Single-cell transcriptomics of human T cells reveals tissue and activation signatures in health and disease

Human T cells coordinate adaptive immunity in diverse anatomic compartments through production of cytokines and effector molecules, but it is unclear how tissue site influences T cell persistence and function. Here, we use single cell RNA-sequencing (scRNA-seq) to define the heterogeneity of human T cells isolated from lungs, lymph nodes, bone marrow and blood, and their functional responses following stimulation. Through analysis of >50,000 resting and activated T cells, we reveal tissue T cell signatures in mucosal and lymphoid sites, and lineage-specific activation states across all sites including distinct effector states for CD8+ T cells and an interferon-response state for CD4+ T cells. Comparing scRNA-seq profiles of tumor-associated T cells to our dataset reveals predominant activated CD8+ compared to CD4+ T cell states within multiple tumor types. Our results therefore establish a high dimensional reference map of human T cell activation in health for analyzing T cells in disease.

Cross-tissue immune cell analysis reveals tissue-specific features in humans

Despite their crucial role in health and disease, our knowledge of immune cells within human tissues remains limited. We surveyed the immune compartment of 16 tissues from 12 adult donors by single-cell RNA sequencing and VDJ sequencing generating a dataset of ~360,000 cells. To systematically resolve immune cell heterogeneity across tissues, we developed CellTypist, a machine learning tool for rapid and precise cell type annotation. Using this approach, combined with detailed curation, we determined the tissue distribution of finely phenotyped immune cell types, revealing hitherto unappreciated tissue-specific features and clonal architecture of T and B cells. Our multitissue approach lays the foundation for identifying highly resolved immune cell types by leveraging a common reference dataset, tissue-integrated expression analysis, and antigen receptor sequencing.

Single-Cell Genomics Reveals a Novel Cell State During Smooth Muscle Cell Phenotypic Switching and Potential Therapeutic Targets for Atherosclerosis in Mouse and Human

BACKGROUND

Smooth muscle cells (SMCs) play significant roles in atherosclerosis via phenotypic switching, a pathological process in which SMC dedifferentiation, migration, and transdifferentiation into other cell types. Yet how SMCs contribute to the pathophysiology of atherosclerosis remains elusive.

METHODS

To reveal the trajectories of SMC transdifferentiation during atherosclerosis and to identify molecular targets for disease therapy, we combined SMC fate mapping and single-cell RNA sequencing of both mouse and human atherosclerotic plaques. We also performed cell biology experiments on isolated SMC-derived cells, conducted integrative human genomics, and used pharmacological studies targeting SMC-derived cells both in vivo and in vitro.

RESULTS

We found that SMCs transitioned to an intermediate cell state during atherosclerosis, which was also found in human atherosclerotic plaques of carotid and coronary arteries. SMC-derived intermediate cells, termed "SEM" cells (stem cell, endothelial cell, monocyte), were multipotent and could differentiate into macrophage-like and fibrochondrocyte-like cells, as well as return toward the SMC phenotype. Retinoic acid (RA) signaling was identified as a regulator of SMC to SEM cell transition, and RA signaling was dysregulated in symptomatic human atherosclerosis. Human genomics revealed enrichment of genome-wide association study signals for coronary artery disease in RA signaling target gene loci and correlation between coronary artery disease risk alleles and repressed expression of these genes. Activation of RA signaling by all-trans RA, an anticancer drug for acute promyelocytic leukemia, blocked SMC transition to SEM cells, reduced atherosclerotic burden, and promoted fibrous cap stability.

CONCLUSIONS

Integration of cell-specific fate mapping, single-cell genomics, and human genetics adds novel insights into the complexity of SMC biology and reveals regulatory pathways for therapeutic targeting of SMC transitions in atherosclerotic cardiovascular disease.

Cardiac recovery via extended cell-free delivery of extracellular vesicles secreted by cardiomyocytes derived from induced pluripotent stem cells

The ability of extracellular vesicles (EVs) to regulate a broad range of cellular processes has recently been exploited for the treatment of diseases. For example, EVs secreted by stem cells injected into infarcted hearts can induce recovery through the delivery of stem-cell-specific miRNAs. However, the retention of the EVs and the therapeutic effects are short-lived. Here, we show that an engineered hydrogel patch capable of slowly releasing EVs secreted from cardiomyocytes derived from induced pluripotent stem (iPS) cells reduced arrhythmic burden, promoted ejection-fraction recovery, decreased cardiomyocyte apoptosis 24 hours after infarction, and reduced infarct size and cell hypertrophy 4 weeks post-infarction when implanted onto infarcted rat hearts. We also show that the EVs are enriched with cardiac-specific miRNAs known to modulate cardiomyocyte-specific processes. The extended delivery of EVs secreted from iPS-cell-derived cardiomyocytes into the heart may help understand heart recovery and treat heart injury.

