Search bioRxiv⌕ Search

Biology subjects

Harmanci, A. S.

Publications and source records attributed to Harmanci, A. S..

5 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↗

Mutant IDH impairs chromatin binding by PDGFB to promote chromosome instability

Non-canonical roles for growth factors in the nucleus have been previously described, but their mechanism of action and biological roles remain enigmatic. Platelet-derived growth factor B (PDGFB) can drive formation of low-grade glioma and here we show that it localizes to the nucleus of human glioma cells where it binds chromatin to preserve genome stability and cell lineage. Failure of PDGFB to localize to the nucleus leads to chromosomal abnormalities, aberrant heterochromatin architecture and accelerated tumorigenesis. Furthermore, nuclear localization of PDGFB is reliant upon the expression levels and mutation status of isocitrate dehydrogenase (IDH). Unexpectedly, we identified macrophages as the predominant source of PDGFB in human, finding that immune-derived PDGFB can localize to the nucleus of glioma cells. Collectively, these studies show that immune derived PDGFB enters the nucleus of glioma cells to maintain genomic stability, while identifying a new mechanism by which IDH mutations promote gliomagenesis.

cancer biology↗

Cell type transcriptional identities are maintained in cultured ex vivo human brain tissue

It is becoming more broadly accepted that human-based models are needed to better understand the complexities of the human nervous system and its diseases. The recently developed human brain organotypic culture model is one highly promising model that requires the involvement of neurosurgeons and neurosurgical patients. Studies have investigated the electrophysiological properties of neurons in such ex vivo human tissues, but the maintenance of other cell types within explanted brain remains largely unknown. Here, using single-nucleus RNA sequencing, we systematically evaluate the transcriptional identities of the various cell types found in six patient samples after fourteen days in culture (83,501 nuclei from day 0 samples and 45,738 nuclei from day 14 samples). We used two pediatric temporal lobectomy samples, an adult frontal cortex sample, two IDH wild-type glioblastoma samples, and one medulloblastoma sample. We found remarkably high correlations of day 14 transcriptional identities to day 0 tissue, especially in tumor cells (r = 0.90 to 0.93), though microglia (r = 0.86), oligodendrocytes (r = 0.80), pericytes (r = 0.77), endothelial cells (r = 0.78), and fibroblasts (r = 0.76) showed strong preservation of their transcriptional profiles as well. Astrocytes and excitatory neurons showed more moderate preservation (r = 0.66 and 0.47, respectively). Because the main difficulty with organotypic brain cultures is the acquisition of human tissue, which is readily available to neurosurgeons, this model is easily accessible to neurosurgeon-scientists and neurosurgeons affiliated with research laboratories. Broad uptake of this more representative model should prompt advances in our understanding of many uniquely human diseases, lead to more reliable clinical trial performance, and ultimately yield better therapies for our patients.

neuroscience↗

NF2 Loss-of-Function and Hypoxia Drive Radiation Resistance in Grade 2 Meningiomas

BackgroundWorld Health Organization Grade 2 meningiomas (G2Ms) exhibit an aggressive natural history characterized by recurrence and therapy resistance. G2Ms with histopathological necrosis have been associated with worse local control (LC) following radiation therapy, but drivers and biomarkers of radiation resistance in these G2Ms remain unknown. MethodsWe performed genetic sequencing and histopathological analysis of 113 G2Ms and investigated the role of intratumoral hypoxia as well as genes of interest through knockdown and clonogenic survival following ionizing radiation. Lastly, we performed transcriptional profiling of our in vitro model and 18 G2M tumors using RNA sequencing. ResultsNF2 loss-of-function (LOF) mutations were associated with necrosis in G2Ms (p=0.0127). Tumors with NF2 mutation and necrosis had worse post-radiation LC compared to NF2 wildtype tumors without necrosis (p=0.035). Under hypoxic conditions, NF2 knockdown increased radiation resistance in vitro (p<0.001). Bulk RNA sequencing of our in vitro model revealed NF2- and hypoxia-specific changes and a 50-gene set signature specific to radiation resistant, NF2 knockdown and hypoxic cells, which could distinguish NF2 mutant and necrotic patient G2Ms by unsupervised clustering. Gene set enrichment analysis of patient tumor and in vitro data revealed downregulation of apoptosis and upregulation of proliferation in NF2-deficient and hypoxic cells, which we validated with functional assays. ConclusionsNF2 LOF in the setting of hypoxia confers radiation resistance through transcriptional programs that reduce apoptosis and promote proliferation. These pathways may identify tumors resistant to radiation and represent therapeutic targets that in the future could improve LC in patients with radiation resistant G2Ms. KEY POINTS1. Spontaneous necrosis with NF2 mutations is associated with radio-resistance in WHO G2Ms. 2. NF2 knockdown in the setting of hypoxia confers radio-resistance to meningioma cells in vitro and is driven by increased cell proliferation and decreased apoptosis. IMPORTANCE OF THE STUDYWorld Health Organization Grade 2 meningiomas (G2M) are often treated with surgical resection followed by radiation, especially in the case of recurrence. However, the mechanisms underlying radiation resistance in G2Ms remain to be identified, and moreover, we lack biomarkers to distinguish G2Ms that will respond to radiotherapy from those that are refractory. In this study we perform histological and molecular analysis of a large cohort of G2Ms to identify predictors of radiation resistance. Using these data and an in vitro model of radiation therapy, we demonstrate that radiation resistance in G2Ms is likely driven by the combination of NF2 gene mutations and the hypoxia that accompanies tumor necrosis. Patients whose tumors bear these two features may therefore benefit from alternative treatments that target specific pathways implicated in radiation resistance.

molecular biology↗

Inhibitory input directs astrocyte morphogenesis through glial GABABR

Communication between neurons and glia plays an important role in establishing and maintaining higher order brain function. Astrocytes are endowed with complex morphologies which places their peripheral processes in close proximity to neuronal synapses and directly contributes to their regulation of brain circuits. Recent studies have shown that excitatory neuronal activity promotes oligodendrocyte differentiation; whether inhibitory neurotransmission regulates astrocyte morphogenesis during development is unknown. Here we show that inhibitory neuron activity is necessary and sufficient for astrocyte morphogenesis. We found that input from inhibitory neurons functions through astrocytic GABABR and that its deletion in astrocytes results in a loss of morphological complexity across a host of brain regions and disruption of circuit function. Expression of GABABR in developing astrocytes is regulated in a region-specific manner by SOX9 or NFIA and deletion of these transcription factors results in region-specific defects in astrocyte morphogenesis, which is conferred by interactions with transcription factors exhibiting region-restricted patterns of expression. Together our studies identify input from inhibitory neurons and astrocytic GABABR as universal regulators of morphogenesis, while further revealing a combinatorial code of region-specific transcriptional dependencies for astrocyte development that is intertwined with activity-dependent processes.

neuroscience↗