A fast and sensitive size-exclusion chromatography method for plasma extracellular vesicle proteomic analysis.

Extracellular vesicles (EVs) carry diverse biomolecules derived from their parental cells, making their components excellent biomarker candidates. However, purifying EVs is a major hurdle in biomarker discovery since current methods require large amounts of samples, are time-consuming and typically have poor reproducibility. Here we describe a simple, fast, and sensitive EV fractionation method using size exclusion chromatography (SEC) on a fast protein liquid chromatography (FPLC) system. Our method uses a Superose 6 Increase 5/150, which has a bed volume of 2.9 mL. The FPLC system and small column size enable reproducible separation of only 50 µL of human plasma in 15 minutes. To demonstrate the utility of our method, we used longitudinal samples from a group of individuals that underwent intense exercise. A total of 838 proteins were identified, of which, 261 were previously characterized as EV proteins, including classical markers, such as cluster of differentiation (CD)9 and CD81. Quantitative analysis showed low technical variability with correlation coefficients greater than 0.9 between replicates. The analysis captured differences in relevant EV-proteins involved in response to physical activity. Our method enables fast and sensitive fractionation of plasma EVs with low variability, which will facilitate biomarker studies in large clinical cohorts.

In-Depth Proteome Profiling of Small Extracellular Vesicles Isolated from Cancer Cell Lines and Patient Serum.

Extracellular vesicle (EV) secretion has been observed in many types of both normal and tumor cells. EVs contain a variety of distinctive cargoes, allowing tumor-derived serum proteins in EVs to act as a minimally invasive method for clinical monitoring. We have undertaken a comprehensive study of the protein content of the EVs from several cancer cell lines using direct data-independent analysis. Several thousand proteins were detected, including many classic EV markers such as CD9, CD81, CD63, TSG101, and Syndecan-1, among others. We detected many distinctive cancer-specific proteins, including several known markers used in cancer detection and monitoring. We further studied the protein content of EVs from patient serum for both normal controls and pancreatic cancer and hepatocellular carcinoma. The EVs for these studies have been isolated by various methods for comparison, including ultracentrifugation and CD9 immunoaffinity column. Typically, 500-1000 proteins were identified, where most of them overlapped with the EV proteins identified from the cell lines studied. We were able to identify many of the cell-line EV protein markers in the serum EVs, in addition to the large numbers of proteins specific to pancreatic and HCC cancers.

Long-read RNA-seq atlas of novel microglia isoforms elucidates disease-associated genetic regulation of splicing.

Microglia, the innate immune cells of the central nervous system, have been genetically implicated in multiple neurodegenerative diseases. We previously mapped the genetic regulation of gene expression and mRNA splicing in human microglia, identifying several loci where common genetic variants in microglia-specific regulatory elements explain disease risk loci identified by GWAS. However, identifying genetic effects on splicing has been challenging due to the use of short sequencing reads to identify causal isoforms. Here we present the isoform-centric microglia genomic atlas (isoMiGA) which leverages the power of long-read RNA-seq to identify 35,879 novel microglia isoforms. We show that the novel microglia isoforms are involved in stimulation response and brain region specificity. We then quantified the expression of both known and novel isoforms in a multi-ethnic meta-analysis of 555 human microglia short-read RNA-seq samples from 391 donors, the largest to date, and found associations with genetic risk loci in Alzheimer's disease and Parkinson's disease. We nominate several loci that may act through complex changes in isoform and splice site usage.

Validation of a Proteomic Signature of Lung Cancer Risk from Bronchial Specimens of Risk-Stratified Individuals.

A major challenge in lung cancer prevention and cure hinges on identifying the at-risk population that ultimately develops lung cancer. Previously, we reported proteomic alterations in the cytologically normal bronchial epithelial cells collected from the bronchial brushings of individuals at risk for lung cancer. The purpose of this study is to validate, in an independent cohort, a selected list of 55 candidate proteins associated with risk for lung cancer with sensitive targeted proteomics using selected reaction monitoring (SRM). Bronchial brushings collected from individuals at low and high risk for developing lung cancer as well as patients with lung cancer, from both a subset of the original cohort (batch 1: n = 10 per group) and an independent cohort of 149 individuals (batch 2: low risk (n = 32), high risk (n = 34), and lung cancer (n = 83)), were analyzed using multiplexed SRM assays. ALDH3A1 and AKR1B10 were found to be consistently overexpressed in the high-risk group in both batch 1 and batch 2 brushing specimens as well as in the biopsies of batch 1. Validation of highly discriminatory proteins and metabolic enzymes by SRM in a larger independent cohort supported their use to identify patients at high risk for developing lung cancer.

