Biology subjects

Levy, S.

Publications and source records attributed to Levy, S..

At least 19 recordsLinked to original sources

bioRxiv2026

CellCage™ Technology Enables High-throughput Isolation of Defined Cell Combinations to Resolve the Determinants of CAR-T mediated Cytotoxicity

Understanding the cellular determinants of CAR-T cell-mediated cytotoxicity at the single-cell level remains a critical challenge in developing effective cellular immunotherapies. Conventional bulk co-culture assays provide only population-level measurements of cytotoxic activity and cannot resolve the functional heterogeneity of individual effector-target interactions. Here, we demonstrate the Cellanome platform, which enables the formation of CellCage Enclosures (CCEs) through spatially controlled photopolymerization, as a tool for high-resolution, high-throughput characterization of CAR-T cell-mediated cytotoxicity. By isolating defined effector-target combinations within individual CCEs and coupling encapsulation with longitudinal time-lapse imaging and automated image analysis, we resolve cytotoxic activity across thousands of individually tracked CCEs. Using anti-CD19 CAR-T effectors and NALM6-GFP targets, we show that individual CD8+ CAR-T effectors exhibit substantial functional heterogeneity, with serial killing capacity, effector-target contact dynamics, and intrinsic motility each independently correlating with cytotoxic potency. Cytotoxic efficacy scaled positively with effector number, and CD4+ T helper cells augmented CD8+-mediated killing in a dose-dependent manner, providing direct evidence for cooperative T cell behavior during tumor cell clearance. Extending the platform to a solid tumor model using AIC100 anti-ICAM-1 CAR-T cells and HeLa-GFP targets, we show that individual AIC100 effectors are insufficient to mediate effective killing, while multi-effector CCEs exhibit robust cytotoxicity scaling with effector abundance and CD4+:CD8+ composition. These data establish the Cellanome platform as a broadly applicable system for dissecting the determinants of CAR-T cell cytotoxicity at single-cell resolution, with direct translational relevance to the optimization of cellular immunotherapy.

immunology↗

bioRxiv2026

Cell-resolved transcriptional responses during heat-induced coral bleaching and recovery

Coral reefs depend on an intracellular symbiosis between cnidarian hosts and photosynthetic dinoflagellates that is disrupted by ocean warming, causing coral bleaching. Yet how these responses are distributed across coral cell types during stress and recovery remains unclear. Here, we generate a temporal single-cell transcriptomic atlas of the reef-building coral Stylophora pistillata, profiling 20,000 cells during seven days at 32 {degrees}C and after 45 days of recovery. Heat stress reduced symbiont abundance and photosynthetic efficiency, followed by partial restoration. Early oxidative-stress, detoxification, and proteostasis programmes gave way to mitochondrial, cytoskeletal, and vesicular remodelling. These shared responses accompanied cell type-specific changes: epidermal cells activated tissue-renewal programmes, calicoblasts repressed calcification genes, and alga-hosting cells progressively lost transport, lipid-biosynthesis, and nutrient-exchange functions. Targeted profiling of sorted alga-positive cells revealed rare transcriptional states, including a population consistent with symbiont engulfment. Our results reveal how bleaching and recovery are coordinated across coral tissues and symbiotic cell states.

genomics↗

bioRxiv2026

Neuronal regulation of infection recovery prevents pathogenic stress and tissue damage

Neural regulation of immunity is increasingly recognized, yet how the nervous system promotes infection resolution after pathogen clearance remains poorly understood. Using Caenorhabditis elegans, we identify the AIA interneurons as key regulators of both infection and recovery. Acute silencing or genetic ablation of AIA neurons during Salmonella enterica infection reduces host survival, causes excessive activation of conserved immune and stress pathways, including PMK-1/p38 MAPK, insulin/IGF-1 signaling, and the XBP-1-mediated unfolded protein response (UPR), and exacerbates intestinal tissue damage. Strikingly, selective silencing of AIA neurons after pathogen clearance markedly impairs recovery, demonstrating that neural activity is required not only for host defense but also for infection resolution. This defect is rescued by xbp-1, but not pmk-1 or daf-16, knockdown, identifying unresolved ER stress as the principal driver of post-infection mortality. Consistently, AIA silencing during recovery sustains UPR activation and worsens epithelial barrier damage. Together, our findings establish AIA interneurons as central regulators of immune homeostasis that promote infection resolution by limiting excessive ER stress and preserving tissue integrity.

