Compartmentalization established by claudin-11-based tight junctions in stria vascularis is required for hearing through generation of endocochlear potential

In multicellular organisms, the internal environment must be isolated and buffered against the external environment, and further divided into various compositionally distinct fluid compartments. This compartmentalization is established by cellular sheets of epithelium delineating the body surface and cavities. Epithelial cellular sheets function as diffusion barriers and are also involved in active transport of materials across the barrier to dynamically maintain the internal environment of each compartment. For cellular sheets to exert these physiological functions, there must be some seal to the diffusion of solutes through the paracellular pathway. Tight junctions (TJs) have been shown to be responsible for this intercellular sealing in vertebrates (for reviews, see Schneeberger and Lynch, 1992; Anderson et al., 1995; Balda and Matter, 1998; Tsukita et al., 2001).

On ultrathin section electron microscopy, TJs appear as a series of discrete sites of apparent fusion, involving the outer leaflet of the plasma membranes of adjacent cells (Farquhar and Palade, 1963). On freeze-fracture electron microscopy, TJs appear as a set of continuous, anastomosing intramembranous particle strands (TJ strands) (for a review, see Staehelin, 1974). Our understanding of the molecular architecture of TJs has been rapidly unraveling in recent years. Three closely related proteins called ZO-1, ZO-2 and ZO-3 were identified as constituting the plaque structures underlying plasma membranes together with various other proteins (for a review, see Mitic and Anderson, 1998; González-Mariscal et al., 2000). TJ strands themselves were recently shown to be mainly composed of linearly polymerized integral membrane proteins called claudins with molecular masses of around 23 kDa (Furuse et al., 1998) (for reviews, see Tsukita and Furuse, 1999; Tsukita et al., 2001). Claudin molecules bear four transmembrane domains with both N- and C-termini located in the cytoplasm. Claudins compose a multigene family consisting of 24 members in human and mouse (Morita et al., 1999a) (for a review, see Tsukita et al., 2001). The expression pattern of claudins varies considerably among tissues. Most cell types express more than two claudin species in various combinations to constitute TJ strands. Within individual single strands, distinct species of claudins are co-polymerized to constitute `heteropolymers' (Furuse et al., 1999). In addition to claudins, two other types of integral membrane proteins have been reported to concentrate at TJs, occludin (Furuse et al., 1993) and JAM (Martin-Padura et al., 1998).

As a focus for a discussion of how individual claudin species are involved in compartmentalization in multicellular organisms, the cochlea provides an advantageous experimental system (for a review, see Wangemann and Schacht, 1996). The cochlea includes two compositionally distinct fluid compartments, the scala vestibuli/tympani and the scala media, which are filled with perilymph and endolymph, respectively (reviewed in Ferrary and Sterkers, 1998; Wangemann, 2002) (Fig. 7A). These fluids are amazingly different in their chemical composition: Perilymph resembles extracellular fluids in general (for reviews, see Wangemann and Schacht, 1996; Ferrary and Sterkers, 1998), but endolymph has the characteristics of an intracellular fluid in that it has high K⁺ and low Na⁺ concentrations (for a review, see Sterkers et al., 1988). Furthermore, the electrical potential of endolymph [endocochlear potential (EP)] is positive by ∼80-90 mV relative to perilymph (Tasaki and Spyropoulos, 1959; Konishi et al., 1978). It is now widely accepted that these characteristics of endolymph (high K⁺ concentration and EP) are indispensable for cochlear hair cells to transduce acoustic stimuli into electrical signals (for a review, see Hudspeth, 1989). To maintain these characteristic features of endolymph, TJs are well developed in the epithelial cellular sheets delineating the endolymph compartment (the scala media) (Janke, 1975a; Janke, 1975b; Gulley and Reese, 1976). In good agreement, it has been reported that loss-of-function mutations of claudin-14, which are highly expressed in the organ of Corti of cochlea, caused hereditary deafness in humans (Wilcox et al., 2001).

