The Cochlear Lateral Wall as a Biological Battery: the Mechanisms Underlying K⁺ Transport and Potential Generation

Abstract

The cochlea possesses an exceptional capacity for rapid sensory transduction, enabling sound perception of at frequencies exceeding 15 kHz. This extraordinary performance depends on a specialized electrochemical environment that maintains a high endocochlear potential (EP) and facilitates efficient K⁺ circulation. The cochlear lateral wall serves as a central component, functioning as a biological battery that generates and sustains the EP via intricate ion transport mechanisms. It consists of two epithelial-like layers—the marginal cell layer and the syncytial layer—which are electrically insulated by tight junctions and interconnected via gap junctions. K⁺ enters hair cells from the K⁺-rich endolymph to initiate sensory transduction and is subsequently recycled through the perilymph and reabsorbed by the lateral wall. Our combined electrophysiological and mathematical modeling studies elucidate that the EP primarily depends on K⁺ equilibrium potentials across the apical membranes of intermediate and marginal cells, primarily mediated by K_ir4.1 and IKs channels, respectively. Pharmacological experiments confirmed that Na⁺,K⁺-ATPase and NKCC are essential for maintaining the K⁺ gradient and EP. Furthermore, our computational model successfully reproduced dynamic changes in ion concentrations and membrane potentials under hypoxic conditions and acoustic stimulation. Genetic studies further reinforce the physiological importance of lateral wall components, as mutations in associated transporters, channels, and structural proteins commonly lead to sensorineural hearing loss. Collectively, these findings underscore the cochlear lateral wall as an integrated electrochemical organ and illustrate the utility of multiscale modeling in bridging molecular mechanisms with systems-level auditory function.

Introduction

The cochlea, a specialized peripheral auditory organ, is notable for its exceptionally rapid sensory transduction compared to other sensory systems. For instance, photoreceptors in the retina, the peripheral organ of vision, respond at approximately 10 to 30 Hz [25], which limits the refresh rate of typical displays to less than 150 Hz. In contrast, cochlear hair cells, the primary auditory receptor cells, can detect sounds reaching up to 20 kHz [53] and exhibit a response speed exceeding 40 kHz. This performance is approximately 100 times faster than the maximal firing rate of neurons (~ 500 Hz) and nearly 1,000 times faster than that of photoreceptors. This characteristic is achieved through specialized mechanisms that distinguish hair cells from conventional excitable cells [53].

In typical excitable cells such as neurons and muscle fibers, excitation is mediated by Na⁺ influx through voltage-gated channels, followed by Na⁺ extrusion via Na⁺,K⁺-ATPase to restore ionic gradients [3, 53]. This coupling between fast ion influx and enzymatic ion extrusion limits the maximum excitation frequency to ~ 500 Hz. In contrast, cochlear hair cells employ ion channels not only for ion influx but also for efflux, enabling significantly faster cycles of excitation [53]. The cochlear architecture and its surrounding electrochemical milieu are optimized to support this ultrafast transduction.

Sound Reception by Hair Cells and the Electrochemical Environment Maintenance System “K⁺ Circulation in the Cochlea”

A defining feature of the cochlea is the unique lymphatic environment surrounding the hair cells. The cochlear duct is filled with two distinct extracellular fluids: endolymph and perilymph [42, 43, 79, 108, 109]. Incoming sound vibrations from the external environment are transmitted to sensory epithelium, which is immersed in both fluids—perilymph below and endolymph above (Fig. 1a). These vibrations are on the nanometer scale, comparable in magnitude to the size of a single water molecule. Located atop the basilar membrane, the organ of Corti houses the sensory hair cells. The basolateral membranes of these cells are bathed in perilymph, whereas their apical hair bundles extend into the K⁺-rich endolymph [53] Endolymph is a highly specialized extracellular fluid characterized by a high potassium concentration of approximately 150 mM [93, 96] and a markedly positive endocochlear potential (EP) of + 80 mV [5, 64, 106] (Fig. 1a).

Fig. 1.

Structure of the inner ear cochlea, cochlear lateral wall, and K⁺ ion environment. (a) Cross-sectional view of the cochlea. The cochlear lateral wall consists of the stria vascularis and the spiral ligament. (b) Magnified view of the cochlear lateral wall (dashed square in (a)) and detailed diagram of the syncytium layer. Transport molecules involved in K⁺ unidirectional transport and the generation of the endocochlear potential (EP) are also shown. Functionally, the cochlear lateral wall consists of two layers: a syncytium layer with fibrocytes forming the spiral ligament and intermediate and basal cells forming the stria vascularis, which are connected by gap junctions. Tight junctions between basal cells isolate the intrastrial space (IS) from the perilymph. The second layer is the marginal cell layer, where tight junctions between marginal cells electrically separate the IS from the endolymph. (c) Changes in the potential and K⁺ ionic environment in the cochlear compartments under various loading conditions.

Adapted from reference [79]. IS intrastrial space, NKCC Na⁺, K⁺, 2Cl⁻ cotransporter, TJ tight junction

When the basilar membrane vibrates, the deflection of hair bundles opens mechanosensitive cation channels at their tips, allowing K⁺ to enter the hair cells from the endolymph. This K⁺ influx depolarizes the hair cells, thereby converting mechanical stimuli into electrical signals that are relayed to the central nervous system via neurotransmitter release from the basolateral membrane [47]. The + 80 mV EP provides a substantial electrochemical gradient across the hair cell membrane, enhancing K⁺ influx and thereby augmenting auditory sensitivity [17, 27, 45]. Disruption or loss of this potential is strongly associated with hearing impairment [27, 45, 50, 61].

