bioRxiv Science⌕ Search

SEARCH · bioRxiv Science

Results for “pathology”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 73 records · Page 4Linked to original sources
bioRxiv2023

LRRK2 kinase inhibition reverses G2019S mutation-dependent effects on tau pathology spread

Mutations in leucine-rich repeat kinase 2 (LRRK2) are the most common cause of familial Parkinsons disease (PD). These mutations elevate LRRK2 kinase activity, making LRRK2 kinase inhibitors an attractive therapeutic target. LRRK2 kinase activity has been consistently linked to specific cell signaling pathways, mostly related to organelle trafficking and homeostasis, but its relationship to PD pathogenesis has been more difficult to define. LRRK2-PD patients consistently present with loss of dopaminergic neurons in the substantia nigra but show variable development of Lewy body or tau tangle pathology. Animal models carrying LRRK2 mutations do not develop robust PD-related phenotypes spontaneously, hampering the assessment of LRRK2 inhibitors efficacy against disease processes. We hypothesized that mutations in LRRK2 may not be directly related to a single disease pathway, but instead may elevate susceptibility to multiple disease processes, depending on the disease trigger. To test this hypothesis, we have previously evaluated progression of -synuclein and tau pathologies following injection of proteopathic seeds. We demonstrated that transgenic mice overexpressing mutant LRRK2 show alterations in the brain-wide progression of pathology, especially at older ages. Here, we assess tau pathology progression in relation to long-term LRRK2 kinase inhibition. Wildtype or LRRK2G2019S knock-in mice were injected with tau fibrils and treated with control diet or diet containing LRRK2 kinase inhibitor MLi-2 targeting the IC50 or IC90 of LRRK2 for 3 to 6 months. Mice were evaluated for tau pathology by brain-wide quantitative pathology in 665 brain regions and subsequent linear diffusion modeling of progression. Consistent with our previous work, we found systemic alterations in the progression of tau pathology in LRRK2G2019S mice that were most pronounced at 6 months. Importantly, LRRK2 kinase inhibition reversed these effects in LRRK2G2019S mice, but had minimal effect in wildtype mice, suggesting that LRRK2 kinase inhibition is likely to reverse specific disease processes in G2019S mutation carriers, but additional work may be necessary to determine the potential effect in non-carriers. This work supports a protective role of LRRK2 kinase inhibition in G2019S carriers and provides a rational workflow for systematic evaluation of brain-wide phenotypes in therapeutic development.

neuroscience↗
bioRxiv2023

RNA aptamer reveals nuclear TDP-43 pathology is an early aggregation event that coincides with STMN-2 cryptic splicing and precedes clinical manifestation in ALS

TDP-43 is an aggregation-prone protein which accumulates in the hallmark pathological inclusions of amyotrophic lateral sclerosis (ALS). However, analysis of deeply-phenotyped human post-mortem samples has shown that TDP-43 aggregation, revealed by standard antibody methods, correlates poorly with symptom manifestation. Recent identification of cryptic-splicing events, such as the detection of Stathmin-2 (STMN-2) cryptic exons, are providing evidence implicating TDP-43 loss-of-function as a potential driving pathomechanism, but the temporal nature of TDP-43 loss and its relation to the disease process and clinical phenotype is not known. To address these outstanding questions, we used a novel RNA aptamer, TDP-43APT, to detect TDP-43 aggregation and used single molecule in situ hybridization to sensitively reveal TDP-43 loss-of-function and applied these in a deeply-phenotyped human post-mortem tissue cohort. We demonstrate that TDP-43APT identifies pathological TDP-43, detecting aggregation events that cannot be detected by classical antibody stains. We show that nuclear TDP-43 pathology is an early event, occurring prior to cytoplasmic aggregation and is associated with loss-of-function measured by coincident STMN-2 cryptic splicing pathology. Crucially, we show that these pathological features of TDP-43 loss-of-function precede the clinical inflection point and are not required for region specific clinical manifestation. Furthermore, we demonstrate that gain-of-function in the form of extensive cytoplasmic aggregation, but not loss-of-function, is the primary molecular correlate of clinical manifestation. Taken together, our findings demonstrate implications for early diagnostics as the presence of STMN-2 cryptic exons and early TDP-43 aggregation events could be detected prior to symptom onset, holding promise for early intervention in ALS. Short AbstractRecent identification of cryptic-splicing events such as the detection of Stathmin-2 (STMN-2) cryptic exons, are providing evidence implicating TDP-43 loss-of-function as a potential driving pathomechanism in amyotrophic lateral sclerosis (ALS). However, the temporal nature of TDP-43 loss and its relation to clinical phenotype is not known. Here, we used a novel RNA aptamer to detect TDP-43 aggregation and used single molecule ISH to sensitively reveal TDP-43 loss-of-function, applying these methods in a deeply-phenotyped human post-mortem tissue cohort. We show that nuclear TDP-43 pathology is an early event, that coincides with STMN-2 cryptic splicing. Crucially, we show that these pathological features of TDP-43 loss-of-function precede the clinical inflection point and are not required for region specific clinical manifestation. Furthermore, we demonstrate that gain-of-function, but not loss-of-function, is the primary molecular correlate of clinical manifestation. Taken together, our findings demonstrate implications for early diagnostics and intervention prior to symptom onset in ALS. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=83 SRC="FIGDIR/small/563701v1_ufig1.gif" ALT="Figure 1"> View larger version (42K): org.highwire.dtl.DTLVardef@1e49c3forg.highwire.dtl.DTLVardef@1ce05b0org.highwire.dtl.DTLVardef@d77205org.highwire.dtl.DTLVardef@7ed2cd_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗
bioRxiv2023

