CoChat

Osteocyte transcriptome mapping identifies a molecular landscape controlling skeletal homeostasis and susceptibility to skeletal disease

Open access

Published version of a preprint

The DOI matches its Crossref record; 133 of the 134 references it cites were found in a registry or an index; 14 of the 48 sources its claims cite were checked for support.

Abstract

Osteocytes are master regulators of the skeleton. We mapped the transcriptome of osteocytes from different skeletal sites, across age and sexes in mice to reveal genes and molecular programs that control this complex cellular-network. We define an osteocyte transcriptome signature of 1239 genes that distinguishes osteocytes from other cells. 77% have no previously known role in the skeleton and are enriched for genes regulating neuronal network formation, suggesting this programme is important in osteocyte communication. We evaluated 19 skeletal parameters in 733 knockout mouse lines and reveal 26 osteocyte transcriptome signature genes that control bone structure and function. We showed osteocyte transcriptome signature genes are enriched for human orthologs that cause monogenic skeletal disorders ( P = 2.4 × 10 −22 ) and are associated with the polygenic diseases osteoporosis ( P = 1.8 × 10 −13 ) and osteoarthritis ( P = 1.6 × 10 −7 ). Thus, we reveal the molecular landscape that regulates osteocyte network formation and function and establish the importance of osteocytes in human skeletal disease.

Overview

CoChat Summary

TL;DR

The paper maps the mouse osteocyte transcriptome across skeletal sites, ages, and sexes and defines a 1,239-gene osteocyte transcriptome signature (OTS) enriched for neuronal/axon-guidance programs. It links OTS genes to skeletal control in knockout mice and to human monogenic skeletal disorders, osteoporosis/eBMD, and osteoarthritis.

Key findings

  • 14,794 genes were actively expressed in osteocytes from tibiae, femora, and humeri; 92% were expressed at all sites and 96% were validated in orthogonal datasets. The OTS comprised 1,239 genes; 77% had no previously known role in the skeleton, and the signature was enriched for genes regulating neuronal network formation/axon guidance.
  • In 733 knockout mouse lines screened, 64 had deletions of OTS genes and 26 (41%) had structural and/or functional skeletal phenotypes; 15 were in genes not annotated with skeletal terms or phenotypes. Examples included Auts2+/− with decreased femoral bone length and cortical diameter, Cttnbp2−/− with increased vertebral BMC and maximum load, Dact3−/− with increased vertebral BMC and yield load, and Ldlrad4−/− with decreased BV/TV and trabecular number and increased trabecular separation.
  • OTS genes were enriched for human orthologs causing monogenic skeletal disorders: 90 OTS genes cause 168 of 612 annotated skeletal disorders (3-fold enrichment, P = 2.4 × 10−22). In the osteogenesis imperfecta/decreased bone density group, 18/33 causal genes were in the signature (P = 7 × 10−13); all 19 known OI genes were actively expressed in osteocytes and 14 were in the OTS.
  • OTS orthologs were associated with eBMD and osteoarthritis: LDSC-SEG showed OTS gene regions explained ~12.5% of eBMD SNP heritability (P = 1.6 × 10−7; 2.2-fold enrichment), competitive gene-set analysis P = 1.8 × 10−13, and 259/992 OTS orthologs had significant gene-level eBMD associations. For OA at any site, P = 1.6 × 10−7, ~11.6% of SNP heritability, and 40/992 OTS orthologs had significant gene-level associations.

Methods

  • Mouse RNA-seq cohorts: bone comparison cohort (16-week-old male tibiae, femora, humeri; n = 8 per bone type), skeletal maturation cohort (4-, 10-, 16-, and 26-week-old male and female humeri; n = 5 per sample type), and osteocyte enrichment cohort (10-week-old male humeri; n = 5 per condition). Osteocyte-enriched bone samples were >80% osteocytes; analyses included de novo transcriptome assembly, activity thresholds, Gaussian mixture modeling, WGCNA, and orthogonal dataset validation.
  • Functional skeletal phenotyping used the Origins of Bone and Cartilage Disease (OBCD)/IMPC resource: 733 knockout mouse lines, 19 structural and functional skeletal parameters, DXA, micro-CT, and biomechanical testing; novel CRISPR-generated knockout lines were also phenotyped.
  • Human genetic analyses used UK Biobank data for eBMD (362,924 individuals) and OA summary statistics (77,052 cases and 378,169 controls), with LDSC-SEG, MAGMA competitive gene-set analysis, and nearest-gene enrichment analysis.

