Immunolocalization of aquaporin CHIP in the guinea pig inner ear

KONSTANTINA M. STANKOVIC, JOE C. ADAMS, AND DENNIS BROWN Division of Health Sciences and Technology, Harvard-Massachusetts Institute of Technology, Eaton-Peabody Laboratory, Massachusetts Eye and Ear Infirmary, Departments of Pathology and of Otology and Laryngology, Harvard Medical School, and The Renal Unit, Massachusetts General Hospital, Boston, Massachusetts 02114

Stankovib, Konstantina M., Joe C. Adams, and Den- nis Brown. Immunolocalization of aquaporin CHIP in the guinea pig inner ear. Am. J. Physiol. 269 (Cell Physiol. 38): C1450-C1456, 1995Aquaporin CHIP (AQP-CHIP) is a water channel protein previously identified in red blood cells and water transporting epithelia. The inner ear is an organ of hearing and balance whose normal function depends critically on maintenance of fluid homeostasis. In this study, AQP- CHIP, or a close homologue, was found in specific cells of the inner ear, as assessed by immunocytochemistry with the use of affinity-purified polyclonal antibodies against AQP-CHIP. AQP- CHIP was predominantly found in fibrocytes in close associa- tion with bone, including most of the cells lining the bony labyrinth and in fibrocytes lining the endolymphatic duct and sac. AQP-CHIP-positive cells not directly apposing bone in- clude cells under the basilar membrane, some type III fibro- cytes of the spiral ligament, fibrocytes of the spiral limbus, and the trabecular perilymphatic tissue extending from the mem- branous to the bony labyrinth. AQP-CHIP was also found in the periosteum of the middle ear and cranial bones, as well as in chondrocytes of the oval window and stapes. The distribu- tion of AQP-CHIP in the inner ear suggests that AQP-CHIP may have special significance for maintenance of bone and the basilar membrane, and for function of the spiral ligament.

cochlea; spiral ligament; basilar membrane; endolymphatic duct and sac; perilymph

WATER CHANNELS, ALSO KNOWN as aquaporins, are pro- teins that facilitate rapid and selective transmembrane movement of water down an osmotic gradient. Aquapo- rin CHIP (channel-forming integral protein of 28 kDa) (AQP-CHIP) is a constitutively active water channel first isolated from human erythrocytes (5), subse- quently purified, and found to be homologous to the major intrinsic protein of the lens, MIP26 (21). AQP- CHIP was first demonstrated to function as a molecular water channel by expression of the cRNA in oocytes (17). This function was verified by reconstitution of highly purified AQP-CHIP protein into proteoliposomes (28), in which AQP-CHIP markedly increased osmotic water permeability but excluded ions and other small mol- ecules. The function of AQP-CHIP in water movement was also confirmed by analysis of membranes partially depleted of other proteins and reconstitution of partially purified protein (25). AQP-CHIP is found in abundance in water-permeable segments of the nephron, where it facilitates reabsorption of water from glomerular fil- trate (16, 19). AQP-CHIP is also abundant in other absorptive or secretory epithelia, such as the choroid plexus, ciliary body, capillary endothelia, hepatobiliary ductules, gall bladder, and male reproductive tract (2-4,

7, 15, 16, 19>, where it is thought to facilitate isosmotic fluid secretion and reabsorption. AQP-CHIP mRNA is transiently observed in periosteum, endocardium, and cornea during fetal development (3).

We examined inner ear tissue for the presence of the AQP-CHIP water channel. The inner ear is an organ containing cells involved in water transport, electrolyte regulation, and sensory transduction. It is thought that the inner ear shares structural and anatomic similari- ties with the kidney (13) and that the two organs may utilize similar physiological mechanisms of electrolyte regulation because certain drugs, such as aminoglyco- side antibiotics and nonsteroidal anti-inflammatory drugs, are both ototoxic and nephrotoxic (6, 27). Using immunocytochemical methods, we found that AQP- CHIP, or a close homologue, is present only in limited types of cells of the inner ear.

