The Quest To Discover the Cell Types of the Brain

The function of the human brain is mediated through its cell types and circuits, and decoding their nature and organization is essential to understanding behavior and thought. The brain contains a discrete but undetermined number of cell types such as neurons and glia. These are further broken down into subtypes: neurons can be distinguished by their neurotransmitters, electrophysiological properties, morphology, connectivity, patterns of gene expression and other functional properties.¹

While the classification of neurons is a prerequisite to manipulating them in controlled ways – and to understanding how they change in brain disorders – information about the types of glial cells, vascular cells and immune cells associated with the nervous system may be more relevant to many brain diseases and to its basic metabolic function.

The Brain Research through Advancing Innovative Neurotechnologies Initiative

On April 2, 2013, the Obama administration proposed a major national project to discover the mysteries of the brain called the Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative. This charge resulted in a working group constructing a comprehensive scientific vision (the Brain 2025 Report) outlining high-priority research areas, including a census of cell types of the brain and characterization of their circuits. In this way, a census can begin by describing large classes of neurons (e.g., excitatory pyramidal neurons of the cortex) and then proceed to finer categories within these classifications. There is not yet complete consensus on what a neuronal type is, since a variety of factors including experience, connectivity and neuromodulators can diversify the molecular, electrical and structural properties of initially similar neurons. In some cases, there may not even be sharp boundaries separating subtypes from each other, and cell phenotypes may change over time. There is, however, general agreement that types can be defined by invariant and generally intrinsic properties, and that this classification can provide a good starting point for a census.

The first phase of the BRAIN Initiative’s approach to classifying and characterizing cell type was a series of pilot investigations called the BRAIN Initiative Cell Census Consortium (2013-2017).² In 2014, the NIH BRAIN Initiative awarded 10 grants to pilot classification strategies for a comprehensive brain cell census. From these pilot projects, multiple brain regions from different organisms were studied using a variety of advanced technologies, and awardees collaborated on defining standards to describe experiments and data sets.

The pilot projects were funded to develop, validate, and scale up emerging genomic and anatomical mapping technologies. These were designed to create a complete inventory of neuronal cell types and their connections in multiple species and during development. Many of these techniques are illustrated in Figure 1. These projects formed the foundation for larger studies by other research groups aimed at developing and deploying new methods to characterize the multitude of cell types and their patterns of connectivity in the mammalian brain.

The lessons learned from this effort informed how best to meet the specific challenges of handling massive sequencing and anatomical datasets, and how to integrate data types such as transcriptomics, epigenomics, physiology, morphology and connectivity. The work also addressed how to harmonize such information from multiple research groups. It also addressed how to disseminate research tools including genetic reagents and animal models for the research community at large, as well as the longer-term effort to generate whole-brain cell atlases in species including mice and humans.

Figure 1: Techniques for cell type profiling in the brain. Techniques for profiling cell types in the brain span a broad range of technologies, including: A) single cell and nucleus RNA-seq, epigenetic profiling such as ATAC-seq, B) identifying neuronal projections through viral tracing and methylation (snmC-seq2), C) single cell full morphology and neuronal reconstruction (fMOST), D) spatial transcriptomics methods such as MERFISH, E) viral tracing methods for anatomy and morphology, F) multimodal profiling for connecting morphology, electrophysiology and transcriptomics such as Patch-seq and G) use of transgenic lines to target specific cell types, H) data is mapped into a common spatial framework (CCF), I) cell type organization is captured through hierarchical taxonomies. Details on the definition and use of these techniques can be found on The BRAIN Initiative Cell Census Network.

The BRAIN Initiative Cell Census Network

With the proof of concept and the feasibility of scaling established, in 2017 the NIH expanded support for the development of cell census data and relevant tools by launching a coordinated set of awards under the auspices of the BRAIN Initiative Cell Census Network (BICCN). The overarching goal of the network was to generate comprehensive 3D common reference brain cell atlases that would integrate molecular, anatomical and functional data for describing cell types in the mouse brain, with prototypes in human and non-human primates (NHP).

BICCN is a collaborative network of centers and laboratories, including data generating centers, data archives and data standards developers, which generate, map and share resources to support several overarching goals. These include generating a high-resolution, comprehensive atlas based on large-scale single-cell transcriptome and epigenome sequencing, along with systematic characterization of neuronal morphology, new genetic tools to experimentally target brain cell types and a prototype atlas of human brain and NHP cell types in both adult and developing human brains. A standard anatomical template for mapping cell types in the mouse brain was also established through completion and validation of a common coordinate framework (CCF).³ BICCN also conducted an initial profiling of cellular diversity in several structures relevant to neurodegenerative and neuropsychiatric disease, including the hippocampus and dorsolateral prefrontal cortex, and importantly, cross-species identification and mapping of cell types between mouse, marmoset and human.

The first phase of the BICCN collaboratively generated a comprehensive multimodal cell census and atlas of the mammalian primary motor cortex (MOp or M1).⁴ This project involved coordinated large-scale analyses of single-cell transcriptomes, chromatin accessibility, DNA methylomes, spatially resolved single-cell transcriptomes, anatomic characterization with morphological and electrophysiological properties and cellular resolution input-output mapping. These results, and their extension to the whole mouse brain and other human regions, are the first step towards creating a catalog or census of all brain cell types.

