The multifaceted role of aquaporins in physiological cell migration

Abstract

Cell migration is an essential process that underlies many physiological processes, including the immune response, organogenesis in the embryo, and angiogenesis, as well as pathological processes such as cancer metastasis. Cells have at their disposal a variety of migratory behaviors and mechanisms that seem to be specific to cell type and the microenvironment. Research over the past two decades has elucidated the water channel protein family of aquaporins (AQPs) as a regulator of many cell migration-related processes, from physical phenomena to biological signaling pathways. The roles that AQPs play in cell migration are both cell type- and isoform-specific; thus, a large swath of information has accumulated as researchers seek to identify the responses across these distinct variables. There does not seem to be a universal role that AQPs play in cell migration; the complex interplay between AQPs and cell volume management, signaling pathway activation, and in a few identified circumstances, gene expression regulation, has shown the intricate, and perhaps paradoxical, role of AQPs in cell migration. The objective of this review is to provide an organized and integrated collection of recent work that has elucidated the many mechanisms by which AQPs regulate cell migration.

NEW & NOTEWORTHY Research has elucidated the water channel protein family of aquaporins (AQPs) as a regulator of many cell migration-related processes, from physical phenomena to biological signaling pathways. The roles that AQPs play in cell migration are both cell type- and isoform-specific; thus, a large swath of information has accumulated as researchers seek to identify the responses across these distinct variables. This review compiles insights into the recent findings linking AQPs to physiological cell migration.

INTRODUCTION TO CELL MIGRATION

Cell migration is an essential process that underlies many physiological mechanisms, including the immune response, wound healing, organogenesis in the embryo, and angiogenesis. Cell migration can also be a hallmark of disease states such as cancer metastasis (1). Our understanding of cell migration, from a phenomenological and mechanistic perspective, has stemmed from observations of different cell types migrating and invading in various natural or engineered environments. Up until the past decade, much of our knowledge about cell migration spawned primarily from experiments on 2-D surfaces, where we have learned that migration-promoting structures polarize to the leading edge of the cell to extend membrane protrusions in the direction of migration (2–4). Membrane protrusions, generated via actin reorganization, consist of spike-like filopodia, used to sense the local environment, or broad lamellipodia, which provide the foundation for cells to move forward (5). These protrusions then adhere to the extracellular matrix (ECM) through integrins, which link intracellularly to the actin cytoskeleton. This ECM-cytoskeletal connection provides the cell with the traction necessary to pull itself forward (6, 7). This “pulling forward” or contraction is mediated by the contractile protein myosin IIa’s interaction with actin. Then, as the cell pulls forward, there is a retraction at the rear of the cell, also mediated by myosin IIa, in concert with disassembly of the trailing edge adhesions and cytoskeleton (8).

As opposed to conventionally understood mechanisms of 2-D migration, single cells migrating in 3-D environments can adopt a variety of migration mechanisms, some of which are highly distinct from those found in 2-D systems. During 3-D migration, cells are exposed to a ubiquitous abundance of complex cues, which the cell can internalize to constantly regulate its movement. Broadly speaking, 3-D migration mechanisms can be split into categories: ameboid, lobopodial, and mesenchymal-like migration. Cells in 3-D environments can also undergo collective migration in large clusters as seen during wound closure or neural crest cell migration in healthy states (9). Developing a better understanding of how cells migrate in these complex environments can aid the development or identification of methods to perturb it for therapeutic benefits by way of genetic, pharmacologic, or materials-based interventions. Although these various 2-D and 3-D cell migration mechanisms have been recently reviewed in detail (10, 11), in this review, we focus on the role of aquaporins in these various modes of cell migration.

Aquaporins (AQPs) are a family of transmembrane proteins whose conventionally known function is facilitating osmotically driven water transport across cell plasma membranes (12). AQPs are known to play a pivotal role in maintaining water homeostasis and solute transfer across the cell membrane. In mammals, there are 13 identified isoforms of the AQP family (AQP0–AQP12). Among these, AQP3, AQP7, AQP9, and AQP10 are also termed “aquaglyceroporins” due to their permeability to both glycerol and water (13–17). AQP3, AQP5, AQP8, and AQP9 have also been implicated in the movement of hydrogen peroxide (H₂O₂) into the cell (18–22). The movement of water, glycerol, and H₂O₂ by AQPs can influence many cellular behaviors, including an array of complex processes that drive cell migration, as we will discuss in this review. Figure 1 provides an overview of which migratory cell types express various isoforms of AQPs in physiological and pathological contexts.

Figure 1.

An overview of AQP isoforms found in various migratory cell types in healthy (A) and pathological contexts (B). AQP, aquaporin. Figure was created using BioRender.com.

Earliest Work Relating Aquaporins to Cell Migration

AQPs were first proposed as a regulator of cell migration in 2002, when Loitto et al. discovered that AQP9 expression was essential for lamellipodia extension and stabilization in neutrophils (23) (Fig. 2). Later, in 2005, Saadoun et al. discovered that AQP1-deficient endothelial cells display impaired migration and abnormal vessel formation in vitro (24). Motivated by these findings, this same group published the first review paper about AQPs and cell migration 3 years after they conducted their initial migration studies (25). This early review primarily described their group’s studies that identified the roles of AQP1 in endothelial cell migration, AQP1 in tumor metastasis, and AQP4 in glial scar formation (24, 26, 27). This review also proposed that the osmotic water flow across the cell membrane, facilitated by AQPs, is an essential step to the formation of migration-dependent protrusions. This proposed mechanism has become essential in our understanding of the role AQPs play in cell migration and has provided a foundation for future experiments. We note that this early review also proposed a mechanism by which AQPs are involved in 3-D migration, however, that mechanism has not yet been confirmed. Since this first review detailing the role of AQPs in cell migration was published, numerous studies have furthered this link between AQPs and cell migration, some confirming movement via the traditional water flux (28, 29). However, others have found that AQPs can also regulate 1) migration by transport of other molecules such as glycerol or H₂O₂, 2) actin polymerization, 3) interactions with other membrane-bound proteins, 4) regulation of migratory direction, and 5) activation of cellular signaling pathways that induce changes in migratory phenotypes (Fig. 2). This review will address the role AQPs play in these mechanisms and will discuss potential directions for future work. Although here we focus on cell migration in normal physiology, we note that our recent review article focuses on how AQPs influence cell migration behaviors across disease states (30).

Figure 2.

Timeline detailing important milestones discovered to enhance the field’s understanding regarding the roles AQPs play in cell migration. AQP, aquaporin. Figure was created using BioRender.com.

AQUAPORINS AND THEIR ROLE IN CELL MIGRATION MECHANISMS

AQPs have been implicated in multiple mechanisms of cell migration, as we discuss in this section. Below, we reveal the dichotomy of AQPs: they are essential for cell migration, yet they do not play a universal role across tissue types, cell lines, or even disease states, likely because their expression and function are dynamic and dependent on a wide array of factors (e.g., cell type, signaling pathways, mechanical environment, pH levels, etc.). In the subsections below, we highlight the multifunctional roles of AQP isoforms across various modes and behaviors of cell migration.