Limitations

  • The paper states: only long bones were studied, not spine or calvaria; cancellous bone was only 1–2% of sampled tissue, so the transcriptome mainly represents cortical osteocytes; contaminating cells cannot be fully excluded; and the study measured transcriptome, not protein expression.
  • Evident from the text: skeletal phenotypes come from mouse knockouts, and human disease links are genetic associations rather than direct causal demonstrations; temporal osteocyte differentiation clusters rely on the in vitro IDG-SW3 cell-line model.

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

Integrity signals

Registry

Crossref · 10.1038/s41467-021-22517-1

Cites

133 of 133 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 The osteocyte transcriptome is broadly conserved across the skeleton. a Schematic of the overall study design used to define the osteocyte transcriptome, to define an osteocyte transcriptome signature, the profile of genes enriched in osteocytes, and to identify their role in skeletal disease. b Diagram illustrating the 3 bone-types in which the osteocyte transcriptome was sequenced and compared. c Representative micro-CT and histological images of bone samples prior to processing (i–ii) and following osteocyte isolation (iii–iv), illustrating the effective removal of bone marrow, muscle and growth plates and enrichment of osteocytes (iv) (n = 8 biological independent mice). Boxes and dotted lines identify areas of high magnification confirming osteocyte isolation. Bone (B), bone marrow (BM), muscle (M) and growth plate (GP) are identified. d Actively expressed genes were identified based on the normalised gene expression distribution (blue histogram) and a hypothetical distribution of active promoters (red dotted line)27 (i). Vertical line denotes the sample-specific threshold of actively expressed genes. Genes were considered active if they were above sample-specific threshold in all replicates, inactive if they were below the threshold in all replicates, or variable if they were above the threshold in at least one but not all replicates (ii). e The number of active genes and biotype composition in the osteocyte transcriptome across skeletal sites (i), and the percentage of total active genes detected in osteocytes in at least one (1+) and two or more (2+) orthogonal datasets (ii). The Any bone column reflects the number of genes actively expressed at any skeletal site. The All bones column reflects the number and proportion (as a percentage) of genes that were actively expressed in all skeletal sites. TEC = To be Experimentally Confirmed. f Pearson correlation of gene expression between individual replicates and bone types (mean of each bone type comparison represented by numbers above heatmap). Tib = tibia, Fem = femur, Hum = humerus, Rep = biological replicate. g Genes that were differentially expressed between osteocytes isolated from different bone types. Homeobox genes and antisense-RNAs are in red (FDR ≤ 0.05, LFC > 0.5). h Methodology for identifying genes expressed in a site-specific manner and a Venn diagram identifying the genes expressed in one bone type but not others.
Fig. 1 The osteocyte transcriptome is broadly conserved across the skeleton. a Schematic of the overall study design used to define the osteocyte transcriptome, to define an osteocyte transcriptome signature, the profile of genes enriched in osteocytes, and to identify their role in skeletal disease. b Diagram illustrating the 3 bone-types in which the osteocyte transcriptome was sequenced and compared. c Representative micro-CT and histological images of bone samples prior to processing (i–ii) and following osteocyte isolation (iii–iv), illustrating the effective removal of bone marrow, muscle and growth plates and enrichment of osteocytes (iv) (n = 8 biological independent mice). Boxes and dotted lines identify areas of high magnification confirming osteocyte isolation. Bone (B), bone marrow (BM), muscle (M) and growth plate (GP) are identified. d Actively expressed genes were identified based on the normalised gene expression distribution (blue histogram) and a hypothetical distribution of active promoters (red dotted line)27 (i). Vertical line denotes the sample-specific threshold of actively expressed genes. Genes were considered active if they were above sample-specific threshold in all replicates, inactive if they were below the threshold in all replicates, or variable if they were above the threshold in at least one but not all replicates (ii). e The number of active genes and biotype composition in the osteocyte transcriptome across skeletal sites (i), and the percentage of total active genes detected in osteocytes in at least one (1+) and two or more (2+) orthogonal datasets (ii). The Any bone column reflects the number of genes actively expressed at any skeletal site. The All bones column reflects the number and proportion (as a percentage) of genes that were actively expressed in all skeletal sites. TEC = To be Experimentally Confirmed. f Pearson correlation of gene expression between individual replicates and bone types (mean of each bone type comparison represented by numbers above heatmap). Tib = tibia, Fem = femur, Hum = humerus, Rep = biological replicate. g Genes that were differentially expressed between osteocytes isolated from different bone types. Homeobox genes and antisense-RNAs are in red (FDR ≤ 0.05, LFC > 0.5). h Methodology for identifying genes expressed in a site-specific manner and a Venn diagram identifying the genes expressed in one bone type but not others.CC BY · PubMed Central
