CoChat
Sign Up
Cited by 809Single-cell and spatial transcriptomicsNeuroscience and Neuropharmacology ResearchNeuroinflammation and Neurodegeneration Mechanisms

The DOI matches its Crossref record; 71 of the 71 references it cites were found in a registry or an index.

Identification of region-specific astrocyte subtypes at single cell resolution

Open access
Batiuk, Mykhailo Y., Martirosyan, Araks, Wahis, Jérôme, de Vin, Filip, Marneffe, Catherine, Kusserow, Carola, Koeppen, Jordan, Viana, João Filipe, Oliveira, João Filipe, Voet, Thierry, Ponting, Chris P., Belgard, T. Grant, Holt, Matthew G.·2020·Nature Communications
DOI 10.1038/s41467-019-14198-8PMC PMC7058027

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Astrocytes, a major cell type found throughout the central nervous system, have general roles in the modulation of synapse formation and synaptic transmission, blood-brain barrier formation, and regulation of blood flow, as well as metabolic support of other brain resident cells. Crucially, emerging evidence shows specific adaptations and astrocyte-encoded functions in regions, such as the spinal cord and cerebellum. To investigate the true extent of astrocyte molecular diversity across forebrain regions, we used single-cell RNA sequencing. Our analysis identifies five transcriptomically distinct astrocyte subtypes in adult mouse cortex and hippocampus. Validation of our data in situ reveals distinct spatial positioning of defined subtypes, reflecting the distribution of morphologically and physiologically distinct astrocyte populations. Our findings are evidence for specialized astrocyte subtypes between and within brain regions. The data are available through an online database (https://holt-sc.glialab.org/), providing a resource on which to base explorations of local astrocyte diversity and function in the brain.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41467-019-14198-8

