DELVE: feature selection for preserving biological trajectories in single-cell data.

Single-cell technologies can measure the expression of thousands of molecular features in individual cells undergoing dynamic biological processes. While examining cells along a computationally-ordered pseudotime trajectory can reveal how changes in gene or protein expression impact cell fate, identifying such dynamic features is challenging due to the inherent noise in single-cell data. Here, we present DELVE, an unsupervised feature selection method for identifying a representative subset of molecular features which robustly recapitulate cellular trajectories. In contrast to previous work, DELVE uses a bottom-up approach to mitigate the effects of confounding sources of variation, and instead models cell states from dynamic gene or protein modules based on core regulatory complexes. Using simulations, single-cell RNA sequencing, and iterative immunofluorescence imaging data in the context of cell cycle and cellular differentiation, we demonstrate how DELVE selects features that better define cell-types and cell-type transitions. DELVE is available as an open-source python package: https://github.com/jranek/delve .

Cell cycle plasticity underlies fractional resistance to palbociclib in ER+/HER2- breast tumor cells.

The CDK4/6 inhibitor palbociclib blocks cell cycle progression in Estrogen receptor-positive, human epidermal growth factor 2 receptor-negative (ER+/HER2-) breast tumor cells. Despite the drug's success in improving patient outcomes, a small percentage of tumor cells continues to divide in the presence of palbociclib-a phenomenon we refer to as fractional resistance. It is critical to understand the cellular mechanisms underlying fractional resistance because the precise percentage of resistant cells in patient tissue is a strong predictor of clinical outcomes. Here, we hypothesize that fractional resistance arises from cell-to-cell differences in core cell cycle regulators that allow a subset of cells to escape CDK4/6 inhibitor therapy. We used multiplex, single-cell imaging to identify fractionally resistant cells in both cultured and primary breast tumor samples resected from patients. Resistant cells showed premature accumulation of multiple G1 regulators including E2F1, retinoblastoma protein, and CDK2, as well as enhanced sensitivity to pharmacological inhibition of CDK2 activity. Using trajectory inference approaches, we show how plasticity among cell cycle regulators gives rise to alternate cell cycle "paths" that allow individual tumor cells to escape palbociclib treatment. Understanding drivers of cell cycle plasticity, and how to eliminate resistant cell cycle paths, could lead to improved cancer therapies targeting fractionally resistant cells to improve patient outcomes.

MRE11 liberates cGAS from nucleosome sequestration during tumorigenesis.

Oncogene-induced replication stress generates endogenous DNA damage that activates cGAS-STING-mediated signalling and tumour suppression1-3. However, the precise mechanism of cGAS activation by endogenous DNA damage remains enigmatic, particularly given that high-affinity histone acidic patch (AP) binding constitutively inhibits cGAS by sterically hindering its activation by double-stranded DNA (dsDNA)4-10. Here we report that the DNA double-strand break sensor MRE11 suppresses mammary tumorigenesis through a pivotal role in regulating cGAS activation. We demonstrate that binding of the MRE11-RAD50-NBN complex to nucleosome fragments is necessary to displace cGAS from acidic-patch-mediated sequestration, which enables its mobilization and activation by dsDNA. MRE11 is therefore essential for cGAS activation in response to oncogenic stress, cytosolic dsDNA and ionizing radiation. Furthermore, MRE11-dependent cGAS activation promotes ZBP1-RIPK3-MLKL-mediated necroptosis, which is essential to suppress oncogenic proliferation and breast tumorigenesis. Notably, downregulation of ZBP1 in human triple-negative breast cancer is associated with increased genome instability, immune suppression and poor patient prognosis. These findings establish MRE11 as a crucial mediator that links DNA damage and cGAS activation, resulting in tumour suppression through ZBP1-dependent necroptosis.

Cellograph: a semi-supervised approach to analyzing multi-condition single-cell RNA-sequencing data using graph neural networks.

With the growing number of single-cell datasets collected under more complex experimental conditions, there is an opportunity to leverage single-cell variability to reveal deeper insights into how cells respond to perturbations. Many existing approaches rely on discretizing the data into clusters for differential gene expression (DGE), effectively ironing out any information unveiled by the single-cell variability across cell-types. In addition, DGE often assumes a statistical distribution that, if erroneous, can lead to false positive differentially expressed genes. Here, we present Cellograph: a semi-supervised framework that uses graph neural networks to quantify the effects of perturbations at single-cell granularity. Cellograph not only measures how prototypical cells are of each condition but also learns a latent space that is amenable to interpretable data visualization and clustering. The learned gene weight matrix from training reveals pertinent genes driving the differences between conditions. We demonstrate the utility of our approach on publicly-available datasets including cancer drug therapy, stem cell reprogramming, and organoid differentiation. Cellograph outperforms existing methods for quantifying the effects of experimental perturbations and offers a novel framework to analyze single-cell data using deep learning.