MATERIALS AND METHODS

Inner ears were obtained from six albino guinea pigs with body weights of 250-300 g. The animals were anesthetized with intraperitoneally delivered urethane (1.5 g/kg) and exsan- guinated by transcardial perfusion with 20 ml warm physiologi- cal saline containing 0.1% sodium nitrite and then 100 ml of fixative. The fixative was a 10% solution of Formalin in saline containing 0.5% zinc dichromate with the pH adjusted to 5.0 with NaOH. Each bulla was then opened rapidly, followed by removal of the stapes, perforation of the round window, and gentle injection of 0.2-0.4 ml of fixative into the Scala vestibuli through the oval window. Specimens were then placed in fixative for 30 min at room temperature or 60 min to 24 h at 4”C, after which they were flushed with saline to remove the excess fixative.

The fixed inner ears were decalcified by immersion in 0.12 M EDTA (pH 7.0), with gentle stirring for lo-30 days at room temperature. The EDTA solution was changed every 2-3 days. The tissues were then dehydrated through a graded series of alcohols, cleared in xylene, and embedded in paraffin.

The paraffin-embedded inner ears were cut serially in 6-pm-thick sections. Every 20th section was stained with 0.1% azure in acetate buffer with pH adjusted to 4.1. Sections from selected regions were deparaffinized, hydrated, rinsed in dis- tilled water, and immunostained with affinity-purified antibod- ies against AQP-CHIP or basolateral intrinsic membrane protein (BLIP) raised in rabbit. BLIP is an AQP-CHIP-related protein that is localized on the basolateral membranes of kidney collecting duct principal cells and gastric parietal cells (24). The characterization and affinity purification of these antibodies have been published previously (19,24).

Two different indirect methods of visualization were used to increase credibility of experimental results; the first method employed a fluorescein-conjugated goat anti-rabbit antibody, and the second method was a peroxidase-staining technique that used biotinylated donkey anti-rabbit secondary antibody.

Cl450 0363-6143/95 $3.00 Copyright o 1995 the American Physiological Society

AQUAPORIN CHIP IN GUINEA PIG INNER EAR Cl451

For immunofluorescence, sections were deparaffinized, rehy- drated through a graded series of ethanol dilutions, washed three times for 5 min in phosphate-buffered saline (PBS), and incubated with 1% bovine serum albumin in PBS for 10 min. The sections were then incubated with anti-AQP-CHIP affinity- purified rabbit anti-serum diluted to 1:800 for 12 h at 4”C, followed by two 5min washes in high-salt PBS (containing 2.7% NaCl to decrease nonspecific antibody binding) and one wash in regular PBS. The sections were incubated for 1 h at room temperature with a fluorescein-conjugated goat anti- rabbit antibody (15 pg/ml in PBS; Calbiochem-Behring, San Diego, CA) and washed twice in high-salt PBS and once in PBS. Counterstaining with Evans blue (0.01% in PBS) for 1 min was then performed, followed by two 5min washes in PBS. Sections were mounted in 50% glycerol in 0.2 M tris(hydroxymethyl)aminomethane l HCl, pH 8.0, containing 1% n-propyl gallate to retard quenching of the fluorescent signal. When a peroxidase-staining technique was used, either a conventional or a biotin-amplified avidin-biotin complex method was used, as previously described (1). Some slides were immunostained with a rabbit anti-Na+-K+-ATPase diluted to 1:300,000 to identify dark cells of the semicircular canals and the utricle and to establish the relationship among AQP-CHIP positive cells and the ATPase rich dark cell epithelia of those organs (23).

Negative control sections were processed in parallel. These sections were treated as described above, except the preimmune serum was used in place of the primary antiserum. An addi- tional control was preabsorption of the AQP-CHIP antibody with human erythrocyte membranes that were enriched in AQP-CHIP by KI and detergent extraction of other proteins (26).

RESULTS

Affinity-purified, polyclonal antibodies against AQP- CHIP and BLIP stained selected cell populations in both the auditory and the vestibular parts of the inner ear, as well as around the endolymphatic duct and sac. A total of nine inner ears from six animals were immunostained with anti-AQP-CHIP antibodies, and six inner ears from four animals were immunostained with anti-BLIP anti- bodies. The same immunostaining patterns were ob- tained with anti-AQP-CHIP and anti-BLIP antibodies, as described below.

