Biology subjects

Chang, D.

Publications and source records attributed to Chang, D..

At least 19 recordsLinked to original sources

bioRxiv2026

Few-shot regulatory sequence design with a genomic language model

A central challenge in generative biology is translating knowledge learned from natural sequences and experiments into designs with desired functions. In-context learning (ICL) is an emergent capability of large language models that allows them to adapt to new tasks from a few examples without additional training, enabling these models to become general-purpose learners. Here, we show that the genome language model, Evo 2, exhibits generative ICL in biological sequence design. Prompting with just tens of examples conditions Evo 2 to generate novel and diverse regulatory DNA sequences without any additional training. Mechanistic analysis suggests that genomic ICL may originate from induction, similar to phenomena observed in natural language. Reinforcement learning leveraging sequence-to-function models as experiment-derived rewards further strengthens ICL. We first validate this approach by generating enhancers that drive cell-type-specific gene expression in three human cell lines. Extending this approach in vivo, enhancers designed using human regulatory element data selectively label layer 5 extratelencephalic neurons in the mouse cortex with high specificity and stronger activity than existing enhancers. Our study provides evidence of emergent properties in genomic sequence models and introduces a promising approach for improving functional conditioning in generative biology.

bioengineering↗

bioRxiv2026

Localis-REX: a general protocol for time-resolved spatiofunctional profiling in live cells and C. elegans

We describe Localis-REX, a method for identifying organ or subcellular locale-specific innate reactive electrophilic metabolite (REM) responsive proteins in live Caenorhabdtis elegans and cultured cells. Localis-REX mimics the natural spatiotemporal production of REMs by expressing HaloTag transiently or stably in a specific organ or tissue of live worms or in a designated organelle or subcellular compartment in cultured cells. Following treatment with a submicromolar photocaged REM and washout, light-triggered uncaging generates brief, localized REM pulses near Halo. REM responsive proteins within the vicinity compete to capture the liberated REM before endogenous clearance processes, including glutathione adduction and redox metabolism, enabling selective identification of the most REM-responsive targets under minimally perturned conditions. Localis-REX can be coupled with 1) proteomics to map chemically-actionable local proteomes or 2) live cell and in vivo functional readouts to define biological responses to locale-specific REM-accumulation, such as translation efficiency monitoring through genetic code expansion. The workflow requires 3-6 h, with downstream data acquisition and analysis requiring an additional 24-48 h, depending on the application 1) or 2. We further describe integration of in vivo Localis-REX with ultraID spatial profiling in live C. elegans, enabling expression normalization across profiled locales. Mechanistic validation of newly-identified responder and REM pairs is facilitated using established T-REX and Z-REX target engagement and on-target REM signaling interrogation protocols. Notably, the workflow was initially established as part of a masters thesis project by a student co-author with no prior experience in C. elegans or cell culture research by the master student, highlighting its accessibility to well-trained junior researchers pursuing interdisciplinary studies.

systems biology↗

bioRxiv2026

GEMS (GWAS Effect Mechanisms via Sequence-to-Function): Unveiling cell type-specific mechanisms for GWAS variants

Genome-wide association studies (GWAS) have linked thousands of loci to common diseases, yet for non-coding variants, which constitute the vast majority of the signal, the cell types and states through which individual risk variants act remain largely unknown. Here, we develop GEMS (GWAS Effect Mechanisms via Sequence-to-Function), a framework that trains supervised classifiers on multimodal variant effect predictions (ATAC-seq and RNA-seq) from the sequence-to-function model Borzoi and applies SHAP-based attribution to associate individual GWAS variants with the cell types most relevant to their disease association. Applied to asthma, LDL cholesterol, multiple sclerosis, and type 2 diabetes, GEMS prioritizes known disease-relevant cell types concordant with stratified LD-score regression. At the variant level, GEMS associates a median of 58.7% of fine-mapped variants with at least one cell type via ATAC-seq and 65.9% via RNA-seq, extending cellular annotation beyond the chromatin accessibility peaks in the single-cell atlas used for model training. ATAC-seq variant-cell-type associations are significantly enriched in cis-regulatory elements from independent experimental studies not used in model training, and enrichment persists after excluding variants overlapping training atlas peaks. For multiple sclerosis, hierarchical biclustering of variant--cell-type associations partitions risk variants into immune and CNS modules driven by distinct transcription factor families, recapitulating the immune-mediated and neurodegenerative components of the disease and demonstrating that integrative analysis of sequence-to-function predictions can stratify phenotype-associated variants into mechanistically distinct, cell-type-specific regulatory programs.

