This Week in The Journal

Stimulating Hair Cell Production in Neonatal Mice

Wenli Ni, Chen Lin, Luo Guo, Jingfang Wu, Yan Chen, et al.

Degeneration of cochlear hair cells causes permanent hearing loss, because hair cells do not regenerate in adult mammals. One potential strategy for restoring hearing after hair cell loss is to stimulate the production of hair cells from cochlear supporting cells. Hair cells and supporting cells arise from the same precursors, and supporting cells can generate replacement hair cells in lower vertebrates. But to be effective, treatments must stimulate supporting cell proliferation as well as hair cell differentiation, because supporting cells have important functions that would be lost if they were all converted into hair cells. Because proliferation and differentiation are regulated by different signaling pathways, a multipronged approach will likely be necessary.

No recently generated cells (red) are present in normal mouse cochlea (top), but overexpressing ß-catenin and knocking out Notch1 (bottom) induced proliferation and increased the number of hair cells (green), some of which (arrows) derived from newly generated cells. See Ni et al. for details.

Ni et al. have made progress toward this goal by simultaneously manipulating Wnt/ß-catenin, Notch1, and Atoh1 signaling. Previous work has shown that ß-catenin acts downstream of Wnt to stimulate proliferation of cochlear precursor cells, Atoh1 specifies hair cell fate, and Notch1 inhibits cell-cycle re-entry in postmitotic cells. Consistent with this, overexpressing ß-catenin or deleting Notch1 in supporting cells induced the generation of new cells in mouse neonatal cochlea. Importantly, the number of newborn cells was significantly higher, and the cells were present in a larger portion of the cochlea, when the two manipulations were combined. Moreover, the combined treatment increased the total number of cells expressing the hair cell marker Myo7a in both the inner and outer hair-cell regions. Few of these cells (<4%) arose from newly generated cells, however, suggesting that transdifferentiation of supporting cells into hair cells accounted for the majority of additional Myo7a-expressing cells in these mice. But adding overexpression of Atoh1 increased the proportion of Myo7a-expressing cells generated from newborn precursors to nearly 14%. Many of these cells expressed other molecular markers of maturing hair cells and had nascent hair bundles.

These results are promising because the combined genetic modifications increase the number of postmitotic hair cells without depleting supporting cells. This provides a strong foundation for future work, which should identify ways to promote full maturation of inner and outer hair cells. In addition, the ability of this strategy to increase hair cell production in the adult cochlea must be investigated.

The Importance of Temporal Sequence in Stimulus Detection

Jorrit S. Montijn, Umberto Olcese, and Cyriel M. A. Pennartz

(see pages 8624–8640)

Neurons in primary visual cortex (V1) are tuned to specific stimulus features, such as orientation. This tuning is broad, however, so many neurons with different preferred orientations respond to a given stimulus. Because different neurons respond with different latencies, the population response typically lasts a few hundred milliseconds. These responses vary across presentations of an identical stimulus. Consequently, each stimulus presentation elicits activity in a slightly different population of differently tuned neurons that fire with varied latencies. How object identity is extracted from this complex activity is unclear, but which neurons are activated and the relative timing of their activity are both thought to be important.

To assess the role of the temporal pattern of neuronal activation in visual stimulus detection, Montijn et al. used calcium imaging to record the activity of ∼100 neurons in layer 2/3 of V1 while mice performed a task requiring the detection of drifting gratings of different orientations and contrasts. They defined short time epochs (“population events” averaging ∼145 ms) during which at least one neuron was active, and then identified clusters of neurons (“assemblies”) that were activated during multiple such events. Each assembly comprised a set of core neurons, with varying participation of additional neurons. Somewhat surprisingly, the core members of assemblies did not necessarily have the same orientation preference, but they tended to be clustered anatomically. Visual stimulus onset temporarily increased the occurrence of population events, and the sequence in which assemblies were activated was more consistent when the stimulus was detected than when it was missed. Similarly, the relative timing of activation of core neurons within population events was more consistent for detected than for missed stimuli. The pattern of temporal activation within and between assemblies appeared to be unrelated to stimulus orientation, however.

These results suggest that the temporal sequence of evoked neuronal activity is important for visual stimulus detection. Although stimulus orientation did not appear to affect assembly composition or temporal sequencing in these experiments, it is important to note that orientation was irrelevant to the task, which simply required stimulus detection. Future studies should determine whether temporal sequencing aids orientation discrimination when such discrimination is necessary.