Noradrenergic signaling controls Alzheimer's disease pathology via activation of microglial β2 adrenergic receptors

In Alzheimers disease (AD) pathophysiology, plaque and tangle accumulation trigger an inflammatory response that mounts positive feed-back loops between inflammation and protein aggregation, aggravating neurite damage and neuronal death. One of the earliest brain regions to undergo neurodegeneration is the locus coeruleus (LC), the predominant site of norepinephrine (NE) production in the central nervous system (CNS). In animal models of AD, dampening the impact of noradrenergic signaling pathways, either through administration of beta blockers or pharmacological ablation of the LC, heightened neuroinflammation through increased levels of pro-inflammatory mediators. Since microglia are the resident immune cells of the CNS, it is reasonable to postulate that they are responsible for translating the loss of NE tone into exacerbated disease pathology. Recent findings from our lab demonstrated that noradrenergic signaling inhibits microglia dynamics via {beta}2 adrenergic receptors ({beta}2ARs), suggesting a potential anti-inflammatory role for microglial {beta}2AR signaling. Thus, we hypothesize that microglial {beta}2 adrenergic signaling is progressively impaired during AD progression, which leads to the chronic immune vigilant state of microglia that worsens disease pathology. First, we characterized changes in microglial {beta}2AR signaling as a function of amyloid pathology. We found that LC neurons and their projections degenerate early and progressively in the 5xFAD mouse model of AD; accompanied by mild decrease in the levels of norepinephrine and its metabolite normetanephrine. Interestingly, while 5xFAD microglia, especially plaque-associated microglia, significant downregulated {beta}2AR gene expression early in amyloid pathology, they did not lose their responsiveness to {beta}2AR stimulation. Most importantly, we demonstrated that specific microglial {beta}2AR deletion worsened disease pathology while chronic {beta}2AR stimulation resulted in attenuation of amyloid pathology and associated neuritic damage, suggesting microglial {beta}2AR might be used as potential therapeutic target to modify AD pathology.

neuroscience↗
bioRxiv2024

TDP-43 pathology is sufficient to drive axon initial segment plasticity and hyperexcitability of spinal motoneurones in vivo in the TDP43-NLS model of Amyotrophic Lateral Sclerosis

A hyperexcitability of the motor system is consistently observed in Amyotrophic Lateral Sclerosis (ALS) and has been implicated in the disease pathogenesis. What drives this hyperexcitability in the vast majority of patients is unknown. This is important to know as existing treatments simply reduce all neuronal excitability and fail to distinguish between pathological changes and important homeostatic changes. Understanding what drives the initial pathological changes could therefore provide better treatments. One challenge is that patients represent a heterogeneous population and the vast majority of cases are sporadic. One pathological feature that almost all ([~]97%) cases (familial and sporadic) have in common is cytoplasmic aggregates of the protein TDP-43 which is normally located in the nucleus. In our experiments we investigated whether this pathology was sufficient to increase neuronal excitability and the mechanisms by which this occurs. We used the TDP-43({Delta}NLS) mouse model which successfully recapitulates this pathology in a controllable way. We used in vivo intracellular recordings in this model to demonstrate that TDP-43 pathology is sufficient to drive a severe hyper-excitability of spinal motoneurones. Reductions in soma size and a lengthening and constriction of axon initial segments were observed, which would contribute to enhanced excitability. Resuppression of the transgene resulted in a return to normal excitability parameters by 6-8 weeks. We therefore conclude that TDP-43 pathology itself is sufficient to drive a severe but reversible hyperexcitability of spinal motoneurones.

neuroscience↗
bioRxiv2024

Divergent and Convergent TMEM106B Pathology in Murine Models of Neurodegeneration and Human Disease

TMEM106B is a lysosomal/late endosome protein that is a potent genetic modifier of multiple neurodegenerative diseases as well as general aging. Recently, TMEM106B was shown to form insoluble aggregates in postmortem human brain tissue, drawing attention to TMEM106B pathology and the potential role of TMEM106B aggregation in disease. In the context of neurodegenerative diseases, TMEM106B has been studied in vivo using animal models of neurodegeneration, but these studies rely on overexpression or knockdown approaches. To date, endogenous TMEM106B pathology and its relationship to known canonical pathology in animal models has not been reported. Here, we analyze histological patterns of TMEM106B in murine models of C9ORF72-related amyotrophic lateral sclerosis and frontotemporal dementia (C9-ALS/FTD), SOD1-related ALS, and tauopathy and compare these to postmortem human tissue from patients with C9-ALS/FTD, Alzheimers disease (AD), and AD with limbic-predominant age-related TDP-43 encephalopathy (AD/LATE). We show that there are significant differences between TMEM106B pathology in mouse models and human patient tissue. Importantly, we also identified convergent evidence from both murine models and human patients that links TMEM106B pathology to TDP-43 nuclear clearance specifically in C9-ALS. Similarly, we find a relationship at the cellular level between TMEM106B pathology and phosphorylated Tau burden in Alzheimers disease. By characterizing endogenous TMEM106B pathology in both mice and human postmortem tissue, our work reveals considerations that must be taken into account when analyzing data from in vivo mouse studies and elucidates new insights supporting the involvement of TMEM106B in the pathogenesis and progression of multiple neurodegenerative diseases.

neuroscience↗
bioRxiv2025

Is there a pathological switch that triggers the onset of renal calcification?

