Fig 2: BRD9 deletion and inhibition in BMDMs leads to overactivated osteoclastogenesis in vitro.a TRAP staining and quantification analysis of BMDMs from 4-week-old LysM-Cre;Brd9fl/fl mice and littermate control mice after 3 days of RANKL-induction. Scale bar, 200 µm. n = 8. b The mRNA expression of osteoclastic-specific genes of Acp5, Ctsk, and Mmp9 in BMDMs from LysM-Cre;Brd9fl/fl mice compared with that from control littermates after 3 days of RANKL-induction, as measured by qPCR. n = 3 for LysM-Cre;Brd9fl/fl mice group. n = 6 for control littermates. c The protein expression of MMP9 and CTSK in BMDMs from LysM-Cre;Brd9fl/fl mice compared with that from control littermates after 3 days of RANKL-induction, as measured by western blot. d Cell viability of RANKL-induced BMDMs with 1 day of iBRD9 at different concentration, shown by cell counting kit 8 assay. n = 10 biologically independent samples. e The mRNA expression of Acp5 and Mmp9 in RANKL-induced BMDMs with 1 day of iBRD9 at different concentration, as measured by qPCR. n = 5 biologically independent samples for Acp5. n = 3 biologically independent samples for Mmp9. f TRAP staining and quantification analysis of RANKL-induced BMDMs with 5 days of iBRD9 at different concentration. Scale bar, 200 µm. n = 5 biologically independent samples. g The protein expression of FOS and CTSK in BMDMs after 3 days of 1 µM iBRD9 or vector treatment with or without RANKL-induction, as measured by western blot. M-CSF, M; RANKL, R. h The mRNA expression of Acp5 and Mmp9 in BMDMs after 3 days of 1 µM iBRD9 or vector treatment with or without RANKL-induction, as measured by qPCR. M-CSF, M; RANKL, R. n = 3 biologically independent samples. All data in this figure are represented as mean ± SD. Two-tailed Student’s t-test for (a) and (b). One-way analysis of variance (ANOVA) with Dunnett’s multiple comparisons test for (d), (e), and (f); ANOVA with Tukey’s multiple comparisons test for (h). All experiments were performed in triplicates unless otherwise stated. Source data are provided in the Source data file.

Fig 3: BRD9-mediated transcriptional activation of interferon-beta signaling negative regulates osteoclastogenesis.a Heatmap hierarchical clustering and the MA plot showing the gene expression profiles in BMDMs after 1 day of BRD9 inhibition during osteoclastic differentiation with M-CSF and RANKL (MR). n = 4. Color scale represents normalized gene FPKM value by z-score scheme. b The top 10 Gene ontology (GO) biological process enriched in the comparison between BMDMs in MR + vector and MR + iBRD9 group. c Gene set enrichment analysis (GSEA) analysis of the top downregulated gene sets in MR + iBRD9 group compared with control group. d GSEA analysis of the top upregulated gene sets in MR + iBRD9 group compared with control group. e The protein expression of IFN-ß, STAT2, and STAT1 in BMDMs at different days after osteoclastic differentiation, as measured by western blot. f GSEA plots and heatmap hierarchical clustering of GO term_0035458 in MR + iBRD9 group and control group. n = 4. g The mRNA expression of Acp5 and Mmp9 in BMDMs treated with 2 days of 1 µM iBRD9 and 0.0625 ng/ml IFN-ß1 during osteoclastic induction, as measured by qPCR. n = 3 biologically independent samples. h TRAP staining of BMDMs treated with 3 days of 1 µM iBRD9 and 0.0625 ng/ml IFN-ß1 during osteoclastic induction. Scale bar, 200 µm. All data in this figure are represented as mean ± SD. Hypergeometric distribution test for (b). Empirical phenotype-based permutation test for (c) and (d). One-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test for (g). All experiments were performed in triplicates unless otherwise stated. Source data are provided in the Source data file.

