Conserved properties of dentate gyrus neurogenesis across postnatal development revealed by single-cell RNA sequencing

Emerging strategies for nerve repair and regeneration in ischemic stroke: neural stem cell therapy

Ischemic stroke is a major cause of mortality and disability worldwide, with limited treatment options available in clinical practice. The emergence of stem cell therapy has provided new hope to the field of stroke treatment via the restoration of brain neuron function. Exogenous neural stem cells are beneficial not only in cell replacement but also through the bystander effect. Neural stem cells regulate multiple physiological responses, including nerve repair, endogenous regeneration, immune function, and blood-brain barrier permeability, through the secretion of bioactive substances, including extracellular vesicles/exosomes. However, due to the complex microenvironment of ischemic cerebrovascular events and the low survival rate of neural stem cells following transplantation, limitations in the treatment effect remain unresolved. In this paper, we provide a detailed summary of the potential mechanisms of neural stem cell therapy for the treatment of ischemic stroke, review current neural stem cell therapeutic strategies and clinical trial results, and summarize the latest advancements in neural stem cell engineering to improve the survival rate of neural stem cells. We hope that this review could help provide insight into the therapeutic potential of neural stem cells and guide future scientific endeavors on neural stem cells.

Single-Cell Transcriptome Reveals Cell Type-Specific Molecular Pathology in a 2VO Cerebral Ischemic Mouse Model

Post-ischemia memory impairment is a major sequela in cerebral ischemia patients. However, cell type-specific molecular pathology in the hippocampus after ischemia is poorly understood. In this study, we adopted a mouse two-vessel occlusion ischemia model (2VO model) to mimic cerebral ischemia-induced memory impairment and investigated the single-cell transcriptome in the hippocampi in 2VO mice. A total of 27,069 cells were corresponding 14 cell types with neuronal, glial, and vascular lineages. We next analyzed cell-specific gene alterations in 2VO mice and the function of these cell-specific genes. Differential expression analysis identified cell type-specific genes with altered expression in neurons, astrocytes, microglia, and oligodendrocytes in 2VO mice. Notably, four subtypes of oligodendrocyte precursor cells with distinct differentiation pathways were suggested. Taken together, this is the first single-cell transcriptome analysis of gene expression in a 2VO model. Furthermore, we suggested new types of oligodendrocyte precursor cells with angiogenesis and neuroprotective potential, which might offer opportunities to identify new avenues of research and novel targets for ischemia treatment.

Sustained generation of neurons destined for neocortex with oxidative metabolic upregulation upon filamin abrogation

Neurons in the neocortex are generated during embryonic development. While the adult ventricular-subventricular zone (V-SVZ) contains cells with neural stem/progenitors' characteristics, it remains unclear whether it has the capacity of producing neocortical neurons. Here, we show that generating neurons with transcriptomic resemblance to upper layer neocortical neurons continues in the V-SVZ of mouse models of a human condition known as periventricular heterotopia by abrogating Flna and Flnb. We found such surplus neurogenesis was associated with V-SVZ's upregulation of oxidative phosphorylation, mitochondrial biogenesis, and vascular abundance. Additionally, spatial transcriptomics analyses showed V-SVZ's neurogenic activation was coupled with transcriptional enrichment of genes in diverse pathways for energy metabolism, angiogenesis, cell signaling, synaptic transmission, and turnovers of nucleic acids and proteins in upper cortical layers. These findings support the potential of generating neocortical neurons in adulthood through boosting brain-wide vascular circulation, aerobic adenosine triphosphate synthesis, metabolic turnover, and neuronal activity.

Transcriptional dynamics orchestrating the development and integration of neurons born in the adult hippocampus

The adult hippocampus generates new granule cells (aGCs) with functional capabilities that convey unique forms of plasticity to the preexisting circuits. While early differentiation of adult radial glia-like cells (RGLs) has been studied extensively, the molecular mechanisms guiding the maturation of postmitotic neurons remain unknown. Here, we used a precise birthdating strategy to study aGC differentiation using single-nuclei RNA sequencing. Transcriptional profiling revealed a continuous trajectory from RGLs to mature aGCs, with multiple immature stages bearing increasing levels of effector genes supporting growth, excitability, and synaptogenesis. Analysis of differential gene expression, pseudo-time trajectory, and transcription factors (TFs) revealed critical transitions defining four cellular states: quiescent RGLs, proliferative progenitors, immature aGCs, and mature aGCs. Becoming mature aGCs involved a transcriptional switch that shuts down pathways promoting cell growth, such SoxC TFs, to activate programs that likely control neuronal homeostasis. aGCs overexpressing Sox4 or Sox11 remained immature. Our results unveil precise molecular mechanisms driving adult RGLs through the pathway of neuronal differentiation.

OmicVerse: a framework for bridging and deepening insights across bulk and single-cell sequencing

Single-cell sequencing is frequently affected by "omission" due to limitations in sequencing throughput, yet bulk RNA-seq may contain these ostensibly "omitted" cells. Here, we introduce the single cell trajectory blending from Bulk RNA-seq (BulkTrajBlend) algorithm, a component of the OmicVerse suite that leverages a Beta-Variational AutoEncoder for data deconvolution and graph neural networks for the discovery of overlapping communities. This approach effectively interpolates and restores the continuity of "omitted" cells within single-cell RNA sequencing datasets. Furthermore, OmicVerse provides an extensive toolkit for both bulk and single cell RNA-seq analysis, offering seamless access to diverse methodologies, streamlining computational processes, fostering exquisite data visualization, and facilitating the extraction of significant biological insights to advance scientific research.

Targeting neuronal epigenomes for brain rejuvenation

Aging is associated with a progressive decline of brain function, and the underlying causes and possible interventions to prevent this cognitive decline have been the focus of intense investigation. The maintenance of neuronal function over the lifespan requires proper epigenetic regulation, and accumulating evidence suggests that the deterioration of the neuronal epigenetic landscape contributes to brain dysfunction during aging. Epigenetic aging of neurons may, however, be malleable. Recent reports have shown age-related epigenetic changes in neurons to be reversible and targetable by rejuvenation strategies that can restore brain function during aging. This review discusses the current evidence that identifies neuronal epigenetic aging as a driver of cognitive decline and a promising target of brain rejuvenation strategies, and it highlights potential approaches for the specific manipulation of the aging neuronal epigenome to restore a youthful epigenetic state in the brain.

Fatecode enables cell fate regulator prediction using classification-supervised autoencoder perturbation

Cell reprogramming, which guides the conversion between cell states, is a promising technology for tissue repair and regeneration, with the ultimate goal of accelerating recovery from diseases or injuries. To accomplish this, regulators must be identified and manipulated to control cell fate. We propose Fatecode, a computational method that predicts cell fate regulators based only on single-cell RNA sequencing (scRNA-seq) data. Fatecode learns a latent representation of the scRNA-seq data using a deep learning-based classification-supervised autoencoder and then performs in silico perturbation experiments on the latent representation to predict genes that, when perturbed, would alter the original cell type distribution to increase or decrease the population size of a cell type of interest. We assessed Fatecode's performance using simulations from a mechanistic gene-regulatory network model and scRNA-seq data mapping blood and brain development of different organisms. Our results suggest that Fatecode can detect known cell fate regulators from single-cell transcriptomics datasets.