Cooperative regulation of AJM-1 controls junctional integrity in Caenorhabditis elegans epithelia

The function of epithelial cell sheets depends on the integrity of specialized cell-cell junctions that connect neighbouring cells. We have characterized the novel coiled-coil protein AJM-1, which localizes to an apical junctional domain of Caenorhabditis elegans epithelia basal to the HMR-HMP (cadherin-catenin) complex. In the absence of AJM-1, the integrity of this domain is compromised. Proper AJM-1 localization requires LET-413 and DLG-1, homologues of the Drosophila tumour suppressors Scribble and Discs large, respectively. DLG-1 physically interacts with AJM-1 and is required for its normal apical distribution, and LET-413 mediates the rapid accumulation of both DLG-1 and AJM-1 in the apical domain. In the absence of both dlg-1 and let-413 function AJM-1 is almost completely lost from apical junctions in embryos, whereas HMP-1 (alpha-catenin) localization is only mildly affected. We conclude that LET-413 and DLG-1 cooperatively control AJM-1 localization and that AJM-1 controls the integrity of a distinct apical junctional domain in C. elegans.

Longitudinal profiling of respiratory and systemic immune responses reveals myeloid cell-driven lung inflammation in severe COVID-19

Immune response dynamics in coronavirus disease 2019 (COVID-19) and their severe manifestations have largely been studied in circulation. Here, we examined the relationship between immune processes in the respiratory tract and circulation through longitudinal phenotypic, transcriptomic, and cytokine profiling of paired airway and blood samples from patients with severe COVID-19 relative to heathy controls. In COVID-19 airways, T cells exhibited activated, tissue-resident, and protective profiles; higher T cell frequencies correlated with survival and younger age. Myeloid cells in COVID-19 airways featured hyperinflammatory signatures, and higher frequencies of these cells correlated with mortality and older age. In COVID-19 blood, aberrant CD163+ monocytes predominated over conventional monocytes, and were found in corresponding airway samples and in damaged alveoli. High levels of myeloid chemoattractants in airways suggest recruitment of these cells through a CCL2-CCR2 chemokine axis. Our findings provide insights into immune processes driving COVID-19 lung pathology with therapeutic implications for targeting inflammation in the respiratory tract.

Generation and persistence of human tissue-resident memory T cells in lung transplantation

Tissue-resident memory T cells (T_RM) maintain immunity in diverse sites as determined in mouse models, whereas their establishment and role in human tissues have been difficult to assess. Here, we investigated human lung T_RM generation, maintenance, and function in airway samples obtained longitudinally from human leukocyte antigen (HLA)-disparate lung transplant recipients, where donor and recipient T cells could be localized and tracked over time. Donor T cells persist specifically in the lungs (and not blood) of transplant recipients and express high levels of T_RM signature markers including CD69, CD103, and CD49a, whereas lung-infiltrating recipient T cells gradually acquire T_RM phenotypes over months in vivo. Single-cell transcriptome profiling of airway T cells reveals that donor T cells comprise two T_RM-like subsets with varying levels of expression of T_RM-associated genes, whereas recipient T cells comprised non-T_RM and similar T_RM-like subpopulations, suggesting de novo T_RM generation. Transplant recipients exhibiting higher frequencies of persisting donor T_RM experienced fewer adverse clinical events such as primary graft dysfunction and acute cellular rejection compared with recipients with low donor T_RM persistence, suggesting that monitoring T_RM dynamics could be clinically informative. Together, our results provide spatial and temporal insights into how human T_RM develop, function, persist, and affect tissue integrity within the complexities of lung transplantation.