IntroductionNephrocalcinosis, nephrolithiasis and Randalls plaque formation are distinct renal pathologies of biomineralization predominantly originating in the renal papillae. Experimental evidence on the events leading to the initial aggregation of nanometer-scale plaque or stone deposits in these regions are limited. Cellular plasticity is a regulatory mechanism of disease progression, and can lead to the transition of epithelial to mesenchymal stem-cell-like phenotypes, and generate macrophages to trigger pathophysiological alterations underlying renal biomineralization. We aim to understand the pathological mechanisms of biomineralization at the renal papillary tip of clinical patient samples and develop functional assays to analyze mechanisms of disease progression within organ-chip devices in vitro. MethodsWe analyzed clinical cohorts of patient renal papillae tissues obtained via nephrectomy (n=34) categorized as stone formers (SF) vs. non-stone formers (NSF). We studied the histopathology and genetic (bulk RNA-sequencing) composition of patient samples in the two groups. We examined the role of primary cells, including peripheral blood mononuclear cells (PBMCs) - progenitors of macrophages, isolated from patient blood samples to differentiate M1 pro and M2 anti-inflammatory macrophage phenotypes for static culture and flow/stretch analyses on organ-on-a-chip devices (Emulate Inc). We stained tissue sections with histology dyes and conducted digital pathology multiplexing analyses via quantitative pathology software (quPath, GitHub) by training an artificial neural network. We conducted fluorescence in situ hybridization (FISH) studies to identify genetic biomarkers of inflammation extracted from the bulk-RNA sequencing data. ResultsBased on the initial results of digital pathology, we identified renal calcium deposits (p value = 0.0017), collagen deposits (p value = 0.0001), fibrosis (p value = 0.0385) and renal casts or inflammatory cells among SF vs NSF cohorts across the cortex-to-tip region of renal papillae. Bulk RNA-sequencing analyses were primarily conducted with DAVID-KEGG and Panther 17.0 classification databases to highlight key regulatory pathways of interest involved at the onset of renal biomineralization, such as the oxidative stress pathway, hypoxia response via HIF activation, and inflammation mediated by chemokine and cytokine signaling. The FISH studies identified genes involved with inflammation; GALNT3, PLEKHO1, SLCO2A1, and VCAM1. We successfully differentiated patient-derived PBMCs to M1 and M2 macrophage lineages to study the impact of oxidative stress by using static 35 mm plate and flow microfluidic organ-chip instruments, to conduct appropriate functional assays in cell culture. ConclusionThe study outcomes provide insights to the precursors of renal biomineralization and delineated the expression of a pathological switch at the onset of hypoxia. The data will provide a fundamental framework to isolate primary cells from patient samples to conduct cell culture studies under static conditions, and translate the outcomes to flow analyses on a Kidney Chip instrument (Emulate. Inc) to mimic pathological conditions in a microphysiological environment in vitro. The ultimate outcome of this project will lead to the development of functional assays that emulate the kidney microphysiology on an organ-chip instrument, suited for clinical translation as a personalized, precision diagnostics and therapeutics platform.

bioengineering↗
bioRxiv2025

Label Noise in Pathological Segmentation is Overlooked, Leading to Potential Overestimation of Artificial Intelligence Models

Artificial intelligence (AI) has transformed medical imaging, driving advancements in radiology and endoscopy. Semantic segmentation, a pixel-level technique crucial for delineating pathological features, has become a cornerstone of digital pathology. Pathology segmentation AI models are often trained using annotations generated by pathologists. Despite the meticulous care typically exercised, these annotations frequently contain empirical label noise. However, the specific types of label noise in pathology data and their impact on AI model training remain inadequately explored. This study systematically investigated the effects of label noise on the performance of pathology segmentation models. Using publicly available datasets and a breast cancer semantic segmentation dataset, modules were developed to simulate four types of artificial label noise at varying intensity levels. These datasets were used to train deep learning models with encoder-decoder architectures, and their performance was evaluated using metrics such as the Dice coefficient, precision, recall, and intersection over union. The results indicated that models were highly susceptible to overfitting label noise, particularly boundary-dependent noise such as dilation and shrinkage. Discrepancies were identified between apparent performance scores obtained under real-world conditions and true performance scores derived using clean test data. This overestimation risk was most pronounced for datasets containing boundary-altering noise. Furthermore, random combinations of noise types and levels significantly impaired model generalization. This study underscores the critical importance of addressing label noise in pathology datasets. It is proposed that future efforts focus on developing standardized methods for quantifying and mitigating label noise, along with creating robust benchmarks using noise-inclusive datasets. Enhancing annotation quality and addressing label noise can improve the reliability and generalizability of AI in pathology, facilitating broader clinical adoption.

