Biology subjects

Stefanovic, B.

Publications and source records attributed to Stefanovic, B..

12 recordsLinked to original sources

bioRxiv2026

Aberrant cortical reorganization and astrogliosis underlie memory deficit post-traumatic brain injury: a longitudinal multiscale multimodal study

Integrating high-dimensional neuroimaging data across spatial scales remains a fundamental bottleneck in quantitative neuroscience. Here, we apply emerging analytical techniques including functional connectome gradient mapping and 3D segmentation of whole-brain cellular resolution microscopy data in a mouse model of traumatic brain injury. At the macroscale, longitudinal functional MRI revealed a transition from frontal hyperconnectivity to hippocampal hypoconnectivity and unimodal-transmodal gradient contraction indicating network dedifferentiation. At the cellular scale, deep-learning based segmentation of whole-brain light-sheet fluorescence microscopy volumes quantified regional astrogliosis, neuronal hyperactivity, and single-axon loss undetected by standard diffusion MRI. Multiscale multimodal data integration not only identified quantitative biomarkers predictive of memory performance post-injury but also enabled validation of in vivo modalities against more sensitive ex vivo histology. This study demonstrates a framework for probing complex multi-network cognitive functions and neurological disorders of network dysfunction.

neuroscience↗

bioRxiv2026

Chronic Stress Exacerbates Long-term Microvascular Network Dysfunction Following Brain Trauma

BackgroundPreexisting factors are among the strongest predictors of recovery following traumatic brain injury (TBI), with chronic stress closely associated with permanent disability and worse long-term outcomes. While chronic cerebrovascular dysfunction is linked to these poor trajectories, as impaired blood flow regulation drives secondary disease progression, the mechanisms regulating this vascular failure remain incompletely understood. Crucially, how premorbid chronic stress and TBI disrupt the fundamental coordination of the cerebrovascular network post-injury remains unknown. MethodsTo elucidate this coordination at rest and in response to increased neuronal activity, we used a model of three repeated moderate closed cortical impacts comorbid with chronic stress induced by social isolation (SI) post-weaning. Our previously developed vascular analysis pipeline (NOVAS3D) was employed to estimate changes in vascular radii across cerebrovasculature proximal to neuronal activation. Arteries and veins were annotated in the imaged volumes to allow for blood flow simulations. ResultsUsing graph-based network analysis, we demonstrate that TBI, when compounded with chronic stress, critically disrupts the long-range coordination of the capillary network. Specifically, the functional coordination of radius changes between nearby, non-adjacent capillaries was reduced by 40{+/-}20% in TBI+SI mice relative to controls. Consequently, simulations estimated that the vascular networks in TBI+SI mice experienced a 68{+/-}7% reduction in arterial red blood cell velocity (VRBC) responses to neuronal activation. These network-wide impairments were fundamentally driven by severely blunted vessel reactivity, including a 28{+/-}7% decrease in the magnitude of arteriolar dilations and a 47{+/-}7% decrease in the magnitude of arteriolar constrictions. ConclusionThese findings provide a mechanistic foundation for worse clinical outcomes seen in TBI patients with comorbid chronic stress, identifying arteriolar reactivity and long-range capillary coordination as critical therapeutic targets to mitigate secondary injury and improve long-term recovery.

neuroscience↗

bioRxiv2025

Stable neuronal representations to repeated stimulation underlie cognitive resilience in Alzheimer's disease pathology

While Alzheimers Disease (AD) typically triggers cognitive decline, some individuals with significant AD pathology maintain normal cognition into late life. Understanding the neuronal underpinnings of such cognitive resilience would propel the development of interventions for delaying dementia. To this end, we used cognitive testing to identify a subset of cognitively resilient 13-month-old TgF344-AD rats (established AD) and their non-transgenic littermates, followed by Neuropixels recording from 8500 neurons during repeated somatosensory stimulation. Cognitively resilient TgF344-AD rats recruited fewer neurons yet displayed more stable neuronal representations during repeated stimulations in cortical excitatory and hippocampal inhibitory ensembles, with reduced excitatory spike burstiness during network activation and a distinct pattern of functional synaptic connectivity. These associations existed independently of amyloid and tau levels. For the first time, our study revealed neuronal population-level hallmarks of maintained cognition that may serve as a novel neurophysiological biomarker of cognitive resilience and a target for stabilizing cognition.