Tumor restriction by type I collagen opposes tumor-promoting effects of cancer-associated fibroblasts

Cancer-associated fibroblasts (CAF) may exert tumor-promoting and tumor-suppressive functions, but the mechanisms underlying these opposing effects remain elusive. Here, we sought to understand these potentially opposing functions by interrogating functional relationships among CAF subtypes, their mediators, desmoplasia, and tumor growth in a wide range of tumor types metastasizing to the liver, the most common organ site for metastasis. Depletion of hepatic stellate cells (HSC), which represented the main source of CAF in mice and patients in our study, or depletion of all CAF decreased tumor growth and mortality in desmoplastic colorectal and pancreatic metastasis but not in nondesmoplastic metastatic tumors. Single-cell RNA-Seq in conjunction with CellPhoneDB ligand-receptor analysis, as well as studies in immune cell-depleted and HSC-selective knockout mice, uncovered direct CAF-tumor interactions as a tumor-promoting mechanism, mediated by myofibroblastic CAF-secreted (myCAF-secreted) hyaluronan and inflammatory CAF-secreted (iCAF-secreted) HGF. These effects were opposed by myCAF-expressed type I collagen, which suppressed tumor growth by mechanically restraining tumor spread, overriding its own stiffness-induced mechanosignals. In summary, mechanical restriction by type I collagen opposes the overall tumor-promoting effects of CAF, thus providing a mechanistic explanation for their dual functions in cancer. Therapeutic targeting of tumor-promoting CAF mediators while preserving type I collagen may convert CAF from tumor promoting to tumor restricting.

Single-Cell Analysis of Regional Differences in Adult V-SVZ Neural Stem Cell Lineages

The ventricular-subventricular zone (V-SVZ) harbors adult neural stem cells. V-SVZ neural stem cells exhibit features of astrocytes, have a regional identity, and depending on their location in the lateral or septal wall of the lateral ventricle, generate different types of neuronal and glial progeny. We performed large-scale single-cell RNA sequencing to provide a molecular atlas of cells from the lateral and septal adult V-SVZ of male and female mice. This revealed regional and sex differences among adult V-SVZ cells. We uncovered lineage potency bias at the single-cell level among lateral and septal wall astrocytes toward neurogenesis and oligodendrogenesis, respectively. Finally, we identified transcription factor co-expression modules marking key temporal steps in neurogenic and oligodendrocyte lineage progression. Our data suggest functionally important spatial diversity in neurogenesis and oligodendrogenesis in the adult brain and reveal molecular correlates of adult NSC dormancy and lineage specialization.

Mizrak et al. performed large-scale, single-cell RNA sequencing of the adult ventricular-subventricular zone neural stem cell niche. They identify regional differences between the lateral wall and septal wall, as well as sex differences in cell types and signaling pathways.

Single-cell transcriptome analysis of lineage diversity in high-grade glioma

BACKGROUND

Despite extensive molecular characterization, we lack a comprehensive understanding of lineage identity, differentiation, and proliferation in high-grade gliomas (HGGs).

METHODS

We sampled the cellular milieu of HGGs by profiling dissociated human surgical specimens with a high-density microwell system for massively parallel single-cell RNA-Seq. We analyzed the resulting profiles to identify subpopulations of both HGG and microenvironmental cells and applied graph-based methods to infer structural features of the malignantly transformed populations.

RESULTS

While HGG cells can resemble glia or even immature neurons and form branched lineage structures, mesenchymal transformation results in unstructured populations. Glioma cells in a subset of mesenchymal tumors lose their neural lineage identity, express inflammatory genes, and co-exist with marked myeloid infiltration, reminiscent of molecular interactions between glioma and immune cells established in animal models. Additionally, we discovered a tight coupling between lineage resemblance and proliferation among malignantly transformed cells. Glioma cells that resemble oligodendrocyte progenitors, which proliferate in the brain, are often found in the cell cycle. Conversely, glioma cells that resemble astrocytes, neuroblasts, and oligodendrocytes, which are non-proliferative in the brain, are generally non-cycling in tumors.

CONCLUSIONS

These studies reveal a relationship between cellular identity and proliferation in HGG and distinct population structures that reflects the extent of neural and non-neural lineage resemblance among malignantly transformed cells.

Human memory T cells with a naive phenotype accumulate with aging and respond to persistent viruses

The number of naive T cells decreases and susceptibility to new microbial infections increases with age. Here we describe a previously unknown subset of phenotypically naive human CD8(+) T cells that rapidly secreted multiple cytokines in response to persistent viral antigens but differed transcriptionally from memory and effector T cells. The frequency of these CD8(+) T cells, called 'memory T cells with a naive phenotype' (TMNP cells), increased with age and after severe acute infection and inversely correlated with the residual capacity of the immune system to respond to new infections with age. CD8(+) TMNP cells represent a potential new target for the immunotherapy of persistent infections and should be accounted for and subtracted from the naive pool if truly naive T cells are needed to respond to antigens.