Role of Aquaporins in Lamellipodia

Lamellipodia are sheet-like projections that contain dense cytoskeleton networks and are located at the leading edge of motile cells (5). Lamellipodia are typical hallmarks of 2-D migration; the cell uses these long and flat structures to adhere to their underlying substrates. However, lamellipodia have also been observed during in vivo migratory events such as neural crest migration (31). Motile cells deploy these projections to probe their surrounding environments and determine the direction of movement while providing a stable connection to drive the cell forward across the ECM.

Conventionally, it was believed that the polymerization of actin at the leading edge of the cell applies pressure to the cell membrane, causing it to extend outward and form the lamellipodia (32); however, this theory was later revised (28). Although actin still plays a large role in lamellipodium geometry and function, it is unable to “push against” the membrane as it polymerizes; rather, it can only fill a pre-existing space (33, 34). This phenomenon is captured in the Brownian Rachet Model, where small Brownian motion-induced undulations of the cell membrane provide the space for actin to polymerize. However, the size and formation time of these undulations are orders of magnitude smaller and slower than the formation of migratory protrusions, while occurring independently of the direction of cell migration (35). Thus, lamellipodia formation involves a spatially directed rapid change in cell volume, which is likely induced by transmembrane water movement; it has been confirmed that this water movement is, in part, facilitated by AQPs.

Building upon Loitto et al.’s early discovery showing that the influx of water is essential for the formation of neutrophil lamellipodia, DiGiusto et al. found that AQP2 forms a larger transporter super structure by localizing with transient receptor potential vanilloid-type 4 (TRPV4) cation channel and sodium/hydrogen exchanger 1 (NHE1) within rat renal cortical collecting duct type (RCCD) cell lamellipodia (23, 28) (Fig. 3). Working together, these transporters provide the necessary exchange of ions and water flux necessary for the formation of lamellipodia and thus cell migration. This super structure is also capable of regulating hyperlocalized NHE1-mediated intracellular pH levels leading to favorable polymerization of F-actin (28). Furthermore, kidney proximal tubule cells with reduced AQP1 exhibit decreased water permeability in lamellipodia-like protrusions, resulting in the reduction of cell motility (36). In colonic epithelial cells, depletion of AQP3 leads to a decrease in the number and sizes of actin-containing lamellipodia (37). Meanwhile, AQP5 overexpression in Madin-Darby Canine Kidney (MDCK) cells promotes actin stress fiber formation (with more dorsal and ventral localization compared with controls) and lamellipodia dynamics, along with a concomitant decrease in cell circularity due to more elongated and protrusive morphologies(38).

Figure 3.

AQP colocalization with migratory-related proteins and downstream signaling. Putative pathways are determined via direct localization or activation of proteins via AQPs or the signals they transport. The speculated downstream pathways have been altered by AQP inhibition or knockdown, yet the direct relationship is not yet defined. AQP, aquaporin. Figure was created using BioRender.com.

Role of Aquaporins in Filopodia

Filopodia are thin, actin-rich protrusions at the leading edge of the cells that, like lamellipodia, also allow the cell to sense its surrounding environment as it migrates. Filopodia are more commonly seen in 3-D cell migration, as their thin and fiber-like structure allows for them to better investigate their more complex surrounding environments (39). In migrating HEK-293 cells, AQP9 accumulates at the leading edge of the cell and facilitates the formation of small blebs in the plasma membrane (29). These blebs continue to expand, forming the filopodia, which is initially devoid of actin. Following formation of these membrane extensions, actin polymerizes in the newly existing space to stabilize the new filopodia. Phosphorylation of AQP9 induces filopodia formation, increased membrane protrusions, and induced actin polymerization via activation of Cdc42 (40); this concept will be further discussed below. Meanwhile, motile cell clusters during neural crest cell migration require AQP1 for filopodia extension, stabilization, and occurrence (41). Inhibition and suppression of AQP1 leads to the random extension and random movements of the cluster’s filopodia.

Role of Aquaporins in Bleb-Like Protrusions/Ameboid Migration

Ameboid migration is a type of cell migration that occurs independently of focal adhesions and proteases. This method of migration is often driven by membrane blebbing polarized to the leading edge of the cell. Following the formation of the bleb, the structure increases in volume, displacing cytoplasm toward the direction of migration (42), thus moving its center of mass forward (43). As blebbing requires a rapid change in volume for its formation, at both the leading and trailing edges, much attention has been placed on elucidating AQPs’ function on this migratory behavior.

AQP1 overexpression, while working in conjunction with NHE1, induces blebbing and increases bleb life spans; this phenomenon occurs across multiple cell lines (44). There is also an influx of Ca²⁺ which causes the formation of migration-related blebs (45). When cells are exposed to an alkaline pH, they display accelerated motility and invasive properties, along with a rapid dynamic blebbing, which is a result of the re-polarization of AQP3 from the nucleus to the blebs at the leading edge of the cell. It was hypothesized that this AQP3 translocation provided the rapid and substantial influx of water to induce the migratory-related blebs (46).

It is also worthwhile to mention that one publication has claimed that AQPs do not play a role in cell blebbing formation. This study only investigated two AQP isoforms that were not strongly expressed in the chosen cell type; furthermore, the cell types used were not of human origin. These limitations could account for the variation in AQP functions found (47).

Role of Aquaporins in the Osmotic Engine Model

The Osmotic Engine Model is an empirically and mathematically supported behavior by which cells migrate through confined spaces in vitro via a mechanism that relies on AQPs, but not necessarily on acto-myosin contractility (48). Meanwhile, amoeboid migration is dependent not only on the formation of blebs during migration, as previously discussed, but also on high acto-myosin contractility (8, 49). We have found that multiple cell types migrating in confined spaces can continue migrating even after the inhibition of actin polymerization via latrunculin A or inhibition of myosin II contractility via blebbistatin (48, 49). Following these discoveries, we proposed a new paradigm of cell migration, which we called the Osmotic Engine Model (48). Notably, this was the first study to propose a mechanism for cell migration that requires ion channel and AQP polarization without the need for actin polymerization or myosin contractility (Fig. 4).

Figure 4.

A graphical representation of the osmotic engine model. A cell is migrating in a confining channel and uses fluxes (J) of ions and water at the front and back of the cell to propel itself forward. The proteins below the fluxes have been implicated in the movement of water or ions at the front or back of the confined cell. AQP, aquaporin. Figure was created using BioRender.com.