genetics↗

bioRxiv2026

Tuning metaplasticity in the adult visual cortex using flickering light

Synaptic connections in the brain are refined by sensory experience during an early postnatal critical period, but by adulthood synaptic connectivity is resistant to further changes. A consequence of lost plasticity is limited recovery from brain injury, disease, and adverse sensory experience. Thus, there is great interest in treatments that can promote synaptic modifications in the adult brain. In a wide variety of contexts, it has been established that the qualities of synaptic plasticity are not fixed but rather vary depending on the recent history of cellular or synaptic activity1. This plasticity of plasticity, or metaplasticity2 explains why temporary manipulations of brain activity (e.g., by drugs3, transcranial stimulation4, or sensory deprivation5) can set the stage for subsequent, potentially therapeutic, long-lasting synaptic modifications6. Here we tested the hypothesis that plasticity in the adult mouse visual cortex is influenced by prior exposure to temporally modulated light and discovered that different flicker frequencies have qualitatively different effects. Exposure to 60 Hz stimulation increased microglia density, depleted perineuronal nets (PNNs), and restored ocular dominance plasticity in response to brief monocular deprivation (MD). Exposure to 40 Hz flicker also enabled ocular dominance plasticity, but it did so in a distinct way and without PNN remodeling. A key distinction is that unlike 60 Hz flicker, which enabled depression of synaptic strength by MD, 40 Hz flicker promoted synaptic strengthening. Indeed, we found that 40 Hz flicker primed a rapid and robust recovery from the effects of long-term MD that failed to occur after 60 Hz flicker. Thus, metaplasticity can be non-invasively "tuned" by light flickering at different frequencies to encourage different forms of synaptic plasticity in the cerebral cortex, including modifications that enable recovery of function.

neuroscience↗

bioRxiv2026

Characterizing cellular subpopulations critical to treatment response in autoimmune diseases

Single-cell RNA sequencing provides a powerful approach for characterizing cell types, states, and lineages within heterogeneous tissues. However, identifying cell subpopulations that drive phenotypes, particularly treatment responses, remains a challenge. In this study, we performed comprehensive analyses to identify treatment response-associated cell subpopulations in autoimmune diseases by mapping bulk response information onto single-cell data. We integrated single-cell and bulk biopsy data from 314 responders and 619 non-responders treated with six therapeutics targeting tumor necrosis factor (TNF), integrin, or interleukin pathways in inflammatory bowel diseases (IBD) and psoriasis (PsO). Our analyses captured 128,428 interactions among 3,617 differentially expressed genes (DEGs), 852 pathways, and nine cell types spanning immune, stromal, and epithelial compartments. The importance of epithelial barrier integrity and enterocyte-mediated permeability in responses and inflammatory signaling in macrophages in non-responses in all tested therapies were highlighted by the presence of shared DEGs and pathways in both Crohns disease (CD) and ulcerative colitis (UC). In PsO, keratinocytes drove non-response to integrin-targeting therapies via disrupted adhesion, migration, and sustained epidermal inflammation. Additionally we introduce SCTRAD (https://immbioinfoabbv.shinyapps.io/SCTRAD/), an online web-based platform that allows exploration of mechanisms that influence the heterogeneity of cellular responses through DEGs and pathways at the single-cell level, as well as analysis of the cell-type-specific drug-related gene network. Our findings provide a comprehensive analysis of the role of cellular heterogeneity in treatment outcomes and for advancing precision medicine strategies in autoimmune diseases.