Feature selection for preserving biological trajectories in single-cell data.

Single-cell technologies can readily measure the expression of thousands of molecular features from individual cells undergoing dynamic biological processes, such as cellular differentiation, immune response, and disease progression. While examining cells along a computationally ordered pseudotime offers the potential to study how subtle changes in gene or protein expression impact cell fate decision-making, identifying characteristic features that drive continuous biological processes remains difficult to detect from unenriched and noisy single-cell data. Given that all profiled sources of feature variation contribute to the cell-to-cell distances that define an inferred cellular trajectory, including confounding sources of biological variation (e.g. cell cycle or metabolic state) or noisy and irrelevant features (e.g. measurements with low signal-to-noise ratio) can mask the underlying trajectory of study and hinder inference. Here, we present DELVE (dynamic selection of locally covarying features), an unsupervised feature selection method for identifying a representative subset of dynamically-expressed molecular features that recapitulates cellular trajectories. In contrast to previous work, DELVE uses a bottom-up approach to mitigate the effect of unwanted sources of variation confounding inference, and instead models cell states from dynamic feature modules that constitute core regulatory complexes. Using simulations, single-cell RNA sequencing data, and iterative immunofluorescence imaging data in the context of the cell cycle and cellular differentiation, we demonstrate that DELVE selects features that more accurately characterize cell populations and improve the recovery of cell type transitions. This feature selection framework provides an alternative approach for improving trajectory inference and uncovering co-variation amongst features along a biological trajectory. DELVE is implemented as an open-source python package and is publicly available at: https://github.com/jranek/delve.

Modeling collective cell behavior in cancer: Perspectives from an interdisciplinary conversation.

Collective cell behavior contributes to all stages of cancer progression. Understanding how collective behavior emerges through cell-cell interactions and decision-making will advance our understanding of cancer biology and provide new therapeutic approaches. Here, we summarize an interdisciplinary discussion on multicellular behavior in cancer, draw lessons from other scientific disciplines, and identify future directions.

Integrating temporal single-cell gene expression modalities for trajectory inference and disease prediction.

BACKGROUND: Current methods for analyzing single-cell datasets have relied primarily on static gene expression measurements to characterize the molecular state of individual cells. However, capturing temporal changes in cell state is crucial for the interpretation of dynamic phenotypes such as the cell cycle, development, or disease progression. RNA velocity infers the direction and speed of transcriptional changes in individual cells, yet it is unclear how these temporal gene expression modalities may be leveraged for predictive modeling of cellular dynamics. RESULTS: Here, we present the first task-oriented benchmarking study that investigates integration of temporal sequencing modalities for dynamic cell state prediction. We benchmark ten integration approaches on ten datasets spanning different biological contexts, sequencing technologies, and species. We find that integrated data more accurately infers biological trajectories and achieves increased performance on classifying cells according to perturbation and disease states. Furthermore, we show that simple concatenation of spliced and unspliced molecules performs consistently well on classification tasks and can be used over more memory intensive and computationally expensive methods. CONCLUSIONS: This work illustrates how integrated temporal gene expression modalities may be leveraged for predicting cellular trajectories and sample-associated perturbation and disease phenotypes. Additionally, this study provides users with practical recommendations for task-specific integration of single-cell gene expression modalities.

The molecular architecture of cell cycle arrest.

The cellular decision governing the transition between proliferative and arrested states is crucial to the development and function of every tissue. While the molecular mechanisms that regulate the proliferative cell cycle are well established, we know comparatively little about what happens to cells as they diverge into cell cycle arrest. We performed hyperplexed imaging of 47 cell cycle effectors to obtain a map of the molecular architecture that governs cell cycle exit and progression into reversible ("quiescent") and irreversible ("senescent") arrest states. Using this map, we found multiple points of divergence from the proliferative cell cycle; identified stress-specific states of arrest; and resolved the molecular mechanisms governing these fate decisions, which we validated by single-cell, time-lapse imaging. Notably, we found that cells can exit into senescence from either G1 or G2; however, both subpopulations converge onto a single senescent state with a G1-like molecular signature. Cells can escape from this "irreversible" arrest state through the upregulation of G1 cyclins. This map provides a more comprehensive understanding of the overall organization of cell proliferation and arrest.

