Data-driven fine-grained region discovery in the mouse brain with transformers.

Technologies such as spatial transcriptomics offer unique opportunities to define the spatial organization of the mouse brain. We developed an unsupervised training scheme and novel transformer-based deep learning architecture to detect spatial domains across the whole mouse brain using spatial transcriptomics data. Our model learns local representations of molecular and cellular statistical patterns which can be clustered to identify spatial domains within the brain from coarse to fine-grained. Discovered domains are spatially regular, even with several hundreds of spatial clusters. They are also consistent with existing anatomical ontologies such as the Allen Mouse Brain Common Coordinate Framework version 3 (CCFv3) and can be visually interpreted at the cell type or transcript level. We demonstrate our method can be used to identify previously uncatalogued subregions, such as in the midbrain, where we uncover gradients of inhibitory neuron complexity and abundance. Notably, these subregions cannot be discovered using other methods. We apply our method to a separate multi-animal whole-brain spatial transcriptomic dataset and show that our method can also robustly integrate spatial domains across animals.

Rabies virus-based barcoded neuroanatomy resolved by single-cell RNA and in situ sequencing.

Mapping the connectivity of diverse neuronal types provides the foundation for understanding the structure and function of neural circuits. High-throughput and low-cost neuroanatomical techniques based on RNA barcode sequencing have the potential to map circuits at cellular resolution and a brain-wide scale, but existing Sindbis virus-based techniques can only map long-range projections using anterograde tracing approaches. Rabies virus can complement anterograde tracing approaches by enabling either retrograde labeling of projection neurons or monosynaptic tracing of direct inputs to genetically targeted postsynaptic neurons. However, barcoded rabies virus has so far been only used to map non-neuronal cellular interactions in vivo and synaptic connectivity of cultured neurons. Here we combine barcoded rabies virus with single-cell and in situ sequencing to perform retrograde labeling and transsynaptic labeling in the mouse brain. We sequenced 96 retrogradely labeled cells and 295 transsynaptically labeled cells using single-cell RNA-seq, and 4130 retrogradely labeled cells and 2914 transsynaptically labeled cells in situ. We found that the transcriptomic identities of rabies virus-infected cells can be robustly identified using both single-cell RNA-seq and in situ sequencing. By associating gene expression with connectivity inferred from barcode sequencing, we distinguished long-range projecting cortical cell types from multiple cortical areas and identified cell types with converging or diverging synaptic connectivity. Combining in situ sequencing with barcoded rabies virus complements existing sequencing-based neuroanatomical techniques and provides a potential path for mapping synaptic connectivity of neuronal types at scale.

In the brain, messages are relayed from one cell to the next through intricate networks of axons and dendrites that physically interact at junctions known as synapses. Mapping out this synaptic connectivity ­ that is, exactly which neurons are connected via synapses ­ remains a major challenge. Monosynaptic tracing is a powerful approach that allows neuroscientists to explore neural networks by harnessing viruses which spread between neurons via synapses, in particular the rabies virus. This pathogen travels exclusively from 'postsynaptic' to 'presynaptic' neurons ­ from the cell that receives a message at a synapse, back to the one that sends it. A modified variant of the rabies virus can therefore be used to reveal the presynaptic cells connecting to a population of neurons in which it has been originally introduced. However, this method does not allow scientists to identify the exact postsynaptic neuron that each presynaptic cell is connected to. One way to bypass this issue is to combine monosynaptic tracing with RNA barcoding to create distinct versions of the modified rabies virus, which are then introduced into separate populations of neurons. Tracking the spread of each version allows neuroscientists to spot exactly which presynaptic cells signal to each postsynaptic neuron. So far, this approach has been used to examine synaptic connectivity in neurons grown in the laboratory, but it remains difficult to apply it to neurons in the brain. In response, Zhang, Jin et al. aimed to demonstrate how monosynaptic tracing that relies on barcoded rabies viruses could be used to dissect neural networks in the mouse brain. First, they confirmed that it was possible to accurately detect which version of the virus had spread to presynaptic neurons using both in situ and single-cell RNA sequencing. Next, they described how this information could be analysed to build models of potential neural networks, and what type of additional experiments are required for this work. Finally, they used the approach to identify neurons that tend to connect to the same postsynaptic cells and then investigated what these have in common, showing how the technique enables a finer understanding of neural circuits. Overall, the work by Zhang, Jin et al. provides a comprehensive review of the requirements and limitations associated with monosynaptic tracing experiments based on barcoded rabies viruses, as well as how the approach could be optimized in the future. This information will be of broad interest to scientists interested in mapping neural networks in the brain.

