Biology subjects

Lewis, J. E.

Publications and source records attributed to Lewis, J. E..

12 recordsLinked to original sources

bioRxiv2025

Hypothalamic deiodinase type-3 establishes the period of circannual interval timing in mammals

Animals respond to environmental cues to time phenological events, but the intrinsic mechanism of circannual timing remains elusive. We used transcriptomic sequencing and frequent sampling of multiple hypothalamic nuclei in Djungarian hamster to examine the neural and molecular architecture of circannual interval timing. Our study identified three distinct phases of transcript changes, with deiodinase type-3 (Dio3) expression activated during the early induction phase. Subsequent work demonstrated that targeted mutation of Dio3 using CRISPR-Cas resulted in a shorter period for circannual interval timing. Hamsters that are non-responsive to short photoperiod and fail to show any winter adaptations do not display changes in Dio3 expression do not show any change in body mass or pelage. Our work demonstrates that changes in Dio3 induction is essential for setting the period of circannual interval timing.

neuroscience↗

bioRxiv2023

Automated Deep Learning-Based Diagnosis and Molecular Characterization of Acute Myeloid Leukemia using Flow Cytometry

Current flow cytometric analysis of blood and bone marrow samples for diagnosis of acute myeloid leukemia (AML) relies heavily on manual intervention in both the processing and analysis steps, introducing significant subjectivity into resulting diagnoses and necessitating highly trained personnel. Furthermore, concurrent molecular characterization via cytogenetics and targeted sequencing can take multiple days, delaying patient diagnosis and treatment. Attention-based multi-instance learning models (ABMILMs) are deep learning models which make accurate predictions and generate interpretable insights regarding the classification of a sample from individual events/cells; nonetheless, these models have yet to be applied to flow cytometry data. In this study, we developed a computational pipeline using ABMILMs for the automated diagnosis of AML cases based exclusively on flow cytometric data. Analysis of 1,820 flow cytometry samples shows that this pipeline provides accurate diagnoses of acute leukemia [AUROC 0.961] and accurately differentiates AML versus B- and T- lymphoblastic leukemia [AUROC 0.965]. Models for prediction of 9 cytogenetic aberrancies and 32 pathogenic variants in AML provide accurate predictions, particularly for t(15;17)(PML::RARA) [AUROC 0.929], t(8;21)(RUNX1::RUNX1T1) [AUROC 0.814], and NPM1 variants [AUROC 0.807]. Finally, we demonstrate how these models generate interpretable insights into which individual flow cytometric events and markers deliver optimal diagnostic utility, providing hematopathologists with a data visualization tool for improved data interpretation, as well as novel biological associations between flow cytometric marker expression and cytogenetic/molecular variants in AML. Our study is the first to illustrate the feasibility of using deep learning-based analysis of flow cytometric data for automated AML diagnosis and molecular characterization.

pathology↗

bioRxiv2022

An Automated Pipeline for Differential Cell Counts on Whole-Slide Bone Marrow Aspirate Smears

Pathologic diagnosis of bone marrow disorders relies in part on microscopic analysis of bone marrow aspirate (BMA) smears and manual counting of marrow nucleated cells to obtain a differential cell count (DCC). This manual process has significant limitations, including analysis of only a small subset of optimal slide areas and nucleated cells, and inter-observer variability due to differences in cell selection and classification. To address these shortcomings, we developed an automated machine learning-based pipeline for obtaining 11-component DCCs on whole-slide BMAs. This pipeline utilizes a sequential process of identifying optimal BMA regions with high proportions of marrow nucleated cells, detecting individual cells within these optimal areas, and classifying these cells into one of 11 DCC components. Convolutional neural network models were trained on 396,048 BMA region, 28,914 cell boundary, and 1,510,976 cell class images from manual annotations. The resulting automated pipeline produces 11-component DCCs that demonstrate high statistical and diagnostic concordance with manual DCCs among a heterogeneous group of testing BMA slides with varying pathologies and cellularities. Additionally, we show that automated analysis can reduce intra-slide variance in DCCs by analyzing the whole slide and marrow nucleated cells within optimal regions. Finally, pipeline outputs of region classification, cell detection, and cell classification can be visualized using whole-slide image analysis software. This study demonstrates the feasibility of a fully-automated pipeline for generating DCCs on scanned whole-slide BMA images, with the potential for improving the current standard of practice for utilizing BMA smears in the laboratory analysis of hematologic disorders.