Thus, the question is how the endolymph compartment with its peculiar chemical compositions and electrical potential is generated and maintained within the cochlea. The stria vascularis is thought to secrete K⁺ from perilymph into endolymph and to generate an EP of +80-90 mV (von Bekesy, 1950; Tasaki and Spyroupoulos, 1959; Konishi et al., 1978; Wangemann et al., 1995; Wangemann, 2002). The stria vascularis is a highly vascularized epithelium and part of the epithelial cellular sheet delineating the endolymph compartment. Interestingly, it consists of two epithelial barriers carrying well-developed TJs - the marginal and basal cell layers (Janke, 1975a; Janke, 1975b) (Fig. 1B). This compartmentalization within the stria vascularis was hypothesized to be essentially important for the K⁺ secretion and the EP generation (for a review, see Wangemann, 2002) but it was difficult to evaluate this hypothesis experimentally. In an earlier study, we found that claudin-11 was primarily expressed in TJs of basal cells in the stria vascularis (Kitajiri et al., 2004). Claudin-11 was previously reported to be expressed in large amounts in the oligodendrocytes in the brain and Sertoli cells in the testis (Morita et al., 1999b; Gow et al., 1999) but, in the inner ear, its expression was restricted to the basal cells. Therefore, in this study, with an expectation that the deficiency of claudin-11 selectively affects compartmentalization in the stria vascularis, we generated claudin-11-deficient mice and examined the cochlea of these mice in detail.

Rat anti-mouse occludin monoclonal antibody (mAb) (MOC37) and rabbit anti-mouse claudin-11 polyclonal antibody (pAb) were raised and characterized as described previously (Saitou et al., 1997; Morita et al., 1999b). Rabbit anti-human ZO-1 pAb was purchased from Zymed (San Francisco, CA). Anti-mouse ZO-1 mAb was produced in rats using the synthesized polypeptides corresponding to the N-terminal half of ZO-1 as an antigen. Rat anti-E-cadherin mAb (ECCD-2) was provided by M. Takeichi (RIKEN Center for Developmental Biology, Kobe, Japan).

The targeting vector was constructed by ligating a 6.0 kb HindIII/SmaI fragment and a 2.8 kb KpnI/ScaI fragment, which are located upstream and downstream of exon 1, respectively, to the pgk-neo cassette. The diphtheria toxin A expression cassette (MC1pDT-A) was placed outside the 3′ arm of homology for negative selection. This targeting vector was designed to delete exon 1 (Fig. 2A), like the previous report of generation of claudin-11-deficient mice (Gow et al., 1999). J1 embryonic stem (ES) cells were electroporated with the targeting vector and selected for ∼9 days in the presence of G418. G418-resistant colonies were removed and screened by Southern blotting with the 3′ external probe (Fig. 2B). Correctly targeted ES clones were identified by an additional 3.8-kb band together with the 5.8-kb band of the wild-type allele when digested with EcoRI. The targeted ES cells obtained were injected into C57BL/6 blastocysts, which were in turn transferred into Balb/c foster mothers to obtain chimeric mice. Male chimeras were mated with C57BL/6 females, and agouti offspring were genotyped to confirm the germ-line transmission of the targeted allele. The littermates were genotyped by Southern blotting. Heterozygous mice were then interbred to produce homozygous mice.

Auditory-brainstem-response (ABR) measurements were performed in a soundproof room according to the method described previously (Zheng et al., 1999). In general, ABR waveforms were recorded for 12.8 milliseconds at a sampling rate of 40,000 Hz using 50-5000 Hz filter settings; waveforms recorded from 1024 stimuli at a frequency of 9 Hz were averaged. ABR waveforms were recorded in decreasing 5-dB sound pressure level (SPL) intervals from the maximum amplitude until no waveforms could be visualized.