Another distinctive feature of the cochlea is the presence of a specialized K⁺ recycling system [108], also known as the “K⁺ circulation” system (Fig. 1a) [77, 79, 119]. In this system, K⁺ exits hair cells into the perilymph through basolateral K⁺ channels and is subsequently taken up by two epithelial-like cell layers in the lateral wall, which recycle it into the endolymph [26, 42, 43, 56, 57, 79, 108, 109]. In typical excitable cells, Na⁺ influx triggers excitation and is counterbalanced by Na⁺/K⁺-ATPase–mediated recirculation. By contrast, the K⁺ circulation in the cochlea operates constitutively, because mechanosensitive channels in hair cells remain partially open even in the absence of acoustic stimulation, allowing continuous cation permeation [22, 119].

The cochlear lateral wall is composed of the stria vascularis, which is densely vascularized, and the spiral ligament as connective tissue, consisting of fibrocytes. Unidirectional K⁺ transport through the lateral wall—particularly involving the stria vascularis—has long been proposed as a key mechanism for maintaining both the EP and K⁺ circulation [61, 63, 102], although the precise molecular underpinnings remained poorly defined. The stria vascularis comprises three main cell types: marginal, intermediate, and basal cells (Fig. 1b). Intermediate and basal cells, along with adjacent fibrocytes in the spiral ligament, form a functional electrochemical syncytium. This syncytium exhibits uniform ionic and electrical properties due to extensive coupling via gap junctions, which likely facilitate intercellular ion flow [56, 57, 65, 94, 98, 107–109]. In contrast, tight junctions between marginal cells and between basal cells serve as electrical barriers, similar to those in renal and intestinal epithelia (Fig. 1b). Therefore, the lateral wall may be functionally conceptualized as comprising two epithelial-like layers: an inner layer of marginal cells and an outer syncytial layer consisting of intermediate cells, basal cells, and fibrocytes (Fig. 1b).

Histological, pharmacological, and electrophysiological studies have revealed that key K⁺ transporters—including Na⁺,K⁺-ATPase and the Na⁺,K⁺,2Cl⁻ cotransporter (NKCC)—are localized to the basolateral membranes of both epithelial-like layers [26, 62, 66, 72]. In contrast, K_ir 4.1 and I_Ks channels are expressed at the apical membrane of the inner and outer layers, respectively [4, 87] (Fig. 1b).

This review article presents recent insights from our experimental and computational studies that describe the roles of K⁺ transport molecules in mediating unidirectional K⁺ movement across the cochlear lateral wall. We further discuss the mechanisms by which these transport processes contribute to generation and maintenance of the EP.

K⁺ Equilibrium Potential

To aid in understanding how the endocochlear potential (EP) is generated, we begin by briefly explaining the concept of the “K⁺ equilibrium potential,” a fundamental electrochemical principle underlying EP formation [53]. As illustrated in Fig. 2a, consider two compartments: space A with a high K⁺ concentration and space B with a low K⁺ concentration. Each compartment contains an equal total number of cations and anions. If the two spaces are separated by an impermeable membrane, no ionic movement occurs (Fig. 2a). However, if the membrane is selectively permeable to K⁺, K⁺ ions will diffuse from A to B along their concentration gradient. As K⁺ moves, positive charge accumulates in space B, creating an electrical potential difference across the membrane. This emergent potential counteracts further K⁺ movement. Eventually, an electrochemical equilibrium is established—when the driving force from the concentration gradient is exactly balanced by the opposing electrical gradient. The resulting potential is called the “K⁺ equilibrium potential” or “K⁺ diffusion potential,” and its magnitude depends on the K⁺ concentration ratio between the two compartments.

$${E_K=\frac{\mathit{RT}}{F}ln([K^{+}]}_A/{[K^{+}]}_B)\to (a)$$

Fig. 2.

Generation of the K⁺ equilibrium potential. (a) When a membrane separates two compartments, space A and space B, with a higher concentration of K⁺ in space A and a lower concentration in space B, and the total number of cations and anions in each space is balanced by ions other than K⁺, selective permeability to K⁺ allows its diffusion down the concentration gradient. This movement results in a net transfer of positive charge, generating a potential difference across the membrane such that space B becomes electrically positive relative to space A. As K⁺ continues to move, the emerging electrical potential increasingly opposes further K⁺ efflux. Eventually, an electrochemical equilibrium is established when the electrical driving force exactly counterbalances the chemical gradient. The resulting membrane potential at this equilibrium is termed the "K⁺ equilibrium potential," and its magnitude is determined by the K⁺ concentration gradient between the two spaces. (b) This membrane potential can be conceptually represented by an electrical circuit model, wherein the membrane functions as a capacitor, and the ion gradient is modeled as a battery with an electromotive force. Ionic permeability is represented as a resistance

Here, E_K denotes the K⁺ equilibrium potential, R is the universal gas constant, T is the absolute temperature, F is the Faraday constant, and [K⁺]_A and [K⁺]_B represent the K⁺ concentrations in compartments A and B, respectively. In biological systems, the K⁺-selective permeability of membranes is conferred by K⁺ channels—transmembrane proteins that allow the selective passage of K⁺. This principle also applies to other monovalent ions. Understanding the quantitative relationship between ion concentration gradients and membrane potentials, as described by this equation, is key to elucidating the mechanisms by which the EP is generated, as detailed in the sections below.

