Refine
Document Type
- Article (3)
Language
- English (3)
Has Fulltext
- yes (3)
Is part of the Bibliography
- no (3)
Keywords
- CRM1 (1)
- Deep learning (1)
- Digital pathology (1)
- Gene regulation (1)
- Glioma (1)
- MEIS2 (1)
- Neural stem cells (1)
- PBX1 (1)
- TALE-homdomain protein (1)
- Transfer learning (1)
Institute
- Medizin (3)
- Georg-Speyer-Haus (1)
- Pharmazie (1)
Adult neurogenesis is regulated by stem cell niche-derived extrinsic factors and cell-intrinsic regulators, yet the mechanisms by which niche signals impinge on the activity of intrinsic neurogenic transcription factors remain poorly defined. Here, we report that MEIS2, an essential regulator of adult SVZ neurogenesis, is subject to posttranslational regulation in the SVZ olfactory bulb neurogenic system. Nuclear accumulation of MEIS2 in adult SVZ-derived progenitor cells follows downregulation of EGFR signaling and is modulated by methylation of MEIS2 on a conserved arginine, which lies in close proximity to nested binding sites for the nuclear export receptor CRM1 and the MEIS dimerization partner PBX1. Methylation impairs interaction with CRM1 without affecting PBX1 dimerization and thereby allows MEIS2 nuclear accumulation, a prerequisite for neuronal differentiation. Our results describe a form of posttranscriptional modulation of adult SVZ neurogenesis whereby an extrinsic signal fine-tunes neurogenesis through posttranslational modification of a transcriptional regulator of cell fate.
PAX6 is a highly conserved transcription factor and key regulator of several neurogenic processes, including the continuous generation of dopaminergic/GABAergic interneurons in the adult ventricular-subventricular (V-SVZ) neurogenic system in mice. Here we report that PAX6 cooperates with the TALE-homeodomain transcription factor PBX1 in this context. Chromatin-immunoprecipitation showed that PBX1 and PAX6 co-occupy shared genomic binding sites in adult V-SVZ stem- and progenitor cell cultures and mouse embryonic stem cells, while depletion of Pbx1 revealed that association of PAX6 with these sites requires the presence of PBX1. Expression profiling together with viral overexpression or knockdown of Pax6 or Pbx1 identified novel PBX1-PAX6 co-regulated genes, including several transcription factors. Computational modeling of genome wide expression identified novel cross-regulatory networks among these very transcription factors. Taken together, the results presented here highlight the intimate link that exists between PAX6 and TALE-HD family proteins and contribute novel insights into how the orchestrated activity of transcription factors shapes adult V-SVZ neurogenesis.
We provide in this paper a comprehensive comparison of various transfer learning strategies and deep learning architectures for computer-aided classification of adult-type diffuse gliomas. We evaluate the generalizability of out-of-domain ImageNet representations for a target domain of histopathological images, and study the impact of in-domain adaptation using self-supervised and multi-task learning approaches for pretraining the models using the medium-to-large scale datasets of histopathological images. A semi-supervised learning approach is furthermore proposed, where the fine-tuned models are utilized to predict the labels of unannotated regions of the whole slide images (WSI). The models are subsequently retrained using the ground-truth labels and weak labels determined in the previous step, providing superior performance in comparison to standard in-domain transfer learning with balanced accuracy of 96.91% and F1-score 97.07%, and minimizing the pathologist's efforts for annotation. Finally, we provide a visualization tool working at WSI level which generates heatmaps that highlight tumor areas; thus, providing insights to pathologists concerning the most informative parts of the WSI.