Activating Notch signaling promoted the proliferation of SCs in the adult mouse utricle in vitro
Previous investigations demonstrated that the overexpression of Notch signaling in the embryonic and neonatal stages is able to promote the formation of extra ectopic HCs in the inner ear in an age-dependent manner15. In the adult stage, the cochlear SCs can only proliferate after co-activating Notch with c-Myc, a powerful cell cycle regulator10. The capacity for HC regeneration was also decreased in the adult utricle9, but limited direct transdifferentiation of HCs was still seen after HC depletion. Thus, we asked if Notch activation alone could promote the proliferation of SCs in adult utricles. Without intervention, the cultured utricular maculae were maintained in a healthy condition and very little loss of sensory cells was seen (Fig. 1A). To activate Notch, we applied similar methods as previously described 10. We created an rtTA/tet-NICD inducible model by crossing RosartTA mice with tetracycline-response-element-controlled NICD mice (tet-NICD). Doxycycline (Dox) application tightly controlled the expression of NICD during the explant culture in a reversible manner. We cultured transgenic utricles with Dox for 3 days before harvesting (Fig. 1B), and EdU was added during the culture to label the mitotic cells in S phase. Our results showed that the number of HCs remained similar to the control groups (Fig. 1B-C). However, compared to the sparse proliferation seen in control cells, many more EdU-positive cells were seen in the SC layer (Fig. 1C-D). The difference was especially obvious in the calbindin-positive striolar region, and in the NICD group around 45.33 Sox2+/EdU + cells per 100 µm2 were observed (N = 3) compared to 0.5 Sox2+/EdU + cells per 100 µm2 in the control group (N = 5) (Fig. 1C-G). In contrast, the numbers were 1.3 and 0 Sox2+/EdU + cells per 100 µm2 for the NICD and control groups in the extra-striolar region, respectively. These results suggested that Notch activation promoted SC proliferation in the adult mouse utricle, especially in the striolar region.
After Damage, Activation Of Notch Promoted Intensive Proliferation In Cultured Adult Utricles
Because robust proliferation was only seen in the striolar region, which was in agreement with previous papers showing that the maculae remain in a relatively stable condition in the absence of damage, we further investigated the effects of Notch activation after HC loss. Adult utricular maculae were cultured with 3 mM neomycin for one day to cause HC damage and then cultured for another 3 days with or without Dox (Fig. 2B). We observed that neomycin killed many of the sensory cells, and few autonomous proliferated cells were seen in the sensory epithelia (Fig. 2A and 2C). The density of HCs decreased from 93.39 to 11.75 HCs per 100 µm2 in the striolar region and from 112.61 to 26.17 HCs per 100µm2 in the extrastriolar region after neomycin treatment. In the Dox-treated group, more proliferated SCs, which were double-labeled with EdU and Sox2, were identified across the whole maculae (Fig. 2C-G). In the extrastriolar region, the density of proliferated SCs was 20.25 SCs per 100 µm2 in the neomycin group, and this increased to 88.68 SCs per 100 µm2 in the neomycin + NICD group. In contrast, in the striolar region the density was 80.25 SCs per 100 µm2 in the neomycin group and 96 SCs per 100 µm2 in the neomycin + NICD group (Fig. 2C). The proliferation of SCs was further confirmed by another marker, Ki67, which was expressed across the whole utricular maculae (Fig. 2J-K). Another HC marker, Parvalbumin, was stained along with Myosin7a, which confirmed the identity of HCs after neomycin and NICD treatments (Fig. 2H-I). In addition, the colocalization of the striolar HC marker Calbindin with Myosin7a showed that HCs tended to adopt the striolar HC fate (Fig. 2F-G). These results suggest that neomycin triggered the maculae to respond to Notch activation, and this significantly promoted the proliferation of sensory cells across the whole maculae.
Notch signaling inhibition after neomycin exposure was able to promote proliferated cells to adopt the HC fate
As demonstrated above, the activation of Notch signaling promoted sensory cell proliferation in the utricles, which was similar to what is seen in Notch-induced pro-sensory domain formation during embryonic development. Thus, the next question we asked was if these proliferated cells had similar capability to differentiate into HCs over the course of development. Given that the expression of Atoh1 has long been known to be able to induce HC formation, we transfected the Notch-activated neomycin exposed maculae with ad-Atoh1 virus overnight and then cultured them for another 14 days (Fig. 3A). Compared to controls (Fig. 3B), the Atoh1 groups had significantly more proliferated EdU+/Myosin7a + HCs (Fig. 3C-D, K-M). Cells with Myosin7a/Sox2/EdU triple labeling indicated the possibility that HCs transdifferentiated from proliferated SCs (Fig. 3D, arrows). To characterize the newly generated HCs, we first used Actin to mark the whole length of the stereocilia and Espin to mark the tip region of the stereocilia. Several HCs showed immature but stereocilia-like structures on their apical surfaces (Fig. 3G). Then, to determine if the stereocilia were functional, we stained the maculae with FM1-43, a fluorescent dye that passes through functional mechanotransduction channels. The co-localization of FM1-43 and Myosin7a indicated that HCs across the maculae were able to transduce ions into the HCs (Fig. 3I). Together, our results indicated that activating Notch after injury in the utricle induced the proliferation of sensory cells and primed the sensory cells to respond to Atoh1 to adopt the HC fate.