Kdm4a is an activity downregulated barrier to generate engrams for memory separation

Memory engrams are a subset of learning activated neurons critical for memory recall, consolidation, extinction and separation. While the transcriptional profile of engrams after learning suggests profound neural changes underlying plasticity and memory formation, little is known about how memory engrams are selected and allocated. As epigenetic factors suppress memory formation, we developed a CRISPR screening in the hippocampus to search for factors controlling engram formation. We identified histone lysine-specific demethylase 4a (Kdm4a) as a negative regulator for engram formation. Kdm4a is downregulated after neural activation and controls the volume of mossy fiber boutons. Mechanistically, Kdm4a anchors to the exonic region of Trpm7 gene loci, causing the stalling of nascent RNAs and allowing burst transcription of Trpm7 upon the dismissal of Kdm4a. Furthermore, the YTH domain containing protein 2 (Ythdc2) recruits Kdm4a to the Trpm7 gene and stabilizes nascent RNAs. Reducing the expression of Kdm4a in the hippocampus via genetic manipulation or artificial neural activation facilitated the ability of pattern separation in rodents. Our work indicates that Kdm4a is a negative regulator of engram formation and suggests a priming state to generate a separate memory.

Epistatic interactions between NMD and TRP53 control progenitor cell maintenance and brain size

Mutations in human nonsense-mediated mRNA decay (NMD) factors are enriched in neurodevelopmental disorders. We show that deletion of key NMD factor Upf2 in mouse embryonic neural progenitor cells causes perinatal microcephaly but deletion in immature neurons does not, indicating NMD's critical roles in progenitors. Upf2 knockout (KO) prolongs the cell cycle of radial glia progenitor cells, promotes their transition into intermediate progenitors, and leads to reduced upper-layer neurons. CRISPRi screening identified Trp53 knockdown rescuing Upf2KO progenitors without globally reversing NMD inhibition, implying marginal contributions of most NMD targets to the cell cycle defect. Integrated functional genomics shows that NMD degrades selective TRP53 downstream targets, including Cdkn1a, which, without NMD suppression, slow the cell cycle. Trp53KO restores the progenitor cell pool and rescues the microcephaly of Upf2KO mice. Therefore, one physiological role of NMD in the developing brain is to degrade selective TRP53 targets to control progenitor cell cycle and brain size.

Autocrine VEGF drives neural stem cell proximity to the adult hippocampus vascular niche

The vasculature is a key component of adult brain neural stem cell (NSC) niches. In the adult mammalian hippocampus, NSCs reside in close contact with a dense capillary network. How this niche is maintained is unclear. We recently found that adult hippocampal NSCs express VEGF, a soluble factor with chemoattractive properties for vascular endothelia. Here, we show that global and NSC-specific VEGF loss led to dissociation of NSCs and their intermediate progenitor daughter cells from local vasculature. Surprisingly, though, we found no changes in local vascular density. Instead, we found that NSC-derived VEGF supports maintenance of gene expression programs in NSCs and their progeny related to cell migration and adhesion. In vitro assays revealed that blockade of VEGF receptor 2 impaired NSC motility and adhesion. Our findings suggest that NSCs maintain their own proximity to vasculature via self-stimulated VEGF signaling that supports their motility towards and/or adhesion to local blood vessels.

Molecular cascade reveals sequential milestones underlying hippocampal neural stem cell development into an adult state

Quiescent adult neural stem cells (NSCs) in the mammalian brain arise from proliferating NSCs during development. Beyond acquisition of quiescence, an adult NSC hallmark, little is known about the process, milestones, and mechanisms underlying the transition of developmental NSCs to an adult NSC state. Here, we performed targeted single-cell RNA-seq analysis to reveal the molecular cascade underlying NSC development in the early postnatal mouse dentate gyrus. We identified two sequential steps, first a transition to quiescence followed by further maturation, each of which involved distinct changes in metabolic gene expression. Direct metabolic analysis uncovered distinct milestones, including an autophagy burst before NSC quiescence acquisition and cellular reactive oxygen species level elevation along NSC maturation. Functionally, autophagy is important for the NSC transition to quiescence during early postnatal development. Together, our study reveals a multi-step process with defined milestones underlying establishment of the adult NSC pool in the mammalian brain.

µPhos: a scalable and sensitive platform for high-dimensional phosphoproteomics

Mass spectrometry has revolutionized cell signaling research by vastly simplifying the analysis of many thousands of phosphorylation sites in the human proteome. Defining the cellular response to perturbations is crucial for further illuminating the functionality of the phosphoproteome. Here we describe µPhos ('microPhos'), an accessible phosphoproteomics platform that permits phosphopeptide enrichment from 96-well cell culture and small tissue amounts in <8 h total processing time. By greatly minimizing transfer steps and liquid volumes, we demonstrate increased sensitivity, >90% selectivity, and excellent quantitative reproducibility. Employing highly sensitive trapped ion mobility mass spectrometry, we quantify ~17,000 Class I phosphosites in a human cancer cell line using 20 µg starting material, and confidently localize ~6200 phosphosites from 1 µg. This depth covers key signaling pathways, rendering sample-limited applications and perturbation experiments with hundreds of samples viable. We employ µPhos to study drug- and time-dependent response signatures in a leukemia cell line, and by quantifying 30,000 Class I phosphosites in the mouse brain we reveal distinct spatial kinase activities in subregions of the hippocampal formation.

Alternative splicing coupled to nonsense-mediated decay coordinates downregulation of non-neuronal genes in developing mouse neurons

BACKGROUND

The functional coupling between alternative pre-mRNA splicing (AS) and the mRNA quality control mechanism called nonsense-mediated decay (NMD) can modulate transcript abundance. Previous studies have identified several examples of such a regulation in developing neurons. However, the systems-level effects of AS-NMD in this context are poorly understood.

RESULTS

We developed an R package, factR2, which offers a comprehensive suite of AS-NMD analysis functions. Using this tool, we conducted a longitudinal analysis of gene expression in pluripotent stem cells undergoing induced neuronal differentiation. Our analysis uncovers hundreds of AS-NMD events with significant potential to regulate gene expression. Notably, this regulation is significantly overrepresented in specific functional groups of developmentally downregulated genes. Particularly strong association with gene downregulation is detected for alternative cassette exons stimulating NMD upon their inclusion into mature mRNA. By combining bioinformatic analyses with CRISPR/Cas9 genome editing and other experimental approaches we show that NMD-stimulating cassette exons regulated by the RNA-binding protein PTBP1 dampen the expression of their genes in developing neurons. We also provided evidence that the inclusion of NMD-stimulating cassette exons into mature mRNAs is temporally coordinated with NMD-independent gene repression mechanisms.