Robust collection and processing for label-free single voxel proteomics.

With advanced mass spectrometry (MS)-based proteomics, genome-scale proteome coverage can be achieved from bulk tissues. However, such bulk measurement lacks spatial resolution and obscures important tissue heterogeneity, which make it impossible for proteome mapping of tissue microenvironment. Here we report an integrated wet collection of single tissue voxel and Surfactant-assisted One-Pot voxel processing method termed wcSOP for robust label-free single voxel proteomics. wcSOP capitalizes on buffer droplet-assisted wet collection of single tissue voxel dissected by LCM into the PCR tube cap and MS-compatible surfactant-assisted one-pot voxel processing in the collection cap. This convenient method allows reproducible label-free quantification of ∼900 and ∼4,600 proteins for single voxel from fresh frozen human spleen tissue at 20 µm × 20 µm × 10 µm (close to single cells) and 200 µm × 200 µm × 10 µm (∼100 cells), respectively. 100s-1000s of protein signatures with differential expression levels were identified to be spatially resolved between spleen red and white pulp regions depending on the voxel size. Region-specific signaling pathways were enriched from single voxel proteomics data. Antibody-based CODEX imaging was used to validate label-free MS quantitation for single voxel analysis. The wcSOP-MS method paves the way for routine robust single voxel proteomics and spatial proteomics.

Computational ranking-assisted identification of Plexin-B2 in homotypic and heterotypic clustering of circulating tumor cells in breast cancer metastasis.

Metastasis is the cause of over 90% of all deaths associated with breast cancer, yet the strategies to predict cancer spreading based on primary tumor profiles and therefore prevent metastasis are egregiously limited. As rare precursor cells to metastasis, circulating tumor cells (CTCs) in multicellular clusters in the blood are 20-50 times more likely to produce viable metastasis than single CTCs. However, the molecular mechanisms underlying various CTC clusters, such as homotypic tumor cell clusters and heterotypic tumor-immune cell clusters, are yet to be fully elucidated. Combining machine learning-assisted computational ranking with experimental demonstration to assess cell adhesion candidates, we identified a transmembrane protein Plexin- B2 (PB2) as a new therapeutic target that drives the formation of both homotypic and heterotypic CTC clusters. High PB2 expression in human primary tumors predicts an unfavorable distant metastasis-free survival and is enriched in CTC clusters compared to single CTCs in advanced breast cancers. Loss of PB2 reduces formation of homotypic tumor cell clusters as well as heterotypic tumor-myeloid cell clusters in triple-negative breast cancer. Interactions between PB2 and its ligand Sema4C on tumor cells promote homotypic cluster formation, and PB2 binding with Sema4A on myeloid cells (monocytes) drives heterotypic CTC cluster formation, suggesting that metastasizing tumor cells hijack the PB2/Sema family axis to promote lung metastasis in breast cancer. Additionally, using a global proteomic analysis, we identified novel downstream effectors of the PB2 pathway associated with cancer stemness, cell cycling, and tumor cell clustering in breast cancer. Thus, PB2 is a novel therapeutic target for preventing new metastasis.

Phosphorylation barcodes direct biased chemokine signaling at CXCR3.

G protein-coupled receptor (GPCR)-biased agonism, selective activation of certain signaling pathways relative to others, is thought to be directed by differential GPCR phosphorylation "barcodes." At chemokine receptors, endogenous chemokines can act as "biased agonists", which may contribute to the limited success when pharmacologically targeting these receptors. Here, mass spectrometry-based global phosphoproteomics revealed that CXCR3 chemokines generate different phosphorylation barcodes associated with differential transducer activation. Chemokine stimulation resulted in distinct changes throughout the kinome in global phosphoproteomics studies. Mutation of CXCR3 phosphosites altered ß-arrestin 2 conformation in cellular assays and was consistent with conformational changes observed in molecular dynamics simulations. T cells expressing phosphorylation-deficient CXCR3 mutants resulted in agonist- and receptor-specific chemotactic profiles. Our results demonstrate that CXCR3 chemokines are non-redundant and act as biased agonists through differential encoding of phosphorylation barcodes, leading to distinct physiological processes.

Phosphorylation barcodes direct biased chemokine signaling at CXCR3.

