A number of ion channels and transporters are expressed in both the inner ear and kidney. In the inner ear, K+ cycling and endolymphatic K+, Na+, Ca2+, and pH homeostasis are critical for normal organ function. Ion channels and transporters involved in K+ cycling include K+ channels, Na+-2Cl−-K+ cotransporter, Na+/K+-ATPase, Cl− channels, connexins, and K+/Cl− cotransporters. Furthermore, endolymphatic Na+ and Ca2+ homeostasis depends on Ca2+-ATPase, Ca2+ channels, Na+ channels, and a purinergic receptor channel. Endolymphatic pH homeostasis involves H+-ATPase and Cl−/HCO3− exchangers including pendrin. Defective connexins (GJB2 and GJB6), pendrin (SLC26A4), K+ channels (KCNJ10, KCNQ1, KCNE1, and KCNMA1), Na+-2Cl−-K+ cotransporter (SLC12A2), K+/Cl− cotransporters (KCC3 and KCC4), Cl− channels (BSND and CLCNKA + CLCNKB), and H+-ATPase (ATP6V1B1 and ATPV0A4) cause hearing loss. All these channels and transporters are also expressed in the kidney and support renal tubular transport or signaling. The hearing loss may thus be paralleled by various renal phenotypes including a subtle decrease of proximal Na+-coupled transport (KCNE1/KCNQ1), impaired K+ secretion (KCNMA1), limited HCO3− elimination (SLC26A4), NaCl wasting (BSND and CLCNKB), renal tubular acidosis (ATP6V1B1, ATPV0A4, and KCC4), or impaired urinary concentration (CLCNKA). Thus, defects of channels and transporters expressed in the kidney and inner ear result in simultaneous dysfunctions of these seemingly unrelated organs.
- vestibular labyrinth
- stria vascularis
- renal tubule
the inner ear is the sensory system for sound, motion, and gravity. It is housed within the temporal bone and consists of the cochlea, vestibular labyrinth, and endolymphatic sac (Fig. 1). The inner ear comprises an array of interconnected fluid compartments that are enclosed by a multitude of highly specialized epithelial cells. The luminal fluid, endolymph, differs in composition between different parts of the inner ear (Table 1). The epithelial cells enclosing the endolymph are highly diverse, as shown in more detail for the cochlea in Fig. 2. Among them are the sensory inner and outer hair cells (Fig. 2, B and C) and stria vascularis (Fig. 2D). Sensory hair cells transduce mechanical stimuli into electrical signals and release neurotransmitters to activate sensory neurons. The stria vascularis is a multilayered epithelium in the cochlea (Fig. 2, A and D), which secretes K+ into the endolymph and generates the endocochlear potential, which contributes significantly to the driving force of sensory transduction. In addition, the inner ear contains and depends on a multitude of highly specialized epithelial cells that control the ionic composition of the endolymph and the magnitude of the transepithelial potential.
Different compartments of the inner ear serve the transduction of specific stimuli. The cochlea transduces mechanical stimuli associated with sound and provides the basis for hearing. The utricle, saccule, and ampullae of the semicircular canals belong to the vestibular labyrinth (Fig. 1), which transduces mechanical stimuli associated with head position and head motion. Vestibular sensory transduction provides input to the vestibular system that controls balance, posture, and eye movements. Sensory transduction in the cochlea and vestibular labyrinth has different electrochemical requirements, although all depend on the cycling of K+ between the endolymph and perilymph. In addition to the fluid compartments that house sensory hair cells, the vestibular labyrinth contains another fluid compartment, the endolymphatic sac, which is devoid of sensory hair cells. The function of the endolymphatic sac is poorly understood, although evidence suggests that it controls endolymph fluid volume (219).
Several epithelia in the cochlea appear to have functional equivalents in the vestibular labyrinth. Among such homology pairs are cochlear (Fig. 2A) and vestibular hair cells, strial marginal cells (Fig. 2D) and vestibular dark cells, outer sulcus cells (Fig. 2A) and vestibular transitional cells, and Reissner's membrane (Fig. 2A) and semicircular canal epithelial cells (163, 303). Several recent reviews have focused on ion transport in different inner ear epithelia (163, 304), hereditary hearing loss (197), cochlear fluid volume regulation (219), and ototoxicity (217).
The similarity between epithelial transport in the inner ear and kidney was first suggested by the observation more than 30 years ago that treatment with high doses of the loop diuretic furosemide causes reversible hearing loss (81, 248). Obviously, the inner ear and kidney have very different functions. Nevertheless, most of the genes encoding the epithelial transporters or channels in the inner ear are similarly expressed and/or similarly sensitive to pharmacological intervention in renal tubular epithelia. More importantly, defects of those genes can lead simultaneously to hearing loss and deranged renal tubular transport. Thus, even though several of the channels and transporters expressed in inner ear epithelia are similarly found in other epithelia or even in excitable tissues such as the heart, the pathophysiologically significant overlap is particularly striking between the inner ear and kidney. In several channelopathies, the renal defect is subtle and clinically overlooked in the face of striking hearing loss or life-threatening cardiac arrhythmia. Closer functional analysis reveals, however, the respective defect in renal function. Thus, much can be learned from a comparison of the transport organization in these two organs. The comparison could further serve as a paradigm that channels and transporters could serve different functions in different organs and that genetic defects or pharmacological inhibition of those channels and transporters could lead to seemingly unrelated functional consequences.
The present review first describes the channels and transporters required for inner ear function. The second part of the review is dedicated to the function of those transporters in renal epithelia. The function and pathophysiological significance of the channels and transporters expressed in both the inner ear and kidney is shown in Table 2.
K+ Cycling in the Inner Ear
Sensory transduction in the cochlea and vestibular labyrinth depends on the cycling of K+ between the endolymph and perilymph (Fig. 2A). K+ cycling in the cochlea consists of K+ flux from the endolymph through sensory hair cells into the perilymph, uptake of K+ from the perilymph into fibrocytes of the spiral ligament, funneling of K+ via gap junctions into basal and intermediate cells of the stria vascularis, efflux of K+ from intermediate cells into the intrastrial fluid, and secretion of K+ by marginal cells of the stria vascularis into the endolymph (163, 304) (Fig. 2A). Similarly, K+ cycling in the vestibular labyrinth consists of K+ flux from the endolymph through hair cells into the perilymph and uptake of K+ from the perilymph and secretion into the endolymph by vestibular dark cells (163, 304). K+ cycling in the cochlea and vestibular labyrinth, however, is not limited to K+ efflux through hair cells and K+ secretion by the stria vascularis and vestibular dark cells. Additional pathways accomplish K+ and Na+ reabsorption from the endolymph. In the cochlea, these additional pathways are provided by Reissner's membrane and by outer sulcus epithelial cells (Fig. 2A) and, in the vestibular system, by semicircular canal epithelial cells and transitional epithelial cells (163, 304).
K+ flux through hair cells.
Sensory transduction in the cochlea and vestibular labyrinth depends on mechanically gated ion channels of hitherto elusive molecular identity that are located in the hair bundles of hair cells (32). Candidates that have been considered and rejected include acid-sensing ion channels (194) and members of the epithelial Na+ channel (ENaC)/degenerin (ENaC/DEG) superfamily, which are widely distributed in the central and peripheral nervous system and are also found in the cochlea (34, 70, 174, 335) and vestibular labyrinth (138, 202, 334). The current view, however, appears to favor the involvement of transient receptor potential (TRP) channels such as TRPN1, TRPV4, TRPML3, and TRPA1 for mechanotransduction (32).
Opening of the so far elusive transduction channels supports influx of K+ from the endolymph into hair cells, which depolarizes the basolateral membrane of the hair cell (Fig. 2, B and C). Influx of K+ from the endolymph into hair cells is balanced by efflux of K+ from hair cells via K+ channels into interstitial spaces that are continuous with the perilymph. The molecular entities of K+ efflux channels depend on the type of hair cell.
