The level of endolymphatic [Ca
2+] is controlled by both secretory mechanisms (e.g., PMCA2 in the apical surface of hair cells [
9]) and absorptive mechanisms. The present paper reports the first evidence for the expression of all genes necessary to constitute a complete transepithelial Ca
2+ absorptive pathway in native tissues of the inner ear. We previously reported that cochlear outer sulcus cells and vestibular transitional cells have apical nonselective cation channels that are likely permeable to Ca
2+ [
10,
11], but their contribution to transepithelial Ca
2+ absorption is not known.
Expression
The transmural absorption of Ca
2+ by a number of epithelia has recently been ascribed to a set of genes that encode the apical entry channels TRPV5 and TRPV6, the cytosolic Ca
2+ buffering proteins calbindin-D9K and calbindin-D28K, and the basolateral Ca
2+-extruding transporters sodium-calcium exchangers (NCX) and plasma membrane calcium ATPases (PMCA) [
3]. In other epithelial, either TRPV5 or TRPV6 is expressed at a higher level, but in spite of the large array of epithelia surveyed previously [
3], expression of this transport system in the inner ear was not reported. We demonstrated here that all components of the epithelial Ca
2+ channel transport system were found to be expressed as transcripts in the vestibular SCCD and cochlea.
Our results validate the primary culture of SCCD in terms of the presence of all the genes that participate in Ca
2+ absorption and regulation, although the relative expression level of TRPV5 and TRPV6 was observed to be substantially reduced in culture compared to the native cells. This apparent discrepancy is consistent with the lack of 1,25-(OH)
2vitamin D
3 in the control culture medium (see below), whereas the native tissues were exposed to intermediate levels of the hormone
in vivo prior to sacrifice. TRPV5 and TRPV6 form homo- and hetero-tetrameric channel complexes and regulation of the relative expression levels of TRPV5 and TRPV6 may be a mechanism for fine-tuning Ca
2+ transport kinetics in TRPV5/6-expressing tissues [
12].
Quantification of expression levels has intrinsic problems. First, there can be large discrepancies between levels of transcripts and levels of the corresponding proteins. Further, the proteins can receive various post-translational modifications and trafficked to different destinations with a consequent multiplicity and altered level of function. Second, seemingly anomalous results due to imprecision of Ct determination in RT-PCR and its logarithmic relation to transcript number can occur, leading to calculated large fold-changes in expression that do not reach statistical significance (e.g., calbindin-D9K in stria vascularis; vide infra).
TRPV5 and TRPV6 gene expression in the inner ear was also demonstrated at the level of protein. Sufficient protein could be obtained from primary cultures of SCCD to allow immunoblot observation of TRPV5. Interestingly, immunoblot data did not show a significant difference in TRPV5 expression between control and 1,25-(OH)
2vitamin D
3 treated SCCD cells. This result is not consistent with the functional data from the same preparation that show an upregulation by 1,25-(OH)
2vitamin D
3 of radiolabeled Ca
2+ fluxes in primary cultures of SCCD [
7]. The apparent discrepancy could be accounted for by 1) the increased transcript expression not affecting protein amount and 2) an upregulation by 1,25-(OH)
2vitamin D
3 of another controlling part of the transport system. The ineffectiveness of PNGase F treatment for TRPV5 suggests the occurrence of other posttranslational modifications than glycosylation or a low abundance of glycosylated protein.
Immunolocalization of both epithelial Ca
2+ channels was consistent with the transcript data for both the vestibular system and for the cochlea. In addition to the cochlear localization of the strial marginal cells and the outer sulcus of the lateral wall, these channels were also observed in the inner sulcus epithelial cells of the rat inner ear. Recent observations in mouse show similar cellular locations, although the preponderance of TRPV5 or TRPV6 in some cells differs [
6]. Our data demonstrate that the earlier findings were not strictly limited to one species. The antibodies used do not restrict their binding to membrane proteins, so that there is staining of protein in the cytosol, preventing the subcellular localization of TRPV5 and TRPV6 to the apical and/or basolateral membranes. In addition, the antibody against TRPV6 did not function in immunoblots.
The other genes required for Ca
2+ absorption--calbindin-D9K, calbindin-D28K, NCX1 and PMCA1b--were previously found to be involved in Ca
2+ absorption by the kidney [
13,
14]. Calbindin-D9K can prevent calcium-dependent inactivation of TRPV5/6 by buffering overloaded Ca
2+ beneath the apical side while calbindin-D28K is freely diffusible in the cytoplasm [
4]. We evaluated all isoforms of NCX1-3 and PMCA1-4 in the inner ear tissues.
PMCA was demonstrated by immunohistochemistry to be present in the basolateral membrane of strial marginal cells and Reissner's membrane epithelial cells [
15]. NCX1 and PMCA1 were the predominant isoforms expressed in the inner ear, as in the kidney. However, other isoforms were also found to be present. PMCA2 was not expressed in either native or cultured SCCD and virtually absent in stria vascularis. Absence of PMCA2 in stria vascularis is compatible with previous reports [
9,
16]. Expression of PMCA2 in the lateral wall fraction might arise from contributions by Reissner's membrane [
16]. PMCA2 is therefore not likely important for inner ear Ca
2+ absorption; rather its function is likely limited to Ca
2+ secretion at the apical membrane of hair cells [
9,
17] and possibly Reissner's membrane [
16].
Our detection of PMCA3 and PMCA4 in all tissue fractions is at variance with previous reports for the cochlea [
9,
16]. It is possible to attribute these discrepancies to species differences. A previous immunohistochemical study showed calbindin-D28k expression in musk shrew spiral ganglion, inner and outer hair cells, but expression in lateral wall was only observed during pre-natal ages [
18]. By contrast, we observed expression of calbindin-D28k in lateral wall fractions of young adult rats.
