CoChat

The DOI matches its Crossref record; 64 of the 65 references it cites were found in a registry or an index; 38 of the 43 sources its claims cite were checked for support.

Modest changes in Spi1 dosage reveal the potential for altered microglial function as seen in Alzheimer’s disease

Open access

Published version of a preprint

Abstract

Genetic association studies have identified multiple variants at the SPI1 locus that modify risk and age of onset for Alzheimer’s Disease (AD). Reports linking risk variants to gene expression suggest that variants denoting higher SPI1 expression are likely to have an earlier AD onset, and several other AD risk genes contain PU.1 binding sites in the promoter region. Overall, this suggests the level of SPI1 may alter microglial phenotype potentially impacting AD. This study determined how the microglial transcriptome was altered following modest changes to Spi1 expression in primary mouse microglia. RNA-sequencing was performed on microglia with reduced or increased Spi1/PU.1 expression to provide an unbiased approach to determine transcriptomic changes affected by Spi1. Reduction in microglial Spi1 resulted in the dysregulation of transcripts encoding proteins involved in DNA replication pathways while an increased Spi1 results in an upregulation of genes associated with immune response pathways. In addition, 194 Spi1 dose-sensitive genes were identified, affecting innate immune and interferon response pathways.

Overview

CoChat Summary

## TL;DR

Modest bidirectional changes in Spi1/PU.1 expression in primary mouse microglia caused large transcriptomic shifts: reduced Spi1 mainly affected cell cycle/DNA repair genes, while increased Spi1 enriched immune, interferon, and antigen-presentation pathways. The authors identify 194 Spi1 dose-sensitive genes and report that a subset of over-expression–linked genes was enriched for AD genetic risk.

## Key findings

  • Spi1 knock-down in sorted microglia changed 1615 genes up and 1462 genes down at adjusted p < 0.05; 230 down-regulated genes exceeded the -2 log2 fold-change cut-off.
  • Spi1 over-expression changed 284 genes up and 62 genes down at p < 0.05; 71 up-regulated and 4 down-regulated genes exceeded a log2 fold-change of 2.
  • Comparing knock-down and over-expression identified 194 Spi1 dose-sensitive genes: 162 decreased with knock-down and increased with over-expression, and 32 showed the opposite pattern; these were enriched for immune response, especially interferon and viral defence pathways.
  • GO analysis found lower Spi1 was most strongly associated with cell cycle and DNA repair pathways, while higher Spi1 was associated with antigen presentation, immune system processes, and response to interferon.
  • In a comparison to IGAP AD GWAS data, a 21-gene set from the over-expression dataset (genes with Benjamini–Hochberg adjusted p ≤ 1 × 10−6 for differential expression) was enriched for AD genetic risk (adjusted p = 0.035).

## Methods

  • Primary mixed glia were prepared from 8-week-old C57BL/6J mouse brains; the cultures were repeated 10 times (n = 10), with 4 used for RNA-seq and 6 for validation experiments.
  • Microglia were infected on day 10 with Spi1 shRNA, control shRNA, Spi1 pSIEW over-expression, or pSIEW control lentivirus; microglia were then sorted by CD11b/CD45 and GFP before RNA-seq.
  • RNA-seq used TruSeq stranded total RNA with Ribo-Zero GOLD, HiSeq 2500 sequencing at 30–40 million pairs, alignment with STAR, differential expression with DESeq2, GO analysis with DAVID, clustering in R, PU.1 binding-site overlap with a published ChIP-seq dataset, and AD-risk enrichment with MAGMA using IGAP data.