Cites

71 of 71 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Single-cell sequencing strategy and cell-type identification. a Whole brains were obtained from C57BL/6J mice at postnatal (P) day 56. Cortical (CX) and hippocampal (HP) astrocytes were prepared separately, using enzymatic digestion followed by mechanical trituration. Two separate batches of astrocytes for each region were prepared. Cortical cell suspensions were prepared from two littermate animals in parallel using separate tubes. Hippocampal cell suspensions were also prepared in parallel using separate tubes; in this case, two different sets of four littermate animals were used. Astrocytes were then specifically labeled with the ASCA-2-PE antibody and single cells were deposited in individual wells of a PCR plate using FACS. Single-cell library preparation was performed using a modified Smart-seq2 protocol. In total, 2976 libraries were prepared and sequenced using a NextSeq 500 system (Illumina). b Each library was sequenced to optimal coverage (on average 1 M reads per library). In total, 2015 high-quality libraries were retained for further analysis. In these libraries, a high fraction of reads mapped to exons (CDS, coding sequence; UTR, untranslated region). Conversely, a low fraction of reads mapped to intronic and intergenic regions. c Visualization of the major higher-order cell types (2015 cells) identified by Seurat using tSNE plots. Each dot represents a single cell. Cells with similar molecular profiles group together; cell types were assigned according to the expression of specific marker genes (and are labeled in different colors). d Gene expression heatmap for higher-order cell types (columns) grouped according to the Seurat classification shown in Fig. 1c. Color-coding from Fig. 1c is retained. Gray, no expression; yellow, low expression; red, high expression, In-normalized gene expression data is shown.
Fig. 1 Single-cell sequencing strategy and cell-type identification. a Whole brains were obtained from C57BL/6J mice at postnatal (P) day 56. Cortical (CX) and hippocampal (HP) astrocytes were prepared separately, using enzymatic digestion followed by mechanical trituration. Two separate batches of astrocytes for each region were prepared. Cortical cell suspensions were prepared from two littermate animals in parallel using separate tubes. Hippocampal cell suspensions were also prepared in parallel using separate tubes; in this case, two different sets of four littermate animals were used. Astrocytes were then specifically labeled with the ASCA-2-PE antibody and single cells were deposited in individual wells of a PCR plate using FACS. Single-cell library preparation was performed using a modified Smart-seq2 protocol. In total, 2976 libraries were prepared and sequenced using a NextSeq 500 system (Illumina). b Each library was sequenced to optimal coverage (on average 1 M reads per library). In total, 2015 high-quality libraries were retained for further analysis. In these libraries, a high fraction of reads mapped to exons (CDS, coding sequence; UTR, untranslated region). Conversely, a low fraction of reads mapped to intronic and intergenic regions. c Visualization of the major higher-order cell types (2015 cells) identified by Seurat using tSNE plots. Each dot represents a single cell. Cells with similar molecular profiles group together; cell types were assigned according to the expression of specific marker genes (and are labeled in different colors). d Gene expression heatmap for higher-order cell types (columns) grouped according to the Seurat classification shown in Fig. 1c. Color-coding from Fig. 1c is retained. Gray, no expression; yellow, low expression; red, high expression, In-normalized gene expression data is shown.CC BY · PubMed Central
Fig. 2 Identification of astrocyte subtypes in adult mouse cortex and hippocampus. Single-cell data were used to identify distinct astrocyte subtypes (AST). a A total of 1811 astrocytes were identified from higher-order clustering. This data was extracted and reclustered using Seurat and five distinct astrocyte subtypes were identified. Clusters are presented in tSNE plots, with each AST color-coded. b Hierarchically clustered average gene expression heatmap for genes overexpressed across subtypes. Rows correspond to cells, columns to genes. Magenta, low gene expression; yellow, high expression. Scaled ln-normalized data are shown. c Astrocytes derived from the cortex (CX) or hippocampus (HP) segregate based on gene expression. d Expression of subtype-specific marker genes selected for in situ hybridization experiments. Markers are classed as absent/low “−” or highly expressed “+,” based on ln-normalized expression data. See also Supplementary Data.
Fig. 2 Identification of astrocyte subtypes in adult mouse cortex and hippocampus. Single-cell data were used to identify distinct astrocyte subtypes (AST). a A total of 1811 astrocytes were identified from higher-order clustering. This data was extracted and reclustered using Seurat and five distinct astrocyte subtypes were identified. Clusters are presented in tSNE plots, with each AST color-coded. b Hierarchically clustered average gene expression heatmap for genes overexpressed across subtypes. Rows correspond to cells, columns to genes. Magenta, low gene expression; yellow, high expression. Scaled ln-normalized data are shown. c Astrocytes derived from the cortex (CX) or hippocampus (HP) segregate based on gene expression. d Expression of subtype-specific marker genes selected for in situ hybridization experiments. Markers are classed as absent/low “−” or highly expressed “+,” based on ln-normalized expression data. See also Supplementary Data.CC BY · PubMed Central