Conversely, in the same equation, if [K⁺]_A is unknown but [K⁺]_B and the K⁺ equilibrium potential are known, one can solve for the unknown concentration, making this equation a practical tool for estimating ion concentrations in biological experiments. In particular, ion-selective electrodes exploit this principle to measure K⁺ concentration based on recorded membrane potentials. Moreover, membrane potentials and associated ionic fluxes can be conceptualized using an equivalent electrical circuit model, wherein the membrane behaves like a capacitor connected to a resistor and an electromotive force (i.e., a “battery”) driven by the transmembrane concentration gradient (Fig. 2b). We adopted this equivalent circuit framework in our previously developed mathematical models of cochlear ion transport [77, 78].

Characteristics of the Cochlear Lateral Wall and Mechanism Underlying the EP

The anatomical and electrochemical features of the cochlear lateral wall can be summarized as follows:

1. Tight junctions connecting the inner and outer epithelial layers, as well as vascular endothelial cells, electrically insulate the stria vascularis from the perilymph, endolymph, and bloodstream.

2. These tight junctions also delineate the basolateral and apical membranes of epithelial cells, enabling the electrical properties of the lateral wall to be described in terms of four distinct membrane domains: the apical and basolateral membranes of both the inner and outer layers.

3. The basolateral membrane of the inner layer and the apical membrane of the outer layer form a highly folded, interdigitating structure that encloses a narrow extracellular compartment termed the “intrastrial space (IS),” approximately 15 nm in width (Fig. 1b)[44]. The IS contains extracellular fluid with a relatively low K⁺ concentration (below 5 mM), similar to other extracellular spaces, but maintains a high electrical potential of approximately + 90 mV—comparable to the endolymphatic potential (Fig. 1c)[48, 76, 88]. This “intrastrial potential” (IS potential), which was suggested to depend primarily on K⁺ permeability on the apical membrane of the outer layer, is widely regarded as the primary electrochemical driver of the endolymphatic potential [88]. This idea is supported by an observation that vascular perfusion of Ba²⁺, a blocker for K⁺ channels, markedly suppressed the EP [72]. The K⁺ channels crucial for formation of the EP was identified as an inwardly rectifying K.⁺ channel K_ir4.1 expressed in stria vascularis [41]. Later, a histochemical study revealed that K_ir4.1 is localized on the apical membrane of the outer layer, suggesting that this channel plays a central role in generation of the IS potential [4, 48, 76, 88, 99]

To examine the mechanisms responsible for generating the high endolymphatic potential, we performed in vivo measurements using double-barreled K⁺-selective ion electrodes inserted into microdomains within and adjacent to the cochlear lateral wall of guinea pigs. This technique allowed for simultaneous recording of K⁺ concentration and electrical potential under various experimental conditions [48, 76, 88].

Our results demonstrated that the highly positive EP is primarily driven by K⁺ equilibrium potentials generated across the apical membranes of both the inner and outer layers (Fig. 1b; syncytial: K_ir4.1; marginal: I_Ks) [76]. These equilibrium potentials arise likely from the electrochemical gradients of K⁺ across the respective apical membranes [99, 101]. Under physiological conditions, the K⁺ concentration is markedly lower in the IS than in the syncytial layer, creating a large K⁺ equilibrium potential at the syncytial apical membrane. Conversely, the K⁺ concentrations in the marginal layer and endolymph are both high, resulting in a negligible potential difference across the apical membrane of marginal cells. Thus, the IS potential substantially accounts for the high endolymphatic potential (Fig. 1c, first row). However, under hypoxic conditions, which mimics experimentally inhibition of Na⁺,K⁺-ATPase in marginal cells [75, 86], we observed K⁺ accumulation in the IS and K⁺ depletion in the marginal cell layer, reducing the outer layer equilibrium potential and increasing that of the inner layer. This shift ultimately reversed the endolymphatic potential to negative values (Fig. 1c, second row) [76].

To illustrate this, we present experimental data in Fig. 3a, depicting simultaneous measurements of electrical potential and K⁺ concentration as a K⁺-selective electrode was advanced from the perilymph into the stria vascularis. As the electrode progressed from the low-K⁺, 0 mV perilymph and passed through the outer layer, which contains numerous fibrocytes, transient and multiple elevations in K⁺ concentration were observed. A distinct region was then identified with a stable K⁺ concentration of 65–85 mM and a slightly positive potential of + 2 to + 3 mV. This domain corresponds to the syncytial layer under baseline or normoxic conditions (see Fig. 1c).

Fig. 3.

Measured cochlear lateral wall environment upon inhibition of inner layer Na⁺,K⁺-ATPase. (a) Measurement of K⁺ concentration and potential in each region of the stria vascularis using a K⁺-selective microelectrode. The red line represents the potential, whereas the blue line shows the change in K⁺ concentration. The Na⁺,K⁺-ATPase in the marginal cell layer was inhibited by anoxia and the infusion of ouabain. The change in K⁺ concentration in the intrastrial space (IS) under anoxia and the measured K⁺ concentration in the syncytium layer (black arrow in (a)) were substituted into the equation for (b) in the text, and compared with the measured IS potential (red and orange lines) [43, 76, 79].

