Biology subjects

Bao, B.

Publications and source records attributed to Bao, B..

At least 19 recordsLinked to original sources

bioRxiv2026

A unique intestinal epithelial cell that expresses canonical lymphoid genes

Enterocytes and intestinal secretory cells harbor innate immunity. However, certain transcriptional programs -even beyond rearranged T- and B-cell receptor genes- are considered the exclusive purview of adaptive immune cells, including abundant and diverse intestinal intra-epithelial lymphocytes (IELs). Here, we identify and characterize columnar intestinal secretory cells that co-express classic epithelial and lymphoid genes, including Ptprc, Cd3e, Cd8a, and Gzma/b but not T- or B-cell receptors. These previously unrecognized epithelium-derived lymphoid-like cells (EDLCs) differ from any known resident cell type, including IELs or tuft cells. Rare in healthy mouse and human intestines, their numbers increase upon IL-25 stimulation or IL-10 receptor deficiency in vivo and upon IL-13 stimulation in vitro. EDLC chromatin is uniquely accessible at thousands of cis-elements, including lymphoid and epithelial loci. These findings reveal a transcriptionally and epigenetically distinct intestinal epithelial derivative that expresses prominent lymphoid character.

cell biology↗

bioRxiv2026

Delineating the Transcriptional and Phenotypic Impact from Biotherapeutic Glycoengineering

Glycosylation is critical to biopharmaceutical activity, stability, and pharmacokinetics. While production cells can be engineered to produce better protein glycoforms, glycans decorate thousands of host cell proteins, and it remains unclear how glycoengineering impacts the host cell. To decipher the cell response to glycoengineering, we studied a library of 166 glycoengineered CHO-K1 cell clones representing 54 different glycosyltransferase modifications. Through integrated analysis of glycomics, RNA-Seq, and phenotypic data, we discovered that glycoengineered mutants clustered into three distinct groups (wild-type-like, Moderate, and Substantial) based on their glycosylation patterns. Different glycosyltransferase families exhibited distinct phenotypic signatures: St3gal modifications increased growth rate and cell density, B4galt knockouts affected cell size, and Mgat knockouts enhanced cell viability. Notably, we found specific cellular reprogramming associated with each glycosyltransferase family, including alterations in energy metabolism, stress responses, and DNA repair mechanisms. These findings were validated in an independent set of 30 glycoengineered CHO-S cell lines, expressing a panel of 10 recombinant proteins. Our extensive analysis reveals phenotypic changes resulting from glycoengineering, identifies their molecular bases, and provides crucial insights for controlling glycosylation during therapeutic protein production.

synthetic biology↗

bioRxiv2026

Geoclimatic oscillations and ancient reciprocal adaptive introgression shape the evolutionary trajectories of threatened Coilia

Geoclimatic oscillations have repeatedly reshaped biodiversity, yet how ancient lineages persist and diversify through profound environmental upheavals remains a central question in evolutionary genomics. By integrating chromosome-level genome assemblies, population genomics, and transcriptomic data across the Coilia species complex, we reconstruct deep evolutionary histories of Coilia along the dynamic East Asian margins. Our analyses reveal divergences dating to the Miocene, coinciding with major tectonic and climatic transitions, including Tibetan Plateau uplift-linked reorganization of the Yangtze drainage toward the East China Sea, and the Middle Miocene Climatic Optimum. Despite long-term separation, we uncover evidence for ancient, reciprocal introgression among Coilia lineages, with gene flow confined to discrete genomic regions rather than distributed genome-wide. These introgressed regions are enriched for loci associated with immune function, vascular development, and osmoregulatory processes, and exhibit population genetic signatures consistent with positive selection. Temporal modelling indicates that this window of adaptive exchange preceded Pleistocene glacial intensification, after which gene flow became effectively restricted. Together, our results suggest that transiently permeable species boundaries historically acted as reservoirs of adaptive variation, shaping evolutionary trajectories and facilitating niche differentiation under repeated geoclimatic oscillations. This study highlights ancient adaptive introgression as an important evolutionary process contributing to the long-term persistence of threatened coastal fishes.

evolutionary biology↗

bioRxiv2026

Mast cell-driven adipocyte lipolysis promotes post-expansive VAT atrophy and ectopic hepatic steatosis

