CoChat

Endosomal TLR3, TLR7, and TLR8 control neuronal morphology through different transcriptional programs

Open access

The DOI matches its Crossref record; 68 of the 69 references it cites were found in a registry or an index.

Abstract

Neuroinflammation is associated with diverse neurological disorders. Endosomal Toll-like receptors (TLRs) including TLR3, TLR7, and TLR8 cell-autonomously regulate neuronal differentiation. However, the mechanisms by which these three TLRs affect neuronal morphology are unclear. In this study, we compare these TLRs in mouse neurons. By combining in vitro neuronal cultures, in utero electroporation, and transcriptomic profiling, we show that TLR8, TLR7, and TLR3 promote dendritic pruning via MYD88 signaling. However, they induce different transcriptomic profiles related to innate immunity, signaling, and neuronal development. The temporal expression patterns and the effects on neuronal morphology are not identical upon activation of these endosomal TLRs. Pathway analyses and in vitro studies specifically implicate mitogen-activated protein kinase signaling in TLR8-mediated dendritic pruning. We further show that TLR8 is more critical for dendritic arborization at a late development stage in vivo. The activation of TLR8, TLR7, or TLR3 results in dendritic shortening, and TLR7 and TLR3 but not TLR8 also control axonal growth. In-depth transcriptomic analyses show that TLRs use different downstream pathways to control neuronal morphology, which may contribute to neuronal development and pathological responses.