immunology↗

bioRxiv2025

Complete genomes of a multi-generational pedigree to expand studies of genetic and epigenetic inheritance

Pedigree analysis remains the gold standard for rare disease diagnostics, yet whole genome sequencing studies typically omit critical regions like centromeres, telomeres, and acrocentric chromosome p-arms. Here, we present telomere-to-telomere (T2T) reference genomes for four self-identified African American individuals of admixed ancestry spanning three generations. Our parent-of-origin assigned, chromosome-level assemblies revealed precise meiotic recombination breakpoints in previously inaccessible regions, including recombination events across acrocentric and subtelomeric sequences. Centromeric regions were highly stable, with multi-megabase arrays inherited intact across three generations, while the position of kinetochore assembly sites remained consistent and predominantly associated with the p-arm proximal region. The relative lengths of telomeres on individual chromosomes were maintained across generations. Using a targeted rDNA assembly approach, we reconstructed a complete megabase-scale ribosomal DNA (rDNA) array corresponding to the paternal chromosome 14. This openly available pedigree provides a benchmark dataset for studying recombination and genetic and epigenetic variation across the complete genome.

genomics↗

bioRxiv2025

A simplified hybrid capture approach retains high specificity and enables PCR-free workflow

Hybrid capture is a critical technology for selective enrichment of genomic regions of interest, enabling cost-effective focused sequencing in both clinical and research applications. We present a simplified hybrid capture approach that eliminates complexities typically associated with hybridization-based selection methods by directly loading the hybridization product onto the sequencing flow cell. The workflow removes bead-based steps, washes, and post-hybridization PCR, while retaining high capture specificity and library complexity. The approach is enabled by the development of a streptavidin flow cell surface, a method to circularize and amplify captured targets on the flow cell, and a fast hybridization protocol. We demonstrate application across targeted panel sizes and improvements to library complexity and variant calling. We also show how the approach can be used to create an entirely PCR-free targeted sequencing workflow that further improves variant calling and enables the detection of repeat expansions.

genomics↗

bioRxiv2025

Functional characterization of immune cells in a cnidarian reveals an ancestral antiviral program

Examining early-branching animal phyla can help reconstructing the evolutionary origins of immune cells. Here, we characterized immune-related cell programs in embryos of the sea anemone Nematostella vectensis, a model of Cnidaria, which diverged [~]600 million years ago from other animals. Using a transgenic Nematostella reporter line expressing mCherry under the RLRb antiviral promoter, we identified a morphologically and transcriptomically distinct cell population activated by the viral mimic poly(I:C). These cells upregulate immune effector and regulator genes and show increased phagocytic activity. Bulk RNA sequencing of RLRb expressing cells and single-cell transcriptomics revealed gene regulatory programs expressed in specialized immune cells under basal conditions and upon activation. Comparing the Nematostella immune expression profile with that of stony corals treated with the immunostimulant 2'3'-Cyclic GMP-AMP demonstrated a conserved immune response across Hexacorallia. This study uncovers a novel cnidarian immune cell type involved in antiviral immunity, providing insights into the evolutionary history of innate immunity.