Adapted from reference

As the electrode advanced further, it entered the IS, which exhibited a high potential of + 70 mV and a low K⁺ concentration of approximately 4 mM (Fig. 3a; see Fig. 1c). Upon induction of hypoxic conditions, the IS potential dropped markedly from + 70 mV to + 22 mV, while K⁺ concentration increased from 4 to 28 mM [76]. Intermediate cells, forming the apical membrane of the syncytial layer, express the K⁺ channel Kir4.1 at high levels. As such, the IS potential is predominantly governed by the K⁺ equilibrium potential across this membrane. Based on the Nernst-like relation in Eq. (a), the IS potential (ISP) can be expressed as:

$$ISP\approx V_{\mathit{SB}}+\frac{\mathit{RT}}{F}ln({{\mathit{aK}}^{+}}_{\mathit{SY}}/{{\mathit{aK}}^{+}}_{\mathit{IS}})\to (b)$$

where V_SB represents the electrical potential at the basolateral membrane of the syncytial layer, and aK⁺_SY and aK⁺_IS correspond to the potassium activities in the syncytial layer and IS, respectively. Because V_SB and aK⁺_SY remain constant under hypoxic conditions, we substituted their measured values—as indicated by the black arrow in the upper panel of Fig. 3a—into Equation (b). We then incorporated the experimentally observed changes in aK⁺_IS under hypoxia and compared the resulting calculated ISP with the observed values (Fig. 3a: predicted ISP). These analyses revealed that the IS potential is predominantly governed by the K⁺ equilibrium potential generated by K_ir4.1 channels expressed in the syncytial layer [100].

Moreover, perfusion with bumetanide, an NKCC inhibitor acting on the basolateral membrane of the marginal cell layer, produced effects similar to those observed under hypoxic conditions [76]. A comparable shift was noted in the potential across the apical membrane of marginal cells, prompting the derivation of the following expression:

$$EP\approx ISP+V_{\mathit{MB}}+\frac{\mathit{RT}}{F}ln({{\mathit{aK}}^{+}}_{\mathit{MC}}/{{\mathit{aK}}^{+}}_{\mathit{EL}})\to (c)$$

Here, V_MB represents the potential at the basolateral surface of the marginal cell layer, whereas aK⁺_MC and aK⁺_EL denote the potassium activities in the marginal cell cytoplasm and the endolymph, respectively.

By integrating Eqs. (b) and (c), and assuming that V_SB and V_MB remain constant under physiological and pathophysiological conditions, we conclude that the EP is governed by two sequential K⁺ equilibrium potentials—one across the syncytial layer and the other across the marginal cell layer. This yields the following comprehensive expression (Fig. 3b) [76]:

$$EP\approx V_{\mathit{SB}}+V_{\mathit{MB}}+\frac{\mathit{RT}}{F}ln({{\mathit{aK}}^{+}}_{\mathit{SY}}/{{\mathit{aK}}^{+}}_{\mathit{IS}})+\frac{\mathit{RT}}{F}ln({{\mathit{aK}}^{+}}_{\mathit{MC}}/{{\mathit{aK}}^{+}}_{\mathit{EL}})\to (d)$$

K⁺ Transport in the Spiral Ligament and the Generation of theHighly Positive EP

The basolateral surface of the syncytial layer in the cochlear lateral wall is composed primarily of fibrocytes within the spiral ligament. Histological analyses have identified several K⁺ uptake transporters in these cells, among which Na⁺,K⁺-ATPase and NKCC likely play prominent roles in K⁺ transport and in generating the high endolymphatic potential [24, 42, 56, 57, 75, 98, 108]. However, the exact contributions of these transporters to EP generation had remained unclear. To address this, we perfused specific transporter inhibitors into the perilymph and transiently inserted a K⁺-selective ion electrode into the syncytial layer to monitor K⁺ concentration and potential changes (Fig. 4). At the same time, a conventional microelectrode was placed in the endolymph to continuously record the EP [1, 113, 114].

Fig. 4.

K⁺ transport in the syncytium layer. (a-d) Measurement of K⁺ concentration and potential in cochlear lateral wall following ouabain perfusion, an inhibitor of Na⁺,K⁺-ATPase. The EP was monitored using a conventional glass electrode (d), while K⁺ concentration and potential were simultaneously measured using a K⁺ electrode temporarily inserted three times from the perilymph towards the endolymph (a, b, c). The timing of each insertion is indicated by the rectangle in (d). The circles in a, b, and c represent the regions that are considered to depict the typical environment within the syncytium. Compared to pre-perfusion (a), the K⁺ concentration in the syncytium layer gradually decreased during infusion (b, c).

Adapted from [114] [114]. (e–h) Effect of bumetanide, an inhibitor of NKCC, infused into the perilymph. The same experimental methodology as in (a-d) was used. Compared to pre-perfusion (e), little change was observed in the K⁺ concentration in the syncytium layer during infusion (f, g). However, it is noteworthy that the EP gradually decreased. (i-l) Effect of low-Na⁺ solution infused into the perilymph. Compared to pre-perfusion (i), the potential in the basal side of the syncytium layer decreased during infusion (j) and recovered (k). Adapted from [113]

We first examined the effect of ouabain, a potent inhibitor of Na⁺,K⁺-ATPase (Fig. 4a-d). Under control conditions using drug-free artificial perilymph, the EP was maintained at approximately + 80 mV (Fig. 4a). Insertion of a K⁺-selective electrode from the perilymph into the cochlear wall revealed multiple peaks in K⁺ concentration corresponding to fibrocyte-rich regions (Fig. 4a). The K⁺ concentration just proximal to the IS was 70.7 mM, consistent with the outer syncytial layer. Following ouabain perfusion, the EP gradually declined and stabilized near 0 mV. Re-insertion of the K⁺-selective electrode revealed a progressive reduction in K⁺ concentration within the syncytial layer, eventually falling to 10.9 mM (Fig. 4b and c). These findings indicate that Na⁺,K⁺-ATPase mediates K⁺ uptake into the syncytial layer, thereby sustaining its high intracellular K⁺ level. Inhibition of this transporter diminished the K⁺ equilibrium potential across the apical surface of the syncytial layer, ultimately abolishing the EP (see Fig. 1c, third row).

