dendritic cells (189). Conversely, increasing pHe in tumors by oral NaHCO3 reduced the formation of spontaneous metastases in mouse models of metastatic breast cancer (483). Since tumor cell migration constitutes one of the hallmarks of cancer and is one of the prerequisites for tumor metastasis (207, 208), the underlying pHi- and pHe-dependent mechanisms offer therapeutic potential.

Similarly, it is well established that inflamed tissues are characterized by an acidic and hypoxic microenvironment. This applies to inflammatory reactions against exogenous pathogens (542) and to autoimmune processes such as rheumatoid arthritis (618). Local hypoxia and acidosis of the inflamed tissue are at least in part due to the massive recruitment of immune cells with high metabolic activity (291). It has profound effects on cells of the innate immune system (104, 363). Like in tumors (164, 635), the acidic and hypoxic microenvironment can impact directly on the cellular migration machinery or elicit indirect effects by altering the secretion of chemokines (44, 364). The hypoxic microenvironment also leads to the upregulation HIF-1α (660). Notably, HIF-1α in turn regulates the expression of the Na+/H+ exchanger NHE1, which is one of the major pH-regulating transport proteins involved in the migration of neutrophils and tumor cells (539).

pHe and pHi have multiple effects on the cellular migration machinery. As already outlined in section IIB, integrin-mediated cell adhesion is pH dependent (321, 436, 554). Studies from our laboratory revealed that it is rather the pericellular pHe within the glycocalyx than the global extracellular pH of the bulk solution surrounding migrating cells that controls adhesion and migration (556, 561). Integrins protude by ∼20 nm above the surface of the plasma membrane (145, 574) so that the interaction of integrins with extracellular matrix proteins takes place within the pericellular nanoenvironment. We could show that the pericellular proton concentration within the glycocalyx is up to twice as high at the leading edge than at the rear part of migrating melanoma cells (299, 556). Molecular dynamics simulations and force measurements with atomic force microscopy indicated that pHe has a strong impact on the conformation of integrins (see FIGURE 10). An acidic pHe facilitates the activation of αvβ3 integrins (436). Thus the gradient of the pericellular proton concentration causes integrins to be in a more activated state at the cell front than at the rear end of migrating cells. This facilitates the formation/stabilization of focal adhesion contacts at the cell front and their release at the rear part, respectively. Moreover, the proton-dependent activation of integrins can account for the increased adhesion of melanoma cells to a collagen matrix in an acid environment (554).

Figure 10.

Model of pH-regulated integrin activation. Following inside-out activation, acidic extracellular pH can promote integrin headpiece opening from the extended-closed to the extended-open conformation. In addition, acidic pH can stimulate headpiece opening on the bent-closed integrin. Thus there are two ways by which acidic extracellular pH can cause integrin activation. Based on the extracellular pH gradient at the cell surface (556, 561), the proton-dependent activation of integrins causes cell adhesion to be stronger at the cell front than at the rear part of migrating cells. [From Paradise et al. (436).]

The extracellular pH gradient within the pericellular space is accompanied by a complementary intracellular pH gradient in melanoma and other cell types such as endothelial cells or in nerve growth cones. pHi is more alkaline at the cell front than at the rear part (361, 487, 543)(see below for a more detailed discussion). The alkaline pHi at the cell front (relative to pHi at the rear part) will lead to a higher focal adhesion turnover at the cell front due to the lower affinity of talin for binding actin and decreased maximal binding at higher pHi (551). Complementary local pH nanoenvironements are also found in invadopodia. Pericellular and intracellular pH of invadopodia are more acidic and alkaline than in neighboring regions, respectively (57, 351). Thus we have a spatial organization of the intracellular proton concentration that is reminiscent of the spatial distribution of the intracellular Ca2+ concentration (625): a global front-rear gradient and local subdomains.

