Statistical inference with a manifold-constrained RNA velocity model uncovers cell cycle speed modulations.

Across biological systems, cells undergo coordinated changes in gene expression, resulting in transcriptome dynamics that unfold within a low-dimensional manifold. While low-dimensional dynamics can be extracted using RNA velocity, these algorithms can be fragile and rely on heuristics lacking statistical control. Moreover, the estimated vector field is not dynamically consistent with the traversed gene expression manifold. To address these challenges, we introduce a Bayesian model of RNA velocity that couples velocity field and manifold estimation in a reformulated, unified framework, identifying the parameters of an explicit dynamical system. Focusing on the cell cycle, we implement VeloCycle to study gene regulation dynamics on one-dimensional periodic manifolds and validate its ability to infer cell cycle periods using live imaging. We also apply VeloCycle to reveal speed differences in regionally defined progenitors and Perturb-seq gene knockdowns. Overall, VeloCycle expands the single-cell RNA sequencing analysis toolkit with a modular and statistically consistent RNA velocity inference framework.

Clinically compliant cryopreservation of differentiated retinal pigment epithelial cells.

BACKGROUND AIMS: Age-related macular degeneration (AMD) is the most common cause of blindness in elderly patients within developed countries, affecting more than 190 million worldwide. In AMD, the retinal pigment epithelial (RPE) cell layer progressively degenerates, resulting in subsequent loss of photoreceptors and ultimately vision. There is currently no cure for AMD, but therapeutic strategies targeting the complement system are being developed to slow the progression of the disease. METHODS: Replacement therapy with pluripotent stem cell-derived (hPSC) RPEs is an alternative treatment strategy. A cell therapy product must be produced in accordance with Good Manufacturing Practices at a sufficient scale to facilitate extensive pre-clinical and clinical testing. Cryopreservation of the final cell product is therefore highly beneficial, as the manufacturing, pre-clinical and clinical testing can be separated in time and location. RESULTS: We found that mature hPSC-RPE cells do not survive conventional cryopreservation techniques. However, replating the cells 2-5 days before cryopreservation facilitates freezing. The replated and cryopreserved hPSC-RPE cells maintained their identity, purity and functionality as characteristic RPEs, shown by cobblestone morphology, pigmentation, transcriptional profile, RPE markers, transepithelial resistance and pigment epithelium-derived factor secretion. Finally, we showed that the optimal replating time window can be tracked noninvasively by following the change in cobblestone morphology. CONCLUSIONS: The possibility of cryopreserving the hPSC-RPE product has been instrumental in our efforts in manufacturing and performing pre-clinical testing with the aim for clinical translation.

Brain matters: unveiling the distinct contributions of region, age, and sex to glia diversity and CNS function.

The myelinated white matter tracts of the central nervous system (CNS) are essential for fast transmission of electrical impulses and are often differentially affected in human neurodegenerative diseases across CNS region, age and sex. We hypothesize that this selective vulnerability is underpinned by physiological variation in white matter glia. Using single nucleus RNA sequencing of human post-mortem white matter samples from the brain, cerebellum and spinal cord and subsequent tissue-based validation we found substantial glial heterogeneity with tissue region: we identified region-specific oligodendrocyte precursor cells (OPCs) that retain developmental origin markers into adulthood, distinguishing them from mouse OPCs. Region-specific OPCs give rise to similar oligodendrocyte populations, however spinal cord oligodendrocytes exhibit markers such as SKAP2 which are associated with increased myelin production and we found a spinal cord selective population particularly equipped for producing long and thick myelin sheaths based on the expression of genes/proteins such as HCN2. Spinal cord microglia exhibit a more activated phenotype compared to brain microglia, suggesting that the spinal cord is a more pro-inflammatory environment, a difference that intensifies with age. Astrocyte gene expression correlates strongly with CNS region, however, astrocytes do not show a more activated state with region or age. Across all glia, sex differences are subtle but the consistent increased expression of protein-folding genes in male donors hints at pathways that may contribute to sex differences in disease susceptibility. These findings are essential to consider for understanding selective CNS pathologies and developing tailored therapeutic strategies.

Self-organizing in vitro mouse neural tube organoids mimic embryonic development.