pathology↗

bioRxiv2021

The neural cell adhesion molecule NrCAM regulates development of hypothalamic tanycytes

Hypothalamic tanycytes are neural stem and progenitor cells, but little is known of how they are regulated. Here we provide evidence that the cell adhesion molecule, NrCAM, regulates tanycytes in the adult niche. NrCAM is strongly expressed in adult mouse tanycytes. Immunohistochemical and in situ hybridization analysis revealed that NrCAM loss of function leads to both a reduced number of tanycytes and reduced expression of tanycyte-specific cell markers, along with a small reduction in tyrosine hydroxylase-positive arcuate neurons. Similar analyses of NrCAM mutants at E16 identify few changes in gene expression or cell composition, indicating that NrCAM regulates tanycytes, rather than early embryonic hypothalamic development. Neurosphere and organotypic assays support the idea that NrCAM governs cellular homeostasis. Single-cell RNA sequencing (scRNA-Seq) shows that tanycyte-specific genes, including a number that are implicated in thyroid hormone metabolism, show reduced expression in the mutant mouse. However, the mild tanycyte depletion and loss of markers observed in NrCAM-deficient mice were associated with only a subtle metabolic phenotype.

neuroscience↗

bioRxiv2021

Relaxin/insulin-like family peptide receptor 4 (Rxfp4) expressing hypothalamic neurons modulate food intake and preference in mice

Relaxin/insulin-like-family peptide receptor-4 (RXFP4), the cognate receptor for insulin-like peptide 5 (INSL5), has been implicated in feeding behaviour as Rxfp4 knockout mice display shorter meal durations and reduced high fat diet (HFD) intake. Here, we generated transgenic Rxfp4-Cre mice to explore Rxfp4 expression and physiology. Using this model, we identified Rxfp4 expression in the central nervous system, including in the ventromedial hypothalamus (VMH). Intra-VMH infusion of INSL5 increased HFD and highly palatable liquid meal intake (HPM) of ad libitum fed wildtype mice. Single-cell RNA-sequencing of VMH Rxfp4-expressing cells (RXFP4VMH) defined a cluster of Rxfp4-labelled neurons expressing Esr1, Tac1 and Oxtr, alongside known appetite-modulating neuropeptide receptors (Mc4r, Cckar and Nmur2). Viral tracing demonstrated RXFP4VMH neural projections to the bed nucleus of the stria terminalis, paraventricular hypothalamus, paraventricular thalamus and central nucleus of the amygdala. Utilising designer receptors exclusively activated by designer drugs (DREADDs), we found that whole body chemogenetic inhibition (Di) of Rxfp4-expressing cells, mimicking native INSL5-RXFP4 signalling, increased intake of HFD and HPM, whilst activation (Dq), either at whole body level or specifically within the VMH, reduced HFD and HPM intake and altered food preference. Ablating VMH Rxfp4-expressing cells recapitulated the lower HFD intake phenotype of Rxfp4 knockout mice, resulting in reduced body weight. These findings identify a discrete Rxfp4-expressing neuronal population as a key regulator of food intake and preference and reveal hypothalamic RXFP4 signalling as a target for feeding behaviour manipulation.

physiology↗

bioRxiv2020

Connectivity for Rapid Synchronization in a Neural Pacemaker Network

Synchronization is a fundamental property of biological neural networks, playing a mechanistic role in both healthy and disease brain states. The medullary pacemaker nucleus of the weakly electric fish is a synchronized network of high-frequency neurons, weakly coupled via gap junctions. Synchrony in the pacemaker is behaviourally modulated on millisecond timescales, but how gap junctional connectivity enables such rapid resynchronization speeds is poorly understood. Here, we use a computational model of the pacemaker, along with graph theory and predictive analyses, to investigate how network properties, such as randomness and the directionality of coupling (bidirectional/non-rectifying versus directional/rectifying gap junctions) characterize the fast synchronization of the pacemaker network. Our results provide predictions about connectivity in the pacemaker and insight into the relationship between structural network properties and synchronization dynamics in neural systems more generally.