If NKCC were similarly involved in K⁺ uptake, its inhibition by bumetanide would be expected to exert effects comparable to those of ouabain. Interestingly, although bumetanide application reduced the EP (Fig. 4h), it did not significantly alter either the potential or K⁺ concentration within the syncytial layer (Fig. 4e-g; see Fig. 1c, fourth row). These observations suggest that NKCC localized at the basolateral surface of the syncytial layer contributes minimally to K⁺ transport. Further analysis revealed that bumetanide applied to the perilymph reached the stria vascularis, where it inhibited NKCC in the marginal cell layer, thereby leading to a loss of EP (Fig. 1c, fourth row). Additionally, perfusion with a low-Na⁺ solution decreased both the basolateral potential of the marginal cell layer and the EP, implying that Na⁺ permeability is substantially greater than that of K⁺ or Cl⁻ (Fig. 4i-l; see Fig. 1c, fifth row). Together, these findings indicate that Na⁺,K⁺-ATPase is the principal mediator of K⁺ uptake and intracellular K⁺ maintenance in the basal region of the syncytial layer. Notably, while it is generally accepted that cells in vivo maintain a negative resting membrane potential, the membrane potential in this layer is constantly positive by several mV with reference to perilymph, owing to a significant Na⁺ conductance, a factor crucial for EP generation [1, 79, 113, 114]. The molecular identification of this Na⁺ conductance represents the final key to fully elucidate the mechanism underlying EP generation. Recent literature suggests that NBCn1 (SLC4A7), NBCn2 (SLC4A10), and NaBC1 (SLC4A11) may be the responsible molecules (Table 1), necessitating further investigation [12, 13, 19, 40, 80].

Table 1.

Genes encoding lateral-wall’s proteins associated with hearing loss in humans and mice

Molecules	Distribution	Human deafness types	Mouse phenotypes (other than deafness)	Reference
KCNE1 (Isk, min K)	MC	Jarvell and Lange-Nielsen syndrome	Collapse of SM and VES, balance disorder	[87, 92, 105]
KCNQ1 (KvLQT1, Kv7.1)	MC	Jarvell and Lange-Nielsen syndrome	Collapse of SM and VES, balance disorder	[15, 92]
KCNJ10 (Kir4.1)	IC	EAST syndrome, SeSAME syndrome	Collapse of SM, loss of EP (~ 0 mV)	[10, 73, 84, 89, 99]
KCNK5 (K2P5.1)	OSC	-	Loss of EP (~ + 50 mV), degeneration of hair cells and spiral ganglions	[16]
ZBTB20	OSC	-	Abnormal root cell differentiation, hearing loss	[112]
BSND (Barttin)	MC	Type 4 Bartter's syndrome	Loss of EP (~ + 17 mV)	[32, 81] [9]
NKCC1 (SLC12A1)	MC, FC	-	Collapse of SM and VES, balance disorder	[29, 31, 36]
Pendrin (SLC26A4)	SPEC, OSC	Pendred syndrome, DFNB4	Loss of EP (~ 0 mV), Gene therapy restored EP and hearing	[2, 33–35, 46, 51, 58, 83, 91, 110]
SLC22A4	StE	DFNB60	-	[6]
KCC3 (SLC12A6)	FC	-	Degeneration of hair cells	[11]
HGF	LW	DFNB39	Defective neural crest cell migration, reduced EP	[74]
CHD7	LW	Enlarged vestibular aqueduct, hearing loss	Structural abnormalities in lateral wall, hearing loss	[82]
GJA1 (Cx43)	FC	AR-nonsydromic deafness	-	[70, 97]
GJB1 (Cx32)	FC	X-linked Charcot-Marie-Tooth and deafness	-	[95]
GJB2 (Cx26)	FC	DFNB1, DFNA3A, hereditary palmoplantar keratoderma with deafness	Loss of EP (~ + 50 mV), degeneration of hair cells	[8, 20, 30, 37, 52, 55, 90, 115, 118]
GJB3 (Cx31)	FC	DFNA2, AR-nonsyndromic deafness	-	[111]
GJB6 (Cx30)	FC	DFNA3B	Loss of EP (~ + 3 mV), degeneration of hair cells	[7, 21, 28, 39, 54, 103]
Panx1	FC, OSC	-	Loss of EP (~ + 60 mV)	[18]
Claudin-11	BC	hypomyelinating leukodystrophy (HLD22)	Loss of EP (~ + 30 mV), hearing loss	[38, 59, 60, 69]
NBCn1 (SLCA7)	FC		Mild hearing loss, degeneration in organ of Corti, (Na+ conductance)	[12, 13, 19]
NBCn2 (SLC4A10)	FC		Early hearing loss, outer hair cell loss, reduced EP	[80]
NaBC1 (SLC4A11)	FC	Harboyan syndrome	Loss of EP (~ + 40 mV),	[40]
S1PR2	LW		EP decline	[49]

EP, endocochlear potential; SM, scala media; VES, vestibular endolymphatic space; FC, fibrocytes; BC, basal cells; IC, intermediate cells; MC, marginal cells; OSC, outer sulcus cells; LC, lateral wall; SC, supporting cell; SPEC, spiral prominence epithelial cells; StE, strial endothelium; DFNA, the locus for autosomal dominant nonsyndromic deafness; DFNB, the locus for autosomal recessive nonsyndromic deafness

Development of the Mathematical Model of Cochlear K⁺ Transport

Multiple models describing EP generation and K⁺ transport have been proposed, originating from Davis’s battery theory, which conceptualized the cochlear lateral wall as a biological battery [27]. Early qualitative models, such as the two-pump and single-pump models (Fig. 5a), posited that marginal cells alone were sufficient to account for EP generation and K⁺ transport [26]. These models were developed at a time when the spatial distribution and functional roles of ion channels and transporters in the lateral wall were not yet fully elucidated. Subsequent identification of inwardly rectifying K⁺ channels, K_ir4.1, in intermediate cells led to the development of the two-cell model (Fig. 5b), which attributes EP generation primarily to intermediate cells, whereas marginal cells are responsible for K⁺ secretion into the endolymph [4, 26, 72, 88, 99, 100, 108, 109]. Although the two-cell model provided a conceptual framework for understanding EP generation and K⁺ recycling (i.e., K⁺ circulation) under physiological conditions, it remained qualitative and lacked the capacity to predict EP dynamics under pharmacological or pathological perturbations.