Prior to our establishment of this model, the theoretical concept of cell movement by water transport was originally proposed by Jaeger et al. in 1999 (50). However, the model by Jaeger et al. neglected to include the transport of ions (50). The Osmotic Engine Model proposes that cells can propel themselves forward in confined spaces by manipulating water and ion fluxes. Importantly, key theoretical predictions of the Osmotic Engine Model were supported by experimental observations in migratory cells. First, we found that knocking down AQP5 significantly reduced the velocity of cells migrating through narrow channels. Next, we found that AQP5 and NHE1 polarize to the leading and lagging edges of migrating cells in these microchannels, with a higher expression of the AQP channels and NHE1 exchangers at the leading edge; the resulting net inflow and net outflow of water lead to a net cell displacement as predicted by the model. Third, the Osmotic Engine Model predicted that osmotic shocks at the leading or trailing end of cells migrating in narrow channels would alter cell migration velocity (including the speed and direction of migration). Indeed, inducing a hypotonic shock at the leading edge or hypertonic shock at the trailing edge led to a counterintuitive reversal of the cells’ migration direction within 3 min after the osmotic shock, followed by later repolarization of NHE1 to the new leading edge within 30–60 min after the osmotic shock. Intriguingly, the Osmotic Engine Model predicted this phenomenon: that an osmotic shock alone could reverse cell migration direction even in the absence of polarization of ion exchangers or aquaporins to the cell’s leading edge. We refer the reader to the original article for details and explanations regarding the theoretical model and experimental results (48).

The Osmotic Engine Model was further supported by another laboratory showing that confined migrating cells polarized AQP4 and SWELL1 to the lagging edge of the cell. This co-localization at the lagging edge allowed for localized cellular volumetric increases and reductions to induce efficient confined cell migration. This effect is postulated to occur as a result of Cl⁻ outflux mediated by SWELL1 (51). Another laboratory developed a confining microchannel device allowing for the electrical manipulation of osmolarity within the samples. Using this device, the group found that confined cell speed is correlated with AQP4 expression (52). This relationship was confirmed across more than 20 cell lines (52). The development and conformation of the Osmotic Engine Model poses three important points for this review: 1) AQPs and ion channels are the two essential proteins for this model. 2) This model adds to the list of how AQPs are essential to every known mechanism of cell migration. 3) Much of the work done to understand cell migration has been conducted in environments that do not effectively replicate in vivo conditions. More work with complex in vitro environments must be conducted, like that of the Osmotic Engine Model, to develop a more holistic view of the cell’s ability to effectively migrate in vivo. However, the model itself still poses several key gaps in understanding that should be noted: 1) Cell migration with the absence of actin polymerization or myosin contractility has yet to be observed in vivo. 2) There remains a lack of information regarding the cellular processes at play, specifically at the lagging edge of the cell, in this model. 3) Other studies have determined that the Osmotic Engine Model alone is not enough to overcome certain microenvironment conditions (like that of increased viscosity) for confined cell migration.

AQUAPORIN-FACILITATED CELL MOVEMENT VIA ACTIVATED SIGNALING PATHWAYS

As discussed above, AQPs are involved in multiple migration mechanisms and have functional roles in the various physical structures that support cell migration. One could then ask if the mode of cell migration influences the specific functional roles of AQPs (i.e., regulating signaling, facilitating water transport, slowing polarization). Although this question remains unanswered, the following sections consolidate what is currently known regarding the mechanisms by which AQPs help cells move. For a more in-depth review on AQP-mediated cell signaling, we refer readers to a previous publication (53).

Aquaporin Active Site

AQPs have active sites that vary by isoform found in their N-terminus, B-Loop, D-Loop, and C-terminus. With these active sites, AQPs can induce a multitude of cell behaviors associated with cell migration: 1) AQPs can stabilize the actin cytoskeleton. 2) AQPs can regulate signaling pathways involved in cell migration. 3) AQPs can co-localize with other membrane-bound proteins to influence function. 4) Activation of AQPs via phosphorylation can regulate AQP water permeability. The specific active sites of AQP isoforms and roles they play in specific pathways have been recently reviewed in detail (54); here, we focus more specifically on the AQP active sites that have been shown to manipulate cell migration.

In 2007, Loitto et al. discovered that AQP9 has a protein kinase C (PKC) phosphorylation site at residue S11 (40). Upon mutation of AQP9’s S11 to alanine, there are dramatic changes in cell morphology and reduction in filopodia size, likely mediated by the decreased interaction between AQP9 S11 and PKCζ. This decreased interaction also leads to reduced levels of small GTPase-Cdc42, which is heavily involved in signaling pathways relevant to cell migration. Furthermore, the D-loop of AQP5 contains a PKA consensus site at S156 that is capable of regulating Ras activation (55) (Fig. 3). Activation of Ras, though not discussed in this particular cited study, could in turn induce cell migration (56).

Role of Aquaporins in Actin Polymerization

Actin is an abundant protein in eukaryotic cells and is an indispensable component of their cytoskeleton. Although actin has multifunctional roles in many cell processes, actin dynamically stabilizes migrating cells through polymerization of new actin filaments at the leading edge of the cell and depolymerization at the trailing end of the cells (7). During this process, AQPs allow for small changes in water flux across the cell membrane, creating small pockets just below the cell membrane where actin can polymerize (28, 29). Aside from the formation of these pockets, AQPs have also been shown to regulate actin polymerization and anchorage to the cytoskeleton:

Reorganization of actin.

Reorganization of the actin cytoskeleton is essential for the dynamic state of the cell. Knockdown of AQP1 in endothelial cells leads to an inability of actin to organize into a structured network polarized to the leading edge of the cells (57). Another study reported a decrease in actin organization within kidney proximal tubule cells derived from AQP1^−/− mice in a wound healing test in vitro (36). As determined from co-immunoprecipitation assays, AQP1 interacts with Lin-7 and β-catenin (Fig. 3). β-Catenin is a transcriptional coactivator that can modulate the expression of matrix metalloprotease (MMPs), chemokines, and cytoskeletal proteins, while regulating cell migration and invasion. Thus, it is possible that the lack of AQP1 leads to the degradation of the Lin-7/β-catenin complex, resulting in decreased organization of actin (57).

From that foundation, it was shown that AQP5 overexpression promotes actin stress fiber formation and lamellipodia dynamics in MDCK cells (38). AQP5, and more specifically, its serine-156 residue (which regulates the Ras pathway), are responsible for these actin morphologies. The formation of C3H10T1/2 cell filopodia occurs upon phosphorylation of AQP9’s serine-11 residue by PKCζ, changing the channel’s structure to induce an influx of water. Concurrently, PKCζ localized in the forming filopodia (following AQP activation) activates Cdc42, resulting in actin polymerization by the downstream activation of Arp2/3 and the freeing of WASP; this Cdc42-mediated actin polymerization is abrogated following mutation of AQP9’s S11 (40). Direct AQP interactions with other transmembrane proteins can also regulate actin polymerization. In RCCD cells, AQP2 interactions with NHE1 in the lamellipodia induce an alkaline pH, providing favorable conditions for actin polymerization (28, 58) (Fig. 3).

Anchorage of actin to the cell membrane.