K+ efflux from cochlear inner hair cells involves the voltage-gated K+ channel KCNQ4 and the large-conductance (BK) Ca2+-activated K+ channel KCNMA1 (a BK channel) (48, 121, 122, 240, 278) (Fig. 2B). In addition, KCNQ4 contributes to the resting membrane potential of inner hair cells and thereby ensures the maintenance of the resting cytosolic Ca2+ concentration (156, 157, 189). Several splice variants of KCNQ4 are expressed in the inner ear, of which one variant, KCNQ4v3, is preferentially expressed in the high-frequency region base of the cochlea (12, 142, 143). Mice lacking KCNQ4 or expressing dominant negative mutations develop normal hearing but later in life suffer progressive hearing loss, which indicates that KCNQ4 is nonessential for basic inner hair cell function but required for the maintenance of hearing (107). Consistently, mutations of KCNQ4 cause progressive high-frequency hearing loss in humans (33, 123, 269). Similarly, mice lacking KCNMA1 develop normal hearing but then suffer progressive hearing loss, indicating that also KCNMA1 is nonessential for basic inner hair cell function (190, 216). Loss of the KCNMB1 β-subunit, which associates with the KCNMA1 α-subunit, appears to have no effect on hearing (216).
The K+ channels KCNQ4 and KCNMA1 also mediate K+ efflux from outer hair cells (Fig. 2C). In contrast to inner hair cells, both channels play a critical role for basic cell function in outer hair cells (24, 87, 155, 158). Mice lacking KCNMA1 or KCNQ4 loose outer hair cells, but not inner hair cells, during the progressive loss of hearing (52, 107, 216). The K+ channels KCNMA1 and KCNQ4 are apparently essential for the survival of outer hair cells.
KCNQ4 associates with the KCNE1 β-subunit and possibly with other KCNE subunits that are expressed in hair cells (255). This interaction may be critical for KCNQ4 function given that the KCNE1 mutation KCNE1(D76N) impairs KCNQ4 function and causes Jervell and Lange-Nielsen syndrome, consisting of life-threatening cardiac arrhythmias and deafness. Other mutations, such as KCNE1(S74L), which do not impair KCNQ4 function, cause Romano-Ward syndrome, consisting of arrhythmias without deafness (255).
K+ buffering near hair cells.
Hair cells and neurons in the cochlea and vestibular labyrinth maintain their resting membrane potential via K+ channels in conjunction with high cytosolic and low extracellular K+ concentrations. Uncontrolled increases of the K+ concentration in the extracellular fluid are expected to affect the membrane potential and responsiveness of hair cells and neurons. Stimulation of cochlear and vestibular hair cells leads to measurable increases in the extracellular K+ concentration in the surrounding perilymph (104, 282). It is conceivable that K+-buffering mechanisms limit the magnitude of these increases. In general, multiple mechanisms have been recognized to limit the amplitude of K+ concentration changes in the extracellular environment near neurons. The predominant mechanism is diffusion into unobstructed open fluid spaces. Current measurements in scala tympani perilymph support the concept that the perilymph serves as an unobstructed open fluid space in the buffering of K+ (336). Furthermore, a strategic localization of K+ channels that differ in their rectification can provide buffering of localized K+ increases. Inward-rectifying K+ channels are well suited as uptake mechanisms in K+ buffering since they conduct K+ influx more efficiently than K+ efflux. A local increase in the extracellular K+ concentration may set the local K+ equilibrium potential below the membrane potential, which promotes K+ influx into the buffering cell. The ensuing elevation of the cytosolic K+ concentration sets the K+ equilibrium potential above the membrane potential and promotes K+ efflux preferentially through less inward-rectifying K+ channels or outward-rectifying K+ channels. Such a mechanism has been described in Müller glia of the retina (116, 208). It is conceivable that a similar mechanism is present in the organ of Corti. Deiter's cells have a membrane potential of −76 mV (186), which is near the K+ equilibrium potential. They express the inward-rectifying K+ channel KCNJ10 (Kir4.1), which is particularly abundant in the membrane area that faces KCNQ4 expressed in outer hair cells (83, 216) (Fig. 2C). K+ exit mechanisms in Deiter's cells may include outward-rectifying K+ channels (181) and the K+/Cl− cotransporters SLC12A6 (KCC3) and SLC12A7 (KCC4) (15, 16) (Fig. 2C). Moreover, Deiter's cells are connected to neighboring supporting cells via gap junctions. K+ could thus be dispersed via gap junctions among epithelial cells that include Deiter's cells (Fig. 2A, DC), Claudius' cells (Fig. 2A, CC), Hensen's cells (Fig. 2A, HC), and outer sulcus cells (Fig. 2A, OC).
SLC12A6 and SLC12A7 may serve as a release mechanism for KCl not only in K+ buffering but also in cell volume regulation (131). The recent claim that SLC12A6 and SLC12A7 serve as a K+ uptake mechanism in Deiter's cells, however, would require an unusually low cytosolic Cl−, which has not been shown thus far (15, 16). Consistent with a role of SLC12A6 and SLC12A7 in cell volume regulation is the finding that mice lacking either transporter hear normal at the onset of hearing but suffer from a more or less early onset of hearing loss (15, 16).
K+ uptake from the perilymph.
K+ released from sensory hair cells may travel along multiple pathways toward the spiral ligament in the lateral wall (246) (Fig. 2A). Pathways that avoid supporting cells or that involve buffering by Deiter's cells or neighboring cells and lead through the open perilymph space of scala tympani are supported by current measurements (336). An additional pathway may involve uptake of K+ into Deiter's cells, dispersion of K+ among Deiter's cells, Hensen's cells, and outer sulcus cells via gap junctions (Fig. 2A), and release of K+ into the interstitial space of the spiral ligament that is continuous with perilymph (110).
Uptake of K+ from the interstitial space of the spiral ligament occurs via specialized fibrocytes (named I–V within the spiral ligament in Fig. 2A). Fibrocyte types II, IV, and V express Na+/K+-ATPase, the Na+-2Cl−-K+ cotransporter SLC12A2 (NKCC1), and the Cl− channels CLCNKA and CLCNKB (36, 154, 177, 206, 230) (Fig. 2D). Although functional data from fibrocytes are lacking, the resemblance of this array of transporters with the basolateral membrane of strial marginal cells and vestibular dark cells suggests that fibrocytes take up K+ from the perilymph. Gap junctions, in particular GJB2 (CX26) and GJB6 (CX30), connect fibrocyte types II, IV, and V among each other as well as to fibrocyte type I cells and basal and intermediate cells of the stria vascularis. Gap junctions form a network in the lateral wall that is thought to provide a pathway for K+ from the sites of uptake into fibrocyte types II, IV, and V to the sites of release from strial intermediate cells into the intrastrial fluid space.
K+ secretion into the endolymph.
Strial marginal cells and vestibular dark cells take up K+ from the intrastrial fluid space and secrete it into the endolymph (Fig. 2D). K+ secretion by strial marginal cells and vestibular dark cells occurs via equivalent mechanisms (311). Both epithelial cells take up K+ across the basolateral cell membrane via the Na+-2Cl−-K+ cotransporter SLC12A2 (NKCC1) and Na+/K+-ATPase and secrete K+ across the apical membrane via the K+ channel KCNQ1/KCNE1 (Fig. 2D). Na+ and Cl− taken up via the Na+-2Cl−-K+ cotransporter is recycled in the basolateral membrane via Na+/K+-ATPase and the Cl− channels ClC-Ka/ barttin (CLCNKA/BSND) and ClC-Kb (CLCNKB/BSND) (Fig. 2D). The following paragraphs focus on the Na+-2Cl−-K+ cotransporter, K+ channels, and Cl− channels that are essential for K+ secretion in strial marginal cells and vestibular dark cells.
NA+/K+-ATPASE AND K+/H+-ATPASE.