Regulation
A wide range of signal pathways converge to control Ca
2+ absorption. An important control mechanism is regulation of the number of expressed apical entry Ca
2+ channels and calbindins by 1,25-(OH)
2vitamin D
3. It is apparently characteristic of different epithelia that 1,25-(OH)
2vitamin D
3 up-regulates predominantly one channel isoform through activation of the vitamin D receptor [
3]. Caco-2 cells up-regulate TRPV6 in a time- and dose-dependent manner [
19]. By contrast, 1,25-(OH)
2vitamin D
3 up-regulates TRPV5, but not TRPV6, in kidney [
3] and SCCD (
vide infra).
Two isoforms of calbindin, calbindin-D9K and -D28K, are expressed in the inner ear epithelia and calbindin-D9K transcript expression is up-regulated by 1,25-(OH)
2vitamin D
3. It has been shown that calbindin-D9K is localized near both the apical and basolateral plasma membranes [
20], while calbindin-D28K is freely diffusible in the cytoplasm [
20] but translocates to the apical membrane to associate with the TRPV5 [
21]. One function of calbindin-D28K is to prevent Ca
2+ from binding to calmodulin, which can block TRPV6 by binding to the COOH-terminal region [
22]. Interaction with either channel is important since the functional channel is likely a tetraheteromer of TRPV5/TRPV6. Both calbindins, NCX1 and PMCA1b genes were up-regulated in hormone-deficient 1α-OHase
-/- mice by 1,25-(OH)
2vitamin D
3 [
20].
NCX and PMCA isoform expression levels were not uniformly changed by 1,25-(OH)2vitamin D3 in the SCCD, lateral wall and stria vascularis. NCX2 in the lateral wall may be important for the Ca2+ absorption pathway because of its up-regulation by 1,25-(OH)2vitamin D3. But the reason why PMCA4 in SCCD and PMCA3 in lateral wall were down-regulated is not clear. The lack of response of all genes in the stria vascularis may indicate that this transport system is not functional in stria. Alternatively, it may represent a vitamin D-insensitive transport system in this tissue.
The activity of the TRPV5 and TRPV6 channels is steeply controlled by both intra- and extracellular pH [
23]. The strong inhibition by acid extracellular pH is of particular interest in view of pathologic conditions leading to lowered pH in both cochlear and vestibular endolymph (see below). Epithelial Ca
2+ channel activity can further be controlled by PIP2, providing a link to a variety of G protein-coupled receptors [
24].
Physiological significance
The question naturally arises whether this transport system is relevant at the whole organ level. Several observations demonstrate the functionality of this system and point to its importance in hearing and balance. It was recently shown that SCCD epithelial cells take up radiolabeled Ca
2+ more rapidly from the apical side than from the basolateral side, that the net uptake is increased in the presence of 1,25-(OH)
2vitamin D
3 and that the apical uptake is inhibited by acidic luminal pH (a hallmark of TRPV5 and TRPV6) [
7]. It was therefore predicted that any condition that results in an acidification of endolymph would also lead to increased luminal [Ca
2+] and a consequent decrease in cochlear and vestibular function, which depend on the normally low endolymphatic [Ca
2+] [
2]. Indeed, mutations or deletion of the bicarbonate-secreting transporter pendrin (expressed in the luminal membranes of cochlear and vestibular epithelial cells) lead to hearing and balance deficits in humans [
25] and mice [
26]. Deletion of pendrin led to acidified endolymph and consequently a dramatic elevation of endolymphatic [Ca
2+] in both the cochlea [
6] and the utricle [
7]. This observation suggests that the TRPV5/6 Ca
2+ absorption system plays a highly significant physiological role in endolymph Ca
2+ homeostasis, even though the endocochlear potential apparently plays a strong role by providing a driving force for passive efflux [
27].
Calcium homeostasis in the inner ear via TRPV5/6 could play an important role in causes of benign paroxysmal positional vertigo (BPPV) and may therefore be an effective drug target. BPPV is characterized by brief episodes of nystagmus and vertigo in response to certain movements of the head. It is thought to be caused by dislodged otoliths (composed primarily of CaCO
3 crystals) from the utricle that enter one of the semicircular canals, leading to inappropriate stimulation of canal. A recent study reported that strong correlations in human patients were observed between diagnosis of BPPV and disturbed calcium homeostasis as reflected in reduced bone mineral density [
28].
Vitamin D has been implicated in hearing function. Ikeda and associates found that vitamin D deficiency resulted in hearing impairment in rats [
29] and that 80% of patients in a study with bilateral sensory neural hearing loss (BSNHL) were found to be deficient in 1,25-(OH)
2vitamin D
3 [
30,
31]. Importantly, the patients with BSNHL and low serum vitamin D had normal serum Ca
2+, consistent with a local effect of vitamin D deficiency in the auditory and vestibular periphery. The vitamin D-deficient rats had a reduced perilymphatic [Ca
2+] level, making the interpretation less clear. It was not known whether the observed effects were due to direct effects on the epithelial calcium channel system (not known at that time), to the lowered systemic [Ca
2+], or to other causes. Nonetheless, the correlations are consistent with a direct action on the system reported here.
The dysfunction of Ca
2+ absorption by mutation or absence of the TRPV5 and TRPV6 genes would be expected to lead to impaired hearing and balance. It is perhaps more than a coincidence, therefore, that the genes for TRPV5 and TRPV6 are located on chromosome 7q35 and 7q33-34 respectively in human [
4] and the locus of the non-syndromic deafness gene [
32] DFNB13 is located at the encompassing region on chromosome 7q34-36 [
33].