Following polymerization, actin anchors to the cell membrane to provide the support and stabilization needed during cell migration (7). The first study relating AQPs and actin found that the C-terminus of AQP2 is capable of binding directly to actin (59); however, this relationship was not further studied. Furthermore, AQP1 can colocalize with ezrin (Fig. 3), a protein that cross links to the actin cytoskeleton and is anchored in the plasma membrane (60). Reduction of AQP1 or ezrin leads to similar decreases in F-actin. AQP2 can also bind with ezrin, indicating that this association could be conserved across other AQP isoforms (61). These studies provide fascinating evidence that AQPs play an unexpected role in actin dynamics; however, the field is still lacking a rigorous exploration of the mechanisms by which AQPs associate with actin and regulate actin dynamics, specifically in facilitating cell migration.

Aquaporin and Adhesion-Related Proteins

Integrins are crucial regulators of cell motility as they physically connect cells with their surrounding matrix. Integrins provide a physical structure across which cell traction forces can be propagated, while also allowing the cell to sense cues from the surrounding environment and ultimately helping to dictate migratory behaviors (62). Interestingly, the second external loop of AQP2 contains an Arginylglycylaspartic acid (RGD) binding domain, allowing AQP2 to colocalize with RGD-binding integrins, especially β1 and α5 (α2β1, α5β1) subunits (63) (Fig. 3). AQP2 is the only AQP isoform containing this loop (63). Upon integrin binding with RGD peptides, AQP2 localizes with integrins, induced by integrin activation of cyclic adenosine monophosphate (cAMP) and calcium influxes. Mutation of the AQP2 RGD motif leads to a significant increase in β1 integrin accumulation on the cell surface, as well as a significant decrease in cell migration. The decrease in β1 turnover and focal adhesion dynamics and disassembly results in an increased AQP2 accumulation on the cell membrane (64).

AQPs are also involved in cell-cell adhesion structures (65). AQP5 over-expression in MDCK and HEK-293 cells induces cell detachment and dissemination from migrating cell sheets (66). The AQP5 COOH-terminal tail domain interacts with ZO-1, plakoglobin, β-catenin, and desmoglein-2, all of which are reduced at cell junctions upon AQP5 over-expression (67). These cell-cell adhesions are dependent on AQP5 serine-156 activation of the Ras signaling cascade. Similarly (but to a lesser extent than AQP5), AQP1 and AQP4 expression are associated with decreased cell-to-cell lateral junctions, whereas AQP3 expression increases the expression of these junction proteins (66).

Ion Channels and Transporters

Ion channels work in tandem with AQPs to regulate cell behaviors, including cell migration. The functions of numerous types of ion channels are closely linked to the actin cytoskeleton; for example, they can regulate polymerization of actin, and in turn the state of actin can regulate the ion channel’s function (68).

NHE1.

Rose et al. (69) reported that there is no significant difference in the migration speed of inner medullary collecting duct cells in isotonic versus hypertonic solution, despite the fact that hypertonic solutions have extensive impacts on cell morphology. They hypothesized that since AQP2 expression modulates NHE1 activity and induces mechanical stability by promoting actin polymerization (Fig. 3), the increase in AQP2 after hypertonic shock leads to a more stable cell. This demonstrates that cells have adapted complex and interlinked mechanisms to continue to migrate and aid in the wound-healing response under pathophysiological conditions.

AE2.

Anion exchangers, such as anion exchanger 2 (AE2), are often concentrated to the leading edge of migrating cells (70). AQPs have previously been shown to co-localize and even work in tandem with AE2 to regulate the water movement within cells (71). Though a direct functional relationship between AQPs and AE2 has not yet been investigated, given the essential role AE2 plays in the regulation of intracellular osmolarity and resulting water flow, it could be hypothesized that these exchangers regulate the formation of migratory structures of motile cells(72).

TRPV4.

TRPV4 is a mechanosensitive Ca²⁺ cation channel extensively expressed in mammalian tissues, which are capable of regulating cell migration (73). In RCCD, AQP2 immunoprecipitates with and induces activation of TRPV4 and the small conductance potassium channel SK3 (74). The co-localization of TRPV4 and AQP2 promotes a dramatic increase in the influx of Ca²⁺ and its accumulation in intracellular stores, resulting in cell volume regulation and cell migration (Fig. 3). The same group also found that AQP2 is essential to produce ATP following the TRPV4-mediated influx of Ca²⁺. This increased production of ATP accelerates the dynamic behaviors of rapid assembly and disassembly of focal adhesions, consequently increasing the migration speed of RCCD cells (75). RCCD lamellipodia formation is heavily regulated by AQP2 and its physically associated TRPV4 activation of NHE1 (Fig. 3). Upon blockage of AQP2 or TRPV4, there is a dramatic reduction in NHE1 lamellipodia activity with ensuing cytoskeletal disorganization and decreased cell migration (28).

AQPs have not yet been linked with the transduction of mechanical signals. However, if AQPs are capable of recruiting TRPV4, it could be concluded that AQP indeed plays a role in stretch-activated ion channel mechanotransduction (76). Interestingly, in Xenopus oocytes, human AQP1 closes with increased membrane tension (77). Hence, one could hypothesize that AQPs also play a similar role to stretch-activated ion channels in transmitting mechanical signals into cellular responses via influx of H₂O, glycerol, or H₂O₂. Based on the study in Xenopus oocytes, we expect that high membrane tension would lead to AQP1 closing while TRPV4 opens; on the other hand, low membrane tension may promote AQP1 opening while TRPV4 closes. This could in turn lead to a feedback loop or cycle where activation of one channel (AQP or TRPV4) induces a response that activates the other, ultimately leading to cyclic and dynamic movement of the cell. These hypotheses have not yet been explored, but studies could be designed to contribute new insights into whether AQPs play a role in stretch-activated ion channel mechanotransduction.

Aquaporins and H₂O₂

H₂O₂ is a stable reactive oxygen species and has been implicated as a mediator of cell migration. Conventionally, it was believed that H₂O₂ enters the cell solely by slow and energetically unfavorable diffusion through the cell membrane. In 2010, Miller et al. showed that AQP3 is a necessary component in (NADPH oxidase 2) NOX2-mediated formation of H₂O₂ and, more importantly, that AQP3 and AQP8 mediate favorable H₂O₂ transport into mammalian cells (20, 21) (Fig. 3).

AQP3 transport of H₂O₂ in relation to cell migration was first identified when Hara-Chikuma et al. determined that AQP3 deficiency impaired H₂O₂ and T cell migration, with ensuing dysregulated F-actin organization and GTPase-Cdc42 response to chemokines (18). Supplementing T-cells with H₂O₂ (which likely enters through passive diffusion) restores the defects that accompanied the AQP3 deficiency, and thus exemplifies the important role that AQP3 plays in the transport and production of H₂O₂ for chemotaxis. The same group found that there is also a reduction in chemotaxis of T-cells from AQP3^−/− mice during the development of psoriasis toward Chemokine (C-C motif) ligand 20 (CCL20) and Chemokine (C-X-C motif) ligand 9 (CXCL9), thereby reducing the immune response(18). Furthermore, the group also discovered that AQP3 transport of H₂O₂ facilitates nuclear factor-κB (NF-κB) signaling in keratinocytes, by inactivating protein phosphatase 2A (78). The model they were investigating did not involve cell migration; however, the relationship between AQPs and NF-κB-mediated migration has been extensively studied across multiple disease states and tissue types (79–82). From initial studies (mostly disease state), AQPs influence NF-κB activation via their active sites to promote migration, but this provides an interesting perspective to the multiple cascades AQPs can utilize to activate migration. Meanwhile, a study conducted outside that group found that AQP3-expressing colon epithelium cells react to the cue of H₂O₂ significantly faster than non-AQP3-expressing cells, which increases migration speed, stress fiber size, lamellipodia size, and number of focal adhesions marked by the proteins focal adhesion kinase (FAK) and paxillin (37).