Strial marginal cells and vestibular dark cells absorb K+ from the intrastrial space and perilymph via Na+/K+ ATPase and the Na+-2Cl−-K+ cotransporter (170, 306). Na+/K+ ATPase takes up K+ and establishes a Na+ gradient that energizes further uptake of K+ via SLC12A2. The Na+/K+-ATPase in strial marginal cells and vestibular dark cells as well as in fibrocytes of the spiral ligament consists of the subunits ATP1A1, ATP1B1, and ATP1B2 (173, 230) (Fig. 2D)
Inhibition of Na+/K+-ATPase with ouabain inhibits K+ secretion and consequently abolishes the endocochlear potential (126, 129). Strial marginal cells appear to express gastric K+/H+-ATPase in addition to Na+/K+-ATPase (135, 235). The functional significance of K+/H+-ATPase for the generation of the endocochlear potential is currently unclear since very high concentrations of K+/H+-ATPase inhibitors were necessary to affect the endocochlear potential.
Strial marginal cells and vestibular dark cells absorb K+ from the intrastrial space and perilymph via the Na+-2Cl−-K+ cotransporter SLC12A2 (NKCC1) (162, 306, 313) (Fig. 2D). SLC12A2 is sensitive to the loop diuretics furosemide and bumetanide and to their analog piretanide (162, 313). SLC12A2 is an essential transporter for K+ secretion and endolymph production. Mice that lack SLC12A2 fail to produce endolymph, which leads to the collapse of Reissner's membrane onto the stria vascularis and organ of Corti (41, 43, 57).
Strial marginal cells and vestibular dark cells secrete K+ into the endolymph via the K+ channel KCNQ1/KCNE1 (165, 306) (Fig. 2D). KCNQ1/KCNE1 is a slowly activating delayed rectifier that carries the slow delayed rectifier K+ current and requires the assembly of the pore-forming KCNQ1 α-subunit with the KCNE1 β-subunit (11, 223). Mice lacking functional KCNE1 or KCNQ1 fail to produce endolymph, which leads to a collapse of Reissner's membrane onto the stria vascularis and organ of Corti due to loss of K+ secretion in the presence of ongoing reabsorptive processes (22, 139, 140, 294). Similar observations have been made in human patients (61). Homozygous or heterozygous compounding mutations of KCNE1 or KCNQ1 lead to Jervell and Lange-Nielsen syndrome, which is characterized by deafness, prolonged cardiac action potentials, and potentially fatal cardiac arrhythmias (103, 182, 231, 233). Consistently, pharmacological inhibition of the KCNQ1/KCNE1 channel leads to hearing loss (79).
The rate of K+ secretion is controlled by the K+ concentration on the apical and basolateral membrane and by cell volume, pH, and a variety of receptors and signaling mechanisms. Transepithelial currents and currents through the apical KCNQ1/KCNE1 K+ channel are enhanced by lowering the apical K+ concentration or increasing the basolateral K+ concentration or by lowering the osmolarity on the basolateral side (161, 306, 308, 317). Furthermore, the rate of K+ secretion is increased by β1-adrenergic receptors via cAMP-dependent stimulation of the KCNQ1/KCNE1 K+ channel (258, 259, 309, 310). Conversely, muscarinic and purinergic receptors suppress K+ secretion (307). Purinergic P2Y4 receptors decrease currents through the KCNQ1/KCNE1 K+ channel via protein kinase C (160, 166). In addition, KCNE1/KCNQ1 K+ channel activity is stimulated by the serum- and glucocorticoid-inducible kinase SGK1 (20, 50), which may contribute to the stimulation of cochlear ion transport and hearing improvement by glucocorticoids and mineralocorticoids (137, 280). The channel is also inhibited by estrogens (298), which may contribute to the inhibitory effect of those hormones on cochlear transport (136).
The rate of K+ secretion may further be regulated by trafficking of KCNE1/KCNQ1 K+ channels to the apical membrane of strial marginal cells, which requires the participation of lysosomal integral membrane protein II (LIMPII) (115). LIMPII is a transmembrane glycoprotein that is mainly located in lysosomal and endosomal membranes (127). Mice lacking LIMPII suffer from progressive hearing loss correlated with a loss of surface expression of KCNQ1/KCNE1 in the apical membrane of marginal cells (115). In addition, mice lacking LIMPII suffer from uni- or bilateral hydronephrosis due to hypertrophy of the smooth muscle layer at the ureteropelvic junction (65).
K+ secretion by strial marginal cells and vestibular dark cells require Cl− to recycle in the basolateral membrane via a major Cl− conductance (306, 314) (Fig. 2D). This Cl− conductance is composed of the Cl− channels CLCNKA/BSND and CLCNKB/BSND (6, 53, 154, 167, 206, 218, 264, 265). Cl− channels CLCNKA/BSND and CLCNKB/BSND consist of the pore-forming CLCNKA α-subunits and CLCNKB and the BSND β-subunit (53, 228). Mutations of BSND reduce channel conductivity and surface expression and thereby cause Bartter's syndrome type 4, which is characterized by deafness and renal salt wasting (14, 53). Similarly, simultaneous mutations of CLCNKA and CLCNKB also lead to Bartter's syndrome type 4, whereas mutations of CLCNKB lead to Bartter's syndrome without deafness (171, 226, 237). This observation is consistent with the finding that CLCNKA and CLCNKB are coexpressed in cells of the inner ear but not the kidney and with the notion that the two channels can substitute for each other in the inner ear (206) but not in the kidney (see below).
K+ and Na+ Reabsorption
Homeostasis of high K+ and low Na+ concentrations in the endolymph is maintained by K+ secretion and Na+ and K+ reabsorption. Reabsorption of K+ is not limited to the pathways through inner, outer, and vestibular hair cells. Indeed, currents generated by the stria vascularis in the cochlea flow not only through hair cells but also through the outer sulcus and through Reissner's membrane (222, 336). Consistently, outer sulcus and Reissner's membrane epithelial cells reabsorb Na+ and K+ from the endolymph (138, 159). Outer sulcus cells take up Na+ and K+ via apical nonselective cation channels, BK channels, and small-conductance K+ channels as well as P2X2 receptor-gated nonselective cation channels. They release Na+ and K+ across the basolateral membrane via Na+/K+-ATPase and K+ channels, respectively (25, 26, 136). Subunits of ENaC (SCNN1) may contribute to the apical nonselective cation channels (Fig. 2A), although the channel involved is not the typical Na+-selective and amiloride-sensitive ENaC channel (25, 70).
Reissner's membrane epithelial cells take up Na+ via the amiloride-sensitive Na+ channel ENaC and extrude Na+ across the basolateral membrane via Na+/K+-ATPase (138) (Fig. 2A). The endocochlear potential contributes to the driving force of cation reabsorption in outer sulcus and Reissner's membrane epithelial cells much like it contributes to the transduction current through inner and outer hair cells.
Na+ and K+ reabsorption has also been found in the vestibular labyrinth. Cation reabsorption in vestibular transitional cells and semicircular canal epithelial cells bear some resemblance to cation reabsorption in outer sulcus and Reissner's membrane epithelial cells, respectively. Vestibular transitional cells reabsorb Na+ and K+ via apical P2X2 receptor-gated nonselective cation channels and extrude Na+ and K+ across the basolateral membrane via Na+/K+-ATPase and K+ channels, respectively (136, 312, 318).
Semicircular canal epithelial cells in the vestibular labyrinth reabsorb Na+ via ENaC, release Na+ across the basolateral membrane via Na+/K+-ATPase, and recycle K+ in the basolateral membrane via K+ channels (201, 202). Na+ reabsorption in semicircular canal epithelial cells is under the control of glucocorticoids but not mineralocorticoids (201, 202). ENaC is activated by the transmembrane serine protease TMPRSS3 (72). A defect of TMPRSS3 leads to deafness (72), which may, however, involve dysregulation of further transporters or channels besides ENaC activity. Loss of function mutations of ENaC do not lead to an inner ear phenotype, which is consistent with the presence of alternative Na+ reabsorption pathways in outer sulcus epithelia cells and transitional cells (136, 159).