Fig. 3 Identification of common and differentially expressed genes in astrocytes. a Chart showing the number of genes expressed in at least 60% of sampled astrocytes (common) and the number of genes specifically enriched in each subtype. b Examples of genes common across astrocyte subtypes, classified by biological function. TFs, transcription factors. c Examples of genes highly enriched in specific astrocyte subtypes, classified by biological function. Note, some genes, e.g. Gabrg1 (γ-aminobutyric acid type A receptor γ1 subunit), could be classified as either an ion channel or as involved in synaptic function/plasticity. Here, classification was based on the principal identified function—ion channel activity.
Fig. 3 Identification of common and differentially expressed genes in astrocytes. a Chart showing the number of genes expressed in at least 60% of sampled astrocytes (common) and the number of genes specifically enriched in each subtype. b Examples of genes common across astrocyte subtypes, classified by biological function. TFs, transcription factors. c Examples of genes highly enriched in specific astrocyte subtypes, classified by biological function. Note, some genes, e.g. Gabrg1 (γ-aminobutyric acid type A receptor γ1 subunit), could be classified as either an ion channel or as involved in synaptic function/plasticity. Here, classification was based on the principal identified function—ion channel activity.CC BY · PubMed Central
Fig. 4 Differential patterning of AST4 and AST5 in adult mouse brain. Multiplexed fluorescence in situ hybridization was used to map locations of AST4 and AST5. a AST4 was identified by high expression levels of Frzb, Ascl1, and Slc1a3. b AST5 was identified by absence/low expression of Ogt and high expression of both Fam107a and Slc1a3. Mapping was performed on three sections obtained from three independent animals aged between P56–P60. Representative images are shown. Top left: low-magnification image of a coronal section. Black dots show the distribution of the astrocyte subtype through one brain hemisphere. Brain regions are defined manually based on definitions from the Allen Brain Atlas. High-magnification images (below) show the localization of markers to specific cells defined on the basis of nuclear (DAPI, blue) staining. Right: bar plots (showing from left to right) fluorescence counts per RNA marker per cell (shown for all cells across all sections analyzed), the distribution of the subtype between brain regions and the distribution of the subtype normalized to the total number of astrocytes per brain region (all Slc1a3 + cells). Astrocytes belonging to the subtype of interest are highlighted by a shaded box (color-coded according to the scheme used in Fig. 2a). Astrocyte numbers across layers are given as average per section analyzed. Error bars are equivalent across the figure and represent SEM. Scale bars, low magnification 1000 µm; high magnification, 10 µm. “+” high gene expression, “−” low or absent gene expression. SO Stratum oriens, SP Stratum pyramidale, SR Stratum radiatum, SG Subgranular zone, \DG Dentate gyrus without SG, SLM Stratum lacunosum-moleculare.
Fig. 4 Differential patterning of AST4 and AST5 in adult mouse brain. Multiplexed fluorescence in situ hybridization was used to map locations of AST4 and AST5. a AST4 was identified by high expression levels of Frzb, Ascl1, and Slc1a3. b AST5 was identified by absence/low expression of Ogt and high expression of both Fam107a and Slc1a3. Mapping was performed on three sections obtained from three independent animals aged between P56–P60. Representative images are shown. Top left: low-magnification image of a coronal section. Black dots show the distribution of the astrocyte subtype through one brain hemisphere. Brain regions are defined manually based on definitions from the Allen Brain Atlas. High-magnification images (below) show the localization of markers to specific cells defined on the basis of nuclear (DAPI, blue) staining. Right: bar plots (showing from left to right) fluorescence counts per RNA marker per cell (shown for all cells across all sections analyzed), the distribution of the subtype between brain regions and the distribution of the subtype normalized to the total number of astrocytes per brain region (all Slc1a3 + cells). Astrocytes belonging to the subtype of interest are highlighted by a shaded box (color-coded according to the scheme used in Fig. 2a). Astrocyte numbers across layers are given as average per section analyzed. Error bars are equivalent across the figure and represent SEM. Scale bars, low magnification 1000 µm; high magnification, 10 µm. “+” high gene expression, “−” low or absent gene expression. SO Stratum oriens, SP Stratum pyramidale, SR Stratum radiatum, SG Subgranular zone, \DG Dentate gyrus without SG, SLM Stratum lacunosum-moleculare.CC BY · PubMed Central