Previous studies demonstrated that inhibition of Notch removes the barrier of lateral inhibition and further promotes more pro-sensory cells to adopt the HC fate during development and regeneration11,17,18. Because the activation of Notch signaling can promote the proliferation of sensory SCs, we wondered whether further inhibition of Notch could induce the proliferated SCs to differentiate into HCs, which would recapitulate the dynamic changes in Notch signaling during inner ear development. We used the Notch inhibitor, DAPT, which is an r-secretase inhibitor that reduces the cleavage of Notch activators in culture and thus inhibits Notch expression19. Similar to Atoh1, DAPT-treated utricles also had significantly increased HCs, especially in the striolar regions (Fig. 3E-F, K-M). Compared to 12.8 HCs per 100 µm2 in the control groups (N = 3), the DAPT group had 28.67 HCs (N = 3) and the Atoh1 group had 38.5 HCs per 100 µm2 in the striolar region (N = 4). Newly generated HCs also expressed Parvalbumin (Fig. 3I). Through the staining of the cilia markers Actin and EspinA, we found that the regenerated HCs had immature bundles on their apical surface. In this case, our results confirmed the hypothesis that sequentially regulating Notch signaling after injury in the utricles recapitulates the developmental processes to promote HC regeneration.
Generation of in vivo models for bi-directional Notch manipulation by localized drug delivery via the round window
Given that our in vitro data clearly demonstrated the potential of Notch signaling alone in HC regeneration in adult utricles, we next investigated in vivo models to test if similar results were obtained in endogenous environments. Similar mouse lines were used as for the explant culture. However, intraperitoneal injection of Dox the in RosartTA/+/tetONICD/+ transgenic mice was lethal. In this case, we tried to apply Dox directly via the round window. Previous groups reported that applying Poloxamer 407 gel on the round window in guinea pigs could maintain at the location for 2 weeks 20,21. Thus, we mixed Dox together with Poloxamer 407 (Dox/P407) for injection through the round window in P30 mice (Fig. 4A-C). These mice were viable, and the sensory utricular maculae were not disturbed by the surgery or by sustained Dox release. Furthermore, we injected EdU daily after the gel was applied until day 20. Our results agreed with a previous paper showing that activating only the Notch signaling pathway in vivo could only stimulate a small amount of spontaneous regeneration in adult mice (Fig. 4D-E). In this case, locally injected Dox/P407 gel had the expected effect in the vestibular system while keeping the mice viable.
We next applied the phased bidirectional regulation of Notch signaling in vivo using the viable mouse model. We applied Dox/P407 gel to the round window 6 days after intraperitoneal injection of IDPN 16,22, which is a vestibular cell toxin, to efficiently kill the HCs and SCs. In addition, EdU injection clearly showed the presence of proliferated sensory cells in the damaged adult mouse utricles (Fig. 5E). Compared to the average of 5.77 EdU/Sox2 double-positive cells in the striola region of control groups, there were an average of 15.03 EdU/Sox2 double-positive cells in the Dox group (Fig. 5I). Upon Notch activation, the density of HCs, especially in the extrastriolar regions, was significantly recovered after injury. Furthermore, to determine if the activated Notch pathway was working normally, we also screened the downstream genes of the Notch signaling pathway, namely Hes1, Hes5, and Hey1, which showed increased expression according to the RT-PCR results (Fig. 5F). Thus, similar to the in vitro data, the activation of Notch after damage is able to promote sensory cell proliferation and HC regeneration.
We next mixed DAPT in the Poloxamer 407 gel and applied it to the round window after IDPN injection. More proliferated SCs were observed as expected (Fig. 5D). The HC density recovered from the IDPN-induced damage, which may be due to direct transdifferentiation of proliferated SCs, which was similar to previous in vitro data. Furthermore, we sequentially activated and inhibited Notch after damage in vivo. Six days after IDPN injection in P30 mice, we applied Dox/P407 gel on the round window via surgical methods, and this was 6 days prior to another round of surgery for DAPT/P407 gel placements. The maculae were harvested one month later. Compared to the single induction of activation or inhibition, the dual regulation of Notch signaling pathways resulted in significantly more SCs across the maculae (Fig. 5G-I). The number of HCs was considerably increased compared to single manipulations. In addition to increased numbers, more Myosin7a and Sox2 double-positive cells were triple-labeled with EdU. The Sox2 and EdU double-labeled cells were mainly found in the striolar region (n = 23.19, Fig. 5I), while the number of sensory cells was increased in the extrastriolar region after Notch manipulation. These results indicated that sensory cells were able to be rejuvenated and reprogrammed via Notch regulation.
Given that the regenerated HCs appear to be immature, we extended the harvest time to around 3 months. Extending the development time led the regenerated HCs with triple labels to express stronger Myosin7a signals (Fig. 6A-B, yellow arrowhead). To further characterize these newly differentiated HCs, we used Phalloidin to label the stereocilia in hair bundles. The newly formed HC-like cells had immature but clearly bundle-like cilia on their apical surface (Fig. 6C-E). These results indicated that the two-step-regulation of Notch signaling in the adult maculae recapitulated the developmental processes that promote SC proliferation after damage and could induce further HC regeneration (Fig. 6F-G).