## Limitations

  • The paper states that the cultures were supplemented with TGF-β, which “partially compensates” for missing environmental cues, but does not fully replicate the in vivo context; it also states there were no suitable transgenic mice available to study Spi1 dose in situ.
  • The authors note that pathway inference from RNA-seq does not prove phenotype and that further work would be required to confirm experimentally whether higher Spi1 truly makes microglia more interferon-responsive.
  • Evident from the text: the study is in primary mouse microglia/mixed glia, so direct applicability to human AD biology is uncertain, and the manuscript itself notes possible species differences between human and mouse microglia.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41598-021-94324-z

Cites

64 of 65 references verified0 citation errors · 1 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Results of RNA-Seq experiment in primary mixed-glial cultures (A) Volcano plots summarising the distribution of genes in the Spi1 knock-down (purple) and Spi1 over-expression (orange). In the Spi1 knock-down dataset a 1462 genes were down-regulated (grey background) and 1615 up-regulated with a p value of ≤ 0.05 (solid lines), as indicated by the numbers on the graph. In the Spi1 over-expression dataset the majority of the genes, 284, were up-regulated and 62 genes had down-regulated expression. (B) Genes that were significantly changed in both the Spi1 knock down (KD, purple) and the Spi1 over expression (OE, orange) using a p ≤ 0.05 threshold. A Plot of adjusted p values from all genes in both datasets. Genes that were below the p ≤ 0.05 threshold in the Spi1 knock-down dataset are highlighted in purple, those that were below the p ≤ 0.05 cut-off in the Spi1 over-expression are within the orange bar. In the bottom left corner the red line surrounds the 196 genes that were significantly changed in both datasets. (C) Venn diagrams summarising the gene expression that were significantly changed (p ≤ 0.05) in either the Spi1 knock-down dataset (purple), the Spi1 over-expression dataset (orange) or changed in both datasets (red). Of interest were the genes with expression that appears to be sensitive to the Spi1 dose in microglia, namely the 194 genes that appear to be expressed relative to the dose of Spi1. (D) The log10 number of Spi1 mRNA fragments per Kilobase of transcript per Million mapped reads (FPKM) shows that Spi1 was lower in Spi1 shRNA infected microglia than in control shRNA infected microglia (One-tailed Unpaired T-test on the log10 transformed data, p = 0.0166, n = 4 for control shRNA and n = 3 for Spi1 shRNA). PU.1 protein expression was normalised to the non-infected (NI) samples in each experiment (as described in methods). Microglia infected with the Spi1 shRNA (solid purple) have reduced PU.1 expression compared to cells infected with a control shRNA (lined purple) (One-tailed Paired T-test, p = 0.0013, n = 3 per group). (E) Spi1 mRNA expression (FPKM) was increased in the samples infected with the Spi1 over-expression construct (filled orange) compared to empty vector control samples (outline orange), (One-tailed Paired T-test on the Log10 transformed data, p = 0.0013, n = 4 per group).The Spi1 over-expression virus (Spi1 pSIEW) increased PU.1 protein expression in microglia by roughly half compared to cells infected with an empty vector control (lined orange). One-tailed Paired T-test, p = 0.0131, n = 4 per group) (D–E) Each dot represents the value from one biological replicate, the means are indicated by the horizontal lines and the error bars display the standard deviation about the mean. All experiments were performed using mixed glia cultures from 8-week-old female C57BL/6J mice. Figures (A–C) were made in GraphPad PRISM 6 (version 3.07) and (D–E) in GraphPad PRISM 8 (version 8.4.3; both GraphPad Software, Inc.).