Fig. 5 Differential patterning of AST1 and AST2 in adult mouse brain. Multiplexed fluorescence in situ hybridization was used to map locations of AST1 and AST2. a AST1 was identified by high expression levels of Gfap, Agt, and Slc1a3. b AST2 was identified by low expression/absence of Agt and high expression of both Unc13c and Slc1a3. Mapping was performed on three sections obtained from three independent animals aged between P56 and P60. Representative images are shown. Top left: low-magnification image of a coronal section. Black dots show the distribution of the astrocyte subtype through one brain hemisphere. Brain regions are defined manually based on definitions from the Allen Brain Atlas. High-magnification images (below) show the localization of markers to specific cells defined on the basis of nuclear (DAPI, blue) staining. Right: bar plots (showing from left to right) fluorescence counts per RNA marker per cell (shown for all cells across all sections analyzed), the distribution of the subtype between brain regions, and the distribution of the subtype normalized to the total number of astrocytes per brain region (all Slc1a3 + cells). Astrocytes belonging to the subtype of interest are highlighted by a shaded box (color-coded according to the scheme used in Fig. 2a). Astrocyte numbers across layers are given as average per section analyzed. Error bars are equivalent across the figure and represent SEM. Scale bars, low magnification 1000 µm; high magnification, 10 µm. “+” high gene expression, “−” low or absent gene expression. SO Stratum oriens, SP Stratum pyramidale, SR Stratum radiatum, SG Subgranular zone, \DG Dentate gyrus without SG, SLM Stratum lacunosum-moleculare.
Fig. 5 Differential patterning of AST1 and AST2 in adult mouse brain. Multiplexed fluorescence in situ hybridization was used to map locations of AST1 and AST2. a AST1 was identified by high expression levels of Gfap, Agt, and Slc1a3. b AST2 was identified by low expression/absence of Agt and high expression of both Unc13c and Slc1a3. Mapping was performed on three sections obtained from three independent animals aged between P56 and P60. Representative images are shown. Top left: low-magnification image of a coronal section. Black dots show the distribution of the astrocyte subtype through one brain hemisphere. Brain regions are defined manually based on definitions from the Allen Brain Atlas. High-magnification images (below) show the localization of markers to specific cells defined on the basis of nuclear (DAPI, blue) staining. Right: bar plots (showing from left to right) fluorescence counts per RNA marker per cell (shown for all cells across all sections analyzed), the distribution of the subtype between brain regions, and the distribution of the subtype normalized to the total number of astrocytes per brain region (all Slc1a3 + cells). Astrocytes belonging to the subtype of interest are highlighted by a shaded box (color-coded according to the scheme used in Fig. 2a). Astrocyte numbers across layers are given as average per section analyzed. Error bars are equivalent across the figure and represent SEM. Scale bars, low magnification 1000 µm; high magnification, 10 µm. “+” high gene expression, “−” low or absent gene expression. SO Stratum oriens, SP Stratum pyramidale, SR Stratum radiatum, SG Subgranular zone, \DG Dentate gyrus without SG, SLM Stratum lacunosum-moleculare.CC BY · PubMed Central
Fig. 6 Differential patterning of AST3 in adult mouse brain. Multiplexed fluorescence in situ hybridization was used to map the location of AST3. Due to technical limitations, AST3 was mapped using a split marker approach. Sections were assessed for (a) low expression/absence of Gfap with expression of Agt and Slc1a3 (to differentiate AST3 from AST1) and (b) low expression/absence of Unc13c with expression of Agt and Slc1a3 (to discriminate between AST3 and AST2). Mapping was on three sections from three independent animals aged P56–P60. Representative images are shown. Top left: low-magnification image of a coronal section. Black dots show astrocyte subtype distribution through one hemisphere. Regions are defined manually based on the Allen Brain Atlas. High-magnification images (below) show the localization of markers to specific cells based on nuclear (DAPI, blue) staining. Right: bar plots (showing left to right) fluorescence counts per marker per cell (for all cells across all sections analyzed), the distribution of the subtype between brain regions and the distribution of the subtype normalized to the number of astrocytes per brain region (all Slc1a3 + cells). Astrocytes belonging to the subtype of interest are highlighted by a shaded box (color-coded according to the scheme used in Fig. 2a). Astrocyte numbers across layers are given as average per section analyzed. Error bars are equivalent across the figure and represent SEM. Scale bars, low magnification 1000 µm; high magnification, 10 µm. “+” high gene expression, “−” low or absent gene expression. SO Stratum oriens, SP Stratum pyramidale, SR Stratum radiatum, SG Subgranular zone, \DG Dentate gyrus without SG, SLM Stratum lacunosum-moleculare.