Enhancer-AAVs allow genetic access to oligodendrocytes and diverse populations of astrocytes across species.

Proper brain function requires the assembly and function of diverse populations of neurons and glia. Single cell gene expression studies have mostly focused on characterization of neuronal cell diversity; however, recent studies have revealed substantial diversity of glial cells, particularly astrocytes. To better understand glial cell types and their roles in neurobiology, we built a new suite of adeno-associated viral (AAV)-based genetic tools to enable genetic access to astrocytes and oligodendrocytes. These oligodendrocyte and astrocyte enhancer-AAVs are highly specific (usually > 95% cell type specificity) with variable expression levels, and our astrocyte enhancer-AAVs show multiple distinct expression patterns reflecting the spatial distribution of astrocyte cell types. To provide the best glial-specific functional tools, several enhancer-AAVs were: optimized for higher expression levels, shown to be functional and specific in rat and macaque, shown to maintain specific activity in epilepsy where traditional promoters changed activity, and used to drive functional transgenes in astrocytes including Cre recombinase and acetylcholine-responsive sensor iAChSnFR. The astrocyte-specific iAChSnFR revealed a clear reward-dependent acetylcholine response in astrocytes of the nucleus accumbens during reinforcement learning. Together, this collection of glial enhancer-AAVs will enable characterization of astrocyte and oligodendrocyte populations and their roles across species, disease states, and behavioral epochs.

Rabies virus-based barcoded neuroanatomy resolved by single-cell RNA and in situ sequencing.

Mapping the connectivity of diverse neuronal types provides the foundation for understanding the structure and function of neural circuits. High-throughput and low-cost neuroanatomical techniques based on RNA barcode sequencing have the potential to map circuits at cellular resolution and a brain-wide scale, but existing Sindbis virus-based techniques can only map long-range projections using anterograde tracing approaches. Rabies virus can complement anterograde tracing approaches by enabling either retrograde labeling of projection neurons or monosynaptic tracing of direct inputs to genetically targeted postsynaptic neurons. However, barcoded rabies virus has so far been only used to map non-neuronal cellular interactions in vivo and synaptic connectivity of cultured neurons. Here we combine barcoded rabies virus with single-cell and in situ sequencing to perform retrograde labeling and transsynaptic labeling in the mouse brain. We sequenced 96 retrogradely labeled cells and 295 transsynaptically labeled cells using single-cell RNA-seq, and 4,130 retrogradely labeled cells and 2,914 transsynaptically labeled cells in situ. We found that the transcriptomic identities of rabies virus-infected cells can be robustly identified using both single-cell RNA-seq and in situ sequencing. By associating gene expression with connectivity inferred from barcode sequencing, we distinguished long-range projecting cortical cell types from multiple cortical areas and identified cell types with converging or diverging synaptic connectivity. Combining in situ sequencing with barcoded rabies virus complements existing sequencing-based neuroanatomical techniques and provides a potential path for mapping synaptic connectivity of neuronal types at scale.

Regional and cell-type-specific afferent and efferent projections of the mouse claustrum.

The claustrum (CLA) is a conspicuous subcortical structure interconnected with cortical and subcortical regions. Its regional anatomy and cell-type-specific connections in the mouse remain not fully determined. Using multimodal reference datasets, we confirmed the delineation of the mouse CLA as a single group of neurons embedded in the agranular insular cortex. We quantitatively investigated brain-wide inputs and outputs of CLA using bulk anterograde and retrograde viral tracing data and single neuron tracing data. We found that the prefrontal module has more cell types projecting to the CLA than other cortical modules, with layer 5 IT neurons predominating. We found nine morphological types of CLA principal neurons that topographically innervate functionally linked cortical targets, preferentially the midline cortical areas, secondary motor area, and entorhinal area. Together, this study provides a detailed wiring diagram of the cell-type-specific connections of the mouse CLA, laying a foundation for studying its functions at the cellular level.

A whole-brain monosynaptic input connectome to neuron classes in mouse visual cortex.