Observation of Individual Microtubule Motor Steps in Living Cells with Endocytosed Quantum Dots

We report the observation of individual steps taken by motor proteins in living cells by following movements of endocytic vesicles that contain quantum dots (QDs) with a fast camera. The brightness and photostability of quantum dots allow us to record motor displacement traces with 300 micros time resolution and 1.5 nm spatial precision. We observed individual 8 nm steps in active transport toward both the microtubule plus- and minus-ends, the directions of kinesin and dynein movements, respectively. In addition, we clearly resolved abrupt 16 nm steps in the plus-end direction and often consecutive 16 nm and occasional 24 nm steps in minus-end directed movements. This work demonstrates the ability of the QD assay to probe the operation of motor proteins at the molecular level in living cells under physiological conditions.

The mitotic kinesin KIF11 is a driver of invasion, proliferation, and self-renewal in glioblastoma

The proliferative and invasive nature of malignant cancers drives lethality. In glioblastoma, these two processes are presumed mutually exclusive and hence termed "go or grow." We identified a molecular target that shuttles between these disparate cellular processes-the molecular motor KIF11. Inhibition of KIF11 with a highly specific small-molecule inhibitor stopped the growth of the more treatment-resistant glioblastoma tumor-initiating cells (TICs, or cancer stem cells) as well as non-TICs and impeded tumor initiation and self-renewal of the TIC population. Targeting KIF11 also hit the other arm of the "go or grow" cell fate decision by reducing glioma cell invasion. Administration of a KIF11 inhibitor to mice bearing orthotopic glioblastoma prolonged their survival. In its role as a shared molecular regulator of cell growth and motility across intratumoral heterogeneity, KIF11 is a compelling therapeutic target for glioblastoma.

An Automated Microwell Platform for Large-Scale Single Cell RNA-Seq

Recent developments have enabled rapid, inexpensive RNA sequencing of thousands of individual cells from a single specimen, raising the possibility of unbiased and comprehensive expression profiling from complex tissues. Microwell arrays are a particularly attractive microfluidic platform for single cell analysis due to their scalability, cell capture efficiency, and compatibility with imaging. We report an automated microwell array platform for single cell RNA-Seq with significantly improved performance over previous implementations. We demonstrate cell capture efficiencies of >50%, compatibility with commercially available barcoded mRNA capture beads, and parallel expression profiling from thousands of individual cells. We evaluate the level of cross-contamination in our platform by both tracking fluorescent cell lysate in sealed microwells and with a human-mouse mixed species RNA-Seq experiment. Finally, we apply our system to comprehensively assess heterogeneity in gene expression of patient-derived glioma neurospheres and uncover subpopulations similar to those observed in human glioma tissue.

Scalable microfluidics for single-cell RNA printing and sequencing

Many important biological questions demand single-cell transcriptomics on a large scale. Hence, new tools are urgently needed for efficient, inexpensive manipulation of RNA from individual cells. We report a simple platform for trapping single-cell lysates in sealed, picoliter microwells capable of printing RNA on glass or capturing RNA on beads. We then develop a scalable technology for genome-wide, single-cell RNA-Seq. Our device generates pooled libraries from hundreds of individual cells with consumable costs of $0.10–$0.20 per cell and includes five lanes for simultaneous experiments. We anticipate that this system will serve as a general platform for single-cell imaging and sequencing.