Temporal bones were removed from 10-week-old mice and the round and oval windows were opened, together with the small holes in the cochlear apical turn and superior semicircular canal. The perilymphatic space was gently perfused with 10% trichloroacetic acid (TCA) from the round to the oval window (Hayashi et al., 1999). Then the samples were immersed in 10% TCA for 1 hour, washed three times with PBS and decalcified with 5% EDTA in PBS for 3 days. Specimens were immersed in 30% sucrose in PBS for 1 day and then frozen in liquid nitrogen. Frozen sections ∼10 μm thick were cut from these samples and air dried on slide glasses. Frozen sections were treated with 0.2% Triton X-100 in PBS for 15 minutes, soaked in 1% bovine serum albumin in PBS and then incubated with primary antibodies for 30 minutes. They were then washed three times with PBS and incubated with secondary antibodies for 30 minutes. Cy3-conjugated anti-rat IgG pAbs or Alexa-488-conjugated anti-rabbit IgG pAbs were used as secondary antibodies. After being washed with PBS, samples were embedded in 95% glycerol in PBS containing 0.1% paraphenylendiamine and 1% n-propylgalate, and observed using a DeltaVision system (Version 2.10; Applied Precision) equipped with an Olympus IX70 microscope (Olympus, Tokyo, Japan) and a cooled charge-coupled device (CCD) system.

Temporal bones obtained from 10-week-old mice were fixed by perilymphatic perfusion as described above with 2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2) overnight at 4°C and washed three times with 0.1 M phosphate buffer. The portion containing the stria vascularis was carefully dissected out, immersed in 30% glycerol in 0.1 M phosphate buffer for 2 hours and then frozen in liquid nitrogen. Frozen samples were fractured at -110°C and platinum shadowed unidirectionally at an angle of 45° in a Balzers Freeze Etching System (BAF060; Bal-Tec). Samples were then immersed in household bleach and replicas floating off the samples were washed with distilled water. Replicas were picked up on formvar-filmed grids and examined with an electron microscope (model H-7500; Hitachi).

Temporal bones were removed from 10-week-old mice and the round and oval windows were opened in PBS containing 1 mM CaCl₂. As mentioned above, the perilymph space was carefully perfused with 100 μl 10 mg ml^-1 EZ-Link™ Sulfo-NHS-LC-Biotin (Pierce Chemical, Rockford, IL) in PBS containing 1 mM CaCl₂ for 5 minutes (Chen et al., 1997) followed by perfusion five times with PBS containing 1 mM CaCl₂. The temporal bones were then fixed by perilymphatic perfusion with 10% TCA for 1 hour and processed for immunofluorescence microscopy. The distribution of injected biotin tracer was visualized by incubating frozen sections with streptavidin/fluorescein-isothiocyanate (FITC) (Oncogene Research Products, Boston, MA) for 30 minutes.

10-week-old mice were anesthetized by an intraperitoneal and an intramuscular injection of the mixture of xylazine (12 mg kg^-1) and ketamine (50 mg kg^-1). The tympanic bulla was exposed using an opisthotic approach and double-barreled K⁺-selective microelectrodes (or a pair of conventional and single-barreled K⁺-selective microelectrodes) were inserted into the endolymph compartment (the scala media) of the basal turn under a stereomicroscope at 80-120× magnification (Model TS; Leitz, Germany). The single- and double-barreled K⁺-selective microelectrodes were fabricated and evaluated as described previously (Fujimoto et al., 1988; Takamaki et al., 2003). The K⁺-selective electrodes were filled with the K⁺-selective liquid ion exchanger (Potassium Ionophore I - Cocktail B #60398; Fluka, Buchs, Switzerland). Each barrel of the microelectrodes for measurement of EP and the potassium potential was connected to a differential electrometer (FD-223; World Precision Instruments, FL) via an Ag/AgCl pellet (MEH1SF; World Precision Instruments), and the outputs of the endolymph were connected to a four-channel recorder (U-638, Nippon Denshi Kagaku, Tokyo, Japan) and a Mac Lab 8s (ADInstruments, New South Wales, Australia). The output of the electromotive force (EMF) from K⁺-selective microelectrodes was recorded by subtracting the EMF from the microelectrode measuring EP with a differential electrometer.

where E150 is the voltage from the output of the K⁺-selective microelectrode in a 150 mM KCl solution, EKe is the voltage difference between the endocochlear voltage K⁺-selective microelectrode and endocochlear potentials (EP), and α is the slope constant of the K⁺-selective microelectrodes. The recording in the endolymph was accepted only when the EP was stable for more than 30 seconds.