neuroscience↗

bioRxiv2025

Quantitative profiling of whole-brain connectomes at single-axon resolution using deep learning and high-resolution light sheet microscopy

Revealing how individual axons create a brain-wide connectome would be indispensable for understanding brain function and behavior, yet remains technically challenging. We introduce MAPL3, an end-to-end pipeline that integrates self-supervised learning with an innovative deep architecture to capture local and global brain-wide axonal projections. MAPL3 enables subject- and population-level quantitative laminar analysis, generalizes across experiments, and outperforms state-of-the-art methods. We showcase its ability to map the circuitry of the orbitofrontal cortex from single axons to whole-brain projectome.

neuroscience↗

bioRxiv2025

Highly effective antifibrotic compound discovered as a contaminant in cefixime preparations: preliminary characterization

There are no approved drugs for specific treatment of excessive type I collagen synthesis in organ fibrosis. The discovery that biosynthesis of type I collagen in fibrosis is regulated by binding of LARP6 to the unique sequence found in type I collagen mRNAs, the 5 stem-loop (5SL), prompt the search for small molecule inhibitors of LARP6 binding. A contaminant in some commercial preparations of the third-generation cephalosporin, cefixime, has been discovered that inhibits LARP6/5SL interaction in vitro and type I procollagen secretion by cells and organoids in culture. The activity, termed ATO-OA, forms stable non-colloidal nano-entities (NE). ATO-OA NE inhibit the binding of LARP6 to 5SL with the IC50 of 3-4 {micro}g/ml and are equally effective in the dissociation of preassembled LARP6/5SL complex as in the inhibition of its formation. ATO-OA NE target the RRM domain of LARP6 to alter its conformation. In cultured cells, ATO-OA NE suppresses type I procollagen secretion into the cellular medium and causes intracellular retention of the protein. The effect of ATO-OA NE in vivo is LARP6 dependent, because the LARP6 knockout cells or the cells which make type I procollagen without 5SL show no effect of ATO-OA. In cultured organoids of human pancreatic adenocarcinoma cells ATO-OA NE diminished type I collagen production at [~]62.5 {micro}g/ml. These results suggest that the discovery of ATO-OA NE as potent and specific inhibitor of LARP6 may be a breakthrough towards development of antifibrotic drugs directly targeting type I collagen biosynthesis.

biochemistry↗

bioRxiv2025

A self-supervised deep learning pipeline for segmentation in two-photon fluorescence microscopy

Two-photon fluorescence microscopy (TPFM) allows in situ investigation of the structure and function of the brain at a cellular level, but the conventional image analyses of TPFM data are labour-intensive. Automated deep learning (DL)-based image processing pipelines used to analyze TPFM data require large labeled training datasets. Here, we developed a self supervised learning (SSL) pipeline to test whether unlabeled data can be used to boost the accuracy and generalizability of DL models for image segmentation in TPFM. We specifically developed four pretext tasks, including shuffling, rotation, axis rotation, and reconstruction, to train models without supervision using the UNet architecture. We validated our pipeline on two tasks (neuronal soma and vasculature segmentation), using large 3D microscopy datasets. We introduced a novel density-based metric, which provided more sensitive evaluation to downstream analysis tasks. We further determined the amount of labeled data required to reach performance on par with fully supervised learning (FSL) models. SSL-based models that were fine-tuned with only 50% of data were on par or superior (e.g., Dice increase of 3% for neuron segmentation and Dice score of 0.88 {+/-} 0.09 for vessel segmentation) to FSL models. We demonstrated that segmentation maps generated by SSL models pretrained on the reconstruction and rotation tasks can be better translated to downstream tasks than can other SSL tasks. Finally, we benchmarked all models on a publicly available out-of-distribution dataset, demonstrating that SSL models outperform FSL when trained with clean data, and are more robust than FSL models when trained with noisy data.