Excessive visceral fat is an essential risk factor of metabolic dysfunction-associated steatotic liver disease (MASLD). Yet, the cellular and molecular mechanisms implicated in the correlation between visceral fat metabolism and hepatic steatosis remain poorly understood. Here we report post-expansive epididymal adipose tissue (EAT) atrophy and hepatic steatosis in mice fed a high-fat diet (HFD) primarily due to elevated adipocyte lipolysis in EAT. Mast cell (MC) accumulation in EAT represents a lipolysis-associated feature. Pharmacological stabilization of MCs suppresses EAT adipocyte lipolysis, and improves EAT atrophy and hepatic steatosis. MC-derived serotonin (5-HT) correlates with visceral adipose tissue (VAT) lipolysis in HFD-fed mice as well as in MASLD patients. Conditional deletion of 5-HT from MCs or its receptor HTR2b on adipocytes demonstrates that MC-derived 5-HT promotes adipocyte lipolysis, EAT atrophy, and hepatic steatosis by binding on adipocyte HTR2b. These results suggest that MCs and MC-derived 5-HT are potential therapeutic targets for obesity-associated MASLD.

cell biology↗

bioRxiv2025

PCNA Inhibition Enhances the Antitumor Activity of KRAS-Targeted Therapies in Pancreatic Cancer

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive disease with a dismal prognosis. More than 90% of PDAC tumors harbor KRAS mutations, and several KRAS inhibitors, such as off-state, on-state, mutation-specific, and pan-RAS inhibitors, are being tested in preclinical and clinical settings. However, the response to these inhibitors as single agents is less than optimal, indicating the need to identify novel combination therapies to improve treatment outcomes. Proliferating cell nuclear antigen (PCNA) is a ring-shaped clamp protein that regulates DNA replication, repair, and resolution of transcription-replication conflict, which are critical processes for pancreatic cancer survival. AOH1996 is a first-in-class, selective PCNA inhibitor in Phase I trials. Here, we found that AOH1996 treatment is efficacious in various PDAC models in vitro. PCNA and KRAS are predicted to be synthetic lethal partners, and RNA sequencing of AOH1996-treated PDAC cells reveals enrichment of MAPK and PI3K signaling pathways. Combination of AOH1996 with KRAS inhibitors demonstrates strong synergy across KRAS G12C and G12D mutant models. Treatment with a combination of AOH1996 and KRAS inhibitors induces cell cycle arrest and apoptosis in PDAC cells. Robust antitumor activity of AOH1996 in combination with RMC-6236 was observed in PDAC tumoroids. In vivo, the combination of AOH1996 with sotorasib or MRTX1133 reduced tumor growth rates compared to single-agent therapy, with no impact on mouse body weight. Residual tumor analysis showed sustained pERK and Myc inhibition in the combination arm. In conclusion, combination of AOH1996 with KRAS inhibitors is a promising therapeutic strategy for KRAS-driven PDAC, warranting further clinical investigation.

cancer biology↗

bioRxiv2025

Co-targeting KRAS and Exportin1 as an effective therapeutic strategy for KRASG12D mutant pancreatic ductal adenocarcinoma