Identification of structural connections between neurons is a prerequisite to understanding brain function. Here we developed a pipeline to systematically map brain-wide monosynaptic input connections to genetically defined neuronal populations using an optimized rabies tracing system. We used mouse visual cortex as the exemplar system and revealed quantitative target-specific, layer-specific and cell-class-specific differences in its presynaptic connectomes. The retrograde connectivity indicates the presence of ventral and dorsal visual streams and further reveals topographically organized and continuously varying subnetworks mediated by different higher visual areas. The visual cortex hierarchy can be derived from intracortical feedforward and feedback pathways mediated by upper-layer and lower-layer input neurons. We also identify a new role for layer 6 neurons in mediating reciprocal interhemispheric connections. This study expands our knowledge of the visual system connectomes and demonstrates that the pipeline can be scaled up to dissect connectivity of different cell populations across the mouse brain.

Cardiovascular baroreflex circuit moonlights in sleep control.

Sleep disturbances are strongly associated with cardiovascular diseases. Baroreflex, a basic cardiovascular regulation mechanism, is modulated by sleep-wake states. Here, we show that neurons at key stages of baroreflex pathways also promote sleep. Using activity-dependent genetic labeling, we tagged neurons in the nucleus of the solitary tract (NST) activated by blood pressure elevation and confirmed their barosensitivity with optrode recording and calcium imaging. Chemogenetic or optogenetic activation of these neurons promoted non-REM sleep in addition to decreasing blood pressure and heart rate. GABAergic neurons in the caudal ventrolateral medulla (CVLM)-a downstream target of the NST for vasomotor baroreflex-also promote non-REM sleep, partly by inhibiting the sympathoexcitatory and wake-promoting adrenergic neurons in the rostral ventrolateral medulla (RVLM). Cholinergic neurons in the nucleus ambiguous-a target of the NST for cardiac baroreflex-promoted non-REM sleep as well. Thus, key components of the cardiovascular baroreflex circuit are also integral to sleep-wake brain-state regulation.

Dense functional and molecular readout of a circuit hub in sensory cortex.

Although single-cell transcriptomics of the neocortex has uncovered more than 300 putative cell types, whether this molecular classification predicts distinct functional roles is unclear. We combined two-photon calcium imaging with spatial transcriptomics to functionally and molecularly investigate cortical circuits. We characterized behavior-related responses across major neuronal subclasses in layers 2 or 3 of the primary somatosensory cortex as mice performed a tactile working memory task. We identified an excitatory intratelencephalic cell type, Baz1a, that exhibits high tactile feature selectivity. Baz1a neurons homeostatically maintain stimulus responsiveness during altered experience and show persistent enrichment of subsets of immediately early genes. Functional and anatomical connectivity reveals that Baz1a neurons residing in upper portions of layers 2 or 3 preferentially innervate somatostatin-expressing inhibitory neurons. This motif defines a circuit hub that orchestrates local sensory processing in superficial layers of the neocortex.

Enhancer viruses for combinatorial cell-subclass-specific labeling.

Rapid cell type identification by new genomic single-cell analysis methods has not been met with efficient experimental access to these cell types. To facilitate access to specific neural populations in mouse cortex, we collected chromatin accessibility data from individual cells and identified enhancers specific for cell subclasses and types. When cloned into recombinant adeno-associated viruses (AAVs) and delivered to the brain, these enhancers drive transgene expression in specific cortical cell subclasses. We extensively characterized several enhancer AAVs to show that they label different projection neuron subclasses as well as a homologous neuron subclass in human cortical slices. We also show how coupling enhancer viruses expressing recombinases to a newly generated transgenic mouse, Ai213, enables strong labeling of three different neuronal classes/subclasses in the brain of a single transgenic animal. This approach combines unprecedented flexibility with specificity for investigation of cell types in the mouse brain and beyond.

Functional enhancer elements drive subclass-selective expression from mouse to primate neocortex.

Viral genetic tools that target specific brain cell types could transform basic neuroscience and targeted gene therapy. Here, we use comparative open chromatin analysis to identify thousands of human-neocortical-subclass-specific putative enhancers from across the genome to control gene expression in adeno-associated virus (AAV) vectors. The cellular specificity of reporter expression from enhancer-AAVs is established by molecular profiling after systemic AAV delivery in mouse. Over 30% of enhancer-AAVs produce specific expression in the targeted subclass, including both excitatory and inhibitory subclasses. We present a collection of Parvalbumin (PVALB) enhancer-AAVs that show highly enriched expression not only in cortical PVALB cells but also in some subcortical PVALB populations. Five vectors maintain PVALB-enriched expression in primate neocortex. These results demonstrate how genome-wide open chromatin data mining and cross-species AAV validation can be used to create the next generation of non-species-restricted viral genetic tools.

RecV recombinase system for in vivo targeted optogenomic modifications of single cells or cell populations.