Overview

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

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

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

Integrity signals

Registry

Crossref · 10.1083/jcb.201712113

Cites

68 of 68 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1. Tlr8is up-regulated in Tlr7−/− neurons and restricts dendritic growth but not axonal extension of cortical and hippocampal neurons. (A and B) Up-regulated Tlr8 RNA levels in Tlr7-deficient neurons. (B) Increased Tlr8 RNA levels in mature neuronal cultures. Q-PCR was performed and normalized with internal control Gapdh. Numbers (N) of independent repeated experiments are indicated. For each experiment, three embryonic cortices and hippocampi were pooled for each set of experiments. (C and D) Different concentrations of CL075 with or without 5 µM poly dT were added to WT and Tlr7−/− cultured neurons at 4 DIV. 1 d later, neurons were harvested for immunostaining using dendritic marker MAP2 antibodies. (C) Dosage effect on WT neurons of CL075 in the presence or absence of poly dT. Based on a previous study (Gorden et al., 2006), CL075 activates TLR7 and CL075/poly dT activates TLR8. (D) Comparison of the responses of WT and Tlr7−/− neurons. (E) Tlr7−/− cultured neurons were treated with CL075 alone or mixed with poly dT at 2 DIV for 24 h and analyzed using axonal marker SMI312 antibodies. (F) Coimmunoprecipitation of Myc-tagged TLR8 and HA-tagged MYD88. The fast-migrating protein species was likely the product of TLR8 proteolysis (Lee and Barton, 2014). IB, immunoblot; IP immunoprecipitation. (G) Dendritic morphology of Tlr7−/−– and Tlr7−/−;Myd88−/−-cultured neurons at 5 DIV. Bars, 20 µm. The sample size (n) indicates the number of examined neurons, which were randomly collected blind from two (C, D, and F) or three (E) independent experiments. Data are presented as mean + SEM (error bars). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Two-tailed nonparametric Mann–Whitney test (A) and one-way (B and E) and two-way (C, D, and G) ANOVA with Bonferroni’s multiple comparisons tests were used.
Figure 1. Tlr8is up-regulated in Tlr7−/− neurons and restricts dendritic growth but not axonal extension of cortical and hippocampal neurons. (A and B) Up-regulated Tlr8 RNA levels in Tlr7-deficient neurons. (B) Increased Tlr8 RNA levels in mature neuronal cultures. Q-PCR was performed and normalized with internal control Gapdh. Numbers (N) of independent repeated experiments are indicated. For each experiment, three embryonic cortices and hippocampi were pooled for each set of experiments. (C and D) Different concentrations of CL075 with or without 5 µM poly dT were added to WT and Tlr7−/− cultured neurons at 4 DIV. 1 d later, neurons were harvested for immunostaining using dendritic marker MAP2 antibodies. (C) Dosage effect on WT neurons of CL075 in the presence or absence of poly dT. Based on a previous study (Gorden et al., 2006), CL075 activates TLR7 and CL075/poly dT activates TLR8. (D) Comparison of the responses of WT and Tlr7−/− neurons. (E) Tlr7−/− cultured neurons were treated with CL075 alone or mixed with poly dT at 2 DIV for 24 h and analyzed using axonal marker SMI312 antibodies. (F) Coimmunoprecipitation of Myc-tagged TLR8 and HA-tagged MYD88. The fast-migrating protein species was likely the product of TLR8 proteolysis (Lee and Barton, 2014). IB, immunoblot; IP immunoprecipitation. (G) Dendritic morphology of Tlr7−/−– and Tlr7−/−;Myd88−/−-cultured neurons at 5 DIV. Bars, 20 µm. The sample size (n) indicates the number of examined neurons, which were randomly collected blind from two (C, D, and F) or three (E) independent experiments. Data are presented as mean + SEM (error bars). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Two-tailed nonparametric Mann–Whitney test (A) and one-way (B and E) and two-way (C, D, and G) ANOVA with Bonferroni’s multiple comparisons tests were used.CC BY-NC-SA · PubMed Central
Figure 2. Tlr8 knockdown promotes dendritic growth in vitro and in vivo. (A) Immunoblotting using HA tag antibody to examine the effect of TLR8 knockdown constructs on TLR8 expression in HEK293T cells. Two expression vectors were used to express miR-Ctrl and miR-Tlr8 as indicated. (B) Tlr7−/−-cultured neurons were transfected with miR-Ctrl and miR-Tlr8 at 2 DIV, and then treated with CL075/poly dT at 4 DIV. Dendritic morphology was analyzed at 5 DIV based on GFP signals. The sample sizes (n) of the examined neurons are indicated. Data were collected from two independent experiments. (C and D) IUE with miR-Ctrl or miR-Tlr8 was performed at E15.5 with WT (C) and Tlr7−/− (D) mice. Layer 2/3 cortical neurons were outlined by GFP signal at P7, P14, and P21 as indicated. N, number of examined mice; n, number of examined neurons. Bars: (B) 20 µm; (C and D) 50 µm. The data are presented as mean + SEM (error bars). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Two-way ANOVA with Bonferroni’s multiple comparisons test (B) and two-tailed nonparametric Mann–Whitney tests (C and D) were used.