Fig. 6 Differential patterning of AST3 in adult mouse brain. Multiplexed fluorescence in situ hybridization was used to map the location of AST3. Due to technical limitations, AST3 was mapped using a split marker approach. Sections were assessed for (a) low expression/absence of Gfap with expression of Agt and Slc1a3 (to differentiate AST3 from AST1) and (b) low expression/absence of Unc13c with expression of Agt and Slc1a3 (to discriminate between AST3 and AST2). Mapping was on three sections from three independent animals aged P56–P60. Representative images are shown. Top left: low-magnification image of a coronal section. Black dots show astrocyte subtype distribution through one hemisphere. Regions are defined manually based on the Allen Brain Atlas. High-magnification images (below) show the localization of markers to specific cells based on nuclear (DAPI, blue) staining. Right: bar plots (showing left to right) fluorescence counts per marker per cell (for all cells across all sections analyzed), the distribution of the subtype between brain regions and the distribution of the subtype normalized to the number of astrocytes per brain region (all Slc1a3 + cells). Astrocytes belonging to the subtype of interest are highlighted by a shaded box (color-coded according to the scheme used in Fig. 2a). Astrocyte numbers across layers are given as average per section analyzed. Error bars are equivalent across the figure and represent SEM. Scale bars, low magnification 1000 µm; high magnification, 10 µm. “+” high gene expression, “−” low or absent gene expression. SO Stratum oriens, SP Stratum pyramidale, SR Stratum radiatum, SG Subgranular zone, \DG Dentate gyrus without SG, SLM Stratum lacunosum-moleculare.CC BY · PubMed Central
Fig. 7 Schematic summary of astrocyte subtype positions in adult mouse brain. Indicated positions are based on in situ hybridization data (Figs. 4–6) and are marked on a representative sagittal section of adult mouse brain (adapted from the Allen Mouse Brain Atlas). Subtypes are color-coded (as in Fig. 2a). Scale bar, 500 µm.
Fig. 7 Schematic summary of astrocyte subtype positions in adult mouse brain. Indicated positions are based on in situ hybridization data (Figs. 4–6) and are marked on a representative sagittal section of adult mouse brain (adapted from the Allen Mouse Brain Atlas). Subtypes are color-coded (as in Fig. 2a). Scale bar, 500 µm.CC BY · PubMed Central
Fig. 8 Astrocyte subpopulations display distinct Ca2+ transient properties. Calcium transients in SR101-labeled astrocytes were detected using Fluo-4. Measurements were made in acute brain slices containing cortical layer 1 (L1), cortical layers 3–5 (L3–5), and the CA1 region of the hippocampus (CA1). Transients were recorded under sequential conditions of baseline activity (BASE), tetrodotoxin (TTX), and TTX plus phenylephrine (PHE). a Representative astrocytes (arrowheads) and the calcium transients recorded from them under each experimental condition. Scale bar, 50 µm. b The total population of active astrocytes was defined as cells responding to application of PHE. The fraction of this population displaying Ca2+ transients under BASE and TTX conditions is shown in blue. c Transient parameters grouped by the brain region recorded. Numerical values are the means for each condition. d Hierarchical clustering of Ca2+ transient parameters after application of PHE. e Proportion of astrocytes from the various brain regions per cluster defined in d. f Astrocyte peak parameters grouped per cluster. One dot equals one cell in c, d, and f. Plots in c and f show mean ± SD. Data normality was tested using a Shapiro–Wilk test. Significant differences were verified using a Kruskal–Wallis test with post-hoc Dunn’s test, with p-values adjusted with the Benjamini–Hochberg method. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. AUC area under the curve. Nine animals were used. In total, 614 cells were analyzed.
Fig. 8 Astrocyte subpopulations display distinct Ca2+ transient properties. Calcium transients in SR101-labeled astrocytes were detected using Fluo-4. Measurements were made in acute brain slices containing cortical layer 1 (L1), cortical layers 3–5 (L3–5), and the CA1 region of the hippocampus (CA1). Transients were recorded under sequential conditions of baseline activity (BASE), tetrodotoxin (TTX), and TTX plus phenylephrine (PHE). a Representative astrocytes (arrowheads) and the calcium transients recorded from them under each experimental condition. Scale bar, 50 µm. b The total population of active astrocytes was defined as cells responding to application of PHE. The fraction of this population displaying Ca2+ transients under BASE and TTX conditions is shown in blue. c Transient parameters grouped by the brain region recorded. Numerical values are the means for each condition. d Hierarchical clustering of Ca2+ transient parameters after application of PHE. e Proportion of astrocytes from the various brain regions per cluster defined in d. f Astrocyte peak parameters grouped per cluster. One dot equals one cell in c, d, and f. Plots in c and f show mean ± SD. Data normality was tested using a Shapiro–Wilk test. Significant differences were verified using a Kruskal–Wallis test with post-hoc Dunn’s test, with p-values adjusted with the Benjamini–Hochberg method. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. AUC area under the curve. Nine animals were used. In total, 614 cells were analyzed.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 71 references