In addition to cell adhesion, pHi affects actin polymerization by regulating the actin-severing protein cofilin whose activity increases with rising pHi (33, 454). The underlying mechanism is a pH-dependent disinhibition of cofilin by a decrease of cortactin binding (351). By severing actin filaments, activated cofilin thereby generates new sites for actin filament assembly at the cell front or in invadopodia (351, 361, 487). Cofilin is inhibited by PI(4,5)P2 binding which in itself is pH dependent. PI(4,5)P2 binding to cofilin is lower at pH 7.5 than at pH 6.5 so that pH-dependent binding of PI(4,5)P2 constitutes a second mechanism to account for the pH sensitivity of cofilin (156, 550). The intracellular alkalinization required for cofilin activation is brought about by NHE1 that is stimulated by many migrational cues such as growth factors (514). Similar pH-driven dynamics of a cytoskeletal motor of migration is found in ameboid sperm of the nematode Ascaris suum. Ameboid locomotion of these cells is based on treadmilling of the socalled major sperm protein (MSP) with polymerization at the cell front and depolymerization at the rear. Ascaris suum sperm establishes a pseudopodial pH gradient, with pHi being 0.15 pH units higher at the leading edge (283). This pHi gradient is of the same order of magnitude as in migrating melanoma cells (361).

pHi also regulates mechanisms underlying the retraction of the rear end of a migrating cell (103, 607) by interfering with Ca2+-dependent signals. While Ca2+ binding to calmodulin does not change when pHi is decreased from pH 6.8 to pH 7.2 (420), the interaction of calmodulin with its effectors such as calcineurin and myosin light-chain kinase is pH sensitive. An acidification increases their Ca2+ sensitivity, thereby allowing enzyme activation in the presence of suboptimal Ca2+ (240). Thus the intracellular pH gradient (361, 487) acts synergistically with the intracellular Ca2+ gradient in activating myosin light-chain kinase and consequently myosin II at the rear part of migrating cells.

Most cells migrate through a three-dimensional network of extracellular matrix proteins. Depending on the cell type, proteolytic digestion of the extracellular matrix is an essential step for efficient migration or invasion. This is particularly relevant for tumor cells (67, 278). The acidic tumor microenvironment and the local nanoenvironment surrounding invadopodia (57), which is generated by the activity of pH regulatory (transport) proteins, facilitates the action of proteases that are mostly secreted by stromal cells. Proteases either have their pH-optimum in the acidic range [e.g., MMP-3 (231), urokinase-type plasminogen activator (281), cathepsin D (333, 577), cathepsin B (572), and cathepsin L (391)], depend on the protonation of their substrate [e.g., MMP-2 (392)], or their conversion from inactive pro-MMPs to active MMPs is promoted by an acidic microenvironment (288). In addition, release and expression of proteases are also upregulated by a low pHe (45, 272, 391). The latter process requires Ca2+ influx that is inhibited by blockers of voltage-gated Ca2+ channels (273). Taken together, these studies clearly indicate that an acidic extracellular pH can activate a whole toolkit of proteases, thereby facilitating invasion and migration through the extacellular matrix.

Finally, many of the transport proteins involved in cell migration are pH sensitive themselves. Modulation occurs in both directions. KCa3.1 channels, for example, are inhibited by an intracellular acidosis (445), while channels like TRPV1 or ASICs are activated by extracellular protons (232). Yet other channels like TRPM7 respond to an extracellular acidification in a more complex way. An extracellular acidification predominantly increases the inwardly directed monovalent cation conductance and not that of divalent cations (262). However, it still remains to be determined whether the pH dependence of ion channels contributes to their role in cell migration under conditions of an increased acid load. To the best of our knowledge, this aspect still awaits clarification. Unpublished observations from our laboratory indicate that acid-mediated inhibition of migration of HepG2 cells cannot be overcome by a concurrent activation of TRPV1 channels. This is in contrast to NHE1 whose activation in response to an intracellular acidification by up to 0.5 pH units renders migration of MDCK-F cells independent of pHi (287).

1. Na+/H+ exchanger NHE1

To date, 10 mammalian isoforms of NHE have been identified (276, 318). The most relevant NHE isoform for migrating cells is NHE1 (558) that is a ubiquitously expressed housekeeping protein involved in maintenance of pHi and cell volume (229). NHE1, in particular its intracellular COOH-terminal tail, is the target for multiple binding partners and signaling cascades critical for its regulation and subcellular localization (269, 380, 381). NHE1 regulates (directed) migration of many different cell types (65, 86, 92, 108, 118, 126, 217, 287, 305, 313, 388, 443, 476, 481, 490, 514, 554). Usually, NHE1 inhibition (418) or knockdown impairs motility. There are several ways by which NHE1 affects cell migration (see Refs. 555, 557, 558 for a more detailed overview). These include effects on cell volume (see sect. IIIB), pHi (361), and thereby the assembly and activity of cytoskeletal elements including focal adhesion proteins (see sect. IIID). It also regulates the pericellular pH at the cell surface and inside the glycocalyx (299, 556), anchors the cytoskeleton to the plasma membrane (380), functions as a plasma membrane scaffold in the assembly of signaling complexes (380, 446), and modulates gene expression (462, 637).