A Proximal-to-Distal Survey of Healthy Adult Human Small Intestine and Colon Epithelium by Single-Cell Transcriptomics.

BACKGROUND & AIMS: Single-cell transcriptomics offer unprecedented resolution of tissue function at the cellular level, yet studies analyzing healthy adult human small intestine and colon are sparse. Here, we present single-cell transcriptomics covering the duodenum, jejunum, ileum, and ascending, transverse, and descending colon from 3 human beings. METHODS: A total of 12,590 single epithelial cells from 3 independently processed organ donors were evaluated for organ-specific lineage biomarkers, differentially regulated genes, receptors, and drug targets. Analyses focused on intrinsic cell properties and their capacity for response to extrinsic signals along the gut axis across different human beings. RESULTS: Cells were assigned to 25 epithelial lineage clusters. Multiple accepted intestinal stem cell markers do not specifically mark all human intestinal stem cells. Lysozyme expression is not unique to human Paneth cells, and Paneth cells lack expression of expected niche factors. Bestrophin 4 (BEST4)+ cells express Neuropeptide Y (NPY) and show maturational differences between the small intestine and colon. Tuft cells possess a broad ability to interact with the innate and adaptive immune systems through previously unreported receptors. Some classes of mucins, hormones, cell junctions, and nutrient absorption genes show unappreciated regional expression differences across lineages. The differential expression of receptors and drug targets across lineages show biological variation and the potential for variegated responses. CONCLUSIONS: Our study identifies novel lineage marker genes, covers regional differences, shows important differences between mouse and human gut epithelium, and reveals insight into how the epithelium responds to the environment and drugs. This comprehensive cell atlas of the healthy adult human intestinal epithelium resolves likely functional differences across anatomic regions along the gastrointestinal tract and advances our understanding of human intestinal physiology.

The consequences of differential origin licensing dynamics in distinct chromatin environments.

Eukaryotic chromosomes contain regions of varying accessibility, yet DNA replication factors must access all regions. The first replication step is loading MCM complexes to license replication origins during the G1 cell cycle phase. It is not yet known how mammalian MCM complexes are adequately distributed to both accessible euchromatin regions and less accessible heterochromatin regions. To address this question, we combined time-lapse live-cell imaging with immunofluorescence imaging of single human cells to quantify the relative rates of MCM loading in euchromatin and heterochromatin throughout G1. We report here that MCM loading in euchromatin is faster than that in heterochromatin in early G1, but surprisingly, heterochromatin loading accelerates relative to euchromatin loading in middle and late G1. This differential acceleration allows both chromatin types to begin S phase with similar concentrations of loaded MCM. The different loading dynamics require ORCA-dependent differences in origin recognition complex distribution. A consequence of heterochromatin licensing dynamics is that cells experiencing a truncated G1 phase from premature cyclin E expression enter S phase with underlicensed heterochromatin, and DNA damage accumulates preferentially in heterochromatin in the subsequent S/G2 phase. Thus, G1 length is critical for sufficient MCM loading, particularly in heterochromatin, to ensure complete genome duplication and to maintain genome stability.

In this study the authors have, for the first time, quantified DNA replication origin licensing dynamics and distribution in single cells at subnuclear resolution. The cell cycle and DNA replication fields have long appreciated that origin licensing is both absolutely essential for replication and that licensing is strictly confined to G1 phase. The biochemical process of origin licensing- which is the DNA loading of MCM complexes- is known in considerable detail. What has never been explored in any system, is the dynamics of origin licensing itself. Here the authors define the dynamics of human MCM loading at different times within G1 in both euchromatin and heterochromatin, and explore the consequences of those dynamics for genome stability.

The structure of the human cell cycle.

Understanding the organization of the cell cycle has been a longstanding goal in cell biology. We combined time-lapse microscopy, highly multiplexed single-cell imaging of 48 core cell cycle proteins, and manifold learning to render a visualization of the human cell cycle. This data-driven approach revealed the comprehensive "structure" of the cell cycle: a continuum of molecular states that cells occupy as they transition from one cell division to the next, or as they enter or exit cell cycle arrest. Paradoxically, progression deeper into cell cycle arrest was accompanied by increases in proliferative effectors such as CDKs and cyclins, which can drive cell cycle re-entry by overcoming p21 induction. The structure also revealed the molecular trajectories into senescence and the unique combination of molecular features that define this irreversibly arrested state. This approach will enable the comparison of alternative cell cycles during development, in response to environmental perturbation and in disease. A record of this paper's transparent peer review process is included in the supplemental information.