Brain circuits comprise vast numbers of interconnected neurons with diverse molecular, anatomical and physiological properties. To allow targeting of individual neurons for structural and functional studies, we created light-inducible site-specific DNA recombinases based on Cre, Dre and Flp (RecVs). RecVs can induce genomic modifications by one-photon or two-photon light induction in vivo. They can produce targeted, sparse and strong labeling of individual neurons by modifying multiple loci within mouse and zebrafish genomes. In combination with other genetic strategies, they allow intersectional targeting of different neuronal classes. In the mouse cortex they enable sparse labeling and whole-brain morphological reconstructions of individual neurons. Furthermore, these enzymes allow single-cell two-photon targeted genetic modifications and can be used in combination with functional optical indicators with minimal interference. In summary, RecVs enable spatiotemporally precise optogenomic modifications that can facilitate detailed single-cell analysis of neural circuits by linking genetic identity, morphology, connectivity and function.

Connectional architecture of a mouse hypothalamic circuit node controlling social behavior.

Type 1 estrogen receptor-expressing neurons in the ventrolateral subdivision of the ventromedial hypothalamus (VMHvlEsr1) play a causal role in the control of social behaviors, including aggression. Here we use six different viral-genetic tracing methods to systematically map the connectional architecture of VMHvlEsr1 neurons. These data reveal a high level of input convergence and output divergence ("fan-in/fan-out") from and to over 30 distinct brain regions, with a high degree (∼90%) of bidirectionality, including both direct as well as indirect feedback. Unbiased collateralization mapping experiments indicate that VMHvlEsr1 neurons project to multiple targets. However, we identify two anatomically distinct subpopulations with anterior vs. posterior biases in their collateralization targets. Nevertheless, these two subpopulations receive indistinguishable inputs. These studies suggest an overall system architecture in which an anatomically feed-forward sensory-to-motor processing stream is integrated with a dense, highly recurrent central processing circuit. This architecture differs from the "brain-inspired," hierarchical feed-forward circuits used in certain types of artificial intelligence networks.

Nontoxic, double-deletion-mutant rabies viral vectors for retrograde targeting of projection neurons.

Recombinant rabies viral vectors have proven useful for applications including retrograde targeting of projection neurons and monosynaptic tracing, but their cytotoxicity has limited their use to short-term experiments. Here we introduce a new class of double-deletion-mutant rabies viral vectors that left transduced cells alive and healthy indefinitely. Deletion of the viral polymerase gene abolished cytotoxicity and reduced transgene expression to trace levels but left vectors still able to retrogradely infect projection neurons and express recombinases, allowing downstream expression of other transgene products such as fluorophores and calcium indicators. The morphology of retrogradely targeted cells appeared unperturbed at 1 year postinjection. Whole-cell patch-clamp recordings showed no physiological abnormalities at 8 weeks. Longitudinal two-photon structural and functional imaging in vivo, tracking thousands of individual neurons for up to 4 months, showed that transduced neurons did not die but retained stable visual response properties even at the longest time points imaged.

Oxytocin signaling in the medial amygdala is required for sex discrimination of social cues.

The neural control of social behaviors in rodents requires the encoding of pheromonal cues by the vomeronasal system. Here we show that the typical preference of male mice for females is eliminated in mutants lacking oxytocin, a neuropeptide modulating social behaviors in many species. Ablation of the oxytocin receptor in aromatase-expressing neurons of the medial amygdala (MeA) fully recapitulates the elimination of female preference in males. Further, single-unit recording in the MeA uncovered significant changes in the sensory representation of conspecific cues in the absence of oxytocin signaling. Finally, acute manipulation of oxytocin signaling in adults is sufficient to alter social interaction preferences in males as well as responses of MeA neurons to chemosensory cues. These results uncover the critical role of oxytocin signaling in a molecularly defined neuronal population in order to modulate the behavioral and physiological responses of male mice to females on a moment-to-moment basis.

Identification of preoptic sleep neurons using retrograde labelling and gene profiling.