bioinformatics↗
bioRxiv2025

Tubulin transforms Tau and -synuclein condensates from pathological to physiological

Proteins phase-separate to form condensates that partition and concentrate biomolecules into membraneless compartments. These condensates can exhibit dichotomous behaviors in biology by supporting cellular physiology or instigating pathological protein aggregation1-3. Tau and - synuclein (Syn) are neuronal proteins that form heterotypic (Tau:Syn) condensates associated with both physiological and pathological processes. Tau and Syn functionally regulate microtubules8-12, but are also known to misfold and co-deposit in aggregates linked to various neurodegenerative diseases4,5,6,7, which highlights the paradoxically ambivalent effect of Tau:Syn condensation in health and disease. Here, we show that tubulin modulates Tau:Syn condensates by promoting microtubule interactions, competitively inhibiting the formation of homotypic and heterotypic pathological oligomers. In the absence of tubulin, Tau-driven protein condensation accelerates the formation of toxic Tau:Syn heterodimers and amyloid fibrils. However, tubulin partitioning into Tau:Syn condensates modulates protein interactions, promotes microtubule polymerization, and prevents Tau and Syn oligomerization and aggregation. We distinguished distinct Tau and Syn structural states adopted in tubulin-absent (pathological) and tubulin-rich (physiological) condensates, correlating compact conformations with aggregation and extended conformations with function. Furthermore, using various neuronal cell models, we showed that loss of stable microtubules, which occurs in Alzheimers disease and Parkinsons disease patients13,14, results in pathological oligomer formation and loss of neurites, and that functional condensation using an inducible optogenetic Tau construct resulted in microtubule stablization. Our results identify that tubulin is a critical modulator in switching Tau:Syn pathological condensates to physiological, mechanistically relating the loss of stable microtubules with disease progression. Tubulin restoration strategies and Tau-mediated microtubule stabilization can be potential therapies targeting both Tau-specific and Tau/Syn mixed pathologies.

biophysics↗
bioRxiv2025

Distinct alpha-synuclein strains derived from Parkinson's disease patient tissues trigger differential inclusion pathology in a novel biosensor cell model

Background-Synuclein (Syn) can misfold and aggregate to form fibrillar {beta}-sheet-rich aggregates ("strains") that are phosphorylated (p-Syn) and deposited into intracellular inclusions in the brain, the pathological hallmark of synucleinopathies including Parkinsons disease (PD), dementia with Lewy bodies (DLB), and multiple system atrophy (MSA). Previously, we reported that seed amplification assays such as real-time quaking-induced conversion (RT-QuIC) amplifies and detects Syn strains from the patient skin. However, whether skin-derived Syn strains induce disease-specific pathological features in a biological system is unknown. MethodsWe generated a U251 human glioblastoma cell line expressing fluorescently tagged Syn carrying the PD-linked A53T mutation and Forster resonance energy transfer (FRET)-based U251 biosensor cells. Using fluorescence microscopy coupled with in situ detergent extraction, FRET-Flow cytometry and high-content confocal imaging, we examined the pathological burden and morphology of p-Syn inclusions seeded by RT-QuIC-amplified patient skin and brain Syn strains in Syn-expressing U251 cells, FRET-based Syn biosensor cells and Syn biosensor cell-derived neurons. ResultsU251 cells allow robust and rapid in situ detection of detergent-insoluble intracellular Syn inclusions triggered by exogenous Syn seeds. In U251 FRET-based biosensor cells, PD skin-amplified strains induce a greater pathological burden and distinct p-Syn inclusion morphology from PD brain-amplified and DLB skin-amplified strains. Inclusion morphology of DLB and MSA skin- and brain-amplified strains are comparable. Furthermore, skin-amplified Syn strains induce neuronal inclusions and cause degeneration of induced neurons reprogrammed from the U251 biosensor cells. Finally, biosensor cell-propagated PD skin Syn strains induce higher seeding activity measured by RT-QuIC than PD brain and DLB skin Syn strains, linking intracellular pathological burden to in vitro seeding activity. ConclusionsWe report the detection of distinct PD strains derived from patient skin and brain tissues that trigger unique pathological phenotypes in U251 Syn biosensor cells and cause degeneration of reprogrammed neurons from the same cell lineage. Moreover, DLB and MSA skin Syn strains mimic pathological features of their brain Syn strains in these biosensor cells. Therefore, the U251 Syn biosensor cell model is a robust tool to potentially discriminate PD and DLB synucleinopathies and to study Syn tissue- and strain-specific etiology and pathogenesis. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=97 SRC="FIGDIR/small/646513v1_ufig1.gif" ALT="Figure 1"> View larger version (28K): org.highwire.dtl.DTLVardef@acf0a5org.highwire.dtl.DTLVardef@1eb0d17org.highwire.dtl.DTLVardef@a89c75org.highwire.dtl.DTLVardef@1b7eca4_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗
bioRxiv2025

Isoliensinine Ameliorates Pathological Cardiac Hypertrophy After Aerobic Exercise by Activating Cytochrome c Oxidase Subunit 6A2

