Search bioRxiv⌕ Search

Biology subjects

Zhu, J.-J.

Publications and source records attributed to Zhu, J.-J..

3 recordsLinked to original sources

GlioVision: A Multi-Modal MRI Framework for Non-Invasive Glioma Molecular Biomarkers Prediction

Gliomas are aggressive primary brain tumors that necessitate critical molecular biomarker predictions for optimal clinical decision-making. Traditional assessment relies on surgical tumor specimens analysis, which carries procedural risks and sampling bias due to tumor heterogeneity. Existing deep learning methods for non-invasive prediction lack real-time applicability, remain resource-intensive, and are frequently trained on narrowly represented datasets. We present GlioVision, a framework built on the MONAI library to process multimodal data, including glioma MRI and molecular labels, to predict and identify, non-invasively, four major glioma molecular biomarkers: IDH mutation, 1p/19q co-deletion, MGMT methylation, and WHO grade. The core architecture comprises Spatially and Channel-wise Recalibrated 3D DenseNet (SCRU-DenseNet), which utilizes a computational attention gate and an Adaptive Contrast-Specific Processing Stream (ACPS) to tackle multi-site, heterogeneous datasets. We introduced the Confidence-Filtered Predictive Manifold (CFPM) to manage uncertainty by excluding predictions with low confidence. GlioVision is trained and validated on the largest multi-cohort datasets, achieving strong biomarker prediction with AUCs of (IDH 0.94, 1p/19q 0.87, MGMT 0.86, WHO grades 0.92), supporting molecularly defined glioma diagnosis under the WHO 2021 classification guidelines. Finally, we provide a Differential Training Integrity Assessment (DTI-A) to analyze routes of MRI data privacy protections through model obfuscation. Our results advance the codebase, model release, and leakage considerations around MRI data analysis literature.

neuroscience↗

Targeting tGLI1, a novel mediator of tumor therapeutic resistance, using Ketoconazole sensitizes glioblastoma to CDK4/6 therapy and Chemoradiation

Glioblastoma (GBM) remains the most aggressive primary brain tumor in adults, with no effective treatments. While cyclin-dependent kinase 4/6 inhibitors (CDK4/6is) show clinical promise in some cancers, they have not significantly improved survival in GBM patients. This lack of response is attributed to the treatment-resistant glioma stem cell (GSC) population. We previously identified truncated glioma-associated oncogene homolog 1 (tGLI1) as a novel transcription factor promoting GSCs; however, its role in CDK4/6i resistance has never been investigated in any cancer type. Here, we found positive correlations between tGLI1 and CDK4/6 therapeutic resistance in patient datasets and in vitro studies. Pharmacological inhibition of tGLI1 using FDA-approved ketoconazole (KCZ), a tGLI1-specific inhibitor, sensitized GBM and GSCs to CDK4/6is. KCZ+CDK4/6i combination therapy demonstrated synergistic anti-proliferative effects, significantly inhibiting GBM stemness and cell cycle progression while increasing apoptosis. The combination was more efficacious than monotherapies in two orthotopic GBM mouse models. tGLI1 promoted GBM resistance to radiation therapy and temozolomide, while KCZ potentiated effects of these treatments. Collectively, we report for the first time that tGLI1 is a novel mediator of GBM resistance to CDK4/6is, and KCZ sensitizes GBM to CDK4/6is, thereby supporting future clinical utility of novel KCZ+CDK4/6i combinatorial therapy for GBM patients.

cancer biology↗

Dynamic tripartite construct of interregional engram circuits underlies forgetting of extinction memory

Fear extinction allows for adaptive control of learned fear responses but often fails, resulting in a renewal or spontaneous recovery of the extinguished fear, i.e., forgetting of the extinction memory readily occurs. Using an activity-dependent neuronal labeling strategy, we demonstrate that engram neurons for fear extinction memory are dynamically positioned in the medial prefrontal cortex (mPFC), basolateral amygdala (BLA), and ventral hippocampus (vHPC), which constitute an engram construct in the term of directional engram synaptic connectivity from the BLA or vHPC to mPFC, but not that in the opposite direction, for retrieval of extinction memory. Fear renewal or spontaneous recovery switches the extinction engram construct from an accessible to inaccessible state, whereas additional extinction learning or optogenetic induction of long-term potentiation restores the directional engram connectivity and prevents the return of fear. Thus, the plasticity of engram construct underlies forgetting of extinction memory.

neuroscience↗