Fig. 2 The osteocyte transcriptome changes with sex and age. a Experimental design used to define the osteocyte transcriptome in each sex during skeletal maturation. b Representative micro-CT images of humeri from female and male mice at different ages. c The number of active genes and biotype composition of the osteocyte transcriptome at different ages in each sex. Any reflects the number of genes actively expressed in any ages or sexes, All reflects the percentage of genes active in any age or sex that are active in all ages and sexes. TEC = to be experimentally confirmed. d Principal component (PC) analysis of samples clustered by age in female (i) and male (ii) mice. The percentage of total variance explained by individual PCs is shown. Dots represent individual biological replicates and ellipses represent 50% confidence intervals for each age. Dist = Euclidean distance between cluster centroids (mean). P-values were calculated using Hotelling’s t-test (one-sided, Bonferroni adjusted), clusters with significant separation (P < 0.05) between centroids are highlighted in green. e Principal component analysis of samples clustered from both sexes at 4 (i), 10 (ii), 16 (iii) and 26 (iv) weeks of age. The percentage of total variance explained by individual PCs is shown. Dots represent individual biological replicates and ellipses represent 50% confidence intervals for each sex. P-values were calculated using Hotelling’s t-test (one-sided, Bonferroni adjusted), clusters with significant separation (P < 0.05) between centroids are highlighted in green.
Fig. 2 The osteocyte transcriptome changes with sex and age. a Experimental design used to define the osteocyte transcriptome in each sex during skeletal maturation. b Representative micro-CT images of humeri from female and male mice at different ages. c The number of active genes and biotype composition of the osteocyte transcriptome at different ages in each sex. Any reflects the number of genes actively expressed in any ages or sexes, All reflects the percentage of genes active in any age or sex that are active in all ages and sexes. TEC = to be experimentally confirmed. d Principal component (PC) analysis of samples clustered by age in female (i) and male (ii) mice. The percentage of total variance explained by individual PCs is shown. Dots represent individual biological replicates and ellipses represent 50% confidence intervals for each age. Dist = Euclidean distance between cluster centroids (mean). P-values were calculated using Hotelling’s t-test (one-sided, Bonferroni adjusted), clusters with significant separation (P < 0.05) between centroids are highlighted in green. e Principal component analysis of samples clustered from both sexes at 4 (i), 10 (ii), 16 (iii) and 26 (iv) weeks of age. The percentage of total variance explained by individual PCs is shown. Dots represent individual biological replicates and ellipses represent 50% confidence intervals for each sex. P-values were calculated using Hotelling’s t-test (one-sided, Bonferroni adjusted), clusters with significant separation (P < 0.05) between centroids are highlighted in green.CC BY · PubMed Central
Fig. 3 Defining an osteocyte transcriptome signature. a Experimental design used to identify genes with enriched expression in osteocytes. Diagram illustrating the strategy used to identify genes enriched for expression in osteocytes in bone samples enriched with osteocytes relative to intact bone samples in which bone marrow and cells lining bone were retained (i). Experimental workflow used to identify genes enriched for expression in osteocytes (ii). b Gene-enrichment in osteocyte-enriched bone samples relative to whole-bone samples distinguished genes known to be expressed in osteocytes (blue) from housekeeping genes (green) and genes expressed in marrow cell populations (yellow) (i). Enrichment for each individual gene is shown as a dot and plotted as a function of normalised gene expression (log2-FPKM). The four component gaussian mixture model fitted to the density distribution of gene-enrichment in (i) used to define the osteocyte enrichment threshold (ii). Each component is denoted by a separate colour. The top two Gene Ontology biological processes associated with genes in each component are illustrated in coloured boxes with p-values (Bonferroni adjusted) calculated by hypergeometric test. (ii). 