Auditory part of the inner ear. The strongest staining in the cochlea was found in mesenchymal cells below the basilar membrane (Fig. lA, arrowheads). Staining was also found in fibrocytes of the spiral ligament adjacent to the bone (Fig. lA, thin arrows), i.e., type III fibrocytes, previously described to be cytochemically different from other fibrocytes of the ligament (8, 22). AQP-CHIP immunoreactivity was clearly associated with cells, rather than the extracellular matrix (Fig. 1, B and C). Within cells, immunoreactivity was associated with the plasma membranes, as illustrated on the image of type III fibrocytes (Fig. 1C). Immunostained type III fibro- cytes were present in all cochlear turns, but there were progressively fewer stained fibrocytes with increasing distance from the basal to the apical turns of the cochlea.

AQP-CHIP was also immunolocalized in mesenchy- ma1 cells lining the Scala tympani and Scala vestibuli (Fig. lA, open arrows). Staining of mesenchymal cells lining the Scala vestibuli was nonhomogeneous and somewhat less intense than that of the Scala tvmpani.

This may reflect the fact that mesenchymal cells lining the Scala vestibuli are widely distributed and have very sparse cytoplasm. A subset of fibrocytes of the spiral limbus, as well as supralimbal cells, was recognized by the antibody (Fig. 2).

A dilution series of the primary antibody revealed that the antigen recognized by the affinity-purified polyclonal anti-AQP-CHIP antibody is most abundant or most tightly bound in cells below the basilar membrane. As the concentration of the primary antibody was de- creased in three steps from 1:2,000 to l:lO,OOO when a biotin-amplified ABC method was used (l), the staining intensity of all cells except subbasilar membrane cells was significantly reduced. At a primary antibody dilu- tion of 1:20,000, the intensity of the subbasilar mem- brane cells was clearly reduced (not shown).

Vestibular part of the inner ear. Many AQP-CHIP positive cells were present in the lining of the vestibular part of the bony labyrinth. Because all parts of endos- teum were not equally well fixed, it could not be established that the entire vestibular part of the bony labyrinth is lined by AQP-CHIP positive cells. However, our data suggest that this is likely to be the case.

Staining was also observed in fibrocytes of the trabecu- lar perilymphatic tissue that extends from the epithe- lium of the membranous labyrinth to the endosteum of the bony labyrinth. This pattern is illustrated in Fig. 3A, which shows a cross section of a semicircular canal. Overall, the staining intensity of fibrocytes of the vestibu- lar system was comparable to that of type III fibrocytes of the spiral ligament in the cochlea.

The most intense staining of the perilymphatic tissue was seen in fibrocytes beneath the dark cell epithelia of the ampullae of the semicircular canals (Fig. 3B). Dark cells of the ampullae were identified by immunostaining adjacent paraffin sections for Na+-K+-ATPase, as previ- ously reported (23).

In the utricle, AQP-CHIP immunostaining was lo- cated in fibrocytes beneath the sensory epithelium (Fig. 3C), as well as in fibrocytes of the endosteum. No staining by anti-AQP-CHIP antibodies was found in the saccule.

Endolymphatic duct and sac. The proximal part of the endolymphatic duct was not stained with anti-AQP- CHIP antibodies. However, fibrocy-tes in the connective tissue around the distal part of the endolymphatic duct approaching the posterior fossa, as well as cells in the connective tissue around the endolymphatic sac, were strongly positive (Fig. 30). The cells that were positively stained were continuous with the cells of the meninges. The epithelial lining of the endolymphatic duct and sac was not stained (arrowheads in Fig. 30).

Middle ear and cranial bones. Sections of the cochlea that also included parts of the middle ear space demon- strated prominent staining of fibrocytes lining the bone of the middle ear (Fig. 1A). AQP-CHIP positive cells were also identified among chondrocytes of the round window, as well as chondrocytes of the footplate and the head of the stapes (Fig. 4), and in periosteum of the stapes. In addition, AQP-CHIP immunoreactivity was present in the endosteum of the internal auditorv

Cl452 AQUAPORIN CHIP IN GUINEA PIG INNER EAR

Fig. 1. Immunostaining for aquaporin CHIP (AQP-CHIP) on 6-pm paraffm sections of guinea pig cochlea. Most intensely stained cells are mesenchymal cells below basilar membrane (arrowheads in A and B) and type III fibrocytes of spiral ligament (thin arrows in A and C). Membrane-associated staining of type III fibrocytes is evident in C -(arrows). Additional stained sites include mesenchymal cells lining Scala tympani (ST) and Scala vestibuli (SV, open arrows), bone-lining cells of middle ear (ME, thick arrow), and a subset of fibrocytes of spiral limbus along with supralimbal cells (boxed region in A). Boxed region is shown magnified in Fig. 2. *Bone. Note absence of staining in epithelial cells facing Scala media (SM). A stylized cochlear cross section is shown at bottom. Dilution of primary antibody is 1:800. Bar, 100 km (A) and 10 km (B and C).