CONCLUSIONS

Our study provides an accessible workflow for the discovery and prioritization of AS-NMD targets. It further argues that the AS-NMD pathway plays a widespread role in developing neurons by facilitating the downregulation of functionally related non-neuronal genes.

Traumatic brain injury promotes neurogenesis at the cost of astrogliogenesis in the adult hippocampus of male mice

Traumatic brain injury (TBI) can result in long-lasting changes in hippocampal function. The changes induced by TBI on the hippocampus contribute to cognitive deficits. The adult hippocampus harbors neural stem cells (NSCs) that generate neurons (neurogenesis), and astrocytes (astrogliogenesis). While deregulation of hippocampal NSCs and neurogenesis have been observed after TBI, it is not known how TBI may affect hippocampal astrogliogenesis. Using a controlled cortical impact model of TBI in male mice, single cell RNA sequencing and spatial transcriptomics, we assessed how TBI affected hippocampal NSCs and the neuronal and astroglial lineages derived from them. We observe an increase in NSC-derived neuronal cells and a concomitant decrease in NSC-derived astrocytic cells, together with changes in gene expression and cell dysplasia within the dentate gyrus. Here, we show that TBI modifies NSC fate to promote neurogenesis at the cost of astrogliogenesis and identify specific cell populations as possible targets to counteract TBI-induced cellular changes in the adult hippocampus.

The meninges host a unique compartment of regulatory T cells that bulwarks adult hippocampal neurogenesis

Our knowledge about the meningeal immune system has recently burgeoned, particularly our understanding of how innate and adaptive effector cells are mobilized to meet brain challenges. However, information on how meningeal immunocytes guard brain homeostasis in healthy individuals remains sparse. This study highlights the heterogeneous and polyfunctional regulatory-T (Treg) cell compartment in the meninges. A Treg subtype specialized in controlling Th1-cell responses and another known to control responses in B-cell follicles were substantial components of this compartment, foretelling that punctual Treg-cell ablation rapidly unleashed interferon-gamma production by meningeal lymphocytes, unlocked their access to the brain parenchyma, and altered meningeal B-cell profiles. Distally, the hippocampus assumed a reactive state, with morphological and transcriptional changes in multiple glial-cell types; within the dentate gyrus, neural stem cells showed exacerbated death and desisted from further differentiation, associated with inhibition of spatial-reference memory. Thus, meningeal Treg cells are a multifaceted bulwark to brain homeostasis at steady-state.

One sentence summary

A distinct population of regulatory T cells in the murine meninges safeguards homeostasis by keeping local interferon-γ-producing lymphocytes in check, thereby preventing their invasion of the parenchyma, activation of hippocampal glial cells, death of neural stem cells, and memory decay.

Comparing the effects of irradiation with protons or photons on neonatal mouse brain: Apoptosis, oncogenesis and hippocampal alterations

BACKGROUND AND PURPOSE

Medulloblastoma (MB) is a common primary brain cancer in children. Proton therapy in pediatric MB is intensively studied and widely adopted. Compared to photon, proton radiations offer potential for reduced toxicity due to the characteristic Bragg Peak at the end of their path in tissue. The aim of this study was to compare the effects of irradiation with the same dose of protons or photons in Patched1 heterozygous knockout mice, a murine model predisposed to cancer and non-cancer radiogenic pathologies, including MB and lens opacity.

MATERIALS AND METHODS

TOP-IMPLART is a pulsed linear proton accelerator for proton therapy applications. We compared the long-term health effects of 3 Gy of protons or photons in neonatal mice exposed at postnatal day 2, during a peculiarly susceptible developmental phase of the cerebellum, lens, and hippocampus, to genotoxic stress.

RESULTS

Experimental testing of the 5 mm Spread-Out Bragg Peak (SOBP) proton beam, through evaluation of apoptotic response, confirmed that both cerebellum and hippocampus were within the SOBP irradiation field. While no differences in MB induction were observed after irradiation with protons or photons, lens opacity examination confirmed sparing of the lens after proton exposure. Marked differences in expression of neurogenesis-related genes and in neuroinflammation, but not in hippocampal neurogenesis, were observed after irradiation of wild-type mice with both radiation types.

CONCLUSION

In-vivo experiments with radiosensitive mouse models improve our mechanistic understanding of the dependence of brain damage on radiation quality, thus having important implications in translational research.

Single-cell long-read sequencing-based mapping reveals specialized splicing patterns in developing and adult mouse and human brain

RNA isoforms influence cell identity and function. However, a comprehensive brain isoform map was lacking. We analyze single-cell RNA isoforms across brain regions, cell subtypes, developmental time points and species. For 72% of genes, full-length isoform expression varies along one or more axes. Splicing, transcription start and polyadenylation sites vary strongly between cell types, influence protein architecture and associate with disease-linked variation. Additionally, neurotransmitter transport and synapse turnover genes harbor cell-type variability across anatomical regions. Regulation of cell-type-specific splicing is pronounced in the postnatal day 21-to-postnatal day 28 adolescent transition. Developmental isoform regulation is stronger than regional regulation for the same cell type. Cell-type-specific isoform regulation in mice is mostly maintained in the human hippocampus, allowing extrapolation to the human brain. Conversely, the human brain harbors additional cell-type specificity, suggesting gain-of-function isoforms. Together, this detailed single-cell atlas of full-length isoform regulation across development, anatomical regions and species reveals an unappreciated degree of isoform variability across multiple axes.

Identifying gene expression programs in single-cell RNA-seq data using linear correlation explanation

OBJECTIVE

Gene expression analysis through single-cell RNA sequencing (scRNA-seq) has revolutionized our understanding of gene regulation in diverse cell types, tissues, and organisms. While existing methods primarily focus on identifying cell type-specific gene expression programs (GEPs), the characterization of GEPs associated with biological processes and stimuli responses remains limited. In this study, we aim to infer biologically meaningful GEPs that are associated with both cellular phenotypes and activity programs directly from scRNA-seq data.

METHODS

We applied linear CorEx, a machine-learning-based approach, to infer GEPs by grouping genes based on total correlation optimization function in simulated and real-world scRNA-seq datasets. Additionally, we utilized a transfer learning approach to project CorEx-inferred GEPs to other scRNA-seq datasets.

RESULTS

By leveraging total correlation optimization, linear CorEx groups genes and demonstrates superior performance in identifying cell types and activity programs compared to similar methods using simulated data. Furthermore, we apply this same approach to real-world scRNA-seq data from the mouse dentate gyrus and embryonic colon development, uncovering biologically relevant GEPs related to cell types, developmental ages, and cell cycle programs. We also demonstrate the potential for transfer learning by evaluating similar datasets, showcasing the cross-species sensitivity of linear CorEx.