G protein-coupled receptor (GPCR) biased agonism, the activation of some signaling pathways over others, is thought to largely be due to differential receptor phosphorylation, or "phosphorylation barcodes." At chemokine receptors, ligands act as "biased agonists" with complex signaling profiles, which contributes to the limited success in pharmacologically targeting these receptors. Here, mass spectrometry-based global phosphoproteomics revealed that CXCR3 chemokines generate different phosphorylation barcodes associated with differential transducer activation. Chemokine stimulation resulted in distinct changes throughout the kinome in global phosphoproteomic studies. Mutation of CXCR3 phosphosites altered ß-arrestin conformation in cellular assays and was confirmed by molecular dynamics simulations. T cells expressing phosphorylation-deficient CXCR3 mutants resulted in agonist- and receptor-specific chemotactic profiles. Our results demonstrate that CXCR3 chemokines are non-redundant and act as biased agonists through differential encoding of phosphorylation barcodes and lead to distinct physiological processes.

Use of Longitudinal Serum Analysis and Machine Learning to Develop a Classifier for Cancer Early Detection.

Early detection of solid tumors through a simple screening process, such as the proteomic analysis of biofluids, has the potential to significantly alter the management and outcomes of cancers. The application of advanced targeted proteomics measurements and data analysis strategies to uniformly collected serum or plasma samples would enable longitudinal studies of cancer risk, progression, and response to therapy that have the potential to significantly reduce cancer burden in general. In this article, we describe a generalizable workflow combining robust, multiplexed targeted proteomics measurements applied to longitudinal samples from the Department of Defense Serum Repository with a Random Forest machine learning method for developing and initially evaluating the performance of candidate biomarker panels for early detection of cancers. The effectiveness of this approach was demonstrated in a cohort of 175 head and neck squamous cell carcinoma patients. The outlined protocols include methods for sample preparation, instrument analysis, and data analysis and interpretation using this workflow.

Targeted Quantification of Protein Phosphorylation and Its Contributions towards Mathematical Modeling of Signaling Pathways.

Post-translational modifications (PTMs) are key regulatory mechanisms that can control protein function. Of these, phosphorylation is the most common and widely studied. Because of its importance in regulating cell signaling, precise and accurate measurements of protein phosphorylation across wide dynamic ranges are crucial to understanding how signaling pathways function. Although immunological assays are commonly used to detect phosphoproteins, their lack of sensitivity, specificity, and selectivity often make them unreliable for quantitative measurements of complex biological samples. Recent advances in Mass Spectrometry (MS)-based targeted proteomics have made it a more useful approach than immunoassays for studying the dynamics of protein phosphorylation. Selected reaction monitoring (SRM)-also known as multiple reaction monitoring (MRM)-and parallel reaction monitoring (PRM) can quantify relative and absolute abundances of protein phosphorylation in multiplexed fashions targeting specific pathways. In addition, the refinement of these tools by enrichment and fractionation strategies has improved measurement of phosphorylation of low-abundance proteins. The quantitative data generated are particularly useful for building and parameterizing mathematical models of complex phospho-signaling pathways. Potentially, these models can provide a framework for linking analytical measurements of clinical samples to better diagnosis and treatment of disease.

Targeted Mass Spectrometry Assays for Specific Quantification of Urinary proPSA Isoforms.

Prostate cancer (PCa) is the second leading cause of male cancer-related deaths in the United States. The pre-mature forms of prostate-specific antigen (PSA), proPSA, were shown to be associated with PCa. However, there is a technical challenge in the development of antibody-based immunoassays for specific recognition of each individual proPSA isoform. Herein, we report the development of highly specific, antibody-free, targeted mass spectrometry assays for simultaneous quantification of [-2], [-4], [-5], and [-7] proPSA isoforms in voided urine. The newly developed proPSA assays capitalize on Lys-C digestion to generate surrogate peptides with appropriate length (9-16 amino acids) along with long-gradient liquid chromatography separation. The assay utility of these isoform markers was evaluated in a cohort of 30 well-established clinical urine samples for distinguishing PCa patients from healthy controls. Under the 95% confidence interval, the combination of [-2] and [-4] proPSA isoforms yields the area under curve (AUC) of 0.86, and the AUC value for the combined all four isoforms was calculated to be 0.85. We have further verified [-2]proPSA, the dominant isoform, in an independent cohort of 34 clinical urine samples. Validation of proPSA isoforms in large-scale cohorts is needed to demonstrate their potential clinical utility.

A streamlined tandem tip-based workflow for sensitive nanoscale phosphoproteomics.