Figure 1 Results of RNA-Seq experiment in primary mixed-glial cultures (A) Volcano plots summarising the distribution of genes in the Spi1 knock-down (purple) and Spi1 over-expression (orange). In the Spi1 knock-down dataset a 1462 genes were down-regulated (grey background) and 1615 up-regulated with a p value of ≤ 0.05 (solid lines), as indicated by the numbers on the graph. In the Spi1 over-expression dataset the majority of the genes, 284, were up-regulated and 62 genes had down-regulated expression. (B) Genes that were significantly changed in both the Spi1 knock down (KD, purple) and the Spi1 over expression (OE, orange) using a p ≤ 0.05 threshold. A Plot of adjusted p values from all genes in both datasets. Genes that were below the p ≤ 0.05 threshold in the Spi1 knock-down dataset are highlighted in purple, those that were below the p ≤ 0.05 cut-off in the Spi1 over-expression are within the orange bar. In the bottom left corner the red line surrounds the 196 genes that were significantly changed in both datasets. (C) Venn diagrams summarising the gene expression that were significantly changed (p ≤ 0.05) in either the Spi1 knock-down dataset (purple), the Spi1 over-expression dataset (orange) or changed in both datasets (red). Of interest were the genes with expression that appears to be sensitive to the Spi1 dose in microglia, namely the 194 genes that appear to be expressed relative to the dose of Spi1. (D) The log10 number of Spi1 mRNA fragments per Kilobase of transcript per Million mapped reads (FPKM) shows that Spi1 was lower in Spi1 shRNA infected microglia than in control shRNA infected microglia (One-tailed Unpaired T-test on the log10 transformed data, p = 0.0166, n = 4 for control shRNA and n = 3 for Spi1 shRNA). PU.1 protein expression was normalised to the non-infected (NI) samples in each experiment (as described in methods). Microglia infected with the Spi1 shRNA (solid purple) have reduced PU.1 expression compared to cells infected with a control shRNA (lined purple) (One-tailed Paired T-test, p = 0.0013, n = 3 per group). (E) Spi1 mRNA expression (FPKM) was increased in the samples infected with the Spi1 over-expression construct (filled orange) compared to empty vector control samples (outline orange), (One-tailed Paired T-test on the Log10 transformed data, p = 0.0013, n = 4 per group).The Spi1 over-expression virus (Spi1 pSIEW) increased PU.1 protein expression in microglia by roughly half compared to cells infected with an empty vector control (lined orange). One-tailed Paired T-test, p = 0.0131, n = 4 per group) (D–E) Each dot represents the value from one biological replicate, the means are indicated by the horizontal lines and the error bars display the standard deviation about the mean. All experiments were performed using mixed glia cultures from 8-week-old female C57BL/6J mice. Figures (A–C) were made in GraphPad PRISM 6 (version 3.07) and (D–E) in GraphPad PRISM 8 (version 8.4.3; both GraphPad Software, Inc.).CC BY · PubMed Central
Figure 2 The 20 most siginificantly changed biological pathways in the Spi1 knock-down (A) and Spi1 over-expression (B) datasets assessed using DAVID. In these graphs the Benjamini–Hochberg adjusted p value(− Log10) is displayed on the x-axis, the Gene Ontology (GO) term listed on the y-axis, the percentage of the gene list in each cluster is denoted by the size of the bubble and the colour denotes the fold-change, where a darker colour indicates a stronger enrichment. The vertical black line indicates the p value cut-off of 0.05. Bubble plots were made using the ‘tidyverse’37 and ‘ggpubr’38 packages in R39,40.
Figure 2 The 20 most siginificantly changed biological pathways in the Spi1 knock-down (A) and Spi1 over-expression (B) datasets assessed using DAVID. In these graphs the Benjamini–Hochberg adjusted p value(− Log10) is displayed on the x-axis, the Gene Ontology (GO) term listed on the y-axis, the percentage of the gene list in each cluster is denoted by the size of the bubble and the colour denotes the fold-change, where a darker colour indicates a stronger enrichment. The vertical black line indicates the p value cut-off of 0.05. Bubble plots were made using the ‘tidyverse’37 and ‘ggpubr’38 packages in R39,40.CC BY · PubMed Central
Figure 3 Absolute Log2 fold-changes in Spi1 dose-sensitive genes. (A) Absolute log2 fold-change values were used the 162 genes where p ≤ 0.05 with negative fold-change values in the Spi1 knock-down dataset (purple) and positive fold-change values in the Spi1 over-expression dataset (orange) were assessed separately from the 32 genes where p ≤ 0.05 and Spi1 dose had the opposite effect on expression (B). Two-tailed Spearman Rank test identified significant correlations between the log2 fold-changes in both graphs (approximate p < 0.0001 and r > 0.8). (C) Top 10 most significantly altered “Biological Processes” according to the Benjamini–Hochberg adjusted p value. The bubble size indicates the percentage of the gene list aligned to this pathway. The colour indicates the fold enrichment, which is the proportion of genes present in this list compared to the background gene expression. The vertical black line indicates the p value threshold of 0.05. Figures (A, B) were made in GraphPad PRISM 6 (version 3.07) and (C) using the ‘tidyverse’37 and ‘ggpubr’38 packages in R39,40.