immunology↗

bioRxiv2025

Self-supervised AI reveals a lethal discohesive phenotype in lung adenocarcinoma

Applications of artificial intelligence (AI) to histopathology are now common, but most require supervision which inherently limits their scope. By using self-supervised learning (SSL), we discover and quantify the full range of histopathological appearances in a disease, and associate them with clinicopathological ground truths such as prognosis. We used this approach to discover under-appreciated morphologies of lung adenocarcinoma (LUAD), using a highly characterised resected tumour cohort of over 4000 slides from over 1000 patients. By constructing an authoritative lexicon of recurrent LUAD appearances, we ab initio discovered several stromal morphologies strongly predictive of outcome. With multimodal data integration and external dataset validation, we propose that epithelial discohesion is lethal, but only in the context of immunologically cold stroma. Both these morphological features are independent of current prognostic schema. Crucially, we describe these features in the context of real-world diagnostic histopathology, giving them immediate clinical translatability. Statement of SignificanceHistopathology relies heavily on epithelial morphology, often neglecting the stroma. We use self-supervised AI to identify under-appreciated morphologies in LUAD linked to poor outcome and validate these observations in an external cohort, demonstrating the utility of self-supervised AI as a powerful biological discovery tool.

cancer biology↗

bioRxiv2025

Self-Supervised AI Reveals a Hidden Landscape of Prognostic Spatial Patterns in Multiplex Immunofluorescence Images

Modern spatial proteomic methods, such as multiplex immunofluorescence (mIF) imaging, offer a data-rich view of spatial biology in intact tissues. However, interpreting its complexity is a major bottleneck, limiting its potential for biological discovery and clinical translation. Current computational methods often rely on segmentation-based approaches that discard crucial morphological information and are limited to testing pre-defined hypotheses. Here, we introduce a self-supervised learning (SSL) framework that enables hypothesis-agnostic, context-aware discovery of biomarkers directly from mIF images. Our approach extracts rich feature representations that capture holistic architectural patterns, which integrate cellular morphology, marker interactions, and microenvironmental context without human supervision. Applying this framework to over 7,000 mIF tissue images from over 1,800 patients in two distinct cancer types, we demonstrate superior prognostic performance over conventional segmentation analyses. The method autonomously identified previously unknown and potentially clinically actionable biological patterns. In lung adenocarcinoma, these include a Ki67-mediated immune evasion phenotype, a sub-cellular pattern of GLB1 expression which aligns with low-grade EGFR-driven tumours, and distinct modes of tumour-immune interaction in PD-L1+ patients. We also find a regulatory T-cell-mediated immunosupressive environment promoting tumour budding in colorectal carcinoma. Our work establishes SSL as a powerful, scalable, and unbiased platform to decode tissue ecosystems while being fully explainable without pre-defined hypotheses. This paradigm shift transforms high-plex imaging from a hypothesistesting tool into a hypothesis-generating engine that can accelerate the discovery of next-generation spatial biomarkers.

cancer biology↗

bioRxiv2025

Mevalonate Biosynthesis is a Metabolic Vulnerability of Gemcitabine-resistant Pancreatic Cancer

Pancreatic ductal adenocarcinoma (PDAC) is a lethal malignancy with a devastating prognosis. Gemcitabine, a pyrimidine anti-metabolite, is a cornerstone in PDAC therapy. However, resistance remains a major hurdle in clinical care. Resistance can arise from microenvironmental metabolites or through direct metabolic reprogramming of pancreatic cancer cells. Here, we generated PDAC models of acquired gemcitabine resistance to determine the relationship between these mechanisms. We observed that physiological levels of exogenous pyrimidines have a diminished ability to impact gemcitabine response in PDAC cells with acquired resistance. This occurs as the metabolic reprogramming of PDAC cells in response to gemcitabine treatment forces a suppression of the pyrimidine salvage pathway. Importantly, this metabolic rewiring renders gemcitabine-resistant PDAC cells highly susceptible to inhibition of the rate limiting enzyme of the mevalonate biosynthesis pathway, 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR), using statins. Notably, statin treatment inhibits the growth of gemcitabine-resistant tumors in immunocompetent mouse models. Through metabolite rescue experiments, we identified geranylgeranyl pyrophosphate as the critical metabolite lost during statin treatment, resulting in reduced protein geranylation in PDAC cells. Finally, as downregulation of the HMGCR is gradually acquired during gemcitabine resistance, we observed that HMGCR expression predicts patient response to gemcitabine. Collectively, these data demonstrate that the mevalonate biosynthesis pathway represents a promising therapeutic target in gemcitabine resistance and may serve as a biomarker to stratify treatment selection in PDAC patients.

cancer biology↗

bioRxiv2025

Granulin+ macrophages promote lineage plasticity in prostate cancer through paracrine signaling loops