The embryonic neural tube is the origin of the entire adult nervous system, and disturbances in its development cause life-threatening birth defects. However, the study of mammalian neural tube development is limited by the lack of physiologically realistic three-dimensional (3D) in vitro models. Here, we report a self-organizing 3D neural tube organoid model derived from single mouse embryonic stem cells that exhibits an in vivo-like tissue architecture, cell type composition and anterior-posterior (AP) patterning. Moreover, maturation of the neural tube organoids showed the emergence of multipotent neural crest cells and mature neurons. Single-cell transcriptome analyses revealed the sequence of transcriptional events in the emergence of neural crest cells and neural differentiation. Thanks to the accessibility of this model, phagocytosis of migrating neural crest cells could be observed in real time for the first time in a mammalian model. We thus introduce a tractable in vitro model to study some of the key morphogenetic and cell type derivation events during early neural development.

Molecular profiling of stem cell-derived retinal pigment epithelial cell differentiation established for clinical translation.

Human embryonic stem cell-derived retinal pigment epithelial cells (hESC-RPE) are a promising cell source to treat age-related macular degeneration (AMD). Despite several ongoing clinical studies, a detailed mapping of transient cellular states during in vitro differentiation has not been performed. Here, we conduct single-cell transcriptomic profiling of an hESC-RPE differentiation protocol that has been developed for clinical use. Differentiation progressed through a culture diversification recapitulating early embryonic development, whereby cells rapidly acquired a rostral embryo patterning signature before converging toward the RPE lineage. At intermediate steps, we identified and examined the potency of an NCAM1+ retinal progenitor population and showed the ability of the protocol to suppress non-RPE fates. We demonstrated that the method produces a pure RPE pool capable of maturing further after subretinal transplantation in a large-eyed animal model. Our evaluation of hESC-RPE differentiation supports the development of safe and efficient pluripotent stem cell-based therapies for AMD.

Sphingolipids control dermal fibroblast heterogeneity.

Human cells produce thousands of lipids that change during cell differentiation and can vary across individual cells of the same type. However, we are only starting to characterize the function of these cell-to-cell differences in lipid composition. Here, we measured the lipidomes and transcriptomes of individual human dermal fibroblasts by coupling high-resolution mass spectrometry imaging with single-cell transcriptomics. We found that the cell-to-cell variations of specific lipid metabolic pathways contribute to the establishment of cell states involved in the organization of skin architecture. Sphingolipid composition is shown to define fibroblast subpopulations, with sphingolipid metabolic rewiring driving cell-state transitions. Therefore, cell-to-cell lipid heterogeneity affects the determination of cell states, adding a new regulatory component to the self-organization of multicellular systems.

Clonal relations in the mouse brain revealed by single-cell and spatial transcriptomics.

The mammalian brain contains many specialized cells that develop from a thin sheet of neuroepithelial progenitor cells. Single-cell transcriptomics revealed hundreds of molecularly diverse cell types in the nervous system, but the lineage relationships between mature cell types and progenitor cells are not well understood. Here we show in vivo barcoding of early progenitors to simultaneously profile cell phenotypes and clonal relations in the mouse brain using single-cell and spatial transcriptomics. By reconstructing thousands of clones, we discovered fate-restricted progenitor cells in the mouse hippocampal neuroepithelium and show that microglia are derived from few primitive myeloid precursors that massively expand to generate widely dispersed progeny. We combined spatial transcriptomics with clonal barcoding and disentangled migration patterns of clonally related cells in densely labeled tissue sections. Our approach enables high-throughput dense reconstruction of cell phenotypes and clonal relations at the single-cell and tissue level in individual animals and provides an integrated approach for understanding tissue architecture.

Confronting false discoveries in single-cell differential expression.

Differential expression analysis in single-cell transcriptomics enables the dissection of cell-type-specific responses to perturbations such as disease, trauma, or experimental manipulations. While many statistical methods are available to identify differentially expressed genes, the principles that distinguish these methods and their performance remain unclear. Here, we show that the relative performance of these methods is contingent on their ability to account for variation between biological replicates. Methods that ignore this inevitable variation are biased and prone to false discoveries. Indeed, the most widely used methods can discover hundreds of differentially expressed genes in the absence of biological differences. To exemplify these principles, we exposed true and false discoveries of differentially expressed genes in the injured mouse spinal cord.

Molecular architecture of the developing mouse brain.