Generation of the Endocochlear Potential
Mechanical stimuli associated with sound, head position, or gravity are transduced into electrical signals by sensory hair cells in the cochlea and vestibular labyrinth. Mechanically induced channel openings permit an influx of K+ from the endolymph into the hair cell. The driving force of this current is roughly the sum of the basolateral membrane potential of the hair cell and the transepithelial potential. The transepithelial potential in the cochlea, called the endocochlear potential, is enormous: +80 mV (Fig. 2A). For cochlear inner and outer hair cells, the driving force for sensory transduction is 120 mV (−40 mV + 80 mV) and 150 mV (−70 mV + 80 mV), respectively (38, 189). Driving forces for sensory transduction in the vestibular labyrinth are smaller due to the smaller endovestibular potential of 3–7 mV (130, 185, 220).
The endocochlear potential is a transepithelial potential that is generated by the stria vascularis (275, 295, 316). The stria vascularis is functionally a two-layered epithelium composed of a layer of marginal cells and a layer of basal cells that is penetrated by a capillary network (99). Marginal cell junctions contain a multitude of different claudins, whereas tight junctions between basal cells contain only claudin 11 (CLDN11) (58, 114). Tight junctions among basal cells define an inner membrane facing the intrastrial space and an outer membrane facing the spiral ligament (Fig. 2). The inner membrane is connected via gap junctions to strial intermediate cells, and strial intermediate cells are connected via gap junctions to strial pericytes and endothelial cells (108, 262). Gap junctions ensure that intermediate cells are electrically a part of the basal cell barrier. The outer membrane of basal cells is connected by gap junctions to type I fibrocytes of the spiral ligament.
The endocochlear potential is essentially a K+ equilibrium potential that is generated by the K+ channel KCNJ10 (Kir4.1) in intermediate cells of the stria vascularis (Fig. 2D) in conjunction with a very low K+ concentration of intrastrial fluid and a high cytosolic K+ concentration in intermediate cells (169, 263). A number of key findings provide support for this model. First, the endocochlear potential and the KCNJ10 K+ channel in intermediate cells share the same sensitivities to a panel of K+ channel blockers (164, 262, 266). Second, the endocochlear potential can be measured across the basal cell barrier (221). Third, the expression of KCNJ10 correlates with the presence of the endocochlear potential in KCNJ10 knockout and pendrin knockout mouse models and in normal development (82, 169, 215, 305, 315). Fourth, loss of claudin 11, which is the only known claudin in basal cell tight junctions, renders the basal cell barrier leaky and leads to a loss of the endocochlear potential (67, 114). Fifth, increases of the K+ concentration in the intrastrial fluid space suppress the endocochlear potential. Such increases can be achieved by vascular perfusion of solutions containing elevated K+ concentrations, inhibitors of Na+/K+-ATPase (ouabain), or inhibitors of the Na+-2Cl−-K+ cotransporter (furosemide or bumetanide) (118, 126, 128, 129, 164). Finally, loss of GJB6 [connexin (Cx)30], which renders capillaries in the stria vascularis leaky to the intrastrial space, leads to a loss of the endocochlear potential (31). Collectively, these findings support the model that the endocochlear potential is a K+ equilibrium potential that is generated by the K+ channel KCNJ10.
Marginal cells of the stria vascularis and fibrocytes of the spiral ligament play important supporting roles in the generation of the endocochlear potential. Fibrocytes of the spiral ligament, which are connected via basal cells to intermediate cells, ensure a high cytosolic K+ concentration in strial intermediate cells. Strial marginal cells reabsorb K+ from the intrastrial fluid spaces and keep the K+ concentration in the intrastrial fluid spaces as low as 2 mM (164, 263, 306).
Gap Junctional Networks
Several major networks of cells that are connected by gap junctions have been recognized in the cochlea (109). Notably excluded from these networks are marginal cells of the stria vascularis, inner hair cells and outer hair cells that are neither connected among each other nor to any of their neighbors. The importance of gap junctions for cochlear function is underscored by the fact that mutations of GJB2 (Cx26) and GJB6 (Cx30) are the most prevalent causes of hereditary childhood deafness consistent with the contribution of GJB2 and GJB6 to all major gap junctional networks in the cochlea (39, 42, 68, 73, 191, 193, 331).
One major network of gap junction interconnected cells links different types of fibrocytes in the spiral ligament as well as basal and intermediate cells, pericytes, and endothelial cells of the stria vascularis (Fig. 2A, spiral ligament). A major purpose of this network is to connect sites of K+ uptake in fibrocyte types II, IV, and V to the site of K+ release in intermediate cells of the stria vascularis. Most gap junctions in this network are formed by heteromeric complexes of GJB2 and GJB6 (2, 59, 109, 134, 325). In addition, endothelial cells of the stria vascularis express GJA1 (Cx43) and GJA7 (Cx45) (29, 134, 260) and fibrocytes of spiral ligament express GJB3 (Cx31) and GJB1 (Cx32) (150, 326). Mutations of GJA1, GJB1, and GJB3 are also associated with deafness (147, 148, 251, 326).
Mice that lack GJB6 are profoundly deaf despite the continued presence of GJB2 (277). The assumed limited gap junction coupling mediated by the remaining GJB2 and other connexins appears to be sufficient for the cycling of K+ but insufficient to prevent leakiness of strial capillaries and breakdown of the endocochlear potential (31). Consequently, mice lacking GJB6 failed to develop an endocochlear potential but had normal endolymphatic K+ concentrations at least at a young age (277). Interestingly, the insufficiency of gap junction coupling, which is associated with leaky capillaries in mice lacking GJB6, can be restored by overexpression of GJB2. Mice lacking GJB6 and overexpressing GJB2 develop a normal endocochlear potential and have normal hearing (3).
Two further networks are formed by epithelial cells in and adjacent to the organ of Corti: the medial and lateral networks (98, 109, 247). Most gap junctions in these networks are formed by GJB2 and GJB6 (2, 59, 109, 134, 325). In addition, some cells express GJA1 (147, 260). The lateral network of gap junction interconnected cells in the organ of Corti includes outer pillar cells, Deiter's cells, Hensen's cells, Claudius' cells, outer sulcus cells, and root cells (98, 109) (Fig. 2A). The major purpose of this network may be metabolic coupling in addition to buffering of K+ that is released from outer hair cells in response to sound stimulation (104).
The medial network includes inner pillar cells, supporting cells of the inner hair cells and interdental cells (Fig. 2, A and B). A major purpose of this network is to buffer glutamate, which is the neurotransmitter released from the inner hair cell. Expression of the glutamate uptake transporter SLC1A3 (GLAST) is limited to the immediate neighbor of the inner hair cell, whereas glutamine synthase, a key enzyme in the detoxification of glutamate, is mainly expressed in adjacent cells but not in SLC1A3-expressing cells (55). Gap junctions between SLC1A3-expressing cells and their glutamine synthase-expressing neighbors may be required for the transcellular metabolism of glutamate. Support for the concept of transcellular glutamate buffering comes from the finding that mice lacking SLC1A3 fail to buffer glutamate, which leads to an accumulation of glutamate in scala tympani perilymph during sound stimulation (76). SLC1A3-expressing supporting cells are also the first cells in the organ of Corti to undergo apoptosis in mice that lack GJB2 in this region of the cochlea (30). Further, several deafness-causing mutations of GJB2 and GJB6 impair the transfer of organic molecules but do not impede ionic coupling, which implies that these mutations do not affect K+ cycling but could impair metabolic coupling and glutamate buffering (31, 332).
Several reasons may account for the intriguing observation that the loss of function of either GJB2 or GJB6 leads to deafness rather than simply being compensated by the remaining connexin-forming homomeric gap junctions (30, 124, 277). First, loss of GJB6 has been shown to suppress the protein expression of GJB2, which reduces intercellular coupling more than predicted by the simple omission of GJB6 (3). Mutations may also exert dominant negative effects on the function of wild-type isoforms (68, 210). Second, heteromeric gap junctions formed from GJB2 and GJB6 have slightly different biophysical properties than homomeric gap junctions (257). The finding that overexpression of GJB2 can rescue hearing of mice lacking GJB6 suggests that the biophysical differences between hetero- and homomeric GJB2 gap junctions are less important than the fact that loss of GJB6 leads to a loss of GJB2 expression and a reduction in intercellular coupling (3).