evolutionary biology↗

bioRxiv2024

FLEXIBLE, PRODUCTION-SCALE, HUMAN WHOLE GENOME SEQUENCING ON A BENCHTOP SEQUENCER

Human whole-genome sequencing (hWGS) provides comprehensive genomic information that can potentially help guide research in disease prevention and treatment and ultimately improve human health. Recent advancements in sequencing technology have improved sequencing quality and further reduced sequencing costs on bench-top sized instruments, making whole-genome sequencing an accessible technology for broader use. Here, we demonstrate the feasibility of a large whole genome sequencing project using a benchtop sequencer in a small laboratory setting, on a scale previously reserved for production-scale factory-sized machines. In this project, 807 samples were prepared and sequenced across 313 flow cells, with high sequencing quality at a median %Q30 of 96.6% and a median %Q40 of 89.31%. To screen library quality and maximize sample yield, we utilized 48-plex sample pre-pool QC runs to provide >1x coverage per sample prior to sample pooling and full-depth sequencing. With this strategy, we consistently achieved >30x human whole genome sequencing of three-plex sample trios with standard settings or up to 4 samples per run with a high-throughput run setting. Additionally, this low-pass data provided valuable sample-level insights, allowing for detection of chromosome copy number variations (CNV) prior to full-depth sequencing. With three instruments running concurrently, >2,800 30x human whole genomes could be sequenced per year. To further demonstrate additional flexibility present in the platform, we also explored two different use cases 1) large insert sizes (1kb+) library to achieve superior genome coverage; 2) a proof of concept for rapid WGS sequencing to minimize sample to answer turnaround time for time-critical sequencing applications. Sequencing of a 2x100 >30x human WGS can be achieved in <12 hours and subsequent generation of fastq, bam and vcf in <1 additional hour. This study provides a cost-effective and flexible real-world demonstration of achieving both high quality hWGS sequencing and instrument flexibility without the need for complex batching schemes or factory-sized sequencers.

molecular biology↗

bioRxiv2024

Gut Bacteria Encode Reductases that Biotransform Steroid Hormones

The metabolism of steroids by the gut microbiome affects hormone homeostasis, impacting host development, mental health, and reproductive functions. In this study, we identify the {Delta}4-3-ketosteroid 5{beta}-reductase, 3{beta}-hydroxysteroid dehydrogenase/{Delta}5-4 isomerase, and {Delta}6-3-ketosteroid reductase enzyme families encoded by common human gut bacteria. Through phylogenetic reconstruction and mutagenesis, We show that 5{beta}-reductase and {Delta}6-3-ketosteroid reductase have evolved to specialize in converting diverse 3-keto steroid hormones into their 5{beta}- and {Delta}6-reduced derivatives. We also find that the novel 3{beta}-hydroxysteroid dehydrogenase/{Delta}5-4 isomerase is fused with 5{beta}-reductase in multiple species, streamlining the multi-step conversion of pregnenolone, a steroid hormone precursor, into epipregnanolone. Through metagenomic analysis, we reveal that these enzymes are prevalent in healthy populations, being enriched in females over males. These findings provide the molecular basis for studying microbial steroid metabolism in the gut, offering insights into its potential impact on hormonal health in hosts, especially in the context of womens health.

microbiology↗

bioRxiv2024

Development and extensive sequencing of a broadly-consented Genome in a Bottle matched tumor-normal pair for somatic benchmarks

The Genome in a Bottle Consortium (GIAB), hosted by the National Institute of Standards and Technology (NIST), is developing new matched tumor-normal samples, the first to be explicitly consented for public dissemination of genomic data and cell lines. Here, we describe a comprehensive genomic dataset from the first individual, HG008, including DNA from an adherent, epithelial-like pancreatic ductal adenocarcinoma (PDAC) tumor cell line and matched normal cells from duodenal and pancreatic tissues. Data for the tumor-normal matched samples comes from seventeen distinct state-of-the-art whole genome measurement technologies, including high depth short and long-read bulk whole genome sequencing (WGS), single cell WGS, and Hi-C, and karyotyping. In future publications, these data will be used by the GIAB Consortium to develop matched tumor-normal benchmarks for somatic variant detection. We expect these data to facilitate innovation for whole genome measurement technologies, de novo assembly of tumor and normal genomes, and bioinformatic tools to identify small and structural somatic mutations. This first-of-its-kind broadly consented open-access resource will facilitate further understanding of sequencing methods used for cancer biology.

genomics↗

bioRxiv2024

High-throughput DNA engineering by mating bacteria

To reduce the operational friction and scale DNA engineering, we report here an in vivo DNA assembly technology platform called SCRIVENER (Sequential Conjugation and Recombination for In Vivo Elongation of Nucleotides with low ERrors). SCRIVENER combines bacterial conjugation, in vivo DNA cutting, and in vivo homologous recombination to seamlessly stitch blocks of DNA together by mating E. coli in large arrays or pools. This workflow is simpler, cheaper, and higher throughput than current DNA assembly approaches that require DNA to be moved in and out of cells at different procedural steps. We perform over 5,000 assemblies with two to 13 DNA blocks that range from 240 bp to 8 kb and show that SCRIVENER is capable of assembling constructs as long as 23 kb at relatively high throughput and fidelity. Most SCRIVENER errors are deletions between long interspersed repeats. However, SCRIVENER can overcome these errors by enabling assembly and sequence verification at high replication at a nominal additional cost per replicate. We show that SCRIVENER can be used to build combinatorial libraries in arrays or pools, and that DNA blocks onboarded into the platform can be repurposed and reused with any other DNA block in high throughput without a PCR step. Because of these features, DNA engineering with SCRIVENER has the potential to accelerate design-build-test-learn cycles of DNA products.