1777 genes above the osteocyte-enrichment threshold (indicated by the red dashed line) were significantly enriched for expression in osteocytes (iii). LFC = log2 fold-change. c Volcano plots comparing the expression of osteocyte-enriched genes (from b) with blood (i), bone marrow (ii) and muscle (iii)13. Dashed lines represent p < 0.05 cutoff. Genes significantly enriched in these tissues relative to osteocytes are identified as red dots. P-values were calculated in the original publishing article13 by F-test with Benjamini–Hochberg adjustment. d Filtering pipeline used to define the 1239-gene osteocyte transcriptome signature (OTS). The number of genes remaining at each filtering stage is indicated. e Expression specificity29 (Tau) of OTS genes relative to other organs and tissues28. Genes with Tau <0.15 = low expression specificity in osteocytes (green), 0.15 ≤ Tau ≤ 0.85 = moderate expression specificity (orange), while Tau > 0.85 = high expression specificity (red). f OTS genes were enriched for expression in osteocytes relative to osteoblasts and bone-lining cells isolated by laser-capture microdissection. Tukey boxplots show a summary of median OTS gene expression values in each cell type calculated across n = 3 biological replicates from a single experiment. Boxes indicate median and interquartile range (IQR) of scaled, normalised gene expression values, whiskers denote values ±1.5 × IQR and outlier values beyond this range are shown as individual points. P-values (two-tailed, Benjamini and Hochberg adjusted) were calculated by competitive gene set test accounting for inter-gene correlation (CAMERA)105.
Fig. 3 Defining an osteocyte transcriptome signature. a Experimental design used to identify genes with enriched expression in osteocytes. Diagram illustrating the strategy used to identify genes enriched for expression in osteocytes in bone samples enriched with osteocytes relative to intact bone samples in which bone marrow and cells lining bone were retained (i). Experimental workflow used to identify genes enriched for expression in osteocytes (ii). b Gene-enrichment in osteocyte-enriched bone samples relative to whole-bone samples distinguished genes known to be expressed in osteocytes (blue) from housekeeping genes (green) and genes expressed in marrow cell populations (yellow) (i). Enrichment for each individual gene is shown as a dot and plotted as a function of normalised gene expression (log2-FPKM). The four component gaussian mixture model fitted to the density distribution of gene-enrichment in (i) used to define the osteocyte enrichment threshold (ii). Each component is denoted by a separate colour. The top two Gene Ontology biological processes associated with genes in each component are illustrated in coloured boxes with p-values (Bonferroni adjusted) calculated by hypergeometric test. (ii). 1777 genes above the osteocyte-enrichment threshold (indicated by the red dashed line) were significantly enriched for expression in osteocytes (iii). LFC = log2 fold-change. c Volcano plots comparing the expression of osteocyte-enriched genes (from b) with blood (i), bone marrow (ii) and muscle (iii)13. Dashed lines represent p < 0.05 cutoff. Genes significantly enriched in these tissues relative to osteocytes are identified as red dots. P-values were calculated in the original publishing article13 by F-test with Benjamini–Hochberg adjustment. d Filtering pipeline used to define the 1239-gene osteocyte transcriptome signature (OTS). The number of genes remaining at each filtering stage is indicated. e Expression specificity29 (Tau) of OTS genes relative to other organs and tissues28. Genes with Tau <0.15 = low expression specificity in osteocytes (green), 0.15 ≤ Tau ≤ 0.85 = moderate expression specificity (orange), while Tau > 0.85 = high expression specificity (red). f OTS genes were enriched for expression in osteocytes relative to osteoblasts and bone-lining cells isolated by laser-capture microdissection. Tukey boxplots show a summary of median OTS gene expression values in each cell type calculated across n = 3 biological replicates from a single experiment. Boxes indicate median and interquartile range (IQR) of scaled, normalised gene expression values, whiskers denote values ±1.5 × IQR and outlier values beyond this range are shown as individual points. P-values (two-tailed, Benjamini and Hochberg adjusted) were calculated by competitive gene set test accounting for inter-gene correlation (CAMERA)105.CC BY · PubMed Central