The role of tumor-infiltrating immune cells in driving phenotype switching remains unclear, despite the well-established association between lineage transition and drug resistance in prostate cancer. This study employed an integrated analysis of single-cell multiomics to investigate the dynamics of immune infiltration, transcriptional programs, and cell-cell communication in prostate cancer. Our results demonstrated that granulin (GRN) positive macrophages facilitated the transition from adenocarcinoma to a multilineage state with mesenchymal and stem-like traits by activating intra-tumoral NF-{kappa}B signaling. Subsequently, the multilineage clones induced macrophages to highly express granulin through the secretion of CSF1, forming a positive feedback cell communication loop. Next, we validated the biological function of granulin in mediating epithelial-mesenchymal transition in vitro. Additionally, organoids drug resistance assay demonstrated that granulin drove resistance to androgen receptor (AR)-targeted drug (enzalutamide). Moreover, pharmacologic blockade of the CSF-1/CSF-1R axis in TRAMP mouse models reduced the expression of GRN in macrophages and suppressed the formation of multilineage subclones in prostate malignant cells. Furthermore, multiplex immunofluorescence staining of tumor samples from TRAMP mouse models revealed the VIM lineages were spatially in close contact with macrophages. Meanwhile, Cytometry by Time-Of-Flight (CyTOF) analysis validated our findings at single-cell protein level in patients with castration-resistant prostate cancer. Besides, three distinct tumor-infiltrating subsets associated with disease relapse were identified, including DCN+ endothelial cells, CCL7+ fibroblasts, and IFIT1+ neutrophils. These results offer potential therapeutic targets to address lineage plasticity-driven resistance to AR-targeted therapy.

cancer biology↗

bioRxiv2025

A global genetic interaction map of a human cell reveals conserved principles of genetic networks

We generated a genome-scale, genetic interaction network from the analysis of more than 4 million double mutants in the haploid human cell line, HAP1. The network maps [~]90,000 genetic interactions, including thousands of extreme synthetic lethal and genetic suppression interactions. Genetic interaction profiles enabled assembly of a hierarchical model of cell function, including modules corresponding to protein complexes, pathways, biological processes, and cellular compartments. Comparative analyses showed that general principles of genetic networks are conserved from yeast to human cells. A genetic interaction network mapped in a single genetic background complements the DepMap gene co-essentiality network, recapitulating many of the same biological connections and also capturing unique functional information to reveal roles of uncharacterized genes and molecular determinants of specific cancer cell line genetic dependencies.

systems biology↗

bioRxiv2025

Discovering Broader Host Ranges and an IS-bound Prophage Class Through Long-Read Metagenomics

Gut bacteriophages profoundly impact microbial ecology and human health, yet they are greatly understudied. Using deep, long-read bulk metagenomic sequencing, a technique that overcomes fundamental limitations of short-read approaches, we tracked prophage integration dynamics in 12 longitudinal stool samples from six healthy individuals, spanning a two-year timescale. While most prophages remain stably integrated into their host over two years, we discover that [~]5% of phages are dynamically gained or lost from persistent bacterial hosts. Within the same sample, we find evidence of population heterogeneity in which identical bacterial hosts with and without a given integrated prophage coexist simultaneously. Furthermore, we demonstrate that phage induction, when detected, occurs predominantly at low levels (1-3x coverage compared to the host region). Interestingly, we identify multiple instances of integration of the same phage into bacteria of different taxonomic families, challenging the dogma that phage are specific to a host of a given species or strain. Lastly, we describe a new class of phages, which we name "IScream phages". These phages co-opt bacterial IS30 transposases to mediate their integration, representing a previously unrecognized form of phage domestication of selfish bacterial elements. Taken together, these findings illuminate fundamental aspects of phage-bacterial dynamics in the human gut microbiome and expand our understanding of the evolutionary mechanisms that drive horizontal gene transfer and microbial genome plasticity in this ecosystem.

microbiology↗

bioRxiv2025

A reference genome assembly of the alpine forage grass Elymus nutans

Elymus nutans Griseb. (Poaceae: Triticeae, 2n=6x=42) is a dominant perennial plant species in the Qinghai-Tibetan Plateau in China, which is an important forage resource in high-altitude and cold regions and is the most popular species used for high yield artificial grassland planting because of its rich nutritional value and good palatability for herbivorous ruminant animals. In this study, using advanced sequencing technology, we generated an allohexaploid reference genome for E. nutans, representing the three sets of chromosomes (subgenomes St, Y, and H). This is the first study to our knowledge to confirm the origin of the Y subgenome, which shares a close relationship with the V subgenome. We predict that E. nutans arose via two-step hybridization and that its speciation occurred [~]3.16 MYA. The whole genome in this study will update the Triticeae family genomics information and provide theoretical basis for the evolutionary of its related speicies.