On the lateral wall of the cochlear duct of wild-type mice, the stria vascularis was easily identified as a band of specialized stratified epithelium (Fig. 1A). As shown in Fig. 1B, the cochlear stria vascularis consists mainly of two epithelial cell layers, marginal and basal. Marginal cells, that face the endolymph compartment (the scala media) with their apical membranes are simple epithelia with well-developed TJs at the most apical portion of their lateral membranes and with a labyrinthine system of narrow cell process at their basal portion. Basal cells also have many cellular processes and establish another barrier by adhering to each other through well-developed TJs. These findings indicate that the stria vascularis itself constitutes an isolated compartment in the cochlea (intrastrial space; II in Fig. 1B).

In an earlier study, we examined the expression and distribution patterns of claudins in the mouse inner ear (Kitajiri et al., 2004): Marginal cells expressed numerous species of claudins such as claudins 1, 2, 3, 8, 9, 10, 12, 14 and 18, whereas only claudin-11 was detectable in basal cells. As shown in Fig. 1C, when frozen sections of stria vascularis were double stained with anti-claudin-11 pAb and anti-occludin mAb, occludin-positive TJs were observed both in marginal and basal cell layers, whereas claudin-11 was exclusively concentrated at TJs of basal cells. Thus, we expected that the deficiency of claudin-11 would selectively affect the barrier of basal cell layers.

We produced mice unable to express claudin-11. Nucleotide sequencing as well as restriction mapping identified three exons that cover the whole open reading frame of Cld11. The putative exon 1 contained the first ATG and encoded the N-terminal portion (amino acids 1-75) of the claudin-11 molecule containing the first transmembrane domain and a part of the first extracellular loop (Fig. 2A). We constructed a targeting vector that was designed to disrupt the Cld11 gene by replacing exon 1 with the neomycin resistance gene (Fig. 2A). Mice were generated from ES cell clones in which the Cld11 gene was disrupted by homologous recombination. Southern blotting confirmed the disruption of the Cld11 gene in heterozygous as well as homozygous mutant mice (Fig. 2B), and a reverse-transcription polymerase chain reaction detected no Cld11 mRNA from the testis of homozygous mutant mice (Fig. 2C).

Claudin-11-deficient (Cld11^-/-) mice were born in the expected Mendelian ratios and looked normal macroscopically. We had previously found that claudin-11 is mainly expressed in oligodendrocytes in the brain and Sertoli cells in the testis (Morita et al., 1999b); in good agreement with this, Gow et al. (Gow et al., 1999) reported that mice lacking the expression of claudin-11 exhibited neurological and reproductive deficits (weakness of hindlimb and male sterility). Our newly generated Cld11^-/- mice also showed these deficits (data not shown).

We then examined the hearing ability of Cld11^-/- mice. The ABR was measured in 10-week-old Cld11^+/+ and Cld11^-/- mice in response to a stimuli with a sound pressure level (SPL) of 0-80 dB (16 kHz). As shown in Fig. 3A, hearing ability was significantly affected in Cld11^-/- mice. The hearing thresholds of 10-week-old Cld11^+/+, Cld11^+/- and Cld11^-/- mice are summarized in Fig. 3B. Cld11^+/+ and Cld11^+/- mice showed normal hearing thresholds (10-20 dB SPL), whereas Cld11^-/- mice showed abnormally increased hearing thresholds (40-60 dB SPL).