neuroscience↗

bioRxiv2025

LARP6 regulates the mRNA translation of fibrogenic genes in liver fibrosis

Metabolic syndrome and excessive alcohol consumption result in liver injury and fibrosis, which is characterized by increased collagen production by activated Hepatic Stellate Cells (HSCs). LARP6, an RNA-binding protein, was shown to facilitate collagen production. However, LARP6 expression and functionality as a regulator of fibrosis development in a disease relevant model remains elusive. By using snRNA-sequencing, we show that LARP6 is upregulated mainly in HSCs of liver fibrosis patients. Moreover, LARP6 knockdown in human HSCs suppresses fibrogenic gene expression. By integrating eCLIP analysis and ribosome profiling in HSCs, we show that LARP6 interacts with mature mRNAs comprising over 300 genes, including RNA structural elements within COL1A1, COL1A2, and COL3A1 to regulate mRNA expression and translation. Furthermore, LARP6 knockdown in HSC attenuates fibrosis development in human liver spheroids. Altogether, our results suggest that targeting LARP6 in human HSCs may provide new strategies for anti-fibrotic therapy. HighlightsO_LILARP6 is upregulated in liver fibrosis, mainly in HSCs. C_LIO_LILARP6 knockdown in human HSCs reduces liver fibrosis development. C_LIO_LIOf the hundreds of gene targets, LARP6 interacts most with collagen mRNAs. C_LIO_LILARP6 regulates mRNA translation via interaction with 5UTRs. C_LI

cell biology↗

bioRxiv2024

Glioblastoma invasion relies on actomyosin contractility and metalloproteinase activity

PurposeGlioblastoma multiforme (GBM) is an aggressive brain tumor with a 5-year survival rate below 7%. Poor outcomes are driven in part by diffuse invasion into surrounding brain tissue, which limits complete surgical resection and promotes recurrence. Improved therapies require a quantitative understanding of the mechanics of GBM invasion. MethodsPatient-derived GBM neurospheres were embedded in Matrigel and analyzed using live imaging, three-dimensional traction force microscopy (3D TFM), and targeted cytoskeletal perturbations. ResultsInvading GBM cells adopted an elongated, protrusion-rich morphology, with F-actin enriched at the cell periphery and microtubules extending along the protrusion shaft, consistent with a mesenchymal-like invasion program. 3D TFM revealed sustained matrix engagement, including progressive bead clustering, increasing cumulative traction forces, and traction hotspots concentrated near protrusion tips. Perturbation studies separated cytoskeletal requirements for invasion and force transmission: actin polymerization was essential for invasion, myosin II activity was required for robust traction generation and efficient invasion, and microtubule polymerization supported directional persistence and maximal traction output. Notably, a low level of invasion persisted under myosin II inhibition despite minimal detectable traction forces, and this residual invasion was not suppressed by pan-MMP inhibition, indicating a traction-poor, MMP-independent invasion component in Matrigel. ConclusionsThese findings establish a quantitative mechanical framework for GBM neurosphere invasion in 3D Matrigel and define distinct contributions of actomyosin contractility and microtubules to invasive progression.

cancer biology↗

bioRxiv2024

A Deep Learning Pipeline for Mapping in situ Network-level Neurovascular Coupling in Multi-photon Fluorescence Microscopy

1Functional hyperaemia is a well-established hallmark of healthy brain function, whereby local brain blood flow adjusts in response to a change in the activity of the surrounding neurons. Although functional hyperemia has been extensively studied at the level of both tissue and individual vessels, vascular network-level coordination remains largely unknown. To bridge this gap, we developed a deep learning-based computational pipeline that uses two-photon fluorescence microscopy images of cerebral microcirculation to enable automated reconstruction and quantification of the geometric changes across the microvascular network, comprising hundreds of interconnected blood vessels, pre and post-activation of the neighbouring neurons. The pipelines utility was demonstrated in the Thy1-ChR2 optogenetic mouse model, where we observed network-wide vessel radius changes to depend on the photostimulation intensity, with both dilations and constrictions occurring across the cortical depth, at an average of 16.1{+/-}14.3 m (mean{+/-}stddev) away from the most proximal neuron for dilations; and at 21.9{+/-}14.6 m away for constrictions. We observed a significant heterogeneity of the vascular radius changes within vessels, with radius adjustment varying by an average of 24 {+/-} 28% of the resting diameter, likely reflecting the heterogeneity of the distribution of contractile cells on the vessel walls. A graph theory-based network analysis revealed that the assortativity of adjacent blood vessel responses rose by 152 {+/-} 65% at 4.3 mW/mm2 of blue photostimulation vs. the control, with a 4% median increase in the efficiency of the capillary networks during this level of blue photostimulation in relation to the baseline. Interrogating individual vessels is thus not sufficient to predict how the blood flow is modulated in the network. Our computational pipeline, to be made openly available, enables tracking of the microvascular network geometry over time, relating caliber adjustments to vessel wall-associated cells state, and mapping network-level flow distribution impairments in experimental models of disease.