Figure 3 Absolute Log2 fold-changes in Spi1 dose-sensitive genes. (A) Absolute log2 fold-change values were used the 162 genes where p ≤ 0.05 with negative fold-change values in the Spi1 knock-down dataset (purple) and positive fold-change values in the Spi1 over-expression dataset (orange) were assessed separately from the 32 genes where p ≤ 0.05 and Spi1 dose had the opposite effect on expression (B). Two-tailed Spearman Rank test identified significant correlations between the log2 fold-changes in both graphs (approximate p < 0.0001 and r > 0.8). (C) Top 10 most significantly altered “Biological Processes” according to the Benjamini–Hochberg adjusted p value. The bubble size indicates the percentage of the gene list aligned to this pathway. The colour indicates the fold enrichment, which is the proportion of genes present in this list compared to the background gene expression. The vertical black line indicates the p value threshold of 0.05. Figures (A, B) were made in GraphPad PRISM 6 (version 3.07) and (C) using the ‘tidyverse’37 and ‘ggpubr’38 packages in R39,40.CC BY · PubMed Central
Figure 4 Hierarchical clustering analyses with the 20 most siginificantly changed biological pathways each cluster assessed using DAVID. (A) Hierarchical clustering of rows, where each row represents the scaled log10 FPKM values for each gene, with dendrogram higlighting cluster boundaries. The Pearson correlation was used to calculate the Z-Scores and UPGMA agglomeration methods were used. These 6 clusters were produced using the cutree function at a height of 1.59. (B) The gene lists were assessed using DAVID and a bubble plot created to highlight the top 5 most significatly enriched pathways (p value threshold of < 0.05) in each gene cluster. Cluster 1 contains genes that had increased expression following a higher Spi1 expression and are mainly linked to immune responses and MHCII antigen presentation. Cluster 2 appears to contain more of the Spi1 dose-sensitive genes and are enriched for viral/immune response signalling pathways. Genes in Cluster 3 had a lower expression in the Spi1 shRNA dataset and are related to immune response and interferon signalling pathways. Cluster 4 contained no significantly enriched pathways using the p ≤ 0.05 threshold. Cluster 5 contains multiple genes related to cell cycle and DNA replication pathways, whose expression was increased in the Spi1 shRNA samples. The bubble size denotes the fold enrichment and the colour Benjamini–Hochberg adjusted p values. Figures were made with the ‘tidyverse’37, ‘gplots’42, ‘plotly’43, ‘dendextend’44 and ‘colourspace’45 packages in RStudio39,40, code adapted from46,47 complete Markdown is available in Supplementary Information.
Figure 4 Hierarchical clustering analyses with the 20 most siginificantly changed biological pathways each cluster assessed using DAVID. (A) Hierarchical clustering of rows, where each row represents the scaled log10 FPKM values for each gene, with dendrogram higlighting cluster boundaries. The Pearson correlation was used to calculate the Z-Scores and UPGMA agglomeration methods were used. These 6 clusters were produced using the cutree function at a height of 1.59. (B) The gene lists were assessed using DAVID and a bubble plot created to highlight the top 5 most significatly enriched pathways (p value threshold of < 0.05) in each gene cluster. Cluster 1 contains genes that had increased expression following a higher Spi1 expression and are mainly linked to immune responses and MHCII antigen presentation. Cluster 2 appears to contain more of the Spi1 dose-sensitive genes and are enriched for viral/immune response signalling pathways. Genes in Cluster 3 had a lower expression in the Spi1 shRNA dataset and are related to immune response and interferon signalling pathways. Cluster 4 contained no significantly enriched pathways using the p ≤ 0.05 threshold. Cluster 5 contains multiple genes related to cell cycle and DNA replication pathways, whose expression was increased in the Spi1 shRNA samples. The bubble size denotes the fold enrichment and the colour Benjamini–Hochberg adjusted p values. Figures were made with the ‘tidyverse’37, ‘gplots’42, ‘plotly’43, ‘dendextend’44 and ‘colourspace’45 packages in RStudio39,40, code adapted from46,47 complete Markdown is available in Supplementary Information.CC BY · PubMed Central