Work connecting AQP8 transport of H₂O₂ is still emerging; the studies have largely focused on immune cells due to their prominent expression of AQP8 (20). The efficient differentiation and activation of B-cells are dependent on AQP8-mediated rapid transport of NOX2-produced H₂O₂ (83). Following this activation, B-cells undergo migration to the T cell/B cell border; however, the exact mechanism for H₂O₂-mediated B cell migration, as facilitated by AQP8, has yet to be investigated. This function also becomes dysregulated for leukemia cells (84, 85), but this topic is further discussed in our recent review article focused on cell migration in pathological situations (30).

AQUAPORINS AND IMMUNE CELL MIGRATION

Cells of the Immune System

Migration is a critical process for the function of immune cells, and the roles AQPs play in facilitating this migration have been extensively studied. For a more in-depth explanation into the functions AQPs play during inflammatory responses, we direct readers to the review by Da Silva et al.(86).

Neutrophils.

Infections and the resulting host signaling lead to a rapid migration of neutrophils from the blood to the inflammatory site to enact their functions (87). As discussed above, the first connection between cell migration and AQP expression was discovered in neutrophils. In 2002, Loitto et al. found that water fluxes are essential to the formation of neutrophil lamellipodium, and that this process is impaired upon the inhibition of AQP9 (23). Following that initial discovery, other groups have made connections between AQP9 and neutrophil migration (Fig. 5A). In a mouse model, it was established that AQP9 expression in neutrophils is required for the sensitization phase of contact hypersensitivity through regulation of neutrophil migration; more specifically, neutrophil migration toward a chemoattractant is significantly impaired in cells from AQP9^−/− mice (88). In addition, there was a reduction in the accumulation of neutrophils within lymph nodes of AQP9^−/− mice.

Figure 5.

AQPs and immune cell responses in contact hypersensitivity models. A: AQP9 regulates neutrophil function: AQP9 regulates numerous steps in the chemoattractant-based migration toward a site of tissue damage. Though not yet determined, we hypothesize that AQP9 can regulate neutrophil transendothelial migration, including the process of cell rolling and extravasation. B: AQP3 regulates CD4+ T-cell function: AQP3 induces activation, transendothelial migration, and chemokine-induced migration for CD4+ T-cells. C: AQP7 regulates dendritic cell function: AQP7 induces DC chemotactic migration specifically toward (CCL21 and CXCL12), antigen uptake, and accumulation in LNs following activation. Though it is not yet determined, we hypothesize that AQP7 may also play a role in DC activation, presentation of uptaken antigens, and migration from the site of inflammation into the lymphatic system. AQP, aquaporin. Figure was created using BioRender.com.

The relationship between neutrophil AQP9 expression and chemoattractant-induced migration was further expanded on when Karlsson et al. demonstrated that AQP9 becomes phosphorylated in the presence of chemokines N-Formylmethionine-leucyl-phenylalanine (fMLF) and phorbol myristate acetate (PMA) (29). Upon alteration of AQP9’s phosphorylation site, which changes residue serine-11 into aspartic acid, neutrophils are unable to polarize AQP9 to their cell membrane, while wild-type cells retain dynamic AQP9 distribution. The phosphorylation of AQP9’s serine-11 by Rac1’s downstream effector PKC leads to AQP9 localization at the leading edge of the cell following chemoattractant stimulation. Further studies could investigate AQPs’ role in other critical steps of the leukocyte adhesion cascade, including cell rolling, extravasation from the bloodstream, and invasion through tissues.

T-cells.

Once activated, T-cells differentiate into sub-cell types that will migrate throughout the body to activate and recruit other immune cells or migrate directly to the site of infection to attack invading pathogens. To reach specific locations within the body, T-cells must use cues from their microenvironment, including chemokines, to migrate to their target location (89).

T-cell migration toward chemokine CXCL12 is dependent on AQP3-mediated H₂O₂ uptake, yet not water or glycerol transport (Fig. 5B). H₂O₂ transport is essential for the activation of the small-GTPase cdc42 and resulting dynamic actin reorganization. Although the exact mechanism behind this H₂O₂-mediated CXCL12-induced cdc42 activation remains unknown, AQP3^−/− T-cells show a reduced Cdc42, Rac1, and Itk activation in response to CXCL12. This change in signaling from AQP3^−/− T-cells results in a significant decrease in chemotaxis compared with wild-type cells. Following CXCL12, AQP3 is polarized to the leading edge of the cell, likely to facilitate this influx of chemokine-produced H₂O₂. The production of H₂O₂ was not investigated in this study, but the authors postulated that it could be result of NOX2 (18).

AQP4 also plays a key role in T-cell chemotaxis. In naïve nontransplanted mice, AQP4 inhibition (with small molecule AER-270/271) reduces the number of circulating CD4+ and CD8+ T-cells. AQP4 knockdown also reduces expression of chemokine receptors Sphingosine-1-phosphate receptor 1 (S1PR1) and Chemokine (C-C motif) receptor (CCR)7, and their master regulator Krüppel-like factor 2 (KLF-2), while also reducing chemotaxis initiated by the chemokines S1P and CCL21 (90). The results here make the claim that the reduced migration is a result of T-cell changes; however, given AQP’s role in DC activation, we speculate that it is possible that the inhibition of AQP4 also reduces DC activation, which may reduce DC migration to the lymphatic system and the ensuing activation of T-cells. Regardless of the exact mechanism, AQP-dependent migration is still essential for this immune response.

Dendritic cells.

Upon pathogenic invasion, dendritic cells (DCs) phagocytose the pathogen, display the pathogen-specific antigens (91), and migrate to the lymphatic system to activate the adaptive immune response.

AQP7, expressed on epidermal and dermal DCs, is involved in antigen uptake and accumulation in the lymph nodes (LNs) (92) (Fig. 5C). AQP7^−/− cells show a reduction in chemokine-dependent cell migration in comparison to wild-type cells in vitro. In vivo, AQP7-deficiency results in a reduced accumulation of antigen-retaining DCs in the LNs. This could be caused by the decreased antigen uptake, which would reduce the activation of migratory pathways in DCs. To our knowledge, this is the only study that investigated the connection between AQPs and DC migration; thus, there is still much that needs to be unraveled about their relationship. For instance, how do AQPs play a role in antigen uptake? Li et al. showed that AQP2 can regulate cell endocytosis (61); could this be a similar result of AQP7 in regulating DC phagocytosis of pathogens? Is the resulting decrease of DCs in the LNs a result of decreased DC activation or a result of reduced DC migration through reduced AQP7 expression?