Diversity and divergence of the glioma-infiltrating T-cell receptor repertoire

Although immune signaling has emerged as a defining feature of the glioma microenvironment, how the underlying structure of the glioma-infiltrating T-cell population differs from that of the blood from which it originates has been difficult to measure directly in patients. High-throughput sequencing of T-cell receptor (TCR) repertoires (TCRseq) provides a population-wide statistical description of how T cells respond to disease. We have defined immunophenotypes of whole repertoires based on TCRseq of the α- and β-chains from glioma tissue, nonneoplastic brain tissue, and peripheral blood from patients. Using information theory, we partitioned the diversity of these TCR repertoires into that from the distribution of VJ cassette combinations and diversity due to VJ-independent factors, such as selection due to antigen binding. Tumor-infiltrating lymphocytes (TILs) possessed higher VJ-independent diversity than nonneoplastic tissue, stratifying patients according to tumor grade. We found that the VJ-independent components of tumor-associated repertoires diverge more from their corresponding peripheral repertoires than T-cell populations in nonneoplastic brain tissue, particularly for low-grade gliomas. Finally, we identified a "signature" set of TCRs whose use in peripheral blood is associated with patients exhibiting low TIL divergence and is depleted in patients with highly divergent TIL repertoires. This signature is detectable in peripheral blood, and therefore accessible noninvasively. We anticipate that these immunophenotypes will be foundational to monitoring and predicting response to antiglioma vaccines and immunotherapy.

PLATE-Seq for genome-wide regulatory network analysis of high-throughput screens

Pharmacological and functional genomic screens play an essential role in the discovery and characterization of therapeutic targets and associated pharmacological inhibitors. Although these screens affect thousands of gene products, the typical readout is based on low complexity rather than genome-wide assays. To address this limitation, we introduce pooled library amplification for transcriptome expression (PLATE-Seq), a low-cost, genome-wide mRNA profiling methodology specifically designed to complement high-throughput screening assays. Introduction of sample-specific barcodes during reverse transcription supports pooled library construction and low-depth sequencing that is 10- to 20-fold less expensive than conventional RNA-Seq. The use of network-based algorithms to infer protein activity from PLATE-Seq data results in comparable reproducibility to 30 M read sequencing. Indeed, PLATE-Seq reproducibility compares favorably to other large-scale perturbational profiling studies such as the connectivity map and library of integrated network-based cellular signatures.

Despite the importance of pharmacological and functional genomic screens the readouts are of low complexity. Here the authors introduce PLATE-Seq, a low-cost genome-wide mRNA profiling method to complement high-throughput screening.

Quantitative assessment of protein activity in orphan tissues and single cells using the metaVIPER algorithm

We and others have shown that transition and maintenance of biological states is controlled by master regulator proteins, which can be inferred by interrogating tissue-specific regulatory models (interactomes) with transcriptional signatures, using the VIPER algorithm. Yet, some tissues may lack molecular profiles necessary for interactome inference (orphan tissues), or, as for single cells isolated from heterogeneous samples, their tissue context may be undetermined. To address this problem, we introduce metaVIPER, an algorithm designed to assess protein activity in tissue-independent fashion by integrative analysis of multiple, non-tissue-matched interactomes. This assumes that transcriptional targets of each protein will be recapitulated by one or more available interactomes. We confirm the algorithm's value in assessing protein dysregulation induced by somatic mutations, as well as in assessing protein activity in orphan tissues and, most critically, in single cells, thus allowing transformation of noisy and potentially biased RNA-Seq signatures into reproducible protein-activity signatures.

Three-dimensional extracellular matrix textured biomaterials

Clinical and experimental investigations have reported that manufactured surface topographies have significant effects on cell adhesion and tissue integration. However, essentially all previously examined topographies bear little relation to cell adhesion substrates found in biological tissues. In vivo, many cells are adherent to extracellular matrices (ECM), which have an extremely complex 3-D topography in the micrometre to nanometre range. In addition, many studies indicate that micro- and nano-scale mechanical stresses generated by cell-matrix adhesion have significant effects on cellular phenotypic behaviour. In this report we describe methodology for the fabrication of topographic replicas of the subendothelial ECM topography with a biomedical polyurethane. Using three-dimensional high resolution scanning electron microscopy, accurate replication of subendothelial ECM topography from the macroscopic to the macromolecular scale is demonstrated. Bovine aortic endothelial cells cultured on the ECM replicas spread more rapidly and had a three-dimensional appearance and spread areas at confluence which appeared more like endothelial cells in native arteries, compared with cells cultured on untextured control surfaces. Since the fabrication process may be used with many different types of materials, including polymers of synthetic and biological origin, these biomimetic ECM-textured surfaces may find both research and clinical applications.

Single-cell characterization of macrophages in glioblastoma reveals MARCO as a mesenchymal pro-tumor marker

BACKGROUND

Macrophages are the most common infiltrating immune cells in gliomas and play a wide variety of pro-tumor and anti-tumor roles. However, the different subpopulations of macrophages and their effects on the tumor microenvironment remain poorly understood.

