Modulation of Wnt Signaling Enhances Inner Ear Organoid Development in 3D Culture

Abstract

Stem cell-derived inner ear sensory epithelia are a promising source of tissues for treating patients with hearing loss and dizziness. We recently demonstrated how to generate inner ear sensory epithelia, designated as inner ear organoids, from mouse embryonic stem cells (ESCs) in a self-organizing 3D culture. Here we improve the efficiency of this culture system by elucidating how Wnt signaling activity can drive the induction of otic tissue. We found that a carefully timed treatment with the potent Wnt agonist CHIR99021 promotes induction of otic vesicles-a process that was previously self-organized by unknown mechanisms. The resulting otic-like vesicles have a larger lumen size and contain a greater number of Pax8/Pax2-positive otic progenitor cells than organoids derived without the Wnt agonist. Additionally, these otic-like vesicles give rise to large inner ear organoids with hair cells whose morphological, biochemical and functional properties are indistinguishable from those of vestibular hair cells in the postnatal mouse inner ear. We conclude that Wnt signaling plays a similar role during inner ear organoid formation as it does during inner ear development in the embryo.

Fig 1. Spatio-temporal changes in expression of pluripotency markers and self-organization of stem cell-derived aggregates in 3D culture.

(A-D) Expression of the pluripotency marker Nanog decreases as vesicles mature between differentiation days 5–8. (A’-D’) Progression of outer epithelium ruffling indicating self-organization. (E-H) Oct4-GFP expressing pluripotent stem cells decrease in number and are confined to the aggregate core by differentiation day 8. Scale bars, 100 μm (D, D’, H).

Fig 2. CHIR treatment exerts time- and dose-dependent effects on otic vesicle derivation in 3D culture.

(A-D’) Pax2⁺/Ecad⁺ otic vesicles in R1 ESC-derived aggregates treated with CHIR for 48hrs starting at day 8 (B, B’), 9 (C, C’) or 10 (D, D’) along with untreated controls (A, A’). (E) Comparison of the number and the area size of Pax2⁺/Ecad⁺ vesicles per aggregate among experimental groups treated with CHIR at different starting times (*P<0.05, **P<0.001; mean ± s.e.m.; n = 42–80 per group). (F) Comparison of the area size of Pax2⁺/Ecad⁺ vesicles per aggregate among experimental groups treated with CHIR at different concentrations. (*P<0.05, **P<0.001; mean ± s.e.m.; n = 52–66 per group). Scale bars, 100 μm (D), 50 μm (D’).

Fig 3. CHIR treatment increases the prevalence and size of Atoh1/nGFP–positive inner ear organoids in 3D culture.

(A-D’) Atoh1/nGFP–positive sensory patches in a protruding vesicle become larger with time in aggregates treated with CHIR. (E-H) There is no noticeable temporal change in the size of Atoh1/nGFP–positive sensory patches in untreated control aggregates. (I) Comparison of the percentage of aggregates containing at least one Atoh1/nGFP–positive vesicles between CHIR-treated aggregates and untreated controls (*P < 0.05; n = 14 per group). (J) The average number of Atoh1/nGFP–positive organoids per aggregate in CHIR-treated aggregates and untreated controls (*P < 0.05; n = 14 per group). (K-L) Immunoflorescence showing Atoh1/nGFP-positive cells express Sox2 and Myo7A. Scale bars, 100 μm (D), 10 μm (D’, H).

Fig 4. CHIR treatment has dose-dependent effects on the number of vesicles containing Myo7a⁺/Sox2⁺ cells.

(A-D’) Representative images show Myo7a/Sox2 expression in day 21 aggregate sensory epithelia that received 0 μM (A, A’), 1 μM (B, B’), 3 μM (C, C’), and 10 μM (D, D’) CHIR between days 8 and 10. (E) The average number of vesicles containing Myo7a⁺/Sox2⁺ cells per aggregate as a function of the CHIR concentration (**P < 0.001; mean ± s.e.m.; n = 55–68 per group). (F) The largest number of Myo7a⁺/Sox2⁺ cells per section as a function of the CHIR concentration (**P < 0.001; mean ± s.e.m.; n = 10 per group). Scale bars, 100 μm (D), 10 μm (D’, H).

Fig 5. CHIR-treated aggregates give rise to inner ear organoids harboring mechanosensitive hair cells.

(A-B) Co-localization of two hair cell markers Calb2 and Myo7a (A) or Sox2 and Myo7a (B) in cells lining the luminal surface of a vesicle. (C) Atoh1/nGFP⁺ cells also express Myo7a and Sox2. (D-F) Cells expressing Brn3C, Myo7a, and Atoh1/nGFP exhibit flask-like morphology with a hair bundle on their apical surface characteristic of vestibular hair cells. Some Myo7a⁺ cells in day 28 samples have faint Atoh1/nGFP expression (arrows), suggesting that these cells are more mature hair cells than cells expressing strong Atoh1/nGFP expression. (G) The hair bundle marker Espin was observed on the apical surface of Calb2⁺ hair cells. (H) Representative voltage-gated currents and mechanosensitive currents recorded from day 25 Atoh1/nGFP⁺ cells in response to voltage injections and hair bundle deflections, respectively. (I) A cluster of Brn3A⁺ neuronal cell bodies were located near Myo7a⁺ hair cells. (J) A TUJ1⁺ neural processes extend and contact Myo7a⁺ hair cells. (K) The ribbon synapse marker CtBP2 was associated with Myo7a⁺ cells and TUJ1⁺ processes. Scale bars, 10 μm (A-G, I-K).

Fig 6. Schematic summary of the proposed role of Wnt signaling in inner ear organoid formation in 3D culture.

Augmentation of canonical Wnt signaling by CHIR99021, a potent Wnt agonist, upon preplacodal formation in stem cell-derived aggregates promotes otic vesicle formation at the expense of other preplacodal derivatives, resulting in the formation of a larger number of hair cells. ne, neural ectoderm. ppe, pre-placodal epithelium. OtP, otic placode. mes, mesoderm. esc, embryonic stem cell. de, definitive ectoderm. nne, non-neural ectoderm.