CONCLUSION

Our findings validate linear CorEx as a valuable tool for comprehensively analyzing complex signals in scRNA-seq data, leading to deeper insights into gene expression dynamics, cellular heterogeneity, and regulatory mechanisms.

Adult Hippocampal Neurogenesis in the Human Brain: Updates, Challenges, and Perspectives

The existence of neurogenesis in the adult human hippocampus has been under considerable debate within the past three decades due to the diverging conclusions originating mostly from immunohistochemistry studies. While some of these reports conclude that hippocampal neurogenesis in humans occurs throughout physiologic aging, others indicate that this phenomenon ends by early childhood. More recently, some groups have adopted next-generation sequencing technologies to characterize with more acuity the extent of this phenomenon in humans. Here, we review the current state of research on adult hippocampal neurogenesis in the human brain with an emphasis on the challenges and limitations of using immunohistochemistry and next-generation sequencing technologies for its study.

Causal gene regulatory analysis with RNA velocity reveals an interplay between slow and fast transcription factors

Single-cell expression dynamics, from differentiation trajectories or RNA velocity, have the potential to reveal causal links between transcription factors (TFs) and their target genes in gene regulatory networks (GRNs). However, existing methods either overlook these expression dynamics or necessitate that cells be ordered along a linear pseudotemporal axis, which is incompatible with branching trajectories. We introduce Velorama, an approach to causal GRN inference that represents single-cell differentiation dynamics as a directed acyclic graph of cells, constructed from pseudotime or RNA velocity measurements. Additionally, Velorama enables the estimation of the speed at which TFs influence target genes. Applying Velorama, we uncover evidence that the speed of a TF's interactions is tied to its regulatory function. For human corticogenesis, we find that slow TFs are linked to gliomas, while fast TFs are associated with neuropsychiatric diseases. We expect Velorama to become a critical part of the RNA velocity toolkit for investigating the causal drivers of differentiation and disease.

Reconstructing developmental trajectories using latent dynamical systems and time-resolved transcriptomics

The snapshot nature of single-cell transcriptomics presents a challenge for studying the dynamics of cell fate decisions. Metabolic labeling and splicing can provide temporal information at single-cell level, but current methods have limitations. Here, we present a framework that overcomes these limitations: experimentally, we developed sci-FATE2, an optimized method for metabolic labeling with increased data quality, which we used to profile 45,000 embryonic stem (ES) cells differentiating into neural tube identities. Computationally, we developed a two-stage framework for dynamical modeling: VelvetVAE, a variational autoencoder (VAE) for velocity inference that outperforms all other tools tested, and VelvetSDE, a neural stochastic differential equation (nSDE) framework for simulating trajectory distributions. These recapitulate underlying dataset distributions and capture features such as decision boundaries between alternative fates and fate-specific gene expression. These methods recast single-cell analyses from descriptions of observed data to models of the dynamics that generated them, providing a framework for investigating developmental fate decisions.

Spatiotemporal analysis of gene expression in the human dentate gyrus reveals age-associated changes in cellular maturation and neuroinflammation

The dentate gyrus of the hippocampus is important for many cognitive functions, including learning, memory, and mood. Here, we investigated age-associated changes in transcriptome-wide spatial gene expression in the human dentate gyrus across the lifespan. Genes associated with neurogenesis and the extracellular matrix were enriched in infants, while gene markers of inhibitory neurons and cell proliferation showed increases and decreases in post-infancy, respectively. While we did not find evidence for neural proliferation post-infancy, we did identify molecular signatures supporting protracted maturation of granule cells. We also identified a wide-spread hippocampal aging signature and an age-associated increase in genes related to neuroinflammation. Our findings suggest major changes to the putative neurogenic niche after infancy and identify molecular foci of brain aging in glial and neuropil enriched tissue.

Analgesic candidate adenosine A3 receptors are expressed by perineuronal peripheral macrophages in human dorsal root ganglion and spinal cord microglia

ABSTRACT

Adenosine receptors are a family of purinergic G protein-coupled receptors that are widely distributed in bodily organs and in the peripheral and central nervous systems. Recently, antihyperalgesic actions have been suggested for the adenosine A3 receptor, and its agonists have been proposed as new neuropathic pain treatments. We hypothesized that these receptors may be expressed in nociceptive primary afferent neurons. However, RNA sequencing across species, eg, rat, mouse, dog, and human, suggests that dorsal root ganglion (DRG) expression of ADORA3 is inconsistent. In rat and mouse, Adora3 shows very weak to no expression in DRG, whereas it is well expressed in human DRG. However, the cell types in human DRG that express ADORA3 have not been delineated. An examination of DRG cell types using in situ hybridization clearly detected ADORA3 transcripts in peripheral macrophages that are in close apposition to the neuronal perikarya but not in peripheral sensory neurons. By contrast, ADORA1 was found primarily in neurons, where it is broadly expressed at low levels. These results suggest that a more complex or indirect mechanism involving modulation of macrophage and/or microglial cells may underlie the potential analgesic action of adenosine A3 receptor agonism.

Dissection and integration of bursty transcriptional dynamics for complex systems

RNA velocity estimation is a potentially powerful tool to reveal the directionality of transcriptional changes in single-cell RNA-sequencing data, but it lacks accuracy, absent advanced metabolic labeling techniques. We developed an approach, TopicVelo, that disentangles simultaneous, yet distinct, dynamics by using a probabilistic topic model, a highly interpretable form of latent space factorization, to infer cells and genes associated with individual processes, thereby capturing cellular pluripotency or multifaceted functionality. Focusing on process-associated cells and genes enables accurate estimation of process-specific velocities via a master equation for a transcriptional burst model accounting for intrinsic stochasticity. The method obtains a global transition matrix by leveraging cell topic weights to integrate process-specific signals. In challenging systems, this method accurately recovers complex transitions and terminal states, while our use of first-passage time analysis provides insights into transient transitions. These results expand the limits of RNA velocity, empowering future studies of cell fate and functional responses.

RNA velocity prediction via neural ordinary differential equation

RNA velocity is a crucial tool for unraveling the trajectory of cellular responses. Several approaches, including ordinary differential equations and machine learning models, have been proposed to interpret velocity. However, the practicality of these methods is constrained by underlying assumptions. In this study, we introduce SymVelo, a dual-path framework that effectively integrates high- and low-dimensional information. Rigorous benchmarking and extensive studies demonstrate that SymVelo is capable of inferring differentiation trajectories in developing organs, analyzing gene responses to stimulation, and uncovering transcription dynamics. Moreover, the adaptable architecture of SymVelo enables customization to accommodate intricate data and diverse modalities in forthcoming research, thereby providing a promising avenue for advancing our understanding of cellular behavior.