The mammalian brain develops through a complex interplay of spatial cues generated by diffusible morphogens, cell-cell interactions and intrinsic genetic programs that result in probably more than a thousand distinct cell types. A complete understanding of this process requires a systematic characterization of cell states over the entire spatiotemporal range of brain development. The ability of single-cell RNA sequencing and spatial transcriptomics to reveal the molecular heterogeneity of complex tissues has therefore been particularly powerful in the nervous system. Previous studies have explored development in specific brain regions1-8, the whole adult brain9 and even entire embryos10. Here we report a comprehensive single-cell transcriptomic atlas of the embryonic mouse brain between gastrulation and birth. We identified almost eight hundred cellular states that describe a developmental program for the functional elements of the brain and its enclosing membranes, including the early neuroepithelium, region-specific secondary organizers, and both neurogenic and gliogenic progenitors. We also used in situ mRNA sequencing to map the spatial expression patterns of key developmental genes. Integrating the in situ data with our single-cell clusters revealed the precise spatial organization of neural progenitors during the patterning of the nervous system.

Spatial tissue profiling by imaging-free molecular tomography.

Several techniques are currently being developed for spatially resolved omics profiling, but each new method requires the setup of specific detection strategies or specialized instrumentation. Here we describe an imaging-free framework to localize high-throughput readouts within a tissue by cutting the sample into thin strips in a way that allows subsequent image reconstruction. We implemented this framework to transform a low-input RNA sequencing protocol into an imaging-free spatial transcriptomics technique (called STRP-seq) and validated it by profiling the spatial transcriptome of the mouse brain. We applied the technique to the brain of the Australian bearded dragon, Pogona vitticeps. Our results reveal the molecular anatomy of the telencephalon of this lizard, providing evidence for a marked regionalization of the reptilian pallium and subpallium. We expect that STRP-seq can be used to derive spatially resolved data from a range of other omics techniques.

Cell type prioritization in single-cell data.

We present Augur, a method to prioritize the cell types most responsive to biological perturbations in single-cell data. Augur employs a machine-learning framework to quantify the separability of perturbed and unperturbed cells within a high-dimensional space. We validate our method on single-cell RNA sequencing, chromatin accessibility and imaging transcriptomics datasets, and show that Augur outperforms existing methods based on differential gene expression. Augur identified the neural circuits restoring locomotion in mice following spinal cord neurostimulation.

Myocardial micro-biopsy procedure for molecular characterization with increased precision and reduced trauma.

Endomyocardial biopsy is a valuable tool in cardiac diagnostics but is limited by low diagnostic yield and significant complication risks. Meanwhile, recent developments in transcriptomic and proteomic technologies promise a wealth of biological data from minimal tissue samples. To take advantage of the minimal tissue amount needed for molecular analyses, we have developed a sub-millimeter endovascular biopsy device, considerably smaller than current clinical equipment, and devised a low-input RNA-sequencing protocol for analyzing small tissue samples. In in vivo evaluation in swine, 81% of biopsy attempts (n = 157) were successful. High quality RNA-sequencing data was generated from 91% of the sequenced cardiac micro-biopsy samples (n = 32). Gene expression signatures of samples taken with the novel device were comparable with a conventional device. No major complications were detected either during procedures or during 7 days' follow-up, despite acquiring a relatively large number of biopsies (median 30) in each animal. In conclusion, the novel device coupled with RNA-sequencing provides a feasible method to obtain molecular data from the myocardium. The method is less traumatic and has a higher flexibility compared to conventional methods, enabling safer and more targeted sampling from different parts of the myocardium.

The emergence and promise of single-cell temporal-omics approaches.

Single-cell transcriptomics enables the measurement of gene expression in complex biological systems at the resolution of individual cells. Multivariate analysis of single-cell data helps describe the variation in expression accompanying cellular processes during embryonic development, disease progression and in response to stimuli. Likewise, new methods have extended the possibilities of single-cell analysis by measuring the transcriptome while simultaneously capturing information on lineage or past molecular events. These emerging approaches have the common strategy of querying a static snapshot for information related to different temporal stages. Single-cell temporal-omics methods open new possibilities such as extrapolating the future or correlating past events to present gene expression. We highlight advancements in the single-cell field, describe novel toolkits for investigation, and consider the potential impact of temporal-omics approaches for the study of disease progression.

A cell fitness selection model for neuronal survival during development.