Fig. 5.

Models of the endocochlear potential and K⁺ transport in the cochlea. (a) Single cell models (the two-pump model and the single pump model) of EP. The EP is generated by an electrogenic K⁺-ATPase in the apical side of the marginal cell layer [26]. (b) The two-cell model of the EP generation. The cochlear lateral wall has two layers, which separate endolymph from the intrastrial space (IS) and IS from perilymph, respectively [100]. (c) Equivalent electrical circuit representing cochlear K⁺ transport (fibrocyte-integrating Nin-Hibino-Kurachi model: fi-NHK model) [78]. The circuit includes not only the lateral wall but also the hair cell layer. NKCC: Na⁺,K⁺,2Cl⁻ co-transporter, ClC: Cl⁻ channel, NSC: non-selective cation channel, MET channel: Mechanotransduction channel. The MET channel, located at the apex of the sensory hair cells, is an ion-permeable channel that opens in response to the bending of the hair cells due to sound stimuli. (d) The calculated dynamic changes in several electrochemical parameters under normal, anoxic, and Ba²⁺ block conditions. Orange and red lines indicate the ISP and EP, respectively. (e, f) illustrates the K⁺ concentration in the intrastrial space ([K⁺]_IS) and marginal cells ([K⁺]_MC), as well as the Cl⁻ concentrations in the same locations ([Cl⁻]_IS and [Cl⁻]_MC). (g) Profiles of the circulation current in the lateral wall. Under physiological conditions, the circulation current is composed purely of K⁺ through Na⁺,K⁺-ATPases and NKCC1 in the basolateral surface of the marginal cell layer. Under anoxic conditions shown in right panel, the carrier of the circulation current switches from K⁺ to Cl⁻ and Na⁺ that permeate the ClC and the non-selective cation channel (NSC), respectively. This change causes an increase in [K⁺]_IS and [Cl⁻]_IS and a decrease in [K⁺]_MC and [Cl⁻]_MC. (h) Difference between fi-NHK model and NHK model in the syncytium layer [77–79].

Adapted from references

Although our experimental findings have clarified the principal mechanisms underlying the generation of the EP under physiological conditions and its changes following transporter inhibition, two unresolved issues remain. The first concerns the reversal of K⁺ concentration gradients under hypoxic conditions, in which the K⁺ concentration in the IS markedly increases (Fig. 1c, second row). The second is the significant decline in K⁺ concentration within the syncytial layer upon inhibition of K⁺ transporters (Fig. 1c, third row)[76]. In most cell systems, inhibition of K⁺ transport would gradually lead to equilibration of intra- and extracellular K⁺ levels. However, a substantial decrease in extracellular K⁺, as observed here, is difficult to reconcile with conventional passive diffusion or equilibrium dynamics. These atypical responses cannot be adequately explained by standard models of ion homeostasis. To resolve these complexities and to integrate changes in electrical potential, ionic currents, and concentration dynamics across compartments, we adopted a computational modeling approach based on equivalent electrical circuitry [77].

As illustrated in Fig. 2b, the dynamics of membrane potential and transmembrane current—governed by ion channels—can be represented by an equivalent electrical circuit and expressed mathematically. Ion transporters, which regulate ion concentrations across membranes, can likewise be formulated within this framework. Accordingly, all membrane-bound transport molecules were incorporated into the circuit model. In this context, hair cells were conceptualized as light bulbs, the EP as a battery, and K⁺ transport as electric current. This framework constitutes a molecular-level mathematical model of cochlear K⁺ circulation (Fig. 5c), conceptually proposed by Davis and partially demonstrated by Brownell and colleagues [27, 119]. The magnitude of the K⁺ transport current is determined by the potential difference between the EP and the hair cell interior, which serves as the driving force for K⁺ influx through apical transduction cation channels of sensory hair cells. Therefore, a reduction in EP directly diminishes the electrochemical driving force, thereby decreasing K⁺ transport current.

Even in the absence of acoustic stimulation, a steady-state transduction current flows through cochlear hair cells [56, 57, 85, 108, 109]. This baseline condition was first modeled to define the resting distributions of membrane potentials, ionic concentrations, and ion fluxes across the cochlear lateral wall. Past experimental results have demonstrated a functional coupling between NKCC and Na⁺,K⁺-ATPase via local Na⁺ recycling, showing that inhibition of one transporter halts the transport activity of the other [76]. Therefore, hypoxic conditions were simulated by functionally inhibiting Na⁺,K⁺-ATPase and NKCC activity in the marginal cell layer [77]. The simulation accurately reproduced the experimentally observed changes in membrane potentials and K⁺ concentrations in both the cochlear lateral wall and the endolymph (Fig. 5d, e, and f) [76]. Our analysis revealed that upon inhibition of Na⁺,K⁺-ATPase and NKCC, the K⁺ currents normally mediated by these transporters were lost and replaced by inward Na⁺ and Cl⁻ currents via co-expressed cation and anion channels in marginal cells (Fig. 5g). Consequently, K⁺ influx into the IS continued unabated, whereas K⁺ efflux from marginal cells was halted, producing a reversal in K⁺ concentration gradients between the two compartments—an outcome consistent with the experimentally observed K⁺ inversion phenomenon (Fig. 1c).