AQUAPORIN CHIP IN GUINEA PIG INNER EAR Cl453

Fig. 2. Immunolocalization of AQP-CHIP in a subset of fibrocytes of spiral limbus (thin arrows), and in supralimbal cells (arrowheads). Boxed region from Fig. 1A is shown at higher magnification. Open arrow points to Reissner’s membrane, which separates Scala media from Scala vestibuli. Primary antibody dilution is 1:800. Bar, 30 Frn.

meatus, in the endosteum of the bony recess in whichthe stapedius muscle is located, as well as in the periosteum between adjacent cranial bones.

Negative controls treated with the preimmune serum showed an absence of staining in cells recognized by anti-AQP-CHIP antibodies. AQP-CHIP staining was completely abolished after preabsorption of the AQP- CHIP antibody with human erythrocyte membranes that were enriched in AQP-CHIP by KI and detergent extraction of other proteins (26) (Fig. 5).

DISCUSSION

Affinity-purified polyclonal antibodies against AQP- CHIP recognized only certain cells of the inner ear. A common feature of these cells is that none are epithelial cells, suggesting that water channels in the inner ear are not involved with transepithelial fluid exchange or that the epithelial cells contain an as yet unidentified mem- ber of the AQP family of proteins. Antibodies against the COOH-terminal peptide of AQP2 did not produce any detectable, specific staining in the inner ear (not shown).

The strongest AQP-CHIP staining was seen in mesen- chymal cells below the basilar membrane, suggesting that the antigen recognized by AQP-CHIP is the most abundant or most tightly bound in these cells. The presence of water channels in these cells suggests that

an ionic gradient may exist across the basilar membrane and that, perhaps, there is a need for a rapid osmotic equilibration across the membrane. Bondy et al. (3) previously suggested that AQP-CHIP may mediate ma- jor changes in hydration accompanying the deposition of the extracellular matrix components. If subbasilar mem- brane cells secrete components of the basilar membrane (10,20>, water channels in those cells could facilitate the deposition of the extracellular matrix and thus partici- pate in maintenance of the basilar membrane.

AQP-CHIP staining was also seen in type III fibro- cytes of the spiral ligament. The precise role of the spiral ligament in cochlear function is not yet fully under- stood. It has long been recognized that the volume of the spiral ligament decreases from basal to apical turns of the cochlea, and the number of fibrocytes within the ligament correspondingly decreases apically. In the pre- sent study, this trend was confirmed for type III fibro- cytes that were immunostained for AQP-CHIP. This finding suggests that the function of AQP-CHIP in type III fibrocytes is closely tied to functions of the spiral ligament. This notion is reinforced by the finding that type III fibrocytes within the spiral ligament form gap junctions with each other and with other fibrocytes within the ligament (12). The gap junction interconnec- tions indicate that the fibrocytes within the spiral ligament share a common intracellular ionic environ- ment. If AQP-CHIP in type III fibrocytes functions as a regulator of cell volume, as it is thought to do for erythrocytes (51, it would appear that the volume being regulated is that of all fibrocytes that are interconnected via gap junctions. Volume regulation may be needed in these cells to protect them from osmotic stress that could occur in response to excessive acoustic stimula- tion, which could induce excessive ionic flux from sen- sory cells or rupture the reticular lamina and cause a mixing of endolymphatic and perilymphatic contents.

Henson et al. (8) suggested that fibroblasts in the marginal region of the spiral ligament in bat and mice have a role in the generation of tension within the basilar membrane-spiral ligament complex. Those ten- sion fibroblasts are thought to be capable of influencing the mechanical properties of the basilar membrane because of their location and attachments, as well as the fact that they contain stress fibers and contractile and contraction-associated proteins (8). Type III fibrocytes of the spiral ligament in the guinea pig are found in the region in which tension fibroblasts were previously described (8, 9). However, the structure of type III fibrocytes in the guinea pig reported here is different from that of tension fibroblasts in the bat (91, suggesting that cells in this location in these two species play different roles in cochlear function.

