Outside the mainstream: novel collecting duct proteins regulating water balance

Abstract

Body water balance is critical to survival and, therefore, very tightly regulated by the hypothalamus and kidney. A key mechanism involved in this process, the arginine vasopressin-mediated phosphorylation and apical membrane insertion of aquaporin 2 in the collecting duct, has been extensively studied; however, with the increased availability of conditional knockout animals, several novel collecting duct proteins have recently been implicated in water homeostasis. In this Mini-Review, we briefly discuss these novel proteins and their roles in the regulation of water homeostasis.

water homeostasis is a critical process involving the coordinated action of the hypothalamus, including the thirst mechanism and control of arginine vasopressin (AVP) release, and the kidneys, where water is either reabsorbed or excreted in the urine. The collecting duct (CD) represents the key site for fine-tuning of water balance by the kidney. Much progress has been made into understanding the molecular mechanisms involved, especially as they relate to the classical AVP-aquaporin 2 (AQP2) axis, as reviewed extensively elsewhere (2, 11, 17, 25). However, recent studies have implicated several novel proteins in the regulation of water balance, and these may or may not act by influencing the components of the classical linear AVP-AQP2 axis upstream of AQP2. In this Mini-Review, we highlight some of the CD proteins implicated in this important area of renal physiology in the last several years (Fig. 1).

Fig. 1.

Novel proteins implicated in water balance in the kidney collecting duct outside of the linear AVP-AQP2 axis. These proteins are localized in different cellular compartments and have a diverse range of functions, such as cell-to-cell adhesion, formation of epithelial polarity (13, 28), transduction of cellular signals (10, 38), recycling of channels/transporters to and from the plasma membrane (23), generation of metabolites (7, 14, 15, 33), posttranslational modifications (as shown by star) of channels/transporters, and transcriptional regulation (16). Proteins specifically discussed in this Mini-Review are highlighted in bold. Abbreviations: AKAP220, A-kinase anchoring protein 220; GATA2, GATA binding protein 2; MPGES-1, microsomal prostaglandin E synthase 1; NFAT5, nuclear factor of activated T cells 5; RhoA, Ras homolog gene family member A; TAZ, transcriptional co-activator with PDZ binding motif.

Membrane Localization of Aquaporins—It’s More than Just cAMP

The presence of aquaporins in the apical and basolateral membranes of select regions of the nephron is critical for water reabsorption from the tubular fluid. The cAMP-dependent protein kinase A (PKA)-mediated phosphorylation of AQP2 downstream of vasopressin V2 receptor (V2R) activation by AVP has long been recognized to promote the forward trafficking of AQP2 into the apical membrane. A-kinase anchoring proteins (AKAPs) can serve as scaffolding proteins to localize PKA and other proteins, such as cyclic nucleotide phosphodiesterases, to specific regions within the cell, thus allowing for spatial control of cAMP signaling events (3). AKAP220 was previously implicated in PKA-dependent phosphorylation of AQP2 (26), but the influence of AKAP220 on AQP2 membrane accumulation, however, appears to be complex. Recently, in vitro studies of three-dimensional polarized spheroids generated from mouse inner medullary collecting duct 3 cells showed that AKAP220 is also important in the organization of the apical actin network in CD cells (36). The same study (36) reported that, while mice with global AKAP220 deletion exhibited similar water intake and urine output compared with wild-type mice under baseline conditions and were able to concentrate their urine during water deprivation, urine osmolality following an acute water load remained higher in AKAP220-null compared with wild-type mice, with more apically distributed AQP2 and reduced Ras homolog gene family member A (RhoA) activity. Inhibition of RhoA activity has been suggested to impair internalization of AQP2 (20), which might account for the phenotype observed in AKAP220-null mice. Interestingly, phosphorylation of AQP2 at Ser256 did not appear to be impaired in the absence of AKAP220 (36). The precise role of AKAP220 in AQP2 trafficking and whether this differs depending on hydration status awaits further clarification.

Several studies have called into question whether cAMP is necessary for apical trafficking of AQP2 downstream of activation of G protein-coupled receptors, such as the V2R and prostanoid receptors EP2 and EP4. A recent study highlighted a disconnect between the time course of cAMP elevation in response to stimulation of these G_s-coupled receptors and the apical membrane targeting of AQP2 (27). Treatment of cells with an adenylyl cyclase inhibitor did not prevent AQP2 membrane targeting (27), and AC6-null mice do not increase cAMP in response to AVP, yet they are still able to concentrate their urine in response to water restriction or dDAVP administration (31). Accordingly, cAMP-independent signaling pathways for AQP2 apical trafficking in response to activation of these receptors await elucidation.