plant biology↗

bioRxiv2025

Genome modeling and design across all domains of life with Evo 2

All of life encodes information with DNA. While tools for sequencing, synthesis, and editing of genomic code have transformed biological research, intelligently composing new biological systems would also require a deep understanding of the immense complexity encoded by genomes. We introduce Evo 2, a biological foundation model trained on 9.3 trillion DNA base pairs from a highly curated genomic atlas spanning all domains of life. We train Evo 2 with 7B and 40B parameters to have an unprecedented 1 million token context window with single-nucleotide resolution. Evo 2 learns from DNA sequence alone to accurately predict the functional impacts of genetic variation--from noncoding pathogenic mutations to clinically significant BRCA1 variants--without task-specific finetuning. Applying mechanistic interpretability analyses, we reveal that Evo 2 autonomously learns a breadth of biological features, including exon-intron boundaries, transcription factor binding sites, protein structural elements, and prophage genomic regions. Beyond its predictive capabilities, Evo 2 generates mitochondrial, prokaryotic, and eukaryotic sequences at genome scale with greater naturalness and coherence than previous methods. Guiding Evo 2 via inference-time search enables controllable generation of epigenomic structure, for which we demonstrate the first inference-time scaling results in biology. We make Evo 2 fully open, including model parameters, training code, inference code, and the OpenGenome2 dataset, to accelerate the exploration and design of biological complexity.

genomics↗

bioRxiv2025

Classification of skin transcriptome reveals two molecular subtypes in hidradenitis suppurativa

Background and AimsHidradenitis suppurativa (HS) is an understudied chronic inflammatory skin disease that is characterized by painful bumps and abscesses. Adalimumab and secukinumab are the only two approved biologics for the treatment of HS. Despite these advances in the treatment of disease, there remains significant unmet medical need, with many patients only achieving moderate improvements in disease, or continuing to experience flares. This suggests that there may be distinct patient subsets with unique molecule drivers of disease. The aim of this study is to identify and characterize the molecular subtypes of HS patients to further understand its heterogeneity. MethodsSix public datasets with a total of 100 skin lesional samples were integrated to identify molecular subtypes and build a 36-gene classifier, which was validated by three independent lesional skin datasets. Two out of six training datasets generated from patients treated with adalimumab were used to identify the relationship between subtypes and response status. ResultsTwo molecular subtypes were identified from the training datasets, in which subtype S1 had a higher response rate of adalimumab than subtype S2 in the two adalimumab treatment datasets. Subtype S1 was characterized by three gene modules associated with keratinization, development, and metabolism, and six cell types related to sebocytes, smooth muscle cells, endothelial cells, schwann cells, basal cells, and proliferating cells. Subtype S2 was associated with three modules related to immune response, wound healing, keratinization, and cell cycle, and three cell types related to T lymphocytes, dendritic cells, and fibroblasts. The two subtypes were replicated in three additional independent datasets. ConclusionsThis study discovered and validated two HS subtypes with different molecular mechanisms and drug response, which may aid interpretation of heterogeneous molecular and clinical information in HS patients.

immunology↗

bioRxiv2025

scCompare: a web app for single-cell RNA sequencing dataset comparisons across multiple auto-immune diseases