Developmental cell death plays an important role in the construction of functional neural circuits. In vertebrates, the canonical view proposes a selection of the surviving neurons through stochastic competition for target-derived neurotrophic signals, implying an equal potential for neurons to compete. Here we show an alternative cell fitness selection of neurons that is defined by a specific neuronal heterogeneity code. Proprioceptive sensory neurons that will undergo cell death and those that will survive exhibit different molecular signatures that are regulated by retinoic acid and transcription factors, and are independent of the target and neurotrophins. These molecular features are genetically encoded, representing two distinct subgroups of neurons with contrasted functional maturation states and survival outcome. Thus, in this model, a heterogeneous code of intrinsic cell fitness in neighboring neurons provides differential competitive advantage resulting in the selection of cells with higher capacity to survive and functionally integrate into neural networks.

Single-cell transcriptional logic of cell-fate specification and axon guidance in early-born retinal neurons.

Retinal ganglion cells (RGCs), cone photoreceptors (cones), horizontal cells and amacrine cells are the first classes of neurons produced in the retina. However, an important question is how this diversity of cell states is transcriptionally produced. Here, we profiled 6067 single retinal cells to provide a comprehensive transcriptomic atlas showing the diversity of the early developing mouse retina. RNA velocities unveiled the dynamics of cell cycle coordination of early retinogenesis and define the transcriptional sequences at work during the hierarchical production of early cell-fate specification. We show that RGC maturation follows six waves of gene expression, with older-generated RGCs transcribing increasing amounts of guidance cues for young peripheral RGC axons that express the matching receptors. Spatial transcriptionally deduced features in subpopulations of RGCs allowed us to define novel molecular markers that are spatially restricted. Finally, the isolation of such a spatially restricted population, ipsilateral RGCs, allowed us to identify their molecular identity at the time they execute axon guidance decisions. Together, these data represent a valuable resource shedding light on transcription factor sequences and guidance cue dynamics during mouse retinal development.

Spatial organization of the somatosensory cortex revealed by osmFISH.

Global efforts to create a molecular census of the brain using single-cell transcriptomics are producing a large catalog of molecularly defined cell types. However, spatial information is lacking and new methods are needed to map a large number of cell type-specific markers simultaneously on large tissue areas. Here, we describe a cyclic single-molecule fluorescence in situ hybridization methodology and define the cellular organization of the somatosensory cortex.

Single-Cell Transcriptomics of Traced Epidermal and Hair Follicle Stem Cells Reveals Rapid Adaptations during Wound Healing.

Epithelial tissues, such as the skin, rely on cellular plasticity of stem cells (SCs) from different niches to restore tissue function after injury. How these molecularly and functionally diverse SC populations respond to injury remains elusive. Here, we genetically labeled Lgr5- or Lgr6-expressing cells from the hair follicle bulge and interfollicular epidermis (IFE), respectively, and monitored their individual transcriptional adaptations during wound healing using single-cell transcriptomics. Both Lgr5 and Lgr6 progeny rapidly induced a genetic wound signature that, for Lgr5 progeny, included the remodeling of receptors to permit interactions with the wound environment, a property that Lgr6 progeny possessed even before wounding. When contributing to re-epithelialization, Lgr5 progeny gradually replaced their bulge identity with an IFE identity, and this process started already before Lgr5 progeny left the bulge. Altogether, this study reveals how different SCs respond and adapt to a new environment, potentially explaining cellular plasticity of many epithelial tissues.

RNA velocity of single cells.

RNA abundance is a powerful indicator of the state of individual cells. Single-cell RNA sequencing can reveal RNA abundance with high quantitative accuracy, sensitivity and throughput1. However, this approach captures only a static snapshot at a point in time, posing a challenge for the analysis of time-resolved phenomena such as embryogenesis or tissue regeneration. Here we show that RNA velocity-the time derivative of the gene expression state-can be directly estimated by distinguishing between unspliced and spliced mRNAs in common single-cell RNA sequencing protocols. RNA velocity is a high-dimensional vector that predicts the future state of individual cells on a timescale of hours. We validate its accuracy in the neural crest lineage, demonstrate its use on multiple published datasets and technical platforms, reveal the branching lineage tree of the developing mouse hippocampus, and examine the kinetics of transcription in human embryonic brain. We expect RNA velocity to greatly aid the analysis of developmental lineages and cellular dynamics, particularly in humans.