Effective phosphoproteome of nanoscale sample analysis remains a daunting task, primarily due to significant sample loss associated with non-specific surface adsorption during enrichment of low stoichiometric phosphopeptide. We develop a tandem tip phosphoproteomics sample preparation method that is capable of sample cleanup and enrichment without additional sample transfer, and its integration with our recently developed SOP (Surfactant-assisted One-Pot sample preparation) and iBASIL (improved Boosting to Amplify Signal with Isobaric Labeling) approaches provides a streamlined workflow enabling sensitive, high-throughput nanoscale phosphoproteome measurements. This approach significantly reduces both sample loss and processing time, allowing the identification of >3000 (>9500) phosphopeptides from 1 (10) µg of cell lysate using the label-free method without a spectral library. It also enables precise quantification of ~600 phosphopeptides from 100 sorted cells (single-cell level input for the enriched phosphopeptides) and ~700 phosphopeptides from human spleen tissue voxels with a spatial resolution of 200 µm (equivalent to ~100 cells) in a high-throughput manner. The new workflow opens avenues for phosphoproteome profiling of mass-limited samples at the low nanogram level.

Mass spectrometry-based targeted proteomics for analysis of protein mutations.

Cancers are caused by accumulated DNA mutations. This recognition of the central role of mutations in cancer and recent advances in next-generation sequencing, has initiated the massive screening of clinical samples and the identification of 1000s of cancer-associated gene mutations. However, proteomic analysis of the expressed mutation products lags far behind genomic (transcriptomic) analysis. With comprehensive global proteomics analysis, only a small percentage of single nucleotide variants detected by DNA and RNA sequencing have been observed as single amino acid variants due to current technical limitations. Proteomic analysis of mutations is important with the potential to advance cancer biomarker development and the discovery of new therapeutic targets for more effective disease treatment. Targeted proteomics using selected reaction monitoring (also known as multiple reaction monitoring) and parallel reaction monitoring, has emerged as a powerful tool with significant advantages over global proteomics for analysis of protein mutations in terms of detection sensitivity, quantitation accuracy and overall practicality (e.g., reliable identification and the scale of quantification). Herein we review recent advances in the targeted proteomics technology for enhancing detection sensitivity and multiplexing capability and highlight its broad biomedical applications for analysis of protein mutations in human bodily fluids, tissues, and cell lines. Furthermore, we review recent applications of top-down proteomics for analysis of protein mutations. Unlike the commonly used bottom-up proteomics which requires digestion of proteins into peptides, top-down proteomics directly analyzes intact proteins for more precise characterization of mutation isoforms. Finally, general perspectives on the potential of achieving both high sensitivity and high sample throughput for large-scale targeted detection and quantification of important protein mutations are discussed.

Understanding of bacterial lignin extracellular degradation mechanisms by Pseudomonas putida KT2440 via secretomic analysis.

BACKGROUND: Bacterial lignin degradation is believed to be primarily achieved by a secreted enzyme system. Effects of such extracellular enzyme systems on lignin structural changes and degradation pathways are still not clearly understood, which remains as a bottleneck in the bacterial lignin bioconversion process. RESULTS: This study investigated lignin degradation using an isolated secretome secreted by Pseudomonas putida KT2440 that grew on glucose as the only carbon source. Enzyme assays revealed that the secretome harbored oxidase and peroxidase/Mn2+-peroxidase capacity and reached the highest activity at 120 h of the fermentation time. The degradation rate of alkali lignin was found to be only 8.1% by oxidases, but increased to 14.5% with the activation of peroxidase/Mn2+-peroxidase. Gas chromatography-mass spectrometry (GC-MS) and two-dimensional 1H-13C heteronuclear single-quantum coherence (HSQC) NMR analysis revealed that the oxidases exhibited strong C-C bond (ß-ß, ß-5, and ß-1) cleavage. The activation of peroxidases enhanced lignin degradation by stimulating C-O bond (ß-O-4) cleavage, resulting in increased yields of aromatic monomers and dimers. Further mass spectrometry-based quantitative proteomics measurements comprehensively identified different groups of enzymes particularly oxidoreductases in P. putida secretome, including reductases, peroxidases, monooxygenases, dioxygenases, oxidases, and dehydrogenases, potentially contributed to the lignin degradation process. CONCLUSIONS: Overall, we discovered that bacterial extracellular degradation of alkali lignin to vanillin, vanillic acid, and other lignin-derived aromatics involved a series of oxidative cleavage, catalyzed by active DyP-type peroxidase, multicopper oxidase, and other accessory enzymes. These results will guide further metabolic engineering design to improve the efficiency of lignin bioconversion.