Reconstruction of macroglia and adult neurogenesis evolution through cross-species single-cell transcriptomic analyses

Macroglia fulfill essential functions in the adult vertebrate brain, producing and maintaining neurons and regulating neuronal communication. However, we still know little about their emergence and diversification. We used the zebrafish D. rerio as a distant vertebrate model with moderate glial diversity as anchor to reanalyze datasets covering over 600 million years of evolution. We identify core features of adult neurogenesis and innovations in the mammalian lineage with a potential link to the rarity of radial glia-like cells in adult humans. Our results also suggest that functions associated with astrocytes originated in a multifunctional cell type fulfilling both neural stem cell and astrocytic functions before these diverged. Finally, we identify conserved elements of macroglial cell identity and function and their time of emergence during evolution.

Deciphering molecular heterogeneity and dynamics of human hippocampal neural stem cells at different ages and injury states

While accumulated publications support the existence of neurogenesis in the adult human hippocampus, the homeostasis and developmental potentials of neural stem cells (NSCs) under different contexts remain unclear. Based on our generated single-nucleus atlas of the human hippocampus across neonatal, adult, aging, and injury, we dissected the molecular heterogeneity and transcriptional dynamics of human hippocampal NSCs under different contexts. We further identified new specific neurogenic lineage markers that overcome the lack of specificity found in some well-known markers. Based on developmental trajectory and molecular signatures, we found that a subset of NSCs exhibit quiescent properties after birth, and most NSCs become deep quiescence during aging. Furthermore, certain deep quiescent NSCs are reactivated following stroke injury. Together, our findings provide valuable insights into the development, aging, and reactivation of the human hippocampal NSCs, and help to explain why adult hippocampal neurogenesis is infrequently observed in humans.

T-Type Ca2+ Channels Mediate a Critical Period of Plasticity in Adult-Born Granule Cells

Adult-born granule cells (abGCs) exhibit a transient period of elevated synaptic plasticity that plays an important role in hippocampal function. Various mechanisms have been implicated in this critical period for enhanced plasticity, including minimal GABAergic inhibition and high intrinsic excitability conferred by T-type Ca2+ channels. Here we assess the contribution of synaptic inhibition and intrinsic excitability to long-term potentiation (LTP) in abGCs of adult male and female mice using perforated patch recordings. We show that the timing of critical period plasticity is unaffected by intact GABAergic inhibition such that 4-6-week-old abGCs exhibit LTP that is absent by 8 weeks. Blocking GABAA receptors, or partial blockade of GABA release from PV and nNos-expressing interneurons by a µ-opioid receptor agonist, strongly enhances LTP in 4-week-old GCs, suggesting that minimal inhibition does not underlie critical period plasticity. Instead, the closure of the critical period coincides with a reduction in the contribution of T-type Ca2+ channels to intrinsic excitability, and a selective T-type Ca2+ channel antagonist prevents LTP in 4-week-old but not mature GCs. Interestingly, whole-cell recordings that facilitate T-type Ca2+ channel activity in mature GCs unmasks LTP (with inhibition intact) that is also sensitive to a T-type Ca2+ channel antagonist, suggesting T-type channel activity in mature GCs is suppressed by native intracellular signaling. Together these results show that abGCs use T-type Ca2+ channels to overcome inhibition, providing new insight into how high intrinsic excitability provides young abGCs a competitive advantage for experience-dependent synaptic plasticity.

Lifelong persistence of nuclear RNAs in the mouse brain

Genomic DNA that resides in the nuclei of mammalian neurons can be as old as the organism itself. The life span of nuclear RNAs, which are critical for proper chromatin architecture and transcription regulation, has not been determined in adult tissues. In this work, we identified and characterized nuclear RNAs that do not turn over for at least 2 years in a subset of postnatally born cells in the mouse brain. These long-lived RNAs were stably retained in nuclei in a neural cell type-specific manner and were required for the maintenance of heterochromatin. Thus, the life span of neural cells may depend on both the molecular longevity of DNA for the storage of genetic information and also the extreme stability of RNA for the functional organization of chromatin.

Topological and geometric analysis of cell states in single-cell transcriptomic data

Single-cell RNA sequencing (scRNA-seq) enables dissecting cellular heterogeneity in tissues, resulting in numerous biological discoveries. Various computational methods have been devised to delineate cell types by clustering scRNA-seq data, where clusters are often annotated using prior knowledge of marker genes. In addition to identifying pure cell types, several methods have been developed to identify cells undergoing state transitions, which often rely on prior clustering results. The present computational approaches predominantly investigate the local and first-order structures of scRNA-seq data using graph representations, while scRNA-seq data frequently display complex high-dimensional structures. Here, we introduce scGeom, a tool that exploits the multiscale and multidimensional structures in scRNA-seq data by analyzing the geometry and topology through curvature and persistent homology of both cell and gene networks. We demonstrate the utility of these structural features to reflect biological properties and functions in several applications, where we show that curvatures and topological signatures of cell and gene networks can help indicate transition cells and the differentiation potential of cells. We also illustrate that structural characteristics can improve the classification of cell types.

Interrogations of single-cell RNA splicing landscapes with SCASL define new cell identities with physiological relevance

RNA splicing shapes the gene regulatory programs that underlie various physiological and disease processes. Here, we present the SCASL (single-cell clustering based on alternative splicing landscapes) method for interrogating the heterogeneity of RNA splicing with single-cell RNA-seq data. SCASL resolves the issue of biased and sparse data coverage on single-cell RNA splicing and provides a new scheme for classifications of cell identities. With previously published datasets as examples, SCASL identifies new cell clusters indicating potentially precancerous and early-tumor stages in triple-negative breast cancer, illustrates cell lineages of embryonic liver development, and provides fine clusters of highly heterogeneous tumor-associated CD4 and CD8 T cells with functional and physiological relevance. Most of these findings are not readily available via conventional cell clustering based on single-cell gene expression data. Our study shows the potential of SCASL in revealing the intrinsic RNA splicing heterogeneity and generating biological insights into the dynamic and functional cell landscapes in complex tissues.

Thyroid hormone action in adult neurogliogenic niches: the known and unknown

Over the last decades, thyroid hormones (THs) signaling has been established as a key signaling cue for the proper maintenance of brain functions in adult mammals, including humans. One of the most fascinating roles of THs in the mature mammalian brain is their ability to regulate adult neurogliogenic processes. In this respect, THs control the generation of new neuronal and glial progenitors from neural stem cells (NSCs) as well as their final differentiation and maturation programs. In this review, we summarize current knowledge on the cellular organization of adult rodent neurogliogenic niches encompassing well-established niches in the subventricular zone (SVZ) lining the lateral ventricles, the hippocampal subgranular zone (SGZ), and the hypothalamus, but also less characterized niches in the striatum and the cerebral cortex. We then discuss critical questions regarding how THs availability is regulated in the respective niches in rodents and larger mammals as well as how modulating THs availability in those niches interferes with lineage decision and progression at the molecular, cellular, and functional levels. Based on those alterations, we explore the novel therapeutic avenues aiming at harnessing THs regulatory influences on neurogliogenic output to stimulate repair processes by influencing the generation of either new neurons (i.e. Alzheimer's, Parkinson's diseases), oligodendrocytes (multiple sclerosis) or both (stroke). Finally, we point out future challenges, which will shape research in this exciting field in the upcoming years.