BackgroundSeveral KRASG12D inhibitors (KRASG12Di) are under clinical evaluation for pancreatic ductal adenocarcinoma (PDAC). However, as seen with other first generation KRAS inhibitors, resistance may limit their long-term efficacy, necessitating combination strategies to enhance therapeutic outcomes. Exportin 1 (XPO1), a nuclear transport protein overexpressed in PDAC, represents a therapeutic vulnerability in KRAS-mutant cancers. Here, we demonstrate that the second-generation XPO1 inhibitor Eltanexor synergizes with MRTX1133 to enhance its efficacy in multiple PDAC models. MethodsWe generated KRASG12Di-resistant PDAC cells and assessed their response to Eltanexor. The antiproliferative effects of MRTX1133 and Eltanexor combinations were evaluated in 2D and 3D in vitro PDAC models. The in vivo efficacy of the combination was tested in KRASG12D-mutant human and murine PDAC xenograft and allograft models. ResultsEltanexor sensitized MRTX1133-resistant PDAC cells to growth inhibition. In both 2D and 3D culture models, the combination of Eltanexor and MRTX1133 significantly reduced cell viability. Mechanistically, the combination treatment suppressed key KRAS downstream signaling molecules, including p-ERK, mTOR, p-4EBP1, DUSP6, and cyclin D1. Kinome analysis further revealed reduced MAPK-related kinase activity. Combining subtherapeutic doses of Eltanexor and MRTX1133 resulted in significant tumor regression and prolonged survival in PDAC xenograft and immunocompetent orthotopic allograft models. Moreover, maintenance therapy with Eltanexor prevented tumor relapse, yielding a durable antitumor response. ConclusionThis study demonstrates that Eltanexor overcomes resistance to MRTX1133 and enhances its efficacy in PDAC. The combination regimen may provide a durable therapeutic response while reducing the required dose of KRASG12D inhibitors, potentially delaying resistance and improving patient outcomes. Statement of Translational RelevancePDAC remains one of the deadliest malignancies, with limited effective therapies and dismal survival rates. The emergence of KRASG12D-selective inhibitors, such as MRTX1133, marks a critical advance for nearly 40% of PDAC patients harboring this oncogenic driver. However, inevitable emergence of adaptive or acquired resistance to KRAS inhibitors remains a major barrier to achieving durable clinical benefit. This study uncovers XPO1 inhibition as a rational and synergistic strategy to augment the antitumor efficacy of MRTX1133. By enhancing KRASG12D inhibitor activity and potentially reducing the required therapeutic dose, this combination approach offers a novel means to delay or overcome resistance. These findings provide a strong preclinical rationale for clinical trials evaluating KRAS inhibitors in combination with XPO1 inhibitors and may significantly improve outcomes for a substantial subset of PDAC patients who currently lack effective targeted treatment options.

cancer biology↗

bioRxiv2025

A niche-dependent redox rheostat regulates epithelial stem cell fate in the distal colon

The niche environment surrounding intestinal stem cells (ISCs) varies along the length of intestine and provides key cues that regulate stem cell fate. Here, we investigated the role of cellular redox balance in colonic ISC function. We show that hypoxia and Wnt signaling synergize to restrict the reactive oxygen species (ROS) generating enzyme NADPH oxidase 1 (NOX1) to the crypt base in the distal colon. NOX1 function maintains a more oxidative cell state that licenses cell cycle entry, altering the balance of asymmetric stem cell self-renewal and directing lineage commitment. Mechanistically, cell redox state directs a self-reinforcing circuit that connects hypoxia inducible factor 1 (HIF1)-dependent signaling with regulation of the metabolic enzyme isocitrate dehydrogenase 1 (IDH1). Our studies show that cellular redox balance is a central and niche-dependent regulator of epithelial homeostasis and regeneration and provide a basis for understanding disease propensity in the distal large intestine. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=176 SRC="FIGDIR/small/634856v2_ufig1.gif" ALT="Figure 1"> View larger version (55K): org.highwire.dtl.DTLVardef@107798aorg.highwire.dtl.DTLVardef@1bc9013org.highwire.dtl.DTLVardef@96d56aorg.highwire.dtl.DTLVardef@95268d_HPS_FORMAT_FIGEXP M_FIG C_FIG HIGHLIGHTSO_LIThe balance of cycling intestinal stem cells (ISCs) versus committed epithelial cells in the uniquely hypoxic niche of the distal colon is regulated by NADPH oxidase 1 (NOX1) dependent H2O2 both at homeostasis and during regeneration. C_LIO_LIPhysiological increase in cellular H2O2 favors maintenance of glycolysis in ISCs for self-renewal through regulation of isocitrate dehydrogenase 1 activity. C_LIO_LIMaintenance of the increased cellular oxidative state stabilizes HIF1 through a re-enforcing metabolic circuit. C_LIO_LIA shift from a relatively oxidative to a reductive cell environment in distal colonic ISCs leads to decreased progression through the cell cycle and altered cell fate determination. C_LI

cell biology↗

bioRxiv2024

Tim-3 promotes viral infection by suppressing USP25-TRAF3-IRF7 signaling pathway

Tim-3, an immune checkpoint inhibitor, plays key roles in maintaining immune homeostasis and is involved in viral evasion. However, the precise role of Tim-3 in viral infection remains to be determined. USP25 is one of the deubiquitinating enzymes that serves to initiate antiviral immunity by deubiquitinating TRAF3 and triggering the anti-viral signaling pathway. Here we found that Tim-3 specific knockout in myeloid cells leads to enhanced anti-viral immunity in mice with VSV encephalitis by increasing type I interferon response. Mechanistically, Tim-3 inhibits the expression of USP25 via STAT1, interacts with USP25 but does not regulate its post-translational modification, and as a result, Tim-3 inhibits USP25 mediated deubiquitination of TRAF3, promotes k48-linked ubiquitination and degradation of TRAF3, and then inhibits the phosphorylation of IRF7, finally down-regulates the interferon response. These findings provide new insights into the function of Tim-3 in antiviral immunity and its related clinical significance.