The contribution of NHE1 to migration relies on its polarized subcellular distribution in migrating cells (see FIGURE 11). NHE1 is concentrated at the leading edge of the lamellipodium (118, 197, 287, 305, 556) where it mediates local salt and water uptake in collaboration with the Cl/HCO3 exchanger AE2 (287), Na+-K+-2Cl cotransporter NKCC1 (200, 527), aquaporin AQP1 (497), and possibly the Na+-HCO3 cotransporter NBC1 (518, 524) as well as carbonic anhydrases (566). Changes of cell volume mediated by NHE1 are also required for the chemotaxis of neutrophil granulocytes. They rapidly swell upon stimulation with the chemoattractant fMLP. This swelling can be partially inhibited by NHE1 blockade, or by replacement of extracellular Na+ (481, 490). While NHE1 blockade does not affect chemokinesis, provided an intracellular acidification is prevented (217), it clearly inhibits fMLP-stimulated chemotaxis of neutrophils (481, 490). Thus stimulated migration in neutrophils requires an active NHE1, while random migration of these cells can be maintained by compensatory mechanisms independent of NHE1 activity. In MDCK-F cells, Na+-HCO3 cotransport (NBC) can partially substitute for the function of NHE1 (524). Inhibition of neutrophil chemotaxis by NHE1 inhibitors can be overcome with hyposmolar media (481, 490). These findings lend further support to the osmoregulatory role of NHE1 in cell migration.

Figure 11.

Polarization of NHE1 in migrating cells. A: topical application of the NHE inhibitor EIPA only impairs migration of transformed renal epithelial MDCK-F cells when it is selectively applied to the cell front. [From Klein et al. (287).] B and C: extracellular pH measurements (pHem) at the surface of the outer leaflet of the plasma membrane of human melanoma cells. Inset indicates the position of the pH indicators (DHPE, 1,2-dihexaecanoyl-sn-glycero-3-phosphoethanolamine; WGA, wheat germ agglutinin). pHem is more acidic at the cell front than at the rear part of the cell. The pericellular pHem gradient largely depends on NHE1 activity. Upon washout of the NHE1 blocker Hoe642, pHem changes much more at the front than at the rear part of the cell. [Modified from Schwab, Stock, and co-workers (299, 554, 556, 561), with permission from S. Karger AG Basel.] D: NHE1 activity also generates an intracellular pHi gradient that is complementary to the extracellular pHem gradient. pHi is more alkaline at the cell front of melanoma cells (MV3, B16), fibroblasts (NIH3T3), and MDCK-F cells. [Modified from Martin et al. (361).] E: morphological correlate of the NHE1-dependent extra- and intracellular pHem and pHi gradients. Immunohistochemistry reveals the concentration of NHE1 at the leading edge of the lamellipodium of fibroblasts (arrows). [From Grinstein et al. (197), with permission from Nature Publishing Group.]

The polarized distribution of NHE1 in migrating cells has a profound impact on local pH homeostasis (FIGURE 11). Despite the fact that glycolytic or ATP generating enzymes, that also produce protons, are enriched within the lamellipodium (260, 302, 405, 586), the pHi of migrating cells is more alkaline at the cell front than in their rear part (361). Acid extrusion at the cell front, which is mediated by NHE1, leads to a complementary extracellular pH gradient at the surface with the more acidic pH at the cell front (556). A similar NHE1-dependent intracellular pH gradient is observed during neurite outgrowth. pHi is more alkaline in growth cones than in the cell soma (543). The pericellular pH nanoenvironment at the cell surface is maintained by the glycocalyx (299). Both intracellular and extracellular NHE1-dependent pH gradients are required for cell migration (557) as described in more detail in section VA. They promote the outgrowth of the lamellipodium by driving actin polymerization and support the formation (at the cell front) and release (at the rear part) of focal adhesion contacts to the extracellular matrix mediated by integrins (433, 623) (see also sect. IIB).