Albeit less well studied, differences in AQP3 expression or localization to the basolateral membrane can also affect water reabsorption across the CD. Mice lacking acyl-CoA binding protein (ACBP), a protein involved in transporting acyl-CoA esters between different enzymes, were found to have slightly higher urine flow under basal conditions and a reduced ability to conserve water during water restriction, but no apparent difference in sodium or potassium handling (18). Expression and basolateral localization of AQP3 were reduced in the ACBP knockout mice, with no apparent changes in the expression and localization of AQP1, 2, or 4 (18). Fatty acyl-CoAs can serve as allosteric regulators of cellular proteins, such as ryanodine-sensitive Ca²⁺ release channels (8), which are in turn important regulators of AQP2 distribution in principal cells (6); however, the mechanism(s) by which ACBP regulates AQP3 expression and localization is unknown.

Transcriptional Regulation—Who Turns the Switch On/Off?

Transcriptional regulation represents an important means of long-term regulation of CD function. Unsurprisingly, deletion of a number of transcription factors results in renal developmental abnormalities, phenotypes of nephrogenic diabetes insipidus (NDI), or urinary concentrating defects. However, several such transcription factors have additionally been shown to play a role in water handling independently of developmental effects. Using an inducible system, postdevelopmental deletion of GATA binding protein 2 (GATA2) from the nephron induced polyuria and revealed an important role of GATA2 as a direct transactivator of AQP2, as well as regulating V2R and AQP3 mRNA levels, suggesting GATA2 promotes water conservation via multiple mechanisms (39). AVP-independent increases of AQP2 by the ligand-activated transcription factor Farnesoid X receptor was also recently shown to contribute to regulation of urine volume (41). Hyperosmotic stress increases the phosphorylation of TAZ (transcriptional coactivator with PDZ-binding motif, also known as Wwtr1), which is followed by increased physical interaction between TAZ and nuclear factor of activated T cells 5 (NFAT5), reducing NFAT5 activity (12). NFAT5 is an osmoregulatory transcription factor that upregulates the expression of AQP2 in response to calcium and osmotic stress (19), suggesting that TAZ may help to fine-tune both NFAT5 and AQP2 expression under conditions of osmotic stress. Together, these recent developments show that there is far more to transcriptional regulation of CD water handling proteins than the classical transcription factor cAMP-responsive element-binding protein. For a recent review describing roles for additional nuclear receptors such as peroxisome proliferator-activated receptor-γ, glucocorticoid, and estrogen receptors, see Ref. 40.

Paracrine Signaling and Cell-Cell/Cell-Matrix Interactions—Love Thy Neighbor

The CD is composed of intercalated and principal cells, and, until recently, water handling within the CD was attributed to the principal cells. However, recent findings also highlight the importance of intercalated cells in this process. The prorenin receptor (PRR) is a component of the local renin-angiotensin system in the kidney but can also activate cell signaling mechanisms, such as ERK1/2 (24), and serves as an accessory protein for vacuolar ATPase (1). Nephron-wide deletion of PRR in mice using an inducible system to avoid confounding developmental abnormalities produced severe polyuria, reduced AQP2 abundance without any change in principal-to-intercalated cell ratio, and produced a 30% higher mortality rate under 24-h water restriction (29). Interestingly, PRR is predominantly localized to the apical membranes of intercalated cells (1, 9, 29). Providing a possible explanation for this apparent paracrine role for PRR in regulating AQP2, the NH₂-terminal extracellular domain of PRR undergoes proteolytic cleavage to form soluble PRR, which was recently shown to activate a receptor component of the Wnt-β-catenin signaling pathway, Frizzled-8, which is expressed by principal cells, promoting AQP2 expression (21). It should also be noted that, in addition to AQP2, PRR has been shown to regulate other proteins, including increasing the activity of the epithelial Na channel (30) and thus may also exert indirect effects on water homeostasis.