immunology↗

bioRxiv2024

Adaptive loss of shortwave cone opsins during genomic evolution in cartilaginous fish

There is ample evidence of the loss of the shortwave visual protein gene in cartilaginous fish. However, the underlying logic and mechanisms by which organisms need to adapt to the environment behind the opsin loss are still unclear. Here, we report the assembly structure of the whole genome of Okamejei kenojei and Prionace glauca, identify TcMar difference between skates/rays and sharks/chimaeras, and analyze the distribution characteristics and intra group differentiation of opsin-related genes in cartilaginous fish. By establishing a zebrafish model with short wave opsin gene deletion, we find blue or violet light via short-wave sensitive opsins SWS1 or SWS2 can cause the photoreceptor layer thinning (the biomarker for age-related macular degeneration in human eye) through enactive the cell aging. The loss of sws is helpful for alleviating the short-wave light damage on the eye. Since the tapetum lucidum in the eye is found broad existence in various cartilaginous fish, a logical hypothesis was creatively proposed that the existence of tapetum lucidum on the retina of cartilaginous fish is interdependent with the loss of shortwave visual protein gene, providing a new perspective and explanatory path for further understanding of the evolution of visual genes in cartilaginous fish.

bioinformatics↗

bioRxiv2024

Protein structure, a genetic encoding for glycosylation

Unlike DNA, RNA, and protein biosynthesis, dogma describes glycosylation as primarily determined by intrinsic cellular limitations, such as glycosyltransferase expression and precursor availability. However, this cannot explain the commonly-observed differences between glycans on the same protein. By examining site-specific glycosylation on diverse human proteins, we detected associations between protein structure and glycan structure, broadly generalizable to human-expressed glycoproteins. Through structural analysis of site-specific glycosylation data, we found protein-sequence and structural features consistently correlated with specific glycan features. To quantify these relationships, we present a new amino acid substitution matrix describing "glycoimpact", i.e., the association of primary protein structure and glycosylation. High-glycoimpact amino acids co-evolve with glycosites, and glycoimpact is high when estimates of amino acid conservation and variant pathogenicity diverge. We report thousands of disease variants near glycosites with high-glycoimpact, including several with known links to aberrant glycosylation (e.g., Oculocutaneous Albinism, Jakob-Creutzfeldt disease, Gerstmann-Straussler-Scheinker, and Gauchers Disease). Finally, glycoimpact quantification is validated by studying oligomannose-complex glycan ratios on HIV ENV, differential sialylation on IgG3 Fc, differential glycosylation on SARS-CoV-2 Spike, and fucose-modulated function of a tuberculosis monoclonal antibody. Finally, to test the causality of protein-glycan associations, we created 5 glycoimpact-designed novel Rituximab variants, 4 of which substantially changed glycoprofiles as predicted. In all, we report that site-specific glycan biosynthesis is influenced by underlying protein structure, enabling glycan structure prediction and genetic sequence-guided glycoengineering.

genetics↗

bioRxiv2024

Decoding glycosylation potential from protein structure across human glycoproteins with a multi-view recurrent neural network

Glycosylation is described as a non-templated biosynthesis. Yet, the template-free premise is antithetical to the observation that different N-glycans are consistently placed at specific sites. It has been proposed that glycosite-proximal protein structures could constrain glycosylation and explain the observed microheterogeneity. Using site-specific glycosylation data, we trained a hybrid neural network to parse glycosites (recurrent neural network) and match them to feasible N-glycosylation events (graph neural network). From glycosite-flanking sequences, the algorithm predicts most human N-glycosylation events documented in the GlyConnect database and proposed structures corresponding to observed monosaccharide composition of the glycans at these sites. The algorithm also recapitulated glycosylation in Enhanced Aromatic Sequons, SARS-CoV-2 spike, and IgG3 variants, thus demonstrating the ability of the algorithm to predict both glycan structure and abundance. Thus, protein structure constrains glycosylation, and the neural network enables predictive in silico glycosylation of uncharacterized or novel protein sequences and genetic variants.

bioinformatics↗

bioRxiv2023

Bacterial Sphingolipids Exacerbate Colitis by Inhibiting ILC3-derived IL-22 Production.