Figure 5 Comparison of Spi1 KD, Spi1 OE and 194 Spi1 Dose-sensitive genes to Spi1/PU.1 Chip-Seq dataset. Spi1 binding sites were determined from PU.1 Chip-Seq published data available online30 as described in methods. Fisher’s exact tests were used to determine if there was a significant overlap between the datasets (A) includes the results from all protein coding genes with a Spi1 binding region (defined by Chip-seq in grey) that are also expressed in KD (purple), OE (orange) RNA-seq datasets or the Spi1 dose-sensitive genes (red). (B) The proportion of genes expressed in the RNA-seq datasets that contained a Spi1 binding site within the promoter region. The KD dataset contained a significant number of genes with a Spi1 binding sequence in the promoter (Fisher’s exact test, overlapping p value = 1.6 × 10−16). No significant enrichment was noted in either the Spi1 OE or dose sensitive genes. Figures were made in PowerPoint based on results from ‘GeneOverlap’48 package in RStudio39,40.
Figure 5 Comparison of Spi1 KD, Spi1 OE and 194 Spi1 Dose-sensitive genes to Spi1/PU.1 Chip-Seq dataset. Spi1 binding sites were determined from PU.1 Chip-Seq published data available online30 as described in methods. Fisher’s exact tests were used to determine if there was a significant overlap between the datasets (A) includes the results from all protein coding genes with a Spi1 binding region (defined by Chip-seq in grey) that are also expressed in KD (purple), OE (orange) RNA-seq datasets or the Spi1 dose-sensitive genes (red). (B) The proportion of genes expressed in the RNA-seq datasets that contained a Spi1 binding site within the promoter region. The KD dataset contained a significant number of genes with a Spi1 binding sequence in the promoter (Fisher’s exact test, overlapping p value = 1.6 × 10−16). No significant enrichment was noted in either the Spi1 OE or dose sensitive genes. Figures were made in PowerPoint based on results from ‘GeneOverlap’48 package in RStudio39,40.CC BY · PubMed Central
Figure 6 Summary of Spi1 over-expression genes that were significantly enriched in the IGAP dataset. (A) Volcano plot of the Spi1 over-expression gene expression changes relative to the empty vector control, highlighting the p ≤ 1 × 10−6 threshold that was shown to be significantly enriched by MAGMA analysis in the IGAP dataset (MAGMA’s empirical multiple testing corrected p value 0.035). The adjusted p value for this gene set was 0.0033 and an enrichment effect size of 0.51. (B) Bubble plot of the genes from the Spi1 over expression dataset that were enriched for AD genetic risk via MAGMA comparison to the human IGAP database, against the Benjamini–Hochberg adjusted p values from the Spi1 over-expression RNA-sequencing data with the fold-change shown in colour. The vertical black line indicates the p value threshold of 0.05. Figure (A) was made in GraphPad PRISM 6 (version 3.07) and (B) using the ‘tidyverse’37 and ‘ggpubr’38 packages in in RStudio39,40.
Figure 6 Summary of Spi1 over-expression genes that were significantly enriched in the IGAP dataset. (A) Volcano plot of the Spi1 over-expression gene expression changes relative to the empty vector control, highlighting the p ≤ 1 × 10−6 threshold that was shown to be significantly enriched by MAGMA analysis in the IGAP dataset (MAGMA’s empirical multiple testing corrected p value 0.035). The adjusted p value for this gene set was 0.0033 and an enrichment effect size of 0.51. (B) Bubble plot of the genes from the Spi1 over expression dataset that were enriched for AD genetic risk via MAGMA comparison to the human IGAP database, against the Benjamini–Hochberg adjusted p values from the Spi1 over-expression RNA-sequencing data with the fold-change shown in colour. The vertical black line indicates the p value threshold of 0.05. Figure (A) was made in GraphPad PRISM 6 (version 3.07) and (B) using the ‘tidyverse’37 and ‘ggpubr’38 packages in in RStudio39,40.CC BY · PubMed Central