synthetic biology↗

bioRxiv2024

Efficient and durable gene activation by Cas9-mediated epigenome editing in vivo

Epigenome editing technology holds great promise for treating diverse genetic disorders. While a series of advances has been made on epigenetic silencing using programmable editors, little progress has been made in leveraging epigenetic activation for therapeutic application. Here we demonstrate epigenetic activation of the LAMA1 gene for the treatment of LAMA2-CMD, a severe congenital muscle dystrophy (CMD) caused by biallelic mutations in the LAMA2 gene. LAMA1 is a sister homologue that is known to compensate for the function of LAMA2. However, supplementing LAMA1 or LAMA2 gene via viral platform is not feasible due to the large size of their coding sequences. Through a single administration of our (Adeno-associated virus) AAV vector encoding all the necessary elements for epigenetic activation, a platform termed CRISPR guide-nucleotide directed modulation (GNDM), we observed significant LAMA1 gene upregulation and phenotype improvements in DyW mice, a severe disease model of LAMA2-CMD. Notably, sustained expression of the GNDM gene and subsequent activation of the LAMA1 gene persisted beyond analyzed period of one year despite immune recognition of the GNDM protein by the host immune system. Regulatory T (Treg) cells appeared to facilitate tolerance to GNDM in the transduced muscle tissue. The muscle-tropic AAV capsid exhibited desired vector biodistribution and promising pharmacodynamics with good safety profiles in adult non-human primates (NHPs). Moreover, administration to juvenile NHPs demonstrated superior pharmacodynamics compared to adults, even at half the adult dose, suggesting safer and more effective therapeutic outcomes in mostly pediatric LAMA2-CMD patients. Our approach holds broad applicability for a range of loss-of-function genetic disorders and could offer a therapeutic breakthrough where active epigenome brings clinical benefit.

bioengineering↗

bioRxiv2024

Pooled PPIseq: screening the SARS-CoV-2 and human interface with a scalable multiplexed protein-protein interaction assay platform

Protein-Protein Interactions (PPIs) are a key interface between virus and host, and these interactions are important to both viral reprogramming of the host and to host restriction of viral infection. In particular, viral-host PPI networks can be used to further our understanding of the molecular mechanisms of tissue specificity, host range, and virulence. At higher scales, viral-host PPI screening could also be used to screen for small-molecule antivirals that interfere with essential viral-host interactions, or to explore how the PPI networks between interacting viral and host genomes co-evolve. Current high-throughput PPI assays have screened entire viral-host PPI networks. However, these studies are time consuming, often require specialized equipment, and are difficult to further scale. Here, we develop methods that make larger-scale viral-host PPI screening more accessible. This approach combines the mDHFR split-tag reporter with the iSeq2 interaction-barcoding system to permit massively-multiplexed PPI quantification by simple pooled engineering of barcoded constructs, integration of these constructs into budding yeast, and fitness measurements by pooled cell competitions and barcode-sequencing. We applied this method to screen for PPIs between SARS-CoV-2 proteins and human proteins, screening in triplicate >180,000 ORF-ORF combinations represented by >1,000,000 barcoded lineages. Our results complement previous screens by identifying 74 putative PPIs, including interactions between ORF7A with the taste receptors TAS2R41 and TAS2R7, and between NSP4 with the transmembrane KDELR2 and KDELR3. We show that this PPI screening method is highly scalable, enabling larger studies aimed at generating a broad understanding of how viral effector proteins converge on cellular targets to effect replication.