The transduction channel in hair cells is a Ca2+-permeable nonselective cation channel. Although K+ is the major charge carrier, the transduction current is in part carried by Ca2+ and the reliability of the transduction process itself depends on the constancy of Ca2+ concentrations in the endolymph (Table 1). Both elevated and reduced concentrations of Ca2+ have been shown to suppress transduction currents and microphonic potentials (187, 271). Furthermore, Ca2+ homeostasis of vestibular endolymph during development affects the formation of otoconia, which are necessary for the detection of gravity and linear acceleration (106, 152). Consistent with the importance of Ca2+ homeostasis in the endolymph are the observations that mice and guinea pigs with reduced or elevated endolymphatic Ca2+ concentrations are deaf and have vestibular deficits (120, 185, 315, 324).
The endolymphatic Ca2+ concentration appears to be controlled by secretory and reabsorptive mechanisms. Ca2+ reabsorption may occur through paracellular and transcellular pathways and may at least in part be driven by the endocochlear potential (91). In general, transepithelial Ca2+ transport may employ Ca2+-permeable channels as Ca2+ uptake mechanisms, Ca2+ binding proteins as Ca2+ buffers in the cytosol, and Ca2+-ATPases or Na+/Ca2+ exchangers as Ca2+ extrusion mechanisms. Ca2+-ATPases appear to be most suitable for Ca2+ extrusion into the endolymph due to the low Na+ concentration in the endolymph (Table 1), which does not provide a driving force for Ca2+ extrusion via Na+/Ca2+ exchangers. Consistently, Ca2+ secretion into the endolymph has been shown to depend on Ca2+-ATPases rather than on Na+/Ca2+ exchangers (92, 324) and loss of function of Ca2+-ATPase ATP2B2 (PMCA2) leads to deafness and to a reduction in the endolymphatic Ca2+ concentration (253, 324).
Among the many different epithelial cells lining cochlear and vestibular endolymph, the cells best understood to be involved in the homeostasis of endolymph Ca2+ include outer hair cells in the cochlea and semicircular canal duct epithelial cells in the vestibular labyrinth, although it is currently unclear whether outer hair cells contribute to the homeostasis of bulk endolymph or only to the homeostasis of the endolymph in the nearest vicinity of the hair bundle. Nevertheless, outer hair cells have been shown to secrete Ca2+ into the endolymph (328). This Ca2+ secretion is required to remove Ca2+ from the cytosol of the hair bundle and to maintain an appropriate Ca2+ concentration in the endolymph surrounding the bundle (7, 84). Outer hair cells express Ca2+-ATPase ATP2B2 in the stereocilia (49, 64) (Fig. 2C), a high concentration of Ca2+ binding proteins in the cytosol (74), and Ca2+-permeable channels in the basolateral membrane including TRPC, TRPV1, TRPV4, and L-type and non-L-type Ca2+ channels (45, 144, 157, 207, 234, 333). Mice lacking TRPV4 develop normal hearing consistent with a redundancy of Ca2+-permeable channels. However, they suffer from a delayed-onset hearing loss and vulnerability to acoustic injury (261). It is currently unclear whether the delayed-onset hearing loss and vulnerability to acoustic injury are due to the loss of TRPV4 in outer hair cells or due to the loss of the channel from other cells including inner hair cells and spiral ganglion neurons (234).
Semicircular canal duct epithelial cells in the vestibular labyrinth have been shown to reabsorb Ca2+ from the endolymph (180). Ductal epithelial cells express Ca2+-permeable TRPV5 and TRPV6 channels, Ca2+ binding proteins, Na+/Ca2+ exchangers, and Ca2+-ATPases (327). Consistent with an apical membrane expression of pH-sensitive TRPV5 and TRPV6 Ca2+ channels is the finding that the transepithelial Ca2+ flux was pH sensitive and that endolymph Ca2+ concentrations were elevated in mice that have acidic endolymph due to loss of pendrin (180).
Other epithelial cells in the cochlea and vestibular labyrinth may be involved in endolymph Ca2+ homeostasis since they express Ca2+-ATPases and Ca2+-permeable channels. Whether these cells secrete or reabsorb Ca2+ is currently not clear. Inner hair cells, in contrast to outer hair cells, may be involved in Ca2+ reabsorption (84). The transduction channel may serve as an uptake channel, and Ca2+-ATPase ATP2B1 (PMCA1) in the basolateral membrane may serve as a release mechanism (49). Similarly, vestibular hair cells express in their basolateral membrane ATP2B1 and ATP2B3 (PMCA3). However, inner and vestibular hair cells also express ATP2B2 in hair bundles (49) (Fig. 2B). It is thus unclear whether inner and vestibular hair cells support a transcellular Ca2+ flux. The stria vascularis expresses ATP2B1 and Ca2+-permeable TRPV4, TRPV5, and TRPV6 channels (1, 37, 144, 268, 315, 324). Reissner's membrane and interdental cells express Ca2+ATPases (37, 64, 324). Outer sulcus cells express ATP2B2 (64), and inner and outer sulcus epithelial cells as well as ductal epithelial cells of the semicircular canals express TRPV5 and TRPV6 Ca2+ channels (180, 315). TRPV5 and TRPV6 Ca2+ channels may be located in the apical membrane of at least some cochlear epithelial cells (315).
The pH of the endolymph varies greatly between different regions in the inner ear. In the cochlea and utricle, the endolymphatic pH is slightly alkaline (pH 7.5) (93, 180, 315). In the endolymphatic sac, on the other hand, the pH is more acidic (pH 6.6–7.1) (35, 281). The functional significance of these differences is largely elusive. The presence of these differences, however, underscores that fluid homeostasis in different compartments of the inner ear is controlled by local ion transport in adjacent epithelia rather than via a fluid flow between different compartments of the inner ear (219).
The homeostasis of endolymphatic pH depends on the secretion of H+ and HCO3−. Epithelial cells that express H+-ATPase in their apical membrane include interdental cells of the spiral limbus (Fig. 2A) and strial marginal cells (Fig. 2D) as well as endolymphatic duct and sac epithelial cells (46, 105, 250). Furthermore, epithelial cells that express in their apical membrane the HCO3−-permeable anion exchanger SLC26A4 (pendrin) include spiral prominence and outer sulcus epithelial cells and spindle cells of the stria vascularis as well as endolymphatic duct and sac epithelial cells (46, 54, 305, 330).
The main buffers, at least in the cochlear endolymph, appear to be CO2 and HCO3−. Glycosaminglycans, which are found in high concentrations in the endolymph of the endolymphatic sac, may contribute to pH buffering (89, 203). Proteins, however, that contribute to the buffering capacity of blood plasma appear to play a lesser role in the buffering of cochlear endolymph due to their low concentration (Table 1). Marginal cells of the stria vascularis are a significant local source of CO2 due to their high metabolic rate and their use of the hexose monophosphate pathway (168). Carbonic anhydrases in the stria vascularis, spiral ligament, and spiral limbus capture metabolically derived CO2 and convert it to HCO3− (146, 188, 245). HCO3− generated within the fibrocyte gap junction network may be secreted into the endolymph via the HCO3−-permeable anion exchanger pendrin (SLC26A4). Consistent with HCO3− secretion into the endolymph is the observation that mice lacking pendrin have an acidic endolymphatic pH (180, 315). Furthermore, increased metabolic rates during acoustic stimulation cause an alkalization of the endolymph, which is consistent with an increased rate of HCO3− secretion (92).
Endolymphatic pH homeostasis is necessary for hearing and the prevention of hearing loss, although effects of pH may be indirect. For example, acidification of the endolymph inhibits Ca2+ reabsorption via pH-sensitive TRPV5 and TRPV6 Ca2+ channels and elevates the endolymphatic Ca2+ concentration, which impairs cochlear function (180, 315). Furthermore, acidification enhances free radical stress and promotes hearing loss (270).