NHE1 and integrins are often colocalized at focal adhesion sites (453). Functionally there is a reciprocal relationship between integrins and NHE1. Integrins trigger signaling cascades including the RhoA-ROCK cascade when binding proteins of the ECM such as fibronectin that eventually lead to increased NHE1 activity (4, 64, 247, 529, 587, 588). Focal adhesion kinase (FAK) is one of the first proteins that becomes phosphorylated in response to adhesion and triggers a signaling cascade promoting cell migration (387). Its phosphorylation relies on the presence of NHE1 (587). Possibly, this is due to a direct effect of protons extruded by NHE1. An NHE1-mediated acidification of the cell surface activates neighboring integrins, i.e., causes an opening of the integrin headpiece and thereby increases the affinity of integrins for their ECM ligands (141, 321, 436, 561). Consequently, adhesion is stronger in an acid environment than in an alkaline one, and migration is optimal at an intermediate adhesion strength (436, 554). Finally, NHE1 regulates the secretion of cell matrix proteins such as fibronectin (271).

NHE1 activity also affects the expression of genes that contribute to cell migration. The comparison of mouse muscle fibroblasts (LAP cells) expressing either wt NHE1 or a transport-deficient NHE1 revealed that NHE1 activity is associated with the regulation of a number of genes that are involved in the organization of the cytoskeleton, cell adhesion, and extracellular matrix assembly. Loss of NHE1 activity leads to downregulation of the expression of myosin regulatory light chain, MMP-9, and merlin, a tumor suppressor protein and member of the ERM family. In contrast, the expression of the actin severing and capping protein gelsolin and that of a number of microtubule-related genes such as Clip170, KIF4, kinesin light chain, and a dynein heavy chain, are upregulated (462).

NHE1-dependent signaling cascades underlying directional cell migration also include the primary cilium, a microtubule-based organelle emerging from the mother centriole (500). The ciliary membrane contains the PDGFRα receptor for platelet-derived growth factor-AA (PDGF-AA) (93, 512), and binding of PDGF-AA to its receptor stimulates NHE1 activity (642). When ciliary assembly in NIH3T3 fibroblasts and mouse embryonic fibroblasts is disturbed or NHE1 is inhibited, chemotaxis towards PDGF-AA is absent, leading to a dramatic delay in wound healing, in vitro and in vivo (511, 514).

NHE1 also elicits nonconductive functions in cell migration. NHE1 is a plasma membrane anchor for cortical actin filaments (26, 119) and a scaffold for signaling complexes (271, 632). In fibroblasts NHE1 attaches the cortical actin cytoskeleton to the leading edge of the lamellipodia through a direct association of its COOH-terminal domain with the ERM (ezrin, radixin, moesin) family of actin-binding proteins that are also concentrated at the cell front. ERM proteins are substrates for Rho kinase 1 (ROCK1) (219) and Nck-interacting kinase (27), both of which stimulate NHE1 activity (66, 587). Thus the colocalization of ERM proteins and NHE1 appears to facilitate this reciprocal modulation of cytoskeletal dynamics required for cell migration.

2. Monocarboxylate transporter

The monocarboxylate transporters (MCTs) mediate the H+-coupled efflux of lactate, pyruvate, butyrate, acetate, and propionate across the cell membrane (203). MCT1 to -4 are widely expressed in many types of cancer (e.g., Refs. 295, 296, 598). They enable cancer cells to cope with the hypoxic conditions they are often exposed to. Silencing of an ancillary MCT subunit, CD147, inhibits MCT1 and MCT4 function and reduces the malignant potential of pancreatic cancer cells (516). Along the same lines, silencing of MCT4 impairs migration of the breast cancer cell line MDA-MB-231 (168). Since MCT4 colocalizes with β1-integrin at the leading edges of the lamellipodia of migrating retinal pigment epithelial (ARPE-19) and MDCK cells (167) and since it is involved in regulating the cellular pH homeostasis, it is conceivable that MCT4 activity contributes to the special pH nanoenvironment at sites of focal contacts (436, 556, 557) (see also sect. IIID). An alternative mechanism proposes that lactate is released by tumor cells (via MCT4), enters endothelial cells through MCT1, and then triggers an autocrine NF-κB/IL-8 pathway driving endothelial cell migration and tube formation (605). In addition, MCTs also affect cell migration by “nonconductive” properties of their ancillary subunit CD147 which promotes cell invasion by inducing matrix-metalloproteases (MMPs) such as MMP1–3, 9, 11, 14, and 15 (641).

