Fig 1.
Lsd1 is abundantly expressed in osteoblasts and important for osteoblast differentiation.
(A, B) RNA-seq analysis was performed from MC3T3-E1 at different stages of differentiation. The differentiation was verified by Von Kossa (top) and ALP staining (bottom) (A) and by expression profiles of osteoblast biomarkers in RNA-seq data (n = 2)(B). (C) mRNA expression of KDMs in MC3T3-E1 cells at different stages of differentiation measured by RNA-seq. Average of FPKM normalized read counts from two replicate samples for each time point are shown. (D) Knockdown of Lsd1 mRNA in MC3T3-E1 cells by stable shRNA expression shown by quantitative qPCR. B-actin normalized Lsd1 mRNA expression relative to shScrambled cells from three replicate samples is shown. (E) shLSD1 knockdown cell and control shScrambled were cultured in standard osteogenic conditions for 14 and 21d and fixed in 10% formalin. Differentiated phenotype was demonstrated by ALP and Von Kossa staining and by (F) quantitation of osteogenic genes mRNA expression at 14D differentiated cells by qPCR (n = 3). (G) Western blot showing Lsd1 protein and methylation state of histone H3 in shLSD1 knockdown cells. (E, F) Results from a representative cell culture experiment from three independent experiments is shown. P-values for statistically significant differences are marked * P<0.05, ** P<0.01, *** P<0.001.
Table 1.
More than two-fold differentially expressed histone demethylases in MC3T3-E1 cells differentiated for 21 days compared to undifferentiated subconfluent cells in RNA-seq analysis.
Fig 2.
Lsd1 inhibition in osteoblasts.
Effect of Lsd1 inhibition in MC3T3-E1 cells by RN-1 (A) and GSK-LSD1 (B) on differentiation demonstrated by ALP and Von Kossa staining of 21d differentiated cells (left), expression of osteoblast related genes in 14d differentiated cells quantitated by qPCR (middle, n = 3) and proliferation rate (right, n = 3). Effect of Lsd1 inhibition in calvarial osteoblasts by RN-1 and GSK-LSD1 (C) demonstrated by ALP staining of 21d differentiated cells (left), expression of osteoblast related genes in 14d differentiated cells quantitated by qPCR (middle, n = 3) and von Kossa stained area (right, n = 3). P-values for statistically significant differences are marked * P<0.05, ** P<0.01, *** P<0.001.
Fig 3.
Lsd1 inhibitor treatment in vivo.
(A) Representative images of Von Kossa stained control (top) and GSK-LSD1 treated (bottom) MMA embedded sections. (B) μCT analysis of trabecular bone. (C) μCT analysis of cortical bone. (D) Histomorphometric analysis of trabecular bone structural parameters. (E) Histomorphometric analysis of osteoblast parameters. (F) Representative images of fluorescent labels in control (top) and GSK-LSD1 treated (bottom) samples. (G) Histomorphometric analysis of dynamic parameters, (H) Bone resorption marker CTX-I and bone formation marker P1NP serum levels. Mice were treated with either 1,5mg/kg GSK-LSD1 or saline on a 4-on-3-off schedule for 4 weeks, n = 10/group). P-values for statistically significant differences are marked * P<0.05, ** P<0.01, *** P<0.001.
Fig 4.
Loss of Lsd1 in long bone mesenchymal cells leads to impaired ossification and growth.
(A) Lsd1 knockout mice have decreased body weight and length of tibia. (B) 2D X-ray data shows osteogenesis imperfecta-like phenotype with increased curvature, impaired secondary ossification and non-union of the fibula to the tibia in the Lsd1-cKO mice (right) compared to control (left). (C) Representative 3D models of μCT data shows similar loss of secondary ossification site and unclosed calvarial sutures in the Lsd1-cKO mice (right, bottom) compared to the control mice (left, top). (D) 3D analysis of trabecular and cortical bone. (E) Histomorphometric analysis of trabecular bone. (F) RT-PCR analysis of RNA extracted from marrow-evacuated bone shows efficiency of Lsd1 knockout (~60%), decrease in osteoblast markers Osx, Runx2 and Col1A1, no significant change of osteoblast marker Trap and adipocyte marker PPARgamma, and a significant increase of AdipoQ expression. A-E: n = 5-7/group, F: n = 5/group. P-values for statistically significant differences are marked * P<0.05, ** P<0.01, *** P<0.001.
Fig 5.
Loss of Lsd1 in long bone mesenchymal cells disrupts growth plate organization but not collagen maturation.
(A) Growth plates of Lsd1 knockout mice have decreased organization and thicker hypertrophic zone (right) compared to control (left) (n = 4–5 per group). (B) Mean cell density of Lsd1 knockout mice was lower in both proliferating zone and hypertrophic zone of the growth plate. (n = 4–5 per group). (C) Chondrocyte column height counting showed impaired cellular organization in the proliferating zone of growth plate. (n = 4–5 per group). (D) Osteoclast counting of subepiphyseal bone showed impaired regeneration seen as decreased relative number of osteoclasts (n = 5 per group).
Fig 6.
(A) Number of Lsd1 peaks detected at different time points. (B) Overlap of Lsd1 peaks between different time points represened by Venn diagram. (C) Distrubution of Lsd1 peaks. (D) Distance of Lsd1 peaks from the TSS at different time points. (E) Top enriched TF consensus motifs in Lsd1 ChIP-seq peak data. (F) Overlap of Lsd1 and published Runx2 ChIP-seq peaks.
Fig 7.
(A) Total number of ChIP-seq peaks. (B) Distribution of ChIP-seq peals on annotated genomic locations. (C) Average distance of H3K4me1/me2/m3 and H3K27me3 peaks from the TSS’s. (D) Overlapping peaks of Lsd1 and histone mark binding sites at 14D differentiated cells. Percentage of Lsd1 peaks overlapping with histone modification peaks are indicated.
Table 2.
Overlapping LSD1 and histone modification peaks.
Fig 8.
Examples of H3K4 methylation patterns at Lsd1 bound promoters and enhancers visualized by integrative genomic viewer (IGV).
View of chromosome 19 with several expressed genes.
Fig 9.
Lsd1 binding correlates with expression status and differential expression.
(A) Determination of gene expression status of RNA-seq data. Genes with log2 expression level <1 were defined as expressed. (B) Correlation between Lsd1 binding and gene transcriotion at 0 and 14 d of differentiation. (C) The average expression level of genes with (blue) or without (red) Lsd1 promoter occupancy at 0D and 14D timepoints. (D) Lsd1 binding was significantly enriched on the promoters of DE genes during differentiation. (E) DE genes with Lsd1 occupancy were equally up- and downregulated. (F) Dynamics of Lsd1 binding was not indicative of up- or downregulation of gene transcription.