Whether mutations of the B1 subunit (ATP6V1B1) or A4 subunit (ATPV0A4) of H+-ATPase cause an alkalinization of endolymph pH is currently unknown. Nevertheless, mutations of either subunit may cause, in humans, a progressive sensorineural hearing loss in addition to renal tubular acidosis (105, 252, 288). The etiologies of these hearing losses, however, are unclear, in particular since mice lacking the B1 subunit (ATP6V1B1) develop normal hearing and show no overt morphological abnormalities in the inner ear (44).
Water transport follows osmotic gradients that are established by the transport, metabolism, or catabolism of solutes (131). Water can permeate most membranes freely with the notable exception of the apical membrane of thick ascending limb and of the cortical collecting duct of the kidney in the absence of the antidiuretic hormone vasopressin. The water permeability of cell membranes depends to a significant extent on the presence of aquaporins, which are water-permeable channels. According to a recent review (184), 13 different aquaporins (AQP0–AQP12) have so far been identified.
A multitude of aquaporins are expressed in the inner ear, including AQP1, AQP2, AQP3, AQP4, AQP5, AQP7, and AQP9 (88, 149, 175, 176, 224, 249, 267, 334). The functional significance of inner ear water channels is largely unclear. Loss of function of AQP1 associated with the Colton blood group does not cause an overt clinical phenotype, although it is associated with a reduction of the urinary concentration capacity (113, 205). Hearing loss or balance difficulties have not been reported in association with the Colton blood group. Neither have hearing loss or balance disorders been reported to be associated with diabetes insipidus due to loss of AQP2. Mice lacking AQP1, AQP3, or AQP5 have normal hearing; however, mice lacking AQP4 have a minor hearing loss of 10 dB at 4–5 wk of age (141) (Fig. 2A). Whether hearing is already impaired earlier is currently unknown.
Functions of Inner Ear Channels and Transporters in Renal Epithelia
Many of the channels, carriers, and pumps accomplishing tranport in the inner ear are similarly expressed in the kidney and participate in renal tubular transport. Accordingly, the hearing loss in patients carrying genetic defects of defined transport molecules may be paralleled by deranged renal acid or electrolyte excretion that affects acid-base or electrolyte homeostasis of the body. Moreover, and possibly related to the different organization and function of the two organs, some of the transport proteins are used for quite different cellular functions in the inner ear and kidney, as discussed below.
The proximal tubule of the kidney (Fig. 3A) reabsorbs ∼60% of filtered NaCl and fluid and most of the filtered amino acids and glucose. K+ channels in the apical cell membrane of proximal tubules (Fig. 3A) contribute to the maintenance of the cell membrane potential during depolarizing Na+-coupled transport (e.g., cotransport of Na+ with amino acids or glucose), thereby stabilizing the electrical driving force for electrogenic Na+ reabsorption.
Henle's loop contributes to the generation of a hypertonic kidney medulla, a prerequisite for urinary concentration. Most importantly, the thick ascending limb of Henle's loop reabsorbs ∼25% of NaCl filtered by the glomeruli without accompanying water reabsorption, thus enhancing interstitial osmolarity (Fig. 3B). The medullary collecting ducts (Fig. 3D) pass the hypertonic kidney medulla. During water retention, water channels allow water to leave the lumen of the collecting ducts, thus leading to urinary concentration.
The distal convoluted tubule (Fig. 3C), connecting tubule (not explicitly shown), and collecting duct (Fig. 3, D and E) allow the fine tuning of renal acid, fluid, and electrolyte excretion. In all nephron segments, the proximal tubule, Henle's loop, distal tubule, and collecting duct K+ channels maintain the cell membrane potential and thus the driving force for electrogenic transport.
K+ channels expressed in both the inner ear and kidney include KCNE1/KCNQ1, KCNJ10, and BK channels (KCNMA1/KCNMB1) (Table 2).
KCNE1 and KCNQ1 have been localized to the brush border of the mid to late proximal tubule (256, 283) (Fig. 3A). Besides their potential role in net K+ secretion into the early proximal tubule (284), they may polarize the brush border membrane and thus maintain the electrical driving force for Na+-coupled transport (132, 133). Studies (283, 284) in knockout mice indeed revealed that lack of functional KCNE1/KCNQ1 K+ channels leads to moderate impairment of electrogenic Na+-glucose cotransport in proximal tubules (see Fig. 3A). KCNE1 may interact with additional K+ channels, especially in the early proximal tubule, where most of the glucose, amino acids, and phosphate are reabsorbed by electrogenic cotransport with Na+ and where KCNE1 but not KCNQ1 was detected. In the early proximal tubule, KCNE1 is likely to coassemble with another KCNQ isoform (Fig. 3A), similar to what has been recently shown for outer hair cells (255). KCNQ1-independent function of KCNE1 may explain the more severe phenotype (e.g., renal Na+ and glucose loss) in mice lacking KCNE1 compared with KCNQ1 (283, 284). Thus, whereas KCNQ1/KCNE1 K+ channels serve to establish high K+ concentrations in the endolymph of the inner ear, they serve the very different function of stabilizing the membrane potential and thus electrogenic reabsorption of Na+ in the proximal tubule of the kidney, with the secreted K+ being subsequently reabsorbed by paracellular routes.
KCNJ10 is expressed in the basolateral cell membrane of renal distal tubules, including the thick ascending limb (97, 151, 272, 273) (see Fig. 3B). These channels are highly sensitive to cytosolic pH and are thus thought to link K+ metabolism with acid-base balance (21). To our knowledge, however, no data are available on K+ or acid-base balance in mice lacking KCNJ10. Whereas KCNJ10 is considered to be of primary importance for the endocochlear potential of the inner ear, its precise role in the kidney remains to be defined.
BK channels (KCNMA1) are expressed in the renal vasculature and tubular system (71, 198) (Fig. 3D). In the latter, they contribute to K+ secretion into the luminal fluid. BK channels in the luminal membrane of the distal nephron (Fig. 3D) are involved in K+ homeostasis in response to a high-K+ diet (9, 179, 212) and mediate renal K+ excretion in response to enhanced tubular flow rates (71, 198–200, 212, 274, 323). In the mouse, the β1-sbunit of KCNMB1 was found exclusively in the connecting tubule (200). Notably, this β1-subunit confers protein kinase G activation of BK channels, dramatically increases the Ca2+ sensitivity of the channel, and leads to the activation of the channel at more negative potentials, thereby presumably enhancing the ability of the pore-forming α-subunit to induce significant K+ excretion in the distal nephron under physiological conditions (for a review, see Ref. 198). Moreover, mice lacking the α-subunit (KCNMA1) (212) but also mice deficient in the β1-subunit (KCNMB1) (199) exhibit blunted flow-induced renal K+ excretion. These studies implied a role for BK channels (KCNMA1/KCNMB1) in flow-induced renal K+ excretion and K+ homeostasis. BK channels are also expressed in other tubular segments, where their function is less clear. Whether circulation or flow of the endolymph similarly affects BK channel activity in the inner ear remains to be determined.
The Na+-2Cl−-K+ cotransporter NKCC1 (SLC12A2) is highly expressed in glomeruli of more mature nephrons (286) and may participate in the macula densa-dependent regulation of renin release (23, 66, 302). A closer look at SLC12A2 knockout mice more recently revealed that they suffer from hypotension, which was proposed to relate in part to an impaired responsiveness of the kidney to aldosterone and vasopressin (302).
The Na+-2Cl−-K+ cotransporter NKCC2 (SLC12A1), which is strongly expressed in the luminal membrane of the thick ascending limb (40) (Fig. 3B), is responsible for most of the NaCl reabsorption in that segment and is a prerequisite for the ability of the kidney to dilute and concentrate urine. Accordingly, genetic defects of SLC12A1 lead to isosthenuria and severe renal salt loss (80, 204, 238, 287). SLC12A1 is not expressed in the inner ear, and lack of functional SLC12A1 does not lead to hearing loss. Conversely, SLC12A2 deficiency, as discussed above, leads to deafness without leading to overt renal salt wasting (43, 57). Importantly, both SLC12A1 and SLC12A2 are inhibited by loop diuretics such as furosemide, and thus inhibition of Na+-2Cl−-K+ cotransport in the inner ear during excessive doses of loop diuretics leads to an accumulation of K+ in the intrastrial space, which abolishes the endocochlear potential (129) and leads to hearing loss (81, 96, 322). Much lower doses are sufficient to inhibit luminal SLC12A1 in the thick ascending limb, since the drug accumulates in the tubular fluid as a consequence of efficient secretion into proximal tubular fluid and fluid reabsorption along the tubule. Thus, natriuretic and diuretic actions can be achieved without hearing loss.