neuroscience↗

bioRxiv2020

Integration of machine learning and genome-scale metabolic modeling identifies multi-omics biomarkers for radiation resistance

Resistance to ionizing radiation, a first-line therapy for many cancers, is a major clinical challenge. Personalized prediction of tumor radiosensitivity is not currently implemented clinically due to insufficient accuracy of existing machine learning classifiers. Despite the acknowledged role of tumor metabolism in radiation response, metabolomics data is rarely collected in large multi-omics initiatives such as The Cancer Genome Atlas (TCGA) and consequently omitted from algorithm development. In this study, we circumvent the paucity of personalized metabolomics information by characterizing 915 TCGA patient tumors with genome-scale metabolic Flux Balance Analysis models generated from transcriptomic and genomic datasets. Novel metabolic biomarkers differentiating radiation-sensitive and -resistant tumors were predicted and experimentally validated, enabling integration of metabolic features with other multi-omics datasets into ensemble-based machine learning classifiers for radiation response. These multi-omics classifiers showed improved classification accuracy, identified novel clinical patient subgroups, and demonstrated the utility of personalized blood-based metabolic biomarkers for radiation sensitivity. The integration of machine learning with genome-scale metabolic modeling represents a significant methodological advancement for identifying prognostic metabolite biomarkers and predicting radiosensitivity for individual patients.

systems biology↗

bioRxiv2020

Personalized Genome-Scale Metabolic Models Identify Targets of Redox Metabolism in Radiation-Resistant Tumors

Redox cofactor production is integral towards antioxidant generation, clearance of reactive oxygen species, and overall tumor response to ionizing radiation treatment. To identify systems-level alterations in redox metabolism which confer resistance to radiation therapy, we developed a bioinformatics pipeline for integrating multi-omics data into personalized genome-scale flux balance analysis models of 716 radiation-sensitive and 199 radiation-resistant tumors. These models collectively predicted that radiation-resistant tumors reroute metabolic flux to increase mitochondrial NADPH stores and ROS scavenging. Simulated genome-wide knockout screens agreed with experimental siRNA gene knockdowns in matched radiation-sensitive and -resistant cancer cell lines, revealing gene targets involved in mitochondrial NADPH production, central carbon metabolism, and folate metabolism that allow for selective inhibition of glutathione production and H2O2 clearance in radiation-resistant cancers. This systems approach represents a significant advancement in developing quantitative genome-scale models of redox metabolism and identifying personalized metabolic targets for improving radiation sensitivity in individual cancer patients.

systems biology↗

bioRxiv2019

Disruption of Redox Balance Enhances the Effects of BRAF-inhibition in Melanoma Cells

SummaryMelanomas harboring BRAF mutations can be treated with BRAF inhibitors (BRAFi), but responses are varied and tumor recurrence is inevitable. Here, using an integrative approach of experimentation and mathematical flux balance analyses in BRAF-mutated melanoma cells, we report that elevated antioxidant capacity is linked to BRAFi sensitivity in melanoma cells. High levels of antioxidant metabolites in cells with reduced BRAFi sensitivity confirm this conclusion. By extending our analyses to other melanoma subtypes in TCGA, we predict that elevated redox capacity is a general feature of melanomas, not previously observed. We propose that redox vulnerabilities could be exploited for therapeutic benefits and identify unsuspected combination targets to enhance the effects of BRAFi in any melanoma, regardless of mutational status.

systems biology↗

bioRxiv2019

Standard housing temperature but not β3-adrenoreceptor agonism drives weight-gain and adipose tissue deposition in rats following diet induced obesity at thermoneutrality