A large proportion of AQP-CHIP positive cells in the inner ear is in close association with bone, suggesting that AQP-CHIP may have special significance for main- tenance of bone. AQP-CHIP-positive bone-lining fibro- cytes include mesenchymal cells lining the Scala tym- pani and Scala vestibuli, fibrocytes lining the vestibular part of the bony labyrinth, and fibrocytes around the endolymphatic duct. Although fibrocytes of the trabecu-

Cl454 AQUAF’ORIN CHIP IN GUINEA PIG INNER EAR

AQUAPORIN CHIP IN GUINEA PIG INNER EAR

Fig. 4. Immunolocalization of AQP-CHIP in chondrocytes of footplate of stapes. Primary antibody dilution is 1:800. Bar, 15 pm.

lar perilymphatic tissue do not directly appose bone, they are connected by gap junction to cells that do (11). AQP-CHIP immunoreactivity in bone-lining fibrocytes of adult animals is a novel observation. Bondy at al. (3) have previously reported that AQP-CHIP is transiently expressed in rat periosteum during fetal development. In those animals, AQP-CHIP mRNA was abundant in condensations of mesenchyme forming the periosteum around sites of endochondral and intramembranous ossification (31, in which it is thought to be involved in deposition of extracellular matrix components. The pre- sent findings are that endosteal and periosteal staining is present in the adult guinea pig temporal bone, as well as in periosteum of other cranial bones that we have examined. Differences between previously reported and the present observations may be, in part, due to method- ological or species differences.

Cells of the middle ear that were immunostained with anti-AQP-CHIP antibodies were almost exclusively asso- ciated with bone, and they included periosteum of the middle ear and the stapes, as well as chondrocytes of the oval window and chondrocytes of the head and the footplate of the stapes. The functional significance of water channels in these cells is also unclear. However, it is possible that water channels may be involved in the production of middle ear effusion leading to a pathologi- cal buildup of fluid in the middle ear and clinical symptoms of otitis media.

Fig. 5. A: immunostaining of cochlea with anti-AQP-CHIP antibodies. B: staining was completely abolished after preabsorption of anti-AQP- CHIP antibody with human erythrocyte membranes enriched in AQP-CHIP. Primary antibody dilution is 1:800. Bar, 40 pm (A and B).

Although the presence of immunostaining with anti- AQP-CHIP serum in the cochlea does not guarantee that the antigen recognized by the antibody is indeed AQP-CHIP, there are several reasons to believe that the inner ear protein is either AQP-CHIP or a close homo- logue. First, AQP-CHIP is well conserved in different tissues (21, and it is likely that the protein recognized in the inner ear is a close homologue or identical to the form found in other tissues. Second, an identical stain- ing pattern was obtained using affinity-purified polyclo- nal antibodies against BLIP, a basolateral integral mem- brane protein. Both BLIP and AQP-CHIP are membrane

Fig. 3. Immunolocalization of AQP-CHIP in vestibular part of inner ear. A: in a semicircular canal, staining is found in fibrocytes of trabecular perilymphatic tissue (arrows), which is continuous with endosteum of bony labyrinth. *Lumen of semicircular canal. B: in ampulla of a semicircular canal, most intense staining is found in fibrocytes underneath dark cell epithelium. Note absence of staining in epithelial dark cells (arrowheads) and in epithelium of crista ampullaris. C: in utricle, AQP-CHIP is immunolocalized in fibrocytes underneath sensory epithelium (arrows). Connective tissue around endolymphatic sac and distal part of endolymphatic duct (D) is intensely stained for AQP-CHIP. *Lumen of endolymphatic duct; arrowheads point to apical surfaces of nonstained epithelial cells. Primary antibody dilution is 1~300. Bar, 25 pm (A-D).