Machine learning-assisted elucidation of CD81-CD44 interactions in promoting cancer stemness and extracellular vesicle integrity.

Tumor-initiating cells with reprogramming plasticity or stem-progenitor cell properties (stemness) are thought to be essential for cancer development and metastatic regeneration in many cancers; however, elucidation of the underlying molecular network and pathways remains demanding. Combining machine learning and experimental investigation, here we report CD81, a tetraspanin transmembrane protein known to be enriched in extracellular vesicles (EVs), as a newly identified driver of breast cancer stemness and metastasis. Using protein structure modeling and interface prediction-guided mutagenesis, we demonstrate that membrane CD81 interacts with CD44 through their extracellular regions in promoting tumor cell cluster formation and lung metastasis of triple negative breast cancer (TNBC) in human and mouse models. In-depth global and phosphoproteomic analyses of tumor cells deficient with CD81 or CD44 unveils endocytosis-related pathway alterations, leading to further identification of a quality-keeping role of CD44 and CD81 in EV secretion as well as in EV-associated stemness-promoting function. CD81 is coexpressed along with CD44 in human circulating tumor cells (CTCs) and enriched in clustered CTCs that promote cancer stemness and metastasis, supporting the clinical significance of CD81 in association with patient outcomes. Our study highlights machine learning as a powerful tool in facilitating the molecular understanding of new molecular targets in regulating stemness and metastasis of TNBC.

An antibody-free platform for multiplexed, sensitive quantification of protein biomarkers in complex biomatrices.

Sensitive, multiplexed protein quantification remains challenging despite recent advancements in LC-MS assays for targeted protein biomarker quantification. High-sensitivity protein biomarker measurements usually require immuno-affinity enrichment of target protein; a process which is highly dependent on capture reagent and limited in capability to measure multiple analytes. Herein, we report a novel antibody-free platform, which measures multiple biomarkers from complex matrices employing a strategically optimized solid-phase extraction cleanup and orthogonal multidimensional LC-MS. Eight human protein biomarkers with different specifications were spiked into canine plasma as a model investigation system. The developed strategy achieved the desired sensitivity, robustness, and throughput via the following steps: (1) post digestion mixed-mode cation exchange-reverse phase SPE enrichment cleaned up the sample initially; (2) rapid, high-pH peptide fractionation further eliminated background components efficiently while selectively enriched signature peptides (SP) to provide sufficient sensitivity for multiple targets; and (3) trapping-micro-LC-MS analysis delivered high sensitivity comparable to a nano-LC-MS method but with much better robustness and throughput for the final analysis. Compared with a conventional LC-MS assay with direct protein digestion and limited clean-up, analysis with this antibody-free platform improved the LLOQ by 1-2 orders of magnitude for the eight protein biomarkers, reaching as low as 5 ng/mL in plasma, with feasible robustness and throughput. This platform was applied for the quantification of biomarkers of respiratory conditions in patients with various lung diseases, demonstrating real-world applicability.

Circulating ACE2-expressing extracellular vesicles block broad strains of SARS-CoV-2.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused the pandemic of the coronavirus induced disease 2019 (COVID-19) with evolving variants of concern. It remains urgent to identify novel approaches against broad strains of SARS-CoV-2, which infect host cells via the entry receptor angiotensin-converting enzyme 2 (ACE2). Herein, we report an increase in circulating extracellular vesicles (EVs) that express ACE2 (evACE2) in plasma of COVID-19 patients, which levels are associated with severe pathogenesis. Importantly, evACE2 isolated from human plasma or cells neutralizes SARS-CoV-2 infection by competing with cellular ACE2. Compared to vesicle-free recombinant human ACE2 (rhACE2), evACE2 shows a 135-fold higher potency in blocking the binding of the viral spike protein RBD, and a 60- to 80-fold higher efficacy in preventing infections by both pseudotyped and authentic SARS-CoV-2. Consistently, evACE2 protects the hACE2 transgenic mice from SARS-CoV-2-induced lung injury and mortality. Furthermore, evACE2 inhibits the infection of SARS-CoV-2 variants (α, ß, and δ) with equal or higher potency than for the wildtype strain, supporting a broad-spectrum antiviral mechanism of evACE2 for therapeutic development to block the infection of existing and future coronaviruses that use the ACE2 receptor.