Therapeutic activation of thermogenic brown adipose tissue (BAT) may be feasible to prevent, or treat, cardiometabolic disease. However, rodents are commonly housed below thermoneutrality ([~]20{degrees}C) which can modulate their metabolism and physiology including the hyperactivation of brown (BAT) and beige white adipose tissue. We housed animals at thermoneutrality from weaning to chronically supress BAT, mimic human physiology and explore the efficacy of chronic, mild cold-exposure and {beta}3-adrenoreceptor agonism under these conditions. Using metabolic phenotyping and exploratory proteomics we show that transfer from 28{degrees}C to 20{degrees}C drives weight gain and a 125% increase in subcutaneous fat mass, an effect not seen with YM-178 administration thus suggesting a direct effect of a cool ambient temperature in promoting weight gain and further adiposity in obese rats. Following chronic suppression of BAT, uncoupling protein 1 mRNA was undetectable in IWAT in all groups. Using exploratory adipose tissue proteomics, we reveal novel gene ontology terms associated with cold-induced weight gain in BAT and IWAT whilst Reactome pathway analysis highlights the regulation of mitotic (i.e. G2/M transition) and metabolism of amino acids and derivatives pathways. Conversely, YM-178 had minimal metabolic-related effects but modified pathways involved in proteolysis (i.e. eukaryotic translation initiation) and RNA surveillance across both tissues. Taken together these findings are indicative of a novel mechanism whereby animals increase body weight and fat mass following chronic suppression of adaptive thermogenesis from weaning. In addition, treatment with a B3-adrenoreceptor agonist did not improve metabolic health in obese animals raised at thermoneutrality.

physiology↗

bioRxiv2019

Exercise does not induce browning of WAT at thermoneutrality and induces an oxidative, myogenic signature in BAT

Background and aimExercise training elicits diverse effects on brown (BAT) and white adipose tissue (WAT) physiology in rodents. However, these animals are typically housed below their thermoneutral zone (i.e. 28-32{degrees}C). In these conditions, BAT is chronically hyperactive and, unlike human residence, closer to thermoneutrality. Therefore, we set out to determine the effects of exercise training in obese animals at 28{degrees}C (i.e. thermoneutrality) on BAT and WAT in its basal (i.e. inactive) state.\n\nMethodsSprague-Dawley rats (n=12) were housed at thermoneutrality from 3 weeks of age and fed a high-fat diet (HFD). At 12 weeks of age half these animals were randomised to 4-weeks of exercise exercise training, i.e. swim-training (1 hour/day, 5 days per week). Metabolic assessment was undertaken during the final 48h and was followed by interscapular and perivascular BAT and inguinal (I)WAT sampling for the analysis of thermogenic genes and the proteome.\n\nResultsExercise attenuated weight gain but did not affect fat mass or general metabolic parameters (i.e. fasting insulin and glucose). Interestingly, although BAT mass was increased, there was no change in thermogenic gene expression. Differentially regulated proteins in BAT enriched gene ontology (GO) terms including 2-oxoglutarate metabolic process, cytochrome-c activity and mitochondrial respiratory chain complex IV. This was accompanied by an upregulation of multiple proteins and GO terms involved in skeletal muscle physiology suggesting an adipocyte to myocyte switch in BAT. UCP1 mRNA was undetectable in IWAT despite an increase in classical browning markers (i.e. PGC1a and ADRB3) with exercise. Enriched GO terms in IWAT included DNA binding and positive regulation of apoptosis. Impact analysis highlighted carbon metabolism and OXPHOS pathways were regulated by exercise in BAT whilst HIF-1 signalling and cytokine-cytokine receptor interaction were among those modified in IWAT.\n\nConclusionExercise training reduces weight gain in obese animals at thermoneutrality and is accompanied by an oxidative, myogenic signature in BAT, rather than induction of thermogenic genes. This may represent a new, UCP1-independent pathway through which BAT regulates body weight at thermoneutrality.

physiology↗

bioRxiv2017

Predicting clinical outcomes from large scale cancer genomic profiles with deep survival models

Translating the vast data generated by genomic platforms into accurate predictions of clinical outcomes is a fundamental challenge in genomic medicine. Many prediction methods face limitations in learning from the high-dimensional profiles generated by these platforms, and rely on experts to hand-select a small number of features for training prediction models. In this paper, we demonstrate how deep learning and Bayesian optimization methods that have been remarkably successful in general high-dimensional prediction tasks can be adapted to the problem of predicting cancer outcomes. We perform an extensive comparison of Bayesian optimized deep survival models and other state of the art machine learning methods for survival analysis, and describe a framework for interpreting deep survival models using a risk backpropagation technique. Finally, we illustrate that deep survival models can successfully transfer information across diseases to improve prognostic accuracy. We provide an open-source software implementation of this framework called SurvivalNet that enables automatic training, evaluation and interpretation of deep survival models.

bioinformatics↗