The Cl− channel CLCNKA/BSND is expressed in the basolateral membrane of thin ascending limbs (not shown), whereas the Cl− channel CLCNKB/BSND is expressed in the basolateral membrane of thick ascending limbs of Henle's loop (53, 297) (see Fig. 3B). In the mouse, ClC-K1 (the rodent ortholog of CLCNKA) is also expressed in the thin ascending limb. Knockout of ClC-K1 in mice results in nephrogenic diabetes insipidus, establishing that ClC-K1 has a role in urine concentration and that the countercurrent system in the inner medulla is involved in the generation and maintenance of a hypertonic medullary interstitium (171). In the thick ascending limb, basolateral CLCNKB/BSND contributes to transcellular NaCl reabsorption. Defects of CLCNKB lead to the renal salt wasting of classical Bartter syndrome without hearing impairment (119, 237). The phenotype of patients suffering from defective CLCNKB (119, 237), however, is less severe than the phenotype of patients suffering from antenatal Bartter syndrome due to defective SLC12A1 (238) or apical K+ channel ROMK (239). Genetic defects of BSND lead to renal salt wasting together with deafness (14) (Table 2).
Voltage-clamp experiments disclosed that a common (prevalence 20% in Caucasians and 40% in Africans) variant of the CLCNKB gene leading to the replacement of threonine by serine at the amino acid position 481 of the ClC-Kb protein (ClC-KbT481S) dramatically increases ClC-Kb Cl− channel activity (100). Expression of the mutated channels should decrease the cytosolic Cl− concentration and thus enhance the driving force and transport rate of Na+-2Cl−-K+ cotransport. As a result, the gene variant may lead to enhanced transport in the inner ear and kidney. The gene variant was associated with increased blood pressure in one study (101) on a population of largely young, healthy individuals but not in two other studies (117, 244) on more elderly populations. The same gain of function mutation was associated with a slight but significant delay of hearing loss in female humans, whereas no significant differences were observed between male carriers and noncarriers of the mutation (60).
Thus, whereas CLCNKA/BSDN and CLCNKB/BSDN serve the recycling of Cl− across the basolateral membrane of marginal cells of the inner ear to maintain the uptake of K+ via SLC12A2, in the kidney the two channels serve the very different function of basolateral net transport of Cl− along the ascending thin and thick limb.
The K+-Cl− cotransporter KCC4 (SLC12A7) is found along the basolateral cell membrane in several nephron segments (289) (see Fig. 3, A and E). KCC4 is colocalized with KCC3 (SLC12A6) in basolateral cell membranes of the proximal tubule (15, 47) (Fig. 3A), where it may contribute to proximal tubular cell volume regulation. This cotransporter is also important for Cl− recycling in type A intercalated cells (Fig. 3E). In accordance with a crucial role of SLC12A7 for KCl release in K+ buffering and volume regulation in both the kidney and inner ear, mice lacking KCC4 suffer from renal tubular acidosis (15, 102, 196) and deafness (15, 16) (Table 2). Whether KCC4 is of similar pathophysiological significance in humans remains elusive.
ENaC (Fig. 3D, SCNN1) is mainly expressed in the luminal membrane of the aldosterone-sensitive distal nephron, where ∼1–3% of filtered Na+ is reabsorbed. ENaC is of critical importance for renal Na+ reabsorption and secondary K+ excretion and, thus, for salt and K+ homeostasis and blood pressure regulation. Patients carrying loss of function mutations of ENaC (“dominant” pseudohypoaldosteronism type 1) as well as knockout mice for ENaC subunits (SCNN1A, SCNN1B, and SCNN1C) suffer from renal salt wasting (69, 90). As mutant mice die soon after birth, it is still elusive whether the mutated gene would induce hearing loss. Patients with gain of function mutations of ENaC (Liddle's syndrome) suffer from hypertension (19, 77, 145, 225, 319, 320) but are not known to suffer from deafness.
Gap junctional channels.
The classic gap junction channels have been shown by a freeze-fracture study (125) in the proximal tubule, and some of the ubiquitous connexin isoforms [GJA4 (Cx37), GJA5 (Cx40), GJA1 (Cx43), and GJA7 (Cx45)] have subsequently been identified in the kidney and localized to mainly vascular and glomerular components (8, 10). In the so-called juxtaglomerular apparatus, Cx40 and Cx43 have been implicated in the regulation of renin secretion (75, 296).
Moreover, GJB6 protein (Cx30), probably in the form of luminal hemichannels, was found to be expressed in renal tubular epithelial cells (Fig. 3E) and inserted into the apical cell membrane, particularly of intercalated cells (172). GJB6 proteins were upregulated by a high-salt diet in the distal nephron (172) (Fig. 3E). It has thus been speculated that GJB6 may function as an apical hemichannel allowing the passage of ATP and having a potential inhibitory role in the regulation of salt reabsorption in the distal nephron (172). Along those lines, a recent study (211) has showed that mice deficient for the ATP receptor P2Y2 present a salt-resistant form of arterial hypertension that is associated with facilitated renal Na+ and fluid reabsorption. It would be interesting to learn whether renal electrolyte excretion and/or blood pressure are altered in patients with defective GJB6 or in GJB6 knockout mice.
Together with the intestine and bone, the kidney is of primary importance for body Ca2+ homeostasis. Hormone-regulated renal Ca2+ reabsorption is mainly localized to the late distal convoluted tubule and connecting tubule, where TRPV5 channels, expressed in the luminal membrane (Fig. 3C), accomplish Ca2+ uptake and Ca2+-ATPases (PMCA1B; Fig. 3C) as well as Na+/Ca2+ exchangers (NCX1) basolateral exit (85). Accordingly, TRPV5-deficient mice suffer from impaired Ca2+ reabsorption leading to renal Ca2+ loss (86, 209). Notably, TRPV6, which mediates Ca2+ reabsorption in the intestine, is also expressed in the luminal membrane of the collecting duct (i.e., downstream of the segments primarily expressing TRPV5; Fig. 3D) and may contribute to renal tubular Ca2+ reabsorption. Accordingly, Ca2+ reabsorption in the collecting duct limits renal Ca2+ loss in mice lacking TRPV5 (86), and mice lacking TRPV6 also lose some Ca2+ into urine (13).
Renal TRPV4 is expressed mainly in the basolateral cell membrane of thin and thick ascending limbs and the distal convoluted tubule (28, 279) (Fig. 3B). TRPV4 channels were presumed to participate in the cellular response to alterations of extracellular osmolarity (195, 254). TRPV4 knockout mice tend to be hypercalcemic, which would indicate that TRPV4 rather decreases net renal Ca2+ reabsorption (63).
Together, these findings suggest that Ca2+ transport pathways that serve to stabilize the Ca2+ concentration of the endolymph in the inner ear are involved in kidney function to regulate the Ca2+ homeostasis of the whole body.
The kidney is of pivotal importance for the regulation of the acid-base balance of the body. Renal regulation of the acid-base balance primarily involves the reabsorption, generation, or excretion of HCO3− as well as the generation and excretion of NH4+. Vacuolar H+-ATPase (ATP6V), H+/K+-ATPase (ATP4A), and the Cl−/HCO3− exchangers pendrin (SLC26A4) and AE1 (SLC4A1) are all expressed in the kidney, where they contribute to acid-base balance. Gastric (and colonic) H+/K+-ATPase is expressed in the collecting duct (Fig. 3E). It is responsible for H+ secretion and K+ reabsorption under normal conditions and may be stimulated by acid-base perturbations and/or K+ depletion. The regulation may be species specific (for a review, see Ref. 236).