Pathological cardiac hypertrophy is an abnormal remodeling process of the heart under chronic pressure overload, characterized by myocardial fibrosis, apoptosis, and progressive decline in cardiac function, ultimately leading to heart failure or death. Although aerobic exercise, particularly swimming, has been shown to alleviate pathological cardiac hypertrophy, its molecular mechanisms remain incompletely understood. In this study, metabolomic and transcriptomic analyses revealed the molecular mechanisms underlying swimming-induced improvement of pathological cardiac hypertrophy. Results showed that swimming training significantly altered the level of isoliensinine in animals. Further studies demonstrated that isoliensinine exerts cardioprotective effects by regulating the expression of Cytochrome C Oxidase Subunit 6A2 (COX6A2), establishing for the first time the critical role of the Isoliensinine-COX6A2 signaling axis in cardiac function regulation. To validate this axis, a pathological cardiac hypertrophy model was constructed in mice through transverse aortic constriction (TAC). Both swimming training and isoliensinine intervention significantly improved cardiac structure and function. In vitro experiments using isoproterenol (ISO)-induced hypertrophy in H9C2 cells further confirmed the essential role of this signaling axis in suppressing pathological hypertrophy and maintaining mitochondrial homeostasis. Our findings demonstrate that swimming mitigates pathological cardiac hypertrophy by enhancing COX6A2 expression by isoliensinine, thereby improving mitochondrial function. This study not only expands the understanding of exercise-mediated mechanisms in pathological cardiac remodeling but also highlights the cardioprotective potential of isoliensinine and COX6A2 as therapeutic targets for preventing and treating heart failure.

molecular biology↗
bioRxiv2018

Evaluating reproducibility of AI algorithms in digital pathology with DAPPER

Artificial Intelligence is exponentially increasing its impact on healthcare. As deep learning is mastering computer vision tasks, its application to digital pathology is natural, with the promise of aiding in routine reporting and standardizing results across trials. Deep learning features inferred from digital pathology scans can improve validity and robustness of current clinico-pathological features, up to identifying novel histological patterns, e.g. from tumor infiltrating lymphocytes. In this study, we examine the issue of evaluating accuracy of predictive models from deep learning features in digital pathology, as an hallmark of reproducibility. We introduce the DAPPER framework for validation based on a rigorous Data Analysis Plan derived from the FDAs MAQC project, designed to analyse causes of variability in predictive biomarkers. We apply the framework on models that identify tissue of origin on 787 Whole Slide Images from the Genotype-Tissue Expression (GTEx) project. We test 3 different deep learning architectures (VGG, ResNet, Inception) as feature extractors and three classifiers (a fully connected multilayer, Support Vector Machine and Random Forests) and work with 4 datasets (5, 10, 20 or 30 classes), for a total 53000 tiles at 512 x 512 resolution. We analyze accuracy and feature stability of the machine learning classifiers, also demonstrating the need for random features and random labels diagnostic tests to identify selection bias and risks for reproducibility. Further, we use the deep features from the VGG model from GTEx on the KIMIA24 dataset for identification of slide of origin (24 classes) to train a classifier on 1060 annotated tiles and validated on 265 unseen ones. The DAPPER software, including its deep learning backbone pipeline and the HINT (Histological Imaging - Newsy Tiles) benchmark dataset derived from GTEx, is released as a basis for standardization and validation initiatives in AI for Digital Pathology.\n\nAuthor summaryIn this study, we examine the issue of evaluating accuracy of predictive models from deep learning features in digital pathology, as an hallmark of reproducibility. It is indeed a top priority that reproducibility-by-design gets adopted as standard practice in building and validating AI methods in the healthcare domain. Here we introduce DAPPER, a first framework to evaluate deep features and classifiers in digital pathology, based on a rigorous data analysis plan originally developed in the FDAs MAQC initiative for predictive biomarkers from massive omics data. We apply DAPPER on models trained to identify tissue of origin from the HINT benchmark dataset of 53000 tiles from 787 Whole Slide Images in the Genotype-Tissue Expression (GTEx) project. We analyze accuracy and feature stability of different deep learning architectures (VGG, ResNet and Inception) as feature extractors and classifiers (a fully connected multilayer, SVMs and Random Forests) on up to 20 classes. Further, we use the deep features from the VGG model (trained on HINT) on the 1300 annotated tiles of the KIMIA24 dataset for identification of slide of origin (24 classes). The DAPPER software is available together with the HINT benchmark dataset.

bioinformatics↗
bioRxiv2019

Alcohol Drinking Exacerbates Neural and Behavioral Pathology in the 3xTg-AD Mouse Model of Alzheimer’s Disease