systems biology↗

bioRxiv2023

Cell2Sentence: Teaching Large Language Models the Language of Biology

We introduce Cell2Sentence (C2S), a novel method to directly adapt large language models to a biological context, specifically single-cell transcriptomics. By transforming gene expression data into "cell sentences," C2S bridges the gap between natural language processing and biology. We demonstrate cell sentences enable the fine-tuning of language models for diverse tasks in biology, including cell generation, complex cell-type annotation, and direct data-driven text generation. Our experiments reveal that GPT-2, when fine-tuned with C2S, can generate biologically valid cells based on cell type inputs, and accurately predict cell types from cell sentences. This illustrates that language models, through C2S fine-tuning, can acquire a significant understanding of single-cell biology while maintaining robust text generation capabilities. C2S offers a flexible, accessible framework to integrate natural language processing with transcriptomics, utilizing existing models and libraries for a wide range of biological applications.

bioinformatics↗

bioRxiv2023

EGFR+ lung adenocarcinomas coopt alveolar macrophage metabolism and function to support EGFR signaling and growth

The limited efficacy of currently approved immunotherapies in EGFR-mutant lung adenocarcinoma (LUAD) underscores the need to better understand mechanisms governing local immunosuppression. Elevated surfactant and GM-CSF secretion from the transformed epithelium induces tumor-associated alveolar macrophages (TA-AM) to proliferate and support tumor growth by rewiring inflammatory functions and lipid metabolism. TA-AM properties are driven by increased GM-CSF--PPAR{gamma} signaling and inhibition of airway GM-CSF or PPAR{gamma} in TA-AMs suppresses cholesterol efflux to tumor cells, which impairs EGFR phosphorylation and restrains LUAD progression. In the absence of TA-AM metabolic support, LUAD cells compensate by increasing cholesterol synthesis, and blocking PPAR{gamma} in TA-AMs simultaneous with statin therapy further suppresses tumor progression and increases T cell effector functions. These results reveal new therapeutic combinations for immunotherapy resistant EGFR-mutant LUADs and demonstrate how such cancer cells can metabolically co-opt TA-AMs through GM-CSF--PPAR{gamma} signaling to provide nutrients that promote oncogenic signaling and growth.

immunology↗

bioRxiv2023

Discovery of the gut microbial enzyme responsible for bilirubin reduction to urobilinogen

The degradation of heme and the interplay of its catabolic derivative, bilirubin, between humans and their gut microbiota is an essential facet of human health. However, the hypothesized bacterial enzyme that reduces bilirubin to urobilinogen, a key step that produces the excretable waste products of this pathway, has remained unidentified. In this study, we used a combination of biochemical analyses and comparative genomics to identify a novel enzyme, BilR, that can reduce bilirubin to urobilinogen. We delineated the BilR sequences from other members of the Old Yellow Enzyme family through the identification of key residues in the active site that are critical for bilirubin reduction and found that BilR is predominantly encoded by Firmicutes in the gut microbiome. Our analysis of human gut metagenomes showed that BilR is a common feature of a healthy adult human microbiome but has a decreased prevalence in neonates and IBD patients. This discovery sheds new light on the role of the gut microbiome in bilirubin metabolism and highlights the significance of the gut-liver axis in maintaining bilirubin homeostasis.

microbiology↗

bioRxiv2022

Druggable Growth Dependencies and Tumor Evolution Analysis in Patient-Derived Organoids of Neuroendocrine Cancer

Neuroendocrine neoplasms (NENs) comprise well-differentiated neuroendocrine tumors and poorly-differentiated carcinomas. Treatment options for patients with NENs are limited, in part due to lack of accurate models. To address this need we established the first patient-derived tumor organoids (PDTOs) from pulmonary neuroendocrine tumors and derived PDTOs from an understudied NEN subtype, large cell neuroendocrine carcinoma (LCNEC). PDTOs maintain the gene expression patterns, intra-tumoral heterogeneity, and evolutionary processes of parental tumors. Through drug sensitivity analyses, we uncover therapeutic sensitivities to an inhibitor of NAD salvage biosynthesis and to an inhibitor of BCL-2. Finally, we identify a dependency on EGF in pulmonary neuroendocrine tumor PDTOs. Consistent with these findings, analysis of an independent cohort showed that approximately 50% of pulmonary neuroendocrine tumors expressed EGFR. This study identifies a potentially actionable vulnerability for a subset of NENs, and further highlights the utility of these novel PDTO models for the study of NENs. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=188 HEIGHT=200 SRC="FIGDIR/small/514549v1_ufig1.gif" ALT="Figure 1"> View larger version (57K): org.highwire.dtl.DTLVardef@14bed0borg.highwire.dtl.DTLVardef@1b7791forg.highwire.dtl.DTLVardef@7120a6org.highwire.dtl.DTLVardef@116b21d_HPS_FORMAT_FIGEXP M_FIG C_FIG HighlightsO_LIPDTOs of pulmonary NETs and LCNEC were established C_LIO_LIPDTOs recapitulate intra-tumoral heterogeneity and evolution of parental tumors C_LIO_LIDrug assays reveal therapeutic vulnerabilities and biomarkers C_LIO_LIPulmonary NET PDTOs are dependent on EGF C_LI