Macrophages.

Along with neutrophils, monocytes are key effector cells in the initial response to the invasion of microorganisms. Monocytes are rapidly recruited to damaged or infected tissues, where stimulation by proinflammatory cytokines causes their differentiation to macrophages (93). Macrophages can shift from their resting (M0) state and respond to the surrounding invaders classically (M1) by phagocytosis, while secreting proinflammatory and antimicrobial cytokines. Upon stimulation of macrophages with anti-inflammatory cytokines, they adopt their alternatively activated (M2) state, which leads to alterations in cell morphology and secretory patterns, both of which aid in tissue reconstruction by promoting ECM development, angiogenesis, and cell proliferation. Macrophages are key players in the immune response, but their function depends on their ability to activate and migrate to the target locations.

Nonstimulated macrophages (M0) isolated from AQP1^−/− mice had a surprisingly increased migration of two- to threefold compared with WT macrophages (94). This enhanced migration from AQP1 knockout is likely mediated by the Src/PI3K/Rac pathway, demonstrating that AQP1 expression is, surprisingly, suppressing M0 macrophage migration, a result that seems to contradict all other studies involving AQP1 (Fig. 6A). In the study’s acute kidney injury (AKI) model, ablation of AQP1 leads to elongation of lamellipodia at the leading edge with an increased arginase activity, indicating an M0–M2 shift. In tandem with this M2 shift, there is a drastic reduction in M1 markers (e.g., NOX and F4/80). This information leads to the unexplained question: why are macrophages the only cell type where AQP1s are preventing migration? The authors hypothesized that this paradoxical finding was a result of the M0 macrophage’s high membrane tension and bending moduli, leading to the closure of the AQP1 channel. They relate this hypothesis to the decreased Rac1 activation in cells with high membrane tension (95), yet they do not connect the finding directly to AQP1. Paradoxically, the ablation of AQP1 from LPS-stimulated macrophages decreased the migration, a result similar to all other known cell types. This again could be related to the decreased membrane tension of M1 macrophages.

Figure 6.

AQPs as a regulators of macrophage activation and cell migration. A: following knock out of AQP1, there is an enhanced M0 to M2 transition, resulting in increased M2 migration. Macrophages are still capable of making the transition while expressing AQP1, yet there is a dramatic change in activation following knockout. The researchers originally believed this inhibitory effect was the cause of high M0 membrane tension closing AQP1s water permeability. This, however, was uninvestigated and does not provide relevant insight into the actual mechanism of inhibition. B: following macrophage M1 activation, AQP1 will regulate the transition to M2 state by activation of various signaling pathways, though the actual mechanisms by which AQP regulates these pathways and how they initiate the transition remain unknown. C: finally, in M1 stimulated macrophages, AQP1 plays a conventional role in cell migration: following its knockout, the cells move slower. Solid arrows represent a transition between activated states, dashed arrows represent cell migration. AQP, aquaporin; WT, wild type. Figure was created using BioRender.com.

Although AQP1 inhibition of macrophage migration seems controversial, other studies have provided supporting evidence for this finding. In another AKI rat model, an M2 macrophage polarization was found to occur by an AQP1-induced M2 phenotype switch, likely mediated by PI3K activation (96) (Fig. 6B). Yet upon AQP1 silencing via siRNA, there was an increase in the M2 markers expressed. In another AKI model from the same group (97), they found that AQP1 inhibits the M1 polarization by suppressing p38 mitogen-activated protein kinases (MAPK) activation and thus NF-κB translocation. Finally, M1 macrophage AQP1 blockage results in reduction of the released proinflammatory cytokine IL-1β and neutrophilic inflammation. This reduction of IL-1β is shown to be, in part, a result of diminished NLR family pyrin domain containing 3 (NLRP3) signaling, triggered by the AQP1 volume restoration.

We have compiled some research questions that could be answered to shed light on this paradoxical finding: What is AQP1’s function in other cell types with high membrane tension? Is AQP1 playing a role in maintaining this high membrane tension? Using a membrane tension probe, like Flipper-TR, is it possible to understand the localized function of AQP1 and membrane tension? Are there any structural changes in AQP1 that occur upon tension that impact the protein’s downstream signaling? Could the activation of specific signaling pathways like the PI3K pathway be mediated by AQP1’s active site that activates Src?

However, AQP3 seems to play an opposite role in macrophage polarization in comparison to AQP1. AQP3-mediated H₂O₂ uptake is involved in NF-κB cell signaling in macrophages, resulting in their classical activation (Fig. 6C). This M1 polarization causes increased release of excess H₂O₂ extracellularly via AQP3 (98). Macrophage phagocytotic function is also impaired in an AQP3^−/− mice model, and this response is accompanied by a reduction in macrophage migration in vitro (99). This change has been hypothesized to occur as a result of AQP3’s glycerol transport, providing the energy needed for protrusion formation; however the exact effect remains unexplored.

Immune Response to Pathogenic Conditions

Wound healing.

To test variations in response during wound closures following different treatments, wounds are often inflicted in vitro or in vivo and the resulting migration is observed. Following mouse bile duct ligation in vivo, there is a significant over-expression of AQP1 in the mouse liver tissue (100). In the same study, AQP1^−/− mice showed a dramatic decrease in vascular remodeling needed for inflammation after the bile duct ligation. Similarly, yet in another system, corneal keratocytes are activated and migrate to the wound site with an increased AQP1 expression after mice underwent corneal debridement (101). In AQP1^−/− mice, there is a significant reduction in keratocyte migration. In addition, corneal epithelial cell migration and proliferation increased significantly during corneal reepithelization and wound healing for wild-type corneal epithelial cells compared with AQP5^−/− cells(102). The impaired wound healing of the stratum spinosum in diabetic mice is related to a decrease in the overall expression of AQP3 (103). The wound healing of AQP3^−/− basal keratinocytes in vitro shows a reduced water and glycerol membrane transport with a twofold decrease in cell speed. This result occurs simultaneously with a reduction in p38 MAPK activation. In vivo, wound healing of AQP3^−/− was 50% complete after 5 days compared with the 80% complete in wild-type mice (104). Although the above studies were conducted on different cell types and explored the role of different AQP isoforms, the general mechanisms of action seemed to be conserved. For a more in-depth review on the role of AQPs in regulating the process of inflammation, beyond cell wound healing migration, we refer readers to the papers by Mariajoseph-Antony et al. and Meli et al.(105, 106)

Glial scar formation.