Alzheimers disease (AD) is a progressive neurodegenerative disorder that represents the most common cause of dementia in the United States. Although the link between alcohol use and AD has been studied, preclinical research has potential to elucidate neurobiological mechanisms that underlie this interaction. This study was designed to test the hypothesis that non-dependent alcohol drinking exacerbates the onset and magnitude of AD-like neural and behavioral pathology. We first evaluated the impact of voluntary 24-h, 2-bottle choice home-cage alcohol drinking on the prefrontal cortex and amygdala neuroproteome in C57BL/6J mice and found a striking association between alcohol drinking and AD-like pathology. Bioinformatics identified the AD-associated proteins MAPT (Tau), amyloid beta precursor protein (APP), and presenilin-1 (PSEN-1) as the main modulators of alcohol-sensitive protein networks that included AD-related proteins that regulate energy metabolism (ATP5D, HK1, AK1, PGAM1, CKB), cytoskeletal development (BASP1, CAP1, DPYSL2 [CRMP2], ALDOA, TUBA1A, CFL2, ACTG1), cellular/oxidative stress (HSPA5, HSPA8, ENO1, ENO2), and DNA regulation (PURA, YWHAZ). To address the impact of alcohol drinking on AD, studies were conducted using 3xTg-AD mice that express human MAPT, APP, and PSEN-1 transgenes and develop AD-like brain and behavioral pathology. 3xTg-AD and wildtype mice consumed alcohol or saccharin for 4 months. Behavioral tests were administered during a 1-month alcohol free period. Alcohol intake induced AD-like behavioral pathologies in 3xTg-AD mice including impaired spatial memory in the Morris Water Maze, diminished sensorimotor gating as measured by prepulse inhibition, and exacerbated conditioned fear. Multiplex immunoassay conducted on brain lysates showed that alcohol drinking upregulated primary markers of AD pathology in 3xTg-AD mice: A{beta} 42/40 ratio in the lateral entorhinal and prefrontal cortex and total Tau expression in the lateral entorhinal cortex and amygdala at 1-month post alcohol exposure. Immunocytochemistry showed that alcohol use upregulated expression of pTau (Ser199/Ser202) in the hippocampus, which is consistent with late stage AD. According to the NIA-AA Research Framework, these results suggest that alcohol use is associated with Alzheimers pathology. Results also showed that alcohol use was associated with a general reduction in Akt/mTOR signaling via several phosphoproteins (IR, IRS1, IGF1R, PTEN, ERK, mTOR, p70S6K, RPS6) in multiple brain regions including hippocampus and entorhinal cortex. Dysregulation of Akt/mTOR phosphoproteins suggests alcohol may target this pathway in AD progression. These results suggest that nondependent alcohol drinking increases the onset and magnitude of AD-like neural and behavioral pathology in 3xTg-AD mice.

neuroscience↗
bioRxiv2026

Preferential generation of pathological tau species in specific subtypes of entorhinal neurons: implications for Alzheimer's Disease.

Alzheimers Disease (AD) is distinguished by the presence of two key pathological features: amyloid plaques, accumulations of proteolytic products of Amyloid Precursor Protein, and neurofibrillary tangles (NFTs), intracellular aggregations of microtubule-associated protein tau. NFTs first appear in particular entorhinal cortex (EC) neurons (called pre-alpha neurons) in asymptomatic patients, then continue to spread through other connected brain regions as the disease progresses. This stereotypical progression of tauopathy through synaptically connected brain regions (i.e. Braak stages) not only suggests that the tauopathy in AD spreads transsynaptically, it also raises the question whether particular neuronal subtypes and/or brain regions are especially prone to tauopathy. We explored this question by overexpressing wildtype human tau protein (hTau) in a variety of entorhinal and neocortical neuronal subtypes. We then compared the tendencies of different neuronal cell types to develop different pathological tau species over time and found that tau pathology does indeed develop at markedly different rates in different neuronal subtypes. Perhaps unsurprisingly, the EC is particularly prone, with the likely rat cognates of the pre-alpha neurons (ECLII fan cells) among the first to express pathological tau label. Fan cells were not, however, the neurons with the greatest vulnerability to generate pathological tau species: subsets of ECLIII neurons were found to express disproportional amounts of pathological tau. This is of particular interest given that the next structures to develop AD-related tauopathy after the EC in patients are CA1 hippocampus and subiculum, where ECLIII neurons project, rather than the dentate gyrus and CA3, where ECLII fan cells project. These results demonstrate differential susceptibility of different neuronal subtypes to pathological tau species and suggest distinct roles for different entorhinal neuronal subtypes in the propagation of the tauopathy underlying AD-related neurodegeneration.

neuroscience↗
bioRxiv2026

Leveraging Large Language Models to Extract Prognostic Pathology Features in Ewing Sarcoma