Proteins involved in cell-cell and cell-extracellular matrix interactions have also been implicated in water handling by the CD. Conditional knockout of the widely expressed integrin-β₁ subunit from the CD of floxed Itgb1 mice using a HoxB7-driven Cre recombinase resulted in significantly lower AQP2 and V2R mRNA expression in the renal medulla and a variety of kidney developmental defects, causing severe polyuria under baseline and water-restricted conditions (37). Consistent with a clear role for the predominantly basolaterally located integrin-β₁ in CD development, in vitro experiments demonstrated that interaction between AQP2 and integrin-β₁ via an integrin-binding RGD motif on AQP2 is crucial for migration and tubulogenesis by LLCPK1 cells, apparently through facilitating focal adhesion turnover (5). The authors (5) speculated that this promigratory and structural effect may not involve AQP2-mediated water permeability. Others reported that peptides containing the RGD sequence, which can stimulate “outside-in” signaling via integrins, increased trafficking of AQP2 to the apical membrane of cultured mouse M1 cells stably transfected with human AQP2 via cAMP and calcium-dependent signaling (34). Separating the potential functional AQP2-mediated effects of integrin-β₁ on water handling from the developmental impact of integrin-β₁ on CD morphology will require further studies with appropriate inducible knockout models. Conditional knockout of integrin-linked kinase induced a polyuric phenotype in mice, in association with reduced total AQP2 expression and a shift of subcellular AQP2 localization away from the apical membrane and into the cytosol (4). Similar effects on AQP2 localization and cell water permeability were seen using integrin-linked kinase small interfering RNA treatment of mouse inner medullary collecting duct 3 cells (4). Together, these data suggest that integrins and similar pathways may increase CD water permeability via effects on AQP2 expression and apical localization.

Conclusion

Recent studies with conditional knockout animals have assisted us in understanding the roles of proteins that function within and outside the linear AVP-AQP2 axis in regulating water homeostasis (Table 1); however, there are more questions that remain unanswered. For example, what cellular signatures are associated with short-term vs. long-term changes in intracellular cAMP levels? Which cAMP-independent pathways are activated downstream of Gs-coupled receptors? How are the short- vs. long-term signaling modalities recognized by the principal cells? There are still gaps in our understanding of the mechanisms and molecules mediating cross talk between intercalated and principal cells, intracellular trafficking of aquaporins, and how developmental cues regulate water-handling properties of the CD. While this review focuses on mechanisms involving the CD, intact neural thirst pathways are essential for proper water balance. Expressions of both AQP2 and AQP3 are increased in response to thirst in rats (35); however, most of the aforementioned studies did not delve into the role of the novel proteins in regulating the thirst pathways. Such questions need to be addressed, especially in animal models of global gene deletion. Whether these proteins might also play a role in the regulation of sodium handling and thus indirectly impact water balance in most cases also remains to be investigated.

Table 1.

Novel collecting duct proteins involved in the regulation of water balance

Novel Regulators of Water Homeostasis	Experimental Models
Direct regulators of AQP2 function
AKAP220	Mice; global gene KO (36)
GATA2	Mice; renal tubular cell-specific KO (39)
Farnesoid X receptor	Mice; global gene KO (41)
NFAT5	Mouse cortical collecting duct cell line (19)
Integrin-β1	Mice; collecting duct-specific KO (37)
Regulators of other CD proteins
ACBP	Mice; global gene KO (18)
GATA2	Mice; renal tubular cell-specific KO (39)
TAZ	Mice; renal tubular cell-specific KO (12)
PRR	Mice; nephron-specific KO (29)

KO, knockout. References for the studies are shown in parentheses.

A clinically-relevant future direction of these studies is to aid in designing treatment options for water imbalance disorders, such as NDI. NDI occurs due to unresponsiveness of the CD to AVP due to either malfunctioning AQP2 or V2R expression and/or trafficking. About 90% of patients with congenital NDI harbor a mutated V2R gene (22). Improved understanding of AVP-independent mechanisms of AQP2 regulation, such as the PRR- Frizzled-8-Wnt-β-catenin pathway, could help in designing drug therapies that could bypass the need to activate V2R in patients with this form of congenital NDI. Acquired NDI, most commonly seen in patients treated with lithium, usually results in dysfunctional water reabsorption due to attenuated AQP2 expression and membrane accumulation. Agonist-mediated activation of novel transcription factors of AQP2, such as Farnesoid X receptor (41), has already shown to be useful in increasing AQP2 expression. Whether this treatment will be effective during lithium-induced NDI remains to be tested. As recently discussed by Sands and Klein (32), understanding the intricacies of these regulatory mechanisms is critical to devising better treatments for NDI and other conditions associated with water imbalance.

GRANTS

An American Heart Association Predoctoral Fellowship (to S. R. Rahman, 15PRE25580003) and a Nebraska Center for Cellular Signaling Pilot Project Grant (to E. I. Boesen) partially supported this work.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

S.S.R. and E.I.B. conception and design of research; S.S.R. prepared figures; S.S.R. drafted manuscript; S.S.R. and E.I.B. edited and revised manuscript; S.S.R. and E.I.B. approved final version of manuscript.