cancer biology↗

bioRxiv2022

Origin and dynamics of Mycobacterium tuberculosis subpopulations that predictably generate drug tolerance and resistance

Initial responses to tuberculosis treatment are poor predictors of final therapeutic outcomes in drug-susceptible disease suggesting that treatment success depends on features that are hidden within a small minority of the overall infecting Mycobacterium tuberculosis (Mtb) population. We developed a multi-transwell robotic system to perform numerous parallel cultures of genetically barcoded Mtb exposed to steady-state concentrations of rifampicin to uncover these difficult to eliminate minority populations. We found that tolerance repeatedly emerged from at least two subpopulations of barcoded cells, one that could not grow on solid agar media and a second that could form colonies, but whose kill curves diverged from the general bacterial population within 4 and 16 days of drug exposure, respectively. These tolerant subpopulations reproducibly passed through a phase characterized by multiple unfixed resistance mutations followed by emergent drug resistance in some cultures. Barcodes associated with drug resistance identified an especially privileged subpopulation that was rarely eliminated despite 20 days of drug treatment even in cultures that did not contain any drug resistant mutants. The association of this evolutionary scenario with a defined subset of barcodes across multiple independent cultures suggested a transiently heritable phenotype, and indeed glpK phase variation mutants were associated with up to 16 % of the resistant cultures. Drug tolerance and resistance were eliminated in{Delta} ruvA mutant consistent with the importance of bacterial stress responses. This work provides a window into the origin and dynamics of bacterial drug tolerant subpopulations whose elimination may be critical to developing rapid and resistance free cures. ImportanceTuberculosis is unusual among bacterial diseases in that treatments which can rapidly resolve symptoms do not predictably lead to a durable cure unless treatment is continued for months after all clinical and microbiological signs of disease have been eradicated. Using a novel steady-state antibiotic exposure system combined with chromosomal barcoding, we identified small hidden Mycobacterium tuberculosis subpopulations that repeatedly enter into a state of drug tolerance with a predisposition to develop fixed drug resistance after first developing a cloud of unfixed resistance mutations. The existence of these difficult to eradicate subpopulations may explain the need for extended treatment regimen for tuberculosis. Their identification provides opportunities to test genetic and therapeutic approaches that may result in shorter and more effective TB treatments.

microbiology↗

bioRxiv2021

Quality-controlled R-loop meta-analysis reveals the characteristics of R-Loop consensus regions

R-loops are three-stranded nucleic acid structures formed from the hybridization of RNA and DNA during transcription. While the pathological consequences of R-loops have been well-studied to date, the locations, classes, and dynamics of physiological R-loops remain poorly understood. R-loop mapping studies provide insight into R-loop dynamics, but their findings are challenging to generalize. This is due to the narrow biological scope of individual studies, the limitations of each mapping modality, and, in some cases, poor data quality. In this study, we reprocessed 693 R-loop mapping datasets from a wide array of biological conditions and mapping modalities. From this data resource, we developed an accurate method for R-loop data quality control, and we reveal the extent of poor-quality data within previously published studies. We then identified a set of high-confidence R-loop mapping samples and used them to define consensus R-loop sites called "R-loop regions" (RL regions). In the process, we revealed the stark divergence between S9.6 and dRNH-based R-loop mapping methods and identified biologically meaningful subtypes of both constitutive and variable R-loops. Taken together, this work provides a much-needed method to assess R-loop data quality and reveals intriguing aspects of R-loop biology.

bioinformatics↗