Fig 4: BRD9 deletion/degrader mitigates LPS-induced bone resorption.a Schematic illustrating the protocol for establishing LPS-induced localized aggressive periodontitis (LAP). b µCT imaging and d quantification of LPS-induced local alveolar bone loss in the maxillary bone from 4-week-old LysM-Cre;Brd9fl/fl mice (n = 4) and littermate control mice (n = 9). Yellow line in b indicates cementum-enamel junction to alveolar bone crest (CEJ-ABC) distance (mm) in d. c H&E staining of LPS-stimulated maxillary bone from 4-week-old LysM-Cre;Brd9fl/fl mice and littermate control mice. Arrows indicate distance between gingival sulcus bottom to alveolar bone crest. e TRAP staining and f quantification of LPS-stimulated maxillary bone from 4-week-old LysM-Cre;Brd9fl/fl mice (n = 3) and littermate control mice (n = 4). g TNF-a, iNOS and RANKL immunofluorescence (green) in LPS-stimulated maxillary bone from 4-week-old LysM-Cre;Brd9fl/fl mice and littermate control mice. h Schematic illustrating dBRD9 degrader treatment. i µCT imaging and j quantification of LPS-induced local alveolar bone loss in BRD9 degrader treated (right) and control (left) lateral of the maxillary bone from 6-week-old mice. n = 6. k H&E staining, l TRAP staining, and m quantification of LPS-induced local alveolar bone in BRD9 degrader treated and control lateral of the maxillary bone from 6-week-old mice. n = 6. n RANKL, TNF-a, and iNOS immunofluorescence (green) in LPS-induced local alveolar bone in BRD9 degrader treated and control lateral of the maxillary bone from 6-week-old mice. o The mRNA expression of TNF-a, iNOS, Il6, Cd86, and Rankl of LPS-induced local alveolar bone in BRD9 degrader treated and control lateral of the maxillary bone from 6-week-old mice. n = 7. All data in this figure are represented as mean ± SD. Two-tailed Student’s t-test for (d), (f), (j), (m), and (o). M1, first molar. M2, second molar. Asterisk in b, c indicates less bone mass. Scale bar in b, i, 1 mm; c, e, k, l, 200 µm; g, n, 100 µm. All experiments were performed in triplicates unless otherwise stated. Source data are provided in the Source data file.

Fig 5: FOXP1 interacts with BRD9 to regulate Stat1 transcription during osteoclastogenesis.a Heatmap showing the average ATAC-Seq signal centered on the transcription start site (TSS) of the nearest genes in osteoclastic induced BMDMs after control and iBRD9 treatment. Color scale represents average signal intensity. b Venn diagram showing the overlapping and discrepant peaks between control and iBRD9 group. c Volcano plot depicting differentially accessible region (DAR) between control and iBRD9 group. d GO terms of genes with gain and loss DAR around TSS after BRD9 inhibition. e Dot bubble plot showing the loss DAR motif enrichment after BRD9 inhibition. f Motifs enriched in BRD9 binding while loss after BRD9 inhibition. g The FOXP1 motif logo. h FOXP1 immunofluorescence (red) and CTSK (green) in the mouse distal femur at 4 weeks. Scale bar, 100 µm. i FOXP1 immunofluorescence (green), TRAP staining (red) and actin (red) in BMDMs after 3 days of RANKL-induction. Scale bar, 100 µm. j Co-IP assay with FOXP1 antibody (or IgG) in BMDMs during osteoclastic induction, followed by immunoblotting of BRD9 and FOXP1. k ChIP assay with FOXP1 antibody (or IgG) in RAW264.7 cell during RANKL-induction treated with iBRD9 or vector. n = 3 biologically independent samples. l The mRNA of Foxp1, Stat1, Myc, Mmp9 and Acp5 in osteoclastic induced RAW264.7 cell treated with Foxp1 knockdown lentivirus (Foxp1-Lenti) or control lentivirus (Control-Lenti). n = 3 biologically independent samples. m The mRNA of Foxp1 in BMDMs from LysM-Cre;Brd9fl/fl (n = 3) and control mice (n = 6) at 4 weeks after 3 days of osteoclast differentiation. n The mRNA of Foxp1 in BMDMs treated with iBRD9 after 3 days of osteoclast differentiation. n = 3 biologically independent samples. o Schematic drawing shows chromatin remodeling mediated by BRD9. All data in this figure are represented as mean ± SD. Negative binomial distribution used for c. Hypergeometric distribution test for d. the one-tailed Fisher’s Exact test for (e) and (f). Two-tailed Student’s t-test for (m), (n), (k), and (l). MR, M-CSF + RANKL. All experiments were performed in triplicates unless otherwise stated. Source data are provided in the Source data file.