22 Dec 2023

The picture of the mouse brain just got exponentially more complex. Using single-cell transcriptomics, epigenomics, and spatial genomics, scientists created an atlas of 32 million brain cells across all regions, distinguishing 5,300 cell types. The findings were reported in 10 articles published in Nature on December 13 by scientists across the U.S.

Sound familiar? The mouse brain atlas builds off the framework used to create the most comprehensive human brain atlas published earlier this year (Oct 2023 news). Like the studies used to make the human version, the mouse work was funded by the NIH’s Brain Research through Advancing Innovative Neurotechnologies (BRAIN) initiative under the Cell Census Network project. BICCN is cataloging and comparing brain cell types from mice, primates, and people.

“This is an important milestone in neuroscience,” Evan Macosko of the Broad Institute in Cambridge, Massachusetts, told Alzforum. “We can only study what we have the tools to detect, but that doesn’t allow us to see the full picture. We haven’t known all the different brain cell types and where they are.”

Bart De Strooper of the U.K. Dementia Research Institute, London, compared this to the completion of the human genome project at the beginning of the millennium. “This is a major resource to start understanding how the brain really works,” he said. “Obviously, this is the end of the beginning. Similar work in humans will be a few orders of magnitude larger, but the path is clear.”

Behold, the Mouse Brain. Spatial transcriptomics and single-cell RNA sequencing drew up the most detailed mouse brain atlas to date. MERFISH spatial transcriptomics of brain slices (outermost ring) visualized the gene expression and distribution of 5,300 cell subtypes. These subtypes fell into 34 cell types (small cell clusters, second ring) that grouped into seven classes (large cell clusters, third ring) based on similar gene expression and location. Cell taxonomy is also shown as a circular dendrogram surrounding an image of a brain. [Courtesy of Cindy van Velthoven, Zizhen Yao, and Hongkui Zeng, Nature, 2023.]

Of the 10 papers, four laid the spatial transcriptomics groundwork. Researchers led by Hongkui Zeng and Zizhen Yao at the Allen Institute for Brain Science, Seattle, used single-cell RNA sequencing (scRNA-Seq) to profile the transcriptomes of 7 million cells from across the entire mouse brain, then followed that with a technique called multiplexed error-robust fluorescence in situ hybridization on 4.3 million cells. MERFISH measures the spatial distribution of transcripts in cells throughout tissue. The cells fell into 34 types, divided into 5,322 subtypes based on gene expression and location (image above). Each cell subtype populated a distinct portion of the brain, and neurons made an exquisitely diverse set of subtype-specific neurotransmitters and neuropeptides, based on the co-expression of enzymes and transporters, highlighting the complexity of the brain. “The 5,000 cell types will keep neuroscientists busy for the next 20 years to figure out what the cells do and how they change under a variety of disease conditions,” Zeng said.

The other three spatial transcriptomics papers followed suit. Scientists led by Xiaowei Zhuang, Harvard University, Cambridge, Massachusetts, analyzed the spatial expression of 1,100 genes in 10 million cells using MERFISH, while Macosko, Fei Chen, and colleagues at the Broad Institute used microscope slide-based RNA-Seq of tissue sections, aka Slide-seq (see companion news story), to map gene expression of 4.4 million cells at 10-micron resolution.

Xiao Wang of the Broad Institute and Jia Liu of Harvard led a project to map 1,022 transcripts among 1.1 million cells at sub-micron resolution using a method called spatially resolved transcript amplicon readout mapping, or STARmap (image below). Both Zhuang’s and Wang/Liu’s analysis identified about 100 transcriptionally distinct tissue regions that refine the brain anatomy.

High-Definition Brain. Twenty-six cell types populate distinct areas of the mouse brain (top images). From these, 230 subtypes, defined by gene expression, delineate 106 tissue regions that sharpen the anatomical definition of the brain (bottom images). [Courtesy of Hailing Shi, Yichun He, Jia Liu, and Xiao Wang. Nature, 2023.]

Four other papers analyzed epigenetics. Researchers led by Joseph Ecker, Salk Institute for Biological Studies, La Jolla, California, used single-nucleus methylome and chromatin conformation/methylome sequencing to identify 301,600 methylomes and 176,000 pairs of open chromatin/methylation locations, respectively. This uncovered 2.6 million differently methylated DNA loci among 274 cell subtypes, indicating crucial areas of gene regulation. Bing Ren and colleagues at the University of California, San Diego, further probed chromatin accessibility with snATAC–seq of 2.3 million cells, capturing transcriptional activity on about 1 million candidate cis-regulatory DNA elements (cCREs). Half had not been annotated before.

For one of the other two epigenetic studies, Ecker collaborated with Edward Callaway at UCSD to analyze single-nucleus methylation-seq of 33,000 neurons from 32 anatomical brain regions. They traced axons with a retrogradely traveling virus that carries a gene for green fluorescent protein. This distinguished gene regulation networks based on neural connectivity. Zeng, Zhigang He, Anne Jacobi, and colleagues at Boston Children’s Hospital also used retrograde labeling, snRNA-Seq, and the spatial transcriptomics atlas to generate a map of 65,000 spinal projecting neurons. They found three distinct types of SPNs based on gene expression and location.

The two remaining papers explored how the complex organization of this mammalian brain evolved. Ecker and Ren teamed up to compare gene regulation in the primary motor cortices of mice, marmosets, macaques, and humans. Transcription factor expression drove species-specific epigenetics, and transposable elements, aka jumping genes, accounted for nearly 80 percent of human-specific cCREs in cortical cells.

Scientists led by Karthik Shekhar at the University of California, Berkeley, and Joshua Sanes at Harvard studied the eye, analyzing transcriptomes of retinal cells from 17 species, including mice, primates, and humans. They traced the origin of human “midget” retinal ganglion cells, which account for 90 percent of RGCs in people, to large cells that comprise just 2 percent of all RGCs in mice. This suggests that these cells became smaller and more numerous as the visual cortex evolved and enlarged.

“These papers provide the AD research community with new foundations for understanding disease mechanisms,” wrote Takaomi Saido of the RIKEN Center for Brain Science in Japan. “Analyses of different AD models in similar ways will make it possible to define essential cellular and molecular processes, for instance, those that play a role in the Aβ amyloidosis-tau pathology axis.” Bruce Lamb, Indiana University School of Medicine, Indianapolis, said this atlas will help the Model Organism Development and Evaluation for Late-onset Alzheimer’s Disease (MODEL-AD) Consortium, which develops new models for sporadic AD. “It should aid our multiomic analyses of the mouse models, in particular, cell types and brain regions impacted,” he told Alzforum.

Catherine Kaczorowski, University of Michigan, Ann Arbor, praised this atlas but noted that it is a map from only one mouse strain. “This is absolutely needed and fundamental, but it’s just the starting point,” she said. She thinks researchers should identify which mouse models have transcriptomic signatures closely aligned with the transcriptomes of AD brain tissue and then deeply map those strains.—Chelsea Weidman Burke

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

News Citations

1. Behold, the Human Brain Like Never Seen Before 12 Oct 2023
2. ‘Slide-tags’ Method Sharpens Spatial Transcriptomics 21 Dec 2023

External Citations

1. BICCN

Further Reading

No Available Further Reading