Commensal bacteria of the Bacteroidetes phylum are the primary producers of sphingolipids in the gut lumen. These lipids serve dual roles as bacterial virulence factors and regulators of the host mucosal immune system, including regulatory T cells and invariant natural killer T cells (iNKT). Sphingolipid composition is significantly altered in fecal samples of patients with inflammatory bowel disease (IBD). However, the specific mechanisms by which bacterial sphingolipids modulate mucosal homeostasis and regulate intestinal inflammation remain unclear. In this study, we investigated the impact of bacterial sphingolipids on intestinal inflammation by mono-colonizing mice with Bacteroides fragilis strains that either express or lack sphingolipids during DSS-induced colitis. We discovered that B. fragilis sphingolipids exacerbate intestinal inflammation. Mice mono-colonized with B. fragilis lacking sphingolipids exhibited less severe DSS-induced colitis. This amelioration of colitis was associated with increased production of interleukin-22 (IL-22) by innate lymphoid cell type 3 (ILC3). Consistent with the inhibitory effect of sphingolipids on IL-22 production, mice colonized with B. fragilis lacking sphingolipids showed enhanced epithelial STAT3 activity, intestinal cell proliferation, and antimicrobial peptide production following DSS treatment compared to those colonized with B. fragilis producing sphingolipids. Additionally, colitis severity in mice colonized with B. fragilis lacking sphingolipids was exacerbated upon IL-22 blockade. Furthermore, our study reveals that bacterial sphingolipids restrict epithelial IL-18 production following DSS treatment and interfere with IL-22 production by a subset of ILC3 cells expressing both the interleukin-18 receptor (IL-18R) and major histocompatibility complex class II (MHC II). These findings indicate that B. fragilis-derived sphingolipids exacerbate mucosal inflammation by impeding epithelial IL-18 expression, resulting in compromised production of IL-22 by ILC3 cells. HighlightsO_LIB. fragilis-derived sphingolipids exacerbate DSS-induced colitis in mono-colonized C57BL/6 mice. C_LIO_LIB. fragilis-derived sphingolipids constrain ILC3-derived IL-22, leading to reduced colonic epithelial cell proliferation and compromised barrier function. C_LIO_LIB. fragilis-derived sphingolipids restrict epithelial NLRC4 inflammasome activation and IL-18 secretion. C_LIO_LIB. fragilis-derived sphingolipids modulate IL-22 production by IL18R+ MHC II+ ILC3s. C_LI

immunology↗

bioRxiv2023

Microbiome induced complement synthesized in the gut protects against enteric infections

Canonically, complement is a serum-based host defense system that protects against systemic microbial invasion. Little is known about the production and function of complement components on mucosal surfaces. Here we show gut complement component 3 (C3), central to complement function, is regulated by the composition of the microbiota in healthy humans and mice, leading to host-specific gut C3 levels. Stromal cells in intestinal lymphoid follicles (LFs) are the predominant source of intestinal C3. During enteric infection with Citrobacter rodentium or enterohemorrhagic Escherichia coli, luminal C3 levels increase significantly and are required for protection. C. rodentium is remarkably more invasive to the gut epithelium of C3-deficient mice than of wild-type mice. In the gut, C3-mediated phagocytosis of C. rodentium functions to clear pathogens. Our study reveals that variations in gut microbiota determine individuals intestinal mucosal C3 levels, dominantly produced by LF stromal cells, which directly correlate with protection against enteric infection. HighlightsO_LIGut complement component 3 (C3) is induced by the microbiome in healthy humans and mice at a microbiota-specific level. C_LIO_LIGut stromal cells located in intestinal lymphoid follicles are a major source of luminal C3 C_LIO_LIDuring enteric infections with Citrobacter rodentium or enterohemorrhagic Escherichia coli, gut luminal C3 levels increase and are required for protection. C_LIO_LIC. rodentium is significantly more invasive of the gut epithelium in C3-deficient mice when compared to WT mice. C_LIO_LIIn the gut, C3-mediated opsonophagocytosis of C. rodentium functions to clear pathogens. C_LI

immunology↗

bioRxiv2022

IL-10 inhibits STAT1-dependent macrophage accumulation during microbiota-induced colitis