B. Na+-K+-2Cl Cotransporter

The Na+-K+-2Cl cotransporter (NKCC1) contributes to the regulation of cell volume and homeostasis of intracellular Cl concentration ([Cl]i). NKCC1 was one of the first transport proteins whose involvement in cell migration was identified. Migration of keratocytes from Xenopus tadpoles (32) and MDCK-F cells (527) is impaired in the presence of the blockers piretanide, furosemide, and bumetanide. It was hypothesized that Na+-K+-2Cl cotransport supports the protrusion of the lamellipodium by inducing localized solute and water uptake, i.e., swelling at the cell front (527). Two recent studies indeed revealed the enrichment of the NKCC1 protein at the leading edge of the lamellipodium of protruding PC12D cell neurite growth cones (404) or of glioma (D54-MG and U87-MG) cells (200). NKCC1 activity stimulated by with-no-lysine kinase 3 (WNK3) (199) not only leads to cell swelling but also to the intracellular accumulation of Cl, which is the prerequisite for channel-mediated Cl efflux and volume loss. Interestingly, WNK3-stimulated NKCC1 activity is required for glioma cell motility only in three-dimensional assays. This points out that a fine-tuned, NKCC1-dependent alternation between very local osmotic swelling and shrinkage is needed to guide invadopodial protrusions through pores or a meshwork of extracellular matrix proteins (620).

C. Na+-K+-ATPase

The Na+-K+-ATPase is composed of a catalytic α and a regulatory β subunit and an FXYD protein (173). The α subunit mediates transport of Na+ and K+, and it binds inhibitory cardiotonic steroids such as ouabain, digoxin, digitoxin, and cardenolides (320, 336). The β subunit is required for normal Na+-K+-ATPase activity. Depending on their expression, FXYD proteins regulate the kinetic properties of the pump in a tissue-specific way (173). Not all Na+-K+-ATPase molecules function as an ion-translocating pump. There is also a nonpumping pool of the Na+-K+-ATPase (332) whose signal-transducing function is activated by cardiotonic steroids (330). It is restricted to caveolae, does not depend on changes of the intracellular Na+ and K+ concentrations, and involves among others the tyrosine kinase Src, the PI3K/Akt pathway, and transactivation of the EGF receptor (290, 614).

Initially, the inhibition of keratocyte migration following blockade of the Na+-K+-ATPase with ouabain was ascribed to the blockade of the pump function (32). However, studies from the last 10 years revealed that the situation is far more complex in that the nonconductive properties of the Na+-K+-ATPase come into play too. In glioblastoma and non-small-cell lung cancer, the α1 subunit of the Na+-K+-ATPase is upregulated and colocalized with caveolin-1 in lamellipodia (320, 382). Inhibition of the Na+-K+-ATPase with the cardenolide UNBS1450 reduces migration in vitro. Part of this effect was ascribed to an UNBS1450-induced intracellular ATP depletion and consecutive disorganization of the actin cytoskeleton. The effects of UNBS1450 are independent of the pump function of the Na+-K+-ATPase since [Ca2+]i and [Na+]i remain unchanged (320). In cytotrophoblast cells, inhibition of EGF-induced chemotaxis/invasion by the endogenous cardiotonic steroid marinobufagenin correlates with reduced ERK1/2 phosphorylation (597). Regeneration of newt retina and lens, a process involving migration of dedifferentiated progenitor cells, coincides with the upregulation of the α1 subunit of the Na+-K+-ATPase (606).

The β subunit of the Na+-K+-ATPase is downregulated in several invasive cancers and in tumor cell lines such as in renal clear cell carcinoma (467) and Moloney sarcoma virus (MSV)-transformed MDCK cells, respectively (468). The causal link between Na+-K+-ATPase β-subunit and cell motility was revealed by showing that restitution of the β-subunit strongly reduces migration of MSV-transformed MDCK cells through interactions with the GTPase Rac1 (468) and proteins involved in PI3K signaling (24). The β2 isoform of the Na+-K+-ATPase, also known as “adhesion molecule on glia” (AMOG), has been linked originally to neuron-glia adhesion and cerebellar granule cell migration (14). Its downregulation is linked to increased migration and invasion of gliomas since its reexpression increases adhesion and reduces migration (532).