Fig. 4 Genes and control pathways identified by the osteocyte transcriptome signature. a The top 26 osteocyte transcriptome signature genes most enriched in osteocytes (outer ring). Genes annotated with either a skeletal biological process (GO database) or skeletal phenotype (MGI database) are highlighted blue; genes reported with experimental evidence of a role in the skeleton are brown, whereas unannotated genes without a skeletal annotation in GO or MGI and no experimental evidence of a role in the skeleton reported in the literature are shown in grey. Numbers denote log2 fold-enrichment (LFC) in gene expression in osteocytes. The proportion of all osteocyte transcriptome signature genes annotated (blue) or unannotated (grey) with a skeletal annotation in GO or MGI is shown in the inner ring. b Clusters (n = 8) of semantically similar biological processes (GO) significantly enriched in the osteocyte transcriptome signature are identified by different colours (Bonferroni-corrected P < 0.05). The top 2 terms (by P-value) in each of the eight clusters are listed. c Heatmap showing clustering of osteocyte transcriptome signature genes based on distinct co-expression patterns during osteocytic-differentiation50 of the IDG-SW3 cell-line, from osteoblast-like cells (day 3) to early (day 14) and mature osteocytes (day 35). The name and number of genes for each cluster are labelled. Cluster colours correspond to subsequent panels (d–e). d Changes in early expression cluster (i), early activation cluster (ii) and maturation cluster (iii) of genes during osteocyte differentiation. Tukey boxplots show the distribution of gene expression values for each cluster in cells from three independent experiments (n = 3). Boxes indicate median and interquartile range (IQR) of scaled, normalised gene expression values, whiskers denote values ±1.5 × QR and outlier values beyond this range are shown as individual points. e Biological processes (BP) enriched in the early expression cluster (i), the early activation cluster (ii) and the maturation cluster (iii). The top 5 GO BP terms in each cluster with P-values (Bonferroni adjusted) calculated by hypergeometric test.
Fig. 4 Genes and control pathways identified by the osteocyte transcriptome signature. a The top 26 osteocyte transcriptome signature genes most enriched in osteocytes (outer ring). Genes annotated with either a skeletal biological process (GO database) or skeletal phenotype (MGI database) are highlighted blue; genes reported with experimental evidence of a role in the skeleton are brown, whereas unannotated genes without a skeletal annotation in GO or MGI and no experimental evidence of a role in the skeleton reported in the literature are shown in grey. Numbers denote log2 fold-enrichment (LFC) in gene expression in osteocytes. The proportion of all osteocyte transcriptome signature genes annotated (blue) or unannotated (grey) with a skeletal annotation in GO or MGI is shown in the inner ring. b Clusters (n = 8) of semantically similar biological processes (GO) significantly enriched in the osteocyte transcriptome signature are identified by different colours (Bonferroni-corrected P < 0.05). The top 2 terms (by P-value) in each of the eight clusters are listed. c Heatmap showing clustering of osteocyte transcriptome signature genes based on distinct co-expression patterns during osteocytic-differentiation50 of the IDG-SW3 cell-line, from osteoblast-like cells (day 3) to early (day 14) and mature osteocytes (day 35). The name and number of genes for each cluster are labelled. Cluster colours correspond to subsequent panels (d–e). d Changes in early expression cluster (i), early activation cluster (ii) and maturation cluster (iii) of genes during osteocyte differentiation. Tukey boxplots show the distribution of gene expression values for each cluster in cells from three independent experiments (n = 3). Boxes indicate median and interquartile range (IQR) of scaled, normalised gene expression values, whiskers denote values ±1.5 × QR and outlier values beyond this range are shown as individual points. e Biological processes (BP) enriched in the early expression cluster (i), the early activation cluster (ii) and the maturation cluster (iii). The top 5 GO BP terms in each cluster with P-values (Bonferroni adjusted) calculated by hypergeometric test.CC BY · PubMed Central
Fig. 5 Deletion of osteocyte transcriptome signature genes affects bone structure and function. a The 26 osteocyte transcriptome signature genes with outlier skeletal phenotypes in the Origins of Bone and Cartilage Disease (OBCD) database of single gene deletions in mice (outer ring). Deletion of genes in blue affected bone structure, genes in brown affected bone function and genes in red both structure and function. Those annotated in either GO or MGI are indicated. Genes in red have a significant outlier phenotype when deleted in knockout mice but are not annotated with skeletal function or phenotype in either GO or MGI databases. The inner ring denotes the number of genes with (green) and without (grey) an outlier skeletal phenotype when deleted in mice in the OBCD database. b Representative quantitative X-ray microradiographic images from the femurs (i) and vertebrae (ii) of adult, female wild-type (WT), Auts2+/−, Cttnbp2−/−, Dact3−/− and Ldlrad4−/− mice (n = 6 mice per genotype). Scale bar = 1 mm. Dot plots illustrate individual parameters. c Representative micro-CT images of trabecular (i) and cortical (ii) bone