Loss of IL-10R function leads to severe early onset colitis and in murine models is associated with the accumulation of immature inflammatory colonic macrophages. We have shown that IL-10R-deficient colonic macrophages exhibit increased STAT1-dependent gene expression, suggesting that IL-10R-mediated inhibition of STAT1 signaling in newly recruited colonic macrophages might interfere with the development of an inflammatory phenotype. Indeed Stat1-/- mice exhibit defects in colonic macrophage accumulation following Helicobacter hepaticus infection and IL-10R blockade, and this was phenocopied in mice lacking IFNGR, an inducer of STAT1 activation. Radiation chimeras demonstrated that reduced accumulation of STAT1-deficient macrophages was based on a cell-intrinsic defect. Unexpectedly, mixed radiation chimeras generated with both WT and IL-10R-deficient bone marrow indicated that rather than directly interfering with STAT1 function, IL-10R prevents the generation of a cell extrinsic signal that promotes the accumulation of immature macrophages. These results define essential mechanisms controlling inflammatory macrophage accumulation in inflammatory bowel diseases. SummaryIntrinsic STAT1-function drives the accumulation of macrophages within the colon following the loss of IL-10R signaling. IL-10R prevents this STAT1-dependent process through a non-cell autonomous mechanism.

immunology↗

bioRxiv2022

Preparing glycomics data for robust statistical analysis with GlyCompareCT

SummaryGlycomics data are rapidly increasing in scale and diversity. Complexities in glycan biosynthesis (hierarchy, competition, and compartmentalization) make preprocessing critical to address resulting sparsity (many similar glycosylation profiles may share few common glycans) and non-independence (substrate-competition in glycan biosynthesis results in non-independence incompatible with many statistical methods). Here, we present GlyCompareCT, a portable command-line tool, to address these challenges thereby facilitating downstream analyses. Given glycan abundances, GlyCompareCT conducts substructure decomposition to quantify hidden biosynthetic intermediate abundance and relationships between measured glycans. Thus, GlyComparCT mitigates sparsity and makes interdependence explicit thereby increasing statistical power. Ultimately, GlyComparCT is a user-friendly implementation of substructure analysis designed to increase accessibility, interoperability, and scope and consistency in glycomics analysis. Availability and implementationSource code, test data, and instructions for GlyCompareCT v1.1.0 are available at: https://github.com/LewisLabUCSD/GlyCompareCT Supplementary informationhttps://github.com/LewisLabUCSD/GlyCompareCT/raw/main/Supplementary%20Material.pdf

bioinformatics↗

bioRxiv2022

Hyperglycemic O-GlcNAc transferase activity drives cancer stem cell induction in TNBC

Enhanced glucose metabolism is a feature of almost all cancers, but downstream functional effects of aberrant glucose flux are difficult to mechanistically determine. The objective of this study is to characterize a mechanism by which elevated glucose level drives a tumorigenic pathway in triple negative breast cancer (TNBC). We used chemical biology methods to track how a metabolite of glucose, N-acetylglucosamine (GlcNAc), is linked to the transcriptional regulatory protein tet-methylcytosine dioxygenase 1 (TET1) as an O-linked GlcNAc post translational modification (O-GlcNAc). In this work, we revealed that intracellular protein glycosylation by O-GlcNAc is driven by high glucose levels in TNBC models, including on TET1. A single enzyme, O-GlcNAc transferase (OGT), is responsible for catalyzing protein modification of O-GlcNAc. We showed that OGT activity is higher in TNBC cell lines compared to non-tumor breast cell lines and is associated with hyperglycemia. Furthermore, enhanced OGT activity activated a pathway for cancer stem-like cell (CSC) reprogramming in TNBC cells. In our model, O-GlcNAcylated TET1 upregulated expression of splicing factor TAR-DNA binding protein (TARDBP), which drives CSC induction as well as higher OGT levels. We show that this OGT-TET1-TARDBP axis "feeds-forward" in hyperglycemic conditions both in cell lines and diet-induced obese mice, which displayed higher blood glucose levels and tumor O-GlcNAc levels than lean littermates. This data converges on a novel pathway whereby hyperglycemia drives aberrant OGT activity, activating a pathway for CSC induction in TNBC. Our findings partially explain a key aspect of how obesity is associated with TNBC risk and negative outcomes. Statement of ImplicationThis work presents a novel mechanism to explain how obesity is a risk factor for triple-negative breast cancer via elevated sugar-transfer activity by O-GlcNAc transferase in hyperglycemia, leading to the induction of a cancer stem-like cells and revealing a targetable pathway in obesity-associated tumors.