MotivationSingle-cell RNA sequencing (scRNA-seq) datasets have been widely used to identify the cell types and marker genes with pivotal roles in driving pathogenesis and progression of auto-immune diseases. Comparative analysis of different cell types or diseases across multiple scRNA-seq datasets can reveal the homogeneous and heterogeneous pathogenesis, and a comprehensive web-based comparison tool that could streamline this process is not yet available. ResultsWe introduce scCompare, a web-based platform for scRNA-seq comparisons in autoimmune diseases. scCompare includes 2,125 differential gene lists from 100 scRNA-seq datasets in 22 auto-immune diseases and a query system supporting 170 standardized keywords from four attributes (disease, cell type, tissue, and treatment). scCompare also provides three modules enabling several comparative analysis and visualization options. geneQuery supports comparisons of queried genes across differential gene lists identified from multiple scRNA-seq datasets. DEGEnricher performs cell-type-specific enrichment analysis across studies based on a user-input gene list. DEGCompare allows interactive comparisons of multiple differential gene lists of many studies and performs pathway enrichment analyses. Using two case studies as examples, we demonstrated that scCompare represents a unique platform for biologists to identify, compare and validate the pathogenesis at the single-cell levels among auto-immune diseases. Availability and implementationscCompare is freely available at https://sccompare.shinyapps.io/main/. The source code is available at https://github.com/abbviegrc/scCompare.

bioinformatics↗

bioRxiv2025

Estimating the effect of tissue- and blood-derived cell reference matrices on deconvolving bulk transcriptomic datasets

BackgroundCell deconvolution is a method used to characterize the composition of the mixed cell population in bulk transcriptomic datasets. Tissue- and blood-derived cell reference matrices (CRMs) are typically used, but their impact on deconvolution has yet to be evaluated. MethodsRecursive feature elimination and random forest methods were applied to build tissue- and blood-derived CRMs using single-cell RNA sequencing (scRNA-seq) datasets of inflammatory bowel disease (IBD) which comprises two subtypes (Crohns disease (CD) and ulcerative colitis (UC)). Combining with three published blood-derived CRMs (IRIS, LM22, and ImmunoStates), public bulk transcriptomes datasets and simulated datasets generated from scRNA-Seq were used to evaluate the deconvolution performance by goodness-of-fit and cell fractions correlation. Additionally, two infliximab-treated bulk datasets were used to compare the treatment-related cell types revealed by tissue- and blood-derived CRMs. Lung adenocarcinoma (LUAD) single-cell and TCGA bulk transcriptomic datasets were also used for evaluation. ResultsFor CD and UC tissue bulk datasets, tissue-derived CRMs showed better deconvolution goodness-of-fit scores compared to blood-derived CRMs. Meanwhile, tissue-derived CRMs represented more accurate cellular proportion estimates for most cell types, such as immune and stromal cells in tissue pseudobulk datasets. They also revealed more treatment-related cell types. Additionally, CRMs derived from the scRNA-seq datasets from CD and UC colon samples had similar performance. In contrast, CRMs derived from CD ileum scRNA-seq dataset had better performance in the ileum bulk datasets. All CRMs yielded consistent deconvolution performance when deconvolving blood bulk transcriptomics. The similar results have also been shown using LUAD datasets. ConclusionsOur results emphasize the importance of selecting appropriate CRMs for cell deconvolution, particularly in bulk tissue transcriptomes in immunology and oncology. Such considerations can be extended to encompass other disease implications. AbbVie Disclosure statementAll authors are current employees of AbbVie. The design, study conduct, and financial support for this research were provided by AbbVie. AbbVie participated in the interpretation of data, review, and approval of the publication. Key PointsO_LITissue-derived CRMs showed higher goodness-of-fit compared to blood-derived CRMs for deconvolving bulk tissue transcriptomics. C_LIO_LIAll CRMs yield consistent goodness-of-fit for deconvolving bulk blood transcriptomics. C_LIO_LITissue-derived CRMs represent more accurate cellular proportion estimates and reveal more treatment-related cell types, and the specific tissue type is relevant to the deconvolution performance. C_LI

bioinformatics↗

bioRxiv2025

A novel DNA repair protein, N-Myc downstream regulated gene 1 (NDRG1), links stromal tumour microenvironment to chemoresistance