Figure 2. Tlr8 knockdown promotes dendritic growth in vitro and in vivo. (A) Immunoblotting using HA tag antibody to examine the effect of TLR8 knockdown constructs on TLR8 expression in HEK293T cells. Two expression vectors were used to express miR-Ctrl and miR-Tlr8 as indicated. (B) Tlr7−/−-cultured neurons were transfected with miR-Ctrl and miR-Tlr8 at 2 DIV, and then treated with CL075/poly dT at 4 DIV. Dendritic morphology was analyzed at 5 DIV based on GFP signals. The sample sizes (n) of the examined neurons are indicated. Data were collected from two independent experiments. (C and D) IUE with miR-Ctrl or miR-Tlr8 was performed at E15.5 with WT (C) and Tlr7−/− (D) mice. Layer 2/3 cortical neurons were outlined by GFP signal at P7, P14, and P21 as indicated. N, number of examined mice; n, number of examined neurons. Bars: (B) 20 µm; (C and D) 50 µm. The data are presented as mean + SEM (error bars). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Two-way ANOVA with Bonferroni’s multiple comparisons test (B) and two-tailed nonparametric Mann–Whitney tests (C and D) were used.CC BY-NC-SA · PubMed Central
Figure 3. TLR8 activation induces dendritic pruning of neurons. (A) TLR8 activation impairs dendritic arborization of WT mature neurons. (B and C) TLR8 activation induces dendritic pruning of WT neurons at 17 DIV (B) and Tlr7−/−-cultured neurons at 7 DIV (C). (A–C) The data were randomly collected blind from two independent experiments and are presented as mean + SEM (error bars). The numbers of examined neurons (n) are indicated. Bars: (A and B) 100 µm; (C) 50 µm. *, P < 0.05; ***, P < 0.001. A two-tailed nonparametric Mann–Whitney test was used.
Figure 3. TLR8 activation induces dendritic pruning of neurons. (A) TLR8 activation impairs dendritic arborization of WT mature neurons. (B and C) TLR8 activation induces dendritic pruning of WT neurons at 17 DIV (B) and Tlr7−/−-cultured neurons at 7 DIV (C). (A–C) The data were randomly collected blind from two independent experiments and are presented as mean + SEM (error bars). The numbers of examined neurons (n) are indicated. Bars: (A and B) 100 µm; (C) 50 µm. *, P < 0.05; ***, P < 0.001. A two-tailed nonparametric Mann–Whitney test was used.CC BY-NC-SA · PubMed Central
Figure 4. TLR8 activation alters dendritic spine density and expression of immediate early genes but not mEPSC. (A) The effect of CL075/poly dT treatment for a day on dendritic spine density of WT cultured neurons transfected with GFP-actin. Data were randomly collected blind from two independent experiments. Cell numbers of examined neurons were 29 for control and 38 for CL075/poly dT treatment. n, number of examined dendrites. (B) The results of mEPSC after CL075/poly dT treatment for 1 d. Number of examined neurons (n) collected from four independent experiments (N) are indicated. (C) Expression of immediate early genes after CL075/poly dT treatment at 18 DIV for 6 h. Q-PCR analysis was normalized with internal control Gapdh. Each dot indicates the result of an independently repeated experiment. Three embryonic cortices and hippocampi were combined for one set of experiments. Means + SEM (error bars) are also indicated in A–C. Bars: (A, original) 50 µm; (enlarged) 1 µm. *, P < 0.05; **, P < 0.01. A two-tailed nonparametric Mann–Whitney test was used.
Figure 4. TLR8 activation alters dendritic spine density and expression of immediate early genes but not mEPSC. (A) The effect of CL075/poly dT treatment for a day on dendritic spine density of WT cultured neurons transfected with GFP-actin. Data were randomly collected blind from two independent experiments. Cell numbers of examined neurons were 29 for control and 38 for CL075/poly dT treatment. n, number of examined dendrites. (B) The results of mEPSC after CL075/poly dT treatment for 1 d. Number of examined neurons (n) collected from four independent experiments (N) are indicated. (C) Expression of immediate early genes after CL075/poly dT treatment at 18 DIV for 6 h. Q-PCR analysis was normalized with internal control Gapdh. Each dot indicates the result of an independently repeated experiment. Three embryonic cortices and hippocampi were combined for one set of experiments. Means + SEM (error bars) are also indicated in A–C. Bars: (A, original) 50 µm; (enlarged) 1 µm. *, P < 0.05; **, P < 0.01. A two-tailed nonparametric Mann–Whitney test was used.CC BY-NC-SA · PubMed Central
Figure 5. Secreted factors are not required for down-regulation of dendritic growth by TLR8 activation. (A and B) The effects of TLR7 and TLR8 activation for 6 h on cytokine and chemokine expressions of cultured neurons. Q-PCR was performed by normalization with internal control Gm10069 or Gapdh as indicated. Three embryonic cortices and hippocampi of the same genotype were combined for one set of experiments. One dot indicates the data of one independent experiment. Data were collected from six experiments (n = 6). (C and D) Experiments of conditioned medium (CM). Experimental flowcharts are summarized in the left panel. The effects of conditioned medium on dendritic length are summarized in the right panel. The sample size (n) indicates the number of examined neurons randomly collected blind from two (D) or three (C) independent experiments. The data are presented as mean + SEM (error bars). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Two-way (A) and one-way (B) ANOVA with Bonferroni’s multiple comparisons test as well as two-tailed nonparametric Mann–Whitney tests (C and D) were used.