Effects of electroconvulsive shock on the function, circuitry, and transcriptome of dentate gyrus granule neurons

Therapeutic use of electroconvulsive shock (ECS) is 75% effective for the remission of treatment-resistant depression. Like other more common forms of antidepressant treatment such as fluoxetine, ECS has been shown to increase neurogenesis in the hippocampal dentate gyrus of rodent models. Yet the question of how ECS-induced neurogenesis supports improvement of depressive symptoms remains unknown. Here, we show that ECS-induced neurogenesis is necessary to improve depressive-like behavior of mice exposed to chronic corticosterone (Cort). We then use slice electrophysiology to show that optogenetic stimulation of adult-born neurons produces a greater hyperpolarization in mature granule neurons after ECS vs Sham treatment. We identify that this hyperpolarization requires the activation of metabotropic glutamate receptor 2 (mGluR2). Consistent with this finding, we observe reduced expression of the immediate early gene cFos in the granule cell layer of ECS vs Sham subjects. We then show that mGluR2 knockdown specifically in ventral granule neurons blunts the antidepressant-like behavioral effects of ECS. Using single nucleus RNA sequencing, we reveal major transcriptomic shifts in granule neurons after treatment with ECS+Cort or fluoxetine+Cort vs Cort alone. We identify a population of immature cells which has greater representation in both ECS+Cort and fluoxetine+Cort treated samples vs Cort alone. We also find global differences in ECS-vs fluoxetine-induced transcriptomic shifts. Together, these findings highlight a critical role for immature granule cells and mGluR2 signaling in the antidepressant action of ECS.

Genetic labeling of embryonically-born dentate granule neurons in young mice using the PenkCre mouse line

The dentate gyrus (DG) of the hippocampus is a mosaic of dentate granule neurons (DGNs) accumulated throughout life. While many studies focused on the morpho-functional properties of adult-born DGNs, much less is known about DGNs generated during development, and in particular those born during embryogenesis. One of the main reasons for this gap is the lack of methods available to specifically label and manipulate embryonically-born DGNs. Here, we have assessed the relevance of the PenkCre mouse line as a genetic model to target this embryonically-born population. In young animals, PenkCre expression allows to tag neurons in the DG with positional, morphological and electrophysiological properties characteristic of DGNs born during the embryonic period. In addition, PenkCre+ cells in the DG are distributed in both blades along the entire septo-temporal axis. This model thus offers new possibilities to explore the functions of this underexplored population of embryonically-born DGNs.

Statistical method scDEED for detecting dubious 2D single-cell embeddings and optimizing t-SNE and UMAP hyperparameters

Two-dimensional (2D) embedding methods are crucial for single-cell data visualization. Popular methods such as t-distributed stochastic neighbor embedding (t-SNE) and uniform manifold approximation and projection (UMAP) are commonly used for visualizing cell clusters; however, it is well known that t-SNE and UMAP's 2D embeddings might not reliably inform the similarities among cell clusters. Motivated by this challenge, we present a statistical method, scDEED, for detecting dubious cell embeddings output by a 2D-embedding method. By calculating a reliability score for every cell embedding based on the similarity between the cell's 2D-embedding neighbors and pre-embedding neighbors, scDEED identifies the cell embeddings with low reliability scores as dubious and those with high reliability scores as trustworthy. Moreover, by minimizing the number of dubious cell embeddings, scDEED provides intuitive guidance for optimizing the hyperparameters of an embedding method. We show the effectiveness of scDEED on multiple datasets for detecting dubious cell embeddings and optimizing the hyperparameters of t-SNE and UMAP.

Quantifying the clusterness and trajectoriness of single-cell RNA-seq data

Among existing computational algorithms for single-cell RNA-seq analysis, clustering and trajectory inference are two major types of analysis that are routinely applied. For a given dataset, clustering and trajectory inference can generate vastly different visualizations that lead to very different interpretations of the data. To address this issue, we propose multiple scores to quantify the "clusterness" and "trajectoriness" of single-cell RNA-seq data, in other words, whether the data looks like a collection of distinct clusters or a continuum of progression trajectory. The scores we introduce are based on pairwise distance distribution, persistent homology, vector magnitude, Ripley's K, and degrees of connectivity. Using simulated datasets, we demonstrate that the proposed scores are able to effectively differentiate between cluster-like data and trajectory-like data. Using real single-cell RNA-seq datasets, we demonstrate the scores can serve as indicators of whether clustering analysis or trajectory inference is a more appropriate choice for biological interpretation of the data.

Generation of adult hippocampal neural stem cells occurs in the early postnatal dentate gyrus and depends on cyclin D2

Lifelong hippocampal neurogenesis is maintained by a pool of multipotent adult neural stem cells (aNSCs) residing in the subgranular zone of the dentate gyrus (DG). The mechanisms guiding transition of NSCs from the developmental to the adult state remain unclear. We show here, by using nestin-based reporter mice deficient for cyclin D2, that the aNSC pool is established through cyclin D2-dependent proliferation during the first two weeks of life. The absence of cyclin D2 does not affect normal development of the dentate gyrus until birth but prevents postnatal formation of radial glia-like aNSCs. Furthermore, retroviral fate mapping reveals that aNSCs are born on-site from precursors located in the dentate gyrus shortly after birth. Taken together, our data identify the critical time window and the spatial location of the precursor divisions that generate the persistent population of aNSCs and demonstrate the central role of cyclin D2 in this process.

DeepVelo: deep learning extends RNA velocity to multi-lineage systems with cell-specific kinetics

Existing RNA velocity estimation methods strongly rely on predefined dynamics and cell-agnostic constant transcriptional kinetic rates, assumptions often violated in complex and heterogeneous single-cell RNA sequencing (scRNA-seq) data. Using a graph convolution network, DeepVelo overcomes these limitations by generalizing RNA velocity to cell populations containing time-dependent kinetics and multiple lineages. DeepVelo infers time-varying cellular rates of transcription, splicing, and degradation, recovers each cell's stage in the differentiation process, and detects functionally relevant driver genes regulating these processes. Application to various developmental and pathogenic processes demonstrates DeepVelo's capacity to study complex differentiation and lineage decision events in heterogeneous scRNA-seq data.