Verification

Claims the paper makes

28 Valid · 2 Not valid · 5 Not checked · 55 Own results and interpretations

59 more claims · 13 with their cited sources checked

Create a free account to see all 90 claims and to verify the 1 reference still open.

See all claims — free account
  • Probably not a substantial claim. 14
  • It may state two separate claims as one. 15
  • Borderline: one of the tests for a substantial claim is unclear. 1

Create a free account to see these claims.

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?

Citations

Cites

Show all 65 references

Research works printed but not linked: 1

  1. Prince, M. et al. Dementia UK: Second Edition - Overview 61 (Alzheimer's Society, 2014).

Cited by

See all 42

Metadata

Authors
Jones, R. E., Andrews, R., Holmans, P., Hill, M., & Taylor, P. R.
Affiliations
  • UK Dementia Research Institute (GB)
  • Cardiff University (GB)
Corresponding author
Taylor, Philip R.
Venue
Scientific Reports
Published
2021-07-22
Volume
11
Issue
1
Pages
14935-14935
ISSN
2045-2322 (ISSN-L)
OpenAlex
W3133041819
Cited by
39
References
62
Citations by year
2026: 6 · 2025: 10 · 2024: 4 · 2023: 12 · 2022: 6 · 2021: 1
FWCI
2.99
Type
journal article
Language
en
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • ORCA Online Research @Cardiff (Cardiff University) (acceptedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Medical Research Council UK
  • Wellcome Trust (107964/Z/15/Z)
  • UK MRC Dementia Research Institute
Topics
  • Life Sciences › Neuroscience › Neurology › Neuroinflammation and Neurodegeneration Mechanisms
  • Life Sciences › Neuroscience › Cellular and Molecular Neuroscience › Alzheimer's disease research and treatments
  • Health Sciences › Medicine › Psychiatry and Mental health › Dementia and Cognitive Impairment Research
Keywords
PU.1, Alzheimer's disease, microglia, gene dosage, microglial transcriptome, interferon response, innate immune response, DNA replication, genetic risk variants, microglial phenotype
MeSH terms
Proto-Oncogene Protein Spi-1, Alzheimer Disease, Animals, Humans, Interferons, Proto-Oncogene Proteins, Signal Transduction, Trans-Activators, Sequence Analysis, RNA, Microglia, Gene Dosage, Polymorphism, Single Nucleotide, Gene Expression Profiling, Mice, Gene Regulatory Networks, Primary Cell Culture

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.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-07, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-07 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Modest changes in Spi1 dosage reveal the potential for altered microglial function as seen in Alzheimer’s disease. https://app.cochat.ai/papers/10.1038/s41598-021-94324-z
The paper
Jones, R. E., Andrews, R., Holmans, P., Hill, M., & Taylor, P. R. (2021). Modest changes in Spi1 dosage reveal the potential for altered microglial function as seen in Alzheimer’s disease. Scientific Reports. https://doi.org/10.1038/s41598-021-94324-z

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.