Lin Wang

Spatial omics technologies provide unprecedented opportunities to investigate how molecular programs are organized within tissues. However, existing analytical frameworks primarily focus on identifying cellular states and characterizing differentially expressed molecular features across these states. The relational spatial organization of molecular features and its dynamics across biological contexts therefore remain largely unexplored. Here we present Gene SPatial Optimal Transport (GeneSPOT), a gene-centric computational framework for characterizing the spatial relationships among genes and quantifying their reorganization across biological conditions. Although gene-centric in design, GeneSPOT generalizes to other molecular features. Across diverse applications, GeneSPOT resolved tissue structures, integrated paired and unpaired spatial multi-omics data, unveiled the spatial convergence of gene programs during embryonic development and identified aging-associated spatial reorganization among genes whose expression did not change significantly with age. GeneSPOT thus complements conventional cell-centric analysis, providing a framework for deciphering how spatial relationships among molecular features are reorganized across biological conditions and over time.

Most small-molecule features detected in untargeted metabolomics remain unannotated, representing the dark matter of the metabolome. Here, we introduce MetaTag, a biochemical-reaction-guided framework that generates high-probability novel metabolites by combining abundant endogenous metabolites with common enzymatic transformations derived from KEGG pathways. In this proof-of-concept implementation, we focused on N-acylation and N-alkylation reactions. We synthesized and characterized a physical library of 1068 compounds, creating a public MS/MS database (https://mtdb-website.vercel.app). Screening against 27 rat tissues matched 180 previously unreported metabolites with 6 at level 1 (confirmed by RT, formula, MS/MS) and 174 at level 2 (confirmed by RT, formula). These metabolites exhibited tissue-specific distributions and participated in the interorgan exchange during fasting. Functional interrogation of one such metabolite, N-propionyltaurine, demonstrated its role as a redox-sensitive regulator in brown adipocytes. This biochemistry-guided strategy provides a scalable platform for the discovery of previously uncharacterized bioactive metabolites.

Untargeted LC-MS metabolomics offers a broad view of the microbial metabolism. However, its application is hindered by two intertwined challenges: distinguishing true biological signals from chemical artifacts and quantifying nutrient partitioning under nutrient-competitive conditions. Here, we present TRACE, an integrated experimental and computational framework that dynamically calibrates mass and retention time tolerances from the data itself to construct isotope-informed peak networks, enabling rigorous discrimination of biological metabolites from artifacts. Across four LC-MS platforms, TRACE reveals that the proportion of high-confidence annotations fell from 2.94 to 1.48%, while the total features increased by 331% from lower- to higher-sensitivity instruments. TRACE also maps nutrient fates into metabolic pathways by detecting isotopic dilution in Saccharomyces cerevisiae cultured with 13C-glucose, 15N-ammonium, and other unlabeled nutrients. Specifically, labeling of glutathione, a linear assembly of three amino acids, accurately reflect direct incorporation from its constituent amino acids; NAD+, whose biosynthesis proceeds through concurrent salvage and de novo pathways, revealed how adenine, tryptophan, and glutamine shaped its final isotopologue pattern. By converting untargeted LC-MS data into functional maps of nutrient flow, TRACE establishes a system-level approach to interrogate microbial metabolism under physiologically relevant competitive conditions.

Mass spectrometry imaging (MSI) has emerged as a powerful analytical tool for visualizing the spatial distribution of endogenous molecules in complex biological tissues. However, achieving authentic and reproducible metabolic maps remains challenging due to pitfalls in sample preparation and instrumental limitations that are often overlooked. Herein, based on practical evaluation, we delineate four critical but frequently overlooked pitfalls in MSI workflows: metabolic quenching, matrix effects, mass resolving power, and quadrupole isolation width. We demonstrate that thaw-mounting during traditional tissue preparation reactivates metabolic enzymes, which can be mitigated by rapid heat stabilization above 100 °C. Region-specific matrix effects causing uneven ionization could be corrected by normalization against isotope-labeled internal standards. Using incremental resolving power, we reveal that a single nominal m/z can represent multiple isobaric species with distinct spatial patterns, underscoring the necessity of ultrahigh resolution. Furthermore, we introduce a spatial correlation analysis that deconvolutes chimeric MS/MS spectra from coisolated precursors, enabling correct assignment of isobaric metabolites (e.g., choline and aminobutyric acid). Collectively, this study identifies four pitfalls and provides actionable strategies to overcome them, offering a practical framework for more reliable spatial metabolomics in preclinical and clinical research.

Pancreatic ductal adenocarcinoma (PDAC) exhibits hyperactive mitochondrial metabolism, yet how this rewiring spatially restricts the availability of metabolites for oncogenic signaling and drives systemic metabolic dysregulation in PDAC remains unknown. Here, we identify enhanced mitochondrial α-ketoglutarate (α-KG) sequestration as a key metabolic vulnerability in PDAC. Using multi-omics, preclinical models, and clinical correlation analyses, we identified elevated mitochondrial metabolic gene expression in PDAC. Moreover, higher expression of dihydrolipoamide succinyltransferase (DLST) correlates with poorer PDAC prognosis, suggesting the role of mitochondrial α-KG sequestration in PDAC progression. Targeting mitochondrial respiration with the complex I inhibitor carboxyamidotriazole orotate (CTO) redirected α-KG flux from mitochondrial sequestration, and increased α-KG-dependent m6A demethylation of MYC mRNA and HIF-1α hydroxylation. Combining CTO or α-KG dehydrogenase complex inhibitor devimistat with an α-KG analog (dimethyl α-KG) amplified c-Myc/HIF-1α suppression. Consequently, prolonged CTO exposure downregulated multiple metabolic pathways (glycolysis, pentose phosphate pathway, fatty acid synthesis) regulated by c-Myc/HIF-1α, and significantly delayed PDAC progression in vitro and in vivo. Our work first identify a novel mechanism whereby mitochondrial metabolism drives systemic metabolic dysregulation in PDAC through the sequestration of α-KG, and establishes “redirecting α-KG flux from mitochondrial sequestration” as a strategy to disable PDAC’s metabolic adaptability. The orotate salt form of carboxyamidotriazole effectively disrupts mitochondrial α-KG sequestration to suppresses PDAC growth at a dose equivalent to the clinically tested level.

