We are born with about 15,000 sensory hair cells in each cochlea. Since these cells cannot regenerate, our hearing relies on the ability of hair cells to survive for decades. Their mechanosensory organelles—the stereocilia bundles—are formed only once, during embryonic development in humans or postnatal development in rodents, and need to be maintained throughout an organism's life. In mature stereocilia, the proteins that make up their cytoskeleton seem to renew at an extremely slow rate (Nature. 2012;481:520; Nat Commun. 2015;6:6855). So what are the molecular mechanisms that maintain these delicate organelles over a lifetime?
MECHANOTRANSDUCTION AND STEREOCILIA STABILITY
In each auditory hair cell, stereocilia are arranged in rows of increasing height to form a precisely shaped hair bundle (Fig.1, left). The stereocilium core is made up of a paracrystal array of parallel actin filaments (J Cell Biol. 1980;86:244). The tips of the shorter-row stereocilia connect to the sides of taller-row stereocilia via fine extracellular connections or “tip links.” Sound-induced deflections of stereocilia bundles increase the tension in the tip links, which in turn gates mechanotransduction channels on stereocilia tips. These channels allow the influx of positive ions, including calcium, into the cell (Nature. 1979;281:675; Proc Biol Sci. 1992;249:185). At resting conditions, the tip links are maintained under a certain level of mechanical tension to guarantee the detection of the smallest bundle deflections—and thus, the detection of very soft sounds. As a result, a small proportion of mechanotransduction channels are open at rest, allowing a constant entry of calcium ions. In a recent study, we discovered that the stability of the actin cytoskeleton in mammalian auditory stereocilia depends on this constant calcium influx (eLife. 2017;10.7554/eLife.24661).
We used scanning electron microscopy to study the effect of pharmacological blockers of the mechanotransduction channels on the morphology of hair cell stereocilia. Blockage of these channels led to a marked decrease in the height of stereocilia, but only in the middle and shorter rows of the hair bundle (Fig. 1, right). The tallest row stereocilia, which do not harbor mechanotransduction channels, remained unaffected (Fig.1, right; Nat Neurosci. 2009;12:553). We also used transmission electron microscopy to evaluate the core of stereocilia during their shortening and observed evidence of remodeling of the actin cytoskeleton at the stereocilia tips but not in the taper region or the rootlets (Fig. 2). In fact, the stereocilia tips are the only sites of the stereocilia cytoskeleton known to undergo constant protein renewal (Nature. 2012; Nat Commun. 2015). We also observed stereocilia shortening after the chemical disruption of tip links, which causes the closure of mechanotransduction channels (eLife. 2017). Our findings were replicated by sequestering calcium ions inside the cell, indicating that stereocilia retraction is likely to be controlled by the calcium influx through mechanotransduction channels (eLife. 2017).
Under experimental conditions, we observed the shortening—and eventual disappearance—of stereocilia after blocking the mechanotransduction channels for several hours or days (eLife. 2017 and unpublished observations). We believe that a similar phenomenon could take place after noise exposure (due to the mechanical breakage of tip links) or in certain types of congenital deafness where mechanotransduction is absent. In fact, abnormalities limited to the transducing stereocilia in the hair cell bundles are a common finding in most mouse models of congenital hearing loss due to defects in some elements of the hair cell mechanotransduction machinery (J Clin Invest. 2011;121:4796; Cell. 2012;151:1283; Neuron. 2014;84:954; Nat Commun. 2017;8:43).
Interestingly, the lack of mechanotransduction hinders the survival of auditory stereocilia, but does not impair their initial development. Hence, it is likely that the stereocilia cytoskeleton becomes susceptible to changes in mechanotransduction levels only after its developmental program is largely complete.
STEREOCILIA REMODELING AND HEARING SENSITIVITY
Short stereocilia could regrow after the reinstatement of mechanotransduction due to blocker washout or the regeneration of tip links, indicating that the stereocilia actin core can dynamically change its shape in response to variations in mechanotransduction levels (eLife. 2017). Since these morphological changes are prominent at the tips of stereocilia—the place where mechanotransduction occurs—it is tempting to speculate that they could have an impact on mechanosensitivity. As previously mentioned, a resting tension at the level of the tip links is required to maintain our hearing sensitivity. In the absence of this resting tip link tension, auditory hair cells could not detect soft sounds. The lack of tension at the tip links would cause mechanotransduction channels to remain closed at rest and therefore inhibit the constant entry of calcium ions (Fig. 3, left). Based on our findings, this decrease in intracellular calcium would cause stereocilia to shorten. We predict that this shortening would stop once the tip links are re-tensioned and the entry of calcium ions is restored (Fig. 3, right). This type of mechanotransduction-dependent regulation of stereocilia bundle mechanics would constitute a novel mechanism of hearing sensitivity regulation at the level of individual stereocilium.
We have yet to discover the exact molecular machinery that senses changes in intracellular calcium levels and controls the remodeling of the stereocilia actin cytoskeleton. Identification of such molecular players would not only provide therapeutic targets to prevent stereocilia degeneration, but might also uncover novel genes that confer susceptibility to age-related or noise-induced hearing loss.
This work was supported by the National Institute on Deafness and Other Communication Disorders (R01DC014658 to G.I.F.) and the American Hearing Research Foundation (2017 grant to A.C.V.).