METHODS

We combined new and previously published single-cell RNA-seq data from 98,015 single cells from a total of 66 gliomas to profile 19,331 individual macrophages.

RESULTS

Unsupervised clustering revealed a pro-tumor subpopulation of bone marrow-derived macrophages characterized by the scavenger receptor MARCO, which is almost exclusively found in IDH1-wild-type glioblastomas. Previous studies have implicated MARCO as an unfavorable marker in melanoma and non-small cell lung cancer; here, we find that bulk MARCO expression is associated with worse prognosis and mesenchymal subtype. Furthermore, MARCO expression is significantly altered over the course of treatment with anti-PD1 checkpoint inhibitors in a response-dependent manner, which we validate with immunofluorescence imaging.

CONCLUSIONS

These findings illustrate a novel macrophage subpopulation that drives tumor progression in glioblastomas and suggest potential therapeutic targets to prevent their recruitment.

Probing dynein and kinesin stepping with mechanical manipulation in a living cell

We report a label-free assay for simultaneous optical manipulation and tracking of endogenous lipid droplets as actively transported cargoes in a living mammalian cell with sub-millisecond time resolution. Using an EM-CCD camera as a highly sensitive quadrant detector, we can detect steps of dynein- and kinesin-driven cargoes under known force loads. We can distinguish single and multiple motor-driven cargoes and show that the stall forces for inward and outward transported cargoes are similar. By combining the stall force observable with the ability to detect individual steps, we can characterize kinesin- and dynein-driven active transport in different force regimes.

Climbing fiber-Purkinje cell synaptic pathology in tremor and cerebellar degenerative diseases

Changes in climbing fiber-Purkinje cell (CF-PC) synaptic connections have been found in the essential tremor (ET) cerebellum, and these changes are correlated with tremor severity. Whether these postmortem changes are specific to ET remains to be investigated. We assessed CF-PC synaptic pathology in the postmortem cerebellum across a range of degenerative movement disorders [10 Parkinson's disease (PD) cases, 10 multiple system atrophy (MSA) cases, 10 spinocerebellar ataxia type 1 (SCA1) cases, and 20 ET cases] and 25 controls. We observed differences in terms of CF pathological features across these disorders. Specifically, PD cases and ET cases both had more CFs extending into the parallel fiber (PF) territory, but ET cases had more complex branching and increased length of CFs in the PF territory along with decreased CF synaptic density compared to PD cases. MSA cases and SCA1 cases had the most severely reduced CF synaptic density and a marked paucity of CFs extending into the PF territory. Furthermore, CFs in a subset of MSA cases formed collateral branches parallel to the PC layer, a feature not seen in other diagnostic groups. Using unsupervised cluster analysis, the cases and controls could all be categorized into four clusters based on the CF pathology and features of PC pathology, including counts of PCs and their axonal torpedoes. ET cases and PD cases co-segregated into two clusters, whereas SCA1 cases and MSA cases formed another cluster, separate from the control cluster. Interestingly, the presence of resting tremor seemed to be the clinical feature that separated the cases into the two ET-PD clusters. In conclusion, our study demonstrates that these degenerative movement disorders seem to differ with respect to the pattern of CF synaptic pathology they exhibit. It remains to be determined how these differences contribute to the clinical presentations of these diseases.

A MYC and RAS co-activation signature in localized prostate cancer drives bone metastasis and castration resistance

Understanding the intricacies of lethal prostate cancer poses specific challenges due to difficulties in accurate modeling of metastasis in vivo. Here we show that NPK EYFP mice (for Nkx3.1 CreERT2/+ ; Pten flox/flox ; Kras LSL-G12D/+ ; R26R-CAG-LSL-EYFP/+) develop prostate cancer with a high penetrance of metastasis to bone, thereby enabling detection and tracking of bone metastasis in vivo and ex vivo. Transcriptomic and whole-exome analyses of bone metastasis from these mice revealed distinct molecular profiles conserved between human and mouse and specific patterns of subclonal branching from the primary tumor. Integrating bulk and single-cell transcriptomic data from mouse and human datasets with functional studies in vivo unravels a unique MYC/RAS co-activation signature associated with prostate cancer metastasis. Finally, we identify a gene signature with prognostic value for time to metastasis and predictive of treatment response in human patients undergoing androgen receptor therapy across clinical cohorts, thus uncovering conserved mechanisms of metastasis with potential translational significance.