Aging involves widespread metabolic dysregulation, including a decline in total nicotinamide adenine dinucleotide (NAD) levels. While NAD precursor supplementation elevates total NAD levels, it does not reveal tissue-specific effects of an altered NADH [reduced form of NAD ⁺ (oxidized NAD)]/NAD ⁺ balance. To address this, we generated transgenic Drosophila expressing the genetically encoded xenotopic enzyme Lb NOX, which converts NADH to NAD ⁺ . Lb NOX expression modulated both NAD(H) and NADP(H) (reduced form of NAD phosphate) metabolites in a sex-dependent manner and rescued neuronal cell death induced by mutant αB-crystallin–associated reductive stress. We demonstrate that tissue-specific targeting of redox NAD metabolism shows distinct outcomes: Muscle-specific Lb NOX expression confers stronger protection against paraquat-induced oxidative stress than whole-body expression, emphasizing tissue-dependent redox sensitivity. Notably, Lb NOX expression in nonneuronal tissues restored youthful sleep patterns in aged flies. Together, these findings establish Lb NOX as an efficient xenotopic tool for in vivo redox manipulation and reveal tissue- and sex-specific NAD(P)H mechanisms underlying aging, stress resilience, and sleep regulation, providing a framework for NAD-based interventions in aging.

Spatially resolved metabolomics plays a critical role in unraveling tissue-specific metabolic complexities. Despite its significance, this profound technology generates thousands of features, yet accurate annotation significantly lags behind that of LC-MS-based approaches. To bridge this gap, we introduce SMART, an open-source platform designed for precise formula assignment in mass spectrometry imaging. SMART constructs a KnownSet database containing 2.8 million formulas linked by ChemEdges derived from repositories such as HMDB, ChEMBL, PubChem, and BioEdges from KEGG biological reactant pairs. Using a multiple linear regression model, SMART extracts formula networks associated with the m/z of interest and scores potential candidates based on several criteria, including linked formulas, database existence score, ChemEdges/BioEdges, and ppm values. Benchmarking against reference data sets demonstrates that SMART achieves prediction accuracy rates of up to 95.8%. Applied to mass spectrometry imaging, SMART successfully annotated 986 formulas in developing mouse embryos. This robust platform enables systematic formula annotation within tissues, enhancing our understanding of metabolic heterogeneity.

Pulmonary fibrosis (PF) is a chronic and progressive lung disease with fatal consequences. The study of PF is challenging due to the complex mechanism involved, the need to understand the heterogeneity and spatial organization within lung tissues. In this study, we investigate the metabolic heterogeneity between two forms of lung fibrosis: idiopathic pulmonary fibrosis (IPF) and silicosis, using advanced spatially‐resolved metabolomics techniques. Employing high‐resolution mass spectrometry imaging, we spatially mapped and identified over 260 metabolites in lung tissue sections from mouse models of IPF and silicosis. Histological analysis confirmed fibrosis in both models, with distinct pathological features: alveolar destruction and collagen deposition in IPF, and nodule formation in silicosis. Metabolomic analysis revealed significant differences between IPF and silicosis in key metabolic pathways, including phospholipid metabolism, purine/pyrimidine metabolism, and the TCA cycle. Notably, phosphocholine was elevated in silicosis but reduced in IPF, while carnitine levels decreased in both conditions. Additionally, glycolytic activity was increased in both models, but TCA cycle intermediates showed opposing trends. These findings highlight the spatial metabolic heterogeneity of PF and suggest potential metabolic targets for therapeutic intervention. Further investigation into the regulatory mechanisms behind these metabolic shifts may open new avenues for fibrosis treatment.

Relapsed/refractory acute myeloid leukemia (rrAML) is a malignant blood cancer with an extremely poor prognosis, largely ascribed to the drug-resistant leukemia stem cells (LSCs). Most patients suffer from a risk of difficult-to-cure as well as severe systemic toxicity when receiving standard chemotherapies. As hyaluronic acid (HA) is a specific ligand of CD44 highly expressed by LSCs, we had HA self-assembled with cisplatin and daunorubicin to form a dual chemodrug nanogel (HA/Cis/Dau) to afford the targeted therapeutic interventions of rrAML. HA/Cis/Dau displayed an extra therapeutic function of inducing the granulocyte-monocyte differentiation in CD44 ⁺ rrAML cells, an rrAML mouse model, and primary blasts isolated from patients with AML. Unlike free drugs directly diffusing and killing rrAML cells, HA/Cis/Dau transported the drugs into lysosomes, causing lysosomal membrane permeabilization, ROS accumulation, and thus a metabolic reprogramming of the rrAML cells. Moreover, HA/Cis/Dau was featured with alleviated side effects, ease of preparation, and cost effectiveness, therefore holding great promises for the targeted treatment of rrAML.