Cited by

Sorted by citations

Items 701–800 of 809

Related

See all

Metadata

Authors
Batiuk, M. Y., Martirosyan, A., Wahis, J., de Vin, F., Marneffe, C., Kusserow, C., Koeppen, J., Viana, J. F., Oliveira, J. F., Voet, T., Ponting, C. P., Belgard, T. G., & Holt, M. G.
Affiliations
  • VIB-KU Leuven Center for Brain & Disease Research (BE)
  • University of Minho (PT)
  • Universidade Politécnica do Cávado e do Ave (PT)
  • Wellcome Sanger Institute (GB)
  • University of Oxford (GB)
Corresponding author
Holt, Matthew G.
ORCID
Venue
Nature Communications
Publisher
Springer Science and Business Media LLC
Published
2020-03-05
Volume
11
Issue
1
Pages
1220-1220
ISSN
2041-1723 (ISSN-L)
DOI
10.1038/s41467-019-14198-8
PMC
PMC7058027
OpenAlex
W3009790191
Cited by
805
References
71
Citations by year
2026: 102 · 2025: 135 · 2024: 137 · 2023: 132 · 2022: 139 · 2021: 112 · 2020: 46
FWCI
34.80
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Lirias (KU Leuven) (submittedVersion, other-oa)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Edinburgh Research Explorer (University of Edinburgh) (submittedVersion, cc-by)
Funders
  • European Commission (604102, 2007–2013, FP7 2007-2013, FP7/2007)
  • Fonds Wetenschappelijk Onderzoek (12V7519N, G066715N, AKUL/13/39, I001818N)
  • KU Leuven
  • Fondation contre le Cancer (2015-143)
  • Medical Research Council (FP7/20072013, MC_UU_00007/15)
  • FP7 Ideas: European Research Council (281961)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Life Sciences › Neuroscience › Cellular and Molecular Neuroscience › Neuroscience and Neuropharmacology Research
  • Life Sciences › Neuroscience › Neurology › Neuroinflammation and Neurodegeneration Mechanisms
Keywords
astrocytes, astrocyte subtypes, forebrain, cerebral cortex, hippocampus, single-cell transcriptomics, mouse brain
MeSH terms
Animals, Astrocytes, Gene Expression Regulation, Mice, Inbred C57BL, Organ Specificity, RNA, Messenger, Calcium Signaling, Cell Shape, Neurogenesis, Single-Cell Analysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review