We compared light microscopic images of toluidine blue-stained sections prepared from Epon-embedded tissues of the inner ear between Cld11^+/+ and Cld11^-/- mice aged 10 weeks, but no obvious gross morphological malformations were observed in Cld11^-/- cochlea (Fig. 4A). On the lateral wall of the Cld11^-/- cochlea, stria vascularis was easily identified normally except for some edematous appearance. We then examined the distribution of claudin-11, occludin, ZO-1 and E-cadherin in the Cld11^-/- stria vascularis by immunofluorescence microscopy (Fig. 4B). TJs in marginal cell layers did not appear to be affected: both occludin and ZO-1 were concentrated at the most apical region of their lateral membranes. By contrast, in basal cell layers lacking expression of claudin-11, neither occludin nor ZO-1 was clearly concentrated at cell-cell adhesion sites, where E-cadherin was concentrated. These findings suggested that, in Cld11^-/- stria vascularis, TJs disappeared from basal cell layers, although the cellular organization of stria vascularis did not appear to be extensively affected. Indeed, this was confirmed by freeze-fracture replica electron microscopy. As previously reported (Janke, 1975a; Janke, 1975b), TJs of basal cells in Cld11^+/+ stria vascularis were characterized by many parallel strands that did not extensively anastomose (Fig. 5a). From low-power freeze-fracture replica images obtained from Cld11^-/- stria vascularis, we successfully identified marginal and basal cell layers (Fig. 5b). Marginal cells bore well-developed TJs (Fig. 5c), whereas no TJ strands were observed between adjoining basal cells (Fig. 5d).

The question then arose of whether the barrier of basal cell layers of stria vascularis is affected in Cld11^-/- mice, and so we performed a tracer experiment. Temporal bones were removed, and the round and oval windows were opened. Then, the perilymph compartment was gently perfused from the round to oval windows with an isotonic solution containing a primary amine-reactive biotinylation reagent (557 Da), which covalently cross-links to an accessible cell surface. After a 5 minute incubation followed by perfusion with PBS, the cochlea was dissected out, fixed and frozen. Frozen sections of stria vascularis were double labeled with anti-E-cadherin pAb (red) and streptavidin (green) to detect cell layers and bound biotin, respectively (Fig. 6). In Cld11^+/+ stria vascularis, the biotinylation reagent diffused freely through the connective tissue, represented as `compartment III' in Fig. 1B, but did not get into the intrastrial space (`compartment II' in Fig. 1B). In sharp contrast to this, the diffusion of the biotinylation reagent in Cld11^-/- cochlea was not stopped at the basal cell layers, but the reagent appeared to pass them to reach the intrastrial space of the stria vascularis. Interestingly, this reagent did not diffuse into the endolymph compartment across the marginal cell layers (`compartment I' in Fig. 1B). These findings clearly indicated that, in stria vascularis, a deficiency of claudin-11 affects the barrier of basal cell layers but not of marginal cell layers.

Stria vascularis has long been believed to be the motor of K⁺ secretion into endolymph and the origin of the endocochlear potential (EP) (for a review, see Wangemann, 2002). Therefore, we next compared K⁺ concentrations of endolymph and EPs between Cld11^+/+ and Cld11^-/- cochlea. For this purpose, a double-barreled K⁺-selective microelectrode or a pair of single-barreled K⁺-selective and conventional microelectrodes was directly inserted into the endolymph compartment (the scala media) of the basal turn of cochlea of Cld11^+/+ and Cld11^-/- mice under anesthetic (Fig. 7A). Interestingly (Fig. 7B), the claudin-11 deficiency (i.e. the destruction of the barrier of basal cell layers) did not change the potassium concentration of the endolymph [Cld11^+/+, 148±7 mM (n=6); Cld11^-/-, 145±6 mM (n=6)] but significantly suppressed the generation of EP [Cld11^+/+, 95±6 mV (n=6); Cld11^-/-, 31±14 mV (n=6)].