Glial scar formation occurs following injury to the nervous system, where scar-forming astrocytes will migrate to the site of injury to protect central nervous system tissue and begin the healing process. Astrocytes were one of the first investigated cell types where a role between AQPs and cell migration was confirmed. From 2005 to 2006, there were a quick succession of publications from A. S. Verkman’s group detailing the relationship between AQP4 expression and astrocyte migration. In 2005, the group showed that there is a decreased Transwell migration for AQP4^−/− astrocytes compared with wild-type cells (27). Although migrating, the AQP4 is dynamically polarized to the leading edge of the astrocyte. In addition, this migration was enhanced by the addition of a small osmotic gradient, suggesting the movement of water is essential to the migratory function. In vivo, the AQP4^−/− mice showed a significantly impaired glial scar formation, in concert with a reduced migration of scar-forming astroglia (27). In 2006, the previous study was expanded upon by the same group, where they injected AQP4-fluorescent astrocytes back into mice brains and observed migration following injury (107). Upon injury, surrounding astrocytes showed an increase in AQP4 expression to aid in their migration toward the wound. In addition, there was greater migration toward the injury site with an increased cell elongation of the AQP4^+/+ astrocytes compared with the AQP4^−/− astrocytes. This was validated and expanded upon more recently, when one study found that there was a disassembly of AQP4 tetramer structures at astrocyte endfeet, with a dynamic polarization of AQP4 monomers toward plasma membrane domains in the neutrophil (108).

Following the quick succession of papers in 2005–2006, the group came out with two review papers, one on AQPs’ role in brain function (109) and the second on AQPs’ role in cell migration (25); these reviews connected their works with others, and ultimately laid the foundation of our knowledge on AQPs’ influence on cell migration. Since this quick burst of papers, the majority of the work relating to brain function has shifted toward understanding how AQPs contribute to the progression of gliomas. Much of that information is further detailed in our recent review article focused on cell migration in pathological situations (30).

Sepsis.

Sepsis is characterized by the excessive activation of inflammatory mediators, leading to extensive inflammation and microvascular malfunction (110). To combat this increased inflammation, the body often overcompensates, leading to a suppressed immune system, thereby leaving the patient at risk for subsequent infections. AQPs have been shown to play a key role in the regulation of sepsis, as thoroughly reviewed by Rump et. al (111); however, this section specifically focuses on the immune cell migration during the disease.

By regulating the migration of immune cells, it is possible to treat the host response to sepsis. In the aforementioned paper, the authors claim that AQP3, AQP5, and AQP9 are the most important isoforms in the regulation of immune cell migration(111). The induction of sepsis and septic shock in a clinical setting was shown to increase AQP1 mRNA expression in leukocytes; this expression level is correlated with the severity of sepsis. Increased leukocyte AQP1 expression is also stimulated in vitro through the addition of LPS (112). It was theorized that increased AQP1 expression may enhance the neutrophil’s ability to migrate to the sites of infection. This has not yet been investigated, but to add to that theory, AQP1 could also influence leukocyte response to chemokines. Like leukocytes, macrophage stimulation by their coculture with bacterium increased AQP9 expression, distribution, and re-organization (113).

A polymorphism in AQP5’s c-allele (i.e., position 1364 is switched from alanine to cystine) is associated with decreased AQP5 expression, along with an increased 30-day survival in patients with severe sepsis. This relationship is related to a reduction in neutrophil migration after the polymorphism. This reduction leads to decreased tissue damage caused by fewer infiltrating neutrophils and proinflammatory mediators (114, 115).

AQUAPORINS AND MESENCHYMAL STEM CELL MIGRATION

Mesenchymal stem cells (MSCs) are multipotent stromal cells that are capable of adhering to tissue culture plastic, express specific cell surface markers and lack others, and have the potential to differentiate to adipocytes, chondrocytes, and osteocytes (116). Interestingly, AQP expression has been shown to play significant roles in the differentiation of MSCs (117–119). Zannetti et al. recently provided a detailed review on the overall role of AQPs in MSC function (120), yet they did not focus on MSC migration. Much of our knowledge of the relationship between AQPs and MSCs stems from their group’s work. Here, we focus our discussion specifically on the role of AQPs in MSC migration.

Migration is an essential process in MSC function. MSCs have a homing function to migrate to a site of injury, where they differentiate into other needed cell types or secrete necessary growth factors. In MSCs, AQP1 modulates β-catenin and FAK expression (121). More specifically, the depletion of AQP1 reduces the activation of β-catenin and increases the degradation of FAK, thereby significantly reducing MSC migration (121). MSCs with increased AQP1 expression are more likely to successfully localize to a fracture site within a mouse model, suggesting that AQP1 also enhances MSC tropism. This pathway is independent of the traditional regulators of MSC migration, specifically stromal cell-derived factor 1 (SDF-1) and CXCR4.

AQPs also seem to be involved in MSC signaling-induced cell migration. MSC-conditioned media (MSC-CM) contains a complex formulation of growth factors, cytokines, and extracellular vesicles (EVs) capable of altering cellular behavior. Two studies by Pelagalli et al. demonstrated that cells cultured with MSC-CM have an increased AQP1 and CXCR4 expression, which results in increased wound healing and cell migration; this is consistent across a variety of cell types. Furthermore, AQP1 and CXCR4 together can regulate MSC migration through AKT and extracellular signal-regulated kinase ½ (ERK) signaling (119, 122). The studies hypothesize that this increase in AQP1 expression could be a complex cross talk capable of inducing cell migration in various tissue types; however, they do not further investigate the mechanism by which MSC-CM increases AQP1 expression.

One well-studied component of the MSC-CM are EVs. These vesicles are membrane-delimited secreted particles involved in paracrine and autocrine signaling and used to transport cargos of nucleic acids, lipids, or proteins, including AQPs. Although it has not yet been investigated, one mechanism behind the MSC-CM increase in AQP expression could be the direct transfer of AQPs via EVs, or the transport of an mRNA cargo coding for AQPs, to the target tissue cells. We suggest that using EVs to induce AQP expression could hold a therapeutic potential in diseases that result from migration deficiencies.

AQUAPORINS AND DEVELOPMENT

Embryogenesis

Cell migration is a key step of embryonic development as tissues undergo indispensable rearrangements that lead to germ layer positioning, patterning, and organogenesis. The roles that AQPs play in cell migration have been investigated in the context of embryogenesis-relevant processes such as neural crest migration, invasive extravillous trophoblast migration, and angiogenesis. It has also been proposed that AQP1 plays a role in the proliferation and migration of granulosa and theca cells, while also regulating the water flux required for follicle development (123). Below, we outline what is known about AQPs in several of these embryogenesis-relevant processes.

McLenna et al. discovered that migrating neural crest cells (NCCs) have a high AQP1 expression (124). The group took this discovery a step further in a subsequent paper and specifically investigated the role that AQP1 plays in NCC motility and invasion (41). They found that AQP1 and EphB receptors co-localize to the invasive front of NCC clusters. AQP1 expression at the leading-edge results in higher NCC migration and invasive speeds in vitro and enhanced invasion in vivo; inhibition of AQP1 with acetazolamide counteracts these responses. Decreased AQP1 expression reduces NCC directionality in response to guidance cues both in vitro and in vivo. Furthermore, there is a significant reduction of number and length of NCC filopodia in vivo after AQP1 knockdown. In vivo, AQP1 expression is associated with NCC matrix degradation via MMP2/9. Finally, during in vitro NCC migration, AQP1 and pFAK colocalize at the leading edge of cells, but upon AQP1 over-expression, there is decreased pFAK and β1 integrins found on the cells’ surface. It was postulated that the AQP-mediated decreased expression of focal adhesion proteins could act similarly to the previously discussed increased turnover. The faster turnover of focal adhesions could lead to a more dynamic cell movement, as opposed to low focal adhesion turnover, where the cells can get “stuck.”