of adult, female wild-type (WT), Auts2+/−, Cttnbp2−/−, Dact3−/− and Ldlrad4−/− mice (n = 6 mice per genotype). Scale bar = 100 µm. Dot plots illustrate bone volume as a proportion of tissue volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), cortical thickness (Ct.Th), internal endosteal diameter and bone mineral density (BMD). d Representative micro-CT images showing large osteocyte lacunae (401–800 µm3) in tibia cortical bone from adult, female WT (n = 11), Auts2+/− (n = 6), Cttnbp2−/− (n = 6), Dact3−/− (n = 6) and Ldlrad4−/− (n = 4) mice. Scale bar = 100 μm. Graphs show osteocyte lacunae number per bone volume (Lc.N/BV) and distribution of lacunae volumes (Lc.V) in WT mice (median and 95% confidence intervals). Violin plot shows distribution of Lc.V in the four knockout mouse lines compared to WT and graphs show relative frequency distribution of Lc.V in Cttnbp2−/− and Ldlrad4−/− mice compared to WT. Kruskall–Wallis test followed by Dunn’s multiple comparison adjustment (two-sided) **P < 0.01, ****P < 0.0001. e Load displacement curves from caudal vertebrae compression testing in adult female WT, Auts2+/−, Cttnbp2−/−, Dact3−/− and Ldlrad4−/− mice. Dot plots show yield load, maximum load and stiffness. For each variable in b, c and e the mean (solid centre lines), ±1.0 SD (dotted lines) and ±2.0 SD. (grey boxes) for WT mice (n = 320) are shown. Individual data-points for each parameter in Auts2+/-, Cttnbp2−/−, Dact3−/− and Ldlrad4−/− lines are shown as green, pink, blue and orange dots, respectively (n = 6 biologically independent mice per genotype). The mean value for each individual phenotype parameter is indicated by a solid bold line. Mean parameters that lie greater or less than two standard deviations from the WT reference mean are considered outliers and are indicated by an asterisk (*), coloured according to the individual mouse line.
Fig. 5 Deletion of osteocyte transcriptome signature genes affects bone structure and function. a The 26 osteocyte transcriptome signature genes with outlier skeletal phenotypes in the Origins of Bone and Cartilage Disease (OBCD) database of single gene deletions in mice (outer ring). Deletion of genes in blue affected bone structure, genes in brown affected bone function and genes in red both structure and function. Those annotated in either GO or MGI are indicated. Genes in red have a significant outlier phenotype when deleted in knockout mice but are not annotated with skeletal function or phenotype in either GO or MGI databases. The inner ring denotes the number of genes with (green) and without (grey) an outlier skeletal phenotype when deleted in mice in the OBCD database. b Representative quantitative X-ray microradiographic images from the femurs (i) and vertebrae (ii) of adult, female wild-type (WT), Auts2+/−, Cttnbp2−/−, Dact3−/− and Ldlrad4−/− mice (n = 6 mice per genotype). Scale bar = 1 mm. Dot plots illustrate individual parameters. c Representative micro-CT images of trabecular (i) and cortical (ii) bone of adult, female wild-type (WT), Auts2+/−, Cttnbp2−/−, Dact3−/− and Ldlrad4−/− mice (n = 6 mice per genotype). Scale bar = 100 µm. Dot plots illustrate bone volume as a proportion of tissue volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), cortical thickness (Ct.Th), internal endosteal diameter and bone mineral density (BMD). d Representative micro-CT images showing large osteocyte lacunae (401–800 µm3) in tibia cortical bone from adult, female WT (n = 11), Auts2+/− (n = 6), Cttnbp2−/− (n = 6), Dact3−/− (n = 6) and Ldlrad4−/− (n = 4) mice. Scale bar = 100 μm. Graphs show osteocyte lacunae number per bone volume (Lc.N/BV) and distribution of lacunae volumes (Lc.V) in WT mice (median and 95% confidence intervals). Violin plot shows distribution of Lc.V in the four knockout mouse lines compared to WT and graphs show relative frequency distribution of Lc.V in Cttnbp2−/− and Ldlrad4−/− mice compared to WT. Kruskall–Wallis test followed by Dunn’s multiple comparison adjustment (two-sided) **P < 0.01, ****P < 0.0001. e Load displacement curves from caudal vertebrae compression testing in adult female WT, Auts2+/−, Cttnbp2−/−, Dact3−/− and Ldlrad4−/− mice. Dot plots show yield load, maximum load and stiffness. For each variable in b, c and e the mean (solid centre lines), ±1.0 SD (dotted lines) and ±2.0 SD. (grey boxes) for WT mice (n = 320) are shown. Individual data-points for each parameter in Auts2+/-, Cttnbp2−/−, Dact3−/− and Ldlrad4−/− lines are shown as green, pink, blue and orange dots, respectively (n = 6 biologically independent mice per genotype). The mean value for each individual phenotype parameter is indicated by a solid bold line. Mean parameters that lie greater or less than two standard deviations from the WT reference mean are considered outliers and are indicated by an asterisk (*), coloured according to the individual mouse line.CC BY · PubMed Central