FXYD5 (also named dysadherin) has been correlated with increased tumor progression and invasiveness. Capan-1 pancreatic cancer cells, expressing only a small amount of endogenous FXYD5, can be transferred to a highly mobile cell line with increased metastatic potential when they are transfected with FXYD5 (537). Similarly, increased expression of FXYD5 in airway epithelial (385) or renal cell carcinoma cells (517) increases migration.


In this section we discuss two pathophysiological examples of great clinical relevance. Research from the last decade clearly revealed that targeting ion transport proteins represents a promising novel therapeutic strategy for the treatment of chronic inflammatory (autoimmune) diseases or cancer. The beneficial effects of targeting ion transport proteins are at least in part due to their impact on the migratory activity of the respective immune and tumor cells.

A. Immune Defense and Inflammation

Migration of cells of the innate and adaptive immune system is a prerequisite for their proper function. Invading pathogens can only be cleared/phagocytosed when the respective immune cells have been recruited and have migrated to such a site of infection. Thus T cells display robust motility in the lymph node, which was suggested to constitute a search strategy of T cells to locate and sample dendritic cells for antigens (384). Similarly, T-cell activation following antigen presentation (383) and immune surveying of the brain by microglial cells (410) are highly dynamic processes. Stated differently, when viewed from a hierarchical perspective, migration of immune cells occupies a superior position compared with other properties of immune cells. Accordingly, primary immunodeficiencies that are often associated with an inability of immune cells to migrate, result in increased susceptibility to (life threatening) infections and inability to repair tissue damage (413, 582).

Ca2+ signaling plays a crucial role in many additional facets of the immune cell function such as phagocytosis, cytokine secretion, and proliferation. Ca2+ signaling in turn strongly relies on the fine-tuned activity of ion channels and transporters regulating the intracellular Ca2+ homeostasis. The important contribution of ion channels, in particular, KV1.3, KCa3.1, and CRAC channels (composed of Orai1 and STIM1), has been studied in great detail in lymphocytes. In fact, one of the earliest studies describing the involvement of K+ channels in cell proliferation was performed in T cells (114). KV1.3, KCa3.1, Orai1, and STIM1 cluster at the immunological synapse when T cells make contact with an antigen-presenting cell. These channels constitute a functional network (transportome) that has a profound impact on intracellular Ca2+ signaling and thereby regulates not only migration but also drives gene expression, proliferation, and cytokine secretion (IL-2) of T cells (see Ref. 59 for review). The physiological significance of ion channel signaling is underlined by the fact that channel inhibition is highly effective in preclinical disease models such as T cell-mediated colitis (124), inflammatory skin diseases (406), asthma (49), type 1 diabetes mellitus or rheumatoid arthritis (29), or EAE (349). It is this integrated function of a distinct set of ion channels not restricted to a selective role in T-cell migration that makes them such promising therapeutic targets.

One further example illustrates that similar considerations also apply to other cells of the immune system and to other ion channels or transporters. Voltage-gated proton channels (VSOP/Hv1) play an important role in activated neutrophils. They are required for the generation of reactive oxygen species during the respiratory burst (423) and maintenance of pH homeostasis during phagocytosis (397). By sustaining the cell membrane potential during the respiratory burst, they also allow continued Ca2+ influx driving cytoskeletal dynamics during migration (144).