The other main mode of migration studied in the context of embryogenesis, in relation to AQPs, is the invasion of extravillous trophoblast cells (EVTs) (125). These cells invade and promote remodeling of the maternal blood vessels to provide an adequate blood supply to the growing trophoblast. EVTs adopt a mesenchymal-like phenotype that allows for migration and invasion through the surrounding endometrium. Although AQP1, AQP3, and AQP9 are expressed on EVT cells, it was found that AQP3 is the key regulator of this mesenchymal-like EVT migration and invasion (126, 127), likely through the regulation of adhesion genes such as PDGF-B, in addition to signaling pathways such as PIK3/AKT/NF-κB and TNF (128). Decreased placental AQP3 expression was related to an increase in preeclampsia caused by shallow EVT invasion (127). Another study has noted the importance of AQP1 in placental development. LPS induces preeclampsia-like phenotypes in mice specifically, insufficient EVT invasion, incomplete remodeling of spiral arteries, and placental lesions, all with a reduction of AQP1. Low-dose aspirin treatments ameliorate these preeclampsia-like symptoms, while restoring the expression of AQP1. Cells revert to a preeclampsia phenotype (even with low-dose aspirin treatments) following AQP1 knockdown (129).

Researchers have made the claim that there are relationships between cancerous epithelial-to-mesenchymal transition (EMT) and EVT invasion/expansion (130); EVT invasion is a highly regulated form of migration capable of being turned off, while mutations that arise from cancer-related EMT cause cells to revert to an unregulated form of EVT. AQPs play a clear role in the progression toward EMT in both cell types, yet little is known about the connection between these two distinct mechanisms. Could AQPs act as the key in this EMT transition, or are they merely symptoms perpetuating the invasion? Further studies should be conducted to explore the potential AQP-bridged connection.

Angiogenesis

Angiogenesis is the growth of new blood vessels, beginning in utero and continuing throughout a person’s life. The process occurs to provide oxygen and nutrients to satisfy the metabolic requirements of nearby tissues. For the formation of new blood vessels, capillary sprouts invade and migrate into their surrounding epithelium. This migration and invasion are heavily regulated by AQPs. Much of the work done to understand the role that AQPs play in angiogenesis has been conducted by observing blood vessel formation in tumors, which is further covered in our recent review article focused on cell migration under pathological conditions (30). Meanwhile, this section will focus on AQPs in nondisease-related angiogenesis.

Treatment of human umbilical vein endothelial cells (HUVECs) with estrogen increases their AQP1 expression in a dose-dependent manner (60). This process is regulated by a functional estrogen response element-like motif, which was identified as the promotor for the AQP1 gene. Upon AQP1 knockdown, estrogen-induced formation of F-actin stress fibers is significantly attenuated, along with significant inhibition of estrogen-enhanced cell proliferation, migration, invasion, and tube formation. Proangiogenic factor Epo is capable of modulating AQP1 within minutes after exposure. Changes in AQP1 localization are accompanied by the rearrangement of F-actin with an influx of calcium ions necessary for endothelial cell migration (131). Upon inhibition of AQP1 with Bacopaside II, there is a reduction in HUVEC migration, tube formation, and cell viability, suggesting the indispensable role of AQP1 in HUVEC angiogenesis (132).

Finally, AQPs are also involved in pulmonary arterial smooth muscle cell migration and proliferation, in a β-catenin-dependent manner (133). Increasing AQP1 expression upregulates β-catenin, while silencing of β-catenin prevents both hypoxia- and AQP1-mediated migration and proliferation. This hypoxia increased AQP1 expression to increase pulmonary arterial smooth muscle migration is regulated by an increased Ca²⁺ influx (134).

INNOVATIONS, FUTURE DIRECTIONS, AND PERSPECTIVES

As extensively discussed above, AQPs play a large role in the migration of many different cell types during normal physiological processes (reviewed here); we refer the reader to our other recent review focused on cell migration in disease conditions (30). Future innovations could focus on AQP-based drugs that increase AQP expression to aid in cell migration-deficient diseases, or that decrease AQP expression or function in diseases involving elevated levels of cell migration. For example, our recent review on pathological cell migration further elaborates on AQP-based drugs, including miRNAs, RNA sponges, and different types of inhibitors (30). To synthetically increase AQP expression, Yan et al. has developed and tested an artificial aquaporin (aAQP) constructed from a pillar[5] arene backbone (135). The artificial channel is selectively permeable to water, while excluding ions and other small molecules. The aAQP increases the wound-healing speed of HUVECs and restores the healing capabilities of cells with loss of functional AQPs. Their design was meant to replicate the properties of AQP1, but future aAQPs could be developed to recapitulate other AQP isoform functions like the transport of H₂O₂ or glycerol. Further exploration in this area could provide an understanding of the mechanisms behind aAQP-modulated cell migration, including if the increased migration speed is solely a result of increased water permeability, or if the aAQP is capable of regulating downstream signaling cascades.

To test hypothetical scenarios of how AQP1 impacts NCC invasion, McLenna et al. developed a hybrid computational model (41). This model investigated parameters to recapitulating five cellular migratory behaviors that AQP1 influences: cell speed, filopodia stability, filopodia polarity, filopodia number, and ECM degradation. From the model, they found that cell speed, filopodia stabilization, and ECM degradation are key parameters for controlling NCC invasion. Interestingly, the model also discovered that stabilization of filopodia, rather than filopodia number or polarity, increases the distance traveled by cells (regardless of the speed); it also reduces the likelihood of cell separation from the collective sheet. However, this model only considered behaviors that were regulated by AQPs, without identifying the roles of the AQPs themselves. Future work should focus on developing more comprehensive computational models that incorporate functions of AQPs in cell migration, including implementing factors regulating AQP expression, downstream signaling, and physical changes from water transport.

In summary, this review is a compilation of comprehensive work connecting AQPs with cell migration through a variety of mechanisms and across a variety of cell types. Two simple points remain clear: 1) AQPs are indispensable to cell migration across all behaviors adopted by motile cells, and 2) our knowledge of AQPs is rapidly expanding, yet we still have much to learn about their role in cell migration and to what degree this knowledge will yield translational benefits for the biomedical field.

GRANTS

The authors acknowledge funding from the National Institute of General Medical Sciences (NIGMS) Maximizing Investigators’ Research Award #R35GM142838 (to K.M.S.) and from the Clark Doctoral Fellowship (to I.M.S.). Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number R35GM142838.

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

I.M.S. and K.M.S. conceived and designed research; drafted manuscript; edited and revised manuscript; approved final version of manuscript.