Fig. 6 The osteocyte transcriptome signature is enriched with genes associated with rare genetic skeletal disorders. a Proportion of genes (%) known to cause genetic skeletal disorders21 that have human orthologs either actively expressed in osteocytes (blue), present in the osteocyte transcriptome signature (OTS) (green) or not active in osteocytes (grey). P-value and fold-enrichment (FE) of osteocyte transcriptome signature genes among disease-causing genes are calculated under the hypergeometric distribution. b Proportion of genes (%) in individual skeletal disease groups that have orthologs actively expressed in osteocytes (blue bars) or present in the osteocyte transcriptome signature (circles). Group size = number of genes in each disease group. Groups with genes significantly enriched in the osteocyte transcriptome signature are coloured according to P-value (Bonferroni adjusted), calculated by hypergeometric test. Only groups with ≥3 genes are represented. N.S. not significant. c Osteocyte-enrichment of genes associated with osteogenesis imperfecta (OI). Genes are coloured according to their identification as being actively expressed in osteocytes (blue) or present in the osteocyte transcriptome signature (green). Bonferroni-corrected P-value and fold-enrichment (FE) of osteocyte transcriptome signature genes among OI-causing genes are calculated under the hypergeometric distribution.
Fig. 6 The osteocyte transcriptome signature is enriched with genes associated with rare genetic skeletal disorders. a Proportion of genes (%) known to cause genetic skeletal disorders21 that have human orthologs either actively expressed in osteocytes (blue), present in the osteocyte transcriptome signature (OTS) (green) or not active in osteocytes (grey). P-value and fold-enrichment (FE) of osteocyte transcriptome signature genes among disease-causing genes are calculated under the hypergeometric distribution. b Proportion of genes (%) in individual skeletal disease groups that have orthologs actively expressed in osteocytes (blue bars) or present in the osteocyte transcriptome signature (circles). Group size = number of genes in each disease group. Groups with genes significantly enriched in the osteocyte transcriptome signature are coloured according to P-value (Bonferroni adjusted), calculated by hypergeometric test. Only groups with ≥3 genes are represented. N.S. not significant. c Osteocyte-enrichment of genes associated with osteogenesis imperfecta (OI). Genes are coloured according to their identification as being actively expressed in osteocytes (blue) or present in the osteocyte transcriptome signature (green). Bonferroni-corrected P-value and fold-enrichment (FE) of osteocyte transcriptome signature genes among OI-causing genes are calculated under the hypergeometric distribution.CC BY · PubMed Central
Fig. 7 Osteocyte transcriptome signature genes are associated with eBMD and osteoarthritis in humans. a The top 100 osteocyte transcriptome signature orthologs with significant gene-wise associations (GWS OTS ortholog) with eBMD25 (ranked by PJOINT < 2 × 10−6). The genome coordinates of each human ortholog are shown. b, The 40 osteocyte transcriptome signature orthologs with significant gene-wise associations (GWS OTS ortholog) with OA (any subcategory)64 (PJOINT < 2 × 10−6). Blue bars show the genome coordinates of significant, conditionally independent GWAS variants associated with eBMD (a) or OA (b) (GWAS lead variants). Genes with significant gene-wise associations and that are located nearest to these loci are shown in red, genes that are not nearest to these loci are shown in black. Genes annotated with a role in the skeleton (either GO biological processes database or skeletal phenotype in the MGI database) are denoted with *, genes in the Origins of Bone and Cartilage Disease (OBCD) database are denoted with $ and genes that cause skeletal genetic disorder when mutated in humans denoted with #.
Fig. 7 Osteocyte transcriptome signature genes are associated with eBMD and osteoarthritis in humans. a The top 100 osteocyte transcriptome signature orthologs with significant gene-wise associations (GWS OTS ortholog) with eBMD25 (ranked by PJOINT < 2 × 10−6). The genome coordinates of each human ortholog are shown. b, The 40 osteocyte transcriptome signature orthologs with significant gene-wise associations (GWS OTS ortholog) with OA (any subcategory)64 (PJOINT < 2 × 10−6). Blue bars show the genome coordinates of significant, conditionally independent GWAS variants associated with eBMD (a) or OA (b) (GWAS lead variants). Genes with significant gene-wise associations and that are located nearest to these loci are shown in red, genes that are not nearest to these loci are shown in black. Genes annotated with a role in the skeleton (either GO biological processes database or skeletal phenotype in the MGI database) are denoted with *, genes in the Origins of Bone and Cartilage Disease (OBCD) database are denoted with $ and genes that cause skeletal genetic disorder when mutated in humans denoted with #.CC BY · PubMed Central