Figure 5. Secreted factors are not required for down-regulation of dendritic growth by TLR8 activation. (A and B) The effects of TLR7 and TLR8 activation for 6 h on cytokine and chemokine expressions of cultured neurons. Q-PCR was performed by normalization with internal control Gm10069 or Gapdh as indicated. Three embryonic cortices and hippocampi of the same genotype were combined for one set of experiments. One dot indicates the data of one independent experiment. Data were collected from six experiments (n = 6). (C and D) Experiments of conditioned medium (CM). Experimental flowcharts are summarized in the left panel. The effects of conditioned medium on dendritic length are summarized in the right panel. The sample size (n) indicates the number of examined neurons randomly collected blind from two (D) or three (C) independent experiments. The data are presented as mean + SEM (error bars). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Two-way (A) and one-way (B) ANOVA with Bonferroni’s multiple comparisons test as well as two-tailed nonparametric Mann–Whitney tests (C and D) were used.CC BY-NC-SA · PubMed Central
Figure 6. Gene list and heat map of the top targets of TLR3, TLR7, and TLR8 activation in neurons. (A and B) Transcriptomic profiles after TLR3, TLR7, and TLR8 activation for 6 h of WT and Tlr7−/−-cultured neurons at 4 DIV were analyzed by RNA-seq. (A) List of the top 10 differentially expressed genes based on the p-values from TLR3-, TLR7-, and TLR8-activated neurons. Red, up-regulated; dark blue, down-regulated. (B) Heat map representing the expression levels of genes across TLR3, TLR7, and TLR8 activation. The top 30 up- and down-regulated genes of each group based on fold change were picked for cross-comparison. The color-coded bar is the z score.
Figure 6. Gene list and heat map of the top targets of TLR3, TLR7, and TLR8 activation in neurons. (A and B) Transcriptomic profiles after TLR3, TLR7, and TLR8 activation for 6 h of WT and Tlr7−/−-cultured neurons at 4 DIV were analyzed by RNA-seq. (A) List of the top 10 differentially expressed genes based on the p-values from TLR3-, TLR7-, and TLR8-activated neurons. Red, up-regulated; dark blue, down-regulated. (B) Heat map representing the expression levels of genes across TLR3, TLR7, and TLR8 activation. The top 30 up- and down-regulated genes of each group based on fold change were picked for cross-comparison. The color-coded bar is the z score.CC BY-NC-SA · PubMed Central
Figure 7. Downstream signal networks of TLR3, TLR7, and TLR8 in neurons differ. (A) Venn diagram showing the overlap between TLR3, TLR7, and TLR8 up-regulated (left) and down-regulated (right) genes. Gene lists showing overlap among and between TLR3, TLR7, and TLR8 are available in Table S1. (B) GO of the biological functions for up-regulated and down-regulated genes from TLR3, TLR7, and TLR8 activation. Only four biological processes in the group of TLR8 up-regulated genes were identified. The top five GO biological processes of the rest of the groups are listed. Gene lists of each GO are available in Table S2.
Figure 7. Downstream signal networks of TLR3, TLR7, and TLR8 in neurons differ. (A) Venn diagram showing the overlap between TLR3, TLR7, and TLR8 up-regulated (left) and down-regulated (right) genes. Gene lists showing overlap among and between TLR3, TLR7, and TLR8 are available in Table S1. (B) GO of the biological functions for up-regulated and down-regulated genes from TLR3, TLR7, and TLR8 activation. Only four biological processes in the group of TLR8 up-regulated genes were identified. The top five GO biological processes of the rest of the groups are listed. Gene lists of each GO are available in Table S2.CC BY-NC-SA · PubMed Central
Figure 8. P38 MAPK and ERK are involved in TLR8 activation in cultured neurons. (A and B) WT neurons at 16 DIV (A) and Tlr7−/− neurons at 5 DIV (B) were stimulated with CL075/poly dT for different time periods as indicated. Phosphorylation levels and total protein levels of ERK, P38, and JNK were assessed by immunoblotting. Results are representative of five independent experiments, and the data of each independent experiment is the mean of experimental duplication. The data are presented as the mean + SEM (error bars). *, P < 0.05; **, P < 0.01; ***, P < 0.001. A one-way ANOVA with Bonferroni’s multiple comparisons test was used.