BackgroundCurrent risk stratification for Ewing sarcoma relies heavily on clinical factors such as metastatic status, failing to capture histologic heterogeneity as a potential prognostic indicator. Although pathology reports contain rich biological data, this information remains locked in unstructured narrative text, limiting large-scale retrospective analyses. We aimed to validate the utility of Large Language Models (LLMs) for scalable data abstraction and to identify prognostic histologic features from a large multi-institutional cohort. MethodsWe conducted a retrospective cohort study using data from six Childrens Oncology Group (COG) clinical trials. We utilized an LLM-based pipeline (OpenAI o3) to extract structured variables--including immunohistochemical (IHC) markers and CD99 staining patterns--from digitized, Optical Character Recognition (OCR)-processed pathology reports. Extraction accuracy was validated against a human-annotated ground truth (n=200) and cross-validated against senior experts (n=48). We assessed the association between extracted features and Overall Survival (OS) using Kaplan-Meier analysis and multivariable Cox proportional hazards regression, adjusting for metastatic status. FindingsWe analyzed 931 diagnostic pathology reports from 185 institutions spanning over 21-years. The LLM achieved a weighted average accuracy of 94% across 17 IHC markers; in a cross-validation subset, the LLM outperformed human annotators (weighted average accuracy over 15 IHC markers: LLM o3: 98.1%, a pediatric resident 91.4%, and a pediatric oncologist 95.9%). Survival analysis identified Neuron-Specific Enolase (NSE) and S100 as significant prognostic biomarkers. After adjusting for metastatic status, NSE positivity was associated with significantly inferior survival (HR 2.15, 95% CI 1.15-4.02, p=0.016); this risk was most pronounced in patients with non-metastatic disease (HR 5.64, p=0.0055). Conversely, S100 positivity was associated with improved survival (HR 0.58, 95% CI 0.34-1.00, p=0.046). InterpretationLLM-assisted extraction of pathology variables is highly accurate and scalable, capable of unlocking "dark data" from historical clinical trials. We identified NSE as a potent risk factor and S100 as a protective marker in Ewing sarcoma, particularly in localized disease. These findings suggest that AI-derived histologic data can refine risk stratification and, if validated, warrant inclusion in future prospective trials. Research in ContextO_ST_ABSEvidence before this studyC_ST_ABSWe searched PubMed and Google Scholar for articles published up to December 2025, using terms such as "Ewing sarcoma", "prognosis", "risk stratification", "large language models", "natural language processing", and "pathology report extraction". Current risk stratification for Ewing sarcoma relies predominantly on clinical variables, specifically the presence of metastatic disease and tumor size. While histologic heterogeneity is well-documented, it is rarely incorporated into risk models because extracting structured data from narrative pathology reports is labor-intensive. Small, single-institution studies have suggested potential prognostic roles for markers like Neuron-Specific Enolase (NSE) or S100, but results have been inconsistent and limited by small sample sizes. Furthermore, while Large Language Models (LLMs) have demonstrated potential for extracting clinical data, their application to "rescue" data from legacy, multi-institutional clinical trial documents (scanned PDF images) for rare disease biomarker discovery has not been validated at scale. Added value of this studyTo our knowledge, this is the largest study to utilize AI to extract histologic data from Ewing sarcoma pathology reports, aggregating 931 patients from six distinct Childrens Oncology Group (COG) clinical trials spanning over 21 years. We validated an LLM-based pipeline that converted noisy, optically character-recognized text from diverse institutions into structured data with 98.1% accuracy, outperforming human annotators. This scalable approach revealed that NSE positivity is a significant independent risk factor for mortality (HR=2.15), particularly in patients with non-metastatic disease where it confers a more than 5-fold increased risk of death. Conversely, we identified S100 positivity as a protective factor associated with improved survival (HR=0.58). Implications of all the available evidenceThis study demonstrates that LLMs can reliably unlock "dark data" from historical clinical trials, rendering vast archives of unstructured medical documents accessible for retrospective analysis. The identification of NSE and S100 as robust prognostic biomarkers suggests that these widely available immunohistochemical stains provide valuable information beyond standard diagnostic information. These findings support the integration of automated data extraction tools in research workflows and suggest that NSE and S100 status should be considered in the design of future risk-stratified clinical trials for Ewing sarcoma.

bioinformatics↗
bioRxiv2026

Separable multidimensional MRI signatures of cellular and structural pathology in Alzheimer's disease

Cognitive decline in Alzheimers disease (AD) reflects progressive disruption of cellular and microstructural organization, yet the biological specificity of MRI signals remains incompletely understood. Multidimensional diffusion-relaxation MRI (MD-MRI) resolves sub-voxel tissue heterogeneity, offering a potential framework to link imaging signals to underlying pathology. We tested the hypothesis that neuronal, glial and white matter pathologies in AD occupy separable regions of diffusion-relaxation space and generate spatially organized imaging signatures linked to cognitive impairment. We integrated ex vivo MD-MRI with co-registered histology from 12 human donors spanning a range of Braak stages and pathological severity. Using nested cross-validated elastic net modeling, we predicted voxelwise A{beta}, pTau, microglia and myelin burden from the multidimensional diffusion-relaxation density distribution. Regional associations were assessed across hippocampal subfields and white matter, and clinical relevance was evaluated by relating MRI-predicted pathology to Mini-Mental State Examination (MMSE) scores. Distinct diffusion-relaxation components were preferentially associated with different pathological markers, indicating separable microstructural signatures. Voxelwise MRI-derived predictions were significantly associated with histological measures of myelin ({rho} = 0.77), pTau ({rho} = 0.62), and microglia ({rho} = 0.61), with weaker correspondence for A{beta} ({rho} = 0.45). Regionally, predicted pathology recapitulated known patterns of selective vulnerability, with elevated pTau and microglial signal in hippocampal subfields and dominant myelin-associated signal in white matter (p < 0.0001). Importantly, higher predicted pTau density in the hippocampus was strongly associated with worse cognitive performance ({rho} = -0.88, p = 0.0014), with a moderate association in white matter ({rho} = -0.66, p = 0.036), suggesting that tau-related microstructural alterations within both gray and white matter contribute to cognitive impairment. By directly linking multidimensional MRI signatures to histologically verified cellular pathology, this study demonstrates that AD-related processes manifest as distinct and spatially organized diffusion-relaxation signatures. These findings provide mechanistic insight into the microstructural basis of MRI contrasts and support the potential of MD-MRI to map regionally specific neuropathological processes in AD. As clinically feasible MD-MRI acquisition protocols continue to emerge, translation of these spectral signatures to in vivo imaging may enable more mechanistically informed assessment of aging and dementia.