Additionally, the mathematical model was refined to replicate the experimentally observed changes in K⁺ transport and membrane potential within the syncytial layer (Fig. 5h) [78]. Based on prior experimental evidence [1, 113, 114], K⁺ transport in the syncytium was modeled exclusively through Na⁺,K⁺-ATPase activity. This updated model successfully reproduced the outcomes of ouabain and bumetanide perfusion experiments applied via the perilymph (Fig. 4) [78].

Changes in the Ionic Environment of the Cochlear Lateral Wall and the EP During Sound Stimulation

Through the integrated theoretical and experimental approaches described above, the underlying ionic transport mechanisms responsible for the observed physiological phenomena have been elucidated with greater clarity. Consequently, mathematical modeling has proven to be an indispensable tool for interpreting results from animal experiments. However, the validity of theoretical predictions critically depends on empirical support; thus, continual refinement of simulations necessitates new in vivo experiments designed specifically to test and update model assumptions.

To investigate how K⁺ circulation is modulated under acoustic stimulation, we measured K⁺ concentration and electrical potential in the IS while applying varying sound pressures and frequencies. We observed that the IS potential decreased in response to increasing sound load, whereas K⁺ concentration increased (Fig. 6a and b) [116]. Notably, the K⁺ profile in the endolymph negligibly changed (Fig. 6a and c) [96]. The potential shift in the IS resembled previously reported changes in the EP; however, analogous fluctuations within the IS had not been documented prior to this study. These findings represent the first direct evidence that sound-induced hair cell currents functionally affect the ionic environment in the cochlear lateral wall. In addition, the measured dynamics of electrochemical properties were primarily replicated by our computational model, reinforcing the idea that the lateral wall is electrically connected to hair cells [116].

Fig. 6.

Potential and K⁺ Concentration in IS and Endolymph during Sound Stimulation. (a) A series of in vivo electrophysiological recordings were performed on a guinea pig using a double-barreled microelectrode sensitive to both K⁺ concentration and potential. The microelectrode was advanced from the perilymph into the cochlear lateral wall and then retracted, with recordings obtained from the IS or endolymph. The red and blue lines represent potential and [K⁺], respectively. After passing through the outer syncytial layer, identified by an increase in K⁺ concentration and a low positive potential, the electrode was held in the IS. Anoxia caused a decrease in potential and an increase in K⁺ concentration. After returning to baseline, the animal was stimulated four times with a 1 kHz, 90 dB SPL tone-burst at 30 s intervals (orange bars). The electrode was then inserted into the endolymph, and the same acoustic stimulation protocol was applied. Expanded views of the recorded responses from the IS and endolymph (dotted boxes) are shown in (b) and (c), respectively. The filled wedge, open wedge, and gray bar above the trace denote periods of electrode advancement, retraction, and holding, respectively.

Adapted from a reference [116]

Physiological Significance of the Molecules Expressed in the Lateral Wall

Over the past three decades, extensive research has demonstrated that genetic disruption or mutation of ion transporters and structural proteins expressed in the cochlear lateral wall is closely associated with hearing loss in both humans and mouse models, as summarized in Table 1. For example, knockout of KCNE1 or KCNQ1 leads to collapse of the scala media, vestibular dysfunction, and ultimately profound deafness [15, 92, 105]. These findings underscore the critical role of KCNQ1/KCNE1 potassium channels in mediating K⁺ transport and maintaining the highly specialized ionic composition of the endolymph. In humans, mutations in either gene result in Jervell and Lange-Nielsen syndrome, a condition characterized by congenital sensorineural hearing loss and cardiac arrhythmias.

Similar pathological features—including collapse of the scala media and vestibular dysfunction—are observed in mice lacking NKCC1 [29, 31, 36]. Moreover, mutations in BSND, which encodes barttin, a subunit essential for Cl⁻ channel function [32], cause Type IV Bartter syndrome, characterized by renal salt wasting and sensorineural hearing loss. In corresponding knockout mice, EP is markedly reduced (~ + 17 mV) [9, 81]. Ablation of KCNJ10, which encodes the inwardly rectifying K⁺ channel K_ir4.1, induces structural disorganization of the cochlea and a near-complete loss of EP [10, 73, 84, 89]. In humans, mutations in KCNJ10 result in EAST/SeSAME syndrome, presenting as a constellation of deafness, epilepsy, ataxia, and renal tubulopathy. Similarly, the two-pore domain K⁺ channel KCNK5, predominantly expressed in outer sulcus cells (OSCs), is critical for cochlear function; its deletion results in degeneration of hair cells and spiral ganglion neurons, along with partial EP loss (~ + 50 mV) [16].

Recent studies have identified additional molecular determinants essential for maintaining cochlear lateral wall integrity [104]. Although KCNJ16, which is localized to the spiral ligament (SL) and supporting cells (SC), appears to be dispensable for normal auditory function [71], disruption of ZBTB20, a gene expressed in outer sulcus cells (OSCs), impairs root cell differentiation and results in hearing loss [112]. Mutations in SLC26A4, which encodes pendrin, are responsible for Pendred syndrome and DFNB4, both characterized by sensorineural hearing loss and thyroid dysfunction [2, 34, 35, 110]. Pendrin-deficient mice exhibit EP collapse (~ 0 mV), cochlear malfunction, and vestibular defects; however, recent gene therapy approaches have successfully restored both EP and auditory function in experimental models [23, 33, 46, 51, 58, 83, 91].