Cyclin F drives proliferation through SCF-dependent degradation of the retinoblastoma-like tumor suppressor p130/RBL2.

Cell cycle gene expression programs fuel proliferation and are universally dysregulated in cancer. The retinoblastoma (RB)-family of proteins, RB1, RBL1/p107, and RBL2/p130, coordinately represses cell cycle gene expression, inhibiting proliferation, and suppressing tumorigenesis. Phosphorylation of RB-family proteins by cyclin-dependent kinases is firmly established. Like phosphorylation, ubiquitination is essential to cell cycle control, and numerous proliferative regulators, tumor suppressors, and oncoproteins are ubiquitinated. However, little is known about the role of ubiquitin signaling in controlling RB-family proteins. A systems genetics analysis of CRISPR/Cas9 screens suggested the potential regulation of the RB-network by cyclin F, a substrate recognition receptor for the SCF family of E3 ligases. We demonstrate that RBL2/p130 is a direct substrate of SCFcyclin F. We map a cyclin F regulatory site to a flexible linker in the p130 pocket domain, and show that this site mediates binding, stability, and ubiquitination. Expression of a mutant version of p130, which cannot be ubiquitinated, severely impaired proliferative capacity and cell cycle progression. Consistently, we observed reduced expression of cell cycle gene transcripts, as well a reduced abundance of cell cycle proteins, analyzed by quantitative, iterative immunofluorescent imaging. These data suggest a key role for SCFcyclin F in the CDK-RB network and raise the possibility that aberrant p130 degradation could dysregulate the cell cycle in human cancers.

Inferring the structures of signaling motifs from paired dynamic traces of single cells.

Individual cells show variability in their signaling dynamics that often correlates with phenotypic responses, indicating that cell-to-cell variability is not merely noise but can have functional consequences. Based on this observation, we reasoned that cell-to-cell variability under the same treatment condition could be explained in part by a single signaling motif that maps different upstream signals into a corresponding set of downstream responses. If this assumption holds, then repeated measurements of upstream and downstream signaling dynamics in a population of cells could provide information about the underlying signaling motif for a given pathway, even when no prior knowledge of that motif exists. To test these two hypotheses, we developed a computer algorithm called MISC (Motif Inference from Single Cells) that infers the underlying signaling motif from paired time-series measurements from individual cells. When applied to measurements of transcription factor and reporter gene expression in the yeast stress response, MISC predicted signaling motifs that were consistent with previous mechanistic models of transcription. The ability to detect the underlying mechanism became less certain when a cell's upstream signal was randomly paired with another cell's downstream response, demonstrating how averaging time-series measurements across a population obscures information about the underlying signaling mechanism. In some cases, motif predictions improved as more cells were added to the analysis. These results provide evidence that mechanistic information about cellular signaling networks can be systematically extracted from the dynamical patterns of single cells.

Autophagy regulates the localization and degradation of p16INK4a.

The tumor suppressor protein p16INK4a (p16) is a well-established hallmark of aging that induces cellular senescence in response to stress. Previous studies have focused primarily on p16 regulation at the transcriptional level; comparatively little is known about the protein's intracellular localization and degradation. The autophagy-lysosomal pathway has been implicated in the subcellular trafficking and turnover of various stress-response proteins and has also been shown to attenuate age-related pathologies, but it is unclear whether p16 is involved in this pathway. Here, we investigate the role of autophagy, vesicular trafficking, and lysosomal degradation on p16 expression and localization in human epithelial cells. Time-lapse fluorescence microscopy using an endogenous p16-mCherry reporter revealed that serum starvation, etoposide, and hydrogen peroxide stimulate autophagy and drive p16 recruitment to acidic cytoplasmic vesicles within 4 hr. Blocking lysosomal proteases with leupeptin and ammonium chloride resulted in the accumulation of p16 within lysosomes and increased total p16 levels suggesting that p16 is degraded by this pathway. Furthermore, autophagy blockers chloroquine and bafilomycin A1 caused p16 aggregation within stalled vesicles containing autophagosome marker LC3. Increase of p16 within these vesicles coincided with the accumulation of LC3-II. Knockdown of autophagosome chaperone p62 attenuated the formation of p16 aggregates in lysosomes, suggesting that p16 is targeted to these vesicles by p62. Taken together, these results implicate the autophagy pathway as a novel regulator of p16 degradation and localization, which could play a role in the etiology of cancer and age-related diseases.