Verification

Claims the paper makes

8 Valid · 0 Not valid · 26 Not checked · 163 Own results and interpretations

117 more claims · 1 with their cited sources checked

Create a free account to see all 197 claims and to verify the 34 references still open.

See all claims — free account
  • It may state two separate claims as one. 47
  • Probably not a substantial claim. 33
  • The quote only partly states this claim. 2

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?

Self-check

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

Each contradiction between two passages listed here was confirmed by an independent check.

3 of 3

Citations

Cites

Show all 133 references

Research works printed but not linked: 1

  1. Warnes, G. R. et al. gplots: various R programming tools for plotting data. R. Package version 2, 1 (2009).

Cited by

See all 133

Metadata

Authors
Youlten, S. E., Kemp, J. P., Logan, J. G., Ghirardello, E. J., Sergio, C. M., Dack, M. R. G., Guilfoyle, S. E., Leitch, V. D., Butterfield, N. C., Komla-Ebri, D., Chai, R. C., Corr, A. P., Smith, J. T., Mohanty, S. T., Morris, J. A., McDonald, M. M., Quinn, J. M. W., McGlade, A. R., Bartonicek, N., Jansson, M., Hatzikotoulas, K., Irving, M. D., Beleza-Meireles, A., Rivadeneira, F., Duncan, E., Richards, J. B., Adams, D. J., Lelliott, C. J., Brink, R., Phan, T. G., Eisman, J. A., Evans, D. M., Zeggini, E., Baldock, P. A., Bassett, J. H. D., Williams, G. R., & Croucher, P. I.
Affiliations
  • Garvan Institute of Medical Research (AU)
  • The University of Queensland (AU)
  • Imperial College London (GB)
  • New York Genome Center (US)
  • Guy's Hospital (GB)
  • Wellcome Sanger Institute (GB)
  • Erasmus MC (NL)
  • Queensland University of Technology (AU)
  • King's College London (GB)
Corresponding author
  • Bassett, J. H. Duncan
  • Williams, Graham R.
  • Croucher, Peter I.
Venue
Nature Communications
Published
2021-05-05
Volume
12
Issue
1
Pages
2444-2444
ISSN
2041-1723 (ISSN-L)
OpenAlex
W3157344667
Cited by
133
References
134
Citations by year
2026: 27 · 2025: 24 · 2024: 21 · 2023: 28 · 2022: 13 · 2021: 14
FWCI
5.93
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)
  • QUT ePrints (Queensland University of Technology) (submittedVersion, cc-by)
  • ResearchOnline@ND (The University of Notre Dame) (submittedVersion)
  • EUR Research Repository (Erasmus University Rotterdam) (submittedVersion, other-oa)
  • RePub (Erasmus University Rotterdam) (submittedVersion, other-oa)
  • RePEc: Research Papers in Economics (submittedVersion)
  • RMIT Research Repository (RMIT University Library) (submittedVersion, other-oa)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Figshare (submittedVersion, cc-by)
  • Research Portal (King's College London) (submittedVersion, other-oa)
  • Site cant be reached (submittedVersion, cc-by)
  • PuSH - Publication Server of Helmholtz Zentrum München (submittedVersion, cc-by)
  • Spiral (Imperial College London) (submittedVersion, cc-by)
  • UNSWorks (University of New South Wales, Sydney, Australia) (submittedVersion, cc-by-nc-nd)
  • Apollo (University of Cambridge) (submittedVersion, cc-by)
  • Apollo (cc-by)
Funders
  • Wellcome Trust (101123/Z/13/A)
  • Mrs Janice Gibson and the Ernest Heine Family Foundation
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bone Metabolism and Diseases
  • Health Sciences › Medicine › Orthopedics and Sports Medicine › Bone health and osteoporosis research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Biochemistry › Bone and Dental Protein Studies
Keywords
osteocytes, transcriptome, skeletal homeostasis, neuronal network formation, osteoporosis, osteoarthritis, bone structure, sex differences, aging
MeSH terms
Age Factors, Animals, Bone and Bones, Bone Diseases, Female, Homeostasis, Humans, Male, Osteocytes, Osteoporosis, Sex Factors, Sequence Analysis, RNA, Mice, Knockout, Computational Biology, Mice, Transcriptome

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 2026-10-04, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-04 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).

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: Osteocyte transcriptome mapping identifies a molecular landscape controlling skeletal homeostasis and susceptibility to skeletal disease. https://app.cochat.ai/papers/10.1038/s41467-021-22517-1
The paper
Youlten, S. E., Kemp, J. P., Logan, J. G., Ghirardello, E. J., Sergio, C. M., Dack, M. R. G., Guilfoyle, S. E., Leitch, V. D., Butterfield, N. C., Komla-Ebri, D., Chai, R. C., Corr, A. P., Smith, J. T., Mohanty, S. T., Morris, J. A., McDonald, M. M., Quinn, J. M. W., McGlade, A. R., Bartonicek, N., . . . Croucher, P. I. (2021). Osteocyte transcriptome mapping identifies a molecular landscape controlling skeletal homeostasis and susceptibility to skeletal disease. Nature Communications. https://doi.org/10.1038/s41467-021-22517-1

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.