Other notable genes affecting cochlear function include SLC22A4, linked to DFNB60-associated deafness [6], and SLC12A6 (KCC3), whose deletion leads to hair cell degeneration [11]. Loss of hepatocyte growth factor (HGF) results in DFNB39, characterized by impaired migration of neural crest–derived cells and a concomitant reduction in EP [74]. Mutations in CHD7, a gene implicated in the formation of enlarged vestibular aqueducts, cause structural abnormalities in the lateral wall and lead to hearing loss [82]. Gap junction proteins are also essential for cochlear homeostasis. Deficiencies in connexins such as Cx26 (GJB2), Cx30 (GJB6), or Cx43 (GJA1) cause reductions in EP and progressive hair cell degeneration, ultimately resulting in hearing impairment [7, 8, 20, 21, 28, 30, 37, 39, 52, 54, 55, 67, 70, 90, 97, 103, 115, 117, 118]. Mutations in Cx31 (GJB3) and Cx32 (GJB1) further underscore the importance of the connexin network, being associated with nonsyndromic hearing loss and X-linked Charcot–Marie–Tooth disease, respectively [95, 111]. Ablation of pannexin-1 (Panx1), expressed in fibrocytes and OSCs, significantly reduces EP (~ + 60 mV), highlighting the role of purinergic signaling in cochlear ion homeostasis [18]. Similarly, disruption of claudin-11, a tight junction protein, impairs EP (~ + 30 mV), emphasizing the necessity of barrier integrity within basal cells of the stria vascularis [38, 59, 60, 69]. In addition, the lack of detectable ATP1A1 and ATP1B1 expression in the human syncytial layer contrasts with functional models derived largely from rodent studies [68]. This discrepancy may reflect species-specific differences or the contribution of alternative Na⁺/K⁺-ATPase isoforms, such as ATP1A2 or ATP1B2, in the human cochlear lateral wall, an issue that remains to be clarified.

Collectively, these findings underscore the multifactorial contributions of ion transport mechanisms, gap junction–mediated intercellular communication, and epithelial barrier integrity to the maintenance of cochlear homeostasis and normal auditory function.

Further Questions, Prospects, and Conclusion

In summary, the intricate interdependence between cochlear K⁺ circulation driven by K⁺ transport and the generation of the EP has been elucidated through a combination of in vivo electrophysiological investigations and quantitative mathematical modeling. Although numerous molecular constituents critical for EP formation and K⁺ homeostasis have been identified, their precise functional contributions and spatial localization within the cochlea remain incompletely characterized.

Future studies should aim to integrate these molecular roles into comprehensive quantitative models to more precisely delineate their contributions to cochlear physiology. From a fundamental biomedical perspective, several key questions remain unresolved—for example, the mechanisms underlying the spatial heterogeneity of EP along the cochlear spiral, and whether the lateral wall exhibits region-specific specializations aligned with frequency tuning. Addressing these challenges will require the development of novel theoretical frameworks capable of capturing phenomena that remain unexplained by current models, through the integration of both established and emerging experimental and computational approaches. Furthermore, the strial vascular system plays an essential supporting role by maintaining metabolic homeostasis and providing the energetic conditions required for sustained ion transport. Disruption of blood flow or oxygen supply is known to cause a rapid decline in EP, indicating that vascular integrity is indispensable for preserving the electrochemical environment of the cochlear lateral wall. Recent genetic and pathological studies further suggest that vascular dysfunction can indirectly compromise EP maintenance. In particular, variants in S1PR2, a receptor with well-established roles in vascular stability and endothelial signaling, are associated with a progressive decline in EP in mice and hearing impairment in humans [49]. In addition, analyses of Norrie disease patients and corresponding animal models demonstrate that abnormalities in cochlear microvasculature precede EP decline and sensory dysfunction [14]. These findings support a model in which vascular pathology acts as a modulatory factor that can secondarily impair EP by disrupting metabolic and ionic balance.

From a clinical otolaryngological perspective, decades of electrophysiological analyses on the inner ear have yielded critical insights into the mechanisms sustaining the EP as an emergent organ-level function. However, these advances have yet to be translated into clinical applications. This translational gap likely stems from the lack of a unifying conceptual framework that bridges molecular mechanisms with whole-organ function.

The integrative strategy proposed in our works—merging experimental and computational methodologies—offers a promising avenue for elucidating organ-level physiological processes in terms of underlying molecular dynamics, thereby providing a theoretical foundation for systems-level interpretation. A comprehensive, mechanistic understanding of normal cochlear function will ultimately facilitate more refined insights into the pathogenesis of auditory disorders and support the development of novel, mechanism-based therapeutic interventions.

Author Contributions

F.N. and H. H. conceived the scope and structure of the review. F.N. and H. H. wrote the original draft. F.N. and H. H. reviewed, edited, and approved the final manuscript.

Funding

Open Access funding provided by Gifu University. This work was partially supported by the following research grants and funds: Grants-in-Aid for Scientific Research (B) 24K02598 (to F.N.), Grants-in-Aid for Scientific Research (A) 24H00798 (to H.H.), AMED-CREST (Multi-sensing: 25gm1510004 and Mechanobiology: 20gm08100004) (to H.H.).

Declarations

Use of AI-Assisted Technologies

During the preparation of this manuscript, the authors used OpenAI ChatGPT (GPT-5.2, OpenAI, San Francisco, CA, USA; used in 2026) solely for assistance with English grammar, clarity, and stylistic refinement of the text. The AI tool did not generate or modify any scientific content, data, analyses, or interpretations. The authors take full responsibility for the integrity and accuracy of the manuscript.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.