Mechanical Control of Regeneration in the Inner Ear

Hundreds of millions of people around the world suffer from disabling hearing loss or have vestibular deficits. A major and irreversible cause of these impairments is the loss of sensory hair cells in the neuroepithelium of the inner ear, which can occur due to loud sounds, ototoxic compounds, infections, and aging. Birds, fish, and amphibians mount a regenerative response to such damage whereby neighboring supporting cells divide and differentiate into replacement hair cells, ultimately leading to recovery of hearing and balance function. In humans and other mammals, these repair processes occur to a limited extent during development, but in adults supporting cells remain largely quiescent. I hypothesized that the regenerative replacement of hair cells in the mammalian inner ear was limited by the development of reinforced apical junctions in mammalian supporting cells, which does not occur in the supporting cells of nonmammals.

The first chapter of this dissertation introduces the inner ear, the public health impact of sensory hair cell losses, and the field of hair cell regeneration. It also briefly introduces the findings that led to for the original research herein.

The second chapter is a comprehensive literature review focusing on cell junctions and mechanical aspects of regeneration in the inner ear. A primary focus is observations that supporting cells in mammals develop uniquely reinforced, E-cadherin-rich apical junctions during postnatal maturation. These junctions are bracketed by thick circumferential bands of filamentous actin that are exceptionally thick and stable. The growth of these F-actin bands correlates to postnatal declines in the rates of spreading and proliferation of supporting cells. In mammals, these reinforced junctions are found both in pillar and Deiters cells within the cochlea, as well as in supporting cells throughout vestibular organs such as the utricle. Supporting cells of birds, fish, and amphibians hardly express E-cadherin and possess thin bands of filamentous actin. Major open questions include the following: (1) Does the reinforcement of junctions in mammalian supporting cells stiffen the epithelium relative to those in nonmammals? (2) Does reversing this junctional reinforcement overcome proliferative quiescence in the mature mammalian utricle? And finally, (3) What signaling pathways in mammalian supporting cells act downstream of junctional reinforcement to maintain their quiescence? These questions are explored in the subsequent chapters.

In the third chapter, I test the hypothesis that the postnatal reinforcement of junctions in mammalian supporting cells mechanically stabilizes the epithelium. For this, I performed mechanical measurements of the utricular epithelia of mice and chickens at various ages. Using a customized micropipette aspiration technique that allows for the visualization of epithelial deformations due to a controlled suction force, I found that the sensory epithelium of the mouse stiffens during postnatal maturation. Pharmacologic disruption of filamentous F-actin made the utricular epithelia of the mice less stiff. In contrast, there was no difference in stiffness between utricles of hatchling chicks and older pullets. I also analyze supporting cell shape as an indirect measure of intraepithelial tension. Supporting cells in the mouse had round apical domains, while supporting cells in the chicken utricle developed progressively more elongated shapes throughout embryonic development, suggesting that they are readily deformable in response to intraepithelial tension. Taken together, the results indicate that the postnatal reinforcement of supporting cell junctions in the mammalian utricle causes an increase in stiffness that restricts responses to cell loss.

The fourth chapter describes the effects of reversing junctional reinforcement in the mouse utricle. An incidental discovery revealed that a combination of growth factors and pharmacologic agents reversed junctional reinforcement in supporting cells of cultured mouse utricles. Detailed time course experiments revealed that the cocktail first thinned the F-actin bands in supporting cells throughout the utricle. Then, levels of E-cadherin declined in supporting cells within the striolar region, located centrally within the utricle. After these changes, the rate of proliferation sharply increased in striolar supporting cells. Epidermal growth factor and a small molecule inhibitor of glycogen synthase kinase 3 were the two components of the cocktail that mediated these effects. This represents the first evidence that the reversal of junctional reinforcement correlates with cell cycle re-entry and proliferation in the long-quiescent supporting cells in a mature mammalian utricle.

It remained unclear what signaling pathways operate downstream of junctional reinforcement to maintain mammalian supporting cells in a state of persistent quiescence. One intriguing candidate was the Hippo pathway and its effector Yes-associated protein (YAP), which respond to changes in mechanical tension at cell junctions and partner with TEAD transcription factors to control epithelial proliferation. In the fifth chapter, I tested the hypothesis that HC loss would activate YAP in the ears of nonmammals, but would not in the ears of mammals which harbor reinforced apical junctions. I found that YAP accumulated in nuclei of supporting cells in the chicken utricle after hair cell loss and mediated regenerative proliferation. In mouse utricles, the YAP remained cytoplasmic after hair cell loss, and little proliferation occurred. I sought to bypass this inhibition using a genetically engineered mouse model that allowed for the expression of an activated variant of YAP. This activated YAP-TEAD transcriptional activity and drove proliferation and mitotic hair cell production in utricles of living mice. Further experiments indicated suggested that inhibitory phosphorylation limited the proliferative effect of YAP in the mouse utricle. Conditional deletion of the Large tumor suppressor (LATS) kinases abolished inhibitory phosphorylation of YAP and led to proliferation of striolar supporting cells, even in adults.

To conclude, I formulate a working model to explain why supporting cells in nonmammals mount a robust proliferative response to hair cell loss, whereas those in mammals remain in a state of proliferative quiescence: Hair cell loss the ears of birds elicits a change in local intraepithelial tension that rapidly propagate through the pliable epithelium and results in physical expansion of neighboring supporting cells. This mechanical signal bypasses or inactivates the Hippo pathway in the supporting cells, leading to the nuclear accumulation of YAP and TEAD-dependent regenerative proliferation. In mammals, thickened circumferential bands of filamentous actin produce a rigid epithelium that impedes the transmission of intraepithelial tension and slows the shape changes that arise upon hair cell loss. The absence of substantial changes in tension and high levels of E-cadherin allow LATS kinases to remain activated, which phosphorylate and inhibit YAP to maintain supporting cells in a state of persistent quiescence. I describe outstanding questions and additional experiments to further interrogate this model. Overall, the work presented here suggests that genetic or pharmacologic targeting of the E-cadherin/LATS/YAP/TEAD axis could represent a therapeutic target for acquired hearing loss and balance disorders by promoting regenerative proliferation in sensory epithelia of the inner ear.