Figure 8. P38 MAPK and ERK are involved in TLR8 activation in cultured neurons. (A and B) WT neurons at 16 DIV (A) and Tlr7−/− neurons at 5 DIV (B) were stimulated with CL075/poly dT for different time periods as indicated. Phosphorylation levels and total protein levels of ERK, P38, and JNK were assessed by immunoblotting. Results are representative of five independent experiments, and the data of each independent experiment is the mean of experimental duplication. The data are presented as the mean + SEM (error bars). *, P < 0.05; **, P < 0.01; ***, P < 0.001. A one-way ANOVA with Bonferroni’s multiple comparisons test was used.CC BY-NC-SA · PubMed Central
Figure 9. P38 and TAK1 inhibitors neutralize the negative effect of TLR8 activation on dendritic growth. (A and D) WT neurons were treated with CL075/poly dT for 30 min at 18 DIV and subjected to immunostaining with phospho-P38, phospho-TAK1, and neuronal marker NeuN antibodies as indicated. Counterstaining with DAPI was performed to label nuclei. (B and E) GFP was transfected at 12 DIV. CL075/poly dT and P38 inhibitor SB203580 (B) and three TAK1 inhibitors, Takinib, 5ZO, and NG25 (E), were added at 17 DIV for 1 d in WT cultured neurons. (C) SB203580 treatment did not influence the effect of CL075 on younger WT neurons at 4 DIV. (B–E) Sample size (n) indicates the number of examined neurons, which were randomly collected blind from two independent experiments. Data are presented as the mean + SEM (error bars). Bars: (A) 25 µm; (B) 50 µm; (D) 20 µm. **, P < 0.01; ***, P < 0.001. One-way (E) and two-way (B and C) ANOVA with Bonferroni’s multiple comparisons test as well as two-tailed nonparametric Mann–Whitney tests (D) were used.
Figure 9. P38 and TAK1 inhibitors neutralize the negative effect of TLR8 activation on dendritic growth. (A and D) WT neurons were treated with CL075/poly dT for 30 min at 18 DIV and subjected to immunostaining with phospho-P38, phospho-TAK1, and neuronal marker NeuN antibodies as indicated. Counterstaining with DAPI was performed to label nuclei. (B and E) GFP was transfected at 12 DIV. CL075/poly dT and P38 inhibitor SB203580 (B) and three TAK1 inhibitors, Takinib, 5ZO, and NG25 (E), were added at 17 DIV for 1 d in WT cultured neurons. (C) SB203580 treatment did not influence the effect of CL075 on younger WT neurons at 4 DIV. (B–E) Sample size (n) indicates the number of examined neurons, which were randomly collected blind from two independent experiments. Data are presented as the mean + SEM (error bars). Bars: (A) 25 µm; (B) 50 µm; (D) 20 µm. **, P < 0.01; ***, P < 0.001. One-way (E) and two-way (B and C) ANOVA with Bonferroni’s multiple comparisons test as well as two-tailed nonparametric Mann–Whitney tests (D) were used.CC BY-NC-SA · PubMed Central
Figure 10. Summary of the TLR3, TLR7, and TLR8 pathways regulating neuronal morphology. TLR3, TLR7, and TLR8 all use MYD88 to control the downstream pathways regulating neuronal morphology, although TRIF is critical for TLR3 to regulate cytokine expression in neurons (Chen et al., 2017). The colored ellipses indicate the major biological processes of target genes downstream of TLR3 (blue), TLR7 (green), and TLR8 (red). Some ellipses overlap because some target genes are shared between different pathways, particularly for inflammatory and innate immune responses downstream of TLR3 and TLR7. For TLR3 and TLR8, only four genes (namely Prex2, Jag1, Cdk6, and Ccnd1) overlap among these top GO processes. The specific agonists and ligands of TLR3, TLR7, and TLR8 are also labeled. The effects of TLR3, TLR7, and TLR8 on neuronal morphology are indicated at bottom. For TLR8, MYD88, TAK1, and P38, MAPKs are required for dendritic pruning. MYD88 and DISC1 are required for TLR3-mediated neuronal morphogenesis (Chen et al., 2017). The biological function of cytokines or innate immune responses induced by the TLR3–TRIF pathway in neurons is unclear. TLR7 uses a traditional pathway involving in MYD88, c-FOS, and IL-6 to negatively regulate dendritic and axonal outgrowth in neurons (Liu et al., 2013).
Figure 10. Summary of the TLR3, TLR7, and TLR8 pathways regulating neuronal morphology. TLR3, TLR7, and TLR8 all use MYD88 to control the downstream pathways regulating neuronal morphology, although TRIF is critical for TLR3 to regulate cytokine expression in neurons (Chen et al., 2017). The colored ellipses indicate the major biological processes of target genes downstream of TLR3 (blue), TLR7 (green), and TLR8 (red). Some ellipses overlap because some target genes are shared between different pathways, particularly for inflammatory and innate immune responses downstream of TLR3 and TLR7. For TLR3 and TLR8, only four genes (namely Prex2, Jag1, Cdk6, and Ccnd1) overlap among these top GO processes. The specific agonists and ligands of TLR3, TLR7, and TLR8 are also labeled. The effects of TLR3, TLR7, and TLR8 on neuronal morphology are indicated at bottom. For TLR8, MYD88, TAK1, and P38, MAPKs are required for dendritic pruning. MYD88 and DISC1 are required for TLR3-mediated neuronal morphogenesis (Chen et al., 2017). The biological function of cytokines or innate immune responses induced by the TLR3–TRIF pathway in neurons is unclear. TLR7 uses a traditional pathway involving in MYD88, c-FOS, and IL-6 to negatively regulate dendritic and axonal outgrowth in neurons (Liu et al., 2013).CC BY-NC-SA · PubMed Central