In humans and other mammalian species, lesions in the preoptic area of the hypothalamus cause profound sleep impairment, indicating a crucial role of the preoptic area in sleep generation. However, the underlying circuit mechanism remains poorly understood. Electrophysiological recordings and c-Fos immunohistochemistry have shown the existence of sleep-active neurons in the preoptic area, especially in the ventrolateral preoptic area and median preoptic nucleus. Pharmacogenetic activation of c-Fos-labelled sleep-active neurons has been shown to induce sleep. However, the sleep-active neurons are spatially intermingled with wake-active neurons, making it difficult to target the sleep neurons specifically for circuit analysis. Here we identify a population of preoptic area sleep neurons on the basis of their projection target and discover their molecular markers. Using a lentivirus expressing channelrhodopsin-2 or a light-activated chloride channel for retrograde labelling, bidirectional optogenetic manipulation, and optrode recording, we show that the preoptic area GABAergic neurons projecting to the tuberomammillary nucleus are both sleep active and sleep promoting. Furthermore, translating ribosome affinity purification and single-cell RNA sequencing identify candidate markers for these neurons, and optogenetic and pharmacogenetic manipulations demonstrate that several peptide markers (cholecystokinin, corticotropin-releasing hormone, and tachykinin 1) label sleep-promoting neurons. Together, these findings provide easy genetic access to sleep-promoting preoptic area neurons and a valuable entry point for dissecting the sleep control circuit.

The genetic basis of parental care evolution in monogamous mice.

Parental care is essential for the survival of mammals, yet the mechanisms underlying its evolution remain largely unknown. Here we show that two sister species of mice, Peromyscus polionotus and Peromyscus maniculatus, have large and heritable differences in parental behaviour. Using quantitative genetics, we identify 12 genomic regions that affect parental care, 8 of which have sex-specific effects, suggesting that parental care can evolve independently in males and females. Furthermore, some regions affect parental care broadly, whereas others affect specific behaviours, such as nest building. Of the genes linked to differences in nest-building behaviour, vasopressin is differentially expressed in the hypothalamus of the two species, with increased levels associated with less nest building. Using pharmacology in Peromyscus and chemogenetics in Mus, we show that vasopressin inhibits nest building but not other parental behaviours. Together, our results indicate that variation in an ancient neuropeptide contributes to interspecific differences in parental care.

Structure of a heparin-dependent complex of Hedgehog and Ihog.

Hedgehog (Hh) signaling molecules mediate key tissue-patterning events during animal development, and inappropriate activation of Hh signaling in adults has been associated with human cancers. Recently, a conserved family of type I integral membrane proteins required for normal response to the Hh signal was discovered. One member of this family, Ihog (interference hedgehog), functions upstream or at the level of Patched (Ptc), but how Ihog participates in Hh signaling remains unclear. Here, we show that heparin binding induces Ihog dimerization and is required to mediate high-affinity interactions between Ihog and Hh. We also present crystal structures of a Hh-binding fragment of Ihog, both alone and complexed with Hh. Heparin is not well ordered in these structures, but a basic cleft in the first FNIII domain of Ihog (IhogFn1) is shown by mutagenesis to mediate heparin binding. These results establish that Hh directly binds Ihog and provide the first demonstration of a specific role for heparin in Hh responsiveness.

The ihog cell-surface proteins bind Hedgehog and mediate pathway activation.

The ihog gene (interference hedgehog), identified by RNA interference in Drosophila cultured cells, encodes a type 1 membrane protein shown here to bind and to mediate response to the active Hedgehog (Hh) protein signal. ihog mutations produce defects characteristic of Hh signaling loss in embryos and imaginal discs, and epistasis analysis places ihog action at or upstream of the negatively acting receptor component, Patched (Ptc). The first of two extracellular fibronectin type III (FNIII) domains of the Ihog protein mediates a specific interaction with Hh protein in vitro, but the second FNIII domain is additionally required for in vivo signaling activity and for Ihog-enhanced binding of Hh protein to cells coexpressing Ptc. Other members of the Ihog family, including Drosophila Boi and mammalian CDO and BOC, also interact with Hh ligands via a specific FNIII domain, thus identifying an evolutionarily conserved family of membrane proteins that function in Hh signal response.

Identification of Hedgehog pathway components by RNAi in Drosophila cultured cells.

Classical genetic screens can be limited by the selectivity of mutational targeting, the complexities of anatomically based phenotypic analysis, or difficulties in subsequent gene identification. Focusing on signaling response to the secreted morphogen Hedgehog (Hh), we used RNA interference (RNAi) and a quantitative cultured cell assay to systematically screen functional roles of all kinases and phosphatases, and subsequently 43% of predicted Drosophila genes. Two gene products reported to function in Wingless (Wg) signaling were identified as Hh pathway components: a cell surface protein (Dally-like protein) required for Hh signal reception, and casein kinase 1alpha, a candidate tumor suppressor that regulates basal activities of both Hh and Wg pathways. This type of cultured cell-based functional genomics approach may be useful in the systematic analysis of other biological processes.