Cl456 AQUAPORIN CHIP IN GUINEA PIG INNER EAR

proteins related to the major intrinsic protein of mamma- lian lens, MIP26; the antiserum to BLIP used in this study is known to cross-react with AQP-CHIP (24), but not vice versa. Third, two different immunostaining techniques were employed, i.e., fluorescence with fluores- cein-conjugated secondary antibody and peroxidase with biotinylated secondary antibody, and both techniques stained the same cell types. Fourth, sections of the inner ear containing parts of the choroid plexus showed the same staining pattern previously reported in the choroid plexus, i.e., staining of the apical but not the basolateral membrane (3). In addition, residual red blood cells were stained in all immunofluorescent sections. A Western blot was not performed due to near impossibility of eliminating contamination of the inner ear tissue by erythrocytes from the vasculature and bone. Neverthe- less, we believe that immunostaining of the inner ear with affinity-purified anti-AQP-CHIP serum identifies specific populations of cells that contain the constitu- tively active aquaporin, AQP-CHIP.

The physiological importance of AQP-CHIP in the inner ear remains to be determined. AQP-CHIP knock- out mutations have recently been described in humans who exhibit no apparent clinical defects (18). It is possible that these AQP-CHIP knockouts express a different protein, possibly from the MIP-related family, that functionally replaces AQP-CHIP. We believe that immunocytochemical localization of AQP-CHIP in the inner ear will prove helpful in designing experiments to decipher a physiological role of AQP-CHIP.

We thank Dr. Shun-ichi Imamura for donating inner ear sections and John Lydon for excellent technical help. We are grateful to Dr. Nelson Y. S. Kiang, Dr. Edmund Mroz, and Dr. William Sewell for constructive criticism of this manuscript.

This work was supported by National Institute on Deafness and Other Communication Disorders Grant T32-DC-00038 and National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-38452.

Address for reprint requests: K. M. Stankovic, Eaton-Peabody Laboratory, Massachusetts Eye and Ear Infirmary, 243 Charles St., Boston, MA 02 114.

Received 20 September 1994; accepted in final form 26 May 1995.

REFERENCES

8.

9.

10.

11.

12.

13.

15.

16.

17.

18.

19.

20.

21.

22.

Adams, J. C. Biotin amplification of biotin and horseradish peroxidase signals in histochemical stains. J. Histochem. Cyto-

23 ’

them. 40: 1457-1463,1992. Agre, P., G. M. Preston, B. L. Smith, J. S. Jung, S. Raina, C. Moon, W. B. Guggion, and S. Nielsen. Aquaporin CHIP: the

24 *

archetypal molecular water channel. Am. J. Physiol. 265 (Renal Fluid Electrolyte Physiol. 34): F463-F476, 1993. Bondy, C., E. Chin, B. L. Smith, G. M. Preston, and P. Agre. Developmental gene expression and tissue distribution of the CHIP28 water-channel protein. Proc. NatZ. Acad. Sci. USA 90: 25. 4500-4504,1993. Brown, D., J.-M. Verbavatz, G. Valenti, B. Lui, and I. Sabolic. Localization of the CHIP28 water channel in reabsorp- 26. tive segments of the rat male reproductive tract. Eur. J. CeZZ Biol. 61: 264-273,1993. Denker, B. M., B. L. Smith, F. P. Kuhajda, and P. Agre. Identification, purification and characterization of a novel Mr 28,000 integral membrane protein from erythrocytes and renal 27. tubules. J. BioZ. Chem. 263: 15634-15642,1988. Gilman, A. G., T. W. Rall, A. S. Nies, and P. Taylor (Editors). The Pharmacological Basis of Therapeutics (8th ed.). 28. New York, NY: Pergamon, 1990. Hasegawa, H., R. Zhang, A. Dohrman, and A. S. Verkman. Tissue-specific expression of mRNA encoding rat kidney water