cancer biology↗

bioRxiv2021

Transfer learning improves outcome predictions for ASD from gene expression in blood

BackgroundPredicting outcomes on human genetic studies is difficult because the number of variables (genes) is often much larger than the number of observations (human subject tissue samples). We investigated means for improving model performance on the types of under-constrained problems that are typical in human genetics, where the number of strongly correlated genes (features) may exceed 10,000, and the number of study participants (observations) may be limited to under 1,000. MethodsWe created train, validate and test datasets from 240 microarray observations from 127 subjects diagnosed with autism spectrum disorder (ASD) and 113 typically developing (TD) subjects. We trained a neural network model (a.k.a., the naive model) on 10,422 genes using the train dataset, composed of 70 ASD and 65 TD subjects, and we restricted the model to one, fully-connected hidden layer to minimize the number of trainable parameters, including a dropout layer to help prevent overfitting. We experimented with alternative network architectures and tuned the hyperparameters using the validate dataset, and performed a single, final evaluation using the holdout test dataset. Next, we trained a neural network model using the identical architecture and identical genes to predict tissue type in GTEx data. We transferred that learning by replacing the top layer of the GTEx model with a layer to predict ASD outcome and we retrained the new layer on the ASD dataset, again using the identical 10,422 genes. FindingsThe naive neural network model had AUROC=0.58 for the task of predicting ASD outcomes, which saw a statistically significant 7.8% improvement from transfer learning. InterpretationWe demonstrated that neural network learning could be transferred from models trained on large RNA-Seq gene expression to a model trained on a small, microarray gene expression dataset with clinical utility for mitigating over-training on small sample sizes. Incidentally, we built a highly accurate classifier of tissue type with which to perform the transfer learning. FundingThis work was supported in part by NIMH R01-MH110558 (E.C., N.E.L.) Author SummaryImage recognition and natural language processing have enjoyed great success in reusing the computational efforts and data sources to overcome the problem of over-training a neural network on a limited dataset. Other domains using deep learning, including genomics and clinical applications, have been slower to benefit from transfer learning. Here we demonstrate data preparation and modeling techniques that allow genomics researchers to take advantage of transfer learning in order to increase the utility of limited clinical datasets. We show that a non-pre-trained, naive model performance can be improved by 7.8% by transferring learning from a highly performant model trained on GTEx data to solve a similar problem.

bioinformatics↗

bioRxiv2020

Elucidating Human Milk Oligosaccharide biosynthetic genes through network-based multi-omics integration

AO_SCPLOWBSTRACTC_SCPLOWHuman Milk Oligosaccharides (HMOs) are abundant carbohydrates fundamental to infant health and development. Although these oligosaccharides were discovered more than half a century ago, their biosynthesis in the mammary gland remains largely uncharacterized. Here, we used a systems biology framework that integrated glycan and RNA expression data to construct an HMO biosynthetic network and predict glycosyltransferases involved. To accomplish this, we constructed models describing the most likely pathways for the synthesis of the oligosaccharides accounting for >95% of the HMO content in human milk. Through our models, we propose candidate genes for elongation, branching, fucosylation, and sialylation of HMOs. We further explored selected enzyme activities through kinetic assay and their co-regulation through transcription factor analysis. These results provide the molecular basis of HMO biosynthesis necessary to guide progress in HMO research and application with the ultimate goal of understanding and improving infant health and development. SO_SCPLOWIGNIFICANCEC_SCPLOWO_SCPCAP C_SCPCAPO_SCPLOWSTATEMENTC_SCPLOWWith the HMO biosynthesis network resolved, we can begin to connect genotypes with milk types and thereby connect clinical infant, child and even adult outcomes to specific HMOs and HMO modifications. Knowledge of these pathways can simplify the work of synthetic reproduction of these HMOs providing a roadmap for improving infant, child, and overall human health with the specific application of a newly limitless source of nutraceuticals for infants and people of all ages.

systems biology↗