Robust reprogramming of glia into neurons by inhibition of Notch signaling and NFI factors in adult mammalian retina.

Generation of neurons through direct reprogramming has emerged as a promising therapeutic approach for neurodegenerative diseases. Despite successful applications in vitro , in vivo implementation has been hampered by low efficiency. In this study, we present a highly efficient strategy for reprogramming retinal glial cells into neurons by simultaneously inhibiting key negative regulators. By suppressing Notch signaling through the removal of its central mediator Rbpj, we induced mature Müller glial cells to reprogram into bipolar and amacrine neurons in uninjured adult mouse retinas, and observed that this effect was further enhanced by retinal injury. We found that specific loss of function of Notch1 and Notch2 receptors in Müller glia mimicked the effect of Rbpj deletion on Müller glia-derived neurogenesis. Integrated analysis of multiome (scRNA- and scATAC-seq) and CUT&Tag data revealed that Rbpj directly activates Notch effector genes and genes specific to mature Müller glia while also indirectly represses the expression of neurogenic bHLH factors. Furthermore, we found that combined loss of function of Rbpj and Nfia/b/x resulted in a robust conversion of nearly all Müller glia to neurons. Finally, we demonstrated that inducing Müller glial proliferation by AAV (adeno-associated virus)-mediated overexpression of dominant- active Yap supports efficient levels of Müller glia-derived neurogenesis in both Rbpj - and Nfia/b/x/Rbpj - deficient Müller glia. These findings demonstrate that, much like in zebrafish, Notch signaling actively represses neurogenic competence in mammalian Müller glia, and suggest that inhibition of Notch signaling and Nfia/b/x in combination with overexpression of activated Yap could serve as an effective component of regenerative therapies for degenerative retinal diseases.

Adversarial Uni- and Multi-modal Stream Networks for Multimodal Image Registration.

Deformable image registration between Computed Tomography (CT) images and Magnetic Resonance (MR) imaging is essential for many image-guided therapies. In this paper, we propose a novel translation-based unsupervised deformable image registration method. Distinct from other translation-based methods that attempt to convert the multimodal problem (e.g., CT-to-MR) into a unimodal problem (e.g., MR-to-MR) via image-to-image translation, our method leverages the deformation fields estimated from both: (i) the translated MR image and (ii) the original CT image in a dual-stream fashion, and automatically learns how to fuse them to achieve better registration performance. The multimodal registration network can be effectively trained by computationally efficient similarity metrics without any ground-truth deformation. Our method has been evaluated on two clinical datasets and demonstrates promising results compared to state-of-the-art traditional and learning-based methods.