B. Tumor Metastasis

Hypoxia (Po2 <2.5 mmHg), usually coinciding with an acidic extracellular pH, is a hallmark of solid tumors (603). There is increasing evidence that an acidic interstitial pHe of tumors gives a selective advantage for tumor progression and metastasis. Thus lactate promotes tumor cell migration but inhibits that of monocytes (180). It drives large changes in gene expression and promotes increased invasion and metastasis (67). This involves alterations of the ECM compartment through pH-dependent upregulation of protease secretion/activation [e.g., matrix metalloproteinases (MMPs), cathepsins], impaired function of the antitumoral immune system (67, 389), or altered growth factor/cytokine secretion (164, 635). In recent years this concept was further refined by showing that pH regulatory transport proteins such as the Na+/H+ exchanger NHE1 are tightly linked to the tumor-promoting effects of extracellular acidity. Examples of NHE1-dependent processes relevant for tumor progression include tumor cell-extracellular matrix interactions (436, 554), migration and invasion (476, 554), MMP activity (57), secretion of ECM proteins (271), and chemoresistance (312). In this context, membrane-bound extracellular-facing carbonic anhydrase (CA) isoforms also need to be considered. CAIX and CAXII have been identified as markers of aggressive carcinomas (87, 569). Tumor spheroids that express CAIX protect their pHi efficiently at the expense of a more acidic pHe. Thus expression of CAIX in conjunction with the relevant acid-base transporters (e.g., Na+-HCO3 cotransport) would be highly beneficial to cancer cells, by maintaining a pHi that is favorable for their growth while creating at the same time a hostile acidic environment for non-tumor cells (87). Tumors lacking extracellular CA isoforms may resort to transmembrane lactic acid secretion by MCT and/or Na+/H+ exchange for protecting their intracellular and acidifying their extracellular milieu (67, 195, 557).

Recent discoveries of the function and expression of ion channels and transporters in tumors clearly indicate that tumor-associated alterations in ion homeostasis affect essentially all transport proteins of cancer cells (178, 439, 460). This generalized alteration of members of all major families of transport proteins in cancer cells is best described with the concept of a tumor cell transportome. The impact of an altered transportome is of crucial importance for almost all hallmarks of cancer. Every cell phenotype is characterized by a specific “ionic signature” depending on the kinetics, magnitude, and subcellular localization of ion signals. Due to the quantitative and functional variations of ion channel activity in cancer, this disease could then be classified as a “channelopathy.” The groundwork for this novel concept was laid in the 1990s by recognizing that K+ channels (16, 491, 527), Cl channels (549), and voltage-gated Na+ channels (196) are required for tumor cell proliferation, migration, and invasion and are activated by integrins during cell adhesion. Another major breakthrough was the discovery that the overexpression of a voltage-gated K+ channel, KV10.1, confers tumorigenicity on non-cancer cells (438). Subsequently, members of the TRP channel family entered the field as important players in cancer cell pathophysiology. In fact, TRPM1 channels were originally identified as tumor suppressors, since their expression is high in benign nevi and low in metastatic melanoma (138).

Several important principles emerged from the studies on ion transport in cancer. It is a general phenomenon that the expression of proteins involved in ion transport markedly differs between normal and the corresponding tumor cells. Mutations of channels or transporters are the exception (266). The correlation of ion channel/transporter expression in tumor samples with the clinical outcome of the respective cancer patient provides diagnostic and prognostic information that can be exploited to improve patients' quality of life. Many of the transport proteins that are involved in tumor cell migration are crucial for other hallmarks of cancer as well (460). Tumor cell proliferation and apoptosis are such examples. The dual inhibition of tumor cell migration and proliferation by blocking, for example, KCa3.1 (151, 523) or KV10.1 channels (6) would of course be very desirable. On the other hand, inhibiting migration and apoptosis by KV1.3 (570) or VRAC blockade (513) would not be advantageous, since it would favor the persistance of tumor cells. These examples indicate that the multiplicity of cancer hallmarks controlled by a given transport protein requires careful consideration of its therapeutic potential. However, such functional constraints of therapeutic options can be overcome when cancer-specifically upregulated ion transport proteins are exploited for targeted therapies. Alternatively, the combined inhibition of several transport proteins involved in complementary mechanisms may be beneficial.


Studying the function of ion channels and transporters in cell migration is still a young field in cellular physiology so that the number of studies mechanistically linking transport proteins to the cytoskeletal motor of migration is still limited. Nonetheless, the field is rapidly expanding and produces an ever-growing list of ion transport proteins contributing to cell migration. Studies from the last ∼15 years provided firm proof of the concept that transport proteins in the plasma membrane and possibly also in intracellular membranes constitute integral components of the cellular migration machinery without which the cytoskeletal migration motor or cell adhesion proteins do not function properly. That members of all major families of transport proteins have been linked to migration of many different cell types is remarkable for two reasons. First, it points towards the universal nature of the involvement of ion transport proteins in cell migration. Second, it lends strong support to the concept of a migration-related transportome. Importantly, “transportome” implies that in each cell a whole array of different ion channels and transporters is involved in controlling cell migration. Ion channels and transporters control migration in multiple ways. These include their core competencies as transport proteins as well as nonconductive properties. The transport function is also exploited in the regulation of the intracellular second messengers Ca2+ and pH. In addition, transport proteins and receptor-activated signaling cascades modulate each other reciprocally. We envision that transport proteins impact on intracellular signaling in a very specialized and localized way that is exemplified by TRPM2 channels that are only required for neutrophil chemotaxis towards fMLP but not CXCL2 (640). However, so far we are not yet in a position to draw a conclusive picture on the role of the transportome in cell migration and in particular in disease states that are associated with “too much” migration.