For homeostasis in multicellular organisms, various compositionally distinct fluid compartments must be established through the barrier function of TJs in epithelial and endothelial cells. Recent identification of claudins, cell adhesion molecules responsible for the TJ barrier, opened a new way to perturb individual compartments and to evaluate the physiological relevance of each compartment at a whole-body level. To date, 24 members of the claudin family have been identified in humans and mice, and these are reported to be expressed in individual cell layers in various combinations and mixing ratios (for reviews, see Tsukita et al., 2001; Gonzalez-Mariscal et al., 2003). It has been shown that a particular compartment can be destroyed when the gene for major species of claudins constituting TJs of its delineating epithelial cell layer is homozygously knocked out in mice. For example, claudin-1 was expressed in large amounts in the epidermis and claudin-1-deficient mice showed severe dysfunction of the epidermal barrier, being dehydrated quickly after birth (Furuse et al., 2002). Endothelial cells of brain blood vessels primarily expressed claudin-5 and, in claudin-5-deficient mice, the blood-brain barrier was severely affected (Nitta et al., 2003). The Cld11 gene, which this study focused on, has already been knocked out by Gow et al. (Gow et al., 1999) and, in these mice, the compartments established by oligodendrocyte myelin sheaths and Sertoli cells were affected, resulting in neurological and reproductive deficits.

From the viewpoint of compartmentalization, the cochlea is most intriguing. In particular, endolymph is unique in its high K⁺ concentration and positive electrical potential, the EP (for a review, see Ferrary and Sterkers, 1998). Accumulated evidence indicates that the stria vascularis is the actual site where K⁺ is secreted into endolymph and the EP is generated (for reviews, see Wangemann et al., 1995; Wangemann, 2002). Interestingly, the stria vascularis itself constitutes a tube-like isolated fluid compartment delineated by two distinct, marginal and basal, cell layers. This tubular compartment runs spirally in tight association with the endolymph compartment, the scala media. The question thus naturally arose about stria vascularis of what, in terms of its K⁺ secretion and the EP generation, is the physiological relevance of the establishment of such a peculiar compartment. One of the most conclusive ways to answer this was selectively to destroy the stria vascularis compartment and to examine the effects on hearing ability itself and the K⁺ concentration and EP of the endolymph. We previously examined the expression patterns of claudins in the cochlea in detail, and found that claudin-11 was the major constituent of TJs of basal cell layers of stria vascularis (Kitajiri et al., 2004). In good agreement with this, TJs of these cells were reported to be characterized by parallel, densely packed strands (Janke, 1975a; Janke, 1975b; Gulley and Reese, 1976), which were also observed in claudin-11-based TJs of Sertoli cells (Gilula et al., 1976; Russell et al., 1985). Considering that the claudin-11 expression was restricted to the basal cell layers of stria vascularis, these findings led us to speculate that, when the Cld11 gene is knocked out, the stria vascularis compartment is selectively destroyed without affecting other cochlear compartments. We thought that this speculation could be evaluated experimentally, because claudin-11-deficient mice were reported to be born alive and grow without severe defects (Gow et al., 1999).

Then, we generated claudin-11-deficient (Cld11^-/-) mice using a conventional homologous recombination method and examined the structure and functions of their cochlea, especially of stria vascularis, in detail. ABR measurements showed that Cld11^-/- mice suffered from deafness. As expected, in Cld11^-/- cochlea no obvious gross morphological malformations were observed and the tracer experiments clearly revealed that the barrier function of basal cell layers, but not marginal cell layers, of stria vascularis was severely affected. Very interestingly, when the K⁺ concentration and EP were measured directly using K⁺-sensitive microelectrodes from the scala media, in Cld11^-/- cochlea, the K⁺ concentration was maintained around a normal level (∼150 mM) but the latter was significantly suppressed down to ∼30 mV. These findings led to two conclusions. The first, that the barrier function of the marginal cell layer is sufficient for generation and maintenance of the high K⁺ concentration of endolymph. This is consistent with previous electrophysiological data (Konishi et al., 1978; Wangemann et al., 1995). The second conclusion is that the basal cell barrier [i.e. the compartmentalization in stria vascularis (intrastrial space)] is indispensable for the generation and maintenance of the EP.