neuroscience↗
bioRxiv2026

Alzheimer disease beta-amyloid pathology accelerates age-related hearing loss: evidence from AppNL-F 'knock-in' mice

Age-related hearing loss (ARHL) is a main acquired risk factor for dementia, including Alzheimer disease (AD), but links are unknown. We are using a mouse model with traits of both aging pathologies to test mechanistic interactions. The "knock-in" AppNL-F mouse reproduces {beta}-amyloid pathology in brain regions homologous to those involved in human AD. Because it was generated from the C57BL/6J mouse, it expresses early signs of ARHL, previously reported in this inbred strain. We found evidence that the early-onset ARHL of the C57BL/6J mouse is accelerated in the AppNL-F mouse. In adult C57BL/6J mice around seven-month-old, there were significant increases in auditory thresholds. In adult age-matched AppNL-F mice, auditory thresholds were significantly more elevated, suggesting acceleration of ARHL. In old mice, past thirteen months of age, hearing thresholds were equally elevated in both strains. Outer hair cell loss was significantly increased in adult AppNL-F relative to age-matched C57BL/6J mice, progressing from basal to apical cochlear turns. Spiral ganglion neuron loss also was larger. In adult AppNL-F mice there was more atrophy and enlarged capillary lumen size in the stria vascularis (SV), supporting accelerated ARHL. These findings suggest that central {beta}-amyloid pathology worsens age-related damage to the auditory receptor, thus accelerating ARHL. Damage to the SV and its capillaries in AppNL-F mice point to exacerbation of strial and vascular pathology in the aging cochlea by central {beta}-amyloid pathology. ARHL acceleration by central {beta}-amyloid pathology may contribute to a vicious circle with implications for prevention and therapies. HighlightsO_LIAge-related hearing loss worsens in a mouse model of Alzheimer {beta}-amyloid pathology. C_LIO_LIHearing thresholds further increase relative to naturally occurring hearing loss. C_LIO_LILoss of outer hair cells and spiral ganglion neurons is larger. C_LIO_LIThe stria vascularis and its microcirculation are more atrophic and damaged. C_LIO_LIAlzheimer disease may potentiate peripheral presbycusis. C_LI

neuroscience↗
bioRxiv2026

The sphingosine-1-phosphate receptor functional antagonist fingolimod exacerbates neuroinflammation and tau pathology in a mouse model of tauopathy

Objective: T cells are increased in the brain in Alzheimer's disease (AD) and primary tauopathies where they correlate with tau pathology. In mouse models, T cells exacerbate tau-related neuroinflammation and neurodegeneration, and depletion of T cells is neuroprotective. Here, we investigate whether the multiple sclerosis drug fingolimod may inhibit infiltration of T cells into the brain in the setting of tau pathology and thereby attenuate tau-mediated neurodegeneration. Methods: P301S-tau transgenic mice expressing human APOE4 (TE4), which develop tau pathology, neurodegeneration, and neuroinflammation with T cell accumulation, were randomly assigned to receive 1 mg/kg/d fingolimod in drinking water or standard mouse drinking water beginning just prior to the accumulation of neurofibrillary tangle pathology at 6 months of age. Mice were then assessed with a battery of cognitive behavioral assays at 9 months of age and euthanized at 9.5 months of age to analyze the blood and brain using flow cytometry, single molecule array (SIMOA) technology, stereotactic brain measurements, and immunohistochemistry. Results: Fingolimod produced the expected lymphopenia in peripheral blood but unexpectedly exacerbated brain T cell accumulation, microgliosis, astrogliosis, and tau pathology, with no significant effect on neurodegeneration or behavioral impairment in TE4 mice. Interpretation: These results indicate fundamental differences in the nature of the adaptive immune response to tau pathology versus other neuroinflammatory diseases such as multiple sclerosis and underscores the importance of understanding the impact of immune modulating therapies on all aspects of AD-related pathology in preclinical models prior to consideration and initiation of clinical trials.

neuroscience↗
bioRxiv2021

The future of artificial intelligence in digital pathology - results of a survey across stakeholder groups

Artificial intelligence (AI) provides a powerful tool to extract information from digitized histopathology whole slide images. In the last five years, academic and commercial actors have developed new technical solutions for a diverse set of tasks, including tissue segmentation, cell detection, mutation prediction, prognostication and prediction of treatment response. In the light of limited overall resources, it is presently unclear for researchers, practitioners and policymakers which of these topics are stable enough for clinical use in the near future and which topics are still experimental, but worth investing time and effort into. To identify potentially promising applications of AI in pathology, we performed an anonymous online survey of 75 computational pathology domain experts from academia and industry. Participants enrolled in 2021 were queried about their subjective opinion on promising and appealing sub-fields of computational pathology with a focus on solid tumors. The results of this survey indicate that the prediction of treatment response directly from routine pathology slides is regarded as the most promising future application. This item was ranked highest in the overall analysis and in sub-groups by age and professional background. Furthermore, prediction of genetic alterations, gene expression and survival directly from routine pathology images scored consistently high across subgroups. Together, these data demonstrate a possible direction for the development of computational pathology systems in clinical, academic and industrial research in the near future.

pathology↗