Dual inhibition of DNA-PK and DNA polymerase theta overcomes radiation resistance induced by p53 deficiency.

TP53 deficiency in cancer is associated with poor patient outcomes and resistance to DNA damaging therapies. However, the mechanisms underlying treatment resistance in p53-deficient cells remain poorly characterized. Using live cell imaging of DNA double-strand breaks (DSBs) and cell cycle state transitions, we show that p53-deficient cells exhibit accelerated repair of radiomimetic-induced DSBs arising in S phase. Low-dose DNA-dependent protein kinase (DNA-PK) inhibition increases the S-phase DSB burden in p53-deficient cells, resulting in elevated rates of mitotic catastrophe. However, a subset of p53-deficient cells exhibits intrinsic resistance to radiomimetic-induced DSBs despite DNA-PK inhibition. We show that p53-deficient cells under DNA-PK inhibition utilize DNA polymerase theta (Pol θ)-mediated end joining repair to promote their viability in response to therapy-induced DSBs. Pol θ inhibition selectively increases S-phase DSB burden after radiomimetic therapy and promotes prolonged G2 arrest. Dual inhibition of DNA-PK and Pol θ restores radiation sensitivity in p53-deficient cells as well as in p53-mutant breast cancer cell lines. Thus, combination targeting of DNA-PK- and Pol θ-dependent end joining repair represents a promising strategy for overcoming resistance to DNA damaging therapies in p53-deficient cancers.

Bistable switches as integrators and actuators during cell cycle progression.

Progression through the cell cycle is driven by bistable switches-specialized molecular circuits that govern transitions from one cellular state to another. Although the mechanics of bistable switches are relatively well understood, it is less clear how cells integrate multiple sources of molecular information to engage these switches. Here, we describe how bistable switches act as hubs of information processing and examine how variability, competition, and inheritance of molecular signals determine the timing of the Rb-E2F bistable switch that controls cell cycle entry. Bistable switches confer both robustness and plasticity to cell cycle progression, ensuring that cell cycle events are performed completely and in the correct order, while still allowing flexibility to cope with ongoing stress and changing environmental conditions.

Genetic determinants of cellular addiction to DNA polymerase theta.

Polymerase theta (Pol θ, gene name Polq) is a widely conserved DNA polymerase that mediates a microhomology-mediated, error-prone, double strand break (DSB) repair pathway, referred to as Theta Mediated End Joining (TMEJ). Cells with homologous recombination deficiency are reliant on TMEJ for DSB repair. It is unknown whether deficiencies in other components of the DNA damage response (DDR) also result in Pol θ addiction. Here we use a CRISPR genetic screen to uncover 140 Polq synthetic lethal (PolqSL) genes, the majority of which were previously unknown. Functional analyses indicate that Pol θ/TMEJ addiction is associated with increased levels of replication-associated DSBs, regardless of the initial source of damage. We further demonstrate that approximately 30% of TCGA breast cancers have genetic alterations in PolqSL genes and exhibit genomic scars of Pol θ/TMEJ hyperactivity, thereby substantially expanding the subset of human cancers for which Pol θ inhibition represents a promising therapeutic strategy.

Chemical screen for epigenetic barriers to single allele activation of Oct4.

Here we utilized the chromatin in vivo assay (CiA) mouse platform to directly examine the epigenetic barriers impeding the activation of the CiA:Oct4 allele in mouse embryonic fibroblasts (MEF)s when stimulated with a transcription factor. The CiA:Oct4 allele contains an engineered EGFP reporter replacing one copy of the Oct4 gene, with an upstream Gal4 array in the promoter that allows recruitment of chromatin modifying machinery. We stimulated gene activation of the CiA:Oct4 allele by binding a transcriptional activator to the Gal4 array. As with cellular reprograming, this process is inefficient with only a small percentage of the cells re-activating CiA:Oct4 after weeks. Epigenetic barriers to gene activation potentially come from heavy DNA methylation, histone deacetylation, chromatin compaction, and other posttranslational marks (PTM) at the differentiated CiA:Oct4 allele in MEFs. Using this platform, we performed a high-throughput chemical screen for compounds that increased the efficiency of activation. We found that Azacytidine and newer generation histone deacetylase (HDAC) inhibitors were the most efficient at facilitating directed transcriptional activation of this allele. We found one hit form our screen, Mocetinostat, improved iPSC generation under transcription factor reprogramming conditions. These results separate individual allele activation from whole cell reprograming and give new insights that will advance tissue engineering.