As regards the mechanisms for generation of EP, two distinct models have been proposed, a `single-cell model' and a `two-cell model' (Fig. 8) (for a review, see Wangemann, 1995). The single-cell model hypothesizes that Na⁺ conductance of the basolateral membrane of marginal cells generates a large positive membrane voltage that is the source for the positive EP (Offner et al., 1987). In this model, the contribution of basal cell layers to the EP generation is not considered. In the two-cell model, K⁺ conductance localized to the inner membrane of basal cells and to intermediate cells that are connected to basal cells through gap junctions, which were assumed to generate the source of EP (Salt et al., 1987; Kikuchi et al., 1995). In this model, the involvement of marginal cells in the generation of EP was limited to the maintenance of the low K⁺ concentration in the intrastrial space. Recent detailed electrophysiological data appear to favor the two-cell model (Wangemann et al., 1995; Takeuchi et al., 1995; Takeuchi et al., 2000; Marcus et al., 2002), but it was technically difficult to evaluate conclusively the importance of the compartmentalization in the stria vascularis for the EP generation.

Set against this situation, the data obtained in this study clearly supported the two-cell model (Fig. 8). It is difficult to explain the downregulation of EP in Cld11^-/- cochlea by the one-cell model. Interestingly, the EP in Cld11^-/- cochlea was not suppressed completely down to 0 mV, but still showed ∼30 mV. It is likely that this voltage simply represents a residual electrical resistance between the intrastrial space and the spiral ligament, because that space is extremely small and tortuous in shape. Cld11^-/- mice would be useful in future experiments aiming to evaluate this speculation.

Claudin-11 was shown to constitute TJ strands between lamellae of myelin sheaths of oligodendrocytes in the brain, and between adjacent Sertoli cells in the testis (Morita et al., 1999b). In claudin-11-deficient mice, as established by Gow et al. (Gow et al., 1999), TJ strands were absent in myelin sheaths of oligodendrocytes and Sertoli cells, conclusively demonstrating that, in these types of cell, TJ strands are mainly composed of a single specific claudin, claudin-11. In this study, we established another line of Cld11^-/- mice and demonstrated that TJ strands in basal cells of stria vascularis in the cochlea were also singly composed of claudin-11. This situation is peculiar, because in most epithelial cellular sheets, TJ strands are composed of more than two distinct species of claudins as heteropolymers (for a review, see Tsukita et al., 2001). Interestingly, in addition to oligodendrocytes, Sertoli cells and cochlear basal cells, claudin-11 is reported to be expressed in renal epithelial cells of the thick ascending limb of Henle, in which, by contrast, claudin-11 appeared to form heteropolymers together with claudins 3, 10 and 16 (Kiuchi-Saishin et al., 2002). Through detailed analyses of claudin-11-deficient mice, we can now state that three important physiological processes, saltatory conduction along axons, spermatogenesis and hearing are fully dependent on the compartmentalization established by TJ strands consisting of a single species of claudin, claudin-11. A question then naturally arises: why is claudin-11 used singly for such important physiological processes, even though there are many other claudin species? It is still too soon to answer this question, but these findings indicate that the physiological relevance of the existence of many claudin species is not a simple safety measure based on functional redundancy. TJs are not a simple barrier: they show ion and size selectivity, and the tightness of their barrier function varies significantly depending on cell type. That cell-type-specific properties of TJ strands are determined by the combination and mixing ratios of claudins within individual TJ strands is widely accepted, so it is fascinating to speculate that TJs in oligodendrocytes, Sertoli cells and cochlear basal cells are highly specialized in terms of their barrier function, which could be why claudin-11 is used singly in these TJs. Cld11^-/- mice will provide a valuable resource in the future, not only for further study of the molecular mechanisms of hearing but also for gaining a better understanding of the physiological relevance of the existence of so many claudin species.

We thank all the members of our laboratory (Department of Cell Biology, Faculty of Medicine, Kyoto University, Japan) for helpful discussions. We also thank T. Nakagawa (Department of Otolaryngology Head and Neck Surgery, Kyoto University, Japan) and H. Takenaka (Department of Otolaryngology, Osaka Medical College, Japan) for helpful discussions. This study was supported in part by a Grant-in-Aid for Cancer Research and a Grant-in-Aid for Scientific Research (A) from the Ministry of Education, Science and Culture of Japan to S.T.