neuroscience↗

bioRxiv2024

Obesity Facilitates Sex-Specific Improvement In Cognition And Neuronal Function In A Rat Model Of Alzheimer's Disease

Obesity reduces or increases the risk of developing Alzheimers disease (AD) depending on whether it is assessed in mid-life or late-life. There is currently no consensus on the relationship between obesity and AD or the mechanism or their interaction. Here, we aim to differentiate the cause-and-effect relationship between obesity and AD in a controlled rat model of AD. We induced obesity in 9-month-old TgF344-AD rats, that is pathology-load wise similar to early symptomatic phase of human AD. To more accurately model human obesity, we fed both TgF344-AD and non-transgenic littermates a varied high-carbohydrate-high-fat diet consisting of human food for 3 months. Obesity increased overall glucose metabolism and slowed cognitive decline in TgF344-AD rats, specifically executive function, without affecting non-transgenic rats. Pathological analyses of prefrontal cortex and hippocampus showed that obesity in TgF344-AD rats produced varied effects, with increased density of myelin and oligodendrocytes, lowered density and activation of microglia that we propose contributes to the cognitive improvement. However, obesity also decreased neuronal density, and promoted deposition of amyloid-beta plaques and tau inclusions. After 6 months on the high-carbohydrate-high-fat diet, detrimental effects on density of neurons, amyloid-beta plaques, and tau inclusions persisted while the beneficial effects on myelin, microglia, and cognitive functions remained albeit with a lower effect size. By examining the effect of sex, we found that both beneficial and detrimental effects of obesity were stronger in female TgF344-AD rats indicating that obesity during early symptomatic phase of AD is protective in females.

neuroscience↗

bioRxiv2020

Characterization of sequence specific binding of LARP6 to the 5' stem-loop of type I collagen mRNAs and implications for rational design of antifibrotic drugs

Excessive synthesis of type I collagen characterizes fibrotic diseases. Binding of LARP6 to the 5 stem-loop (5SL) of collagen mRNAs regulates their translation and the high rate of biosynthesis in fibrosis. LARP6 needs two domains to form stable complex with 5SL RNA, the La-domain and the juxtaposed RRM domain (jointly called the La-module). We describe that the La-domain of LARP6 is necessary and sufficient for recognition of 5SL in sequence specific manner. The three amino acid motif, RNK, located in the flexible loop which connects the second -helix to the {beta}-sheet of the La domain is critical for binding. Mutation of any of these three amino acids abolishes the binding of La-domain to 5SL. The major site of crosslinking of LARP6 to 5SL RNA was mapped to this motif. The RNK motif is not found in other LARPs, which can not bind 5SL. Presence of RRM increases the stability of complex between La-domain and 5SL RNA and RRM domain does not make extensive contacts with 5SL RNA. We propose a model in which the initial recognition of 5SL by LARP6 is mediated by the RNK epitope and further stabilized by the RRM domain. This discovery suggests that the interaction between LARP6 and collagen mRNAs can be blocked by small molecules that target the RNK epitope and will help rational design of the LARP6 binding inhibitors as specific antifibrotic drugs.

molecular biology↗

bioRxiv2018

Acute and chronic stage adaptations of vascular architecture and cerebral blood flow in a mouse model of TBI

The 3D organization of cerebral blood vessels determines the overall capacity of the cerebral circulation to meet the metabolic requirements of the brain. This study used Arterial Spin Labeling (ASL) MRI with a hypercapnic challenge and ex vivo Serial Two-Photon Tomography (STPT) to examine the relationship between blood flow and 3D microvascular structure following traumatic brain injury (TBI) in a mouse. Mice were exposed to a controlled cortical impact TBI and allowed to recover for either 1 day or 4 weeks. At each time point, ASL MRI was performed to quantify cerebral perfusion and the brain vasculature was imaged in 3D with STPT. Registration of ASL to STPT enabled flow changes to be related to the underlying microvascular structure in each ASL voxel. Hypoperfusion under rest and hypercapnia was observed both 1 day and 4 weeks post-TBI. Vessel density and vascular volume were reduced 1 day post-TBI, recovering by 4 weeks; however, the reorganized vasculature at the latter time point possessed an abnormal radial pattern. Our findings demonstrate functionally significant long-term changes in the vascular architecture following injury and illustrate why metrics beyond traditional measures of vessel density are required to understand the impact of vascular structure on function.

neuroscience↗