Verification

Claims the paper makes

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

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

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 68 references

Cited by

Metadata

Authors
Hung, Y.-F., Chen, C.-Y., Shih, Y.-C., Liu, H.-Y., Huang, C.-M., & Hsueh, Y.-P.
Affiliations
Institute of Molecular Biology, Academia Sinica (TW)
Corresponding author
Hsueh, Yi-Ping
Venue
Journal of Cell Biology
Published
2018-05-18
Volume
217
Issue
8
Pages
2727-2742
ISSN
  • 0021-9525 (ISSN-L)
  • 1540-8140
OpenAlex
W2804100187
Cited by
86
References
69
Citations by year
2026: 13 · 2025: 5 · 2024: 5 · 2023: 17 · 2022: 9 · 2021: 16 · 2020: 13 · 2019: 6 · 2018: 2
FWCI
4.16
Type
journal article
Language
en
Open access
Locations
  • The Journal of Cell Biology (publishedVersion, cc-by-nc-sa)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc-sa)
  • PubMed Central (publishedVersion, cc-by-nc-sa)
  • PubMed (publishedVersion)
Funders
  • Academia Sinica
  • Ministry of Science and Technology (MOST 106-2321-B-001-019, 105-2311-B-001-061-MY3)
  • Simons Foundation (SFARI#388449)
Topics
  • Life Sciences › Neuroscience › Neurology › Neuroinflammation and Neurodegeneration Mechanisms
  • Life Sciences › Immunology and Microbiology › Immunology › Immune Response and Inflammation
  • Life Sciences › Neuroscience › Cellular and Molecular Neuroscience › Neuroscience and Neuropharmacology Research
Keywords
Toll-like receptors, TLR3, TLR7, TLR8, neuronal morphology, dendritic pruning, dendritic branching, axon growth, MyD88 signaling, MAPK signaling, neuroinflammation, neuronal development, gene expression profiling
MeSH terms
Animals, Female, Gene Expression Regulation, Male, Membrane Glycoproteins, Mice, Inbred C57BL, Neuronal Plasticity, Neurons, Endosomes, Signal Transduction, Electroporation, Gene Expression Profiling, Cell Enlargement, Toll-Like Receptor 3, Toll-Like Receptor 7, Toll-Like Receptor 8, Mice, Myeloid Differentiation Factor 88

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

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.