Facile One-Pot Nanoproteomics for Label-Free Proteome Profiling of 50-1000 Mammalian Cells.

Recent advances in sample preparation enable label-free mass spectrometry (MS)-based proteome profiling of small numbers of mammalian cells. However, specific devices are often required to downscale sample processing volume from the standard 50-200 µL to sub-µL for effective nanoproteomics, which greatly impedes the implementation of current nanoproteomics methods by the proteomics research community. Herein, we report a facile one-pot nanoproteomics method termed SOPs-MS (surfactant-assisted one-pot sample processing at the standard volume coupled with MS) for convenient robust proteome profiling of 50-1000 mammalian cells. Building upon our recent development of SOPs-MS for label-free single-cell proteomics at a low µL volume, we have systematically evaluated its processing volume at 10-200 µL using 100 human cells. The processing volume of 50 µL that is in the range of volume for standard proteomics sample preparation has been selected for easy sample handling with a benchtop micropipette. SOPs-MS allows for reliable label-free quantification of ∼1200-2700 protein groups from 50 to 1000 MCF10A cells. When applied to small subpopulations of mouse colon crypt cells, SOPs-MS has revealed protein signatures between distinct subpopulation cells with identification of ∼1500-2500 protein groups for each subpopulation. SOPs-MS may pave the way for routine deep proteome profiling of small numbers of cells and low-input samples.

Antibody-Free Mass Spectrometry Identification of Vascular Integrity Markers in Major Trauma.

Antibody mediated strategies for protein biomarker detection are common, but may limit discovery. We hypothesized that the use of antibody-free proteomics is feasible for detecting protein biomarkers in plasma of patients sustaining major trauma. A subset of subjects with major trauma from a prospective observational trial were analyzed. Patients were assigned to one of four groups based on their presenting Abbreviated Injury Severity Score (AIS). Sensitive, antibody-free selective reaction monitoring (SRM) mass spectrometry (MS), with spiked-in isotopically labeled synthetic peptides, was used for targeted protein quantification of a panel of 10 prospective targets. An overall tiered sensitivity analytical approach was used for peptide detection and quantification based upon plasma immunoaffinity depletion and PRISM fractionation. Forty-four patients were included in the analysis, of which 82% were men with a mean age of 50 (±19) years. Half had isolated head injury (n = 22), with the remaining patients experiencing multiple injuries or polytrauma (n = 14), isolated body injury (n = 2), or minor injury (n = 6). Peptides from 3 proteins (vascular adhesion molecule 1 [VCAM1], intercellular adhesion molecule 1 [ICAM1], and matrix metalloproteinase 9 [MMP9]) were detected and quantified in non-depleted processed plasma. Peptides from 2 proteins (angiopoietin 2 [Ang2] and plasminogen activator inhibitor-1 [PAI1]) were detected and quantification in depleted plasma, whereas the remaining 5 of the 10 prospective targets were undetected. VCAM1 (p = 0.02) and MMP9 (p = 0.03) were significantly upregulated in in the major trauma groups (1-3) versus mild injury group (4), whereas the others were not. There were no differences in protein expression between patients with traumatic brain injury (TBI; groups 1 and 2) versus those without TBI (groups 3 and 4). We detected non-specific upregulation of proteins reflecting blood-brain barrier breakdown in severely injured patients, indicating label-free MS techniques are feasible and may be informative.

The AML microenvironment catalyzes a stepwise evolution to gilteritinib resistance.

Our study details the stepwise evolution of gilteritinib resistance in FLT3-mutated acute myeloid leukemia (AML). Early resistance is mediated by the bone marrow microenvironment, which protects residual leukemia cells. Over time, leukemia cells evolve intrinsic mechanisms of resistance, or late resistance. We mechanistically define both early and late resistance by integrating whole-exome sequencing, CRISPR-Cas9, metabolomics, proteomics, and pharmacologic approaches. Early resistant cells undergo metabolic reprogramming, grow more slowly, and are dependent upon Aurora kinase B (AURKB). Late resistant cells are characterized by expansion of pre-existing NRAS mutant subclones and continued metabolic reprogramming. Our model closely mirrors the timing and mutations of AML patients treated with gilteritinib. Pharmacological inhibition of AURKB resensitizes both early resistant cell cultures and primary leukemia cells from gilteritinib-treated AML patients. These findings support a combinatorial strategy to target early resistant AML cells with AURKB inhibitors and gilteritinib before the expansion of pre-existing resistance mutations occurs.