In pancreatic ductal adenocarcinoma cancer (PDAC) drug resistance is a severe clinical problem and patients relapse within a few months after receiving the standard-of-care chemotherapy. One contributing factor to treatment resistance is the desmoplastic nature of PDAC; the tumours are surrounded by thick layers of stroma composing up to 90% of the tumour mass. This stroma, which is mostly comprised of extracellular matrix (ECM) proteins, is secreted by cancer-associated fibroblasts (CAFs) residing in the tumour microenvironment. However, the mechanistic basis by which the tumour stroma directly contributes to chemoresistance remains unclear. Here, we show that CAF-secreted ECM proteins induce chemoresistance by blunting chemotherapy-induced DNA damage. Mechanistically, we identify N-myc downstream regulated gene 1 (NDRG1) as a key protein required for stroma-induced chemoresistance that responds to signals from the ECM and adhesion receptors. We further show that NDRG1 is a novel DNA repair protein that physically interacts with replication forks, maintains DNA replication and functions to resolve stalled forks caused by chemotherapy. More specifically, NDRG1 reduces R-loops, RNA-DNA hybrids that are known to cause genomic instability. R-loops occur during replication-transcription conflicts in S-phase and after chemotherapy treatments, thus posing a major threat to normal replication fork homeostasis. We identify NDRG1 as highly expressed in PDAC tumours, and its high expression correlates with chemoresistance and poor disease-specific survival. Importantly, knock-out of NDRG1 or inhibition of its phosphorylation restores chemotherapy-induced DNA damage and resensitizes tumour cells to treatment. In conclusion, our data reveal an unexpected role for CAF-secreted ECM proteins in enhancing DNA repair via NDRG1, a novel DNA repair protein, directly linking tumour stroma to replication fork homeostasis and R-loop biology, with important therapeutic implications for restoring DNA damage response pathways in pancreatic cancer. Summary paragraphDrug resistance is a severe clinical problem in stroma-rich tumours, such as pancreatic ductal adenocarcinoma (PDAC), and patients often relapse within a few months on chemotherapy1-9. The stroma, comprised of extracellular matrix (ECM) proteins, is secreted by cancer-associated fibroblasts (CAFs) residing in the tumour microenvironment10-13. Prior work show that ECM proteins provide survival benefits to cancer cells14,15. However, the precise role of CAF-secreted ECM in resistance to DNA damaging chemotherapies remains poorly understood. Here, we link ECM proteins to chemoresistance by enhanced DNA damage repair (DDR). Mechanistically, we identify N-myc downstream-regulated gene 1 (NDRG1) as a key effector downstream of ECM and the integrin-Src-SGK1-signalling axis that mediates enhanced DDR. We show that NDRG1 loss, mutation of conserved His194, or inhibition of NDRG1 phosphorylation by SGK1 lead to replication fork stalling, increased R-loops, and higher transcription-replication conflicts, resulting in genomic instability and sensitivity to chemotherapies. Our analysis of PDAC patient cohorts16 found that high NDRG1 expression correlates with chemoresistance and poor patient survival. In conclusion, we uncover an unexpected role for CAF-secreted ECM proteins in promoting therapeutic resistance by enhancing DDR and establish NDRG1 as a novel DNA repair protein directly linking tumour stroma to DDR.

cancer biology↗

bioRxiv2024

Potential mechanisms and effects of melatonin-regulated Nrf2/HO-1 pathway on acute lung injury due to formaldehyde exposure

Acute lung injury is a topic of great interest in critical care medicine due to its high mortality rates. The lungs are the immediate target organ for formaldehyde inhalation damage. Lung damage and fibrosis are the most important outcomes of severe and acute lung disease and pose a serious threat to human health. Melatonin (MT), a natural bioactive compound with anti-inflammatory and antioxidant properties, However, it is not clear whether MT can prevent FA-induced acute lung injury (ALI). Therefore, in this study, we aimed to evaluate the protective effects of MT and the potential mechanisms against FA-induced ALI. An environmental exposure bin was used to inhale 3 mg{middle dot}m3 FA-induced ALI, which was given intraperitoneally with different doses of MT (5/10/20 mg/kg) after successful modeling. In addition, rats were treated with Nrf2 inhibitor (ML385) to validate the signaling pathway. Lung function was measured, histopathological/morphological changes in lung tissue were assessed, and inflammatory expression and oxidation levels in lung tissue were detected. We observed that MT greatly alleviated the lung dysfunction, pathological lung injury, pulmonary edema and inflammatory response after successful modeling of FA. In additional, MT played a role in modulating the Nrf2/HO-1 signaling pathway, which effectively inhibit oxidative stress caused by FA-induced lung tissue injure. Moreover, we found that activation of the NF-{kappa}B pathway is associated with inflammation caused by this injury. Overall, our data suggest that MT inhibits the expression of oxidative stress and inflammation in lung tissue through the institutional or Nrf2/HO-1 pathway, alleviating FA-induced ALI.

developmental biology↗