channel CHIP28k by in situ hybridization. Am. J. Physiol. 264 (Cell Physiol. 33): C237-C245, 1993. Henson, M. M., K. Burridge, D. Fitzpatrick, D. B. Jenkins, H. C. Pillsbury, and 0. W. Henson, Jr. Immunocytochemical localization of contractile and contraction associated proteins in the spiral ligament of the cochlea. Hear. Res. 20: 207-214, 1985. Henson, M. M., and 0. W. Henson, Jr. Tension fibroblasts and the connective tissue matrix of the spiral ligament. Hear. Res. 35: 237-258,1988. Keithley, E. M., A. F. Ryan, and N. K. Woolf. Fibronectin-like immunoreactivity of the basilar membrane of young and aged rats. J. Comp. NeuroZ. 327: 612-617, 1993. Kikuchi, T., J. C. Adams., D. L. Paul, and R. S. Kimura. Gap junction systems in the rat vestibular labyrinth: immunohisto- chemical and ultrastructural analysis. Acta Oto-Zaryngol. 114: 520-528,1994. Kikuchi, T., R. S. Kimura, D. L. Paul, and J. C. Adams. Gap junction systems in the rat cochlea: immunohistochemical and ultrastructural analysis. Anat. Embryol. 191: 101-119, 1995. Kronenberg, J., and G. Leventon. Histology of the endolym- phatic sac of the rat ear and its relationship to surrounding blood vessels: the “endolymphatic glomerulus.” Am. J. OtoZogy. 7: 130-133,1986. Nielsen, S., B. L. Smith, E. I. Christensen, and P. Agre. Distribution of the aquaporin CHIP in secretory and resorptive epithelia and capillary endothelia. Proc. NatZ. Acad. Sci. USA 90: 7275-7279,1993. Nielsen, S., B. L. Smith, E. I. Christensen, M. A. Knepper, and P. Agre. CHIP28 water channels are localized in constitu- tively water-permeable segments of the nephron. J. CeZZ BioZ. 120: 371-383,1993. Preston, G. M., T. P. Carroll, W. B. Guggino, and P. Agre. Appearance of water channels in Xenopus oocytes expressing red cell CHIP28 protein. Science Wash. DC 256: 385-387, 1992. Preston, G. M., B. L. Smith, M. L. Zeidel, J. J. Moulds, and P. Agre. Mutations in aquaporin-1 in phenotypically normal humans without functional CHIP water channels. Science Wash. DC 265: 1585-1587,1994. Sabolic, I., G. Valenti, J.-M. Verbavatz, A. N. Van Hoek, A. S. Verkman, D. A. Ausiello, and D. Brown. Localization of the CHIP28 water channels in rat kidney. Am. J. Physiol. 263 (CeZZ PhysioZ. 32): C1225-C1233, 1992. Santi, P. A., J. T. Larson, L. T. Funcht, and T. S. Econo- mou. Immunohistochemical localization of fibronectin in the chinchilla cochlea. Hear. Res. 39: 91-102, 1989. Smith, B. L., and P. Agre. Erythrocyte Mr 28,000 transmem- brane protein exists as a multisubunit oligomer similar to channel proteins. J. BioZ. Chem. 266: 6407-6415, 1991. Spicer, S. S., and B. A. Schulte. Differentiation of inner ear fibrocytes according to their ion transport related activity. Hear. Res. 56: 53-64, 1991. Spicer, S. S., B. A. Schulte, and J. C. Adams. Immunolocaliza- tion of Na+,K+-ATPase and carbonic anhydrase in the gerbil’s vestibular system. Hear. Res. 43: 205-218, 1990. Valenti, G., J.-M. Verbavatz, I. Sabolic, D. A. Ausiello, A. S. Verkman, and D. Brown. A basolateral CHIP28/MIP26- related protein (BLIP) in kidney principal cells and gastric parietal cells. Am. J. Physiol. 267 (Cell Physiol. 36): C812-C820, 1994. Van Hoek, A. N., and A. S. Verkman. Functional reconstitu- tion of the isolated erythrocyte water channel CHIP28. J. BioZ. Chem. 267: 18267-18269,1992. Verbavatz, J. M., D. Brown, I. Sabolic, G. Valenti, D. A. Ausiello, L. A. Van Hoek, T. Ma, and A. S. Verkman. Tetrameric assembly of CHIP28 water channel in lyposomes and cell membranes: a freeze fracture study. J. CeZZ BioZ. 123: 605-618,1993. Walker, E. M., M. A. Fazekas-May, and W. R. Bowen. Nephrotoxic and ototoxic agents. CZin. Lab. Med. 10: 323-354, 1990. Zeidel, M. L., S. V. Ambudkar, B. L. Smith, and, P. Agre. Reconstitution of functional water channels in liposomes contain- ing purified red cell CHIP28 protein. Biochemistry. 31: 7436- 7440,1992.