Vacuolar H+-ATPase, H+/K+-ATPase, and Cl−/HCO3− exchangers SLC26A4 and SLC4A1 are all expressed in the intercalated cells of the kidney (Fig. 3E), which are critically involved in acid-base balance (for reviews, see Refs. 4, 232, and 299). The localization of H+-ATPase in the apical or basolateral membrane can vary between cortical intercalated cells, indicating that subpopulations of these cells have opposite polarities of an H+-ATPase, consistent with the presence of both H+- and HCO3−-secreting cells (17, 18). Along those lines, type A, type B, or non-A, non-B intercalated cells are defined according to the presence or absence of the Cl−/HCO3− exchanger SLC4A1 and the subcellular distribution of H+-ATPase (5, 111). Type A intercalated cells mediate net secretion of H+ through apical H+-ATPase (ATP6V), which functions in series with basolateral SLC4A1 (51, 276, 321). Particularly during metabolic alkalosis, type B intercalated cells mediate the secretion of HCO3− by employing the apical Cl−/HCO3− exchanger pendrin (SLC26A4), which functions in series with basolateral H+-ATPase (5, 51, 62, 111, 112, 214, 243, 276, 300, 321). Non-A, non-B intercalated cells may be HCO3−- or H+-secreting cells or may interconvert between the two functions (111, 276). They express both pendrin as well as H+-ATPase in the apical membrane (300).
Under basal unstimulated conditions, persons with genetic disruption of pendrin (SLC26A4; Pendred syndrome) and mice lacking SLC26A4 exhibit no change in arterial pH, renal function, or fluid balance (213). Under conditions of dietary NaCl restriction or administration of mineralocorticoids, however, pendrin expression is increased in type B intercalated cells in rodents, and, under these conditions, mice lacking SLC26A4 show evidence for impaired renal HCO3− excretion as well as impaired Cl− retention, which results in elevated arterial pH and serum HCO3− and lower blood pressure compared with wild-type mice (213, 293, 301). In contrast to the inner ear, where SLC26A4 mutations lead to hearing loss in humans and mice (54), in the kidney pendrin is essential for a normal response to low-salt conditions (Table 2), indicating that pendrin may be a new target for antihypertensive therapy. In humans, loss of pendrin leads, in addition, to defective iodide uptake into thyroid glands and thus in later life to goiter (54). The latter is not found in mice and is in humans of much later onset than the hearing loss.
Autosomal dominant and recessive forms of distal tubular acidosis are caused by mutations in ion transporters of acid-secreting type A intercalated cells (for a review, see Ref. 4). These include at least two subunits of apical H+-ATPase (Fig. 3E and Table 2). Loss of function mutations of the genes encoding for the B1 subunit of H+-ATPase lead to recessive distal tubular acidosis with sensorineural hearing loss (105, 241). HCO3− therapy successfully treats systemic symptoms of distal renal tubular acidosis but fails to correct deafness, suggesting that transepithelial acid secretion is required for normal cochlear development and hair cell survival. Mice lacking the B1 subunit (ATP6V1B1) have preserved hearing but exhibit impaired maximal urinary acidification (56). Although patients with distal renal tubular acidosis due to homozygous B1-subunit mutations typically present as infants with spontaneous metabolic acidosis and failure to thrive (105), mice lacking ATP6V1B1 raised on a standard rodent diet were healthy, grew normally, and did not develop metabolic acidosis (56). The phenotypic discrepancy may be related to dietary differences, since a standard rodent diet provides a large net dietary alkali load, whereas the typical Western human diet, which has higher protein content, imposes a net acid load (56). On the other hand, apical expression of the alternative B-subunit isoform B2 is increased in the medulla of mice lacking ATP6V1B1 and may partially, although not completely, compensate for the loss of ATP6V1B1(56).
Mutations in ATP6N1B, encoding a new kidney vacuolar H+-ATPase subunit, which was also localized to the apical membrane of type A intercalated cells, cause recessive distal renal tubular acidosis with preserved hearing (241). Similarly, multiple mutations have been described for the Cl−/HCO3− exchanger SLC4A1, which are associated with distal renal tubular acidosis in the absence of deafness (for a review, see Ref. 4).
In conclusion, many transport proteins that stabilize the pH of the endolymph of the inner ear are also involved in the renal transport of HCO3− and H+ and, thus, contribute to the acid-base homeostasis of the body.
Aquaporins AQP1–AQP4 play a central role in water reabsorption of the kidney (for a review, see Ref. 183). AQP1 is particularly expressed along the proximal tubule (Fig. 3A), and near-isosmolar fluid reabsorption, a hallmark of proximal tubular function, is dramatically impaired in mice lacking AQP1, indicating that proximal tubular fluid reabsorption is largely due to transcellular water movement through AQP1 (227, 285). AQP1 is the principal water channel in the thin descending limbs of Henle's loop and is also expressed in the outer medullary descending vasa recta, where it facilitates water transport and is thus an important component of the urinary concentrating mechanism (27, 153, 192, 291). A very recent study (78) has provided evidence for the involvement of AQP1 in the migration of proximal tubule cells and possibly in the response of the proximal tubule to injury. In comparison, AQP2 is exclusively expressed in the principal cells of the connecting tubule and collecting duct and is the predominant vasopressin-regulated water channel (Fig. 3D). AQP3 and AQP4 are both present in the basolateral plasma membrane of collecting duct principal cells (Fig. 3D) and represent exit pathways for water reabsorbed apically via AQP2. Studies in patients have demonstrated that AQP2 is essential for urinary concentration (183, 229, 290). Loss of function mutations of AQP2 cause nephrogenic diabetes insipidus (329). The inheritence is usually autosomal recessive but may, in some patients, be autosomal dominant (autosomal dominant nephrogenic diabetes insipidus) (329) and could be related to a dominant negative monomer that leads to a missorting of AQP2 to the basolateral instead of apical plasma membrane of collecting duct cells (242) (Table 2). Moreover, mice lacking AQP2, AQP3, or AQP4 suffer from various degrees of nephrogenic diabetes insipidus (292).
Epithelial transport in the inner ear and kidney is critical for the function of both organs. Many of the proteins accomplishing ion transport within the inner ear and kidney are encoded by the same genes, as shown in Table 2. Most of the transporters are involved in K+ cycling within the inner ear and simultaneously participate in the renal tubular transport of Na+ and K+. Other transport systems are involved in regulating and stabilizing the Ca2+ concentration or pH of the endolymph and in the regulation of renal tubular transport of Ca2+, HCO3−, and H+ and thus participate in Ca2+ homeostasis and acid-base balance of the whole body. The different arrangement of channel proteins may lead to completely different transport functions. For example, KCNQ1/KCNE1 K+ channels serve to establish high K+ concentrations in the endolymph of the inner ear, whereas in proximal renal tubules, they stabilize the membrane potential across the apical cell membrane and thus contribute to the maintenance of the electrical driving force for Na+-coupled electrogenic transport. The comparison of the transport processes in the inner ear and kidney thus illustrates the amazing versatility of biology in the use of individual molecules. Moreover, the comparison leads to pathophysiological insight into syndromal genetic disease as well as into side effects of drugs targeting those channels and transporters, and it may provide clues to new therapeutic approaches. Our knowledge, though, is still far from complete, and many mechanisms are a matter of speculation. It is an aim of this brief synopsis to stimulate future interdisciplinary research in this exciting and clinically important area of physiology.
Work in the laboratory of the authors was supported by the Deutsche Forschungsgemeinschaft (to F. Lang and M. Knipper), American Heart Association Grant-In-Aid 655232Y, the Department of Veterans Affairs, and National Institutes of Health (NIH) Grants DK-56248, DK-28602, DK-70667, and GM-66232 (to V. Vallon). The support by NIH Research Grant R01-DC-01098 (to P. Wangemann) NIH Grant P20-RR-017686 is gratefully acknowledged.
The authors acknowledge the meticulous preparation of the manuscript by Jasmin Bühringer.
- Copyright © 2007 the American Physiological Society