Transcriptional dynamics orchestrating the development and integration of neurons born in the adult hippocampus

The adult hippocampus generates new granule cells (aGCs) that exhibit distinct functional capabilities along development, conveying a unique form of plasticity to the preexisting circuits. While early differentiation of adult radial glia-like neural stem cells (RGL) has been studied extensively, the molecular mechanisms guiding the maturation of postmitotic neurons remain unknown. Here, we used a precise birthdating strategy to follow newborn aGCs along differentiation using single-nuclei RNA sequencing (snRNA-seq). Transcriptional profiling revealed a continuous trajectory from RGLs to mature aGCs, with multiple sequential immature stages bearing increasing levels of effector genes supporting growth, excitability and synaptogenesis. Remarkably, four discrete cellular states were defined by the expression of distinct sets of transcription factors (TFs): quiescent neural stem cells, proliferative progenitors, postmitotic immature aGCs, and mature aGCs. The transition from immature to mature aCGs involved a transcriptional switch that shutdown molecular cascades promoting cell growth, such as the SoxC family of TFs, to activate programs controlling neuronal homeostasis. Indeed, aGCs overexpressing Sox4 or Sox11 remained stalled at the immature state. Our results unveil precise molecular mechanisms driving adult neural stem cells through the pathway of neuronal differentiation.

ZEBRA: a hierarchically integrated gene expression atlas of the murine and human brain at single-cell resolution

The molecular causes and mechanisms of neurodegenerative diseases remain poorly understood. A growing number of single-cell studies have implicated various neural, glial, and immune cell subtypes to affect the mammalian central nervous system in many age-related disorders. Integrating this body of transcriptomic evidence into a comprehensive and reproducible framework poses several computational challenges. Here, we introduce ZEBRA, a large single-cell and single-nucleus RNA-seq database. ZEBRA integrates and normalizes gene expression and metadata from 33 studies, encompassing 4.2 million human and mouse brain cells sampled from 39 brain regions. It incorporates samples from patients with neurodegenerative diseases like Alzheimer's disease, Parkinson's disease, and Multiple sclerosis, as well as samples from relevant mouse models. We employed scVI, a deep probabilistic auto-encoder model, to integrate the samples and curated both cell and sample metadata for downstream analysis. ZEBRA allows for cell-type and disease-specific markers to be explored and compared between sample conditions and brain regions, a cell composition analysis, and gene-wise feature mappings. Our comprehensive molecular database facilitates the generation of data-driven hypotheses, enhancing our understanding of mammalian brain function during aging and disease. The data sets, along with an interactive database are freely available at https://www.ccb.uni-saarland.de/zebra.

STAB2: an updated spatio-temporal cell atlas of the human and mouse brain

The brain is constituted of heterogeneous types of neuronal and non-neuronal cells, which are organized into distinct anatomical regions, and show precise regulation of gene expression during development, aging and function. In the current database release, STAB2 provides a systematic cellular map of the human and mouse brain by integrating recently published large-scale single-cell and single-nucleus RNA-sequencing datasets from diverse regions and across lifespan. We applied a hierarchical strategy of unsupervised clustering on the integrated single-cell transcriptomic datasets to precisely annotate the cell types and subtypes in the human and mouse brain. Currently, STAB2 includes 71 and 61 different cell subtypes defined in the human and mouse brain, respectively. It covers 63 subregions and 15 developmental stages of human brain, and 38 subregions and 30 developmental stages of mouse brain, generating a comprehensive atlas for exploring spatiotemporal transcriptomic dynamics in the mammalian brain. We also augmented web interfaces for querying and visualizing the gene expression in specific cell types. STAB2 is freely available at https://mai.fudan.edu.cn/stab2.

A relay velocity model infers cell-dependent RNA velocity

RNA velocity provides an approach for inferring cellular state transitions from single-cell RNA sequencing (scRNA-seq) data. Conventional RNA velocity models infer universal kinetics from all cells in an scRNA-seq experiment, resulting in unpredictable performance in experiments with multi-stage and/or multi-lineage transition of cell states where the assumption of the same kinetic rates for all cells no longer holds. Here we present cellDancer, a scalable deep neural network that locally infers velocity for each cell from its neighbors and then relays a series of local velocities to provide single-cell resolution inference of velocity kinetics. In the simulation benchmark, cellDancer shows robust performance in multiple kinetic regimes, high dropout ratio datasets and sparse datasets. We show that cellDancer overcomes the limitations of existing RNA velocity models in modeling erythroid maturation and hippocampus development. Moreover, cellDancer provides cell-specific predictions of transcription, splicing and degradation rates, which we identify as potential indicators of cell fate in the mouse pancreas.

Deep generative modeling of transcriptional dynamics for RNA velocity analysis in single cells

RNA velocity has been rapidly adopted to guide interpretation of transcriptional dynamics in snapshot single-cell data; however, current approaches for estimating RNA velocity lack effective strategies for quantifying uncertainty and determining the overall applicability to the system of interest. Here, we present veloVI (velocity variational inference), a deep generative modeling framework for estimating RNA velocity. veloVI learns a gene-specific dynamical model of RNA metabolism and provides a transcriptome-wide quantification of velocity uncertainty. We show that veloVI compares favorably to previous approaches with respect to goodness of fit, consistency across transcriptionally similar cells and stability across preprocessing pipelines for quantifying RNA abundance. Further, we demonstrate that veloVI's posterior velocity uncertainty can be used to assess whether velocity analysis is appropriate for a given dataset. Finally, we highlight veloVI as a flexible framework for modeling transcriptional dynamics by adapting the underlying dynamical model to use time-dependent transcription rates.

Automatic cell-type harmonization and integration across Human Cell Atlas datasets

Harmonizing cell types across the single-cell community and assembling them into a common framework is central to building a standardized Human Cell Atlas. Here, we present CellHint, a predictive clustering tree-based tool to resolve cell-type differences in annotation resolution and technical biases across datasets. CellHint accurately quantifies cell-cell transcriptomic similarities and places cell types into a relationship graph that hierarchically defines shared and unique cell subtypes. Application to multiple immune datasets recapitulates expert-curated annotations. CellHint also reveals underexplored relationships between healthy and diseased lung cell states in eight diseases. Furthermore, we present a workflow for fast cross-dataset integration guided by harmonized cell types and cell hierarchy, which uncovers underappreciated cell types in adult human hippocampus. Finally, we apply CellHint to 12 tissues from 38 datasets, providing a deeply curated cross-tissue database with ∼3.7 million cells and various machine learning models for automatic cell annotation across human tissues.

GID complex regulates the differentiation of neural stem cells by destabilizing TET2

Brain development requires a delicate balance between self-renewal and differentiation in neural stem cells (NSC), which rely on the precise regulation of gene expression. Ten-eleven translocation 2 (TET2) modulates gene expression by the hydroxymethylation of 5-methylcytosine in DNA as an important epigenetic factor and participates in the neuronal differentiation. Yet, the regulation of TET2 in the process of neuronal differentiation remains unknown. Here, the protein level of TET2 was reduced by the ubiquitin-proteasome pathway during NSC differentiation, in contrast to mRNA level. We identified that TET2 physically interacts with the core subunits of the glucose-induced degradation-deficient (GID) ubiquitin ligase complex, an evolutionarily conserved ubiquitin ligase complex and is ubiquitinated by itself. The protein levels of GID complex subunits increased reciprocally with TET2 level upon NSC differentiation. The silencing of the core subunits of the GID complex, including WDR26 and ARMC8, attenuated the ubiquitination and degradation of TET2, increased the global 5-hydroxymethylcytosine levels, and promoted the differentiation of the NSC. TET2 level increased in the brain of the Wdr26+/- mice. Our results illustrated that the GID complex negatively regulates TET2 protein stability, further modulates NSC differentiation, and represents a novel regulatory mechanism involved in brain development.