Active BICCN Working Groups continue to extend and integrate new and existing data across labs towards an integrated transcriptomic and epigenomic atlas of the entire mouse central nervous system. BICCN reflects the increasingly collaborative nature of modern neuroscience and has accomplished the deepest coordinated characterization of cell types in any organ to date. Consortia such as the Human Cell Atlas (HCA) and Human Biomolecular Atlas Program (HuBMAP) are also key representatives of this community and are leading molecular profiling in other organs.   

A major publication with 11 companion manuscripts summarizing this work was published in Nature demonstrating:

• A unified molecular genetic catalog of cortical cell types that integrates transcriptome, spatial transcriptomic, open chromatin and DNA methylation maps.
• A cross-species analysis that provides a unified taxonomy of transcriptomic types and their hierarchical organization that are conserved from mouse to marmoset and human.
• Development of imaging tools for morphometry at whole brain scale, cross modality brain mapping and resources of single cell morphological cell types.
• An extensive genetic toolset for targeting and fate mapping glutamatergic projection neuron types aimed at linking their developmental trajectory to their circuit function.
• Brain cell type ontologies and nomenclature systems for describing and manipulating cell types.

Together, the results establish a unified and mechanistic framework of neuronal cell type organization that integrates multimodal molecular, genetic and spatial information with multi-faceted phenotypic properties. The results of this work are summarized in an online application, the Cell Type Knowledge Explorer shown in Figure 2.

Figure 2: Cell type knowledge explorer is a multimodal cell census and atlas of primary motor cortex, developed in collaboration with the BICCN and described in detail in the  BICCN flagship publication. This resource is the entry point for exploration of individual cell type taxonomies for human, marmoset, and mouse primary motor cortex. High-throughput, single-cell transcriptomic and epigenomic profiling demonstrates broad conservation of molecular cell type identities across these three species. The tree on the left A) illustrates a cross-species consensus classification of cell types, and each colored block represents an individual cell type in human, marmoset or mouse in the context of this consensus. Through the taxonomies on the right B) users can learn more about primary motor cortex cell type characteristics and their linkages across mammalian species.

The BRAIN Initiative Cell Atlas Network

Currently, the NIH BRAIN Initiative Cell Census Program is broadening and deepening the systematic cell census and atlas efforts with a new emphasis on the human brain. This work has established a network of projects that will work cooperatively to generate comprehensive and high-resolution brain cell atlases that encompass molecular, anatomical and functional annotations of brain cell types (neurons, glia and other non-neuronal cells) across the lifespan in human and other species.

The goal is to uncover fundamental knowledge on diverse cell types and their three-dimensional organizational principles in these species’ brains. The approach includes generating comprehensive molecular taxonomies of brain cell types in humans, NHP and mice, as well as open-access digital brain cell reference atlases of humans, NHP and mice. As part of the final phase of the BRAIN Initiative, The BRAIN Initiative Cell Atlas Network (BICAN) presents novel challenges essential to humans such as developing optimal tissue preparation and stabilization methods, creating brain tissue sample standards and increasing diversity of donors. As shown in Figure 3, six major projects are now underway in BICAN, toward constructing large-scale cell type profiling in human and NHP, as well as in mice and humans.

Figure 3: BICAN approaches to cell type profiling and atlas construction. A) Center for Multiomic Human Brain Cell Atlas samples cells from 30 donors across the whole lifespan and 50 brain regions using single-cell snmC-seq to capture chromatin organization and DNA methylation information and spatial transcriptomic expression (MERFISH). B) An Atlas of Human Brain Cell Variation profiles human variation in 200 donors and 50 regions using multiome and spatial techniques. C) A Multidisciplinary Center for Developing Human and Non-human Primate Brain Cell Atlases profiles development in human, macaque and marmoset in 30 brain regions.D) Comprehensive single-cell atlas of the developing mouse brain profiles a complete developmental series in the mouse. E) Spatiotemporal epigenomic and chromosomal architectural cell atlas of developing human brains investigates human development using a combination of single cell transcriptomic and epigenomic techniques. F) Functionally guided adult whole-brain cell atlas in human and NHP constructs a functionally (fMRI) driven sampling to construct a deep molecular census of adult brain cell types.

The ultimate expectation of the BRAIN Initiative is to achieve a full census of neuronal and glial cell types in mouse, human and NHP, an intellectual framework for cell type classification and to provide experimental access to the different brain cell types to determine their roles in health and disease.

Taxonomies of putative types and representative cells in the mammalian brain will provide a frame of reference for studies across labs, and possibly in different organisms, allowing cross‐comparison. The data and resources under development in the combined BICCN/BICAN ecosystems will provide researchers with unprecedented tools to address the challenges of understanding the brain.

References:

1. Zeng H. What is a cell type and how to define it? Cell. 2022;185(15):2739-2755. doi: 10.1016/j.cell.2022.06.031

2. Ecker JR, Geschwind DH, Kriegstein AR, et al. The BRAIN Initiative cell census consortium: Lessons learned toward generating a comprehensive brain cell atlas. Neuron. 2017;96(3):542-557. doi: 10.1016/j.neuron.2017.10.007

3. Wang Q, Ding SL, Li Y, et al. The Allen mouse brain common coordinate framework: A 3D reference atlas. Cell. 2020;181(4):936-953.e20. doi: 10.1016/j.cell.2020.04.007

4. Callaway EM, Dong HW, Ecker JR, et al. A multimodal cell census and atlas of the mammalian primary motor cortex. Nature. 2021;598(7879):86-102. doi: 10.1038/s41586-021-03950-0