The overwhelming majority of studies on members of the transportome in cell migration investigates only the function of one transport protein in one particular cell type at a time. Nonetheless, the multiplicity of these studies allows some generalization so that it can be taken for granted that NHE1 and aquaporins play important roles in cell migration. However, this approach, as valid as it was in initially cataloging the relevant members of the transportome, does not come close to reflecting the physiological complexity. NHE1 may serve as an example to explain the future needs in the field to validate the integrated function of individual transportome members. NHE1 is crucial in maintaining intra- and extracellular pH homeostasis, and its role in cell migration also depends on the pH micorenvironment of the cell. Yet, so far it is entirely unknown how NHE1 and its impact on pH regulation affect the role of other pH-sensitive ion channels in cell migration such as KCa3.1 (445). To the best of our knowledge, all studies revealing a role of KCa3.1 in cell migration were performed at ∼pH 7.4. However, in tumor cells that usually express KCa3.1 channels, NHE1 and other pH regulatory proteins may produce a pH microenvironment that is not necessarily permissive for KCa3.1 activity. This example clearly indicates that future studies need to address the mutual impact of members of the transportome on each other and the resulting consequences for cell migration derived thereof. The demonstration of the functional interdependence between KCC2 expression, GABA receptors, and voltage-gated Ca2+ channels in cortical interneuron migration is a good example for such studies (43). Thus a more integrative view on the transportome in cell migration combined with the use of the appropriate genetic mouse models and modern intravital or live-cell imaging technologies will be the next steps to further advance this new and exciting field in cellular physiology.


Our work was supported by Deutsche Forschungsgemeinschaft, European Commission (ITN “IonTraC”), IZKF Münster, and IMF Münster.


No conflicts of interest, financial or otherwise, are declared by the authors.


We thank all present and past members of the cell migration laboratory without whose enthusiastic commitment much of the work from our laboratory would not have been possible. We thank Thomas Fortmann for preparing FIGURE 6A.

Address for reprint requests and other correspondence: A. Schwab, Institut für Physiologie II, Robert-Koch-Str. 27b, D-48149 Münster, Germany (e-mail: aschwab{at}


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84a.
  85. 84b.
  86. 85.
  87. 86.
  88. 87.
  89. 88.
  90. 89.
  91. 90.
  92. 91.
  93. 92.
  94. 93.
  95. 94.
  96. 95.
  97. 96.
  98. 97.
  99. 98.
  100. 99.
  101. 100.
  102. 101.
  103. 102.
  104. 103.
  105. 104.
  106. 105.
  107. 106.
  108. 107.
  109. 108.
  110. 109.
  111. 110.
  112. 111.
  113. 112.
  114. 113.
  115. 114.
  116. 115.
  117. 116.
  118. 117.
  119. 118.
  120. 119.
  121. 120.
  122. 121.
  123. 122.
  124. 123.
  125. 124.
  126. 125.
  127. 126.
  128. 127.
  129. 128.
  130. 129.
  131. 130.
  132. 131.
  133. 132.
  134. 133.
  135. 134.
  136. 135.
  137. 136.
  138. 137.
  139. 138.
  140. 139.
  141. 140.
  142. 141.
  143. 142.
  144. 143.
  145. 144.
  146. 145.
  147. 146.
  148. 147.
  149. 148.
  150. 149.
  151. 150.
  152. 151.
  153. 152.
  154. 153.
  155. 154.
  156. 155.
  157. 156.
  158. 157.
  159. 158.
  160. 159.
  161. 160.
  162. 161.
  163. 162.
  164. 163.
  165. 164.
  166. 165.
  167. 166.
  168. 167.
  169. 168.