A high-resolution transcriptomic and spatial atlas of cell types in the whole mouse brain

The mammalian brain consists of millions to billions of cells that are organized into many cell types with specific spatial distribution patterns and structural and functional properties1-3. Here we report a comprehensive and high-resolution transcriptomic and spatial cell-type atlas for the whole adult mouse brain. The cell-type atlas was created by combining a single-cell RNA-sequencing (scRNA-seq) dataset of around 7 million cells profiled (approximately 4.0 million cells passing quality control), and a spatial transcriptomic dataset of approximately 4.3 million cells using multiplexed error-robust fluorescence in situ hybridization (MERFISH). The atlas is hierarchically organized into 4 nested levels of classification: 34 classes, 338 subclasses, 1,201 supertypes and 5,322 clusters. We present an online platform, Allen Brain Cell Atlas, to visualize the mouse whole-brain cell-type atlas along with the single-cell RNA-sequencing and MERFISH datasets. We systematically analysed the neuronal and non-neuronal cell types across the brain and identified a high degree of correspondence between transcriptomic identity and spatial specificity for each cell type. The results reveal unique features of cell-type organization in different brain regions-in particular, a dichotomy between the dorsal and ventral parts of the brain. The dorsal part contains relatively fewer yet highly divergent neuronal types, whereas the ventral part contains more numerous neuronal types that are more closely related to each other. Our study also uncovered extraordinary diversity and heterogeneity in neurotransmitter and neuropeptide expression and co-expression patterns in different cell types. Finally, we found that transcription factors are major determinants of cell-type classification and identified a combinatorial transcription factor code that defines cell types across all parts of the brain. The whole mouse brain transcriptomic and spatial cell-type atlas establishes a benchmark reference atlas and a foundational resource for integrative investigations of cellular and circuit function, development and evolution of the mammalian brain.

Adversarial training improves model interpretability in single-cell RNA-seq analysis

Motivation

Predictive computational models must be accurate, robust, and interpretable to be considered reliable in important areas such as biology and medicine. A sufficiently robust model should not have its output affected significantly by a slight change in the input. Also, these models should be able to explain how a decision is made to support user trust in the results. Efforts have been made to improve the robustness and interpretability of predictive computational models independently; however, the interaction of robustness and interpretability is poorly understood.

Results

As an example task, we explore the computational prediction of cell type based on single-cell RNA-seq data and show that it can be made more robust by adversarially training a deep learning model. Surprisingly, we find this also leads to improved model interpretability, as measured by identifying genes important for classification using a range of standard interpretability methods. Our results suggest that adversarial training may be generally useful to improve deep learning robustness and interpretability and that it should be evaluated on a range of tasks.

Availability and implementation

Our Python implementation of all analysis in this publication can be found at: https://github.com/MehrshadSD/robustness-interpretability. The analysis was conducted using numPy 0.2.5, pandas 2.0.3, scanpy 1.9.3, tensorflow 2.10.0, matplotlib 3.7.1, seaborn 0.12.2, sklearn 1.1.1, shap 0.42.0, lime 0.2.0.1, matplotlib_venn 0.11.9.

A toolbox of astrocyte-specific, serotype-independent adeno-associated viral vectors using microRNA targeting sequences

Astrocytes, one of the most prevalent cell types in the central nervous system (CNS), are critically involved in neural function. Genetically manipulating astrocytes is an essential tool in understanding and affecting their roles. Adeno-associated viruses (AAVs) enable rapid genetic manipulation; however, astrocyte specificity of AAVs can be limited, with high off-target expression in neurons and sparsely in endothelial cells. Here, we report the development of a cassette of four copies of six miRNA targeting sequences (4x6T) which triggers transgene degradation specifically in neurons and endothelial cells. In combination with the GfaABC1D promoter, 4x6T increases astrocytic specificity of Cre with a viral reporter from <50% to >99% in multiple serotypes in mice, and confers astrocyte specificity in multiple recombinases and reporters. We also present empty vectors to add 4x6T to other cargo, independently and in Cre/Dre-dependent forms. This toolbox of AAVs allows rapid manipulation of astrocytes throughout the CNS, is compatible with different AAV serotypes, and demonstrates the efficacy of using multiplexed miRNA targeting sequences to decrease expression in multiple off-target cell populations simultaneously.

Models of KPTN-related disorder implicate mTOR signalling in cognitive and overgrowth phenotypes

KPTN-related disorder is an autosomal recessive disorder associated with germline variants in KPTN (previously known as kaptin), a component of the mTOR regulatory complex KICSTOR. To gain further insights into the pathogenesis of KPTN-related disorder, we analysed mouse knockout and human stem cell KPTN loss-of-function models. Kptn -/- mice display many of the key KPTN-related disorder phenotypes, including brain overgrowth, behavioural abnormalities, and cognitive deficits. By assessment of affected individuals, we have identified widespread cognitive deficits (n = 6) and postnatal onset of brain overgrowth (n = 19). By analysing head size data from their parents (n = 24), we have identified a previously unrecognized KPTN dosage-sensitivity, resulting in increased head circumference in heterozygous carriers of pathogenic KPTN variants. Molecular and structural analysis of Kptn-/- mice revealed pathological changes, including differences in brain size, shape and cell numbers primarily due to abnormal postnatal brain development. Both the mouse and differentiated induced pluripotent stem cell models of the disorder display transcriptional and biochemical evidence for altered mTOR pathway signalling, supporting the role of KPTN in regulating mTORC1. By treatment in our KPTN mouse model, we found that the increased mTOR signalling downstream of KPTN is rapamycin sensitive, highlighting possible therapeutic avenues with currently available mTOR inhibitors. These findings place KPTN-related disorder in the broader group of mTORC1-related disorders affecting brain structure, cognitive function and network integrity.

Neurogenesis in primates versus rodents and the value of non-human primate models

Neurogenesis, the process of generating neurons from neural stem cells, occurs during both embryonic and adult stages, with each stage possessing distinct characteristics. Dysfunction in either stage can disrupt normal neural development, impair cognitive functions, and lead to various neurological disorders. Recent technological advancements in single-cell multiomics and gene-editing have facilitated investigations into primate neurogenesis. Here, we provide a comprehensive overview of neurogenesis across rodents, non-human primates, and humans, covering embryonic development to adulthood and focusing on the conservation and diversity among species. While non-human primates, especially monkeys, serve as valuable models with closer neural resemblance to humans, we highlight the potential impacts and limitations of non-human primate models on both physiological and pathological neurogenesis research.